[House Hearing, 109 Congress]
[From the U.S. Government Publishing Office]


 
  2+2 SHOULD NEVER EQUAL 3: GETTING INTERCENSAL POPULATION ESTIMATES 
                          RIGHT THE FIRST TIME

=======================================================================

                                HEARING

                               before the

                       SUBCOMMITTEE ON FEDERALISM
                             AND THE CENSUS

                                 of the

                              COMMITTEE ON
                           GOVERNMENT REFORM

                        HOUSE OF REPRESENTATIVES

                       ONE HUNDRED NINTH CONGRESS

                             SECOND SESSION

                               __________

                           SEPTEMBER 6, 2006

                               __________

                           Serial No. 109-244

                               __________

       Printed for the use of the Committee on Government Reform


  Available via the World Wide Web: http://www.gpoaccess.gov/congress/
                               index.html
                      http://www.house.gov/reform

                     U.S. GOVERNMENT PRINTING OFFICE

36-680 PDF                 WASHINGTON DC:  2007
---------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office  Internet: bookstore.gpo.gov Phone: toll free (866)512-1800
DC area (202)512-1800  Fax: (202) 512-2250 Mail Stop SSOP, 
Washington, DC 20402-0001






































                     COMMITTEE ON GOVERNMENT REFORM

                     TOM DAVIS, Virginia, Chairman
CHRISTOPHER SHAYS, Connecticut       HENRY A. WAXMAN, California
DAN BURTON, Indiana                  TOM LANTOS, California
ILEANA ROS-LEHTINEN, Florida         MAJOR R. OWENS, New York
JOHN M. McHUGH, New York             EDOLPHUS TOWNS, New York
JOHN L. MICA, Florida                PAUL E. KANJORSKI, Pennsylvania
GIL GUTKNECHT, Minnesota             CAROLYN B. MALONEY, New York
MARK E. SOUDER, Indiana              ELIJAH E. CUMMINGS, Maryland
STEVEN C. LaTOURETTE, Ohio           DENNIS J. KUCINICH, Ohio
TODD RUSSELL PLATTS, Pennsylvania    DANNY K. DAVIS, Illinois
CHRIS CANNON, Utah                   WM. LACY CLAY, Missouri
JOHN J. DUNCAN, Jr., Tennessee       DIANE E. WATSON, California
CANDICE S. MILLER, Michigan          STEPHEN F. LYNCH, Massachusetts
MICHAEL R. TURNER, Ohio              CHRIS VAN HOLLEN, Maryland
DARRELL E. ISSA, California          LINDA T. SANCHEZ, California
JON C. PORTER, Nevada                C.A. DUTCH RUPPERSBERGER, Maryland
KENNY MARCHANT, Texas                BRIAN HIGGINS, New York
LYNN A. WESTMORELAND, Georgia        ELEANOR HOLMES NORTON, District of 
PATRICK T. McHENRY, North Carolina       Columbia
CHARLES W. DENT, Pennsylvania                    ------
VIRGINIA FOXX, North Carolina        BERNARD SANDERS, Vermont 
JEAN SCHMIDT, Ohio                       (Independent)
------ ------

                      David Marin, Staff Director
                Lawrence Halloran, Deputy Staff Director
                      Benjamin Chance, Chief Clerk
          Phil Barnett, Minority Chief of Staff/Chief Counsel

               Subcommittee on Federalism and the Census

                   MICHAEL R. TURNER, Ohio, Chairman
CHARLES W. DENT, Pennsylvania        WM. LACY CLAY, Missouri
CHRISTOPHER SHAYS, Connecticut       PAUL E. KANJORSKI, Pennsylvania
VIRGINIA FOXX, North Carolina        CAROLYN B. MALONEY, New York
BRIAN P. BILBRAY, California

                               Ex Officio

TOM DAVIS, Virginia                  HENRY A. WAXMAN, California
                     John Cuaderes, Staff Director
            Ursula Wojciechowski, Professional Staff Member
                         Juliana French, Clerk
          Mark Stephenson, Minority Professional Staff Member


























                            C O N T E N T S

                              ----------                              
                                                                   Page
Hearing held on September 6, 2006................................     1
Statement of:
    Kincannon, Charles Louis, Director, U.S. Census Bureau.......     8
    Swanson, David A., professor of sociology, director of the 
      Center for Population Studies, Chair of the Department of 
      Sociology and Anthropology, University of Mississippi; Joy 
      E. Phillips, Ph.D, associate director, State Data Center, 
      District of Columbia Office of Planning; Ken Hodges, chief 
      demographer, Claritas Inc.; and Warren A. Brown, senior 
      research associate, director, program on applied 
      demographics, research director, New York Census Research 
      Data Center, Cornell Institute for Social and Economic 
      Research, Cornell University...............................    32
        Brown, Warren A..........................................    62
        Hodges, Ken..............................................    56
        Phillips, Joy E..........................................    40
        Swanson, David A.........................................    32
Letters, statements, etc., submitted for the record by:
    Brown, Warren A., senior research associate, director, 
      program on applied demographics, research director, New 
      York Census Research Data Center, Cornell Institute for 
      Social and Economic Research, Cornell University, prepared 
      statement of...............................................    64
    Clay, Hon. Wm. Lacy, a Representative in Congress from the 
      State of Missouri, prepared statement of...................    81
    Hodges, Ken, chief demographer, Claritas Inc., prepared 
      statement of...............................................    58
    Kincannon, Charles Louis, Director, U.S. Census Bureau, 
      prepared statement of......................................    10
    Phillips, Joy E., Ph.D, associate director, State Data 
      Center, District of Columbia Office of Planning; Ken 
      Hodges, chief demographer, Claritas Inc., prepared 
      statement of...............................................    43
    Swanson, David A., professor of sociology, director of the 
      Center for Population Studies, Chair of the Department of 
      Sociology and Anthropology, University of Mississippi, 
      prepared statement of......................................    35
    Turner, Hon. Michael R., a Representative in Congress from 
      the State of Ohio, prepared statement of...................     3


  2+2 SHOULD NEVER EQUAL 3: GETTING INTERCENSAL POPULATION ESTIMATES 
                          RIGHT THE FIRST TIME

                              ----------                              


                      WEDNESDAY, SEPTEMBER 6, 2006

                  House of Representatives,
         Subcommittee on Federalism and the Census,
                            Committee on Government Reform,
                                                    Washington, DC.
    The subcommittee met, pursuant to notice, at 2 p.m., in 
room 2247, Rayburn House Office Building, Hon. Michael R. 
Turner (chairman of the committee) presiding.
    Present: Representatives Turner, Norton, Maloney, and Dent.
    Staff present: John Cuaderes, staff director; Ursula 
Wojciechowski, professional staff member; Juliana French, 
clerk; Jon Heroux, counsel; Peter Neville, fellow; Kimberly 
Trinca, minority counsel; Mark Stephenson and Adam Bordes, 
minority professional staff members; and Cecelia Morton, 
minority office manager.
    Mr. Turner. Good afternoon. A quorum being present, this 
hearing of the Subcommittee on Federalism and the Census will 
come to order.
    Welcome to the subcommittee's hearing entitled ``2+2 Should 
Never Equal 3: Getting Intercensal Population Estimates Right 
the First Time.'' This is the fifth in a series of hearings on 
Census Bureau programs. Today's hearing will examine the 
Bureau's intercensal population estimates program.
    Intercensal population estimates are vital to the accurate 
allocation of hundreds of billions of Federal dollars. 
Unfortunately, intercensal estimates are often inaccurate. As a 
result, the allocation of these Federal dollars is also often 
inaccurate.
    Washington, DC, for example, has disputed the Bureau's 
estimates every year since 2002, and has noted in a July 22, 
2006 Washington Post article that the Bureau recently 
acknowledged that they had missed about 6 percent of D.C.'s 
population.
    Cities, counties and States can challenge the Bureau's 
estimates and have done so 91 times since the 2000 census. 
However, I am concerned that numerous other local governments 
may not have the resources or expertise necessary to challenge 
the estimates or are unaware of the opportunity to challenge. 
As a result, these governmental units live with what may be 
inaccurate estimates and inaccurate Federal funding 
allocations.
    During this hearing, we will examine whether the Bureau's 
methodology results in reasonably accurate estimates and 
equitable allocation of Federal grants; the importance of 
accuracy; opportunities for improving the estimates; whether 
the Bureau's challenge process is transparent, fair and takes 
into account a community's ability to challenge; and whether 
the Bureau strives to continuously improve the estimates.
    Before we move on, I would like to recognize Representative 
Eleanor Holmes Norton of the District of Columbia, who 
certainly has an interest in this issue. And I yield to her for 
any comments she may have.
    [The prepared statement of Hon. Michael R. Turner follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Ms. Norton. Thank you, Mr. Chairman. I have only brief 
comments, beginning with thank you for both inviting the 
District of Columbia to testify here and for inviting me, 
although a member of the full committee, not of this 
subcommittee, to sit with you in this very important hearing. 
It is always timely, but particularly timely now, because while 
citizens may not know it, the census of course has to begin 
preparing full throttle now for the regular census.
    Mr. Chairman, your initiative here today may not produce 
what most of us would like, an immediate increase in the per 
capita that we receive, because now that you know there are 
more people, OK, where's the moola. But it is very important 
that the census not have to essentially wait until 10 years to 
try to figure it all out, and what is supposed to happen is 
greater accuracy because they are doing it all along and we 
want to find out why that has not occurred in so many 
instances.
    Mr. Chairman, if I may say so, we may not get money or any 
benefit from an intercensal new estimate. But it can do huge 
harm. Everybody looks at these census figures to see whether, 
for example, businesses should move here, whether schools are 
growing or losing population. These figures from the granddaddy 
of all figures, the Census Bureau, all of us rely upon the 
Census Bureau and honor the Census Bureau for its work. But to 
the extent that these intercensal estimates are inaccurate, 
they can do huge harm to local jurisdictions, which is why you 
see local jurisdictions rising up in arms, not because they 
expect any change, immediate change in Federal disbursements, 
but because they know everybody who has anybody to do with 
their city in any fashion will look first here.
    Now, when people rise up in arms about whether they are 
losing or gaining population, everybody says, yeah, yeah, yeah, 
you're never going to agree that you're losing. But we find out 
that the Census, in its honesty, and that is what you have to 
congratulate the Census most for, that it is non-partisan, it 
acts on the figures. The problem we have is perhaps with its 
methodology. We knew in the District of Columbia that we were 
one of the hottest cities in the United States. We knew that 
people were moving into the District, and we had to try to 
figure out why the Census said we were continuing to lose 
population at a huge rate.
    Who gets hurt, finally, Mr. Chairman? Think about who gets 
money. Most people in the middle class do not get Federal 
grants, ultimately. Who depends upon this per capita, and we 
are basically talking about per capita grants, who depends upon 
it are the people who get food stamps, people who need Section 
8 housing, so the accuracy, particularly since much of the 
money goes directly to needy recipients, could not be more 
important.
    I thank you and congratulate you on starting us off with a 
serious look at this issue.
    Mr. Turner. Thank you.
    I want to recognize Carolyn Maloney of New York, who just 
got here. If you'd like, I can read the witnesses that are 
going to be testifying as you get settled, and then return to 
you for opening comments.
    Mrs. Maloney. That would be great. Thank you.
    Mr. Turner. On our first panel, we have the Honorable 
Charles Louis Kincannon, Director of the U.S. Census Bureau. On 
our second panel we have four witnesses. They are: Dr. David 
Swanson, director of the State Center of Mississippi and 
professor at the University of Mississippi; Dr. Joy Phillips, 
associate director of the Washington, DC Data Center; Mr. Ken 
Hodges, assistant vice president and chief demographer for 
Claritas Inc.; and Dr. Warren Brown, research director of the 
Cornell Institute for social and economic research at Cornell 
University. We look forward to each of the witnesses and their 
expert testimony today.
    And with that, then, I would like to recognize Carolyn 
Maloney of New York for her opening statement.
    Mrs. Maloney. I really want to thank the chairman and the 
ranking member, Ms. Norton, for holding this. First of all, Mr. 
Chairman, let me say that I am delighted to see so many New 
Yorkers represented at the hearing today. Warren Brown from 
Cornell and Ken Hodges from Claritas are but two of the 
excellent demographers we have in the city of New York. And I 
am proud that both the State of New York and the city of New 
York have been at the forefront of the fight for an accurate 
census.
    Absent today is Joe Salvo from the city planning department 
of New York city. While Dr. Salvo is not in the room today, we 
cannot discuss census accuracy without recognizing the 
leadership he and the entire team at the Department of City 
Planing have demonstrated on this issue. Both Mayor Michael 
Bloomberg and Giuliani before him have recognized the 
importance of an accurate count. And the planning department 
has worked very hard to help ensure it.
    The 2000 census would have missed over a quarter of a 
million people in New York City if the city had not forced the 
Census Bureau to correct its address list for the city. Since 
then, the planning department has repeatedly challenged the 
Census Bureau's estimates for New York City, and won every 
single challenge. The citizens of New York City are deeply 
indebted to the hard-working experts at the New York City 
planning department.
    Unfortunately, the issue of the accuracy of census counts 
and estimates is not new. Following the 1990 census, the Census 
Bureau made a political decision not to adjust the intercensal 
population estimates for the net under-count observed in the 
1990 census. The 2000 census results proved the folly of that 
decision.
    The GAO reported that the 1999 estimates underestimated the 
total population of the United States by 6.8 million people. 
Half of that error would have been eliminated had the Census 
Bureau corrected these estimates for the observed net under-
counted.
    The 2000 census which is used as the base for the 
intercensal estimates since then is equally flawed. We are 
repeatedly told that the net error in the 2000 census is almost 
zero. While that statistic may be true, it hides the huge 
errors in that census. While the Census Bureau has refused to 
produce estimates of the gross errors in the 2000 census, that 
is the number of people missed in the census and the number of 
people counted twice, we know from its own research that the 
gross error is at least 12 million.
    The Census Bureau gets to zero error by letting the people 
counted twice substitute for those missed. That may work for 
statistics, but it does not work for representation. The people 
counted twice are not like the people missed. They are not the 
same color, they do not have the same income, and they do not 
live in the same places. Those errors are built into the 
estimates for 2001 through 2005.
    As I mentioned earlier, New York's 2000 census count is 
considerably higher than it would have been if the Census 
Bureau had been left on its own. Instead, the city of New York 
took advantage of a program mandated by Congress to review and 
correct the Census Master Address File for the 2000 census. The 
city identified for the Census Bureau thousands of addresses 
that were not on the census address list. Addresses not on the 
address list don't get counted. Making sure those addresses 
were on the list gave New York City a more accurate count.
    The story does not end there, as we will hear from the 
witnesses today. The Census Bureau estimates are worse in 
cities like New York, where IRS records do not adequately 
reflect changes in the city's population. As a result, New York 
City has repeatedly been forced to challenge the Census Bureau 
estimates to get a fair count.
    However, the fair count for New York City comes at the 
expense of other cities in the State. While the Census Bureau 
has repeatedly been forced into cooperating with State and 
local governments, it has done so reluctantly. GAO reviewed the 
cooperative effort to correct the census address list prior to 
the 2000 census and found that effort was weak at best. Plans 
for a local updating of census addresses for 2010 remains 
sketchy at best.
    Congress was told that the American Community Survey would 
allow the Census Bureau to produce more accurate population 
estimates. We repeatedly asked for the details of just how that 
would be done. But no details were forthcoming.
    Congress has invested nearly half a billion dollars in the 
American Community Survey. I hope the Director's testimony 
today will tell us how that investment has improved the 
population estimates. And Mr. Chairman, I am pleased that you 
are holding this hearing today. The Census Bureau should be 
held accountable. However, this problem is not new. When the 
results of the 2000 census were announced in December 2000, one 
of the startling facts was that the population estimates were 
seriously flawed. They were flawed 6 years ago, and they were 
equally flawed today.
    Thank you.
    Mr. Turner. I want to thank both Mrs. Maloney and Ms. 
Norton for their participation today, Ms. Norton for her 
representation of our Nation's capitol and Carolyn Maloney for 
her representation of New York City. Both cities that we will 
see in our testimony that have challenged the estimates, both 
of which have been under-counted. And the impacts on those 
communities are something we are trying to address.
    With that, I want to proceed with turning to Mr. Kincannon. 
As everyone is aware, each witness has kindly prepared written 
testimony which will be included in the record of this hearing. 
Witnesses will notice that there is a timer light on the 
witness table. The green light indicates that you should begin 
your prepared remarks, and the red light indicates that your 
time has expired. The yellow light will indicate when you have 
1 minute left in which to conclude your remarks.
    It is the policy of this committee that all witnesses be 
sworn in before they testify. So Director Kincannon, will you 
please rise and raise your right hand.
    [Witness sworn.]
    Mr. Turner. Please note that the witness has responded in 
the affirmative. And Director Kincannon, we will turn to you 
for your statement.

  STATEMENT OF CHARLES LOUIS KINCANNON, DIRECTOR, U.S. CENSUS 
                             BUREAU

    Mr. Kincannon. Thank you, Mr. Chairman. Good morning, and 
on behalf of the U.S. Census Bureau, I want to thank you, 
Chairman Turner, and the members of the subcommittee for the 
opportunity to discuss our intercensal estimates program.
    Population estimates provide important information for 
decisionmakers at the national, State and local levels, and are 
essential to many Federal programs and initiatives. They are 
used to determine program eligibility and to ensure equitable 
funding for a large number of Federal programs, including 
Federal transit grants, the Crime Victim Assistance Program, 
Community Development Block Grant program, medical assistance 
programs and many others.
    The estimates are also used by the Census Bureau to provide 
population controls for the American Community Survey, which 
was fully implemented only last year with the support of this 
committee and the Congress. The American Community Survey 
replaces the long form of the census with annual relevant 
information for local communities. It provides vital 
information that describes how the population is actually 
changing, not just population decline or growth. It also 
provides vital information on the occupancy rate, persons per 
household, international migration and domestic migration, all 
of which will help improve the estimates program. But this is 
the first year in which we have had nationwide data covering 
all of the country that we will be able to use that.
    The estimates are produced in cooperation with 
representatives from each of the States. Each year, we produce 
more than 39,000 sub-national estimates of total population, 
including estimates for every State, county and incorporated 
place or city. We also report measures of the population by 
age, sex, race and Hispanic origin for the Nation, States and 
counties.
    The population estimates released annually by the Census 
Bureau are the product of a longstanding, valued partnership 
with the States. This 40-year old cooperative is alive and 
thriving today in every State as well as the District of 
Columbia, and Puerto Rico is a participating member. We work 
together each year to produce a set of consistently, timely 
estimates for every State, county and incorporated place.
    Local areas also have an opportunity to revise their 
estimates through a challenge process using one of several 
established approaches that may require additional data and 
take into account local information. We provide guidance and 
assistance to help any local community through that process. 
Normally, a community contacts us by phone, e-mail or letter to 
express their concern and request a challenge package. The 
local government is invited to provide source material, such as 
building permits, utility hookups and other data to supports 
its request for an alternative estimate. If the data support an 
alternative estimate, we will send an acceptance letter to the 
local government, notify relevant agencies and post the new 
estimate on the Census Bureau's Web site. This new estimate 
becomes the official estimate and we notify the relevant 
agencies of Government of this change.
    Of the 39,400 official population estimates released for 
2004, there were only 38 challenges, most of which were 
accepted. The challenges have been described in the media as 
Census Bureau mistakes that it has been forced to correct. It's 
not really that simple. The population estimates are the result 
of a cooperative effort. We rely on information from other 
Federal agencies, as well as the States, to produce one set of 
official estimates intended to serve all customers. If 
documentation is provided that supports an alternative 
estimate, the original is revised, not because we didn't get it 
right the first time, but because working with the States and 
tracking into account their additional documentation, we can 
arrive at a more accurate estimate.
    As we look to the future, we are re-examining our 
assumptions, methodologies and source data and discussing these 
efforts with our partners, especially the States that rely on 
the estimates for funding allocations. Alaska, California and 
North Dakota are particularly interested in working with the 
Census Bureau and have contributed to our research efforts.
    During the past 6 months, we have sponsored two research 
conferences on population estimates, with participants from 
Federal agencies, State partners and academia. In these 
conferences we examined how we could better improve the 
estimates for international migration, current assumptions and 
methodologies and better meet user needs. Many users believe we 
should examine new approaches and perhaps embrace alternative 
or multiple methods. It is important as we proceed to consult 
with our State partners and data users.
    As Members of Congress, you are also an important partner, 
because your decisions affect what we do and affect every 
community in America. As we study alternatives, I hope we will 
have a chance, Mr. Chairman, to brief you in the future.
    Thank you very much, and I would be happy to answer any 
questions.
    [The prepared statement of Mr. Kincannon follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Turner. Thank you, Director Kincannon.
    In looking at your testimony, there is one part of it that 
kind of strikes me. In looking at the testimony that is to 
follow you and in looking at several of the news articles that 
describe estimates that have been made in estimates as a result 
of protests from cities, it's clear that some of the issues 
that cities face, or communities that face that challenge of 
this data, is that once a challenge has been made, once 
additional information is provided, it doesn't appear that the 
Census Bureau adequately reflects that change that has 
occurred, or the change in methodology that would then reflect 
more accurate estimates in the future.
    And your statement here, knowing that's part of the issue, 
that it's not just that you didn't get it right the first time, 
but that it's not right the first time and then people have to 
repeatedly go back to readdress the issue, your statement here 
strikes me, because you said ``many of the challenges we 
ultimately accepted were described in the media as Census 
Bureau mistakes.'' And it's not that simple, you say.
    Well, if the data that you have is wrong and the estimates 
that you had were ultimately proven to be not accurate, then it 
would seem that they were mistaken, and if they were mistaken, 
that would appear to be a mistake. And once that is identified, 
it would seem that there would be a process by which the Census 
Bureau would then modify its procedures and processes, so that 
communities would not have to continually be in a process of 
coming to the Census Bureau and providing other data to reflect 
changes in the estimation process.
    Could you please describe, once a community objects, 
provides you that additional data and it is accepted by you, 
what is done by the Census Bureau to make certain that is 
adequately reflected in the future?
    Mr. Kincannon. Well, what we can do, and do, if we accept a 
challenge, based on the evidence submitted by a city, that 
becomes the new basis for this set of estimates that we have 
next year. So we don't throw away what they've given us. But 
new changes, additional housing units, additional developments 
of some kind, new changes in group quarters, are not any more 
known to us in the year plus one than they were in the base 
year.
    So we really need the assistance of State and local 
officials in making sure that the changes that don't show up in 
the large scale record data, administrative record data, for 
States and for counties, can be taken into account.
    Mr. Turner. I have a letter that we've received from the 
North Dakota census committee member, Katherine Strumbeck. And 
she states in here, ``North Dakota has experienced declining 
annual estimates in 12 of the past 15 years. The first 10 years 
of declines were later invalidated by Census 2000, which showed 
we actually experienced significant population gains.'' Despite 
proven flaws in the methodology, the 2001 estimates conflicted 
with the Census 2000 growth by dropping the North Dakota 
population below the 1990 estimate as if the growth in Census 
2000 had not occurred. So there are many individuals and 
groups, organizations, communities, that believe that current 
data is not necessarily reflected in those estimates.
    Mr. Kincannon. Well, I explained face to face with Governor 
Hoeven and with Ms. Strumbeck that the 2001 estimates were not 
changed to accept the results of the 2000 Census, because the 
process really takes longer to feed in. They were corrected for 
the next years and were backward changed for 2001, I believe. 
Just a matter that between December of 1 year and we don't 
have, perhaps we should, but we did not have at that time a 
method of carrying it down to the local level. So that lags by 
a year in getting incorporated.
    Mr. Turner. OK, I am sorry, I am confused. Because I had 
thought that you had just said previously that the Census 
Bureau does, when they get information that conflicts with 
their annual estimates, modify their estimate process to make 
sure that you don't have inaccurate estimates.
    Mr. Kincannon. The new census results are incorporated in 
this estimate process as soon as they can be, which is not 6 
months, but I am sorry, for the year following that. The 2002 
and the whole set of records is corrected.
    I understood you to be speaking of evidence that a city had 
given to us or a town about new housing units or additional 
group quarters or other things that we do not measure directly 
ourselves, and that we do incorporate in the base for the next 
year's estimate.
    Mr. Turner. But the decennial census, you do not?
    Mr. Kincannon. We do, but it takes more than 6 months to 
get it incorporated in there.
    Mr. Turner. I see, it's an issue of time.
    Mr. Kincannon. At least it did in 2000. I will not be here 
in 2010, 2011, to make sure that they are incorporated then, 
but you and Ms. Norton and Mrs. Maloney probably will be here 
and can be vigilant and remind the then-director to make sure 
that preparation is made to take account of that.
    Mr. Turner. When we were putting this hearing together, 
there was a sense that our staff received from local 
communities that the process, the ability to object, to 
challenge, and the time periods for filing a challenge and the 
technical expertise in able to be order to do that may be 
something that is not widely understood by local government 
communities. Can you please explain the census outreach and its 
ability to work with communities as they find that the numbers 
don't reflect what is accurate?
    Mr. Kincannon. Well, I believe I am correct in saying that 
when we transmit estimates for an area to the official 
representative of that jurisdiction, first at the State level, 
that we include first of all, they have reviewed and consulted 
with us about that estimate at the State level, but we still 
include in the materials that we send to them, and they are 
also available on our Web site, not too difficult to find, I 
think, how you can register a challenge or a proposal for new 
evidence to be considered.
    I believe the same thing is true when we publish the other 
estimates that we make clear that there is a process available 
if there is local information that we don't have that could 
improve our estimates. It may be that we need to review those 
lists and undertake a direct mailing to targeted 
representatives in each county and in each incorporated place. 
We do make contacts with representatives of local government at 
other times, and if we are not reaching everyone so that they 
understand this process, I will undertake to do so.
    Mr. Turner. One of the themes that seems constant is the 
conflict between the component method and the housing unit 
method for determination of the estimates. Can you please 
describe that for us and its application?
    Mr. Kincannon. Well, the component method of administrative 
records is based, as its name applies, on administrative 
records to the extent possible. It starts with the preceding 
estimate or census, if that was the last global estimate for 
the local area. It adds births and deaths from vital statistics 
that come largely, but not exclusively, from States and are 
augmented by information from Federal agencies, and are also 
estimated, because there are time lags that are in the vital 
statistics process, the way they're carried out in this 
country, based on local administration of the program.
    As an aside, one lag is introducing new racial categories 
that are standard in the census and have been standard, some of 
them for 20 years, getting those introduced into particularly 
birth records at the local level is difficult. That's not the 
Census Bureau that does that. And it's a difficult problem to 
deal with it, but it is a concern.
    In addition, we make an estimate of internal migration 
between the States, using tax records. That is not perfect, but 
it's a very strong representation and was used in past years 
for very significant allocations of resources when there was a 
revenue sharing program.
    We also make an estimate of net international migration. 
The administrative records on migration, as you know, are 
imperfect in a number of ways. This is the most difficult 
component of our estimate. I am happy to say that with the 
American Community Survey, directly asking people where they 
lived 1 year ago, we will have a new and independent, truly 
independent measure, of where people lived the year before, 
whether in another State or in another country. That should be 
a basis for improving our estimates of international migration.
    Mr. Turner. One more followup question, then I am going to 
turn to Ms. Norton. So knowing though that D.C. and New York, 
two dynamic urban areas, have both had estimates that have not 
accurately reflected their populations, what do you believe 
needs to be done to improve either the methodology or the 
process that you're undertaking?
    Mr. Kincannon. We have held a couple of conferences talking 
about the methodology used. And it seems clear that there's 
more openness now amongst users and our partners in making the 
estimates to considering whether different methods can be used 
for different areas. It has been a hallmark of our practice 
that we need to be consistent and use the same methodology 
applying to D.C. as to Utah. I could argue that case. But if we 
can make better estimates overall, modifying the estimates, the 
methods based on locally available records, we are willing to 
do that. We would proceed prudently on that basis, but we are 
open to doing that.
    Mr. Turner. Is that your recommendation, that you believe 
that should occur?
    Mr. Kincannon. I believe we should examine it thoroughly in 
cooperation with our partners and with users of the data and of 
course with the Congress.
    Mr. Turner. Ms. Norton.
    Ms. Norton. Thank you, Mr. Chairman. Just fascinating.
    I must say that what you just said about D.C. and Utah, I 
am glad you used that example, we have a bill on the floor, Mr. 
Chairman, I wish you would look at, that would give both Utah 
and D.C. a vote. And each would gain by what they learned, by 
what their constituents learn and what people who want to do 
business there learn. So you handed me that plug, sir, no, I'm 
sorry, our witness handed me that plug.
    But therein lies one of the great flaws. I discovered it 
only, sir, when I asked another agency known for the same 
professionalism that has surrounded you since you were created 
by our Constitution. I asked that GAO look at the structural 
imbalance problem in the District of Columbia. And what they 
initially came back with was the comparison between the 
District of Columbia, yes, and the functional equivalent of 
Utah. And you've got the most skewed results.
    So we said, could there be a problem in that there is a 
uniqueness to a city-state that means that the statistical 
model has to reflect that. That is hard to do. But you know 
what? The GAO went back and with a great deal of statistical, 
use of the statistical science and with fresh eyes about the 
huge differences between every State and a state-city, they 
came back with an entirely different result. I just put that on 
the table, and would invite you to speak with the demographers 
and others who worked on this very challenging issue and the 
landmark report they did on the structural imbalance in the 
District of Columbia. Because I think it points, indeed, I have 
huge problems with the statistics that say that D.C., for 
example, which never had any factors of the highest this or 
that disease rate, well, is that really right? Sometimes you 
think, well, you compare city to city, but it still doesn't 
come out right, because it's really not a city.
    Anyway, this is fascinating to a non-statistician who still 
can think analytically and wonder what is going on here. So I 
am really fascinated by the work you are doing.
    Let me just ask you an obvious question. Given what you 
said to the chairman about even the time lag it takes to 
correct, you said, well, you can't do it in 6 months, you spoke 
of a year. Why did it take the District of Columbia 4 years to 
get an acknowledgement from the Census Bureau that its 
estimates, which showed us losing large amounts of population, 
were inaccurate? Why couldn't that have happened earlier? This 
was not a ``lag by a year,'' as you said earlier in your 
testimony. This was a lag by at least 4 years. And it did harm 
to the District of Columbia.
    Mr. Kincannon. First of all, I would like to say I know 
what you mean that it does harm beyond the distribution of 
Federal funds. Attracting business, attracting tourists, 
attracting all kinds of interest and investment in a locality 
depends on reliable statistics about the population. And to 
show a shrinking population can send a message that is not 
consistent with the vitality of the city.
    Shortly, well, I am not sure exactly, but some time in my 
first year after taking office in March 2002, I read a story in 
the Washington Post where Mayor Williams was critical of the 
Census Bureau's latest population estimates. And they were 
words that many a mayor has spoken, some with greater 
justification than others. So I was interested in this, since I 
had the new responsibility, and I lived in a city where I and 
my bosses read that same paper. I asked the estimates folks, 
had D.C. appealed this or had they supported the claim, what 
evidence had been submitted. I was told that D.C. had not been 
participating in the process.
    I may have misunderstood that, but my understanding was 
that D.C. was not participating in the meetings of the Federal-
State cooperative program and had not formally challenged.
    Ms. Norton. So although they challenged for 4 years, they 
really weren't participating in the process? So it is their 
fault that they didn't get these census updates until 4 years 
later?
    Mr. Kincannon. Madam, I don't know whose fault it was.
    Ms. Norton. Well, I need to know that, because if it is 
something that--look, we are not really here to lay blame, but 
in understanding how the responsibility is allocated. We need 
to look to the States and localities to do what they are 
supposed to do, and of course, we look to the Federal 
Government.
    So your testimony seems to be that the District, for 4 
years, was not participating, not that it didn't provide the 
information you needed but were absent, and the mayor was 
complaining but not having his people participate in the 
process you required in order to get an update?
    Mr. Kincannon. I asked that our staff work directly with 
the staff of the Government of the District of Columbia to 
engage them in the process and to see if there was evidence 
that would enable us to correct our estimates if justified. And 
that proved to be beneficial. And we came up with a 
significant----
    Ms. Norton. And 4 years later, I am just trying to find out 
what accounted for 4 years.
    Mr. Kincannon. I don't know what accounted for 4 years. I 
don't know how long it took to get the evidence and how long it 
took to evaluate it. I will provide you with that information 
if I can.
    Ms. Norton. I would appreciate it. Because obviously I want 
to make sure that the District by the next census in fact has 
accurately submitted its own data. So if I don't know, from 
your point of view, what happened, then it is going to be hard 
for me, the Federal official, to say to them, you better make 
sure you do X, Y and Z. Because they are going to be here 
later, and I will ask them. But they will give me their 
version. I need your version of why it took 4 years and your 
testimony here as of now is that it is their fault.
    Now, just let me ask you this. If you don't challenge, 
since they did challenge for 4 years, is there any way to 
assure corrections will be made at all? Here are the big cities 
who, after all, will find the wherewithal. So suppose you are a 
small or mid-level city. And you say, we know that is wrong, 
but when we look at our budget, geez, to get the experts that 
can go up against a Census Bureau, boy, those are expensive, 
even if we get them from our State university. So they say, 
well, we just hope they get it right the next time.
    How are we to sense the purpose of the intercensal 
estimates is to assure that by the time you get to the annual 
census, it will virtually all be right? What happens to the 
great majority of jurisdictions who do not have the 
wherewithal, whatever that might be, to challenge what they 
think are incorrect census figures?
    Mr. Kincannon. Well, first of all, I would not wish to 
characterize the District of Columbia as being at fault in 
this. The practice of participation, to the best of my 
knowledge, had fallen into dis-use. And we revived that use, 
there were new staff members, as I understand it, in the 
District of Columbia. And when they engaged in that process, we 
learned a great deal from it, and were able to make improved 
population estimates.
    I am pleased at that. I don't know if I may be misinformed 
or I certainly don't know how the situation came to pass that 
there was not active participation. But there was no reason for 
their not to be active participation, since we are all living 
in the same metropolitan area, and we can reach out 
particularly to the D.C. Government and ensure that you are 
cooperating with us.
    Ms. Norton. We wouldn't even have to pay transportation to 
get our folks here. So I am really concerned about those of you 
who live a few miles away.
    Let me move on.
    Mr. Kincannon. I did want to say, you asked about smaller 
cities, but 60 percent of our challenges came from cities with 
population, under 50,000 population.
    Ms. Norton. Say that again?
    Mr. Kincannon. Sixty percent of our challenges came from 
cities with a population of 50,000 or less. So cities about a 
size are able to organize their records about housing starts 
and about utility connections and new construction and school 
records or tax records, and make a case to us. They know a lot 
about their local data that we don't know and can't know. So 
smaller cities are managing to do that.
    Ms. Norton. Well, it should be a collaborative process. But 
it sounds like the functional equivalent of get yourself a 
lawyer, file a challenge and if he is good enough, he may 
prevail.
    I was looking at page 9 of your testimony. In any vastly 
important enterprise like the Census, you are going to find the 
Congress looking for who is ultimately accountable, and you are 
going to find Chairman Turner going home and saying to his 
constituents in the Dayton area what they are going to have to 
do. But ultimately, we have to find some objective source to 
indicate who is responsible. The Census is known for its 
objectivity, at least when it comes to accounting.
    But I noted on page 9 that you indicated, and I certainly 
accept this, that you have to rely on information from other 
Federal agencies. Fine. So of course, if you do census, I am 
sure if I came to you with some data, you wouldn't say, thank 
you, Eleanor, let me see if I can incorporate this. I am sure 
you use the Census expertise to get it, so OK, they provided 
it, I want you to run down who those agencies are and tell me 
what you do about them. And then you say, if the documentation 
is provided that supports an alternative estimate, the original 
revised, not because we didn't do it right here, is what the 
chairman referred to the first time, but because, and here is 
the operative language, working with the States and taking into 
account their additional documentation, we can arrive at a more 
accurate estimate.
    So from that I get, who got it wrong were the Federal 
agencies who didn't always provide you with the best 
information. That's No. 1. And No. 2, States who had not 
submitted the additional documentation, but now that they have, 
we can arrive at a more accurate estimate. Who is coordinating 
this process so that up front, without challenges, the 
legendary Census Bureau can assure that in fact the data 
provided by the States is accurate, is responsibility to lay 
nowhere ultimately in this system? Or if it is everybody's job 
then it is nobody's job and nobody needs to feel responsible.
    Is there anybody over this process? Is that the Census 
Bureau or does Congress need to make clear, clearer than the 
Constitution, clearer than the additional statutes, that 
accuracy lies someplace and should we then spell out who it is?
    Mr. Kincannon. If the Congress wishes to make the Census 
Bureau and not local governments accountable for detailed 
knowledge about local construction, utility connections, school 
enrollment and the other, tax return filings and things like 
that, then I think this Congress would have to support some 
kind of a census county agent program with staffing in each 
locality to look at that.
    Ms. Norton. So you don't, even with all your vaunted 
expertise, you do not regard yourself as the ultimate arbiter 
of the accuracy of the information you receive and how it is 
used?
    Mr. Kincannon. No, ma'am. We do regard ourselves as the 
arbiter of information we receive. But if we don't receive it, 
we stand with the estimates we are able to make from----
    Ms. Norton. We are going around in circles. You do receive 
it, you don't accept it, you go 4 years when it comes to the 
District of Columbia before any changes is made.
    Mr. Kincannon. Well, let's see what the records show.
    Ms. Norton. If you are going to tell me, don't think I am 
not going to ask the District of Columbia, you mean 4 years you 
just refused to submit data, even though you were screaming 
that you weren't receiving--it just doesn't make any sense. As 
long as there is a Federal agency charged by the Congress of 
the United States with doing an accurate census that can affect 
everything about your economy, you are going to find the 
Congress laying blame with the agency it funds. Now, that 
agency better find a way, since it is getting our funding, not 
the States, not the localities, to make sure that its peer 
agencies give it the right information and that it gets the 
right information elsewhere. But you cannot rely on Congress to 
say it is a shared responsibility, because there is nothing we 
can do about it. Every Member of Congress is affected by what 
we are discussing here today.
    Let me go on to another question.
    Mr. Turner. I do have some additional questions, I will get 
back to you for a second round if this is an appropriate time.
    Ms. Norton. I have had my 10 minutes, OK.
    Mr. Turner. I will ask a couple of questions, then we will 
go back to Ms. Norton. I don't want to limit the substance of 
her questions. But she is on a topic that I also wanted to jump 
in on. And first off, let me say that in the interest of time, 
I have questions about the Master Address File and the TIGER 
enhancements and the ACS and how it might be able to be used 
and/or is impacted negatively by this process. And also 
questions concerning the funding that the Bureau has received 
to improve its estimates of migration and immigration that I 
will submit for written responses.
    Following on Ms. Norton's theme of the task that you have 
in front of you, I have two things that I want to highlight and 
then I will turn it to you. One of which is, out of all the 
examples we got, this one is the one that was the most striking 
to me. We got a letter from Sugar City, Idaho. Pretty small 
town. And this is the letter that we received. It says, ``I 
have been asked to write this letter concerning the city's need 
to challenge the Census Bureau's 2004 intercensal population 
estimate. In the year 2000, the population of Sugar City was 
1,242. In 2004, the population estimate was 765. After 
challenging the population, it was at 1,448. If we had left the 
estimate in 2004, the city would have only received 53 percent 
of its various State funds,'' so basically 47 percent of the 
overall population of this town was missed. And clearly, this 
is a town that is not going to have the resources that D.C. 
has.
    I wondered if the Census Bureau either does have or should 
have a trip wire, if you will, in instances where the decennial 
census has a number that shows a significant increase or the 
intercensal estimates have been decreasing that trip wire would 
cause an expedited review of the next estimate. And in this 
instance where the estimate would appear that 47 percent of the 
population of this small town was no longer there that the 
Census Bureau would take an affirmative step, not just the step 
of waiting for challenges, but an affirmative step of saying, 
wait a minute, this isn't 5 or 10 percent of a place that has 
large numbers, but this is a large percentage of a population 
that suddenly has gone missing. And then perhaps we should work 
in partnership with those communities. That's the first one.
    The second one is, to acknowledge that it is my 
understanding, and you can correct me if I am wrong, that you 
do over 39,000 estimates with a staff of about 76 and that the 
Bureau has new management in charge of population estimates. 
What I would like to do is if you could tell us about the new 
management team that is in place, what Bureau changes might be 
occurring in the population estimates area, and have you 
considered reorganizing the population division or the 
estimates branches, and if you could comment on the enormity of 
the task that you have with the resources that you have.
    Mr. Kincannon. Well, the first question, should we have a 
trip wire for very small cities with large changes, like Sugar 
City, I will take a look at that and see. I am not familiar 
with the Sugar City case. It does appear that Sugar City 
mustered the resources to challenge it and we accepted that 
challenge.
    I guess we have a lot of more or less self-enforcing 
features in our form of Government, and that depends on both 
the willingness of people to obey certain structures and their 
vigilance in terms of their self-interest. And even very small 
towns are even quite vigilant and effective in asserting their 
self-interest.
    Mr. Turner. Right, but their success, of course, can't be 
considered as proof that therefore the system works. Because 
there are a number of communities who either, by the fact that 
we don't have their data, it doesn't mean that they have the 
expertise and merely desired not to proceed. It is an example 
of a community that clearly had small resources and had a 
significant miss in the estimate.
    So I wouldn't say that the fact that they did go through 
the process shows blanketly across the country that every 
community therefore has the ability to do it. But it clearly 
shows that some communities have a significant miss. And my 
question is, shouldn't that cause something in the Census 
Bureau light bulb department to go off and say, perhaps this is 
something where we should be in greater partnership with the 
community?
    Mr. Kincannon. As I said, we will take a look at the trip 
wire notion and see if it is applicable in a way that we could 
achieve it within our current resources. Congress has not been 
un-generous with its resources on this program, but at the same 
time, to assertively acquire or demand certain information from 
each and every city and town, all 39,400 units, is a very big 
workload and is a burden on those bodies as respondents.
    Maybe you are suggesting we could survey those with a 
certain percentage decline or a certain percentage increase. 
Although we have never had a challenge about an estimate being 
higher than expected, we would have to be vigilant on the other 
side, I suppose, of that, if we had a trip wire.
    But we can take a look at that. I don't believe, myself, 
that we should mount a new data collection program from State 
and local governments in that regard. But if the Congress 
thinks we should, then we will take a look at it.
    Mr. Turner. Since you are aware, as I indicated, that there 
is a new team in place for the population estimates on the 
management side, is there a reorganization that is anticipated, 
or do you not see changes occurring in that division?
    Mr. Kincannon. I think what I believe will occur and hope 
will occur is a thorough reexamination of the methods used in 
collaboration with academic experts, with our partners in the 
Federal-State program and with users to see if there is a 
better way of changing the methodology to produce better 
estimates. And that I hope to have occurring in the new future. 
So we are going to take a good look at it.
    I don't anticipate ex ante a reorganization of staff or 
that sort of thing. We will see what methods may need to be 
changed, and then if organizational changes make sense, that 
would go forward.
    Mr. Turner. Ms. Norton.
    Ms. Norton. Thank you, Mr. Chairman. I appreciate your 
indulgence. I have only two more questions.
    One has to do with a change that is so clear and obvious, 
resulting in such huge miscalculations and over-payment of some 
jurisdictions and under-payment of others that--challenge is 
not what this is about. This is about changes in our country 
regarding incarceration, where because so much of incarceration 
is now out-sourced and therefore private prisons are used, or 
State prisons are used from one part of the country to the 
other, some of the most astounding, perhaps the most astounding 
figures I have seen have come from a paper. I teach a seminar 
at Georgetown where I'm still a tenured professor.
    And a student did a paper on the way in which the census 
counts inmates. It was astounding. It was much worse than 
anything we have discussed here today. What it means is that 
there are rural counties in the United States that are barely 
populated which offer no services, which have the imported 
prison population counted in their numbers. And so though it 
will be easy, especially for the census to figure out a way of 
finding where these inmates come from, instead it uses a kind 
of blanket, easy for the census to do. But taking real funds 
from where they are needed to say, county X, you now have X 
more people than you thought you had, because we count those 
people as residing there. Of course, they don't get any 
services. But somehow their population is increased.
    Is that true, and would you explain that, how the census 
deals with inmates and with the changes in incarceration in our 
country?
    Mr. Kincannon. We deal with prison and jail inmates as we 
deal with residents generally in the community. That is, we 
count them as we have for every census in the place where they 
usually reside most of the time.
    Ms. Norton. And they reside while they are in prison all of 
the time where they are incarcerated.
    Mr. Kincannon. Yes.
    Ms. Norton. You see, this is my problem with the census. If 
I have any problem with the census, it is my skepticism about 
whether the census moves with the times and has updated its 
processes to reflect what is happening in our country. Now, 
would you not say, sir, that it defies the intent of Congress 
that there would be small counties that could register a 300 or 
400 percent increase in population to then have funds 
distributed that could not possibly go to the prison population 
as they do not? And to have other places under-counted?
    I would almost rather have a neutral way of doing it than 
to have funds to go where there are no people or very few 
people. And what the student's paper revealed was astonishing 
to me, because it was not what the District and New York have 
been complaining about, which is, let's sit down and figure 
this out. It was like in your face, anybody knows that this 
small county, which is essentially going out of business, and 
that is why they have a prison in the first place, it doesn't 
have any population. And the census says, that is all right, 
that is the way we have been doing it since 1700 and I will be 
darned if we are going to change.
    Don't we depend upon you to figure out how to keep 
taxpayers' money for food stamps, for Section 8 housing, from 
not going where there are people but where there are prisoners 
who can't possibly under law take advantage of those Federal 
funds? Can't you do something about that?
    Mr. Kincannon. I hope that the Congress does not depend 
upon the Census Bureau to ensure that there is not mis-
allocation of program money. The Congress establishes the 
rules----
    Ms. Norton. It is not program money. This is done on a per 
capita basis.
    Mr. Kincannon. Well, Congress still establishes those 
rules. And if it doesn't want the census numbers to include 
certain portions of them, it can so do.
    Ms. Norton. Sir, this was thoroughly researched. This 
student got a very good mark. This was a matter of your 
regulations. The Congress leaves to you these matters. You are 
the expert agency. I am going to ask you if you would at least 
promise me this. Because this kind of gross, palpable 
overpayment of money that could be used for food stamps and for 
other necessities to jurisdictions where there is not 
population cannot possibly have been the intent of Congress. 
You are not arguing that is the intent of Congress, are you?
    Mr. Kincannon. I am arguing that the Congress established 
the rules for funding and how they incorporate----
    Ms. Norton. No, sir, we----
    Mr. Kincannon [continuing]. Statistics of one order or 
another----
    Ms. Norton. No, sir, this was researched, there was nothing 
in the record to show that Congress had established this. These 
were regulations of the Census Bureau. That is why I am putting 
the question to you.
    Mr. Turner. Ms. Norton, perhaps since we are going to leave 
the record open and give the Director opportunity to answer 
written questions, perhaps this is one that would be a great 
topic to provide additional data, and in a written question 
allow him to answer.
    Ms. Norton. Thank you very much, Mr. Chairman. I don't want 
to go on with this, because I mean, the figures are clear. If 
he wants to indicate that there are some changes in law that we 
should try to get, I would do that. But I would ask you to look 
into the regulations themselves, and if you would not mind 
getting back to me your basis for continuing these huge 
overpayments, resulting of course in underpayments elsewhere. I 
am aware that sometimes it is more difficult to find, or you 
can argue that, where the prisoner really was from.
    But you can designate, just like you designate that 
prisoner is from some place 1,000 miles away, you designate 
that. Because it happens to be there and because your 
regulations say, not the statute, say in the place where you 
spend most of your time, the Congress hasn't said that, so I'm 
asking you as a regulatory matter to look into that, get back 
to me with your sense of why you continue to do that, whether 
you think legislation is necessary.
    Finally, Mr. Chairman, I do want to say that on the 
District of Columbia, and I am going to ask them to provide me 
with what they submitted, they are going to have testimony here 
that talks about an under-count of 572,000 for 2000, a reported 
population count of 572,000 some in 2000. Then showing 
decreases in 2005, they are going to talk about this beginning 
in 1990. What they are discussing is the actual count in 2000 
shows a population of 572,059 with a difference of 53,000 
people. When you consider we are talking about a city with not 
600,000 people in it, you can see the very serious concern we 
have.
    I do promise you this, that I am going to do whatever it 
takes to get the District to do what it must do if you will 
promise that you will make sure that you do what you need to 
do. When the District tells me for example that one of the 
things you weren't doing was counting charter schools, all you 
have to do is read the Washington Post to know that we have the 
largest number of charter schools per capita in the country and 
that they are 15 percent. I mean, your folks reading the 
newspapers could have gotten that one. But they reported a drop 
in public school population because they ignored what was 
happening in charter schools.
    That is what I mean by professionalism. We expect the 
Census Bureau to be able to find those kinds of facts out, 
before it issues estimates. Yes, sir.
    Mr. Turner. And that can be a great segue into our moving 
into the second panel, if that is OK with you.
    Mr. Kincannon, do you have anything you would like to give 
us as a concluding remark, knowing of course that you are going 
to be receiving additional questions in writing from us, and we 
do greatly appreciate your time.
    Mr. Kincannon. I will only say, in reference to Ms. Norton, 
that we did do a report about counting prisoners at the request 
of the Appropriations Subcommittee. And I will include that in 
my response. Thank you.
    Mr. Turner. Thank you very much.
    We will then turn to the second panel, if they would please 
come forward. Thank you, Director Kincannon.
    We would like to welcome our second panel, which includes 
Dr. David Swanson, director of the State Data Center of 
Mississippi; Dr. Joy E. Phillips, associate director of 
Washington, DC Data Center; Mr. Ken Hodges, assistant vice 
president and chief demographer, Claritas Inc.; and Dr. Warren 
A. Brown, research director, Cornell Institute for social and 
economic research, Cornell University.
    We do have a policy that all the witnesses be sworn in. 
Before I proceed with the oath, I want to confirm that on your 
table there is the witness timing light that will give you the 
green light for you to proceed. Yellow means that you will have 
1 minute to conclude your remarks and red is to ask that you 
kindly conclude your remarks as we move forward then to the 
next witness or to the question portion of the hearing.
    Since it is our policy that we do swear in our witnesses, I 
would ask that you please rise and raise your right hands.
    [Witnesses sworn.]
    Mr. Turner. Please let the record show that all witnesses 
have responded in the affirmative. Mr. Swanson, we will begin 
with you.

    STATEMENTS OF DAVID A. SWANSON, PROFESSOR OF SOCIOLOGY, 
  DIRECTOR OF THE CENTER FOR POPULATION STUDIES, CHAIR OF THE 
    DEPARTMENT OF SOCIOLOGY AND ANTHROPOLOGY, UNIVERSITY OF 
 MISSISSIPPI; JOY E. PHILLIPS, PH.D, ASSOCIATE DIRECTOR, STATE 
   DATA CENTER, DISTRICT OF COLUMBIA OFFICE OF PLANNING; KEN 
HODGES, CHIEF DEMOGRAPHER, CLARITAS INC.; AND WARREN A. BROWN, 
    SENIOR RESEARCH ASSOCIATE, DIRECTOR, PROGRAM ON APPLIED 
DEMOGRAPHICS, RESEARCH DIRECTOR, NEW YORK CENSUS RESEARCH DATA 
  CENTER, CORNELL INSTITUTE FOR SOCIAL AND ECONOMIC RESEARCH, 
                       CORNELL UNIVERSITY

                 STATEMENT OF DAVID A. SWANSON

    Mr. Swanson. Mr. Chairman Turner and honorable Members of 
Congress, my name is David A. Swanson. My Ph.D is in sociology 
with a concentration in population studies. I do research in 
the field of applied demography and teach at the University of 
Mississippi where I am a professor of sociology and also serve 
as the chair of the sociology and anthropology departments and 
director for the Center for Population Studies.
    I have been involved in applied demography more than 30 
years. My time as an applied demographer includes nearly 4 
years with the Population Enrollment and Economic Studies 
Division with the Washington State Office of Financial 
Management, where I learned the basics of the craft of applied 
demography by doing city and county local censuses, developing 
city, county, municipal and special area estimates and 
generating population forecasts.
    Drawing on my experience, I am pleased to provide my 
observations on the Census Bureau's estimation program. My 
testimony covers three areas: first, the major challenge faced 
by the Census Bureau in providing timely, accurate and cost-
effective estimates; second, a suggestion for dealing with this 
challenge; and third, the issues presented by my suggestion 
that need to be resolved.
    First, the major challenge. Fueled by the proliferation of 
Federal programs distributing benefits using decennial census 
data, and the knowledge that Federal courts were now willing to 
consider apportionment cases, several lawsuits were filed 
against the Census Bureau following the 1970 census, a practice 
that itself has proliferated over the past 30 years and now 
threatens to move into other areas of the Census Bureau's work, 
such as the annual estimates program.
    The reason for much of this conflict is clear. Hundreds of 
billions of Federal dollars are allocated each decade to States 
and local governments using census counts and intercensal 
estimates, and these funds are allocated in a zero sum fashion. 
The situation will lead to even more litigation and other forms 
of conflicts as the States, cities and counties struggle to get 
their populations counted in the decennial census and estimated 
during the intercensal periods.
    This atmosphere of conflict is, I believe, the major 
challenge facing the Census Bureau's decennial census and 
intercensal estimates programs. Within the Census Bureau, it 
not only serves to foster a ``defensive'' working environment, 
but also takes important resources away from production and 
research activities. As the defensive climate within the Bureau 
hardens, States and local governments feel even more 
frustration in their attempts to work cooperatively with the 
Bureau, and often turn to more confrontational forms of 
communication. This is particularly attractive for local 
governments in States lacking strong demographic centers.
    Second, my suggestion. Breaking with the past, the Census 
Bureau decided to retain and update its Master Address File, 
the MAF, for the 2000 census. The MAF is a critical resource 
for the American Community Survey and its retention facilitates 
the planning and conduct of an accurate and cost-effective 2010 
census.
    Importantly, the continuously updated MAF also represents 
an untapped resource for intercensal estimates and leads 
directly to the potential to have timely, accurate and cost-
effective estimates done using a method that is not only simple 
to apply and explain, but one that offers the potential for a 
meaningful role for States and local governments to play in the 
development of this estimates. What is this method? It is the 
well-known housing unit method. To be successful, however, this 
approach needs a nationwide system of State demographic centers 
that participate in a meaningful partnership with the Census 
Bureau.
    Third, issues that need to be resolved. Turning now to the 
obstacles associated with my proposal for population estimates 
based on the MAF and a set of strong State demographic centers, 
I will begin with the issue of confidentiality. I believe this 
problem is not insurmountable in regard to my proposal. The 
National Research Council has issued recommendations to 
reconcile access and confidentiality. The Census Bureau itself 
has appointed a chief privacy officer and worked to put 
effective procedures in place regarding this reconciliation.
    Another important obstacle is the financial cost of 
developing a national system of State demographic centers, such 
that each State center functions according to accepted 
standards. States need to shoulder a share of these costs. 
After all, it is to their benefit to have high quality State 
demographic centers. As such, I propose that a funding 
mechanism involving Federal-State matching funds be considered.
    What about accuracy? Can the proposed MAF-based population 
estimation system provide accurate data? In a recent report, 
the GAO identified MAF/TIGER problems that needed to be solved 
in order to have a good census in 2010. The first problem is to 
resolve the address issues such as duplication, omission, 
deletion, incorrect locations in the MAF, and second, 
implementing GPS-based geocoding of housing units. These same 
two problems represent sources of error in the proposed MAF-
based system for population estimates. Consequently, if the 
Census Bureau solves these problems, in regard to the 2010 
census, it will do much in regard to the accuracy of the 
proposed MAF-based population estimation system.
    Given the experience gained by the Census Bureau in regard 
to MAF/TIGER and widespread knowledge and use of the housing 
unit method, and the capabilities of the best State demographic 
centers, Alaska, California, Florida, Texas and Washington, for 
example, I believe that the timeliness and accuracy of MAF-
based population estimates based on a comprehensive system of 
State demographic centers, functioning at the level of the best 
State demographic centers, would be sufficient for purposes of 
resource allocation, research, decisionmaking and planning for 
the national, State and local levels. I believe it would also 
prove to be cost-effective and equitable.
    I note that the conflict-free system used in Finland to 
produce annual population data has the type of State and 
national participation and cooperation that I propose. I 
believe that this arrangement goes a long ways toward keeping 
the Finnish system of producing annual population data both 
equitable and conflict-free, even though, as in the case in the 
United States, these data are used to distribute funds and 
other resources to regional and local governments in a zero sum 
fashion.
    With the exception of confidentiality, all the challenges 
facing the development of a national MAF-based population 
estimation system are in the form of cost, technical problems 
or a combination of both. The major technical tasks in building 
and maintaining a MAF-based population estimation system come 
down to two areas: address data collection and updates. The 
feasible way to effect a solution to these problems is to 
enhance the Federal-State cooperative programs already a part 
of the Census Bureau activities such that local entities are 
compensated for helping to maintain the system. There are data 
collection activities in the United States that already follow 
this model, vital registration systems, for one.
    In conclusion, what I am proposing is that the Master 
Address File be more fully exploited by using the housing 
method as a universal means of population estimation for all 
areas of geography, and that State demographic centers be 
developed to uniform level of capability, so that States and 
local governments play a integral part in this system. I 
suggest that this proposal be supported by State-Federal 
matching funds. This would lead not only to timely, accurate 
and cost-effective intercensal estimates, but also to less 
conflict, greater equity and there would be a uniformly higher 
level of demographic human capital in the country. It also 
would be of great use in developing information about the 
extent of large scale natural disasters, such as Hurricane 
Katrina.
    Thank you, I would be happy to answer any questions from 
members of the subcommittee.
    [The prepared statement of Mr. Swanson follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Turner. Thank you. For the rest of the individuals, I 
would ask that you do try to keep it within the 5-minutes 
allocated. We do have your written testimony, which we have had 
the opportunity to review. And the reason I am asking is so 
that we can get to the opportunity for asking questions, which 
of course we expect that you would have the opportunity to 
embellish your answers perhaps with some of the text that you 
are unable to present in your testimony. But we do have it in 
writing and it will be part of the complete record.
    Dr. Phillips.

                  STATEMENT OF JOY E. PHILLIPS

    Ms. Phillips. Good afternoon, Chairman Turner, 
Representative Norton, members of the subcommittee and all in 
attendance. My name is Joy Phillips. I am the Associate 
Director of the State Data Center with the District of 
Columbia's Office of Planning.
    Thank you for the opportunity to be part of this hearing. I 
am here to testify about the District of Columbia's experience 
with the U.S. Census Bureau and its use pertaining to the 
intercensal population estimates.
    My remarks will address the following areas: the challenge 
process, challenge problems, impact of underestimation and 
recommendations for improved estimates.
    The challenge process. The intercensal population estimates 
and its related products released by the Census Bureau each 
year are undoubtedly invaluable pieces of data for various 
levels of government, businesses and of course the public. For 
the District of Columbia, the release by the Census Bureau of 
each intercensal population estimate since 2001 brought much 
concern. Mayor Williams was concerned; so too were other 
District officials, managers and members of the general public. 
In summary, the thinking was that the census got it all wrong.
    Published reports show that over the past 5 years, 
Washington, D.C. has emerged as the strongest and most 
resilient economy in the country, with over $13 billion worth 
of projects completed since 2001, and $7 billion more under 
construction. In essence, the District is in a state of 
economic renaissance with commercial and residential 
construction appearing in almost every neighborhood.
    Against this backdrop, after the Census reported a 
population count of 572,000 in 2000, it continued to report 
population decreases of approximately 4,000 persons each year, 
down to 550,000 in 2005. The pattern was similar to the 1990's. 
The District government was not about to sit silently by for 
another 10 years and experience, among other things, an erosion 
of Federal funding due to incorrect population numbers. With 
approval from the Mayor, the District of Columbia Office of 
Planning decided to challenge the census 2005 population 
estimates by using the housing unit method. Building permits 
and demolition data from 1999 through 2005 were collected and 
analyzed thoroughly. Our calculations resulted in a population 
estimate of 577,900 persons for 2005, some 27,000 persons more 
than the Census estimated.
    To supplement this information, we also submitted 
information on residential utility connections, school 
enrollment, individual tax filers, all pointing to either 
stability or increasing of the population, and not a decrease 
to the magnitude suggested by the Bureau. By now, everyone 
knows that we were successful in our bid to revise these 
estimates.
    However, our successful challenge was not without problems. 
Among our many problems in preparing for the challenge were 
lack of information sharing and labor intensiveness. On lack of 
information sharing, the gist of it was that the Census Bureau 
did not agree to share their list of group quarter entities 
with us at the State Data Center. We were quite surprised, 
given our signed agreements as partners to, among other things, 
collect and disseminate population related data.
    The Bureau invoked the agreements of Title 13, forbidding 
them to share the list of group quarter statistics they used in 
the 2000 census count. Thus, the additional 7,000 persons in 
group quarters in 2005 over 2000 were not included in the 
Census Bureau revision of the estimates.
    The labor intensiveness problem involved the amount of 
resources used up in legwork, new agreements, manipulating 
systems not built for retrieval of the data needed, and 
convincing external agency managers of the critical nature of 
our data needs and so on.
    The impact of underestimation. The impact of 
underestimating the population in any given area goes beyond 
the importance of fair representation of residents in Federal 
and State legislatures. There is also a fiscal impact that is 
quite challenging to quantify. To date, we have not identified 
any national data that exists on the fiscal impact of 
population underestimation on States, cities or local 
governments. However, drawing from an analysis of information 
obtained from a survey of 34 cities, and data gathered from 
various grants awarded to the District of Columbia from 2000 
through 2006, some conclusions could be derived on the fiscal 
impact of population underestimates. Using either scenario, and 
being extremely conservative, the District of Columbia seems to 
have lost at least $5 million of direct funding over the 
period. And this does not take into consideration possible 
funding loss through competitive grants which may turn out to 
be quite substantial.
    The impact of under-estimation in the District is of 
particular concern since it affects resources available to our 
children. According to the Census Bureau, the District lost 
almost 19,000 children between 2000 and 2005, and 
underestimating of children means that a significant number of 
kids that need assistance are not even included in the data 
used to distribute public funds. City governments must 
therefore maintain educational and social service programs to 
serve these children in the absence of financial support from 
Federal agencies.
    How can the Census Bureau improve on the reporting of our 
population estimates? Our recommendations are for more direct 
communication between the Census Bureau and the entities that 
collect and maintain key data items used in the tabulation of 
population estimates. Second, an approach using a compilation 
of population component method and the housing unit method.
    In closing, we urge the Congress to support a more accurate 
population estimate by committing the necessary support and 
funding now and in the future to the Census Bureau. The 
benefits of accurate population counts and estimates are clear.
    Thank you.
    [The prepared statement of Ms. Phillips follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Turner. Thank you, Ms. Phillips.
    Mr. Hodges.

                    STATEMENT OF KEN HODGES

    Mr. Hodges. Thank you. I am a demographer with a company 
that provides demographic information to businesses. For 24 
years, I have produced population estimates and been a user of 
the Census Bureau's estimates.
    Businesses are big users of demographic data, and they 
require estimates for very small geographic areas, such as 
census block groups. Census Bureau does not estimate population 
for block groups, so we do that in the private sector.
    Error can be high for block group estimates. It is about 15 
percent on average. But it often drops to less than 5 percent 
with aggregation to larger areas, such as a trade area around a 
store or a service area for a telecommunications provider. The 
reduction of error based on aggregation depends on the accuracy 
of estimates for larger areas, such as counties and cities. At 
these levels, most private suppliers defer to the Census 
Bureau's estimates, which are produced with resources not 
available to us.
    So businesses have a sizable stake in the accuracy of the 
Census Bureau's population estimates. The consequences of error 
are difficult to quantify, but with things like commercial 
development and media measurement involved, large numbers of 
dollars and even jobs can be at stake.
    Revisions to some estimates are a reminder that they do 
involve error. Some recent revisions have been large enough to 
raise questions about the Census Bureau's estimates program. 
The Census Bureau is aware of these concerns and recently held 
a conference where external experts recommended flexibility in 
the use of new data and methods and collaboration with local 
experts. The Census Bureau is receptive to these 
recommendations and motivated to improve its program.
    Currently, error is addressed through the challenge 
process, in which a revised estimate can be issued for 
government units that provide sufficient evidence of 
documentation for revision. But while revisions can suggest how 
large estimation error can be, one cannot judge the entire 
estimates program based on a few high profile revisions.
    Most users have a rather narrow perspective on the Census 
Bureau's estimates. For example, local governments focus on a 
single estimate, and if they dispute that one number, the 
estimates program is a problem in their view, and very 
understandably so. The challenge process responds to these 
concerns of local governments, but can give the impression of 
an estimates program that is in serious trouble.
    In my work, we use the Census Bureau's estimates for all 
cities, counties and towns and we use them every year. From 
this broader perspective, the estimates have actually served us 
very well. For example, they are the starting point for our own 
county estimates, where back in 1990, we had an average error 
of about 4.1 percent. By 2000, that error was down to about 3.4 
percent. I would love to take all the credit for that 
improvement in accuracy, but most of it goes to the improvement 
in the Census Bureau estimates that we are using as input.
    The continued existence of problem estimates, though, 
points to the need for further improvements. We do need to go 
beyond just the improvements achieved through revisions based 
on challenges. Governments do not challenge to get a lower 
population estimate. All 89 revisions identified now in the 
Census Bureau's Web site are increases over the original 
estimate. So the revisions that we have just now only correct 
for one type of error and certainly not all of those.
    And because the Census Bureau estimates all areas 
nationwide, they face a zero sum constraint and cannot simply 
add population based on local data and input. The Census 
Bureau's estimates have to sum to a realistic national total 
for U.S. population and therefore have to correct for estimates 
that are too low and for estimates that are too high.
    To conclude, some promising resources and new methods are 
being proposed. The Census Bureau is determined to move forward 
with its program. But users need to maintain realistic 
expectations. Even with improved accuracy, the Census Bureau's 
population estimates will still be estimates and subject to 
error. The frequency of challenges and the magnitude of 
revisions may be reduced, but some areas will still dispute 
their population estimates, and there will be continued 
pressure for revisions. This is the imperfect nature of 
population estimates, and from that broader perspective, those 
imperfect estimates will be of continued value to many business 
applications.
    Thank you very much for the opportunity to testify, and I 
look forward to your questions.
    [The prepared statement of Mr. Hodges follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Turner. Thank you so much.
    Dr. Brown.

                  STATEMENT OF WARREN A. BROWN

    Mr. Brown. Chairman Turner, Ms. Norton, thank you for the 
opportunity to be part of this hearing on the Census Bureau's 
population estimates. I direct Cornell's program on applied 
demographics. I am a research director for the New York Census 
Research Data Center, which is a consortium of research 
institutions in the New York Metropolitan Area and upstate New 
York. And I represent New York State in the Federal-State 
cooperative for population estimates. I am a past chair of that 
group's steering committee.
    I appreciate the opportunity to share my observations 
regarding the Census Bureau's population estimates activities, 
and I'm drawing upon a number of years of experience working 
with the folks at the Census Bureau and advocating on behalf of 
communities and counties in New York State as well as aware of 
what is going on in other States around the country.
    My written testimony and a copy of a paper I co-authored 
with Joe Salvo of the New York City Department of City Planning 
for a Census Bureau conference on population estimates has been 
referred to numerous times so far today have been entered into 
the record, or I have submitted them for the record.
    There are two things from my written testimony that I would 
really like to focus on. The first is error, then correcting 
those errors. Population estimates are not perfect. And as a 
decade goes along, typically the errors mount up, and we get 
further and further away. I think it is unrealistic to expect 
an estimates process to nail the subsequent decennial census 
right on the head.
    That said, the Census Bureau has produced a research file 
that allowed their current major methodology to work without 
the benefit of challenges, without the benefit of review from 
the States. It was an opportunity to see how good is the 
administrative records component method. It produced estimates 
for April 1, 2000. The estimates for the Nation were low, they 
were 6.8 million low. That number or that percentage was a 2.4 
percent error. If that error had been equally distributed 
across all of the local areas, the cities, the counties, the 
municipalities, then we might have had error in the census, but 
at least they would have been equitable and they would have 
been evenly distributed. They were not.
    The greatest underestimate was the District of Columbia at 
minus 9.3 percent, Nevada at minus 6.8, Arizona at minus 5.4, 
and Rhode Island at minus 5.1. Clearly that methodology does 
not do a very good job with areas as diverse as the District of 
Columbia, Nevada, Arizona and Rhode Island.
    The challenge process, it has been observed that there are 
a number of challenges that have taken place, and yet of the 
almost 40,000 areas of local government, the fact that there 
were 38 challenges, successful challenges in 2004, it is an 
infinitesimal proportion of all area. Obviously there are 
errors out there.
    When a challenge is successful, what are the benefits? What 
happens? The graph that I have submitted here shows the 
changing nature of State population. This is a State as large 
as New York State. I have expanded the vertical axes to 
accentuate the differences, if you will. But you can see that 
the estimates that were produced in 2005 incorporated the 
challenges that were made in 2004. They incorporated them for 
2004. So if you focus on 2004, you will see that the number is 
substantially larger in the 2005 estimates than it was in the 
2004 estimates. It reflects that adjustment.
    The 2005 estimate was based on the same methodology and 
shows a downturn in the population. As we have talked about the 
perception of growth, change or decline, it is very important, 
and what this tends to show is that we have a declining 
population in New York State. We think the challenge that we 
are currently engaged in will correct that.
    In summary, getting estimates right the first time or at 
least with maximum accuracy, precision and equity, calls for 
fuller partnership between Federal, State and local 
governments. Thank you for inviting me to participate in this 
oversight hearing.
    [The prepared statement of Mr. Brown follows:]
    [GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
    
    Mr. Turner. Thank you. First for questions we will turn to 
Ms. Norton.
    Ms. Norton. Thank you, Mr. Chairman. Let me very quickly 
ask a set of questions to the District of Columbia.
    You say in your testimony, and here I am quoting from page 
2, ``The District of Columbia was not about to sit silently by 
for another 10 years and experience, among other things, an 
erosion of Federal funding.'' And then you say, after release 
of the 2005 population estimates in December 2005, the District 
``was ready to act.''
    Now, you heard me be very tough on Mr. Kincannon.
    Ms. Phillips. Yes, I did. [Laughter.]
    Ms. Norton. And you heard me promise I was going to be 
fair, and be equally tough on the District of Columbia. Perhaps 
you left out something. But I asked him about 4 years and your 
testimony seems to indicate a challenge to a place in 2005. 
Would you care to comment on the disparity there, if any?
    Ms. Phillips. Yes, I can. I can only speak for when I took 
the job as----
    Ms. Norton. Well, that won't do. He couldn't get away with 
that, and I have to tell you, all of us have to take, we can't 
say, the member who was here last messed up.
    Ms. Phillips. Right.
    Ms. Norton. So don't blame me. This happened on the watch 
of the Mayor of the District of Columbia.
    Ms. Phillips. Yes.
    Ms. Norton. He has been a good mayor. And I am asking you, 
let me be specific, when was the first challenge filed to the 
2000 census?
    Ms. Phillips. From my research, the first challenge was 
filed in 2005. When I took the job in February 2005, we began 
looking at the information to file the challenge. When the 2005 
population estimates came out in December, then we got the 
official piece together and the challenge was actually filed, I 
believe, in March 2006.
    Ms. Norton. OK, you know what, was it done with any 
guidance from the Census Bureau?
    Ms. Phillips. It was done with guidance, and that is why I 
am going back to, I can speak to when I came. When I came on 
board, I realized there was a problem. I contacted the Census 
Bureau, and they did engage in discussions with us as to what 
it entails in filing a challenge.
    As you have seen in my stated testimony, we did have that 
discussion at the regional office, the Washington Council of 
Governments. And we were still not pleased with the results of 
that discussion, and that is why we went ahead and did the 
challenge.
    Ms. Norton. I need you to go back and speak with the Mayor. 
When was the Mayor elected?
    Ms. Phillips. Well, the Mayor has----
    Ms. Norton. He's completing an 8-year term. When was he 
elected?
    Ms. Phillips. In 1999.
    Ms. Norton. This is a mayor that has a very good reputation 
in the Congress and he has a very good reputation with me.
    Mr. Turner. I can add to that, he has an excellent 
reputation.
    Ms. Norton. We have seen palpable improvements in the 
government. We are really not in the blame game, we are in the 
making sure that we get everybody doing his or her part. And 
then seeing, because you heard who I am going to hold 
responsible. I don't fund you to do this.
    Ms. Phillips. Right.
    Ms. Norton. We fund Mr. Kincannon.
    Ms. Phillips. That's why I made sure when I came on board--
--
    Ms. Norton. So I need you to go back--I only have so much 
time--I need you to go back and I need you to come into my 
office with people who were there at the time, so that I can 
understand, since the Mayor has been complaining about this 
virtually since he took office, why no challenge was made. And 
the reason I want you to do that really goes to my next 
question, which goes to our various experts.
    Suppose the challenge isn't made and the corrections made. 
However, perhaps the corrections are made without the 
challenge. I wasn't able to find that to be the case. But if 
you have an accumulation of time without corrections, does that 
mean, what does that mean concerning accuracy or inaccuracy of 
the next decennial census? Do any of the experts care to focus 
on this?
    The reason, the predicate for that question is that 
Congress is spending money to have it done on a virtually, 
well, not an annual basis, but to have it done often. And I 
thought the premise was, the more often, the closer to the data 
and the change in the data, the closer you get to accuracy, so 
by the time you get to the 10-year period, you say, well, let's 
look back, when all the witnesses are gone and you get what Ms. 
Phillips just said, well, I wasn't there then, Congresswoman, 
you can then speak with the people who have the responsibility.
    So I am asking you what the net effect is, in your 
judgment, on accuracy, if corrections are not appropriate made 
on a timely basis by whatever, on whatever basis, challenged or 
not?
    Ms. Phillips. From the information we have, we can see that 
it affects numerous programs.
    Ms. Norton. No, please, I'm asking this--they are going to 
call time on me. I am asking this, I am not directing this to 
you, Ms. Phillips, and I understand what the difficulty is. You 
are going to get, you all are going to come in and see me. I 
don't want a letter. You don't have to come far. And I am here 
to try to work with--we are having a new mayor, Mr. Chairman. 
So I want to work with whoever that new mayor is.
    But I want to know from the experts now, the people who are 
outside experts, objective experts, what is the cumulative 
effect if the updates don't occur on an active fashion? Yes, 
sir, Mr. Brown?
    Mr. Brown. Let me point to two areas. The address list that 
has already been described as New York City was very successful 
in getting a more accurate count by vigorously developing its 
own address list to compare against the Census Bureau's address 
list, part of the housing unit method and what David Swanson 
has been referring to as an ongoing registry of housing units 
and those addresses to be used in the estimates process is 
probably the single most important thing you can do to ensure 
an accurate census count.
    So if you will, the estimates are like a dress rehearsal 
for the decennial census.
    Ms. Norton. As often as you can do them?
    Mr. Brown. Get your act together, get your resources 
together and maintain a complete and accurate listing.
    Ms. Norton. So in your judgment, you will have a more 
accurate census at the end of the period if you did that?
    Mr. Brown. Absolutely. And the second, which is almost like 
closing the barn door after the cows have left, is during the 
review process, did we get good coverage, because the States 
participated with the Census Bureau in reviewing the Census 
2000, is to compare the estimates to the count. If the 
estimates are low and the count was low, it is not going to be 
surprising. But if the estimates are high and the count was 
low, then it is going to cause concern that maybe we missed 
something here. So getting that estimate correct is incredibly 
important.
    Ms. Norton. Yes, sir, Mr. Hodges?
    Mr. Hodges. Put another way, with some of David Swanson's 
recommendations for the housing unit method and the use of an 
Master Address File, if one goes that route with the estimates, 
you are then building estimates with sources that are more 
similar to those that contribute to the census count itself. 
Whereas right now, with the component method, as well as it 
works in some areas, it is a very different process than that 
for the census.
    So whether the ultimate census counts themselves are more 
accurate or not, you would have results that are more similar.
    Ms. Norton. Interesting.
    Mr. Swanson. If I could add one other point.
    Ms. Norton. Yes, sir, Mr. Swanson.
    Mr. Swanson. This goes back to my experience in Washington 
State, which started its estimates program in 1942. The benefit 
of having a housing unit method and having a registry for 
housing that I suggested is that I have seen in action, when 
you use housing unit method estimates for cities, and there 
were 265 in Washington State when I worked there, is that when 
you had a dispute with the local government, you had a method 
of self correction. The elements that go into making the 
housing unit method work are exactly the elements you need when 
you do a census count. And the cities would provide the housing 
unit numbers to the State, the State demographic center, as we 
started to develop the estimates for the cities.
    So the disputes usually, they would boil down to two 
issues, what was the vacancy rate and how many persons per 
household were there in each of them. And we could solve those 
easily with the census and/or if it was a larger town where 
census was costly and prohibitive from that regard, for 
example, Seattle, we never did a census in Seattle, but we 
certainly did a lot of sample surveys to update the information 
that we used in the housing unit method.
    So what you get is a system that is self-corrective. In 
that regard, I am in full agreement with what Warren said and 
what Ken said. It is one of the reasons I recommended looking 
into the math, is that if you are using something like that, 
you have a head start for having a pretty clear idea of what 
the census is going to come about with.
    Ms. Norton. And that is a concern of the Congress, the 
accuracy concern is, it is not a member, particularly in the 
House, which is done by reapportionment.
    Mr. Chairman, I just have two questions. One, I really am 
interested in, Mr. Brown, in page whatever it is in your 
testimony. As I say to my students, number my pages, after all, 
I don't count, your grade is not based on the number of pages 
but on the content here.
    You name the jurisdictions that had the greatest 
underestimates, the District of Columbia led. And I kept 
looking for what these jurisdictions might have in common, 
District of Columbia, Nevada, Arizona and Rhode Island, in that 
order. Do they have anything themselves in common, or how would 
you, what would be your professional judgment for the reason?
    Mr. Brown. Nevada and Arizona, rapidly growing, the two 
fastest growing States in the country. So the growth is just 
outpacing our ability to estimate the population.
    Ms. Norton. Oh, well, that is what we pay the statisticians 
for. But I hear you. I hear you, this is what, when I asked Mr. 
Kincannon, I said, are you keeping up with a society that 
changes overnight, and yes, we are going to expect the 
statisticians to be smart enough to figure that out. But I hear 
you. That is not a criticism of you. But go ahead.
    Mr. Brown. The District of Columbia I think is sort of 
encapsulates everything that is a weakness in the 
administrative records component method. College students, 
immigrants from abroad.
    Ms. Norton. Transients who come here to work.
    Mr. Brown. Transient population that is present perhaps in 
Providence, Rhode Island, and since Rhode Island is a small 
State, Providence dominates it. And we know that it is present 
in New York City. So I think you have large urban center 
problems, you have rapidly growing State problems. Not a common 
thread between the four States, but two threads that I would 
lump them into.
    Ms. Norton. Thank you. That is very helpful. This is what 
I'm going to press the Census Bureau to understand. And I 
understand all that you all have been saying about conformity, 
apples to apples. But then the science and the skill is in 
making that happen while acknowledging differences. I think 
that is hard, and I say so from the work the GAO did on the 
District and the structural imbalance. So I don't see this as 
anything but very difficult, but a wonderful challenge, it 
seems to me, for somebody in your profession.
    Could I ask, experts, hey, look, is it almost always an 
underestimate? Do they ever go the other way? I know one time 
they do is when they give people who don't have a ny population 
but have a lot of prisoners a lot of money for Title I, people 
who are not going to school and can't. But I am saying, in the 
general course of things, is the basic rule that if they are 
going to err, they are going to err on the side of what they 
consider caution, which is low, rather than high? Is that how 
it happens?
    Mr. Hodges. Very briefly, there are a great many over-
estimations among those of us who do population estimates. 
Warren Brown mentioned the roughly 6 million----
    Ms. Norton. I am trying to find the frequency. Is the 
frequency with, because I am sure everybody can find somebody. 
Is the frequency with too low rather than too high and how 
often would you rate how often, knowing full well that you 
can't give me a statistically accurate number. Is it rare for 
there to be more population assigned than----
    Mr. Hodges. Let me suggest, without having notes with me, 
that probably varies from year to year. In the 2000 census, the 
estimates were lower than the count. In the 1990 census, if I 
recall correctly, the count came in lower than the estimates. 
So that while with 2000, there more likely would have been 
under-estimates, in 1990, quite the opposite may have been 
true.
    Ms. Norton. It certainly wasn't true for the District of 
Columbia.
    Final question, Mr. Chairman, then I am through. I wonder 
whether particularly the experts have anything to say about 
what Ms. Phillips said. She did indicate that there were 
difficulties with data sharing. And I don't know whether 
changes in statute might be necessary or whether this data 
sharing, for example, on group quarters was a matter of 
privacy, was it statutory, does it need to be changed? I was 
not sure the meaning of that data sharing and why, when the 
Census wants to get a collegial process going, there would be 
that problem.
    Mr. Swanson. I can't speak exactly to the Census Bureau, 
but I know in some cases where I've dealt with population 
estimates, some group quarters populations are extremely 
sensitive. For example, shelters for abused women. So you can 
understand why they don't want to share information.
    Ms. Norton. Just a number?
    Mr. Swanson. Well, I can just say for some types, there are 
issues of sensitivity.
    Ms. Norton. But wait a minute. They are sharing it with 
people in their own jurisdiction.
    Mr. Swanson. I have had people tell me in other States the 
same thing, not the Census Bureau, but in my job as a State 
demographer in other places where people were hesitant to give 
that information to me.
    Ms. Norton. Yes, that is a demographer. I am saying, is 
there any valid reason, perhaps statutory, that you know of why 
the Census would be cautious about sharing information such as 
group quarters?
    Mr. Swanson. Yes. What we do as State demographers in the 
cooperative is we develop a list of every group quarters 
facility in our State. We don't track them all, but we track 
the major ones. We identify where they are located and we give 
them the number of people that are there.
    The information flows to the Census Bureau. Once it becomes 
part----
    Ms. Norton. So you know it, you can go in those places 
yourself, then.
    Mr. Swanson. Absolutely.
    Ms. Norton. So you know what is in there. So what is the 
problem with data sharing, then, about group quarters? You give 
them the information to say, in this homeless shelter there 
are, in this abused, whatever, there are? That is where they 
get it from in the first place.
    Mr. Swanson. The Census Bureau is quite proud of the fact 
that it is a one way relationship, that because of Title 13, 
once data comes into their records and becomes part of their 
information, they cannot share it back.
    Ms. Norton. With the jurisdiction that provided the data?
    Mr. Swanson. That is correct.
    Ms. Norton. Ms. Phillips, is that what they told you?
    Ms. Phillips. Yes, the Title 13, I wanted that list of 
group quarters so that I can compare 2000 with 2005. And they 
could not share that list with me. So when I submitted my list, 
what they were saying is my list was different from the list 
that they had in 2000, so they could not use the updated 
numbers.
    Ms. Norton. Mr. Chairman, I really would ask that matter be 
looked at very--I did not have this question to ask the Census 
Bureau. That is why I asked about privacy and the rest of it. 
But I am astonished that this kind of information, which the 
District of Columbia can walk into any facility, all of them 
have to be licensed, all of them have to comply with the law. I 
don't see the privacy issue. I don't see any--the only thing I 
could see is there could be something statutory. I will look at 
that.
    But other than that, this is the kind of thing, if all I 
know is what you all are telling me, it is what yields 
disrespect for Government. If you say, you provide the data and 
then we challenge the ultimate data and say, we want to 
specifically see how you are counting X data, which we earlier 
provided you, and we say, oh, no, no, we can't let you see 
that, even though we got it from you, that is nonsense. And 
considering the testimony we have received about the importance 
of group quarters, and the housing component of the census, it 
does seem to be a central question of data sharing for the 
Census Bureau that says above all, it wants a collegial 
relationship.
    And I thank you, Mr. Chairman.
    Mr. Turner. Thank you, Ms. Norton.
    For the record, I want to note that my district, which is 
anchored predominantly by Dayton, OH, according to the last 
estimates, lost 4.4 percent, reducing the community to 158,000. 
My district also reaches down toward Cincinnati, and according 
to the Cincinnati Post, they began with this headline that the 
city leads in population loss. It says that in the last 5 
years, it has the highest rate of any major American city as a 
result of recent estimates of losses in population.
    So we have identified the issue of Federal funds. We have 
identified the issue of businesses that locate in the area. But 
there is also the overarching issue of perceptions of a 
community and the impacts that it can have in long term 
strategic planning, based on estimates that may or may not be 
accurate.
    And I am going to conclude, I also want to recognize that 
Charlie Dent of Pennsylvania has joined us, who has been very 
active on this committee and in the issue of the census.
    One of the issues that Director Kincannon seemed 
comfortable with was the ability of communities to object, both 
their knowledge and awareness of the process, their access to 
data and expertise that could result in a successful challenge. 
I would like if each of you would, in conclusion, as a 
conclusionary question, but I am going to give you one 
opportunity also to add anything to the record that you want, 
tell us if you believe that communities have the accurate 
information of the process and the expertise generally to go 
through the processes that New York City and D.C. have been 
successful in.
    And I want to highlight just one other thing that Director 
Kincannon had said, that I think was perhaps not accurate, and 
that is, he concluded that by the fact that there were 
communities that were going through the process that therefore 
it must be available and accessible. We don't necessarily have 
data about, unless you can contribute that in this hearing, of 
communities that have not gone through the process that have 
wanted to.
    But if you would please comment on that.
    Mr. Brown. I would be happy to start. I think the State 
representatives in the Federal-State co-op need to do a better 
job educating their counties and their communities. I think the 
Census Bureau charges us with that responsibility. We want to 
develop a relationship with our counties and our communities 
that, I talk more communities out of challenging that I do into 
challenging, even though New York is one of the most active 
challenge States, that I am sure that my friends at the Census 
Bureau think I go around drumming up challenges, that we get 
communities that are upset. We walk through the data with them 
and say, you don't have any grounds.
    So anyway, I would say that if a community is not, in New 
York State, is not aware, then it is my failing as much if not 
more so, than the Census Bureau's.
    Mr. Turner. Mr. Hodges.
    Mr. Hodges. I would hope that communities are aware of the 
opportunity to challenge and do have equal access to that 
opportunity. But I would be surprised if the ability to 
challenge and the inclination were equal across the board for 
39,000 government units. That would be surprising.
    Mr. Turner. Ms. Phillips.
    Ms. Phillips. I would agree with what the past two speakers 
have said. The information is published on the Census Bureau's 
Web site, in terms of the challenge process. So anyone can 
access it. But I think the problem lies with the resources that 
are available within each one of those jurisdictions. It does 
take a lot of work to get the challenge through.
    Mr. Swanson. The key to answering that problem, I think, 
lies with the State demographic centers. And there is a lot of 
weak State demographic centers. In some States, they don't 
exist. I think what Warren said about New York puts that in a 
totally different situation, where they are working actively 
with the local governments. That wouldn't happen in 
Mississippi. We don't have that kind of activity.
    Texas is a much better situation, because it has the 
capabilities to do that. They are active partners in the 
Federal-State cooperative program for population estimates, and 
they have experienced people there. But Alabama and Mississippi 
do not. And I think that reflects down to the fact that the 
cities are not in a very good position to challenge because 
they lack that kind of a local contact.
    Mr. Turner. For each of you, before we conclude this 
hearing, you have heard both the questions that we have had, 
the testimony of Director Kincannon and the testimony of each 
of you, I want to give you an opportunity, if there is anything 
you want to add to the record before we close. Dr. Swanson, I 
will begin with you, if there is anything that you would like 
to add to the record as a concluding remark.
    Mr. Swanson. The MAF, the MAF, the MAF. I think it is one 
of the keys to having a good estimation system, and I think it 
will tend to resolve lots of problems, including the one I am 
most concerned about, and that is the escalation of conflict 
between the Census Bureau and all these other stakeholders.
    Mr. Turner. Ms. Phillips.
    Ms. Phillips. What I wanted to add is in terms of my first 
recommendation, and that is for more direct communication. I 
refer to it as that trigger that should indicate when there is 
some dramatic increase or decrease in the numbers, that the 
Census needs to reach out to the local community to sort of 
understand what is going on at the grass roots level. I think 
that is quite important. You referred to it as a trip wire. And 
I think if that something is put in place to ensure that when 
these triggers occur, that there is a followup with the local 
community, then we can arrive at better estimates in the 
future. Thank you.
    Mr. Turner. Mr. Hodges.
    Mr. Hodges. Nothing to add. Thank you very much.
    Mr. Turner. Mr. Brown.
    Mr. Brown. In the 1980's, the Census Bureau and individual 
States jointly selected methods most appropriate for that 
State. There was not a one size fits all approach. In the 
1990's, there was pressure to streamline the process, to do it 
more quickly. I would like to see us return to a more full and 
open partnership with the States to find methods and input data 
that work best for the people of that State and do it State by 
State.
    Mr. Turner. I would like to thank each of you for 
participating today, both in the preparation that you have put 
forth in your written comments and also your participation 
here. I appreciate your willingness to share your expertise and 
your thoughts on population estimates.
    I would also like to thank all of my colleagues for their 
participation. In the event that there might be additional 
questions that we did not have time for today, the record shall 
remain open for 2 weeks for submitted questions and answers.
    I want to thank you all, and with that we will be 
adjourned.
    [Whereupon, at 4:02 p.m., the subcommittee was adjourned.]
    [The prepared statement of Hon. Wm. Lacy Clay and 
additional information submitted for the hearing record 
follow:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]

                                 
