Since 1970 American incomes have become strikingly less equal. Living standards of poor and lower middle-class Americans have fallen while those of affluent Americans have continued to improve. And the trend toward inequality has not been confined to the United States. Nations throughout the industrialized world have seen income disparities rise since the late 1970s.
Many people blame rising income inequality on the growing importance of trade, especially trade with nations in the developing world, in the past quarter century. In The Trap, a bestseller in Western Europe, Sir James Goldsmith argues that free trade with low-wage countries has harmed and threatens to impoverish low-skilled and middle-class workers in the advanced industrial countries. A similar argument was used by Ross Perot and other U.S. opponents of the North American Free Trade Agreement, who warned that freer trade with Mexico would eliminate industrial jobs and reduce the wages of semi- skilled U.S. workers. More recently, Republican presidential aspirant Patrick Buchanan has called for an "equalization tax" on imports from third world countries to protect American workers against competition from Asian and Latin American workers who may be paid one-tenth the U.S. industrial wage. How well does the case against free trade stand up to the facts?
Trends in U.S. Income Inequality
There is no disputing the worsening trend in U.S. income inequality. Figure 1 documents the rate of change in income for Americans divided into five quintiles of the income distribution. Income changes are calculated after taking account of changes over time in the price level and changes in the size of families in different parts of the income distribution. (Average family size shrank after 1969, so it took less income to support families at the same standard of living.) The black bars represent income changes during 1969-79, the red bars, changes between 1979 and 1993. During 1969-79, Americans in all quintiles made income gains, though people in the lowest income quintile made the smallest percentage gains. After 1979, incomes fell in each of the three bottom quintiles while continuing to grow in the top two. Compounded over the 24-year period, the differences in the rate of change in income imply dramatic movements in relative well-being. At the 5th percentile, income fell 34 percent; at the 95th percentile, it rose 43 percent. In 1969, income at the 95th percentile of adjusted personal income was a little less than 12 times income at the 5th percentile. By 1993, income at the 95th percentile was more than 25 times income at the 5th percentile.
Several developments lie behind the widening income gap, and most of them have little direct or indirect link to liberalized international trade. Significant gains in capital income during the decade of the 1980s, for example, caused unearned income to grow strongly in the top part of the income distribution. And at the same time, changes in the pattern of government transfersþwith growing cash transfers going to the elderly, many of whom tend to be well up the income distribution, and shrinking cash transfers to the poorþreduced the effectiveness of those transfers in combating poverty. Changes in the structure of U.S. households also added to income inequality. Single-parent families are more likely to be poor than families that have two parents, and a much higher percentage of Americans now lives in single-parent families. Finally, the dramatic increase in paid employment among American women has tended to boost inequality since the late 1970s. In the 1950s and 1960s, families with a well-paid male earner were less likely than average to have a well-paid female earner. By 1993, families with a highly paid male earner were more likely than average to have a highly paid female earner. These non-trade-related economic and demographic trends account for more than half the growth in overall U.S. income inequality since 1969.
Even if trade is not to blame for trends in unearned income or changes in the composition of American households, it could still be an important source of growing wage inequality. Figure 2, based on annual earnings reports in the Census Bureauþs Current Population Survey, shows that between 1969 and 1993, earnings fell for men in the bottom 40 percent of the earnings distribution, remained unchanged for men in the middle quintile, and rose for men at the top. The disparate trends in wage earnings became more pronounced after 1979. Earnings fell sharply in low-wage groups, and wage disparities between well-paid and poorly paid men widened at an accelerated pace. Although overall wage trends have been much healthier for women (black bars in figure 2), women have also experienced widening earnings disparities, especially in recent years. After 1979 women in the top quintile saw their earnings climb more than 25 percent. For women at the bottom, annual earnings fell after 1979.
Does Trade Harm Unskilled U.S. Workers
The argument that trade is to blame for U.S. earnings inequality rests on the assumption that trade hurts U.S. workers with skills similar to those of workers in developing countries. The intuition behind this view is straightforward: very poorly paid unskilled workers overseas take away job opportunities and drive down the wages of unskilled American workers.
It is easy to see in theory how surging exports from developing countries could harm less-skilled U.S. workers in the trade-affected industries. To counter the competition from cheap unskilled labor abroad, American employers must reduce the wages, or make less intensive use, of unskilled labor if they wish to remain in business. Presumably, some employers who continue to rely heavily on unskilled workers will go bankrupt, others will move production overseas, others will adopt new technologies that permit them to dismiss some unskilled workers, and still others will specialize in new products where relative wages and factor prices favor production in the United States. No matter which alternative they choose, the demand for less-skilled workers in the traded-goods industries will fall. Shrinking demand will reduce the relative wage of less-skilled workers in comparison with highly skilled workers.
But if trade is the main factor behind the growing woes of unskilled workers in the traded-goods industries, then firms that do not produce internationally traded goods and services should take advantage of the shrinking wage of less-skilled workers by hiring more of them. If, instead, they also begin to pare back use of unskilled labor, it must be something other than (or in addition to) trade that is lowering the demand for less-skilled workers.
Figure 3 helps show whether trade is behind the drop in relative demand for less-skilled labor. It compares wage inequality trends among male workers in two broad classes of U.S. industries—one (including manufacturing, mining, and agriculture) that is highly trade-affected and another (including construction, retail trade, personal services, and public administration) that is not trade-affected. (An excluded group of industries, including transportation, wholesale trade, nance, and insurance, falls in an intermediate category.) Earnings inequality is calculated as the ratio of annual earnings at the 90th percentile of the earnings distribution to earnings at the 10th percentile. Male inequality is growing in both the most and the least trade-affected industries, and it is growing at the same rate—47 percent between 1969 and 1993. Although wages are more equal among women in trade- affected industries than among women in the least-affected industries, wage inequality among women has grown faster in the non-trade industries since 1979— the very period in which U.S. trade problems and manufactured imports were concentrated. When the data for men and women are combined, the earnings ratio in the most trade-affected industries rose 29 percent between 1969 and 1993—exactly the same as the rise in inequality across all industries.
The same pattern of relative earnings change is apparent in trends among workers with different levels of schooling. Educational pay premiums have risen since 1969 for every industry and for both sexes. But the premiums have not risen any faster in the industries most affected by trade than they have in other industries. For working men as a group, the premium for post-college education rose 36 percent between 1969 and 1993; for men in trade-affected industries, the premium rose 33 percent. And the gap in pay between high school dropouts and men with some college rose exactly as fast among men in trade-affected industries as it did among men as a whole. Women in the trade-affected industries had a somewhat larger rise in the post-college pay premium than women in other industries, but the difference is comparatively small.
Even though wage inequality and educational pay premiums moved in the same pattern across different industries, liberalized trade may still explain the pronounced shift toward greater inequality. In a competitive and efficient labor market, pay premiums for skill and education should eventually rise and fall together across industries, whatever the reason for the change in pay premiums.
But if trade from newly industrializing countries in Asia and Latin America is placing special pressure on producers in trade-affected industries, we would expect these industries to shed low-wage workers faster than industries where competitive pressure comes exclusively from other domestic RMS. To what extent was this the experience in the United States?
Between 1969 and 1993, trade-affected industries did indeed reduce the percentage of less educated workers on their payrolls (figure 4). In 1969, 42 percent of male and 45 percent of female workers in trade-affected industries had no high school degree. By 1993, those figures had fallen to 18 percent for men, 17 percent for women. These trends certainly seem consistent with the view that liberal trade has deprived less-skilled workers of job opportunities in the traded- goods sector. But employment patterns in industries unaffected by trade moved in exactly the same direction. The percentage of male workers without a high school degree in the industries least affected by trade fell from 36 percent in 1969 to 13 percent in 1993. If anything, industries unaffected by trade cut their use of low-skill workers even faster than trade-affected industries—a pattern that is extremely hard to square with the claim that foreign trade is the main factor behind soaring wage inequality.
Beset on All Sides
Over the past quarter century, the nation has seen a dramatic shift in the pattern of demand for workers with different levels of skill. Job opportunities for the less skilled have shrunk, and relative wages for unskilled and semi-skilled workers have plunged. But these trends are not conned to the traded-goods sector. They are also apparent in industries, such as construction and retail trade, where international trade is a minor concern. International trade, it seems, has not been the decisive factor in the trend toward greater earnings inequality. Other developments have been at least as influential, if not more so.
Among economists, the leading explanation for increased wage inequality is changes in the technology of production. Such innovations as the personal computer or new forms of business organization have favored workers with greater skill and reduced the value of unskilled labor.
But other developments are also at work. Economic deregulation, new patterns of immigration into the United States, declining minimum wages, and the dwindling influence of labor unions have also contributed to the job woes of unskilled and semi-skilled workers. Liberal trade with the newly industrializing countries of the world has certainly played a part in worsening the job prospects of Americas unskilled workers. But if we follow the advice of Ross Perot and Patrick Buchanan and erect a new wall of trade protection, we would do little to ease the plight of less-skilled workers. Too many other forces are conspiring to push their wages down.