Prepared
Statement OF
Charles
Louis Kincannon
Director
Before the Subcommittee on
Science,
the Departments of State,
Justice and Commerce, and Related Agencies
Committee on Appropriations
27 July 2006
Good morning.
On behalf of the U.S. Census Bureau, I want to thank Chairman Wolf and
the members of the Subcommittee for the opportunity to be here today. I’ve been invited to discuss planning for the
reengineered 2010 Decennial Census Program.
The reengineered 2010 Decennial Census
Program comprises three integrated components.
The American Community Survey, which replaces the census long form, will
provide timely, accurate data for states, towns, and even neighborhoods. Second, the Master Address File (MAF) and
TIGER Enhancement Program will serve our nation by updating the address list
and modernizing the electronic maps by which we collect and disseminate data.
Third, and most important, is the 2010 Census, a short-form only census, which
is the Census Bureau’s core constitutional responsibility. The primary goals of the reengineered 2010
Decennial Census Program are: 1) to improve the relevance and timeliness of
census long-form data; 2) reduce operational risk; 3) improve the accuracy of
census coverage; and 4) contain costs.
It will enable us to fulfill our mission in the most efficient,
cost-effective, and accurate manner possible.
The 2007 Budget
Request and Work-to-Date
Before
we discuss the ramifications of the recent House and Senate actions, it is
worth briefly reviewing the budget request.
The President’s FY 2007 budget request for the Census Bureau was
approximately $878 million, including $184 million for salaries and expenses,
as well as $182 million for other economic and demographic programs conducted
by the Census Bureau. The majority of
the budget request— $512 million— was designated for the decennial census
program, an increase of $64 million from last year.
This
increase is primarily due to preparations for the 2008 Census Dress Rehearsal,
including development of the handheld computers to be used in field operations
and the data capture system to collect all types of survey responses. The decennial request included $180 million
for the American Community Survey; $74 million for MAF/TIGER; and $258 million
for the 2010 Census.
It
is important to put these costs in context, both in terms of the unique
function of the census in our government and society, and in terms of the total
cost of the mandate over the course of a decade. The Census is one of the first tasks
explicitly mentioned in the Constitution, directly following the election of
representatives. Article I, Section 2 of
the Constitution states, “the actual enumeration shall be made within three
years after the first meeting of the Congress of the
To
conduct Census 2000, the Census Bureau canvassed six and one half million miles
of roads in 66,000 census tracts in 3200 counties in 50 states and the
Conducting
an accurate census poses a formidable challenge. The Census Bureau must strive to count each
member of the nation’s increasingly diverse population once and only once and
in their precise geographic location.
Accuracy is fundamental. The
census determines the composition of the U.S. House of Representatives. It also determines the representation for
fifty state legislatures and 39,000 municipalities. The census also directs the distribution of
nearly $200 billion of federal funds annually, and many more billions in other
state and local funding, as well as private investment.
I assure you the Census Bureau is prepared to meet this challenge. Early in this decade, the Census Bureau introduced and sought support for the reengineered 2010 Decennial Census Program. This plan is responsive to the concerns and suggestions expressed during Census 2000. Many, including Congress and the Government Accountability Office, expressed concern about the risks and costs associated with the census, urging the Census Bureau to incorporate earlier planning and testing in the design for 2010.
At this point in the decade, the
Administration and the Congress have committed more than $1.3 billion toward
these goals and to the reengineered census.
Therefore, it is important to review the progress we have made thus far
with regard to the decennial census.
·
In 2002, the Harris Corporation was
awarded an eight-year contract, valued over $200 million, to work with the
Census Bureau to match the street centerlines of the TIGER database with
GPS-coordinates. This will support our
efforts to use GPS-equipped handheld computers in the field for data collection
in 2010. To date, we have aligned
streets and roads for approximately 1,700 of the nation’s counties, with about 1,600
remaining to reach completion by April 2008.
·
In 2003, the Census Bureau conducted a
national test to study alternative self-response options and contact
strategies. We also looked at
alternative presentations of the race and Hispanic origin questions as part of
the planning and testing efforts for the short-form only census in 2010.
·
In 2004, a major field test was
conducted in two locations, Southwest Georgia and
·
In 2005, the Census Bureau began full
implementation of the American Community Survey with a sample of nearly three
million households per year. This is a
major undertaking which provides timely data for states and local communities,
replacing the old system that delivered data only once a decade. It is one of the principal goals of the
reengineered census. The American
Community Survey will provide timely, accurate information for every county,
city, and neighborhood—the level where the most crucial decisions affecting
American communities are made. These
data are required to carry out an array of federal mandates, including the Voting
Rights Act. The responses to population
and housing questions support programs such as No Child Left Behind, Low Income
Home Energy Assistance Program, and community block grants.
·
Also in 2005, we conducted a second
national mailout test to study such things as new coverage questions; wording
and presentation of residence rules instructions; design, layout, wording, and
presentation of the race and ethnicity questions and other short form content;
and replacement questionnaire strategies.
This year, the Census Bureau has just
completed initial data collection for the second major field test for the 2010
Census in Travis County, Texas and on the Cheyenne River Reservation in South
Dakota. This is the final opportunity to test methods and technologies in the
field before the 2008 Census Dress Rehearsal.
The Census Bureau is also on schedule to deliver the first release of
annual data from the American Community Survey next month. For the first time, we will provide social
and economic data for every congressional district and every government with a
population of 65,000 or more. Each of
these activities is consistent with the Administration’s priorities and
received Congressional approval. They
are an important part of the decennial lifecycle and lay the groundwork for
success in the 2010 Census.
We
have also made progress with respect to automation and infrastructure. These efforts support several reengineered
census goals, including improved accuracy and reduced risk. Part of our effort has centered on two major
systems, the 2010 Decennial Response Integration System (DRIS) and the Field
Data Collection Automation (FDCA) system.
Both of these are large information technology contracts, totaling
together over one billion dollars. The purpose of the DRIS contract, which was awarded
last year to Lockheed Martin Corporation, is to
ensure accurate and protected collection and storage of Americans’ data whether
by paper form, handheld computer, or telephone. We are currently
involved in Phase I of this program, which includes design and implementation
of the system for the 2008 Census Dress Rehearsal. We also awarded the FDCA contract, which
provides comprehensive up-to-date technology for our field data collection and
our field offices. The handheld
computers being developed as part of the FDCA contract will be used in the
field for two main decennial operations: the Address Canvassing activity that
ensures the currency and completeness of the Master Address File and the
collection of census data by personal visits to households that have not
responded by other means. The contract
for handheld computers, software, telecommunications, desktop environments, and
support services was awarded to the Harris Corporation in March of this year.
At
this stage of the decennial lifecycle we are poised to make final the
fundamental methodological and operational design decisions about the
collection of data in the 2010 Census.
These decisions have assumed the use of GPS-equipped handheld computing
devices for data collection. However,
recent congressional actions, if enacted, would require us to reevaluate the
plans for the reengineered 2010 Decennial Census Program. As you are aware, the recent House action is
primarily targeted at the 2010 decennial census, prescribing a reduction of
$53.3 million for that program alone.
The total reduction to our periodic programs request is $58.3
million.
Difficult Decisions: Impacts of the Current House Mark
The
current House mark is problematic for two reasons: the size of the reduction,
and the nature of the cut, which does not provide us the flexibility to spread
the cut to non-decennial items. A
reduction of this magnitude in FY 2007 funding would have a major impact on the
2010 Decennial Census Program and could prevent the Census Bureau from making long-sought
improvements and meeting the agreed-upon goals of the reengineered census. Many of the planned efforts are targeted at
improving coverage, reducing the undercount, correcting geographic
misallocations, and collecting data with the least possible burden on
respondents.
In addition, I believe that
with the current House bill, we would not be able to proceed with the data
collection automation efforts.
Specifically, I believe the Census Bureau would need to drop its
automation efforts to develop and implement GPS-equipped handheld computers in
the field. We have evaluated the plans
for the decennial census looking for alternative reductions. Alternative
reductions, however, would affect a broader array of activities, thereby
undercutting our ability to conduct the Dress Rehearsal, which in turn will
jeopardize our ability to implement new methods, systems, and operations
designed to meet the strategic goals of the 2010 Census—to improve coverage,
contain costs, and reduce risk. We
believe it is imperative to conduct a dress rehearsal that allows us to
implement these new components (methods, operations, and systems—including
system interfaces) and ensure they will work—and work together—for the 2010
Census. Major Dress Rehearsal operations
that are affected by these new components include such things as address list
development, Group Quarters enumeration, military enumeration, service-based
enumeration, Update-Leave, the second mailing, data capture, and coverage
measurement field operations. If we do
not have the opportunity to do this, it would be ill-advised to introduce these
components for the first time in the 2010 Census, so opportunities for coverage
improvement, cost reductions, and risk reduction would be lost.
Dropping
the plans to use handheld computers would force Census enumerators to revert
back to a paper-based census, ultimately increasing costs. We would begin planning immediately to use
paper questionnaires, rather than handheld computers in the field, to conduct
the 2008 address canvassing operation and the non-response follow-up (NRFU) in
preparation for the 2010 Census.
Significantly more space, more staff, and additional resources would be
needed to manage the paper process and to extend and expand our paper capture capabilities. Reverting to a paper census would also have a
deleterious effect on coverage. We would
no longer be able to plan to electronically remove late mail returns from the
enumerator assignments, which would preclude enumerators knocking on the doors
of people who already responded. Also
without the GPS-equipped handheld computers, corrections to geographic
additions in the field would be done manually, which is more error-prone and
would negatively impact the geographic accuracy of the census.
The
House mark would also significantly impact the American Community Survey. We would eliminate the Group Quarters data
collection operation. Without the
detailed social and economic information about people living in Group Quarters,
including prisons, nursing homes, and college dormitories, the American
Community Survey would result in incomplete information for certain populations
such as the elderly, who are more likely to live in Group Quarters.
The cut would also affect
other critical activities, particularly the Community Address Updating System
(CAUS), which is part of the MAF/TIGER Enhancement Program. Without CAUS, we cannot work with local
communities to update addresses in rural
Impacts of an Alternative Funding Proposal
The current House mark would force us to make
fundamental changes to our plans for 2010 Census, including the elimination of
the use of handheld computers for field data collection or acceptance of other
highly undesirable risks. The risk of
failure and the associated costs would be dramatically increased.
Therefore, we have explored the option of spreading
a portion of the $53.3 million reduction to programs outside of the Decennial
Census, even though the current House mark does not give us that
flexibility. This alternative allocation
would reduce funding for programs in both the salaries and expenses and periodic
programs accounts. If we are allowed to
prioritize and propose alternative reductions to the Census Bureau’s programs,
which would result in our continuing the reengineering program, we can better
minimize the effects on the decennial census.
These alternative reductions, however, will have substantial impacts on
other programs at the Census Bureau while continuing to affect the decennial
census and the American Community Survey.
We would still be forced to cancel Group Quarters data collection, and
the CAUS activities. We would also delay
the remaining work to align street centerlines.
We
have examined economic and other demographic programs to try to identify
activities that might be reduced without significant impact to the nation’s
data infrastructure. In the economic
area, we would potentially eliminate, delay, and reduce programs such as the
Current Industrial Reports, Quarterly Financial Reports, Quarterly Residential
Improvements and Repair Data, and Quarterly Services Survey. In selecting the programs to reduce, we
strived to preserve the programs and content that serve as benchmarks for
current surveys, as well as composite measures of economic activity, including
GDP, the Producer Price Index, and the Index of Industrial Production. We must also retain the Principal Economic
Indicators, as these provide GDP source data, and the annual surveys that are
used in the Bureau of Economic Analysis’ National Income and Product Accounts
and the Federal Reserve Board’s Flow of Funds.
Finally, it is vital that we protect core infrastructure, such as the
Business Register, that serves as the foundation upon which other programs
operate.
The Importance
of Census Data
While
we recognize the current fiscal climate and the need to conserve resources, the
House mark, however allocated, would lead to reductions that would harm key
Census Bureau programs and the accuracy of our Nation’s statistical data. In an era of dramatic economic activity and
change, the House proposed funding level would result in less information
available for use by decision makers to assess the health of the American
economy. They would lead to less
accurate data for determining Congressional representation and implementing
Federal assistance programs. Among
Federal stakeholders, these data are critical to the Congress, the Federal
Reserve Board, the Bureau of Economic Analysis, the Department of the Treasury,
and the Bureau of Labor Statistics to evaluate economic well-being or potential
growth of the economy. The data are also
used in trade negotiations and to assess the impact of global markets on the
I
am grateful to the Chairman and this Committee for the opportunity to discuss
the impact of proposed reductions to the Census Bureau’s budget. I strongly urge your colleagues to support
the President’s 2007 request of $878 million for the Census Bureau. This funding is necessary if we are to
continue a decennial census reengineering effort, while also producing timely
and accurate economic and demographic data for the Nation. Short of that, we are asking for the
discretion to absorb reductions to minimize impairments to our mission and the
priorities mandated over the past several years with regard to the decennial
census. With sufficient funding, we
believe it is possible to conduct an accurate, improved census in 2010 that
incorporates long-sought coverage improvements such as a second mailing,
bilingual questionnaires, and the use of GPS-enabled handheld computers that
should improve accuracy and reduce lifecycle costs. In contrast, the risks and costs associated
with the decennial census are likely to grow if we do not remain on track. If we are required to cease our handheld
computer and automation preparations in FY 2007, we would not be able to start
them again and catch up in later years.
These opportunities for 2010 would be lost.
For
the past five years, the Census Bureau has proceeded along a clearly defined
plan, with specific testing goals and results.
But now, two years from the 2008 Census Dress Rehearsal, we are forced
to question key operational and design decisions. The National Academy of Sciences, in its
appraisal of the 2000 census, warned against this predicament, recalling the
increased costs, confused constituencies, and added tensions. The National Academy of Sciences said, and I
quote:
“The 2000 census planning began in a climate of
concern about the perceived failures of the 1990 census—one that saw a
substantial decline in public cooperation and, despite higher per household
costs than in the 1980 census, resulted in worse coverage of minorities,
renters, and children relative to other population groups. The Census Bureau’s initial design to remedy
these problems for 2000 relied on much greater use of statistical techniques in
the census enumeration, but this plan encountered opposition from members of
Congress and others. As a result the
Bureau had to contend with externally imposed last-minute changes in design,
delayed budget decisions, consequent changes in plans, and insufficient time
for operational trials. All of these
problems increased not only the costs of the 2000 census but also the risk that
it could have been seriously flawed in one or more respects. In light of this experience, the panel
recommends that the Census Bureau, the administration, and Congress agree on
the basic design for the 2010 census not later than 2006 in order to permit an
appropriate, well-planned dress rehearsal in 2008.” (End quote.)[1]
I
am concerned about the uncertainty at this stage of the 2010 decennial
lifecycle and how it bears resemblance to the environment of earlier censuses. We must resolve the key operational and
design decisions before proceeding.
Every delayed decision will result in further delays in testing—and
possibly even no testing. Delayed
decisions increase both operational risks and costs, and threaten the accuracy
of the data.
We
believe the President’s 2007 Budget request addresses these risks, contains
costs, and provides timely, more relevant data for the nation, and we ask for
support of the President’s budget request.
We believe the American Community Survey is a success and will provide
the nation’s states, cities, and communities with the data to make informed
decisions. We believe the MAF/TIGER
Enhancement Plan will enhance a national resource, improving the accuracy of
geospatial data for all users. And
finally, we believe the short-form only 2010 Census will provide an accurate
count, consistent with the Constitution’s promise.
Mr.
Chairman, with the support of Congress, I believe we can succeed, and I would
be happy to answer your questions.