2000 Census
Lessons Learned for Planning a More Cost-Effective 2010 Census Gao ID: GAO-03-40 October 31, 2002GAO reviewed the funding of 2000 Census planning and development efforts and the impact it had on census operations. Total funding for the 2000 Census, referred to as the life cycle cost, covers a 13-year period from fiscal year 1991 through fiscal year 2003 and is expected to total $6.5 billion adjusted to 2000 year dollars. This amount was almost double the reported life cycle cost of the 1990 Census of $3.3 billion adjusted to 2000 year dollars. Considering these escalating costs, the experience of the U.S. Census Bureau in preparing for the 2000 Census offers valuable insights for the planning and development efforts now occurring for the 2010 Census. Thorough and comprehensive planning and development efforts are crucial to the ultimate efficiency and success of any large, long-term project, particularly one with the scope, magnitude, and the deadlines of the U.S. decennial census. For fiscal years 1991 through 1997, $269 million was requested in the President's Budgets for 2000 Census planning and development and the program received funding of $224 million by Congress, or 83 percent of the amount requested. According to U.S. Census Bureau records, the bulk of the $86 million in funding received through the end of fiscal year 1995 was obligated for program development and evaluation methodologies, testing and dress rehearsals, and planning for the acquisition of automated data processing and telecommunications support. The U.S. Census Bureau was responsible for carrying out its mission within the budget provided and bureau management determined the specific areas in which available resources were invested. GAO could not determine what effect, if any, that higher funding levels might have had on bureau operations as this is dependent upon actual implementation and the results of management decisions that may or may not have occurred. According to bureau officials, early planning and development efforts for the 2000 Census were adversely affected by lower funding than requested for fiscal years 1993 through 1997. They identified 10 areas where additional funding could have been beneficial. These included difficulties in retaining knowledgeable staff, scaled back plans for testing and evaluating 1990 Census data, delays in implementing a planning database, and limited resources to update address databases. The bureau's experience in preparing for the 2000 Census underscores the importance of solid, upfront planning and adequate funding levels to carry out those plans.
RecommendationsOur recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.
Director: Team: Phone:GAO-03-40, 2000 Census: Lessons Learned for Planning a More Cost-Effective 2010 Census
This is the accessible text file for GAO report number GAO-03-40
entitled '2000 Census: Lessons Learned for Planning a More Cost-
Effective 2010 Census' which was released on November 07, 2002.
This text file was formatted by the U.S. General Accounting Office
(GAO) to be accessible to users with visual impairments, as part of a
longer term project to improve GAO products‘ accessibility. Every
attempt has been made to maintain the structural and data integrity of
the original printed product. Accessibility features, such as text
descriptions of tables, consecutively numbered footnotes placed at the
end of the file, and the text of agency comment letters, are provided
but may not exactly duplicate the presentation or format of the printed
version. The portable document format (PDF) file is an exact electronic
replica of the printed version. We welcome your feedback. Please E-mail
your comments regarding the contents or accessibility features of this
document to Webmaster@gao.gov.
Report to Congressional Requesters:
October 2002:
2000 CENSUS:
Lessons Learned for Planning a More Cost-Effective 2010 Census:
GAO-03-40:
Letter:
Results in Brief:
Background:
Total Funding for Bureau Planning Was Lower Than Requested:
Funding and Other Factors Affected Planning Efforts:
2000 Census Planning Provides Lessons Learned for The 2010 Census:
Conclusions:
Recommendation for Executive Action:
Agency Comments and Our Evaluation:
Appendixes:
Appendix I: Objectives, Scope, and Methodology:
Appendix II: Analysis of Funding by Fiscal Year for Planning and
Developments of the 2000 Census:
Appendix III: Comments from the Department of Commerce:
GAO Comments:
Tables:
Table 1: 2000 Census Planning and Development Funding Requested and
Received for Fiscal Years 1991 Through 1997 (dollars in millions):
Table 2: 2000 Census Planning and Development Funding for Fiscal Year
1992 (dollars in millions):
Table 3: 2000 Census Planning and Development Funding for Fiscal Year
1993 (dollars in millions):
Table 4: 2000 Census Planning and Development Funding for Fiscal Year
1994 (dollars in millions):
Table 5: 2000 Census Planning and Development Funding for Fiscal Year
1995 (dollars in millions):
Table 6: 2000 Census Planning and Development Funding for Fiscal Year
1996 (dollars in millions):
Table 7: 2000 Census Planning and Development Funding for Fiscal Year
1997 (dollars in millions:
AIR: American Indian Reservation:
DADS: Data Access and Dissemination System:
DSF: Delivery Sequence File:
FTE: full-time equivalent:
ICM: Integrated Coverage Measurement:
MAF: Master Address File:
OMB: Office of Management and Budget:
TIGER: Topologically Integrated Geographic Encoding and Referencing
system:
Letter October 31, 2002:
The Honorable William Lacy Clay, Jr.
House of Representatives:
The Honorable Carolyn B. Maloney
House of Representatives:
This report responds to your request to review the funding of 2000
Census planning and development efforts and the impact it had on census
operations. Total funding for the 2000 Census, referred to as the life
cycle cost, covers a 13-year period from fiscal year 1991 through
fiscal year 2003 and is expected to total about $6.5 billion adjusted
to 2000 year dollars. This amount was almost double the reported life
cycle cost of the 1990 Census of $3.3 billion adjusted to 2000 year
dollars. Current life cycle unadjusted cost estimates of the 2010
Census range from $10 billion to $12 billion. Considering these
escalating costs, the experience of the U.S. Census Bureau in preparing
for the 2000 Census offers valuable insights for the planning and
development effort now occurring for the 2010 Census. This report is
part of a series of GAO studies on lessons to be learned from the 2000
Census that can help bureau efforts to conduct an accurate and cost-
effective census in 2010.
We classified 2000 Census efforts into the following three phases.
* The planning and development phase consisted of a variety of
activities to prepare for the actual decennial census count on April 1,
2000. Although planning and development efforts continued into
subsequent phases of the census, this phase involved 7 fiscal years
from 1991 through 1997 with actual funding of about $238 million
adjusted to 2000 year dollars, or about 4 percent of the total
decennial life cycle cost.
* The implementation phase included conducting the 1998 dress
rehearsal, establishing 511 temporary local census offices in the 50
states, hiring and training over 500,000 temporary personnel, and then
conducting the actual census count. This phase involved 3 fiscal years
from 1998 through 2000 with actual funding of about $5.6 billion
adjusted to 2000 year dollars, or about 86 percent of the total
decennial life cycle cost.
* The postenumeration phase involves compiling and checking the census
counts for public release by April 1, 2001, for subsequent use in
congressional redistricting, and for other postcensus studies. This
phase covers 3 fiscal years from 2001 through 2003 with estimated
funding of about $686 million adjusted to 2000 year dollars, or about
10 percent of the total decennial life cycle cost.
As agreed with your offices, our review focused on the planning and
development phase of the 2000 Census and (1) the funding requested,
received, and obligated, with funding received and obligated by major
planning category,[Footnote 1] (2) funding and other factors that
affected planning efforts, and (3) lessons learned for the 2010 Census.
Our analysis was based on unaudited budget and financial data provided
by the U.S. Census Bureau. We performed our work between January 2001
and July 2001 at which time our review was suspended due to an
inability to obtain access to certain budget records. After lengthy
discussions with senior officials of the bureau, the Department of
Commerce, and the Office of Management and Budget (OMB), and
consultation with your staffs, this access issue was resolved in May
2002 and we completed our analysis in June 2002. Our work was done in
accordance with U.S. generally accepted government auditing standards.
Further details on our scope and methodology are presented in appendix
I.
The Department of Commerce provided written comments on a draft of this
report, including two attachments. We reprinted the comments in
appendix III, except for the second attachment, Potential Life-Cycle
Savings for the 2010 Census, which the bureau stated is currently under
revision and is outside the scope of our review.
Results in Brief:
Thorough and comprehensive planning and development efforts are crucial
to the ultimate efficiency and success of any large, long-term project,
particularly one with the scope, magnitude, and deadlines of the U.S.
decennial census. For fiscal years 1991 through 1997, about $269
million was requested in the President‘s Budgets for 2000 Census
planning and development and the program received funding of about $224
million by the Congress, or about 83 percent of the amount requested.
For fiscal years 1991 and 1992, OMB deemed the Department of Commerce
requests to fund early census reform as insufficient and doubled the
amounts to $1.5 million and $10.1 million, respectively. These amounts
were included in the President‘s Budgets and the Congress concurred by
authorizing the full amount requested. However, funding reductions by
the Congress occurred every subsequent fiscal year from 1993 through
1997. According to the bureau, these reductions resulted in the
elimination, deferral, or scaling back of certain projects in planning
for the 2000 Census. The bureau obligated 99 percent of its
appropriated 2000 Census funding through fiscal year 1997.
According to U.S. Census Bureau records, the bulk of the $86 million in
funding received through the end of fiscal year 1995 was obligated for
program development and evaluation methodologies, testing and dress
rehearsals, and planning for the acquisition of automated data
processing and telecommunications support. For fiscal years 1996 and
1997, bureau records indicated the bulk of $138 million of decennial
funding received was obligated for planning for the establishment of
field data collection and support systems, refining data content and
products, evaluating test results, and procuring automated data
processing and telecommunications support. For the planning and
development phase, personnel costs consumed about 53 percent of total
costs; contractual services consumed 16 percent; and space, supplies,
travel, and other expenses consumed the remaining 31 percent.
The U.S. Census Bureau was responsible for carrying out its mission
within the budget provided and bureau management determined the
specific areas in which available resources were invested. We could not
determine what effect, if any, that higher funding levels might have
had on bureau operations as this is dependent upon actual
implementation and the results of management decisions that may or may
not have occurred. According to bureau officials, early planning and
development efforts for the 2000 Census were adversely affected by
lower funding than requested for fiscal years 1993 through 1997. They
identified 10 areas where additional funding could have been
beneficial. These included difficulties in retaining knowledgeable
staff, scaled back plans for testing and evaluating 1990 Census data,
delays in implementing a planning database, and limited resources to
update address databases. While lower funding may have affected these
areas, information from various bureau and GAO reports and testimony,
as well as our current review, showed that operational, methodological,
and other factors also contributed to weaknesses in the bureau‘s
planning efforts.
The bureau‘s experience in preparing for the 2000 Census underscores
the importance of solid, upfront planning and adequate funding levels
to carry out those plans. As we have reported in the past, planning a
decennial census that is acceptable to stakeholders[Footnote 2]
includes analyzing the lessons learned from past practices, identifying
initiatives that show promise for producing a better census while
controlling costs, testing these initiatives to ensure their
feasibility, and convincing stakeholders of the value of proposed
plans. A major contributing factor to the funding reductions was the
bureau‘s persistent lack of comprehensive planning and priority
setting, coupled with minimal research, testing, and evaluation
documentation to promote informed and timely decision making. Over the
course of the 1990s, the Congress, GAO, and others criticized the
bureau for not fully addressing such areas as (1) capitalizing on its
experiences from past decennial censuses to serve as lessons learned in
future planning, (2) documenting its planning efforts, particularly
early in the process, (3) concentrating its efforts on the few critical
projects that significantly affected the census count, such as
obtaining a complete and accurate address list, (4) presenting key
implementation issues with decision milestones, and (5) identifying key
performance measures for success.
In light of the challenges facing the bureau as it prepares for the
next decennial census in 2010, we are recommending to the Secretary of
Commerce that the bureau‘s requests for funding of planning and
development activity provide comprehensive information supported by
sound data. This information would include, but is not limited to, such
information as:
* specific performance goals for the 2010 Census and how bureau
efforts, procedures, and projects would contribute to those goals;
* detailed information on project feasibility, priorities, and
potential risks;
* key implementation issues and decision milestones; and,
* performance measures.
The department agreed with our recommendation and stated that the
bureau is expanding documents justifying its budgetary requests.
Background:
As a result of controversy and litigation surrounding the 1990
Decennial Census, the U.S. Census Bureau recognized the need for a
full-scale review of its decennial census program. The Congress, OMB,
and GAO also agreed that this review was needed and that it must occur
early in the decade to implement viable actions for the 2000 Census and
to prepare for the 2010 Census.[Footnote 3] Early in the 1990s, in
reports and testimonies, we stressed the importance of strong planning
and the need for fundamental reform to avoid the risk of a very
expensive and seriously flawed census in 2000.[Footnote 4]
To address a redesign effort, in November 1990 the bureau formed the
Task Force for Planning the Year 2000 Census and Census-Related
Activities for 2000-2009. The task force was to consider lessons
learned from the 1990 Census, technical and policy issues,
constitutional and statutory mandates, changes in U.S. society since
earlier decennial censuses, and the most current knowledge of
statistical and social measurement. The bureau also established a Year
2000 Research and Development Staff to assist the task force and
conduct numerous research projects designed to develop new approaches
and techniques for possible implementation in the 2000 Census. In June
1995, the task force issued its report, Reinventing the Decennial
Census.
Concerns about the 1990 Census also led the Congress to pass the
Decennial Census Improvement Act of 1991 (Public Law 102-135) requiring
the National Academy of Sciences to study the means by which the
government could achieve the most accurate population count possible
and collect other demographic and housing data. The academy established
a panel on methods to provide an independent review of the technical
and operational feasibility of design alternatives and tests conducted
by the U.S. Census Bureau. The panel issued its final report in
September 1994. A second academy panel on requirements examined the
role of the decennial census within the federal statistical system and
issued its final report in November 1994.
In March 1995, the bureau conducted the 1995 Census Test which provided
a critical source of information to decide by December 1995 the final
design of the 2000 Census. These efforts resulted in a planned approach
for reengineering the 2000 Census which was presented in a May 19,
1995, U.S. Census Bureau report, The Reengineered 2000 Census.
In October 1995, we testified on the bureau‘s plans for the 2000
Census.[Footnote 5] In that testimony, we concluded that the
established approach used to conduct the 1990 Census had exhausted its
potential for counting the population cost-effectively and that
fundamental design changes were needed to reduce census costs and to
improve the quality of data collected. We also raised concerns about
the bureau proceeding with design plans for the 2000 Census without
input from the Congress. In the intervening months, the bureau was
unable to come to agreement with the Congress on critical design and
funding decisions. In February 1997, we designated the 2000 Decennial
Census a new high-risk area because of the possibility that further
delays could jeopardize an effective census and increase the likelihood
that billions of dollars could be spent and the nation be left with
demonstrably inaccurate census results.[Footnote 6] In July 1997, we
updated our 1995 testimony on bureau design and planning initiatives
for the 2000 Census and assessed the feasibility of bureau plans for
carrying out the 2000 Census.[Footnote 7]
To respond to Title VIII of Public Law 105-18, which required the
Department of Commerce to provide detailed data about the bureau‘s
plans by July 12, 1997, the bureau issued its Report to Congress, The
Plan For Census 2000. This plan also incorporated the bureau‘s Census
2000 Operational Plan that was updated annually. In November 1997,
Public Law 105-119 established the Census Monitoring Board to observe
and monitor all aspects of the bureau‘s preparation and implementation
of the 2000 Census.[Footnote 8] Section 209 (j) of this legislation
also required the bureau to plan for dual tracks of the traditional
count methodology and the use of statistical sampling to identify
historically undercounted populations of children and minorities.
Census Appropriations:
As 1 of 13 bureaus within the Department of Commerce, the U.S. Census
Bureau must submit its annual budget for review and inclusion in the
department‘s budget. The department must then make choices in
consideration of its overall budget to OMB and will therefore make
adjustments to bureau-requested budgets as deemed necessary. OMB will
review and further adjust department and bureau budgets to consider the
programs and priorities of the entire federal government that become
the President‘s Budget. The Congress may then adjust the President‘s
Budget through the appropriation process that becomes the budget of the
departments and the bureaus after signature by the President. The
appropriations for decennial census are no-year funds that are
available until expended, rescinded, transferred, or until the account
is closed.
Total Funding for Bureau Planning Was Lower Than Requested:
As shown in table 1, the Department of Commerce requested a total of
$268.7 million for 2000 Census planning and development in the
President‘s Budgets for fiscal years 1991 through 1997. The program
received total funding of $223.7 million from the Congress, or about 83
percent of the amount requested. Although the 2000 Census received all
of the funding requested in the President‘s Budgets for fiscal years
1991 and 1992, it received reduced funding for each fiscal year from
1993 through 1997. According to the bureau, these reductions resulted
in the elimination, deferral, or scaling back of certain projects in
planning for the 2000 Census. The bureau subsequently obligated 99
percent of its appropriated 2000 Census funding through fiscal year
1997.
Table 1: 2000 Census Planning and Development Funding Requested and
Received for Fiscal Years 1991 Through 1997 (dollars in millions):
Fiscal year: 1991; 2000 Census requested funding (President‘s
Budget)[A]: $1.5; 2000 Census appropriated funding by Congress[B]:
$1.5; 2000 Census requested funding not appropriated: $0; Percent of
requested annual funding received: 100.
Fiscal year: 1992; 2000 Census requested funding (President‘s
Budget)[A]: 10.1; 2000 Census appropriated funding by Congress[B]:
10.1; 2000 Census requested funding not appropriated: 0; Percent of
requested annual funding received: 100.
Fiscal year: 1993; 2000 Census requested funding (President‘s
Budget)[A]: 19.4; 2000 Census appropriated funding by Congress[B]:
13.7; 2000 Census requested funding not appropriated: 5.7; Percent of
requested annual funding received: 71.
Fiscal year: 1994; 2000 Census requested funding (President‘s
Budget)[A]: 23.1; 2000 Census appropriated funding by Congress[B]:
18.7; 2000 Census requested funding not appropriated: 4.4; Percent of
requested annual funding received: 81.
Fiscal year: 1995; 2000 Census requested funding (President‘s
Budget)[A]: 48.6; 2000 Census appropriated funding by Congress[B]:
42.0; 2000 Census requested funding not appropriated: 6.6; Percent of
requested annual funding received: 86.
Fiscal year: 1996; 2000 Census requested funding (President‘s
Budget)[A]: 60.1; 2000 Census appropriated funding by Congress[B]:
51.3; 2000 Census requested funding not appropriated: 8.8; Percent of
requested annual funding received: 85.
Fiscal year: 1997; 2000 Census requested funding (President‘s
Budget)[A]: 105.9; 2000 Census appropriated funding by Congress[B]:
86.4; 2000 Census requested funding not appropriated: 19.5; Percent of
requested annual funding received: 82.
Fiscal year: Total; 2000 Census requested funding (President‘s
Budget)[A]: $268.7; 2000 Census appropriated funding by Congress[B]:
$223.7; 2000 Census requested funding not appropriated: $45.0; Percent
of requested annual funding received: 83.
[A] Amounts include prior year recoveries of $1.6 million.
[B] Amounts include prior year recoveries and carry-in of $16.1
million.
Source: U.S. Census Bureau.
[End of table]
Bureau records indicated that the bulk of $86 million of decennial
funding received through the end of fiscal year 1995 was obligated for
program development and evaluation methodologies, testing and dress
rehearsals, and planning for the acquisition of automated data
processing and telecommunications support. For fiscal years 1996 and
1997, bureau records indicated that the bulk of $138 million of
decennial funding received was obligated for planning the establishment
of field data collection and support systems, refining data content and
products, evaluating test results, and procuring automated data
processing and telecommunications support. For the planning and
development phase, personnel costs consumed about 53 percent of
planning and development funds; contractual services consumed 16
percent; and space, supplies, travel, and other expenses consumed the
remaining 31 percent.
Because of different major program categories used by the bureau from
fiscal years 1991 through 1997, we could not present a comprehensive
table of funding for the period. However, we were able to analyze the
funding by fiscal year and a detailed analysis of funding requested,
received, and obligated, and funds budgeted by major program category
for fiscal years 1991 through 1997 are presented in appendix II.
Funding and Other Factors Affected Planning Efforts:
The U.S. Census Bureau was responsible for carrying out its mission
within the budget provided and bureau management determined the
specific areas in which available resources were invested. We could not
determine what effect, if any, that higher funding levels might have
had on census operations as this is dependent upon actual
implementation and the results of management decisions that may or may
not have occurred. However, according to bureau officials, lower than
requested funding levels for fiscal years 1993 through 1997 adversely
affected the bureau‘s planning and development efforts for the 2000
Census. As examples, they cited the following 10 areas where reduced
funding levels caused the bureau to curtail planning initiatives.
Although lower funding levels may have affected these areas,
information from previous bureau and GAO reports and testimony
indicated that operational, methodological, and other factors also
contributed to weaknesses in the bureau‘s planning efforts.
1. Difficulties in retaining knowledgeable staff. Although many key
bureau personnel and project managers involved with the 2000 Census had
also worked on the 1990 and earlier decennial censuses, bureau
officials stated that many experienced people retired or left the
bureau after the 1990 Census. According to the bureau, a contributing
factor was lower funding levels to pay personnel compensation and
benefits, which in turn affected the number of personnel with
institutional knowledge of the decennial census to lend support to the
2000 Census planning and development effort.
We noted that soon after a major event such as the decennial census
count, it is not unusual for personnel to leave the bureau, as did
three senior executives after the 2000 Census. In addition, Office of
Personnel Management data indicated that over half of the bureau‘s
full-time, nonseasonal work force of 5,345 employees as of March 2002
is eligible for retirement by 2010. Thus, the human capital issue will
remain a key planning area to ensure that the bureau has the skill mix
necessary to meet its future requirements.
2. Scaled-back plans for testing and evaluating 1990 Census data. A
bureau official stated that the amount of qualitative and quantitative
data from the 1990 Census was limited and hampered the quality and
results of planning and development efforts for the 2000 Census.
Additionally, many opportunities were lost in capitalizing on the 1990
Census data that did exist and more funding to evaluate this data could
have facilitated 2000 Census research and planning efforts. Bureau
officials stated that as they moved forward with planning for the 2000
Census, they had to scale back plans for testing and evaluating 1990
Census data because of a lack of funding. For example, they cited the
inability to update a 1990 Census study of enumerator supervisor
ratios.
3. Delays in implementing a planning database. Bureau officials stated
that they were unable to implement an effective planning database in
the early years of the 2000 Census. In one of its first plans, the
bureau conceived of a planning database that would capture data down to
very small geographic levels and would be continuously updated over the
decade for a number of census purposes. This database would have
enabled the bureau to target areas where language resources were
needed, identify areas where enumeration and recruiting could be
difficult, and position data capture centers to support the most cost-
efficient and effective infrastructure. However, according to bureau
officials, with lower funding through fiscal year 1995, the planning
database was put on hold. Later in the decade, the bureau resurrected
the planning database but did not develop and use it fully.
4. Limited resources to update address databases. According to bureau
officials, sufficient resources to update and coordinate large
databases of addresses and physical locations provided a continuous
challenge to the bureau. At the end of the 1990 Census, the bureau‘s
database contained 102 million addresses, each assigned to the census
block area in which it was located. At that point, the U.S. Census
Bureau‘s Geography Division initiated discussions with the U.S. Postal
Service to utilize its Delivery Sequence File (DSF) that contained
millions of addresses used to deliver the U.S. mail.[Footnote 9] The
bureau planned to use the DSF in updating its address database which
became the Master Address File (MAF).[Footnote 10] With lower funding
through 1995, bureau officials cited limited resources to update the
MAF database and to assess the quality of entered information.
5. Program to identify duplicate responses was not fully developed.
Bureau officials stated the program to identify duplicate responses was
not fully developed for the 2000 Census and more emphasis and funding
were needed to develop appropriate software and procedures. It is
important to be able to identify duplications in the MAF and multiple
responses from a person or household that contribute to a population
overcount. This includes operations to identify multiple responses for
the same address and computer matching of census responses received
against all other people enumerated in the block. Duplications also
occurred due to college students counted both at school and at home,
people with multiple residences, and military personnel residing
outside their home state.
6. Abandoned plans to use administrative records. In early planning for
the 2000 Census, the bureau funded efforts to use records from
nonbureau sources of information (such as driver licenses, voter
registrations, and other government programs) to supplement the census
count. This administrative records project was the result of extensive
research studies conducted by the bureau beginning in 1993 that focused
on initial plans for three uses of nonbureau information to:
* derive census totals for some nonresponding households,
* enhance the coverage measurement operations, and:
* help provide missing content from otherwise responding households.
Although bureau officials determined that administrative records had
the potential to improve coverage, the bureau abandoned plans to fund
and more fully develop an administrative records database in February
1997. While the lack of funding may have been a contributing factor,
bureau documents indicated that this action was primarily due to
questions about the accuracy and quality of administrative records and
issues of privacy protection.
7. Problems with multiple language questionnaires. Bureau officials
cited several funding and operational problems with census
questionnaires in the five languages that were used other than
English.[Footnote 11] In 1995, the bureau planned to mail forms in both
Spanish and English to areas with high concentrations of Spanish
speakers and produce forms in other languages as needed. In March 1997,
in response to requests for forms in other languages, the bureau
announced its intent to print questionnaires in multiple languages in
an effort to increase the mail response rate. The bureau selected four
additional languages as a manageable number based upon a perceived
demand. However, the bureau could not determine how to pinpoint the
communities that needed the non-English questionnaires. Instead, the
bureau indicated in a mailing that the questionnaires were available in
five languages and if an individual wanted a questionnaire in a
language other than English, the individual had to specifically request
the questionnaire in that language. As a result, the bureau did not
know the number of questionnaires to print in the five languages until
late in the process. Finally, the bureau did not have the time to
comprehensively assess the demand for questionnaires in other
languages.
8. Cost-effective use of emerging data capture technology. Early bureau
research assessed current and emerging data capture technologies, such
as electronic imaging, optical mark recognition, and hand-held devices,
which offered the potential for significant cost reductions in
processing large volumes of data. Bureau officials indicated they were
unsure of their exact requirements for the emerging data capture
technologies, and this resulted in most contracts being cost-
reimbursement contracts[Footnote 12] that required more funding than
planned. The bureau estimated that it ultimately spent about $500
million on contracts to improve the data capturing process.
Bureau officials also stated that they did not have the time to fully
develop and test the data capture systems or data capture centers, both
of which were contracted for the first time in the 2000 Census. For
example, the bureau said it could not adequately prepare for the full
development and testing of the imaging contract. As a consequence, the
first imaging test did not occur until 1998, and bureau officials
stated that it became clear that imaging was not working due to
technical and implementation problems. To some extent, this is not
unexpected when implementing new technologies. Although the contractor
and the bureau felt the system was not ready, it was tested anyway due
to the short time frame and major problems developed. Even though the
system eventually became operational in time for the 2000 Census count,
bureau officials indicated that this occurred at a higher than
anticipated risk and cost.
9. More use of the Internet. In the early 1990s, the full impact of the
Internet as a global communications tool was not yet envisioned.
Officials indicated that the bureau did not have sufficient time and
funding during the planning phase to fully understand and test all the
implications of using the Internet as a vehicle for census responses.
In addition, the bureau‘s major concern was that computer security
issues had not been adequately addressed, particularly since census
information must be protected and significant penalties may be imposed
for unauthorized disclosure.[Footnote 13] Also, the public perception
of using the Internet as a response medium had not been fully explored.
Nevertheless, in February 1999, the bureau established a means for
respondents to complete the 2000 Census short forms on the Internet
protected by a 22-digit identification number. According to bureau
officials, they received about 60,000 short forms via the Internet. The
rapid evolution of the Internet has the potential to significantly
reduce bureau workload and the large volume of paper forms for the 2010
Census.
10. Preparation for dress rehearsals. Bureau officials cited many
problems during the fiscal year 1998 dress rehearsals for the 2000
Census that were a direct result of funding levels in the early
planning and development years. They stated that because of delays in
receiving funding in the fall of 1997, they had to delay the dress
rehearsal census day from April 4 to April 18, 1998. In addition,
because many new items were incomplete or still under development, the
bureau said it could not fully test them during the dress rehearsals
with any degree of assurance as to how they would affect the 2000
Census.
However, despite these problems, the bureau testified in March 1998
that all preparatory activities for the dress rehearsal--mapping,
address listing, local updates of addresses, opening and staffing
offices, and printing questionnaires--had been completed.[Footnote 14]
In 1999, the bureau issued an evaluation that concluded that all in
all, the Census 2000 dress rehearsal was successful.[Footnote 15] The
evaluation also stated that the bureau produced population numbers on
time that compared favorably with independent benchmarks. It also
acknowledged some problems, but devised methods to address those
problems. Although the bureau conceded that planning efforts could be
improved, the lack of funding did not appear to be a significant issue,
except as it affected the ability to earlier plan the dress rehearsal.
2000 Census Planning Provides Lessons Learned for The 2010 Census:
The bureau‘s experience in preparing for the 2000 Census underscores
the importance of solid, upfront planning and adequate funding levels
to carry out those plans. As we have reported in the past,[Footnote 16]
planning a decennial census that is acceptable to stakeholders includes
analyzing the lessons learned from past practices, identifying
initiatives that show promise for producing a better census while
controlling costs, testing these initiatives to ensure their
feasibility, and convincing stakeholders of the value of proposed
plans.
Contributing factors to the funding reductions for the 2000 Census were
the bureau‘s persistent lack of comprehensive planning and priority
setting, coupled with minimal research, testing, and evaluation
documentation to promote informed and timely decision making. Over the
course of the decade, the Congress, GAO, and others criticized the
bureau for not fully addressing such areas as (1) capitalizing on its
experiences from past decennial censuses to serve as lessons to be
learned in future planning,
(2) documenting its planning efforts, particularly early in the
process,
(3) concentrating its efforts on a few critical projects that
significantly affected the census count, such as obtaining a complete
and accurate address list, (4) presenting key implementation issues
with decision milestones, and (5) identifying key performance measures
for success.
* Capitalizing on experiences from past censuses. In a fiscal year 1993
conference report,[Footnote 17] the Congress stated that the bureau
should direct its resources towards a more cost-effective census design
that would produce more accurate results than those from the 1990
Census. Further, the Congress expected the bureau to focus on realistic
alternative means of collecting data, such as the use of existing
surveys, rolling sample surveys, or other vehicles and that cost
considerations should be a substantial factor in evaluating the
desirability of design alternatives.
In March 1993 we testified[Footnote 18] that time available for
fundamental census reform was slipping away and important decisions
were needed by September 1993 to guide planning for 1995 field tests,
shape budget and operational planning for the rest of the census cycle,
and guide future discussions with interested parties. We noted that the
bureau‘s strategy for identifying promising census designs and features
was proving to be cumbersome and time consuming, and the bureau had
progressed slowly in reducing the design alternatives for the next
census down to a manageable number.
* Documenting early planning efforts. It is particularly important
early in the planning process to provide a roadmap for further work. We
found that the bureau did not document its 2000 Census planning until
late in the planning phase. While the U.S. Census Bureau prepared a few
pages to justify its annual budget requests for fiscal years 1991
through 1997, it did not provide a substantive document of its 2000
Census planning efforts until May 1995, and this plan was labeled a
draft.[Footnote 19] Finally, the Congress mandated that the bureau
issue a comprehensive and detailed plan for the 2000 Census within 30
days from enactment of the law.[Footnote 20] On July 12, 1997, the
bureau issued its Report to the Congress--The Plan for Census 2000,
along with its Census 2000 Operational Plan.[Footnote 21]
* Concentrating efforts on a few critical projects. While the bureau
required many activities to count a U.S. population of 281 million
residing in 117.3 million households, a few critical activities
significantly affected the Census 2000 count, such as obtaining a
complete and accurate address list. Although the bureau was aware of
serious problems with its address list development process, it did not
acknowledge the full impact of these problems until the first quarter
of 1997. Based upon its work with the postal service database, the 1995
Census Test, and pilot testing at seven sites, the bureau had gained
sufficient evidence that its existing process would result in an
unacceptably inaccurate address list due to:
* inconsistencies in the quality of the postal service database across
the nation;
* missing addresses for new construction;
* difficulties in identifying individual units in multiunit structures,
such as apartment buildings; and:
* inability of local and tribal governments to provide usable address
lists.
In September 1997, the bureau acknowledged these problems and proposed
changes.[Footnote 22] However, we believe that this action occurred too
late in the planning process and was not given a higher priority to
benefit the 2000 Census enumeration.
* Presenting key implementation issues and decision milestones. The
bureau discussed program areas as part of its annual budget requests
for fiscal years 1991 through 1997, but the requests did not identify
key implementation issues with decision milestones to target its
planning activities. Decision milestones did not appear until July
1997, when the bureau issued its Census 2000 Operational Plan.
Stakeholders such as the Congress are more likely to approve plans and
funding requests when they are thoroughly documented and include key
elements such as decision milestones.
* Identifying key performance measures. Census planning documents
provided to us through fiscal year 1997 did not identify key
performance measures. We believe that identifying key performance
measures is critical to assessing success in the planning phase of the
census and can provide quantitative targets for accomplishments by
framework, activity, and individual projects. Such measures could
include performance goals such as increasing mail response rates,
reducing population overcount and undercount rates, and improving
enumerator productivity rates.
Conclusions:
The lessons learned from planning the 2000 Census become even more
crucial in planning for the next decennial census in 2010, which has
current unadjusted life cycle cost estimates ranging from $10 billion
to $12 billion. Thorough and comprehensive planning and development
efforts are crucial to the ultimate efficiency and success of any
large, long-term project, particularly one with the scope, magnitude,
and deadlines of the U.S. decennial census. Initial investment in
planning activities in areas such as technology and administrative
infrastructure can yield significant gains in efficiency,
effectiveness, and cost reduction in the later implementation phase.
The success of the planning and development activities now occurring
will be a major factor in determining whether this large investment
will result in an accurate and efficient national census in 2010.
Critical considerations are:
* early planning;
* a comprehensive and prioritized plan of goals, objectives, and
projects;
* milestones and performance measures; and:
* documentation to support research, testing, and evaluation.
A well-supported plan early in the process that includes these elements
will be a major factor in ensuring that stakeholders have the
information to make funding decisions.
Recommendation for Executive Action:
As the U.S. Census Bureau plans for the 2010 Census, we recommend that
the Secretary of Commerce direct that the bureau provide comprehensive
information backed by supporting documentation in its future funding
requests for planning and development activities, that would include,
but is not limited to, such items as:
* specific performance goals for the 2010 Census and how bureau
efforts, procedures, and projects would contribute to those goals;
* detailed information on project feasibility, priorities, and
potential risks;
* key implementation issues and decision milestones; and:
* performance measures.
Agency Comments and Our Evaluation:
In commenting on our report, the department agreed with our
recommendation and stated that the bureau is expanding the documents
justifying its budgetary requests. For example, the bureau cited a
document which outlines planned information technology development and
activities throughout the decennial cycle of the 2010 Census. The
bureau also included a two-page document, Reengineering the 2010
Census, which presented three integrated components and other plans to
improve upon the 2000 Census. In this regard, it is essential that, as
we recommended, the bureau follow through with details and
documentation to implement these plans, define and quantify performance
measures against goals, and provide decision milestones for specific
activities and projects.
As agreed with you office, unless you announce its contents earlier, we
plan no futher distribution of this report until 7 days after its
issuance date. At that time, we will send copies of this report to the
Chairman and Ranking Minority Member of the Senate Committee on
Governmental Affairs, the House Committee on Government Reform, and the
House Subcommittee on Civil Service, Census, and Agency Organization.
We will also send copies to the Director of the U.S. Census Bureau, the
Secretary of Commerce, the Director of the Office of Management and
Budget, the Secretary of the Treasury, and other interested parties.
This report will also be available on GAO‘s home page at http://
www.gao.gov.
If you or your staffs have any questions concerning this report, please
contact Gregory D. Kutz at (202) 512-9095 or kutzg@gao.gov, Patricia A.
Dalton at (202) 512-6806 or daltonp@gao.gov, or Roger R. Stoltz,
Assistant Director, at (202) 512-9408 or stoltzr@gao.gov. Key
contributors to this report were Corinne P. Robertson, Robert N.
Goldenkoff, and Ty B. Mitchell.
Gregory D. Kutz
Director
Financial Management and Assurance:
Signed by Gregory D. Kutz:
Signed by:
Patricia A. Dalton
Director
Strategic Issues:
[End of section]
Appendixes:
Appendix I: Objectives, Scope, and Methodology:
The objectives of our review focused on the planning and development
phase of the 2000 Census that we classified as covering fiscal years
1991 through 1997 and (1) the funding requested, received, and
obligated with funding received and obligated by major planning
category, (2) funding and other factors that affected planning efforts,
and (3) lessons learned for the 2010 Census.
To determine the amount of 2000 Census planning and development funding
requested, received, and obligated, we obtained and analyzed annual
decennial census budgets included in the President‘s Budgets for fiscal
years 1991 through 1997, budgets subsequently received after
appropriation by the Congress, and amounts later obligated for the
purchase of goods and services by the bureau against those budgets. We
then obtained explanations from senior bureau officials for significant
variances in these budgets and the effect on decennial planning and
development. However, we did not assess the efficiency of the budgeting
process and the validity, accuracy, and completeness of obligations
against budgeted amounts received.
To determine the funding received and obligated by major planning
category for 2000 Census planning and development, we obtained and
analyzed funding requested, received, and obligated by framework,
activity, project, and object class and examined annual operational
plans. However, our analysis was hampered by the bureau‘s inconsistent
use of categories that evolved from 1 activity of general planning in
fiscal year 1991, 8 major study areas in fiscal years 1992 and 1993,
and 8 to 15 broad categories called frameworks beginning in 1994. For
internal management and reporting, the bureau further identified
program efforts by activities and projects that have varied since
fiscal year 1991.[Footnote 23] Additionally, the bureau expanded,
contracted, or modified program names and descriptions making
comparisons more difficult. We also obtained explanations from bureau
officials for significant efforts and variances in its funding received
and obligation of planning and development funding for the 2000 Census.
However, we did not assess the merits of budgeting by program and the
subsequent validity, accuracy, and completeness of obligations.
To identify funding and other factors that affected planning efforts,
we analyzed significant changes in funding requested, received, and
obligated at the framework level; identified initiatives that were
reduced, eliminated, or severely curtailed; discussed the effect of
these areas with bureau officials; and evaluated bureau responses. We
also reviewed various reports, testimony, and supporting documents
prepared by the bureau, GAO, and others. However, we could not
determine what effect, if any, that higher levels of funding might have
had on 2000 Census operations. These factors are dependent upon actual
implementation and the results of management decisions that may or may
not have occurred.
To provide lessons learned for the 2010 Census, we identified areas for
improvement and obtained support from bureau, GAO, and congressional
reports, testimony, interviews, and other documents.
Our work was performed in Washington, D.C. and at U.S. Census Bureau
headquarters in Suitland, Maryland between January and July 2001 when
our review was suspended due to an inability to obtain access to
certain budget records. After lengthy discussions with senior officials
of the bureau, Department of Commerce, and OMB, and consultation with
your staffs, this access issue was resolved in May 2002 and we
completed our analysis in June 2002. Our work was done in accordance
with U.S. generally accepted government auditing standards, except that
we did not audit budget and other financial data provided by the U.S.
Census Bureau.
On October 16, 2002, the Department of Commerce provided written
comments on a draft of this report, including two attachments. These
comments are presented in the ’Agency Comments and Our Evaluation“
section of the report and are reprinted in appendix III, except for the
second attachment, Potential Life-Cycle Savings for the 2010 Census,
which is currently under revision and is outside the scope of our
review.
[End of section]
Appendix II: Analysis of Funding by Fiscal Year for Planning and
Development of the 2000 Census:
This appendix includes our analysis of 2000 Census funding requested,
received, and obligated, and funding received and obligated by major
planning category for fiscal years 1991 through 1997. Our analysis was
hampered by the bureau‘s inconsistent use of major planning categories
that evolved over the period as follows:
* 1 activity of general planning in fiscal year 1991,
* 8 major study areas in fiscal years 1992 and 1993, and:
* 8 to 15 broad categories called frameworks beginning in 1994.
For internal management and reporting, the bureau further identified
program efforts by activities and projects that have varied since
fiscal year 1991. In addition, the bureau expanded, contracted, or
modified program names and descriptions making comparisons more
difficult.
Fiscal Year 1991 and 1992 Funding:
In March 1991 we testified[Footnote 24] that fundamental census reform
was needed because escalating costs and the apparently increased
undercount of the 1990 Census suggested that the current census
methodology may have reached the limits of its effectiveness. Of three
principles we presented, the last was that the Department of Commerce
must be willing to invest sufficient funds early in the decade to
achieve cost savings and census improvements in 2000. In fact, OMB
deemed some of the Department of Commerce requests to fund early census
reform as insufficient and doubled the department‘s requested amounts
to $1.5 million for fiscal year 1991 and $10.1 million for fiscal year
1992. These amounts were included in the President‘s Budgets and the
Congress concurred by authorizing the full amount requested. Census
planning officials said that if OMB had not augmented the department‘s
request, testing of reform options for 2000 would have been
constrained.
For the first year of the 7-year 2000 Census planning and development
phase, the fiscal year 1991 funding received was $1.5 million and the
bureau obligated the entire amount. The funding contained only one
category of general planning for the 2000 Census with funds to be used
for:
* completion of detailed cost-benefit studies of alternatives designs
for conducting the decennial census;
* exploration of new technologies to improve the 2000 Census;
* establishment of research and development efforts for administrative
methods and modeling and estimation techniques; and:
* planning of field tests in fiscal year 1993 to include new census
content, methods, technologies, and field structures.
Because total amounts were small and involved only general planning,
there were no significant variances. We noted that about 46 percent of
the funding was obligated for personnel costs relating to 19 full-time
equivalent (FTE) staff, 29 percent for services including consultants;
and the remaining 25 percent for space, supplies, travel, and other
costs.
Fiscal year 1992 funding received was $10.1 million and the bureau
obligated $9.4 million against it. The funding now identified eight
major study areas for the 2000 Census as indicated in table 2.
Table 2: 2000 Census Planning and Development Funding for Fiscal Year
1992 (dollars in millions):
Major study area: 1; Description: Simplify questionnaire; Amount: $1.0.
Major study area: 2; Description: Administrative records, modeling, and
estimation; Amount: 2.8.
Major study area: 3; Description: Technology options for questionnaire
distribution and collection; Amount: 3.5.
Major study area: 4; Description: Various uses of sampling; Amount: .9.
Major study area: 5; Description: Techniques for special areas and
subpopulations; Amount: .3.
Major study area: 6; Description: TIGER[B] enhancements; Amount: [A].
Major study area: 7; Description: Address list maintenance; Amount: .8.
Major study area: 8; Description: Administration; Amount: .8.
Description: Major study area : Total; Amount: Major study area :
$10.1.
[A] Activity or amounts were less than $.1 million.
[B] Topologically Integrated Geographic Encoding and Referencing
system.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1992, the bureau experienced almost a six-fold increase
in its funding received of $10.1 million over the $1.5 million for
fiscal year 1991. About half of the fiscal year 1992 funding was
obligated for personnel costs as a result of almost a five-fold
increase in FTE staff from 19 in fiscal year 1991 to 111 in fiscal year
1992 to work on decennial planning and development issues. Services,
including consultants, accounted for another quarter of the obligations
with the remaining quarter for space, supplies, travel, and other
costs. Technology options included a $1.7 million services contract to
develop emerging data capture technology to compile census statistics.
Fiscal Year 1993 Funding:
For fiscal year 1993, the Congress reduced the President‘s Budget
request of $19.4 million for 2000 Census planning and development to
$13.7 million,[Footnote 25] for a reduction of about 29 percent. As a
result of this $5.7 million reduction, the bureau made significant cuts
in its funding of techniques for special areas and subpopulations by
$2.2 million, or about 70 percent, and also eliminated activities to:
* establish contacts with state and local government budgeted for $1.6
million,
* assess customer needs budgeted for $1.0 million,
* survey public motivation budgeted for $.8 million, and:
* prepare infrastructure for a 1995 Census Test budgeted for $.5
million.
In a fiscal year 1993 conference report,[Footnote 26] the Congress
stated that the bureau should direct its resources towards a more cost-
effective census design that will produce more accurate results than
those from the 1990 Census. For example, the bureau‘s research in
fiscal year 1992 indicated that reducing the number of questions on the
census form is an important way to increase response, thereby
increasing accuracy and reducing cost.
Therefore, the Congress expected the bureau to focus on realistic
alternative means of collecting data, such as the use of existing
surveys, rolling sample surveys, or other vehicles and that cost
considerations should be a substantial factor in evaluating the
desirability of design alternatives.
In March 1993 we testified[Footnote 27] that time available for
fundamental census reform was slipping away and important decisions
were needed by September 1993 to guide planning for 1995 field tests,
shape budget and operational planning for the rest of the census cycle,
and guide future discussions with interested parties. The bureau‘s
strategy for identifying promising census designs and features was
proving to be cumbersome and time consuming, and the bureau had
progressed slowly in reducing the design alternatives for the next
census down to a manageable number.
Fiscal year 1993 funding received was $13.7 million and the bureau
obligated $13.5 million against it. The budget continued to identify
eight major study areas for the 2000 Census as indicated in table 3.
Table 3: 2000 Census Planning and Development Funding for Fiscal Year
1993 (dollars in millions):
Major study area: 1; Description: Simplify questionnaire; Amount: $1.2.
Major study area: 2; Description: Administrative records, modeling and
estimation; Amount: 4.0.
Major study area: 3; Description: Technology options for questionnaire
distribution and collection; Amount: 4.3.
Major study area: 4; Description: Various uses of sampling; Amount:
1.6.
Major study area: 5; Description: Techniques for special areas and
subpopulations; Amount: .9.
Major study area: 6; Description: TIGER[B] Support; Amount: .4.
Major study area: 7; Description: Address list maintenance; Amount:
[A].
Major study area: 8; Description: Administration; Amount: 1.3.
Description: Major study area : Total; Amount: Major study area :
$13.7.
[A] Activity or amounts were less than $.1 million.
[B] Topologically Integrated Geographic Encoding and Referencing
system.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1993, the bureau experienced a 36 percent increase in
its funding received of $13.7 million over the $10.1 million for fiscal
year 1992. About 53 percent of the fiscal year 1993 funding was
obligated for personnel costs as a result of a 48 percent increase in
FTE staff from 111 in fiscal year 1992 to 164 in fiscal year 1993 to
work on decennial planning and development issues. Services, including
consultants, accounted for about 11 percent of the funding with the
remaining 36 percent used for space, supplies, travel, and other costs.
Fiscal year 1993 was identified by the bureau as the beginning of a 3-
year period to identify the most promising changes to be integrated in
the 1995 Census Test.
Fiscal Year 1994 Funding:
For fiscal year 1994, the Congress reduced the President‘s Budget
request of $23.1 million for 2000 Census planning and development to
$18.7 million,[Footnote 28] for a reduction of about 19 percent. As a
result of this $4.4 million reduction, the bureau eliminated decennial
operational preparation for $2.5 million, and reduced funding for
questionnaire design and cost modeling by $1.6 million or 70 percent.
In May 1993 we testified[Footnote 29] that the U.S. Census Bureau had
altered its decision-making approach and refocused its 2000 Census
research and development efforts. Driven by its impending September
1993 deadline for deciding which designs to test in 1995 for the 2000
Census, the bureau recommended rejecting all 14 design alternatives
that had formed the framework of its research program that was under
study for a year. Instead, the bureau reverted to an earlier approach
of concentrating favorable features into the design for application in
the 2000 Census.
A fiscal year 1994 House Appropriations Committee report[Footnote 30]
cited our May 1993 testimony and stated that it was unacceptable for
the bureau to conduct the 2000 Census under a process that followed the
general plan used in the 1990 Census. A fiscal year 1994 conference
report[Footnote 31] expressed concern that the U.S. Census Bureau had
not adequately addressed cost and scope issues for the 2000 Census and
expected the Department of Commerce and OMB to take a more active role
in planning for the decennial census to ensure that data requirements
for federal agencies and state and local government were considered in
the planning effort.
In October 1993 we testified[Footnote 32] that the U.S. Census Bureau‘s
research and development efforts had been slowed by its changing
planning strategy and that the bureau still faced the difficult task of
integrating its Test Design Recommendation proposals into a detailed
implementation plan for the 1995 census test. We noted that the
bureau‘s plans to conduct research and evaluations for such promising
proposals as the one-number census,[Footnote 33] sampling for
nonresponse, and defining the content of the census were in a state of
flux. Other important research and planning activities, such as
improving the address list and using new automated techniques to
convert respondent answers to machine-readable format, were behind
schedule. Funding for research and test census preparation in fiscal
years 1994 and 1995 was in doubt as evidenced by the budget cuts
proposed by the House Appropriations Committee and the opinions
expressed in its report accompanying the fiscal year 1994
appropriations bill.[Footnote 34]
The bureau obligated the entire amount of its fiscal year 1994 funding
received of $18.7 million. Funding originally contained 6 design areas
for 2000 Census research and development, the 1995 Census test, and
decennial operational preparation but was later revised to present
funds received and obligated in 13 frameworks of effort as indicated in
table 4.
Table 4: 2000 Census Planning and Development Funding for Fiscal Year
1994 (dollars in millions):
Framework number: 1; Description: Program development and management
information; Amount: $1.1.
Framework number: 2; Description: Content requirements and public use
forms; Amount: .7.
Framework number: 3; Description: Test censuses & dress rehearsal;
Amount: 5.5.
Framework number: 4; Description: Decennial geographic support; Amount:
[A].
Framework number: 5; Description: Evaluation & development; Amount:
7.1.
Framework number: 6; Description: Address list compilation; Amount: .1.
Framework number: 7; Description: Precensus day operations & support
systems; Amount: [A].
Framework number: 8; Description: Stakeholder education and
consultation; Amount: .2.
Framework number: 9; Description: Puerto Rico & outlying areas; Amount:
[A].
Framework number: 10; Description: Tabulation, publication, & data user
services; Amount: [A].
Framework number: 11; Description: Automated/telecommunication
support; Amount: 4.0.
Framework number: 12; Description: Year 2001 and beyond; Amount: [A].
Framework number: 13; Description: Cooperation with the U.S Postal
Service; Amount: [A].
Description: Framework number : Total; Amount: Framework number :
$18.7.
[A] Activity or amounts were less than $.1 million.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1994, the bureau experienced a 36 percent increase in
its funding received of $18.7 million over the $13.7 million for fiscal
year 1993. About 44 percent of the fiscal year 1994 funding was
obligated for personnel costs as a result of a 34 percent increase in
FTE staff from 164 in fiscal year 1993 to 220 in fiscal year 1994 to
work on decennial planning and development issues. Services, including
consultants, accounted for another 13 percent of obligations with the
remaining 43 percent for space, supplies, travel, and other costs. We
noted that six frameworks received little or no funding and three
frameworks accounted for 89 percent of the fiscal year 1994 funds
received and obligated as follows:
Framework 5 - Evaluation and development consumed $7.1 million or 38
percent of funding received and obligated for research and
developmental work to support the 1995 census test. This included:
* research on the use of matching keys beyond just a person‘s residence
address to develop matching procedures that would allow the bureau to
make use of person-based administrative records files that do not have
a current residential address;
* research on various uses of sampling including technical and policy
issues on conducting the entire census on a sample basis and conducting
only the nonresponse follow-up portion of the census on a sample basis;
and:
* race and ethnicity studies including extensive consultation with
stakeholders, focus group testing, and planning of field tests.
Framework 3 - Test census and dress rehearsal consumed $5.5 million or
29 percent of funding received and obligated to increase 1995 Census
Test activities from preliminary studies and planning to the full-scale
preparatory level program. These included such activities as:
* completion of questionnaire content determination,
* analysis of a database of population characteristics by geographic
area to make selections of test sites,
* determination of evaluation program objectives for the test, and:
* determination of objectives for and design stakeholder consultation.
Framework 11 - Automation/telecommunication support consumed $4.0
million or 21 percent of funding received and obligated for automated
systems design and acquisition of data capture technology to upgrade
the 1990 Census system (FACT90) to a 2000 Census system (DCS 2000).
Fiscal Year 1995 Funding:
For fiscal year 1995, the Congress reduced the President‘s Budget
request of $48.6 million for 2000 Census planning and development to
$42.0 million for a reduction of about 14 percent. As a result of this
$6.6 million reduction, the bureau eliminated $9.0 million for
decennial operation preparation and $.8 million for 1996 testing while
increasing funding for program development and other areas by $3.2
million.
In January 1994 we testified[Footnote 35] that while we were encouraged
by the U.S. Census Bureau‘s recent focus on testing specific proposals
to modify the census methodology, we believed that the bureau must
aggressively plan for and carefully implement its research, testing,
and evaluation programs. Further, the results of those efforts must be
available to make fully informed and timely decisions and build needed
consensus among key stakeholders and customers for changes in the 2000
Census.
A fiscal year 1995 Senate Appropriations Committee report[Footnote 36]
strongly recommended that the bureau adopt more cost-effective means of
conducting the next census as the budgetary caps and strict employment
ceilings adopted by the President and the Congress would not
accommodate a repeat of the process used in the 1990 Census.
Fiscal year 1995 funding received was $42.0 million and the bureau
obligated $40.9 million against it. The number of frameworks increased
to 15 as indicated in table 5.
Table 5: 2000 Census Planning and Development Funding for Fiscal Year
1995 (dollars in millions):
Framework number: 1; Description: Program development and management
information; Amount: $2.5.
Framework number: 2; Description: Content requirements and public use
forms; Amount: 1.7.
Framework number: 3; Description: Test censuses & dress rehearsal;
Amount: 30.0.
Framework number: 4; Description: Decennial geographic products &
services; Amount: 1.3.
Framework number: 5; Description: Evaluation & development; Amount:
5.1.
Framework number: 6; Description: Address list compilation; Amount:
[A].
Framework number: 7; Description: Precensus day operations & support
systems; Amount: [A].
Framework number: 8; Description: Postcensus day operations; Amount:
[A].
Framework number: 9; Description: Census marketing, communications, &
partnership; Amount: .5.
Framework number: 10; Description: Puerto Rico & other island
territories; Amount: [A].
Framework number: 11; Description: Tabulation, dissemination, &
customer services; Amount: [A].
Framework number: 12; Description: Automation/telecommunication
support; Amount: [A].
Framework number: 13; Description: Year 2001 & beyond; Amount: [A].
Framework number: 14; Description: Cooperation with the U.S. Postal
Service; Amount: [A].
Framework number: 15; Description: Follow-on surveys; Amount: .9.
Description: Framework number : Total; Amount: Framework number :
$42.0.
[A] Activity or amounts were less than $.1 million.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1995, the bureau experienced a 125 percent increase in
its funding received of $42.0 million over the $18.7 million for fiscal
year 1994. About 51 percent of the fiscal year 1995 funding was
obligated for personnel costs as a result of a 211 percent increase in
FTE staff from 220 in fiscal year 1994 to 685 in fiscal year 1995 to
work on decennial planning and development issues. Services, including
consultants, accounted for about 7 percent of the obligations with the
remaining 42 percent for space, supplies, travel, and other costs. We
noted that eight frameworks received little or no funding and Framework
3 accounted for over 70 percent of fiscal year 1995 funds received and
obligated. The main focus of Framework 3 was conducting the 1995 Census
Test, in order to select by December 1995 the features to be used for
the 2000 Census. According to census plans and our discussions with
officials, the bureau focused on the following major areas.
* Complete preparation for the 1995 Census Test, conduct the test, and
begin evaluations in order to select the features to be used for the
2000 Census. In addition, the bureau would conduct a full-scale census
test in four district office areas that would be the culmination of the
research and development program.
* Investigate, develop, test, and evaluate components of a continuous
measurement system as a replacement for the 2000 Census sample data
questionnaire.
* Develop, test, and evaluate various matching keys for the automated
and clerical matching and unduplicating[Footnote 37] systems developed
under the direction of the matching research and specifications working
group.
* Conduct activities independent of the research and development
program; these are preparatory activities required to implement the
2000 Census regardless of the design. This included such activities as
planning the address list update activities as necessary to supplement
the Master Address File (MAF) for use in the 2000 Census and begin
initial planning of the field organization structure for the 2000
Census.
* Recommend the broad scope of content that should be included in the
2000 Census questionnaire based on consulting with both federal and
nonfederal data users, and begin planning for small special purpose
tests to supplement or follow up on the 1995 Census Test.
Fiscal Year 1996 Funding:
For fiscal year 1996, the Congress reduced the President‘s Budget
request of $60.1 million for 2000 Census planning and development to
$51.3 million,[Footnote 38] for a reduction of about 15 percent. As a
result of this $8.8 million reduction, the bureau reduced funding for
field data collection and support systems by $9.9 million or 43 percent
while increasing funding in other areas.
In October 1995 we testified[Footnote 39] that the U.S. Census Bureau
had decided to make fundamental changes to the traditional census
design such as shortening census questionnaires, developing an accurate
address list, and sampling households that fail to respond to
questionnaires. However, we noted that successful implementation of
these changes would require aggressive management by the bureau and
that the window of opportunity for the Congress to provide guidance on
these changes and applicable funding was closing.
A fiscal year 1996 conference report[Footnote 40] continued to express
concern about progress related to the next decennial census. It
cautioned the bureau that the cost of the 2000 Census had to be kept in
check and only through early planning and decision making could costs
be controlled. The report further recognized that fiscal year 1996 was
a critical year in planning for the decennial census, and that numerous
decisions will be made and preparations taken which will have a
significant bearing on the overall cost of conducting the census, as
well as the design selected.
The bureau obligated the entire amount of its fiscal year 1996 funding
received of $51.3 million. Beginning with fiscal year 1996, the number
of frameworks was reduced to eight as indicated in table 6 below.
Table 6: 2000 Census Planning and Development Funding for Fiscal Year
1996 (dollars in millions):
Framework number: 1; Description: Program development and management;
Amount: $8.2.
Framework number: 2; Description: Data content and products; Amount:
9.6.
Framework number: 3; Description: Field data collection and support
systems; Amount: 13.3.
Framework number: 4; Description: Address list development; Amount:
2.3.
Framework number: 5; Description: Automated data processing and
telecommunications support; Amount: 6.8.
Framework number: 6; Description: Testing, evaluation, and dress
rehearsal; Amount: 9.4.
Framework number: 7; Description: Puerto Rico, Virgin Islands, and
Pacific areas; Amount: .3.
Framework number: 8; Description: Marketing, communications, and
partnerships; Amount: 1.4.
Description: Framework number : Total; Amount: Framework number :
$51.3.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1996, the bureau experienced a 22 percent increase in
its funding received of $51.3 million over the $42.0 million for fiscal
year 1995. About 44 percent of the fiscal year 1996 funding was
obligated for personnel costs as a result of a 5 percent decrease in
FTE staff from 685 in fiscal year 1995 to 653 in fiscal year 1996 to
work on decennial planning and development issues. Services, including
consultants, accounted for about 13 percent of the obligations with the
remaining 43 percent for space, supplies, travel, and other costs.
Three frameworks incurred over 60 percent of funding received and
obligated for the following.
* Framework 3 - Field data collection and support systems incurred
costs of $13.3 million including $4.4 million to develop personnel and
administrative systems for field office enumeration; $3.1 million for
precensus day data collection activities; and $2.0 million for
automation acquisition and support for field offices.
* Framework 2 - Data content and products incurred costs of $9.6
million including $4.4 million to develop and produce questionnaires
and public use forms for the census including conduct of a National
Content Test; $2.9 million for race and ethnicity testing of concepts
and respondent understanding and wording of the race and ethnicity
questions; and $1.6 million for continued work with federal and
nonfederal data users in the content determination process to prepare
for the congressional submission by April 1, 1997.
* Framework 6 - Testing, evaluations, and dress rehearsals incurred
costs of $9.4 million including $3.3 million for an Integrated Coverage
Measurement (ICM) special test;[Footnote 41] $2.6 million for research
and development on sampling and sampling methods for the 2000 decennial
count; and $2.1 million for 1995 Census Test coverage and evaluation.
Fiscal Year 1997 Funding:
For fiscal year 1997, the Congress reduced the President‘s Budget
request of $105.9 million for 2000 Census planning and development to
$86.4 million,[Footnote 42] for a reduction of about 18 percent. As a
result of this $19.5 million reduction, the bureau reduced funding for
marketing, communications, and partnerships by $14.4 million or 76
percent, and field data collection and support systems by $23.6 million
or 53 percent, while increasing amounts in other areas by $18.5
million.
A fiscal year 1996 House Appropriation Committee report[Footnote 43]
expressed concern that the bureau appeared not to have developed
options and alternative plans to address issues of accuracy and cost.
In addition, sufficient progress had not been made on issues the
committee had highlighted many times--the number of questions on the
long-form and reimbursement from other agencies for inclusion of such
questions to assure that the question is important.
The bureau obligated the entire amount of its fiscal year 1997 budget
of $86.4 million. Planning continued in eight frameworks as indicated
in table 7.
Table 7: 2000 Census Planning and Development Funding for Fiscal Year
1997 (dollars in millions):
Framework number: 1; Description: Program development and management;
Amount: $5.4.
Framework number: 2; Description: Data, content, and products; Amount:
12.3.
Framework number: 3; Description: Field data collection and support
systems; Amount: 20.9.
Framework number: 4; Description: Address list development; Amount:
2.5.
Framework number: 5; Description: Automated data processing and
telecommunications support; Amount: 20.2.
Framework number: 6; Description: Testing, evaluation, and dress
rehearsal; Amount: 19.8.
Framework number: 7; Description: Puerto Rico, Virgin Islands, and
Pacific areas; Amount: .8.
Framework number: 8; Description: Marketing, communications, and
partnerships; Amount: 4.5.
Description: Framework number : Total; Amount: Framework number :
$86.4.
Source: U.S. Census Bureau.
[End of table]
For fiscal year 1997, the bureau experienced a 68 percent increase in
its funding received of $86.4 million over the $51.3 million for fiscal
year 1996. About 63 percent of the fiscal year 1997 funding was
obligated for personnel costs as a result of a 36 percent increase in
FTE staff from 653 in fiscal year 1996 to 891 in fiscal year 1997 to
work on decennial planning and development issues. Services, including
consultants, accounted for about 25 percent of the obligations with the
remaining 12 percent for space, supplies, travel, and other costs.
The bureau viewed fiscal year 1997 as pivotal, since this was the year
when research and testing activities culminated into operational
activities and marked the end of the planning and development phase of
the 2000 Census. For the fiscal year, four frameworks incurred about 85
percent of funding received and obligated as follows.
Framework 3 - Field data collection and support systems incurred $20.9
million for activities under precensus day operations and support
systems, and postcensus day operations. Projects included:
* $4.1 million for geographic patterns including questionnaire delivery
methodologies by area and corresponding automated control systems;
* $4.0 million for planning of data collection efforts including
activities for truncation and/or the use of sampling for nonresponse
follow-up and increased efforts to develop procedures for enumerating
special populations such as the military, maritime, institutional,
migrant, reservation, and those living in other than traditional
housing units;
* $3.8 million for direction and control by 12 regional offices that
would provide logistical support and direct enumeration efforts by
local census offices; and:
* $3.1 million for planning and developing personnel and administrative
systems to support 2000 Census data collection and processing
activities, such as types of positions, pay rates, personnel and
payroll processes, and systems, space, and security requirements.
Framework 5 - Automation/telecommunication support incurred $20.2
million for activities to include evaluating proposals for the
acquisition of automation equipment and related services, funding the
development of prototype systems, and moving toward awarding contracts
to implement such systems for the 2000 Census. Projects included
setting up data capture systems and support to process census
questionnaire responses and telecommunication systems required to
provide nationwide toll-free 800 number services to answer respondent
questions and to conduct interviews.
Framework 6 - Testing, evaluation, and dress rehearsal incurred $19.8
million for the following activities:
* $3.7 million to begin gearing up for the 1998 dress rehearsal in
order to prepare personnel to conduct the census testing efficiently
and effectively; and:
* $7.0 million to conduct activities for ICM special testing and
American Indian Reservation (AIR) test census such as:
* questionnaire delivery and mail return check-in operations,
* nonresponse followup,
* data capture operations,
* ICM computer-assisted personal visit interviews,
* computer and clerical matching,
* follow-up and after follow-up matching, and:
* evaluation studies.
Framework 2 - Data content and products incurred $12.3 million for
activities related to the development of computer programs and systems
for data tabulation and for the production of paper, machine-readable,
and on-line data products. Projects included:
* $4.5 million to move from research in fiscal year 1996 to
implementation in fiscal year 1997 of the Data Access and Dissemination
System (DADs), including development of the requirements for Census
2000 tabulations from DADs, and development of computer programs and
control systems that will format the processed Census 2000 data for use
in DADs; and:
* $2.2 million towards development of a redistricting program for
Census 2000.
[End of section]
Appendix III: Comments From the Department of Commerce:
THE SECRETARY OF COMMERCE Washington, D.C. 20230:
OCT 16 2002:
Mr. Gregory D. Kutz Director:
Financial Management and Assurance General Accounting Office
Washington, DC 20548:
Dear Mr. Kutz:
The Department of Commerce appreciates the opportunity to comment on
the General Accounting Office draft document entitled 2000 Census:
Lessons Learned for Planning a More Cost-Effective 2010 Census. The
Department‘s comments on this report are enclosed.
Sincerely,
Donald L. Evans:
Signed by Donald L. Evans:
Enclosures:
Comments from the U.S. Department of Commerce U.S. Census Bureau:
U.S. General Accounting Office draft report entitled 2000 Census:
Lessons Learned for Planning a More Cost-Effective 2010 Census:
Comments on the Text of the Report:
The U.S. Census Bureau has reviewed this report carefully and
appreciates this opportunity to respond prior to its publication.
We agree with the General Accounting Office (GAO) that Census 2000 was
very successful. However, this was not attributable to a sound planning
and development effort. Rather, it was due to a tremendous effort at
the time of execution, which resulted in a very costly census conducted
at an unacceptably high level of risk. Accordingly, we also agree with
the fundamental conclusion of the report, that ’thorough and
comprehensive planning and development efforts are crucial to the
ultimate efficiency and success of any large, long-term project,
particularly one with the scope, magnitude, and deadlines of the U.S.
decennial census.“:
We believe that our plan for reengineering the decennial census design
for 2010 directly addresses these issues (see Appendix 1,
’Reengincering the 2010 Census“). However, as GAO stresses, the success
of this effort hinges on our ability to conduct early planning,
development, and testing of all components of the census design. The
GAO points out that the Census Bureau only directed 4 percent of the
life-cycle costs for Census 2000 to early planning and development. For
the 2010 census, this has been increased to 8 percent of the life-cycle
costs of conducting a short-form-only census. This figure increases to
17 percent when the cost of implementing the American Community Survey
and the Master Address File/Topologically Integrated Geographic
Encoding and Referencing (MAF/TIGER) enhancements program is taken into
consideration. (These costs are summarized in Appendix 2, ’Life-Cycle
Savings of the 2010 Census,“ which is currently under revision.):
The GAO recommends that the Census Bureau ’provide comprehensive
information backed by supporting documentation in its future funding
requests for planning and development activities.“ The Census Bureau
wholeheartedly agrees with this recommendation. Consistent with
Department of Commerce and Office of Management and Budget guidelines,
the Census Bureau is expanding the documents justifying budgetary
requests. A primary example of this is the ’Capital Asset Plan and
Business Case“ (Exhibit 300), which outlines the Census Bureau‘s
planned information technology development and activities throughout
the decennial census cycle.
Appendix 2:
Reengineering the 2010 Census U.S. Census Bureau:
September 2002:
The reengineered 2010 census consists of three highly integrated
activities designed to dramatically improve upon an already very good
Census 2000. We will accomplish this by taking advantage of
opportunities for innovations made possible through the expanded use of
technology and targeting of coverage improvement procedures that will
enable the U.S. Census Bureau to:
1. Improve the relevance and timeliness of census long form data.
2. Reduce operational risk.
3. Improve the accuracy of census coverage.
4. Contain costs.
The three integrated components are:
1. Collect and tabulate long-form data every year throughout the decade
through a large household survey (the American Community Survey [ACS]).
2. Enhance and improve our existing address lists and geographic
information system (GIS) data base (Topologically Integrated Geographic
Encoding and Referencing [TIGER]) by bringing them into alignment with
true global positioning system (GPS) coordinates and converting our
TIGER to a commercial off-the-shelf (COTS) database environment.
3. A program of early planning, development, and testing designed to
completely restructure the management and conduct of a short-form-only
census in 2010. That will further reduce the differential undercount
measured since the 1940s without resorting to statistical adjustment
and will provide the savings needed to support this initiative.
Adopting the ACS as the planned replacement for the census long form
will allow the short-form-only census to focus more directly on meeting
the legally mandated collection and issuance of the apportionment and
redistricting data. This will transfer to the ACS the responsibility to
provide estimates of detailed demographic and housing data throughout
the decade. This more timely and therefore more relevant data will
greatly enhance the information currently provided by the once-in-a-
decade long form.
An updated master address file (MAF) and an accompanying improved TIGER
data base with GPS positional accuracy will allow the Census Bureau to
maintain the inventory and location of addresses and features. In
addition, we will greatly expand our ability to improve the accuracy
and completeness of our census GIS systems that process these data.
These MAF/TIGER enhancements are key to allowing the Census Bureau to
adopt the technology necessary to fully utilize GPS-equipped, hand-held
mobile computing devices to find, interview, and update data on persons
and housing units for the short-form-only census in 2010, thereby,
achieving the:
Appendix 2:
Census Bureau‘s constitutionally mandated objectives at a greatly
reduced cost.
These two components (the ACS and the MAF/TIGER enhancements) are truly
exciting and innovative in their own right. But unless they can be
translated into an improved 2010 census and done so without expanding
the cost of census taking, the goals of the reengineered census will
not be met.
The third component (early 2010 planning, development and testing) is
essential to complete the picture. The new short-form-only census,
which is the end goal of this component, is, in fact, the key component
to the success of this reengineering effort. Without it, we are left
with a census that improves relevance but at a greatly expanded cost
and with no serious improvements in operational risk or coverage
accuracy.
By taking advantage of no long-form requirements for the 2010 census
coupled with access to current data from the ACS for targeting areas
requiring special attention to improve coverage, the potential cost and
accuracy of the 2010 short-form-only census can be greatly improved.
Add to this the availability of a fully GPS-aligned MAF/TIGER system
allowing for a dramatic reduction in field infrastructure costs
resulting mostly from a near elimination of paper and the huge staff
and space required to handle that paper, we have the potential to
completely restructure the data collection, data capture, and data
processing of the 2010 census. This will result in a census that is
more focused on coverage issues, less operationally risky and less
costly.
This, however, will not happen automatically. A decennial census is a
very complex task involving hundreds of thousands of temporary staff
over a very short period of time and costing several billions of
taxpayer dollars.
To do this successfully, procedures must be fully tested under real
life conditions and refined well in advance of Census Day. You only get
one chance to get it right.
The early years of this component involve extensive planning,
development, testing, revising, and retesting of literally thousands of
procedures needed to complete a successful census. We are planning to
restructure many of these procedures to reduce costs and improve
accuracy while keeping operational risk to a minimum. To do this, we
plan a major field test in 2004, focusing primarily on improved
methodologies for data collection and coverage. In 2006, we plan a
second major field test. This time the test will be focused primarily
on the systems integration needed to carry out this new census design.
In 2008, we plan for a full dress rehearsal of the new census, setting
the stage for a 2010 census that delivers on all the goals of 2010
census reengineering.
In addition to improved accuracy and reduced operational risk, we
expect cost reductions in this component to be sufficient to pay for
all three components of the reengineered census. That is, all three
components will be carried out at a cost that is no greater and
probably somewhat less than repeating the process of Census 2000.
The following are GAO‘s comments on the letter dated October 16, 2002,
from the Department of Commerce.
GAO Comments:
1. The objectives of our report did not include assessing the degree of
success of the 2000 Census.
2. See ’Agency Comments and Our Evaluation“ section of this report.
[End of Section]
FOOTNOTES
[1] The bureau‘s use of major categories evolved from 1 activity of
general planning in fiscal year 1991, to 8 major study areas in fiscal
years 1992 and 1993, to 8 to 15 broad categories called frameworks
beginning in 1994. For internal management and reporting, the bureau
further identified program efforts by activities and projects that have
varied since fiscal year 1991.
[2] Stakeholders include the Congress, federal agencies, state and
local governments, the public, demographers, and others who rely upon
census information.
[3] Global Report of the Task Force for Planning the Year 2000 Census,
Reinventing the Decennial Census (Washington, D.C.: June 1995).
[4] U.S. General Accounting Office, Census Reform Needs Attention Now,
GAO/T-GGD-91-13 (Washington, D.C.: Mar. 12, 1991); Decennial Census:
1990 Results Show Need for Fundamental Reform, GAO/GGD-92-94
(Washington, D.C.: June 9, 1992); Decennial Census: Fundamental Reform
Jeopardized by Lack of Progress, GAO/T-GGD-93-6 (Washington, D.C.: Mar.
2, 1993); and Decennial Census: Focused Action Needed Soon to Achieve
Fundamental Breakthroughs, GAO/T-GGD-93-32 (Washington, D.C.: May 27,
1993).
[5] U.S. General Accounting Office, Decennial Census: Fundamental
Design Decisions Merit Congressional Attention, GAO/T-GGD-96-37
(Washington, D.C.: Oct. 25, 1995).
[6] U.S. General Accounting Office, The High-Risk Series, GAO/HR-97-2
(Washington, D.C.: Feb. 1997) is a special effort to review and report
on the federal program areas we have identified as high risk because of
their vulnerability to waste, fraud, abuse, or mismanagement.
[7] U.S. General Accounting Office, 2000 Census: Progress Made on
Design, but Risks Remain, GAO/GGD-97-142 (Washington, D.C.: July 14,
1997).
[8] Public Law 105-119, § 210, 111 Stat. 2483 (Nov. 26, 1997).
[9] The Census Address List Improvement Act of 1994 (Public Law 103-
430), required the U.S. Census Bureau to use the DSF and address lists
from local and tribal governments to build the Census 2000 address
list.
[10] MAF is a computer database of household addresses contained in the
database from the 1990 Census, various versions of the U.S. Postal
Service‘s DSF, and update information provided by state, local, and
tribal governments. The MAF was to be updated throughout the decade to
provide a basis for producing address labels needed to deliver 2000
Census questionnaires, to track questionnaires returned, and to
identify addresses for later site visits by enumerators.
[11] Questionnaires were available in Spanish, Chinese, Korean,
Vietnamese, and Tagalog.
[12] Cost-reimbursement contracts allow for payment of all allowable
incurred costs within a predetermined ceiling set in the contract.
Cost-reimbursement contracts place less cost and performance risk on
the contractor, as opposed to fixed-price contracts, which place more
responsibility on the contractor for performance costs and resulting
profit or loss.
[13] Title 13 U.S.C. § 9 prohibits (1) use of the information furnished
under Title 13 for any purpose other than the statistical purposes for
which it is supplied, (2) making any publication whereby the data
furnished by any particular establishment or individual under Title 13
can be identified, or (3) permitting anyone other than sworn officers
and employees of the Department of Commerce or agency thereof to
examine individual reports. Title 13 U.S.C. § 214 and 18 U.S.C. § 3551,
et seq., provides for a fine of not more than $250,000 or imprisonment
for not more than 5 years, or both, for disclosure of census
information prohibited by 13 U.S.C. § 9.
[14] U.S. Census Bureau, Prepared Statement of James F. Holmes, Acting
Director (Washington, D.C.: Mar. 26, 1998).
[15] U.S. Census Bureau, Census 2000 Dress Rehearsal Evaluation Summary
(Suitland, MD: 1999).
[16] GAO/GGD-97-142.
[17] U.S. House of Representatives, Making Appropriations for the
Departments of Commerce, Justice, and State, the Judiciary, and Related
Agencies for the Fiscal Year Ending September 30, 1993, Report 102-918
(Washington, D.C.: Sep. 28, 1992).
[18] GAO/T-GGD-93-6.
[19] U.S. Census Bureau, The Reengineered 2000 Census (Suitland, MD:
May 19, 1995).
[20] Public Law 105-18, Title VIII (June 12, 1997).
[21] The bureau distributed a revised and reissued version of this
report in August 1997.
[22] U.S. Census Bureau, Census 2000 Address List Reengineering,
(Suitland, MD: Sept. 24, 1997).
[23] For example, in fiscal year 2000, within the 8 decennial census
frameworks, the bureau identified 23 activities and 119 projects.
[24] GAO/T-GGD-91-13.
[25] $13.0 million of funding plus $.7 million for prior year
recoveries and carry-in.
[26] U.S. House of Representatives, Making Appropriations for the
Departments of Commerce, Justice, and State, the Judiciary, and Related
Agencies for the Fiscal Year Ending September 30, 1993, Report 102-918
(Washington, D.C.: Sep. 28, 1992).
[27] GAO/T-GGD-93-6.
[28] $8.1 million of funding plus $10.6 million for prior year
recoveries and carry-in.
[29] GAO/T-GGD-93-32.
[30] U.S. House of Representatives, Departments of Commerce, Justice,
and State, the Judiciary, and Related Agencies Appropriations Bill,
Fiscal Year 1994, Report 103-157 (Washington, D.C.: June 24, 1993).
[31] U.S. House of Representatives, Making Appropriations for the
Departments of Commerce, Justice, and State, the Judiciary, and Related
Agencies for the Fiscal Year Ending September 30, 1994, H.R. Conference
Report 103-293 (Washington, D.C.: Oct. 14, 1993).
[32] U.S. General Accounting Office, Decennial Census: Test Design
Proposals Are Promising, But Fundamental Reform is Still at Risk, GAO/
T-GGD-94-12 (Washington, D.C.: Oct. 7, 1993).
[33] A one-number census combines the features of both the traditional
head count and statistical methods to produce a single population
count.
[34] U.S. House of Representatives, Departments of Commerce, Justice,
and State, the Judiciary, and Related Agencies Appropriations Bill,
Fiscal Year 1994, Report 103-157 (Washington, D.C.: June 24, 1993).
[35] U.S. General Accounting Office, Decennial Census: Promising
Proposals, Some Progress, But Challenges Remain, GAO/T-GGD-94-80
(Washington, D.C.: Jan. 26, 1994).
[36] U.S. Senate, Departments of Commerce, Justice, and State, the
Judiciary, and Related Agencies Appropriation Bill, 1995, Report 103-
309 (Washington, D.C.: July 14, 1994).
[37] The bureau uses unduplication to refer to the process of detecting
duplicate census responses that contribute to a population overcount.
Causes of duplication include multiple responses for the same address,
college students counted both at school and at home, people with
multiple residences, and military residing outside their home state.
[38] $48.8 million of funding plus $2.5 million for prior year
recoveries and carry-in.
[39] GAO/T-GGD-96-37.
[40] U.S. House of Representatives, Making Appropriations for the
Departments of Commerce, Justice, and State, the Judiciary, and Related
Agencies for the Fiscal Year Ending September 30, 1996, H.R. Conference
Report 104-378 (Washington, D.C.: Dec. 1, 1995).
[41] The ICM Special Test used cognitive questionnaire testing methods
to refine the count interview instrument and to gather data from the
1995 Census Test to conduct estimation research.
[42] $84.1 million of funding plus $2.3 million for prior year
recoveries and carry-in.
[43] U.S. House of Representatives, Departments of Commerce, Justice,
and State, the Judiciary, and Related Agencies Appropriations Bill,
Fiscal Year 1997, Report 104-676 (Washington, D.C.: July 16, 1996).
GAO‘s Mission:
The General Accounting Office, the investigative arm of Congress,
exists to support Congress in meeting its constitutional
responsibilities and to help improve the performance and accountability
of the federal government for the American people. GAO examines the use
of public funds; evaluates federal programs and policies; and provides
analyses, recommendations, and other assistance to help Congress make
informed oversight, policy, and funding decisions. GAO‘s commitment to
good government is reflected in its core values of accountability,
integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony:
The fastest and easiest way to obtain copies of GAO documents at no
cost is through the Internet. GAO‘s Web site ( www.gao.gov ) contains
abstracts and full-text files of current reports and testimony and an
expanding archive of older products. The Web site features a search
engine to help you locate documents using key words and phrases. You
can print these documents in their entirety, including charts and other
graphics.
Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as ’Today‘s Reports,“ on its
Web site daily. The list contains links to the full-text document
files. To have GAO e-mail this list to you every afternoon, go to
www.gao.gov and select ’Subscribe to daily E-mail alert for newly
released products“ under the GAO Reports heading.
Order by Mail or Phone:
The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or
more copies mailed to a single address are discounted 25 percent.
Orders should be sent to:
U.S. General Accounting Office
441 G Street NW,
Room LM Washington,
D.C. 20548:
To order by Phone:
Voice: (202) 512-6000:
TDD: (202) 512-2537:
Fax: (202) 512-6061:
To Report Fraud, Waste, and Abuse in Federal Programs:
Contact:
Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov
Automated answering system: (800) 424-5454 or (202) 512-7470:
Public Affairs:
Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 U.S.
General Accounting Office, 441 G Street NW, Room 7149 Washington, D.C.
20548:
