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 productsí 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 Presidentís 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 Presidentís 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 bureauís 

planning efforts.



The bureauís 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 

bureauís 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 bureauís 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 bureauís 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 bureauís 

plans by July 12, 1997, the bureau issued its Report to Congress, The 

Plan For Census 2000. This plan also incorporated the bureauís 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 bureauís 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 

departmentís 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 Presidentís Budget. The Congress may then adjust the Presidentís 

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 

Presidentís 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 Presidentís 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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 (Presidentís 

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 bureauís 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 bureauís 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 bureauís 

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 bureauís 

database contained 102 million addresses, each assigned to the census 

block area in which it was located. At that point, the U.S. Census 

Bureauís 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 bureauís 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 bureauís 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 bureauís 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 

bureauís 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 GAOís 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 Presidentís 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 bureauís 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 ďAgency Comments and Our EvaluationĒ 

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 bureauís 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 departmentís requested amounts 

to $1.5 million for fiscal year 1991 and $10.1 million for fiscal year 

1992. These amounts were included in the Presidentís Budgets and the 

Congress concurred by authorizing the full amount requested. Census 

planning officials said that if OMB had not augmented the departmentís 

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 Presidentís 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 bureauís 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 bureauís 

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 Presidentís 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 Bureauís 

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 

bureauís 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 personís 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 Presidentís 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 Bureauís 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 Presidentís 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 Presidentís 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 

Departmentís 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 ďthorough 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.Ē:



We believe that our plan for reengineering the decennial census design 

for 2010 directly addresses these issues (see Appendix 1, 

ďReengincering the 2010 CensusĒ). 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, ďLife-Cycle 

Savings of the 2010 Census,Ē which is currently under revision.):



The GAO recommends that the Census Bureau ďprovide comprehensive 

information backed by supporting documentation in its future funding 

requests for planning and development activities.Ē 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 ďCapital Asset Plan and 

Business CaseĒ (Exhibit 300), which outlines the Census Bureauís 

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 Bureauís 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 GAOís 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 ďAgency Comments and Our EvaluationĒ section of this report.



[End of Section]



FOOTNOTES



[1] The bureauís 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, ß 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 

Serviceís 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. ß 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. ß 214 and 18 U.S.C. ß 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. ß 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).



GAOís 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. GAOís 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. GAOís 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 ďTodayís Reports,Ē 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 ďSubscribe to daily E-mail alert for newly 

released productsĒ 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: