The decennial Census count has been a staple of American life since the early days of the republic, but at a time when public funding is being slashed and scientific data questioned, Census-watchers fear the 2020 count is heading toward a crisis.
In addition, recent rhetoric from the Trump administration has left some groups fearful of sharing personal information with the government, further threatening the success of the count.
If the Census Bureau does not receive significantly more than the White House has proposed, it could be “catastrophic” for the 2020 count, said Terri Ann Lowenthal, a former staff director of the House census oversight subcommittee.
“I think Congress is taking a grave risk right now,” she said, adding that the bureau is in danger of not being able to afford a comprehensive “dress rehearsal” slated for 2018. “What Congress does in the next few weeks will really determine how much confidence both the Census Bureau and the public can have in the bureau’s ability to take an inclusive Census in 2020.”
The Census, which every ten years counts every person in the United States, was mandated by the Founding Fathers and has been called a keystone of American democracy. The government uses it to allocate public resources; businesses use it to choose where to invest; and the count affects congressional redistricting.
In announcing its budget last month, the administration touted its proposed $1.5 billion for the Census Bureau as a $100 million “increase.” Critics called that smoke and mirrors, however, noting that the White House was comparing it to the 2016 budget of $1.37 billion. The proposed amount is about $140 millionless than the $1.64 billion, or 20 percent increase, requested for 2017.
In past decades, the ramp-up in the years leading up to the count has been precipitous, with budgets sometimes doubling between the year ending in 7 and the year ending in 8.
Congress did not pass a budget last year, opting instead for a continuing resolution that kept most federal agencies funded at 2016 levels. They are slated to pass the 2017 budget by April 28; debate on new funding for 2018 could continue into this fall.
Many agree that the bureau could stand to trim its budget. The 2010 count was the biggest non-wartime mobilization in American history and the most expensive Census ever. With over 500 field offices and 550,000 staff deployed around the country, it cost $13 billion over ten years.
To carry out a similar operation this time would cost $17.8 billion. But five years ago, Congress ordered the Census Bureau not to spend more in 2020 than it had in 2010. So the bureau announced changes designed to achieve “the most automated, modern, and dynamic decennial census in history.”
Instead of sending workers to physically walk 11 million blocks, it would compile new addresses using geographic information systems and aerial imagery. It would encourage respondents to fill out questionnaires online rather than by mail. It would use data it already had from the public and data available from commercial sources.
The new approaches would allow the bureau to halve the number of field offices and reduce the number of workers to 300,000 or less, keeping costs down at around $12.5 billion over ten years.
The problem, critics say, is that the new methods must be tested well before 2020. And that takes money. Instead, the bureau is so short on funds that it has cancelled tests that were planned for this year and suspended development of a communications campaign.
Failing to beef up funding now reflects a pound-foolish, penny-wise mentality, said Arturo Vargas, executive director of NALEO Educational Fund, a nonprofit that promotes Latino participation in the political process.
“The bureau needs the money now,” he said. “It can’t be doing 2020 blindly. It should be taking the time to do the testing now, to work out all the kinks (and) make sure the bureau gets it right, because there are no do-overs after 2020.”
Funding for the Census Bureau does not all go to the decennial count. Among other things, the bureau puts out the annual American Communities Survey, a more detailed questionnaire sent to a smaller sample of respondents, and the Economic Census, conducted every 5 years, including this year.
For the 2020 Census, the strain is already showing. In October, citing uncertainty about funding, the bureau said it was canceling 2017 testing in Puerto Rico, the Standing Rock Reservation in North and South Dakota, and the Colville Reservation and Off-Reservation Trust Land in Washington State. Earlier this year, the government accountability office placed the Census on its High Risk list.
Part of the problem, critics say, is that Congress is not good at longterm planning.
“The Census is not the sort of legislative advantage that Congressmembers can see on their radar screens,” Lowenthal said. “It’s not a new highway that warrants a press release, it’s not a health clinic for senior citizens and veterans in the home district. But it is the foundation of democracy…and it’s at the point where starving the Census Bureau of sufficient funding will have the certain consequence of a less accurate count.”
Either that or, lacking money to test the cost-cutting measures, the Census might have to fall back on old methods, necessitating requests for emergency funds and ultimately driving costs higher.
Wilbur Ross, the newly-appointed secretary of the Commerce Department, which oversees the Census, has positioned himself as an advocate for the count and he has vowed to push back against efforts to cut its funding (he also worked as a Census taker while in business school.) His office did not respond to a request for comment.
Census bureau officials did not respond to specific questions about funding concerns; a spokesman said in a statement that they were “confident in our ability to complete an accurate Census in a timely fashion.”
Underfunding is not the only Census-related concern in the Trump era. In January, a leaked draft executive order suggested adding a question about immigration status; such a move would likely discourage undocumented immigrants from filling it out. And last month, LGBTQ advocates protested after a list of planned subjects for the 2020 ACS was released that for the first time included sexual orientation and gender identity, then quickly re-released with that topic missing. The bureau said the first version was an error.
A proposal to add a dedicated "Middle Eastern or North African" option to the race/ethnicity terminology in 2020 has tested well, but since the 2016 presidential campaign some who might have checked this category say they worry about being targeted by the government.
For several years some Republican lawmakers have also sought to make the ACS voluntary, which critics say would decrease participation, reduce accuracy, and cost more because it would be necessary to send out many more forms to get a big enough sample.
But these concerns pale next to the prospect of insufficient funding. A less accurate Census could mean a greater undercount of traditionally undercounted groups such as poor people, young children, immigrants, and non-English speakers, reducing public services and congressional representation for them.
It could also greatly impact businesses. For example, a company deciding whether to build anything from a giant shopping mall to a small shop needs to know how many prospective customers and employees live in the vicinity, information they have typically gotten from Census data. A less accurate count would compromise this, especially in less populated areas, said Howard Fienberg, director of government affairs at Insights Association, which represents the marketing research and analytics industry.
“It’s very easy to open a new business in New York City, but putting it in some small town in West Virginia is much more difficult,” Fienberg said. “You have to have really rock solid data to be able to make the case,” he said. “When we have uncertainty, business goes nowhere.”
Congress has in the past turned down requests for funding from the Census Bureau, sometimes with dire consequences, said Andrew Reamer, a research professor at the George Washington Institute of Public Policy at The George Washington University.
“The consequence of not providing the funding was that GDP estimates before the Great Recession were far too high, leading policymakers to underrespond,” he said. “If we have an inaccurate Census, then businesses will make poorer decisions and there will be a cost.”