[House Hearing, 113 Congress]
[From the U.S. Government Publishing Office]



 
            ENSURING AN ACCURATE AND AFFORDABLE 2020 CENSUS

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

                                HEARING

                               before the

                   SUBCOMMITTEE ON FEDERAL WORKFORCE,

                    US POSTAL SERVICE AND THE CENSUS

                                 of the

                         COMMITTEE ON OVERSIGHT

                         AND GOVERNMENT REFORM

                        HOUSE OF REPRESENTATIVES

                    ONE HUNDRED THIRTEENTH CONGRESS

                             FIRST SESSION

                               __________

                           SEPTEMBER 11, 2013

                               __________

                           Serial No. 113-58

                               __________

Printed for the use of the Committee on Oversight and Government Reform


         Available via the World Wide Web: http://www.fdsys.gov
                      http://www.house.gov/reform




                  U.S. GOVERNMENT PRINTING OFFICE
82-932                    WASHINGTON : 2013
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing 
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; DC 
area (202) 512-1800 Fax: (202) 512-2104  Mail: Stop IDCC, Washington, DC 
20402-0001



              COMMITTEE ON OVERSIGHT AND GOVERNMENT REFORM

                 DARRELL E. ISSA, California, Chairman
JOHN L. MICA, Florida                ELIJAH E. CUMMINGS, Maryland, 
MICHAEL R. TURNER, Ohio                  Ranking Minority Member
JOHN J. DUNCAN, JR., Tennessee       CAROLYN B. MALONEY, New York
PATRICK T. McHENRY, North Carolina   ELEANOR HOLMES NORTON, District of 
JIM JORDAN, Ohio                         Columbia
JASON CHAFFETZ, Utah                 JOHN F. TIERNEY, Massachusetts
TIM WALBERG, Michigan                WM. LACY CLAY, Missouri
JAMES LANKFORD, Oklahoma             STEPHEN F. LYNCH, Massachusetts
JUSTIN AMASH, Michigan               JIM COOPER, Tennessee
PAUL A. GOSAR, Arizona               GERALD E. CONNOLLY, Virginia
PATRICK MEEHAN, Pennsylvania         JACKIE SPEIER, California
SCOTT DesJARLAIS, Tennessee          MATTHEW A. CARTWRIGHT, 
TREY GOWDY, South Carolina               Pennsylvania
BLAKE FARENTHOLD, Texas              MARK POCAN, Wisconsin
DOC HASTINGS, Washington             TAMMY DUCKWORTH, Illinois
CYNTHIA M. LUMMIS, Wyoming           ROBIN L. KELLY, Illinois
ROB WOODALL, Georgia                 DANNY K. DAVIS, Illinois
THOMAS MASSIE, Kentucky              TONY CARDENAS, California
DOUG COLLINS, Georgia                STEVEN A. HORSFORD, Nevada
MARK MEADOWS, North Carolina         MICHELLE LUJAN GRISHAM, New Mexico
KERRY L. BENTIVOLIO, Michigan
RON DeSANTIS, Florida

                   Lawrence J. Brady, Staff Director
                John D. Cuaderes, Deputy Staff Director
                    Stephen Castor, General Counsel
                       Linda A. Good, Chief Clerk
                 David Rapallo, Minority Staff Director

 Subcommittee on Federal Workforce, U.S. Postal Service and the Census

                   BLAKE FARENTHOLD, Texas, Chairman
TIM WALBERG, Michigan                STEPHEN F. LYNCH, Massachusetts, 
TREY GOWDY, South Carolina               Ranking Minority Member
DOUG COLLINS, Georgia                ELEANOR HOLMES NORTON, District of 
RON DeSANTIS, Florida                    Columbia
                                     WM. LACY CLAY, Missouri


                            C O N T E N T S

                              ----------                              
                                                                   Page
Hearing held on September 11, 2013...............................     1

                               WITNESSES

Mr. John Thompson, Director, U.S. Census Bureau
    Oral Statement...............................................     5
    Written Statement............................................     8
Mr. Robert Goldenkoff, Director, Strategic Issues, U.S. 
  Government Accountability Office
    Oral Statement...............................................    14
    Written Statement............................................    17
Ms. Carol Cha, Director, Information Technology, U.S. Government 
  Accountability Office
    Oral Statement...............................................    15
    Written Statement............................................    17

                                APPENDIX

The Honorable Stephen Lynch, a Member of Congress from the State 
  of Massachusetts, Opening Statements...........................    56


            ENSURING AN ACCURATE AND AFFORDABLE 2020 CENSUS

                              ----------                              


                     Wednesday, September 11, 2013

                  House of Representatives,
    Subcommittee on Federal Workforce, U.S. Postal 
                           Service, and the Census,
              Committee on Oversight and Government Reform,
                                                   Washington, D.C.
    The subcommittee met, pursuant to call, at 9:31 a.m., in 
Room 2154, Rayburn House Office Building, Hon. Blake Farenthold 
[chairman of the subcommittee] presiding.
    Present: Representatives Farenthold, Walberg, Collins, 
Lynch, Norton, and Clay.
    Also Present: Representative Maloney.
    Staff Present: Molly Boyl, Parliamentarian; Lawrence J. 
Brady, Staff Director; Caitlin Carroll, Deputy Press Secretary; 
John Cuaderes, Deputy Staff Director; Adam P. Fromm, Director 
of Member Services and Committee Operations; Linda Good, Chief 
Clerk; Michael R. Kiko, Staff Assistant; Jeffrey Post, Senior 
Professional Staff Member; Laura Rush, Deputy Chief Clerk; 
Scott Schmidt, Digital Director of Digital Strategy; Peter 
Warren, Legislative Policy Director; Jaron Bourke, Minority 
Director of Administration; Courtney Cochran, Minority Press 
Secretary; Kevin Corbin, Minority Professional Staff Member; 
Yvette Cravins, Minority Counsel; Devon Hill, Minority Research 
Assistant; Jennifer Hoffman, Minority Press Secretary; and Mark 
Stephenson, Minority Director of Legislation.
    Mr. Farenthold. The subcommittee will come to order. As is 
normal, we will begin the hearing by reading the Oversight 
Committee mission statement. We exist to secure two fundamental 
principles: First, that Americans have a right to know the 
money Washington takes from them is well spent; and second, 
Americans deserve an efficient, effective government that works 
for them. Our duty on the Oversight and Government Reform 
Committee is to protect these rights. Our solemn responsibility 
is to hold the government accountable to taxpayers, because the 
taxpayers have the right to know what they get from their 
government.
    We will work tirelessly in partnership with citizens 
watchdogs to deliver the facts to the American people and bring 
genuine reform to the Federal bureaucracy. This is the mission 
of the Oversight and Government Reform Committee.
    At this time, I'll recognize myself for an opening 
statement. The costs of the decennial census has gone up 680 
percent per household since 1960, and the 2010 Census was 
almost twice as expensive per household as the 2000 Census. The 
2010 Census reached $14.7 billion in cost. If the recent trends 
hold true, the 2020 Census could cost the American taxpayers 
$25 billion or more. This is simply unacceptable.
    As government watchdogs, it's our job to make sure the 
American taxpayers' money is being used appropriately and 
efficiently.
    The purpose of today's hearing is to find ways the Bureau 
can achieve significant cost savings without impacting the 
quality and integrity of the count. The most obvious of these 
is adding an Internet response option to the Census, which is 
now being offered on a nationwide basis for the American 
Community Survey. Without getting into the pros and cons of the 
community survey, it does show that technology can be better 
used to gather Census data.
    We should also look at the digitalization of follow-up 
efforts for those who do not self-respond to the Census. In 
fact, new technology has created an opportunity to completely 
redesign the model. Through the Bring Your Own Device model, 
the Census Bureau has the opportunity to have its temporary 
field staff simply download a secure application to a smart 
phone that will not only allows them to record response 
information, but also has live updates with which house they 
should visit next and possibly GPS data, directions, adjusting 
for traffic and many other things. Collecting Census data, 
there should be an app for that.
    Another issue with respect to cost savings is expanding the 
use of administrative records. What are the cost savings of 
this and what are the privacy issues associated with inter-
governmental data sharing? Administrative records do help 
improve data accuracy and reduce the need for multiple 
nonresponse follow-ups.
    A redesign mapping project focusing on continuously 
tracking changes and partnering with the public and private 
geospatial entities and perhaps even the Postal Service could 
meet--I'm sorry--could eliminate the need for a nationwide 
canvass in upcoming years.
    We also face challenges from public distrust of the 
government in light of recent scandals currently being 
investigated by the full committee. The IRS targeting of groups 
based on political ideology; the NSA's rampant data collection 
and leaks; the administration spokeswoman, Susan Rice, 
misleading the public after the Benghazi attack all have had 
effects making the American public more distrustful of the 
government and also less likely, I believe, and based on the 
information that's coming into my office, less likely to 
participate and cooperate with the government.
    Take Lois Lerner. She's probably costing the taxpayers 
money because people are fearful how their personal information 
revealed to the Census service might be used against them, so 
they're not filling out their survey, meaning the government 
has to send out people to follow up with them. How will all of 
these affect the response rate and accuracy in 2020? Add to 
that the detailed and, in my opinion, intrusive questions asked 
on the American Community Survey, and we're facing additional 
challenge that potentially add costs.
    Finally, we must address in a cost-effective manner the 
question of accuracy, especially dealing with non-traditional 
households: the homeless, those living in the shadows and non-
English speakers. We cannot craft solutions to problems like 
homelessness and undocumented residents if we do not have 
accurate data.
    New cost saving ideas are attainable, while regaining the 
public trust might be more difficult. A National Academy of 
Sciences panel has stated that a realistic goal is for the 2020 
Census to cost less than the 2010 Census on a per-household 
basis. I think that's doable and I support this goal, and I 
challenge the Census Bureau and the entire government to make 
this happen. The Census coming in under budget would send a 
strong signal to the American people about the stewardship of 
their tax dollars, and perhaps be a way we can regain some of 
the trust between the government and the governed.
    With this in mind, I hope this hearing starts a dialogue 
between the committee and the Census Bureau so that we can make 
the 2020 Census a cost-effective success.
    And with that, I'd like to welcome our witnesses and thank 
you for appearing before us. We will now recognize the 
distinguished ranking member, the gentleman from Massachusetts, 
Mr. Lynch, for his opening statement.
    Mr. Lynch. Thank you, Mr. Chairman. At the outset, I'd also 
like to congratulate Director Thompson on being confirmed as 
the 24th director of the United States Census Bureau. We look 
forward to working with you and your staff as we continue to 
examine Census operations.
    I'd like to also thank Mr. Goldenkoff and Ms. Cha of the 
Government Accountability Office for helping this committee 
with its work today. Thank you.
    This is a critical moment for the United States Census 
Bureau. While the next decennial Census may be 7 years away, 
the agency is already in the midst of researching, testing and 
implementing a variety of technological and procedural reforms 
designed to better ensure a 2020 Census that maximizes cost 
containment without compromising data quality.
    I would note that these important initiatives come on the 
heals of a 2010 Census that according to GAO, and as the 
chairman has pointed out, was the most expensive U.S. Census in 
history at a cost of approximately $13 billion. Now, that was a 
56 percent increase and an $8 billion cost increase from the 
2000 Census, so the rate of increase is alarming.
    Chief among the lessons learned from the 2010 Census that 
GAO has consistently identified as key areas for reform is the 
need for the Bureau to reexamine its approach to how it takes 
the Census. In particular, 2010 witnessed the Bureau employing 
the same mail-out, mail-back and door visit Census form process 
that it has been using since 1970, despite a population that 
has become increasingly demographically diverse and 
significantly concerned about personal privacy. If I were a 
cynic, I would--I would suggest that we might shift the 
responsibility of the Census over to the NSA since they seem to 
be taking--keeping track of everybody anyway.
    But according to the GAO, this basic design of the 
decennial Census is no longer capable of a cost-effective 
enumeration, and without changes, future enumerations could be 
fiscally unsustainable.
    To its credit, the Census Bureau is already making changes 
to reform its approach to Census taking, and I'm greatly 
encouraged by its responsiveness to the recommendations issued 
by GAO. Specifically the Census Bureau is currently examining 
whether it could achieve meaningful cost savings by utilizing 
State and local agency administrative records, including 
driver's license and school documentation to identify persons 
associated with a particular household address.
    In addition, the Census Bureau is researching the viability 
of electronic data collection and technology receptive 
promotional methods, including the use of email, text messaging 
and social networking sites. Moreover, the agency is testing 
the feasibility of using its monthly housing and population 
``American Community Survey'' data as a basis for conducting 
smaller and more frequent Census-related tests throughout the 
decade in place of broader, more expensive and one-time-only 
tests that have defined previous decennial Census-taking 
cycles.
    I would also interject that I know Mr. Chaffetz and I had 
discussed in the previous Congress the opportunity to utilize 
the United States Postal Service, because postal workers go to 
each and every home in America 6 days a week, and that does 
replicate in a way the door-to-door survey that the Census 
conducts, albeit we would be using existing employees who go 
door to door on a regular basis and might have a better feel 
for the number of residents in a particular household.
    As reported by GAO in January of this year, the various 
reform initiatives underway at the Census Bureau and 
particularly the agency's expanding use of technology must be 
accompanied by a corresponding effort to implement effective 
information security policies and protocols. I share the 
chairman's concern regarding the current lack of comprehensive 
security programs at the Census Bureau that includes identified 
security risks, up-to-date security management program 
policies, meaningful information, security training for Census 
employees, and effective incidence response guidance. However, 
I am confident that the Bureau will continue to work with GAO 
to take the necessary steps to enhance its agency-wide security 
program and implement proper safeguards against the 
unauthorized use or disclosure of sensitive personal 
information.
    Additionally, I'd like to note that the Census Bureau, like 
every other Federal agency, is already feeling the impact of 
the sequestration and is currently operating with a budget that 
is 13 percent less than the administration's request for 2013. 
Coupled with chronic underfunding over the past several years, 
the agency is facing a variety of difficult decisions that pose 
a challenge to its mission, from reduction in field tests and 
increased reliance on existing data, to delays in planning for 
the 2020 decennial Census and the suspension of the country's 
economic Census. Importantly, these and other Census activities 
are mission critical and result in data that is used to 
determining funding allocations for community services, 
infrastructure and neighborhood improvements as well as public 
health and educational program and forecasting economic and 
employment conditions.
    These efforts are essential to our national economy and 
policymaking, and I'd urge my colleagues to keep the importance 
of the Census Bureau's mission in mind as we continue to 
examine Census operations.
    Mr. Chairman, I thank you for holding this important 
hearing and I look forward to the discussion these and other 
issues will come up today with today's witnesses. I yield back 
the balance of my time.
    Mr. Farenthold. Thank you, Mr. Lynch. Our other members 
will have 7 days to submit opening statements for the record. 
We will now introduce and recognize our panel.
    The Honorable John Thompson is the director of the United 
States Census Bureau. I add my congratulations as well on your 
confirmation. Mr. Robert Goldenkoff is the director of 
Strategic Issues Group at the U.S. Government Accountability 
Office. And our technical expert here is Carol Cha. She's the 
director of IT Group at the United States Government 
Accountability Office. Welcome.
    Pursuant to the committee rules, all witnesses will be 
sworn in before they testify. Will you all please stand up and 
raise your right hand. Do you solemnly swear or affirm that the 
testimony you are about to give will be the truth, the whole 
truth and nothing but the truth?
    Let the record reflect that all the witnesses answered in 
the affirmative. You all may be seated.
    In order to allow time for discussion, we'd particularly 
like to ask our questions up here and feel like we get the 
information we're most interested in, we would ask that you 
would limit yourself to around 5 minutes with your statement. 
We have in the record and hopefully everybody up here has 
reviewed the written material that was provided to us. And so 
if you would limit yourself to 5 minutes. We've got the little 
traffic light system in front of you. The green light means 
you're good to go, the yellow light means you're running out of 
time, so hurry up, and the red light means stop. There will be 
no tickets issued, but I will start looking sternly at you when 
the red light comes on. So we will get underway.
    We will now start with Mr. Thompson. Mr. Thompson.

                       WITNESS STATEMENTS

                   STATEMENT OF JOHN THOMPSON

    Mr. Thompson. Thank you, Chairman Farenthold, Ranking 
Member Lynch, and members of the subcommittee. I am pleased to 
have the opportunity to appear today before the subcommittee to 
discuss 2020 Census operations. I'll make a brief statement. I 
will ask that my full statement be added for the record. Let me 
also say that I am delighted to return to the Census Bureau as 
director, and I'm looking forward to working with the 
committee.
    A democracy needs credible, objective and timely 
information, and the decennial Census is one of the important 
sources of this information. The 2020 Census will continue this 
tradition, but as with each Census, we must consider new 
opportunities with respect to information technology.
    I believe current plan activities will support a Census 
that is significantly less costly and will deliver faster 
results than the 2010 Census. I can pledge the best efforts of 
the dedicated public servants of the agency, but we must also 
ask the Congress for its support to ensure the Census Bureau 
can face these challenges.
    Among the most promising options we are researching to 
accomplish our goals are, first, re-engineering the field data 
collection operations; second, making better use of information 
previously collected by Federal and State agencies; third, 
using the Internet as the primary self-response option; and 
fourth, drawing on geographic tools and data sets to eliminate 
the need to physically canvass large portions of the United 
States.
    One of the major opportunities to reduce costs lies with 
applying operations research methods to work management and 
route planning. We are exploring techniques to take advantage 
of automation to conduct our field work more efficiently, 
including routing our interviewers, accounting for optimum 
times to visit, tracking travel distances and traffic, and 
other factors to minimize extensive travel and wasted visits.
    We are also exploring adaptive design techniques that will 
help us supply statistical methodology to do our work smarter. 
The objective of this effort is to re-engineer our field 
management processes, thereby reducing both the hours that our 
interviewers will spend on collecting the data from households 
that do not take advantage of the self-response option and 
reducing the infrastructure required to support these efforts.
    In addition to our efforts to re-engineer the field data 
collection, we are looking at other alternatives to reduce the 
non-response follow-up workload. One of the promising 
innovations is the use of data people have already given to the 
government to enumerate households that do not return Census 
questionnaires. These data range from information about vacant 
units collected by the U.S. Postal Service, to information that 
is collected by various Federal and State agencies to 
administer and support the programs.
    We also have important opportunities to reduce the non-
response follow-up workload by improving self-response. The 
traditional way of responding through the mail, then following 
up through an interview is inefficient. The Internet in 
contrast is becoming an increasing important tool for self-
response.
    For the 2020 Census, the Census Bureau is actively testing 
strategies to encourage and motivate the use of the Internet as 
the primary response option. The geographic support services 
initiative at the Census Bureau is directed at obtaining the 
highest quality address and geospatial data.
    We are looking to expand partnerships with tribal, State 
and local governments as well as with the private sector. These 
partnerships will be critical in allowing the Census Bureau to 
develop a more focused address canvassing program that will 
allow the Census Bureau to avoid re-canvassing large areas of 
the country for which there are no changes or which--or for 
which information can be obtained from an alternative source.
    Finally, I would like to note that all these options are 
buttressed by the Census Bureau's efforts to more effectively 
integrate IT services and systems. Instead of building 
different IT systems that serve single programs or even single 
Census operations, we are building systems to share across the 
enterprise.
    We have also virtualized our servers, building a private 
cloud environment which we're using to deliver a virtual 
desktop infrastructure, or VDI, that allows us to decouple a 
user's device from any sensitive data. This capability allows 
us to dramatically increase our telework program. Now employees 
can use their devices and we do not need to furnish equipment. 
We believe this capacity puts us in the strong footing to 
consider Bring Your Own Device, or BYOD, as a real possibility 
for the 2020 Census.
    In the 2010 Census, we used a custom-built mobile device 
and custom-built software to run on that device. Now we have a 
team working not only on the technological considerations and 
security requirements, but also the personnel policies that 
would have to accompany a BYOD approach for 2020 based on 
utilization of off-the-shelf technology and software.
    Finally, the budget requested for FY 2014 includes an 
increase of $150.7 million for the 2020 Census. The Census 
Bureau will need this to conduct the necessary research and 
testing efforts to prepare for 2020 at this point. The work in 
fiscal 2014 is critical to meet our schedule to produce 
analyzed data in time for key decision-making. The next 2 years 
represent an extremely critical period in the 2020 planning 
development cycle. We cannot delay the work to begin developing 
our major systems in fiscal year 2016, therefore, our 
preparations for this effort over the next 2 fiscal years are 
of the highest priority.
    To meet these challenges will require the best efforts of 
the Census Bureau, and we are looking forward to working with 
the Congress so they can provide the support to meet these 
challenges successfully.
    Thank you very much, and I would be happy to respond to 
questions.
    Mr. Farenthold. Thank you, Mr. Thompson.
    [Prepared statement of Mr. Thompson follows:]


    [GRAPHIC] [TIFF OMITTED] 82932.001
    
    [GRAPHIC] [TIFF OMITTED] 82932.002
    
    [GRAPHIC] [TIFF OMITTED] 82932.003
    
    [GRAPHIC] [TIFF OMITTED] 82932.004
    
    [GRAPHIC] [TIFF OMITTED] 82932.005
    
    [GRAPHIC] [TIFF OMITTED] 82932.006
    
    Mr. Farenthold. Before we go on to Mr. Goldenkoff, I see 
the gentlelady from New York, Ms. Maloney, has joined us. She 
is not officially a member of this subcommittee, so at this 
point, I'd like to ask unanimous consent that she be authorized 
to participate in this hearing. Without objection, so ordered.
    We do have two other speakers at the table. It's my 
understanding that on behalf of the GAO, Mr. Goldenkoff will 
speak and Ms. Cha is here because they realized some of us up 
here are tech geeks and might want to get a little more into 
the weeds than might otherwise happen. So at this point we'll--
we'll recognize Goldenkoff from the GAO.

                 STATEMENT OF ROBERT GOLDENKOFF

    Mr. Goldenkoff. Thank you. Chairman Farenthold, Ranking 
Member Lynch, and members of the subcommittee, I appreciate the 
opportunity to be here today to discuss the progress the Census 
Bureau is making in controlling the cost of the 2020 
enumeration.
    As you well know, the fundamental challenge facing the 
Census Bureau is how to cost-effectively count a population 
that is growing steadily larger, more diverse, increasingly 
difficult to find, and less willing to participate in the 
decennial.
    The cost of counting each housing unit is no longer 
fiscally sustainable. In 2010 dollars, the Bureau spent around 
$16 to enumerate each housing unit in 1970, compared to $98 in 
2010. Much like running up a down escalator, with each decade, 
the Bureau has had to invest substantially more resources 
simply to match the results of previous enumerations.
    In our past work, we noted that to control costs while 
maintaining accuracy, actions were needed in at least three 
areas: first, the Bureau needs to transform itself into a high 
performing organization; second, the Bureau needs to re-
engineer key Census-taking operations; and, third, the Bureau 
needs to strengthen its IT management and security practices.
    Joining me this morning, as was already mentioned, is Carol 
Cha, director of GAO's IT acquisition management team. And our 
remarks will focus on the status of the cost containment 
initiatives within each of those three areas. Specifically, 
we'll highlight the progress the Bureau has made to date and 
what still needs to be done going forward.
    I would also like to take this opportunity to welcome John 
Thompson back to the Census Bureau as its newest chief 
executive and wish him every success going forward. John's 
experience, both inside and outside the Bureau combined with 
the new 5-year fixed term of office for the Census director 
should help provide the expertise and continuity of leadership 
that will be important for keeping decennial preparations on 
track in the years ahead.
    Overall, we found that the Bureau's plans for controlling 
enumeration costs show promise if effectively implemented. At 
the same time, the Bureau's plans contain a number of open 
questions that will need to be addressed for a successful head 
count in 2020. For example, in the area of organizational 
transformation, the Bureau has restructured its entire 
decennial directorate in order to improve collaboration and 
communication across its divisions, improve operational 
efficiencies, and instill a more innovative culture.
    The Bureau is also developing agency-wide standards, 
guidance and tools in such areas as risk management and IT 
investment to reduce duplicative efforts across the Bureau.
    These and other actions the Bureau has taken are all 
important steps forward, however, several initiatives require 
additional attention. For example, in response to our 
recommendations, the Bureau has created a cost estimation team 
reporting to the director and recently hired an individual to 
lead that group. Going forward, it will be important for the 
Bureau to finalize its cost estimating policies, procedures and 
guidance, as we have already recommended, in order to develop 
reliable cost estimates for 2020.
    With respect to re-engineering key Census-taking 
activities, the Bureau is researching a number of operational 
changes that may yield significant savings. As John already 
mentioned, key among these are using the Internet as a self-
response option and replacing certain enumerated collected data 
with administrative records.
    To be sure, the new Census-taking methods the Bureau is 
considering have the potential to reduce costs. As one example, 
the Bureau estimates that administrative records could save up 
to $2 billion by reducing the need for certain labor intensive 
door-to-door visits by Census workers. However, the Bureau has 
never used these methods at the scale needed for the decennial, 
if at all, which entails a degree of operational risk; thus, as 
the countdown to 2020 continues, it will be important for the 
Bureau to effectively design, test and implement these new 
approaches on schedule with an eye toward ensuring they will 
generate the needed cost savings, function in concert with 
other Census operations, and work at the scale needed for the 
national head count.
    I will now turn to my colleague, Carol, who will discuss 
the Bureau's progress in strengthening IT management and 
security practices.

                     STATEMENT OF CAROL CHA

    Ms. Cha. Chairman Farenthold, Ranking Member Lynch, members 
of the subcommittee, thank you for inviting us to testify 
today.
    The Bureau currently has a number of research and testing 
activities that are underway, and plans to use those results to 
select the IT investments to support the 2020. The technology 
options being explored collectively represent a dramatic leap 
from 2010 and, thus, a greater amount of risk for the Bureau. 
And at this time, the Bureau has not yet achieved the level of 
institutional maturity needed to reliably bring these solutions 
to bear.
    Accordingly, we have recently initiated work to evaluate 
the status and plans for the various IT options, which include, 
as Robert had mentioned, the use of the Internet, a possible 
use of a Bring Your Own Device model, or BYOD, to enable 
enumerators to use their own smart phones or other mobile 
devices to perform field data collection and other activities, 
as well as the implementation of enterprise-wide IT services 
such as delivering Bureau-wide storage servers and 
communication services via the cloud to--in an effort to 
improve scalability and reduce complexity.
    And while these options offer the potential for greater 
efficiency and effectiveness, that potential is dependent in 
large part on the Bureau's having well-established IT 
management and security controls. IT controls that, if 
effectively implemented, significantly enhance the ability to 
deliver these solutions within cost, schedule and performance 
targets, as well as to ensure the protection of the information 
that they contain.
    Our work on the 2010 Census highlighted the mismanagement, 
major cost, schedule and performance issues associated with the 
acquisition of several critical IT investments. Many of these 
problems were caused by weaknesses in IT governance, 
requirements management and IT workforce planning. For example, 
in the case of the FITCA handheld computers, functional 
requirements increased by thousands due in part to a lack of a 
robust requirements process.
    Additionally, our work earlier this year on the Bureau's IT 
security posture showed that while the Bureau had taken steps 
to safeguard the information and systems that support its 
mission, it had not effectively implemented appropriate access 
controls to protect those systems from intrusion. Accordingly, 
we have made numerous recommendations aimed at strengthening 
and improving the Bureau's IT management and security capacity.
    To the Bureau's credit, it has made measurable progress to 
address the areas of IT governance and security; however, more 
work remains.
    It will be critical for the Bureau to stay aggressive in 
its push to fully implement our remaining open recommendations. 
By doing so, the Bureau will be better positioned to deliver 
its planned IT solutions in a predictable and consistent manner 
and to ensure the adequate protection of these systems.
    In summary, on behalf of Robert and myself, the Bureau is 
making progress along a number of fronts to secure a more cost-
effective enumeration, with efforts aimed at transforming the 
organization, improving the cost-effectiveness of Census-taking 
methods, as well as strengthening its IT practices; however, a 
high degree of risk and uncertainty exists, and it will be 
critical for the Bureau to further define its roadmap for 2020 
and to set clear executive-level decision points to improve its 
ability to manage those risks as well as achieve desired 
outcomes.
    In addition, your continued oversight will also remain 
vital to help ensure that the Bureau's on path to a more 
efficient Census.
    Thank you, and we look forward to addressing your 
questions.
    [Prepared statements of Mr. Goldenkoff and Ms. Cha follow:]
    [GRAPHIC] [TIFF OMITTED] 82932.007
    
    [GRAPHIC] [TIFF OMITTED] 82932.008
    
    [GRAPHIC] [TIFF OMITTED] 82932.009
    
    [GRAPHIC] [TIFF OMITTED] 82932.010
    
    [GRAPHIC] [TIFF OMITTED] 82932.011
    
    [GRAPHIC] [TIFF OMITTED] 82932.012
    
    [GRAPHIC] [TIFF OMITTED] 82932.013
    
    [GRAPHIC] [TIFF OMITTED] 82932.014
    
    [GRAPHIC] [TIFF OMITTED] 82932.015
    
    [GRAPHIC] [TIFF OMITTED] 82932.016
    
    [GRAPHIC] [TIFF OMITTED] 82932.017
    
    [GRAPHIC] [TIFF OMITTED] 82932.018
    
    [GRAPHIC] [TIFF OMITTED] 82932.019
    
    [GRAPHIC] [TIFF OMITTED] 82932.020
    
    [GRAPHIC] [TIFF OMITTED] 82932.021
    
    [GRAPHIC] [TIFF OMITTED] 82932.022
    
    [GRAPHIC] [TIFF OMITTED] 82932.023
    
    [GRAPHIC] [TIFF OMITTED] 82932.024
    
    [GRAPHIC] [TIFF OMITTED] 82932.025
    
    [GRAPHIC] [TIFF OMITTED] 82932.026
    
    [GRAPHIC] [TIFF OMITTED] 82932.027
    
    Mr. Farenthold. Thank you very much, and I'll take the 
first stab at it, recognizing myself for 5 minutes of 
questioning.
    Mr. Thompson, the number we hear is $16 per household. I 
assume the typical suburban or urban household is not your 
problem in that dollar amount. I got my Census form in 2010, 
filled it out and mailed it back. I couldn't have cost $16.
    Where are your high dollar problems and what--do you have 
any specific thoughts on how you address those?
    Mr. Thompson. I do, Congressman. A lot of the cost in the 
Census is built up in the infrastructure required to collect 
the information from respondents that, not like you, don't 
respond and are very difficult to get to respond. So to the 
degree that we can reduce that infrastructure by using some of 
the operations research methods, by using administrative 
records, as you suggested before, that's where we're going to 
really reduce the cost, and then we can apply some additional 
methods to make the way that the interviewers work more 
efficient and effective.
    Mr. Farenthold. Now, is the $16 a current number, or wasn't 
2010 substantially higher than that? Do you know?
    Mr. Thompson. My good colleague here, who estimated the 
cost, said it was much higher.
    Mr. Goldenkoff. It was $100.
    Mr. Thompson. Yeah.
    Mr. Goldenkoff. About $98 per housing unit.
    Mr. Farenthold. So it's close to $100 in--on the last 
Census, when historically it's been more like 16.
    Mr. Goldenkoff. Yes. That's the average overall, but it's 
been going up every single decade.
    Mr. Farenthold. Okay. Is your microphone on?
    Mr. Goldenkoff. Sorry. It was $100 on average per housing 
unit in 2010, and I think that was up from around $70 per 
housing unit in--in 2000.
    Mr. Farenthold. Okay. All right. Let's talk--you're talking 
a little bit about technology. I'm happy to wee you brought a 
laptop, Mr. Thompson. Maybe you're atypical of a government 
that doesn't seem to be able to compute its way out of a paper 
bag.
    You're talking about a cloud-based system. Are you 
developing your own cloud-based system or are you outsourcing 
to a private company that has an existing cloud infrastructure?
    Mr. Thompson. Right now we are building our own cloud 
system, but we are in consultation with private sector experts 
to build that. We have to--right now we have to build our own, 
because we believe that is the best way to maintain the 
security of the Census information.
    Mr. Farenthold. In that case, so that brings up my next 
question, actually, with respect to security. As you talk about 
using other records that the government has to ease your job, 
that creates a level of information sharing between--between 
agencies and the obvious technical and privacy associate--
issues associated with that. In light of the information that 
is leaking out from Mr. Snowden at the NSA, how--are we opening 
a can of worms there with respect to privacy that we need to be 
worried about, and do you feel you can access that data from 
other agencies without specific congressional authorization?
    Mr. Thompson. That's--those are certainly good points. So 
let me start by saying that we are actively researching the use 
of administrative records to improve the Census. We'll take in 
information from other agencies. We give no information back 
out of the Census Bureau. Once we have the information, we 
protect it internally and nothing--nothing leaves the Bureau.
    Mr. Farenthold. Now, that's what the NSA thought, too.
    Mr. Thompson. Well, let me also say that you mentioned 
security issues. And security is--I think is one of the most 
important issues in assuring the public that their information 
is confidential, and that's one of the reasons I was really 
glad to see when I came on that the GAO was looking at the 
Bureau's IT security. IT security is an ongoing challenge.
    Mr. Farenthold. I'm kind of short on time. So do you feel--
do feel like you're going to need congressional authorization 
to do some of this stuff, yes or no?
    Mr. Thompson. I think we'll need the help of the Congress 
in some instances.
    Mr. Farenthold. All right. And let me go to Ms. Cha for a 
second on a technical issue. We hear a lot of discussion about 
needing to gear up now, but in terms of Internet time, if we're 
looking at a do-it-yourself, do--how much of this work can we 
do now, how much do we have to wait until we're a little closer 
because we don't know what devices are going to be in vogue or 
what--you know, what new technology will be developing in that 
time frame?
    Ms. Cha. Mr. Chairman, thank you for the question. Well, as 
I mentioned earlier, the Bureau is undertaking a number of 
exploratory measures to look at what technology is available. I 
think what's important to be mindful of is that the Bureau 
focus on institutionalizing those IT management and security 
controls at this time in order to be in a position to 
effectively deliver those solutions, whether--my understanding 
from Director Thompson is that the current plan is to start the 
development planning work for systems----
    Mr. Farenthold. And do you agree with--I'm sorry. Do you 
agree with Director Thompson that an in-house system is 
necessary and as cost-effective as, say, outsourcing to Amazon 
or RackSpace or some company like that for the cloud services 
that already has a big infrastructure in place?
    Ms. Cha. Well, Mr. Chairman, we haven't done the work 
associated with that, so it's hard for me to say. Once we get 
that information and get a clearer sense of this--of the 
roadmap that they intend to--to craft, we'll be in a better 
position to comment.
    Mr. Farenthold. Yeah. I just worry about large capital 
expenditures in a high tech area that changes for a Census that 
is done once every 10 years.
    I've already gone a little bit over my time, so in 
fairness, I will recognize Mr. Lynch for 6 minutes and 17 
seconds.
    Mr. Lynch. Thank you, Mr. Chairman. I don't think I'll need 
that, but I do want to ask, the impact of sequestration, I 
realize that you do work over the 10-year Census period to get 
us to a point that we deliver the Census at the end of that 
period as accurately as possible. And I know in March of 2013, 
we got a letter from Secretary of Commerce Rebecca Blank, who 
warned that--she was Acting Secretary, actually, but she warned 
that sequestration could, and I want to quote from her letter, 
she said, ``it could force the agency to delay the economic 
Census,'' which is the once-every-5-year survey that forms the 
basis for a wide range of economic indicators, from GDP to 
unemployment rates.
    Sequestration also could force the Bureau to conduct fewer 
field tests and increase their reliance on existing data, and 
delay evaluation programs and operational design decisions 
until 2015. Census advocates also say that the budget cuts will 
lead to a less-informed government and create a more expensive 
2020 enumeration and endanger data that business, researchers 
and State and local governments utilize.
    So let me ask you, Director Thompson, do you agree with the 
advocates that the across-the-board cuts to the Census can 
yield some damaging and negative results that can affect not 
only the government policymaking, but also economic decision-
making in the private sector?
    Mr. Thompson. Congressman, I do agree with that. The 
Bureau, because of the limitations in 2013, had to move several 
tests back to 2014. These include testing adaptive design 
methodologies, which are statistical ways to make the non-
response follow-up more efficient, as well as ways to minimize 
the number of visits that you would make to households, and to 
test administrative records and reducing the workloads. That's 
been moved back to 2014. Also some of the products on the 
economic Census have actually been delayed.
    So 2014 is a critical year, and without adequate funding, 
we'd be forced to make decisions between what to do, what not 
to do. These would involve the tests I just mentioned; it could 
also involve tests that would allow us to reduce--to reach our 
goal of doing a targeted canvassing. So funding for the Census 
Bureau in this time is very critical.
    Mr. Lynch. All right. I understand it will accommodate some 
of these sequestration effects. The Bureau discussed reducing 
contract work and discussed not filling vacancies that are 
being created through attrition. Can you further describe the 
Bureau's plan to address those type of cuts and reductions 
going forward?
    Mr. Thompson. Well, I could, but I think it would take a 
long time as--at the hearing. As we went over the various 
scenarios, it could----
    Mr. Lynch. Well, let's talk about the trends.
    Mr. Thompson. But--but in a general sense, as I said 
before, the big issue would be without the adequate funding, we 
would be in a position where we would be deciding what research 
and development to do for 2020 and what research and 
development not to do. And 2016 is the year--is a key year for 
us to begin developing the systems for the 2020 Census, and we 
would be developing those systems without as much information 
as we would use, we would not be building in the cost-saving 
methods that we would. So that we're putting the program more 
and more at risk by pushing things further into the decade.
    Mr. Lynch. Let me ask you, I know from my own experience 
that especially with phone lines, you know, fewer and fewer 
people have phone lines, and I know that you--the traditional 
way has been to, you know, call in, follow up and--but now 
place is less important, but no one--you know, we have one 
because we just don't get rid of it, but a lot of people don't 
have phone lines anymore and so place, address is less 
important to the American consumer.
    Are you having--are you having trouble with that or have 
you developed a--a way of dealing with that that doesn't reduce 
your accuracy in terms of tracking people?
    Mr. Thompson. Yeah. This is--this is another topic we could 
talk about for a long time, because I used to deal with this 
with my previous company that I worked at. We did the biggest 
telephone survey for the Federal Government.
    Mr. Lynch. Were you with the CIA?
    Mr. Thompson. No. This was the scientific survey. I'm 
sorry.
    Mr. Lynch. Oh, okay.
    Mr. Thompson. And--but these--these are--these are--these 
are really, really important issues, because cell phones, as 
you say, are not address-based, so you have to do more probing 
interviews to determine where a residence is. And there are 
ways you can do this. They're more expensive. You also can't 
use auto dialing techniques to call cell phones, because 
they're protected under various acts, so that makes it more 
expensive, too, but--but there--there--there's methodology to 
deal with it; it's just more expensive and more time-consuming, 
and the response for dialing cell phones is lower than for land 
lines.
    Mr. Lynch. Yeah. Yeah. Okay. With the remaining time, if I 
could ask representatives from GAO, do you see what the 
implications could be for the Census by way of the 
sequestration cuts?
    Mr. Goldenkoff. Yeah. It's--it's the sequestration cuts 
themselves, the--as well as uncertainty in funding that comes 
from a continuing resolution and, you know, a lot of this is 
congressional policy, so I don't want to weigh in there, but 
the fact is is that the Census Bureau has essentially scrapped 
its approach, its old approach for taking the Census, these 
tried and true methods, and it's trying something new for 2020.
    And so past experience has shown that upfront research and 
testing, these early investments in research and testing are 
critical to stave off cost increases down the road. So to the 
extent that budget cuts and uncertainty in funding forces the 
Bureau to put off a lot of the decisions and testing into the 
future, it just puts the entire operation at risk.
    Mr. Lynch. Thank you. Mr. Chairman, I yield back.
    Mr. Farenthold. Thank you very much. At this time we'll 
recognize the vice chair, Mr. Walberg, for 5 minutes.
    Mr. Walberg. Thank you, Mr. Chairman. And thanks to the 
panel for being here.
    Certainly around Census time, probably one of the greatest 
glut of emails, phone calls, letters, contacts we receive is 
about, why, you know, why does the Census do this to us? And 
specifically about questions.
    But let me go back to a train of questioning the chairman 
began to some degree, and that's on the historic cost, $16 per 
household. And you've indicated that that has been growing 
consistently over the time, and this past being $100 per 
household.
    What's the driving force behind, Mr. Thompson, the cost, as 
you see it? And then Mr. Goldenkoff, I want to ask you the 
same. What's the driving cost behind that consistent increase 
in the cost per household?
    Mr. Thompson. Let me take this from a couple perspectives. 
In some censuses, the 2000 Census, which I was intimately 
involved in, a big component of the increase in costs was a 
lack of agreement on the design until late in the game. And if 
you couple that--late in the decade. And if you couple that 
with the growing pressures on the Census Bureau to count 
everyone as accurately as possible, or to count them--the 
population very accurately, that's driving the cost up.
    For the 2010 Census, looking at that, there are forces on 
the Census Bureau to provide an accurate count, the population 
is becoming much more diverse, and the big problem is, at least 
in my view, was that the Bureau was using methods to do this, 
to try to produce this count, that were essentially generated 
in 1970, and it's--and as--as Mr. Goldenkoff said, as previous 
Director Grove says, those methods are unsustainable for taking 
censuses in the future.
    So it's a combination of the population becoming much more 
complex, much more diverse, pressures on the Census Bureau to 
produce extremely accurate data, and using methods that are no 
longer up to the task.
    Mr. Walberg. Mr. Goldenkoff, your response.
    Mr. Goldenkoff. Those were all accurate statements and 
things that GAO has said in years past. Basically as the 
population has become more diverse, the Census Bureau has to 
conduct more operations to ensure everyone is counted 
accurately. You have people living in basements and converted 
attics and in sheds in backyards, people living in cars. If 
everybody filled out their Census form when they received it, 
we could probably--the Census Bureau could do a very cost-
effective Census, but of course they don't, and it's the field 
operations, it's the constant follow-up operation, a lot of 
redundant operations that are needed that if you don't catch 
someone in one operation, you'll get them in a second operation 
or a third or a fourth.
    So everything that John has said is true, but I'll add one 
more aspect to this. It's all well and good to have these 
redundant operations, but what I think the Bureau needs to do a 
better job of is determining the return on investment of each 
one of those Census-taking operations. For example, to develop 
an accurate address list, something in the neighborhood of 
around a dozen separate operations were used. Some of those add 
accuracy, some of those actually may create more noise.
    Mr. Walberg. Could the--just jumping in there. With the 
lists and maps as well, would the private sector be able to 
accomplish an assist for us at a far more accurate and reduced 
cost simply because we know there are lists out there? I mean--
--
    Mr. Goldenkoff. Well, yes and no. I mean, first of all, 
probably the most accurate lists in terms of where mailboxes 
are located comes from the Postal Service, and the Bureau is 
already using that.
    Where the challenge comes in is the people who live in 
unconventional housing units, and that's where some amount of 
address--where the address canvassing comes in where the Bureau 
goes door to door looking for clues that people might be living 
in a basement or in an attic or in the shed in the backyard or 
possibly in the car. There are homeless people out there. 
People are very mobile. So it's really not clear if the private 
sector has a better address list.
    Mr. Walberg. Are we asking too much, our questioning 
process. Are we obtaining a greater amount of information than 
really is required? Is that part of the cost?
    Mr. Goldenkoff. Well--go ahead.
    Mr. Walberg. Mr. Thompson.
    Mr. Thompson. I--well, I believe that the information we 
collect on the decennial Census is extremely critical, and I 
could go over the uses of it. It's used for basically enforcing 
civil rights. It's used for redistricting purposes.
    I will also say that we provide that information, the 
topics in the 7 year to the Congress, and then we provide the 
actual questions in the 8 year to make sure that we are in 
agreement with the Congress before we take the Census on the--
on the content of it.
    Mr. Walberg. Well, I guess that--that's--that's, again, a 
great concern of my constituents and many people, that we are 
committing acts of intrusion beyond what seems to be necessary 
to people. And, of course, when we're talking about costs, I 
think it's a question that ought to be addressed as well. Are 
we collecting absolutely necessary information, are we going 
overboard in certain cases, and are we providing information 
that actually gives help?
    So I see my time has expired. I yield back.
    Mr. Farenthold. Thank you very much. We'll now go to Ms. 
Maloney for 5 minutes.
    Mrs. Maloney. Thank you very much. And congratulations Mr. 
Thompson, on your new role as Director of the Census Bureau. I 
was very pleased to have sent a letter supporting your 
nomination during Senate confirmations, and I was very pleased 
that Congress last year passed legislation that included a 
provision that I had authored under separate legislation for a 
long time that makes the director of the Census a presidential 
term appointment of 5 years, subject to the confirmation of the 
Senate, and this is very important, so that the director is 
tied to the needs of the Census and not to an election 
calendar.
    And I agree with your testimony today that credible, 
objective and timely data on population growth is fundamental 
to a democracy, incredibly important for fair representation 
and fair distribution of hundreds of billions of dollars in 
Federal aid.
    And your testimony today on the escalating cost, now up to 
roughly $100 a survey, I'm really pleased to hear that you're 
looking at new initiatives to cut those costs, including using 
the Internet as a self-response option and replacing the door-
to-door collection with administrative records under certain 
circumstances.
    But relying on the Internet, it potentially could save a 
great deal of money, but not everyone has access to an Internet 
or--and people don't know how to use it, particularly the 
elderly and certain populations. So is there a risk of an 
undercount in specific populations as we rely or move to 
relying on an Internet response?
    Mr. Thompson. Well, Congresswoman, you're hitting a very 
important point in terms of how we have to take the Census in 
the United States; that is, there's no one method that will be 
the right method for each segment of our diverse population. 
The Internet is a--is a vehicle that can be used to enumerate a 
large part of our population, in my opinion, and it will save--
it will result in a lot of resource savings. That doesn't mean 
we're going to do away with the mail for everyone, or we're 
going to not have to go visit them and talk to them with 
interviewers that speak in language or various other methods, 
but we have to realize that if we can make savings in one area, 
we can use those savings to target methods to make sure that 
every part of our diverse population is counted.
    Mrs. Maloney. And what's the update on these operations? 
Are you--do you have a lesson plan or a plan of how you're 
going to use the Internet that's been produced and are you 
putting that into effect, or where does it stand?
    Mr. Thompson. So right now we are in the process of doing 
research on a small scale that's aimed at how we can motivate 
our response to the Internet. We're looking at some various 
other options, which would include a pre-registration so people 
might sign up in advance to do the Census, but we're--we're in 
the beginning stages of doing that work, and that's what we 
really need to--we need to focus on that if we're going to be 
able--and get the funding to do it.
    Mrs. Maloney. Okay. Can you elaborate on your testimony on 
the $2 billion projected savings if you use administrative 
records in 2020 to reduce the door to door? And exactly how 
would that work? And certainly saving money is necessary, but--
but are there risks to increasing your reliance on 
administrative records, and what can you do about these 
records, and what particular types of records would the Bureau 
seek? Who would have access to these records, and what type of 
information would the Bureau seek to verify on relying on 
administrative records?
    Mr. Thompson. Well, I think the 2 billion figure was 
mentioned by Robert, and I need to get straight with him on 
exactly where it came from, but--but I would--I would stipulate 
that there are significant savings to be realized by using 
administrative records. And the Census Bureau has used--we have 
used administrative records before on a small scale, as was 
noted, we've used Postal Service records, we've used military 
records to count the military overseas. And the Secretary of 
Commerce has really broad discretion in acquiring administrate 
records working with other agencies.
    The big thing is--one of the big things is deciding on some 
issues as to would you use administrative records in lieu of at 
least trying one contact on non-response follow-up? That's 
something that we need to talk about, need to really understand 
and get widespread buy-in on as to whether that--that would be 
acceptable. That would probably offer the greatest savings. 
There are other options for using administrative records.
    I will--I will say that when the Census Bureau gets records 
from anywhere, the Postal Service, any source of administrative 
records, we keep those records confidential. We do not give 
information back out of the Census Bureau. In fact, within the 
Census Bureau, only those individuals that need to actually 
access the records have access to them, and we--to the extent 
we can, we take identifying information off of the records.
    We have used records in our research so far from a variety 
of sources, from HUD, from Social Security, the Indian Health 
Service, the CMS. We--we are looking to really work with the 
Congress to get endorsement that this is the right methodology, 
that there is support for this methodology throughout the 
decade.
    Mr. Farenthold. Thank you very much. We'll now go to the 
gentleman from Georgia, Mr. Collins.
    Mr. Collins. Thank you, Mr. Chairman. This--being new to--
from this position, looking at this the first time as I've 
started digging into this with my committee in the committee 
side, but also my staff, this is an interesting area that often 
gets overlooked with all the headlines and everything else in 
the world. And as I--as I made a statement during all this, I 
said a lot of times the big stuff will take care of the big 
stuff, but it's the things like this that really matter to 
people, because it's amazing how many questions we get on this 
and how many issues, and especially from a cost perspective.
    I want to ask some questions that sort of--are not random, 
but sort of rapid fire, and then some just from my own case. 
Mr. Goldenkoff, you had said that people are less willing to 
participate. I'm going to come back to that statement here in 
just a moment, but I have a question that I just haven't found. 
What is the percent of non-response follow-up? What is--what is 
the percent of those who--Ms. Cha, anybody who wants to answer 
this real quickly.
    Mr. Goldenkoff. In 2010, it was--the response rate was in 
the low 60 percent, I believe.
    Mr. Collins. So 60 percent did it the normal, proper way. 
They get the envelope in the mail, they respond, they turn it 
back in and----
    Mr. Goldenkoff. Exactly. It was like 60-plus. Something 
like 63.
    Mr. Thompson. You have a nice chart in your--in your 
prepared statement.
    Mr. Goldenkoff. Yeah, I do.
    Mr. Collins. And I may--I may just be missing the number. 
Out of curiosity, I've asked.
    Mr. Goldenkoff. 63.
    Mr. Collins. Just curious. And--because the question was 
just raised on the Internet. What is the highest percentage of 
demographics that actually turned it in the proper way?
    Mr. Goldenkoff. I----
    Mr. Collins. Age demographics.
    Mr. Thompson. So the characteristics of those that don't 
return the questionnaires----
    Mr. Collins. No. I want to know the ones that do.
    Mr. Thompson. That don't return the questionnaire, tend to 
be more minority, more----
    Mr. Collins. No, no. I want to know those who do.
    Mr. Thompson. They tend to be more non-minority, more owner 
than----
    Mr. Collins. Elderly? Senior adult, 45 and over? 55 and 
over? 65 and over? Because I have--I have a purpose of the 
question. I'm not trying to lead you----
    Mr. Thompson. No, no, no, no.
    Mr. Collins. --a rabbit hole here.
    Mr. Thompson. I believe--I believe in the research that I--
that I've done that the elderly do respond at a higher rate 
than--than--than the young. That----
    Mr. Collins. Because one of the concerns is if we use the 
Internet, the elderly will be, you know, disenfranchised, then 
we'll--then we're turning it on the proper way and they're 
going--they tend to be more responsive to this. And--and that's 
why--I'm not trying be tricky. I was just trying to--you know, 
as we look at the honest answers here and looking at it as we 
go forward.
    Mr. Goldenkoff. The Internet is just an additional method, 
it's not that the other methods are being tossed. It's just one 
more method that people can use.
    Mr. Collins. But many times----
    Mr. Goldenkoff. It's another option.
    Mr. Collins. But many times when this conversation comes 
up, we throw--this comes up in a lot of other areas that we're 
undercounting, we're doing these kind of things based on--and 
so I just wanted to get that out there.
    I want to follow up also--I wish I had far more time than 
this, but we'll get into, I'm sure, more----
    Mr. Farenthold. The second round.
    Mr. Collins. Mr. Chairman, we'll get this. I want to go 
back to the one that we get a lot, and it's balancing privacy 
and effectiveness. And especially in the ACS, this is where we 
get a lot of questions. And I follow up on the--Congressman 
Walberg's questions as well. Are we asking, you know, too much? 
And I think--well, I'm interested in what you said, that you're 
going to bring the questions to Congress. Do we see these? Does 
that come in the form of legislation or is that just merely 
informational?
    Mr. Thompson. We submit the questions to the Congress. 
It's--if the Congress wants to take action, they would have to 
enact legislation.
    Mr. Collins. Okay. So you're just bringing this for 
informational purposes, and if there--if there was need to be 
changed, then--okay. Thank you for that.
    Mr. Goldenkoff. Mr. Collins, on the ACS, those questions at 
least initially, you know, it evolved from the Census long 
form. They were--they were tied to legislation.
    Mr. Collins. Oh, I get it. But mostly we're not using the 
long form anymore, we're using the short form, we're taking it 
more, you know, frequently. And these are questions that come 
in, and then, of course, you get into the, as you said earlier, 
less willing to participate, and especially in this environment 
that we're currently in, which will probably get not a lot 
better as far as people trusting the government and trusting 
the information to be held. These are the kind of questions we 
get in our office all the time.
    And, you know, one of the issues that, you know--you know, 
given the public's increased skepticism here, I think we've got 
to do a better job of how we, you know, ask the questions and 
are we properly asking, and then also doing away or de-
emphasizing, because I've looked into this, there's been one 
prosecution for not turning in your Census data, yet we have it 
on there and it is in law and I understand that, and the threat 
is if you don't turn this in, we're going to prosecute you. I 
think a little less emphasis on that when people call, it would 
probably help a great deal, because the next thing they do is 
they call my office and say, I'm not sending this stuff in. 
They don't need to know this. Are they going to prosecute me?
    So maybe more of a follow-up than anything else. The 
questions that come from there, however, as we do this, the 
ability to save money, the ability to look at this, I am very 
concerned, as the chairman is, that we're developing software, 
we're developing the platform process now that could really be 
outdated in 6 to 7--you know, 6 years, 7 years as we start 
again. What's the thought process that went into that, not just 
laying groundwork, as you said, but I'm concerned that even the 
answers you get now may be outdated even in 2 or 3 years. So 
what is being taken into account there?
    Ms. Cha. Well, I can't speak on the Bureau's behalf, but I 
can tell you that high performing organizations, what they do 
is they typically have an internal team that is tasked with 
looking at and keeping up with the key technological trends, 
and to anticipate and identify those disruptive technologies 
that may affect their core business, and then develop options 
for--for actually implementing them into their business.
    Ms. Cha. So, and that is occurring in both the Federal and 
the private sector space. So that is just one example in the 
Federal sector. When you look at the Department of 
Transportation, they do have certain major programs, for 
example, that have a technology scan and assessment group that, 
for example, looks at technology options for deploying things 
such as connected vehicles, for example. So those things are 
occurring, but I think Director Thompson would be in a better 
position to talk about what they are doing to ensure that they 
stay abreast of the current----
    Mr. Collins. Mr. Chairman, we'll come back to that but this 
is the question because before this committee in particular, we 
hear more problems, and I think, Mr. Chairman will agree and 
the ranking member as well, we hear more concerns on IT 
mismanagement and expenses almost than any other thing that we 
do. And so it is really concerning to me when I say we're going 
to build our own system. So I will refer back to the chairman 
and we'll continue on.
    Mr. Farenthold. Thank you very much. You know the Lord 
works in mysterious ways. Mr. Collins brought up the American 
community survey. I realize this committee had a hearing on 
that last year, but oddly enough, guess what I got in the mail 
last month? Guess what phone call I got last week, very 
threatening nature, by the way, because I had not yet filled it 
out because I wanted to go through this hearing first.
    I do think at some point we'll review that, and I expect 
Mr. Collins and I would love to visit with you specifically 
about that, and if we determine that everything wasn't covered 
last year, we may jump into that again. It may be time to 
review what we are asking. And when you are asked questions 
like what your mortgage is, how much your electric bill is? I 
understand some of the purposes of these and the committees 
that would want them, but it's not raised well to the American 
people, and it's an issue that I get a lot. That is obviously 
out of scope of this hearing. We were talking about costs. But 
I do think it remains an issue that a lot of our constituents 
are concerned about.
    So let me go on with a couple of questions. We talked a 
little bit, Mr. Thompson, and actually I'm going to take this 
to the GAO because this was before your time, the electronic 
slate program that was, we paid to develop custom hardware and 
software for the enumerators last time around in 2010. It's my 
understanding that that never really worked beyond them being 
able to enter their time. From practical purposes, they weren't 
able to put a lot of data in that. Is that consistent with your 
findings or have you investigated the success of that at all?
    Ms. Cha. Well, in looking at the major systems that were 
acquired for 2010, that does seem to be consistent. So for 
example with the FITCA handheld computers which was customed 
developed hardware and software, ultimately the Bureau 
delivered a half-baked solution that ran into significant cost 
overruns and ultimately at the end of the day required that 
manual workaround in order to meet the immutable 2010 deadline, 
and that ultimately led to a $3 billion increase to the 
ultimate cost.
    Mr. Farenthold. And I commend that mistake to you, Mr. 
Thompson and Ms. Cha as a consideration with respect to really 
something cloud-based and less done in house. Computer 
technology doesn't have a 10-year life span. Technology changes 
so quickly that you really are potentially buying a computer 
system that will be used for one Census when maybe renting 
something through a cloud service might be worth looking at.
    You talked a little bit about statistical methodology. 
There is no proposal in your statistical methodology that does 
away with we're going to try to count everyone. We're not going 
to make assumptions like when somebody does a phone poll of X 
number of likely voters, they extrapolate the whole country. We 
really are going to try to do what the Constitution requires, 
is that correct?
    Mr. Thompson. Yeah, exactly, Congressman. I was very clear 
in my confirmation hearing about that is that that is not a 
technique that I think is appropriate for taking the decennial 
Census.
    Mr. Farenthold. And then with respect, you also talked 
about changing methodology with respect to how you get hold of 
nonresponders. What do you see as your number one or number two 
failures in areas and how do you--groups meet, bridge, you 
know, people who live in cars or undocumented residents, where 
are your big failures and what are your top one or two ideas 
for addressing that?
    Mr. Thompson. So our research has showed that the people 
that are less likely to be counted or harder to count tend to 
be minority, they tend to be highly mobile, they tend to be 
maybe more renters and in unusual living situations. So the 
effort has to be--and they are and it is more expensive to 
count those individuals. So my goal is to find ways to count a, 
the big part of the population which is relatively easier to 
count, a lot more cost effectively, thereby freeing up 
resources that we can then target to counting.
    Mr. Farenthold. At some point, I would like to get from 
you, and if you could maybe do that in writing based on time 
considerations, what you consider to be your top five counting 
problems and your top five proposed solutions, and they may not 
necessarily jibe. It is kind of the 80/20 rule you might be 
able to, your solutions might be for a problem that is a little 
further down the line, but it is an easier solution. But I want 
to see what your biggies are. And can you tell me a little 
about, you're spending a lot of time and effort on geospatial 
data, there is a lot of that already out there and it doesn't 
change that much, obviously you get new houses built, but they 
don't move, and at what point are we doing something that there 
is no longer a need for but there is such a marketplace for 
with GPS's in everybody's cars and the prevalence of Google 
Maps, at what point is this geospatial data collected and 
maintained by the government a worthwhile deal or something 
that the private sector has taken it over, we have ceded it and 
it is done.
    Mr. Thompson. Well, you are talking about something that is 
very dear to my heart in terms of targeted canvassing. That is 
we don't have to build everything. We can buy it, we can get it 
off the Internet, we can work with private companies. We don't 
have to do it all ourselves and that is why it's exciting now 
to look to foreign partnerships with the private sector to get 
information that will--I'm stumbling and I'm sorry--that will 
allow us to not have to canvass all of the country.
    Mr. Farenthold. And finally, I hope Mr. Lynch will explore 
some of the postal, he mentioned in his opening statement, the 
possibility nobody knows the area better than the postal worker 
who walks it every day. I think there is a huge opportunity for 
that. Understanding that postal workers are paid a lot more 
than Census workers, so we've got to factor that into the 
costwise equation. But I know Mr. Lynch brought it up. If he 
doesn't ask the question I will.
    And then finally my last question is, you mentioned the 
need for Congressional action with respect to possibly 
authorizing sharing of data, and Mr. Collins brought up the 
possibility of amending the questions both on the short form 
Census and the American Community Survey.
    The sooner we do that the better I think in your preparing.
    At some point, could you also let us, send a letter to this 
committee with some of the changes that you think you would 
like to see in legislation? Because we actually probably move 
slower than you do.
    With that, I will recognize Mr. Lynch for a second round of 
questioning.
    Mr. Lynch. Thank you, Mr. Chairman. I will follow up on 
that point. Back in September, 2011, GAO recommended that the 
Secretary of Commerce and the United States Postal Service 
consider a couple of things: One, expanding their current 
collaborative efforts to include recruiting mail carriers or 
retired mail carriers for the 2020 Census. And the second thing 
that they talked about was assessing whether strategies can be 
developed to reduce undeliverable as addressed mail in the 
Census.
    So, Mr. Goldenkoff, you want to take a crack at this and 
just see where we are, what those--I did read the GAO reports 
but I want to get this on the record. Where are we with those 
efforts of collaboration?
    Mr. Goldenkoff. Well they are ongoing it is my 
understanding as was already mentioned, the partnerships that 
the Bureau wants to form for the geospatial data, the GSSI 
which will allow for a rolling updating of the address list 
throughout the decade, instead of doing it in the 2 years 
before the Census basically taking one large bite all at once, 
do it in smaller increments, and that will be very helpful for 
targeted address canvassing. There are other opportunities for 
collaboration with the Postal Service that the Bureau will need 
to explore, particularly as it gets several year down the road.
    For example, of the roughly 600,000 enumerators that were 
hired to do nonresponse follow-up, only 2,400 were postal 
workers. And if the Bureau had perhaps done a more targeted 
approach to actually, a targeted recruitment of postal workers 
they could have gotten a lot more. And as you said, these are 
people who know the neighborhoods, they know how to deliver 
stuff. This is what they do for a living. They know the 
complexities of the different neighborhoods. They already have 
the background checks done. So that's an option there as well.
    Use of postal facilities that the Postal Service doesn't 
need, the Postal Service has a lot of extra space. Perhaps the 
Census Bureau can use some of those for their local Census 
office. There might be some opportunities there as well.
    Mr. Lynch. Right. I know that in the GAO report that I 
read, it had a dollar figure of $41 an hour for an urban postal 
worker--excuse me, an urban mail carrier, a letter carrier, and 
I think it was $34 for a rural letter carrier. But the 
retirees, obviously, would be a lesser cost. They've already 
got their pension, they've already got retiree health benefits. 
I'm just curious if, you know, in looking--the Postal Service 
has already put out several early retirement programs, 
incentives to get postal workers to retire.
    Many of those workers have expressed a desire to continue 
working part-time or in some capacity. This might offer us an 
opportunity if we worked through the postal workers with their 
unions to actually recruit some of those postal workers, letter 
carriers that they might want to supplement their income but 
they wouldn't have the full range of benefits that you know an 
active letter carrier currently requires.
    Mr. Goldenkoff. And that's where the trade-offs seem to be 
taken into account, the cost benefit. You mentioned the hourly 
wage rates for the mail carriers of $34 and $41 depending if 
they were rural or urban. That compares with $15 per hour on 
average for a Census enumerator. So you see there's a huge cost 
difference. And to the extent that some enumeration takes place 
in the evenings, on weekends, postal workers I guess would be 
making additional money for overtime, for example, working on 
Sundays.
    So you can see how the cost increase can be significant if 
they are doing it as part of their postal job as opposed to 
being employed directly as a temporary employee by the Census 
Bureau.
    Mr. Lynch. I just question the efficiency of someone 
walking into a neighborhood they have never been in before and 
they're supposed to find people as opposed to the efficiency of 
a letter carrier that has been down that street, knocked on 
that door, been in that house a thousand times.
    Mr. Goldenkoff. You're absolutely right and that's why if 
they were hired as supplemental income, just sort of on their 
own, not through their postal employment, but just as anybody 
else, but it was targeted recruitment, there could be both not 
only cost savings there, but also it could be done more 
efficiently.
    Mr. Lynch. What about reducing the number of undeliverable 
as addressed mailings? Mr. Thompson or Ms. Cha, as well, if you 
any thoughts on how we go about that, or Mr. Goldenkoff.
    Mr. Goldenkoff. What we were getting at there is some forms 
just because of the nature of the address, the Postal Service 
can't find the particular address, and so the, using Postal 
Service records, for example, those can be eliminated before 
they get sent out.
    The other thing I want to say too is that to the extent 
that there is this partnership going on, it doesn't necessarily 
have to be a one-way flow of benefits. The Postal Service can 
make use of Census information to improve their operational 
efficiencies as well, and there has been some cases of that in 
2010.
    Mr. Lynch. Thank you. I see my time has expired. I yield 
back.
    Mr. Farenthold. Thank you and we are bumping up against the 
time we need to get to the 9/11 ceremony remembrance. I know 
Mr. Collins had about 1 minute more he wanted to take and Ms. 
Norton you walked in.
    Are you guys able to wait until we are done with the 9/11 
remembrance on the Capitol steps and we'll reconvene then? We 
can't miss the 9/11 event. I believe the House Members were 
supposed to report there at 10:45 so. We'll, if you guys are 
willing to hustle over there, we'll go ahead.
    And Mr. Collins, if you can keep it to a minute and with 
your closing we'll go and give Ms. Norton and Mr. Clay a 
chance.
    Mr. Collins. I'm going to give a lesson, this is something, 
as I said a few minutes ago, it was very interesting and my 
office we are going to be very involved, I want to echo the 
chairman's sentiment. Think now about what you need from 
Congress. Think now about what you want to bring before us to 
change. Get it to us now so we can have the input and we are 
not in a last minute. That is the main thing that I want to 
emphasize and my office and my staff will be willing to work 
with you. Please do so. Thank you and I yield back.
    Mr. Farenthold. Thank you. We'll recognize Ms. Norton now, 
the gentlelady from the District of Columbia.
    Ms. Norton. I simply want to ask about the value or lack 
thereof of the community survey. Has it increased the accuracy 
of the Census, this monthly survey? You use it as a base for 
the Census.
    Does it increase the costs if you didn't have it or if it 
is voluntary, would it be valuable any longer? Would it cost 
any less?
    Mr. Thompson. So, thank you for the question, Congresswoman 
Norton.
    The American Community Survey replaced the decennial Census 
long form to provide that valuable information.
    Because the American Community Survey is taken on a rolling 
basis, the information it provides is more current and 
therefore more accurate than the decennial Census long form. 
The American Community Survey is therefore a very valuable 
resource for planning programs, administering programs. And it 
is used by a variety of stakeholders, city planners, the 
private sector.
    Ms. Norton. Does it increase the cost significantly, I 
mean, if it were voluntary?
    Mr. Thompson. Oh--yeah, if the American Community Survey 
was made voluntary I think Dr. Groves--it would increase the 
cost. My recollection is that Dr. Groves in his testimony last 
July indicated that the costs might go up by as much as $60 
million a year.
    Ms. Norton. Has it decreased the undercount? Has it had an 
effect on the undercount?
    Mr. Thompson. The American Community Survey Census is 
providing more current information. The information can be used 
for better planning and thus reduce the undercount to some 
degree.
    Ms. Norton. Thank you, Mr. Chairman.
    Mr. Farenthold. And we'll go to Mr. Clay now.
    Mr. Clay. Thank you so much, Mr. Chairman. I will be as 
brief as possible.
    Mr. Thompson, on March 15th of this year, I, along with 
several Members of Congress, transmitted a letter to the Census 
Bureau urging the Bureau to define a national solution to 
ensure that incarcerated individuals are counted at their last 
legal residence. The U.S. Supreme Court issued a decision in 
June of 2012 upholding a Maryland law that tabulates prisoners 
as residents of their home addresses for the purpose of 
redistricting. However, since the current Census Bureau policy 
designates a prison cell as a residence it prevents populations 
of more than 1,500 Federal and State prisons that are largely 
male, urban and African American or Latino from being counted 
as residents of their home community.
    These incarcerated individuals normally have no ties to the 
prison location, cannot vote, and most often, return to their 
home communities upon release. Our letter urged the Bureau to 
count these individuals at their last legal home address to 
prevent districts housing prisoners from being overrepresented.
    Can you share your thoughts on counting the prison 
population?
    Mr. Thompson. Congressman, at least I, at the Census 
Bureau, take the rules that we use to count individuals very 
seriously. After each Census and before each Census, the Census 
Bureau works with the variety of stakeholders to try to ensure 
that the rules they use to count people are the most 
appropriate for that Census. And for the 2020 Census, we have 
not made our decision yet. And my opinion on this will be based 
on consultations with stakeholders to try to adopt rules that 
will be most appropriate to count people. So I'm looking 
forward to talking with you and other stakeholders on how to 
count not just the prison population but other components for 
population.
    Mr. Clay. And be aware that several States have implemented 
laws like the one upheld by the U.S. Supreme Court for Maryland 
that allows for you to count prisoners from their home 
districts. Because when you think about it, it skews the 
redistricting process when you count prisoners in prisons who 
are not there permanently.
    I know in my home State of Missouri, they are able to 
create State representative districts because of prison 
populations in rural areas, and it skews the entire process. So 
I want you to be aware, and I look forward to working with you 
and sharing our views about this issue.
    So Mr. Chairman and Mr. Thompson thank you and I yield 
back.
    Mr. Farenthold. Thank you very much. I apologize. We are 
going to have to rush out here. We really do want to 
participate in the 9/11 ceremony. I would like to thank the 
panel and the witnesses, and we are adjourned.
    [Whereupon, at 10:50 a.m., the subcommittee was adjourned.]


                                APPENDIX

                              ----------                              


               Material Submitted for the Hearing Record

[GRAPHIC] [TIFF OMITTED] 82932.028

[GRAPHIC] [TIFF OMITTED] 82932.029

                                 
