Additional Actions Would Help the Bureau Realize Potential Administrative Records Cost Savings
GAO-16-48: Published: Oct 20, 2015. Publicly Released: Oct 20, 2015.
- Highlights Page:
- Full Report:
- Related WatchBlog Post:
What GAO Found
Increased reliance on administrative records—information already provided to the government as it administers other programs—has been discussed since the 1970s as a possible way to improve the quality or reduce the cost of the decennial census, and it may finally play a significant role in the decennial census in 2020. The U.S. Census Bureau (Bureau) estimates that it can save $1.4 billion using administrative records, compared to relying solely on traditional methods.
The Bureau recently completed its 2015 Census Test in Maricopa County, Arizona—a major test involving administrative records. The Bureau used this census test to demonstrate the feasibility of using administrative records to reduce the cost of its largest decennial field operation, following up door to door to enumerate households that do not respond to the census. Yet turning this estimated savings—and the potential savings from other uses of the records, such as using administrative records to help validate and update the address list rather than having to send temporary workers to every housing unit in the country—into a real cost reduction for the taxpayer will require detailed planning that includes milestones for ensuring outstanding challenges are addressed. This would include preventing disclosure of records and addressing concerns the public may have over their use, and obtaining access to remaining records. The Bureau has not set deadlines for deciding which records it will use and for which purposes, but doing so will help the Bureau complete needed activities on time and prioritize which activities—or records—to abandon if time and resources become a constraint.
Bureau officials said they consider the test a large success because it demonstrated a variety of new methods and advanced technologies that are under consideration for the 2020 Census. The test also demonstrated the feasibility of a prototype system for managing the field operation, yet implementation issues with some of the prototype technology were not systematically reported or tracked, and may have affected the usefulness of test data. Systematic problems arising during test interviews can affect key test measures, such as the number of hours spent going door to door. Knowing which cases experienced such problems can help link cost estimates to specific design features and prioritize future research, development, and acquisition efforts.
Key assumptions the Bureau used in estimating potential cost savings from administrative records are logical, and the Bureau plans to provide additional support for them. For example, the Bureau's assumption that it could reduce its follow-up workload follows clearly from the Bureau's use of administrative records to remove vacant units from among those housing units needing follow-up because people did not respond to the census, reducing that workload by 11.6 percent. This assumption was also validated by the Bureau's experience in its recent test, and the Bureau plans further testing of this assumption during future tests in 2016 and beyond. The Bureau released an updated life cycle cost estimate in October 2015, and GAO anticipates reviewing its reliability after the Bureau makes available support for the estimate.
Why GAO Did This Study
The cost of the decennial census has steadily increased during the past 40 years, prompting the Bureau to reengineer key census-taking methods for the 2020 Census, including making greater use of information from administrative records. Given the potential cost savings associated with the use of administrative records, GAO reviewed (1) the Bureau's plans for using them and what opportunities and challenges the Bureau faces going forward; (2) the extent to which the Bureau's key 2015 test of them was implemented in accordance with objectives; and (3) the key assumptions supporting estimates of expected cost savings. To meet these objectives, GAO reviewed Bureau planning documents and test plans, interviewed Bureau officials, and observed implementation of the 2015 Census Test in Arizona. GAO also relied on its Schedule Assessment Guide.
What GAO Recommends
GAO recommends that the Census Director ensure that resources focus on activities with promise to reduce cost by documenting milestones related to deciding which records to use for which purposes and by systematically recording better information about implementation issues affecting specific cases in future tests. The Department of Commerce concurred with GAO's findings and recommendations, and provided minor technical comments, which were included in the final report.
For more information, contact Robert Goldenkoff at (202) 512-2757 or firstname.lastname@example.org.
Recommendations for Executive Action
Status: Closed - Implemented
Comments: In its December 2015 action plan, the Bureau provided the deadlines for making final decisions about which records to use for what purposes. All of these decision points and dates were included in the 2020 Census Operational Plan, which we released on October 6, 2015. Those decisions with deadlines included: For the Local Update of Census Addresses (LUCA) operation, to what extent can administrative records and third-party data be used to validate addresses submitted by LUCA participants? For the Non-ID Processing operation, how will administrative records and third-party data be used to improve matching in Non-ID Processing? For the Nonresponse Followup operation, what is the final set of administrative records and third-party data (including state-level data sources) that are necessary to support the 2020 Census NRFU operation? For the Nonresponse Followup operation, for each of the final administrative record and third-party datasets, what are the allowable use, required timing, and acquisition approach for the data? Additionally, Bureau officials stated that no later than September 2018, the Census Bureau will determine the final set of administrative records and third-party data sets used to support the 2020 Census NRFU. After September 2018, the Census Bureau will not pursue the acquisition of further sources.
Recommendation: To help ensure the Bureau focuses its resources on those activities that show promise for substantially reducing enumeration cost, the Secretary of Commerce should direct the Under Secretary of the Economics and Statistics Administration and the Director of the U.S. Census Bureau to establish clearly documented deadlines for making final decisions about which records to use for what purposes, particularly for purposes not yet demonstrated as feasible or involving records it does not already have access to, such as the National Directory of New Hires and KidLink.
Agency Affected: Department of Commerce
Comments: The Census Bureau informed us in December 2015 that no later than the end of December 2015, it would document how these matters have been addressed in the enumerator training (or in help screens on their mobile device) planned for the 2016 Census Test, and that it would use results and observations from that test to further refine such information for future tests and for the 2020 Census. We are carrying out an audit that examines the Bureau's implementation of its 2016 Census Test and we will look at these and related issues.
Recommendation: To help ensure the Bureau focuses its resources on those activities that show promise for substantially reducing enumeration cost, in advance of the 2016 Census Test and later tests, the Secretary of Commerce should direct the Under Secretary of the Economics and Statistics Administration and the Director of the U.S. Census Bureau to, ensure systematic capture of information about fieldwork cases that experience problems by including information in enumerator training about where to record the issues, who to contact, what details to include, and the importance of doing so.
Agency Affected: Department of Commerce