Richard Katz

A decade ago, the great American jobs train fell off its tracks. Traditionally, boosts in private-sector employment have accompanied recoveries from economic downturns. In the first seven years after the beginning of the 1980 and 1990 recessions, for example, the number of private-sector jobs increased by 14 percent. Yet in January 2008, seven years after the previous pre-recession peak and before the most recent recession began, private-sector jobs were up only four percent. Today, for the first time in the postwar era, there are fewer of these jobs than there were ten years before.

Ignoring the overall dearth of jobs, Michael Spence (“The Impact of Globalization on Income and Unemployment,” July/August 2011) singles out the fraction of employment in sectors related to trade. He claims that China and other developing countries have taken U.S. jobs and blames globalization for the substantial increase in income inequality across the country. It is misleading, he says, to argue that “the most important forces operating on the structure of the U.S. economy are internal, not external.” He is wrong: the fault lies not in China or South Korea but at home.

It is true, as Spence claims, that the tradable sector of the economy has lost many jobs in the past several decades. (According to Spence, the tradable sector produces goods and services that can be consumed anywhere, such as manufacturing, whereas the nontradable sector produces goods and services that must be consumed domestically, such as health care). But that is like saying that children with big feet are better readers than children with small feet. It would help to know that the kids with big feet are fifth graders and that those with smaller feet are second graders. The lion’s share of tradable products consists of manufactured goods, leading Spence to conflate the fate of manufacturing jobs with that of jobs in the tradable sector. The United States lost 4.4 million manufacturing jobs between 1990 and 2008. Since Spence says that the total number of tradable-sector jobs grew by 600,000 during this period, then the United States must have gained five million jobs in the tradable sector outside of manufacturing. That number represents about a fifth of all private-sector job growth. In short, the key determinant of whether jobs grew or fell was not whether they were in the tradable sector but whether they were in manufacturing -- just as the key determinant of reading ability is not foot size but grade level.

As economies mature, manufacturing jobs sooner or later shrink, and this happens with or without globalization. This pattern holds true across a spectrum of rich countries, including those with significant trade surpluses. Even in Germany, where, according to Spence, the government protected factory jobs from international competition, such jobs fell by 27 percent between 1990 and 2008. During the same period, factory jobs fell by 26 percent in Japan, 24 percent in France, and 24 percent in the United States (although U.S. manufacturing output grew by 74 percent). In the former West Germany, factory jobs began falling back in 1971, meaning that the disruption of German reunification in the 1990s merely exacerbated a long-standing trend.

The explanation for the shrinkage of factory jobs is well known. As economies mature, people tend to spend less on goods and more on services. Meanwhile, productivity growth tends to be higher in manufacturing than in services. Suppose demand for manufacturers goes up by 15 percent but output per worker goes up even more, say, by 25 percent. In that case, jobs will drop by ten percent. Manufacturing mirrors farming: just a tiny sliver of the work force can now feed the entire country.

Meanwhile, if China is stealing U.S. manufacturing jobs, why are there 12 million fewer factory jobs in China today than there were in 1996, the peak year for such jobs, even as manufacturing output has tripled? The reason is that reforms boosted efficiency, thereby making cheaper goods available.

Spence’s inaccurate diagnosis has led him to some prescriptions that could make the patient even sicker. His proposal that the United States should protect jobs at the expense of efficiency and wage growth -- as if it could not promote all three at the same time -- essentially advises the United States to copy the approach that Japan took in the 1970s and 1980s. This is what ultimately led to the two decades of anemic growth in Japan that began in the early 1990s.

Although some of Spence’s remedies, such as improving education, are absolutely correct, restricting trade to preserve jobs would be counterproductive. For any given level of demand for manufactured goods, the only way to create more jobs is to lower each worker’s output. Between 1990 and 2008, manufacturing output per worker increased by 135 percent in the United States, but it increased by only half that in Germany and Japan. In those two last countries, lower productivity growth made goods more expensive, thereby hurting real wages across their economies. It also pressured employers to restrain wages in order to compete in international markets. That is part of the reason why real wages fell by 4.5 percent in Germany and by five percent in Japan over the past decade. Slower productivity growth also translates into slower overall GDP growth -- one of the primary causes of Japan’s two decades of economic lethargy.

The primary culprit in the United States’ rising inequality level is not globalization, as Spence argues, but political and social conditions at home. Although economy-wide efficiency and growth are necessary for raising living standards, they do not guarantee it. Half of all single mothers without a high school diploma in the United States live in poverty. Workers’ bargaining power has been reduced by long spells of high unemployment (over seven percent for 14 of the past 36 years), a plunge in unionization, and a 30 percent drop in the inflation-adjusted minimum wage since 1970. Real wages have grown, but by far less than the national income has. Conversely, the country’s tax policies, excessive financial deregulation, and corporate pay trends have shifted income to the richest of the rich. The share of total income taken by the top one percent of taxpayers (1.5 million) more than doubled, from ten percent to 21 percent, between the 1970s and 2008; for the top one-hundredth percent of taxpayers (a mere 15,000), that share rose from one percent to five percent. If globalization had caused the United States’ enormous disparity between the income of the bottom 20 percent and that of the top one percent, the disparity would have also occurred elsewhere. But no other rich country has come close to approaching it.

Globalization undeniably plays some role (although not the primary role) in job and wage trends for workers in the tradable sector, but the U.S. government’s policies compound its harm. U.S. workers who lose their jobs due to globalization have to take a 20 percent pay cut, on average, to get their next job, the worst decline of any country in the Organization for Economic Cooperation and Development (OECD). Spence contends that in the new globalized world, U.S. workers have no choice but to sacrifice wages to preserve jobs and that the country as a whole must sacrifice some efficiency and growth to equalize incomes. Both propositions are dubious. Cross-country data on rich countries show no correlation between trade-to-GDP ratios and growth in real wages.

Moreover, look at the Scandinavian countries, which suffered a deep recession and structural problems similar to those of Japan during the 1990s, from banking crises to inefficient domestic sectors. Scandinavia has since recovered, partly by becoming even more globalized. Sweden and Denmark increased their trade-to-GDP ratios by almost 40 percentage points between 1990 and 2008. Just before the most recent global recession, Danish workers enjoyed the highest real wages in Europe, the second-lowest unemployment rate in the OECD, and the highest income equality in the OECD. The Swedes, who came in second in income equality, enjoyed one of the highest rates of real wage growth in the decade prior to the global slump and saw their country’s unemployment rate drop from 10.2 percent to 6.2 percent during that time. To combat the negative effects of globalization, the United States should expand trade and then use part of the resulting addition to GDP to aid those who have been harmed by open trade.

RICHARD KATZ is Editor of The Oriental Economist Alert, a semiweekly report on Japan.

Robert Z. Lawrence

Michael Spence faults globalization for rising unemployment and income inequality in the United States over the past two decades. But he fails to prove that globalization is to blame and overlooks the true culprits responsible for the decline in U.S. jobs.

To make his case, Spence divides the economy into tradable and nontradable sectors. He finds that between 1990 and 2008, value added per employee in tradable products grew by 40 percent more than value added per employee in nontradables. He reasons that if value added per worker in a sector rises rapidly, real incomes per worker must rise, too. As a result, he concludes, real income growth in the tradable sector has outpaced income growth in the nontradable sector, thereby leading to inequality. Since Spence argues that global supply chains are the reason for the increase in value added per worker in tradables, he holds globalization responsible for rising income inequality in the United States.

Spence’s contention that increases in value added per worker lead to higher incomes would be true if value added were measured in terms of increases in actual dollars per worker. But the measure that Spence uses is designed to measure not income but output; Spence measures only how the value added per worker in a given industry shifts based on changes in the prices of goods and services produced in that industry.

A rise in wages in terms of what workers produce -- the measure that Spence uses -- does not necessarily translate into a rise in workers’ purchasing power. For example, suppose that productivity rises rapidly in the computer industry and the improvement is passed on to consumers in the form of lower prices. The actual wages of computer industry workers would not rise in such a scenario. Spence’s measure would show higher value added per worker, but workers who produced computers, and workers employed elsewhere with similar wages, would benefit only if they chose to spend more of their money on computers. Thus, the growth of output per worker does not necessarily lead to higher incomes and, as a result, higher wage inequality. This is why the data actually contradict Spence’s inference that incomes in the tradable sector have mirrored value-added growth and thus outpaced income growth in the rest of the economy. The increases in his measure of value added per worker have been concentrated in manufacturing -- yet between 1990 and 2008, average weekly wages in manufacturing increased by four percent less than did wages in the economy as a whole.

Spence is also wrong to blame the slow employment growth in the tradable sector over the last two decades on the outsourcing of lower-value-added manufacturing activities, such as clothing production and electronic assembly, to other countries. To be sure, this movement has lowered some salaries and eliminated some jobs in the United States, but its impact on overall rates of employment and wages is far smaller than Spence implies. In fact, he completely ignores a far more important factor in promoting wage inequality: technological innovation.

For the past 50 years, technological development has helped productivity grow more rapidly in manufacturing than in the rest of the economy. On the one hand, this growth could lead to less employment, since it allows the production of a given quantity of goods with fewer workers. On the other hand, it results in cheaper products, creating an incentive for consumers to buy more goods, which could increase employment. Yet in practice, the consumer response to cheaper goods has been insufficient to offset the job losses associated with higher productivity. Just as rapid productivity growth in agriculture has led to fewer jobs on farms, rapid productivity growth in manufacturing has led to fewer jobs in factories.

Higher productivity has thus contributed far more than outsourcing to the decline of U.S. jobs in the tradable sector. Indeed, the share of U.S. employment in manufacturing has been falling steadily at a rate of almost half a percentage point per year for the past 50 years. And the decline in that share from 1990 onward, when global supply chains began to flourish, has not been any greater than it was in the 1960s, when the economy was relatively closed.

Income inequality and high unemployment are serious problems. But Spence’s flawed reasoning, which blames them on U.S. trade with emerging economies, will divert policymakers from understanding their true causes -- and from devising appropriate responses.

ROBERT Z. LAWRENCE is Albert L. Williams Professor of International Trade and Investment at the Harvard Kennedy School and a Senior Fellow at the Peterson Institute for International Economics.


My essay was based on my analysis of trends in employment and value added in the U.S. economy between 1990 and 2008. Having found that the tradable sector generated almost none of the jobs added in the United States during that period, I argued that the shifting structure of the global economy, driven by the growth and changing structure of the emerging economies, is having a major impact on U.S. employment and income distribution. The United States needs to understand those forces and then respond.

The most striking feature of the 18 years that I analyzed is that virtually all the employment growth during that time occurred in the nontradable sector (such as government services and health care). Although the United States may have had a problem with unemployment in the private sector during this time, it did not, as Richard Katz suggests, have employment troubles overall. Absent employment growth in government services and health care, which accounted for almost 40 percent of job gains during that period, higher unemployment numbers could have appeared much earlier -- due to the fact that employment dropped considerably in the manufacturing subset of the tradable sector of the economy.

Katz writes that in arguing that tradable-sector jobs disappeared, I failed to recognize that manufacturing represents a greater share of the tradable sector than other industries and that employment dropped more in manufacturing than in any other tradable industry. Katz argues that since manufacturing is the largest segment of the tradable sector and employment fell in manufacturing not because of globalization but because of technological development, my overall analysis is incorrect. Yet it is not true that manufacturing represents the majority of the tradable sector. In 2008, for example, manufacturing jobs comprised fewer than 15 million of the 34.3 million jobs in the tradable sector. Between 1990 and 2008, manufacturing employment did fall, but value added grew and value added per worker grew even faster. But whereas high-value parts of the supply chain, such as design and marketing, thrived domestically, the lower-value-added parts, such as product and component assembly, tended to move to other countries. Combined with the fact that the nonmanufacturing elements of the tradable sector (such as finance and consulting) grew, these conditions produced essentially no overall growth in the U.S. economy.

According to Katz, I wrongly single out the effects of external factors on U.S. employment and income distribution and believe that economic policies in China and South Korea are to blame for the changes in employment and wage equality in the United States. But I do not fault those countries for harming U.S. jobs and wages. If the United States has a structural or productivity challenge, it is clearly rooted in domestic problems. But these domestic issues are set in an open economy, which means that they are thus subject to the effects of globalization.

Katz next points out that due to technological advancement, productivity growth has generally been higher in manufacturing than in services, meaning that fewer workers are producing more. But manufacturing is not declining in absolute terms; instead, it is declining relatively, as a result of the fact that the service industry is making up a greater share of the total U.S. economy, as is the case when economies experience an increase in income.

Finally, Katz criticizes my suggestions for how to alleviate U.S. unemployment and inequality and wrongly suggests that I advocate protecting jobs, companies, and industries from external competition. In advanced and emerging economies alike, such actions interfere with structural change, productivity, and income growth. Instead, I call for strengthening the competitiveness and inclusiveness of the U.S. economy by improving education to meet the premium on highly educated workers, investing in infrastructure, and reforming the tax structure.

Robert Lawrence, meanwhile, notes that my measurement of value added per employee assesses industry output and that such output does not necessarily translate into purchasing power for workers (although the correlation is high). Ideally, one would measure productivity growth in specific components along value-added chains, but current data can measure productivity growth only at the industry-wide level. When measured this way, it is a combination of two forces: productivity gains in various parts of the value-added chain and shifts in the composition of the domestic component of the value-added chain toward the higher-value-added and higher per capita income components. In manufacturing industries, which have long, relatively complex value-added chains, the shift between 1990 and 2008 was toward higher-value-added components. The manufacture of lower-value-added parts moved abroad, causing a significant amount of the decline in employment.

Lawrence correctly points out that technological innovation has reduced the need for labor and raised productivity across the economy. The U.S. economy is certainly subject to powerful internal technological forces. But globalization is increasingly making its mark. Tracking the growing impact of the global economy on the U.S. market is vital to understanding whether the United States is on a path toward growth and employment -- and, if not, what the U.S. government and the business community can do to correct course.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now