Posted 12/12/2002 10:47 PM Updated 12/12/2002 10:47 PM

Endangered manufacturing jobs

Percentage of non-farm workers employed in manufacturing and service jobs (All 2002 figures are to date):

Manufacturing jobs

1950 34%

2002 13%

Service jobs

1950 59%

2002 82%

Examples of manufacturing job losses. Number employed in thousands:

Textiles

1960 924

2002 433

Apparel

1960

1,233

2002 522

Metal1

1960 1,185

2002

593

Examples of service job gains. Number employed in thousands:

Educational2

1960 616

2002 2,510

Business services3

1960

656

2002 9,301

Health

1960 1,548

2002

10,661

1- primary metal industries such as steel; 2 - private education institutions; 3 - services to businesses such as advertising, data processing, credit reporting, etc.

Source: USA TODAY analysis of Bureau of Labor Statistics data

BEYOND WORDS

Audio

USA TODAY's Barbara Hagenbaugh interviews Rochester, N.Y., Mayor William Johnson

List Names, positions and brief resumes for key employees. You can also include pictures, credentials, and affiliations with professional or other organization.

U.S. manufacturing jobs fading away fast

By Barbara Hagenbaugh, USA TODAY

ROCHESTER, N.Y. — Charles Seitz remembers when Rochester was a bustling manufacturing town. Now, all the 58-year-old unemployed engineer sees is a landscape of empty buildings.

"There's nothing made here anymore," the former Eastman Kodak employee says, his eyes welling with tears as he talks about his struggle to find a new job. "Wealth is really created by making things. I still adhere to that."

It's a situation that's been playing out across the country for decades but has received increased attention in recent years.

Fifty years ago, a third of U.S. employees worked in factories, making everything from clothing to lipstick to cars. Today, a little more than one-tenth of the nation's 131 million workers are employed by manufacturing firms. Four-fifths are in services.

The decline in manufacturing jobs has swiftly accelerated since the beginning of 2000. Since then, more than 1.9 million factory jobs have been cut — about 10% of the sector's workforce. During the same period, the number of jobs outside manufacturing has risen close to 2%.

Many of the factory jobs are being cut as companies respond to a sharp rise in global competition. Unable to raise prices — and often forced to cut them — companies must find any way they can to reduce costs and hang onto profits.

Jobs are increasingly being moved abroad as companies take advantage of lower labor costs and position themselves to sell products to a growing — and promising — market abroad. Economy.com, an economic consulting firm in West Chester, Pa., estimates 1.3 million manufacturing jobs have been moved abroad since the beginning of 1992 — the bulk coming in the last three years. Most of those jobs have gone to Mexico and East Asia.

By Carlos Ortiz, AP

Eastman Kodak workers in Rochester, N.Y., have been losing jobs as the film maker closes plants.

Last month, film giant Eastman Kodak — the largest employer in Rochester and the central focus of the community since the company was founded by George Eastman in 1888 — announced it was shutting down an area plant and laying off the 500 employees who make single-use, sometimes called "throw-away," cameras. The work will now be done in China or Mexico, two countries where the company already has operations.

The movement of jobs to other countries angers Seitz the most.

"The United States got to where it is today by making things," he says. "People are suffering, and communities are suffering."

Four years ago, the materials engineer was laid off after 26 years at Kodak. It was right before Christmas, and he was two months away from being eligible for full retirement benefits.

Seitz's father had worked as a toolmaker at Bausch & Lomb, another major Rochester employer. After leaving Kodak, Seitz found work at another manufacturer, but was laid off after 11/2 years. He's since worked a bit for his son-in-law's plumbing business and is now collecting unemployment benefits .

Fearing for his job, Bill Williams, 56, took a buyout package from Kodak in January after 27 years with the company.

"I don't see much future in manufacturing," says Williams, who started out mopping floors and later ran a machine that coated film. "It's too easy for them to move their plants overseas."

Changing, not dying

Many in manufacturing disagree that the sector is dying. They say it's just changing. The sector's output grew for a decade through 2000 before weakening during the economic downturn in 2001 that swept across the economy but hit the manufacturing sector hardest.

And economists say the change in manufacturing, albeit painful, is healthy for the sector and for the overall economy in the long run.

"It's good for us to displace low-wage, manual kinds of labor with higher-skill, higher-tech, higher-education-content labor," says Federal Reserve Bank of St. Louis President William Poole, who compares what's happening with the decline in agricultural employment of the early 20th century.

"It's an ongoing process, but it obviously means that people who fall behind in the accumulation of skills find their incomes advance more slowly, and they suffer a greater risk of unemployment," he says.

The Manufacturers Alliance, an Arlington, Va.-based group that represents mostly large manufacturers across the USA, is publishing a book next year: U.S. Manufacturing: The Engine for Growth in a Global Economy. They argue the sector has strong footing.

"Our economy is flexible enough, inventive enough, that we'll find a way to keep a vibrant manufacturing sector," Manufacturers Alliance President Thomas Duesterberg says.

What the future looks like, according to the group, economists and other industry watchers:

* High technology. Companies will constantly be coming up with new products and new high-tech ways to cut costs, with U.S. manufacturers leading the way. Budgets will intensely be geared toward research and development, a trend already underway.

At Kodak, a machine now mixes filmmaking ingredients with precision. Ten years ago, it took 14 workers to do the job — in the dark — a repetitive process asking for on-the-job injuries and mistakes.

Few workers are seen at all in "Building 38," which employs the highest level of technology at Kodak Park — a sprawling 1,300-acre, 170-building complex dotted by smoke stacks. Most of the people who are there watch computer screens and closed-circuit TV monitors behind protective goggles in control rooms.

* Development in the USA. Most of the innovation will still happen at home, taking advantage of the nation's highly skilled and educated workforce.

* Production abroad. Actual production of those products will likely happen in other countries. That's not just to take advantage of the lower labor costs but also to position products in global markets. For example, many U.S. companies are making or preparing to make products in China, because they expect increased demand for consumer goods in the quickly developing country as well as in other parts of Asia. Making the products closer to the point of sale reduces transportation costs.

* Fewer U.S. factory workers. Many economists believe most of the manufacturing jobs that have been lost are gone for good, especially repetitive work on the factory floor. More job losses in the sector can be expected.

Source: 1/ US Bureau of Labor Statistics; 2/ Bureau of Economic Analysis.

Now let’s shift our focus to the period from 1990 to 2004, when the major

buildup in China trade occurred. During these most recent 14 years, the US lost 3.3

million manufacturing jobs (Table 1). Meanwhile, total output in manufacturing in the

US was increasing by more than 50% (i.e., the index went from 75.0 to 117.0), and

output per worker in manufacturing was increasing by 73% (i.e., the index of output per

Page 7

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 7

worker in US manufacturing went from 100.00 to 173.34; see Table 2, below). To better

understand these changes, I apply "Job Shift Analysis", a simple technique patterned

after shift-share analysis.

Job Shift Analysis of US Manufacturing Employment Losses 1990-2004

In the Job Shift Analysis model, we are trying to address four sets of changes that

are affecting manufacturing job growth/decline in a country (in the present paper, for the

USA in Section I and for China in Section III):

1. Static job losses in the manufacturing sector from productivity growth;

2. Implied, potential manufacturing job gains from GDP growth;

3. Manufacturing job losses from structural changes such as the shift to

producing less labor-intensive goods, the increasing demand for services

relative to goods, and the related shift of existing jobs and of GDP-

induced job growth to other sectors of the economy;

4. Gains (losses) of jobs due to competitive manufacturing advantages

(disadvantages) experienced by the home country.

Competitive advantages in a particular manufactured good can arise from a

number of sources, only some of which are directly related to labor. These non-labor

advantages can include intellectual property attributes and the protections that one

national environment (such as the US) provides compared to another national

environment (such as China or Argentina). They can entail access to final markets, such

as—for example—BMW’s desire to assemble automobiles within the country

representing its most important final market. They can include external economies of

agglomeration or of localization (in the contemporary vernacular, "clustering") arising

from the presence of supply chain partners and other related firms. Likewise, advantages

can grow out of traditional advantage such as access to important natural resources. One

such "natural resource" can be labor related—e.g., productive and/or inexpensive

workers, or a particularly-educated workforce (including innovative and creative people).

And, finally, all such advantages as these can both enjoy and enhance the advantages

provided by a favorable rate of exchange for the national currency.

Labor-based competitive advantages on the cost side (much the objective of data

collection and analysis efforts feeding off of the US Department of Labor’s Foreign

Page 8

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 8

Labor Statistics web site http://www.bls.gov/fls/home.htm, from which I take part of my

data for this paper) derive from the interplay between the following components: (1)

Output per hour by manufacturing workers; (2) Total worker costs per hour, denominated

in the local currency; and (3) The effective exchange rate between the local currency and

foreign exchange (often normalized into US Dollars). Together, these are the components

of the labor cost per unit of manufacturing output stated in the common denominator of

US Dollars.

As we discuss again later, productivity can have a negative, direct effect on labor

employment. And productivity can simultaneously have a positive, indirect

effect—subject, of course, to what is happening with local wage costs denominated in the

local currency and the exchange rate that links local costs to the global economy.

Because technology spreads rapidly in a globalized manufacturing economy, it is

possible for every nation to be affected by the direct effect which acts to reduce the

overall number of manufacturing workers needed. With the competitiveness factor, on

the other hand, there will be both gainers and losers. With rapid productivity growth,

only a very few countries will be able to overcome the employment reducing effects with

sufficient competitive gains to make up for those effects. It is my judgment that,

generally, the "winners" in manufacturing job creation/retention will fall into two

categories: (1) Small countries with productive and well-managed economies; and (2)

Previously-inefficient national economies in which market liberalization is making

available large numbers of workers at very low opportunity cost.

With Job Shift Analysis, we can divide the US manufacturing job gains and losses

over time into three groups: (1) Job gains (losses) one would expect from productivity

growth; (2) Job gains (losses) one would expect from growth in GDP; and (3) A residual

category intended to capture job gains (losses) from the combination of structural and

competitive changes outlined above and not included in the first two parts of the Job

Shift Analysis calculation. As I demonstrate below, this paper’s application of Job Shift

Analysis suggests that productivity growth in manufacturing dominates the overall

manufacturing job losses by the US during the period 1990-2005.

Page 9

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 9

Table 2. Index of Output per Worker in Manufacturing in the US—1990 to

2005 (January to January)

Manufacturing Productivity

Index

Year

(1990=100)

1990

100.00

1995

118.06

2000

144.29

2001

146.46

2002

157.13

2003

164.96

2004

173.34

2005

182.92

Source: US Bureau of Labor Statistics.

We start by asking the simple question: "Given the productivity growth from

1990 to 2004, how many manufacturing workers would be needed in 2004 to produce the

same output as 1990?" We can use this simple question to calculate the "Productivity

Factor" effect on job growth (loss) during the period:

Productivity Factor = 17.695 x (100.00/173.34) – 17.695 = – 7.5 mil. jobs

What this first calculation tells us is that, in static terms, the US lost 7.5 million jobs to

productivity growth during the period 1990-2004. This amounted to more than 40% of

the manufacturing jobs that existed in the US in 1990.

If the US economy of 2004 otherwise looked just like the US economy of 1990,

that would be fine. But in that interim, GDP grew by 56% (the same rate as the growth

in manufacturing output during that same period, incidentally). Based on the 56% growth

in GDP, the economy should have added back 5.7 million manufacturing jobs (based on

the ‘new’ manufacturing productivity levels of 2004):

GDP Growth Factor = (17.695 – 7.5) x 0.56 = + 5.7 mil. jobs

In other words, after the effect of productivity growth, we should have had

(17.695 million – 7.5 million = 10.2 million) manufacturing jobs to supply the 1990 level

of US manufacturing output. Growth of 56% in GDP should have added back (at the

new productivity levels) by a factor of (10.2 million x 0.56 = 5.7 million). Combine the

Page 10

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 10

GDP growth effect with the job losses from productivity growth and you get a net loss of

–1.8 million manufacturing jobs:

GDP Growth Factor – Productivity Growth Factor = (5.7 – 7.5) = – 1.8 mil jobs

The above calculation tells us that during this 14-year period, GDP did not grow

sufficiently to add back the manufacturing jobs in the US that were lost to productivity

growth during that same period. But the actual job losses were greater than the 1.8

million calculated above, raising the question of what happened to the other 1.5 million

(of the 3.3 million) manufacturing jobs that were lost from 1990-2004? In this analysis, I

attribute those losses to Competitive and Structural Factors affecting the US economy.

Competitive & Structural Factors = – 3.3 million – (– 1.8 million) = – 1.5 million

jobs

One of these competitive and structural factors recognized by others, of course, is

the changing composition of Personal Consumption Expenditures (PCE) in the GDP.

Table 3 shows a continuing and long-developing shift in the balance of PCE from

expenditures on "goods" to expenditures on "services". This shifting balance would

affect manufacturing employment adversely, since the manufacturing sector produces the

bulk of the goods that go into that part of the PCE accounting. Because of the shifting

balance in PCE, one should not expect all of GDP growth effect to go straight into

creating new manufacturing jobs, since ‘services’ production and consumption in the

GDP accounts grew faster than ‘goods’ production and consumption in those same

accounts.

Page 11

Indeed, there is some evidence from our calculations of US competitive gains in the late-1990s that

3

disappears or goes the other way after 2000.

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 11

Table 3. Personal Consumption Expenditures for "Goods" versus "Services" in the

National Income and Product Accounts of the United States—1950 to 2004

Year

Goods

Services

1950

67.1%

32.9%

1960

59.1%

40.9%

1970

55.1%

44.9%

1980

51.8%

48.2%

1990

44.9%

55.1%

2000

41.7%

58.3%

2004

41.0%

59.1%

Source: National Income and Product Accounts, 1950 to 2004. Bureau of Economic

Analysis.

Still another matter is the competitive effect of growing global manufacturing

competition. Just as in the Shift-Share Analysis alluded to in passing, above, it is

possible to gain an increasing share of jobs with an overall sector or industry that actually

is in secular decline globally.

Had the US manufacturers been gaining significant global competitive advantage

during the period we are analyzing, we might have seen net positive numbers resulting

from our Competitive and Structural Factor calculation. We did not, so we might

3

surmise that the US did not secure sufficient competitive gains during full period 1990-

2004 to make up for the net manufacturing job losses shown for the sum or our

Productivity and GDP Growth Factor calculations. Of course, the global competitiveness

of US manufacturers is more complicated than worker productivity alone (including

negotiated wage agreements, taxation, and exchange rate issues as well, for example).

Meanwhile, compare the United States economy’s 1.5 million manufacturing job

losses from competitive and structural factors to the 7.5 million lost to productivity

growth. From this comparison, I conclude that, during the period 1990 to early-2005, US

manufacturing productivity growth cost the US several times more manufacturing jobs

than all other factors combined—including global competition.

Page 12

Further buttressing confidence in our calculation is the fact that the Index of Manufacturing Output (Table

4

1) stood at about the same level at the end of 2004 as in 2000 (more on this point below).

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 12

Job Shift Analysis of US Employment Changes 2000-2005

Before we go on to look at other issues such as manufacturing sector trends in

other countries, let us refine our Job Shift Analysis to look at the period from 2000 to

early-2005, the period during which 83% of the overall US job losses in manufacturing

alluded to above occurred. This particular application of Job Shift Analysis reveals some

striking outcomes on the interplay between productivity growth and economic

adjustments in the US economy.

Using the same data tables and Job Shift Analysis algorithms as previously, we

get the following results:

Productivity Growth Factor (2000 to 2005): – 3.005 million jobs

GDP Growth Factor (2000 to 2005): + 1.831 million jobs

Total Actual Job Gains or Losses (2000 to 2005): – 3.005 million jobs

Competitive & Structural Factor (2000 to 2005): – 1.831 million jobs

This application of Job Shift Analysis tells us something very interesting about

the period after 2000, because the actual job losses in manufacturing during that period

match up closely with losses we would have expected from productivity growth alone.

4

In addition to being a period of rapidly rising manufacturing exports by China, the

2000-2005 interval also was a period of particularly high productivity growth in

manufacturing in the US (and, as we show in Section III, in the larger world as well—See

Table 2 and Table 9). According to this second application of Job Shift Analysis, 100%

of the US manufacturing job losses from the base point of 2000 were due to productivity

growth. Meanwhile, 100% of the GDP Growth Factor went into creating income and

jobs in OTHER sectors of the US economy (i.e., other than manufacturing) after 2000.

Thus, the period 2000-2005 was a period that was particularly marked by competitive

and structural adjustments in the US economy.

Meanwhile, during the post-2000 period, US GDP rose by about 13%. And the

US index of manufacturing output shown in Table 1 that stood at 117.3 in 2000 was back

Page 13

Appendix Table 1 shows actual net job creation overall in the US economy during 2000-2005 (January to

5

January) of 1.7 million jobs, very close to the 1.831 million jobs that our Job Shift Analysis suggested

should have developed outside of manufacturing but within the overall US economy.

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 13

again to 117 (after declining in 2001-2003 and then rising again through 2004).

Productivity growth in that period took away three million manufacturing jobs—the

number of manufacturing jobs actually lost in the period. The GDP Growth Factor

should have created 1.831 million new jobs—but they were not in manufacturing . So,

5

where were they?

Section II. Index of Job Quality Changes in the US

Table 4 shows that the job losses in the US economy between 2000 and early 2005

occurred in manufacturing (about 3 million jobs), in "information" (434,000), in

wholesale trade (279,000), in retail trade (134,000), in transport and warehousing

(50,000), and in utilities (31,000). Meanwhile, 2.2 million net jobs were developing in

education and health services, 1.1 million in government (presumably paralleling the

overall average of two-thirds of them being in local government), and almost a million in

Leisure and Hospitality. All-in-all, the Bureau of Labor Statistics recorded net job

growth of 1.7 million in the US economy as counterpart to the GDP growth of 13%

between 2000 and early-2005. This suggests that the two largest employment substitutes

for manufacturing jobs lost during 2000-2005 were (a) Education and Health Services,

and (b) Government. Take Government out of the job accounts, and fewer than 0.6

million private sector net new jobs were created in the US after January 2000 (Therefore,

the application of the term "jobless recovery" following the recovery from the 2001

economic downturn). US GDP grew by 13% after 2000, and the US civilian workforce

grew by about 3.3% (i.e., from 142.6 million in 2000 to 147.4 million in 2004, according

to BLS data). However, US civilian employment grew less than one-half of one percent

(i.e., 600,000 divided by 137 million) during that period.

Page 14

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 14

Table 4. US Job Gains and Losses by Sector (January to January), 1990 to 2005

Jobs Gained (Lost) in ‘000s

Sector

2000-2005

1995-2000

1990-1995

Government

1,129

1,162

1,211

Education & Health Services

2,214

1,836

2,321

Financial Activities

472

853

230

Information

(434)

766

110

Leisure & Hospitality

971

1,305

985

Other Services

305

619

301

Professional and Business Services

222

3,723

1,778

Transport and Warehousing

(50)

569

348

Utilities

(31)

(68)

(61)

Wholesale Trade

(279)

581

87

Retail Trade

(134)

1,347

520

Construction

329

1,534

(186)

Natural Resources and Mining

13

(59)

(110)

Manufacturing

(2,999)

47

(516)

Total

1,728

14,215

7,018

Source: Calculated in Appendix Table 1. Data from USDOL, BLS.

To see how the pattern of job gains and losses affected the overall quality of net new

employment created in the US economy from January 2000 to January 2005, I created an

"Index of Job Quality Change" in which job changes in each (private) sector were

multiplied by the average hourly compensation for that respective sector, and the sum of

these products was then divided by the product of total private sector job change

multiplied by the private sector average hourly compensation (for January 2005, as

reported by BLS). I then calculated the Index for each of the intervals we have been

analyzing and present the Index calculation for the post-2000 period in Table 5. In Table

6, the results of calculations of the Index of Job Quality Changes for the 1990-1995 and

the 1995-2000 periods also are presented.

Page 15

See the papers and related citations in Economic Perspectives, Federal Reserve Bank of Chicago, Vol. 28,

6

No. 2, (2 Quarter 2004).

nd

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 15

Table 5. US Private Sector Worker Average Hourly Compensation (January 2005)

and Calculation of Index of Job Quality Change for 2000-2005 Period

Net Jobs

Hourly

Sector

2000-2005

Compensation

Product

Private Sector Category

(in ‘000s)

in 2005

in Index

Education & Health Services

2,214

$16.16

35,778

Financial Activities

472

$17.53

8,274

Information

(434)

$21.42

(9,296)

Leisure & Hospitality

971

$8.91

8,652

Other Services

305

$13.98

4,264

Professional and Business Services

222

$17.46

3,876

Transport and Warehousing

(50)

$16.43

(822)

Utilities

(31)

$25.62

(794)

Wholesale Trade

(279)

$17.66

(4,927)

Retail Trade

(134)

$12.08

(1,619)

Construction

329

$19.23

6,327

Natural Resources and Mining

13

$18.08

235

Manufacturing

(2,999)

$16.14

(48,404)

Sum of the Sector Products

1,544

Total Private

599

$15.67

9,386

Index of Job Quality Change (2000-2005): (1,544/9,386) = 0.16

The Index of Job Quality Changes for the periods 1990-1995 and 1995-2000

presented in Table 6 show striking differences between these two earlier intervals versus

the post-2000 period results that were calculated in Table 5. Indeed, the positive job

quality performance of the 1995-2000 period (Index of 1.03) stands in stark contrast to

the very negative job quality performance of the US economy after 2000 (Index of 0.16).

Driving this dramatic change in my Index for 1995-2000 versus the Index for 2000-2005

was the decline in relative importance of newly-tradable services such as Professional

and Business Services and the increase in relative importance of less-tradable services

such as Education and Health Services, and Leisure and Hospitality between these two

periods. This suggests to me two things: (1) The importance of the "sectoral reallocation"

modeling that is being done by a number of analysts, particularly in conjunction with the

Chicago Fed; and (2) The importance of broadening that work to include global data and

6

Page 16

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 16

analysis on employment change. My Section III should make clear the importance of

this second suggestion.

Table 6. Index of Job Quality Change for Five-Year Intervals During 1990-2005

Index of

Interval

Job Quality Change

1990-1995

0.95

1995-2000

1.03

2000-2005

0.16

1990-2005

0.97

Source: Author’s calculations from data in Tables 4 and 5.

Section III. Manufacturing in China and the Rest of the World

Let us begin Section III with a look at the World Bank’s World Development

Indicators 2005 data on "Industry" employment as a percent of total employment for the

most industrialized of the Bank’s member countries. From the WDI 2005 data, we see a

dramatic shift in the proportion of workers employed in industry versus services in the

middle- and high-income countries during the decade ending in 2000-2002. In the

middle-income countries, the proportion of workers employed in industry declined by a

third during that decade, while in the high-income countries the proportion declined by

half (Table 7).

Page 17

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 17

Table 7. Employment by Economic Activity, High-Income and Upper-

Middle-Income Countries (1990-1992 and 2000-2002)

"Source: World Bank, World Development Indicators 2005.

(1) Agriculture, forestry, hunting and fishing are included in "agriculture".

(2) Manufacturing, mining, and construction are included in "industry".

(3) Transportation, communication, public utilities, trade, finance, public administration, private

household services, and miscellaneous services are included in "services".

From the notes to Table 7, you can see that "Industry" includes more than

"manufacturing" alone, though manufacturing is the dominant sector in that category. To

examine more closely what was happening to manufacturing employment, per se, in the

closest competitor economies of the US between 1990 and 2004, we look to data

compiled by the US Department of Labor and summarized in Table 8, below (For

comparative data going back to 1960, see Appendix Table 3).

Page 18

"In addition, the share of unemployed individuals who have been seeking work for longer than one year is

7

among the highest in the OECD, at about 50 per cent", according to the OECD’s Employment Outlook

2005.

Judith Bannister, "Manufacturing Employment and Compensation in China" (December 2004).

8

Consultant's report to US Department of Labor, Bureau of Labor Statistics.

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 18

Table 8. Manufacturing as Percent of Total Civilian Employment, Ten

Countries, 1990 and 2004

% Employed

% Employed

Country

in Mfg in 1990

in Mfg in 2004

United States

18.0%

11.8%

Canada

15.7%

14.4%

Australia

15.0%

11.3%

Japan

24.3%

18.3%

France

21.0%

16.3%*

Germany

31.6%

22.7%

Italy

22.6%

21.8%

Netherlands

19.1%

14.0%**

Sweden

22.3%

14.9%*

United Kingdom

22.3%

14.9%*

Note: * Indicates 2003 data, while ** indicates 2002 data.

Source: "Comparative Civilian Labor Force Statistics, 10 Countries, 1960-2004". US Department

of Labor, Bureau of Labor Statistics, Office of Productivity and Technology (May 13, 2005).

http://www.bls.gov/fls/flslforc.pdf

Manufacturing as a percent of total civilian employment declined in every

country presented in Table 8. The two smallest percentage declines occurred in Canada

and in Italy and for two very different reasons. Canada seems to have experienced some

competitive gains in manufacturing employment (a point to which we turn below), while

Italy slowed somewhat the decline in its own manufacturing employment by creating

distortions that leave the country with an unemployment rate (8.1% in 2004) that stands

above the OECD average. Italy’s policy experiments seem also to be reflected in its

7

productivity growth in comparison to other countries profiled in Table 9, (the lowest

productivity growth amongst the countries covered).

In addition to the productivity growth shown for the countries in Table 9, Judith

Bannister suggests that manufacturing productivity grew by about 60% in China

8

between 1995 and 2001. In Table 10, we reproduce Bannister’s estimates of changes in

Page 19

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 19

manufacturing employment in China from 1978 through 2002. Note the large difference

between Bannister’s manufacturing employment estimates for China compared to those

reported by the ILO and shown in Appendix Table 4. In both data sets, major declines in

manufacturing employment in China are reported between 1995 and 2002.

Table 9. Productivity Growth and Employment Change in Manufacturing in

Fourteen Countries, 1992-2003

% Growth

% Change

In Output per Worker

in Employment

In Manufacturing

In Manufacturing

Country

(1992-2003)

(1992-2003)

Canada

34.5

+1.1%

Australia

42.0

-25.7%

Japan

54.3

-25.7%

Korea

155.3

-11.8%

Taiwan

76.1

-2.7%

Belgium

44.0

-16.8%

Denmark

36.0

-12.6%

France

58.0

-10.9%

Germany

35.1

-21.0%

Italy

10.9

-2.9%

Netherlands (*1990-2002)

35.2*

-12.7%

Norway

13.5

-1.5%

Sweden

101.5

-3.6%

United Kingdom

34.9

-18.1%

Source: Bureau of Labor Statistics, US Department of Labor.

ftp://ftp.bls.gov/pub/special.requests/ForeignLabor/prodsuppt01.txt

Table 9 juxtaposes data on manufacturing productivity and manufacturing

employment changes for fourteen countries between 1992 and 2003. In all of the

reporting countries, except Canada, productivity growth was associated with employment

decline.

Meanwhile, Bannister’s estimate of productivity growth (60%) and employment

decline (from 98 million down to 80 million jobs) in manufacturing in China (Table 10)

between 1995 and 2002 are consistent with the results for the other fourteen countries

shown in Table 9. They are also consistent with the results for the US that were presented

in Section I of this paper.

Page 20

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 20

Table 10. Reported Manufacturing Employment in China (Urban and

Rural), 1978-2002

China

Urban

Rural

Index of

Prepared by Judith Bannister. These Figures refer to the mainland provinces of the PRC not

including Hong Kong, Macao, or Taiwan. These data are from China's annual yearend reporting

system, not from census data and not adjusted to agree with census data. The column on

‘Derived urban manufacturing employment’ is calculated from national manufacturing employment

minus rural manufacturing employment."

Source: This table is reproduced with little alteration from the report by Judith Bannister,

"Manufacturing Employment and Compensation in China" (December 2004). Consultant's report

to US Department of Labor, Bureau of Labor Statistics. Bannister reports her sources as "China

NBS & Labour, 2003, pp. 8, 10, 13, 16, 21, 23-26, 171, 473."

The fact that China’s manufacturing employment did not decline by 37.5% (i.e.,

one minus the reciprocal of 1.60) suggests that there were GDP Growth and/or

Competitive and Structural effects upon manufacturing employment in China during the

period 1995-2002 that had some degree of offsetting effect. Productivity Growth should

have reduced China’s manufacturing employment from 98 million workers in 1995 down

to 61 million workers by 2002. However, Bannister reports that China’s manufacturing

employment declined "only" to 80 million workers in 2002 (before recovering a bit to 83

Page 21

Justin Yifu Liu, "Is China’s Growth Real and Sustainable?" China Center for Economic Research, Peking

9

University. Working Paper No. E2004003 (February 26, 2004). http://www.ccer.edu.cn/download/3024-

1.pdf

Note that ILO data in Appendix Table 4 show manufacturing employment in China as "only" 29 million,

10

while Bannister (2004) shows reported manufacturing employment in China to be more than 80 million

workers (out of a workforce estimated in WDI 2004 to be 770 million). Likewise, ILO data in Appendix

Table 4 also show "only" 6.68 million manufacturing workers in India, which would amount to about 1.4%

of that country’s estimated workforce of 470 million. I conclude that ILO data seem to undercount actual

employment in a number of important cases.

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 21

million by 2003). This suggests that China’s GDP Growth, and Competitive and

Structural Factors combined to offset the Productivity Factor effects by an amount

equivalent to 19 million manufacturing jobs.

With an average annual GDP growth rate of 7.8%, the GDP Growth Factor for

9

China should have added back an astounding 42 million manufacturing jobs between

1995 and 2002 (based on Bannister’s data). Yet, as we already indicated, the country

added back "only" 19 million manufacturing jobs. What happened to the remaining 23

million jobs that one would have expected from this combination of manufacturing

productivity growth and GDP growth between 1995 and 2002? According to our Job

Shift Analysis model, we attribute the difference to Competitive and Structural changes

in China’s economy.

Because China is such a large part of the global manufacturing picture, the

absolute magnitude of the numbers in our China Job Shift Analysis is astonishing. The 17

million manufacturing jobs that Bannister’s data indicate were actually lost by China

during 1995-2002 are equivalent to the total US manufacturing employment at that time.

Pulling together national manufacturing employment data from the ILO and other

sources, I estimate that global manufacturing employment totaled between 150 million

and 200 million workers in 2002, with China employing somewhere between one-fourth

and one-half the global total (within these totals, China employed more than 80 million,

the EU-25 approximately 30 million, and Japan and the US together about 25 million).

10

If we assume a global average manufacturing productivity growth of 30%

between 1995 and 2002 (i.e., half that estimated for China), then the global economy

might have lost as many as 60 million manufacturing jobs to productivity growth during

that period. Assuming global GDP growth of approximately the same amount adding

Page 22

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 22

some jobs back combined with a pattern of structural changes similar to those affecting

the US and Chinese economies, one might expect global experience to be something like

that of the US and China (except, of course, that the "competitive" effect nets out on the

global level). Such intuitive calculations would suggest net manufacturing job losses

globally in the 20 million job range. Can we find the numbers to test that intuition?

I have used Appendix Table 4 to cobble together estimates of global

manufacturing employment for 1995 and for 2002 (using ILO data as the building

blocks). In between my 1995 and 2002 global manufacturing employment estimates

shown in Appendix Table 4, I calculate a global net loss of 23 million manufacturing

jobs. From that and related data and calculations, I fudge that calculation with the

statement that the global economy seems to have experienced net manufacturing job

losses between 20 million and 30 million during the period 1995 to 2002.

A bit of a rebound in manufacturing output has occurred as the post-2001

recovery has taken hold, as reflected, for example, in the US Index of Manufacturing

Output shown in Table 1 (and in additional data released after I compiled my data tables

for this paper). However, no comparable rebound is reflected in the manufacturing

employment index numbers shown for the fourteen countries profiled in Appendix Table

2 (nor in ensuing employment data releases).

We all would like to know what has actually happened to manufacturing

employment globally since 2002—particularly in light of the fact that the US economy

lost another million manufacturing jobs after that date. But I simply do not have the data

at this point in time (July 2005) to render a global estimate beyond 2002. Judith

Bannister’s data suggest that China might have regained two million manufacturing jobs

between 2002 and 2003 (and others suggest that more competitive job gains followed in

2004 and 2005). Still, I do not expect that a global rebound of more than 5 million

manufacturing jobs occurred after 2002. Therefore, I estimate that the global economy

lost, at a minimum, 15 million 25 million net manufacturing jobs between 1995 and

2005.

Page 23

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 23

Section IV. Conclusions

The evidence I am able to assemble suggests that manufacturing productivity is

growing rapidly on a global basis—not just in the US. Sure, Italy amongst the industrial

countries did not show high levels of productivity growth in recent years. And we don’t

have manufacturing productivity numbers for the least developed of the world’s

countries. But we do have data for the countries that make up more than half of the

manufacturing output and manufacturing employment in the global economy—China, the

EU-25, and the remaining OECD countries. With few exceptions, they show dramatic

increases in manufacturing productivity and depressing losses in manufacturing jobs

between 1995 and 2002—even in the face of often-gratifying levels of GDP growth.

Data I am able to assemble suggest that global manufacturing employment was

between 150 million and 200 million workers in 2002, with one-fourth to one-half of

those being in China. My estimates lead me to believe that global manufacturing

employment was 20-30 million lower in 2002 than in 1995. The economic recovery after

2001 probably added back no more than 5 million jobs after 2002, with far greater than

half of those add-backs apparently occurring in China. Thus, in the ten years 1995-2005,

the global economy appears to have suffered net losses of 15-25 million manufacturing

jobs. This is significant when one considers that US manufacturing job losses during that

period amounted to about 2 million and that China’s manufacturing job losses during the

period appeared to be significantly greater than total US manufacturing employment at

the beginning of 2005.

I conclude that manufacturing productivity growth and structural shifts in demand

are so great in the global economy that only two kinds of countries will be able to add net

new manufacturing jobs in the second five years of the 21 century: (1) Small countries

st

who can achieve competitive gains at a sufficient rate to make up for productivity growth

and structural change, and (2) Emerging market countries undergoing substantial market

liberalization and, thus, having access to large amounts of resources (particularly labor)

with very low opportunity costs.

As many other analysts already have pointed out, manufacturing will not be the

job creator in the early years of the 21 century in the US that it was during much of the

st

Page 24

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 24

20 century. In addition "manufacturing competitiveness" is not likely to mean the same

th

thing as "job creation" in coming years, as much of that competitiveness will require cost

reductions (read ‘labor reduction’) and switching to production of higher valued products

that do not entail much labor per unit of output. The era of job creation through

"industrialization" (i.e., manufacturing) in the US appears to be over. Again, as others

have suggested, we must now find new ways to achieve employment and income growth.

Not just because we are in the US—a high-income country with a powerful currency.

But also because we are part of a broader economy in which dramatic manufacturing

productivity growth and structural changes in demand and in production organization are

global—not just national phenomena. On the academic research side, our models must

begin to reflect these realities, and (as the BLS is beginning to do) we must focus greater

attention to building the global datasets that we will need to accomplish such changes.

Section V. Epilogue

I end this paper with an appeal for the multilateral organizations to commit

themselves to the task of providing researchers and policy analysts with datasets capable

of reflecting the full extent of global economic relations. As I wrestled in this paper with

the simple model that I have dubbed "Job Shift Analysis", I really wanted to do a

standard shift-share analysis particularly for the USA but also for a number of countries,

using a global dataset made up of fungible national data. As I was beginning thinking

about this paper, Ben Bernanke was making a speech (March 2005) in which he argued

that developing countries as a group had emerged as net exporters and that, as a result,

savings were now flowing in the opposite direction compared to preceding decades. He

could say this from cobbled-together data that (like my own cobbling in this paper) had a

large measurement error associated with it. Nevertheless, because the change he was

chronicling was much larger than the huge error that remained in the summary data, it

proved to be a valuable exercise that helped all of our thinking on the US trade deficit

and on US and international interest rates.

While there is widespread recognition that we now face a globalized and deeply

interrelated economy, much of the analysis from which we build our understanding of

that economy continues to use national datasets and to look at national or—at best—bi-

lateral economic relations. Using national datasets alone and listening to others who

Page 25

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 25

were doing the same thing, I was finding it difficult to understand what the factory and

the company level managers were saying about the globalization process that was/is

leading to massive restructuring of their supply chains. I finally concluded that I could

not really understand those adjustments until I had some understanding of what was

happening at the aggregate, global level. I don’t think I am the only person experiencing

this problem. All of us could do better analysis, if we had addable data that were

(decently) consistent country-to-country and country-to-globe.

I know that lot of work is required to make one country’s data series comparable

to and addable with those of another country. The OECD has put a lot of resources into

creating comparable datasets between its member countries. The Foreign Labor

Statistics group at the US Department of Labor continues to work hard at comparing

labor and employment data between the US and other countries. And the Growth and

Development Centre at Groningen University in the Netherlands has expended

tremendous energy and resources in developing international comparisons of

productivity. It is hard work, requiring lots of resources.

So, who should provide these globally-addable data? The usual suspects are the

International Monetary Fund for macroeconomic data and the World Bank for global

microeconomic data. I would be pleased if the annual edition of World Development

Indicators, for example, would routinely provide the capacity to get global sums for

every measure in that dataset. Country-by-country index numbers are fine for some

kinds of work. But for many types of analysis, there is no real substitute for raw numbers

that can be added up globally.

Page 26

W. Ward

CIT Working Paper 052507 (August 4, 2005)

Page 26

Appendix Table 1. Employment by Sector in US, 1990 to 2005 (January data, in thousands of employees)

Educ.

THE LIBERAL NEWS™ © Assisting single mothers by our 441 society plan. The Gospel Followers of JESUS CHRIST[sm]© Editor: Dr. Stephen-James Warner

Saving the World; One Person At A Time[sm] = Make Every Day Christmas; Every Night Christmas Eve!

FRONTPAGE

GOSPEL FOLLOWERS OF JESUS

PROTECT OUR TRADEMARK

Preface

Trustworthys

HONORABLE TRUST SITES

HON DYLAN RATTIGAN&CHENK

KEITH OLBERMANN

HONORABLES 2011

>>>>>WORTHY OF TRUST

HonorAwards

THE 441 SOCIETY

Financial

>>>>>OUR RESEARCH

Statistics=Factoids

SITE MISSION MAP CONTENT

GAO,CBO,CENSUS

>>>>>OUR BOOK REVIEWS

>>>>>WHAT ARE THE ISSUES

Opinion=Remarks

NegativeViews2Depressing

Gloom and Doom Grimms

theliberalnews.org!

the prophet?

The Dishonorables

DEMAGOGUE = BECK

Site Map

TV COMMERCIAL 4 REFORMS

ADVERTISING HONOR SYSTEM

911

BLOGS BLOGGER.COM

HEALTH-CARE PROFITEERING

STOP HEALTH MONOPOLY

HEALTH WAGE PRICE CONTROL

21ST CENTURY POL PARTY

PREJUDICE>FREE-MASONS

CYNIC'S CORRUPTION LIST

STOP SYSTEMIC CORRUPTION

NEED NATIONAL PROTESTS

DC MARCH LIVING WAGE JOB

UNIONS=LABOR ALLIANCES

RIGHT TO LIVING WAGE

BUY AMERICAN MOVEMENT

ECONOMIC CONVENTION PLAN

2011=USA MUST START OVER

OUTLAW OUTSOURCING

START REBUILD AMERICA

AlternativeEnergy=PickOne

Quick Use Energy Sources

CUTTING CARBON ILLUSION

Clean Coal Slurry

Coal Gasification Clean

High-Octane Furnaces

Co-generation Plants

Underground Nuclear

Uniform Nuclear Design

Windmill Design Invention

WINDMILL INVENTION NOW!

NEED FORBES FLAT TAX NOW!

CREATE NEW MANUFACTURING

BusinessIndustrialComplex

BANKS INVEST USA OR TAXED

STOP EXPORT US CAPITAL

AMERICA FIRST= INVESTMENT

SaveUSCapitalFutureInvest

USA REFORMS 2011

SOLUTIONS-REFORMS

Specific Solutions

Robotics

ANTI-TRUST LAWS> MONOPOLY

MONOPOLYvsFREE ENTERPRISE

CORP. MONOPOLIES RUN USA

USA A TWO-CLASS SOCIETY

TOP 10% GET 50% INCOME

NEW PARTY DEMS & REPS

NO REPUBLICANS OF OLD

DEBT DEFICIT FALSEHOOD

DEFICIT? TAX THE RICH