Manufacturing Trends: Long-Term Context for Today’s Policy Issues
Congressional research reportSep 9, 2005
Ask Donna
What actually matters in this document.
Text
Order Code RL32179
CRS Report for Congress
Received through the CRS Web
Manufacturing Trends:
Long-Term Context for
Today’s Policy Issues
Updated September 9, 2005
Stephen Cooney (Coordinator)
Industry Specialist
Resources, Science, and Industry Division
Bernard A. Gelb
Industry Specialist
Resources, Science, and Industry Division
Robert Pirog
Specialist in Energy Economics and Policy
Resources, Science, and Industry Division
Congressional Research Service ˜ The Library of Congress
Manufacturing Trends: Long-Term Context for Today’s
Policy Issues
Summary
Although the U.S. economy has emerged from a relatively shallow recession in
2001, by mid-2005 there was still no recovery in the number of manufacturing jobs.
Nevertheless, real output in U.S. manufacturing in early 2005 stood more than 30%
higher than at a similar point after the recession of 1991, even with 2.5 million fewer
manufacturing employees. Members of Congress are increasingly concerned as to
how these developments affect manufacturing employment in their states and
districts. This report revises a CRS report released in early 2004 to provide a
baseline study for Members of Congress on long-term trends in manufacturing
output, productivity and employment, including some extensions of time-series data.
Employment in manufacturing has declined as a share of overall employment
since 1960 at about the same rate of decline as the current-dollar share of GDP
accounted for by manufacturing output. Both measures have fallen from 30% to less
than 15%. Only about 14 million people were employed in manufacturing by mid2005. The all-time peak was more than 19 million in 1979, and each successive
cyclical economic peak since then has seen fewer persons so employed. But
measured on a real basis, manufacturing output kept pace with total output in other
sectors, despite its shrinking share of employment. The explanation seems to lie in
rising labor productivity, which grew 50% more quickly for manufacturing than for
the total economy between 1960 and 2000.
These overall trends in manufacturing mask highly divergent performances
among individual industries. Output and productivity among individual sectors do
not tend to cluster around overall average levels of performance. This report
examines in more detail three specific manufacturing sectors: information technology
industries, which have been high-growth areas of the economy and internationally
competitive; the automotive sector, which has been affected by high levels of import
penetration and is divided between the “Big Three” U.S.-based manufacturers and
the U.S. facilities of internationally based firms; and, textiles and apparel, which are
facing a high level of import competition and have experienced large numbers of job
losses.
Globalization, meaning the increased internationalization of markets, inputs and
investment, has had a major impact on U.S. manufacturing. Since 1980, the U.S.
trade balance in manufactured goods has gone from a surplus to a deficit of nearly
$500 billion. The deficit is concentrated in consumer and automotive products, but
even capital goods moved into deficit in 2003. While foreign outsourcing (also
known as offshore outsourcing or offshoring) has become a major issue, the report
notes evidence that nearly all major industrial countries, including China, have lost
manufacturing jobs since 1995. Changes in the dollar exchange rate and U.S.
international trade agreements may also have affected domestic manufacturing.
The report concludes by examining various approaches to industrial and
industrial competitiveness policies. This CRS report will not be updated.
Contents
Overview of Manufacturing in the U.S. Economy . . . . . . . . . . . . . . . . . . . . . . . . 1
Employment Trends in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Manufacturing Output and the U.S. Economy . . . . . . . . . . . . . . . . . . . . . . . . 4
Productivity and Manufacturing Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Performance Varies Widely by Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Output Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1960 to 1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1990 to 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Productivity Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1960 to 1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1990 to 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Employment Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1960 to 1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1990 to 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Performance of Selected Major Industrial Sectors . . . . . . . . . . . . . . . . . . . . . . . . 17
Information Technology Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Decline in Information Technology Industries . . . . . . . . . . . . . . . . . . 18
Semiconductors Bolster U.S. International IT Competitiveness . . . . . 20
IT Manufacturing Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Automobiles and Light Trucks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Production Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Employment Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Textiles and Apparel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
The Economics of Textile and Apparel Production . . . . . . . . . . . . . . . 30
U.S. Textile and Apparel Production, Trade, and Employment . . . . . 31
Textile Trade Policy and Agreements . . . . . . . . . . . . . . . . . . . . . . . . . 32
Globalization: Impact on U.S. Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
The Manufactures Trade Balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Manufactured Imports and Exports . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Foreign Outsourcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
The Dollar Exchange Rate and U.S. Manufacturing . . . . . . . . . . . . . . . . . . 40
Market-Based Exchange Rate Systems . . . . . . . . . . . . . . . . . . . . . . . . 40
The Dollar, the Trade Balance and Employment in Manufacturing . . 41
Asian Economies and the Dollar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
The Impact of U.S. Trade Agreements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Effects of NAFTA on US. Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
China’s Accession to the WTO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Industrial Policy and Industrial Competitiveness Policy . . . . . . . . . . . . . . . . . . . 56
Planning in Market Economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Industrial Competitiveness Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Conclusion and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Appendix: Change in Industrial Classification System . . . . . . . . . . . . . . . . . . . 65
List of Figures
Figure 1. Manufacturing Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Figure 2. Manufacturing Share of U.S. GDP and Employment . . . . . . . . . . . . . . . 5
Figure 3. Real GDP and Manufacturing Output Growth . . . . . . . . . . . . . . . . . . . . 6
Figure 4. Manufacturing Employment, Output and Labor Productivity . . . . . . . . 8
Figure 5. U.S. Information Technology Market Change . . . . . . . . . . . . . . . . . . . 19
Figure 6. Total U.S. Automobile and Light Truck Production . . . . . . . . . . . . . . 26
Figure 7. Employment in Motor Vehicle and Equipment Industry . . . . . . . . . . . 28
Figure 8. U.S. Manufactured Exports and Imports . . . . . . . . . . . . . . . . . . . . . . . 35
Figure 9. U.S. Trade Balances by “End-Use” Sectors . . . . . . . . . . . . . . . . . . . . . 37
Figure 10. Dollar Index: Price-Adjusted Major Currencies . . . . . . . . . . . . . . . . . 42
List of Tables
Table 1. Trends in Output, Labor Productivity, and Employment
in Manufacturing Industries: 1960-1990, 1990-2000 . . . . . . . . . . . . . . . . . 14
Table 2. U.S. Trade in Information Technology Products . . . . . . . . . . . . . . . . . 22
Table 3. Distribution of U.S. Motor Vehicle Production . . . . . . . . . . . . . . . . . . . 27
Table 4. Output, Productivity, and Employment
in U.S. Textile and Apparel Manufacturing Industries . . . . . . . . . . . . . . . . 31
Table 5. Details of U.S. Automotive Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Manufacturing Trends: Long-Term Context
for Today’s Policy Issues
Overview of Manufacturing in the U.S. Economy1
Members of Congress have become increasingly concerned about a perceived
decline in U.S. manufacturing, particularly as it may affect employment levels in
their states or districts. For two years after the economy emerged from a relatively
shallow recession in 2001, there was a continuing loss in manufacturing jobs. And
even as the recovery progresses, many fear that new non-manufacturing jobs will not
have the same relatively high levels of wages and benefits traditionally associated
with manufacturing. By contrast, despite the recession of 2001, the level of real
(inflation-adjusted) U.S. manufacturing output as of mid-2005 stood almost 30%
higher than at a similar point after the recession of 1991, with about 2.5 million fewer
manufacturing employees. This is more than three times real manufacturing output
in 1960, when a comparable number of people were employed in manufacturing.2
Employment Trends in Manufacturing
Figure 1 shows that, as compiled by the Department of Labor’s Bureau of Labor
Statistics (BLS), the number of manufacturing employees in the United States in
2002 was about the same as it was in 1960. The total was about 15 million, which
is also about the same as in the early 1950s, during the Korean War boom. But, of
course, the U.S. economy and population are now much larger, so that a far smaller
share of the workforce is directly employed in manufacturing. The level of
manufacturing employment has tended to move up and down with the business cycle,
but with successively lower “peaks”after the three recessionary periods of the 1980s
and 1990s (“recessionary years,” defined here as real growth of 1.0% or less in gross
domestic product, are marked in the figure). Thus, the secular trend in manufacturing
employment has been down since 1979, when 19.4 million persons were so
employed in the United States. The number fell by 2.4 million over the next three
years with the impact of the “double-dip” recession in 1980-82, and recovered to its
next cyclical peak of 18 million jobs in 1989. Following the recession of 1990-91,
the next peak employment year was 17.7 million in 1998, after which manufacturing
employment began to fall, well in advance of the 2001 recession.
The August 2003 manufacturing employment level of 14.6 million persons was
3 million lower than the 1998 peak. It was also 1.25 million lower than when the
1
2
This section was written by Stephen Cooney.
U.S. Dept. of Labor. Bureau of Labor Statistics (hereafter BLS). “National Employment,
Hours and Earnings” and “Output: Manufacturing” series (data may be subject to revision).
CRS-2
latest recession ended in November 2001. Thus, manufacturing accounted for all of
the net total nonfarm job loss of 1.1 million in 2001-03.
Figure 1. Manufacturing Employment
20
15
Manufacturing employment (millions)
10
1960
1965
1970
1975
1980
1985
1990
1995
2000
Years with dashed lines indicate recessions or real growth less than 1%.
Source: U.S. Dept. of Labor. Bureau of Labor Statistics (BLS). “National Employment Hours
and Earnings” data from BLS website (www.bls.gov – proceed from directions on the site).
Definition of manufacturing based on North American Industrial Classification System (NAICS).
Both the fall in manufacturing employment before the 2001 recession, and the
slowness of the recovery in both manufacturing and total employment after a
recovery began in late 2001, also suggest that we may be experiencing structural
changes in the economy. Possibly, long-term structural changes to increase output
per hour of labor are overwhelming the normal employment trends related to the
business cycle.3 Research at the New York Federal Reserve Bank found that even
the “jobless recovery” after 1991 began to create jobs by 15 months after it began.
By contrast, the U.S. economy was still shedding jobs in mid-2003, almost two years
after economists decided that the recovery had begun. Indeed, the unprecedented
slowness of the employment recovery in this case caused a long delay in the decision
of National Bureau of Economic Research, the accepted arbiter for declaring when
recessions begin and end, to determine that November 2001 marked the beginning
of a recovery. The Fed researchers also noted that layoffs are now increasingly
permanent, not temporary, as in the past, when employers wanted to keep
experienced employees around to be rehired when the “slowdown” in orders picked
up again. They noted some industries closely associated with high technology (such
3
This concept is explored in Federal Reserve Bank of New York. Erica L. Groshen and
Simon Potter, “Has Structural Change Contributed to a Jobless Recovery?” Current Issues
in Economics and Finance, IX:8 (August 2003), which provides the data and analysis for
the following comments. For a more general analysis of employment decline and the slow
recovery in employment from the 2001 recession, see CRS Report RL32047, The “Jobless
Recovery” from the 2001 Recession: A Comparison to Earlier Recoveries and Possible
Explanations, by Marc Labonte and Linda Levine; and, CRS Report RL30799, Corporate
Downsizing and Other Mass Layoffs, by Linda Levine.
CRS-3
as electronic equipment) or considered high fliers in the 1990s (communications, and
securities and commodity brokers) are downsizing post-2001 at a faster rate than the
general cyclical trend, suggesting a structural shift and permanent downsizing.4
Employer forecasts and actions in late 2003, as reported in Business Week,
predicted a strong recovery ahead in the U.S. labor market. The article also stated
that the slow and uncertain start of the present recovery discouraged employers from
rushing too quickly to make permanent new hires. It forecast that an employment
recovery could start in late 2003, though it might not be robust until mid-2004, when
the growth cycle might be firmly re-established. On the other hand, even this
optimistic analysis found that “... new jobs may not be coming back soon in the
hardest-hit corners of the economy. Some companies, especially in manufacturing,
are still concentrating on cost-cutting rather than expansion.” The article cited recent
large job cuts announced at paper, apparel and information technology equipment
manufacturing companies. However, the article did also emphasize that productivity
gains would also mean an increase in U.S. jobs in the longer term, not a decrease.5
It may be that, as argued by Federal Reserve Board Chairman Alan Greenspan,
“the U.S. economy has been gradually moving toward an economy based on
‘conceptual’ assets, such as ideas protected as intellectual property, and away from
‘physical’ assets, such as plants and production machinery.”6 Nevertheless, there is
an increasing perception that manufacturing employment and the manufacturing role
in the economy are under pressure, and that this may adversely affect U.S. standards
of living. For example, the Washington Post editorialized about “The Lost Factory
Job,” while the Greensboro News & Record, in the heart of the hard-hit textile
manufacturing belt, headlined an editorial, “Saving Factory Jobs Critical to the
Economy.”7 In a more sweeping fashion, Louis Uchitelle, chief economics
correspondent of the New York Times, wrote, “Manufacturing is slowly disappearing
in the United States ... [T]he essence of a great world power is its edge in producing
not services but manufacturing products that other people want ...”8 The National
Association of Manufacturers (NAM) sponsored a study that warned:
If the U.S. manufacturing base continues to shrink at its present rate and [its]
critical mass is lost, the manufacturing innovation process will shift to other
4
Ibid., esp. Chart 1.
5
Business Week, “Ready to Say ‘Help Wanted’?” (September 22, 2003), pp. 36-37, which
includes quotation.
6
As summarized by Brett Ferguson, “Jobs Lost Since 2001 Recession May Not Return, New
York Fed Study Says,” Daily Report for Executives, Sept. 8, 2003, p. EE-6. The idea of
“conceptualization” of the U.S. economy was first laid out by Chairman Greenspan in
“Market Economies and the Rule of Law,” remarks to a Federal Reserve Bank of Atlanta
conference (Sea Island, GA), April 4, 2003.
7
Washington Post, September 1, 2003; Greensboro News & Record; September 3, 2003.
8
New York Times, August 17, 2003.
CRS-4
global centers. Once that happens, a decline in U.S. living standards in the future
is virtually assured.9
The report, however, emphasized that the U.S. manufacturing sector continues
to make positive and beneficial contributions to the U.S. economy, in part through:
Manufacturing productivity gains ... historically higher than those of any other
economic sector — over the past two decades, manufacturing averaged twice the
annual productivity gains of the rest of the private sector. These gains enabled
Americans to do more with less, increase our ability to compete, and facilitate
higher wages for all employees.10
This ability “to do more with less,” a sign of strength in manufacturing,
paradoxically means that fewer manufacturing employees may be needed, if overall
demand for U.S. manufactured goods, domestically and abroad, is not rising fast
enough. Then-president of NAM Jerry Jasinowski, reflected on this paradox:
[He] is resigned to the fact that many of the factory jobs cut will not reappear.
But he is also proud of the fact that productivity growth in manufacturing has
consistently outstripped that of the rest of the U.S. economy, yielding great
benefits to the nation overall.11
The NAM study further emphasized that the manufacturing sector has a stronger
multiplier effect than any other major sector of the economy. It cited Commerce
Department calculations released at the end of 2002, which showed that each
additional dollar in U.S. final demand for manufactured products required $1.43 in
intermediate goods and services, with associated employment. For natural resources
products, each additional dollar of demand required $1.22 in intermediate goods and
services. For the output of all other sectors, each $1 of final demand required less
than $1.00 of intermediate goods and services. This ranges from transportation, at
about 90¢, to financial and business services, which generate only 50¢ in
intermediate goods and services.12
Manufacturing Output and the U.S. Economy
Does the decline in manufacturing employment indicate a general decline in
manufacturing’s relative role in the economy? Manufacturing’s share of employment
and of current dollar gross domestic product (GDP) have indeed fallen over the long
term. Manufacturing accounted for less than 15% of total employment and GDP in
2002, compared with about 30% in 1950 (Figure 2). This decline in manufacturing’s
9
Joel Popkin & Co. (for NAM Council of Manufacturing Associations). Securing America’s
Future: The Case for a Strong Manufacturing Base (Washington, June 2003), p. 3.
10
Ibid., p. 1.
11
John M. Berry, “Some Lost Jobs May Never Come Back,” Washington Post (Nov. 29,
2003), p. E2.
12
Popkin, Securing America’s Future, pp. 4-10 and Chart 1, based on Commerce
Department 1997 benchmark input-output tables, the latest available data.
CRS-5
share was neither cyclical nor unique to the business cycle subsequent to 2000, as
Figure 2 shows.13
Figure 2. Manufacturing Share of U.S. GDP and Employment
35%
30%
25%
20%
15%
10%
5%
Manufacturing share of GDP
Manufacturing share of total employment
0%
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
Sources: GDP data: Economic Report of the President, Feb. 2003. Table B-12 for 1959-1998; U.S. Dept. of
Commerce. Brueau of Economic Analysis (BEA). Robert E. Yuskavage and Erich H. Strassner, “Gross
Domestic Product by Industry for 2002.” Table 1 for 1999-2002 estimates; and unpublished data from BEA
website “Gross Domestic Product by Industry” (www.bea.doc.gov), as viewed on September 16, 2003.
Workforce data: Same as Figure 1.
This long term relative decline does not necessarily indicate a weakness in
manufacturing, but reflects the sector’s faster gains in efficiency relative to those of
the rest of the economy. This has tended to lower the prices of manufactured goods
relative to the prices of goods and services produced in other sectors,14 “causing”
manufacturing’s share of current-dollar GDP to decrease noticeably.15
13
The U.S. Dept. of Commerce, Bureau of Economic Analysis (BEA) issued revised
estimates of gross domestic product and related national income and product data shortly
after research for this report was completed. In its release, BEA states that “the picture of
the economy shown in the revised estimates is very similar in broad outline to the picture
shown in the previously published estimates.”
14
The average price for manufactured goods as whole rose 62% between 1977 and 2001,
compared with 143% for GDP as a whole, based upon price indexes derived by the BEA
(www.bea.doc.gov/bea/dn2/gpo.htm).
15
GDP in current dollars is a measure of the total value of goods and services produced in
the economy at prevailing prices at a given point in time. The “weight” of a component of
GDP thus is based upon what purchasers actually pay for a product or group of products in
the current period, not what they might have paid using a past period’s prices. BEA says
that “current-dollar shares of GDP provide a more accurate measure of the relative
importance of components” than shares based upon chained dollars. (Chained-dollar
estimates are a variant of constant dollar output.) See J. Steven Landefeld, Brent R.
Moulton, and Cindy M. Vojtech, “Chained-Dollar Indexes: Issues, Tips on Their Use, and
Upcoming Changes,” Survey of Current Business (Nov. 2003), pp.8-16. However, a sector’s
share of GDP in current dollars in a later period may be understated if the prices of its goods
or services have declined relative to those of other sectors, as for many manufactured
products. Calculating a sector’s share using constant dollars eliminates such an effect,
(continued...)
CRS-6
Figure 3. Real GDP and Manufacturing Output Growth
5
4
3
2
1
0
-1
Real GDP
Real Mfg. Output
-2
-3
1960-69
1969-79
1979-90
1990-2000
2000-02
Source: Overall GDP data: Economic Report of the President (2003). Table B11; and, BEA. Yuskavage
and Strassner for 2002 data.
Manufacturing Output: BLS. “Major Sector Production and Costs Index.” (www.bls.gov) as viewed on
September 16, 2003.
If we compare the growth of manufacturing output and total GDP on a real
(constant dollar) basis (Figure 3), we find that manufacturing has performed about
as well as the rest of the economy, and better in the 1990-2000 growth cycle. A
“peak-to-peak” comparison (using the last full year before subsequent recessions) of
manufacturing and GDP in constant dollars (removing the effects of price changes
but not removing quality improvements) shows that manufacturing output grew 3.3%
per year on average between 1960 and 2000, versus 3.4% for GDP.16 Real
manufacturing output grew nearly as fast or faster than real GDP during three of the
four economic expansions during that period, the exception being the 1980s. In the
1990s, real output in manufacturing increased 3.6% per year on average, whereas real
GDP rose 3.2%; the share of real GDP accounted for by manufacturing increased
from 16.4% in 1990 to 17.2% in 2000.17
In 2001-02, manufacturing output in real terms fell at an average annual rate of
3.0%, while the economy as a whole grew by 1.3%, indicating much stronger real
15
(...continued)
although it has some drawbacks. This basis is used in the following paragraph and in figure
3.
16
To compare the rates of growth, Figure 3 uses data for real GDP calculated by BEA and
for constant-dollar manufacturing output, data published and used by BLS in its
measurements of productivity change. BLS data are used for manufacturing real output,
because BEA does not have a series of constant-dollar output by industry for the years
before 1977. For further information, see “GDP by Industry” at [http://www.doc.gov/bea].
17
See Fig. 4 in U.S. Dept. of Commerce, Manufacturing in America: A Comprehensive
Strategy to Address the Challenge to U.S. Manufacturers (Jan. 2004), which shows
essentially the same relationship between real manufacturing output and the total economy
between 1977 and 2000.
CRS-7
growth over the past two years for the non-manufacturing sectors. However, this is
only a two-year period, and not measured on the same “peak to peak” basis as the
other periods in the figure. And, as the Department of Commerce 2004 report noted,
manufacturers are usually hit hardest during a recession: manufacturing output on
average has fallen by 7% against an average 2% fall in total GDP in eight recessions
since 1950. In the “relatively mild” recession of 2001, however, that study found that
the manufacturing sector was not only hit harder than the overall economy, but was
affected by a particularly slow recovery. It ascribes this combination of effects to
four causes:
!
!
!
!
A “significant retrenchment in business technology following a
surge in such investment” in the 1990s. (The impact on U.S. high
technology, especially information technology manufacturing, is
explored in a later section of the present report.)
An especially sharp drop in inventories; the Commerce Department
found that “inventory liquidation was much more severe in the 2001
recession than it was in the 1990-1991 recession.”
“... Uncertainty caused by the events of September 11, 2002, which
depressed investment and demand.”
“... The extent to which slower growth at home was compounded by
the effects of slower growth abroad, particularly the dramatic drop
in U.S. manufacturing exports to our principal export markets.”18
The details and impact of slow manufactured export growth are
discussed in a later section of the present report.
Productivity and Manufacturing Output
Real output in manufacturing kept pace with the real gains of the economy as
a whole despite a continually declining share of total employment, because of
superior labor productivity gains, as measured by the BLS series on output per person
per hour. Over the entire period 1960-2000, the average annual increase in
productivity was 50% greater for manufacturing than for total nonfarm business
activity: 3.0% against 2.0%. Manufacturing labor productivity improvement in the
1970-90 period was also higher than total nonfarm business productivity, and in the
1990s, it increased to double the overall business average.
Presenting these data for the same periods as in Figure 3, there was little
difference between the two measures in the 1960s, as both improved at an annual
average of 2.9%. In the 1970s and 1980s, both rates of labor productivity
improvement slowed, although manufacturing productivity performed better.
Manufacturing productivity increased by an annual average of 2.4% in the 1970s and
2.8% in the 1980s. By contrast, the productivity performance of total nonfarm
business gained 1.9% annually in the 1970s and 1.4% in the 1980s. By the 1990s,
manufacturing output per person-hour increased twice as fast as total nonfarm
productivity, 4.0% to 2.0%.19 While manufacturing output declined by an average
18
These causes are listed and discussed in ibid., pp. 19-21.
19
“Output per person-hour” or “output per hour” is the traditional measure of labor
(continued...)
CRS-8
of 3% per year in 2000-02, hourly output per employee increased by an average of
3.3% per year, a level that nearly equals the performance of all nonfarm business.
These strong productivity numbers help explain why U.S. manufacturing has been
able to maintain high levels of output with a declining employment base.20
Figure 4 summarizes the data related to manufacturing from this section. It
compares changes in manufacturing employment, real output and productivity
measures since 1960. Employment, output measured in real terms, and output per
hour all increased in the 1960s and ‘70s. However, the trends diverged significantly
starting in the 1980s, and more dramatically in the 1990s. From 1960 to 1969,
employment grew 20%, productivity by nearly 30% and total real manufacturing
output by more than 50%. Despite a major recession in the middle of the decade,
employment continued to grow in the 1970s, reaching its all-time peak in 1979, more
than 25% higher than the 1960 level, while productivity was two-thirds higher and
total output more than double the level of 1960. To this point, the gains in the three
measures were synchronous, if not completely parallel.
Figure 4. Manufacturing Employment, Output and Labor
Productivity
400
350
Indices of Manufacturing:
Employment
Output
Productivity
300
250
200
150
100
1960
1965
1970
1975
1980
1985
1990
1995
2000
Source: U.S. Department of Labor. Bureau of Labor Statistics. “National Employment, Hours and Earnings,”
Employment in Manufacturing, viewed on September 15, 2003; “Major Sector Productivity and Costs Index,”
Output in Mfg and Output per Hour in Mfg., viewed September 16-17, 2003. All data from www.bls.gov
But employment never regained the 1979 peak in the 1980s, and ended the
decade only 15% higher than in 1960. Productivity and real output both continued
to grow. After the recession of 1991, manufacturing employment grew marginally
until 1998, then began falling — ending in 2002 at a lower absolute level than in
19
(...continued)
productivity in the economy. Hereafter, this report will use the shorter term.
20
Manufacturing in America on this point summarizes, “Because productivity gains in
manufacturing have outstripped the growth in demand for manufactured goods,
manufacturing employment has been falling for the past three decades;” see pp. 17-18.
CRS-9
1960. Labor productivity during the same period accelerated to reach nearly 3.5
times the rate of 1960, and continued to improve during the period of recession and
slow growth in 2001-02. The real value of total manufacturing output in 2002 was
also about 3.5 times the 1960 level, despite a dip in 2001. These data suggest a longterm trend growth in real manufacturing output, driven by improved productivity, and
that the decline in output in 2000-2 was only cyclical in nature.
Thus far, the analysis has dealt with U.S. “manufacturing” as a single,
undifferentiated sector. But, of course, a wide range of products and processes are
covered by the term, however it may be defined. The next sections of the report
examine divergences of performance within the manufacturing sector, and look in
more detail at selected specific industries.
Before examining these divergent trends, the report will look briefly at one more
topic that affects manufacturing across the board, unionization. Manufacturing has
traditionally been one of the most highly unionized (“organized” in labor union
terms) sectors of the U.S. economy. The decline of the share of workers in
manufacturing jobs has thus contributed to the widely noted decline of union
membership within the total workforce. By 2004, 12.5% of all wage and salary
workers in the United States were union members, the lowest percentage since this
statistical series was begun by BLS on its current basis in 1983, when the level was
20.1%. Within manufacturing, however, union membership rates have fallen even
more precipitously. It was 27.8% when the series began in 1983, and declined to
14.3% in 2002, barely more than half the earlier level — and only one point higher
than the overall average of the total workforce. Manufacturing remains one of the
more highly unionized sectors in the private sector economy, with only construction,
and transportation and public utilities ranking higher, though both have also
registered large declines over the past 20 years (by contrast, government employees
have maintained a relatively high and stable level of union representation, at just
under 40% throughout the period).21
Performance Varies Widely by Industry22
While analysis of the overall manufacturing sector in relation to the economy
as a whole is instructive, examination of trends of individual industries within the
sector can provide further insight. Different industries experienced considerable
variations in output growth, productivity gains, employment ups and downs; and
patterns of change over time. Moreover, instances of similar growth rates or patterns
may be due to different causes.
This section describes and analyzes the increases and/or decreases over two
discrete periods in output, labor productivity, and employment observed for the “twodigit” industries that comprise the manufacturing sector under the Standard Industrial
21
BLS. “Union Members in 2004,” annual press release and table (Jan. 27, 2005), and
unpublished time series from Current Population Survey.
22
This section was written by Bernard A. Gelb.
CRS-10
Classification system (SIC).23 The two-digit level is the first level of disaggregation
“below” that of manufacturing as a whole, containing a diversity of “sub-industries”
in many cases. It is used in this report because BLS does not have output and
productivity data for years prior to 1987 at a more disaggregated level.
The analysis is based upon indexes of real output and output per hour developed
by BLS, and the BLS’s employment by industry data. The periods covered are 1960
to 1990 and 1990 to 2000. The earlier period begins sufficiently later than the end
of World War II so as not to be significantly affected by that war, and is long enough
to capture secular trends. The latter period includes the unusually strong economic
years of the second half of the 1990s. For reasons of comparability, the analysis is
on a “peak-to-peak” basis; 1960, 1990 and 2000 are each a final year of an economic
expansion, as measured by annual totals of real GDP.24
This section of the report does not attempt to analyze trends in every individual
industry. This is partly because trends at the two-digit level are a composite of
developments in the sub-industries, which would necessitate analysis of each of the
sub-industries. Trends in information technology industries, in motor vehicle
manufacturing, and in textile and apparel manufacturing are examined in greater
detail later in this report under “Selected Major Industries.”
Output Trends
Perhaps the most striking findings of this analysis are the degrees of divergences
among rates of output growth by individual industries, and between those of
individual industries and that for manufacturing as a whole. One of the notable
aspects of this is the lack of concentration around an average rate of output growth
for all manufacturing.
The disparities that will be described reflect industries’ production method(s),
places in the chain of production, access to material inputs, product mix, distance
from markets, growth of market, technology level, rate of innovation, and degree of
exposure to foreign competition. And these aspects affect and are affected by the
evolution of the economy toward a smaller proportion of goods production being
accounted for by basic materials. Because the two-digit level is fairly aggregated,
analysis with respect to these factors can be clouded by differences in production
processes, product type and product technology level among parts of an “industry.”
As indicated in the first part of this report, real output by the manufacturing
sector increased at an average annual rate of 3.1% per year between 1960 and 1990.
However, only six of the 18 two-digit manufacturing industries for which BLS has
23
Although no longer used for data collection and compiling, the SIC system is used here
because output, productivity, and employment data developed by the Bureau of Labor
Statistics on the new North American Industrial Classification System are not yet available
for years preceding 1987. See the Appendix to this report for further discussion of the
change to the new system.
24
CRS believes that an annual, rather than a monthly, criterion for determining business
cycle peaks is the appropriate one for this report, inasmuch as the analysis uses annual data.
CRS-11
developed indexes of output had rates of growth during this period within 20% of
that for manufacturing as a whole; five had growth rates 40% or more below that of
manufacturing as a whole; and four had growth rates 40% or more above that of
manufacturing. For the 1990-2000 period, only three industries had growth rates
within 20% of that of manufacturing as whole; ten had growth rates 40% or more
below that for total manufacturing; and two had growth rates two and a half to four
times as fast as that for manufacturing as a whole.25 These two industries are the very
large machinery manufacturing groups, which notably include information
technology and electronic products. (See Table 1 for the data underlying this
paragraph and most of the rest of this section.)
1960 to 1990. In view of the factors listed above, it is not surprising that
output by the primary metals industry grew very slowly between 1960 and 1990, as
U.S. metal ore reserves became depleted and world production of primary materials
using energy-intensive processes tended to shift overseas, partly attracted by lower
energy costs abroad. Output by stone, clay, and glass manufacturers increased much
faster than that by primary metals producers even though they, too, produce basic
materials using energy-intensive processes. Losses by U.S. producers of stone, clay
and glass manufactures to overseas producers probably were limited partly due to the
low average unit value of some products, combined with the difficulty of transporting
many others, and the availability of these common materials close to their markets.
This raises the cost of transportation relative to production cost and limits the
geographic size of the market. Both industries’ output grew much slower than the
all-manufacturing average.
Industries with very rapid output growth between 1960 and 1990 include rubber
and plastic products, chemicals and allied products, industrial and commercial
machinery (including computers),26 electronic and other electrical equipment, and
instruments and related products. These industries are adept at innovation: entire or
large parts of these industries employ a high level of technology and/or employ
material inputs that are easily modified and/or combined to develop new products.
And, in many cases, their products substitute easily for higher value materials. Over
the 30-year period, these industries’ output rose at average annual rates between 4.0%
and 5.8%.
Industries with growth of about the same rate as total manufacturing in the
1960-1990 period include lumber and wood products, furniture and fixtures, paper
and allied products, and printing and publishing. The markets for goods that these
industries produce are predominantly linked to general economic and income growth.
While part of the paper industry performs energy-intensive basic material processing,
most of the value added in the two-digit industry is accounted for by the entities that
“convert” — through coating, cutting, forming, etc. — raw paper and paperboard into
25
Given the 1960-1990 average annual output increase of 3.1% for total manufacturing, an
industry would be within 20% if its average increase was between 2.4% (0.8 x 3.1) and
3.7% (1.2 x 3.1).
26
Computers and computer equipment were classified in this two-digit industry under the
SIC. Computer components were classified under Electronic and Other Electrical
Equipment and Components
CRS-12
the wide variety of paper and paperboard products used by businesses and
households. In the case of lumber and wood products, a big part of demand growth
stemmed from rising incomes that spurred a marked increase in the size of new
homes constructed, rather than by an increase in the number of homes built.
1990 to 2000. Between 1990 and 2000, output developments for most
individual industries were markedly different than their 1960-1990 experience. One
aspect is that divergences in production growth rates among individual industries,
when compared with the average for manufacturing as a whole, were even greater in
the 1990-2000 period than between 1960 and 1990. In the later period, only three of
the 18 two-digit manufacturing industries had rates of growth within 20% of that for
manufacturing as a whole (compared with six in the earlier period).
A second aspect is the difference between the growth rates of the individual
industries in the two periods. Production increases for the two fastest growing
industries (industrial and commercial equipment, electronic and other electrical
equipment) accelerated from their already rapid increases, as computerization and
telecommunication usage by businesses and households mushroomed, helped by
continuing product innovation. These industries’ average annual output growth rates
jumped from 5.1% and 5.8% to 9.0% and 14.3%, respectively. Output by the “subindustry” making computers and computer equipment rose at an average annual rate
of 29% between 1990 and 2000.
The slowest-growing industry in terms of output in 1960-1990 experienced
faster growth in 1990-2000, although not up to the all-manufacturing average.
Primary metals benefitted from modernization of the integrated mill portion of the
steel industry, accelerated increase in market share produced by steel minimills, and
computerization of the machine tools used by the fabricated metals industry.
Among industries that saw output growth slow in the 1990-2000 period were
paper and allied products, chemicals and allied products, and printing and publishing.
All three had been growing as fast as or faster than the all-manufacturing average.
The first two decelerated partly because imports started to make noticeable inroads
into domestic markets, in contrast with the earlier period. The third suffered from the
widespread penetration of in-house computer-enabled printing.
The 1990-2000 period also was characterized by the fact that twice as many
individual industries experienced slower output growth compared with 1960-1990
than industries that had faster output growth compared with 1960-1990. That total
manufacturing production rose faster between 1990 and 2000 than between 1960 and
1990 is because the industries with faster output growth were large enough and their
output increases were rapid enough to more than offset the slow production growth
of the other industries.
Productivity Trends
Trends in labor productivity among manufacturing industries in the two periods
covered in this report were at least as complex and diverse as the output trends,
although there was somewhat more concentration of rates of gain in the 1960-1990
period. Divergences in rates of gain in output per hour in the 1960-1990 period
CRS-13
among individual industries were less extensive than those for output, described
earlier. Nine of the 18 two-digit manufacturing industries had rates of increase in
output per hour within 20% of the annual average for manufacturing as a whole
(2.7% per year); four had growth rates 40% or more higher than that of
manufacturing as a whole; and two had growth rates 40% below that of
manufacturing. But the disparities were greater in the 1990-2000 period, during
which only three industries had growth rates within 20% of the all manufacturing
average (3.8% per year); eight had growth rates 40% or more below that for total
manufacturing; but two had growth rates two and a half to four times as fast as that
for manufacturing as a whole.
CRS-14
Table 1. Trends in Output, Labor Productivity, and Employment
in Manufacturing Industries: 1960-1990, 1990-2000
Output
Employment
(thousands)
Average Annual %
Change
Industry
‘60‘90
‘90‘00
‘60‘90
‘90‘00
1960
Total manufacturing
3.1
3.6
2.7
3.8
15,438 17,695 17,263
Food & kindred products
2.2
1.8
2.5
1.5
1,790
1,661
1,687
Textile mill products
3.2
1.0
4.2
3.6
924
691
531
Apparel, other fabricated
textile products
1.8
1.6
2.4
6.6
1,233
1,036
634
Lumber & wood products
3.0
1.1
2.7
- 0.2
670
733
830
Furniture & fixtures
2.8
3.9
1.8
2.7
365
506
556
Paper & allied products
3.3
1.5
2.8
2.2
597
697
656
Printing & publishing
3.0
0.5
1.1
0.7
911
1,569
1,547
Chemicals, allied products
4.0
2.2
3.1
2.6
828
1,086
1,034
Petroleum & coal products
1.9
1.2
2.7
3.6
212
157
127
Rubber & miscellaneous
plastic products
4.9
4.8
2.3
3.4
413
888
1,011
Stone, clay, glass, concrete
1.6
1.2
1.7
1.6
572
556
579
Primary metals
0.5
1.7
1.8
2.0
1,185
756
699
Fabricated metal products
1.9
3.2
1.5
2.1
1,230
1,419
1,539
Industrial & commercial
machinery, computers1
5.1
9.0
3.9
8.9
1,496
2,095
2,121
Electronic & other
electrical equipment2
5.8
14.3
4.7
14.0
1,221
1,673
1,726
Transportation equipment
3.0
3.6
2.4
4.0
1,668
1,989
1,852
Instruments, related
products
5.6
2.7
4.0
4.2
632
1,006
845
3.3
2.4
2.3
2.1
390
375
392
Miscellaneous mfg.
1
Output
per Hour
2
1990
2000
Includes computer equipment.
Includes communication equipment.
Source: Data in this table were derived from the following BLS sets of data: “Major Sector
Productivity and Costs Index” series, “Major Sector Multifactor Productivity Index” series, and
“National Employment, Hours, and Earnings” series.
CRS-15
The fairly high level of aggregation at the two-digit level combined with
considerable heterogeneity in some cases makes the challenge of explaining rates of
gain in productivity and changes in those rates over time particularly great for
industries such as transportation equipment, electronic and other electric equipment,
and rubber and miscellaneous plastic products.27 For example, in addition to motor
vehicles, products of the transportation equipment industry include guided missiles
and space vehicles, travel trailers and campers, and motorcycles and bicycles, among
others. And among the products of the electronic and other electric equipment
industry are electric distribution equipment such as transformers and switch gear,
household refrigerators and freezers, and electronic components and accessories such
as semiconductors and printed circuit boards.28
1960 to 1990. Probably indicating a connection between productivity
improvement and commercial success,29 the industries with the highest rates of
increase in output per hour between 1960 and 1990 by and large are those with the
fastest output growth. These industries included chemicals and allied products,
industrial and commercial machinery, electronic and other electrical equipment, and
instruments and related products, with average annual rates of gain in productivity
of 3.1%, 3.9%, 4.7%, and 4.0%, respectively. Manufacturing as a whole had an
average annual gain of 2.7%. One of the attributes of rapid output growth is that the
need for additional production capacity presents more opportunities for the
introduction of state-of- the- art production facilities and new production processes
than in the case of less rapidly growing industries. At least parts of the industries
cited above are characterized by rapid technological innovation, which tends to create
new products faster and generates increased demand.
Correspondingly, if the above assumption about a connection between
productivity improvement and commercial success is valid, it is not surprising that
four out of the five industries with the slowest gains in productivity (furniture and
fixtures; stone, clay, and glass; primary metals, and fabricated metal products) had
slower than average rates of increase in output during the 1960-1990 period.
Industries with average rates of increase in both output per hour and production in
the 1960-1990 period were lumber and wood products, paper and allied products, and
transportation equipment. Demand for the products of these seven slow- or averagegrowing industries tends to be related to general economic growth and/or population
growth. Less rapid output increase affords fewer opportunities for the introduction
of state-of-the-art production facilities and processes.
1990 to 2000. Between 1990 and 2000, productivity developments for most
individual industries were markedly different than their 1960-1990 experience —
similar to the changes in output. As noted above, divergences in output per hour
increase among industries were even greater in the 1990-2000 period than between
27
The same comment holds for analyses of production increases.
28
Such groupings have evolved from industrial classification concepts of several decades
ago, when the nature and range of products and production processes were considerably
different.
29
A suggestion of “connection” does not imply 100% predictability.
CRS-16
1960 and 1990. Also, productivity growth rates of the individual industries differed
markedly in the two periods.
In the 1990-2000 period, the rise in output per hour of the two fastest growing
industries (industrial and commercial machinery, electronic and other electrical
equipment) accelerated sharply from their already rapid 1960-1990 increases — from
3.9% and 4.7% to 8.9% and 14.0%, respectively. Productivity of the “sub-industry”
making computers and computer equipment rose at an average annual rate of 32%
between 1990 and 2000. The extraordinary gains of the above-mentioned two-digit
industries and marked improvement by several others offset the slower than average
productivity gains by nine other industries.
The “several others” were apparel, furniture, petroleum and coal products,30
rubber and miscellaneous plastic products, and transportation equipment. Eight of
the nine industries with slower gains in output per hour in the1990-2000 period than
in 1960-1990 had average annual increases of 2.5% or less. These were food and
kindred products, lumber and wood products, paper and allied products, printing and
publishing, stone-clay-glass, primary metals, fabricated metal products, and
miscellaneous manufacturing.
Employment Trends
As discussed in the section on the manufacturing sector as a whole, the trend
(though not the level) in employment in an industry to a great extent is an outcome
of changes in the industry’s production level combined with changes in labor
productivity in the industry. Given the diversity of trends in production and in labor
productivity among the two-digit industries, it is not unexpected that trends in
employment among the industries were diverse as well during the two periods
examined. Only four of the 18 two-digit manufacturing industries had percentage
changes in employment between 1960 and 1990 within 10 percentage points of the
increase for manufacturing as a whole (14.6%), whereas 11 industries had
employment percentage changes 20 points or more higher or lower than the
percentage change for manufacturing as a whole. Taking into account that 1990 to
2000 was one third as long as 1960 to 1990, the disparities appear greater in the
1990-2000 period: only two industries had percentage changes within three
percentage points of the all manufacturing average (-2.5%); and ten had changes 20
percentage points or more higher or lower than the percentage change for total
manufacturing.
1960 to 1990. Employment in the manufacturing sector as a whole increased
by about 3¼ million people between 1960 and 1990. But, as can be concluded from
the data on divergences in the previous paragraph, the gain in overall sector
employment is the net result of many increases and decreases in absolute numbers
among the individual industries. Not surprisingly, the largest absolute increases
tended to occur in large rapidly growing industries. Thus, the industrial and
commercial machinery, electronic and other electrical equipment, rubber and
miscellaneous plastic products, and instruments and related industries experienced
30
Petroleum refining accounts for about 90% of the output of this two-digit industry.
CRS-17
gains in employment of 600,000, 450,000, 475,000, and 470,000, respectively. But
two not particularly fast-growing industries, printing and publishing and
transportation equipment, registered employment increases of about 640,000 and
320,000, respectively.
Offsetting part of these large increases plus several smaller ones, were declines
in employment of varying magnitudes in seven other industries between 1960 and
1990. These include a 430,000 decrease in primary metals and a 130,000 decrease
in food and kindred products, in addition to employment decreases totaling 430,000
in the textile mill and the apparel industries (discussed later).
1990 to 2000. Given the fact that the 1960-1990 period was much longer than
the 1990-2000 period, the amounts of employment decreases and increases in
individual industries in the later period would be expected to be smaller. This was
the case. The largest decrease was that of 400,000 in apparel manufacturing, where
very rapid productivity increases (averaging 6.6% per year) and import competition
probably were factors. Drops of 160,000 were experienced both by textile mills and
by the instruments industry, and there was a 135,000 employment decrease in the
transportation equipment industry. The otherwise fast-growing instruments industry
was hit by a sharp decrease in defense spending on search and navigation equipment.
As for transportation equipment, employment losses in aircraft and parts more than
offset gains in motor vehicles and equipment, which will be discussed in a later
section. The total of the deceases in these industries was about 850,000;
miscellaneous others saw employment drop a combined 280,000. Increases of
120,000, 120,000, and 100,000 in employment in the rubber and plastics, fabricated
metals, and lumber and wood products industries, respectively, plus a number of
small increases in employment among other industries offset some of the total
decreases to yield an employment decline of about 430,000 in the manufacturing
sector as a whole.31
From examination of trends in output, labor productivity, and employment of
individual manufacturing industries, it appears that, notwithstanding their
commonalities, they differ sufficiently in many characteristics and in the markets they
serve. Therefore, they experience vastly different outcomes over the course of time.
Performance of Selected Major Industrial Sectors
Having noted a wide divergence of performance among different sectors of the
U.S. manufacturing economy, this report will now investigate three important sectors
in more detail. The three specific sectors selected are all important to the overall
U.S. economy, in terms of output and employment, and epitomize the divergence of
experiences within U.S. manufacturing.
!
31
Information technology (including especially computers,
communications equipment and semiconductor components) was a
The employment decrease and increase figures in this discussion of employment trends
are based upon unrounded numbers for the beginning and ending years of the two periods.
CRS-18
!
!
high-growth area of the economy in the 1990s, and has remained
internationally competitive.
The U.S. automotive sector, which has historically expanded more
by investment abroad than by exports, faced increasing pressure
from imports in the 1980s. Now many foreign-based manufacturers
have manufacturing operations in the United States, leaving an
industry notably divided between the traditional “Big Three”
(including one now foreign-owned company) and automobile
manufacturers that were originally internationally based.
Textiles and apparel are traditional U.S. industries that have been
struggling to compete against imports. Their output growth slowed
and employment fell in the 1990s, despite a strong overall economy.
Now they face the challenge of the elimination on January 1, 2005
of the remaining import quotas.
Information Technology Industries32
Decline in Information Technology Industries. Information technology
(IT) industries led U.S. economic growth in the 1990s, but their decline was a major
cause of recession in 2001. Afterwards, the IT sectors have lagged the economy,
rather than leading it, though by 2003 it had begun growing again.33 From the
perspective of the industrial economy and IT businesses, the recession hit hardest at
manufactured hardware, particularly in low levels of demand for computers and
telecommunications equipment in which semiconductors are the key component.
This section will focus particularly on trends affecting these product groups.
Figure 5 shows the strong growth especially of business investment in the two
largest product groups that are intensive users of semiconductors, computers and
communications equipment, for most of the period following the recession of the
early 1990s. The annual value of private business investment in computers and
peripherals more than doubled from $44 billion to $101 billion between 1992 and
2000. Communications equipment purchases by business increased at an even faster
pace, from $46 billion to $124 billion. Together with software, these items
accounted for about 50% of the entire increase in U.S. business equipment
investment in the 1990s. If we refer to the inflation-corrected values reported by the
Commerce Department Bureau of Economic Analysis (BEA), information-processing
equipment and software accounted for 68% of the total real increase in business
investment between 1992 and 2000, and nearly 20% of all real U.S. growth.
However, BEA cautions that the inflation-corrected estimates are questionable
32
33
This subsection was written by Stephen Cooney.
Tom Runiewicz argued that “high tech” again led the recovery for two reasons: the “short
life cycle for computers and related equipment” as systems ordered for Y2K became
outdated, and corporations look at IT systems as a quick way to boost productivity, as the
economy recovers; Global Insight: Perspectives, viewed online Nov. 18, 2003.
CRS-19
because prices for computing power and performance have been changing so
rapidly.34
The rapid rate of growth in the capital-goods end markets for IT products
reversed sharply in 2001. As clearly shown in Figure 5, business demand for
computers and communications equipment fell by a combined total of $30 billion.
Consumer demand for computers, mostly PCs for home use, held up better than
business investment, falling only marginally, but this is a much smaller market, less
than $50 billion per year, including software. Also displayed in Figure 5 is consumer
demand for video and audio equipment, a broader consumer market of more than $70
billion, but less intensive in its use of semiconductors and other IT inputs; also, many
of these types of products are imported. The market for these consumer products did
not grow as strongly after 2001, but it never seriously declined. In 2002, business
and consumer demand for computers stabilized in dollar value, but business
investment in communications equipment fell again, from $110 billion to $89 billion.
However, business investment and consumer purchases of computers, as well as
business investment in communications equipment, turned up in 2003, suggesting
that IT products contributed a stronger economic recovery.
Figure 5. U.S. Information Technology Market Change
140
120
$ billions
100
80
60
40
20
0
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Business Investment: Computers & Peripheral Equipment
Business Investment: Communications Equipment
Consumer Goods: Video & Audio Goods
Consumer Goods: Computers & Peripherals (incl. Software)
Source: U.S. Dept. of Commerce. Bureau of Economic Analysis. National Income and Product
Accounts, Tables 2.4.5 and 5.5.5 (August 2004).
34
BEA. NIPA Table 5.9, published August 2002, fn. 1 states, “because of rapid changes in
relative prices, the [real] estimates for computers are especially misleading as a measure of
the contribution or relative importance of this component.”
CRS-20
Semiconductors Bolster U.S. International IT Competitiveness. The
IT industry is a global industry, with respect not only to sourcing but also to
standards and applications. The relative international competitiveness of U.S. IT
equipment manufacturing sectors, widely considered as “sunrise” industries, has
become an issue for the U.S. employment and manufacturing base. With slow or
unsteady growth in markets outside the United States, there has also been a negative
impact on U.S. exports, including IT products.
Semiconductors are the key IT product in which there is a U.S. comparative
advantage. The United States has recovered the overall technological and
competitive leadership position in semiconductors, which it once was in danger of
losing, in terms of dollar value of products sold by U.S. companies. U.S. companies
had lost their initial preeminence in the global market to Japanese companies by the
mid-1980s. When U.S. industry threatened a massive antidumping case against
Japanese producers, the Reagan Administration negotiated the U.S.-Japan
Semiconductor Agreement of 1986. The Japanese industry and government agreed
to accept as a target a minimum penetration level of 30% of its market from all
foreign-owned companies, along with a monitoring mechanism.35 A second initiative
was establishment of a private-public consortium between the federal government
and U.S.-based semiconductor and manufacturing industry producers, known as
“SEMATECH” (for Semiconductor Manufacturing Technology). In addition,
macroeconomic factors favored U.S. producers. The fall in the dollar exchange rate,
as noted elsewhere in this report, helped U.S.-based producers regain their market
competitiveness, while slow domestic growth in Japan in the 1990s held down
demand in Japanese producers’ home market.
Subsequently, U.S. companies recovered overall dominance of the world
semiconductor industry. In 1988-89 Japanese manufacturers held more than 50% of
the global semiconductor market, the only time they have done so. At that point,
Japan’s companies controlled 88% of their domestic market, about 35% of the
market elsewhere in Asia and the Pacific, and more than a quarter of the U.S.
domestic market. Because of the opening of Japan’s own market, prodded by careful
monitoring of the Semiconductor Agreement in the early 1990s, Japanese companies’
share of their domestic market declined to a consistent level of about 70% in the late
1990s. U.S. companies’ share increased from 13% to 23% of the Japanese market
over the same period.36 Meanwhile, by 2001, the Japanese share of the U.S. market,
declined by half to 11%, while the U.S. companies’ share was greater than 70%.
Japanese producers share of the Asia-Pacific market outside Japan fell by half, to less
than 19%, while U.S. companies held more than 50% of that market. In Europe,
35
The history of this agreement is reviewed in CRS Report 96-486, The U.S.-Japan
Semiconductor Agreement: Should It Be Renewed? by William H. Cooper (May 30, 1996).
For an account from the perspective of a U.S. negotiator, see Clyde V. Prestowitz, Jr.,
Trading Places (New York: Basic Books, 1988), ch. 2.
36
All regional figures are from Semiconductor Industry Association (SIA), World
Semiconductor Statistics, as cited in CRS Report RL31708, Semiconductors: The HighTechnology Downturn and Issues in the 108th Congress. The market share figures are
counted according to the home country of companies, regardless where the semiconductors
were manufactured.
CRS-21
where Japanese companies never have gained a large share of the market, their role
declined to 14%, while U.S.-based companies again had more than 50% of the
market.37
U.S. exports and world trade in IT products were encouraged further by the
Information Technology Agreement (ITA), negotiated in 1996. Most of the countries
that produce IT equipment agreed to eliminate all tariffs on such products and
components, including computers, telecommunications equipment and
semiconductors (but not consumer electronics). There were originally 43 signatory
countries, which accounted for more than 90% of all trade in IT products; the total
value of trade covered in the original agreement was more than $650 billion. For the
United States and Japan, signature of the ITA in principle involved relatively little
change, since neither country had a tariff on semiconductors, and most other
electronics tariffs were minimal. The European Union eliminated its substantial
tariffs on semiconductors and other electronic products, while China and India both
joined the agreement, after negotiating transitional “staging” periods.38 The major
trading countries not included in the ITA bloc are the nations of Latin America,
notably Brazil, Argentina and Mexico. Though U.S. chip exporters have free access
to Mexico through NAFTA, Mexico still applies tariffs to chips imported from nonfree-trade partner countries.
With rapid economic growth, renewed U.S. industry competitiveness and a more
open global trade environment, both U.S. exports and imports of most IT products
grew rapidly in the 1990s, as shown in Table 2. But domestic growth was stronger
than foreign growth, leading to U.S. imports increasing at a higher rate than exports.
This was especially the case for computer peripherals, which the United States has
always tended to import. Also, the table shows that U.S. computer exports increased
only marginally in the 1990s, an indication that the U.S. comparative advantage is in
semiconductors, the memory and logic of IT products. But the comparative
advantage in semiconductors has been so strong that even with increased global twoway sourcing, the U.S. trade surplus in these products grew along with increased
domestic growth and trade, and at least partially offset the deficit in peripherals. As
exports grew from $16 billion to $60 billion between 1992 and 2000, the U.S. trade
position in semiconductors changed from a balanced position to a U.S. surplus of $12
billion. Moreover, the United States maintained a balanced status in trade in
communications equipment.
37
It is possible that some of this market-share gain in the Asia-Pacific region was due to
increased outsourcing by U.S. companies of printed circuit board assembly operations,
relative to similar activities by Japanese companies. But the sizes of overall gains of U.S.
companies against Japanese companies in all markets make it unlikely that this is solely an
aspect of U.S. outsourcing.
38
See CRS Report 98-376, The Information Technology Agreement (ITA): Background on
a Proposal to Expand the Scope of the Multilateral Trade Agreement, by Glennon J.
Harrison.
CRS-22
Table 2. U.S. Trade in Information Technology Products
(All figures in billions of dollars)
1992
2000
2001
2002
X
M
Bal
X
M
Bal
X
M
Bal
X
M
Bal
Semiconductors
16
16
0
60
48
12
45
30
15
42
26
16
Computers
9
5
4
11
14
-3
11
13
-2
9
16
-7
Computer
Access.
20
27
-7
44
76
-32
37
61
-24
29
59
-30
Telecom.
Equipment
12
11
1
31
33
-2
28
25
3
22
23
-1
X = exports; M = imports.
Source: Department of Commerce. Bureau of the Census, Foreign Trade Division.
U.S. International Trade in Goods and Services (FT900), “Annual Revisions” for
1993, 2001 and 2002, Exhibits 6-7 (www.census.gov/foreign-trade/Press-Release).
The bottom fell out of the global semiconductor business in 2001. Total
worldwide sales fell from more than $200 billion to about $140 billion in 2001 and
2002, though stronger growth has recently been reported.39 The surplus in
semiconductors marginally increased in 2001-02, however, even though
semiconductor exports fell to $42 billion. The picture for computers was not so
positive. After a decade of roughly balanced trade, imports of computers nearly
doubled exports in 2002: $16 billion to $9 billion. After 2000, telecommunications
equipment trade has roughly stayed in balance, as exports and imports have fallen
proportionately.
IT Manufacturing Employment. The employment impact of these
trade and product demand developments in general mirrors that discussed earlier
regarding manufacturing employment in general. In 1990, 1.9 million people were
employed in manufacturing computer and electronic products, which includes all the
product groups discussed here, plus other electronic products such as instruments and
medical equipment. Despite the sector’s strong contribution to U.S. growth,
employment hit a plateau of about 1.7-1.8 million in the late 1990s through 2000.
Afterwards, the employment level dropped dramatically, reaching an annual rate of
1.5 million reported for 2002, and a 2003 total of less than 1.4 million.
About one-third of the employment in this sector is in semiconductors and other
electronic components, where the number increased from 574,000 in 1990 to a peak
of 676,000 ten years later, then fell to 525,000 in 2002, and 461,000 in 2003.
39
SIA. World Semiconductor Statistics, “World Semiconductor Shipments.” For the 2003
upturn, see SIA press release, “Global Chip Sales Reach $13.42 Billion in August 2003,”
which noted that this improvement was the sixth consecutive monthly increase.
CRS-23
Employment rose in communications equipment at a slower rate in the 1990s,
peaking at 247,000 in 2000, but had fallen to 191,000 in 2002. Employment in
manufacturing computers in the 1990s never reached the 1990 level of 367,000; it
peaked at 322,000 in 1998, and had fallen to 249,000 by 2002 with the industry
downturn. Instruments and other electronic products recorded a major and steady
loss of employment right through the decade, from 626,000 in 1990, through the
500,000 level in 1998, to 478,000 by 2000 and 450,000 in 2002 — a nearly 30% fall
in employment. Altogether, despite the high growth in IT’s role in the U.S. economy
and the robust international competitive position of the U.S. industry, employment
gains were absent or relatively modest, and disappeared with the onset of the
recession in 2001.
The emergence of China as both a large market and a major producer of
electronics could also significantly alter the conditions for technology dominance and
site selection for chip manufacturing locations. Already, the development of China
as a major market helps explain why chip sales in Asia and the Pacific (outside
Japan) increased by almost 30% in 2002, when they declined everywhere else.40 In
2000, according to SIA figures, the “Americas” region (comprising the entire
Western Hemisphere, but dominated by the U.S. market) represented the world’s
leading market region for semiconductor shipments: $64 billion, or almost a third of
the global total. By 2002, Asia-Pacific (ex-Japan), with $51 billion in shipments
accounted for more than one-third of the global total, $20 billion ahead of the
Americas. By 2006, SIA forecast that Asia-Pacific will account for more than 40%
of semiconductor shipments, and Japan, which is already recovering due to
aggressive efforts of its companies to sell into China, is projected to move into
second place.41
One aspect of China’s rapid integration into the world economy is that Chinese
production may create permanent downward pressure on prices, because of its
virtually inexhaustible supply of low-cost labor. This phenomenon may especially
affect pricing for those electronics products and components, such as consumer
electronics and printed circuit boards, where labor cost is critical. There is also some
question as to whether China will ever be a net exporter of semiconductors,
especially the more sophisticated variety, in the foreseeable future.42
40
SIA. “World Market Shares, 1991-2001,” as updated for 2002 final data (provided
courtesy of SIA). Business Week has particularly focused on the phenomenon of Chinese
development, as in “High Tech in China: Is It a Threat to Silicon Valley?” (Oct. 28, 2002)
and “Greater China” (Dec. 9, 2002). See also Electronic News, “China Gains as U.S.
Economy Struggles,”(Sept. 23, 2002).
41
“SIA Projects Robust Growth for Semiconductor Industry,” released Nov. 5, 2003 on its
website, [http://www.semichips.org]. The Japanese newspaper Mainichi Shimbun has
claimed that global industry monitoring data (World Semiconductor Trade Statistics) will
show that Japan will again overtake the United States in total semiconductor shipments in
2003; noted in Dow Jones International News, “Japan to Top US in 2003 Chip Shipments,”
(Sept. 28, 2003).
42
Business Week, “How Low Can Prices Go?” (Dec. 2, 2002); Robert Samuelson,
“Deflation Out of China?” in Washington Post (Dec. 4, 2002); Dow Jones International
(continued...)
CRS-24
Nevertheless, an SIA executive noted that manufacturing and R&D expenditures
in his industry tend to follow market location, and believed that the booming market
development in China and elsewhere in Asia are challenging for production located
in the United States.43 There was concern, voiced by U.S. trade officials and the
industry, that China maintained a discriminatory value-added tax regime that directly
promoted domestically manufactured semiconductors against imports.44 The United
States challenged China’s policy in a WTO case, and China agreed to eliminate the
discriminatory features of its tax policy on semiconductors. A report prepared for
SIA maintains that China’s increasing capability as a competitor in semiconductor
manufacturing is due not to low labor costs, which may be the case for other
electronic products, but to tax policies, and other measures to promote local
production. China has moved away from reliance on state-owned companies to
liberalization of foreign investment inflows, explicit cooperation with Taiwaneseowned companies and adoption of other aspects of Taiwan’s high technology
development model.45
Automobiles and Light Trucks46
The U.S. automotive industry is the largest in the world. In 2002, the United
States ranked first in vehicle production, with total production of 12.3 million
vehicles. Japan ranked second, with a production of 10.2 million vehicles, and
Germany third with 5.5 million vehicles. The United States has the largest national
market in the world for total vehicle sales. In 2002, the U.S. market, with sales of
16.8 million vehicles, was almost three times larger than the next largest market in
Japan.47 In the 1980s, tensions with Japan over automotive trade led to “voluntary”
42
(...continued)
News, “China’s Reliance on Chip Imports to Continue, Study Says” (July 24, 2003). See
also Economist Intelligence Unit — Business Asia. “Misincentives in China,” (October 6,
2003).
43
He also clarifies that the official SIA numbers may understate the large and growing role
of semiconductor foundries, based largely in Taiwan, which produce chips for sale by other
companies. Daryl Hatano, SIA. “Fab America — Keeping U.S. Leadership in
Semiconductor Technology,” presentation May 9, 2003. Market share data from SIA,
“World Market Shares 1991-2001” and “Semiconductor Forecast Summary, 2003-2006”
(June 2003).
44
U.S. Trade Representative. National Trade Estimates Report on Foreign Trade Barriers
(2003), pp. 49, 53-54.
45
“SIA Report Details Growth of China Chip Industry; Tax on Semiconductor Trade
Penalizes Importers, Distorts Investments,” Business Wire (Oct. 29, 2003). Details are in
China’s Emerging Semiconductor Industry, by Thomas Howell et al. of the law firm Dewey
Ballantine, for SIA (Oct. 2003), especially ch. 3.
46
This subsection was written by M. Angeles Villarreal. The data and analysis have been
updated and expanded in CRS Report RL32883, U.S. Automotive Industry: Policy Overview
and Recent History, by Stephen Cooney.
47
Center for Automotive Research (CAR), Economic Contribution of the Automotive
Industry to the U.S. Economy - An Update, study prepared for Alliance of Automobile
(continued...)
CRS-25
Japanese export restraints negotiated by the Reagan Administration, and later
continued unilaterally by the Japanese companies. One result of these voluntary
quotas was to encourage the major Japanese companies (Honda, Toyota, Nissan) to
invest in automotive manufacturing facilities in the United States, the so-called
“transplants.” These have been followed by major manufacturing investments by
Daimler Benz and BMW of Germany, and by Hyundai of Korea.48
The “Big Three” U.S. -based automakers, General Motors Corp. (GM), Ford
Motor Co., and Chrysler Group (now a part of DaimlerChrysler AG), together lead
the market in U.S. passenger car sales, although their share of the market has
declined to less than 50%: 48.4% of U.S. sales in 2002, down from 55% in 2000 and
51.4% in 2001. In light trucks, which include sport utility vehicles (SUVs), the
market share of Big Three sales is much higher, but it has also been in decline in
recent years. In 2002, the three U.S. automakers accounted for 76.6% of light trucks
sold in the United States, down from 79% in 2000 and 77.2% in 2001. The
remainder of cars and light trucks sold in the United States were produced by foreignbased companies. Most of these vehicles were produced in transplant facilities.49
Production Trends. Figure 6 shows total motor vehicle production in the
United States between 1978 and 2002. In 1978, U.S. motor vehicle production was
12.8 million, but decreased to a low of 7.0 million in 1982. In the early 1990s, motor
vehicle production grew considerably, partially due to the strengthening U.S.
economy, and also to an increase in light truck production. Production of light trucks
increased from 3.7 million in 1990 to 7.1 million in 2002. In comparison, production
of cars decreased from 6.1 million in 1990 to 5.0 million in 2002. The Alliance of
Automobile Manufacturers50 gives three factors for the resilience of the U.S. motor
vehicle industry after 1990. First, the overall size of the U.S. automotive market was
driven by growth in personal income and the formation of U.S. households. Second,
auto manufacturers were quick to respond after the events of September 11, 2001 by
providing price incentives, which helped sustain motor vehicle demand. Third,
automakers have been very responsive to the growing demand for light trucks and
SUVs, which comprised almost 59% of U.S. light vehicle production.51
47
(...continued)
Manufacturers (Fall 2003).
48
An early “transplant” investment by Volkswagen in New Stanton, Pennsylvania was
closed. The Daimler Benz investment in Vance, Alabama, preceded the company’s merger
with Chrysler.
49
Standard & Poor’s, Industry Surveys: Autos & Auto Parts, December 26, 2002.
50
The Alliance of Automobile Manufacturers is a coalition of 10 car and light truck
manufacturers, including BMW Group, DaimlerChrysler, Ford Motor Company, General
Motors, Mazda, Mitsubishi Motors, Nissan, Porsche, Toyota, and Volkswagen.
51
Study prepared for the Alliance of Automobile Manufacturers, Economic Contribution
of the Automotive Industry to the U.S. Economy — An Update (Fall 2003), p. 3.
CRS-26
Figure 6. Total U.S. Automobile and Light Truck Production
(Millions)
14.0
7.4
3.9
5.7
12.0
3.5
2.2
3.2
3.5
3.8
4.1
4.1
8.9
9.2
2.4
1.2 1.2
8.0
6.8 6.7
7.1
6.5
4.9
3.7
10.0
7.2
5.6 5.8 6.2 6.5
4.0
3.4
7.8
8.2
7.8
7.1 7.1
1.9 6.8
6.8
6.6
6.1
6.0
5.1
5.4 5.7
6.0
6.3
6.1 5.9
5.5 5.6 5.5
4.9 5.0
4.0
2.0
0
1978
1982
1986
Light Trucks
1990
1994
1998
2002
Cars
Sources: Chart prepared by CRS based on data from the following: 1978-1989: Ward’s
Automotive Yearbook, 1989 and 1990; 1990-2002: Alliance of Automobile Manufacturers.
The share of motor vehicles produced by transplant facilities in the United
States has increased significantly. Between 1990 and 2002, the share of cars
manufactured in the United States by U.S. manufacturers decreased from 78.3% to
63.1%, while that of foreign manufacturers increased from 21.7% to 36.9% (see
Table 3). The Big Three are still among the top four car manufacturers in the United
States, with GM ranking first. In 2002, GM produced 1.7 million cars, and was
followed by Ford, with a production of 1.1 million cars; Honda, with a production
of 641,000 cars; and the Chrysler Group, with a production of 420,000 cars. In 2003,
Honda and Toyota each produced more cars in the United States than did the
Chrysler Group of DaimlerChrysler (Honda: 593,000; Toyota: 436,000; and,
Chrysler: 363,000).52 Global Insight, the econometric forecasting firm formerly
known as DRI, estimated that U.S. passenger car production by Toyota and Honda
would significantly outpace the Chrysler Group output through 2008.53
In light truck and sport utility vehicle (SUV) production, foreign producers have
also markedly increased their share of total U.S. production. Between 1990 and
2002, their share of light trucks and SUVs manufactured in the United States
increased from 5% to 14%. However, the Big Three remain by far the largest
producers of light trucks, with GM ranking first. In 2002, GM produced 2.4 million
units of light trucks or SUVs, followed by Ford Motor Co., with 2.3 million units;
and the Chrysler Group, with 1.3 million units. Of the foreign manufacturers, Toyota
ranked first, with a production of 280,848 units in 2002. Nissan, Honda, and the
52
Automotive News, October 13, 2003, p. 42.
53
Global Insight. World Car Industry Forecast Report (September 2003), pp. 200-201.
CRS-27
GM-Toyota joint venture New United Motor Manufacturing, Inc. (NUMMI) each
produced more than 100,000 units.
Table 3. Distribution of U.S. Motor Vehicle Production
Car Production
1990
Light Truck/SUV Production
2002
1990
2002
Units
(Mils.)
Share
(%)
Units
(Mils.)
Share
(%)
Units
(Mils.)
Share
(%)
Units
(Mils.)
Share
(%)
GM
2.65
43.6
1.67
33.3
1.47
41.3
2.42
34.2
Ford
1.38
22.7
1.07
21.4
1.39
39.0
2.34
33.1
Chrysler1
0.73
12.0
0.42
8.4
0.53
14.8
1.33
18.8
Total Big
Three
4.76
78.3
3.17
63.1
3.38
95.1
6.09
86.1
ForeignBased Mfrs.
1.32
21.7
1.85
36.9
0.17
4.9
0.98
13.9
Total U.S.
Production
6.08
—
5.02
—
3.55
—
7.07
—
1
Chrysler Corp. prior to 1999, now Chrysler Group of DaimlerChrysler.
Source: Ward’s Automotive Yearbook, 2003.
Employment Trends. Figure 7 shows employment in the motor vehicles and
equipment industry from 1978 to 2002, based on BLS data. Unlike the general
patterns of employment in manufacturing, peak employment in the auto industry,
defined broadly, occurred in 1999, while that in manufacturing occurred in 1979. In
1978, the total number of employees in the auto industry was just over one million.
In 1980, the number of jobs in the auto industry decreased considerably due to the
1980 recession. However, during most of the 1990s, employment in the auto sector
increased consistently and reached a peak of more than one million jobs in 1999.
Since then, employment dropped to 911,000 in 2002.
CRS-28
Figure 7. Employment in Motor Vehicle and Equipment Industry
(Thousands)
1200
1000
800
600
400
200
0
1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
Source: BLS (SIC Code 371) at www.bls.gov.
Employment indirectly related to automotive manufacturing is much larger.
However, the majority of jobs indirectly related to the auto industry are not directly
tied to manufacturing. The 2003 study by the Center for Automotive Research cited
above found that while the automobile manufacturing was the largest U.S.
manufacturing activity, directly accounting for 1.2 million jobs by its estimate, the
majority of jobs related to the industry are in supplier and related industries - such as
dealerships, auto repair and maintenance, plastics, and rubber, trucking, etc. - and not
in direct manufacturing. The study estimated that employment associated with total
automotive industry activity in the United States was about 3.5 million jobs, and 6.6
million when all related activities are included. Annual compensation that was
directly attributable to the industry was estimated to be about $152 billion.54 The
study estimates that about four in ten indirect jobs generated by the auto industry are
in manufacturing, and most of them are in durable goods.55
The overall direction of motor industry employment was also affected by the
2003 industry collective bargaining agreement reached between the Big Three and
the United Auto Workers union (UAW). Acknowledging that U.S. automakers were
facing difficulties in increasing competition from foreign manufacturers, the UAW
made a number of concessions to the automakers in the new contracts. Unlike the
54
Compensation as used here is defined as the total contribution of the automotive industry,
including wages and benefits, to U.S. private sector income.
55
CAR study, p. 17.
CRS-29
1999 contracts, the 2003 contracts did not ban plant closings.56 They were designed
to help the Big Three respond to increasing competition:
!
GM sought to save retiree costs, especially in rapidly rising
healthcare outlays, because it has a higher ratio of retirees to current
employees. Any increase in retiree pension and healthcare costs thus
affects GM more than the other two major U.S. automakers. The
UAW and GM reached an agreement in which GM would shut down
three plants employing 1,500 workers, in exchange for assurances
that it would continue to choose its unionized former affiliate Delphi
for parts, as opposed to nonunion suppliers.57
!
Ford sought to reduce excess capacity by shutting down five vehicle
production plants protected by the previous contract. The UAW
agreement with Ford let it close or sell four plants, to eliminate
4,600 manufacturing jobs through plant closures in the United
States. There was to be a further reduction of 3,000 salaried
positions in North America, and cuts of more than 4,000 jobs in
Europe. With the plant closings, Ford could meet its goal of
trimming North American production capacity by nearly one million
units.58
!
Chrysler wanted concessions from the union to help in vehicle
assembly labor productivity so that it would become more
competitive. In exchange, the company sought to enhance wages or
post-retirement benefits for UAW members. The UAW said
Chrysler identified nine plants that it wanted to close or sell, but that
four would remain covered by the new agreement. Chrysler
confirmed that it would close two parts plants, eliminating 1,580
jobs, and sell three others under the ratified contract.59
The effort between auto producers and the UAW in negotiating the new
contracts was designed to help U.S. manufacturers address problems facing the
industry. One of the issues is that the U.S. market may be facing overcapacity as
foreign competitors continue to build plants in North America. The UAW agreed to
certain plant closures in exchange for a preservation of wages and benefits. Some
analysts have stated that while the new contracts offered hope for the Big Three, the
union’s concessions on plant closings may not have been enough and that the
56
Chicago Tribune, “UAW Oks Pact with DaimlerChrysler,” September 27, 2003, p. 2.
57
St. Louis Post-Dispatch, “GM Contract Would OK Closings,” Sept. 22, 2003, p. A5.
58
Automotive News, “Ford Trims Capacity, Still Has Problems,” October 6, 2003, p. 8.
59
Automotive News, “More Cuts for a Bloody Chrysler,” October 6, 2003, p. 6.
CRS-30
industry may need to close additional plants.60 Most transplants are not unionized.
Toyota, Nissan, Honda, BMW, and Mercedes employed about 48,000 non-union
workers, with considerably lower total benefit costs.61
Textiles and Apparel62
Steeply rising imports and steeply declining employment in U.S. textile and
apparel manufacturing63 have brought considerable attention to and concern about
these industries. Because of their importance to the U.S. economy, to certain U.S.
geographic regions, and to many U.S. trade partners, textiles and apparel have been
major issues in U.S. trade relations with a number of countries, leading to the signing
of bilateral and multilateral agreements generally restricting the quantities of textiles
and apparel traded, including the establishment of quotas.
The Economics of Textile and Apparel Production. Textile and apparel
manufacture, and international trade in those products, have been important elements
of economic activity and growth since the Industrial Revolution. Major reasons for
this are (1) textiles and apparel are basic items of consumption in all countries, and
(2) textile manufacture to some extent, and apparel manufacture in particular, are
labor-intensive, requiring relatively little fixed capital for entrepreneurs to establish
production facilities. Thus, these industries are major generators of employment.
Modest capital requirements contributed to textiles and apparel becoming major
industries at the start of the Industrial Revolution and remaining important to
developing countries now. The percentage of total manufacturing value added
accounted for by textile and apparel production among developing countries, for
example, was triple the percentage among industrialized countries in 2000.64
Lower wage rates in developing countries together with the labor-intensiveness
of apparel manufacture tend to give these countries a comparative advantage in
apparel manufacture and a locational advantage for textile manufacture. Thus, textile
and apparel manufacture is tending to shift to developing countries, with textiles and
apparel constituting large portions of their exports. Textile and apparel manufacture
(measured by constant-dollar value added) in industrialized countries declined
between 1980 and 2000, whereas textile and apparel manufacture in developing
60
International Herald Tribune, “Big 3’s Labor Deals Offer Hope But Union’s Concessions
on Plant Closings May Not Be Enough,” September 25, 2003, p. 25.
61
USA Today, “Foreign Companies Cast Long Shadow on UAW Negotiations,” August 6,
2003, p. 1B.
62
This subsection was written by Bernard A. Gelb.
63
Common usage often includes apparel and other fabricated textile products under the
general term “textiles.” For greater precision, this report uses the more specific terms, with
“textiles” generally meaning fibers and fabrics, and “apparel” meaning items of clothing
other than footwear. Exceptions are where industry production and trade data are reported.
64
United Nations, Industrial Development Organization. International Yearbook of
Industrial Statistics 2002. Vienna: 2002. p. 55.
CRS-31
countries increased.65 Between 1980 and 1999, textile and apparel exports of
developing economies (in nominal dollars) rose 500% while developed economies’
textile and apparel exports rose 125%. Textiles and apparel comprised 13% of
developing economies’ exports in 1999, versus 4% for developed economies.66
U.S. Textile and Apparel Production, Trade, and Employment.
Textile and apparel manufacturing were two very large industries when the 1960s
began; and U.S. textile and apparel manufacturing output rose respectably between
1960 and the early 1990s. Textile manufacturing production tripled between 1960
and 1994; apparel manufacturing production doubled. Since 1994, however, output
by both industries has fallen (Table 4). In contrast, total U.S. manufacturing output
nearly tripled between 1960 and 1994, and rose 20% more between 1994 and 2002,
despite the recent recession.67
Table 4. Output, Productivity, and Employment
in U.S. Textile and Apparel Manufacturing Industries
Apparel and Other Fabricated
Textile Products
Textile Mill Products
Years
Output
Productivity Employment
(1996 = 100) (1996 = 100) (thousands)
Output
(1996 =
100)
Productivity Employment
(1996 = 100) (thousands)
1960
33.6
23.1
924
50.4
37.1
1,233
1973
65.7
40.6
1,010
75.5
48.5
1,438
1987
86.6
72.7
725
93.5
76.0
1,097
1994
101.5
91.2
676
100.8
90.0
974
2000
95.4
112.0
531
102.4
142.1
634
2002
85E
n.a.
432
98E
n.a.
521
E - CRS estimate based upon Federal Reserve indexes of industrial production.
n.a. - Not available.
Source: Data in this table were obtained from the following BLS sets of data — “Major Sector
Multifactor Productivity Index” series, and “National Employment, Hours, and Earnings” series.
More significant to many in the U.S. textile and apparel industries, employment
in those industries has decreased markedly in the last three decades — by 57% and
64%, respectively, between 1973 and 2002. The two industries together employed
about 950,000 in 2002, or 6% of total manufacturing employment, compared with
65
United Nations. op. cit. p. 58-59.
66
United Nations. 1994 International Trade Statistics Yearbook, Vol. II. New York: 1995.
p. S-20, 76, 92; 1999 International Trade Statistics Yearbook, Vol. II. New York: 2000. p.
S-42, 98, 114.
67
Production change data for U.S. industries used hereafter in this report are based upon
indexes of real gross output derived by BLS — designed to reflect changes in the constantdollar value of production — rather than in constant-dollar value added (used by the U.N.).
CRS-32
2.4 million, or 16% of total manufacturing employment in 1973. And the 700,000decline in the two industries’ employment between 1994 and 2002 equaled about
40% of the drop in total manufacturing employment.
Some of the decline in U.S. textile and apparel employment is linked to gains
in productivity, and some to increases in importation of textiles and apparel. Output
per hour in textile manufacturing more than tripled between 1960 and 1987, then rose
more than 50% between 1987 and 2000. Apparel manufacturing output per hour
doubled in the earlier period, and nearly doubled again in the shorter later period.
Thus, gains in textile manufacturing productivity were more rapid between 1960 and
1987; and those in apparel manufacturing the most rapid more recently (Table 4).
U.S. imports of textiles in 2002 (in current dollars) ran more than three times
their 1980 level; and 2002 apparel imports were more than ten times their 1980 level.
Imports of all textiles and apparel exceeded exports by an estimated $62 billion in
2002. To a great extent, the increase in imports of apparel over the years reflects a
number of sharp increases in imports. For example, there were two such increases
in the 1960s, one in the 1970s, one in the 1980s, and one in the 1990s.
As can be seen from the import data above, U.S.-made textiles have fared less
badly with respect to trade than U.S.-made apparel. Textile production is less laborintensive, more easily automated, and, as a major input to apparel, can be exported
to serve as inputs to foreign-made apparel that then is exported to the United States.
The U.S. textile manufacturing industry also has been helped by requirements in
many trade agreements and trade preference programs that U.S.-made fibers and
fabrics be used to produce the apparel made abroad.
The considerable extent of U.S. textile and apparel trade with developing
countries is indicated by the following 2002 data. For imports, 4 of the top 10 fiber
and fabric import sources, 4 of the top 10 non-apparel textile product sources, and 8
of the top 10 apparel sources were developing countries. For U.S. exports, 7 of the
top 10 fiber and fabric destinations, 4 of the top 10 non-apparel product destinations,
and 7 of the top 10 apparel destinations were developing countries. Mexico was
among the top five in all the above import and export product groups; and China was
first as a textile product exporter and first as an apparel exporter to the United
States.68
Textile Trade Policy and Agreements. In attempts to resolve conflicts
between the interests of exporters and importers, the United States has signed a
number of agreements (multilateral and bilateral) over the years generally restricting
the quantities of textiles and apparel imported, including the establishment of quotas.
The international Agreement on Textiles and Clothing (ATC) provided for the
phasing out of those quotas by January 1, 2005. It placed trade in textiles and apparel
under the rules governing other products, but with a provision allowing importing
countries to impose transitional safeguard mechanisms to protect against damaging
surges of imports of products not under quota and not yet integrated under World
68
The trade data used in this section are from the Dataweb database compiled by the U.S.
International Trade Commission from U.S. Departments of Commerce and Treasury data,
obtained stepwise October 2, 2003 from [http://dataweb.usitc.gov/scripts/INTRO.asp].
CRS-33
Trade Organization (WTO) rules. Developing countries, whose exports have been
limited, considered the phase-out procedure as unfair, and pressed for accelerated
implementation of the phase-out. They contended that the United States and other
developed countries delayed import liberalization.
The United States has entered into several bilateral trade agreements in recent
years. Because apparel and textile mill products account for 9% of exports by China
to the United States, the agreement with China, reached November 15, 1999,
probably is most important with respect to textiles and apparel. Among the wide
range of issues covered, it incorporated the 1997 textile and apparel agreement
between the two countries. Major elements of the 1999 agreement were (a) China,
upon accession to the WTO, would “catch up” to the ATC schedule of quota phaseouts by 2005 for other WTO members, but the United States retained the right to
impose safeguard measures through the end of 2008, allowing continuation of some
quotas under some conditions, and (b) China would significantly lower its tariffs on
a wide range of textile and apparel products, and not impose new nontariff barriers.
U.S. textile and apparel importers praised the agreement, especially regarding
the quick ending of quotas. U.S. textile manufacturers were disappointed that the
agreement did not continue the quotas on Chinese textiles and apparel for 10 years,
a phase-out duration faced by other WTO members; and the industry trade group
expressed concern over projected U.S. job and production losses.69 U.S. labor, as
represented by the AFL-CIO, criticized the agreement as failing to protect workers’
and human rights. P.L. 106-286 created mechanisms to monitor China’s compliance,
and authorized the President to grant China permanent normal trade relations (PNTR)
status after it joined the WTO. The President granted PNTR status to China on
December 27, 2001, after it officially joined the WTO on December 11, 2001.
Imports of textiles and apparel from China subsequently rose 13% between 2001
and 2002, and increased 29% in the first seven months of 2003 over the same period
in 2002. Increases in imports of three groups of textile and apparel products from
China since the full phase-out of import quotas in January 2005 have been so rapid
that the United States has put caps on imports from China of some of these items, as
permitted under the China WTO accession agreement.70 Analysis of the possible
effects of China’s accession to the WTO can be found in the section on trade
agreement impacts later in this report.71
Recent bilateral trade agreements with Vietnam and Singapore also have
negative significance for U.S. textile and apparel manufacturing. The July 2000
agreement with Vietnam was followed up in December 2001 with the U.S. granting
69
American Textile Manufacturers Institute, “Statement by Doug Allen, President,
Regarding the U.S.-China WTO Agreement,” Nov. 15, 1999. U.S. Association of Importers
of Textiles and Apparel, “Importer Association Hails U.S.-China Agreement on WTO
Accession,” Nov. 15, 1999.
70
Inside US Trade, “Commerce’s CITA Approves Three China Safeguard Provisions,”
November 18, 2003.
71
For more on U.S.-China trade relations in general and textile and apparel trade in
particular, see CRS Issue Brief IB91121, China-U.S. Trade Issues.
CRS-34
normalized trade status to Vietnam, conditional on annual review under terms of the
Jackson-Vanik amendment.72 Such status significantly cut U.S. tariffs on most
imports from Vietnam, leading to a very large increase in Vietnamese exports of
textiles and apparel to the United States. An April 2003 agreement put quotas on 38
categories of Vietnam’s clothing exports. The December 2002 trade agreement with
Singapore makes Singaporean exports of textiles and apparel to the United States
duty free if made from U.S. yarn or from materials further along the production chain
that originate in the United States or Singapore. A limited amount of apparel exports
from Singapore will be exempt for eight years, and tariffs on those exports will be
phased out over five years. The United States commits to more liberal rules of origin
once further liberalization of such rules is achieved in the WTO.
The U.S. Congress has made efforts to spur economic growth in poorer regions
of the world, to some extent by providing textile and apparel trade benefits. Among
the provisions of those measures, Congress has eased trade terms in stages on such
goods from Andean, Caribbean, and sub-Saharan region countries.73 The extent of
these trade benefits is constrained by concerns that growth of textile and apparel
production in the above regions has caused and could cause further difficulty for
segments of the U.S. textile and apparel industries. Thus, as noted above, many of
the trade preferences for textiles and apparel tend to require that U.S.-made fibers and
fabrics be used to produce the apparel that is made in the beneficiary countries.74
Globalization: Impact on U.S. Manufacturing
The Manufactures Trade Balance75
Today “globalization” is a well-known phenomenon, and has been for nearly a
generation. In the strictly economic sense, it means a focus on world markets as a
source of inputs and as a place to sell goods and services, and as a destination for
investment capital. In terms of direct impact on everyday lives, the world today is
more globalized, both in a broader and a deeper sense, than ever before. This can be
seen in terms of the amount of U.S. exports and imports expressed as a percent of
GDP. After two generations of protectionist policies worldwide, an international
economic depression and two world wars, U.S. imports and exports together equaled
less than 10% of U.S. GDP in 1960, and just a little more than that 10 years later.
72
The Jackson-Vanik provision of the Trade Act of 1974 is found at 19 USC 2432. For
details on U.S.-Vietnam trade relations, see CRS Issue Brief IB98033, The Vietnam-U.S.
Normalization Process, by Mark E. Manyin, esp. p. 5.
73
See also CRS Issue Brief IIB95050, Caribbean Basin Interim Trade Program:
CBI/NAFTA Parity; CRS Report RL30790, The Andean Trade Preference Act: Background
and Issues for Reauthorization; and CRS Report RS21772, AGOA III: Amendment to the
African Growth and Opportunity Act.
74
For discussion and description of rules of origin in U.S. trade preference programs and
free trade agreements, see CRS Report RL31934, Textile and Apparel Rules of Origin in
International Trade.
75
This subsection was written by Stephen Cooney.
CRS-35
Then the value of total trade compared to GDP doubled to more than 20% by 1980.
In 1990 it was still at the same ratio, but by 2000 it increased to 26%.76
Figure 8. U.S. Manufactured Exports and Imports
U.S. Manufactured Exports and Imports
1200
$ billons
Exports
Imports
1000
800
600
400
200
0
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004*
* January-October annual rate.
Manufactures trade from Standard Industrial Trade Classification 5-9, including manufactured re-exports. Imports at
customs value.
Sources: U.S. Dept. of Commerce. Bureau of the Census. Foreign Trade Division. 1989-2003data based on FT900: U.S.
International Trade in Goods and Services, annual reports; 2004 data, FT900, October 2004;1980-88 data from National
Association of Manufacturers, based on Census Bureau sources.
Manufactured Imports and Exports. Figure 8 illustrates the relative
growth of manufactured exports and imports since 1980. The overwhelming
majority of U.S. trade in physical goods is in manufactured goods, which therefore
account for most of the changes in the overall U.S. trade balance in physical goods.
In 2003, for example, manufactures accounted for 77% of all U.S. goods exports (by
comparison, agricultural commodities accounted for 8.5%); manufactures also
accounted for nearly 82% of all U.S. goods imports (by comparison, mineral fuels
accounted for 12%).77
In the early 1980s, as the figure shows, the United States still had a trade surplus
in manufactured goods, but that surplus was declining. The U.S. economy grew
76
Calculated from Economic Report of the President, 2003. Table B1. Addition of imports
and exports is used only to illustrate the relative importance of trade in the total economy.
In arriving at the calculation of total GDP, imports, that component of national consumption
sourced abroad, are subtracted. For a detailed analysis of the macroeconomic aspects of the
overall U.S. trade deficit and the deficit on the balance of payments on current account, see
CRS Report RL31032, The U.S. Trade Deficit: Causes, Consequences and Cures, by Craig
Elwell; and, CRS Report RL30534, America’s Growing Current Account Deficit: Its Cause
and What It Means for the Economy, by Marc Labonte and Gail E. Makinen.
77
U.S. Dept. of Commerce. Bureau of the Census. Foreign Trade Division. FT900: U.S.
International Trade in Goods and Services, 2002 Annual Report, Exhibit 14.
CRS-36
strongly after a recessionary period in 1980-82, the value of the dollar was rising,
discouraging U.S. exports, and foreign markets, notably in Europe, were struggling
with the effects of “stagflation.” By 1983, the surplus turned into a $24 billion
deficit. The deficit rose to more than $100 billion in 1986-88. It then declined to
less than $100 billion for five years from 1989 through 1993, probably due to a
reversal of exchange rate and growth trends, as well as the effect of a recession and
a slow recovery in the early 1990s on U.S. domestic demand. Exports grew 8.8% per
year during this period, and increased by $133 billion; imports of manufactures grew
by $119 billion, or only 5.8% annually. Both export and import growth accelerated
in 1993-7: manufactured exports grew by more than $200 billion, or more than 11%
per year; but imports grew by $250 billion, or more than 13% annually.
Manufactured exports stopped growing after 1997. In 2002 and 2003 data,
exports (including re-exports) were a little more than $600 billion, not substantially
higher than the 1997 level of $593 billion. During the intervening years, there was
only one strong export growth year: 2000 with a 12.5% rise to nearly $700 billion,
but then exports fell by more than $50 billion the next year.78 By contrast,
manufactured imports continued to increase strongly, by nearly $300 billion, to more
than $1 trillion in 2000. Manufactured imports then fell because of slower U.S.
growth, but not as fast as exports. The manufactures trade deficit was more than $300
billion in each year after 1999, and reached $400 billion in 2003, an amount greater
than one-quarter of estimated U.S. manufacturing output of $1.4 trillion.
Figure 9 illustrates the U.S. industrial goods trade balances according to “enduse” sectors. This is a different data set, so the data include some non-manufactured
products, especially in industrial supplies and materials (though petroleum products
have been subtracted in this calculation). This figure is not meant to imply that U.S.
trade should be balanced in all sectors, but rather to indicate relative U.S.
comparative advantage, and where the deficit has increased most sharply.
The figure shows that out of a total U.S. trade deficit of $548 billion in 2003,
almost $380 billion was due to the deficits in consumer goods and automotive
vehicles — and these deficits showed the strongest increases during the high-growth
1990s (a $270 billion combined negative change since 1992). While many complain
about the large quantities of imported consumer goods in U.S. retail stores, the fact
is that the United States has long had a deficit in manufactured consumer goods.
Similarly, despite the rapid growth in transplant automotive manufacturing
establishments discussed earlier, the automotive trade deficit increased by $84 billion
between 1992 and 2003.
Capital goods — products used for the production of other goods and services
— long the sector where the U.S. has had its greatest comparative advantage —
turned negative by $2 billion in 2003, falling from a $41 billion surplus earlier. Non78
Michael Mandel calculates that if total U.S. exports had grown by 18%after the 2001
recession, as they did in the equivalent time period after the recession in 1990-91, instead
of the actual levels, total U.S. exports today would be $170 billion higher than they are and
would have created a conservatively estimated 850,000 additional U.S. jobs; in “So Where
Are the Jobs?: They’re on the Way ... Or Maybe Not,” Business Week (January 26, 2004),
pp. 39-40.
CRS-37
oil industrial supplies and materials, used as inputs by U.S. manufacturers, showed
a relatively smaller change in the 1990s, from a 1992 surplus of $15 billion to a slight
deficit in 2003.
Figure 9. U.S. Trade Balances by “End-Use” Sectors
Foreign Outsourcing. In view of the negative movement in the U.S. trade
balance and the loss of domestic manufacturing jobs, foreign outsourcing of jobs has
become a major issue. The loss of jobs in product groups where the United States
may have lost much of its comparative advantage has long been understood, though
not happily accepted (see the textiles and apparel discussion earlier, for example).
But alarm has increased over apparent losses of jobs related to high-technology
industries and services, for example, software programming and call centers. Two
private research firms have reportedly estimated that large numbers of such servicesector jobs will move overseas: Forrester Research reportedly estimated that 400,000
such job displacements would occur in 2003, and Gartner Dataquest estimated that
500,000 IT vendor and services jobs would go by the end of 2004.79 According to
a report in the Toronto Globe and Mail on a McKinsey Global Institute report,
Ireland was the largest country for U.S. IT and business-process outsourcing in 2002
79
Steven Greenhouse, “IBM Explores Shift of Some Jobs Overseas,”New York Times, July
22, 2003, p. C1; Eric Auchard, “One in Ten Technology Jobs May Move Overseas,”Chicago
Tribune, August 3, 2003, p. 5.
CRS-38
($8.3 billion), followed closely by India at $7.7 billion. Canada ranked a distant third
with $3.7 billion.80
Official estimates of overseas investment by U.S. multinational companies show
some increase in overseas commitments, but do not support contentions of largescale or rapid growth in overseas outsourcing in recent years. For example, the
number of persons directly employed at all foreign affiliates in which U.S. parents
have a direct investment interest (greater than 10%) was stable at around 6.6 million
persons between 1982 and 1993, according to Commerce Department data. That
number had increased more than 40%, or almost 3 million employees, by 2000, the
latest date for which such data are available. But the total employment reported by
U.S. multinational parent firms remained concentrated in the United States.
Employment at foreign affiliates of such firms rose only from 26% of their total
employment in 1982, to 27.5% in 1993, to 29.3% by 2000. However, this number
does not include overseas outsourcing by “contract manufacturers.” Such firms as
Solectron and Flextronics have become important manufacturers in the electronic
businesses, while the 1990s also saw the rise of semiconductor manufacturing
“foundries,” such as Taiwan Semiconductor Manufacturing Company and United
Microelectronics Corporation, and Singapore-based Chartered Semiconductor. These
are all now multibillion-dollar enterprises.81
Nor does an examination of capital outflow numbers (equity increases and
reinvested earnings) appear to indicate any recent surge of direct investment by U.S.
manufacturers in Asia, in comparison with other regions. Total U.S. direct
investment capital outflow in 2002 was $120 billion, little more than half the level
of $209 billion in 1999. Investment in manufacturing affiliates was also down, from
$40 billion to $30 billion. This was especially true in computers, electronics and
components, where overseas investment by U.S. parents declined from $11 billion
in 1999 and $17 billion in 2000, to about $1 billion in 2002, although this may well
reflect that domestic IT manufacturers had a low level of earnings to invest
anywhere. Overall, between 1999 and 2002, U.S. companies’ cumulative direct
investment in manufacturing affiliates increased $37 billion in Europe (22%), $20
billion in Canada (40%), and $15 billion in Asia and the Pacific (36%), while falling
$5.5 billion (11%) in Latin America.82
Nevertheless, U.S. executives have stressed the attractiveness of Asian business
locations for future investment and growth. Intel CEO Craig Barrett, recently
opening a new semiconductor test facility in Sichuan, China, said:
Most countries other than the [United States] provide incentives for the future ...
People in these countries are fully capable of doing any engineering job a U.S.
employee can do. New talent in these workforces is going to have a massive
80
David Ticoll, “Wise Up about Offshore Outsourcing,” Toronto Globe and Mail, October
16, 2003.
81
On the reviving boom in chip foundries, see Business Week, “A Chip Boom? In Asia, at
Least,” (Nov. 3, 2003).
82
BEA “U.S. Direct Investment Abroad,” Survey of Current Business (Sept. 2003), by
Jeffrey H. Lowe, Tables 5 and 11.
CRS-39
effect on where jobs are created ... Look at how Taiwan affected the global
electronics industry, and then imagine that on ten times the scale.83
In a later interview, when he indicated that Intel is unlikely to increase hiring or
to expand plant capacity in California, Barrett said that global competition and the
maturing U.S. electronics market now mean that 70% of Intel’s markets lie outside
the United States; “Our investments are really following our customers,” he
concluded.84
But whether the location of new manufacturing jobs is driven by lower
production costs or the need to be close to the customer base, it is clear that the recent
U.S. experience is not unique: manufacturing employment is declining almost
everywhere as productivity improves. Moreover, it is especially occurring in China,
where the inefficient state-owned economic sector is being opened to domestic and
international competition, resulting in large-scale job losses. According to a report
by Alliance Capital economist Joseph G. Carson, manufacturing employment in
China between 1995 and 2002 fell from 98 million to 83 million, a loss of 15 million
jobs. Thus, China lost more manufacturing jobs during the past seven years than the
total number of manufacturing jobs left in the United States.85 Over the same period,
the other largest manufacturing nations lost a total of 7 million manufacturing jobs.
Of the 20 largest economies reported in Carson’s analysis, only five showed any
gains in manufacturing employment (Canada, Mexico, Spain, Taiwan and the
Philippines). But their gains were only marginal and hardly offset the net total loss
of 22 million manufacturing jobs among all major industrial countries, including
China, during the period — a decline of 11% in employment, while output increased
30%.86
The Commerce Department’s 2004 report, Manufacturing in America, indicated
that the trend in foreign outsourcing, such as it is, may be one aspect of “structural
changes shaping the competitive environment” of U.S. manufacturing companies.
It ascribes these changes to three fundamental trends. First, there has been a
technological revolution in manufacturing resulting from improvements in
computing, communications and distribution. Moreover, this revolution has been
global, and not just restricted to U.S.-based companies. The Commerce report notes,
for example, that in the 1960s, 60-70% of global research and development activities
occurred in the United States and were largely funded by the federal government.
Now the U.S. private sector finances twice the level of R&D of that financed by the
government, and the U.S. share of the global total is only 30%.
Second, as mentioned earlier in the present report, successive reductions of tariff
and non-tariff barriers through international trade negotiations have substantially
83
Quoted from Xinhua Financial Network. “Global Job Shifts Toward Asia Threaten Future
US Employment — Intel CEO” (Aug. 28, 2003).
84
Reuters. “Intel CEO Says California Has Lost Its Luster,” (October 21, 2003).
85
Some would observe that comparing the situation of state-owned enterprises in China,
many of which were never viable on a market basis, to the situation of American
manufacturers is a major simplification.
86
Joseph G. Carson. “AllianceBernstein US Weekly Economic Update,” October 10, 2003.
CRS-40
increased the role and impact of international trade in the U.S. economy. And, third,
the end of the Cold War and important policy changes in many major countries have
led to an emergence of new participants in the international economy, notably China,
as indicated just above, and countries formerly comprising the Soviet Union and its
bloc in eastern Europe — participants which are also international competitors. In
such an environment, the Commerce Department argues, competition is increasingly
not horizontal competition between individual companies, but rather competition
between different vertically integrated supply chains.87
The Dollar Exchange Rate and U.S. Manufacturing88
With the slowdown in U.S. manufacturing growth at the end of the 1990s, the
impact of a relatively high value of the dollar in foreign exchange again became a
major concern for many. This section of the paper examines the case for a causal
linkage between exchange rates and manufacturing employment. While, in principle,
exchange rate variations could affect both a nation’s imports and exports, and hence,
employment, the empirical case is mixed. First, neither in the short or the long term
are exchange rates the singular influence on manufacturing employment. Swings in
the business cycle account for significant short term employment variation, while in
the long term the effects of productivity growth exert an important influence.
Second, data from 1980 to 2005 show periods where the variation manufacturing
employment is consistently predicted by changes in the exchange rate, but also
periods where the two move in an inconsistent manner. Finally, expert opinion
reflects the mixed state of empirical evidence. For example, while some observers
believe an upward revaluation of the Chinese yuan would provide a significant boost
to manufacturing employment in the U.S. by increasing the competitiveness of our
exports and reducing our imports, others believe the effect would be minimal, merely
increasing our imports from countries other than China, with little effect on
employment or our aggregate trade balance.
Market-Based Exchange Rate Systems. Since the early 1970s, the value
of the U.S. dollar in terms of other currencies has been determined through a system
of flexible exchange rates. In this system, governments typically do not control the
value of their currencies as a matter of policy, but let the market forces of supply and
demand determine the value. In principle, the value of a nation’s currency is directly
related to the balance of trade, as well as flows of international capital into and out
of the country. By affecting relative costs, and hence sales, of both imported and
exported goods, the exchange rate may also affect both sectoral and general
macroeconomic activity. As a result, macroeconomic variables such as employment,
inflation, and the rate of economic growth may all be influenced by the exchange
rate.89
87
Manufacturing in America, pp. 22-30.
88
This subsection was written by Robert Pirog.
89
The U.S. dollar plays a key role in the world monetary system, because it serves as an
international means of payment, as well as being the standard to which some nations peg
their exchange rate, if they are not market-determined. Since the dollar serves these
functions, a variety of benefits, costs and uncertainties are created.
CRS-41
The degree to which exchange rate variations affect sectors of the economy, like
manufacturing, or even the overall economy, depends on how open that sector, or
economy, is to world markets. The openness of an economy depends on the level of
tariffs, capital controls, and the degree to which competitive products are produced
in other nations.
Researchers measure the degree of openness by adding an industry’s exports and
imports, and expressing the sum as a proportion of the industry’s domestic sales,
exports and imports. By this measure, the openness of the U.S. manufacturing sector
has grown sharply since the early 1970s. The growth rate of openness in U.S.
manufacturing has averaged over 5% per year from 1972 to 2003. The growth has
not been uniform, however. Industries that were the most open to the world economy
in 1972 have shown the highest growth of openness. This heterogeneous pattern
implies that some sectors of the manufacturing industry are experiencing substantial
international competition, while other sectors might remain relatively isolated.90
On a bilateral basis, internationally traded currencies bear a symmetrical
relationship to one another. When one currency is appreciating, this is equivalent to
saying the other currency is weakening, and vice versa. For example, if the dollar is
being compared to the euro, a declining value of the dollar is equivalent to an
increasing value of the euro. Exchange rates can only be expressed relative to
another currency. However, for many purposes it is more useful to express the value
of a currency relative to the value of a basket of currencies of the nation’s important
trading partners. In this case, the value of a particular currency rises or falls against
the weighted average of the nation’s trading partners, with the same symmetrical
value relationship holding.
When the value of a currency is market determined, the event that initiates a
change in the exchange rate is usually a change in the demand and/or supply of
assets(the international capital market), or a change in the demand and/or supply of
goods and services in international trade. As a nation’s currency is rising in value,
or appreciating, the nation will typically move toward a trade deficit, as imports
become cheaper and exports more expensive, as well as experiencing a net inflow
of foreign capital. When the nation’s currency is falling in value, or depreciating, the
nation will typically move toward a trade surplus and experience a net outflow of
capital.
Is it then better for a nation to have a currency whose value is rising or falling,
high or low? The answer depends on where you stand in the economic system. A
high, or appreciating, dollar is good for U.S. buyers of imported goods, many of
which seem cheap in comparison to domestic goods. The resulting inflow of lower
price imported goods also keeps domestic inflation low. The same high, or
appreciating, dollar is bad if you are the owner of, or a worker in, a firm that is in an
import competing or export oriented industry that is losing sales, profits and jobs to
foreign competition. A low, or depreciating, dollar should reverse the relative
90
Michael W. Klein, Scott Schuh, and Robert K. Triest, “Job Creation, Job Destruction, and
the Real Exchange Rate,” Journal of International Economics LIX:2 (March 2003), pp. 245248.
CRS-42
positions of consumers and workers. In reality, these relationships are neither so
clear, or simple. A manufacturer who competes in the export markets may rely on
imported production inputs or components that rise in cost as the dollar falls.
The Dollar, the Trade Balance and Employment in Manufacturing.
Because the dollar exchange rate is determined in an open currency exchange
market, its value can be expected to change over time, reflecting, among other
things, changing conditions of demand and supply. Changes in the value of the dollar
are typically measured in terms of a trade-weighted index, generally including a set
of nations that are important industrial trading partners whose currencies are widely
traded on international markets.91 Figure 10 shows the movement of the nominal
major currencies dollar index from 1980 to July 2005.
Figure 10. Dollar Index: Price-Adjusted Major Currencies
140
130
120
Index*
110
100
90
80
70
60
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2005
*March 1973=100
Source: Board of Governors, Federal Reserve System, Monthly Price-adjusted Major
Currencies Dollar Index Series (G-5).
On the basis of this index, the dollar reached its highest level in the history of
the data set in March 1985. The lowest level was attained ten years later in April
91
Economists also use real exchange rate indexes that control for the relative differential
inflation rates in trading nations. The Federal Reserve real index of the dollar exchange rate
against major foreign currencies has generally paralleled the nominal index for the past 20
years, as most major trading nations have set a priority on bringing down rates of inflation.
Over the period covered in this report, the nominal and real major currencies indexes convey
a consistent picture of changes in the value of the dollar. The Federal Reserve reports the
value of both indexes on a monthly basis.
The data are available at
[http://www.federalreserve.gov/releases/H10/Summary].
CRS-43
1995. From that low point, the dollar rose by approximately 50%, achieving a peak
value in February of 2002. By the end of 2004, much of that appreciation had been
reversed, leaving the dollar less than 10% higher than at the low point in 1995.
During the first six months of 2005 the dollar appreciated by about 5% overall.
During the period January 1981 until the dollar peak of March 1985, the dollar
appreciated by approximately 47%. Manufacturing employment declined by about
4%. During the same period, the goods balance of payments moved from a $28
billion deficit to a $120 billion deficit, over a fourfold increase. Goods exports fell
by about 2%, goods imports rose by 32%.92
The ten year period of general decline of the dollar after the peak of March 1985
until the low point of April 1995 resulted in a dollar depreciation of about 45%.
Manufacturing employment again declined by approximately 4%. The goods balance
of payments varied over the period, deteriorating from 1985 to 1987 and then
improving from 1987 to 1992. Large, increasing, deficits characterized 1993-95,
reaching $174 billion in 1995. Goods exports increased by more than 160% over the
period, but goods imports rose by 121%, from an initial base of imports which were
more than 50% larger than exports in 1985.
Although the time-frames are different for the 1980-85 period of dollar
appreciation and the 1985-95 dollar depreciation, the magnitudes of dollar variation
are similar. Also similar is the fact that manufacturing employment declined in both
circumstances, albeit more slowly in the period of the falling dollar. On the basis of
this evidence, it would be hard to make the case that the exchange rate had the
theoretically predicted effect on manufacturing exports, production and employment
over these periods, although it is possible that a falling dollar exchange rate and
higher exports contributed to the modest increase in manufacturing employment in
the early 1990s. It is also possible that the effects of rising productivity, shifting
demand patterns and international capital mobility more than cancelled the effects of
exchange rate changes during the 1990s.
The dollar again appreciated from April 1995 until it peaked in February of
2002, increasing in value by about 40%. Manufacturing employment declined by
approximately 10% over the period. The goods balance of payments deficit grew
sharply over this period, increasing from $174 billion in 1995 to $482 billion in
2002. Goods exports increased from 1995 to 2000 by about 34%, but then declined
in both 2001 and 2002, falling by about 12% in those two years. Goods imports
increased by 55% over the period.
The period from January 2002 to November 2003 showed the overall value of
the dollar depreciating by about 17%. Manufacturing employment declined by an
additional 7% or 1.13 million jobs over the period. In at least one case, potential jobs
were lost because the value of the dollar declined. As reported in the New York
Times, Daimler-Chrysler announced the cancellation of a $750 million investment
92
Import, export and trade balance data are from the U.S. Census Bureau, Foreign Trade
Division website, [http://www.census.gov/foreign-trade]. Goods trade is dominated by
manufactured goods which in 2002 represented 80% of U.S. goods exports and 84% of U.S.
goods imports.
CRS-44
in Georgia which might have generated 3000 jobs, at least in part because of dollar
depreciation which made the importation of components from Europe too expensive,
reflecting the point made earlier about the effect of a depreciating dollar on imported
production imports.93
While the exchange rate may be an important factor in determining the
competitiveness of U.S. manufacturing, it is hardly the only factor. Economic
recession, changes in productivity, changes in technology, as well as structural
changes in domestic and international business all play a role. Comparing Figures
1 and 10 shows that there are periods when the linkage between exchange rates and
employment appears consistent with exchange rate theory. For example, the decline
of the dollar after the peak attained in March 1985 is followed by a period in which
manufacturing employment stabilized and increased. However, the most recent data
seem, at this point, inconsistent with the theory. The decline in the dollar since
February 2002 has been accompanied by a sharp decline in manufacturing
employment, consistent with the domestic economic slowdown. Moreover, the data
on manufacturing employment as a share of total employment, shown in Figure 2,
suggest that manufacturing’s share of all employment has declined both in periods
when the dollar was appreciating and when it was depreciating. Although the speed
of the decline in employment might be affected by the exchange rate, its overall
trend has been downward since the 1960s.
Several research efforts have examined the forces influencing manufacturing
employment and have reached mixed conclusions. John A. Tatom, of the Federal
Reserve Bank of St. Louis, found little relationship between the exchange rate and
manufacturing employment. In fact, he found that in the 1980’s manufacturing
output expanded more during periods of a strengthening dollar than during periods
when the dollar was weakening. For Tatom, the principal factors in the decline in
manufacturing employment are productivity improvements and the sensitivity of
demand for manufactured goods during declines in over-all economic activity.
Productivity improvements tend to reduce the relative price of manufactured goods
and require fewer workers for any given level of output. The enhanced productivity
of workers suggests that firms should expand employment, while the decreasing
relative price of manufactured goods suggests that firms should reduce employment.
A fall in the relative price of manufactured goods raises the cost of manufacturing
labor relative to its productivity — any given wage is set against lower revenue due
to decreased price. In effect, a decline in the relative price of manufactured goods is
equivalent to an increase in the manufacturing wage rate relative to nonmanufacturing labor, which reduces the incentive to hire labor. As per the next
paragraph, the falling price of manufactured goods relative to other goods and
services stimulates demand for manufactured goods, but this increase must overcome
the wage/productivity effect for more workers to be hired.
For the productivity effect to offset the price effect, the demand for
manufactured goods should expand more than proportionately to a given fall in price.
93
“DaimlerChrysler Drops Plan for Van Plant,” New York Times, (Sept. 24, 2003), p. C16.
CRS-45
Tatom finds that this is not the case, and the price responsiveness (elasticity) of
manufactured goods demand is low.94
Michael Klein, Scott Schuh and Robert Triest of the Federal Reserve Bank of
Boston examined the relationship between job creation, job destruction and the real
exchange rate in the manufacturing industries. They found, using econometric
analysis applied to a data set from 1973-1993, a number of important relationships
between the exchange rate and manufacturing employment. First, the authors found
that the sensitivity of job destruction to the exchange rate depends on the openness
of that particular industry to trade, either in exports or in import competition. This
result is expected. Second, they found that while job destruction is related to the
exchange rate, job creation is significantly less related. This asymmetric relationship
implies that when the dollar appreciates the rate of job loss is accelerated, but when
the dollar depreciates, jobs are not necessarily created. The rate of manufacturing job
destruction merely returns to its historic trend. This relationship appears to be
continuing in more recent data where it consistently appears that higher values of the
dollar may contribute to job losses, but a decline in the dollar rarely reverses the
situation.95
If the Klein, Schuh and Triest results are an accurate description of the forces
facing the manufacturing industry, policy based on realigning exchange rates can be
expected to have only limited effectiveness. From this perspective, an appreciation
of the Chinese yuan, or the other Asian currencies against the dollar might slow the
rate of job loss in American manufacturing, but would be unlikely to cause net job
creation. A different perspective is provided by C. Fred Bergsten of the Institute for
International Economics. He feels that an appreciation of the Chinese yuan on the
order of 20-25% would also encourage a number of other East Asian countries to
allow their currencies to appreciate. Such a regional currency appreciation could lead
to the creation of as many as 500,000 U.S. manufacturing jobs, in Bergsten’s view.96
Asian Economies and the Dollar. Considerable concern has been
expressed that many Asian countries have been deliberately keeping their currencies
undervalued against the U.S. dollar, in order to promote their exports. This view has
been emphasized by the Coalition for a Sound Dollar, an advocacy group made up
of more than 80 trade associations, who believe the strong dollar is still damaging
American manufacturing.97 They feel that the real problem lies mainly with four
nations, China, Japan, Korea and Taiwan, who, the Coalition charges, have taken
actions to keep the values of their currencies artificially low to gain an export
advantage, or to impair or negate trade concessions. The Coalition asserts that such
policies could be in violation of World Trade Organization and International
Monetary Fund rules.
94
John A. Tatom, “Why Has Manufacturing Employment Decreased?” Federal Reserve
Bank of St. Louis Review, December 1986, pp. 15-25.
95
Klein, Schuh, and Triest, pp. 255-259.
96
C. Fred Bergsten, comments on PBS Newshour with Jim Lehrer (Sept. 3, 2003).
97
Coalition for a Sound Dollar, The Overvalued Dollar — Six Years Later, available at the
website [http://www.ssci.org/images/mfg-report.pdf].
CRS-46
As evidence for their charges, the Coalition cited seven instances of currency
intervention in 2002 by Japan, for a total of $33 billion in exchange markets, that
were designed to keep the value of the dollar above ¥115. They cite similar
interventions, designed to weaken the yen in 2001, totaling $28 billion. The
Coalition also cites numerous threats of intervention by Japanese government
officials lauding a weaker yen and threatening currency market intervention. In
March 2004, it was reported, however, that Japan had ended its campaign to weaken
the yen against the dollar. A recovering domestic economy and a strong outlook for
exports were cited as reasons for the policy shift.98
For Taiwan and Korea similar interventions are described, with these nations
buying $60 billion to weaken their currencies against the dollar. But other observers
do not attribute such currency actions to a desire to manipulate exchange rates in
order to increase exports. They have characterized the Japanese intervention and
other measures merely as a means to manage the decline of the dollar to prevent it
from falling too quickly, which could have damaging effects on the Japanese
economy.99
Until very recently, and then only with respect to the Japanese yen, there has
been very little movement of the dollar against the currencies of Japan, Taiwan, and
South Korea. Over the period August 2002 to August 2003 the dollar declined by
almost 9% as measured by the major currencies index. Over the same period, the yen
itself changed little against the dollar. In August 2002, the yen traded at 118.99 to
the dollar and in August 2003 it traded at 118.66 to the dollar, a variation of less than
0.3%. The dollar weakened by 1.5% against the Korean won and appreciated by 1%
against the Taiwan dollar over the same period.
During the period August 2003 to November 2003 the dollar declined by 6%
against the major currencies index. The dollar also declined, by 8%, against the yen.
The dollar has stayed much steadier against the other two East Asian currencies: from
August 2003 to November 2003 it appreciated by 0.6% against the won and
depreciated by 1% against the Taiwan dollar.
Problems exist in establishing clear causality between these currency
interventions and fluctuations in the exchange rate. In 2001, foreign exchange
transactions were $1.4 trillion per day on world currency markets.100 Total reserves
held in central banks worldwide were $1.7 trillion. Over 90% of foreign exchange
market transactions are carried out by foreign exchange dealers and financial
institutions. Currency interventions on the scale of those cited by the Coalition for
a Sound Dollar by Japan and other nations may not have been sufficient to stabilize
the value of the yen and other currencies at a value that was far from a market
98
Leo Lewis, Richard L. Perry and Robert Thomson, “Japan Ends Its $150 Billion Currency
Intervention as Economy Firms,” Times Online (U.K.), Mar. 29, 2004.
99
Christian E. Weller and Laura Singleton, Reining in Exchange Rates: A Better Way to
Stabilize the Global Economy, Economic Policy Institute, Briefing Paper #131, Sept. 2002,
p.5.
100
The amount of currency that actually changes hand on a settlement basis is a fraction of
this total.
CRS-47
equilibrium. An unmeasurable influence is the degree to which the Japanese actions
signaled the market, altering the actual market outcome.
The case of China has been different from that of Japan and many other Asian
exporters. China maintained a fixed exchange rate of 8.28 yuan per dollar from 1994
to July 2005. The value of the yuan did not float against the dollar and China
maintained controls on capital. China has a surplus in its trade balance with the
United States, the U.S. deficit totaling $72 billion in the first five months of 2005,101
and holds foreign exchange reserves in excess of $300 billion. Many observers agree
that the yuan has been undervalued, which has contributed to the trade imbalance and
the resulting buildup of financial surpluses in China. The extent to which the yuan
continues to be undervalued is open to question. The Economist magazine estimates
the level at 56%, while an economist at the Manufacturers Alliance trade association
estimates 40%, and other analysts suggest a level of 10-15%.102 Secretary of the
Treasury John Snow undertook a trip to China in September of 2003 to discuss the
exchange rate imbalance with Chinese leadership. The Chinese offered to allow the
yuan’s value to float in the market at some time in the future, but made no definite
commitment. In July 2005, the Chinese government announced that the yuan would
no longer be linked to the dollar. Instead, it would be tied to a basket of currencies.
At the same time, they allowed a 2% increase in the dollar value of the yuan.103
The Chinese may have good reasons not to be eager to jump into the world of
market determined exchange rates. The fixed value of the yuan was looked at as an
important pillar of stability during the Asian financial crisis of 1997-98. China’s
current account trade surplus is based largely on U.S. trade, and its surplus with the
rest of the world has been declining. With capital constraints in place it is difficult
to assess the true value of the yuan. If capital were allowed to flow freely, it is not
inconceivable that capital might flow out of China on a net basis which would tend
to push the value of the yuan lower. The fixed exchange rate policy of China reflects
its desire for stability. Until the economic structure of China is judged to be capable
of withstanding the volatility of world markets, the likelihood of a change in China’s
exchange rate policy may be low.104
On October 30, 2003 Treasury Secretary John Snow presented the Treasury
Department view on this issue, when he discussed the semiannual report to Congress
on exchange rates and international economic issues. The report’s purpose is partly
to evaluate the exchange rate policies of many trading nations. The report, which
reviewed developments in the first half of 2003, concluded that no nations, including
China and Japan, were manipulating the value of their currencies to obtain unfair
trade advantage. Secretary Snow acknowledged that, while China continued to peg
its currency to the dollar and Japan had been intervening in currency markets, those
101
FT900 (July 2005), Exhibit 14.
102
“Flying on One Engine: A Survey of the World Economy,” The Economist, Sept. 20,
2003, p.25.
103
James T. Areddy et al., “China Lets Yuan Rise vs. Dollar, Easing Trade Tensions
Slightly,” Wall St. Journal (Jul. 22, 2005), p. A1.
104
For more detail, see CRS Report RS21625, China’s Currency Peg: Implications for the
U.S. and Chinese Economies, by Wayne M. Morrison and Marc Labonte.
CRS-48
actions in themselves did not meet the technical requirements under the Omnibus
Trade and Competitiveness Act of 1988 to support a finding of unfair currency
manipulation.105
Secretary Snow reiterated the Bush Administration’s belief that currencies of
major nations, especially China, should be market determined. He also pointed out
that the Administration believes that bilateral financial diplomacy was the course of
action most likely to result in movement toward the goal of market- determined
currency values.106 However, Secretary Snow observed in May 2005 that if China did
not alter its policies, they would likely meet the statutory technical requirement for
designation as manipulating their currency to gain a trade advantage. China’s
alteration of its exchange rate mechanism was likely influenced by pressure from
both the Bush Administration and Congress.107
Federal Reserve Board Chairman Alan Greenspan pointed out in a 2003 speech
that if, as most observers seem to agree, the value of the Chinese yuan is out of
alignment with market valuation, this might very well have adverse effects on world
capital markets and, at least indirectly, affect U.S. jobs and employment. He did not,
however, see a revaluation of the yuan as having much effect on aggregate U.S.
employment. He stated that if currency realignment reduced imports from China,
they would tend to be replaced by imports from other nations that compete with
China. He further noted that a rapid exposure of the yuan to market forces would
have uncertain consequences for China’s fragile banking system as well as providing
no guarantee that the yuan would actually rise in value. He also stated that the
potential long term benefits to the world trading system of a prosperous China are
important enough that a careful approach to currency valuation and market
determined capital flows should be followed.108
The Impact of U.S. Trade Agreements109
The United States has been actively pursuing trade agreements over the last two
decades. It has had free trade agreements with Canada and Israel since the 1980s.
In 1994, it implemented the North America Free Trade Agreement (NAFTA) with
Mexico and Canada. Since then, it has concluded several agreements and is actively
pursuing others. Free trade agreements (FTAs) raise some important policy issues
for Congress as it considers the implementing legislation, including the impact on
U.S. manufacturing.
105
U.S. Dept. of the Treasury. “Testimony of Treasury Secretary John Snow Before the
Senate Committee on Banking, Housing and Urban Affairs” (Oct. 30, 2003), pp. 1, 3-4; and,
Report to Congress on International Economic and Exchange Rate Policies, JS-954 (Oct.
30, 2003), pp. 1, 6-7.
106
Snow, “Testimony,” p. 4.
107
U.S. Dept. of the Treasury. Report to Congress on International Economic and Exchange
Rate Policies, JS-2448 (May 2005), p. 2.
108
Federal Reserve Board of Governors. “Remarks by Chairman Alan Greenspan Before the
World Affairs Council of Greater Dallas,” (Dallas, TX, December 11, 2003).
109
This section was written by M. Angeles Villarreal and Stephen Cooney.
CRS-49
In 2000, the United States completed negotiations with Jordan for a bilateral
free trade agreement, which went into effect in September 2001. Subsequently, the
United States completed free trade agreements with Chile, Singapore, Morocco and
Australia, which were approved by Congress and signed into law.110 The Bush
Administration has also submitted to the 109th Congress an FTA that it has negotiated
with five Central American nations and the Dominican Republic.111 The broadest
initiative has been the multilateral trade negotiations in the World Trade
Organization (WTO), known as the Doha Round. Another major multilateral
initiative is the Free Trade Area of the Americas (FTAA), in which the United States
is continuing trade negotiations with 33 other Western Hemispheric countries. Other
initiatives include FTAs with Thailand, South Korea, and the Southern Africa
Customs Union.
Effects of NAFTA on US. Industry. In the early 1990s, the United States
entered into trade negotiations with Mexico, which later included Canada, to form
the North American Free Trade Agreement, effective in January 1994. NAFTA is
the first major trade agreement the United States has had with a developing country
where per capita income is much lower than in the United States. The difference in
income levels raised concerns that Mexico’s lower wages would lead to a large
number of U.S. jobs being relocated to Mexico as industries adjusted to the changes
in trade and investment regulations. While the integration of the U.S., Mexican, and
Canadian economies was expected to make the U.S. economy more productive and
globally competitive, the adjustment costs were expected to be more concentrated in
communities with a large number of manufacturing jobs, such as the automotive,
textile, and apparel industries.
The full effects of NAFTA on the U.S. economy are still unclear. Proponents
of NAFTA claim that the agreement has increased U.S. trade with Mexico and
Canada, and benefitted the U.S. economy. They believe that NAFTA has had a
positive impact on U.S. trade and investment with Canada and Mexico, and that
NAFTA has increased U.S. exports to Canada and Mexico.112 Critics of NAFTA
argue that hundreds of thousands of U.S. jobs have been lost because of the
agreement. NAFTA critics generally base their arguments on the increasing trade
deficit with Mexico, stating that increasing U.S. imports from Mexico have caused
plant closures and job losses in the United States.
There are a number of reasons why the overall effects of NAFTA are not easily
measured. First, it is difficult to isolate the effects of NAFTA because of other
110
For details on the agreements see CRS Report RL31789, U.S. Singapore Free Trade
Agreement; CRS Report RL31144, U.S.-Chile Free Trade Agreement: Economic and Trade
Policy Issues; CRS Report RS21464, Morocco-U.S. Free Trade Agreement; and CRS Report
RL32375, The U.S.-Australia Free Trade Agreement: Provisions and Implications. For
issues related to labor see CRS Report RS21560, Free Trade Agreements with Singapore
and Chile: Labor Issues, by Mary Jane Bolle.
111
See CRS Report RL31870, The Dominican Republic-Central America-U.S. Free Trade
Agreement; and CRS Report RS22159, DR-CAFTA Labor Rights Issues.
112
See Council of the Americas and the U.S. Council of the Mexico-U.S. Business
Committee, NAFTA at Five Years, prepared by the Trade Partnership, Washington, D.C.,
January 1999.
CRS-50
variables affecting trade and investment such as economic growth and exchange
rates, both of which affect consumer spending and the demand for imports, foreign
investment, employment levels, and relative wages. Second, some of the market
opening measures in Mexico that resulted from NAFTA were already taking place
prior to the agreement and NAFTA may have only accelerated the process. While
trade expansion arguably has benefitted the overall U.S. economy in terms of
improved production processes, and the increased availability of goods and services
for U.S. consumers at lower cost, there also have been job losses associated with
NAFTA. Two of the U.S. workforce sectors that have been most affected are the
textiles and apparel industry, and the automotive industry.
The overall effect of NAFTA on the U.S. economy has been relatively small,
primarily because two-way trade with Mexico amounts to less than three percent of
U.S. GDP. Therefore, any changes in trade patterns with Mexico would not be
significant in relation to the overall U.S. economy. In some sectors, however, traderelated effects could be expected to be more significant, especially in those industries
that were more exposed to the removal of tariff and non-tariff trade barriers, such as
the textile, apparel and automotive industries. Most of the trade-related effects of
NAFTA may be attributed to changes in U.S. trade and investment patterns with
Mexico. At the time of NAFTA implementation, the U.S.-Canada Free Trade
Agreement already had been in effect for five years and some industries in the United
States and Canada were already highly integrated. Most tariffs on industrial products
traded between the United States and Canada were zero at the time of NAFTA
implementation. In contrast, Mexico had followed an aggressive import-substitution
policy for many years prior to NAFTA in which it had sought to develop certain
domestic industries through trade protection.
The Department of Labor NAFTA-Trade Adjustment Assistance113 program
provides some data on the number of workers covered by certification. The number
of certified workers is not the same as the number of jobs lost due to NAFTA, but it
provides some indication of the adjustment costs of NAFTA.114 Between January
1994 and December 2001, 415,371 workers were covered by NAFTA-TAA
certification. The industry with the highest number of NAFTA-TAA certified jobs
was the combined textile and apparel industry, with 34% of total NAFTA-TAA
certifications, followed by the automotive industry, with 6% of the total. NAFTATAA certification figures may have overestimated job losses among certified workers
because not all certified workers may have actually lost their jobs. Data from the
Department of Labor suggest that as few as 20-30% of certified workers collect
NAFTA-TAA benefits. Certified workers may not have actually lost their jobs, may
113
Congress included a NAFTA Transitional Adjustment Assistance (NAFTA-TAA)
Program in the NAFTA implementing legislation to address concerns regarding worker
dislocations. The NAFTA-TAA Program, which was later consolidated with the former
Trade Adjustment Assistance (TAA) program and is now part of a new reformed TAA
Program, provided assistance to workers who lost their jobs due directly to import
competition or production shifts to Mexico or Canada.
114
CRS Report RS20229, p. 6.
CRS-51
have found another job, or may not have collected benefits for other reasons.115 On
the other hand, the actual number of jobs lost may be higher than the number of those
certified, because not all workers who have lost their jobs due to import competition
or production shifts have necessarily applied for, or would qualify for, certification.
The main NAFTA provisions related to textiles and apparel were the elimination
of U.S. tariffs and quotas for goods coming from Mexico, and elimination of
Mexican tariffs on U.S. textile and apparel products. Goods are required to meet the
rules of origin provision, assuring that apparel products traded among the three
NAFTA partners are made of yarn and fabric made within the free trade area.
Without a rules of origin provision, apparel companies would have been able to
import fabrics from Asia at lower cost and export the final product to the United
States under the free trade provisions. Textile and apparel quotas, as noted earlier in
this report, were eliminated in January 2005, and all exporters of such products to the
U.S. market, as well as domestic producers, are faced with intensified competition
from China.116
Between 1993 and 2001, U.S. trade in textiles and apparel with Mexico
increased more rapidly than U.S. trade with all countries, suggesting that Mexico
may have begun supplying the U.S. market with goods that would have otherwise
been supplied by Asian countries. The share of U.S. trade with Mexico in textiles
and apparel increased from 8% of U.S. world trade in textiles and apparel in 1993 to
16% percent in 2001. In comparison, the share of U.S. trade with Asia in textiles and
apparel decreased from 55% in 1993 to 45% in 2001. The trade deficit with Mexico
in textiles and apparel also increased more rapidly than the total U.S. deficit in these
products, from $760 million in 1993 to $4.2 billion in 2001. While the large increase
in U.S. imports from Mexico may have displaced U.S. workers in the textiles and
apparel industries, some studies have suggested that NAFTA may have helped the
U.S. textile industry by shifting production from Asian countries to North America.
For example, one study reported that U.S. imports from NAFTA countries tend to
have a higher U.S. content than imports from outside the region, such as China, Hong
Kong, and Taiwan.117
NAFTA automotive provisions include the phased elimination of tariffs, gradual
removal of many non-tariff barriers to trade, rules of origin provisions, enhanced
protection of intellectual property rights, less restrictive government procurement
practices, and the elimination of performance requirements on investors from other
NAFTA countries. Its significance is summarized in Table 5.118 Because the U.S.
115
CRS Report 98-782 E, NAFTA: Estimated U.S. Job “Gains” and “Losses” by State Over
5 ½ Years, updated February 2, 2000, pp. 2-3.
116
CRS Report RS20889, Textile and Apparel Quota Phaseout: Some Economic
Implications.
117
U.S. International Trade Commission. Impact of the North American Free Trade
Agreement on the U.S. Economy and Industries: A Three Year Review, Publication 3045
(July 1997), p. 82.
118
An updated and more detailed discussion of U.S.-Mexican automotive trade
developments is in CRS Report RL32883, pp. 19-23 and 53; Table 5 is taken from that
(continued...)
CRS-52
and Canadian automotive industries were already highly integrated following the
U.S.-Canada Free Trade Agreement and an earlier U.S.-Canada Auto Pact in 1965,
most of the impacts of NAFTA relate to trade liberalization with Mexico. In
particular, NAFTA required the removal of Mexico’s restrictive trade and investment
policies. Mexican tariffs on all types of motor vehicles and parts, which were as high
as 20% for some goods, were phased out by 2003. In addition, Mexico agreed to
lower or entirely remove investment restrictions in the automotive sector, which
provided an incentive to increase U.S. investment in Mexico. The United States
eliminated the 2.5% tariff on motor vehicles manufactured in Mexico and phased out
the 25% tariff on Mexican light trucks. The United States also phased out tariffs on
most Mexican auto parts. Mexico’s importance to the U.S. motor vehicle industry
has increased as a result.
U.S. motor vehicle exports to Mexico increased from almost nothing in 1990
to $4.1 billion in 2004. But imports from Mexico increased from $2.9 billion in 1990
to $19.1 billion in 2004, representing an increase of 658%, faster than any other
major source. By value, motor vehicle imports from Mexico increased from less than
5% to 13.5% of the U.S. total. However, the combined share of Mexico and Canada
in the total vehicle and parts import market increased less dramatically, from 40% to
48%. As for U.S. exports the North American partners have been the biggest U.S.
market for vehicles and parts, both before and after NAFTA, as their share was 73%
in both 1990 and 2004. The major change, as mentioned above, is in the increased
role for Mexico as a market for U.S. vehicle exports.
Table 5. Details of U.S. Automotive Trade
(All totals in billions of dollars)
1990
2000
2004
Trading Partners
Exports
Imports
Exports
Imports
Exports
Imports
Canada — Vehicles
8.1
20.4
14.9
40.9
18.4
46.5
— Auto Parts
13.7
9.1
29.6
17.6
29.9
19.2
Mexico — Vehicles
0.3
2.9
3.8
21.0
4.1
19.1
— Auto Parts
4.3
4.5
12.6
18.2
11.3
21.4
NAFTA Totals:
— Vehicles
8.4
23.3
18.7
61.9
22.5
65.6
— Auto Parts
18.0
13.6
42.2
35.8
41.2
40.6
European Union*
— Vehicles
1.7
9.3
2.2
22.2
6.1
31.0
— Auto Parts
1.9
4.6
4.8
7.7
4.6
10.9
Japan — Veh
This text is long and has been trimmed here. Open the source document for the complete record.
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.