Download PDF Chairman Gowdy, Ranking Member Davis, and Members of the Subcommittee:
Thank you for the opportunity to testify with regard to the American Community Survey (ACS), and in particular the legal requirement that Americans participate in the ACS.
This issue involves important questions of both individual privacy and lawmakers’ need for accurate data upon which to make important policy decisions. In the United States, we have sought to achieve an appropriate balance between these two needs. It is my opinion that mandatory participation in the ACS, coupled with legal protections for privacy of ACS respondents, maintains that balance in a reasonable way.
The American Community Survey replaced the Census long form, which previously had gathered detailed information on a subset of the U.S. population. Roughly one-in-six Census respondents were required to fill out the long form in addition to the standard Census questionnaire.
Researchers have pointed out technical pros and cons of the ACS versus the Census long form. The annual sample size of the ACS is smaller than for the Census long form, but the ACS is produced every year whereas the long form was generated only every 10 years. For that reason, the ACS allows for better real-time analysis and better tracking of trends from year to year. These abilities clearly would be of interest to policymakers in Congress and the administration.
But the ACS and the Census long form are similar in that participation in both is mandated by law. Like for the long form, mandatory participation in the ACS is controversial and raises legitimate privacy concerns of which policymakers should remain cognizant.
However, for several reasons I believe that mandatory participation in the ACS remains a reasonable policy.
First, the greater detail of information captured by the ACS has allowed the standard Census questionnaire to become less detailed. Thus, for the typical American, the Census process may have become less intrusive over time.
"It is likely that with voluntary participation data will fall short most for the individuals and households on whom government policy is most focused, including the poor, the less educated, and those with poorer language skills." Second, the same law that mandates individual participation in the ACS also makes it illegal for the Census Bureau to release data in such a way that an individual's privacy might be violated. Any Census employee who violates the privacy of Census data faces significant jail time and large monetary fines. I am not aware of any instance in which ACS respondents – or, for that matter, respondents to any Census survey – have had their privacy violated.
Third, and most importantly, without good data policymakers are essentially flying blind, lacking solid knowledge of the Americans they are seeking to assist. We already suffer too much from what might be referred to as “policymaking by anecdote,” where lawmakers seek to pass legislation before sufficiently examining the severity – or sometimes even the existence – of a perceived problem. Reducing the quantity and quality of data available to policymakers, analysts and researchers threatens to exacerbate this problem.
Moreover, it is likely that with voluntary participation data will fall short most for the individuals and households on whom government policy is most focused, including the poor, the less educated, and those with poorer language skills.
In my own research, I have found that the ACS filled gaps in existing data sets and allowed for analysis that would have been difficult or impossible to conduct in its absence. For instance, I am currently using the ACS in ongoing research on public sector compensation, some of which has been presented in hearings before the full Oversight Committee. For much of that research, we used the Census Bureau's Current Population Survey. However, the ACS contains more detailed information that has allowed us to better control for the different skills of public and private sector employees. Setting public-sector compensation at appropriate levels impacts the quality of the government workforce at the federal, state and local levels and can have fiscal repercussions potentially worth hundreds of billions of dollars per year. Without good data, though, this kind of analysis is extremely difficult to undertake.
Those who wish to make participation in the ACS voluntary raise important points, and we should not allow our concern for individual privacy to fade even if we judge that mandatory participation is the best policy course. In the United States, the government exists to serve the people, not vice versa. Nevertheless, I believe that government can best serve the American people by continuing to gather high quality survey data.
Andrew G. Biggs is a resident scholar at AEI.