By Sheldon Danziger and Deborah Reed(*)
From Foresight, Vol. 7, No. 3
published 2000
Income inequality in the United States has grown substantially in the past quarter century. Even the current period of sustained economic growth has done little to stem the tide. Market forces are unlikely to alleviate the hardship for low-income families who have borne the brunt of economic changes in the near future. As inequality remains high, we propose several public policy reforms that would raise living standards for low-income families and workers.
Census Bureau statistics report that family income inequality (according to numerous summary measures) reached a postwar low in the late 1960s and climbed almost continuously from that time. Inequality has been higher in the 1990s than in any decade since the end of the second World War. Most industrialized nations have also experienced growing income inequality, but the rise in the United States has been more rapid and started from a higher level. This differential experience is due primarily to the fact that other countries have pursued labor market and tax and transfer policies that have done more to counter the increased inequality generated by market forces.
We measure inequality by the relative income gap between the upper and lower “middle class,” that is, the ratio of income at the 75th percentile to that of the 25th percentile— the 75/25 ratio. The black line in Figure 1 traces that ratio from 1973-1997 and shows that inequality in family income rose over that time. In 1973, an upper-middle income family had 2.4 times the income of a lower-middle income family. By 1997, the income ratio had increased to 3.0, only slightly lower than the highest value of 3.1 in 1993.
Although inequality has increased over the entire period, different parts of the income distribution had differing experiences during each decade. During the 1970s, inequality rose because the income of families in the upper-middle grew faster than the income of those in the lower-middle of the distribution. Between 1973 and 1979, income at the 75th percentile grew by 9 percent from $64,200 to $70,000 while income at the 25th percentile grew by 4.5 percent from $26,300 to $27,500. However, during the 1980s and 1990s, increased inequality resulted from rising incomes for families at the upper middle but falling incomes for those at the lower middle. By 1997, family income at the 75th percentile had grown to $80,500 while income at the 25th percentile had fallen to $26,900 (incomes in 1998 dollars, adjusted to represent a family of four).
Looking at the extremes of the distribution, family income fell at the bottom and grew at the top over the last quarter century. The income of families at the 10th percentile fell 7 percent between 1973 and 1997 (see Figure 2). At the 90th percentile income grew 38 percent. Figure 2 also shows that inequality increased throughout the distribution, as the size of each bar increases from the lowest through the median to the highest income families.
Figure 2: Percentage Change in Real Family Income & Male Weekly Wages
Changes in inequality can also be evaluated by comparing the percentage of people who are poor with the percentage who are “rich.” We define as “rich” persons living in families with incomes more than seven times the poverty line (about $105,000 for a family of four in 1997 using an alternative price index). Between 1973 and 1997, the percent of poor increased by 1.2 percentage points to 11.8 percent of all persons while the percent of rich increased by 8.1 percentage points to 14.3 percent of persons. This change further indicates that economic growth has raised incomes at the top of the distribution, while absolute incomes at the bottom have stagnated or fallen.
The robust economic recovery of the 1990s has produced the lowest unemployment and inflation rates in 30 years and has lifted living standards across the income distribution. However, families below the median have not yet fully recovered from the income stagnation and recessions of the early 1980s and early 1990s. Family income in the lower middle of the distribution remains a few percentage points below its 1989 and 1979 peak levels, whereas the upper middle has increased a few percentage points since 1989 and 15 percent since 1979. As Figure 1 shows, as the economy recovered from the recession of the early 1990s, family income inequality fell slightly. But as of 1997, the 75/25 ratio was still higher than the levels of the last three business cycle peaks (1973, 1979 and 1989).
Male earnings are the largest single component of family income, and changes in their distribution account for much of the increased family income inequality. The current economic expansion brought some real growth to the bottom of the male wage distribution between 1996 and 1998. However, the 10th and 25th percentiles of weekly wages for full-time male workers, $300 and $432 respectively, remain more than 6 percent below their level at the 1989 business cycle peak and more than 15 percent below that of the 1979 peak. The 75/25 ratio for male weekly wages was just below 2.2 in 1998, slightly under its highest value, attained in 1997, but well above the levels of the 1979 and 1989 business cycle peaks (1.7 and 2.0, respectively; see Figure 1).
Male earnings inequality rose over the 1979-1997 period because real earnings fell for workers at the median and below and grew for those at the top of the distribution. Figure 2 shows a decline of 22 percent in male earnings at the 10th percentile. The decline was 16 percent at the 25th percentile and 7 percent at the median. Over the same period, male earnings increased 6 percent at the 75th percentile and 13 percent at the 90th percentile.
Most economists agree that the main cause of growing earnings inequality is the rising value of worker skill. For example, in 1979, the median weekly wage of full-time male wage and salary workers with college degrees was 29 percent higher than the wage of men with high school diplomas only. By 1998, college graduates had a 68 percent edge. Over this period, real wages increased by 8 percent for male college graduates but declined by 18 percent for high school graduates. For women, the weekly wage gap between college and high school graduates increased from 43 percent in 1979 to 79 percent in 1998. There was also an increase in the experience differential: Earnings of workers with substantial labor market experience increased relative to those of new labor market entrants.
There is some disagreement about both the relative importance of various causes of the rising value of skill and how much other factors contribute to rising inequality. No single factor can account for most of the inequality increases, but several factors are important. Labor-saving technological changes have increased the demand for skilled workers who can run sophisticated equipment and simultaneously reduced the demand for less-skilled workers, many of whom have been displaced by automation. Global competition has increased worldwide demand for the goods and services produced by skilled workers in high-tech industries and financial services. Lower-skilled workers increasingly compete with low-wage production workers in developing countries. Immigration has increased the size of the low-wage workforce and competition for low-skilled jobs. Institutional changes, such as the decline in the real value of the minimum wage and shrinking unionization rates, also moved the economy in the direction of higher earnings inequality.
Factors other than rising earnings inequality also contributed to the increase in family income inequality. In particular, changes in family structure, especially the growing share of female-headed households, increased the number of low-income families. The increased tendency for high-earning men to be married to high-earning women further separates the incomes of dual-earning couples from those of female-headed households.
Income inequality has trended upwards for a quarter century. The family income gap and the earnings gap between more-educated and less-educated workers are much higher now than they were in 1973. The long and robust economic recovery has generated real income growth across the distribution but has done little to reverse the trend. Because growing inequality is predominantly due to long-term structural changes in labor markets that we expect will continue—especially technological changes and globalization—income inequality is likely to remain high in the United States in the coming decade.
The current recovery produced a small decline in the income gap between families in the upper middle of the distribution and those in the lower middle. The greatest single year of improvement occurred between 1993 and 1994. The 1997 income gap is so large, however, that it would take nine additional years of declining inequality of this magnitude for family income inequality to return to 1979 levels and almost twelve years to reach 1973 levels.
Because the economic returns to skills have increased so much, the labor market now provides incentives for workers and young people to upgrade their skills through education and training. Indeed, the percentage of high school graduates entering college has increased in the last few years. The resulting growth in supply of skilled workers will eventually reduce labor market inequities somewhat. However, it will take many years for such an adjustment to have a large effect. Thus, the wages of less-skilled workers are likely to remain low in the interim.
Furthermore, those still in school and young workers are the most likely to respond to the incentives due to wage premia. Prime-age workers who have been hurt by changes in the labor markets over the last quarter century are unlikely to undertake substantial investments in education or training programs because they have relatively few years of work left before retirement. In addition, children from the poorest families and racial/ethnic minorities who are concentrated in the inner cities typically attend lower-quality schools, are more likely to drop out of high school, and are less likely to attend college than are children from higher-income families. Taken together, the increased inequality of the last quarter century and pervasive inequalities in school quality suggest that the supply responses of disadvantaged young people to the increased wage premia are likely to be smaller than average.
We are concerned about increased inequality over the last quarter century because of its effects on both the absolute and relative well-being of low-income families and workers. Even without a decline in income for those at the bottom, we care about rising inequality because, as Adam Smith noted, the minimum acceptable standard of living tends to be higher the richer the society. Moreover, according to recent Nobel laureate Amartya Sen, the absolute well-being of the poor in terms of their ability to "participate in the standard activities of the community" depends on their relative well-being in terms of resources.
Rising income inequality is likely to make equality of opportunity harder to attain. The children of the poor are increasingly subject to lower-quality education, lower-quality health care, and more dangerous communities. Concerns about equal opportunity are particularly relevant for children from female-headed households and those who are racial and ethnic minorities because they are far more likely to grow up in poor families.
Many of the economic forces that contributed to the rising income inequality have also brought positive changes in our economy. Technological changes and increased globalization raise the average standard of living by bringing new goods to consumers and producers and by reducing their prices. The rising value of skill provides incentives for people to upgrade their own skills, which can be financially and personally rewarding. We see no reason to attempt to lower inequality by slowing technological changes or adopting protective barriers to trade, even if there were feasible options to do so.
The United States has pursued policies to promote free trade and technological advancement in the interest of growth and efficiency. Those policies have produced winners and losers. Government policies should play a greater role in reducing the resulting inequities by aiding the low-income families and less-skilled workers who have borne the brunt of labor market changes.
Reasoned policy in this area must take into account two realities. First, the recent welfare reform experience demonstrates that the public favors policies that promote work. Welfare reform has dramatically reduced cash assistance for people who are capable of working, but it has offered expanded wage and child care subsidies. Second, education and well-designed training programs can improve the earnings of workers and promote economic opportunity in the long run. However, the effects of these policies will be slow and are unlikely to substantially improve the well-being of low-income workers in the coming decade. Thus, we support an expansion of labor market-oriented antipoverty policies that raise the incomes of the least-skilled workers.
For people who are able to find jobs, the key elements of support are expanded wage supplements and refundable child care tax credits. The Earned Income Tax Credit (EITC), substantially expanded in 1993, has done much to offset the decline in real wages for workers at the bottom of the earnings distribution who work year round and who have children. However, only a small percentage of low-wage workers who do not have children receive the EITC, and their maximum credit is only a few hundred dollars. Almost all working poor and near-poor families with children are now eligible for the EITC, and they are eligible for substantial credits. For example, a single mother with two children who works year round, full time at the minimum wage receives about $3600. The EITC for childless workers should be raised substantially. This change would make the federal income tax more progressive and increase their living standards without taking them through the welfare system. Several states have also adopted their own EITCs for families with children, something other states should consider, especially those that continue to impose income taxes on the working poor.
In addition, even though the employment rate and earnings of single mothers have increased substantially in the past five years, many, especially former welfare recipients, have great difficulty working full-time, full-year when their children are young. As a result, many remain poor despite their increased earnings and the expanded EITC. An increase in the minimum wage would be particularly beneficial for this group, as would increases in public subsidies for child care. Many of the working poor spend a substantial portion of their earnings on child care. The Dependent Care Credit (DCC) provides tax relief in the form of a nonrefundable credit that depends on family earnings and the amount spent on child care. However, it only benefits families with positive income tax liabilities. If the DCC were made refundable, it would raise the disposable income of single mothers and other low-income working families who spend substantial sums on child care but do not owe federal income tax. This reform would also make the federal income tax more progressive.
Finding a job has become more difficult for less-skilled workers over the past quarter century. In mid-1999, even with the lowest unemployment rate in 30 years, many less-skilled workers were unable to find work. For those who want to work but are unable to find regular employment, transitional public service "jobs of last resort" at wages just below the minimum can provide the basis for a work-oriented safety net. Such "jobs of last resort" are more important now that we have "ended welfare as we know it." During the next recession, many former welfare recipients will find themselves out of work and without recourse to cash assistance because of time limits, sanctions, and other aspects of welfare reform.
The good news is that the current economic recovery seems to have slowed and might have ended the quarter century trend toward rising income and earnings inequality. The bad news is that there is little evidence that inequality will return to the level of the late 1970s any time soon, much less to the lower levels of the late 1960s and early 1970s. Given that this era of inequality continues, we have offered several suggestions for work-oriented policy reforms, which at a relatively modest cost, could greatly benefit those workers and families who have been most disadvantaged by the structural economic changes we documented.
© 1999 Brookings Institution
This article originally appeared in the Fall 1999 issue of Brookings Review.
* Sheldon Danziger is Henry J. Meyer collegiate professor of social work and public policy at the University of Michigan, Ann Arbor. Deborah Reed is a research fellow at the Public Policy Institute of California. Return to text.