[Senate Hearing 110-336]
[From the U.S. Government Publishing Office]



                                                        S. Hrg. 110-336
 
   PREPARING FOR 2010: IS THE CENSUS BUREAU READY FOR THE JOB AHEAD?

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

                                HEARING

                               before the

                FEDERAL FINANCIAL MANAGEMENT, GOVERNMENT
                   INFORMATION, FEDERAL SERVICES, AND
                  INTERNATIONAL SECURITY SUBCOMMITTEE

                                 of the

                              COMMITTEE ON
                         HOMELAND SECURITY AND
                          GOVERNMENTAL AFFAIRS
                          UNITED STATES SENATE


                       ONE HUNDRED TENTH CONGRESS

                             FIRST SESSION

                               __________

                             JULY 17, 2007

                               __________

        Available via http://www.access.gpo.gov/congress/senate

       Printed for the use of the Committee on Homeland Security
                        and Governmental Affairs


                    U.S. GOVERNMENT PRINTING OFFICE
37-360                      WASHINGTON : 2008
_____________________________________________________________________________
For Sale by the Superintendent of Documents, U.S. Government Printing Office
Internet: bookstore.gpo.gov  Phone: toll free (866) 512-1800; (202) 512ï¿½091800  
Fax: (202) 512ï¿½092104 Mail: Stop IDCC, Washington, DC 20402ï¿½090001

        COMMITTEE ON HOMELAND SECURITY AND GOVERNMENTAL AFFAIRS

               JOSEPH I. LIEBERMAN, Connecticut, Chairman
CARL LEVIN, Michigan                 SUSAN M. COLLINS, Maine
DANIEL K. AKAKA, Hawaii              TED STEVENS, Alaska
THOMAS R. CARPER, Delaware           GEORGE V. VOINOVICH, Ohio
MARK L. PRYOR, Arkansas              NORM COLEMAN, Minnesota
MARY L. LANDRIEU, Louisiana          TOM COBURN, Oklahoma
BARACK OBAMA, Illinois               PETE V. DOMENICI, New Mexico
CLAIRE McCASKILL, Missouri           JOHN WARNER, Virginia
JON TESTER, Montana                  JOHN E. SUNUNU, New Hampshire

                  Michael L. Alexander, Staff Director
     Brandon L. Milhorn, Minority Staff Director and Chief Counsel
                  Trina Driessnack Tyrer, Chief Clerk


FEDERAL FINANCIAL MANAGEMENT, GOVERNMENT INFORMATION, FEDERAL SERVICES, 
                AND INTERNATIONAL SECURITY SUBCOMMITTEE

                  THOMAS R. CARPER, Delaware, Chairman
CARL LEVIN, Michigan                 TOM COBURN, Oklahoma
DANIEL K. AKAKA, Hawaii              TED STEVENS, Alaska
BARACK OBAMA, Illinois               GEORGE V. VOINOVICH, Ohio
CLAIRE McCASKILL, Missouri           PETE V. DOMENICI, New Mexico
JON TESTER, Montana                  JOHN E. SUNUNU, New Hampshire

                    John Kilvington, Staff Director
                  Katy French, Minority Staff Director
                       Liz Scranton, Chief Clerk


                            C O N T E N T S

                                 ------                                
Opening statements:
                                                                   Page
    Senator Carper...............................................     1
    Senator Coburn...............................................     9

                               WITNESSES
                         Tuesday, July 17, 2007

Louis I. Kincannon, Director, U.S. Census Bureau.................     3
Mathew J. Scire, Director, Strategic Issues and David A. Powner, 
  Director, Information Technology, U.S. Government 
  Accountability Office..........................................    22
Andrew Reamer, Fellow, Metropolitan Policy Program, The Brookings 
  Institution....................................................    23
Maurice P. McTigue, Vice President of the Mercatus Project, 
  George Mason University........................................    26

                     Alphabetical List of Witnesses

Kincannon, Louis I.:
    Testimony....................................................     3
    Prepared statement...........................................    39
McTigue, Maurice P.:
    Testimony....................................................    26
    Prepared statement...........................................    78
Powner, David A.:
    Testimony....................................................    22
    Prepared joint statement.....................................    43
Reamer, Andrew:
    Testimony....................................................    23
    Prepared statement...........................................    65
Scire, Mathew J.:
    Testimony....................................................    22
    Prepared joint statement.....................................    43

                                APPENDIX

Letter from Senator Don Nickles, dated May 18, 2007, and copy of 
  a June 18, 1982 Congressional Record...........................    82
Charts submitted for the Record by Senator Carper................    84


   PREPARING FOR 2010: IS THE CENSUS BUREAU READY FOR THE JOB AHEAD?

                              ----------                              


                         TUESDAY, JULY 17, 2007

                                   U.S. Senate,    
          Subcommittee on Federal Financial Management,    
                Government Information, Federal Services,  
                                and International Security,
                            of the Committee on Homeland Security  
                                          and Governmental Affairs,
                                                    Washington, DC.
    The Subcommittee met, pursuant to notice, at 2:33 p.m., in 
Room SD-342, Dirksen Senate Office Building, Hon. Thomas R. 
Carper, Chairman of the Subcommittee, presiding.
    Present: Senators Carper and Coburn.

              OPENING STATEMENT OF SENATOR CARPER

    Senator Carper. The hearing will come to order. Welcome, 
everyone. Mr. Kincannon, thank you for being our lead-off 
witness. We are going to be joined shortly by Senator Coburn, 
but I am going to go ahead and start.
    We have a vote scheduled for 2:45, and my hope is that we 
may be able to have opening statements and to actually get 
through your statement, Mr. Kincannon, and then break, go vote, 
and then come back and just grill you for about the next 5 
hours.
    No, I am just kidding. We will not. It will seem like 5 
hours, but it will not be, I promise you. Thanks for coming.
    The hearing today is not our first on the census, but it is 
one that we hope will be the beginning of our efforts to 
exercise effective oversight with respect to the Census 
Bureau's preparations for the 2010 census.
    My thanks to our Ranking Member, Senator Coburn, for his 
commitment to this oversight work and for making it clear that 
the arrival of the latest decennial census does not mean that 
American taxpayers should write out a blank check. And I agree 
with Dr. Coburn on that score.
    The requirement that the Federal Government conduct a 
census every 10 years is enshrined in Article 1, Section 2 of 
our Constitution. It is something that we have to do and we 
have to do it right. Innumerable programs at all levels of 
government depend on an accurate census, as does the work of a 
number of academics and others out side of government. The 
make-up of the U.S. House of Representatives also depends on 
its outcome--except in States like Delaware where we only have 
one Representative. But in a lot of other States, especially 
States like California, where I think they have 53, it is real 
important.
    So I would count myself among those who would tell the 
Census Bureau to do what they need to do to get it right. But 
getting it right should not be an excuse to break the bank.
    According to data provided by the Census Bureau and listed 
on the chart that we have on display.\1\ We can look at the 
cost of conducting the census all the way back to 1790, and we 
find that the cost of the first census was about a penny per 
capita. For 2010, we are looking for the cost of that census to 
be right around $11 billion, and the per capita cost of 
conducting the count will surge to something like $36.
---------------------------------------------------------------------------
    \1\ Charts submitted for the Record appears in the Appendix on page 
84.
---------------------------------------------------------------------------
    Now, I will grant that counting every man, woman, and child 
in our country is a bigger and more complicated task in 2010 
than it was in 1790--or a bigger task than it was even in 2000 
or 1990. But with the advances in technology that we have had 
over time, I personally do not understand why the price tag for 
the 2010 count is so high. Maybe we will get some answers to 
that today.
    I am also concerned that the price tag could go higher at 
the end of the day, perhaps significantly higher. I think that 
was the case with the 2000 census, so we look with some concern 
at the estimate for 2010 as a result.
    The cost of the 2010 census is now projected, as I said, at 
about $11.5 billion. GAO has said, however, that this number 
may be based on outdated projections that do not take into 
account the results of testing that is currently ongoing.
    I am also concerned that the handheld computers that census 
takers will be using to count these households that do not 
return their census forms right away have not always worked as 
well as expected. These computers are a big part of the Census 
Bureau's projected cost savings this time around. If they do 
not work as well as they should, I can see us spending more 
money than we planned between now and 2010 on staff, on paper, 
and on office space.
    Finally, we will hear from GAO today that some key systems 
that the Census Bureau will be heavily relying on in the coming 
years are not being tested now during the so-called dress 
rehearsal that is traditionally used to troubleshoot before the 
decennial census actually begins.
    I mention all of this because it sounds a lot like what 
happened 10 years ago. The cost of the 2000 census ultimately 
hit $6.5 billion; that was 30 percent higher than originally 
projected, according to GAO. This increase was due in part to 
some of the same kinds of problems that we see today as the 
preparations for 2010 ramp up. I believe we need to work hard 
in the coming months to ensure that the mistakes and cost 
overruns of the past are not repeated this time around. And I 
suspect that most of us in this room and on this panel agree 
with that.
    Dr. Coburn is apparently at another meeting, and we will 
offer him the opportunity, once he arrives, to offer whatever 
opening statement he wishes to make. But rather than to delay, 
why don't we just go ahead and I am going to ask our first 
witness to just hold your horses for just a moment because I 
want to give you a little bit of an introduction here.
    Mr. Kincannon was confirmed in his current job in March 
2002, a little over 5 years ago. He began his career as a 
statistician at the Census Bureau in 1963--at the age of 4.
    [Laughter.]
    In 1963, after graduating from the University of Texas in 
Austin. He held a number of positions in the Census Bureau 
before leaving in 1975 during the Ford Administration to join 
the staff of OMB, where he worked on statistical and regulatory 
policy. He also served as the statistical liaison to Vice 
President Nelson Rockefeller's office.
    Mr. Kincannon returned to the Census Bureau in September 
1981. He was appointed Deputy Director and Chief Operating 
Officer in January 1982 by President Reagan's first Director of 
the Census Bureau, Bruce Chapman. Mr. Kincannon has served as 
Deputy Director to John Keane in the Reagan Administration and 
Barbara Everitt Bryant in the George H.W. Bush Administration.
    Mr. Kincannon, you have probably a longer bio than almost 
anybody I have ever introduced. This is pretty impressive.
    Mr. Kincannon also served as Acting Director of the Census 
Bureau from July 1983 to March 1984 and again from January to 
December 1989, during which time he directed the final 
preparation for the 1990 census. So you have had a chance to do 
this before.
    In October 1992, Mr. Kincannon was appointed as the first 
chief statistician in the Organization for Economic Cooperation 
and Development in Paris. That sounds like a pretty good job. 
He coordinated the Organization's statistical programs and 
advised the OECD Secretary General on statistical policy and he 
left this post in June 2000 to return to the United States.
    I might add he is one of the few witnesses we have ever had 
before this panel who knows where Flower Bluff, Texas, is, 
which is where I lived when I was stationed in the Navy at 
Corpus Christi Naval Air Station.
    Mr. Kincannon, we are delighted that you are here. We look 
forward to your testimony and the opportunity to ask some 
questions. You may proceed. Your entire statement will be 
entered in the record, and I will ask you to summarize as you 
deem appropriate.

   TESTIMONY OF LOUIS I. KINCANNON,\1\ DIRECTOR, U.S. CENSUS 
                             BUREAU

    Mr. Kincannon. Thank you very much, Mr. Chairman. It is a 
pleasure to be here, and I am sorry. I thought you got the 
concise C.V., and I could have shortened your time a little bit 
by leaving out some of the repeat kind of assignments.
---------------------------------------------------------------------------
    \1\ The prepared statement of Mr. Kincannon appears in the Appendix 
on page 39.
---------------------------------------------------------------------------
    Thank you for the opportunity to discuss the progress of 
the 2010 census, the reengineered decennial census. Census Day 
is now less than 3 years away. As we look forward, we should 
note that the success of the short-form census in 2010 also 
depends on the success of all other components of the 
reengineered decennial census program. And even though Delaware 
does not have to go through redistricting, there are a lot of 
people in Delaware, State government and businesses alike, who 
are hungering after the detailed data from the ACS and the 
update for benchmarks from the census.
    The Census Bureau's overall request for discretionary 
funding in 2008 totals $1.2 billion. The request for $797 
million for the decennial census, our highest priority, 
accounts for nearly two-thirds of the budget. The overall cost 
of the decennial census, its life-cycle costs, is $11.5 
billion. That has changed a little bit from that chart, but it 
is the right ballpark. And that includes the cost of the annual 
American Community Survey and the MAF/TIGER Enhancements 
Program, both key to a successful short-form-only census in 
2010.
    This figure represents a slight saving to the American 
taxpayer based on if we had started out on the pathway of 
repeating the 2000 census with the same methods, almost the 
same cost envelope, but somewhat different, and giving 10 times 
the information coming from the American Community Survey. 
Furthermore, at this point of the decade, if we were forced to 
replicate the design of the 2010 census, it would cost us $1.4 
billion more than the current decennial program that includes 
the American Community Survey.
    To examine the progress we have made, it may be useful to 
describe briefly the status of key activities, including the 
MAF/TIGER Enhancements Program.
    The MAF/TIGER Enhancements Program is a multi-year effort 
to collect and correct the locations of streets and other 
geographic information. We are working with the Harris 
Corporation to realign street centerlines for every one of the 
3,232 counties in this country. This initiative is on schedule 
and within budget. The Census Bureau's budget request for next 
year includes $59 million to complete the final 367 counties in 
time to conduct Address Canvassing Operations, which is the 
first major field activity nationwide for the decennial census.
    During this operation, listers will canvass blocks and 
conduct brief interviews to verify or update address 
information against the address information on the Census 
Bureau's lists and maps, including the information provided by 
tribal, State, and local governments as part of the Local 
Update of Census Addresses (LUCA), program. The LUCA program 
provides every tribal, State, and local government the 
opportunity to review our address list and to submit either 
corrections or additions. It is the most important single role 
that State and local governments can play in improving the 
results of the census in their areas.
    The accuracy of the census address list and the map are 
vital because the census must fulfill two principal 
requirements: To count every person living in America, once and 
only once, we hope, and to count every person at the correct 
address because the statistics are only useful in their detail, 
not in their totality. Therefore, the accuracy and ultimate 
success of the census--our constitutional obligation--depend 
upon the accuracy of the MAF and TIGER systems.
    Our plans for 2008 demonstrate our commitment to achieving 
this constitutional responsibility, and we are requesting $551 
million to sustain the continuing activities associated with 
the short-term census, including the 2008 dress rehearsal.
    The sites for the dress rehearsal are in San Joaquin 
County, California, and Fayetteville and surrounding counties 
in eastern North Carolina near Fort Bragg. In April, we opened 
Local Census Offices (LCOs), in both locations and started 
hiring approximately 1,300 people in preparation for address 
canvassing, which began in May.
    The dress rehearsal is our last opportunity before the 
census to ensure planned procedures and operations will 
function as designed once they are integrated. While it is 
still too early to evaluate the dress rehearsal, we completed 
the address canvassing operation on June 26, 2007, on schedule 
and can report its success, as well as some challenges with the 
software on the handhelds used for this. However, it is 
important to note that these challenges are being addressed and 
corrected and do not pose serious challenges to the use of 
handheld computers in the 2010 census.
    We are pleased in general with the performance of the 
handheld computers whose overall durability and usability were 
affirmed during the address canvassing operation. In fact, out 
of almost 1,400 handheld computers, only five had problems out 
of the box, and two were dropped and broken in the course of 
the exercise, and all of those were replaced under warranty.
    Based on the dress rehearsal experience, as well as our 
ongoing planning efforts, we are confident that we can and will 
effectively implement the use of handheld computers for the 
2010 census as well as other planned improvements.
    Finally, we will implement the planned improvements we have 
tested throughout the decade through the short-form-only 
census, ranging from improved questionnaire content to a 
replacement second mailing, which could well increase the 
census response rates as much as 7 to 10 percent, and 
dramatically, therefore, increase the efficiency of our field 
operations.
    We believe these planned and tested improvements are 
vitally important to the accuracy of the 2010 census.
    Thank you for your support in the past and in the future. I 
will be happy to answer questions when the time comes. Thank 
you, sir.
    Senator Carper. The time has come. Let us just start off by 
going back almost 220 years, and we mentioned earlier that the 
cost of the 1790 census was, I think, about a penny per person. 
And I realize we are a far different country today and we are 
looking for a different kind of information. But why did it 
cost so little?
    Mr. Kincannon. Well, I am not sure, Mr. Chairman. I know 
the $6.4 billion in the 2000 census was in 2000 dollars, and if 
the 1790 census is in 1790 dollars, then it is hard to make a 
good comparison. But a penny went a good deal farther, I think. 
I do not know what the marshals were paid and gasoline costs 
were extremely modest.
    In all seriousness, we did not really collect as much 
information. We only collected the name of the householder and 
the number of other people in the household, by free and slave. 
There was very little detail on that questionnaire, and that 
makes it easier.
    In addition, we were not concerned about confidentiality in 
those days, and the results were posted in local areas to see 
if anyone had been missed. So it was a collaborative effort. 
There were not so many people, apparently, as we perceive today 
less enthusiastic about being reported to the government. And 
we used U.S. Marshals to collect the census results, which may 
have added a more urgent tone to their visits.
    Senator Carper. OK. Correct me if I am wrong, but let us go 
back to 2000, and I think I said in my opening statement that 
there was a cost estimate for the census in 2000, and as it 
turned out--and you may have still been in your previous job 
over in Paris at the time. But I think I indicated that the 
actual cost of the census exceeded the forecast by some 20 or 
30 percent. And that sort of has us uneasy, looking to the 2010 
census.
    Just go back with us in time to 2000 and tell us what 
happened. I can understand missing the estimate by 2 percent, 4 
percent, 6 percent, but not by 20 or 30 percent. What happened?
    Mr. Kincannon. Well, you are correct. I was still in Paris 
at that time, and I remember reading in the last 1990s, even in 
the Herald Tribune--it made news, the change, the revision in 
cost estimates. The principal reason, I believe, was a court 
decision in 1999 which required the Census Bureau to redesign 
major elements of the plans for the enumeration in 2000. That 
was very costly----
    Senator Carper. Wait a minute. Can you just sort of flesh 
that out for us a little bit?
    Mr. Kincannon. The plan was to conduct something--I believe 
it was called a ``single-number census,'' and that required 
using a sample non-response follow-up in order to estimate 
those still outstanding. That would compress the amount of time 
needed and, of course, save costs.
    A lawsuit was brought. I believe it was by one of the 
political parties, but I don't remember which. I was not paying 
such close attention to it at that time. At any rate, the 
Supreme Court said that sample-based figures could not be used 
in the enumeration in order to apportion seats of Congress. 
This was based on a law passed, I believe, in 1975 that was to 
have facilitated a mid-decade census, really a sample exercise, 
and the Congress at that time did not want to undergo 
reapportionment throughout the decade, so they put a 
prohibition specifically against using sample-based figures to 
do that.
    Senator Carper. Now, let us fast forward to today and then 
on to 2010. Given what happened to the costs that ultimately 
were incurred on the 2000 census as compared to what was 
anticipated, what costs are we looking at for 2010? And can you 
give us some comfort as to why we should not be concerned that 
those costs might be exceeded by some substantial amount 2 
years hence, 3 years hence?
    Mr. Kincannon. Well, I do not see that in the forecast, and 
I would be very concerned if I did, and I would tell you that. 
Of course, I will not be held accountable, and you, with the 
good will of the people of Delaware, will be around to try to 
understand that and explain it.
    There is a big difference in this past decade in that we 
were strongly advised following the 2000 census by the GAO and 
the Inspector General to plan and test the operations we were 
going to use in the census, to decide on that plan as early as 
we could, to test it at various phases, to incorporate the 
findings from tests in a revised plan for the census, test it 
again and so on until we got to the census time and had a 
thoroughly tested plan.
    We were, by and large, more successful in doing that than 
certainly in the 2000 census cycle, and my own recollection of 
the 1990 census cycle was that we were less successful than in 
this decade.
    As a result, we are going forward with a plan for 2010 that 
has been pretty carefully tested and evaluated. The plans for 
use of technology have proven very successful. We have just 
used the handheld computers in the address canvass part of the 
dress rehearsal for 2010, and they worked very well. So we are 
quite confident that our plans for questionnaire design, for 
short-form-only census, and for the automation that we have 
planned will work well. And I think the variation is probably 
at the lower or middle limit of the range that you said you 
found tolerable. Of course, that is uncertain but that is what 
I see.
    Senator Carper. Talk to us a little bit about these 
handheld computers. A good deal has been said about them, and 
recently we have heard--I think in the media this week--some 
cause for concern. What do they do? Just in simple terms, what 
do these computers do for us?
    Mr. Kincannon. These computers do a great deal.
    Senator Carper. And give us some idea about what they cost, 
if you would.
    Mr. Kincannon. I think they cost about $400 for each one. 
We will buy a half a million or more of those.
    Senator Carper. What do we do with them when we have 
finished with them? Sell them on eBay?
    Mr. Kincannon. Well, they will not have any data in them. 
The data are encapsulated and separated and destroyed. So we 
could sell them on eBay or maybe GSA would have to sell them on 
eBay. But we may find other things to do with them. I am not 
aware of plans. That has not been my focus. It should be a 
focus of somebody, but not me.
    Senator Carper. I ask the question only half seriously. But 
is it possible that other countries facing a census of their 
own might want to buy the computers from us without the 
information they have collected?
    Mr. Kincannon. It is possible. The experience of the 
contractor that did the DRIS contract work for us in 2000 went 
on to do that same kind of work in at least two other 
countries, and the same contractor won the award for us this 
time. So whether this breakthrough will prove appealing to 
other countries is a question that I cannot answer, but I am 
sure that the contractor's business agents are considering 
whether there is an after-market for their skills and 
equipment.
    What these handheld computers will do is to collect the 
information from households that have not returned their census 
questionnaire. They will do much more than that, though, 
because they will be used in the address canvass. They will 
receive the maps and address lists that we have, their work 
assignments for the day, by wireless signal or land signal, 
depending on the part of the country where they are working. 
They will carry out that work and return their updated changes 
that same way.
    They will use the handheld computers to convey their 
information about hours worked and units of work completed and 
their travel, and that will be the basis for calculating their 
weekly compensation.
    The same thing when they start out on non-response follow-
up, they will receive their assignments on the handheld 
computers. Those assignments will be grouped in a way that is 
orderly for them to follow geographically, and they will be 
updated on a continual basis, based on late receipts in the 
office. This is a major cost-saving effect because when we are 
doing it with paper, there is a big gap between when we have to 
shut down and print an assignment to go to all the non-response 
enumerators in the field, and a number of late questionnaires 
are received. And this costs us about $75 million for every one 
percentage point that we follow up on when we did not have to, 
and it irritates a lot of citizens as well.
    When they collect their information, it will be each day, 
or even a part of a day, relayed back to the data processing 
center by wireless, all encrypted and protected properly, or 
over land line if they are in a part of the country without 
cell phone service.
    As soon as this receipt is verified as complete, then the 
data remaining on the handheld will be blanked out and will not 
be susceptible to somebody intercepting it or using it in some 
way.
    So it handles guidance to where they are going, their 
assignments both for address canvassing and for non-response 
follow-up, and payrolling and other administrative work--all 
handled by paper before.
    Senator Carper. Alright. Thank you. I want to focus a bit 
on the dress rehearsal that is underway. Did you say in North 
Carolina and in California? Or is it in several counties of 
each of those States?
    Mr. Kincannon. It is in San Joaquin, California, one county 
in California, and in several counties around Fayetteville, 
North Carolina.
    Senator Carper. Roughly how many people are involved in the 
population that is being serviced there?
    Mr. Kincannon. I do not know. Tens of thousands.
    Senator Carper. OK. Fair enough. Thank you.
    Staying for a moment on the dress rehearsal, so far you 
spoke a little bit to this, but let me ask you to come back and 
I will ask you this directly. What problems are you running 
into so far during the dress rehearsal? How do you plan to 
manage your risks going into 2010, especially since some of 
your systems may not be tested as rigorously or in as timely a 
manner as you might initially have hoped?
    Mr. Kincannon. Well, first, let me go through both 
successes and problems in the dress rehearsal. We completed the 
address canvassing operation on time. We started on time and we 
ended on time. The performance of the handhelds, I have 
mentioned that they were durable and so forth and so on. They 
were physically very good.
    The biometric identification and replacing passwords--that 
is, a fingerprint is used for the enumerator to open and access 
the services of the handheld--that worked very well. In the 
2006 test, done with a handheld that required a password, 40 
percent of the Help Desk calls were to unlock a forgotten 
password. We did not have that incidence. We had a very small 
percentage of people where there was a malfunction. It was less 
than 2 percent, if I recall correctly. So that works very well, 
and except under strange circumstances you cannot leave your 
fingerprints at home. So that is a great gain.
    We did have some software problems, operating software 
problems with the handheld relating to particularly the 
capacity for handling a large number of addresses where we 
perhaps had not explained carefully enough to the contractor 
how wide the scope of addresses could be in a single day's 
assignment. That is going to have to be modified.
    There were other software problems as well, some of them 
remedied by transmitting patches to the handhelds in the field, 
and others that will require some more detailed changes before 
we go out for the data collection in the dress rehearsal even.
    So the significance of that is it is relatively easy to 
make software corrections from a central point and apply it to 
all users. If we had significant hardware problems, it is a 
much more difficult problem to solve. So I think that--I do not 
like to have any problems, but I am glad we found them in the 
dress rehearsal, and we will meet them.
    Another problem we had was in the Help Desk function, which 
is also handled by the contractor.
    Senator Carper. Mr. Kincannon, I am going to ask you to 
hold up for a moment, if you will, and we have less than 5 
minutes to go on this vote. I am not as fast as I used to be, 
so I need to head over to the floor quickly to vote.
    I would say to our staffs, if Dr. Coburn arrives while I am 
away, he is welcome to reconvene the hearing and offer whatever 
statement he wants and to begin to ask questions, and I will 
finish up on this question with you when I return.
    But for now, let's just stand in recess until either Dr. 
Coburn or I return. Thank you very much.
    [Recess.]

              OPENING STATEMENT OF SENATOR COBURN

    Senator Coburn [presiding]. In the hope for some efficiency 
in Congress, which is rare, we will start again, and I will 
hold the gavel until Senator Carper comes back. I appreciate 
you, Mr. Kincannon, coming before our Subcommittee again. We 
have expressed to you some serious concerns and reservations 
that we have. That is not to belittle or demean anybody that 
works for your agency or you. We just have realistic concerns.
    You have a trial run starting next year, and I still think 
we have 2\1/2\ years until you are really into this full thing, 
and my hope is that we help make the appropriate--or ask the 
appropriate questions so that we can be as successful as we 
need to be concerning the census.
    I am deeply worried about this new cost estimate, which is 
$200 million more than what we had the last time we had this 
conversation, and I am worried that is an underestimate. I hope 
you can reassure me today that it is not. That figures out, I 
believe--and you can correct me if I am wrong--to about $90 a 
household in this country for doing the census. I think that is 
on individuals rather than per household, which to me seems 
extremely high.
    I also continue to believe now, with almost 70 percent of 
the American people online, 60 percent paying their taxes 
online, 50 percent banking online, that we are missing a great 
opportunity in terms of not doing some type of online census. 
And I understand your position on that, and although I disagree 
with it, I take your position. And I know that you are trying 
to use some technology in terms of handheld devices that are 
going to be doing it. My hope is that they work very well and 
that the $90 is not anywhere close to what we think it was 
going to be.
    I guess the other thing is I would just like to hear from 
you on the problems that you see in front of you and how they 
have changed since the last time we had the hearing and what 
you think needs to happen between now and 2008 when you do your 
test runs and where we can be of assistance, either in terms of 
appropriations or in terms of oversight.
    So with that, I will let you comment.
    Mr. Kincannon. Well, maybe this is the statement I really 
wish I had made but did not have a chance to with all the 
clearance process and so on. I would say I appreciate your 
view. I know you disagree with us about the view of the 
Internet. We are cautious about change at this stage and have 
concerns that I think are well based. I hope you keep raising 
it in the future, including for my successor and for out-years. 
But I think that we have made the right decision.
    What do I see as problems ahead? I recognize that you and 
other Members of Congress have not been as closely associated 
with the process of testing and evaluating as I have been over 
the last 5 years and some months. And no wonder you do not feel 
as comfortable as I do because of the experience in 2010 with 
the sudden, sharp increase in cost. But I do feel comfortable, 
seeing how we have planned, tested, evaluated, modified, and 
moved forward, that we have a process that is working well.
    Two big dangers are significant changes in methodology at 
this stage. This is what caused the big 30-percent increase in 
cost of the 2000 census because of a court decision that ruled 
the Census Bureau could not use sample-based non-response 
follow-up for figures that would be used for apportionment. 
Then a different and significant change had to be made, and it 
cost about $1.7 billion.
    A late change is very inefficient and costly, and I know it 
is frustrating to Members of Congress when some of them may not 
have been elected when we were formulating these basic plans, 
and they see something that they think would be an improvement, 
and it might be, and we do not want to make that change because 
we cannot keep it under control.
    A second danger that I think is quite important and it has 
been a real problem in the past years, and that is the 
continuing resolution. Continuing resolutions present a problem 
for programs that are irregular in the government. If you have 
a constantly funded program, a continuing resolution keeps that 
work going on very well. We have a lot of that kind of work at 
the Census Bureau. But the census, neither the economic census, 
which will be taken covering this year, nor the census of 
population and housing are evenly funded. So that when we are 
going up the scale, failing to get an appropriation for 2 or 3 
months or more is quite destructive. If you miss 3 months' 
work, 3 months' hiring, you cannot hire 25 percent more people 
than you plan, or actually a third in that case, and catch up 
and then fire the people at the end of the fiscal year or 
something of that sort. So it is just lost time, and that would 
be a major concern on my part.
    I am not concerned--our experience in the dress rehearsal 
which began this past spring, when we were conducting the 
address canvass portion of that, convinced me that our 
handhelds are working. There are problems in the software, but 
not with the hardware. And the problems with the software, some 
were remedied by patches transmitted to the handhelds, which 
worked very well. Others will be remedied before we start the 
non-response follow-up phase next spring.
    The handhelds themselves, out of about 1,400, only five 
were flawed coming out of the box, and two more were damaged in 
the course of rather rugged operation in the field. And those 
are tolerable levels of shortcoming. I am very pleased with 
that and feel quite confident about the technology moving into 
the future.
    Senator Coburn. We are, what economists would say, at full 
employment with a fairly low unemployment rate, and I know as 
you gear up for the census, you are going to be hiring a 
significant additional number of people. Have you all 
anticipated the degree of difficulty that you will have now 
versus 2000 in terms of the difference in terms of employment 
levels and underemployment that might not be out there today 
that you utilized in 2000?
    Mr. Kincannon. The employment levels were pretty high in 
2000, actually, and if I could tell you now what they would be 
in 2010, I would be probably making some money on----
    Senator Coburn. Have you anticipated that it is going to be 
much more difficult to find part-time work and full-time 
workers for the census in 2010 given the employment level that 
we have today?
    Mr. Kincannon. I am not sure that it will be more 
difficult, but we have flexibility in setting wage rates by 
local area, which helps us respond to that difficulty. There 
are large areas of the country where the labor market is not 
tight.
    Senator Coburn. Where is that?
    Mr. Kincannon. In some of the Midwestern States, the 
employment is not particularly tight. There are people looking 
for work and available for work, and there are other areas of 
the country as well. But labor markets are very localized----
    Senator Coburn. I do not want to debate that issue with 
you. The fact is that we are at an all-time high employment, we 
are at an all-time low unemployment statistics. We are at an 
all-time low in terms of underemployment statistics. If that 
persists, will you have time and will that change your cost 
numbers significantly if you are going to have to pay a 
significant increase over what you might think today? Will that 
significantly change this $11.2 billion or $11.4 billion?
    Mr. Kincannon. It is important also to factor in the fact 
that the baby-boom generation is coming to retirement now and 
will be in many cases quite interested in short-term temporary 
employment, and a lot of them are very well qualified, not just 
to use the handhelds, which actually have seemed quite usable 
by people that are not particularly trained in technology.
    I am not particularly worried about that. We have to keep 
our eye on it.
    Senator Coburn. Do you have a planned strategy in place in 
the department now if, in fact, you were to run into those type 
of problems?
    Mr. Kincannon. Yes. We are prepared to recruit more than a 
million workers to meet our half a million required workforce.
    Senator Coburn. Alright. And have you tested the phone 
system yet?
    Mr. Kincannon. Tested the phone system? I am not sure I 
understand.
    Senator Coburn. In terms of the census and how you are 
going to do that.
    Mr. Kincannon. Certain phone functions will be the same as 
we have used in the past, the caddy interviewing of people who 
telephone in and want to report on the phone to us. We will 
take those down with caddy type reporting that we use month in 
and month out and have used even in recent censuses. So those 
have not been tested. Again, they are based on technology that 
is proven.
    The interactive voice response method has been tested. It 
is not being tested in the dress rehearsal, but it has been 
tested in other means, and we are prepared, I think, to go with 
the telephone systems that we have, the voice-based telephone 
systems. And, of course, the telephone systems used to relay 
the information, the encrypted information, from the handheld 
computers have been tested.
    Senator Coburn. Well, the GAO had commented that certain 
parts of the phone system have not yet been tested. That was 
the purpose for my question based on their testimony.
    Mr. Kincannon. I think they may have been referring to 
testing in the dress rehearsal, but I am not sure.
    Senator Coburn. Mr. Chairman, I would like to put into the 
record a letter from former Senator Don Nickles dated May 18, 
2007, and also put a copy of a June 18, 1982, Congressional 
Record as to Senator Nickles' amendment in terms of English 
well as limited English proficient, and I would like to quote 
it.\1\ ``I appreciate you bringing me up to date with the 
current interpretation by the Census Bureau, which includes 
mandating a bilingual assistance for people who say they speak 
English well. This is a direct contradiction of the amendment I 
offered and that was passed to include persons who say they 
speak English well as limited English proficient is a needless 
waste of time and resources.''
---------------------------------------------------------------------------
    \1\ The letter from Senator Don Nickles and copy of a June 18, 1982 
Congressional Record appear in the Appendix on pages 82 and 83 
respectively.
---------------------------------------------------------------------------
    I will quote further from his letter. ``It is embarrassing 
to see that the Director of the Census Bureau state in your 
letter to change the definition of limited English proficient 
would need to pass an amendment similar to my amendment, which 
passed in 1982.''
    Senator Carper [presiding]. Without objection.
    Senator Coburn. The reason I bring that up is that the very 
intent of Senator Nickles' amendment is what you say needs to 
happen to change what you are doing, and here is the author of 
the amendment saying you have totally misread what he said in 
his amendment. And I think the record needs to reflect that 
because that was not his intention. That was not the amendment 
that was passed. And the Congressional Record which I 
introduced will support that with the statements on the floor.
    And with that, I will limit my questioning, and we will go 
on.
    Senator Carper. Thank you, Dr. Coburn.
    Senator Coburn. Mr. Chairman, could we have one of your 
staff stay for the next session just so they can hear what we 
are doing, if you wouldn't mind?
    Senator Carper. That would be good. I would appreciate that 
too.
    I was asking a question and had to run off for the vote, 
and I did not give you a chance to fully answer it. But the 
question that I was asking was: What problems are you running 
into so far during your dress rehearsal? And the dress 
rehearsal is actually next year, isn't it, in those two areas?
    Mr. Kincannon. It began in May with the address canvass 
portion of the dress rehearsal in those dress rehearsal areas. 
So we are in the middle of it. The most exciting and well-known 
part of the dress rehearsal is non-response follow-up, but 
there is a lot more----
    Senator Carper. What problems are you running into so far 
during the dress rehearsal? Which I guess we are, what, 2 
months into now?
    Mr. Kincannon. The address canvass portion began in May and 
was completed at the end of June, on schedule. I went through 
some of these items, problems that we had encountered. There 
were some software problems which were being--some of them were 
being corrected by patches transmitted from the Harris 
Corporation to the handhelds. That worked very well. Others 
problems will need different kinds of fixes than those we 
worked on before. We have to go out in the field for the non-
response follow-up portion of the dress rehearsal late next 
spring.
    There were some problems and challenges in the contractor's 
Help Desk process. They had envisioned a different flow of 
concerns partly because of those software problems. And they 
are now readjusting to make sure up front there is enough 
people to respond to that. That is critical that workers not be 
discouraged by hanging on the line or getting a busy signal.
    But there have been a lot of successes as well. I mentioned 
that we replaced the password protection of the security of the 
data on the handhelds with a biometric measure, a fingerprint. 
And whereas 40 percent of the calls to the Help Desk during the 
2006 test census were about people that had forgotten their 
password and needed to have it unlocked, that dropped to near 
zero because you cannot forget your fingers mostly. That was 
very good.
    The handhelds I mentioned, I think to both of you, proved 
quite durable. Only a handful, 5 out of 1,400, were defective 
and 2 more were damaged, and all were immediately replaced out 
of adequate reserves.
    We completed 100 percent of the address canvass and on 
time, so that went pretty well. Those are the main problems 
that I think we encountered with the handhelds.
    Senator Carper. Alright.
    Senator Coburn. Just a couple other questions. My staff 
showed me an article that was in Government Executive yesterday 
where they talked about some of the older employees having 
trouble with the technology on the handheld. Is that a big 
problem or a small problem? This is from Government Executive 
yesterday: ``. . . one of the leaders she trained--an older 
woman--quit because the technology was too intimidating.'' Is 
that a small problem or is that a bigger problem?
    Mr. Kincannon. I think it is a small problem. That is our 
experience in the test in New York in 2004, and in Texas and in 
the other test----
    Senator Coburn. And it was tested with older individuals 
who are not necessarily computer savvy.
    Mr. Kincannon. Yes, it was tested with people who applied 
for jobs and got them.
    Senator Coburn. Yes, OK. And then one other thing. The cost 
of these handhelds is about $400, correct?
    Mr. Kincannon. Yes. I have an exact figure here, $411.43.
    Senator Coburn. And you are also paying for a wire service 
on top of that, Internet service?
    Mr. Kincannon. Yes, that is correct.
    Senator Coburn. Just by comparison, an iPhone costs $400, 
and we bought how many thousands of these?
    Mr. Kincannon. We will buy at least half a million.
    Senator Coburn. So we are going to buy 500,000 at $410. It 
ought to be great.
    Mr. Kincannon. It ought to fit its need perfectly and yet 
not be desirable for anybody else because it will not have any 
other use than collecting census information.
    Senator Coburn. Alright. Thank you very much.
    Mr. Kincannon. If we had 500,000 iPhones, if we could get 
them, I am not sure how long they would stay in our hands, 
frankly.
    Senator Carper. Mr. Kincannon, in 2004, GAO recommended 
that the Census Bureau develop a comprehensive, integrated 
project plan for managing decennial operations complete with 
milestones, complete with, I think, itemized cost estimates and 
risk and mitigation plans. I understand that this document has 
yet to be produced despite the fact that Census Day is less 
than 3 years away and we are in the middle, as you said, of the 
2010 dress rehearsal as we speak.
    When do you expect the Bureau's strategic plan will be 
finalized? And what are the impediments to completing this 
plan?
    Mr. Kincannon. We did submit to GAO in December of last 
year our research and development management plan, and they 
have found that useful, I believe. They can corroborate that or 
deny----
    Senator Carper. When you say ``they,'' what does that mean?
    Mr. Kincannon. The General Accounting Office, yes. Let the 
record show I pointed to my friends and colleagues over there. 
And there are plans in order to develop other versions and next 
levels. They understand that and have, as I believe, agreed 
that we are proceeding in the proper manner to get that done.
    Some of that is supposed to be August. It is not another 
version of this research plan, but it is the next layer of 
planning that will be available toward the end of August.
    Senator Carper. What are the impediments to completing the 
plan?
    Mr. Kincannon. Well, there is a lot of work to get it 
completed, and we are working on that diligently and we will 
complete it.
    Senator Carper. By when?
    Mr. Kincannon. End of August.
    Senator Carper. OK. Thank you.
    Mr. Kincannon. I answered somewhat speculatively about the 
population in the areas covered by the dress rehearsal.
    Senator Carper. Oh, yes.
    Mr. Kincannon. I said several tens of thousands. Actually, 
in the Fayetteville area, it is 334,000 households, and in San 
Joaquin County, California, 231,000 households. So it is a much 
more substantial area.
    Senator Carper. What is the rationale for having chosen 
those particular locales with that kind of population?
    Mr. Kincannon. The areas were selected based on having some 
language diversity, some group quarters, and some military 
bases and personnel who present unique kinds of enumeration 
problems.
    Senator Carper. Alright. As you know, there are segments of 
the population that are always harder or easier to count than 
others. What are some of the reasons for that problem? What is 
the Census Bureau planning for 2010 to make it easier to 
capture those who have been difficult to reach in the past?
    Mr. Kincannon. We have found over the years that difficulty 
to count or reluctance to respond correlates with low-income 
levels, low education levels, and youthfulness. Older people, 
those with an education, and those with higher incomes----
    Senator Carper. When you say ``youthfulness,'' just 
describe what you mean by ``youthfulness''?
    Mr. Kincannon. Say up to 25, maybe younger than that. 
Rather young people. I have a very broad spectrum of what I see 
as young nowadays, but there would be a few young people in the 
second row up there. Some of them are not quite so young as 
that, I guess.
    Senator Carper. They look pretty young to me.
    Mr. Kincannon. Yes. I find it difficult adapting to the age 
of my own children, but now that they have children themselves, 
I have to face up to their adulthood.
    So it is very young people, teenagers, maybe early 20s, 
people who have lower incomes, people who have less educational 
attainment. That often in our society, but not always, 
correlates with minority group status. But poorer white people 
also do not always have good response cooperation. It has to do 
with your involvement in society, your comfort with society, 
and the stakes you have in society, I suppose.
    What are we doing to address this? Of course, we have a 
large and growing immigrant population, legal and illegal. That 
is not a technical use of the term ``illegal,'' but you know 
what I mean. This means that we have much more language 
diversity, and we pay close attention to that. This year, or 
this census, because we have a short-form-only census, we will 
be able to mail in certain areas a bilingual, side-by-side, 
English-Spanish questionnaire. That is going to be very helpful 
with the largest language minority group in the country.
    We will target the neighborhoods where we send that based 
on the results from the American Community Survey, which tells 
us where there are neighborhoods with people who do not speak 
English very well. It is the same standard we use with the 
Voting Rights Act. And there we will mail bilingual 
questionnaires. That has tested very well, and we believe it 
will have a positive effect on response.
    Other languages will be provided. There will be some 
translated questionnaires for the five largest other languages 
than English and Spanish, and then we will have questionnaire 
guides, that is, a translation of the questionnaire but not a 
full questionnaire, so that you will have to look at an English 
questionnaire and look at the number on the translation guide 
and understand it and be able to fill it out that way. That 
will be for about 30 other languages.
    There are other things, too, of course. We have a 
partnership program planned. This was very successful in the 
2000 census. We formed partnerships with national organizations 
like NALEO and the NAACP and so forth, as well as groups that 
are not concerned particularly with minority groups but with 
other parts of society.
    We also have a cadre of partnership workers in the regions 
working with grass-roots leaders in their area. This raises the 
awareness of people who are leaders in all the communities in 
our country and who have, let's face it, more credibility than 
someone coming out from the regional office, and certainly from 
Washington, to say this is important to you and it is safe to 
report because we hold it confidential. And these groups not 
only convey a sense of confirmation about the importance and 
safety of responding to the census. They also are able to 
secure the cooperation of these local groups in providing space 
for our use for recruiting, for training, and for promoting the 
census. And an evaluation in 2000 indicated that the value of 
space provision alone more than paid for the partnership 
program. And we are quite confident that GAO is positive about 
this, too, that the partnership program did improve the 
cooperation and turnout of people who otherwise might not have 
answered the census.
    We will have a promotion, an advertising campaign. The 
advertising will be paid advertising, as it was so very 
successful in 2000. We will have a private contractor. That 
contract will be awarded probably at the end of August, early 
September, for a single integrated plan for all of our 
promotional activities--advertising, partnership, other kinds 
of promotional things.
    Senator Coburn. Will that be competitively bid?
    Mr. Kincannon. Yes, sir. It has been competitively bid. It 
has been out for some time, and as I say, we are in the stage 
of approaching the award decision now. That contractor will 
produce a plan for all these activities in consultation with us 
by February of next year. Then we will be able to get moving on 
all this.
    Senator Coburn. Is that a fixed-price contract?
    Mr. Kincannon. I think they have bid, and we look at the 
bid made and the value we get in return.
    Senator Coburn. Is it a fixed-price or a cost-plus 
contract?
    Mr. Kincannon. Cost plus.
    Senator Coburn. OK.
    Senator Carper. I have just one more issue to raise, and I 
think Senator Coburn raised this while I was voting. The issue 
deals with the option that we have chosen not to pursue, at 
least this time, and that is the option of doing at least a 
portion of the census on the Internet, an online approach. 
Apparently, you looked at it, you thought about it, and 
decided, at least this time, not to do it.
    What would have had to be different for you to have come to 
a different conclusion?
    Mr. Kincannon. What we would have to have is some test 
results that showed we gained from it, that we increased the 
response rate overall, that some people replied that otherwise 
would not; or a large enough proportion of the population 
replied that we would not have to print so many questionnaires; 
or could otherwise save on processing costs. And we would have 
to have a comfortable feeling in our anatomy about the security 
of that and the control of risk of phishing and other kinds of 
dangers that occur on the Web.
    The nature of the census is such that you have a very 
limited period to get things in. And if a rumor starts about 
identity theft through the census response, we are concerned 
about that.
    Senator Carper. In Delaware, our State slogan in the First 
State of Delaware is ``It is good being first.'' I can tell you 
from experience there are some things you do not want to be 
first in.
    Mr. Kincannon. Yes, sir.
    Senator Carper. But there are other countries, as I recall, 
that have actually done an online census. In fact, has Canada 
done--what are some other----
    Mr. Kincannon. They offer an online option in their census, 
as did Australia and New Zealand.
    Senator Carper. Those are countries with which we have 
actually a fair amount in common and a lot of affinity for. 
Canada is our neighbor to the north and we have a lot of 
interaction with that country, especially.
    Why do you suppose they found value in and decided to use 
the Internet as an option and we have not? Why does it seem to 
work for them but not for us?
    Mr. Kincannon. Why did the Canadians try it? I think their 
law requires them to try it, requires all government services 
to have a Web-based way of using that service. I do not want to 
suggest that at this stage of the decade, but I think that is a 
fact.
    There are a number of differences in the Canadian census. 
These are people that we work with closely, we admire them very 
much, but they have different requirements laid on them. They 
do not have any statutory deadline by which they produce 
results, so that if some failure occurs, they can take the time 
to recoup. We have to report the results by December 2010. I am 
sure we could suggest a change in that, but we would not like 
to have to do that.
    Also, Canada does not produce nearly the geographic detail, 
so you can also take more leisure and have less demanding 
requirements. They produce data for the provinces and for 
around 45 metropolitan areas and about 5,000 census 
subdivisions. They produce track data in the metropolitan areas 
only. They have about 5,000 tracks, I think it is. California 
alone has 7,000 tracks, and in the country as a whole, we have 
over 8 million tracks. So there is quite--I am sorry, 8 million 
blocks and a large number of tracks as well. So we produce a 
lot more geographic detail, and it may be that our population 
is--I do not know. Our response rate was in between New Zealand 
and Australia. One of those was at 9; we had a test at 6, 7, 
and one at 7.2, and the second of those was at a little over 6, 
I believe.
    Senator Carper. I am sorry. Just a clarification. What is 
the significance of----
    Mr. Kincannon. The percentage of the population replying by 
Internet. In Canada it was higher, about 18 percent, and so 
although they did not save any money in 2000--I am sorry, their 
2006 census it would have been. They believe they have 
confidence that they can print somewhat fewer questionnaires in 
their 2011 census.
    Senator Carper. I do not know how long in this country we 
have had the option of filing our taxes online, but my guess is 
that the first time we did it, the number or the percentage of 
people who elected to do so was not so great. I think now it is 
probably over half.
    Senator Coburn. Sixty percent.
    Senator Carper. Yes, about 60 percent.
    Your successor has been nominated, and so there is, I would 
say, a fair chance that you will not hold the same position 10 
years from now that you hold today. Senator Coburn will 
probably still be around. I am not sure where I will be in 10 
years.
    But looking down the line 10 years from now, somebody else 
will be sitting in your seat. Do you think 10 years from now we 
will still be debating whether or not it makes sense to do a 
portion or have as an option an Internet alternative as part of 
the census?
    Mr. Kincannon. I do not know any way to answer that. I 
cannot tell whether we will have a more secure Internet with 
less cyber crime than we do now. I do not know whether there 
will be broader uses by the Census Bureau in household surveys 
with the Internet and whether that will justify the investment 
in a very secure system that might be robust under those 
circumstances. Trying to foresee what will happen with 
technology is difficult.
    There was a film 20 years or so ago called ``A Clockwork 
Orange,'' and it looked into the future, and the future of 
sound reproduction was a very fancy tape cassette. Even by the 
time the film came out, CDs had replaced that, and now we see 
CDs have been replaced by DVDs and super DVDs and HD DVDs and 
all kinds of things so that one's head almost spins at the 
evolution of technology.
    We will continue exploring and testing the Internet, not 
for 2010, but to see how it evolves and whether we can make use 
of it. It is very attractive in theory, and if we can make it 
work for us, then that will be----
    Senator Coburn. I am just appalled. If the Internal Revenue 
Service that has 10 million pages worth of regulations can have 
a secure Internet service where 60 percent of the people in 
this country can file online, all the businesses have to file 
online, to tell me that we are waiting on technology to be able 
to catch up in the census, I just do not buy that, Mr. 
Kincannon. I am sorry. But that technology is out there today. 
It is the fact that we have not had the vision to go get it, 
and it does not sound like we are going to have the vision to 
do it in 2010.
    I would just say--and I told you this at the last hearing. 
I am going to do everything I can to force Internet census down 
your throat, and I am going to do it with amendments on the 
floor. They may lose, but the American people are going to 
say--when 70 percent of the people in this country are online 
and the head of the Census Bureau is saying we do not think we 
can do this safely or appropriately, they will not buy that 
either.
    And so I think you all are living in the past instead of 
the future, and I would recommend heartily to you that you get 
on board so that when we are doing this in 2020 that we are 
online and that everything has been planned now to make sure we 
get there. It is appalling that we are not doing the American 
Community Survey online right now. I spent 30 minutes on the 
phone with one of your people answering questions that I did 
not want to answer just out of the American Community Survey. I 
could have done the whole thing on the Internet in 10 minutes. 
But I spent 20 minutes filling it out on paper and then another 
30 minutes with your agent. It is impossibly inefficient, and 
it needs to change.
    And so I am just offering a challenge to you today that I 
am going to be there--you know it. I am very plain-spoken. I am 
very forthright in what I am going to do. It is unconscionable 
that we are not doing some of this on the Internet, and it does 
not have anything to do with security, and it does not have 
anything to do with technology that is out there today. It has 
everything to do with the lack of vision of getting it done. 
And I know you have run a test and you were not happy with the 
test. But nobody gets to see that test.
    One other thing. Why in the world are you doing a cost-plus 
contract on the promotion for the census rather than a fixed-
price contract?
    Mr. Kincannon. Well, I do not know the answer to that.
    Senator Coburn. Well, you are in charge of it. Why are we 
not doing a fixed-price contract for something you know what is 
going to do, rather than have a cost-plus contract that people 
are--whatever it costs, we are going to pay it plus. You cannot 
manage that at the same time you manage the census. Why is 
there not a fixed-price contract so that the American people 
know what we are going to get, here is the value of what we are 
going to get, and here is what we are going to pay? Because we 
have a terrible record on cost-plus contracts throughout this 
government. So why would we not have a fixed-price contract?
    Mr. Kincannon. I will send you an answer about that. I am 
not sure I agree with you, but I appreciate your point.
    Senator Coburn. Well, I tell you, this Subcommittee has 
looked at a ton of cost-plus contracts, and we have not seen 
many that have been very beneficial to the American taxpayer. 
Most of them have been very beneficial to the contractor. Most 
businesses would not do it in a cost-plus. They would do it 
with a fixed price. Why aren't we? And I would be happy to have 
your answer. And I have three other questions to submit for the 
record.
    Mr. Kincannon. Well, I did not exactly answer some of your 
implied questions about that.
    On the IRS--and I do not manage the IRS. I am a taxpayer. I 
have once filed online and once not filed online and had the 
option to do so. The IRS does not operate a secure website that 
60 million people can file on. You buy software from a private 
company, send it to them, and they relay it to the----
    Senator Coburn. Why couldn't the Census do that? The 
technology is out there. As I said, the technology is there. It 
is the vision of using the technology to get us to where we 
want to go.
    Senator Carper. Mr. Kincannon, you may want to have a chat 
with your nominated successor to say to get ready to answer 
that question when we have our hearing for his or her 
nomination.
    Mr. Kincannon. I think my nominated successor is familiar 
with the Senator from Oklahoma's views on this matter, and I do 
not know what his views are, and I do not intend to discuss 
them with him prior to his confirmation.
    Senator Carper. Alright. I think that pretty well wraps it 
up for today with respect to your testimony. Anything you want 
to add before you----
    Mr. Kincannon. I defer to the authority of the Subcommittee 
and the full Committee, but I hope that you are assiduous in 
conducting hearings on my successor and free me to maybe do one 
of these cost-plus contracts with----
    Senator Coburn. Fixed price.
    Mr. Kincannon. Oh, OK. Fixed price. that would be alright.
    Senator Carper. I will say on that subject, we had a full 
Homeland Security and Governmental Affairs Committee hearing 
today on, among other things, contracting out. And one of the 
questions I asked of our panel was: Give us some examples of 
where cost-plus or no-bid contracts make sense. They struggled 
to come up with the answers.
    How many years of service do you have to the people of this 
country?
    Mr. Kincannon. Thirty-five years of Federal service and 8 
years at the OECD, so that is enough, I think.
    Senator Carper. That is a lifetime. And while we may have 
some disagreement with you with respect to our use of the 
Internet as an option in regard to the 2010 census, I think I 
speak for all of us in thanking you for your service to the 
people of our country. We will try not to prolong too long the 
consideration of your successor and to give you the opportunity 
to head on to the rest of your life. Thank you very much.
    Mr. Kincannon. Thank you both.
    Senator Carper. You bet. With that, let me invite our 
second panel to join us, please.
    Senator Coburn, you are recognized.
    Senator Coburn. While you are coming forward, I just wanted 
to put a couple of things up that we have looked at. First, the 
one that is up there now shows per capita cost of the census in 
terms of individuals. With technology, we are getting behind. 
The next sign is after inflation cost of the census, which I 
think is very telling. These are in 2000 dollars, I believe. In 
1970, we spent $900 million, 1980, $2.2 billion; 1993, $3.3 
billion; 2000, $6.6 million; and in constant dollars, we are 
going to spend $9.3 billion this year on census, which is 
essentially 6 percent more people. So we are going to have at 
least a minimum, in constant dollars, 50-percent increase in 
dollars for a 6-percent increase in the number of people.
    The cost of the census shows the absolute total cost in 
thousands and increase percentage from 40 percent--from 2000 to 
2010, we are anywhere from 79 percent or above, 2000 was 152 
percent above the one in 1990, 120 percent, 1980, 350 percent. 
The point is we cannot afford to keep growing as a population 
because we are going to go bankrupt counting it.
    Finally, to make my point, 74 percent of the American 
public adults are online right now; 60 percent of the people 
and 100 percent of business pays their taxes online. So I have 
a challenge, and the challenge is to the American public. Help 
us make the census better. Take the challenge. Can you beat 
Uncle Sam? Figure it out. Less than $90 for a household to get 
all the counts that we need. And Mr. Kincannon is not here 
anymore, but he missed my point on the piece of equipment. For 
$400, you can get an iPhone that does 20 things. And we got a 
piece of equipment that is going to be good for one census, 
$220 million, and then we are going to throw it away. It is not 
going to be good for anything.
    So, first of all, we have overpaid for what we have bought 
in terms of technology, which was probably another cost-plus 
contract instead of a fixed-price contract. And those are the 
kind of things that we need to be paying attention to here.
    I again will say--and I said this to Mr. Kincannon--the 
American public will not buy the fact that we are not doing 
this online. There is no excuse for us not to do a portion of 
it online, even at this late date. I know they are risk averse 
because of the criticisms they get from this body. But this is 
not a hard thing to do, and the technology is out there, and we 
just need the leadership and vision to do it.
    I thank the Chairman.
    Senator Carper. You bet. Thank you, Dr. Coburn.
    We have our four witnesses here. I am not going to give a 
full introduction, but I would like to just mention that Mathew 
Scire, who I understand will be delivering the oral statement 
for both of our GAO witnesses, is Director of GAO's Strategic 
Issues team. Welcome, or as we say in Delaware, ``Bienvenue.''
    Mr. Powner is Director of GAO's Information Technology 
team. I believe you are here to respond to questions but not to 
testify. Is that right? Welcome. Thank you.
    Mr. Reamer, a fellow at The Brookings Institution, 
Metropolitan Policy Program, thanks for joining us.
    And, finally, Mr. McTigue comes to us from the Mercatus 
Center at George Mason University, where he works as Vice 
President of the Center and Director of the Government 
Accountability Project. Welcome.
    We are delighted that you are all here. Your entire 
testimonies will be entered into the record, and we will 
recognize each of you for roughly 5 minutes. If you go a little 
bit long, we will not make a big deal of it, but try to keep 
pretty close to 5 minutes. Thank you.
    Mr. Scire, you are on.

 TESTIMONY OF MATHEW J. SCIRE,\1\ DIRECTOR, STRATEGIC ISSUES, 
  AND DAVID A. POWNER, DIRECTOR, INFORMATION TECHNOLOGY, U.S. 
                GOVERNMENT ACCOUNTABILITY OFFICE

    Mr. Scire. Mr. Chairman, Senator Coburn, my colleague and I 
thank you for the opportunity to be here today to discuss 
preparations for the 2010 census. Thorough planning is 
important to the success of any large, long-term project. To 
enhance the Census Bureau's performance and accountability, we 
recommended that it develop a comprehensive project plan and 
annually updated life-cycle cost estimates. The Bureau is now 
developing such a plan and has updated its cost estimates. 
However, more robust information on the likelihood of key 
assumptions and their impact on life-cycle costs could help 
inform the Congress not only what the census is likely to cost, 
but also the confidence of that estimate. One key assumption is 
the productivity of field workers.
---------------------------------------------------------------------------
    \1\ The prepared joint statement of Mr. Scire and Mr. Powner 
appears in the Appendix on page 43.
---------------------------------------------------------------------------
    Mr. Chairman, Census 2010 relies as never before on the use 
of contractor-developed automation and technology. At the 
request of this Subcommittee, we are assessing four key 
technology investments. Thus far, we see mixed progress. For 
example, while the Decennial Response Integration System is 
expected to meet cost targets during the dress rehearsal, the 
Field Data Collection Automation program is projected to 
experience cost overruns. Also, the Bureau has delayed some key 
functionality that was expected to be part of the dress 
rehearsal and did not complete plans for end-to-end testing 
that is critical to understanding the performance of 
interrelated systems. Finally, we believe project teams could 
do more to identify risks, establish mitigation plans, and 
report risk status to higher-level officials.
    It is important today for the Bureau to monitor closely the 
costs, schedule, and performance of its IT acquisitions and 
aggressively manage the risks that they face. Mr. Chairman, we 
have entered a new and critical stage in the planning and 
operations of the decennial census. Dress rehearsal operations 
are well underway, and the very first operation of Census 2010 
in which the Bureau enlists the help of local governments has 
been launched. More recently, we observed the first use of the 
handheld computers by field workers in the address canvassing 
operation of the dress rehearsal. These devices are keystone to 
the reengineered census. We observed technical difficulties 
with the devices, however. Without correction, these 
inefficiencies can affect worker productivity and ultimately 
the cost of the census.
    Finally, I would like to draw attention to Bureau plans for 
enumeration in the Gulf Coast region. The effects of Hurricanes 
Katrina and Rita are still visible today, and numerous housing 
units have been or will be lost. Conversely, in some 
jurisdictions, there is new development. This continuing change 
in housing stock may affect census operations. For example, the 
productivity of Bureau field staff conducting address 
canvassing could be affected as they potentially face 
challenges of distinguishing between occupied, uninhabitable, 
as well as temporary housing units doubled up on lots. On the 
other hand, non-response workload could be increased if the 
Bureau mails questionnaires to housing units that are vacant on 
Census Day.
    In summary, we believe that the challenges highlighted 
today require careful monitoring and oversight. More 
transparent planning and cost reporting will help. Likewise, 
the costs, schedule, and performance of key technology 
investments demand greater attention. As in the past, we look 
forward to supporting the Subcommittee's oversight efforts to 
promote a timely, complete, accurate, and cost-effective 
census.
    This concludes my opening remarks. Thank you again for the 
opportunity to speak today. My colleague and I would be glad to 
take any questions that you may have.
    Senator Carper. Thank you, Mr. Scire.
    Mr. Reamer, you are recognized. Your full statement will be 
entered into the record.

  TESTIMONY OF ANDREW REAMER,\1\ FELLOW, METROPOLITAN POLICY 
               PROGRAM, THE BROOKINGS INSTITUTION

    Mr. Reamer. Chairman Carper, Senator Coburn, I am pleased 
to be here. My role today is twofold: First, I will describe 
the extraordinary return on investment the Nation gets from the 
decennial census; and, second, I will review key issues with 
regard to the Census Bureau's readiness to conduct the census.
---------------------------------------------------------------------------
    \1\ The prepared statement of Mr. Reamer appears in the Appendix on 
page 65.
---------------------------------------------------------------------------
    The census is fundamentally important to the Nation in 
terms of our democracy and conducting public policy, as well as 
to the functioning of our $13.6 trillion economy. The 
architecture of our representative democracy rests on the 
foundation provided by the census. The House is apportioned 
according to the census. The Electoral College votes, 
therefore, are based on the census. And because the President 
is selected by the Electoral College, the judiciary is, in 
fact, affected by the census as well. So every Federal branch 
is affected by the census.
    State legislatures rely on the census to redraw 
congressional and State districts. Local governments use the 
census to create county council districts and school board 
districts and voting precincts.
    The decennial census is also critically important for the 
effective performance of government. When I discuss the value 
of the decennial census, by extension I am including two 
additional census programs: One is the Annual Population 
Estimates Program, which uses the decennial census counts as a 
basis to provide annually updated population estimates; and the 
second is the American Community Survey, which provides on an 
annual basis detailed characteristics of population down to the 
neighborhood level.
    The Federal Government relies on the census data in three 
ways.
    First, the census data guide the distribution of hundreds 
of billions in Federal financial assistance. In fiscal year 
2004, I estimate that at least $287 billion across 75 grant 
programs were allocated across the country on the basis of 
census numbers or census-derived numbers. That is about 62 
percent of the total, and I have given each of you a packet 
that has the figures for your State.
    Second, census data provide key benchmarks for Federal 
enforcement of civil rights and anti-discrimination laws and 
court decisions in voting and in the workplace.
    Third, census data play an important role in informing the 
design, implementation, and evaluation of a variety of Federal 
efforts beyond financial assistance and regulation, including, 
for instance, adult education, small business development, 
veterans' health, affordable housing, transportation planning, 
disabled students, and even groundwater contamination.
    And the census provides the basis for giving Members of 
Congress up-to-date profiles of their constituent population 
through the American Community Survey.
    State and local governments rely heavily on census data to 
make on-the-ground investment decisions across all domains of 
government, including education, highway transportation, 
affordable housing, access to health care, workforce training, 
criminal justice, and, very importantly, responses to and 
planning for natural and manmade disasters.
    The influence of the census is also pervasive across the 
private sector. Businesses of all types--retail, manufacturing, 
services--and all sizes--from J.C. Penney, Wal-Mart, and 
Target, down to sole proprietorships--use census data to 
identify markets, select business locations, make investment 
decisions in plant and equipment and new product development, 
determine goods and services offered, and assess labor markets. 
Nonprofits such as hospitals and community service 
organizations use them, and the public and private sector work 
together in local economic development using census data to 
create jobs and expand the tax base.
    Fundamentally, then, in my view, census data are essential 
for the effective operation of the entire $13.6 trillion 
economy.
    Now, with regard to preparedness for 2010, clearly we need 
a complete and accurate census for all these public purposes. 
Achieving such a census requires an accurate Master Address 
File. We need to know every address in the country, we need to 
get a questionnaire to each address, have them return it, and 
capture the information provided accurately.
    There are several issues in this regard. One is the Census 
Bureau needs adequate funding. It is the largest peacetime 
operation that the Nation undertakes, and it is important--
whatever the appropriate cost is, it is important for Congress 
to recognize that preparations take several years in advance 
before the count and that the count does not start in 2010.
    The second issue is the management of the Local Update of 
Census Addresses program (LUCA). This program is critical to 
having an accurate Master Address File (MAF). This Census 
Bureau program gives States and localities the list of 
addresses currently in the MAF, and the localities can update 
it. Clearly, LUCA is important because it affects the flow of 
Federal funds, apportionment, and business investment 
decisions. In 2000, New York City added over a third of a 
million households through the LUCA process, so it is vital to 
localities. The 2010 program seems to be designed very well. 
However, there seem to have been some problems in the 
implementation in the 2008 dress rehearsal, and also in terms 
of actually getting OMB approval and asking for public comment. 
I encourage the Subcommittee to look into these issues of 
implementation.
    The third issue, mentioned earlier in your conversation 
with Mr. Kincannon, is the community partnership program. It is 
one thing to know where people live. It is another thing to get 
them to answer. And it is important that the Bureau have the 
funding it needs in a timely way to create a community outreach 
program to reach the hard-to-count.
    For fiscal year 2008, the Administration did not provide 
the Census Bureau with money for a community partnership 
program. The House Appropriations Committee added those funds, 
and I encourage this Subcommittee to support your colleagues in 
the conference committee to see that those funds are provided.
    There are several other issues listed in my testimony: An 
important way to count people called ``Update Enumerate,'' 
which I would be happy to talk about in the Q&A period; the 
process of training half a million temporary workers; managing 
technology contracts; back-up and contingency plans; and then, 
last, as Mr. Kincannon mentioned, a lot of people with senior 
experience are retiring, and how does the Census Bureau capture 
the knowledge that they have regarding the proper conduct of a 
census.
    I hope you have found my remarks of value, and I would be 
happy to answer any questions.
    Senator Carper. I think we did. Thank you very much. Dr. 
Coburn.
    Senator Coburn. I just might note for our audience that Mr. 
McTigue is a former Minister of Labor from New Zealand and a 
Member of the Parliament of New Zealand.
    Senator Carper. Is that right? Which island are you from?
    Mr. McTigue. I am from the south.
    Senator Carper. One of my favorite places on Earth is the 
South Island of New Zealand, especially Queenstown.
    Mr. McTigue. Thank you very much, Senator.
    Senator Carper. What a great place.
    Mr. McTigue. We are still accepting migrants.
    [Laughter.]
    Senator Carper. In a couple of years, I might be tempted. 
Every now and then, I think the voters of Delaware would 
probably like for me to look in that direction.
    Well, we are delighted that you are here. You come from a 
beautiful place.

   TESTIMONY OF MAURICE P. McTIGUE,\1\ VICE PRESIDENT OF THE 
   MERCATUS CENTER, AND DIRECTOR, GOVERNMENT ACCOUNTABILITY 
                PROJECT, GEORGE MASON UNIVERSITY

    Mr. McTigue. Thank you very much, Mr. Chairman, and thank 
you for the invitation to present testimony in front of this 
Subcommittee again.
---------------------------------------------------------------------------
    \1\ The prepared statement of Mr. McTigue appears in the Appendix 
on page 78.
---------------------------------------------------------------------------
    My expertise is not as a statistician but, rather, in 
organizational performance, and in that organizational 
performance work, trying to understand how organizations might 
be able to improve their level of performance. So my comments 
are really couched in those terms.
    While it is a constitutional requirement to gather the 
census each 10 years, it should not be considered to be a 
bureaucratic process. In my view, it should be considered to be 
a dynamic process, because its real function is to enable 
better decisionmaking by decisionmakers at the level of 
Congress and right down through the private sector and the 
public sector. That means that there are two criteria that 
become very important for the Census Bureau in conducting the 
census. The first, of course, is accuracy, but the second is 
the utility of the information. How useful is this information 
going to the parties that use it in decisionmaking? And, of 
course, if you look at utility, one of the questions then is 
timeliness. And timeliness, of course, is something that we 
have become much more critically aware of as we move into the 
information revolution. So improving the quality of information 
and improving the access to it would indeed improve the quality 
of decisionmaking.
    Electronic measures are the best way to do that, and some 
of the experience that I have had with the IRS here in the 
early stages of their online work was that response rates were 
low. It might come as a surprise to both of you that the public 
are not normally very happy about responding to government 
requests. In fact, from time to time they are even suspicious 
about it, and they need some encouragement.
    The early response rates for the IRS were quite low, but 
they did a lot of work on identifying those who did not file 
online and trying to find ways of encouraging them to do just 
that. It is wrong for the Census Bureau to think that you just 
offer the option and it will be taken up. You have to do 
something to quell the concerns of those people who are 
currently non-responsive. And I think that is one of the 
challenges for the Census Bureau going through to 2020.
    Even at this late stage, with the quality of technology 
today, an option is something that, in my view, could easily be 
provided or manufactured. It is also, in my view, wrong to 
consider that the security issues are too difficult. The short 
census form this time contains very little information that is 
sensitive. Well, maybe some people care a lot about their age, 
but the rest of it really just identifies our name, our birth 
date, where we reside, and our background in terms of 
ethnicity--not things that are very sensitive. It was different 
when you had the long form. So I think that some of the 
arguments put up by the Census Bureau do not really hold water 
as far as that is concerned.
    The last comment that I want to make in that area, though, 
was that one of the discoveries of revenue organizations around 
the world as they move to online filing was not just the huge 
cost reduction of online filing, but it was the dramatic 
improvement in accuracy. The error rate for the IRS went down 
to 1/20 of what it was before. With the best will in the world, 
all of those people who voluntarily respond to the census are 
not going to get it right. More of them would get it right if 
they were doing it online, and that would improve the quality 
of that information for decisionmakers.
    Another point that I picked up in my research was this: 
There seems to be an undue concentration of the efforts of the 
whole process on the non-responders. One of the things that the 
IRS did very well was recognize that the more people that they 
could get to respond online, the more effective their tax 
collection process was. So they put a lot of effort into trying 
to get voluntary compliance and voluntary filing. There does 
not seem to be the same effort going on in the strategy for the 
census to encourage people to respond first so you have a lower 
cohort or a smaller cohort of people that you have to follow up 
with the very expensive non-response process. And concentrating 
on that, in my view, would reduce costs; it would also improve 
accuracy.
    I hope these comments are helpful to the Subcommittee, and 
I would be very happy to answer questions. Thank you.
    Senator Carper. Thanks very much to all of you.
    I want to stick with this idea, for at least a little bit, 
of an online option. When was the first time New Zealand used 
it?
    Mr. McTigue. New Zealand conducts census every 5 years, 
even though I know that is an incorrect interpretation of the 
word, but it does it every 5 years because that also matches up 
with Boundary Review Commissions for setting electoral 
boundaries. I think that they have done one census.
    There would also be some differences in making comparisons. 
If you look at Dr. Coburn's figures here, the percentage of 
adults online in New Zealand is probably closer to 50 percent, 
and the saturation of broadband access on the Internet would be 
even lower than it is in the United States, and some of these 
things require broadband access to be able to conduct the 
returns online effectively. So those might be two of the 
criteria.
    But I would expect that the first time you do something, 
you have to expect a low response and gradually build 
confidence in people being able to trust the process and to use 
it.
    Senator Carper. Mr. Powner, let me ask you a question, and 
this requires really some judgment, and I know you cannot 
answer this definitively. But if you were running the census, 
if you were our nominee for--and I do not think you are going 
to be, but if the President should call you tonight and say, 
``I want you to run the census for 2010. Can I submit your name 
to Coburn and Carper and that crew?'' and for some reason you 
would say yes and you got confirmed, do you think you would 
have time to alter the game plan for 2010 so that at least we 
would have some mechanism that would help inform the process 
going forward to suggest--to learn from the experience, to see 
if it is a total disaster or if there is some virtue to it so 
that when we approach 2020 we will at least have had a head 
start?
    Mr. Powner. Well, a couple of key points. I think that when 
you look at the feedback so far with all the testing that has 
occurred, the dress rehearsal is what is key. So we really want 
to look at the dress rehearsal, and you will be able to 
determine whether the handhelds are going to be successful at 
this point in time.
    As my colleague here mentioned, we are very concerned about 
the performance of those handhelds based on our observations to 
date. So I think going forward what we would want to do is 
place a lot of faith in that dress rehearsal.
    Now, here are some concerns with the acquisitions. Some of 
those acquisitions we have backed off in terms of including 
them in that dress rehearsal. There is a dissemination system 
referred to as DADS II. The contract was delayed one year. That 
will not be part of the dress rehearsal. The system that 
integrates Internet, phone--was to include Internet, phone, and 
the paper responses, there is some functionality that is being 
deferred, and that will not be included.
    So one of the key things is when we have this dress 
rehearsal, that will inform us a fair amount, but we also have 
to look at what is being deferred, and that raises the level of 
importance of future testing between 2008 and 2010, especially 
when you look at interfaces and all the interactions with the 
systems and the business processes.
    So dress rehearsal will give us a fairly informed position 
at that point in time, but that is not the end game, and we 
have got to look at that testing and take that very seriously 
going forward.
    Senator Carper. Alright. You may have answered my question. 
I am not sure that you did. Is there some way to modify over 
the next year or so, as we approach the 2010 census, to include 
in the process at least a demonstration that would help inform 
us going forward as to whether or not the idea of doing an 
online census is something that we should consider expanding in 
2020?
    Mr. Powner. Well, in terms of the online census, if we were 
going to reconsider that, I think a key point--if you recall, 
the DRIS contract, which included the Internet response, was 
let--I believe it was October 2005. It included the Internet 
response at that point in time, and it was somewhere around 
spring of 2006 that we got word that the Internet was now not 
included.
    I think a key question, if you want to revisit the online 
response, is to go back to the contractor who has that 
contract--at one time it was in the contract, and they were 
planning for it: Internet, mail, phone. How easy or how 
difficult would that be to reinsert that, and what are the 
associated costs? That has been kind of a blind spot to us 
because usually when there is a contract modification, even 
pulling things out, I am not aware that the cost actually went 
down when we took the Internet response out.
    Senator Coburn. Twenty-two million dollars is the cost of 
that contract.
    Senator Carper. Anybody else want to respond to the same 
question?
    Mr. Scire. Well, I think I would just add that it is a 
little bit unclear what the process was in terms of deciding 
whether or not to include or exclude the Internet----
    Senator Coburn. Is your microphone on? I am not sure I can 
hear you.
    Mr. Scire. It is not clear what the process was for the 
Bureau as it went through and made the decision on the 
exclusion of the Internet. I think that is part of the point 
that Mr. Powner was trying to make here. And so it is tough to 
say what the analysis was at that point in time. The analysis 
that they have done most recently, which is about a month old, 
they look at the $22.5 million cost, and they look at various 
potential savings and conclude that the savings do not come 
anywhere near the cost. But part of the assumptions here have 
to do with the conditions under which they test what the 
Internet response would be, but these tests did not include 
advertising, for example. In a real use of the Internet, you 
would have advertising, which would promote the use of the 
Internet and so forth.
    So I do not believe that we really know what the Internet 
response would really be if it were offered.
    Senator Coburn. One of the things they could do is you 
could run the Internet census first, advertise and promote it, 
and then re-engage all your address book and everything else to 
lessen that cost.
    Can I follow up with a question or do you want to finish?
    Senator Carper. Let me ask one more question, and then I 
will pass it over to you.
    Mr. Scire, I gather from your testimony that you are not 
entirely confident that the 2010 census will actually cost 
$11.5 billion, as Mr. Kincannon said that he expects that it 
will. What were some of the factors that led to the last-minute 
cost increases that we saw in 2000? I asked Mr. Kincannon that 
question. Have you seen anything in your oversight that tells 
you that we are not about to see the same kind of escalation 
this time? And what, if anything, can we do to prevent the cost 
rise in 2010 that we experienced a decade or so earlier?
    Mr. Scire. OK. Part of the explanation for the change in 
cost, the actual cost in 2000 was the late change in the form 
of the census, that they had the Supreme Court decision which 
took sampling off the table, and so they had to go to complete 
enumeration. So I think that is a major cost driver for the 
increase that they experienced in 2000.
    You asked about the $11.5 billion and the confidence that 
we have in the $11.5 billion estimate for 2010, and I would say 
that the Census Bureau does not know what confidence it has in 
the $11.5 billion estimate. One of the things that we had 
recommended is that the Bureau in its life-cycle cost 
estimation do sensitivity analysis, which would take a look at 
the factors that drive costs and to describe what the 
likelihood is of those particular assumptions.
    So productivity, for example, is a major assumption and 
driver for cost. If productivity is not what you would expect 
and the cost is going to increase, the Bureau could be 
reporting to the Congress what a 1-percent difference in 
productivity might translate into in terms of the ultimate 
life-cycle cost. And by doing that and looking at all the 
different factors that go into the estimate of cost, it could 
provide you an estimate of $11.5 billion, plus or minus the 
range of whatever they think it might actually be. So you can 
get sort of a range.
    I heard Mr. Kincannon talking about not seeing a forecast 
of large cost increases. He is not worried about wage rates. I 
think we would say we are more interested in an objective fact-
based assessment of the likelihood of those various outcomes, 
and how a percentage difference in what they assume might 
translate into the ultimate cost.
    Senator Carper. Alright. I am going to go back--and you may 
have answered this but I missed it. What, if anything, can we 
do in our oversight role and the Census Bureau in their role as 
the operational manager, what can we do to prevent costs in 
2010 from going up dramatically, as they did a decade earlier?
    Mr. Scire. Well, the one thing that I wanted to mention in 
part of my response was the census could provide better 
information to you about what these cost drivers are. So, for 
example, if the performance of the handheld computers really is 
what is going to ultimately drive the cost of the census, then 
the focus of oversight should really be on that. If it is wage 
rates--you talked about that earlier--then focus on that.
    But that is something that I do not see, is information 
that would permit you to identify the areas where there is the 
greatest sensitivity and the greatest influence on cost. That I 
think would help with oversight.
    Mr. Powner. Chairman Carper, if I could also add from a 
technology point of view that we are spending $3 billion of the 
$11.5 billion on new technologies. And if you look historically 
at what has happened, a major cost driver is requirements 
creep. This is nothing new. We testified before you on the 
high-risk list and watchlist. That is a big reason for cost 
growth on many programs throughout the Federal Government.
    There is a concern about requirements creep. On the FDCA 
contract, which includes the handhelds and some of the other 
contracts, we are seeing evidence of requirements creep, and 
this is not anything new. Several years ago, we recommended to 
the Census Bureau that they ought to define those requirements 
as completely as possible up front so that there is not this 
ambiguity going forward. And sure enough, that is coming back 
to bite them at this time.
    Senator Carper. OK.
    Senator Carper. I believe its $220 million?
    Senator Coburn. Yes.
    Senator Carper. We are doing our math up here on how much 
of the $11.5 billion is going to be spent for these handheld 
devices, and Senator Coburn says it is about $220 million, 
which is actually a fairly small percentage of the overall 
cost. It is, what, about 2 percent or something, I think.
    Senator Coburn. But the whole contract is $600 million.
    Senator Carper. Alright. OK, Dr. Coburn, you are on.
    Senator Coburn. Thank you. A couple of things. Mr. Scire, 
you were talking about field data and cost overruns. Would you 
describe that a little bit more for me? In your opening 
statement you talked about you were worried about cost overruns 
on field data. Please go into a little more depth on that.
    Mr. Scire. Yes, there are a couple of things, and I will 
ask my colleague to join in.
    First, what we observed at the dress rehearsal and the 
reason that we went out to do this was we realized that the 
handhelds are a keystone to the reengineered census, and so we 
are very interested in seeing how they perform. And, again, our 
observations are preliminary. I have to stress that. But what 
we did see is the inefficiencies in terms of being able to link 
multiple addresses to a single map spot, a slowness of the 
devices when it had a large assignment area or data for large 
assignment areas. Those affect efficiency. That affects 
productivity. And so the advantages that you would get--or part 
of the cost advantages that you would get from the introduction 
of this technology would be lost. So that is why we looked at 
that.
    We are working with the Bureau and with Harris Corporation 
to understand----
    Senator Coburn. Mr. Kincannon says that is all software, 
not hardware. Is that right?
    Mr. Scire. I would say we are still working to figure out 
what are the explanations for these.
    Senator Coburn. You do not know the answer to that. OK.
    Mr. Scire. I also want to raise another point. Mr. 
Kincannon said that the address canvassing dress rehearsal 
started and finished on time. I think what we are interested in 
looking at is the productivity. How did these devices perform? 
Did the Bureau have to bring in additional resources in order 
to complete the address canvassing on time? So it is not just 
the starting and completing. It is also the conduct and----
    Senator Coburn. What was the cost per contact?
    Mr. Scire. We are looking at all those, but we also in the 
statement talk about cost overruns in the FDCA contract and the 
dress rehearsal, and I think my colleague can add to that, if 
you would like.
    Mr. Powner. Yes, Dr. Coburn, we actually look at earned 
value data. I think you are familiar with that----
    Senator Coburn. Yes.
    Mr. Powner [continuing]. Where it is a requirement of the 
Administration; all contractors are required to provide that on 
major IT acquisitions. You can actually take that earned value 
data and project trends based on historically what has 
happened. And we are starting to see increases with the field 
data contract. It is projected right now only about $20 
million, but there is still a ways to go. And given the 
uncertainty with requirements, as our written statement 
mentions, we are concerned about additional requirements growth 
and even more increases with that contract.
    The other thing to keep in mind, too, the DRIS contract, 
that is on schedule and within cost right now, but they are 
delaying functionality. So sometimes when you start delaying 
that functionality, it will catch up to you eventually. You 
will start seeing those.
    Senator Coburn. Does that belie the fact that it is 
supposedly on schedule and--that they are delaying the 
functionality, is there a problem?
    Mr. Powner. Then it is not really on full schedule, 
correct.
    Senator Coburn. That is what I am saying. So you are saying 
opposite things with the same statement: Yes, we are on 
schedule and under budget, but we are delaying functionality, 
which means it is not working.
    Mr. Powner. Yes, exactly. And that is why that earned value 
data--I mean, that is a common technique where you can say, 
hey, we are on schedule and budget. But with that earned value 
data, you get that third leg on whether you are on track with 
the delivery of functionality. And so that is the criticality, 
and we will continue to review that for your Subcommittee going 
forward.
    Senator Coburn. Your observations of the question Senator 
Carper asked about this plan from the census and contingency 
plan and here is what we are going to be and he said it was 
going to be available by, I believe, August. Do they have the 
planning in place and ready to go for options and contingencies 
that are not expected today but they have sat down and talked 
about what they are going to do? Do they have that plan? Is 
that there? Are they working toward what if this happens? Do 
they have the plans in--or are they going to have to spend a 
ton more money to throw money at it because they have not 
planned? And I know that is general, so----
    Mr. Powner. Historically, what has happened is because we 
have the immovable deadline, you throw more money at it. 
Historically, that is what has happened.
    Senator Coburn. If you go back to this number from the 2000 
census--did you say it was $1 billion, the reformatting because 
of the Supreme Court decision cost, a portion of that $1.4 
billion or----
    Mr. Scire. The total additional cost is $1.5 billion, I 
think.
    Senator Coburn. OK. So you take $1.5 billion away from $6.3 
billion, you have $4.8 billion. So we are talking about 
almost--that is not going to happen this time. We know what is 
out there. So what you are really talking about is a 300-
percent increase for this census over the one from 2000. And 
that is in spite of advanced technology in this country.
    By canceling the Lockheed Martin Internet contract, that is 
the reverse of requirement creep. That goes the opposite. Yet 
we did not see any savings from it. We are seeing increased 
costs; whereas, we save money here.
    I am going to go back to Senator Carper's question. Can you 
perceive--kind of like our challenge to the American public, 
can you come up with a way to collect the census for less than 
$90? Go to my website, coburn.senate.gov/ffm, sign on, and we 
will give you the instructions on how you help us do oversight 
in the Federal Government.
    Can you perceive of a way where we could interject, either 
in a pilot study or another study, where we could have secure, 
advanced Internet responses to the census that is promoted 
prior to 2010? Can you imagine that in the realm of 
possibilities?
    Mr. Powner. Absolutely. I mean, security, that is not an 
issue. There are various methods of encrypting that----
    Senator Coburn. Technology is not a----
    Mr. Powner. The technology is there. How much time do you 
need to set it up, to test various interfaces and those types 
of things that will need to be put in place, especially with 
integrating that with the other data. But, sure, that is a 
possibility.
    Senator Coburn. All these firms that Mr. Reamer talked 
about who use census data to make business and economic 
decisions also employ a slew of private contractors who do 
exactly the same thing as the census. In other words, they are 
data collectors, they are screeners, they are survey takers. 
And, Mr. McTigue, you might respond to this. Is it off the wall 
to think that we couldn't do this in the private sector more 
efficiently, cheaper, and better, and in an accurate way if, in 
fact, it was legal to do so?
    Mr. McTigue. Well, certainly I am prepared to respond and 
say yes, of course, there are many organizations in the private 
sector that gather large quantities of highly secure data and 
handle it very securely. We would all be very unhappy if our 
banking information went astray, but much of it is handled 
online. Insurance information is handled online. A lot of 
medical information is sent around electronically, and it is 
done very securely.
    But there are other functions as well. For example, credit 
rating agencies gather huge quantities of information, and they 
send that product on to others who use it, and it is done in a 
very secure way. And that is information that we would be very 
unhappy if it became public, and it is very rare that it 
actually does. So the capability is out there. Do they actually 
do censuses? No. But they do lots of other counting.
    Even if you looked at a major grocery chain, every night 
they take the information from the cash registers, which 
supplies the information to restock all of the chain the 
following day, and that is extremely precise.
    So there are systems and capabilities out there and the 
experience to be able to do it. Can you do it between now and 
2010 if you are allowed to? I think the answer is no, it is 
getting too close. But certainly that capability would exist in 
the private sector.
    Senator Coburn. Let me ask our GAO panelists, given what 
you are looking at, given the canvassing, the runs that we have 
seen, when should we look at this again as a Subcommittee, your 
recommendation, to be able to have the most impact on the 
Census Bureau, knowing that we are going to look at it? In 
other words, when would you recommend we come back up and have 
this discussion again to see if we are meeting any of the 
earmarks that you all see as deficit now?
    Mr. Scire. Well, I think that Mr. Kincannon mentioned this 
upcoming August operational plan. What we had seen back in 
December was a research and development plan, and that had 
detailed milestones. It went through, I think, 2006, actually, 
not beyond. And that is why I have been looking forward to this 
kind of operational plan that would lay out the 
interrelationships among the operations. Looking at that might 
provide a baseline for understanding what are the key points, 
what are the key risks that the Bureau faces in Census 2010. 
And so that might help provide a road map, and looking at that 
road map itself I think would be a valuable oversight exercise.
    Senator Coburn. Thank you, Mr. Chairman.
    Senator Carper. You bet.
    Mr. Powner, let me just go back and ask you a question, and 
maybe a question or two for Mr. Reamer.
    Mr. Powner, correct me if I am wrong, but I believe you 
worked in the private sector and have some experience managing 
large procurements. Is that a fair statement?
    Mr. Powner. Yes, Mr. Chairman.
    Senator Carper. I also believe you have spent some time 
examining large IT projects in the Federal Government. Is that 
correct?
    Mr. Powner. Correct.
    Senator Carper. OK. Thinking back on some of that 
experience, some of what you have seen elsewhere, what 
mistakes--when you look at what the Census Bureau is doing as 
we approach the 2010 census, any mistakes that--even if we have 
already talked about them, or if we have not--some mistakes 
that you think they may be making that we just ought to stop 
for a moment and say, ``This is a mistake''?
    Mr. Powner. Well, I have mentioned five areas that we would 
like to see greater rigor. One is requirements. We still have 
moving baselines and requirements creep. We need to solidify 
those requirements as soon as possible on these major 
acquisitions. That is one.
    The other item is there are many interfaces involved with 
all of these systems. It is going to be very important that we 
identify all these interfaces and ensure that those interfaces 
are tested adequately. And then from a testing perspective, 
because not everything is going to be tested in the dress 
rehearsal, we still do not see some of the post-dress rehearsal 
test plans in place. It is not only--I mean, we are concerned 
about the execution of the plans, but we ought to have those 
plans in place now.
    Executive level involvement, one of the major issues where 
IT projects go awry is the executives are not as engaged in 
mitigating risks. And as you well know, we are looking at risk 
management of these major acquisitions for you, and there are a 
couple items with risk management that we are concerned about. 
One is, are all the risks clearly identified? I can tell you 
right now on the FDCA contract, the performance issues on the 
handhelds.
    Senator Carper. Which contract?
    Mr. Powner. This is the Field Data Collection, which 
includes the handhelds. If you look at the risk logs, the 
software problems associated with the handhelds are not listed 
on their risk log. So there is a concern about the completeness 
of those risks. There is a concern about having the appropriate 
mitigation plans in place. And, finally, getting those key 
risks reported to the executives, we do not see evidence on all 
the projects that the executives are going through the 
appropriate reviews.
    Senator Carper. Alright. Thank you.
    Mr. Reamer, are you satisfied with the steps that the 
Census Bureau has taken to address some of the problems in the 
account that showed up in 2000? Do you think they are doing 
what they need to do to make the 2010 census more accurate and 
more inclusive of those segments of our population that are 
historically undercounted? Any ideas where they might need to 
make some improvement in this regard?
    Mr. Reamer. I am not a methodological expert in terms of 
the collection process, but from what I have read and what I 
understand, yes, I think they are making some improvements.
    In any census there is an undercount and a double count, 
and I think they are taking steps now to remove the double 
count where people can be counted in two----
    Senator Carper. How do they do that?
    Mr. Reamer. I am actually not familiar with the details. I 
think a different approach in terms of the application of 
residence rules to help people understand who they are and who 
they are not to include as they fill out the form. The Bureau 
can give you great detail on this.
    Senator Carper. OK.
    Mr. Reamer. So that leaves the undercount. I think the 
Community Partnership Program was quite successful in trying to 
reduce the undercount. There is a need for that program going 
forward.
    Senator Carper. OK.
    Mr. Reamer. Sorry. One other thing.
    Senator Carper. Go ahead.
    Mr. Reamer. Which is improvements in the LUCA program. The 
first time the Census Bureau was required to do LUCA was in 
2000, and it was an uneven process. You needed a lot of 
resources as a city or a town to actually comply with the 
Census Bureau process. You could not call in the State 
government to help you out, and so I think the participation in 
LUCA was lower than it could have been. Places like New York--I 
mentioned New York added a third of million plus households 
through the LUCA process--were able to do it. They have the 
resources. But towns of 10,000 could not.
    This time around, I think, the Bureau has learned from its 
experience and has significantly improved the LUCA process, 
opened up the options that localities can use. State 
governments can come in and help. State governments have lots 
of resources, such as drivers' records, that can help augment 
the review of addresses. And the Bureau I think has improved 
training, as I understand it, and lengthened the time. So my 
expectation--and I think the expectation of observers--is that 
LUCA will do better this time around. This is an instance where 
the Bureau has learned a lot from its first experience and so 
should provide a more accurate Master Address File.
    Senator Carper. OK. Let me just say to each of our 
witnesses, if anybody has a short closing thought you would 
like to leave with us, that would be fine. Anybody at all?
    Mr. McTigue. Mr. Chairman, if I could just make one 
additional comment, and it would be that it would be nice if 
some entrepreneurial thinking at the Census Bureau looked 
either at the Community Survey or at the 2010 census on the 
presumption that in 2020 most of us will be filling in our 
census forms online, and using that as an experimental basis, 
either to run in parallel, the experiment would identify these 
are the things that we need to answer between now and 2020 if 
we are going to do that. But even more importantly, being able 
to get better responses from the Community Survey, which is 
conducted every month online, would be a good harbinger of 
having the potential capacity to do all of the census on line 
in 2020.
    Senator Carper. Thank you for that thought.
    Anyone else a closing thought?
    Mr. Scire. Yes, if I might add, part of what we are seeing 
is that the introduction of technology also introduces risks, 
and it is not enough to identify those risks. It is also 
necessary to develop risk mitigation strategies and 
alternatives should that technology not perform as expected.
    The other is that in terms of oversight, it is important to 
make more transparent what the plans are and what the costs are 
and how sensitive they are to underlying assumptions.
    Senator Carper. Alright. Anyone else? Mr. Reamer.
    Mr. Reamer. Just a suggestion, that a lot of the 
conversation here has focused on cost, and it is appropriate to 
focus on cost. I believe one of my roles here today was to 
place cost in the context of return on the Nation's 
investment--to encourage people to think about that return in 
terms of political governance of the country and in terms of 
the entire economy. It is quite remarkable the return we get on 
the census, and clearly the cost should be appropriate, but I 
want to suggest keeping the ROI in mind as well.
    Senator Carper. Alright. Thank you. Mr. Powner.
    Mr. Powner. I will just further comment on what Mr. Scire 
said. I think it is about risk mitigation at this point and 
really managing and tracking those costs down to fine details, 
so that when we see little blips in the process that we 
effectively jump on those risks and attempt to move them in the 
other direction.
    Senator Carper. To my colleague, Dr. Coburn, any closing 
thought?
    Senator Coburn. No. I just would submit for the record that 
the $600 million contract for the handheld device, it was 
another cost-plus contract, which I think we are going to see--
I hate to be the prophet of doom in terms of success or 
increased cost, but had we had one out there where we said here 
is the performance, you get paid when it works, and here is the 
fixed price you get paid, we would have seen faster response, 
better quality, rather than on the concept of cost-plus, 
because people do not have to be responsible when it is cost-
plus.
    Senator Carper. Well, I hope experience proves you wrong. I 
will be delighted--and I am sure you will be, too--if that does 
not turn out to be the case.
    Let me again thank our witnesses. I have been calling Dr. 
Kincannon, Mr. Kincannon, most of the afternoon. Who is here 
from the Census Bureau? Somebody is still here? Which is it--
Dr. Kincannon or Mr. Kincannon? Mister. So I was right? 
Everybody else was wrong. That is a first. Mister or Doctor--we 
are glad he was here.
    Our thanks to each and every one of you for being here with 
us today and for your testimony, actually quite helpful 
testimony, and for your responses to our questions. The census 
is a big deal, and we are reminded of that every 10 years that 
it is important to our country, not just for the folks in the 
U.S. House of Representatives or those who might want to run 
for the U.S. House of Representatives, not only for governors 
and legislators, State legislators, that are trying to figure 
out how to apportion their State's voting districts, but also 
for our Nation's economy and for a whole host of other reasons. 
It is important we get it right. It costs a lot of money, and 
it is important that we spend that money wisely.
    This Subcommittee has a number of responsibilities, but one 
of them is to try to make sure that we get it right, and we 
appreciate your help in enabling us to meet our 
responsibilities in that regard.
    There may be a question or two that will follow in writing, 
and if there is, we would appreciate very much your timely 
response to those questions.
    Good to be with all of you this afternoon. Thank you for 
joining us, and with that, this hearing is adjourned.
    [Whereupon, at 4:44 p.m., the Subcommittee was adjourned.]


                            A P P E N D I X

                              ----------                              

[GRAPHIC] [TIFF OMITTED] T7360.001

[GRAPHIC] [TIFF OMITTED] T7360.002

[GRAPHIC] [TIFF OMITTED] T7360.003

[GRAPHIC] [TIFF OMITTED] T7360.004

[GRAPHIC] [TIFF OMITTED] T7360.005

[GRAPHIC] [TIFF OMITTED] T7360.006

[GRAPHIC] [TIFF OMITTED] T7360.007

[GRAPHIC] [TIFF OMITTED] T7360.008

[GRAPHIC] [TIFF OMITTED] T7360.009

[GRAPHIC] [TIFF OMITTED] T7360.010

[GRAPHIC] [TIFF OMITTED] T7360.011

[GRAPHIC] [TIFF OMITTED] T7360.012

[GRAPHIC] [TIFF OMITTED] T7360.013

[GRAPHIC] [TIFF OMITTED] T7360.014

[GRAPHIC] [TIFF OMITTED] T7360.015

[GRAPHIC] [TIFF OMITTED] T7360.016

[GRAPHIC] [TIFF OMITTED] T7360.017

[GRAPHIC] [TIFF OMITTED] T7360.018

[GRAPHIC] [TIFF OMITTED] T7360.019

[GRAPHIC] [TIFF OMITTED] T7360.020

[GRAPHIC] [TIFF OMITTED] T7360.021

[GRAPHIC] [TIFF OMITTED] T7360.022

[GRAPHIC] [TIFF OMITTED] T7360.023

[GRAPHIC] [TIFF OMITTED] T7360.024

[GRAPHIC] [TIFF OMITTED] T7360.025

[GRAPHIC] [TIFF OMITTED] T7360.026

[GRAPHIC] [TIFF OMITTED] T7360.027

[GRAPHIC] [TIFF OMITTED] T7360.028

[GRAPHIC] [TIFF OMITTED] T7360.029

[GRAPHIC] [TIFF OMITTED] T7360.030

[GRAPHIC] [TIFF OMITTED] T7360.031

[GRAPHIC] [TIFF OMITTED] T7360.032

[GRAPHIC] [TIFF OMITTED] T7360.033

[GRAPHIC] [TIFF OMITTED] T7360.034

[GRAPHIC] [TIFF OMITTED] T7360.035

[GRAPHIC] [TIFF OMITTED] T7360.036

[GRAPHIC] [TIFF OMITTED] T7360.037

[GRAPHIC] [TIFF OMITTED] T7360.038

[GRAPHIC] [TIFF OMITTED] T7360.039

[GRAPHIC] [TIFF OMITTED] T7360.040

[GRAPHIC] [TIFF OMITTED] T7360.041

[GRAPHIC] [TIFF OMITTED] T7360.042

[GRAPHIC] [TIFF OMITTED] T7360.043

[GRAPHIC] [TIFF OMITTED] T7360.044

[GRAPHIC] [TIFF OMITTED] T7360.045

[GRAPHIC] [TIFF OMITTED] T7360.046

[GRAPHIC] [TIFF OMITTED] T7360.047

                                 
