[House Hearing, 107 Congress]
[From the U.S. Government Publishing Office]
BEA: IS THE GDP ACCURATELY MEASURING THE U.S. ECONOMY?
=======================================================================
HEARING
before the
SUBCOMMITTEE ON THE CENSUS
of the
COMMITTEE ON
GOVERNMENT REFORM
HOUSE OF REPRESENTATIVES
ONE HUNDRED SEVENTH CONGRESS
FIRST SESSION
__________
APRIL 5, 2001
__________
Serial No. 107-8
__________
Printed for the use of the Committee on Government Reform
Available via the World Wide Web: http://www.gpo.gov/congress/house
http://www.house.gov/reform
U.S. GOVERNMENT PRINTING OFFICE
75-327 WASHINGTON : 2001
____________________________________________________________________________
For Sale by the Superintendent of Documents, U.S. Government Printing Office
Internet: bookstore.gpr.gov Phone: toll free (866) 512-1800; (202) 512�091800
Fax: (202) 512�092250 Mail: Stop SSOP, Washington, DC 20402�090001
COMMITTEE ON GOVERNMENT REFORM
DAN BURTON, Indiana, Chairman
BENJAMIN A. GILMAN, New York HENRY A. WAXMAN, California
CONSTANCE A. MORELLA, Maryland TOM LANTOS, California
CHRISTOPHER SHAYS, Connecticut MAJOR R. OWENS, New York
ILEANA ROS-LEHTINEN, Florida EDOLPHUS TOWNS, New York
JOHN M. McHUGH, New York PAUL E. KANJORSKI, Pennsylvania
STEPHEN HORN, California PATSY T. MINK, Hawaii
JOHN L. MICA, Florida CAROLYN B. MALONEY, New York
THOMAS M. DAVIS, Virginia ELEANOR HOLMES NORTON, Washington,
MARK E. SOUDER, Indiana DC
JOE SCARBOROUGH, Florida ELIJAH E. CUMMINGS, Maryland
STEVEN C. LaTOURETTE, Ohio DENNIS J. KUCINICH, Ohio
BOB BARR, Georgia ROD R. BLAGOJEVICH, Illinois
DAN MILLER, Florida DANNY K. DAVIS, Illinois
DOUG OSE, California JOHN F. TIERNEY, Massachusetts
RON LEWIS, Kentucky JIM TURNER, Texas
JO ANN DAVIS, Virginia THOMAS H. ALLEN, Maine
TODD RUSSELL PLATTS, Pennsylvania JANICE D. SCHAKOWSKY, Illinois
DAVE WELDON, Florida WM. LACY CLAY, Missouri
CHRIS CANNON, Utah ------ ------
ADAM H. PUTNAM, Florida ------ ------
C.L. ``BUTCH'' OTTER, Idaho ------
EDWARD L. SCHROCK, Virginia BERNARD SANDERS, Vermont
------ ------ (Independent)
Kevin Binger, Staff Director
Daniel R. Moll, Deputy Staff Director
James C. Wilson, Chief Counsel
Robert A. Briggs, Chief Clerk
Phil Schiliro, Minority Staff Director
Subcommittee on the Census
DAN MILLER, Florida, Chairman
CHRIS CANNON, Utah WM. LACY CLAY, Missouri
MARK E. SOUDER, Indiana CAROLYN B. MALONEY, New York
BOB BARR, Georgia DANNY K. DAVIS, Illinois
------ ------
Ex Officio
DAN BURTON, Indiana HENRY A. WAXMAN, California
Jane Cobb, Staff Director
Erin Yeatman, Professional Staff Member
Dan Wray, Clerk
David McMillen, Minority Professional Staff Member
C O N T E N T S
----------
Page
Hearing held on April 5, 2001.................................... 1
Statement of:
Dennis, Bob, Congressional Budget Office, Assistant Director,
Macroeconomic Analysis; Richard Berner, president, NABE;
Diane Swonk, chief economist, Bank One, Inc.; Gordon
Richards, economist, National Association of Manufacturers;
and Dr. Ernst R. Berndt, MIT, chair of the Federal Economic
Statistics Advisory Committee.............................. 34
Landefeld, J. Steven, Director, Bureau of Economic Analysis;
and Frederick Knickerbocker, Associate Director for
Economic Programs, Bureau of the Census.................... 5
Letters, statements, etc., submitted for the record by:
Berndt, Dr. Ernst R., MIT, chair of the Federal Economic
Statistics Advisory Committee, prepared statement of....... 81
Berner, Richard, president, NABE, prepared statement of...... 54
Clay, Hon. Wm. Lacy, a Representative in Congress from the
State of Missouri, prepared statement of................... 98
Dennis, Bob, Congressional Budget Office, Assistant Director,
Macroeconomic Analysis, prepared statement of.............. 37
Knickerbocker, Frederick, Associate Director for Economic
Programs, Bureau of the Census, prepared statement of...... 21
Landefeld, J. Steven, Director, Bureau of Economic Analysis,
prepared statement of...................................... 9
Miller, Hon. Dan, a Representative in Congress from the State
of Florida, prepared statement of.......................... 3
Richards, Gordon, economist, National Association of
Manufacturers, prepared statement of....................... 68
Swonk, Diane, chief economist, Bank One, Inc., prepared
statement of............................................... 62
BEA: IS THE GDP ACCURATELY MEASURING THE U.S. ECONOMY?
----------
THURSDAY, APRIL 5, 2001
House of Representatives,
Subcommittee on the Census,
Committee on Government Reform,
Washington, DC.
The subcommittee met, pursuant to notice, at 2:02 p.m., in
room 2247, Rayburn House Office Building, Hon. Dan Miller
(chairman of the subcommittee) presiding.
Present: Representative Miller.
Staff present: Jane Cobb, staff director; Chip Walker,
deputy staff director; Michael Miguel, senior data analyst;
Erin Yeatman and Andrew Kavaliunas, professional staff members;
Daniel Wray, clerk; David McMillen, minority professional staff
member; and Teresa Coufal, minority staff assistant.
Mr. Miller. Good afternoon. The subcommittee will come to
order. We will proceed. I will have a brief opening statement
and then we will go with our first panel. I called this hearing
to examine the function and needs of a relatively small but
significant Federal player in providing the policymaker and the
public a timely and accurate picture of national and
international economic activity. The Bureau of Economic
Analysis [BEA], is a statistical agency within the Commerce
Department's economic and statistics administration. It has a
budget of close to $50 million and employs approximately 445
people. It produces, among other things, one of our Nation's
primary economic indicators, the Gross Domestic Product [GDP],
something we will be looking at closely today.
BEA also produces estimate of analyses of personal income
population and employment for regions, States, metropolitan
areas and countries. BEA helps define the international
economic picture by producing the U.S. balance of payments.
Additionally, it measures U.S. direct investment abroad and
foreign direct investment in the United States. In information
provided to the subcommittee by BEA, it is clear that BEA's
statistics are heavily relied on by government and industry.
For example, the Congressional Budget Office and Office of
Management and Budget rely on BEA estimate of economic growth
to make Federal budget projections. BEA's regional income and
product estimates are used to allocated more than $100 billion
annually in Medicaid and other Federal grants to States.
Virtually, all States use BEA data in their tax projections
infrastructure planning and allocations of State funds to
counties. BEA's national, international and regional estimates
are essential inputs to private sector business forecasts and
production and investment plan. Business associations use BEA's
national and regional data by industry to gauge the economic
health of association members. Financial planners use BEA's
income and saving data, as well as the growth of GDP and its
components, to develop and assess investment and retirement
planning strategies.
Today we will examine DEA to give Congress and the public a
better understanding of this agency's important functions, with
a particular focus on the accuracy of the Gross Domestic
Product. We also hope to learn of some of the issues BEA faces
in its challenge to produce vivid, accurate and timely
snapshots of our rapidly changing economy.
[The prepared statement of Hon. Dan Miller follows:]
[GRAPHIC] [TIFF OMITTED] T5327.001
[GRAPHIC] [TIFF OMITTED] T5327.002
Mr. Miller. We have invited a number of witnesses to help
us look at BEA today. On panel one we will hear from the
Director of BEA, Mr. Steven Landefeld and Mr. Frederick
Knickerbocker of the Census Bureau, a key survey taker and data
provider to the BEA. On panel two we will hear from economists
and officials in business government and academia who have been
asked to speak to BEA's role the accuracy of GDP and the issues
they see are important to this agency. I welcome and thank you
for joining us today and look forward to your testimony, so we
will proceed immediately with the first panel.
We are delighted that both of you have joined us here
today. We will start with Dr. Landefeld. He is the Director of
the Bureau of Economic Analysis. Dr. Landefeld has been the
Director of BEA since 1995. Prior to becoming Director, he
served as Deputy Director and Associate Director of economics
at BEA. Joining Dr. Landefeld on panel one is Frederick
Knickerbocker, the Associate Director for economic programs at
the Census Bureau. Mr. Knickerbocker became the Associate
Director for economic programs in 1995. As such, Mr.
Knickerbocker is responsible for approximately 100 economic and
business surveys as well as preparation of many of the Nation's
principal economic indicators.
Mr. Landefeld.
STATEMENTS OF J. STEVEN LANDEFELD, DIRECTOR, BUREAU OF ECONOMIC
ANALYSIS; AND FREDERICK KNICKERBOCKER, ASSOCIATE DIRECTOR FOR
ECONOMIC PROGRAMS, BUREAU OF THE CENSUS
Mr. Landefeld. Thank you, Mr. Chairman. Also thank you for
doing a good part of my testimony today. I was just able to cut
out a whole bunch of things I was going to say. But I did want
to thank you for the opportunity to appear before you to
discuss the Bureau of Economic Analysis. As you and the Census
Subcommittee know, and as you indicated, Mr. Chairman, we are
the other statistical bureau in the Commerce Department.
Although we are small in size relative to our sister agency
Census--our staff is about 400 people now, not 450-something--
we are, as you noted, one of the Nation's most important
statistical agencies. Our signature products are the GDP and
the national income and product accounts, which were developed
in the late 1930's by the Nobel Laureate, Simon Kuznets, and
which are regarded as the mainstay for analyzing the U.S.
economy.
Although you reviewed a number of functions, I thought it
would be useful to describe how we do what we do, which is, in
essence, we are the Nation's economic accountant. That is, we
obtain and interpret large volumes of diverse data from both
government and private sources, such as the Census Bureau and
then organize, combine and transform these data into a
consistent and comprehensive set of economic accounts for the
Nation as a whole. BEA's accounts provide a full detailed
picture of economic activity and include such widely watched
statistics as GDP, corporate profits and some of the other
series you have noted. These data have a large impact on
interest rates, stock prices and exchange rates and are vital
ingredients for public policy and business planning and
investment decisions. As a result, they affect every American
who runs a business, saves for retirement or takes out a
mortgage.
In your wonderful summary, there was one area I noted that
was not mentioned--and it certainly does deserve mention,
especially as people worry about the new economy,--which is our
industry accounts. In addition to our national, regional and
international accounts you described, we have industry
accounts, which include gross product by industry, which
measures the contribution of private industry and government to
GDP, and the input-output tables, which show the linkages
between industries. These data are important because they
provide policymakers and business planners with critical
information to assess such issues as the impact of taxes in a
particular industry on other industries or the indirect impact
of growth in one industry on other industries.
I will now turn to one of the major topics you asked us to
discuss today, which is the accuracy of BEA's estimates.
Although our estimates of GDP and related measures are regarded
among the most accurate and timely in the world, they are not
without error. In order to provide timely estimates within 1
month of the end of the quarter, BEA must use partial data and
estimate missing source data in inventories, merchandise trade,
things of that sort. As more complete and accurate source data
become available in the following months, BEA revises the
estimates. In general, one finds that BEA's early GDP estimates
do a relatively good job of providing a general picture of
economic activity. In particular, the estimates can generally
tell you if the economy is expanding or contracting, something
of relevance right now; if growth is accelerating or
decelerating; if growth is high, average or low relative to
trend; what components of the U.S. economy are the main sources
of growth--consumer spending, investment spending,
inventories--or what is going on; what the general trend and
patterns are for key variables such as investment, saving
rates, or government share of GDP; and the timing of components
contributing to recessions and economic expansions. Where the
estimates have been subject to greater uncertainty is in the
measurement of longer-term growth rates.
Unfortunately in recent years, there has been a persistent
difference between growth as measured by production, or GDP,
and growth as measured by the incomes earned in production, or
gross domestic income. In concept, the two measures should be
equal, but in recent years the income measure has been growing
at a 4.9 percent annual rate while growth as measured by the
product side has grown at a 4\1/2\ percent annual rate, a 0.4
percentage point difference.
While there has always been uncertainty about trend growth
in the economy, the difference between the two measures is not
only larger than in the past, but the impact of such a
discrepancy seems to have a larger pocketbook effect. The
larger effect is due to the importance of BEA's estimates for
long-term budget projections and the reliance on BEA data for
the allocation of Federal funds to State and local governments.
The discrepancy also has had a larger effect on the economy
because of the increasing impact of BEA's data on financial and
foreign exchange markets. The impact of BEA's data on these
markets is more widely felt than in the past because almost
half of U.S. households now hold stock in one form or another,
an increasing share of loans are indexed, and with the
globalization of the U.S. economy, an increasing share of
businesses and households are affected by exchange rates.
In my written testimony, I focus on three examples of
challenges that BEA confronts in keeping up with the rapidly
changing economy. The first example deals with measuring GDP as
we move from an industrial economy to the new economy. The
second example deals with measuring the balance payments, which
as highlighted by the Trade Deficit Review Commission has
become increasingly difficult because of rapid changes in size
and complexity of international trade and financial
transactions. And the third is the need to better explain the
sources of the precipitous decline in the U.S. personal saving
rate through an integrated statistical treatment that focuses
on the impact of changes in the stock market and household
finances on personal savings. However, in the interest of time,
I will discuss just the first of these examples, the challenges
in measuring GDP.
One of the most difficult issues confronting public and
private decisionmakers is the uncertainty over the rates of
inflation and growth in the U.S. economy over the last 5 years
and their likely rates of change over the next 5 to 10 years.
BEA has had difficulty in keeping up with the changing economy,
and as I noted, errors have been creeping into BEA's measures
of trend growth in real GDP, incomes and inflation. Upward
visions in estimated tax receipts, or the ``tax surprises''
seen in recent years, have been, in part, the result of a
pattern of upward revisions in BEA estimates. BEA estimates are
an important factor in policy decisions that have a lasting
impact on the economy. Not only do BEA's estimates form the
baseline for the projections, but most long-term projections
assume that future growth will resemble the recent trends
published by BEA.
As Federal Reserve Board Chairman Alan Greenspan noted in a
recent speech, the biggest payoffs in efforts to improve
economic forecasts are likely to come from raising the quality
of data collected rather than improving forecasting techniques.
Small errors in real GDP can have such a large impact on long-
term budget projections that they can swamp differences in
proposed policy initiatives. Understatement of the growth rate
of real GDP associated with a given rate of inflation may lead
monetary policy officials to understate the rate of growth that
can be sustained without sparking higher inflation. Business
planners are also affected as they try to determine whether the
performance of the economy over the last 5 years is real and
permanent, the so-called ``new economy.''
BEA has worked hard in recent years to keep up to date with
the rapidly changing economy. Using resources made available at
BEA by eliminating programs, such as the leading indicators,
and utilizing improved data developed by BEA and its source
data agencies, the Bureau has been able to make a number of
advances. These include new price and output indexes that
better measure things such as banking services, cell phones,
computer software and the Internet. These accomplishments
notwithstanding, scarce resources and gaps in the source data
have prevented us from fully keeping up with changes in the
economy. The remaining gaps have a direct impact in the quality
of estimates. They include, first, for over 20 percent of real
GDP, mainly in services, there are no price indexes to produce
inflation-adjusted estimates, and the estimates are based on
measures of physical inputs and outputs or cost-based deflators
resulting in an understatement of GDP and productivity growth
and an overestimate of inflation for these components.
Second, for 20 percent of nominal GDP, also in services,
BEA has developed estimates using a broad range of source data
that differ significantly in coverage, concept, level of
detail, classification and timing. These inconsistencies
contribute to our persistent inability to keep up with changes
in this rapidly growing sector.
Third, the source data used in BEA's quarterly estimates
focus on the old industrial economy and cover only the wages
and salaries of production and nonsupervisory workers, thereby
missing over 40 percent of compensation in the U.S. BEA must
estimate the wages and salaries of these missing supervisory
and professional workers and estimate the impact of stock
options, in-kind benefits and other new forms of compensation
using a patchwork of partial data.
And finally, BEA lacks quality-adjusted price indexes for a
number of key products in telecommunications and other IT
areas, resulting in an understatement of real GDP and an
overstatement of inflation.
In summary, while BEA is doing a good job of measuring
today's economy, significant challenges remain. Discussing the
problems that new technologies and changes in the structure of
output pose for the measurement of GDP, Chairman Greenspan
recently noted, ``Certainly statistical systems in the United
States, both public and private are world class, and indeed, in
many respects, set the world standard. But given the rapidly
changing economic structure, one could readily argue that more
statistical resources need to be applied to understanding the
complexities of the newer technologies that confront
analysts.''
In the current fiscal year, BEA received its first real
increase in funding in nearly 8 years. The President's budget
blueprint for fiscal year 2002 proposes a $9 million, or 18
percent, increase in BEA's budget to extend the work begun in
fiscal year 2001. These funds would enable BEA to begin to fill
the gaps in BEA's estimates outlined above by developing new
price and output indexes for services and high-tech products,
new measures of compensation that measure the stock options and
rapidly growing forms of compensation that I mentioned, updated
measures of international trade and finance and integrated
measures of change in the real and financial economy.
Second and equally important, it would help us to upgrade
BEA's IT infrastructure so as to raise the efficiency and
accuracy of BEA's estimates, upgrade BEA's ability to
disseminate its data to its customers, and introduce electronic
reporting to reduce the respondent burden on the 40,000
companies reporting on BEA's surveys.
Thank you, Mr. Chairman, for this opportunity.
[The prepared statement of Mr. Landefeld follows:]
[GRAPHIC] [TIFF OMITTED] T5327.003
[GRAPHIC] [TIFF OMITTED] T5327.004
[GRAPHIC] [TIFF OMITTED] T5327.005
[GRAPHIC] [TIFF OMITTED] T5327.006
[GRAPHIC] [TIFF OMITTED] T5327.007
[GRAPHIC] [TIFF OMITTED] T5327.008
[GRAPHIC] [TIFF OMITTED] T5327.009
[GRAPHIC] [TIFF OMITTED] T5327.010
[GRAPHIC] [TIFF OMITTED] T5327.011
Mr. Miller. Thank you. We will proceed with Mr.
Knickerbocker. And everybody's written statement will be
included in the record. You may proceed.
Mr. Knickerbocker. Mr. Chairman, thank you for the
opportunity to participate in today's hearing on the activities
of the Bureau of Economic Analysis and the challenges BEA
faces. We in the economic programs part of the Census Bureau
collaborate with BEA in many different ways and very
frequently. While the data we collect are used by practically
all Federal agencies and are closely monitored by the Federal
Reserve Board, we regard BEA as our most important government
customer. A high proportion of all the data we collect serves
as source data for BEA. We are the principal source of the data
BEA uses to develop its product side estimates of the gross
domestic product.
Close collaboration between BEA and the Census Bureau means
that the two agencies share a common view of the most promising
opportunities for the improvement of economic statistics.
Two examples of how basic data are organized illustrate
this point. First, until a few years ago, the Federal
statistical system operated with an antiquated industry
classification system, the 60-year-old Standard Industrial
Classification system. In the last decade, a team established
by the Office of Management and Budget of Federal statistical
agencies designed a new, up-to-date and flexible industry
classification system. The result, it is called the North
American Industry Classification System, provides statistics,
profiling the American economy as it enters the 21st century,
not as it was at the time of World War II. The Census Bureau,
in cooperation with BEA and Bureau of Labor Statistics has led
the effort to introduce the new classification industry system
into Federal economic statistics.
Second, while the updating of the industrial classification
system represents a significant step forward, more needs to be
done. Firms and manufacturing industries make quite specific
products. Firms in service industries deliver quite specific
services. To generate the statistics that will support analyses
of many economic policy issues, for example, the sources of
productivity growth in the economy--data at the detailed
product level are required. This is especially true for
services where measuring the output of service providers is
particularly difficult. The Census Bureau, again, in
collaboration of BEA and the Bureau of Labor Statistics, is
developing a product classification system that will provide
the framework for the collection of substantially more product
level data then has been available in the past. The collection
task will fall to the Census Bureau. The task of putting the
more abundant data to work will fall to BEA.
Of late, officials at BEA has devoted much time to
measuring, describing and putting into perspective the new
economy. The one feature of the new economy that has attracted
much attention is E-business. The Census Bureau has pioneered
the collection of official statistics on E-business starting in
late 1999 with a collection of quarterly data on retail sales
over the Internet. This was followed by collecting annual data
on E-commerce activity in the manufacturing, retail, wholesale
and services sector. Detailed data on the E-businesses
processes used in manufacturing plants were collected at the
same time. The results of these collections have been released
in recent weeks with more results scheduled for release in May.
Our efforts at collecting data on E-business are in their
early stages. Still, our early efforts will give BEA some
baseline statistics from which it can develop its own measures
on the role of E-business in the economy. Looking forward, the
Census Bureau believes it can contribute to further
understanding of E-business by enhancing its collection of data
on business purchases of information, technology hardware and
software, the infrastructure of E-business.
Currently, the Census Bureau captures much of its data on
business expenditures for plant equipment through the Annual
Capital Expenditures Survey. Without too much change, we
believe this survey can be modified to pick up more specific
data on E-business infrastructure, an advance that should help
BEA perfect in its own investment statistics, a key element in
GDP, and these improvements in investment statistics would
certainly be welcomed by private industry.
Another feature of the new economy where BEA and the Census
Bureau have a common interest is in the increasing reliance by
business on leasing. Once upon a time, companies bought their
plants and bought the equipment they put in the plants. Once
upon a time, companies hired the workers that worked in the
plants. The company, its assets and its work force were all
under the same control. That simple world made it relatively
easy to collect data for a company and its operations. Now more
and more companies are leasing their assets and leasing their
employees.
These changes generate questions that make collecting data
more difficult. For example, who owns the assets? For example,
who is the employer of record for the employee? These and many,
many other sorts of questions are those that have to be
resolved by the Census Bureau to produce good data. The Census
Bureau is devoting substantial attention to developing
strategies to cope with leasing in its data collection efforts.
To the extent that we are successful, we should be able to give
BEA better data to factor this new business practice into its
picture of the economy.
At the Census Bureau, we also collect data via information
technology, and this approach has direct consequences for the
completeness and quality of the data we provide to BEA. For
close to a decade, we have collected some data through early
stage electronic means, but now we hope to take the next
obvious step, that is to say, offering the opportunity to
report over the Internet to the 5 million companies that we
will contact directly in the 2002 economic Census.
From experience, we know that electronic collection of data
pays off. For example, an increasing proportion of the data
required to be filed with the government at the time goods are
exported is now filed over electronic networks. About 50
percent of the paper documents, the paper documents that were
filed at the time of exporting, contained at least one error.
Today, the error rate for documents filed electronically runs
at 5 percent. The Census Bureau devotes substantial energy to
inspecting and correcting incoming data to assure the accuracy
of the data we release. Clearly, the cleaner the incoming data
we receive, the more we will be able to concentrate our efforts
to correcting the most troublesome data and the happier our
customers, including BEA, will be.
Finally, Mr. Chairman, there are some data projects that
the Census Bureau will work on as we gain in the productivity
of our programs. The projects would make the data that the
Census Bureau provides to BEA more useful. I have in mind
improved data on nonmerchant wholesalers, broader coverage of
service sector industries, more timely data on capital
expenditures by State and local governments, and more accurate
valuation of export statistics.
Mr. Chairman, that concludes my testimony. I thank you for
this opportunity to appear before you.
[The prepared statement of Mr. Knickerbocker follows:]
[GRAPHIC] [TIFF OMITTED] T5327.012
[GRAPHIC] [TIFF OMITTED] T5327.013
[GRAPHIC] [TIFF OMITTED] T5327.014
[GRAPHIC] [TIFF OMITTED] T5327.015
[GRAPHIC] [TIFF OMITTED] T5327.016
[GRAPHIC] [TIFF OMITTED] T5327.017
Mr. Miller. I thank you both for your statements, and I
appreciate you being here giving us a chance to talk about
this. I'm sorry some of my colleagues--because we adjourned
yesterday afternoon--have left town already. Let me start off,
first of all, about data collection and the quality of the
data. You say you use 5 million, you mention 5 million
businesses will be in next year's----
Mr. Knickerbocker. The Economic Census, sir, is conducted
every 5 years. It is conducted for the years ending in 2 and 7.
At the time of the Economic Census, we collect data from 22
million business locations in the United States. We collect
data on between 15 and 16 million business locations basically
through extracting certain data from tax records. We also
contact firms directly. By ``contact directly,'' we send out
questionnaires and/or we will deliver Internet questionnaires
to between 5 and 6 million companies. So that was the 5 that I
was referring to.
Mr. Miller. How about small business versus large business
as the cooperation and the quality of data. Small business is a
significant portion of our economy, of course, and the growth
of our economy, too. What is the challenge of small business
data collection?
Mr. Knickerbocker. That is one of the reasons that we make
such extensive use of tax records. Tax records give us the
name, the location, the nature of the activity and the revenue
of the business. And then to flesh out detail on small
businesses, we send out samples, let us say, of 60,000 firms,
in particular categories of small businesses to get the
details, like the typical purchase patterns of business, the
typical customer, and things like that.
So our first line of activity is basically to send as few
questionnaires as possible to small business, to try to use
what we refer to as administrative record, tax records, as an
alternative source of data simply so that we don't have to
pester small business persons. Then we use, as I say, sampling
techniques to gather a rich sense of some of the subsidiary
details of the small business.
Mr. Miller. What about the monthly quarterly data? You
don't use IRS data for that?
Mr. Knickerbocker. No. Once every 5 years.
Mr. Miller. Let switch over now to the monthly quarterly
annual data, the sources of that data, say, for small business.
How do you collect that data?
Mr. Knickerbocker. We do not collect data on small
businesses per se. We include small businesses in our samples,
for example, our monthly collection of data on manufacturing or
retail sales or wholesale, and in those cases, our sample
frames are built up to reflect the composition of those
industries, the number of small, medium and large size firms
incorporated in those sample frames, pro rata in their shares
of activity.
Mr. Miller. How about underground economy? The nonreported
income. Is that changing much in this country?
Mr. Knickerbocker. I would have to defer to my colleague to
the right because they have, for 10 or 15 years, been the most
venturesome in trying to come to grips with that very difficult
problem.
Mr. Miller. Mr. Landefeld.
Mr. Landefeld. By the way, I would say one thing about the
small businesses. In days gone by, when I first started in
statistics, you know you could collect a lot of dollars for the
economy by going to three major auto companies. But when you
begin to talk about things like auto repair services and other
services, it is much more expensive in terms of number of
firms. You have to survey to get that, which I think is one of
the reasons why we still lack data, so intensively, as I said,
in the services sector. For both the Census and BLS, those tend
to be sectors that are hard to measure and part of the reason
why they are not in our regular source data.
With respect to the underground economy, what we generally
do is try to measure just the portion of it which is not
reported to the IRS authorities. That is one of our major data
sources. So we use various data to estimate that. For example,
proprietors' income, according to the last taxpayer compliance
measurement program, which unfortunately is also known as the
``tax audits from hell'' program, which was abolished by the
Congress, but that was our last read on it. For every dollar
proprietors reported to the IRS, there was another dollar they
did not report.
So we carry forward a lot of those incomes that are
underground or simply not reported to the IRS in our estimates.
And we currently have no estimate of that, and one would think
that with the increasing reporting of everything from video
store receipts, etc., that would have some impact on
compliance. So that raises a lot----
Mr. Miller. So those tax audits from hell were a good
source of information for you that you are going to be lacking.
So that was your source of----
Mr. Landefeld. Right. Because the only way you can really
find out that information is through a lifestyle audit, that is
to find out if the person's receipts were far more than they
reported.
Mr. Miller. Talk about this sharing of data, and I know
when we went through the whole issue of the decennial census,
and the confidentiality of the data is absolutely crucial, as
the Census Bureau believes, for the participation in the
decennial. How much data sharing occurs now and how much needs
to be made additional, and comment about that. A couple people
mentioned data sharing in their statements, and then any impact
that would have on the ability to collect accurate data.
Mr. Landefeld. Perhaps I can comment first. From our
viewpoint, where we are integrating all this data, it would be
tremendously important because if you look at the data, for
example, from the Bureau of Labor Statistics, which collects
its own data and doesn't share it with the Census Bureau,
versus the Census for the very same industry, same time period,
significant differences in things such as sales and employment
occur in those industries. As we try to piece together our
picture of the economy, because most of our measures on one
side are based on income, the other based on Census type data,
we have very large problems in trying to integrate those
various data sources, and it would go a long way toward solving
many of the problems, including the discrepancies in the growth
rate on the two sides and a number of issues we confront.
Mr. Miller. What sources of data would you want to share?
Does the IRS share as much as you want to share? Whether they
should or not is another question.
Mr. Landefeld. I think the first piece of information we
would be interested in having shared would be the Census data
and the BLS data, which are integral to our input and output,
our national accounts, because we get different reads based on
that data. And by looking inside it and seeing how companies
are differently classified or what the differences in reporting
are, we believe we could fix a lot of problems in our
estimates. I mentioned that discrepancy where we have an income
measure growing at 4.9 percent and a product-side measure
growing at 4\1/2\, which causes no end of problems for
forecasts. Those are the kind of things we would hope to be
able to address. IRS data, we only can look at it selectively
for corporate profit returns. Census can look at it more
broadly than we can.
Mr. Knickerbocker. We at the Census Bureau have been in
support of the concept of data sharing. There have been, as I
am sure you know, several bills introduced to effect that in
the last several sessions and we have been quite supportive of
that. The classic example would be that we at the Census Bureau
maintain a business register of essentially every business
place, the basic facts on every business place in the United
States. At BLS they maintain a business register. Each of these
are complicated files of 7 or 8 million firms with are all
sorts of data on those. These are two parallel registers. To be
sure, they do serve somewhat different purposes. I don't think
if we had data sharing we could simply shut down one of the two
registers, but I think there is no question but that there
could be significant efficiencies gained in terms of how these
two registers would go on because there is certainly some
proportion of duplication right now.
So I cite that as an obvious example of some of the gains
from data sharing. We think that the quality of samples could
be improved. That is to simply say by sharing information one
could get an additional data point or two incorporated in our
data that would help us generate better samples and, vice
versa, for the agency to whom we might supply data. We should
be able to quit asking companies the same data, the same
questions, over and over again. Every questionnaire that goes
outs requires the respondent to give us the name of the
company, the location of the company, its EIN, plus five or six
basic facts. How many times does the company have to keep
saying the same thing over and over again? There ought to be
one repository in government that has all the basic facts on
companies, eliminating repetitive requests for data.
I would make this point, sir. We are very attracted to data
sharing. We should, however, mention IRS. Practically all the
data that the Census Bureau has--I should say the economic
program has on businesses is either directly or indirectly
derivative of certain IRS records and/or there is some IRS
content in those records. IRS, I think for perfectly
understandable reasons, has concerns about sharing, meaning
that it takes a much more restrictive view toward the sharing
of records than we do.
So here is a consideration should Congress pass data
sharing. My point is that Congress is going to have to
confront, to find some way to conform IRS regulations to data
sharing if data sharing is to be as fruitful as it might
otherwise be.
Mr. Miller. I guess it is also true with Census data that
other agencies want to use to project into the future. Did you
want to add something else?
Mr. Landefeld. I will add an example. Congress once passed
a piece of legislation that allowed BEA, BLS and Census to
share data on foreign direct investment, and as a result of
that sharing we were able to go, using our enterprise and their
establishment data sets, from having data by State for 66
industries to over 500 industries, a creation of a huge data
set on foreign direct investment with no additional respondent
burden, very little cost to the agencies overall. And that is
one example of the type of advantage you can get out of sharing
this kind of data.
Mr. Miller. You are familiar with the American Community
Survey. If it replaces the long form, it will be done on an
annual basis. What impact will that have on your data?
Mr. Landefeld. We mainly use that type of information on
our regional accounts, and it is our hope that with that
regular ongoing surveying that will go on as part of the
American Community Survey--I must say I am no expert at all on
this subject--but that regular surveying of larger geographic
areas, we think we will be able to get much better, up-to-date
types of information which we use in allocating data to the
regions, States, municipalities in the United States.
Mr. Miller. One of the things about data is the timeliness
of the data, as you know there was a discussion with Mr.
Greenspan, about how fast he can react and how accurate the
data is and you come up with the best estimates you can and
then you revise them. In our next panel I would like to talk
about this, as well, is what happened in the 1990 recession
period and the data and how the data changed. Would you comment
about that? I know we are going through economic times now that
Mr. Greenspan wants accurate data.
Mr. Landefeld. One hates to extrapolate from that one
episode. For most of the postwar period we have done a pretty
good job, but that is indeed one of the misses we had in terms
of the particular timing of that business cycle. We did show a
turndown at that time, a slowdown in economic activity--but not
nearly the decline that we had then. And I think that is
somewhat worrisome because as I look right now, for example, at
the data, one of the most important components of our estimates
that is helping to hold up the economy in the current period is
investment in computer software. And while the annual data on
that are pretty good, I do worry about the quality, and we are
working to try to improve the quality of the quarterly
estimate. If the slowdown we saw in computers were also
reflected in software, we would have seen several tenths at
least taken off the real GDP growth rate in the last quarter,
which I think psychologically would have been important because
it would have put us below 1 percent growth rate in our
estimates of the slowdown.
So there are a number of components of that sort and
services in many of the industries I have mentioned where we
are using very crude extrapolators for a lot of components that
are either new economy or in services. And that does worry you
because it is only when we get the annual surveys, and in the
case of many of those services only once every 5 years do we
get data on all service industries as part of the quinquennial
census. So there is an awful lot of extrapolation going on with
all kinds of partial data that does worry you in terms of our
ability to capture the timing changes in the U.S. economy.
Mr. Miller. We had the problem with the CPI and the market
basket problem and adjusting to that with the new economy, and
they are making the adjustments and proceeding. You mention
about changes taking place. Are you able to adjust quickly
enough to changes in the economy? We are going through this
change and I think Mr. Greenspan said we are perhaps 25 percent
through this technology revolution. And I don't know whether we
are at 50 percent or 10 percent or 75 percent, but obviously
there are many changes going on. Are you able to quickly
react--I shouldn't say quickly, but react properly to that type
of change? As you say, there are new industries new products,
everything.
Mr. Landefeld. I don't mean to be a two-handed economist,
but the answer is yes and no. We were one of the leaders in
developing price indices and quantity indices where the weights
changed every quarter, eliminating some of the biases that were
and are now being addressed in the CPI. So with respect to that
the Bureau was one of the leaders, and it actually eliminated a
very large bias in real GDP. That was much larger than the bias
we all heard about in the Consumer Price Index.
So on that score the answer is yes, but in a very important
way the answer is no, because for a lot of high-tech products
and services that use high-tech products--insurance, the
securities industry, the data we are using are those that I
described as input-based or output-based estimates. And as a
result, if we count output based on input, we get zero
productivity growth by design and understate the rate of growth
in real GDP in that industry and also overstate inflation in
those industries. So we still have serious problems in keeping
up with changes in the economy and high-tech sectors. We don't
have quality-adjusted prices for local area networks and all
kinds of things of that sort. We are working very hard at
developing, as I mentioned in terms of cell phones and others,
but an awful lot of work remains. The President of the American
Economic Association, Dale Jorgenson, has made this point in a
number of his papers in assessing the new economy, that a major
part of the problem in assessing the new economy is the fact
that there are so many sectors that are major users of IT and
also products that are produced that are high-tech that are not
appropriately measured, and that tends to bias the results one
gets in looking at the, ``new economy.''
Mr. Knickerbocker. If I could speak to that point. I
mentioned in my testimony about e-business. Certainly the
concept of the Internet was known throughout all the 1990's,
but really the Internet as a way of doing commerce really took
off in 1998. By late 1999 we were, as I indicated, gathering at
least the first sorts of data on activity over the Internet,
retail sales over the Internet. Were we gathering data on day
one when it became important to gather data on the Internet?
No, but we gathered data on it within a year of the time when
it surfaced as an important element in our economy.
So we are in the lead in gathering data, and we are
certainly very mindful of the task. We are also aware of
changes in business practice and of our obligation to generate
some data on them as quickly as possible.
Mr. Miller. How much of a problem does making adjustments
in your data over time cause you? And to the comparability of
the data?
Mr. Landefeld. That is a major concern as one compares
current periods to past periods. We at the BEA have prided
ourselves in keeping a nice consistent time series. Every time
we do a revision we go back to 1929. But I must say it is
getting more and more difficult to do. You can only extend the
series back so far. That is a major part of our job. The Bureau
of Labor Statistics just introduced a new price index for
securities brokers and dealers at our request. Unfortunately,
they only gave us 6 months of data because they are in the
current process of estimating current prices, and we have got
to work to extend those backward. But we are finding
increasingly our ability to do so is limited.
Thank goodness, some of these products did not exist in the
past so you only have to extend it so far back. But there is
the whole question that many academics have pointed out, Bob
Gordon in particular of Northwestern, that there were a lot of
innovations back then that we may not have fully captured the
impact of. So there may be some things we are missing in the
past. Some of the examples like computers are so egregious you
had to do something with them. And I think that is what we have
tried to address, that is the examples where we really
absolutely must do something because the rate of decline in
both the price per unit of computers and the quality-adjusted
price is so large you have to estimate for that. But we are not
about to go out trying to adjust every price that is out there.
Mr. Miller. Looking down the road when you start projecting
5, 10 years in to the future, right now there is a lot of
debate about tax cuts 10 years in the future in Congress, as
you know, and 10 years ago what was the projection? How far
would you have been off 10 years ago, from 1991 to today? Maybe
the next panel would be able to answer that.
Mr. Landefeld. I really can't tell you. All I can say right
now, and I think Dick Berner may address this and certainly Bob
Dennis from CBO, but most rules of thumb say over 10-year
forecasts about a 0.1 percentage point error in real GDP can
produce errors in 10-year projections of $200 billion or more,
depending on whose rules of thumb you are using, CBO or OMB.
That is the reason the differences in the growth rate are so
important. It is just one-tenth of 1 percentage point that has
those kind of $200 billion effects over time. That is why we
are particularly worried about this 0.4 percentage point
discrepancy between our two measures of growth.
Mr. Miller. One more last question, because we need to go
on to the next panel. An area that I have a great interest in
is what is going on in biotechnology. How do you plug that into
longevity, life expectancy, I mean, revolutionizing--the impact
on the economy, on trade?
Mr. Landefeld. Gee, I am kind of boggled. We are having
enough problems just measuring pharmaceutical prices.
Mr. Miller. But that is the future.
Mr. Landefeld. Clearly that is another form of information
technology investment which is becoming increasingly important.
Our first crack at this kind of thing was the capitalization of
computer software, but it obviously influences the market
valuation of firms, that kind of biotechnology. So it is
something we can and should be measuring. It is on our long-
term agenda. I think there is a recent Brookings study on
exactly this issue of what those kinds of things are worth and
their market value. I think that study panel urges us to move
forward on that, but I must say our current concerns are so
large that is a little down the road for us.
Mr. Knickerbocker. If I could add to that, sir. What Steve
is saying is what I see is our greatest challenge. It is
relatively easy to collect data on physical capital, bricks,
buildings, equipment, things like that, but today horsepower is
becoming less important and brain power is becoming more
important. Human capital, intellectual capital, and how we
measure human capital, which is the driving force in business
today, explaining human capital and collecting the basic facts
on human capital that has got to be the No. 1 challenge that we
have before us.
Mr. Miller. It affects trade data significantly, too,
doesn't it? We are a major exporter of that.
Mr. Knickerbocker. If we knew what our exports statistics
were to the nearest 7 percent, we would be better off, sir.
Mr. Miller. Let me thank you all. Do either of you want to
make a concluding comment?
Then we will move on to the next panel. It is a huge
challenge you all have and you have got a great deal of
credibility and respect. And I think the recognition that
Congress finally gave you, an increase last year, and certainly
my understanding is President Bush's budget will include a
generous one next year, shows the recognition that we need to
continue to work to improve, and it is an amazing challenge you
have. So thank you all very much for being here. I look forward
to working with you.
Mr. Landefeld. Thank you, Mr. Chairman.
Mr. Knickerbocker. Thank you, Mr. Chairman.
Mr. Miller. We will take a second to allow you all to move
and we will let the next panel have a seat.
Welcome. Our second panel includes representatives of the
Congressional Budget Office and industry associations who are
active data users and advocates of the Federal statistical
system. We have Bob Dennis, who is the Assistant Director of
Macroeconomic Analysis of CBO, the primary source of budget
information for Congress. Richards Berner is the current
president of National Association of Business Economists, whose
members have a vested interest in accurate and timely economic
statistics. Diane Swonk is the chief economist and senior vice
president for Bank One and the immediate past president of the
NABE. Gordon Richard is an economist representing the 14,000
member National Association of Manufacturers. And Professor
Ernie Berndt joins us from MIT, Sloan School of Management.
Professor Berndt also chairs an advisory committee to the
Bureau of Economic Analysis, Bureau of Labor Statistics and the
Census Bureau.
I thank all of you for being here today. We will start with
Mr. Dennis.
STATEMENTS OF BOB DENNIS, CONGRESSIONAL BUDGET OFFICE,
ASSISTANT DIRECTOR, MACROECONOMIC ANALYSIS; RICHARD BERNER,
PRESIDENT, NABE; DIANE SWONK, CHIEF ECONOMIST, BANK ONE, INC.;
GORDON RICHARDS, ECONOMIST, NATIONAL ASSOCIATION OF
MANUFACTURERS; AND DR. ERNST R. BERNDT, MIT, CHAIR OF THE
FEDERAL ECONOMIC STATISTICS ADVISORY COMMITTEE
Mr. Dennis. Good afternoon, Mr. Chairman. Mr. Chairman and
members of the subcommittee, I am pleased to be here today to
discuss some of the major issues affecting the Bureau of
Economic Analysis, which is the enormously respected keeper of
the national income and product accounts. In my testimony I
will focus on the crucial role that those accounts play in
shaping public understanding of the U.S. economy and helping
the Congressional Budget Office to construct its baseline
budget projections. I will also note several ways in which
BEA's data might be improved.
It is not too much to say that the national income and
product accounts are what make modern empirical macroeconomics
possible. Those accounts are the organizing principle that
enables us to see how the parts of the economy fit together.
The accounts are also the foundation of CBO's economic
forecast, which underlies the baseline budget projections that
the Congress needs to do its work. We use those accounts both
to track what has happened in the past and to ensure that our
assumptions for the future are internally consistent.
The economy of course does not stand still but keeps
changing its structure. In the past decade, forecasters and
analysts have had to cope with the sets of changes that have
come to be called the new economy. And as we have heard, those
changes have posed special challenges to the statisticians at
BEA, who have done an excellent job of meeting them. However,
CBO believes that some further progress can be made, and in my
testimony I will suggest some areas for improvement. Many of
those improvements would require changes in procedures not only
at BEA but also at the agencies that provide BEA's source data.
As we have heard, BEA is not by and large a data gathering
agency but gets its data from the surveys and economic censuses
at the Census Bureau, from the Bureau of Labor Statistics
[BLS], from administrative records such as tabulations of the
IRS, and from various private sources. Some data improvements
may also require additional reporting by businesses. In those
cases, of course, it would be necessary to assess any
additional burdens that those requirements would impose, and we
have not made any such assessments.
Let me first briefly describe how CBO uses BEA data. Those
data play a large role in CBO's budget projections because they
provide the foundation of the economic projections, which in
turn underlie both the revenue and outlay projections. BEA
data, along with data from BLS or the Bureau of the Census, are
the key supply-side inputs used to explain economic growth.
Besides contributing to CBO's economic projections, BEA
data also helps more directly in CBO's projections of revenues.
Revenues are sensitive to the distribution of national income
between wages and salaries and corporate profits. BEA provides
measures of those incomes, and CBO projects those measures
forward as part of its overall economic projections. BEA's
estimates of the capital stock, moreover, which determine how
much corporate income must be assigned to depreciation, also
have an important influence on the relationship between output
and revenues.
Now let me turn briefly to the challenges of the new
economy for forecasters and statisticians. What people mean by
the new economy is a complex of developments, particularly over
the last decade, including rapidly falling costs for
information technology [IT] and consequently for information
itself, changes in the organization of production as firms take
advantage of the lower cost of information, and the
proliferation of new companies doing new things, which are
always among the hardest to track.
To understand what is happening, forecasters need a
statistical system that can keep pace with the changes in the
economy. One of the main tasks of the statistical system is to
separate economic growth into the share that reflects price
changes and the remaining share, which reflects the real growth
of the economy. Developing good price indexes is often
difficult, however. The quality of most goods and services
changes over time, and price indexes must take those changes
into account.
For example, even though a computer now may sell for
roughly the same price as a computer last year, few people
would be happy to purchase last year's model rather than this
year's. The same number of dollars this year buys vastly more
computing power than it did last year, and that improvement in
quality has to be reflected in the price index. BEA has led the
way in improving estimates of the contribution of computers.
The estimates are often rough, but they are generally
preferable to ignoring all of the available information about
changes in quality.
Nevertheless, there are still important areas where further
improvements in the measurement of prices and quality could
greatly improve our understanding of the new economy. One such
area is communications equipment. According to a forthcoming
CBO analysis, the lack of good quality adjustments for that
same equipment may have resulted in an underestimate of real
investment growth of about 0.6 percentage points per year, on
average, between 1996 and 2000.
There are also places outside the IT sector where current
techniques could represent what is going on in the economy. For
example--this has already been mentioned--two Federal Reserve
economists found that reported productivity growth in many
service industries was persistently negative between 1977 and
1999, even though firms in the industries remained profitable.
They found that if they replaced those unexpected negative
productivity growth rates for several service industries with
an estimate of zero, the overall productivity growth would then
be reported about 0.3 percentage points higher. That is overall
productivity growth.
Finally, let me mention a couple of ways in which the
statistical system could be even more helpful to CBO in doing
its economic and revenue projections.
First, we could use better and more current estimates of
wages and salaries under withheld income and payroll taxes.
Steve Landefeld mentioned the problem of data on supervisory
and professional employees. Other problems arise from the
exercise of certain stock options, which ought to be part of
wages and salaries but which are not currently captured by any
government statistics. The lack of data on stock options
distorts our understanding both of the growth of wages and of
tax trends. We understand that BEA is investigating ways to
improve those data, and we look forward to its results.
Second, contemporaneous information on the sources of
withheld tax payments would be very helpful to CBO as well as
to BEA. Employers are not asked to report contemporaneously on
how much of the tax they withhold is due to payroll taxes, even
though they have to calculate payroll taxes and income taxes
separately in order to know how much to remit. As a result, BEA
and tax analysts have to make do for more than a year with
estimates of that split, which complicates the tracking of tax
credits. Technological advances, however, may have made it
cheaper for businesses to give us those data in real time.
I have some additional discussions of these suggestions and
others in my written testimony. BEA is already working on most
of them, and indeed it has a much better and more comprehensive
list than we do.
I would just like finish with the following thought. The
new economy poses severe problems for national income
statisticians, but it may also offer an opportunity. The IT
revolution has lowered the cost of information, and that is
having dramatic effects on the way businesses produce and use
information. The IT revolution also offers the opportunity for
government statisticians to gather more useful data without
intruding into or imposing excessive burdens on private
business.
Mr. Chairman, I will be glad to answer any questions.
[The prepared statement of Mr. Dennis follows:]
[GRAPHIC] [TIFF OMITTED] T5327.018
[GRAPHIC] [TIFF OMITTED] T5327.019
[GRAPHIC] [TIFF OMITTED] T5327.020
[GRAPHIC] [TIFF OMITTED] T5327.021
[GRAPHIC] [TIFF OMITTED] T5327.022
[GRAPHIC] [TIFF OMITTED] T5327.023
[GRAPHIC] [TIFF OMITTED] T5327.024
[GRAPHIC] [TIFF OMITTED] T5327.025
[GRAPHIC] [TIFF OMITTED] T5327.026
[GRAPHIC] [TIFF OMITTED] T5327.027
[GRAPHIC] [TIFF OMITTED] T5327.028
[GRAPHIC] [TIFF OMITTED] T5327.029
[GRAPHIC] [TIFF OMITTED] T5327.030
[GRAPHIC] [TIFF OMITTED] T5327.031
Mr. Miller. Thank you.
Mr. Berner.
Mr. Berner. Mr. Chairman, thank you for this opportunity to
appear before you. Today I am here in my role as president, as
you indicated, of the National Association for Business
Economics [NABE]. We are a professional organization for people
who use economics in their work, and our mission is to provide
leadership in the use and understanding of economics.
As you have heard from some of the other people in this
room, the national income and products accounts are really
critical for evaluating the forecasting and understanding the
U.S. economy. And I just want to leave you with the point that
from our perspective it is essential that these data faithfully
portray the rhythm of economic activity as well as the separate
parts of a very complex $10 trillion economy. As Bob Dennis has
noted and as Steve Landefeld also noted, these data are
essential for your policy deliberations, particularly with
regard to the budget. Steve and Bob have talked about some of
the improvements that have been made in our Federal statistical
infrastructure as they are used by BEA.
I want to emphasize the fact that, as has already been
said, our economy is constantly changing. The industrial
economy of the past has given way to the very different
knowledge-based information economy, and that constant
evolution obviously requires both new sources of data and
resources for agencies to collect and analyze them. While our
statistics remain among the best in the world, lack of
investment in our infrastructure has left us with a system that
still does a better job of measuring infrastructure activity
than information-based output.
The new data initiatives that have already been discussed
cover services and high tech industries more comprehensively
and more accurately than only 4 years ago, yet major gaps
remain. The most important industry in some statistical tables
is still the one labeled ``all other.'' While BEA makes every
effort to ensure that its four major set of accounts, national,
industry, regional and international, tell consistent stories,
holes in the data often make that impossible.
Steve did not tell you, I don't think, that statisticians
must estimate from a patchwork quilt source data roughly 20
percent of the GDP. Moreover, it has been discussed already
that data on prices that enable us to separate inflation from
real growth are often lacking. Steve did mention the software
investment is one area where he has incomplete data and where
he has to make estimates. At my firm, Morgan Stanley, we have
surveys of businesses that may tell a somewhat different story
from the extrapolations that the BEA has to make.
Now, here is the punch line: More and better data obviously
require more funding. And you have heard that before. I want to
tell you that business people and policymakers increasingly
recognize that funding improved statistics in general, and the
GDP accounts in particular will pay huge dividends. My friend
to my right, predecessor as NABE president, Diane Swonk, will
recount for you in a moment the broad support that these
efforts have in the business community.
For his part, Fed Chairman Greenspan also supported that in
his comments last week.
You asked a question just a moment ago about biotechnology.
Fed Chairman Greenspan indirectly addressed that by asking
whether or not when we consider the cost of medical procedures,
how we should measure prices of those procedures given the
advances in technology that have been made. And that is a
question that Director Landefeld, Nick Knickerbocker, and
others in our agencies will have to grapple with.
Personally, we agree with Fed Chairman Greenspan that
greater payoffs will probably come from better data than from
more technique and so does our membership at NABE. Our members
recognize the importance of funding constraints on enhanced
data gathering. That fits our longstanding support for
maintaining fiscal discipline. Our members consistently
supported moving to a balanced budget since we began polling
them on policy issues 25 years ago. However, we also recognize
that the costs of incomplete and inaccurate information far
exceed the combined budgets of our major statistical agencies.
In a survey published just last week, 70 percent of NABE
respondents favored increasing spending on economics
statistics. They ranked such increases first among seven
alternatives for increased Federal spending including education
and infrastructure. Don't get us wrong, those are important.
But these investments will pay huge dividends. That is not
surprising. We have long been concerned about improving the
quality and timeliness of these data. In 1985, NABE created a
statistics committee, chartered to work for the improvement of
the national statistical system. Along with Chairman Greenspan,
we supported efforts to reduce bias in the consumer price
index. And working closely with the Council of Economic
Advisors, the committee developed recommendations for data
improvement.
I would add, Mr. Chairman, that we would welcome the
opportunity to work with you toward that end.
NABE believes that our national data collection efforts
should be as efficient as possible. You will hear from me and
others that toward that end we believe that Congress should
mandate data sharing among the agencies solely for statistical
purposes. As you know, confidentiality statutes that permit
data to be seen only by the employees of a single agency
present a formidable barrier to effective working relationships
among the agencies. They virtually guarantee duplication of
efforts and inconsistencies among related data sets that you
have already heard about. Moreover they deny, in effect,
agencies' resources from undertaking new analyses that could
improve the information available to policymakers. This is not
a cost-effective way to run any business--either public or
private.
Federal statistical agencies and others such as the Federal
Reserve are already cooperating in several ways to improve our
statistical infrastructure. But I believe that permitting data
sharing would take that cooperation to a new level.
Consequently NABE supports reintroduction of the Statistical
Efficiency Act of 1999. It was passed unanimously by the House.
This legislation would permit exchange of statistical
information under specific statutory controls. In summary, Mr.
Chairman, NABE supports enhanced funding for improved economics
statistics; and we also support the efficient use of those
funds through data sharing among Federal agencies. I would be
happy to answer any questions you may have.
Mr. Miller. Thank you.
[The prepared statement of Mr. Berner follows:]
[GRAPHIC] [TIFF OMITTED] T5327.032
[GRAPHIC] [TIFF OMITTED] T5327.033
[GRAPHIC] [TIFF OMITTED] T5327.034
Mr. Miller. Diane Swonk, please.
Ms. Swonk. Thank you for allowing me to speak on something
that is so close and dear to my heart given the work that both
Dick and I did last year to try to get people to recognize the
issue on the U.S. statistical agencies and the funding that
they need. I commend their efforts to try to improve the data
in what was a harsh funding environment for so long. I am just
going to provide some summary comments from my remarks as you
already have them on file. And I am dyslexic so I am really bad
at reading them any ways. Dyslexic economists are kind of
dangerous since we flip numbers around as well.
I would like to start with my view that economics is at its
very heart the study of collective human behavior, one of the
hardest concepts for us to even imagine measuring. I think, to
paraphrase Chairman Greenspan, which all of us are doing since
he gave such a timely speech last week at NABE meeting, he did
talk about an economy that is increasingly dominated by ideas
instead of material inputs or manual labor, as one that is
putting significant stresses on our ability to--on our
statistical systems. With that said the U.S. economic
statistics many times represent our only true light in what is
becoming an increasingly dense forest of global economic
information. Business leaders and the press have already begun
to recognize the magnitude of the issue and they realize that
statistics shape everything from our own strategic risk
assessment at the banks, strategic planning, to portfolio
management. And just the rumor of one of these statistics being
out of kilter from where many are expecting, we know can move
billions of dollars around the world in a split second now.
Moreover the gap left by what has been taken away in terms
of what is now faulty or incomplete data provided by the U.S.
statistical agencies has left many of us to rely on private-
sector information. Dick pointed out that his firm now does its
own surveys which are commendable but there are many a survey
that provide a sliver of information in what is really only a
piece of a much larger, more complex puzzle. I think of things
like the National Association of Purchasing Managers index--
which before the last Fed meeting just because it happened to
come out before the January 3rd surprise inter-meeting Fed
meeting, people all now think that is what moves the Fed which
is utterly ridiculous that one number would move the Fed to do
an inter-meeting move like that. Especially one number that is
not held accountable to the same kind of accountability our
U.S. statistical agencies are held accountable for.
There is also today the Challenger, Gray & Christmas survey
was released recording lay offs. These surveys never state when
the lay offs are going to occur, whether they are due to
attrition, how much they are going to show up in the
unemployment statistics, and really tell us much more about
structural change in the large corporate sector rather than, as
you pointed out earlier, what is so importantly going on in the
small business sector. Small businesses don't have to name how
many people they hire or how many people they are able to hire
now after complaining in other surveys they have not been able
to hire for years and now finally have some workers to hire. So
I find that an important point to make as well.
At worse, some of these issues, in terms of these private
surveys that are now becoming so popular, 15 years ago nobody
even paid attention to some of these surveys, I might add, that
are out there. They give a distorted, inaccurate view of the
macro economy. It is not to say that they were put in
unscrupulous private sectors hands. I represent the private
sector, and I know incentives well. And knowing that your
statistic might happen to move a market is an incredible
temptation to take a position on before it actually comes out.
That is one reason why I believe in the U.S. statistical
agencies and that the data should come from the government. I
don't believe a lot of things should from the government, but I
believe in fiscal discipline but certainly with prudence
funding the statistical agencies.
In response to all these issues, businesses have taken
things into their own hands investing in extraordinary
information technologies. My own company, Bank One Corp., is
now looking to increase its investment in the Intranet and
Internet to be able to know real-time information on anything
that is going on in any 1 of our 14 states of dominance in any
one of our business lines. That is very important to us, but
severely compromised now is our ability to be able to forecast
some of the trends that helped shape the strategy of the bank
when I first started.
My first forecast that I ever made was in 1986 for the
renaissance in the Midwest economy trying to get the bank
focused on looking to the Midwest rather than New York to be a
bank and looking at its own comparative advantage. I am not
sure I can make that same forecast today given the lack of
regional data and the lack of quality in the regional data that
is now available, because as the statistical agencies have had
to make cutbacks in their priorities, prioritize what they do
cover, regional has often gotten short shifted. We do not know
retail sales in any State in the country, your State, we do not
know the retail sales in your State. It seems so utterly
ridiculous when you are thinking about I helped many a State
and local government try to forecast revenues and understand
their economic environment with fewer and fewer economic
information on that front is--I think is a huge problem.
Also I think why shouldn't the statistical agencies have
the same ability that we have given the private sector to
automate and aggregate data that is now being collected in the
private sector. This would far increase efficiencies and
sometimes inaccuracies filled out by the wrong people by
surveys in the private sector. I am very much in support of
increased investment in infrastructure in the statistical
agencies. This goes far beyond just supporting data collection
and quality data. It is talking about really raising the bar on
the kind of information we can collect in a new information
world. And if we don't make those kinds of investments, the
kind of data we are going to be getting is yesterday's data at
best rather than today's data which is so critical to
policymaking and other issues.
I have already talked about some of the issues that we
face. I think underscoring the risks, I think you referred to
it a bit earlier, of faulty or lagging economic information you
noted the 1990 situation where as late as October 1990 Chairman
Greenspan was trying on record to reassure an increasingly
skeptical public based on data that said we were still in a
slow but economic expansion, not a recession. It wasn't until 2
years later upon the revision of that data in 1992, that we
actually saw in the data a recession acknowledged. A recession
that actually began 2 months before Greenspan was making
comments on record that he thought the economy was still
expanding given the economic data.
We don't know what history would have changed if that
information had been available, but clearly it points out the
need and the need for continual increases in the accuracy of
the data.
I also note the importance of the 1997 and 1998 financial
crises that rocked global markets around the world certainly
required the Fed and the Treasury to intervene in 1998 to
stabilize what had been a liquidity freeze in our own financial
markets in the United States because of, in part, faulty
information around the world. The information that we see in
the United States is the best in the world, is the most
transparent, and the most accountable. Other countries that do
not have or are not as well funded as we are even with our
needs for funding have far less credible data and the
transparency issues are clearly not there. People were making
investments without clear information of what those investments
were assuming they had the same kind of information that we had
here and we got caught very hard by that issue.
Also as has been already mentioned is the budget debate and
how important the source data that goes into the debate is.
Just having that data--know that it is going to be revised in
and of itself makes this question the outlook. We could spend
all day debating the assumptions on the forecast, but I think
we would all agree at the end of the day that having the best
source data possible is the only way to possibly get to any
kind of a close and accurate end point in the data.
I will return to where I started to some extent and say
that efforts to improve the quality of U.S. statistics are
commendable but still fall far short in catching what I think
is a moving target: A rapidly evolving information-based
economy.
The statistical agencies have suffered from neglect and a
lack of advocates. I noted in my comment that the word ``data''
appears to be the most uninteresting four-letter word in the
human vocabulary, not attracting much attention out there. NABE
has certainly, I hope, changed that. It was our goal when Dick
and I sat down last year, our goal starting back in the mid-
1980's to make this a more national debate on statistics to
underscore the importance. And I think we have raised the
volume if nothing else.
Dick pointed out how our diverse multinational membership,
70 percent, agree. Do you know how hard it is to get 70 percent
of economists to agree on anything? That is a really remarkable
thing when--and it has been the same every year. The
overwhelming majority of our members choosing that as their
most important objective.
Moreover I think what I have been stunned by is our allies
in every corner. I mention in my comments, last year when I was
working on the lobbying effort to increase funding for
statistical agencies, actually had one CEO return a call from
his vacation because he thought it was so important to get back
to me to be included on a list of people writing letters in
support of statistical funding. It spans party lines. I have a
list that I started--actually I couldn't finish in time to get
here because of my travel schedule but just in 1 day I was able
to get seven CEOs that I called around and actually got a hold
of personally to say please include me on the list, Diane.
Every single person we have approached has come back to us
with, of course, we support you. And many of these CEOs have
also gone to great lengths to write many a letter to many a
Congressperson in order to keep that support out there. I think
that is really important. Our only--there are no enemies in
this game of the statistics. There are no people out there
against us. We are an advocate, but we don't have enemies.
Our only true enemy is complacency. I urge--and this is
certainly following after some of the things that you have
heard from other people, but funding on quality and timeliness
of the economic statistics really includes everything from
funding for infrastructure to funding for competitive pay
packages in this economy. Despite the slowdown, I was just out
yesterday at a customer in Springfield, IL, who is paying
$700,000 a month to temporary workers, double the wages of
their existing workers, just to fill positions. This is very
important to continue to have quality people to be able to work
and fund in the funding of these agencies.
I also encourage investments in infrastructure. Why
shouldn't we share data between agencies? And investments in
infrastructure could make the sharing of that data much more
rapid, much more efficient, and much more accurate, frankly.
Also I think it would also make--investments in infrastructure
could make the collection of data much more accurate.
Finally, I think it is important to point out funding for
research techniques as well. One of the things that we do in
this country better than others is we actually know how to
survey for statistical information better. And not only do we
need to continue to improve upon that, especially in an idea-
based economy, I think we also have a responsibility for
ourselves that would payoff not only for the United States but
for the global economy for many decades to come to continue to
invest in the quality of research on statistics and export that
technique abroad so that other economies we deal with are
playing in the same playing field we are. This would mean
enormous returns for our own financial markets and could add
much stability where we have seen instability in the recent
past.
I guess my last quote from Chairman Greenspan, which has
been quoted very much today because of his support on this, he
said about a little over a year ago in a question asked of him
to a Senate panel, that when it comes to statistical funding, I
am extraordinarily reluctant to advocate any increase in
spending so it has to be either very small and/or very
formidable argument that is involved. He said, and I find in
this case, regarding U.S. statistical agencies, both conditions
are more than met.
And I think that sums up our membership and certainly those
of us who have to deal with this on a daily basis, and also
every
CEO I talk to feels it is very important to their business
lines and their conduct of business not only in this country
but abroad. Thank you.
Mr. Miller. Thank you.
[The prepared statement of Ms. Swonk follows:]
[GRAPHIC] [TIFF OMITTED] T5327.035
[GRAPHIC] [TIFF OMITTED] T5327.036
[GRAPHIC] [TIFF OMITTED] T5327.037
Mr. Miller. Mr. Richards.
Mr. Richards. Thank you, Mr. Chairman. As a professional
statistician, I think that the BEA has done an excellent job;
and as a statistician in the business world, we really put our
money where our mouth is. We get inquiries from the
manufacturing sector all the time as to what statistics they
should be looking at, what time series they should be using for
particular problems. And I always refer them to the government
agencies.
It is not always the BEA, as sometimes we think they should
look more at the index of industrial production compiled by the
Federal Reserve or the shipments data compiled by the Census
Bureau, but I invariably tell them to rely on the government
data. I get quite a few inquiries about some of the private-
sector surveys. I wasn't going to put this in my written
statement, but I think many of the private-sector surveys
provide misleading and inaccurate information. The government
agencies have made much more of an effort to make the data
accurate, reliable, and it is actually quite user friendly. The
problem for the private sector is getting enough non-economists
out there to be aware of the data sources and to give them some
guidance as how to use that. In fact, this is one area in which
the government data is vastly superior to most of the
alternatives.
As far as what I think the BEA has been doing right for the
last 10 years, let me cite three examples. First of all is the
adoption of chain-weighting in GDP, which is a major
innovation. And we certainly see this in terms of say the
relative difference between growth and inflation, that is, the
share of nominal output that is compromised by growth and
compromised by inflation. If we hadn't had chain-weighting we
would be reporting a higher rate of inflation at a lower rate
of growth. This has very clear policy implications for the
Federal Government because transfer payments were indexed to
the Consumer Price Index. As a result, the Federal Government
ended up spending more than was absolutely necessary on these
income transfers.
The second big innovation the BEA has made is the
redefinition of GDP to include software. And as any computer
programmer can tell you, software should not be treated as an
intermediate input such as raw material. It is a valuable
productive tool which in turn can be used to generate value
added.
The third major innovation that the BEA has engaged in,
which again we agree with completely, is the imputation of
quality improvement to the computer sector. The way that they
have done this is to take a weighted average of computer
processing and capacity and add that to the real value of
computers. If you do not do this quality imputation, which has
been somewhat controversial, you end up with extremely low
estimates for the rate of growth; and this, in turn, has very
significant implications for policy decisions. Of those three
innovations, I think the two most significant are the quality
imputation to computers and the redefinition of GDP to include
software.
Throughout the 1990's, there has been a debate in which we
have participated as to how fast the economy can grow at a
stable inflation rate. For a long time, we had this situation
in which the growth rates that were being reported were
relatively low but the inflation rate was continuing to decline
and national income was growing faster than national product.
Now, the increase in income relative to product suggests
that the cause might be hidden productivity. It was a paradox
that was resolved when the BEA adopted these innovations and we
discovered that the missing output wasn't missing at all.
Rather it was the output that was being generated by the
quality of computers and by the inclusion of software in the
national income accounts.
We have also done our own production function studies on
this issue, and what we find is that using reasonable measures
of technology, a theme I would like to return to in just a
moment, we get estimates suggesting that the productivity trend
in the United States could be sustained at something like 3
percent per year over the next 10 years. There is quite a
debate going on right now as to whether or not the increase in
productivity that we have had since the mid-1990's is just a
one-time event or is sustainable in the long term. And the
econometric models that we have developed and in some instances
had published in the journals clearly indicate that this is a
long-term development. The BEA's innovations in compiling
better GDP data were instrumental in deriving these estimates
that indicate the trend in productivity is sustainable.
One issue that has come up recently and certainly in this
hearing is the difficulty involved in measuring intellectual
capital. So I would like to suggest one possible approach to
this. This is more a suggestion than anything else. I think it
is going to need to be debated. Right now there is very unusual
discrepancy in the national income accounts. Research and
development spending is counted in GDP if it is done by the
government, it falls under government purchases, but if it is
done by private industry, R&D is counted as an intermediate
input and netted out.
In my view, R&D can be taken as one measure of the
increasing intellectual capital that is becoming increasingly
important in the economy. In fact, if you add R&D spending into
GDP and you also put R&D in as a production function you can
explain an additional 0.6 percentage points per year of
productivity growth. And of course that is quite an important
issue from our point of view because productivity or output per
hour we know has to come from physical capital and technology
but the technology, component is poorly measured.
So one thing that BEA should probably consider doing is
redefining output to include R&D under business-fixed
investment.
Finally, I would like to conclude with one comment about
the recent debate on income versus product and how serious the
current economic slowdown is. The problem--the discrepancy
between income and product during the mid-1990's was really
resolved in favor of higher output. We saw income rising faster
than product and it turned out that we were growing much more
rapidly than we expected.
Now, however, we have a situation in which the product side
is reporting a pretty serious slowdown, growth is 1 percent in
the most recent quarters, it will probably come in about 1
percent when BEA releases it, and yet national income has been
rising by more than $70 billion faster than national product
for the last 2 years. So we are seeing again some indication
that there may be hidden productivity out there, that there
may, in fact, be higher out there.
We don't know the source of the output. But there is
clearly an indication that the American economy has a good deal
of resiliency. There may be additional technical advance,
additional productivity that isn't being measured in the
product-side but is showing up on the income-side. That in turn
suggests that once we are out of the current slowdown, we see a
recovery in demand, that we can actually sustain the current
expansion for a long period of time.
Thank you, Mr. Chairman. I will be happy to answer any
questions.
Mr. Miller. Thank you very much.
[The prepared statement of Mr. Richards follows:]
[GRAPHIC] [TIFF OMITTED] T5327.038
[GRAPHIC] [TIFF OMITTED] T5327.039
[GRAPHIC] [TIFF OMITTED] T5327.040
[GRAPHIC] [TIFF OMITTED] T5327.041
[GRAPHIC] [TIFF OMITTED] T5327.042
[GRAPHIC] [TIFF OMITTED] T5327.043
[GRAPHIC] [TIFF OMITTED] T5327.044
[GRAPHIC] [TIFF OMITTED] T5327.045
[GRAPHIC] [TIFF OMITTED] T5327.046
[GRAPHIC] [TIFF OMITTED] T5327.047
Mr. Miller. And finally, Professor Berndt.
Mr. Berndt. Thank you. I thank the Chair for inviting me to
appear today. Although I currently serve as chair of the
Federal Economic Statistics Advisory Committee, called FESAC
for short, I have not had the opportunity to share these
remarks with them; and so my remarks today should be
interpreted as my own and not necessarily those of my fellow
FESAC members.
As we all know, the last few decades have been marked by
dramatic technological and economic changes. To make important
decisions wisely within such a speedily changing environment,
businesses, government policymakers, employees, retirees,
students, homemakers, and even academic researchers rely very
critically on data and information provided by our statistical
agencies. A major challenge facing these agencies, as a number
of speakers have already emphasized, is to track this moving
target of current economic activity reliably, efficiently, and
promptly.
Let me begin with FESAC and the role FESAC plays in this.
FESAC is an interagency advisory committee to three economic
statistical agencies: BLS, the BEA, and Census. FESAC's mandate
is to analyze issues involved in collecting, tabulating, and
publishing Federal economic statistics, but particularly those
issues that cut across the three statistical agencies and that
could benefit from enhanced interagency cooperation and
coordination.
A goal of FESAC, therefore, is to foster greater efficiency
within the Federal statistical system and thereby enable it to
provide higher quality statistics in support of more informed
economic and social decisionmaking.
Let me now turn to the BEA which is the focus of today's
hearing. Although probably best known for publishing our
Nation's GDP data, the BEA is, in fact, a key provider of a
wide variety of national, industry, regional, and international
economic data on income, production, prices and international
trade. In carrying out its mission, as a number of speakers
have emphasized, the BEA relies on data from the Census and the
BLS and, in turn, provides the BLS with data it needs in
fulfilling its own responsibilities.
In my brief remarks today, I would like to discuss with you
several important issues facing the BEA. But I want to focus on
issues that involve not just the BEA but also the Census and
the BLS. Since my time is short, to illustrate the points I
want to make, I want to focus on a measurement of but one
important and widely observed economic indicator, labor
productivity. And being an academic, I naturally had to put
something on a blackboard exhibit.
As can be seen in this exhibit, labor productivity is a
simple ratio. In the numerator, we have inflation adjusted, or
a real measure of output; and in the denominator on the bottom
we have some measure of hours of labor input. BEA publishes the
numerator and BLS publishes the denominator. And BLS computes
the ratio and publishes the ratio as well. So you can think of
it as BEA over BLS. Let's look at the numerator and denominator
a little more carefully.
First on the numerator, in producing its measure of real
output, the BEA relies on Census to provide output figures in
current dollars. In turn, Census collects sales data from a
representative sample of establishments which it identifies
utilizing a comprehensive register listing of establishments
that serves as a sampling frame for all of the Census Bureau's
business surveys. As an aside, what an establishment is, in a
digital economy with increasing e-commerce, presents ever more
complex issues. But we leave that for another day.
To convert the Census sales figures in nominal dollars into
real inflation-adjusted output data, which is what we need in
that numerator, the BEA deflates them using a combination of
price indexes provided by the BLS and in some cases those that
it has constructed on its own. I might add that BEA was a
pioneer in developing deflators for computers in collaboration
with private-sector firms such as IBM, and more recently for
software, in collaboration with a variety of academic and
private-sector vendors.
So in summary and referring still to the numerator, how one
constructs reliable deflators and thereby measures real output
for diverse industries such as banking, consulting, tax
preparation, investment advice, and health care raises very
challenging issues for all three agencies. FESAC is focusing
considerable attention on such output measurement challenges.
Let's briefly turn to the bottom to the denominator of
labor productivity, the measures of hours worked by employees
and by the self-employed. Like the Census, BLS has a list of
establishments from which it selects those asked to provide
essential economic data. Unfortunately, the universe list of
establishments at the BLS and at the Census do not match
precisely; and currently, data sharing is not permitted. More
on that in a minute. Although BLS measures of hours worked by
production and nonsupervisory workers are likely to be very
reliable, those types of production workers are now a minority.
A very distinct minority in our changing economy.
Hours worked by others such as entrepreneurs and Internet
startups, by telecommuting consultants, by sales reps and
office workers using cell phones while driving to and from work
and utilizing fax and modems at home are very difficult to
measure reliably. Currently the BEA and BLS are both expending
considerable efforts on creating better measures of hours
worked and on how individuals allocate their time. These topics
will be addressed in detail at our next FESAC meeting. A
related set of issues on how one measures, and values, labor
compensation when you have stock options, other deferred
compensation and important non-wage benefits such as health
insurance, are also of great concern to all three agencies and
to FESAC.
This simple example of this ratio of output over labor
input illustrates, I think, some of the complexity involved in
putting together the Nation's economic statistics. Clearly,
constructing and publishing a measure such as labor
productivity involves a great deal of coordination across our
Federal statistical agencies. By and large Mr. Chairman, I
believe this coordination works quite well. Each of the three
principal economic statistical agencies has a reasonably well-
defined set of responsibilities. And each is committed to
working collaboratively with the others to address issues of
mutual interest such as those I have identified above. At the
same time, I believe current arrangements do seem occasionally
to involve some needless duplication and burden on the public.
So let me conclude with an unabashed and blatant plea to
this subcommittee. Current U.S. laws restrict agencies' ability
to share information with one another even for only statistical
purposes. These data-sharing restrictions and especially the
inability of the agencies to share their business register
lists with each other are very costly to our economy. Both
Census and the BLS have universe lists of establishments, but
these do not always agree, particularly in the context of a
very rapidly changing economy when even the notion of what is
an establishment can be called into question.
BEA relies on both Census and BLS establishment data and
must make refereeing choices when these data do not appear to
agree with each other. I believe the sharing of universe lists
and other data among appropriate Federal statistical agencies
would not only achieve budget savings, greater efficiency, and
increased accuracy, but that this would also reduce the
reporting burden on the public and in small business in
particular. Moreover the data sharing could be carried out in
ways that protected the important confidentiality interests of
those providing information.
I strongly urge this subcommittee to support passage of
legislation enabling the appropriate sharing of information
among statistical agencies for statistical purposes. A good
basis for such legislation would be the Statistical Efficiency
Act of 1999 which was passed by the House in the last Congress
as H.R. 2885 but was not considered by the Senate. Passage of
such legislation would be an important good government victory
in my view. Thank you.
[The prepared statement of Mr. Berndt follows:]
[GRAPHIC] [TIFF OMITTED] T5327.048
[GRAPHIC] [TIFF OMITTED] T5327.049
[GRAPHIC] [TIFF OMITTED] T5327.050
[GRAPHIC] [TIFF OMITTED] T5327.051
[GRAPHIC] [TIFF OMITTED] T5327.052
[GRAPHIC] [TIFF OMITTED] T5327.053
[GRAPHIC] [TIFF OMITTED] T5327.054
[GRAPHIC] [TIFF OMITTED] T5327.055
[GRAPHIC] [TIFF OMITTED] T5327.056
Mr. Miller. I thank all of you for being here today and
especially those who came from out of town. I appreciate it. I
found all your statements very interesting.
Let me start off, it has been such a dramatic change,
historic change in our economy during the past decade. How do
you rate the quality of the data you are getting today, many of
you all have been doing this for a few years I know, to 10
years ago or even 20 years ago? Especially as you know this
economy is expecting--going through this technological
revolution, whatever you want to call it, do you feel the data
is as good today or is it better today than 10 years ago or 20
years ago?
Mr. Berner. Mr. Chairman, why don't I start. I think that
the data today really suffers from fact that, as I mentioned in
my prepared remarks, we do have a big hole in the data; 20
percent are based on estimates, and unfortunately that 20
percent often comes from the area that is most dynamic and most
rapidly changing. We talked about software a little bit and
talked a little bit about the surveys that not only my firm but
others do in the private arena to try to get a better
understanding of what is happening in that area.
Just to mention it, 43 cents of every IT spending-dollar is
now, according to Steve Landefeld's statistics, accounted for
by software and it has been growing like a weed. So it is a
very important component of capital spending. And it is a very
important innovation to include that in our data.
Having better data on software outlays, particularly in the
wake of what we have seen with the preparation for Y2K and its
aftermath and other areas would be very important. So that the
challenge is not that the quality of the data have
deteriorated, the challenge comes from the things that we have
all talked about, namely that the economy is changing far more
rapidly today and requires a much more flexible statistical
infrastructure in order to deal with it.
Mr. Miller. You mentioned 20 percent is estimated. Is that
what it was 10, 20 years ago and has that 20 percent changed?
Mr. Berner. Well Steve Landefeld can talk about that, but
let me answer the second question. The composition of the 20
percent has really changed. But there is an important area
besides software, obviously that is critical, and that is in
the service arena. And BEA and the other two major statistical
agencies and the Federal Reserve have made major efforts to
expand their coverage of services and to develop new concepts
and new metrics for gauging what is going on in services and to
try to improve the measurement of productivity in that arena.
But it is a constant challenge because services are broad,
diverse, and certainly cannot be lumped into any one category;
and that diversity obviously has to be dealt with in coming up
with these, both concepts and metrics, in measuring this part
of our economy.
Ms. Swonk. I would like to add to that. I certainly echo
the issue that services are one of the areas where we are not
measuring things as we could. And it has always been a problem.
I was talking to you before about the size of my economics
department and its shrinkage and how much we still produce
relative to its prior size. So obviously we have had some major
productivity gains within our department and certainly within
our bank we now handle more assets with fewer people than ever
before and do it effectively, I hope, depending on which day
you look at my stock price.
With that said, one of the things that is my concern is as
much as there were faults in a lot of regional data, and my hat
originally before being just the chief economist at Bank One
was being the regional economist at Bank One, is the gaps that
are left because of priority choices that had to be made. And
there is a lot of data that is not being collected now. And for
all of its faults it was all we had. It was not perfect by any
means.
And I know that priority decisions had to be made given the
budget cuts. But to not be able to as a large regional firm
that crosses many regions in this country, to not be able to
assess the characteristics of the consumer or business climate
in an accuracy level that you feel confident with that we are
now turning to our own information which is, frankly, faster
and more real-time information than I can get from the
government, that is a real problem.
And it also means that I can't share all of my analysis as
I want to with some of the policymakers that I talk to because
much of it is private, our own private analysis of our own
economic information inside the bank. And that is just not the
direction we want these things to go. It has really left many
people at the regional level scrambling for ways to figure out
what their revenues are going to be, what you know retail sales
revenues are going to be.
Many have tried to make up different kinds of measurements,
many of the regional Feds have tried to make up different
measurements of retail sales. I am using that as one example.
But clearly we have lost some things in the mix. I won't even
begin to go into the mortgage data and how important that has
become. Here we are in the mist of another mortgage refinancing
boom and our group has done significant work on mortgage
refinancing and its contribution to the U.S. economy which is
not included in income but, boy, it is spent. The mortgage data
is very compromised at this point in time because of priority
decisions that had to be made earlier on.
Mr. Berner. Mr. Chairman, if I could, let me just mention
one other area and that is the international arena. Nick
Knickerbocker mentioned as an aside that if we knew how to
measure our exports to within the tolerance of 7 percent we
would be much better off.
You can imagine what the discrepancies are in the service
areas of our international accounts which are perhaps even more
compelling at this point in time. And that is because not only
does our economy have a more global look to it but obviously
the huge wave of foreign investment in the United States in the
last several years has made the sharing of data and the sharing
of information about the income exchanges from that direct
investment much more important.
It now appears, for example, that the European economies
are slowing down much more rapidly than most people had
anticipated. One reason for that may be that European
corporations are responding to the slowdown and the results
that they are seeing in the United States and that is having an
impact on their business.
If we had better data on foreign direct investment on the
reported income flows associated with those, and the BEA and
other agencies make every effort to improve those data, then we
would be able to analyze that better. And I think that points
to one other issue which is data sharing perhaps at least
cooperation across borders. And as I am sure you know, both
Steve and Nick and Cathy Abraham at BLS are making every effort
to do that and to cooperate. But obviously more resources would
permit them to cooperate more effectively with their
counterparts overseas.
Ms. Swonk. I wanted to add one extra point too and that is
one of the things that we have seen is because of gaps left in
the data this reliance on more private sector unreliable data.
And I am rather stunned. I have been on many a talking head
show where an economist or economic analyst, whatever they may
title themselves, tries to analyze data that they don't
understand.
Not only do they not understand it because they haven't
researched it, and haven't been taught it because we have lost
much of that, but also it is private-sector data that doesn't
have the same accountability. If you really want to understand
the flaws and the gaps in the U.S. statistical agency's data,
you can understand that. They provide that for you. They tell
you. They are accountable. So you could say this could be a
seasonal factor. This could be because it snowed last month.
They tell you that information.
Where on the private-sector data that is coming to dominate
some of the financial market moves, there is no accountability
whatsoever. I really fear that some of the gaps that are left
are being filled by the private sector. As much as I believe in
the private sector, this is just not one place they belong.
They don't have the same incentives. They can discontinue data
series if they go out of business. There are all kinds of areas
where there are some real severe problems.
Mr. Miller. What competition is there for BEA? I mean you
mentioned the private sector. Is there potential for someone to
offer competing data?
Ms. Swonk. I don't think there is any way that a private
sector could get the kind of confidential information that a
U.S. Government agency could get to provide overall economic
data. But I am stunned in the last decade to see how many
reports come by or people trying to sell me their information
of their particular survey on the world and what the
information--trying to tell me what that information provides.
I look at it and realize it doesn't provide what they are
telling me it provides. So I don't think there is any real
competition in the sense that I don't think any private-sector
firm would be trusted with the kind of, you know, intimate data
that corporations provide and small businesses provide to the
U.S. Government under confidentiality agreements.
However, it is amazing how much is even worse in terms of
the private-sector data that is coming out, how much is being
pedaled out there in terms of more economic information to try
to fill in this picture of the economy that is increasingly
finding gaps in it.
Mr. Richards. I would like to pick up on something that
Diane said about the unreliability about private-sector data. A
lot of the data that is being held up as competitive with BEA's
data is not very reliable.
Here is an example: In November and December, consumer
confidence as measured by some surveys dropped by about 17
percent which seemed to imply that consumer spending was poised
for a major slowdown but what we actually observed in January
was a significant rebound in consumer spending. So that not
only did the consumer confidence data give a false reading it
could not even call the direction of change correctly.
Nonetheless, it is clear that the stock markets were reacting
to the consumer confidence numbers.
There are many other examples of private sector surveys
that are poorly put together and contain false and misleading
information, but unfortunately that false and misleading
information is moving the markets in a significant way.
Mr. Berndt. Let me just add to that, if I may disagree
slightly with some of my colleagues. I think there are some
industries that have much deeper coverage from private-sector
sources than the government, because of the govdernment's
sampling procedures. Let me take an area that I know
particularly well, which is health care. And there are a number
of--for example pharmaceutical industry data sources which have
samples of products that are in the hundreds of thousands each
month whereas the BLS's sampling procedures can only be about
500 products a month. So it varies, I think. But certainly
there is nothing in the private sector that can rival the
comprehensiveness that the accounts from the BEA and BLS and
Census Bureau provide.
Mr. Miller. Don't individual States--a lot of times the
State universities--I remember when I was back in graduate
school, they would have their own departments generating that
type of information. For those individual States, talking about
Florida and Louisiana, two of the States where I went to
school. But it seems like they still crank out the data. How
reliable is that State-type of data?
Ms. Swonk. It is interesting because, on a regional basis,
I rely more and more on those kinds of departments to get a
feel for--Florida is a big State for us, for Bank One. I rely
more and more on that information and what the Federal Reserve
puts out to get a feel for economic information.
The problem is even there much of that State, the business
departments or the business research groups, they base their
information off of employment data coming out by the State or
by the Federal Government and I have seen gaps in their data
sources as well. So they are now having to make assumptions on
top of assumptions to get to those conclusions.
And again there is no consistency across States. You are
getting to issues--I mean I want to compare data that is in
Florida produced for Florida that compares to data in Michigan
prepared for Michigan. And when you get to the individual
research institutions, although they are extremely valuable and
I rely on them very heavily when I do regional analysis, there
is not the--they are not always comparable in terms of what it
is they are analyzing, what their purposes are. Some of them
have more purposes to advise State government, some of them
have purposes to attract more investment to the State. So the
inconsistencies there just again make the problem more complex
in terms of what the information is actually telling us.
Mr. Miller. You mentioned about the BLS and BEA and census
and you talked about the funding. You know this is an
authorizing committee not an appropriation committee. I happen
to sit on both. Actually I sit on both appropriation committees
that fund BLS and Census Bureau and BEA. It is hard always to
explain how the government operates in a way because I sit on
the Labor HHS subcommittee which is where BLS is funded. But I
happen to sit on the Commerce, Justice, State, and the
Judiciary appropriation subcommittee which is where the Census
Bureau, BEA is.
And you mention about--there has always been the question
of consolidation of statistical agencies. We are not here to
discuss that, debate that issue specifically. But there is--
when you have different appropriations subcommittees, you have
different authorizing committees, and yet there is competition
between agencies collecting data. I think you--Mr. Berndt, you
mentioned the problems of not sharing the data. But yet, there
is somehow the advantage of having competing sources of data
are there? And what would you recommend? Do you think--that
is--I am interested. I was not fully aware that there is an
advisory committee that represents that cross of all of the
agencies between departments and how that operates too.
Mr. Berndt. Let me start to answer that. But you open up a
wide topic on which we could have hearings for some time. There
are historical reasons why we have the different agencies. I
think in general I agree with you that having some competition
among agencies is, in some sense, a good thing. I think,
however, in quite a few instances, there really is actual
duplication and replication. I think we could proceed quite
wisely and prudently by defining, identifying some of those
areas and without getting into a big argument of whether we
want to have a statistics United States like Statistics Canada,
but rather are there opportunities where we can efficiently
share data and avoid duplication and use our public-sector
dollars more prudently. That would be important first steps to
take. There are those opportunities now, particularly as we
have the information technology revolution where we have common
standards of collecting and reporting data, and it makes it
much easier now to do that. So I want to shy away from your big
question.
Mr. Miller. I really don't even want to bring that one up
either, I guess. But----
Mr. Berndt. I would like to suggest I think there are
enormous numbers of small steps that together could improve our
interagency coordination and make our public-sector dollars for
data collection spent more wisely.
Mr. Miller. One of the concerns has been about
confidentiality of data, whether it is just basic census data
or financial information. We are in an age where with the
technology revolution going on that access to data but then
confidentiality of it and being able to--what impact that has
on participation and supplying data. What is the challenge
there about the--I mean, one of things--I am a former
businessman. I remember getting forms in the mail. The
University of Florida would send me something or the State. And
you know I was a relatively small business back home; we didn't
have an economics department certainly.
Ms. Swonk. We hardly do too.
Mr. Miller. So how do you complete that data? Of course
when I was in the business we--technology has made it a little
easier to generate that data. But this whole issue of
confidentiality and willingness to participate on small
business is a real challenge, I think. How do you overcome all
that?
Ms. Swonk. You know I agree 100 percent with that. My
husband is actually a small business owner and just completed
one of the forms that he had to complete for a survey actually.
I asked him if he did complete it himself because he is the CEO
of his small firm, and he said he did. I said good for you
because often it is passed down to someone's secretary, and
that is where a lot of the problems are. He didn't find the
questions that intrusive. He thought there could have been more
questions. Of course, he has got a little bias in his
background given his marriage to me.
But I think one of the things he did say, he said why isn't
this automated. I could have just e-mailed it back. Why
couldn't I have done this? Or why couldn't I have done that? We
do have small business surveys out there like the National
Federation of Independent Business, another one of our former
presidents, Bill Dunkelberg, heads up that survey. And small
businesses are very willing to share information when they--
when it is very narrowly defined and also when they see a
benefit that it could help them. And I think again, making
this, data, the least interesting word in English language and
making more people aware of how important that is to policy
would help.
Education is one of the key issues here in terms of the
small business sector. And there are many organizations that
represent small business that can be friends to the statistical
agencies to try to then help them, I think, in that arena. You
are right, the ease with which these forms can be filled out
even in a large corporation, I am appalled at some times some
of the stuff that comes in. We were asked to be part of
something that the Fed was encouraging our organization to
become a part of and they called and asked me should we do
this. I said, are you kidding? Of course, we should do this.
And then they were going to try to put it on a low level
person. I said no it has to be by someone at a high enough
level that knows the information. These are always challenges
and the more that we can make these automated and easier and
simpler and blind, more of a feeling of blindness in terms of
aggregating the data back to the government I think the more
participation you will have.
Mr. Miller. Let me go back to my first question as we
conclude here. That is the quality of the data you get today
and the ability to do forecasting versus 10 years ago. Have we
improved?
Mr. Richards. Mr. Chairman, I think that we have improved
in very significant ways. Ten years ago we did not have chain-
weighting in the national income accounts. Ten years ago we
were not including software. Ten years ago we did not have
quality imputations for computers. So it is a question of is
the glass half full or half empty. I think it is half full. But
there are still some improvements that we have to make. It is
not so much the manufacturing sector which I represent which is
covered very well, it is in the service sector where, in many
cases, industries like banking finance and real estate there is
no direct measure of output. So the BEA has no choice except to
develop some kind of imputation. And that is very difficult to
do.
And I think you know there clearly is room for quality
improvement. I have been critical of the private-sector data
but some of data that is collected directly by the private
sector such as transactions conducted at ATM machines which, of
course, are recorded by banks could be given to government
agencies which could then develop better measures of what the
service sector is doing. I think the data has improved
significantly, but there is room for further improvement.
Ms. Swonk. I would echo that. We are chasing a moving
target. So no matter how much you improve the data you have to
improve it more to catch this moving target. The clear issue
that I leave is with catching that moving target some pieces of
data have been left behind.
Mr. Miller. One question for Professor Berndt about the
cooperation between the three agencies that you work with. How
often does your advisory committee meet?
Mr. Berndt. Our advisory committee was formed last year. We
meet twice a year.
Mr. Miller. So it is fairly new, then, the creation of it.
Have you seen any improvement because of the short time you
have been existence? Or has there been a problem? Is that the
reason it was created between the agencies?
Mr. Berndt. Each of the agencies had their own advisory
committees in the past. I believe this was recognized: there
were significant opportunities for coordinating better, and it
was under that sort of a rationale that this particular
committee was created.
Mr. Miller. OMB do you--is OMB involved in this loop? She
is nodding yes. Were they the impetus that created this?
Mr. Berndt. They were part of the impetus, yes. But it was
the agencies themselves that also recognized that it is time to
do this.
Mr. Miller. What is the objective of this?
Mr. Berndt. The objective, I think, is to find some issues
on which all three agencies need better data and can work
together on putting together survey forms that match their
common needs better, that reduces their reporting burden on the
public, that reduces the duplication. And that what some of the
folks here have talked about utilize some of the state-of-the-
art thinking in how do you measure some of these difficult
concepts, like how do you measure output in our health care
sector where we have improved outcomes and extended life spans.
So it is issues like that that cut across the various agencies
that this subcommittee or this advisory committee is trying to
address. We will be happy to report back to you in the future.
Mr. Berner. If I could, I think one of the things we are
learning here is not only do we endorse data sharing among our
statistical agencies but perhaps we should have data sharing
among the panels who advise and oversee them in their work. So,
you know, Professor Berndt and I will probably get together
after this meeting and talk about ways that we can cooperate
because we have a statistics committee at our organization that
obviously has provided advice in the agencies in the past and
will continue to do so in the future. And to the extent that we
overlap, we can make a much more efficient set of
recommendations to the agencies.
Mr. Miller. All right. Let me once again thank you all for
participating here today. I find it very informative and
enlightening to have this. As I mentioned earlier, I am
delighted that the administration's budget proposal--I assume
that is coming from the work of Kathy Wallman over there--
allowed for the increase that was--you know, shows the
attention and interest and now the commitment of government to
that. This information is very valuable for the future of our
country.
So I thank you very much for your contribution and your
support for it and the information provided here today. So on
behalf of the subcommittee, I say thank you for appearing here
today.
I ask unanimous consent that all Members and witnesses that
have written opening statements be included in the record. And
without objection so ordered.
[The prepared statement of Hon. Wm. Lacy Clay follows:]
[GRAPHIC] [TIFF OMITTED] T5327.057
[GRAPHIC] [TIFF OMITTED] T5327.058
Mr. Miller. In case there are additional questions Members
may have for our witnesses, I ask unanimous consent that the
record remain open for 2 weeks for Members to submit questions
for the record and that witnesses submit written answers as
soon as practicable. Without objection so ordered.
Thank you all very much for being here today. We stand
adjourned.
[Whereupon, at 3:59 p.m., the subcommittee was adjourned.]