« July 2010 | Main | September 2010 »

August 19, 2010

Is U.S. Manufacturing Disappearing?

By William Strauss

When discussing the health of the manufacturing sector, one major issue is whether we should assess the number of people employed in the sector or focus on the amount of output created in manufacturing. Interestingly, each leads to the opposite conclusion about the strength of manufacturing in the United States.[1]

Manufacturing employment as a share of total employment in the United States has been declining over the past 60 years. In 1950, nearly 31% of nonfarm workers were employed in manufacturing. Since then, the share has been dropping three or four percentage points per decade, falling to 28.4% in 1960, 25.1% in 1970, 20.7% in 1980, 16.2% in 1990, 13.1% in 2000, and 9.1% in 2009. Even with this downward trend in manufacturing’s share of jobs, employment in manufacturing has on average been fairly stable over the past 60 years, averaging a decline of –0.1% per year. In contrast, the growth of nonfarm employment averaged 1.9% per year, and this led to the reduction in manufacturing’s share of jobs.[2]

By 2006, the U.S. economy employed about as many workers in manufacturing as in 1950, just over 14 million. And so, looking at manufacturing employment alone leads one to believe that the sector is in decline or at best stagnant.

However, a very different conclusion emerges if you focus on the amount of goods being produced by the manufacturing sector. While employment has changed very little over the past 60 years,[3] output in manufacturing has increased at an annual rate of 3.4%. Manufacturing output in 2007 (the recent peak in manufacturing output) was over 600% higher than in 1950.

Productivity Is the Key

So how was manufacturing output able to surge over the past 60 years with little change in the sector’s employment? The answer can be found by looking at productivity. The increase in both the number and quality of machinery over time, along with technological improvements in production processes and inventory management, have given rise to greater manufacturing sector output at lower unit cost. Productivity growth in the manufacturing sector has averaged 2.9% over the past 60 years. In essence, this means that manufacturing sector output has risen each and every year by around 2.9%. What took 1,000 workers to produce in 1950 could be produced with 184 workers in 2009.

Between 1950 and 1979, productivity growth in the manufacturing sector was matched by the productivity growth of the nonfarm economy—both averaged a rate of 2.5% each year. Since then, with the adaptation of CNC machinery (i.e. Computer Numerical Control) manufacturing during the late 1970s (and other subsequent innovations), productivity growth in the manufacturing sector accelerated to average 3.3% per year between 1980 and 2009, while in the nonfarm economy productivity growth eased to an average of 2.0% per year.

Manufacturing output itself grew on average by 4.2% per year between 1950 and 1979 and then slowed between 1980 and 2009 to an average of 2.2% per year. So, over the past 30 years, relatively slower manufacturing output growth and faster productivity growth led to a declining manufacturing labor force.

Efficiency Leads to Lower Prices

Another observation about manufacturing’s long-term behavior in the U.S. economy is that, between 1950 and 2007 (prior to the severe recession), manufacturing output was just over 600% higher while over the same period growth in real GDP of the U.S. was only a slightly lesser 560%. Yet, the manufacturing share of GDP declined markedly over this period as measured in current dollar value of output. In 1950, the manufacturing share of the U.S. economy amounted to 27% of nominal GDP, but by 2007 it had fallen to 12.1%. How did a sector that experienced growth at a faster pace than the overall economy become a smaller part of the overall economy? The answer again is productivity growth. The greater efficiency of the manufacturing sector afforded either a slower price increase or an outright decline in the prices of this sector’s goods. As one example, inflation (as measured by the Consumer Price Index) averaged 3.7% between 1980 and 2009, while at the same time the rise in prices for new vehicles averaged 1.7%. So while the number (and quality) of manufactured goods had been rising over time, their relative value compared with the output of other sectors did not keep pace.[4] This allowed manufactured goods to be less costly to consumers and led to the manufacturing sector’s declining share of GDP.

Strong Productivity Gains Expected to Continue

Since much of the gains in U.S. manufacturing have been due to strong productivity, a natural question to raise is whether these gains will continue. Often, advancement in technology leads to productivity gains. Accordingly, U.S. spending on research and development can be used as a proxy for the effort being devoted to developing new technology. On this front, the U.S. appears to be in relatively good shape as we continue to invest heavily in research and development. Research and development averaged 2.5% of our GDP between 1953 and 2008. Between 1999 and 2008, it averaged 2.7%, with 2008 at 2.8%.

The private sector has played an ever-increasing role in research and development spending, suggesting that these resources are well-directed in raising productivity. Fifty years ago the majority of research and development was being funded by the government, much of it in support of public sector programs. More recently, the private sector has become the major funder—the privately funded share of research and development averaged 36% during the 1960s; 47% in the 1970s; 54% in the 1980s; 66% in the 1990s; and 72% between 2000 and 2008. [5]

Every two years, Chicago hosts one of the premier manufacturing shows in the world, the International Manufacturing Technology Show IMTS. It is amazing to see the cutting-edge technologies that are becoming available to manufacturers. I typically ask exhibitors of common manufacturing equipment to explain to me the differences between their new equipment and what was displayed two years earlier. The response is almost universal: The new pieces of equipment are more accurate, faster, more versatile, and less expensive than their predecessors.

I often have the opportunity to tour manufacturing production facilities, and I am impressed by the continuous improvements in technology that companies employ. I always ask these producers the following question: Can they envision being able to be even more productive? Nearly all of them tell me without hesitation that they absolutely can become even more efficient, and many then launch into a description of the near-term plans that will make them so.


The changes in manufacturing output, productivity, and labor have not been occurring just over the past few years but over decades. Using recent history is a guide, we can look forward to an industry that will continue to produce more, contributing to a stronger U.S. economy, with manufacturing employment representing a smaller share of the overall U.S. labor market.

The manufacturing sector remains vibrant and innovative. Manufacturing output has been rising at a solid pace over time. Most of this growth, especially over the past 30 years, has been achieved by improving productivity. Of course, for some workers and towns, this increase in productivity has been a double-edged sword, since highly productive operations can achieve their output goals using fewer workers. Nonetheless, higher productivity has fostered a globally competitive U.S. manufacturing sector with the ability to produce more goods with relatively lower price increases, which has benefited U.S. households and the overall economy.

[1] For a further discussion, see William Strauss, CFL 2003. (Return to text)

[2] There is a break that occurs during this period. Between 1950 and 1979, manufacturing employment increased on average by 1.4% per year (over the same period nonfarm employment was rising on average by 2.4% per year), and between 1980 and 2009 manufacturing employment declined on average by 1.6% per year (over the same time nonfarm employment growth slowed, rising on average by 1.3% per year). (Return to text)

[3] Between 1950 and 1979 productivity growth rates in both durable manufacturing and nondurable manufacturing were nearly identical, averaging 2.6% and 2.7% per year, respectively. However, between 1980 and 2009 productivity growth for durable manufacturing surged, to average 3.9% per year, and productivity growth for nondurable manufacturing declined, to average 2.4% per year. Durable goods manufacturing makes greater use of machinery, and the sector was aided by the advancements in the capabilities of machines over this period.(Return to text)

[4] Low elasticity of demand for many manufactured goods contributes to a falling nominal share of the manufacturing sector in GDP. Over time, elasticity of demand for manufactured products by domestic and overseas markets was not responsive enough to offset falling prices.(Return to text)

[5] As reported by the National Science Foundation. (Return to text)

Please enter the security code you see here

Posted by Testa at 11:16 AM | Comments (0)

August 4, 2010

A continued role for central cities?

Bill Testa and Bill Sander

The rapid flight of jobs and people from central cities, such as New York and Chicago, during the 1970s called into question what role, if any, they could play in metropolitan area economies. Today, many central cities of the Northeast and Midwest continue to be significant centers of commerce and employment in their metropolitan areas[1]. Over the past four to five decades, the manufacturing production sector has invariably shifted its activities from central cities to the suburbs (or moved away from metropolitan regions altogether). While doing so, this sector has shed many jobs. Yet certain important service sectors continue to find central cities to be desirable and hospitable locations in which to conduct their business. These include, for example, higher education, specialized medical services, and some types of government services that require face-to-face service delivery[2]. For such businesses, central cities remain accessible destinations for customers drawn from broad surrounding market areas. Customers are still willing to travel to service providers’ places of business located in cities.

In addition, a central city location can be highly productive for those businesses in which frequent face-to-face communication must take place among their workers or nearby business-to-business customers. The high density of central business districts can be very productive for such industries: Deals are struck, creative ideas are born, or temporary partnerships are formed as a result of frequent (and often unplanned) personal interactions that a central city facilitates. Such businesses often include law, financial services, advertising/marketing, public relations, management consulting, R&D, headquarters, and accounting (see figure below).

Click to enlarge.

Workers in these “city-centric” businesses are generally those who engage in creative, administrative, and highly skilled work activities[3]. In no small part, the rising share of high-skill jobs in the overall U.S. economy has helped to boost this economic role of central cities. As the rate of technological change has increased (and as global markets have become more open) over the past four decades, the wages and incomes of highly skilled and educated workers have grown considerably. More generally, a study by E. Glaeser and A. Saiz finds that, compared with other regions, “skilled cities” have been growing more rapidly for at least a century because they are more productive and also because they adapt more easily to economic shocks, such as changing industry fortunes.

Drawing on data from the U.S. Census Bureau’s American Community Survey (ACS), the table below presents the proportion of the city workforce with four-year college degrees (or higher) as compared with the rest of the metropolitan area (suburbs). The workers are reported here by their place of work although they may reside anywhere in the metropolitan area, be it city or suburb. Although educational attainment is far from a comprehensive measure of skill, it is a widely used measure since the level of educational attainment correlates strongly with wages and income. Generally, it is seen that the central city workers of the Northeast and Midwest tend to have greater educational attainment than their suburbs[4]; Detroit and Philadelphia are exceptions.[5].

Click to enlarge.

Although central cities tend to draw workers specializing in education-intensive occupations, the city's share of such jobs in the metropolitan area, as compared with the surrounding suburbs, varies greatly. The first two columns of the table below show the shares of a metropolitan area’s college-educated workers employed in the city and in the suburbs. At one end of the spectrum, the city of Detroit accounts for only 12 percent of employed workers who have college degrees in the Detroit metropolitan area. At the other end of the spectrum, New York City hosts 50 percent of employed workers with a college degrees in the metropolitan region. In the industrial belt, central city Chicago leads with 38 percent; Milwaukee and Pittsburgh are not far behind. The central city of Detroit does lie far behind with 12 percent of the metropolitan area’s college educated workers.

Click to enlarge.

The choice of residential location by those adults who are college-educated also has important consequences for central cities because such households generally contribute heavily in their payment of local taxes. Older cities of the Northeast and Midwest often have a disproportionate amount of low-income households for whom public services such as education must be financed through local taxes. Having more highly educated, high-income households living in the city boosts the local tax base, providing more revenue for local governments to provide these vital public services.

And as shown by various studies, where workers choose to live can affect where they choose to work—and vice versa. Accordingly, by attracting more highly educated (and other high-income) households, a city can draw in employers that are searching for available high-skill workers. Recognizing these incentives, many central city mayors have fashioned public amenities and services, such as parks, public art, and high-quality schools, to attract highly educated, high-income households in recent years.

The final two columns of the table above show the shares of college-educated households in each metropolitan area residing in the central city and in the nearby suburbs. Although gentrified neighborhoods can be found in virtually all cities, some central cities—for example, Detroit and St. Louis—can be seen as having fared poorly overall at attracting residents with high educational attainment. In addition, there are cities that have performed well at attracting jobs requiring a high level of education, but have done less well at drawing highly educated, high-income households such as Boston, Pittsburgh, and Philadelphia. Chicago and Milwaukee are seen to be fairly balanced and successful in attracting both college-educated workers and households.

Central cities in the Northeast and Midwest continue to struggle to find new and sustainable economic roles within metropolitan areas. Almost without exception, the city’s economic base has historically centered on manufacturing production and goods transportation—industries that have since diminished or re-located to the suburbs or elsewhere. Meanwhile, the nation’s economy has been restructuring to favor occupations requiring highly skilled and educated individuals. Fortunately for some central cities, their high population density, among other amenities, provides favorable conditions for these types of jobs. Some cities of the Northeast and Midwest have been successful at not only attracting highly skilled jobs, but also at attracting highly educated, high-income households. In addition to boosting the local tax base, such success may be highly desirable because employers tend to follow the available local work force.

Click to enlarge.

Click to enlarge.

[1] This is not to say that employment is centralizing. A recent study of 98 metropolitan areas by Elizabeth Kneebone (link) observes that employment decentralized in almost every major industry between 1998 and 2006. E. Glaeser and M. Kahn (link) find that, on average, less than 16 percent of metropolitan area employment locates within three miles of a city center. (Return to text)

[2] Tim Bartik and George Erickcek characterized the economic roles of these industries for central cities’ economies, see reference (link) (Return to text)

[3] One exception has been the centralization of entertainment and recreational activities in some central cities. Such activities—involving tourism, hotels, eating and drinking places, music and theatre venues, and sporting events-- may not employ highly skilled or educated workers to the same degree. (Return to text)

[4]The ACS enumerates city-based workers by their place of work; it also separately estimates workers by place of their residence. (Return to text)

[5] Note that a city’s downtown or “central business district” often differs markedly in its share of college-educated workers, compared with the remainder of the city—this is illustrated by Manhattan versus the rest of New York City in table 1 . In addition, a study (link) by E.L. Birch of 44 selected cities from 1970 to 2000 finds that young and educated households are trending toward living downtown, but that downtowns are also often the domiciles of some of the most and least affluent population groups. (Return to text)

Please enter the security code you see here

Posted by Testa at 10:19 AM | Comments (0)