A critical Census Bureau program is in trouble, and how it might be fixed

The Census Bureau might be best known for the every-ten-year count. But its American Community Survey is also a vital instrument for planners and researchers throughout the United States. The ACS is in trouble thanks in part to declining response rates and the pandemic. With why this matters and how the Commerce Department can shore up the ACS, the Federal Drive with Tom Temin turned to George Washington University research professor and co-author of a recent...

READ MORE

The Census Bureau might be best known for the every-ten-year count. But its American Community Survey is also a vital instrument for planners and researchers throughout the United States. The ACS is in trouble thanks in part to declining response rates and the pandemic. With why this matters and how the Commerce Department can shore up the ACS, the Federal Drive with Tom Temin turned to George Washington University research professor and co-author of a recent call to action, Dr. Andrew Reamer.

Interview transcript:

Tom Temin: Dr. Reamer, good to have you on.

Andrew Reamer: Nice to be with you.

Tom Temin: Just review for us briefly the ACS and what it is and why it’s such an important part of what the census puts out.

Andrew Reamer: Sure, the ACS is new, and it’s very old, the American Community Survey. The idea was planted by James Madison in 1790. It grows out of the decennial census, the innovation in the Constitution was that every 10 years you have a census, counting the population so that House of Representatives could be apportioned appropriately across the states. And Madison had this bright idea, look, if we’re going to count people, why don’t we ask them some questions about who they are and what they do so Congress can be more informed about our constituents and legislate better. So Congress agreed, and for 140 years, from 1790 to 1930, there were additional questions on the Census about race, about age, about gender, about occupation, and eventually about how much income you made and what language you spoke. And in the Great Depression, people figured out how to do surveys. And so in 1940, 1950, 1960, a lot of the questions moved to a sample. So every 10 years from 1940, to 2000, these additional questions got asked of a subset of the population. Very, very powerful data drove public policy, exactly what Madison wanted, drove the allocation of federal funding humongous amounts of funding. And the data are so useful that Census Bureau and Congress agreed in 2005, to not wait once a decade to ask these questions, but to ask them every month. And so that’s the American Community Survey is like the fourth iteration of Madison’s bright idea. But now the questions are asked of about two and a half percent of the population every year. And the data are incredibly important for both the public sector and the private sector. Your radio station in terms of understanding your penetration of listeners, and the profile is based in part on the ACS. With no ACS, you would be in the dark about what the Federal News Network is achieving.

Tom Temin: So what is the issue now with the ACS?

Andrew Reamer: The issue was funding and the ability to kind of produce a more current accurate picture of America, so businesses and the government can make good decisions. So just a little bit of background, 10 years ago, there was a big effort by the Republicans to either eliminate the American Community Survey or to make it voluntary. It is mandatory to fill the thing out. And that’s important so that there is a good response rate. The problem is that the funding is insufficient to have a sample that’s big enough to produce accurate small area statistics on a timely basis. So there’s a need to expand the sample. And that takes money. It’s an expensive program. It’s about, I think, over 200 million bucks a year to run, but it drives a $21 trillion economy. So you know, the return on investment for increasing funding a bit to expand the sample and improve the data products is worth it. And that’s what the report calls for. One can look at the report, you immediately jump into the weeds regarding the nature of the recommendations. It’s not easy soundbites for radio.

Tom Temin: Well, we’ll make a try. We’re speaking with Dr. Andrew Reamer. He’s a research professor at the George Washington University, and a co-author of a study that has been published by the Census Project, and we’ve had the group on quite a bit.

Basically, then you’re calling for more money so that they can get sufficient samples done each month. And surveying is expensive to get an accurate picture down to the granularity that people are used to with Census.

Andrew Reamer: Yes, the ACS, unlike any other survey the government does goes down to the block level, the neighborhood level. The ACS is used, among other things for drawing, the redistricting the congressional districts. So the data are very, very granular. And the sample is only good enough now to do five year averages. And a lot of changes in five years. If you think back to, you know, 2017 is like a different world. And so the expanding the sample can make the data more current, and also make it more, because this is a survey, it’s an estimate. And the estimate, what the Census Bureau will do is say, OK, you know, we think that 25% of the people in this neighborhood have a bachelor’s degree, but it’s an estimate. It’s the truth is somewhere between 20 and 30%. And you know, if it’s a wide range in there, then that is less useful to your radio station. So what a bigger sample does will tighten up the range. They say we think the estimate’s 25%, but it’s between 23% and 27%, not 20% to 30%. That’s what increased funding for a bigger sample would do. It would make the results more timely in that rather than five years, you could do three years, and they would tighten up the estimates. So you’re a little more confident that the estimate actually reflects reality.

Tom Temin: And recently, the Census Bureau was not able to actually publish full data, again, because of declining response rates. And that really set off alarm bells for people that care about this, like the Census Project.

Andrew Reamer: Yes. And that was in large part because of the pandemic. The pandemic just threw off a lot of things it threw off the decennial census, but the decennial census was able to recover by extending the collection period. But for the ACS, right, people were not responding to the request for information. And where the Census Bureau normally would send someone to a door and knocking knock on the door or call, people were not answering. And so the response rates really dramatically declined in 2020. And the result was that for the first time in the ACS’s history, the data that came out just a couple of weeks ago, are considered, quote unquote, experimental because they had to go to plan B to come up with the estimates for 2020, because the response rate was lower. And you want to have enough funding to be able to handle future unexpected interruptions and data collection like COVID for climate change.

Tom Temin: Are there new methodologies they could use to ensure a better response rate? I don’t know what their methodology is now. But more and more activity like this is going online, less dependent on mail, and filling out forms and so forth.

Andrew Reamer: So several things. One is a real innovation that is in place and really made a big difference was it sounds funny: internet collection. Up until, I don’t know, 10 years ago, it was all by mail. And even the decennial census in 2010, was by mail, there was no internet option. So the ACS internet option has really improved the quality of the data. And people’s willingness to respond.

Some new techniques that they’re considering are the use of administrative records to kind of fill in the blanks and/or remove questions from the survey, just try to reduce the respondent burden. So if there’s information in data that the government collects, for other purposes, the Census Bureau is in a much better position than it’s ever been to use these records from other federal programs, from state programs, even from commercial databases to either replace questions asked on the ACS and fill in the blanks. So there are experiments going on, but that takes money. So that is one angle, instead of making the ACS solely based on a survey, merge it with like survey and administrative records, and they get kind of a dual methodology.

Tom Temin: All right, so your recommendations that are aimed, I guess, both at the bureau and at Congress, primarily, correct?

Andrew Reamer: Yes. And everyone in between. Because the bureau, as you’re listeners well know, there is a ladder of you know, OK, that goes from the the ACS office to the Census Bureau to the Department of Commerce, and the budget office there to OMB. And everyone along the way has to say yes, fine. And then we saw the release of the president’s budget. OMB is approving the request for the ACS, which have asked for some more money, and then that goes to Congress.

Tom Temin: Got it. Well, maybe by the middle of 2023. They’ll have the money.

Dr. Andrew Reamer is a research professor at the George Washington University and co-author of a census study published by the Census Project. Thanks so much for joining me.

Andrew Reamer: My pleasure.

Related Stories