Publisher Council on Foreign Relations Press
Release Date October 1999
Price $19.95 paper
Despite recent record growth and employment statistics in the United States, this book appears at the right time. Begun around the last presidential election, when concerns about low American wages and booming international trade first surfaced, it is being published before the next--when the United States must finally reckon with its large and growing external deficit. U.S. attitudes toward globalization will shape the policy debate going into the next election, especially if a slowdown in growth is associated with the rebalancing of global demand.
After the failed efforts by the International Monetary Fund to prevent financial crises in Russia and Brazil, discussion in the United States and abroad has arisen on Asia, the role the lending institution should be playing, and the suitability of the international financial architecture in general. Despite the remarkable resilience in the face of a large external shock, the American public remains skeptical about the trend toward globalization. One small sign of the doubt is the fate of "fast-track" authority in the Congress. It was decisively defeated in October 1998, despite support by a Democratic president and by the Republican majority. Even the North American Free Trade Agreement is now supported by less than a majority of voters, even though it reflects a vast increase in trade for the United States.
Attitudes within the other industrial countries mirror this ambivalence. Look at the recent cases of conflict that have emerged in the World Trade Organization. On one side, there is talk of a new effort to ensure free trade; on the other, bananas became a major issue between the United States and the European Union. Developing countries, afflicted by external shocks and internal financial difficulties, are more doubtful about globalization as well.
The message of this book is that there is no simple link between the forces of globalization and increased wage inequality around the world. Certainly, the global integration of goods and financial markets has facilitated the spread of skill-biased technology and increased the potential returns to investment. These influences, together with higher immigration, may have contributed to increased wage dispersion in the United States. However, it is difficult, if not impossible, to disentangle the influence on wages of expanded foreign investment and trade, more rapid technology diffusion, and changes in labor market structure. These developments are interrelated and, indeed, mutually reinforcing aspects of a broader process of market integration and innovation.
Even if one could identify "globalization" as the cause of increased wage dispersion, the appropriate solution is not to restrict international commerce. The evidence shows that expanded trade and competition at the global level raises living standards overall, boosting productivity and creating more high-wage jobs. However, ensuring continued public support for a liberal international system will require greater efforts to help workers adapt through better education and measures to alleviate the burden of adjustment on those who cannot easily adapt.
The authors greatly appreciate the contributions made by participants in the study group to this volume. The creative leadership of the group's co-chairs, Jessica Einhorn and John Lipsky, was especially valuable.
Albert Fishlow and Karen Parker
In the fall of 1996, the Council on Foreign Relations convened a study group to examine the nature, causes, and consequences of growing wage disparities in the United States. At that time, the presidential election season was in full swing. The issue of wage inequality had gained considerable salience as a result of presidential contender Patrick Buchanan's rhetoric against immigration, free trade, and other features of the global economy. Indeed, so central was the issue of income distribution to the 1996 campaign that the two leading parties executed a role reversal: The Republican candidates addressed the electorate's concerns about wages and job insecurity in a way they had not done for decades, while the Democratic president took a far more sanguine view of developments in the U.S. labor market.
By the time the study group completed its deliberations, fears of job losses and pay cuts had receded. As a result of a robust seven-year expansion, the U.S. economy was producing jobs at a record pace. Unemployment has fallen close to 4 percent, the lowest level in more than two decades. Labor productivity and wages began to recover from the stagnation of the 1970s and 1980s. Moreover, the incomes of the poorest Americans were growing relatively rapidly--and thus the income gap was beginning to narrow. Yet inflation remained quiescent. In that favorable environment, some participants in the study group questioned the group's guiding premise: that an increasingly unequal wage distribution risked creating a backlash against open trade and investment flows and America's traditionally liberal immigration policy.
The global economy is now emerging from a major global slowdown. The U.S. economy has emerged seemingly unscathed from the biggest global slowdown this decade. Consumption and business investment remain strong, productivity is surging and inflation remains subdued. Indeed, the most notable development of the past decade has been the striking divergence in performance between the United States and other industrial countries. Why, then, should one still be concerned about the issue of wage inequality--and its perceived link to the process of globalization?
The debate over foreign trade, investment, and wages may have been hushed for the time being, but it is certainly not over. The fact that commercial disputes are still so prevalent, despite the strong performance of the U.S. economy, should alert us to the reality that domestic support for free trade is rather shallow. If and when the U.S. economy begins to slow down--under the burden of a huge external deficit and a strong dollar--the concern over competition from abroad will intensify. Indeed, it would not be surprising if globalization were to re-emerge as a political issue in the next presidential election. Discussion has begun on the redesign of the global financial system; less attention has been devoted to the question of how open the global trading system will remain--and how committed the United States is to its maintenance.
Even prior to the global financial crisis of 1997≠98, U.S. support for globalization was tenuous. The failure of President Clinton to secure "fast-track" authority to negotiate free trade agreements with Latin America revealed that the decades-old bipartisan consensus in favor of gradual trade liberalization no longer exists. The defeat of fast track was all the more striking given that the legislation was put forward by a Democratic president and supported, albeit halfheartedly, by the Republican leadership of the House and Senate. The initiative failed, moreover, at a time when the U.S. economy was experiencing its best performance in decades, when consumer confidence was at an all-time high, and when American businesses enjoyed a preeminent position in the global market. This watershed event revealed that, in the eyes of many Americans, free trade poses a threat to--rather than an opportunity for--higher U.S. living standards.
The United States is not alone in its ambivalence toward globalization. The election of left-leaning governments in France, Germany, and the United Kingdom, the uneasiness of some Europeans toward the European monetary unit (EMU), and the slow pace of deregulation and liberalization in Japan underscore the tenuous support for integration among our industrial partners. Backing for open trade and capital flows among the denizens of emerging markets--who stand to gain from increased global commerce in labor-intensive manufactures--has also become more tentative. While Asian financial leaders have endorsed the principles of free trade at meetings of the Asia-Pacific Economic Cooperation (APEC) forum--and have begun to liberalize their financial sectors to foreign investment--there also have been calls for greater regulation of global financial flows. Indeed, Malaysia's decision in 1998 to impose capital controls and the subsequent Russian default have led other countries to question the desirability of unfettered capital flows.
Even as the global economy recovers, the nation's trade deficit approaches record levels. The economy will likely slow as a result while unemployment will rise from its current low levels. Although the substantial scope for countercyclical monetary and fiscal policy will cushion the slowdown, the halcyon days of falling unemployment and rising wages have passed.
America's trade deficit, which was running at an estimated $145 billion in 1997, will more than double by the end of 1999. Moreover, the sharp reduction in local currency costs that has resulted from the currency devaluations and disinflation in Asia will enhance the incentives for U.S. multinationals to relocate production facilities abroad. The competitive challenges posed by these developments are likely to reignite public concern over job insecurity and slow wage growth. America's leadership role in pressing for open global trade and financial flows could be threatened by a deterioration in domestic public support--at just the time when U.S. leadership is most needed.
The first step toward grappling seriously with these issues is to understand the causes of and remedies for increased wage inequality. This volume draws together the research of a number of distinguished scholars who have focused their attention on this problem. Chapter 2, by Richard Freeman, provides a detailed and comprehensive analysis of recent wage and employment trends in the United States. In Chapter 3, David Blanchflower and Matthew Slaughter examine the phenomenon of rising wage disparities through the lens of the contrasting experiences of other industrial countries. The next three chapters consider possible non-trade explanations for the growing U.S. wage gap. Theodore Moran, in Chapter 4, reviews the literature on foreign direct investment abroad and the wage and employment patterns of American multinationals. In Chapter 5, J. Bradford Jensen and Kenneth Troske exploit new plant-level data on U.S. manufacturing firms to evaluate alternative explanations for rising wage inequality. Steven Camarota and Mark Krikorian assess the impact of immigration on the distribution of U.S. wages in Chapter 6. Lisa Lynch concludes the volume by analyzing, in Chapter 7, the effectiveness of alternative remedies--particularly education and training--in arresting wage declines among low-income Americans.
These researchers find no "smoking gun" in the forces of globalization that can explain why low-income Americans--alone among residents of the major industrial countries--have experienced such a sharp decline in wages, even as workers at the upper end of the income ladder have enjoyed unprecedented gains. However, the chapters do suggest that technological progress, foreign direct investment, and global trade are inextricably linked and may be associated with a rising share of high-wage, skilled workers in the U.S. labor force.
It may never be possible for researchers to identify separately the precise contributions of trade, technology, foreign investment, education, and immigration to changing wage patterns. Indeed, the usefulness of such distinctions probably is limited. Each of these influences appears to be part of a more complicated process in which advanced technology, capital, and managerial know-how spread ever more rap-idly throughout the global economy. It is a process that amplifies the potential market for--and thus the rewards for--skill. As a result, relatively small differences in resources or ability, augmented by good strategic planning, can lead to large disparities in outcomes for firms and their workers.
In the following sections, we identify some of the principal findings that emerged from this rich set of chapters. Each chapter addresses a complex set of issues worthy of in-depth analysis and exploration. Careful review will lead readers to appreciate the diversity of facts and interpretations that have evolved from this expanding literature. Through greater understanding of these complex issues, we may be able to identify remedies that are both feasible and appropriate to the challenges at hand. The need is urgent, for the problem of wage inequality threatens not only our economic future, but also the social compact that has helped this country to adapt and excel through a period of extraordinary change.
Stagnant overall wages and increasing wage disparities in the United States constitute a major change in the economic landscape of the last two decades.
From 1900 to 1973, real average hourly earnings rose by about 2 percent per year in the United States. Two percent annual wage growth, when compounded, results in a doubling of income every 35 years, guaranteeing that living standards increase from one generation to the next. By contrast, between 1973 and 1995, the real median earnings of salaried men fell by about 0.5 percent per year. While other earnings series show continued gains in real wages in the post-1970 period, they also depict a marked slowdown from historical trends. The real earnings of women rose by 7 percent over the same period, both in absolute terms and relative to men's salaries. However, women's wages also became more unequal.
Total hourly compensation rose modestly between 1973 and 1995--by about 10 to 16 percent. However, these increases largely reflected increased nonwage compensation, especially medical benefits. As a result, the gross domestic product (GDP) per family rose by nearly a quarter, and average family incomes increased by 10 percent, between 1979 and 1994. However, the median family income remained roughly constant; most of the gains were achieved by higher-income households.
Thus it seems that most American families have had to run faster to stay in the same place: Household incomes have been sustained by extended working hours and through increased labor force participation by women. Women's labor force participation rose from about 55 percent in 1985 to 59 percent in 1995. U.S. workers spend more hours on the job than do employees in most other industrialized countries--five to ten full weeks more than their counterparts in Europe.
U.S. families have saved less and borrowed more in order to preserve their living standards. Contrary to popular perception, much of this borrowing has been used to finance consumption of nonessentials, such as travel and entertainment. Installment debt has grown at a rapid pace and is now at record levels. While some of the increase may reflect greater reliance on credit cards for transactions, the steady rise in household debt service, despite lower interest rates, indicates that real borrowing has grown. Household net worth nevertheless has reached record levels, as broader participation in the equity markets and rising stock prices have boosted wealth. However, aggregate measures of household wealth obscure two divergent trends: Equity market participation and the associated income gains remain highly skewed toward higher-income households, while middle- to lower-income families carry a substantial amount of debt. Measures of living standards that rely on consumption data may be misleading, to the extent that the rise in consumer spending--which has now driven the saving rate below zero--is unsustainable.
According to research by Jeffrey Williamson, the surge in wage inequality over the past two decades is comparable to that which occurred in the United States in the early years of this century.1 However, in that earlier period, average wages continued to rise, so that few workers experienced an absolute decline in their incomes. Nevertheless, the decades that followed saw a surge in isolationist and protectionist sentiment. America's liberal immigration policy was reversed in the 1920s, and the imposition of high tariff barriers and successive rounds of competitive devaluation contributed to a global depression.
There have been growing wage disparities among and within all classes of workers.
Men in the top 10 percent of the wage distribution received 4.75 times the pay of those in the bottom 10 percent in 1995, a 20 percent increase in the multiple since 1979. For women, the 90/10 pay gap increased by 47 percent.
Much of the rise in inequality has resulted from declining wages of low-income workers. While the incomes of those in the top 20 percent of the wage distribution remained broadly stable between 1979 and 1995, workers in the lowest quintile suffered a marked fall in earnings. Regardless of which price deflator is used, the real earnings of low-wage workers fell sharply. Wage declines were largest for male high school dropouts, who suffered a 20 percent decline in their real earnings between 1979 and 1993. Falling wages are not confined to a small group of workers. Even experienced, college-educated men suffered a decline in earnings over this period.
Compensation in the form of benefits also has become more unequal. Whereas in 1979, 57 percent of high school graduates and 39 percent of high school dropouts received employer-provided pensions, in 1993 these figures had fallen to 45 and 21 percent, respectively. By contrast, there was little change in pension provision for college graduates. The distribution of medical benefits has followed a similar pattern, as Freeman points out in Chapter 2.
About half of the increase in inequality can be accounted for by differences in skills, as indicated by education, age, and work experience. Returns to schooling increased sharply after 1979 and have now reached a postwar high. As the study by Jensen and Troske in Chapter 5 reveals, inequality rises during recessionary periods, in which the ratio of skilled to unskilled workers in the labor force ratchets upward. This pattern mirrors that of European unemployment, which rises during recessions and diminishes only modestly during expansions.
Falling relative wages for less skilled workers has not led to higher participation rates for these Americans. On the contrary, the rise in the skill premium has been accompanied by a declining share of less skilled workers in the labor force--in all occupations. From the 1970s through the 1990s, annual hours worked for men in the bottom deciles of the pay distribution fell while those in the upper deciles rose or were steady. Thus, inequality in hours worked increased along with inequality in hourly pay, producing even greater income disparities.
More than half of the growth in inequality remains unexplained. Even among workers with observationally equivalent skills, wage disparities have risen. For example, among college graduates, the wage ratio between the 90th and 10th percentiles rose by 22 percent between 1979 and 1997.
The distribution of family income has followed that of wages. The top 20 percent of families obtained virtually all of the gain in income over the past 20 years, and within that group most of the gains went to the top 5 percent. As it turns out, the wives of high-income men tend to be high earners themselves.
There is no evidence that the growing disparity of wages has been offset by increased labor mobility or that mobility is much higher in the United States than in Europe. On the contrary, the wage gains that used to accrue from job experience have diminished over the past 15 years. Among recent immigrants, mobility is likely to be an even greater problem, as 40 percent are high school dropouts. (Twenty-nine percent of dropouts are immigrants.)
Moreover, poverty and poor educational training seem to be becoming more entrenched in urban areas. Research by Chris Mayer has documented that the share of the poor living in census tracts with a poverty rate of more than 40 percent increased from 16 to 28 percent, while the share living in tracts with poverty exceeding 20 percent rose from 55 to 69 percent.2 This greater concentration of poverty is likely to hamper individuals' access to good primary and secondary schools and the financial support to pursue higher education.
Despite the general dynamism of the U.S. labor market, there are some signs that structural employment has increased. As Lynch indicates in Chapter 7, the duration of unemployment has increased, and the share of unemployed workers who are permanently displaced in recent years is high in comparison to similar points in earlier business cycles. Job losses due to the abolition of positions or shifts have increased in recent years, largely among older, white-collar, and more educated workers. The costs of job loss are large and enduring; recent studies suggest that, even six years after job displacement, earnings remain about 9 percent lower than before.
Slow aggregate wage growth and rising income disparities have been most pronounced in the United States.
Inequality rose to some degree in most of the advanced market economies in the 1980s and 1990s. However, the widening in the wage distribution has been most pronounced in the United States (which had the largest wage disparities to start with), began the earliest, and has continued into the 1990s, as Blanchflower and Slaughter demonstrate in Chapter 3.
Some European countries have experienced rising unemployment rather than higher wage inequality. However, this pattern is not uniform. The United Kingdom, Canada, and New Zealand faced both higher earnings inequality and rising unemployment. On the other hand, the Netherlands, Japan, Austria, and Sweden experienced low or declining unemployment and only a small rise in earnings inequality. France and Germany have seen substantially higher unemployment but little change in earnings inequality.
Overall wage growth in the United States has been slower than in other countries. In the 1980s average real hourly compensation of blue-collar manufacturing workers rose in the United Kingdom (by 2.6 percent), Japan (by 1.6 percent), France (by 0.9 percent), and Germany (by 1.3 percent). In contrast, pay for these workers fell by 12 percent in the United States. As a result, low-paid workers in America earn less than their counterparts abroad, even though average incomes in the United States are significantly higher. Low-paid German workers earn roughly twice as much as low-paid Americans. Even in the United Kingdom, whose GDP per capita is two-thirds that of the United States, low-income workers earned 32 percent more than their American counterparts in 1995. Overall, one-third of American workers earn less in purchasing-power-parity terms than comparable workers overseas. Adjusted for hours worked, the disparities are even greater.
On the other hand, the U.S. economy has created many more jobs than European countries. In 1974 the ratio of employed workers to the population aged 15 to 64 was the same in the United States and Europe (about 65 percent). By 1995 the ratio stood at 73.5 percent in the United States and 60 percent in Europe. However, U.S. job growth has been concentrated in high-wage and low-wage occupations, making it harder for American workers to move up the income scale.
This rise in U.S. wage disparities coincides with "globalization" of the American economy, which, most economists agree, has been beneficial for U.S. living standards.
The labor market trends just described have coincided with a rapid expansion in global trade, especially trade in manufactures between industrial and developing nations. As a share of industrial country output, imports of manufactured goods from developing countries rose fivefold in the two decades after 1970.
Labor and capital flow increasingly across national borders. U.S. foreign direct investment and portfolio capital flows have surged over the past two decades, with a growing share directed toward emerging markets. At the same time, immigrant inflows into the United States have reached record levels. An estimated 1.2 million immigrants (about 900,000 legal and 300,000 illegal) entered the United States in 1996. As a result, the foreign-born share of the American population roughly doubled between 1970 and 1997 to about 10 percent.
Not surprisingly, the global integration of goods, capital, and labor markets has been implicated in the slow wage gains and rising inequality of the past two decades. Some have called for renewed protection of American industry, sometimes in the guise of labor and environmental standards. However, most economists still emphasize the overall benefits of economic integration.
And, in fact, the benefits are many. U.S. consumers have enjoyed lower-cost imported goods. Television sets, microwaves, automobiles, and computers have become less expensive and more reliable. Were it not for job creation in the high-wage export and technology sectors, the slowdown in U.S. productivity and earnings would likely have been greater. The evidence suggests that foreign direct investment has contributed to the growth of U.S. exports, which are produced with more advanced technologies by higher-skill, better-paid workers. To the extent that trade augments competition and expands potential markets, productivity is enhanced, although economists debate the degree of change.
The availability of low-wage (and often skilled) immigrant labor has reduced the cost of doing business in the United States, while increased competition has contributed to the process of disinflation. It is no accident that the U.S. economy is enjoying an unprecedented combination of low inflation, low unemployment, and rapid productivity growth. Immigration has also lengthened the age profile of the American labor force and may ease the pressures on the Social Security system in future years. For these and other reasons, the U.S. performance is the envy of many industrial countries.
There is less consensus among economists about the distributional impact of globalization. However, few see a "smoking gun" in trade with developing countries.
Much of the research on the rise in wage inequality has focused on the "demand side" of the labor market. Since both employment and the relative wages of less skilled workers have fallen since 1980, economists have concluded that the demand for their services has declined. Among the demand-side explanations, attention has been focused on the effects of foreign trade, investment (including outsourcing of production), and technological progress.
According to classical trade models, expanded commerce with low-wage countries ought to augment wage disparities since trade rewards those factors of production that a country has in abundance. Generally speaking, the relative wages of less skilled workers in poorer countries are expected to rise, as are the wages of skilled workers in the more advanced countries. Conversely, the rate of return to capital in low-income countries should fall (relative to labor), whereas wealthier nations should experience a rise in capital income.
The debate over globalization and its impact on wage inequality has received considerable attention--despite the limited empirical evidence in support of a link--because theory suggests there should be a connection. However, if there is a link, it is considerably more complicated than conventional models would suggest.
One strand of research analyzes the impact of trade on the employment of less skilled workers by measuring the quantity of labor that is displaced. This research indicates that commerce with developing countries can account for no more than 10 to 20 percent of the rise in wage inequality in the United States during the 1980s. However, these models are poorly specified, in that the volume of trade is determined endogenously and depends on many factors. In early studies researchers identified the growing U.S. trade deficit as a key contributor to rising wage disparities, even though the deficit primarily reflected macroeconomic phenomena (the low U.S. saving rate and strong dollar) rather than microeconomic factors such as the distribution of skilled labor and/or changes in labor institutions. Wage inequality continued to rise in the late 1980s, even as the deficit narrowed. As economist Paul Krugman has asserted, the amount of imports from developing countries is simply too small to have induced the sizable wage disparities now evident in the United States.3
More recent research recognizes that it is the potential rather than the actual supply of foreign goods that influences prices of traded products and the wages earned by their producers. The threat of competition from abroad--either through imports or outsourcing of production--may exert a strong influence on prices and wages. To the extent that such pressures influence relative wages, a corresponding change in the relative prices of goods produced with more skilled labor vis-ŗ-vis those produced with less skilled labor would be expected. This body of research, which is surveyed in Chapter 3, finds trade to have little or no impact on relative wages--even controlling for the impact of technological improvements on the prices of some goods, such as computers.
In principle, price-based studies can account for the "threat effect" of foreign competition--even when no trade actually occurs--on wages in the industries exposed to it. However, their accuracy is dependent on the proper specification of the sectors exposed to import competition; the import penetration ratio may not be the best measure of such competition. These studies also suffer from compositional and quality biases in the relative price indices as well as nontrade influences on prices that may not be adequately controlled for. Both the price- and quantity-based studies have shortcomings, but they reach the same conclusion: Trade with developing countries has had only a limited impact on the relative wages of U.S. workers.
Closely related to foreign trade is foreign direct investment (FDI) abroad. There are two potential threats to low-skilled U.S. workers from such investment: the shifting of manufacturing jobs overseas via outsourcing and the "hoarding" of high-wage jobs in the home countries of offshore multinationals that invest in the United States. Research on these influences, which is thoughtfully summarized in Chapter 4, suggests that FDI abroad by U.S. multinationals actually enhances U.S. exports, thus contributing to the creation of higher-wage jobs in this country.
As Chapter 4 points out, jobs in export industries pay 13 to 15 percent more than those in nonexporting firms, provide 11 percent higher benefits, experience 20 percent faster job growth, and are 9 percent less likely to go out of business. Andrew Bernard and Brad Jensen have confirmed that when other plant characteristics that are correlated with wages (e.g., size, capital intensity, and productivity) are controlled for, exporting is associated with higher wages for both production and nonproduction workers.4 Other research has shown that outsourcing leads to increased demand for skilled workers but not higher relative wages for them.
There is little evidence that FDI, by itself, triggers increased imports of goods from emerging markets that compete with those produced by less skilled U.S. workers. However, the threat to relocate production abroad, regardless of whether the firm faces competitive pressures to do so, can be used to negotiate smaller wage increases. As Dani Rodrik points out in his book, Has Globalization Gone Too Far?, the elasticity of demand for labor would tend to rise as a result of such threats as well as the volatility of earnings and the incidence of nonwage costs.5 While such influences are difficult to measure, research by Slaughter and others suggests that the elasticity of labor demand is higher in industries where there is greater international competition.6 Moreover, trade with other industrial countries can be just as potent in restraining wages as trade with emerging markets.
The principal alternative explanation for the rise in demand for skilled labor is skill-biased technology. However, the technology story, by itself, is not persuasive.
For many economists, the most plausible explanation for rising wage inequality is technological progress that demands more skilled workers. Typically, the impact of technology is measured as the residual--that which cannot be explained by other factors. Research by Alan Krueger and others has attempted to identify a direct link between technological progress and relative wages.7 These studies have found that the demand for skilled workers has grown more rapidly in industries that make larger capital investments, undertake greater research and development expenditures, and use more business equipment, such as computers. However, the direction of causality is not clear: The presence of skilled labor may facilitate the adoption of more advanced technology.
Another strand of research attempts to differentiate the impact of technology-related changes in the labor market from those induced by shifts in product demand by decomposing changes in the ratio of production to nonproduction workers (and their relative earnings) into those that occur between and within industries. Technological change is thought to change the relative proportions of skilled and unskilled workers within a given industry or firm. On the other hand, if trade is the cause of increased wage dispersion, it is assumed that the reallocation of labor should occur mainly between industries (or firms); those industries that use relatively more unskilled labor would shrink in response to import competition. In an influential 1994 paper, Eli Berman, John Bound, and Zvi Griliches found that most of the increased demand for skilled workers occurred within industries, which they interpreted as evidence in favor of a technology-based explanation for rising skill premiums in the U.S. labor market.8
The pervasive shift toward more skilled labor across U.S. industries--and in many other industrial and developing countries--suggests to many economists that a broad process of technological innovation is under way.9 If globalization were behind the decline in wages of less skilled workers, some sectors (or industries, or countries) might be expected to use their services more intensively. However, this appears not to be the case.
As Chapter 5 demonstrates, aggregate studies miss much of the story, since most of the increase in wage dispersion has occurred within industries. Even among plants in the same detailed industry classification, there is tremendous heterogeneity along a number of dimensions, such as the rate of technology adoption, foreign investment, export performance, productivity, and employment patterns. Research by John Haltiwanger has shown that industry attributes at the four-digit Standard Industrial Classification (SIC) level can account for less than 10 percent of the variation over time in the growth of employment, output, total factor productivity, and investment.10
Plant-level studies find that most of the increase in the demand for skilled labor has occurred within plants in the same industry. However, no single variable (technology-related or otherwise) can account for much of this trend, nor its cyclical component. Moreover, between-plant changes in the relative wages of skilled workers have been the primary cause of the workers' increased earnings. The reallocation of labor among plants with different characteristics is an important source of aggregate changes in relative wages. Most of the increase in wage inequality appears to be the result of growing disparities among firms.
Plants that hire more skilled workers are likely to see a rise in productivity and average wages, with little or no shift in relative earnings if differently skilled workers play complementary roles. On the other hand, the closure of less productive plants that employ more unskilled labor would result in a fall in their relative earnings. It is not clear whether these within- and between-plant changes in employment and wages should be attributed to trade, technology, or both.
In Chapter 5, Jensen and Troske note that, even controlling for other attributes, there is a positive correlation between export performance and the use of advanced technology.11 However, the causal mechanism is not defined. It may be that the use of advanced technology or the decision to export leads firms to use more capital, employ more skilled workers, and pay higher wages. Alternatively, large capital-intensive firms may be more likely to adopt advanced technologies or to export as a result of economies of scale or other factors.
If technology is the main reason relative wages have risen so sharply in the United States, it is a puzzle why productivity has not grown more rapidly. Nor is it obvious that technology's impact should have been greater in the 1980s than in the 1970s, and in the United States rather than in other industrialized countries.
Labor institutions have also influenced the extent and pattern of inequality.
All major industrial countries have experienced large shifts in the industrial and occupational structure toward sectors and jobs that use a greater proportion of skilled workers. Moreover, the share of manufacturing in total employment has declined everywhere but in Japan. There is no evidence that the expansion of trade or technological progress occurred more slowly in other countries than in the United States.
Even so, the United States experienced a much sharper, and earlier, rise in inequality. America also has faced a larger increase in inequality among groups of workers with similar skills. These contrasting experiences reflect differences in the supply of labor across countries as well as changes in labor institutions.
As Chapter 3 points out, the distribution of wages in countries with strong labor institutions is almost always more compressed than elsewhere. Unions tend to reduce inequality by standardizing pay rates among workers within a given establishment or across establishments. The threat of unionization also forces nonunion employers to raise pay and/or benefits in order to keep unions out. The swift decline in union strength in the United States, along with the fall in the real minimum wage, is correlated with the early and sharp widening of the U.S. wage gap. Union density has declined dramatically in this country since 1970--from about 27 percent of the labor force to 15 percent at present. America's unionization rate is the lowest in the world. (The figure for France, although lower, is not comparable.)12 After 1990 union density also fell markedly in the United Kingdom, a country that has seen a sharp increase in wage inequality as well.
Minimum wages also compress the pay structure, at least for those who are employed. The sharp fall in the real U.S. minimum wage over the past two decades and the weakening of wage councils in the United Kingdom fit the pattern of rising wage inequality in these two countries. In contrast, increases in France's minimum wage appear to have prevented an erosion in real wages at the low end of the pay scale.
There appears to be a trade-off between wage inequality and unemployment. However, the pattern is not clear-cut. Some countries (the Netherlands, Austria, Japan, and Sweden) enjoy both low unemployment and moderate wage dispersion, while others ( Australia, Canada, the United Kingdom, and New Zealand) suffer from high unemployment and large earnings disparities. Most Europeans enjoy more generous social benefits than do Americans, mitigating the impact of unemployment on household income and the social stigma associated with joblessness. To discern the full impact of prolonged unemployment, all sources of income and the lifetime earnings profile should be examined.
Labor institutions clearly matter, but it is not obvious why the regulations protecting workers have declined in many countries. Weakening labor standards may be a result of increased global competition. However, if global trade and investment played a role in the demise of labor institutions, their impact should be evident from the price and wage studies just cited.
Trade, foreign investment, and technology have not, by themselves, caused the dramatic widening of the U.S. wage structure. However, these may be mutually reinforcing elements of a broader process of market integration that has increased the demand for skilled workers.
Limitations in data, interpretation, and testing have prevented the identification of individual causes of rising wage inequalities. Measuring skills, the impact of trade, and technological progress is quite difficult, as is devising price indices that are free of compositional and quality biases. The large increase in residual inequality should make people wary of skill-based explanations for growing wage disparities. Ultimately it may prove impossible to disentangle the various influences as they are closely related and, often, mutually reinforcing.
As Chapter 4 and 5 demonstrate, a firm's size, participation in export markets, foreign direct investment abroad, and technological progress go hand in hand. Those firms that pay the highest wages and hire relatively more white-collar workers are those that export, invest overseas, employ the latest technologies, and are large.
As Adrian Wood has suggested, "defensive innovation" may be a response to competition in export and import markets.13 Global trade may facilitate the transfer of skill-biased technologies among industrial countries and toward emerging markets, which would account for the increased demand for skilled labor everywhere. The expansion of markets increases the incentives for innovation and rewards small (perhaps unmeasureable) differences in skill with large payoffs. The attributes of workers may matter less than the productivity of the industries or firms in which they work. Strategic decisions by firms are extremely important for worker outcomes, particularly when there is limited mobility of labor across industries.
Similarly, foreign investment may be "defensive" in character. If so, trade must be cited as a fundamental cause of wage disparity, insofar as it creates competitive pressures and facilitates the segmentation of the production process. Then, the activities of multinational corporations do no more than hasten the inevitable.
Changes in labor institutions may be a consequence of expanded trade and foreign investment rather than an independent cause of increased wage disparities. Indeed, labor institutions may change in response to wage inequality, insofar as the interests of workers would tend to diverge.
As these examples suggest, it is time for a broader conceptualization of the process of globalization and how it can affect the demand for labor. The conventional framework of analysis is too narrow, resting as it does on the restrictive assumptions of perfect competition, diversified product markets, and intersectoral factor mobility. Models that incorporate threat effects, strategic advantage, increasing returns to scale, and dynamic technological progress are better suited to explaining the phenomenon of rising wage inequality in the current global environment.
Even if these influences were conceptually separable, it would be a challenge to devise tests that could accurately distinguish between them, as the studies in this volume demonstrate.
Changes in the relative supply of less skilled workers as a result of immigration also may have played a contributing role in the rise in wage inequality.
The character of U.S. immigration has changed over the past two decades, as Chapter 6 points out. The educational attainment of newly arrived legal immigrants has diminished over time; among those who immigrated in the 1990s, 39 percent are high school dropouts. (Immigrants make up 11 percent of the workforce and 29 percent of those without a high school diploma.) In contrast, 34 percent of 1980s arrivals and 26 percent of those who arrived before 1980 were dropouts. The educational achievement of illegal immigrants may be even lower; among those who were granted amnesty under the 1986 Immigration Reform and Control Act, 74 percent lacked a high school diploma.
As a result of their more limited skills, recent immigrants' earnings have fallen farther behind those of their native counterparts. Whereas in 1970 the average annual earnings of recent male immigrants was 81 percent that of natives, by 1990 that figure had declined to 65 percent.
Lower earnings of less skilled immigrants have contributed to a widening of the wage distribution in some parts of the country. Immigration has not only skewed the distribution of income by increasing the number of workers in the lowest-paying jobs. It appears that immigrants also have depressed the earnings of low-skilled native workers. Chapter 6 estimates that the wages of less skilled workers were reduced by an estimated 0.7 percent for every 1 percent increase in the immigrant composition of a given occupation. Since low-skilled native workers are employed in occupations whose immigrant ratio averaged 15 percent, these workers suffered an estimated 10.5 percent drop in wages relative to others with the same individual and occupational attributes.14
George Borjas, Richard Freeman, and Lawrence Katz also have found that post-1979 immigration can explain between 27 and 55 percent of the decline in relative wages of high school dropouts. 16However, no more than 10 percent of the fall in wages of high school graduates relative to college graduates could be attributed to immigration.
The American educational system, which has failed to meet the demand for skilled workers, also must bear some of the responsibility for the decline in relative wages.
Variations in skills are considerably less in other countries than in the United States, as noted in Chapter 7. Moreover, international variation in skills has tended to match changes in wage inequality from 1979 to 1990. In Germany and other countries, there are high minimum educational standards and clear incentives for students to do well in school--even for students who have no plans to attend college. Apprenticeship training is available for those who do not go on to university, leading to higher skill attainment among those in the bottom half of the ability distribution.
Most puzzling of all is why parents and students have not responded to the rise in skill premiums by attaining more education. As noted earlier, the wage gap between more and less skilled workers has nearly doubled over the past 15 years, while college enrollments have fallen. Part of the problem may be the rising cost of tuition; the General Accounting Office (GAO) estimates that tuition at four-year public colleges and universities has risen three times faster than median household income between 1980 and 1995. Student aid has not kept pace with tuition levels, so families are relying more on loans and personal finances to attend college. Access to educational financing may be limited for students from poor families, particularly in urban areas, making inequality ever more entrenched.
In the face of rising costs and limited financing, many students are forced to delay entry into college, temporarily suspend training, or drop out of school despite the high wage premium.
On-the-job training contributes to significantly higher business productivity, according to research surveyed in Chapter 7. Yet American businesses underinvest in training, due to high turnover rates and the large number of small firms in this country. (The cost of training, including downtime, is high.) Japan and Germany seem to have overcome these challenges through programs that facilitate joint investment in on-the-job training, certification of employer-provided skills acquisition, employment guarantees for workers, and extensive government financial support.
Funding for and the quality of education at the primary and secondary grades in the United States also have diminished. The portion of the federal budget allocated to education and training has fallen sharply over the past 20 years. Per capita spending on education and training (as a share of the population aged 5 to 24) has declined from over $4,700 in the 1970s to less than $3,500 at present, in inflation-adjusted terms. Of the more than $500 billion currently spent on education by federal, state, and local governments, some $211 billion is allocated to higher education (for which students can obtain their own funding), leaving relatively little money for publicly financed primary and secondary education.
The time to implement remedies is now.
Identifying precisely the causes of rising wage inequality is desirable but not essential to devising appropriate remedies. There are three possible approaches to the problem: (1) attempt to reverse the underlying causes of the wage gap; (2) boost workers' skills; and (3) implement ameliorative programs for those who cannot easily adjust. The ideal remedies are neutral with respect to productivity or are growth-enhancing (i.e., skills training that raises the productivity of less skilled labor).
Regardless of the evidence, in many people's minds globalization will remain inextricably linked to the problem of job insecurity and falling wages. Public concerns must be addressed somehow, or the benefits of freer international trade and financial flows will be lost as a consequence of poorly conceived remedies. There is clear evidence that trade, technology, and foreign direct investment raise overall wages, not just those of skilled workers. Rather than restrict incentives for productivity-enhancing external commerce, government policy should seek to preserve and expand access to overseas markets through reciprocal reductions in tariff and nontariff barriers. Labor standards can be incorporated into trade agreements, provided these do not merely serve as obstacles to further progress. Regional initiatives to liberalize trade may be an effective interim strategy toward a more liberal and equitable global trading system.
To ease the adjustment process, policymakers might consider taking steps to improve the skills profiles of new immigrants through skills-based admission criteria. Attention needs to be given to improving the quality of education in the United States, especially at the precollegiate level. The establishment of high minimum educational standards, augmented funding for primary and secondary education, increased opportunities for self-financing of higher education, and the creation of effective worker retraining and adjustment programs should be high on the agenda.
Some training programs have a proven track record. The Job Training Partnership Act (JTPA) programs for disadvantaged adults; residential programs for at-risk youths; the San Jose Center for Employment and Training; some welfare-to-work programs; and job search assistance all show returns of 40 percent or more for each dollar spent. Small and medium-size employers could be given tax credits for formal training programs and/or allowed to treat training expenditures as investments for tax and accounting purposes. Finally, it would be wise to improve the unemployment insurance system, which has been geared toward temporary job losses, not permanent job shifts and retraining needs.
For those who are too old or otherwise unable to adjust through retraining efforts, steps must be taken to ameliorate the burden of adjustment via transfer programs such as the earned-income tax credit, trade adjustment assistance, food stamps, and the like.
Can we afford it? Some estimates of the cost of ameliorating wage inequality range as high as $400 billion. However, not all of this money would need to be spent by the government, or in a single year. A well-targeted 10 percent annual increase in spending on education and training could go a long way toward raising the earnings of those with a high school education or less. These costs should be weighed against the risks to the economy and society of a break in the social compact that has yielded decades of unprecedented prosperity. If appropriate remedies to rising wage disparities are not identified and implemented, the backlash against globalization now spreading around the world will find willing supporters in the United States.
1. Jeffrey G. Williamson, Industrialization, Inequality and Economic Growth (Brookfield, Vt.: Edward Elgar, 1997).
2. Chris Mayer, "Does Location Matter?" New England Economic Review, Special issue (May/June 1996): 26≠40.
3. Paul Krugman, "Growing World Trade: Causes and Consequences," Brookings Papers on Economic Activity no. 1 (1995).
4. Andrew B. Bernard and J. Bradford Jensen, "Exporters, Skill Upgrading, and the Wage Gap," Journal of International Economics 42 (February 1997): 3≠31.
5. Dani Rodrik, ed., Has Globalization Gone Too Far? (Washington, D.C.: Institute for International Economics, 1997).
6. Matthew Slaughter, "International Trade and Labor-Demand Elasticities," National Bureau of Economic Research Working Paper no. 6262 (November 1997).
7. See for example, Alan A. Krueger, "How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984≠1989," Quarterly Journal of Economics 108, no. 1 (February 1993): 33≠60.
8. Eli Berman, John Bound, and Zvi Griliches, "Changes in the Demand for Skilled Labor within U.S. Manufacturing: Evidence from the Annual Survey of Manufactures," Quarterly Journal of Economics 109, no. 2 (May 1994): 367≠97.
9. For example, Robert Lawrence and Slaughter find that the ratio of (higher paid) non-production to production workers tended to increase both in the industrial country parent company and in its developing country affiliate. See "International Trade and American Wages in the 1980s: Giant Sucking Sound or Small Hiccup?" in Martin Neil Baily and Clifford Winston (eds.), Brookings Papers on Economic Activity: Microeconomics 2 (1993): 161≠211.
10. John Haltiwanger, "Measuring and Analyzing Aggregate Fluctuations: The Importance of Building from Microeconomic Evidence," Federal Reserve Bank of St. Louis Review (May/June 1997): 57≠78.
11. These controls are important, since large plants are more likely to use advanced technologies, to export, to be relatively capital intensive, to be more productive, and to pay higher wages.
12. World Labor Report 1997≠98 (Geneva: International Labour Organization, 1998), table 1.2.
13. Adrian Wood, "How Trade Hurt Unskilled Workers," Journal of Economic Perspectives 9, no. 3 (Summer 1995): 57≠80.
14. The authors employed a number of tests to establish that immigrants are not simply concentrated in the lowest-paying jobs.
15. George Borjas, Richard Freeman, and Lawrence Katz, "Searching for the Effect of Immigration on the Labor Market," National Bureau of Economic Research Working Paper no. 5454 (February 1996).