2010 Census
Costs and Risks Must be Closely Monitored and Evaluated with Mitigation Plans in Place Gao ID: GAO-06-822T June 6, 2006The decennial census is a constitutionally mandated activity, with immutable deadlines. It produces data used to allocate about $200 billion yearly in federal financial assistance, reapportion the seats of the House of Representatives, and provide a profile of the nation's people to help guide policy decisions. The U.S. Census Bureau (Bureau) estimates the 2010 Census will cost $11.3 billion, making it the most expensive census in the nation's history, even after adjusting for inflation. Based primarily on GAO's issued reports, this testimony addresses the extent to which the Bureau has (1) developed detailed and timely cost data for effective oversight and cost control, (2) reduced nonresponse mail follow up costs, and (3) produced risk mitigation plans to address identified challenges.
The Bureau's most recent life-cycle cost estimate for the 2010 Census does not reflect the most current information from testing and evaluation nor provide complete information on how changing assumptions may affect cost. As GAO reported in January 2004, the Bureau derived its initial cost estimate by considering the cost of the 2000 Census along with certain assumptions that drive costs, such as staffing needs, the nonresponse rate for mailing back the census questionnaire, census worker productivity and pay rates, and inflation; however, GAO's ongoing work has found that the most recent (September 2005) estimate does not incorporate current information on certain 2001 assumptions. For example, the 2004 Census Test suggests some assumptions about staffing and space associated with new technology have changed. Specifically, Bureau evaluations indicate that more staff at the local census office was needed to support the use of the new hand-held mobile computing device (MCD) and additional storage space was needed for the MCDs. Since 2000, the Bureau has reengineered the decennial census and has begun new initiatives to reduce nonresponse follow up costs. Key to the Bureau's steps to reduce the costs of nonresponse follow up is successfully using the MCDs to eliminate millions of paper questionnaires and maps. Importantly, the Bureau must first resolve the MCD's technological challenges. During 2004 and 2006 tests, the MCDs had significant reliability problems. For example, in the 2004 test the MCDs experienced transmission problems, memory overloads, and difficulties with the map ping feature. Bureau officials have contracted the design and implementation for a new MCD that will not be ready until the 2008 Dress Rehearsal. If after the Dress Rehearsal the MCD is found not to be reliable, the Bureau could be faced with the remote but daunting possibility of having to revert to the costly paper-based Census used in 2000. The Bureau does not have risk mitigation plans to address certain identified challenges to a cost-effective census. Most notably, the Bureau does not have a plan to assess additional resources that may be needed to update the address and map file for areas affected by hurricanes Katrina and Rita. Moreover, the Bureau has not yet assessed whether new procedures will be necessary nor whether local partners will be available to assist in updating address and map data. Updating address files to reflect the changes caused by the hurricanes will be formidable, in part because, according to Red Cross estimates, nearly 525,000 people were displaced in a 90,000 square mile area. Another risk to be mitigated stems from the need to closely monitor the performance of about $1.9 billion in contracts. The Bureau has agreed to take steps to mitigate some of those risks. For example, the Bureau has said it will enhance the ability of key contract project offices to better manage contracts through such actions as developing mitigation plans with milestones for key activities and regularly briefing senior managers.
GAO-06-822T, 2010 Census: Costs and Risks Must be Closely Monitored and Evaluated with Mitigation Plans in Place
This is the accessible text file for GAO report number GAO-06-822T
entitled '2010 Census: Costs and Risks Must be Closely Monitored and
Evaluated with Mitigation Plans in Place' which was released on June 6,
2006.
This text file was formatted by the U.S. Government Accountability
Office (GAO) to be accessible to users with visual impairments, as part
of a longer term project to improve GAO products' accessibility. Every
attempt has been made to maintain the structural and data integrity of
the original printed product. Accessibility features, such as text
descriptions of tables, consecutively numbered footnotes placed at the
end of the file, and the text of agency comment letters, are provided
but may not exactly duplicate the presentation or format of the printed
version. The portable document format (PDF) file is an exact electronic
replica of the printed version. We welcome your feedback. Please E-mail
your comments regarding the contents or accessibility features of this
document to Webmaster@gao.gov.
This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed
in its entirety without further permission from GAO. Because this work
may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this
material separately.
Testimony:
Before the Subcommittee on Federal Financial Management, Government
Information, and International Security, Committee on Homeland Security
and Governmental Affairs, U.S. Senate:
United States Government Accountability Office:
GAO:
For Release on Delivery Expected at 2:30 p.m. EST:
Tuesday, June 6, 2006:
2010 Census:
Costs and Risks Must be Closely Monitored and Evaluated with Mitigation
Plans in Place:
Statement of Brenda S. Farrell Acting Director, Strategic Issues:
GAO-06-822T:
GAO Highlights:
Highlights of GAO-06-822T, a testimony before the Subcommittee on
Federal Financial Management, Government Information and International
Security, Committee on Homeland Security and Governmental Affairs, U.S.
Senate.
Why GAO Did This Study:
The decennial census is a constitutionally mandated activity, with
immutable deadlines. It produces data used to allocate about $200
billion yearly in federal financial assistance, reapportion the seats
of the House of Representatives, and provide a profile of the nation‘s
people to help guide policy decisions. The U.S. Census Bureau (Bureau)
estimates the 2010 Census will cost $11.3 billion, making it the most
expensive census in the nation‘s history, even after adjusting for
inflation. Based primarily on GAO‘s issued reports, this testimony
addresses the extent to which the Bureau has (1) developed detailed and
timely cost data for effective oversight and cost control, (2) reduced
nonresponse mail follow up costs, and (3) produced risk mitigation
plans to address identified challenges.
What GAO Found:
The Bureau‘s most recent life-cycle cost estimate for the 2010 Census
does not reflect the most current information from testing and
evaluation nor provide complete information on how changing assumptions
may affect cost. As GAO reported in January 2004, the Bureau derived
its initial cost estimate by considering the cost of the 2000 Census
along with certain assumptions that drive costs, such as staffing
needs, the nonresponse rate for mailing back the census questionnaire,
census worker productivity and pay rates, and inflation; however, GAO‘s
ongoing work has found that the most recent (September 2005) estimate
does not incorporate current information on certain 2001 assumptions.
For example, the 2004 Census Test suggests some assumptions about
staffing and space associated with new technology have changed.
Specifically, Bureau evaluations indicate that more staff at the local
census office was needed to support the use of the new hand-held mobile
computing device (MCD) and additional storage space was needed for the
MCDs.
Since 2000, the Bureau has reengineered the decennial census and has
begun new initiatives to reduce nonresponse follow up costs. Key to the
Bureau‘s steps to reduce the costs of nonresponse follow up is
successfully using the MCDs to eliminate millions of paper
questionnaires and maps. Importantly, the Bureau must first resolve the
MCD‘s technological challenges. During 2004 and 2006 tests, the MCDs
had significant reliability problems. For example, in the 2004 test the
MCDs experienced transmission problems, memory overloads, and
difficulties with the mapping feature. Bureau officials have contracted
the design and implementation for a new MCD that will not be ready
until the 2008 Dress Rehearsal. If after the Dress Rehearsal the MCD is
found not to be reliable, the Bureau could be faced with the remote but
daunting possibility of having to revert to the costly paper-based
Census used in 2000.
The Bureau does not have risk mitigation plans to address certain
identified challenges to a cost-effective census. Most notably, the
Bureau does not have a plan to assess additional resources that may be
needed to update the address and map file for areas affected by
hurricanes Katrina and Rita. Moreover, the Bureau has not yet assessed
whether new procedures will be necessary nor whether local partners
will be available to assist in updating address and map data. Updating
address files to reflect the changes caused by the hurricanes will be
formidable, in part because, according to Red Cross estimates, nearly
525,000 people were displaced in a 90,000 square mile area. Another
risk to be mitigated stems from the need to closely monitor the
performance of about $1.9 billion in contracts. The Bureau has agreed
to take steps to mitigate some of those risks. For example, the Bureau
has said it will enhance the ability of key contract project offices to
better manage contracts through such actions as developing mitigation
plans with milestones for key activities and regularly briefing senior
managers.
What GAO Recommends:
The Bureau is taking action on several of GAO‘s recommendations to
reduce nonresponse time and mitigate contract-related risks. A January
2004 report contained recommendations to the Bureau for improving the
transparency of the 2010 Census‘ life-cycle costs. While the Bureau did
not agree with this recommendation, the Bureau stated that in response
it would develop a comprehensive project plan that would include
milestones, itemized estimated costs, and measurable goals.
[Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-822T].
To view the full product, including the scope and methodology, click on
the link above. For more information, contact Brenda S. Farrell at
(202) 512-6806 or farrellb@gao.gov.
[End of Section]
Mr. Chairman, Mr. Carper, and Members of the Subcommittee:
Thank you for the opportunity to be here today to discuss the life-
cycle costs of the 2010 Census as well as the actions that the U.S.
Census Bureau (Bureau) is taking to contain those costs. The Bureau
estimates the 2010 Census will cost $11.3 billion, which would make it
the most expensive census in our country's history, even after
adjusting for inflation. Since the 2000 Census, we have monitored how
the Bureau has incorporated lessons learned from the 2000 Census into
its planning for the next decennial census, as well as its cost and
design. My overall point today is that the decennial's cost and risks
must be closely monitored and evaluated, with mitigation plans in place
to help ensure that accurate results are delivered on time and within
projected costs. Based primarily on our issued reports, this testimony
addresses the extent to which the Bureau has (1) developed detailed and
timely cost data for effective oversight and cost control, (2) reduced
nonresponse mail follow-up costs, and (3) produced risk mitigation
plans to address identified challenges, such as assessing the resources
that may be needed to update address files and maps in areas affected
by hurricanes Katrina and Rita. I will also present the preliminary
results of ongoing work--on which we plan to issue a report later this
month--on the Bureau's efforts to build a complete and accurate address
list, the foundation of a successful census.
As you know, Mr. Chairman, the decennial census is a crucial,
constitutionally mandated activity undertaken by the Bureau. The stakes
for a successful census are very high. The data that the census
produces are used to reapportion the seats of the U.S. House of
Representatives; realign the boundaries of the legislative districts of
each state; allocate about $200 billion dollars each year in federal
financial assistance; and provide a social, demographic, and economic
profile of the nation's people to guide policy decisions at each level
of government. Further, businesses use census data to target new
services and products and to tailor existing ones to demographic
changes.
Mr. Chairman, I would like to commend the subcommittee for calling
today's hearing, as past experience has shown that strong and
continuing congressional involvement--especially while there is still
time to make cost-effective decisions and influence the direction of
the decennial census--is essential to the decennial's ultimate success.
Today's hearing is particularly timely because the Bureau is currently
holding the 2006 Census Test in the central portion of Travis County,
Texas, and at the Cheyenne River American Indian Reservation and Tribal
Trust Lands in South Dakota, where the Bureau is evaluating key
operations and equipment it plans to employ for the full enumeration in
2010. After this test, the Bureau will have only one more opportunity
to assess its census-taking-procedures--a "Dress Rehearsal" scheduled
for 2008. Moreover, after the Dress Rehearsal, the Bureau will begin to
transition from preparatory to operational activities, leaving little
room for delays or design changes, which at that point could
significantly increase the cost of 2010 Census.
Importantly, for decades we have been reviewing the national
enumeration on behalf of Congress. Over the years, through a series of
reports and testimonies, we have acquired broad institutional knowledge
that gives us a historical view of the census. I want to highlight
several broad themes that have emerged from our work.
First, completing the decennial census is a monumental undertaking, and
the Bureau recognizes that streamlined and efficient operations are
critical for the census' cost-effectiveness. The Census' sheer size and
complexity make it a risky and fragile enterprise. The 2000 Census, for
example, involved the hiring of more than 500,000 enumerators on a
temporary basis, opening 511 local census offices nationwide and 24,000
questionnaire assistance centers, processing 1.5 billion sheets of
paper, and in 10 weeks following up with 42 million nonrespondent
households. The size of the census means that small problems can
magnify quickly, and big problems could be overwhelming. For example,
60 seconds might seem like an inconsequential amount of time, but in
2000, if enumerators had spent just 1 minute more at each household
during nonresponse follow-up, it could have added almost $10 million to
the cost of the census, assuming a pay rate of around $13 per hour
(wages ranged from $8.25 to $18.50 per hour for enumerators in 2000,
depending on location).
Second, sound risk management is important to a successful census
because many risks are interrelated, and a shortcoming in one operation
could cause other operations to spiral downward. For example, a low
mail response rate would drive up the follow-up workload, which in turn
would increase staffing needs and costs. (Of course, the reverse is
also true, where a success in one operation could have a number of
positive downstream impacts.) Rigorous up-front planning and testing,
and where needed, risk mitigation plans are the best ways to stave off
these problems. In the 2000 Census, the Bureau successfully planned and
mitigated risk in recruiting and hiring workers by using management
information systems capable of tracking key operations with real-time
measures. To recruit the vast army of people needed to fill the ranks
of its workforce for the 2000 Census, the Bureau set a recruitment goal
of 2.4 million qualified applicants. Because the Bureau tracked the
progress local census offices were making in meeting their individual
goals, it was able to mitigate risk by quickly raising pay rates and
taking other actions at those offices where recruitment was lagging. In
the end, the Bureau exceeded its recruitment goal by 100,000 people.
Third, the census is conducted against a backdrop of immutable
deadlines, and the census' elaborate chain of interrelated pre-and post-
Census Day activities are predicated upon those dates. The Secretary of
Commerce is legally required to (1) conduct the census on April 1 of
the decennial year, (2) report the state population counts to the
President for purposes of congressional apportionment by December 31 of
the decennial year, and (3) send population tabulations to the states
for purposes of redistricting no later than 1 year after the April 1
census date. To meet these legally mandated reporting requirements,
census activities need to take place at specific times and in the
proper sequence. Bureau officials have recently stated, and we agree,
that the design and plans being implemented are too far down the road
and time is too short to allow for significant adjustments. In fact, as
Census Day approaches, the tolerance for any operational delays becomes
increasingly small. Indeed, considerable risk and cost increases could
accompany design changes that occur late in the decade. This requires
the Bureau to have risk-based mitigation plans in place now to ensure
that 2010 Census operations are ready and that few, if any, changes to
the fundamental design happen after the 2008 Dress Rehearsal.
Based on the Bureau's desire to address the issues associated with the
2000 enumeration, in designing the 2010 Census the Bureau had four
goals in mind: (1) increase the relevance and timeliness of data, (2)
reduce operational risk, (3) increase coverage and accuracy, and (4)
contain costs. To achieve these goals, three components--all new
operations--are important to the Bureau's plans for 2010:
* enhancing procedures for its address list (the MAF--Master Address
File) and the associated geographic information system (the TIGER®--
Topologically Integrated Geographic Encoding and Referencing
database[Footnote 1]),
* replacing the census long-form questionnaire with the American
Community Survey (ACS)[Footnote 2], and:
* conducting a short-form only decennial census that is supported by
early research and testing.
My remarks today are based primarily on reports that GAO issued from
2002 through May 2006 on the planning and development of the 2010
Census. These reports are listed in appendix I. We analyzed Bureau
documents and data and interviewed key Bureau officials regarding the
2004 and 2006 Census Tests. In that regard, we visited the Texas and
South Dakota test sites; Queens, New York; and several counties in
rural south-central Georgia, where an earlier field test was held in
2004. During these visits we observed the address canvassing operation-
-where workers go door to door verifying addresses and updating maps as
part of the Bureau's effort to build a complete and accurate address
list, and we observed the nonresponse follow-up operation--where
enumerators collect information from those households that do not
return their initial questionnaire. We conducted our work in accordance
with generally accepted government auditing standards.
The Bureau's $11.3 Billion Cost Estimate for 2010 Census Lacks Timely
and Complete Data:
The Bureau's $11.3 billion life-cycle cost estimate for the 2010 Census
lacks timely and complete supporting data. The supporting data of the
estimate is not timely because it does not contain the most current
information from testing and evaluation. Also, the supporting data of
the estimate is not complete because it does not provide sufficient
information on the how changing assumptions could affect cost.
Cost for Each Decennial Census Continues to Significantly Increase:
In January 2004, we reported that the Bureau's cost projections for the
2010 decennial census continue an escalating trend.[Footnote 3] As
noted above, the Bureau now estimates the 2010 Census will cost $11.3
billion, making it the most expensive in history, even after adjusting
for inflation. Although some cost growth can be expected, in part
because the number of housing units--and hence the Bureau's workload--
has become larger, the cost growth has far exceeded the increase in the
number of housing units. The Bureau estimates that the number of
housing units for the 2010 Census will increase by 10 percent over 2000
Census levels. At the same time, as shown in figure 1, the average cost
per housing unit for 2010 is expected to increase by approximately 29
percent from 2000 levels (from $56 per housing unit to $72 per housing
unit in 2000 inflation-adjusted dollars).[Footnote 4]
Figure 1: Decennial Census Average Cost per Housing Unit (Fiscal Year
2000 Inflation-Adjusted Dollars):
[See PDF for image]
[End of figure]
The risk exists that the actual, final cost of the census could be
considerably higher. Indeed, the Bureau's initial cost projections for
previous censuses proved to be too low because of such factors as
unforeseen operational problems or changes to the fundamental design.
For example, the Bureau estimated that the 2000 Census would cost
around $4 billion if sampling was used, and a traditional census
without sampling would cost around $5 billion. However, the final price
tag for the 2000 Census (without sampling) was over $6.5 billion, a 30
percent increase in cost. Today's climate of large federal deficits and
other fiscal challenges requires holding the decennial's costs as low
as possible, while promoting an accurate, timely census.
2010 Cost Estimate Lacks Timely and Complete Information:
Despite a history of cost increases, the Bureau's most recent cost
estimate is not based on timely and complete information. Table 1 shows
the Bureau's latest revised estimate that was released in September
2005. Based on this table, the bulk of the funds will be spent between
fiscal years 2007 through 2013.
Table 1: Bureau's Revised September 2005 Estimate of Life-cycle Costs
for the 2010 Decennial Census Program (in millions of dollars,
nominal):
Program component: American Community Survey;
FY 2001: $23.6;
FY 2002: $29.0;
FY 2003: $56.8;
FY 2004: $64.1;
FY 2005: $144.1;
FY 2006: budget request: $169.9;
Subtotal: FY01-06: $487.5;
FY 2007-: FY 2013: (est.) $1,219.8;
Total: (est.) $1,707.3.
Program component: MAF/TIGER Enhancements Program;
FY 2001: $0;
FY 2002: $15.0;
FY 2003: $47.0;
FY 2004: $82.4;
FY 2005: $81.2;
FY 2006: budget request: $79.8;
Subtotal: FY01-06: $305.4;
FY 2007-: FY 2013: (est.) $228.9;
Total: (est.) $534.3.
Program component: Short Form 2010 Census;
FY 2001: $0;
FY 2002: $21.0;
FY 2003: $41.6;
FY 2004: $106.0;
FY 2005: $163.0;
FY 2006: budget request: $214.5;
Subtotal: FY01-06: $546.1;
FY 2007-: FY 2013: (est.) $8,466.8;
Total: (est.) $9,012.9.
Program component: Total;
FY 2001: $23.6;
FY 2002: $65.0;
FY 2003: $145.4;
FY 2004: $252.5;
FY 2005: $388.3;
FY 2006: budget request: $464.3;
Subtotal: FY01-06: 1,339.0;
FY 2007-: FY 2013: (est.) $9,915.5;
Total: (est.) $11,254.6.
Source: U.S. Census Bureau.
Note: These figures have not been audited by GAO.
[End of table]
As we stated in our January 2004 report[Footnote 5], in June 2001, the
Bureau derived its 2010 cost estimate by using the actual cost of the
2000 Census combined with assumptions about cost drivers, such as (1)
staffing needs, (2) enumerator productivity, (3) pay rates for census
workers, (4) the nonresponse rate for mailing back the
questionnaires[Footnote 6], and (5) inflation. However, the most recent
life-cycle cost estimate[Footnote 7] does not incorporate current
information about those 2001 assumptions. One key assumption, that has
not been updated pertains to the use of a new technology--specifically,
new hand-held, GPS-enabled mobile computing devices (MCDs)--that would
be important to the success of the 2010 census by automating and
streamlining address canvassing, nonresponse follow-up, coverage
measurement, and payroll operations. The Bureau anticipated that the
use of MCDs would facilitate reductions in administrative and support
costs in the Bureau's field offices, including a 50 percent reduction
in clerical and administrative local census office staff costs and a 50
percent reduction in space at each local census office. However, the
Bureau's existing assumptions about the use and reliability of the MCD
were not updated to reflect information from the 2004 test, which
showed that assumptions about staffing and space associated with the
new technology had changed since the June 2001 estimate. The Bureau's
evaluations about those test results indicate that more help desk staff
at the local census office were needed to support the use of the MCD,
and additional storage space was needed for the devices. However, the
Bureau did not use this information when revising its cost estimate in
2005 because, according to Bureau officials, they conduct field tests
for operational purposes only--not to inform the cost estimates. In our
view, revising cost estimates on the most recent information--including
test results that are pertinent to cost assumptions--can assist the
Bureau and external decision makers to oversee costs and make necessary
resource allocations to help ensure a successful, cost-effective,
census.
The Bureau's cost estimate lacked complete information, such as a
sensitivity analysis regarding assumptions that could affect cost
drivers. OMB Circular A-94 provides guidelines for cost-benefit
analysis of federal programs and recommends that agencies develop a
sensitivity analysis for major projects with significant uncertainty,
like the decennial census. The circular provides a method for
determining how sensitive outcomes are to changes in assumptions. In
January 2004, we reported that the Bureau could provide more robust
information on the likelihood that the values the Bureau assigned to
key cost drivers could differ from those initially assumed and be
timelier--previously the life-cycle cost estimate had been provided at
2-year intervals.[Footnote 8] The Bureau's latest life-cycle cost
document does not contain a sensitivity analysis on assumptions that
impact cost; it did, however, indicate that the life-cycle cost would
be updated annually.
Having transparent information about cost estimates is especially
important because decennial costs are sensitive to many key
assumptions. In fact, for the 2000 Census, the Bureau's supplemental
funding request for $1.7 billion in fiscal year 2000 primarily involved
changes in assumptions related to increased workload, reduced employee
productivity, and increased advertising. Given the cost of the census
in an era of serious national fiscal challenges, it would be beneficial
for the Bureau and Congress to have sensitivity information about the
likelihood--high, medium, or low--that certain assumptions would drive
costs. By providing this information, the Bureau would better enable
Congress to consider funding levels in this uncertain environment.
Our January 2004 report also highlighted the challenge that the Bureau
would have in containing the cost of the 2010 Census. To increase the
transparency of the census' life-cycle costs for Congress, we
recommended that Office of Management and Budget (OMB) establish
triggers that would signal when the annual 2010 Census costs and/or
life-cycle 2010 Census costs exceeded some predetermined amount. We
also recommended, among other things, that OMB ensure the Bureau
analyzes the sensitivity of the cost figures to specific assumptions.
However, OMB disagreed with our recommendation, because it said it
already has internal procedures within its budget reviews to monitor
2010 Census costs. OMB shared our view that the costs and risks
associated with the 2010 Census must be carefully monitored and
evaluated throughout the decade. OMB also agreed that it is essential
to understand the key cost drivers and said that it is working with the
Bureau to ensure that the Bureau develops high-quality, transparent
life-cycle cost estimates.
In addition, we recommended in our 2004 report that the Bureau develop
a comprehensive project plan that would be updated as needed to (1)
include milestones for completing key activities; (2) itemize the
estimated cost of each component; (3) articulate a clear system of
coordination among project components; and (4) translate key goals into
measurable, operational terms to provide meaningful guidance for
planning and measuring progress. Some, but not all, of this information
is available in various documents, and to be useful, it would need to
be pieced together. As a result, we recommended that the Bureau combine
this information into a single, comprehensive document. The Bureau
disagreed with our recommendation, although it said it would develop
such a plan nonetheless and provide it to GAO, Congress, and other
stakeholders. The Bureau has not yet issued such a document.
Bureau Has Taken Steps to Reduce Nonresponse Follow-up Costs, But
Challenges with Technology Remain:
Since 2000, the Bureau has reengineered the decennial census and has
begun to implement new initiatives. These include (1) using a short-
form-only census questionnaire; (2) automating field operations; and
(3) using a targeted second mailing to households that fail to respond
to the initial census questionnaire, instead of sending an enumerator
to visit houses that have not responded. These initiatives could reduce
the workload and cost of nonresponse follow-up. While these initiatives
show promise, the Bureau will need to address technological challenges
with the MCD that will be used to collect data for nonresponse follow-
up.
The Bureau is finding it increasingly difficult to locate people and
get them counted in the census. As in previous censuses, the major cost
for the 2010 Census is what the Bureau calls "field data collection and
support systems," accounting for over half of the life-cycle costs of
the decennial census.
First, the Bureau plans to contain the cost of nonresponse follow-up by
increasing mail response through a short-form-only census. The overall
mail response rate has been declining steadily since 1970. In the 1980
Census, the mail response rate was 75 percent, 3 percentage points
lower than it was in the 1970 Census. In the 1990 census, the mail
response rate dropped to 65 percent and, in 2000, appeared to be
leveling off at about 64 percent. Contributing to this decline is the
public's unwillingness to complete the long form. Specifically, the
response rates in 1990 and 2000 to the short form have been higher than
the response rate to the long form. Bureau data suggest a 1 percent
increase in the mail response rate would result from conducting a short-
form-only census.
Secondly, by using the MCD, the Bureau plans to automate field data
collection to contain the cost of nonresponse follow-up. The MCD allows
the Bureau to automate operations and eliminate the need to print
millions of paper questionnaires and maps used by census workers to
conduct address canvassing and nonresponse follow-up, as well as
managing field staff's payroll. As stated above, the benefits of using
the MCD have been tested in the 2004 and 2006 tests. For example,
during the 2004 Census Test, the MCD allowed the Bureau to successfully
remove over 7,000 late mail returns from enumerators' assignments,
reducing the total nonresponse follow-up workload by nearly 6 percent.
The ability to remove late mail returns from the Bureau's nonresponse
follow-up workload reduces costs, because census workers no longer need
to make expensive follow-up visits to households that return their
questionnaire late, after the mail-back deadline. If the Bureau had
possessed this capability during the 2000 Census, it could have
eliminated the need to visit nearly 773,000 late-responding households
and saved an estimated $22 million (based on our estimate that a 1
percentage point increase in workload could add at least $34 million in
direct salary, benefits, and travel costs to the price tag of
nonresponse follow-up[Footnote 9]). Moreover, operations that
traditionally had to be done in sequence, such as nonresponse follow-up
and then verifying the housing unit status for addresses marked vacant,
can now be performed simultaneously by using the MCD, which may shorten
the time needed for local census offices to stay open.
However, the Bureau's ability to collect and transmit data using the
MCD is not known and, at this point, constitutes a risk to the cost-
effective implementation of the 2010 Census. During the 2004 test of
nonresponse follow-up and the 2006 test of address canvassing, the MCDs
experienced significant reliability problems.
During the 2004 Census Test, the MCDs experienced transmission
problems, memory overloads, and difficulties with a mapping feature--
all of which added inefficiencies to the nonresponse follow-up
operation.[Footnote 10] During the 2006 Census Test, for address
canvassing, the device was slow to pull up and exit address registers,
accept the data entered by the census workers, and link map locations
to addresses for multiunit structures. Furthermore, the MCDs would
sometimes lockup, requiring workers to reboot them.
Census workers also found it difficult to transmit an address and map
location that were identified for deletion. Because the Bureau could
not fix this problem, workers returned to the local census office so
technicians could address the problem. The MCD's global positioning
system (GPS) receiver, a satellite-based navigational system to help
workers locate street addresses and collect coordinates for each
structure in their assignment area, was also unreliable. Some workers
had trouble receiving signals, and when a signal was available, the
receiver was slow to find assignment areas and correct map locations,
according to Bureau officials. The Bureau extended the operation 10
days and still was unable to complete the job, leaving census blocks in
Austin, Texas and on the Cheyenne River Reservation, South Dakota,
unverified.
The Bureau has acknowledged that the MCD's performance is an issue but
believes it will be addressed through a contract awarded on March 30,
2006, to develop a new MCD. However, the new MCD will not be tested
until the 2008 Dress Rehearsal, and if problems do emerge, little time
will be left to develop, test, and incorporate refinements. Given that,
it will be important that the Bureau have a risk mitigation plan in
place to help ensure the successful testing of the MCD at the Dress
Rehearsal. In our May 2006 report, we highlighted the tight time frames
to develop the MCD and recommended that systems being developed or
provided by contractors for the 2010 Census--including the MCD--be
fully functional and ready to be assessed as part of the 2008 Dress
Rehearsal.[Footnote 11] The Department of Commerce, the Census Bureau's
parent agency, noted in its comments on our draft report that the
Bureau provided competitors for the contract with information about the
design, requirements, and specification for the 2006 test in the
request for proposals. Commerce also noted that the Bureau would share
preliminary results from the 2006 test with the firm that was awarded
the contract, upon the availability of those results. The Bureau,
however, did not specify when preliminary results would be available.
However, if after the 2008 Dress Rehearsal the MCD is found not to be
reliable, the Bureau could be faced with a remote but daunting
possibility of having to revert to the costly, paper-based census used
in 2000.
Finally, a targeted second mailing to households that fail to respond
to the initial census questionnaire could reduce the workload and cost
of nonresponse follow-up. According to Bureau studies, sending a second
questionnaire could yield a gain in overall response of 7 to 10 percent
from non-responding households. In reports, we have highlighted how a
second mailing could boost the mail response rate by several percentage
points, which in turn would result in considerable savings by reducing
the number of costly personal visits enumerators would need to make to
non-responding households. The Bureau has never before included this
operation as part of a decennial census and over the decade has been
testing its feasibility. The targeted second mailing is a part of the
2006 test, the results of which will allow the Bureau to identify and
resolve any operational issues; to demonstrate a more refined plan as
part of the 2008 Dress Rehearsal; and, ultimately, to increase the
likelihood that the second mailing will produce the desired cost
savings and other benefits in 2010.
Bureau Lacks Risk Mitigation Plans for Certain Challenges:
Recent work that we have conducted has identified several challenges
that, if not properly managed, could increase the cost of the 2010
Census. As the Bureau moves from testing to demonstrating the design in
the Dress Rehearsal, it will be important for the Bureau to have risk
mitigation plans in place to reduce the severity of challenges to a
cost-effective census. These challenges include (1) overseeing
contractors responsible for conducting key census-taking operations,
(2) successfully updating address and map files, and (3) assessing the
resources that will be needed to update the address files and maps for
areas affected by hurricanes Katrina and Rita.
Increased Reliance on Contractor Support for the 2010 Census Introduces
Risk:
The Bureau is relying extensively on contractors to supply mission-
critical functions and technologies for the 2010 Census. The Bureau
estimates that they will spend $1.9 billion, or nearly 17 percent, of
the Bureau's overall decennial costs to award seven major contracts for
the 2010 Census. To date, the Bureau has awarded three of its seven
major contracts. These three contracts support (1) MAF/TIGER
modernization; (2) the development and operation of the Decennial
Response and Integration System (DRIS)--a system planned to integrate
paper, Internet, and telephone responses; and (3) the Field Data
Collection Automation (FDCA) program--a system designed to provide
field staff with the equipment and infrastructure needed to collect
census data.
Contractors can help the Bureau address the challenges it faces as it
plans for and implements the 2010 Census, especially as it becomes
increasingly difficult for the Bureau to count the nation's population
with its in-house staff and capabilities. The contractors that the
Bureau relied on to perform major decennial activities during Census
2000 generally performed well.[Footnote 12] However, increased reliance
on contractors entails certain management challenges, including the
oversight of contractors to ensure that they meet the Bureau's needs in
an effective, economical, and timely manner. For example, according to
the Department of Commerce Office of Inspector General, the Bureau did
not have sufficient program management staff to efficiently acquire
systems and manage complex, high-dollar contracts during Census
2000.[Footnote 13] As a result, the cost of the Bureau's data capture
system increased from $49 million to $238 million by the end of that
decennial.
As we noted in our May 2006 report, the Bureau has not yet awarded four
other major contracts for the 2010 Census, but has already pushed back
the award dates of two of the remaining contracts because of changes in
its acquisition approach. The Bureau's tight schedule for systems
development and testing as well as the interdependence of decennial
systems could affect its ability to develop fully functional and
sufficiently mature systems that can be demonstrated in concert with
other operations during the 2008 Dress Rehearsal. We previously
reported that during the 1998 Dress Rehearsal for the 2000 Census, a
number of new features were not test-ready; as a result, the Bureau
said it could not fully evaluate them with any degree of assurance as
to how they would affect the census.[Footnote 14] These late design
changes and untested systems resulted in additional costs to the
census.
Closely monitoring major contracts continues to be important. In March
2006, we testified that while project offices responsible for the DRIS
and FDCA contracts had carried out initial acquisition management
activities, neither office had the full skill sets needed to
effectively manage the acquisitions.[Footnote 15] For DRIS, the
Bureau's project office had established baseline requirements, but the
Bureau had not validated the requirements and had not implemented a
process for managing them. Also, the project office had identified the
project's risks but had not written mitigation plans or established
milestones for completing key risk mitigation activities. As for FDCA,
the Bureau again had specified baseline requirements but had not
validated them. While, the project office had begun to oversee the
contractor's performance, it had not determined which performance
measures it would use, and the office had not implemented a risk
management process. Until these basic management activities are
implemented, both systems could face increased risks of cost overruns,
schedule delays and performance shortfalls. We have made
recommendations addressing those issues, such as developing mitigation
plans with milestones for key activities and regularly briefing senior
managers. The Bureau has agreed to complete these activities as soon as
possible.
As part of its effort to allow respondents to use the Internet during
the decennial census, the Bureau proposed to develop the use of the
Internet under the DRIS contract. However, in May 2006, Bureau
officials informed us that the Internet response option was no longer a
contract requirement and that they are uncertain whether Internet
response would be an option for the 2010 Census. The removal of the
Internet from the DRIS contract is an unexpected change, because just 3
months earlier in our March 2006 testimony,[Footnote 16] we reported
that the DRIS contract was expected to process Internet responses for
the 2010 Census.
High-level Bureau officials explained that they made the decision to
remove the Internet from the contract partly because of the potential
risks associated with computer security attacks. In addition, according
to a Bureau official, the Bureau's testing to date showed nothing to
indicate that offering an Internet response option would improve
overall response rates or save any money. According to Bureau
officials, if the Internet response option is included in the design,
it will be developed in-house by Bureau staff. Bureau officials
emphasized that they only have one chance every 10 years to collect
this information; moreover, any public perception of an unsecured
Internet Web site could result in residents not responding to the
census, and in the long term could cost more than if the Internet had
not been used. It should be noted that there are security techniques to
address Internet attacks, and other federal agencies use the Internet
to successfully meet many missions. According to a Bureau official, the
Bureau believes it made a sound business decision by removing the
Internet from the DRIS contract requirements. Further, the official
told us that the Bureau did not develop a formal business case document
on this decision.
Address and Mapping Challenges Pose a Risk to a Cost-Effective Census:
To contain decennial costs, long-standing and emerging issues related
to the Bureau's address lists and maps need to be addressed. A complete
and accurate address list is the cornerstone of a successful census
because it identifies all households that are to receive a census
questionnaire and serves as the control mechanism for following up with
households that fail to respond. Although the Bureau went to great
lengths to build a complete and accurate MAF for the 2000 Census, of
the 116 million housing units contained in the database, the Bureau
estimates it incorrectly included 2.3 million housing units and missed
another 2.7 million housing units. In light of these and other
problems, the Bureau concluded that enhancements to MAF/TIGER were
necessary to make census data more complete and accurate.
The Bureau has conducted research and testing to help resolve each of
the problems experienced in the 2000 Census, including addresses that
were duplicated, missed, deleted, and incorrectly located on a map (a
problem known as "geocoding error"). For example, the Bureau is
researching ways to capture missed addresses for housing units that
were hard to find--often associated with apartments in small multiunit
structures. However, some deadlines for completing research are not
firm, while other deadlines that have been set continue to slip. As a
result, it is not known whether the research and evaluation efforts
underway will be completed in sufficient time to allow the Bureau to
develop new methodologies and procedures for improving the MAF by June
2007--the Bureau's announced deadline for determining the baseline for
all program requirements.
In addition, one major research effort using software to identify
duplicate addresses (an estimated 1.4 million duplicate addresses were
removed during the 2000 Census) did not work and will not be used in
2010. As a result, duplicate addresses may still be a problem for the
2010 MAF, and if not detected, can result in increased cost when
nonresponse enumerators attempt to collect data from a duplicate
address incorrectly listed in the MAF.
New issues surrounding the schedule of address activities have emerged.
One such issue revolves around the planning and development of the 2010
Census amid tight and overlapping schedules for updating addresses and
map files. For example, Bureau officials estimate that TIGER maps for
600 to 700 counties of 3,232 counties in the United States will not be
updated in time to be part of local update of census address (LUCA)--
the Bureau's program to give local, state, and tribal government
officials the opportunity to review the address lists and maps and
suggest corrections.[Footnote 17] LUCA participation is important
because local knowledge contributes to a more complete and accurate
address file. Not having the most current TIGER maps could affect the
quality of a local government's review and could potentially increase
the cost of conducting the census. For example, to the extent LUCA
participants are not able to use the maps to identify duplicate and
nonexistent addresses, and if subsequent address operations also fail
to identify those same addresses, then nonresponse follow-up
enumerators would make unnecessary and costly attempts to locate these
incorrectly included addresses.
Bureau Does Not Have a Plan to Assess Resources Needed to Update
Address and Map Files in Areas Affected by Hurricanes Katrina and Rita:
The Bureau does not have a plan to assess additional resources that may
be needed to update the address and map file for areas affected by
hurricanes Katrina and Rita. The task of updating Census address files
to reflect the changes caused by the hurricanes will be formidable and
possibly costly, as much has changed to the landscape since the 2000
Census. On August 29, 2005, hurricane Katrina devastated the coastal
communities of Louisiana, Mississippi, and Alabama. A few weeks later,
hurricane Rita hit the border areas of Texas and Louisiana. Damage was
widespread. For example, the Red Cross estimated that nearly 525,000
people were displaced as a result of hurricane Katrina and
approximately 90,000 square miles were affected. In some places, entire
communities were obliterated. Homes were declared uninhabitable, and
streets, bridges, and other landmarks were destroyed.
For the 2010 Census, locating housing units and the people who reside
in them will be critical to accurate population counts of places hit by
the hurricanes, especially since it is estimated that hundreds of
thousands of people have--either temporarily or permanently--migrated
to other areas of the country. The Bureau anticipates that by 2009,
residents will have decided whether to return to the region. However,
Bureau officials have not provided information regarding the basis of
this conclusion. Given the magnitude of the area, population, and
infrastructure affected, it would be prudent for the Bureau to begin
assessing whether new procedures will be necessary, determining whether
additional resources may be needed, and identifying whether local
partners will be available to assist the Bureau in its effort to update
address and map data, as well as other census-taking activities.
Without having done a resource analysis, the Bureau remains uncertain
about whether additional funds will be needed to help locate and count
residents affected by the hurricanes.
In summary, the 2010 Census is an expensive but vitally important
undertaking, the success of which is needed to meet the information
requirements of policymakers at all levels of government, as well as
business interests, and academic researchers. The Bureau responded to
concerns about the accuracy, completeness, and cost-effectiveness of
the 2000 Census by reengineering the heretofore paper-based processes
used in all previous censuses.
At the same time, the projected life-cycle cost of $11.3 billion makes
the next decennial census the most expensive in our history, and many
factors can cause the 2010 Census to be more expensive. It is important
to consider that some factors that may increase the costs of the
census--such as counting more people than ever who do not speak English
or who live in alternative, hard-to-find housing--are inherent in the
characteristics of the population that needs to be counted. Largely,
demographically related cost factors will continue to exist, regardless
of actions taken by the Bureau, and must be treated as givens by Bureau
planners. Still, other factors that can cause cost increases can and
should be mitigated. While needed, the reengineering introduced by the
Bureau presents new challenges and increased risks. The Bureau needs to
ensure that its new MCDs work as designed, and that contractors perform
according to requirements, on schedule, and at cost. Moreover, the
Bureau still needs to fully resolve preexisting issues related to the
accuracy and completeness of the address list.
Overall, we have long recognized that redesigning massive enterprises
entail risks and uncertainties. Such risks and uncertainties need to be
managed through the use of adequate planning and risk management by
Bureau management. Such tools also serve the oversight requirements of
external stakeholders--most notably Congress, which is being asked to
authorize and appropriate more funds than ever to pay for the census.
In January 2004, recognizing the cost escalation risks of the 2010
Census, we concluded that the Bureau's plans for 2010 lacked the needed
budgetary supporting detail, supporting analysis, and other
information, making it difficult for Congress and us to oversee the
Bureau's operations and assess the feasibility of the Bureau's design
and the extent to which it would lead to greater cost-effectiveness.
While the Bureau has made progress in planning and designing the 2010
Census, the Bureau will need to continue to take steps to manage and
mitigate risks for a comprehensive, accurate, and cost-effective
population count in 2010.
That concludes my statement, Mr. Chairman. I would be pleased to
respond to any questions you or other members of the Subcommittee may
have.
Contacts and Acknowledgements:
For questions regarding this testimony, please contact Brenda S.
Farrell, on (202) 512-6806, or by email at farrellb@gao.gov Individuals
making contributions to this testimony include Betty Clark, Robert
Goldenkoff, Ernie Hazera, Shirley Hwang, Krista Loose, Lisa Pearson,
Scott Purdy, Cynthia Scott, and Tim Wexler.
[End of section]
Appendix I: Related Products by GAO:
GAO Products:
2010 Census: Census Bureau Generally Follows Selected Leading
Acquisition Planning Practices, but Continued Management Attentions Is
Needed to Help Ensure Success. GAO-06-277. Washington, D.C.: May 18,
2006.
Census Bureau: Important Activities for Improving Management of Key
2010 Decennial Acquisitions Remain to Be Done. GAO-06-444T. Washington,
D.C.: March 1, 2006.
2010 Census: Planning and Testing Activities Are Making Progress. GAO-
06-465T. Washington D.C.: March 1, 2006.
Information Technology Management: Census Bureau Has Implemented Many
Key Practices, but Additional Actions Are Needed. GAO-05-661.
Washington, D.C.: June 16, 2005.
2010 Census: Basic Design Has Potential, but Remaining Challenges Need
Prompt Resolution. GAO-05-09. Washington, D.C.: January 12, 2005.
Data Quality: Census Bureau Needs to Accelerate Efforts to Develop and
Implement Data Quality Review Standards. GAO-05-86. Washington, D.C.:
November 17, 2004.
Census 2000: Design Choices Contributed to Inaccuracies in Coverage
Evaluation Estimates. GAO-05-71. Washington, D.C.: November 12, 2004.
American Community Survey: Key Unresolved Issues. GAO-05-82.
Washington, D.C.: October 8, 2004.
2010 Census: Counting Americans Overseas as Part of the Decennial
Census Would Not Be Cost-Effective. GAO-04-898. Washington, D.C.:
August 19, 2004.
2010 Census: Overseas Enumeration Test Raises Need for Clear Policy
Direction. GAO-04-470.Washington, D.C.: May 21, 2004.
2010 Census: Cost and Design Issues Need to Be Addressed Soon. GAO-04-
37. Washington, D.C.: January 15, 2004.
Decennial Census: Lessons Learned for Locating and Counting Migrant and
Seasonal Farm Workers. GAO-03-605. Washington, D.C.: July 3, 2003.
Decennial Census: Methods for Collecting and Reporting Hispanic
Subgroup Data Need Refinement. GAO-03-228. Washington, D.C.: January
17, 2003.
Decennial Census: Methods for Collecting and Reporting Data on the
Homeless and Others Without Conventional Housing Need Refinement. GAO-
03-227. Washington, D.C.: January 17, 2003.
2000 Census: Lessons Learned for Planning a More Cost-Effective 2010
Census. GAO-03-40. Washington, D.C.: October 31, 2002.
The American Community Survey: Accuracy and Timeliness Issues. GAO-02-
956R. Washington, D.C.: September 30, 2002.
FOOTNOTES
[1] The TIGER database is a mapping system that identifies all visible
geographic features, such as type and location of streets, housing
units, rivers, and railroads. To link TIGER to the master address file
(MAF), the Bureau assigns every housing unit in the MAF to a specific
location in the TIGER, a process called "geocoding." TIGER is a
registered trademark of the U.S. Census Bureau.
[2] ACS is intended to be a monthly survey of 250,000 households that,
under the Bureau's plans, will replace the long-form census
questionnaire.
[3] GAO, 2010 Census: Cost and Design Issues Need to Be Addressed Soon,
GAO-04-37 (Washington, D.C.: Jan. 15, 2004).
[4] These figures include the 10-year costs for ACS replacement for the
census long form and the costs of MAF/TIGER.
[5] GAO-04-37.
[6] Lower mail-back response rates increase costs by necessitating
costly follow-up visits by enumerators to nonresponding households and/
or the mailing of a follow-up questionnaire.
[7] U.S. Census Bureau, Census Bureau Estimated Life Cycle Costs for
Reengineering the 2010 Decennial Census Program (Washington, D.C.:
Sept. 2005).
[8] GAO-04-37.
[9] GAO, 2000 Census: Contingency Planning Needed to Address Risks That
Pose a Threat to a Successful Census, GAO/GGD-00-06 (Washington, D.C.:
Dec. 14, 1999).
[10] GAO, 2010 Census: Basic Design Has Potential, but Remaining
Challenges Need Prompt Resolution, GAO-05-9, (Washington, D.C.: January
12, 2005).
[11] GAO, 2010 Census: Census Bureau Generally Follows Selected Leading
Acquisition Planning Practices, but Continued Management Attention Is
Needed to Help Ensure Success, GAO-06-277 (Washington, D.C.: May 18,
2006).
[12] For example, the data capture system exceeded its performance
goals for accuracy, and the advertising campaign blanketed the country
with more than 250 advertisements in 17 languages, which helped boost
the response rate higher than the Bureau had expected.
[13] Department of Commerce Office of Inspector General, Improving Our
Measure of America: What Census 2000 Can Teach Us in Planning for 2010,
OIG-14431 (Washington, D.C.: Spring 2002).
[14] GAO-05-9.
[15] GAO, Census Bureau: Important Activities for Improving Management
of Key 2010 Decennial Acquisitions Remain to be Done, GAO-06-444T ,
(Washington, D.C.: Mar. 1, 2006).
[16] GAO-06-444T.
[17] In the Census Address List Improvement Act (Pub. L. No. 103-430,
Oct. 31, 1994), Congress required the Bureau to develop a local address
review program giving local governments and tribal governments greater
input into the Bureau's address list development process.
GAO's Mission:
The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people.
GAO examines the use of public funds; evaluates federal programs and
policies; and provides analyses, recommendations, and other assistance
to help Congress make informed oversight, policy, and funding
decisions. GAO's commitment to good government is reflected in its core
values of accountability, integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony:
The fastest and easiest way to obtain copies of GAO documents at no
cost is through GAO's Web site (www.gao.gov). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site.
To have GAO e-mail you a list of newly posted products every afternoon,
go to www.gao.gov and select "Subscribe to Updates."
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. Government Accountability 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:
Congressional Relations:
Gloria Jarmon, Managing Director, JarmonG@gao.gov (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125
Washington, D.C. 20548:
Public Affairs:
Paul Anderson, Managing Director, AndersonP1@gao.gov (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548:
