Is America facing the real risk of the first failed census?

Is America facing the real risk of the first failed census?
© Getty Images

Many have debated whether adding a citizenship question to the decennial census is good or bad policy. However, little attention has been paid to whether the 2020 census is capable of withstanding the additional uncertainty a change of this magnitude at this late stage in the process would bring. 

The constitutionally mandated count of every person living in the United States was already facing significant strain. Costs have continued to escalate. In the 1970’s it cost $16 to count each household. In the last census it cost $70 per household. 

This inflation was not the result of government ineptitude. As the country’s diversity increased, it became more difficult to count the population. People also became more aware of how their personal data could be used, and more reluctant to share any information, even when its confidentiality is guaranteed under federal law. This necessitated a larger, more diverse army of census takers going door-to-door, at considerable cost to the taxpayer. 


Congress had pushed to contain costs and not spend more than it did during the last round. The result has been the most radical change in the process of counting the country’s population in generations.

For example the program which creates the database of households to be counted has changed from 100 percent in-person data collection to 70 percent use of administrative data sources. The bureau has put its faith in electronic devices over pen-and-paper. They assume a majority of Americans will fill out their census forms online, an assumption that is one data-breach away from suffering serious adverse challenges.

The result was predictable. Costs still escalated and uncertainty of adequate funding resulted in implementation delays. Test programs were scaled back or eliminated entirely. The Government Accountability Office had already classified the 2020 Census as having a “high risk” of failure.

Members from both parties recognized this real risk when Congress recently increased funding for the Census to $2.8 billion for the current fiscal year, nearly double the originally requested amount. 

A few days later it was announced the census would ask every respondent their citizenship status for the first time since 1950. What is most notable is not whether the question may be asked, but rather, what it will do to the overall count.

Normally new census questions are approved three years before they are deployed. This allows them to be tested and refined so problems can be identified and resolved. It ensures concerns are not hypothetical but justified.

The citizenship question was not put forward three years ago: it was first proposed just over three months ago. While some argue that this question has been asked on other surveys, the methodology of asking some people some questions while relying on statistical estimates to fill in the gaps is not the same as asking every person the same question with the goal of a 100 percent response rate. 

Many of the changes for 2020 are untested outside of small "dress-rehearsals" and questions persist. Will governmental records be accurate? Will computers be reliable? Will respondents be comfortable submitting personal information about their families online? 

Many are worried whether the 40 million foreign born residents of the United States will be willing to be counted at all. But before one even approaches that question, there is a more fundamental one. Will this additional uncertainty be too much for an already stressed system? Is America facing the real risk of the first failed census? Will this be the proverbial straw that breaks the camel’s back?

The impacts would be real and devastating for everyone living in this country, irrespective of citizenship status. The census determines how more than $675 billion dollars in federal funding are distributed annually 435 Congressional seats are distributed based on its results; Tens of thousands of state and local districts are tied to its outcome. 

In an increasingly hyper-partisan and polarized country too many politicians seem uninterested in whether government actual works. The only question that matters to them is will my side get more or less. 

It does not need to be this way. I served as the chief demographer for California’s Citizens Redistricting Commission, established by voters who decided politicians shouldn’t choose whom they represent.

Borrowing a line from the West Wing, shouldn’t we be able to come together to say no partisan agenda is more important than being able to answer the simple question of this is who we are, these are our numbers? 

Karin Mac Donald is the director of the statewide database, the redistricting database for the State of California, at the University of California, Berkeley, Law school. She serves as the state's liaison to the Census bureau's Redistricting and Voting Rights Data program. With her consulting firm, Q2 Data & Research, LLC, she served as the principal redistricting consultant to the state's Citizens Redistricting Commission which constructed California's congressional and legislative districts in 2011.