The day that President Donald Trump fired FBI Director Jim Comey, a second political drama unfolded. It received far less media attention, but in the long run might have even more severe implications for the health of the country’s democratic processes. That day, Census Bureau Director John Thompson, who had recently agreed to remain at his post for an additional year, suddenly announced his resignation, his departure slated to take effect in early June.
Although Thompson wasn’t formally a “political appointment,” he had been hired under President Barack Obama. What was not made clear in the announcement of his resignation was whether the director had resigned voluntarily or was pushed out by an administration so determined to sweep the federal system of all traces of Obama appointees that it had sought to replace everyone from U.S. district attorneys through to the White House head usher.
Either way, Thompson’s departure came at a perilous moment, leaving the Census Bureau leaderless just when it needed to fight tooth and nail for adequate funding to conduct the 2020 census.
As of the middle of July, the Trump administration has not nominated a successor to Thompson; nor has it yet reached out to any of the groups – demographers’ associations, business groups, and so on – usually consulted during the search for a new head of the Census Bureau, leaving the impression among observers that it is content to see the office rudderless.
Thompson himself was not available for comment for this article. Acting Director Steven Buckner also did not respond to requests to comment. And, despite repeated phone and email requests, sent over a period of several weeks, no one else from the Census Bureau responded either.
But a May 31 letter written to the Trump administration, and also sent to the U.S. Commerce Secretary and the leaders of Congress, by The Leadership Conference on Civil and Human Rights, and signed by 168 civil rights, academic and demographers’ organizations from around the country, shows the depth of concern about this situation.
Thompson’s resignation, the authors wrote, creates a “leadership vacuum less than three years before the start of the nation’s largest, most complex peacetime mobilization and less than one year before a critical ‘dress rehearsal’ of the full census process…. The strength of our civic and economic institutions depends on the success of these activities and other irreplaceable surveys and data programs.”
Few things in American political culture involve as much long-term planning as does the diennial census.
As soon as one census is complete, Congress has to create a 10-year funding plan for the next one, a plan that envisions a gradual acceleration in effort in the first few years of the 10-year cycle, followed by a massive acceleration toward lift-off in the last few as systems are tested, and surveyors hired and trained. Half of all census spending occurs in the final year of the 10-year process.
Demographers, over this decade-long cycle, have to start developing questions and survey methods likely to best capture key data about the roughly 330 million people who reside in this country.
And technologists have to work on the most effective ways to utilize rapidly evolving computer and communications systems to allow the hundreds of thousands of people hired to do the surveys effective means of inputting their data, in real time, into the Census Bureau databases.
Getting the census right is vital to the nation’s political processes: working out who lives where allows for a fair apportionment of political representation, both between and also within states – which is why the Founding Fathers, as they tried to craft a viable democracy that would survive the ages, embedded the requirement for the census into the new constitution.
Historically, census numbers have almost always been accepted by Congress as accurate when working out reapportionment – although when the 1920 one showed that, for the first time, America was now more urban than rural, the Congress, dominated by rural representatives, did refuse to use its numbers to reapportion power toward the cities and away from more conservative rural districts.
But the power of the census data goes far beyond the drawing of political districts: hundreds of billions of dollars of federal money are allotted each year based on complex population formulas.
If the census over-counts certain groups, then those people and the locations they live in disproportionately benefit from government expenditures; and, by contrast, if the census under-counts groups, those groups are systematically short-changed.
And, since census data are used to work out where people of varying ethnic and national backgrounds live, getting the census right is also vital for a fair enforcement of civil rights laws, ranging from protecting equal access to housing through to education investments and desegregation of public schools.
Beyond that, because census data are seen as the statistical gold standard, all other large-scale surveys – on housing, health, transport, and so on – are weighted to reflect census numbers.
If the Census Bureau undercounts a particular group, Ken Prewitt, professor of public affairs at Columbia University and director of the federal agency from 1998 through early 2001, explains, then all other surveys carried out by the government over the next 10 years “will reproduce that statistical error.”
In the 2000 to 2010 cycle, the Census Bureau spent roughly $12.5 billion on the 2010 count. In 2011, Congress, which had newly-elected Tea Party members, froze the 2020 Census budget at the same level, $12.5 billion, taking no account of inflation during the period.
The working assumption used to justify this was that technology could be used to substitute for man-hours, meaning that hundreds of thousands fewer people would have to be hired to generate comprehensive survey data. Current projections assume that up to 65 percent of census responses will be gathered via the Internet in 2020, and that many gaps, created by people’s failure to respond, will be filled in by extensive mining of school, public health, housing and other online records.
These were, at best, optimistic scenarios, since any new technology has to be comprehensively field-tested to make sure that, during the few months of the actual data-collection, during the spring and early summer of 2020, the systems work as designed.
But what makes it even more perilous is that the Congressional budget serves as a ceiling rather than a floor – in other words, Congress and the administration can take money away from this process in their annual appropriations bills. Which is precisely what is happening today.
Trump’s proposed budget allocates only $1.5 billion to the Census Bureau for 2018, a mere $27 million more than was spent in 2017, at a point in the cycle when, historically, the agency test-runs its methods in several locations nationally, and fine-tunes the list of questions that it sends to Congress for approval – all of which requires that spending increases by several hundred million dollars year-on-year.
The shortfall in 2018 is approximately $300 million. It will, if current budgeting trends continue, be an order of magnitude worse again in 2019 and 2020.
In fact, the startling scale of the shortfall in these final years leading up to the census is so extreme that the Bureau will not pursue “peak operations” tests, which are essentially dress rehearsals, in all three locations chosen for trial runs. In 2018, those areas selected are Bluefield-Beckley-Oak Hill, West Virginia (the state is home to one of the lowest income earning counties in the U.S.); urban Providence, Rhode Island; and Pierce County in Washington state. All told, 700,000 people would be surveyed.
This means that many new technologies will be used for the first time during the actual census itself. The Bureau has also had to largely put on hold its “advertising and partnership program,” which uses targeted media ads to convince traditionally hard-to-reach populations, such as immigrants, low-income communities of color, Native Americans and the rural poor, to participate in the census and to accurately answer the questions asked by surveyors.
The advertising agency Young and Rubican has a $400 million-plus contract with the Bureau to run this campaign; so far, Prewitt has heard, it has received less than $30 million of this, meaning it has not been able to design its ads, translate them into multiple languages, book slots in local media and so on.
So devastating could this be that the General Accounting Office has now labeled the 2020 Census a “high risk endeavor,” one of four or five major federal programs now listed as being particularly vulnerable to wholesale failure.
“We could have a historically disastrous Census,” says Phil Sparks co-director of The Census Project, a coalition of organizations from around the country that works to ensure the Census is conducted fairly.
“It’s not a robust test of new features,” Prewitt worries. “You can’t do that in the final year. The most important consequence of an underfunded Census is that it will likely be inaccurate,” says Prewitt. “If it’s poorly funded it cannot approximate accuracy.”
If Trump’s budget request isn’t challenged in Congress – and observers are pessimistic that a distracted, polarized Congress has the foresight to think three years out about the Census – then, Prewitt believes, “it will be the most seriously underfunded Census since the Second World War. We are in new territory.”
Statisticians who cross-check census returns, have time and again found that middle-class Whites are over-counted in the census by about 0.8 percent – partly because they are more likely to cooperate with surveyors; and partly because many middle-class White families own more than one home and are frequently double-counted if they answer the doors to Census Bureau takers at more than one of their residences.
By contrast, African-Americans and Latinos are, historically, significantly under-counted. In fact, African-Americans are twice as likely to be under-counted as Whites are to be over-counted; and Latinos are three times more likely to be under-counted.
Under-fund the census, and all of these trends get magnified. Statistically, America post-2020 might well look Whiter, more affluent and more suburban than it actually is.
“There’s a number of groups at disproportionate risk,” explains Terri Ann Lowenthal, a consultant to The Leadership Conference on Civil and Human Rights, and a former staff director for the U.S. House of Representatives Subcommittee on the Census and Population.
“There historically has been a persistent undercount of communities of color – especially of Blacks, Latinos, American Indians living on reservations and tribal lands and young children under the age of 5, particularly children of color.”
Lowenthal also worries that, with Trump as president and with an administration openly hostile to immigrants, many noncitizen immigrants will be less likely to cooperate with Census Bureau takers. “Immigrants will be fearful of how the data they provide will be used. There are many mixed-status households.”
In such a climate, defunding the advertising and partnership program can only serve to further dilute immigrant participation, and, by extension, ultimately limit the amount of federal dollars that flow to immigrant-heavy communities.
“With the underfunding of the census, what Congress and the new administration are doing is stripping the census down to the bare bones. And that will weaken operations designed to make the census better in undercounted communities. Targeted advertising, linguistic help, and so on,” Lowenthal concludes.
“You get what you pay for. It is hard to avoid the conclusion, given the budget request for ’18, that a fair and accurate census for vulnerable communities is not a priority for this administration.”
Sasha Abramsky is a freelance journalist and book author. His last Equal Voice News article was “Trump’s 100 Days: Workers Get Pummeled. People Fight Back.” His new book, “Jumping at Shadows: The Triumph of Fear and the End of the American Dream,” is being published by Nation Books in September. His 2013 book, “The American Way of Poverty: How the Other Half Still Lives,” was listed by The New York Times as one of the 100 books of the year. He is a senior fellow for democracy at the Demos think tank. All original and contracted Equal Voice News content – articles, photos and videos – can be reproduced for free, as long as proper credit and a link to our homepage are included. This story has been updated.
2017 © Equal Voice for America’s Families Newspaper
Published by Marguerite Casey Foundation