STATE OF WORKING AMERICA 2000-01
INTRODUCTION AND EXECUTIVE SUMMARY
Table of Contents
Introduction: The living standards debate
Documentation and Methodology
Slower growth, greater inequality and much more work
Unexceptional median income growth, but glimpse of change in late 1990s
An income ‘generation gap’
Among racial/ethnic groups, African Americans make relative gains
For family types, strong growth in 1990s among dual-earner couples
Growing inequality of family income
Counter-arguments to the evidence on income trends
Are taxes the reason for rising inequality and disappointing growth in family incomes?
Is the increase in inequality sensitive to income definitions?
Alternative explanations: mobility and demographics
Growth in inequality narrows pathways to prosperity
Expanding capital incomes
The ‘time crunch:’ married-couple families with children working harder than ever
Turnaround in the late 1990s
Contrasting hours and hourly wage growth
Contrasting compensation and wage growth
Wages by occupation
Wage trends by wage level
The male/female wage gap
Shifts in low-wage jobs
Trends in benefit growth and equity
Explaining wage inequality
Productivity and the compensation-productivity gap
Rising education wage differentials
Young workers’ wages
The importance of within-group wage inequality
School quality and tests
Wage growth by race and ethnicity
The shift to low-paying industries
Trade and wages
The union dimension
An eroded minimum wage
Summarizing the role of labor market institutions
The technology story of wage inequality
Information technology workers
Executive pay soars
What does the future hold?
Sustained low unemployment key to recent progress
Unemployment and underemployment
Unemployment and the earnings distribution
Job stability and job security
Declining job stability
The contingent workforce
Nonstandard work: widespread and often substandard
More than one job
Deeper in debt
Wealth and worth
Low net worth
The roles of measurement, growth, family structure, and work
Who are the poor?
Alternative approaches to measuring poverty
Poverty, overall growth, and inequality
The role of demographics and inequality
The changing effects of taxes and transfers
The increasing prevalence of working poor
The low-wage labor market: workers’ characteristics and earnings
Variations across the country
Median family income grows in Midwest and South
The growth of income inequality by state
Trends in employment and unemployment and their impact on wage growth
Poverty rates vary greatly by region and area
The regressivity of state tax liabilities
Incomes and productivity: United States loses edge
Workers’ wages and compensation: slow, unequal growth
Household income: slow, unequal growth
Employment and hours worked: strength of the U.S. model?
Evaluating the U.S. model
Appendix A: Family income and poverty data
Appendix B: Wage analysis computations
Appendix C: Information technology and productivity
The living standards debate
For most workers the early part of the economic expansion that began in 1991 was disappointing: incomes continued to decline and poverty increased during what came to be called the “jobless” recovery. This was also a period of great worker insecurity, downsizing, and low productivity growth. But economic conditions have improved remarkably since 1995. Persistent low unemployment helped to boost wages, brought new workers into the labor market, and allowed workers to move from substandard jobs temporary, part-time, or without benefits to better, more regular jobs. The best news is that low-wage workers and low-income families have benefited most, thereby reversing nearly 20 years of declining wages and incomes.
The turnaround from widespread wage decline between 1979 and 1995 to widespread wage growth since is a significant new development for working Americans at the start of the 21st century. But the late 1990s were remarkable for two other stories as well:
- The growth of labor productivity the value of the goods and services that an average worker produces in an hour of work has accelerated. Labor productivity is probably the best indicator of an economy’s capacity to provide for the material needs of its population, and since 1995 it has grown about 2.5% per year, well above the 1.4% rate that prevailed from the mid-1970s through the mid-1990s.
- The shape of wage inequality changed in a subtle but important way. In the 1980s, inequality “fanned out” the top pulled away from the middle, and the middle pulled away from the bottom. In the 1990s, however, wages at the bottom and the middle grew closer, while the top pulled further away from the middle. Nevertheless, inequality in the U.S. remains high, and the gap between the best-off and those in the middle continues to grow.
New trends in the 1990s
Broad-based wage growth. From 1 995 to 1998, the average hourly wage, adjusted for inflation, grew 2.6% a year, far better than the 0.6% annual growth during the 1989-95 period or in the two prior business cycles of 1973-79 and 1979-89. Because the value of employer-paid health insurance and pensions fell, though, recent overall compensation growth of 1.9% has been less than the 2.6% in wage growth. The real wage growth in the late 1990s was enjoyed across the wage structure, by low-, middle-, and high-wage earners and by both men and women. Wage growth has also been greater among lower-wage workers than among high-wage (or middle-wage) workers, marking a dramatic turnaround from the early 1990s and the prior two decades, when low-wage workers not only fared worse than other workers but actually saw their wages continuously decline after inflation. For instance, the wages of low-wage men ($8.02 in 1989) fell 0.9% a year from 1989 to 1995
and then grew 1.7% a year from 1995 to 1999. Moreover, the growth for low-wage men was faster than for middle-wage men (1.3%) and comparable to that of high-wage men (1.8%). Similarly, the wages of low-wage women grew faster (2.0%) than they did for middle-wage (1.4%) or high-wage (1.9%) women.
The quality of jobs has also brightened as unemployment has remained low. The share of workers who are “involuntary” part-timers—working part time but wanting a full-time job—dropped to 2.6% in 1999, down from 3.7% in 1995 and lower than at any time over the last few decades. This is equivalent to shifting 1.4 million involuntary, part-time workers into full-time jobs. At the same time, the number of workers counted as “marginally attached” to the labor force—those wanting work but who had stopped looking several weeks prior to the unemployment survey—dropped by 400,000.
Moreover, the attributes of jobs have improved remarkably since 1995. The share of workers with employer-provided health insurance—a clear dividing line between a “good” and a “bad” job—has grown since 1995, reversing the downward slide in health insurance coverage during the 1980s and early 1990s. The growth of part-time jobs relative to full-time jobs slackened, to the point that there were proportionately fewer voluntary part-time workers in 1999 than in 1995.
Work through temporary help agencies continued to grow throughout the recovery, but the growth slackened considerably after 1995—the share of workers employed by temporary agencies grew 60% from 1991 to 1995 but by just 26% from 1995 to 1999. Had the 1991-95 growth continued, there would have been 826,000 more people employed through temporary agencies in 1999. In addition, during the 1995-99 period the share of the workforce working more than one job—multiple job-holders—fell from 6.2% to 5.8%, reflecting fewer people working two part-time jobs or supplementing their full-time job with a part-time one. In terms of all types of nonstandard work—including regular part-time, temporary help agency, on-call, independent contracting, and contract firm work—the share of workers in these arrangements fell from 26.4% to 24.8% of total employment during 1995-99.
The tight labor markets have not only boosted pay and pushed job quality higher, but they have also allowed greater participation in the labor market, in particular by female heads of household affected by welfare reform. Both the employment rates and the hours worked by low-income single-mother families rose to historically unprecedented rates over the recovery, with much of the growth coming in the latter half of the 1990s. For example, the employment rates of mother-only families with some income from welfare grew from 40.4% in 1995 to 56.0% in 1998. Among poor families with children headed by single mothers, hours of work grew 40% in the 1990s, far more than for any other family type.
The shift to better job quality and the demand for the work of previously unattached or marginally attached persons reflect an improved bargaining position for workers, both individually and through unions. In other words, as a result of low unemployment workers have been able to obtain jobs that better fit their needs.
Productivity growth. The 2.5% annual growth in labor productivity since 1995 compares favorably to the 1.4% productivity growth of the early 1990s as well as the rates of the 1970s and 1980s. Productivity growth has been as fast in other recent decades, but, unlike in this instance, the fast growth tended to occur in the early years of recoveries.
The new shape of inequality. One feature of the economic landscape that has not changed is the unequal distribution of income and wage gains among the different economic classes and the continuing rise in these disparities. However, in the case of wages, the growth of inequality has shifted: the top earners are still pulling away from the middle class and the working class, but now low earners and the poor are closing the gap. The wage gains described above bear this out: low-wage men and women posted gains as high as those posted by 90th percentile earners, while the smallest gains went to workers in the middle. The widening of the wage gap between the top and the middle is even greater than that illustrated in the accompanying table, because the highest wage earners (say, those at the 95th percentile, who earn more than 95% of all workers) have had better wage growth than the merely high-wage earners (90th percentile). Climb even further up the ladder and into the ranks of the corporate chief executive officers and one will find pay growth that is indeed extraordinary—62.7% for CEOs versus 12.9% for the 90th percentile earner over the 1989-99 period.
At the bottom of the wage scale, the share of workers earning poverty-level wages dropped considerably during the 1995-99 period. Among black and Hispanic men, for instance, 5.4% of the workforce left poverty-level jobs for higher wages. Among women, 4.3% shifted to higher-than-poverty-level wages.
Income inequality, unlike wage inequality, continued its historic pattern even through the late 1990s boom, with the top pulling away from the middle and the middle pulling away from the bottom (although somewhat more moderately for the latter than in the past). This continued rise in income inequality is evident no matter how one calculates income—using the conventional Census measure of money income; measuring on an after-tax basis; adjusting for family size; or including government non-cash transfers (e.g., food stamps, Medicaid, and so on). Although estimates of income inequality in a particular year will differ depending on the definition used, all income measures show a significant rise in inequality over the 1990s.
The pattern of this new inequality is somewhat different than that for wages because family income comprises more than wage income. A significant amount of income at the bottom comes from government assistance (Temporary Assistance for Needy Families, or TANF, Social Security, unemployment insurance), and the top derives a huge share from capital income (dividends, capital gains, interest). As with wages, the incomes of those at the top—the upper 1%, 5%, or 20%—have grown significantly faster than those in the middle. For instance, the median family’s income grew 3.9% from 1989 to 1998, while that of a high-income family (at the 95th percentile) grew 11.6%. This divergence would be even wider if these numbers included the sizeable capital gains of recent years, gains that primarily accrue to the best-off families. For instance, capital gains made up about 7% of income in 1998-99, about double the share a decade earlier. The most comprehensive income measure available shows the ratio of the incomes of the upper 5% to the middle fifth rising continuously over the 1990s, from 6.7 in 1989 to 7.8 in 1999. By this measure, the growth of inequality at the top in the 1990s is about half as strong as in the 1980s.
Another benefit of low unemployment has been a lessening of the black/white wage and income gaps. For instance, the median family income among African Americans grew 1.1% a year from 1995 to 1998, more than double the 0.4% growth among whites. Correspondingly, the median hourly wage of black men and black women grew faster than the wage of their white counterparts. Given these impressive wage and income gains, it is not surprising that poverty fell after 1995, especially among minorities: it was down by 4.7 percentage points among Hispanics and 3.2 points among African Americans, but by just 0.7 points among whites. Even more impressive is that, after 1995, poverty rates fell by over 5 percentage points among minority children and by 1.1 points among white children.
Explanations for the recent trends
One major factor behind faster wage and income growth has been persistent low unemployment, which has helped spur pro
ductivity growth and has given workers leverage to obtain better wages and better jobs and to shift away from part-time and irregular, contingent-type work. Low unemployment has also been an important factor in lessening inequality at the bottom. Information technology (IT) is associated with these new trends in some ways, but there are also some trends—such as the new patterns of inequality—that are not associated with IT. For instance, the increased use of information technology has been a major factor propelling faster productivity growth, and yet high-tech-sector employers have not been wage leaders. Nor has technological change been associated with the rise of wage inequalities in the late 1990s or over the last few decades. Rather, globalization, deunionization, and the shift to low-paying industries have kept wage inequality at the top growing, while increases in the minimum wage have helped lessen wage inequality at the bottom in the 1990s.
Persistent low unemployment. The economy has not only achieved a relatively low unemployment rate—just 4.2% in 1999—but unemployment has remained low for several years. In fact, the unemployment rate has stayed at or below 5.6% since 1995, the first time since 1970 that unemployment has remained so low for more than two years in a row. This is what we call “persistent low unemployment,” unemployment that remains low enough long enough to require economic actors—businesses, workers, unions, and others—to adjust to the new environment.
Low unemployment is both a consequence and a cause of the higher productivity growth since 1995. The greater productivity growth offset the greater wage growth that accompanied low unemployment, thereby keeping inflation stable and forestalling any Federal Reserve Board action to slow the recovery’s growth rate. On the other hand, low unemployment likely led to higher productivity, because employers, unable to find or hire new workers in the tight job market, were forced to produce more with the same workforce, i.e., by increasing investment, upgrading worker skills through training, reorganizing the work process, and introducing new technologies.
Low unemployment has helped fuel faster growth and has lessened inequality at the bottom. But what is responsible for current low unemployment? Part of the answer is that inflation hasn’t accelerated, for two reasons. First, continued high levels of worker anxiety and corporate policies have suppressed wage growth, even among white-collar workers. Second, no major “supply shocks,” such as those that led to soaring energy and food prices in the 1970s, have beset the U.S. economy. In fact, the rapidly falling prices of computer-related equipment and the import price declines following the “East Asian” financial crisis might be considered positive supply shocks that helped reduce inflation. With inflation low, the Federal Reserve Board did not radically raise interest rates and end the recovery.
Another reason for the low unemployment is that the Federal Reserve Board has not followed the traditional policy of ending a recovery when unemployment fell below some benchmark, such as 5.5%. Instead, its wait-and-see policy has allowed unemployment to fall and, when no inflation ensued, the recovery to continue.
Rapid growth in demand for goods and services, due to faster consumption based on consumer debt and a higher stock market as well as a pickup in the growth of computer-related equipment investment, has also helped to lower unemployment.
Low unemployment has clearly contributed to the greater wage growth among low-wage earners relative to other workers. Low-wage earners benefit more from low unemployment, as reflected in a greater growth of annual earnings: this occurs because they experience both greater gains in employment and hours worked and attain a greater boost to their hourly wages.
As discussed above, this period of persistent low unemployment has allowed workers to obtain the types of jobs that better fit their needs. This shift to better job quality reflects an improved bargaining position for workers, both individually and through unions, and establishes that more leverage for workers, as well as greater productivity, is associated with the recent acceleration in wage growth.
The ‘new economy.’ This section explores the role of information technology in generating faster productivity, as well as one other dimension of what is referred to as the “new economy”: the rapid rise in stock market values and increased ownership of stocks. We also explore the role of the IT sector on labor market trends. We conclude that investments in information technology—hardware and software and the high productivity of sectors producing information technology—are a major reason for the higher productivity growth. But, we note that the IT sector has not contributed significantly to job growth, nor has it demonstrated any wage leadership. In fact, IT wages have failed to rise any more quickly than those of other sectors with similarly skilled workers. We also find that technological change has not been a significant factor in the growth of wage inequality in either the 1980s or the 1990s. As for the stock market, its growth in the late 1990s has been no more than a minor feature of economic life for most families, primarily because most families have little or no stock holdings.
Technological change, and particularly the increased use of information technology equipment and software, has been a major reason for the recent acceleration of productivity growth, accounting for between a third and a half of the faster productivity growth in the late 1990s, relative to that of the 1974-90 period (see Appendix C). Why there was a surge in IT investment starting in 1996 is something our analysis cannot answer and remains to be addressed by other researchers. It is our opinion that the 1995-99 IT surge was driven more by technological developments rather than any proximate economic policy (budget or tax policy, interest rates, deregulation). Other factors that help explain the recent productivity acceleration include: fast demand growth at low unemployment forcing greater efficiency; faster improvements in the organization of work; and, a faster growth of workforce skills. Contrary to the popular perception, the IT sector has not played a leadership role in the labor market. Although IT-producing industries generated jobs at a substantially faster rate than other industries in the 1990s and in the 1992-99 recovery, IT industries contributed a small share of total job growth—about 7.5% of all new jobs. IT occupations, such as programmers, systems analysts, and so on, made up 2.0% of all employment in 1999, up from just 1.3% in 1989. Thus, the IT sector contributed disproportionately to job growth, but was still not a major job generator.
Another common misperception regarding IT workers is that their wages are skyrocketing, reflecting the high demand for their skills (see Chapter 2). Yet a comparison of the wage growth among IT workers relative to comparable workers—those with similar education, experience, and occupation—finds that IT sectors were not wage leaders: among men, IT wages were stable relative to those of comparable workers since the mid-1980s; among women, IT wages fell relative to similar workers. Thus, the IT sector apparently is not experiencing any labor shortage and has not contributed directly (except through the overall productivity effect) to wage acceleration.
There is no evidence to support the notion that the growth of wage inequality reflects primarily a technology-driven increase in demand for “educated” or “skilled” workers. Indeed, technological change has led to a demand for education and skill during the entire 20th century, and probably no more during the 1980s or 1990s than in the 1970s. Without any evidence of “acceleration” of change, there is no reason to believe that technological innovati
on shifted in a way that led to the post-1979 growth of wage inequality. Second, skill demand and technology have little relationship to the growth of wage inequality within the same groups of workers (i.e., workers with similar levels of experience and education), which was responsible for half of the overall growth of wage inequality in the 1980s and 1990s. Technology has been and continues to be an important force, but there was no “technology shock” in the labor market in the 1980s or 1990s, and no ensuing demand for “skill,” that was not satisfied by the continuing expansion of the educational attainment of the workforce.
Moreover, the conventional story about technology leading to the increased demand for skills and the erosion of wages among the “less-skilled” does not readily explain the 1990s pattern of growth of wage inequality: neither the trends in education and experience differentials nor the trends in wage inequality among workers of similar skills. In particular, the 1990s are seen as a period of rapid technological change. Yet, this was the period of the lessening of wage inequality at the bottom. Similarly, education differentials grew slowly among men during most of the 1990s—it was the growth of other dimensions of wage inequality, not easily linked to technology, that kept male wage inequality at the top growing in the 1990s. During the technology-led boom of the late 1990s, there was a growth of education differentials, but other differentials, such as for experience, declined; so there was no growth of “skill differentials” overall during the technology boom.
Another widely heralded feature of the “new economy” is the rising stock market. Indeed, the last decade or so has witnessed a breathtaking run-up in the price of stocks. For instance, the inflation-adjusted value of the Standard & Poor’s 500 index of stocks tripled between the beginning and the end of the 1990s. These increases have focused enormous media and public attention on the stock market. Some pundits have even suggested that what happens to a family’s stock portfolio is more important than what happens to its paychecks. Yet, while stock ownership has spread and a small number of individuals have ridden the stock market boom to great personal wealth, data on stock ownership establish that the stock market, in practice, is of little or no financial importance to the vast majority of U.S. households.
Even well into the stock market boom of the 1990s, a majority of U.S. households had no stock holdings of any form, either direct or indirect (in 401(k)s, defined-contribution pension plans, or IRAs). In 1998, just under half (48.2%) of households owned stock in any form and only about one-third (36.3%) held stock worth $5,000 or more. In fact, the distribution of stockholdings is more unequal than the distribution of wealth in general. While the wealthiest 1% of all households control about 38% of national wealth, the top 1% of stockowners hold almost half (47.7%) of all stocks. In contrast, the bottom 80% of households (in terms of stockholding, many of whom own no stock) hold just 4.1% of total stock. The value of stockholdings grew across the board in the 1990s, but in dollar terms—and relative to the typical household’s retirement needs—the increases were small for 80% of households. For example, the value of the stock held by the middle 20% of households grew only $5,500 between 1989 and 1998 (and this growth includes any shifts they made from non-stock assets into stocks).
The high concentration of stock ownership means that the gains associated with the recent stock boom have been highly concentrated as well. Between 1989 and 1998, almost 35% of the growth went to the wealthiest 1% of households, and almost 38% of the total increase went to the next 9% of households, meaning that 73% of the gains from the stock market accrued to the upper 10%. The middle 20% of households accounted for only 2.8% of the rise in the overall value of stockholdings over the period.
The rising stock market may or may not be a consequence of an information technology “new economy,” but in any case its gains bypassed most households.
Rising minimum wage, contracting retail sector. Two developments in the 1990s help explain the better wage and job quality trends for women, especially those in low-wage employment. First, increases in the minimum wage in 1991-92 and again in 1996-97 raised its value in real terms by 14.4% over the 1989-99 period, sharply reversing its continuous erosion during the 1980s. Second, retail trade, the lowest-wage sector and an extensive user of part-time, mostly female workers, expanded in the 1980s but contracted (relative to total employment) in the 1990s. Together, these factors—along with low unemployment—help explain why low-wage women fared far better in the 1990s than in the 1980s and why wage inequality at the bottom among women declined in the 1990s.
The relatively continuous growth of wage inequality at the top, reflected in the wage gap between the 90th or 95th percentile and the median worker, appears to result from trends that have been ongoing since the early 1980s—continued globalization, a shift toward lower-paying industries (except among women), and a weakening of unionism.
Getting better versus being good. In spite of the widespread living standard improvements accompanying the strong recovery of the late 1990s, the fundamental economic situation, given the dramatic, broad-based wage erosion and rising inequalities of the 1979-95 period, can still not be considered “good.” For instance, over the 1989-99 period productivity, which might be considered the economy’s “ability to pay,” or its yardstick to measure good growth, rose 20.5%. However, typical workers received virtually none of this increase—the median hourly wage among men was slightly less in 1999 than in 1989, while for women it was up just 4%.
The divergence between the wage growth of typical workers and productivity growth arises because of two phenomena—rising wage inequality and a shift of income from workers to owners of capital. Of course, the wage gap between those at the top and those in the middle has grown steadily. In addition, a growing share of corporate income is paid to owners of capital, with a corresponding lower share paid out as compensation. Consequently, the returns to capital have soared in the 1990s to historically high levels. Without this ratcheting-up of profitability, average compensation could have been 4.3% higher in 1999.
Also, while a middle-class, married-couple family’s income grew 9.2% from 1989 to 1998, a substantial part of this growth reflected a growth in family work hours, up 182 hours to 3,600 total, or about 4.5 extra full-time weeks a year since 1989.
The continued growth of inequality over the 1990s, compounding the dramatic inequality surge of the 1980s, leaves the nation with very high levels of inequality both in historic terms and when compared to other industrial countries. Despite a higher average-income level in the U.S., low-wage workers here earn substantially less than those in other advanced countries. In fact, the typical low-wage worker in an advanced European economy earns 44% more than in the United States, and the percentage of U.S. children who are poor is twice as high as in other advanced countries. In 1998, the U.S. poverty rate was 12.7%, comparable to the 12.8% rate of 1989 and higher than the 11.7% rate of 1979—despite more than 20 years of growth. Almost one in five (18.9%) children were poor in 1998, including more than a fourth of African American and Hispanic children.
Other persistent economic problems include a low share of the workforce having employer-provided health insurance coverage—62.9% in 1998 compared to 70.2% coverage in 1979—and the share of the workforce—nearly half—without pension coverage.
although the current income boom has generated substantial improvements, by many measures of adequacy, inequality, and income, the current economy still does not match up to reasonable expectations. We now turn to two particular stress points for working families—increased work hours and increased indebtedness.
The time-squeeze on working families. Family earnings growth over the past few decades has come increasingly from greater work effort—a rise in the number of earners per family and in the average weeks and weekly hours worked per earner. Along with the income generated by more work come greater costs for transportation, clothing, and child care. And more paid work comes at the expense of leisure and time for “household production”—laundry, car-pooling, cleaning, and other maintenance activities. Generally, the increase in work hours has made the ability to balance both work and family a major challenge of family life today.
Over the 1980s the increased work effort of families occurred simultaneously with a fall in real hourly wages for men and for some groups of women. While this pattern changed in the 1990s as husbands’ wages rebounded, the fact remains that over the 19-year period from 1979 to 1998, increases in annual family earnings were primarily achieved through more work rather than through higher hourly wages.
Much of this increased work effort has, of course, come from wives. The ability of women to enter the labor market and reduce decades of gender discrimination represents a positive social and economic change. At the same time, though, for many families the increased work effort of their female members has been forced upon the family as the only way to keep income growing. (“Work” in this section refers exclusively to labor market work. Of course, women have long been the major contributors to non-market work.) In this regard, families are clearly worse off if their primary means to obtain higher incomes is more hours of work rather than regular pay increases.
The average family, with the combined help of all of its members, now works 83 weeks a year, up from 68 weeks in 1969, or an additional quarter-time worker in every married-couple family with children. The greatest increase in weeks worked has been among middle-class families, whose weeks worked grew by the equivalent of a person working more than one-third of a year (19 weeks). The growth in weeks worked among higher-income households, by contrast, was only half as much. We can track trends in family hours, but only back to 1979. As would be expected, the pattern for hours is similar to that of weeks, with smaller increases in the 1990s relative to the 1980s. While the number of hours worked by married couples in the lowest fifth is consistently lower than those of better-off fifths, their increase in hours over the 1980s—11%—was greater than that of higher income families. Middle-income families experienced the greatest percent growth of hours in both periods: 11.2% over the 1980s and 6.8% in the 1990s. The result: middle income families have added 12.5 weeks (just over three months) and 613 hours since 1979.
Some critics have argued that the increase in inequality has been generated simply by those with higher incomes working longer and harder while other less well-off families did not increase their work effort. But the data on weeks and hours worked belie this claim: the increases in hours and weeks occurred throughout the income distribution, with the greatest increases among middle-income families.
In a similar vein, work hours have increased about as quickly among families headed by someone with more (“at least some college”) as with less (“high school degree or less”) education. Among middle-income families, hours are far higher among families with less education (4,156 per year, or two full-time workers) than with more education (3,712 hours), and hours growth since 1979 has been greater among the less-educated families. Thus, again we see that problems balancing work and family are not “upper-income” problems and that families that are falling behind (those headed by someone with less education) have done so despite a greatly increased work effort. Hours of work have grown especially quickly among black families, particularly among those in the lowest fifth of the income distribution. In fact, average work effort of African American married-couple families with children grew more than it did for either whites or Hispanics in both the 1980s and the 1990s. Although they started out the period working fewer hours, by 1998 low-income black families worked more hours than either white or Hispanic low-income families. Among middle-income families with children, African American families worked more hours than families in other racial/ethnic groups in 1998. Their 4,278 annual hours were 489 more than for white families and 228 hours more than Hispanics. Therefore, an average middle-income black family with children needed over 12 more weeks of work than the average white family in order to reach the middle-income ranks.
High-income minority families consistently had hours of work above 4,500, suggesting full-time work by both parents, with contributions by other family members as well. Thus, these data suggest that, due to their lower wage levels, for minority families to make it to the top or middle of the overall income distribution they have to put in extremely long hours relative to whites. By 1998, the highest-income minority families were working at least 500 more hours per year than white families with comparable incomes. Across education and racial groups, and across the income scale, hours spent in the labor market have clearly increased for prime-age, married-couple families with children. By 1998, the average middle-income family was spending 87 weeks and 3,600 hours at work, compared to 75 weeks and 3,041 hours in 1979. Clearly, the time crunch is a real phenomenon, and it is one experienced by the majority of working families, not just the elite lawyers, money managers, or “knowledge workers” of the new economy. These data also deny the notion that increased income inequality among families is due to a greater increased work effort among upper-income families.
Greater debt. In 1999, the total value of all forms of outstanding debt was greater than the total disposable income of all households. In contrast, household debt was about 20% of total disposable income after World War II and 60% in the early 1960s.
For typical households, debt levels are high compared to the value of assets. In 1998, the average outstanding debt (typically, outstanding mortgage debt plus credit card debt) for households in the middle 20% was $45,800, about five times greater than the corresponding $9,200 average for stockholdings and about half the total value of other assets (overwhelmingly the family home). The increase between 1989 and 1998 in the average household debt held by the middle 20% (up $11,800) was much larger than the corresponding increase in both stocks ($5,500) and non-stock assets ($8,500). While middle-class households captured 2.8% of the total growth in stock market holdings between 1989 and 1998, they absorbed 38.8% of the unprecedented rise in household debt over the same period.
In and of itself, debt is not a problem for households. In fact, credit generally represents a tremendous economic opportunity for households, since they can use it to buy houses, cars, and other big-ticket consumer goods that provide services over many years; to cope with short-term economic setbacks such as unemployment or illness; or to make investments in education or small businesses. Debt becomes a burden only when required debt payments begin to crowd out other economic obligations.
Fortunately, the average household’s debt burden—the minimum required payments on outstanding debt as a share of disposable income (13.4% in 1999)—has not changed much over the last two decades, primaril
y because lower nominal interest rates have offset the higher debt. Consequently, though, households are much more vulnerable to high interest rates than in the past.
An increasing portion of households, however, are experiencing financial hardship from their debts. One indicator of “hardship” is service payments equal to more than 40% of household income. By this measure, about 14% of middle-income households and 20% of households in the $10,000-24,999 income range experienced financial hardship. Despite the strong recovery of 1995-99, the share of households with high debt-service payments increased significantly in the 1990s. Between 1989 and 1998, for example, the share of households facing high debt burdens increased 4.7 percentage points among households in the $25,000-49,999 range and 4.9 percentage points among households in the $10,000-24,999 range. The ultimate indicator of debt-related difficulties is personal bankruptcy, and in 1999 about six out of every 1,000 adults declared personal bankruptcy, almost twice as many as in the last business cycle peak in 1989. Despite the strong economic recovery during the second half of the 1990s, personal bankruptcy increased continuously.
The dramatic improvement of trends over the 1995-99 period raises the question of what will happen in the future. Are we now enjoying a temporary reprieve from the wage declines and surging inequalities of the 1980s and early 1990s? Are we in a “new economy” that will generate continued prosperity?
If we are right about the explanations for the recent prosperity, then the future is partly dependent on the policy choices we make and partly dependent on the unfolding logic of today’s technologies. A major part of the recent good news regarding wage growth is that it is connected to faster productivity growth, which was brought about, in large part, by the increased use of information technology hardware and software. Precisely what brought about this technology surge is beyond the scope of this work, yet it is safe to say that technological change is partly the result of economic trends and public policies (government research and development, education policy) and partly of the result of a host of factors—invention, or pure accident—that are beyond prediction. But if the technology trend continues, then productivity growth will remain strong, although probably not as strong as in recent years. The result will be continued real wage growth, at least on average.
The persistent low unemployment of recent years has generated much of the gains documented in this book, especially the improved quality of jobs and the stronger-than-average wage and income growth at the bottom. Although some external (to the economy and to policy makers) factors affect unemployment levels, the achievement of low unemployment—below the 6% rate many economists previously recommended—is primarily a policy outcome. At some point the recovery will end and a recession will ensue, and how fast and by what means we reestablish low unemployment will be a major public issue, particularly for low-income families. In the wake of welfare policy changes, low-income families now depend much more on wage income and less on government assistance than before. When unemployment rises in the next downturn, low-wage workers will certainly experience the greatest rise in unemployment, and their families will have a weaker safety net to fall to. Therefore, limiting the next downturn as well as ameliorating its effect need to be on the public agenda.
Another policy, the raising of the minimum wage, has helped promote growth at the bottom. Policy decisions on further increases will determine how low-wage workers, especially women, fare in the future.
Thus, how the future gains from productivity are shared is still an open question. Unbridled globalization, continued deunionization, an eroded minimum wage, and further privatization and deregulation will not only reinforce today’s high level of inequality but expand upon it. Keeping unemployment low, providing income supports for the bottom (through the minimum wage, tax credits, housing subsidies, child care, and so on), facilitating collective bargaining by workers, providing national health insurance, and strengthening private pension coverage will lead to a sharing of productivity growth and a lessening of inequality. Whether the inequalities that have emerged are reversed in the near future will greatly affect the future living standards of working families.
A comprehensive review of the vast amount of data that describe the working world at the turn of the new century reveals three important new developments in the state of working America.
The first development is rapid wage growth. After more than 15 years of stagnation and decline, inflation-adjusted wages began to rise in 1995. What’s more, these increases have been largest for workers at the bottom, reversing, at least for now, a long-term trend toward widening wage inequality.
The second new development is an acceleration in the growth of labor productivity—the value of the goods and services that an average worker produces in an hour of work, probably the best indicator of an economy’s capacity to provide for the material needs of its population. Since 1995, labor productivity has grown about 2.5% per year, well above the 1.3% rate that prevailed from the mid-1970s through the mid-1990s.
The final development is a subtle but important change in the shape of wage inequality. In the 1980s, inequality “fanned out”—the top pulled away from the middle, and the middle pulled away from the bottom. In the 1990s, however, wages at the bottom and the middle grew closer, while the top pulled further away from the rest.
One of the major tasks of this book is to document and explain these important new features of the economy. Persistent low unemployment—the unemployment rate has remained below 5.5% since February 1996—plays a key role in the explanation of these recent and dramatic improvements in the labor market. When unemployment is low, workers can more easily press for higher pay, better benefits, and improved working conditions because employers cannot easily replace dissatisfied workers with less demanding ones from the pool of the unemployed. When unemployment remains low for a prolonged period, the work environment changes. Workers feel more and more empowered and employers become more and more sensitive to workplace issues that can affect recruitment and retention of workers.
Sustained low unemployment helps to explain all three of the new features of the economy at the turn of the century. Both unionized and non-unionized workers have translated the increased bargaining power that comes with low unemployment into higher wages, a move that helps to explain the reversal of long-term wage stagnation. The tight labor market also means that the economy is operating closer to its full capacity, where efficiency is often highest, and this capacity utilization can account for at least part of the acceleration in productivity. Sustained low unemployment has also probably contributed to the new shape of inequality. A tight labor market most positively affects the workers who have the fewest skills, the least attachment to work, the most exposure to discrimination, and typically the lowest wages. As a result, the tight labor market of the late 1990s boosted the wages of workers at the bottom even more than it did those of workers at the middle and the top.
Another factor that contributes to understanding these recent developments is the “new economy.” Our reading of the available data concludes that investments in information technology (hardware and software)—and the high productivity of sectors producing information technology—are the major reasons for the rece
nt acceleration in productivity growth. Taken together, various aspects of the investment in and deployment of information technology account for about two-thirds of the acceleration in average labor productivity between 1974-90 and 1996-99.
At the same time, though, the information technology (IT) sector has not contributed significantly to job growth or wage improvements in the 1990s. Although information-technology-producing industries generated jobs at a substantially faster rate than did other industries in the 1990s, information technology industries still contributed only a small share to total job growth—about 7.5% of all new jobs. Information technology occupations, such as programmers, systems analysts, and so on, made up just 2.0% of all employment in 1999, up from 1.3% in 1989. IT sectors were not wage leaders: among men, IT wages were stable relative to those of comparable workers since the mid-1980s; among women, IT wages fell relative to those of similar workers. Nor did information technology drive wage growth in the 1990s. Among men, workers in information technology jobs did no better than workers in other kinds of jobs with comparable education and experience; among women, information technology workers actually lost some ground relative to comparable workers in other sectors.
Two other developments help to explain recent improvements in wages, especially wages for women. One is the series of increases in the federal minimum wage implemented in 1990, 1991, 1996, and 1997. These increases represented a sharp reversal of the situation in the 1980s, when the minimum wage’s purchasing power eroded continuously. The other development was the decline in the employment share of retail trade, a sector known for its low wages and extensive use of part-time and women workers. Any shift of employment out of retail trade into other sectors would help to raise overall wages; given the high concentration of women in the sector, women’s wages benefited most from the shift.
Do these many improvements mean that we are living in the “best economy ever”? Any serious analysis of the condition of today’s workers must distinguish between “things getting better” and “things being good.” The last five years have seen significant improvements over a broad range of economic indicators important to workers, including unemployment, wages, and incomes. But these improvements follow a long period of stagnation and decline in living standards— wage and income inequality remain high; tens of millions of workers still lack health insurance, pension plans, and vacation and sick pay; families are working more hours and feeling the “time squeeze” more acutely than at any point in the postwar period; and households are burdened with the highest levels of debt in history. Recent improvements are heartening, but much remains to be done.
Incomes rising, with slower growth in income inequality
Despite strong growth in inflation-adjusted incomes in the last half of the 1990s, real income growth over the entire decade was slow and unequally distributed. Between 1989 and 1998, the inflation-adjusted income of the median, or typical, family grew just 0.4% per year—identical to the disappointing average growth rate of the 1980s and well below the 2.6-2.8% average rate from 1947 to 1973. Even the income gains of the late 1990s—median inflation-adjusted income grew 2.5% a year—fell short of the average rate achieved during the first 30 years of the postwar period.
Income inequality continued to grow in the 1990s, though at a slower rate than in the 1980s. Between 1989 and 1998, the share of total income received by the bottom 20% of households fell 0.4 percentage points, while the share received by the top 5% grew from 17.9% percent in 1989 to 20.7% percent in 1998.
Income inequality continued to grow even during the boom in the second half of the 1990s. From 1995 to 1998, the real incomes of low-income families, or families in the 20th percentile, grew 1.9% each year. But this was slower than the 2.3% growth rate for families in the middle (60th percentile) and far slower than the 3.2% rate for families at the top (95th percentile).
Blacks and Hispanics made only mixed progress in the 1990s. Over the entire decade, the average income of black families grew 1.1% per year, more than double the rate for whites (0.4% per year). During the boom years of 1995-98, however, the growth in black family earnings (1.9% per year) trailed that of whites (2.4% per year). For Hispanics, the story was reversed. Over the full decade of the 1990s, Hispanic family incomes fell an average of 0.4% per year, but Hispanic families did much better than whites during the boom years, when average Hispanic family income grew 4.1% per year.
For the first time in the postwar period, the division of total corporate income between income paid to workers and income paid to owners of capital shifted strongly in favor of owners of capital during the 1990s. In 1999, owners of capital received 20.5% of the income paid out by the corporate sector, up from 18.2% in 1989. This 2.3 percentage-point rise in the “profit share” was more than four times larger than the 0.5 percentage-point increase between 1979 and 1989.
The general conclusions about slow income growth and rising inequality stand even after we consider changes in family composition (the increase in numbers of women raising families alone), changes in family size (the long-term decline in average family size), different definitions of income, and changes in tax structures.
The most important factor contributing to the income growth of the last decade was the increase in the number of hours that families worked each year. In 1998, for example, the typical middle-income married-couple family worked a total of 3,885 hours (adding spouses’ total annual hours together). This represented an increase of 247 hours—or about six weeks more than a similar family worked in 1989.
African American and Hispanic families worked far more hours than white families in the same income range. White married-couple families in the middle 20% of the overall income range, for example, worked an average of 3,789 hours per year. Black families in that same income range, however, worked an average of 4,278 hours per year, a difference of almost 500 hours per year. Hispanic families in the middle-income range worked an average of 4,050 hours per year, or about 140 hours more than white families with similar incomes.
Wages rise sharply in late 1990s
After more than 15 years of stagnant or declining wages at the middle and bottom of the wage range, the purchasing power of wages at all levels grew rapidly between 1995 and 1999. After adjusting for inflation, the median wage for all workers grew 7.3% during 1995-99. For male workers, the median rose 5.5%; for female workers, it rose 5.8%.
Wage growth during those years was even stronger for workers at the bottom, as the real wage of the 10th percentile worker rose 9.3% (9.8% for men and 9.1% for women). Wages for workers at the top grew even more: at the 95th percentile, real wages jumped 8.5% between 1995 and 1999 (11.7% for men and 7.8% for women). Workers in the middle experienced the smallest gains (median male earnings actually fell 1.2% during the 1990s). Wage inequality in the 1990s, then, involved the bottom and the middle growing closer, and the top pulling farther away from the rest. This pattern of wage growth represents a sharp break with inequality trends of the 1980s, when the top pulled away from the middle and the middle pulled away from the bottom.
Two forces seem to have influenced this new shape of inequality. The first is the series of increases in the federal minimum wage in 1990, 1991, 1996, and 1997, which boosted wages at the bottom but had less impact on wages in the middle. The second is the sustained low unemployment of the late 1990s, which had a beneficially disproportionate impact on the wage
s of workers at the bottom, where unemployment fell the most.
Other important developments in the wage and compensation distribution include the following:
- The real wage of the median CEO rose 62.7% during 1989-99, helping the average CEO to earn 107 times more than the typical worker. This ratio of CEO to worker pay was almost double the ratio of 56 in 1989 and more than five times greater than the ratio of 20 in 1962. According to the most recent data, U.S. CEOs also earn about 2.5 times more than their foreign counterparts.
- The wage gap between men and women narrowed in the 1990s. In 1999, the median woman earned 76.9% of what the median man earned, up from 73.1% in 1989.
- In 1999, just over one in four (26.8%) U.S. workers earned poverty-level wages—the wage required to lift out of poverty a family of four headed by a full-time, full-year worker (about $8.19 per hour in 1999). Women are much more likely to work in poverty-level jobs than men: about one in five (20.7%) men and one in three (33.4%) women earned poverty-level wages in 1999. Poverty-level work was especially common among racial and ethnic minorities. In 1999, 29.5% of black men, 40.7% of black women, 40.3% of Hispanic men, and 51.8% of Hispanic women were in jobs that paid poverty wages. At the national level, however, the share of workers earning poverty-level wages declined—among both white and black workers—as wages grew in the 1990s.
- In 1998, 62.9% of private sector workers had employer-provided health insurance, about the same rate of coverage as in 1989 (63.1%). Coverage is highly unequal across wage levels. In the same year, 82.3% of workers in the top fifth of the wage distribution had coverage, compared to just 29.6% of workers in the bottom fifth.
- In 1999, fewer than half (49.2%) of private sector workers had employer-provided pension plans. As with health insurance, the rate of pension coverage varies considerably across wage levels. Almost three-fourths of workers in the top fifth of wages had pension coverage in 1999, compared to less than one in five workers (17.9%) in the bottom fifth.
- In 1999, 27.3% of workers had a four-year college degree or more. Of these workers, about one-third (8.6% of the total workforce) had an advanced degree. Just over 10% of the workforce had less than a high school diploma. About one-third (32.3%) had a high school diploma or equivalent, but no further education. About the same number of workers (29.6%) had some college, but had not earned a four-year degree.
- Entry-level wages increased substantially in the second half of the 1990s. Between 1995 and 1999, real wages of young high school graduates increased 6.3% for men and 6.2% for women. Among young college graduates, real wages rose 14.9% for men and 9.4% for women.
- Between 1995 and 1999, real wages grew across almost all race and ethnic groups. Among men, median wages grew 6.2% for whites, 8.1% for blacks, and 11.8% for Asians. The median wage for Hispanic men, however, fell 3.5% over the same period. Among women, median wages grew 6.5% for whites, 5.5% for blacks, 8.2% for Hispanics, and 9.1% for Asians.
- Total compensation for union workers substantially exceeds that of non-union workers. Even after controlling for differences in the characteristics of union and non-union workers, union workers’ total compensation is about 27.8% higher than that of non-union workers.
- Even after four increases in the federal minimum wage in the 1990s, its inflation-adjusted value was more than 20% lower in 2000 than in 1979. The erosion of the real value of the minimum wage over the 1980s and 1990s was responsible for a significant share of the increase in women’s wage inequality over that period.
Sustained low unemployment key to improvements
One of the keys to recent improvements in hourly wages and family incomes is the sustained low unemployment that began in 1996 and continues through 2000. In mid-2000, the unemployment rate stood at about 4.0%. Since 1996, the unemployment rate has remained below 5.5%, a development that has significantly boosted the bargaining position of workers, especially those at the bottom and middle. The current period of sustained low unemployment is unprecedented in recent economic history: at no other time since 1970 has the unemployment rate remained below 5.5% for more than two consecutive years.
The national unemployment rate was low in 1999, but unemployment rates vary significantly by race and ethnicity. In 1999, unemployment for whites was 3.7%, less than half the rate for blacks (8.0%) and well below the rate for Hispanics (6.4%). Between 1989 and 1999, however, the unemployment situation improved more for blacks (down 3.4 percentage points) and Hispanics (down 1.6 percentage points) than it did for whites (down 0.8 percentage points).
The underemployment rate (which includes unemployed and discouraged workers, workers with only a marginal attachment to the labor force, and involuntary part-time workers) was 7.5% in 1999, about 3.3 percentage points higher than the standard unemployment rate.
All measures of employment—total non-farm payroll, total civilian employment, total hours worked, and full-time equivalent employment—show that employment grew more slowly over the 1990s than it did in earlier postwar business cycles. The share of working-age women with jobs grew significantly in the 1990s, from 54.9% of all working-age women in 1989 to 58.5% in 1999. Corresponding employment rates for men, however, declined slightly, from 74.5% in 1989 to 74.0% in 1999.
One welcome feature of the expansion of the late 1990s is the apparent reversal in the long-term trend toward a greater share of nonstandard jobs, or jobs that aren’t regular full-time positions. Between 1995 and 1999, the share of regular full-time employment rose from 73.6% to 75.1% of all jobs. During that same period, the prevalence of various forms of nonstandard work declined: part-time work dropped from 16.5% to 15.5% of all jobs; self-employment from 7.5% to 6.7%; and independent contracting from 6.7% to 6.3%. Other forms of nonstandard work remained flat over the period: temporary help agency employment held steady at 0.9-1.0%; on-call work at 1.5-1.7%; and work with contract firms at 0.5-0.6%.
Nonstandard workers, however, continue to suffer relative to their regular full-time co-workers. While many nonstandard workers prefer the flexibility of their arrangements, they generally receive lower pay than their full-time counterparts. Nonstandard workers are also far less likely than standard workers to have health insurance or pension benefits.
Involuntary part-time work declined substantially over the 1990s, from 4.3% of all employees in 1989 to 2.6% in 1999. The share of workers holding two or more jobs also fell in the 1990s, from 6.2% in 1989 to 5.9% in 1999.
Household wealth mostly untouched by stock boom
Wealth is the total value of a household’s assets minus its debts. The stock market boom of the 1990s left the impression that most Americans were experiencing an unprecedented growth in wealth. The truth, however, is that most Americans have no economically meaningful stake in the stock market. The most recent government data show that less than half of households hold stock in any form, including mutual funds and 401(k)-style pension plans. The same data reveal that 64% of households have stock holdings worth $5,000 or less.
The distribution of wealth remains highly unequal. The wealthiest 1% of households control about 38% of national wealth, while the bottom 80% control only 17%. The ownership of stocks is particularly unequal. The top 1% of stock owners hold almost half (47.7%) of all stocks, by value, while the bottom 80% own just 4.1% of total stock holdings.
The total wealth of the typical American household rose only marginally during the 1990s. The net worth of the average household i
n the middle 20% of the wealth distribution rose about $2,200 in the 1990s—from $58,800 in 1989 to $61,000 in 1998. Over that same period, the value of the stock holdings of the typical household grew by $5,500 and the value of non-stock assets grew by $8,500. Meanwhile, typical household debt increased $11,800. The relatively modest gains in stock and non-stock assets, combined with the explosion in household debt, meant that the 1990s were far less generous to typical households than business-page headlines often suggest.
For the typical household, rising debt, not a rising stock market, was the big story of the 1990s. While households in the middle 20% of the wealth distribution captured 2.8% of the total growth in stock market holdings between 1989 and 1998, these same families were saddled with 38.8% of the unprecedented rise in household debt. While low nominal interest rates have made it easier for households to carry this greatly expanded debt, many households appear to be straining. About 14% of middle-income households have debt-service obligations that exceed 40% of their income; 9% have at least one bill that is more than 60 days past due. Meanwhile, despite the robust state of the economy, personal bankruptcy rates reached record highs in the late 1990s.
Poverty remains high despite recovery
Despite the economic boom in the second half of the 1990s, the national poverty rate in 1998 was 12.7%, just one-tenth of a percentage point less than in 1989 and a full percentage point higher than in 1979.
Child poverty has remained stubbornly high. In 1998, 18.9% of American children—almost one in five—lived in poverty. The 1998 child poverty rate was an improvement over the 19.6% rate for 1989, but still well above the 16.4% rate for 1979.
In 1998, more than one in four African Americans (26.1%) and Hispanics (25.6%) lived in poverty, a rate about two-and-a-half times greater than that for whites (10.5%). Racial disparities in poverty rates were even worse for children. In 1998, more than one in three African American (36.7%) and Hispanic (34.4%) children were growing up in poverty, compared to 15.1% of white children.
During the 1990s, however, overall poverty rates declined more for blacks (down 3.2 percentage points) and Hispanics (down 4.7 percentage points) than they did for whites (down 0.7 percentage points). The rate of child poverty among blacks fell dramatically in the 1990s (down 7.0 percentage points). Child poverty also declined among Hispanics (down 1.8 percentage points) but rose slightly among white families (up 0.3 percentage points).
The official poverty measure almost certainly understates the true level of poverty. One alternative estimation procedure implemented by the Census Bureau suggests that, during the 1990s, the true share of the population living in poverty was an average of 3.6 percentage points higher than suggested by the official estimate.
The poor, like all workers, are working harder than ever. In 1998, for example, poor families with children worked an average of 1,213 hours per year, about 140 hours (almost three full-time weeks) more than in 1989. The increase was greater for poor minority families than for poor white families. Between 1989 and 1998, the average annual hours worked rose by 152 for poor African American families and by 175 for poor Hispanic families; over the same period, poor white families worked an average of only 3 hours longer.
Improvements in the national economy on the one hand and reductions in the generosity of public assistance on the other have combined to increase the share of poor families’ income earned through work and to decrease the share received through public assistance. Between 1989 and 1998, for example, among poor working families with children, the share of earned income rose from 58% to 71%, while the share of public assistance income fell from 25% to 11%.
South and Midwest make big gains
Trends in wages, incomes, poverty, employment, unemployment, and other economic measures vary widely across the country’s four major regions (Northeast, South, Midwest, and West) and the 50 states. In general terms, during the 1990s the South and Midwest were the biggest beneficiaries of new economic developments. By several important measures, economic conditions actually worsened in the Northeast and West.
Between 1989 and the end of the 1990s, for example, median hourly wages and median family income grew fastest in the South and Midwest. Meanwhile, median hourly wages fell in both the Northeast and the West, while median family income declined in the Northeast and remained unchanged in the Midwest. Over the same period, unemployment rates and poverty rates fell most in the South and Midwest, but poverty rates rose substantially in the Northeast.
Between 1989 and 1999, growth in hourly wages also varied greatly across regions. For low-wage workers (those in the 20th percentile of each region’s wage distribution), inflation-adjusted hourly wages rose 13.1% in the Midwest and 7.5% in the South. Over the same period, hourly wages fell 2.3% in the Northeast and 0.2% in the West. The pattern of wage growth was similar for workers in the middle of the wage distribution. Median hourly wages rose 4.9% in the Midwest and 4.7% in the South, but fell 2.4% in the Northeast and 2.7% in the West.
Regional differences in wage growth carried over into regional differences in income growth. Over the 1990s, median family income rose 1.0% annually in the Midwest and 0.7% annually in the South. But it did not grow at all in the West and fell at an average rate of 0.3% per year in the Northeast.
In 1998, poverty rates were highest in the West (14.0%) and South (13.7%) and lowest in the Northeast (12.3%) and Midwest (10.3%). But again, during the 1990s the biggest improvements occurred in the South and Midwest. Between 1989 and 1998, the poverty rate fell 1.6 percentage points in the Midwest and 1.7 percentage points in the South, while it rose 1.5 percentage points in the West and 2.3 percentage points in the Northeast.
In 1999, when the national unemployment rate stood at 4.2%, unemployment rates differed significantly from state to state—from 2.5% in Minnesota, Nebraska, and South Dakota to over 6.0% in New Mexico, West Virginia, and the District of Columbia. As with many other trends in the 1990s, the jobless situation improved most in the Midwest, where the unemployment rate fell 1.9 percentage points between 1989 and 1999, and the South, where it fell 1.7 percentage points. Improvements were much smaller in the Northeast (down 0.2 percentage points) and the Midwest (down 0.4 percentage points).
Over the 1990s, employment growth varied from region to region. Between 1989 and 1999, employment grew rapidly in the South (25.7%), West (24.4%), and Midwest (18.3%), but was almost stagnant in the Northeast (4.1%).
For the world, a poor role model
The United States is often offered as an economic model for the rest of the world, but the analysis presented here argues for caution. While the United States has succeeded in lowering unemployment and has been particularly successful in incorporating women into the labor force, it lags behind the rest of the developed world in many other important measures of economic performance. Several other rich countries have managed to achieve low unemployment rates and to incorporate women into the labor market without the high—and rising—level of inequality that has characterized the United States over the last two decades and longer.
In 1998, the United States had one of the highest per capita incomes of all the rich industrialized economies. Using market exchange rates, only Norway, Switzerland, Denmark, Finland, and Sweden had a higher per capita annual income than the United States ($32,051). Using purchasing-power-parity exchange rates, which take into consideration differences in the relative prices in differ
ent countries, the United States actually had the highest per capita income in 1998.
Despite strong economic growth in the second half of the 1990s, growth in per capita income in the United States landed in the middle of the range for the rich industrialized economies. Between 1989 and 1998, per capita income in the United States grew at a 1.6% annual rate, above the rates in Germany, Japan, and France but below those in Norway, Denmark, the Netherlands, Australia, and several other advanced economies.
The United States has led the world in productivity (the amount of goods and services produced in an hour of work) for most of the postwar period. During the 1990s, however, several countries (Belgium, western Germany, France, the Netherlands, and Norway) reached and even exceeded U.S. productivity levels.
The United States has the most unequal income distribution and one of the highest poverty rates among all the advanced economies in the world. The U.S. tax and benefit system is also one of the least effective in reducing poverty (though the U.S. Social Security system compares favorably with other countries when it comes to reducing poverty among the elderly).
Contrary to widely held perceptions, the United States offers less economic mobility than other rich countries. In one study, for example, low-wage workers in the United States were more likely to remain in the low-wage labor market five years longer than workers in Germany, France, Italy, the United Kingdom, Denmark, Finland, and Sweden (all the other countries studied in this analysis). In another study, poor households in the United States were less likely to leave poverty from one year to the next than were poor households in Canada, Germany, the Netherlands, Sweden, and the United Kingdom (all the countries studied in this second analysis).
The United States had a low unemployment rate (4.2%) in 1999. But so did Austria (3.7%), Denmark (5.2%), Japan (4.7%), the Netherlands (3.3%), Norway (3.2%), Portugal (4.5%), and Switzerland (3.5%).
The international data also show that workers in the United States put in more hours per year (an average of 1,966 in 1998) than do workers in every other advanced economy except Portugal. The average hours worked in other large industrialized economies—even Japan (at 1,898), the previous world leader in hours worked—are typically well below those of the United States.