Politicizing Census puts crucial data at risk

One of the U.S. government’s most important tasks is conducting the decennial Census in which it attempts to count every person in the United States.

This is a monumental undertaking: In 2010, U.S. Census Bureau workers sent out questionnaires to 120 million households by mail, hand-delivered 12 million questionnaires in rural locations or those affected by natural disasters, and went door-to-door to nearly 47 million households that did not send back their responses.

All told, the last Census was the largest mobilization of a civilian workforce by the federal government in history and cost approximately $13 billion. With so much at stake, it is important to get this right.

Unfortunately, a draft executive order, if enacted, would politicize the 2020 census, thereby jeopardizing the integrity of its results and driving up costs for taxpayers.

The draft executive order would direct the Census Bureau to include questions about immigration status in the decennial census, an idea that Republican members of Congress previously proposed before the 2010 Census.

{mosads}While the Census Bureau does ask questions in the American Community Survey about place of birth, citizenship and year of entry, as the Census Bureau plainly states, “The Census Bureau does not collect data on the legal status of the foreign born.”

The reason it does not ask this question is that immigrants who are in the country without authorization are unlikely to return census forms or answer questions in person to census takers if they believe their answers may be used against them.

Such a requirement would thus significantly reduce response rates and create at least two problems.

First, it will reduce the quality of Census data, such as information about the 4.5 million American children living with at least one unauthorized parent, and thus harm many programs that depend on this information.

Census data is used to apportion congressional seats, redraw lines for legislative districts and allocate billions of dollars of federal assistance. In addition, both the public and private sector use social, economic and demographic data from the Census to make countless decisions affecting nearly every aspect of society, such as forecasting future housing needs, planning the location of hospitals and medical clinics, finding locations for factories and distribution centers, and enforcing fair lending practices.

Undercounting certain populations can have a significant impact on their livelihood.

For example, in the 2010 Census, the undercount rate for Latino children was 7.1 percent, compared to 4.3 percent for non-Latinos. This means that these communities had less access to federal funds for programs on maternal and child health services, early childhood education, and nutrition.

Moreover, undercounting undermines the value of government data as an authoritative source of information for rooting out biases elsewhere.

For example, a growing number of companies use data in innovative ways to deliver better services to individuals, such as to expand credit opportunities and housing. They can ensure these practices are fair, but doing so requires conducting a disparate impact analysis, a method that is only useful if it is based on accurate data.

Second, lower response rates will increase the cost of data collection. The most significant cost in conducting the Census is the non-response follow-up where Census takers go door-to-door to get information from those who do not respond to initial inquiries sent by mail.

While the Census Bureau is implementing a number of new initiatives to modernize its data collection process during the next Census, such as allowing responses by Internet, supplementing its data collection by using administrative records from other government agencies and third-party sources, and using imagery and mapping software to update its master address list rather than sending out staff to walk every block, a significant increase in non-response would cut into these savings.

The success of the decennial Census depends on public cooperation, and the Census Bureau has strived to maintain public trust so that individuals will share accurate information. Politically charged questions like this threaten to undermine the goodwill the Census Bureau has earned from past outreach.

The Census Bureau has worked hard to rehabilitate its reputation that previously suffered because of questionable activities in its past, such as divulging information about Japanese-Americans during World War II so that they could be imprisoned in internment camps, or more recently, providing statistics to the Department of Homeland Security about where Arab-Americans were living.

A public misstep of this magnitude could not only disrupt the 2020 Census, but have longer-term negative consequences on public trust in the Census Bureau.

In an era of “fake news” and “alternative facts,” the Census Bureau plays a valuable role as a trusted provider of accurate and reliable information. Policymakers would be wise to shield the Census Bureau from any measures that would undercut its credibility.

Daniel Castro is director of the Center for Data Innovation, a think tank studying the intersection of data, technology and public policy. Follow him on Twitter @CastroTech.

The views of contributors are their own and not the views of The Hill.