Redesigned Approach Holds Promise, but Census Bureau Needs to Annually Develop and Provide a Comprehensive Project Plan to Monitor Costs
GAO-06-1009T: Published: Jul 27, 2006. Publicly Released: Jul 27, 2006.
The U.S. Census Bureau (Bureau) estimates that the 2010 Census will cost over $11.3 billion, making it the most expensive in our history. The U.S. House of Representatives and Senate appropriation bills propose to reduce the Bureau's fiscal year 2007 budget request, raising questions about the Bureau's design of the 2010 Census and associated costs. Based on issued GAO work, this testimony addresses the extent to which the Bureau has (1) made progress redesigning its approach, including nonresponse follow-up, a key cost driver; and (2) developed a comprehensive project plan for the 2010 Census, as well as timely, detailed cost data for effective oversight and cost control.
Since 2000, the Bureau has made significant progress in redesigning the 2010 Census. Preparations for the 2010 Census appear to be further along than at a similar point of the 2000 Census; the Bureau plans to make the most extensive use of contractors in its history to implement such mission-critical tasks as data collection and processing, and updating addresses and maps; and it has developed new initiatives, such as changing to a short-form-only census and automating field operations to reduce nonresponse follow-up costs. Still, the Bureau will have to resolve challenges that could increase the costs of the census. For example, the Bureau will need to effectively monitor contracts, as $1.9 billion of the $11.3 billion life-cycle costs will be spent on seven major contracts. The Bureau has agreed to take steps to mitigate some of these challenges, such as enhancing the ability of key contract project offices to better manage contracts through such actions as developing action plans with milestones for key activities and regularly briefing senior managers. Also, the use of hand-held mobile computing devices (MCD) to help reduce nonresponse follow-up costs by automating operations and managing the agency's payroll is a key component of the redesigned census. However, the MCDs experienced reliability problems during testing. The Bureau maintains that those problems will be fixed by developing a new MCD through a contract awarded in March 2006; however, the new MCD will not be tested until the 2008 Dress Rehearsal, and little time will remain to develop, test, and incorporate refinements if the MCDs do not perform as expected. If after the Dress Rehearsal the MCD is found to be unreliable, the Bureau could be faced with the remote but daunting possibility of having to revert to the costly paper-based census used in 2000. The Bureau has not developed and provided a comprehensive, integrated project plan that details milestones, itemized costs, and measurable goals for completing key activities. Also, the Bureau's $11.3 billion life-cycle cost estimate lacks timely and complete supporting data, because it does not contain the most current information from testing and evaluation nor does it provide sufficient information on how changing assumptions could affect costs. For example, one key assumption that has not been updated pertains to the use of the MCDs. The Bureau anticipates that their use could reduce administrative and support costs in its local census offices, including 50 percent cost reductions for staff and office space. However, the 2004 Census Test showed that more help desk staff and more storage space would be needed to support the devices. The Bureau did not change the life-cycle cost estimate because, in the view of Bureau managers, field tests are for operational purposes, not to inform cost estimates. However, using test results to update cost assumptions could assist the Bureau and external policymakers to oversee costs and make necessary resource allocations. Furthermore, absent a comprehensive plan and updated cost information, the effect of proposed 2007 budget reductions on the overall design and life-cycle costs of the 2010 Census cannot be determined.