A More Vital Economy
A website devoted to strategies to revitalize the U.S. economy

Browne's observations on decline of U.S. Manufacturing


China's growing share of exports - March 25, 2013

 Manufacturing compensation costs - March 14, 2013
Manufacturing Job Creation in the Recovery - January 22, 2013


China’s Growing Share of Global  Exports

March 25, 2013

An interesting paper by Steven Husted and Shuichiro Nishioka on “China’s Fair Share? The Growth of China’s Exports in World Trade” shows how China’s gains in global market share have been accompanied by reductions in the market shares of the United States, Japan and other developed countries. Developing countries, in contrast, have not lost ground to China. This note summarizes my interpretation of several key findings.

In 1998, China accounted for less than 2 percent of world exports.  In 2010, it accounted for over 10 percent.  Husted and Nishioka’s analysis shows that the United States, Japan and Europe lost market share to China, while developing countries generally did not.  Japan and Europe were the big losers in the late 1990s, with the United States actually gaining share in this period, along with China.  After 2000, however, the United States experienced the largest loss of market share.  China’s gains were broadly based, with its share of exports increasing in almost every market.

The authors analyze the growth in exports between 1995 and 2010 for over 90 countries.  They decompose the change in market share of each country into five components:
1) changes in market share in individual products exported to individual countries from the base period (market share)
2) global shifts in demand for different exported commodities (whether demand has shifted towards or away from the products exported by a country)
3) changes in the composition of a country's export products relative to changes in other countries' exports
4) global shifts in demand among importers (whether demand has grown faster or slower among the markets served by a country)
5) changes in the composition of a country's export markets relative to changes for other countries.

Among the 90-plus countries studied, China’s market share increased 13 percentage points between 1995 and 2010.  Japan’s market share fell by 5 percentage points; U.S. market share fell by 4 percentage points, and Europe’s share by 3 percentage points.

The United States experienced a substantial loss of market share in individual products and countries (-6 percentage points,) which was partially offset by shifts in demand towards products exported by the United States (1 percentage point) and by the United States shifting towards products where demand was growing (1 percentage point.)

Most of China’s gain was attributable to gains in market share in individual products exported to individual countries (12 percentage points).  The composition of China’s exports also shifted to products where demand was strong more than other countries (3 percentage points.) Meanwhile, demand shifted away from the products they initially exported (-1 percentage point) and the markets they served (-1 percentage point.)

Basically, China gained market share across the board.  It also shifted its export mix towards products that were more in demand, offsetting a modest shift in demand away from their initial export products and markets.  One consequence of these shifts is that China’s exports have become more technologically sophisticated and more like those of an advanced country.

Most of China’s exports are manufactured products.  In 1995, miscellaneous products, which include apparel, toys and other exports often associated with developing countries, accounted for almost half of China’s exports.  Machinery and transport equipment, which are more commonly exports of advanced countries, accounted for one-quarter.  In 2010, those shares had reversed; and machinery and transport equipment accounted for over half of China’s exports while miscellaneous manufactures were about a quarter.  At the same time, China’s share of world exports in both categories grew dramatically, so that China accounted for a third of world exports of miscellaneous manufactured products and almost 20 percent of exports in machinery and transport equipment.

Manufacturing Compensation Costs 
and Improved Competitiveness

March 14, 2013

The decline in U.S. manufacturing employment since 2000 occurred despite a decline in U.S. compensation costs compared to costs in many of its trading partners.

In December 2012, the Bureau of Labor Statistics released estimates of hourly manufacturing compensation costs from 1997 to 2011 in the United States and 33 other countries (BLS News Release, “International Comparisons of Hourly Compensation Costs in Manufacturing , 2011.)  Compensation in almost all the other countries increased relative to that in the United States.  Not included among the countries analyzed is China, which emerged as a powerful global competitor over this period. Nevertheless, compared to almost all its other important trading partners, U.S. labor cost competitiveness improved.

This note summarizes key points of the BLS compensation data.  It then comments on a separate BLS analysis of compensation costs in China. (Banister and Cook, “China’s employment and compensation costs in manufacturing through 2008.”)

Manufacturing compensation costs over time

The BLS has estimated manufacturing compensation costs for the United States and 33 other countries since the late 1990s.  Missing from the list are China, India and Hong Kong, as well as major oil suppliers, Saudi Arabia and Venezuela.  However, all the other countries ranking among the 15 largest exporters of goods to or importers of  goods from the United States in 2012 are included, among them Canada (ranked 1 in total goods trade), Mexico (3), Japan (4), Germany (5), United Kingdom (6), South Korea (7), and Brazil (8).

In 1997, eight of the 33 countries had higher manufacturing compensation costs than the United States. All were in the northern portion of Western Europe.  Switzerland, at 132 percent of the U.S. costs, was highest, followed by Germany (127 percent.)  By 2001, largely because of an appreciating dollar, only one country, Switzerland, had higher manufacturing compensations costs; and these were only modestly higher, at 107 percent of U.S. costs.  Over the next ten years, however, U.S. costs fell sharply relative to other countries.  In 2011, 15 of 33 countries had higher compensation costs than the United States, and all but one country had seen their compensation costs increase relative to the United States.

Most of the countries with higher compensation costs than the United States in 2011 were in Western Europe, but not all.  Norway ranked highest with compensation costs 181 percent of those in the United States; other Scandinavian countries were also well above the United States.  Switzerland had the second highest level of compensation at 170 percent.  Germany stood at 133 percent.  Ireland and Italy, both of which have faced severe economic and financial difficulties in recent years, had compensation levels above the United States - 112 percent for Ireland and 102 percent for Italy.  But the group of higher wage countries also included three non-European nations: Australia (120 percent of the U.S.), Canada (103 percent) and Japan (101 percent).

Of course, some important trading partners have compensation levels much lower than those in the United States.  Brazil’s compensation was only a third that in the United States, Mexico’s was less than 20 percent.  Korea’s compensation was just over half the U.S. level and Taiwan’s just over a quarter.   Most eastern European countries also have low compensation costs.  However, even among this group, compensation costs have increased relative to the United States.  The sole exception among the 33 countries is Taiwan, where compensation fell from 31 percent of the U.S. level in 1997 to 26 percent in 2011.  The increase in compensation was also modest in Brazil and the United Kingdom (87 percent in 2011.)

Drivers of changes in compensation costs

Changes in compensation are driven by changes in compensation within individual countries and by changes in exchange rates.

The value of the U.S. dollar rose sharply against most currencies in the late 1990s, peaking in 2001.  Depending upon the index used, the increase in the value of the dollar on a trade-weighted basis was between18 and 25 percent.  The appreciating dollar raised the cost of U.S. products in global markets and effectively increased U.S. manufacturing compensation costs relative to those in most other countries.  Since then, the value of the dollar has fallen steeply, except for a surge late in 2008, after the Lehman Brothers financial crisis.  Again, depending upon the index, the decline in value from 2001 to 2011 was 25 to 33 percent. The cheaper dollar effectively reduced manufacturing compensation costs in the United States relative to those elsewhere and should have improved the international competitiveness of U.S. manufactured products.

Compensation costs were also affected by differences in rates of increase within individual countries.  In the United States, manufacturing compensation costs increased at an annual rate of 3.1 percent over the period 1997 to 2011.  Twenty-two countries had higher rates of compensation growth, as measured in their own currencies, while ten had lower rates of compensation growth.  The highest rates of compensation growth were in countries with quite low levels of compensation internationally.  Thus, Mexico and Brazil had compensation growth, in their own currencies, of 8 and 7 percent per year, respectively.  Argentina had the highest rate of domestic compensation growth – 17 percent per year.  Following an economic and financial crisis in 1999-2002, in which the value of the Argentine currency collapsed, domestic wages and prices rose very rapidly.

Countries with low rates of compensation growth tended to be ones with relatively high compensation internationally.  Thus, compensation grew 2 percent per year in Germany and 1.4 percent in Switzerland.  But some relatively high wage countries had quite rapid compensation growth; and, at the low end, Japan saw compensation increase only 0.5 percent per year, even though compensation is about on par with that in the United States.  

The bottom line is that the United States has seen its compensation costs decline relative to compensation in many countries through a combination of a depreciating U.S. dollar and comparatively low domestic (U.S.) compensation growth.  In countries with much lower compensation costs, increases in domestic compensation costs was the dominant factor driving up costs, although appreciating currencies often contributed.  Where currencies depreciated, the increase in domestic compensation costs tended to be larger.  In a number of the higher wage countries, increases in the values of their currencies was the driving force, raising compensation costs relative to the United States even when domestic compensation growth was more moderate.


Although China is the United States’ second largest trading partner and the second largest economy in the world, China was not included among the 33 countries for which BLS calculated manufacturing compensation costs because of the inadequate quality of its data.  However, BLS performed a separate analysis of compensation costs of manufacturing employees in China and found they were about 4 percent of those in the United States in 2008.  This is substantially lower than all of the 33 countries, with the exception of Philippines (5 percent in 2008).

Compensation has been growing rapidly in China, but from a very low level.  From 2002 to 2008, hourly manufacturing compensation measured in yuan doubled.  In addition, starting in the summer of 2005, China allowed a gradual appreciation of the yuan relative to the dollar; this raised compensation measured in U.S. dollars by roughly another 20 percent.  However, in response to the financial and economic crisis and the resulting decline in world trade, China slowed the appreciation of the yuan; so the yuan rose only 6 percent between 2008 and 2011.  If manufacturing compensation continued to grow internally in China at the same rate as before (12 percent per year from 2002 to 2008), the combination of domestic compensation growth and exchange appreciation would mean that compensation in China 2011 was just over $2.00 per hour, about the same as in the Philippines and 6 percent of compensation costs in the United States.


Over the past decade, manufacturing compensation costs in the United States have fallen relative to those in almost all its major trading partners.  A number of its important partners, including Canada, Japan, Germany , France and the Netherlands, have costs that are the same as or higher than those in the United States.  However, manufacturing compensation costs in China are far, far lower, despite rapid internal growth in compensation and some currency appreciation.

Manufacturing Job Creation in the Recovery
 (from June 2009 to December 2012)

January 2013

It has been close to a year since I focused intently on the issues addressed in this website.  Instead, I concentrated on preparing and teaching a seminar on central banking for the International Business School at Brandeis University, as well as several other projects. Nevertheless, I read with a combination of interest, enthusiasm, and skepticism, various reports arguing that manufacturing is making a come-back in the United States.  Enthusiasm - because I think a strong manufacturing sector is important for U.S. economic health, for reasons covered elsewhere in this website.  Skepticism – because the job numbers seemed rather modest.

So now that I am back, focused on manufacturing and how it can contribute to a more vital U.S. economy, I thought I should look at what has been happening recently in manufacturing, particularly with respect to job creation.

As can be seen from the table below, which compares the loss of manufacturing jobs in each recession since the late 1960s with the subsequent recovery, recent job growth has not been impressive.  Over 2 million manufacturing jobs were lost between the National Bureau of Economic Research’s pre-recession peak and recession trough.  As of December 2012, three and a half years after the recession trough, only 263,000 jobs had been gained – an increase of just over 2 percent.

Table 1. Cyclical Declines & Recoveries in Manufacturing Employment (000’s)

NBER      NBER                                   Peak        Decline                    Cumulative change from trough after

Peak        Trough                                 Jobs          Peak- trough            1 year             2 yrs           3 yrs              3.5 yrs

Dec 69

Nov 70







Nov 73

Mar 75







Jan 80

Jul 80







Jul 81

Nov 82







Jul 90

Mar 91







Mar 01

Nov 01







Dec 07

Jun 09







                                                                                          % decline               Cumulative % change from trough

Dec 69

Nov 70






Nov 73

Mar 75






Jan 80

Jul 80






Jul 81

Nov 82






Jul 90

Mar 91






Mar 01

Nov 01






Dec 07

Jun 09






Source: www.bls.gov/webapps/legac/cestab1.htm, as of January 21, 2013

In contrast, the number of manufacturing jobs lost in the 1970 and 1973-75 recessions had been regained three and a half years after the trough.  The recovery after the 1981-82 recession was also much more vigorous, although incomplete.  On the other hand, the recent recovery compares favorably with the experience following the recessions of 1990-91 and 2001.  Manufacturing job losses in both the 1990-91 and 2001 recessions were considerably smaller than in 2007-2009, but in neither case did manufacturing jobs recover.  Manufacturing employment continued to fall after the NBER trough and in the “recovery” from the 2001 recession, it kept falling.  Indeed, manufacturing jobs fell more in the three and a half years after the 2001 recession trough than in the official recession, and the combination of the jobs lost in the recession and three and a half years of recovery now surpasses the job losses in the 2007-2009 recession and subsequent recovery.  So by the standard of 2001 and its aftermath, the recent recovery in manufacturing looks good; but by any other standard it is disappointing.

In contrast to the recession, which affected almost all manufacturing industries, the recovery has been narrowly based.  Only nine out of 22 manufacturing industries (I have split transportation equipment into two - motor vehicles and parts and “other,” which is dominated by aerospace) had more jobs in December 2012 than at the trough of the recession.  The remaining 13 had the same or fewer jobs.  Moreover, most of the nine manufacturing industries that added jobs did not add very many. The recovery to date has been dominated by motor vehicles, fabricated metals, machinery, and primary metals.  All of these industries suffered very severe job losses in the recession; and the recovery, while substantial, represents only 40 percent to somewhat over 50 percent of the jobs lost.

Table 2. Recovery in Manufacturing Jobs by Industry, Jun 2009 to Dec 2012

Largest gains

                                                Thousands                                                                   Percent

Motor vehicles & parts


Motor vehicles & parts


Fabricated metal products


Primary metals






Primary metals


Beverages & tobacco


Plastics & rubber


Fabricated metals


Source: www.bls.gov/webapps/legacy/cestab1.htm. As of January 21, 2013


Nevertheless, other manufacturing industries lost jobs in the recovery, including some that did not fare so poorly in the recession.   Over the five years from December 2007, the peak of the recession, to December 2012, the largest job losses were in furniture, motor vehicles, printing, wood products and computers and electronics, all of which lost between 160,000 and 166,000 jobs. In percentage terms, the biggest losers were furniture, wood, apparel, textile mills, and printing, all of which saw employment fall from a quarter to one-third. No manufacturing industry can be said to have done well in this period. Of the larger industries, employment in food manufacturing held up the best.

Geographically, the strength of the recovery in manufacturing reflects the location of the industries that have enjoyed the most robust recovery.  The manufacturing recovery has been strongest in Michigan and Indiana, and considerably further back Washington, Iowa, Tennessee, and Ohio.   Most of these states, the exceptions being Washington and Iowa, also experienced very steep job cuts in the recession.   Michigan saw its manufacturing jobs fall by 27 percent in the recession and Indiana by 22 percent. In the recovery to date, manufacturing jobs have grown 20 percent in Michigan and 15 percent in Indiana; so even with a strong recovery, manufacturing jobs in Michigan and Indiana are still well below their peak levels.  Of states with large manufacturing bases (more than 50,000 manufacturing jobs), Washington and Indiana have fared the best, losing 3 percent and 5 percent of manufacturing jobs over the five years December 2007 to December 2012.  In contrast, New Jersey, West Virginia and Florida lost roughly 19 percent of their manufacturing jobs over these five years, while Puerto Rico lost more than 25 percent.

In sum, the recovery in manufacturing looks good only in comparison to the very weak recoveries in the 1990s and 2000s.  However, a few industries that fared very poorly in the recession have come back relatively strongly, while others that fared relatively better in the recession have continued to struggle. 

Decline in Manufacturing

June 8, 2011


            In the past year the number of manufacturing jobs in the United States has increased by roughly 200,000. Not surprisingly this increase has been welcomed as an encouraging sign of U.S. economic recovery.  What is a bit surprising is the lack of commentary about the extent of the manufacturing decline in the United States over the past decade, against which the increase of the past year pales into near insignificance.

             Since spring 2000, the number of manufacturing jobs in the United States has fallen by a third - from 17.3 million to 11.7 million.  In contrast, manufacturing jobs fell by less than 1 million in the 1990s and by about 1 ½ million in the 1980s. 

             The loss of manufacturing jobs in the 1980s provoked considerable anxiety. It gave rise to many columns and articles about the hollowing out of the American economy. The simultaneous growth in employment in service-producing activities, and especially fast-food establishments, led to derogatory comments that America was becoming a nation of hamburger flippers.  Concern was expressed that the new jobs did not offer the wages and benefits of the lost manufacturing jobs. Additionally, some feared that loss of manufacturing jobs would lead to loss of technological leadership, as manufacturing was seen as one of the most productive and innovative sectors of the economy.

             Contributing to the alarm were the timing of the job losses and their proximate source. Manufacturing job losses took place in the very severe recession of the early 1980s and many of these losses appeared attributable to import competition, particularly from the Japanese. Not only were the Japanese producing products that were lower cost than those made in the United States, but also Japanese products were increasingly higher quality and technologically sophisticated.

             In the debate over the implications of these developments, I tended to side with those who did not see the situation as so alarming.  I noted that while some service sector jobs paid much less than manufacturing jobs, many service sector jobs, particularly those requiring more formal education, offered high wages – and often a pleasanter working environment. I also observed that the loss of manufacturing jobs did not necessarily entail a loss of innovative capacity in manufacturing or more generally. I referred to the experience of New England, which was then undergoing something of an economic renaissance.  The region had become a center for computer and other high technology industries, even as textiles, shoes and other older manufacturing industries had moved first south and now abroad.  As production processes standardized, it was probably inevitable that those manufacturing activities would move to lower cost locations. Meanwhile, however, the combination of engineering and research skills, entrepreneurial spirit, social and business networks, and risk-oriented financial institutions that had given rise to earlier manufacturing leadership would continue to foster new ideas, new industries and new jobs.

             Now, however, I find myself increasingly concerned about the decline in U.S. manufacturing.  Yet despite the dramatic falloff in manufacturing employment in the past decade, the public reaction has been muted – certainly in comparison with the 1980s.  Political figures still talk about the importance of creating manufacturing jobs, but the passion of the past is lacking.

             Perhaps, as with the boy who cried wolf, people are tired of hearing about the issue.  After all, the decline in manufacturing employment in the 1980s was not, in the end, that large. Manufacturing came back from the steep declines in early years of the decade. Not necessarily the same industries.   Primary metals and the leather goods industries were devastated, but instruments, rubber and plastics, chemicals and printing and publishing all increased over the decade. The beleaguered motor vehicle industry ended up about where it started, although down from the late 1970s.

             Furthermore, the most visible threat to U.S. manufacturing prowess – Japan – experienced its own economic challenges in the following decade.  During the 1980s, Japan seemed a competitive juggernaut, setting new standards for quality and for efficient production practices.  A host of business school programs and business organizations were formed to promote the adoption of Japanese management techniques in the United States.  Throughout the 1990s, however, Japan found itself mired in recession, while losing competitive ground in key industries to its Asian neighbors, especially Korea and increasingly China.

  So perhaps Americans have become skeptical of warnings about the loss of manufacturing jobs and the challenges posed by other countries. Why then have I changed my tune?

Essentially, for the same reasons that people were alarmed in the 1980s – the loss of jobs that provide a reasonably good wage to those who lack college degrees and the potential loss of technological leadership, along with productive capacity. What is different from the 1980s is that the process of de-industrialization has gone much further in the United States, educational attainment in the United States appears to be leveling off, and more and more countries have become global competitors in the sophisticated manufacturing industries that were traditionally seen as areas where the United States had a comparative advantage.

Thus, from a jobs perspective, the simple fact that the decline in manufacturing has been so sharp since 2000, has meant a dramatic deterioration in employment opportunities for those lacking college. Further, in recent years, construction employment has also fallen sharply. Although a  smaller industry, construction also offers relatively good wages for workers lacking college degrees and in the early years of the decade, its expansion provided an alternative for some of those displaced from manufacturing.  While the decline in construction has a large cyclical component, recovery seems distant. Men have been particularly affected by the declines in these two industries.

 While some of the industries that have grown in the past decade offer very attractive earnings, achieving these high earnings requires considerable formal education. In some cases, not just a BA but a professional or master’s degree is expected for the more lucrative opportunities. Examples are finance and insurance; professional scientific and technical services; and, of course, health care.

Meanwhile, educational attainment is leveling off among younger adults.  After rising rapidly from 1950 to 1980, the educational attainment of those in their late 20s and early 30s has increased more slowly in the past two decades.  This leveling off is particularly pronounced among men. The 2010 Census shows that 70 percent of the male population between 25 and 34 did not have a BA degree or better – essentially the same fraction as for males between 45 and 54.  In other words, we are not offsetting the loss of relatively high wage jobs that did not require a college degree by increasing educational levels. Despite advances, black and Latino men lag in educational attainment.

 The poor earnings prospects of men who lack college have significant social as well as economic consequences.  Women, who have made major educational advances in recent decades, are reluctant to make permanent commitments to partners whose economic future lacks promise. The result is more households lacking the stability and the institutional support that existed for the traditional family. 

 The bottom line is that the loss of well-paying (manufacturing) jobs, which seemed a serious challenge in the early 1980s, has grown. We have neither found alternative opportunities for those who prefer physical work and do not have college degrees nor dramatically altered either skill levels or attitudes about what constitutes meaningful work

 Even more disturbing is the possibility that those who warned about the loss of U.S. technological leadership, along with manufacturing capacity might have been right.   As noted, New England’s economic history suggested that the standardization of production processes would result in the migration of manufacturing jobs to lower cost locations, but that a region could still retain a competitive edge in research and more technologically advanced manufacturing industries. Further, the feedback between new scientific and technological discoveries, on the one hand, and the production experience, on the other, could lead to important advances.  The region’s success in information technology in the 1970s and 1980s and in biotechnology and pharmaceuticals in the 1990s and 2000s shows the plausibility of the argument.

However, the loss of manufacturing in the United States has been so deep that it calls into question whether sufficient production remains to provide the synergies with research that lead to new breakthroughs.  A critical element of the New England story – and also that of Silicon Valley - was the importance of a diverse mix of activities in close physical proximity, such that insights from one activity could cross-pollinate other activities resulting in new and better ways of doing things. As the nation loses ground in manufacturing industry after manufacturing industry, it loses the opportunity to take advantage of such synergies.

Clearly, there are exceptions. The United States remains a leader in biotech and pharmaceuticals, where research process continues to support a meaningful manufacturing presence.  And medical instrumentation continues to draw from strengths in information technology, precision manufacturing and medical research. On the other hand, in green technologies, production has moved overseas, especially to China, just as we are finally seeing some of the technological breakthroughs that advocates have long promised.  And these technological breakthroughs are increasingly coming from the overseas producers. 

 Dependence on foreign sources for rare earths, while not a manufacturing example, illustrates the challenge. Rare earths are used in many high technology products, such as superconductors, lasers, aerospace products, and batteries, including batteries for electric vehicles.  According to a recent article in MIT’s Technology Review (Bourzac, May/June 2011), the technology for extracting rare earths was developed in the United States and the biggest mine in the world was once here. Now, however, almost all rare earth production is in China.  Admittedly, China’s dominance reflects a disregard for the environmental consequences of rare earth mining. But not only does China control the world’s rare earth production, it is also a leader in the production of many highly sophisticated products that are based upon rare earths – products that one might otherwise expect to be made in a more advanced country.

When one looks at trade patterns, it is striking the degree to which the United States imports more of what were once considered high technology products than it exports. In many cases, production processes have become standardized and these are no longer the cutting edge products they once were. Even so, the patterns suggest that other countries have made important advances in areas that were once U.S. specialties. U.S. exports are still more oriented to capital goods than its imports.  Capital goods account for over a third of U.S. exports and just under a quarter of  U.S. imports.  However, total imports are much larger than total exports; so the dollar values of capital goods imports and exports are roughly equal.  Within the capital goods category, the United States is a net importer of computers and computer accessories and telecommunications equipment. It is also a net importer of pharmaceuticals, which is in the consumer products category. It is a net exporter of semiconductors, civilian aircraft and aircraft engines, and industrial machines and engines.

Not so long ago, one might have taken comfort from claims that the U.S. comparative advantage has shifted away from high technology manufacturing to financial expertise. The United States is a net exporter of services, and the perceived expertise of U.S. financial institutions and the liquidity and depth of U.S. financial markets have caused global investors to put their funds in U.S. assets.  This has created jobs in financial services and industries, like construction, that benefited from these investments. Capital inflows also boosted the value of the dollar, with a side effect being a weaker manufacturing sector.  The 2007-2008 financial crisis has now called into question whether such reliance on financial expertise and innovation has served the nation well.  I fear that the financial activities of the past decade may have resulted in a misallocation of resources rather than a more efficient allocation, one that has undermined U.S. manufacturing competitiveness and that may have long term consequences in terms of future productivity growth and advances in economic well-being.

Reviving the Manufacturing Sector in New England  

June 16, 2011


For at least 30 years, many economists and economic development specialists have been concerned about the decline in manufacturing in the United States.  From 1980 to 2000, the decline was largely relative to the growth in the rest of the economy. The actual number of manufacturing jobs did not fall all that much – from 19 million to 17 million.  Today, however, fewer than 12 million workers are employed in manufacturing.

Concern over the decline in manufacturing focuses on two issues. Manufacturing provides relatively high wage employment for workers who lack college degrees. Additionally, some, including the writer, are concerned that loss of production capacity will undermine U.S. leadership in research and development.  In essence, they argue that the production process provides opportunities for learning by doing and that at giving up “doing” risks ceasing to be on the cutting edge of learning.

How then might one revive manufacturing? Two reports have recently addressed this question from a New England perspective – one by Deloitte Consulting for the New England Council (Reexamining advanced manufacturing in a networked world: Prospects for a resurgence in New England) and one by the Center for Urban and Regional Policy at Northeastern (Staying Power: the Future of Manufacturing in Massachusetts) and come to relatively positive conclusions.


Rather surprisingly, in light of the dramatic job losses in manufacturing in the past decade, both reports argue that a high priority must be developing the skills needed for a manufacturing workforce and attracting more young people into manufacturing.

Given job losses, one might think that there would be plenty of former manufacturing workers available to work in the sector. But surveys and interviews with manufacturing executives indicate that is not the case.  Admittedly, these interviews took place before the recent Great Recession; but even before then, manufacturing employment had been falling.

Some of the problem may be a skills mismatch, with successful manufacturers requiring higher skills than possessed by workers displaced from manufacturing.  However, the Northeastern report found that most manufacturing executives think that a high school degree is sufficient for a majority of their workers.  That said, workers still need strong vocational skills. The manufacturers interviewed think that young people are being discouraged from taking vocational training because so much emphasis is placed on attending college.  While a college degree is associated with a substantial boost to earnings, many young adults, increasingly men, do not succeed in acquiring a degree.

Recruiting skilled craftsmen is especially difficult. Often, these individuals have acquired their skills through on-the-job experience; so there may be a large industry-specific or even firm-specific component to their knowledge.  Further, since many of the more highly skilled manufacturing workers are older, they may prefer retirement to making the adjustments that a new firm is likely to require.

Nevertheless, to this writer, it seems probable that the shrinkage of the manufacturing sector and a loss of critical mass are also factors. The pool of workers with manufacturing experience is simply not as deep as it once was and referral networks have become thin. Further, manufacturing today tends to be located in suburban locations, where many young people do indeed aspire to careers requiring college degrees and may regard manufacturing as an unappealing option. At the same time, less advantaged young people in city locations who might be attracted to manufacturing may have difficulty learning about and accessing these jobs.

 A project at the Federal Reserve Bank of Boston on the challenges facing the city of Springfield, Massachusetts, highlights some of the issues.  Interviews and a survey of Greater Springfield employers  asked about their experience hiring entry-level workers.  These employers included manufacturers, but they were not limited to this sector. The employers did not face a shortage of candidates. However, many candidates were not ready to work.  Attendance problems and poor work attitudes were common.  Although the employers generally did not require more than a high school degree or a GED, many candidates lacked the skills such credentials should confer.  In response, employers placed high premiums on prior work experience and on referrals from existing workers. This reliance on prior experience and referrals put potential job candidates in Springfield’s downtown neighborhoods at a disadvantage in competing for these openings since employment rates in these neighborhoods were low.  Downtown residents were less likely to have prior experience or to know someone who could refer them. Transportation to work was also a problem.  A car was needed, since entry-level openings are often on off-shifts.

 The basic point is that with manufacturing accounting for only 10 percent of employment rather than 20 or 30 percent, traditional ways of recruiting or looking for jobs – often reliant on friends and family members – are not going to work as well as the once did. Thus, part of a strategy to revitalize manufacturing has to involve recruitment and training.


             The New England Council study, which focused on advanced manufacturing, stressed the importance of networks as an important competitive advantage of the New England region.  Larger manufacturers are able to draw upon a web of smaller suppliers of components, parts and services located in the region. These networks, the report argues, help develop products faster and at lower cost. The web of inter-connections tends to be self-sustaining, since re-creating those relationships in a different location would be difficult.

             However, while networks can be self-sustaining, logic argues that the loss of critical elements of the networks can have ramifications far beyond the individual firms.  Most obviously, if the large manufacturer at the center re-locates, it will have adverse consequences for all its regional suppliers. But additionally, the loss of several important suppliers in a network might be the tipping point for a large firm that was contemplating alternatives. More generally, networks mean that once the productive capacity is gone, rebuilding will be very difficult.  It is not just one firm that must be recruited or re-created, but many. 

Thus, a strategy to revitalize manufacturing in a field where networks are important must keep these linkages in mind.  Developing inventories of high quality suppliers, making introductions and helping smaller firms market themselves to potential customers may be part of the answer.  Cooperation among neighboring states seems essential to developing a sufficiently large network of suppliers. Furthermore, for small and medium-sized family-owned firms, succession planning can be an issue.  Helping to find potential buyers for family-owned businesses could help preserve important elements of the network.


             Both the New England Council and Northeastern studies place considerable emphasis on the importance of marketing manufacturing and branding New England and Massachusetts respectively as “manufacturing friendly” places.  To a large degree, the argument for branding is linked to workforce challenges.  Public perceptions of manufacturing as a dirty and declining industry make recruiting workers, particularly younger workers, difficult.  Manufacturing CEOs seem to think that negative or, at best, passive governmental attitudes towards manufacturing have contributed to this image and they would like to see a more positive portrayal of their sector.

             The Northeastern report observes an additional phenomenon, however. Manufacturing CEOs resent what they perceive as a lack of respect or interest from government officials.  They feel they are being written off in favor of other, more glamorous sectors – like the life sciences or financial services. Compared to the direct costs of doing business and difficulties attracting workers, it seems doubtful that perceptions of government attitudes could have a major effect on the future of manufacturing in the region.  But CEOs are people and at the margin, a sense of being unappreciated could tip the balance towards relocating or expanding elsewhere.  And certainly, perceptions of hostility or indifference on the part of government officials could kill off any consideration of moving into an area. 

  One interesting observation in the Northeastern report was that business CEOs may persist in holding their own negative views even after concerns have been addressed.  In their survey of manufacturing CEOs, the Northeastern researchers found that workers’ compensation was ranked very high as a problem, even though the state had made major changes that had greatly reduced costs.  This suggests that frequent interactions between government officials and business leaders may be important.  Not only might such forums address CEOs’ views that business does not receive sufficient respect, but they might also provide an opportunity to confront misperceptions about the true state of affairs.

 The Northeastern report also suggests a note of caution. Business leaders, especially those from smaller firms, do not take advantage of training and other opportunities.  One cannot assume that if the state offers forums that they will come – at least not without aggressive outreach.  This should not be so surprising.  Smaller manufacturers have thin management staffs and their time is devoted to running the business. But it makes outreach difficult.


Productivity and the Decline of Manufacturing:

Why worry?”

June 23, 2011


Some analysts are more sanguine about the decline in manufacturing jobs, seeing these losses as the inevitable consequence of large productivity gains in manufacturing.  They point out that manufacturing’s share of value added has not fallen as much as jobs.  They also note large manufacturing profits.  In some cases, they draw comforting comparisons with the decline in agricultural employment in the first half of the 20th century.  Advances in agricultural productivity have allowed the United States not only to feed a much larger population but also to supply export markets, even though the fraction of the workforce employed in agriculture has fallen from 40 percent in 1900 to about 2 percent today.

Productivity gains in manufacturing have been large, and real value added has not fallen much. Manufacturing’s share of real value added declined from 12.1 percent in 1998 to 11.4 percent in 2009, and all of this decline can be attributed to the recent recession.  Manufacturing’s share of total real value added was 12.8 percent in 2007. However, this picture of stability is very strongly influenced by the computer and electronic products industry, for which the share of real value added is estimated to have increased from 0.4 percent in 1998 to 2.3 percent in 2009.  Almost every other manufacturing industry saw its share of total value added decline; one exception was petroleum and coal.


Table 1.  Comparing Manufacturing’s Share of Employment and Real (2005$) and Nominal Value Added (percent)


                                                1998                        2000                        2007                        2009


Manufacturing share of           

Employment                             13.3                        12.6                         9.7       8.7

Real value added                        12.1                        12.4                       12.8                    11.4

Nominal value added            15.1                        14.2                        12.1                    11.2           


Computer & electronic share of

Employment                              1.4                        1.3                          0.9                        0.8           

Real value added                        0.4                        0.7                          1.9                        2.3

Nominal value added            1.7                        1.7                          1.4                         1.5


Source: Author’s calculations from Bureau of Economic Analysis (BEA) Industry Economic Accounts, GDP by Industry




The rise in value added in manufacturing is due in large part to major quality improvements in computers.  These “output” gains are characterized as a decline in computer prices.  Because the capabilities of the computers of the past were much lower than those of computers of today, we treat them as much higher cost and deflate historic computer output by a much higher price index to derive a measure of real value added that is comparable to today.  If we look at nominal value added, we see a modest decline in the computer industry’s share of value added, from 1.7 to 1.5 percent, rather than a sizable increase, and a much more pronounced decline in the value added of the total manufacturing sector – from 15.1 percent of value added in 1998 to 12.1 percent in 2007 and 11.2 percent in 2009.

 The point is not that the quality improvements and productivity gains in computers and some other manufacturing industries are not real.  But the real value added figures, which show a very large increase in real computer output and a modest decrease in total manufacturing output, present a somewhat misleading impression of the health of the industry.  One would generally associate increases in real value added with an expanding industry.  One would normally expect to see strong growth in dollar sales.  But that is not the case for computers.


Table 2. Components of Value Added for All Industries, Manufacturing, and Computers and Electronic Products, 1998 and 2009 (billions of dollars)


                                                1998                        2009                        Percent change


All industries

  Gross output                               15987                        24804                                   55

  Intermediate inputs                      7194                        10685                                   49

  Value added                                  8794                        14119                                   61

  Price index                                      85.5                        109.6                                    28

  Real value added                       10284                       12881                                   25



  Gross output                                 3840                        4522                                    18

  Intermediate inputs                      2513                        2938                                    17

  Value added                                 1327                        1585                                    19

  Price index                                    106.5                       107.8                                    1

  Real value added                        1246                       1470                                    18


Computers & electronic products

  Gross output                                     430                          353                                    -18

  Intermediate inputs                          276                          147                                   -47

  Value added                                     154                          207                                     34

  Price index                                        397.7                        70.3                                 -82

  Real value added                                39                          294                                  654


Source: Author’s calculations from Bureau of Economic Analysis (BEA) Industry Economic Accounts, GDP by Industry



The dollar value of gross computer output was almost 20 percent less in 2009 than in 1998.  Gross output was down substantially even before the recession.  Nominal value added did not actually fall, however.  A second unusual feature of the computer industry and one that may distort impressions of the industry’s contribution to the U.S. economy is that the dollar value of intermediate inputs to computers has fallen dramatically.   Value added equals gross output less intermediate inputs.  In the case of computers, nominal gross output has declined by about 20 percent since the late 1990s but the nominal value of intermediate inputs has fallen close to 50 percent. This results in an increase in nominal value added of roughly a third.

  The increase in nominal value added in the computer industry has gone disproportionately to “gross operating surplus” including profits, with a small increase in compensation in the form of higher wages.  The number of workers in the computer industry has fallen sharply – from 1.8 million in 1998 to 1.1 million in 2009.

These patterns are generally consistent with much of the conventional wisdom about the U.S. computer industry: focusing more on research and development in this country and relying on low cost imports for more routine components.  Thus, Apple imports much of its product from abroad and makes large profits from being able to sell a high quality product that is produced at low cost abroad. U.S. manufacturing gets credited for big productivity gains for the increase in output delivered by the Apple product.  But our system does not fully recognize that a lot of the profit comes about because low cost foreign producers are substituting for higher cost American suppliers.  A missing piece in the puzzle is that we are attributing most of the productivity gain from higher quality to Apple and other firms headquartered here, whereas key enablers are the foreign producers of important components. This may be a correct allocation; but over time, the source of productivity gains may shift as foreign producers improve their own innovation capabilities. 

 The sharp fall-off in spending on intermediate inputs is not characteristic of manufacturing generally. A few industries show similar, but much less pronounced tendencies, and for others, the value of intermediate products has increased relative to gross output.

The point is simply that we should not dismiss concerns about the decline of manufacturing because real value added has not declined.  These figures are strongly influenced by developments in one particular industry that has very unusual features that may present a misleading picture of what is actually taking place in the manufacturing sector as a whole.  Even in computers, we should fear complacency.  The industry continues to make impressive productivity gains, which should provide many benefits to users, but the computer industry’s direct contributions to U.S. employment, compensation, and even profits are modest given its contribution to real value added. 


 Observations on Science and Engineering Indicators 2010

August 15, 2011


The National Science Board's Science and Engineering Indicators 2010  (SEI) contains a wealth of information on science and engineering activity in the United States and other countries – statistics on R&D expenditures, educational achievement, science and engineering occupations, patents, scholarly articles, and a great deal more.  This is a very valuable resource.

I read SEI looking for explanations for the decline in manufacturing employment in the United States and for what I fear may be an erosion of U.S. technological leadership.  I did not find a smoking gun, although changes in the composition of R&D may have affected the sectors in which jobs are created.  It is also striking the degree to which the United States looks to those born abroad for its advanced science and engineering (S&E) workforce. Making it easier for foreign-born S&E workers to become permanent residents and citizens would allow them to more fully develop their talents – to the benefit of all. 

The most disturbing aspects of SEI are the international comparisons of educational attainment and achievement. The United States’ educational advantage over the rest of the world is fast diminishing. This bodes poorly for future U.S. technological leadership, not just in manufacturing.  And even if the United States is able to retain its innovative edge, by virtue of its economic and political systems, the poor performance of U.S. students on international tests and the lack of progress in raising educational attainment suggests that the fruits of U.S. innovative leadership are likely to be captured by relatively small numbers of highly trained professionals and executives and not by the public at large.

No Smoking Gun

SEI does not show major changes in U.S. science and engineering activity that one would expect to cause a decline in manufacturing prowess.  R&D spending has not fallen relative to GDP or relative to expenditures in most other countries.  Nor does one see a falloff in science and engineering graduates at U.S. colleges and universities.  Although a substantial fraction of science and engineering students are foreign-born, particularly at the graduate level, most of these foreign students plan to stay in the United States – and seem to succeed in doing so.  But some shifts provoke questions.

Research and Development

 U.S. R&D spending was just under 3 percent of GDP in 2008.  This share approaches or equals past highs in the 1960s, 1980s and late 1990s.  The federal government’s share of total R&D has fallen dramatically, however.  Whereas the federal government accounted for about two-thirds of R&D spending in the 1960s, and close to half in the 1980s, it now accounts for about a quarter.  Business funds most of the rest of R&D. 

Most business spending on R&D is carried out by business. Additionally, some business R&D is funded by the federal government.  (In 2008, the federally supported share was less than 10 percent – down from about half in the 1960s.) The big R&D performers in the business sector are chemicals (especially pharmaceuticals), computers and electronic products (especially semiconductors and instruments), software and computer services, aerospace and defense manufacturing, R&D services, and automotive. 

   Most business R&D is development, with some applied research and very little basic research.  Despite the growth in business funding of R&D, however, the share of total R&D going to basic research has not declined over the past 40 years. Indeed, at 17 percent in 2008, the share of basic R&D was high by historic standards.  

About 60 percent of federal spending on R&D is currently going to defense.  The defense share was much higher in the 1960s and somewhat higher in the 1980s. It was lower in the 1970s and 1990s. The shares of nondefense (and total) federal R&D spending going to health have increased sharply over the past 30 years.  The shares going to economic development R&D, including energy, have fallen sharply.

   The U.S. ratio of R&D to GDP, at 2.7 percent in 2007, is higher than ratios in most counties, but not as high as the ratios in Japan and Korea.  In the past ten or so years, South Korea’s R&D share has risen rapidly to about 3.5 percent of GDP.  R&D has also risen rapidly in China but from a much lower level. R&D is now about 1.5 percent of GDP in China.  In terms of the dollar value of R&D spending (as measured by purchasing power parity), the United States ranked first by a very large margin in 2007. Japan was second and China was third. Excluding defense, the U.S. margin of superiority over most countries was not as wide. U.S. non-defense R&D spending was 2.2 percent of GDP in 2007, compared to 3.4 percent for Japan. (Non-defense R&D figures are not available for China and may not be relevant, given the importance of state–owned enterprises with links to the military.) 

   On the surface, at least, these trends do not seem hostile to innovation and technological progress in manufacturing in the United States.  Indeed, one could argue that shifts to more business funding of R&D and away from defense, should result in more commercially practical applications.  (I recall discussions in the late 1980s in which it was argued that Japan’s greater focus on business research, as compared to defense, was a competitive advantage.)  At the same time, the shift to more business R&D has not caused the share of R&D going to basic research to fall, suggesting that we are not trading away opportunities for major research breakthroughs in exchange for incremental gains.

The increasing focus of federal research on health is potentially meaningful. It makes sense for the federal government to emphasize research where the social payoff is large relative to private gain – although large R&D expenditures by the pharmaceutical industry indicate that private gains are also substantial.  And it should not be surprising if the returns come in the form of health-related outcomes and not necessarily in employment-generating activities.  However, it seems plausible that we have seen a payoff in employment generation. Surely, some of the growth in employment in the health care industry is due to innovative treatment options that have been made possible through health-related R&D.  The job creation may even extend to manufacturing, as employment has held up better in medical devices and pharmaceuticals than in many other manufacturing industries.   But the way in which health-related R&D affects the economy and generates employment is probably different from the stimulative effects of past R&D expenditures on, for example, information technology or aerospace.  In particular, we tend to view the expansion of the healthcare system as a double-edged sword.  While we value the improvements in health outcomes, we consider increasing health care costs to be a burden.

It is also rather striking how much the share of federal R&D going to economic development including energy has fallen, given all the rhetoric about the need for developing new energy sources over successive presidential administrations.  In 1981, 36 percent of federal nondefense R&D and just over 15 percent of total federal R&D went to economic development.  In 2007, 10 percent of nondefense and 4 percent of total R&D went to this category.  Meanwhile, health and the environment increased from 31 to 55 percent of nondefense R&D (and from 14 to 23 percent of total.)

In sum, as a nation, we have not cut back spending on R&D. The federal government is doing relatively less and the business sector is doing relatively more. To the degree that business spending is more focused on commercially practical activities, this could be a positive for manufacturing.  Federal spending has shifted towards health. Although job creation is not the objective, increased federal support for health research has probably contributed to the growth of the health care industry.

Science and Engineering Workforce

 Explanations for the blossoming of the information industry in New England and its transformation of the regional economy emphasize three critical elements:  new technologies that grew out of defense-related research in WWII and the Cold War; elite research universities where the brightest students from across the country performed much of this research; and the coming-of-age of the highly educated baby boom generation, which provided workers with cutting-edge skills for the new firms and industries built on the emerging technologies and which also contributed to the expansion of the research and education infrastructure.

Since a key part of the narrative is the entry into the labor force of large numbers of young adults trained in state-of-the-art science and engineering (S&E) disciplines, a natural question for someone concerned about the nation’s innovative capacity is the current and future supply of scientists, engineers, and others with the skills needed by technologically advanced industries. 

According to SEI, the number of BAs in S&E relative to the population of 20-24 year-olds has increased modestly since the 1970s, as has graduate enrollment in S&E relative to the population aged 25 to 29.  Since the early 1990s, S&E BAs have accounted for just under a third of all degrees awarded.  Engineering accounts for about 5 percent, natural sciences about 10 percent and the social sciences including psychology 16 percent.  Biology (natural science) has become more popular. Computer science has had its ups and downs, increasing in popularity around the century date change and then decreasing. Foreign students on temporary visas account for about 4 percent of S&E BAs, although more in some fields.

The number of Master’s degrees in S&E has increased since the early 1990s, but not as fast as Masters in other fields.  Engineering has increased relatively slowly; biology and psychology have increased relatively fast.  Foreign students on temporary visas account for about one-quarter of all science and engineering Masters (although just over 10 percent of all Master’s degrees.) Almost 40 percent of engineering and computer science degrees go to foreign students on temporary visas. The number of doctoral degrees in S&E has also increased.  Growth has been especially rapid in medical and life science, a category for which data only exist at the doctoral level.  Interestingly, this field is made up almost entirely of U.S. citizens and permanent residents, with women outnumbering men two-to-one. In contrast, 60 percent of those receiving doctorates in engineering are foreign students on temporary visas.  For all science and engineering doctoral degrees, a third go to foreign students on temporary visas.

A potential concern is that these students on temporary visas, who account for disproportionate shares of the most highly trained scientists and engineers produced by U.S. universities, will return to their countries of origin and the United States will not benefit from their advanced learning.  However, surveys indicate that roughly three-quarters of foreign recipients of doctoral degrees planned to stay in the United States in 2007 and about half had job offers or post-doctoral research opportunities. Foreign students have accounted for a majority of post-doctoral fellows in the United States since the early 1990s. 

About 5 million college educated workers are employed in science and engineering occupations –  4 percent of the total workforce.  This is roughly one-third the number of workers who have degrees in science and engineering.  The disparity between S&E occupations and degree holders is most pronounced for those in the social sciences, least for engineers.  However, many S&E degree holders in non-S&E occupations are engaged in activities related to their degree, such as management or sales or other non-S&E activities for which their training was relevant.

 About 25 percent of college-educated S&E workers are foreign and 40 percent of S&E workers with doctorates are foreign. Immigration of S&E workers trained abroad, as well as U.S. training of native and foreign-born students, has contributed to the number of people in S&E occupations, which has grown faster than the overall workforce since the 1950s.

Women and minorities account for increasing shares of the workforce with S&E degrees.  In 2006, only 35 percent of workers with S&E degrees under the age of 30 were white males, while almost 60 percent of those with S&E degrees over age 50 were white males.

Almost half of all workers with S&E degrees work for for-profit firms. Interestingly, the next largest category – 17 percent – is self-employment.  This is larger than the self-employed share of the total workforce (11 percent).  Additionally, the self-employed with S&E degrees are more likely to have incorporated their businesses.  Over half of S&E self-employed are incorporated, compared to one-third for all self-employed.

SEI points out that it takes a long time for foreign-born S&E workers to become U.S. citizens. Most of those who stay in the United States eventually do. In  2003, about 90 percent of foreign S&E workers who had arrived before 1980 were citizens; almost all the rest were permanent residents. However, only 42 percent of those who had arrived in 1993 had become citizens by 2003; 46 percent were permanent residents and 12 percent were on temporary visas.  Among those who arrived in 1998, just 9 percent had become citizens by 2003, 43 percent were permanent residents and 49 percent were on temporary visas.

As in the case of R&D spending, these data on science and engineering workers do not seem especially ominous.  Certainly, the ability of the United States to attract students in science and engineering, as well as trained professionals, from all over the world should be favorable to our technological leadership. On the other hand, the prominence of foreign-born workers in S&E occupations raises the question of why native-born workers do not pursue these careers to a greater degree.

Additionally, given U.S. dependence upon foreign-born scientists and engineers, the time required for these workers to become citizens or even permanent residents is excessive.  Being in a temporary visa status not only limits the opportunities available to these individuals but also prevents the nation from fully benefiting from their talents.  Most obviously, as SEI points out, many S&E jobs in the federal government require U.S. citizenship. Some private firms involved in defense and other classified work may also be restricted to hiring U.S. citizens. Additionally, foreigners who want to work in the United States and who do not have family members already in the country depend upon employers to apply for their visas. This dependence upon the employer for visa applications is an impediment to foreign workers on temporary visas changing jobs and restricts their advancement and productivity growth.  Some may eventually return home. (Changing jobs is possible if the new employer makes a new visa application, but this is a serious barrier to securing employment and likely an impediment to looking.) Foreign-workers on temporary visas are also prevented from starting their own businesses – unless they have substantial financial resources and can take advantage of a special visa for immigrant investors.  Yet, as noted above, being an entrepreneur is an attractive career choice for many S&E workers – one that may have substantial spillovers in terms of innovation and job creation. The bottom line is that the United States imposes significant hurdles to foreign S&E workers becoming permanent residents and citizens even though it has become dependent upon their skills.

A Future Smoking Gun

International Educational Attainment and Achievement 

  International comparisons of educational attainment and achievement are the truly scary message in Science and Engineering Indicators.  Given U.S. global leadership in many indicators of technological achievement, such as patents, scientific articles, trade in intellectual intangibles – all documented in SEI, it is almost shocking to see how poorly U.S. students perform in international comparisons of proficiency in mathematics and science. Tests of students in Grades 4 and 8 (Trends in International Mathematics and Science Study – TIMSS) show the United States ranking in the lower middle of participating countries, while tests of 15 year olds (Program for International Student Assessment – PISA) show the United States ranking close to the bottom.  TIMSS shows a slight relative improvement over time, while PISA does not. The United States is not the only country with a reputation as a high technology center that ranks poorly in these tests.  Israel is another.  But for the most part, most countries that the United States would regard as competitors in knowledge-intensive manufacturing and services industries (and many that it would not) rank higher in international educational performance. Interestingly, the United States rates worse on the PISA, which is said to place more emphasis on problem-solving skills than rote learning.

  The United States has an advantage over most other countries in terms of the fraction of the population with college degrees.  However, that margin has substantially diminished – and will diminish further - as can be seen by comparisons of educational attainment for younger adults (those aged 25 to 34) and the entire population of normal working age (25 to 64.)

 In 2006, 29.9 percent of U.S. residents 25 to 64 had BAs (or its equivalent) or better, compared to an average of 19.3 percent for OECD nations. Only one nation, Norway, surpassed the United States, with 30.5 percent of its population having the equivalent of a BA or better. Israel at 29.8 percent was close.

For younger people (25 to 34), the United States has a lot more company. College educational attainment in the United States is the same for those 25 to 34 as it is for those 25 to 64. This means that younger people are no more likely to have a college degree than their parents. In contrast, in most other countries, the younger group is substantially more educated than their elders. A particularly striking case is South Korea. Only 23.5 percent of Koreans age 25 to 64 have BAs; but among younger adults (25 to 34), 32.9 percent have BAs – a higher fraction than in the United States.  Thus, even if Korean young people make no further educational gains, the educational attainment of the Korean workforce will rise with the passage of time as older people retire and are replaced by more highly educated younger workers.  Meanwhile, educational attainment in the United States stays the same.

 For OECD countries, the average share of 25 to 34 year-olds with a BA was 24.6 percent in 2006 compared to 29.9 percent in the United States. – half the gap for the population 24 to 65.  In six countries the educational attainment of younger adults surpassed that in the United States – Norway (39.8), Israel (34.8), Netherlands (34.3), Korea (32.9), Denmark (31.7) and Sweden (30.6). Another five countries were close, with over 29 percent of younger adults having BAs. 

If one includes those with associates degrees and advanced technical skills, the picture is essentially the same, although the list of countries surpassing the United States is somewhat different as some countries put more emphasis on technical training than others.

 The challenge for the United States is that a major competitive advantage – a highly educated population - is eroding.  Other countries are catching up in terms of educating younger adults, while the United States has not made much progress since the baby boom generation. At the same time, it appears that K-12 students in the United States are not mastering skills in mathematics and science as well as students in many other countries.  This does not bode well for future U.S. leadership in knowledge intensive industries.  But even if the United States is able to maintain its leadership position - perhaps because its economic system is more supportive of innovation or because of the quality of its research universities – these patterns are troubling. A population that is not improving its skill levels is unlikely to reap the financial rewards from these innovations to the degree that American workers did in the past. Rather, the fruits of innovation will go disproportionately to a relatively small elite of professionals and senior executives and the nation will continue to debate what to do about the loss of middle class jobs and rising income inequality. 


 For anyone interested in technology and innovation in the United States, the NSF’s annual Science and Engineering Indicators is a valuable resource, containing information on a host of indicators relevant to U.S. performance over time and in comparison to other countries.  This note summarizes my observations in three areas – R&D expenditures, the science and engineering workforce, and international educational comparisons. However, there is a lot more useful material, including data on individual states.

I was somewhat reassured by R&D patterns: there has not been a decline in R&D spending compared to GDP.  Business is doing relatively more and the federal government is doing relatively less.  Federal R&D spending has also shifted substantially towards health.  These trends do not seem inherently worrisome, although the shift towards health seems likely to have affected the sectors that benefit most from federal R&D. Other countries, notably China and South Korea, have substantially increased R&D spending in the past two decades, but the United States still ranks relatively high.

Nor has there been a decline in the science and engineering workforce or in degrees granted. However, foreign students on temporary visas compromise a large fraction of the graduate student body in S&E fields and foreign workers make up a disproportionate share of workers in S&E occupations.  Given the nation’s dependence upon foreign-born scientists and engineers, U.S. immigration policy should be more welcoming.  These highly educated workers, who are critical to U.S. future technological leadership, should be able to become permanent residents and citizens more rapidly.

What I did find disturbing were international comparisons of educational attainment and performance.  While the United States still ranks high in terms of the fraction of the population with college degrees, its standing is not impressive if one just looks at younger adults.  Other nations have made major advances, while the United States has stood still. Further, in tests of science and mathematics, U.S. teenagers compare poorly with their counterparts in many countries.  

Web Hosting Companies