Summary

A reduction of this magnitude at a critical juncture in preparing for the 2010 Census will have a major impact on the core mission of the Census Bureau. It will make it impossible for the Census Bureau to implement substantial automation improvements of a reengineering effort that has been underway since 2001.

As a result, this $53.3 million reduction in FY2007 will trigger a 20-fold increase in costs for FY2008-2011 of at least $1 billion.

In addition to these outyear cost increases, dropping these components of the reengineering effort also means that hundreds of millions of dollars now have been wasted on five years of research, development, and testing of these new methods and technologies.

Further, reengineering efforts devoted to improving coverage (reducing undercounts, overcounts, and geographic misallocations) will also be compromised, as will efforts to reduce the risk of operational failures during the 2010 Census.

Thus, this cut will make it impossible to meet three of the four strategic goals that the Census Bureau and the Department of Commerce set for the 2010 Decennial Census Program (specifically, reducing coverage errors, reducing risk, and containing cost).

The fourth goal—improving the timeliness of data—also will be affected. Although the Census Bureau can continue its American Community Survey at this funding level, it cannot sustain the recently introduced addition of a monthly sample of people who live in group quarters (such as the elderly, college and university students, and prisoners). This means that the ACS will no longer be a full replacement for the decennial census long form data.

Details

This funding reduction means the Census Bureau will have to drop all planned efforts to develop and implement GPS equipped Hand Held Computers for conducting major field data collection activities for the 2010 Census. Reverting to paper-based operations will add at least $1 billion to the overall cost of the 2010 Census:

· We will have to print paper address listing books and at least 30 million paper maps in order to conduct the Address Canvassing operation (for approximately 10 million blocks). In addition, significantly more space and staff will be needed to digitize the map updates so they can be inserted into the TIGER database in order to appear on subsequent paper products, such as the final maps provided to the states for redistricting.

· We will have to print at least 40 million more paper questionnaires, and at least 30 million paper maps to use during Nonresponse Followup (NRFU) visits to households who do not return their completed form by mail.

· We will no longer be able to electronically remove late mail returns from the NRFU assignment lists after the operation is underway (as we could have done using Hand Held Computers). This means that an estimated 3 million households will be burdened with completing a second questionnaire, and just this additional workload will add an estimated $200 million to the cost of the operation.

· We will need more and larger temporary census offices because of the increased space and staff that will be required to store and manage the additional paper forms and maps.

· Without the GPS locater functionality on a Hand Held Computer, the Census Bureau will be unable to make its planned improvements in the accuracy of address geocoding. Geographic coding errors (e.g., addresses being tabulated in the wrong political jurisdiction) was the largest remaining source of coverage errors in the decennial census.

The Census Bureau will delay efforts to realign its TIGER database with GPS coordinates. The number of counties to be completed in FY2007 will be cut by at least half, and this work will no longer be completed for all counties by mid 2008. That target date was driven by the need to have the realignment completed before conducting the 2010 Census Address Canvassing operation. However, because GPS equipped Hand Held Computers now will not be used for that operation, there is less urgency to complete the realignment in April 2008.

The Census Bureau also will eliminate the entire Group Quarters operation for the American Community Survey (ACS). Without the GQ data collection, the ACS cannot fully represent the total population of the U.S. Further, the elimination of this program will result in incomplete information for critical population groups (i.e., the elderly, the population in prisons). Ultimately not including the GQ population in the ACS means the ACS cannot fully be the replacement for the long form in 2010.

We will cancel the Community Address Updating System component of the ACS. This will reduce the overall coverage completeness of the ACS until updates can be incorporated from the 2010 Census Address Canvassing operation (which now will be a paper based operation in FY2009).

All of these late, and major, changes to the overall design and methods for the 2010 Census will add significant risk of operational or methodological failures that could compromise the coverage accuracy of the results. This in turn means we must direct more resources to our coverage measurement program.

The following is the impact statement for reduction to GSS—It can’t be used here since it is in a different subactivity of the budget (Demographic Programs).

We also will need to make a reduction to our Geographic Support System activities which would halt state, local and county geographic liaison activities and would stop all field work on Master Address File Geocoding Office Resolution (MAFGOR) operations in the Regional Offices. This will eliminate the Census Bureau's ability to locate new structures obtained through the USPS Delivery Sequence File, which must be done in time for the 2010 Census. This reduction would negatively impact the effectiveness of all Census geographic operations and reduce the quality of the Master Address File (MAF) and housing unit coverage in the sampling frame for the American Community Survey. This will also impact the Boundary and Annexation Survey for American Indian Areas that is supported by Field staff.