The State of Working America 1998-99

This is an earlier edition of The State of Working America. Visit for more information on the latest edition.

The State of Working America 1998-99

EPI’s flagship book The State of Working America 1998-99 was released by Cornell University Press on January 1, 1999.

Purchase this publication.

Purchase this publication in paperback.

Read the Press Release

Read the Table of Contents

Read the Executive Summary

Read the Introduction


New Report on state of working America finds little evidence of “New Economy” in 1990s as wage, income, inequality trends of 1980s continue. Despite 2.6% annual gain since 1996, median wages still trail 1989 peak; typical family working six weeks more to keep pace; earnings for new college grads down

Washington, D.C . – Seven years into the recovery, the economy has recently begun to produce broad-based wage gains, as real wages have risen 2.6 percent annually for the typical American worker since 1996. Tight labor markets, low inflation and a higher minimum wage helped to spur this recent growth, but wages of many workers still have not been restored to 1989 levels. Amidst positive overall growth, significant economic disparities persist as trends in wages, income and inequality in the 1990s continue to follow patterns set in the 1980s, according to a new study released this Labor Day by the Economic Policy Institute (EPI).

The State of Working America 1998-99, by economists Lawrence Mishel, Jared Bernstein and John Schmitt, provides a comprehensive study of the changing living standards of working Americans. The 414-page book presents new data on family incomes, taxes, wages, jobs, unemployment, wealth and poverty, as well as state-by-state, regional and international comparisons of key indicators.

Putting recent economic gains in historical context, the study finds that the living standards of most working families still have not recovered from the recession of the early 1990s, nor have their wages kept pace with the growth in productivity. The income growth that has been generated among middle-income families has been driven largely by an increase of working hours – an additional six weeks annually for the typical family since 1989 – to make up for the long-term deterioration of wages. The economic realities facing the typical American family over the 1990s include, increased hours of work, stagnant or falling income, and less secure jobs offering fewer benefits.

New groups of workers have experienced wage declines in the 1990s, including recent college graduates and many information-technology and other white-collar workers. Women workers in the middle and upper-middle part of the wage distribution, whose real wages rose significantly in the 1980s, have experienced a sharp deceleration in the 1990s.

The report’s key findings include:

  • The inflation-adjusted earnings of the median worker in 1997 were 3.1 percent lower than in 1989. Over the same period, real hourly wages stagnated or fell for the bottom 60 percent of workers, except for low-wage workers, whose wages rose 1.4 percent during that time.
  • Median family income was $1,000 (2.3 percent) less in 1996 (the most recent year for which data are available) than in 1989. By this point in every prior recovery, the income of the typical family had surpassed its previous peak. However, data for 1997 (available in October) likely will show that the median family has finally returned to its 1989 income level.
  • The typical married-couple family worked 247 more hours (over six weeks) per year in 1996 than in 1989, despite an 8 percent growth in the economy’s productive capacity over the same period.
  • Income inequality has continued to grow rapidly in the 1990s, but at a somewhat slower rate than in the 1980s.
  • Jobs have grown more insecure in the 1990s as the share of workers in “long-term jobs” (those lasting at least 10 years) fell from 41 percent in 1979 to 35.4 percent in 1996, with the worst deterioration having taken place since the late 1980s.
  • The typical middle-class family had nearly 3 percent less wealth in 1997 than in 1989, despite the stock market boom. This is because the richest 10 percent of households in the U.S. have reaped 85.8 percent of growth in the stock market since 1989.
  • CEO pay continues to skyrocket, having more than doubled between 1989 and 1997, and rising to 116 times the pay of the average worker – an almost eightfold increase since 1965. The average CEO’s salary, bonus and returns from stock plans grew 100 percent between 1989 and 1997.
  • The record profitability of companies in the 1990s has come partly at the expense of their workers. Had profitably grown at historically normal levels, hourly compensation could have been 7 percent higher in 1997 than it actually was.

“The recent gain in wages is a welcome reversal of long-term wage decline, but most working families are still playing catch-up,” says economist Jared Bernstein. “As economic growth slows, they are likely to fall further behind,” he adds.

“The wage erosion that many white-collar workers and young college graduates have experienced over the 1990s suggests that the ‘new economy’ has still not delivered for knowledge workers,” says Lawrence Mishel.

“The stock market boom has not rescued working families,” says John Schmitt. “Most Americans own no stock in any form and those who do, typically own very little.”

Published biennially, The State of Working America has become over the past decade a respected source for the latest available data on changes in the economic well being of working Americans, as well as an overview of economic trends since World War II. The latest edition contains substantial original research, including new analyses of data collected by the Bureau of Labor Statistics and Census Bureau, as well as other government and academic sources.

The following is a summary of selected chapters characterizing the economic realities facing American workers in the final years of the 20th Century:

WAGES | Americans work longer for less

Wages and salaries make up about three-quarters of family incomes and are the primary determinant of the recent slow growth in income and accompanying rise in inequality. Recent increases have been insufficient to counteract a 20-year trend of stagnant and declining wages.

  • Wages for the bottom 80 percent of men were lower in 1997 than in 1989, with the median male worker’s real wage having fallen 6.7 percent.
  • Women’s wages rose at most level
    s of the wage distribution, with 2.7 percent growth in real wages for low-paid women contrasting a steep decline of 18.2 percent in the 1980s. But median wages for women grew just 0.8 percent in the 1990s, a considerably lower rate than the 5.7 percent growth during the previous decade.
  • Wage declines have been worst among entry-level workers. Between 1989 and 1997, real hourly wages for such positions fell 7.4 percent among men and 6.1 percent among women.
  • Young college graduates with one to five years’ experience have suffered as well, seeing their earnings fall 6.5 percent for men and 7.4 percent for women between 1989 and 1997 – erasing more than half of women’s gains in the 1980s.

Factors contributing to these wage declines include: a steep drop in the number of and bargaining power of unionized workers; erosion in the value of the minimum wage, only partially corrected by recent increases; a decline in manufacturing jobs and the corresponding expansion of lower-paying service-sector employment; globalization; and increased nonstandard work, such as temporary and part-time employment.

Family Incomes | Slow, unequal growth persists

Family income growth remains slow, and it has been slower in the 1990s than in the previous business cycles. It has taken the median family longer to regain its pre-recession income level in this recovery than in any other since World War II.

  • Younger families have been hit hard by overall slow family income growth and widening inequality. Recent groups of young families have started out at lower incomes and obtained slower income gains as they approach middle age. This has constrained income mobility, which previously had worked to offset income inequality.

Wealth | The rich get richer, the rest get poorer

A family’s ability to plan for the future and cope with financial emergencies is strongly affected by its wealth – tangible assets such as a house and car, plus financial assets like stocks and bonds. Distribution of wealth remains more concentrated at the top than distribution of income, and in fact wealth inequality has worsened in the 1990s.

  • Projections for 1997 indicate that since 1989 the share of wealth held by the top 1 percent of households grew from 37.4 percent of the national total to 39.1 percent.
  • Over the same period, the share of all wealth held by families in the middle fifth of the population fell from 4.8 percent to 4.4 percent. After adjusting for inflation, the value of this middle group’s holdings actually fell nearly 3 percent, primarily due to increased indebtedness.

The highly publicized stock-market boom has had little impact on majority of Americans because most working families own little or no stock.

  • While the share of households owning stock has risen in the 1990s, the most current available data show that in 1995, almost 60 percent of households owned no stock in any form, including mutual funds and defined-contribution pension plans.
  • In 1995, less than one third of American households had stock holdings greater than $5,000 and 90 percent of the value of all stock was in the hands of the wealthiest 10 percent of households.

Jobs | Growth down, insecurity up

The average unemployment rate during the current business cycle has been lower than during any such cycle since 1967-73, with joblessness falling to about 4.5 percent in mid-1998. But even this historic low has not fully restored workers’ sense of job security or reduced the share of workers in contingent and other nonstandard jobs.

  • Displaced workers face difficulties finding new employment, with more than one-third out of work when interviewed one to three years after their displacement. When they do find work, their new jobs pay, on average, about 13 percent less than the jobs they lost, and more than one-fourth no longer have employer-provided health insurance.
  • Work in the 1990s is of an increasingly contingent nature, with almost 30 percent of workers employed in situations that were not regular full-time jobs in 1997.

Poverty | Rates remain high despite economic expansion

The slow growth, heightened inequality of income distribution, and falling wages that mark the 1990s and ’80s have contributed to poverty rates that are both high and unresponsive to economic growth.

  • The most recent poverty rate – 13.7 percent in 1996 – is 0.9 percentage points above the 1989 rate of 12.8 percent.
  • More than one in five children (20.5 percent) were poor in 1996, up from 19.6 percent in 1989 and 16.4 percent in 1979. Childhood poverty rates are especially high for minorities, with 39.9 percent of black children and 40.3 percent of Hispanic children living in poverty in 1996.

Taxes | A further cause of worsening inequality

Average tax rates have changed little since 1979, but effective tax rates have fallen sharply for the richest taxpayers. Nonetheless, the increase in inequality and decline of living standards is largely attributable to what employers put into paychecks – not what the government takes out.

  • The wealthiest 1 percent of families has seen their tax bills fall by $36,710 since 1977 as a result of changes in tax law.

Lawrence Mishel  is EPI’s research director and co-author of each previous edition of The State of Working America. He specializes in the field of productivity, competitiveness, income distribution, labor markets, education, and industrial relations. Mishel holds a Ph.D. in economics from the University of Wisconsin, Madison.

Jared Bernstein  is a labor economist at EPI and co-author of three previous versions of The State of Working America. He specializes in the analysis of trends in employment and compensation, and tracks developments in family income inequality and poverty with an emphasis on low-wage labor markets. Bernstein served as deputy chief economist for the U.S. Department of Labor between 1995 and 1996. He holds a Ph.D. in social welfare from Columbia University.

John Schmitt  is a labor economist at EPI and co-author of The State of Working America 1996-97. He has written for general and academic publications on wage inequality, the minimum wage, unemployment and economic development. Schmitt’s Ph.D. in economics is from the London School of Economics.

The Economic Policy Institute is a nonprofit, non-partisan economic think tank based in Washington, D.C. Founded in 1986, EPI seeks to widen the debate about policies to achieve healthy economic growth, prosperity and opportunity in the United States. Institute founders include Lester Thurow, Robert Reich,
Ray Marshall, Barry Bluestone, and EPI president Jeff Faux.




FAMLY INCOME: Slower growth, grea
ter inequality

Median income grows slowly in 1980s, declines in 1990s
An economic ‘generation gap’
Income growth among racial/ethnic groups
Only dual-income, married couples gain
Growing inequality of family income
Counter-arguments to the evidence on income trends
How much has inequality really grown in the 1990s?
Inequality as measured by consumption
The impact of demographic changes on income

The ‘hollowing out’ of the middle class
Greater capital incomes, lower labor incomes
Increased work by wives cushions income fall and counteracts inequality
Falling behind the earlier generations
Income mobility


TAXES: Burden on the wealthy greatly diminished
The tax burden: still light overall
Despite progressive changes, an increase in regressivity since 1970s
The diminished progressivity of federal tax rates
What federal tax changes mean in dollars
The causes of changes in the federal tax burden
Changes in corporate taxation: the shift to untaxed profits
The shift to state and local taxes


WAGES: Long-term erosion and growing inequality
More hours and stagnant wages
Contrasting compensation and wage growth
Wages by occupation
Wage trends by wage level
The male-female wage gap
The expansion of low-wage jobs
Trends in benefit growth and inequality
Dimensions of inequality
Productivity and the compensation-productivity gap
Rising education-wage differentials
Young workers have been hurt most
Decomposing the growth in 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?


JOBS: Falling unemployment with increasing insecurity
Unemployment and underemployment
Unemployment and the earnings distribution
Job growth slows since the 1970s

Job stability and job security
Declining job stability
Job security

The contingent workforce
Nonstandard work: widespread and often substandard
Long-term growth in part-time work
Growth in temping
More than one job


WEALTH: Concentration at the top intensifies
Aggregate household wealth: financial assets boomed, tangibles failed to grow
Wealth inequality exceeds income gap
Growing wealth inequality
Who gains from the stock market boom?
Growing debt


POVERTY: Increasing inequality undermines connection between growing economy and lower poverty rates
Who are the poor?
Alternative approaches to measuring poverty
The depth of poverty
Poverty, overall growth, and inequality
The role of demographics and inequality
The changing effects of taxes and transfers
Income, wage, and employment trends among the poor
Poverty and the low-wage labor market


REGIONAL ANALYSIS: Tighter labor markets, but income growth stagnant
Median family income declines in most states
The growth of income inequality by state
Job growth, but falling median wages
Employment and unemployment
Wage trends

Poverty rates vary greatly by region and area
The regressivity of state tax liabilities


INTERNATIONAL COMPARISONS: Less-than-model behavior
Incomes and productivity: United States loses the 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 dat
APPENDIX B: Wage analysis computations
APPENDIX C:The measurement of inflation

Table notes
Figure notes
EPI Publications
About EPI
About the Authors

Executive Summary

Using a wide variety of data on family incomes, taxes, wages, unemployment, wealth, and poverty, The State of Working America 1998-99 examines the impact of the economy on the living standards of the American people. The story we tell is one of great disparities.

As this book goes to press, the economy is in an expansion, but many of the economic problems first evident in the 1980s continue to be felt. For example, despite growth between 1989, the year of the last business cycle peak, and 1996 in gross domestic product, employment, and hours worked by the typical family, median family income in 1996 was still about $1,000 lower than it was in 1989. The significant improvements in 1997 and 1998 in wages for most workers have still left wage trends in the 1990s no better than they were for most workers in the 1980s. Wage declines have also pulled down new groups of workers in the 1990s, including many white-collar workers and recent college graduates. Women workers in the middle and upper-middle part of the wage distribution, who saw real wages rise significantly in the 1980s, have experienced a sharp deceleration in the 1990s.

At the same time, jobs have become less secure and less likely to offer health and pension benefits. Middle-class wealth (the value of tangible assets such as houses and cars, plus financial assets, minus debts) has also fallen. These same factors have kept economically less-advantaged families in poverty despite an extended economic recovery.

American workers might be able to take some solace if their sacrifices now would eventually guarantee an improved standard of living for themselves or their children. Unfortunately, the country has little to show for the belt-tightening of the last two decades: productivity growth has been lackluster; only corporate profits, the stock market, and top-executive pay are doing better than in the past.

To be sure, some bright spots have appeared. The unemployment rate in mid-1998 stood at about 4.5%. Inflation had fallen to below 2% per year. Changes to the tax code in 1993 reversed some of the inequities built into the federal tax structure in the 1980s. After a decade of neglect, four increases in the minimum wage in the 1990s have boosted the earnings of millions of low-wage workers. The simultaneous expansion of the earned income tax credit has further improved the earnings of low-wage workers in the poorest families.

Nevertheless, the typical American family is probably worse off near the end of the 1990s than it was at the end of the 1980s or the end of the 1970s, despite an increase in the productive capacity of the overall economy. To the extent that the typical American family has been able to hold its ground, the most important factor has been the large increase in the hours worked by family members.
The following is a summary of the economic realities that characterize the state of working America.

Family incomes: slow a
nd unequal growth
Since 1979, the most important development regarding American incomes has been slow growth and increasing inequality. In the most recent period for which we have data, 1989-96, median family income fell by over $1,000, or 2.3%. While it is likely that the data for 1997 (available in October 1998) will show that the typical family has finally regained the after-inflation income level that it had achieved in 1989, income stagnation of this magnitude is unprecedented in the postwar era. In every other postwar expansion, the income of the typical family had, at this point, already surpassed the level reached in the preceding peak.
In any event, the restoration by 1997 of a family income level obtained in 1989 is disappointing on two further counts: first, the typical married-couple family with children worked 247 more hours (about six more full-time weeks) per year in 1996 than in 1989 and, second, the productive capacity of the economy improved about 8% over the same period. American families are working harder to stay in the same place and are seeing little of the gains in the overall economy.

Why have income trends continued to be so poor in the 1990s? Along with overall slow growth, the primary reason is the continuing wage deterioration among middle- and low-wage earners, now joined by white-collar and even some groups of college-educated workers. Despite a reversal over the last two years in these long-term trends, a longer-run view underscores continuing problems in the 1990s. Over the full 1989-97 period, for example, wages fell faster for the median worker (-0.4% per year) than they did in the 1979-89 period (-0.2%) per year.

Another key factor in understanding recent trends is the deceleration in the growth in the hours of paid work performed by members of working families. In the 1980s, many families compensated for falling hourly compensation, which was particularly steep for male workers, by working more hours. Some families increased the number of family members in the paid workforce. In other families, the number of hours worked by members already in the labor force increased. The annual hours worked by all family members in the typical married-couple family with children grew 368 hours per year (more than nine weeks of full-time work), from 3,236 hours per year in 1979 to 3,604 hours per year in 1989. After a large increase in the annual hours of paid work in 1979-89, many working families had less scope for increasing the number of hours of additional work that they could provide. The annual hours of paid work performed by the typical family grew an additional 247 hours, to 3,851 per year in 1996, a substantial extra time commitment for working families, but not enough to keep pace with the hours growth of the 1980s or enough to counteract the simultaneous squeeze on wages.

Younger families have been especially hard hit by overall slow family income growth and widening inequality. A cohort, or intergenerational, analysis of income growth shows that recent groups of young families have started out at lower incomes and obtained slower income gains as they approached middle age. One result of this trend has been to constrain income mobility, which had worked in the past to offset increasing income inequality.

Another major factor fueling growing inequality in the 1990s has been the acceleration of capital income growth, which resulted from a surge in profitability in the 1990s. This growth has generated a stock market boom that has overwhelmingly benefited the richest families. The increase in the rate of profit (the return to capital or interest and profits per dollar of plant and equipment assets) has also squeezed wage growth since 1979, but especially since 1989. Had profitability grown only at historically normal levels, then hourly compensation could have been 7% higher in 1997 than it actually was.

Some analysts have suggested that changes in the demographic composition of American families have been a major cause of the income problems documented above, implying a lesser role for economic causes such as wage decline. While the increased share of economically vulnerable families (e.g., female-headed families with children) has without a doubt put downward pressure on income growth, this process is a dynamic one that has not been constant over time. Moreover, some demographic factors, such as the increase in educational attainment, have led to increased family income. Contrary to the conventional wisdom, which has typically assigned the primary role to changes in family type, we find clear evidence that, while the shift to less well-off family types has put downward pressure on incomes, educational upgrading has more than compensated for this effect. Most importantly, on net, during the 1979-96 period, when income inequality was increasing most quickly, the demographic factors we consider (i.e., age, education, and race of the household head, along with family type) led to rising, not falling, household incomes. Thus, we should discount explanations that depend on demographic change to explain income decline.

Taxes: a further cause of worsening inequality
The effective federal tax rate for a middle-class family of four has changed little since 1980. In that year, this family paid about 23.7% of its income in federal income tax and Social Security and Medicare contributions. By 1985, the contribution had increased slightly to 24.4%, a level maintained through 1995.

While average federal tax rates for most Americans have changed little since 1979, effective tax rates have changed substantially for those with the highest incomes. Between 1977 and 1985, for example, changes in tax laws reduced the tax bill for the wealthiest 1% of families by an average of $97,250 per family relative to what these families would have paid in the absence of those changes. Meanwhile, these same changes increased the tax payments of the bottom 80% of families by an average of $221 per family relative to what they would have paid without the new tax code. Progressive tax changes in 1986 and again in 1993 partially reversed some of these inequities. On net, however, the wealthiest 1% of families have seen their tax bills fall by $36,710 since 1977, thanks to changes in the law.

The sharp reduction in the effective federal tax rates facing the richest 1% of taxpayers has contributed to the rise in income inequality since 1979. Nevertheless, since the typical family faces the same effective tax rates in the mid-1990s as in the late 1970s, changes in tax policy cannot account for the decline in living standards of the broad middle class. Most of the rise in inequality and the fall in living standards, then, reflects what employers are putting into paychecks, not what the government is taking out.

Wages: working longer for less
Since wages and salaries make up roughly three-fourths of total family income (more for the broad middle class), wage and salary trends are the primary determinant of the recent slow growth in income and the accompanying rise in income inequality. While the last two years have seen significant growth in real wages at all levels, especially at the bottom, these increases have generally not yet been sufficient to counteract the two-decade-long pattern of stagnant and declining wages. After adjusting for inflation, hourly wages stagnated or fell between 1989 and 1997 for the bottom 60% of all workers (wages over the 1990s did increase 1.4% for workers at the 10th percentile). In real terms, earnings of the median worker in 1997 were about 3.1% lower than they were in 1989.

As in the 1980s, men generally experienced more difficulties than women. Wages for the bottom 80% of men were lower in 1997 than in 1989. Median male workers’ real wage fell about 6.7% over the 1989-97 period, a rate of decline was almost as rapid as that of the 1980s. Women’s wages, however, rose between 1989 and 1997 at all levels of the wage distribution (with the exception of a
slight decline at the 20th percentile). The growth in real wages for low-paid women (up 2.7% between 1989 and 1997) stood in strong contrast to the steep declines over the 1980s (down 18.2% between 1979 and 1989). Wages for women at the middle and the top grew in the 1990s, but at rates that were far slower than those achieved in the 1980s. The share of jobs paying less than a “poverty-level wage” (i.e., less than the hourly wage that is required to keep a full-time, full-year worker’s annual income at the poverty line for a family of four) did not change between 1989 and 1997. This stagnation in the distribution of wages suggests either that job creation between 1989 and 1997 largely followed that of the 1989 wage distribution, or, if new jobs were somehow “better” than average, as some have claimed, that the wages paid on “existing” jobs deteriorated.

Including nonwage fringe benefits, such as employer health care and pension costs, does not change the overall picture. The hourly cost of benefits grew slightly faster than wages in the 1980s, but slightly slower than wages in the 1990s (primarily due to health care cost containment). Moreover, analyses of the average costs of health care benefits can conceal both the decline in the share of workers receiving employer-provided health care (down 7.6 percentage points between 1979 and 1997) and the disproportionate loss of benefits among low-wage workers (10.7 percentage points of high-school-educated workers lost health care coverage between 1979 and 1997, compared to only a 4.6 percentage-point drop among of college-educated workers).

The worst declines in wages have been for entry-level jobs. Between 1989 and 1997, for example, the average hourly wage for men with a high school degree and one to five years of work experience fell 7.4%; among comparable women, real wages fell 6.1%. Even young college graduates have suffered. Male college graduates with one to five years’ experience earned 6.5% less in 1997 than in 1989. Their female counterparts were earning 7.4% less in 1997 than in 1989 (after an 11.2% increase in the 1980s).

Meanwhile, corporate chief executive officers (CEOs) have seen their pay skyrocket. In 1965, the typical CEO made about 20 times more than the average production worker; in 1989, the ratio had almost tripled to 56; by 1997, relative CEO pay had more than doubled again to 116 times the pay of the average worker. A separate estimate of CEO pay shows that the salary, bonus, and returns from stock plans of the average CEO grew 100% between 1989 and 1997. Extraordinarily high CEO pay appears to be a uniquely American phenomenon, with U.S. CEOs earning, on average, more than twice as much as CEOs in other advanced economies.

While economists continue to grapple with explanations for falling wages and widening wage inequality, a number of factors appear to account for most of the shifts in the wage structure. These include severe drops in the 1980s and 1990s in the number (and bargaining power) of unionized workers; an erosion through the 1980s in the inflation-adjusted value of the minimum wage, which has only been partially corrected in the 1990s; the decline in higher-wage manufacturing jobs and the corresponding expansion of low-wage, service sector employment; the increasing globalization of the economy through immigration and trade; and the growth in contingent (temporary and part-time) and other nontraditional work arrangements.

Many policy makers have cited a technology-driven increase in demand for “educated” or “skilled” workers as the most important force behind wage inequality. The evidence suggests that the overall impact of technology on the wage and employment structure was no greater in the 1980s and 1990s than in the 1970s. Productivity growth, for example, was lackluster in the 1980s and 1990s, not what we would expect if technology were inducing a widespread restructuring of the economy. It is also difficult to reconcile the idea that technology is bidding up the wages of “more-skilled” and “more-educated” workers, given the stagnation since 1989 in the wages of many college graduates and white-collar workers. Technology has been and continues to be an important force in shaping the economy, but no evidence exists that a “technology shock” during the 1980s and 1990s created a demand for “skill” that could not be satisfied by the ongoing expansion of the educational attainment of the workforce.

Jobs: slow growth and greater insecurity
The good news is that the average unemployment rate since the beginning of the business cycle in 1989 has been lower than during any business cycle since 1967-73, and, by mid-1998, the unemployment rate stood at about 4.5%. While falling unemployment has undoubtedly helped boost wages in the last two years, even current low levels of unemployment have not fully restored workers’ sense of job security or reduced the share of workers in contingent or nonstandard jobs.

Data through the mid-1990s show that job stability (based on objective measures of job duration) and job security (based on more subjective measures) have been deteriorating over the last two decades. This conclusion generally reflects a decline in job stability for men and simultaneous gains (from low levels) for women. The median time that a 35-44-year-old male worker has been with his current employer, for example, fell from 7.6 years in 1963 to 6.1 years in 1996, with most of the decline (nine-tenths of a year) taking place between 1987 and 1996. The corresponding female worker saw her time with the same employer rise from 3.6 years to 4.8 years between 1963 and 1996, with most of the increase (eight-tenths of a year) taking place before 1987.

The share of workers in “long-term jobs” (those lasting at least 10 years) fell sharply between 1979 and 1996. Long-term jobs accounted for 41.0% of all jobs in 1979, but just 35.4% in 1996. Again, the worst deterioration has taken place since the end of the 1980s. The decline in long-term jobs affected men most. Just under half (49.8%) of men held long-term jobs in 1979, but this proportion had fallen 9.8 percentage points to 40.0% by 1996. Gains for women over the same period were much smaller, rising 1.2 percentage points, from 29.1% in 1979 to 30.3% in 1996.

Another measure of job stability, involuntary job loss (not for cause), was higher in the economic recovery years of 1993-95 than it was in the period 1991-93, which included the trough of the current business cycle. In 1996, 11.4% of the working-age population reported losing a job at some point in 1993-95, when the unemployment rate averaged 6.2%. In 1993, the share that reported losing a job sometime in 1991-93, when the average unemployment rate was 7.3%, was lower (10.9%).

With more than one-third of current workers with their current employers for at least 10 years, long-term jobs continue to be an important part of the economic landscape. And with only about 10% of workers losing their jobs in any given three-year period, most workers appear isolated from the threat of losing their job. Nevertheless, the sharp decline in the share of long-term jobs and the persistent high rate of job displacement despite a fall in the national unemployment rate have understandably affected workers’ perceptions of job security.

Survey data show rising feelings of job insecurity through 1996, despite economic growth and falling unemployment. In 1989, only 8.0% of workers thought that it was very or fairly likely that they would lose their jobs in the next 12 months; by 1996, the figure had risen to 11.2%, despite the almost identical unemployment rate in the two years (5.3% in 1989, 5.4% in 1996). Over the same period, the share of workers who reported that it would be “very easy” to find other jobs with the same income and benefits as their current jobs fell 7.1 percentage points, from 34.2% to 27.1%.

Data on the economic consequences of job
loss justify workers’ anxieties. Displaced workers face difficulties finding new employment (more than one-third were out of work when interviewed one to three years after their displacement). When they do find work, their new jobs pay, on average, about 13.0% less than the jobs they lost, and more than one-fourth of those who had health insurance on their old jobs don’t have it at their new ones.

Given that the unemployment rate is relatively low, we should probably look elsewhere for the source of workers’ insecurity. One of the prime suspects is the increasingly contingent nature of much of the work available in the 1990s. Almost 30% of workers in 1997 were employed in situations that were not regular full-time jobs. These “nonstandard” work arrangements ranged from independent contractors and other self-employed workers to workers employed by temporary agencies or as day laborers. While many of these workers appreciate the flexibility of their current arrangements, nonstandard workers generally earn less than workers with comparable skills and backgrounds who work in regular full-time jobs. Nonstandard workers are also far less likely than regular full-time workers to have health or pension benefits.

Wealth: the rich get richer, the rest get poorer
Stagnant and falling wages and incomes tell only part of the story of rising inequality. A family’s ability to plan for the future and to cope with financial emergencies is strongly affected by its wealth (tangible assets such as a house and car plus financial assets such as stocks and bonds).

he distribution of wealth is even more concentrated at the top than is the distribution of income, and wealth inequality has grown worse in the 1990s. Between 1989 and 1997 (projected), the share of wealth held by the top 1% of households grew from 37.4% of the national total to 39.1%. Over the same period, the share of all wealth held by families in the middle fifth of the population fell from 4.8% to 4.4%. What is even more disturbing is that, after adjusting for inflation, the value of this middle group’s wealth holdings actually fell between 1989 and 1997, due primarily to a rise in indebtedness. Between 1989 and 1995 (the latest year for which data are available), the share of households with zero or negative wealth (families with negative wealth owe more than they own) increased from 15.5% to 18.5% of all households. By 1995, almost one-third (31.3%) of black households had zero or negative wealth.

The stock market boom of the 1980s and 1990s has had little or no impact on the vast majority of Americans for the simple reason that most working families do not own much stock. While the share of households owning stock has risen in the 1990s, by 1995 almost 60% of households still owned no stock in any form, including mutual funds or defined-contribution pension plans. Moreover, many of those new to the stock market have only small investments there. In 1995, for example, fewer than one-third of all households had stock holdings greater than $5,000. In the same year, almost 90% of the value of all stock was in the hands of the best-off 10% of households. Not surprisingly, then, projections through 1997 suggest that 85.8% of the benefits of the increase in the stock market between 1989 and 1997 went to the richest 10% of households.

Poverty: rates remain high despite economic expansion
Since the mid-1980s, poverty rates in the United States have failed to respond to economic growth. The most recent poverty rate – 13.7% in 1996 – is 0.9 percentage points above the 1989 rate of 12.8%. Even with an economy that grew between 1979 and 1996, poverty rates in those 17 years were high by historical standards, averaging 13.6% for the period 1979-89 and 14.1% for the period 1989-96.

Poverty rates for minorities and children are well above the national average. More than one-quarter of blacks (28.4%) lived in poverty in 1996 (not too far below the 30.7% rate in 1989 and the 31.0% rate in 1979). More than one in five (20.5%) children were poor in 1996, up from 19.6% in 1989 and 16.4% in 1979. For minority children, poverty rates are especially high: 39.9% of black children and 40.3% of Hispanic children under 18 were poor in 1996. The poor also appear to be poorer now than at any time in the last 20 years: in 1996, the share of people in poverty whose incomes were below 50% of the poverty line was 39.5%.

Some argue that these rates are artificially high due to erroneous measurement. But a study by a nonpartisan panel of poverty experts shows that an updated measure of poverty would actually increase the number of poor by about 9 million persons (with most of the increase among the working poor). Regardless of the poverty definition used, poverty rates have been growing faster than economic conditions would predict.

The conventional wisdom typically defines the problem in terms of the supposedly counterproductive behavior of poor people themselves, implying that, with more effort, the poor could lift themselves up by their bootstraps. Recent trends in family structure and low-wage labor markets, however, contradict this analysis. The role of family structure (the shift to family types more vulnerable to poverty) has become increasingly less important since the 1970s. The role of family structure (the shift to family types more vulnerable to poverty) is typically cited as the key reason that poverty rates have been unresponsive to economic growth. While it is true that the increase in female-headed families has consistently put upward pressure on poverty rates, its impact has fallen considerably over time. And a countervailing demographic change – the rising educational attainment of heads-of-households in poor families – should have led to consistently larger declines in poverty. A full accounting of the demographic and economic forces responsible for recent poverty problems assigns a relatively minor role to family structure.

In fact, the problems analyzed throughout this book – slow growth, heightened inequality of the income distribution, and, in particular, falling wages – all conspired to keep poverty rates historically high throughout the 1980s and into the 1990s.

Variations across regions
Trends in the nation’s 50 states and various regions, in broad terms, mirror those at the national level. Nevertheless, important regional differences exist in the trends for family income, employment, wages, and poverty. These different experiences underscore another dimension of inequality in the United States, one that flows from regional disparities in wage levels and job opportunities.

Over the 1980s, the Northeast outperformed the rest of the country with respect to most important economic indicators, including median hourly wages, median family incomes, poverty, and unemployment. However, despite low unemployment, low-wage workers in some Northeastern states (New York and Pennsylvania, for example) still lost ground. States in the West, particularly California, experienced almost no growth in employment or median incomes, and wages declined for workers at the median and below.

Family income inequality increased persistently at the state level in both the 1980s and 1990s. Over the 1980s, the top-fifth/bottom-fifth ratio grew 2.4 points in New York and 2.1 points in California. Other states where income inequality grew faster than the national average in the 1980s included Indiana, Michigan, Missouri, West Virginia, Mississippi, Louisiana, and Hawaii. Inequality continued to grow in most states in the 1990s, with faster growth in both New York and California. By the end of the period, the average income of the richest fifth of New York families was 13.7 times that of the poorest families in that state. The Southwestern states of New Mexico and Arizona also saw relatively fast growth in inequality over the 19
90s; these states ended the period with levels similar to New York (13.0 in New Mexico, 13.8 in Arizona).

California and New York suffered most acutely in the 1990s recession, as these states’ incomes and employment contracted and poverty grew. The most recent data show incomes of working families in these large states to be lower than at the previous business cycle peak in 1989. Many other states, however, have clearly benefited from the recent tightening of labor markets. In 1997, unemployment was below 4% in many states (especially in the Midwest), and, thanks in part to increases in the minimum wage, the real wages of low-wage workers in these states have grown over the recovery.

International comparisons: falling behind in wages, productivity
A comparison of the recent economic performance of the United States and other advanced, industrialized countries sheds important light on the U.S. economy over the last two decades. Over the postwar period, the United States has consistently found itself in the middle or the bottom of advanced countries with respect to growth in national income per person. Even in the 1990s, when the United States has been heralded as a model “new economy,” national income per person in the United States has grown at only about the same rate as it has in France, Italy, and the United Kingdom and more slowly than it has in Japan, Germany, and the average rate for advanced economies.

A similar story holds for the most important long-run determinant of living standards: growth in labor productivity, or the production of goods and services in an average hour of work. Since at least the early 1960s, productivity growth rates in the United States have averaged only half the rate of other advanced economies. For many years, economists dismissed the more rapid growth in productivity in other countries as evidence only that it is easier for countries with lower levels of output per hour to play “catch-up” with the United States. A new development in the 1990s, however, is that at least four European economies (Belgium, France, the Netherlands, and western Germany) appear to have finally caught-up to average U.S. productivity levels. Thus, it seems that the alleged inefficiencies of more regulated economies have apparently not prevented them from narrowing – and in several cases closing – the productivity gap with the United States.

As productivity differences narrow between the United States and the rest of the advanced economies, the U.S. position at or near the top of the world income chart relies increasingly on working longer, not more efficiently. Between 1990 and 1995, a rise in the average hours worked per year in the United States and an even larger decline in the average hours worked per year in Japan have given the United States the dubious distinction of being the advanced economy with the longest work year. An important contributor to the much longer work schedule in the United States is the lack of legally mandated, employer-paid vacation time, which is typically three to five weeks per year for all workers in most European economies.

Along with slower growth, the U.S. economy has consistently produced the highest levels of economic inequality among the advanced economies. The United States had the highest overall poverty rate among 16 advanced economies in the late 1980s and 1990s. High-income families (those in the 90th percentile of family income) in the United States earn almost six times more than their low-income counterparts (those in the 10th percentile). The average ratio for other advanced economies is under four, with only the United Kingdom (with a ratio of about five) anywhere near the U.S. level. In fact, U.S. inequality is so severe that low-income families in the United States are worse off than low-income families in the 12 other advanced economies for which comparable data exist, despite the higher average income level in the United States. (The United Kingdom is the only country where low-income families are worse off than in the United States). Inequality in the United States (along with the United Kingdom) has also shown a strong tendency to rise over the last two decades, even as inequality was relatively stable or declining in most of the rest of the advanced economies.

Finally, economic mobility for those at the bottom – a factor that, in principle, could counteract the effects of inequality – is actually lower in the United States than in other wealthy economies. The United States, for example, had a lower transition rate out of poverty than France, Germany, Ireland, the Netherlands, and Sweden in the mid-1980s. (Only Canada, which had a lower overall poverty rate, had a worse transition rate than the United States.) Low-wage workers in the United States also appear to be less likely than workers in other economies to move on to higher-wage employment. Among low-wage workers in eight advanced economies in 1986, for example, U.S. low-wage workers had the lowest probability of having moved to high-wage jobs and the highest probability of being unemployed five years later.


By many important indicators, the American economy is soaring. Unemployment in early 1998 fell to its lowest point in 30 years. In a reversal of the trend of the previous two decades, real wages for most workers were finally on the rise, and productivity – a broad measure of the efficiency of the labor force – was picking up speed. What’s more, these positive developments have been accompanied by a higher level of confidence about the economy across a broad cross-section of the American public. In sum, it looks as if the economy of 1996-98 has delivered broad-based growth to most workers and their families.

But how significant are these recent changes? Do they represent a reversal of decades of stagnant family income growth and real wage losses for most workers? Or are the fundamental problems that have beset working families over the long term simply on pause, due to current low unemployment and the increased demand for work?

The evidence presented in this 1998-99 edition of The State of Working America suggests that a marked transformation in the U.S. economy has yet to occur. When we put recent gains in their historical context, it is clear that the living standards of many working families have neither fully “recovered” from the early 1990s recession nor benefited from the overall growth in productivity. Moreover, whether recent gains continue depends upon whether the factors responsible for the long-term erosion of wages continue to be offset by tight labor markets, unexpectedly low inflation, and further minimum wage increases.
This introductory essay spells out our assessment of the living standards of America’s working families, both past and present, and attempts to address the current issues in the public discussion. It begins with a short-term view that enumerates the impressive gains of the past few years. The second part, which compares the current 1990s business cycle with that of the 1980s, finds that, by many measures, workers and their families in the 1990s have yet to recover the ground they gained in the 1980s but lost in the 1990s recession. Finally, viewing the long term, we ask whether the gains of the late 1990s are evidence of a new economic order, the “reward” for two decades of economic pain in the form of falling wages and unequal growth. We find no evidence that real wage losses, the increase in economic inequality, and the heightened insecurity of working families are part of some sacrifice that has led to higher productivity, compensation, or per capita income.

The good news about recent wage trends
One of the most important and troubling economic phenomena of the last two decades
has been declining wages and stagnant family incomes amidst positive overall economic growth. Even when unemployment was beginning to fall in the middle part of this decade, the economic gains continued to elude most working families. Wages continued to decline through 1995, and family incomes were still far below the level they had reached before the 1990-91 recession.

The 1996-98 period, however, is different. Over the last year and a half, as shown in Table A, low unemployment has persisted, and wages have not only grown faster than inflation, they have grown faster at the bottom of the wage scale than at the top.

Table A Wage Growth in the 1990s

For example, male wages at the 20th and 50th percentiles fell by about 1% per year in the 1989-96 period (the 20th percentile worker earns less than 80% of the workforce; the 50th percentile, or median, worker earns less than half of the workforce). But from 1996 through the first half of 1998, real wages grew 4.1% per year for low-wage male workers and 2.0% per year for the median male worker, indicating a narrowing of the wage gap between middle- and low-wage men. Wages fell for low- and middle-wage female workers at an annual rate of 0.2% from 1989 to 1996, but reversed course and grew at annual rates of 2.7% and 2.6% respectively from 1996 to mid-1998. Wages for high-wage male and female workers also grew relatively quickly in the 1996-98 period, but, in both cases, wage growth at the 90th percentile was slower than at the 20th. This pattern of growth indicates a narrowing of the wage gap between the top and the bottom of the wage scale, an uncharacteristic pattern given inequality’s persistent increase over the past two decades.

The other panels in the table show the primary factors responsible for the recent growth spurt of real wages: falling unemployment, the increase in the real value of the minimum wage, and the decline in the growth of inflation.

The role of low unemployment. The overall unemployment rate was 4.5% in the first half of 1998, down almost a full percentage point since 1996. Moreover, this 0.9-point decline overall was accompanied by larger declines among groups of workers who are traditionally further down the hiring queue. For example, the unemployment rate for African Americans and Hispanics fell by 1.5 and 2.0 points over this period. Looking at a particularly disadvantaged group – young (18-35), minority high school graduates – reveals an even larger decline of 3.5 percentage points for young blacks. Of course, even with these large declines, unemployment rates for young minority workers are still many times higher than the overall rate (e.g., the unemployment rate for young African Americans with a high school degree was more than three times that of whites in each of the years shown). Nevertheless, these improvements provide clear evidence that persistently tight labor markets have greatly increased the employment opportunities of the least well-off.

The role of a higher minimum wage. The Congress mandated two $0.90 increases in the federal minimum wage in the 1990s, the first of which was implemented in 1990/91, the second in 1996/97. By 1996 inflation had eroded much of the value of the first increase, but the second increase, which raised the real value of the minimum wage 9% per year over the 1996-98 period, clearly helped to lift wages at the lower end of the wage scale.

The role of lower inflation. Inflation, as measured by the annualized growth rate of the consumer price index, grew by 3.4% per year in the 1989-96 period but slowed to 2.3% in 1996-98. This unexpected deceleration in inflation means that wage increases given by employers simply to offset anticipated higher inflation translated into real wage increases. Thus, some of the wage growth over the 1996-98 period is a result of inflation temporarily growing more slowly than expected. As inflationary expectations begin to conform more closely to the actual path of price growth, the contribution of lower prices to real wage growth is likely to diminish.

These recent wage and employment trends are a welcome reversal of the long-term trend toward rising inequality. The problem of inequality’s persistent growth has led numerous analysts to view low unemployment levels as well as increases in the minimum wage as desirable policies that could be implemented without adversely affecting the economy or hurting the workers they are meant to help. As the data in the table show, low unemployment has not led to runaway inflation, nor has the increase in the minimum wage hurt the job prospects of low-wage workers.

A broader look at the 1990s
The above examination of the past few years provides important insights into the short-term condition of the economy and underscores the importance of persistent low unemployment. But to accurately assess the economy, it is best to examine an entire business cycle (in this case, 1989-98) and compare it to trends over other cycles. We now have enough years of data to evaluate most of the economic landscape of the 1990s.

Unfortunately for working Americans, the 1990s have been, in many ways, an extension of the 1980s. Income and wage inequality have increased (though at a slower rate), families are working longer for less, wealth has become even more concentrated, and poverty has not fallen much in response to overall economic growth. And the 1990s have introduced some new problems with regard to living standards: an increase in job insecurity; a decline in wages for white-collar and entry-level, college-educated workers; wage stagnation for middle-wage females; and, at least through 1996, worse income growth for the typical family.

Family income
By 1997, the most recent year for which data are available as this book goes to press, the income of the median American family was only slightly ($285) higher than it was at the peak of the last business cycle peak in 1989. The initial decline was attributable to the recession that began in 1990, but the median continued to fall as the recovery got under way in 1991 and 1992. In 1994, median family income finally responded to overall growth, and it has increased each year since. The fact that it took eight years for median family income to reach its prerecession level is unprecedented in the postwar era. In every prior recovery, the income of the typical family had, by this point in time, far surpassed its level of the prior peak.

Examining the continued growth of income inequality in the 1990s has been made more difficult because of changes in the survey instrument used by the Census Bureau to track family income. However, we use a specially constructed data set (explained in Appendix A) that allows for a consistent comparison of 1989 to 1996. These data clearly reveal that slow income growth is not the only problem: family income has also become increasingly unequal during this recovery. The inflation-adjusted average income of the top 1% of families, for example, grew by 10% from 1989 to 1996, while the income of the middle fifth fell by 2.1% and that of the lowest fifth fell by 4.0%. Though this is a slower increase in income inequality than occurred over the 1980s, it nevertheless represents considerable growth of income inequality.

Working longer
The primary factor driving these income problems is the continuing wage deterioration among non-college-educated workers, joined in the 1990s by white-collar and even some groups of college-educated workers. This has meant that families have had to constantly boost their hours of work outside the home in order to keep their </

Wage erosion
The deterioration in hourly wages is a long-term trend that continued in the 1990s. Between 1989 and 1997 (our last full year of wage data), the median male wage fell 0.9% per year, and the hourly wages of the bottom 80
% of male workers were lower in 1997 than in 1989. Wages for females, which grew 0.6% per year at the median in the 1980s, were flat in the 1990s, rising just 0.1% per year from 1989 to 1997.

Adding fringe benefits to these wage data does not brighten the picture (Figure A). Since the average value of the benefit package fell relative to wages over the 1990s, the typical worker actually lost more in terms of compensation (wages and benefits) than in wages. From 1989 to 1997, compensation fell 4.2% for all workers, 7.8% for males. By contrast, wages alone fell 3.2% for all workers and 6.7% for males.

Figure A Productivity and compensation growth in the 1990s

The disappointing path of compensation growth is especially clear when compared to productivity, which grew 9% from 1989 to 1997. If economic growth were distributed as equally as in the past, then the growth in productivity would have lifted the income of the typical, middle-class family, whose 4,000 hours of work in 1996 surely made a nontrivial contribution to the rise in this important indicator. The fact that most families are simply striving to get back to the living standards they enjoyed in 1989 is a telling measure of our diminished expectations.

Much of the conventional economic analysis of the 1990s argues a college degree is a prerequisite to participating in the new technology-driven economy. If this is so, then we would expect the wages of new college graduates to be on the rise, or at least immune from the trend of long-term wage deterioration. But this is not the case. The 1990s have been a bad decade for young, college-educated workers. The hourly wages of entry-level college graduates fell about 7% for both males and females from 1989 to 1997, a sharp reversal particularly for young women graduates, whose hourly wages grew 11.2% in the 1980s. Even young college graduates in scientific and engineering occupations earned lower wages in 1997 than their counterparts earned in 1989; those in computer occupations ended the period with only modestly higher wages.

Thus, while the tight labor market of the last few years has delivered real gains to those who have fallen behind over the last two decades, these gains have been too small to reverse long-term wage and income losses. The real wage of the median worker was 4% lower in the first half of 1998 than in 1979; for the median male worker it was down 15%. On the other hand, the post-1979 period has been a good one for those at the top of the wage and income scales.

There is no “smoking gun” to account for these long-term income and wage losses. Instead, a variety of related factors have, since the late 1970s, interacted to increase inequality and reduce the wages of most workers. All of these factors share a common characteristic: they reflect the general deregulatory, laissez-faire shifts in the economy and forces that have weakened the bargaining power of workers, both union and non-union and both white and blue collar. For instance, the long-term decline in labor market institutions – the falling real value of the minimum wage along with continuing deunionization – can explain one-third of the growing wage inequality among prime-age workers. The expansion of low-wage service-sector employment has contributed perhaps another 20-30%. This shift to lower-paying industries is causally linked to another important contributor to the economic problems of working families: the increasing globalization of the economy through immigration and trade. By itself, globalization may account for 10-20% of the increase in wage inequality (with immigration playing an even larger role in the wage erosion of low-wage workers); the combined effects of globalization and the shift to lower-paying industries can conservatively account for 30% to 40% of the growth of wage inequality. Thus, the weakening of labor market institutions, the impact of globalization, and the shift to low-wage service industries can together account for two-thirds to three-fourths of the growth in wage inequality.

Job security
Another new and disturbing feature of the 1990s recovery has been the decline in job security, defined here as the sense among workers that, as long as their job performance is adequate, their employment situation will remain unchanged.
Workers appear less and less likely to be able to count on the long-term employment attachments that in the past provided opportunities for steady wage growth, fringe benefits, and long-term job security. Between the 1980s and the 1990s, for example, the share of workers who have been at their current jobs for at least 10 years has fallen. Involuntary job loss (layoffs, “downsizing,” and other job displacements not for cause) actually increased between the recession of 1992 and the recovery through 1995. These objective measures of job stability may have contributed to workers’ subjective perceptions that jobs were less secure through most of the 1990s recovery than they were in past recoveries, including those with weaker labor markets. Survey data through 1996 indeed show workers feeling less optimistic than in the past that their jobs would last and more pessimistic about their employment prospects if they lost their jobs. It is not unreasonable to conclude that job insecurity can help explain why wage growth was slow to respond to falling unemployment throughout most of the 1990s recovery.

Given that the unemployment rate is relatively low, we should probably look elsewhere for the source of workers’ insecurity. One of the prime suspects is the increasingly “contingent” nature of much of the work available in the 1990s. Almost 30% of workers in 1997 were employed in situations that were not regular full-time jobs. These “nonstandard” work arrangements ranged from independent contracting and other forms of self-employment to work in temporary agencies or as day labor. The most readily documented indicator of this trend is the near doubling of the share of workers employed by temporary help agencies, from 1.3% in 1989 to 2.4% in 1997. While many of workers in nonstandard jobs appreciate the flexibility of their current arrangements, they generally earn less than workers with comparable skills and backgrounds who work in regular full-time jobs. Nonstandard workers are also far less likely than regular full-time workers to have health or pension benefits.

The theme of a growing, even booming, economy leaving families behind is perhaps nowhere more evident than in a discussion of American poverty. The fact that poverty did not fall between 1995 and 1996 (the most recent year for which data are available) is one of the many contradictions of the 1990s recovery.

In 1997, 13.3% of the population, or 35.6 million Americans, were poor, a higher poverty rate than in 1989 (12.8%) or 1979 (11.7%). Conventional explanations that blame demographic trends for the uncoupling of the historical relationship between economic growth and falling poverty rates fall short in the 1990s, since these trends, which include not only the shift to mother-only families but also the educational attainment of family heads, worked to lower poverty over this period. Also, including the value of benefits provided to the poor (yet not typically counted in their incomes) fails to explain the disconnection. Rather, the increase in both economic inequality and the share of jobs that pay poverty-level wages have kept the poverty rate from falling as the economy has expanded.

This lack of access to jobs in the 1990s that could lift the poor out of poverty challenges recent shifts in U.S. anti-poverty policy. The welfare-to-work component of welfare reform partially reflects American values regarding the integrity of work as well as voters’ distaste for dependence on government support of low-income families. However, the wage and employment opportunities facin
g poor persons will have to expand considerably before anyone can reasonably expect the poor to work their way out of poverty.

International comparisons
The 1990s have been promoted as a stellar period for the American economy relative to that of other industrialized economies. While this is generally true for unemployment rates, other broad economic indicators, including growth in per capita income and labor productivity, fail to show that the U.S. is economically dominant among the advanced countries (defined as those in the Organization for Economic Cooperation and Development, or OECD). With regard to wage growth and inequality, measures that are more directly relevant to living standards of working families, the U.S. is clearly behind.

Per capita income has historically been higher in the U.S. than in other OECD countries. But in the 1990s, average annual growth in this broad measure of prosperity was just 1.0% in the U.S, compared to 1.3% among the other OECD countries. A similar comparison for productivity shows the U.S. growing more slowly than the OECD average in both the 1980s and 1990s, so that by 1995 the U.S. no longer led the world in productivity levels; by that year, West Germany, France, the Netherlands, and Belgium all had productivity levels slightly above or comparable to those of the U.S.

Enriching the rich: surging profitability and CEO pay
During the 1990s, in the midst of slow income growth, widespread wage erosion, heightened job insecurity, and stubbornly high poverty, corporate profitability and compensation of top executives has surged. And while this surge increased the wealth of the richest households, that of middle-income households was no greater in 1997 than in 1989.

By 1997, the net worth (assets minus debts) of the top 1% was about $10 million, up 11.3% from 1989 (Table C). Over the same period, the net worth of middle-class families (those in the middle fifth of the wealth distribution) fell by 2.9%. This pattern of wealth accumulation, facilitated in large part by the stock market boom, enabled the top 1% of households to control 39.1% of total net wealth by 1997, an increase of 1.8 percentage points over 1989 and 5.3 points over 1983.

Table 3 Wealth, profitability, and CEO pay trends, 1979-97

Note that the stock market boom of the 1990s has not enriched the middle-class household. While the share of households owning stock has risen in the 1990s, by 1995 (the most recent year for which these data are available) almost 60% of households still owned no stock in any form, including mutual funds and pensions, and fewer than one-third of all households had stock holdings greater than $5,000. In the same year, almost 90% of the value of all stock was in the hands of the best-off 10% of households. Not surprisingly, then, estimates through 1997 suggest that 85.8% of the benefits of the increase in the stock market between 1989 and 1997 went to the richest 10% of households.

This increase in wealth concentration stems in part from an economic phenomenon that was a much larger factor in the 1990s than in the 1970s or 1980s: the increase in the “return on capital,” or corporate profitability. As shown in Table C, the pre-tax profit rate (the profits and interest income earned from plant and equipment assets) grew 3.3 percentage points from 1989 to 1997, to stand at a 30-year high. As a result, the share of income derived from profits and interest (capital income) has also reached historic highs, climbing 3.2 percentage points in the 1990s. Had the growth in profitability been more modest (achieving, for example, the long-term average of the 1960-80s), average labor compensation in 1997 could have been 7.7% higher. Reflecting this increased profitability and the stock market boom, Table C also shows that compensation of chief executive officers (CEOs) doubled from 1989 to 1997 and grew 71% over the 1992-97 recovery.

Whether there has been a payoff to the economy, or to a typical working family, from downsizing, restructuring, deunionization, and globalization is uncertain. What is clear, however, is that CEO compensation, business profitability, and the income and wealth of the top 1% owe much to the economy of the 1990s.

Summing up the 1990s
From the perspective of working families, the economy of the 1990s is woefully similar to that of the 1980s:

  • Income growth remains slow, and it has been slower in the 1990s than in previous business cycles; it has taken the median family longer to regain its pre-recession income level in this recovery than in any other since World War II;
  • Income inequality has continued to grow in the 1990s, although it has done so more slowly (at about two-thirds the rate) than in the 1980s;
  • Despite the 1996-98 spurt in wages, wage growth in the 1990s has been no better than in the 1980s; due to the slowdown in the growth of benefits over the 1990s, median hourly compensation, which was flat over the 1980s, declined in the 1990s.
  • Many workers who were able to avoid wage losses in the 1980s succumbed to the long-term trend in the 1990s. Among them were new college graduates, including those in technologically advanced fields (such as engineering) and male white-collar workers.
  • The wage premium enjoyed by college-educated workers grew strongly in the 1980s but grew little in the 1990s for men and fell for women. This development challenges the notion that the 1990s was a period in which technological advances led to accelerated demand for highly educated workers.
  • Jobs grew more insecure in the 1990s, as downsizing diminished the job stability of white-collar workers.
Is there a ‘new economy’?
The economy over the 1979-89 and 1989-97 business cycles brought only modestly higher incomes for middle-income families and lower incomes for those at the bottom of the income scale. Whatever income growth has been generated among middle-income families, particularly married-couple families, has in large part been driven by more family members working and working more hours each year. Thus, American families still face the consequences of a long-term erosion in wages, deteriorating job quality, and greater economic insecurity. To some analysts, however, these are the unfortunate but unavoidable costs associated with a transition to a “new economy,” whose promise is expanding living standards for all. In some analyses, the wage and productivity growth of 1996 and 1997 are cited as evidence of a successful transition to a new economy.

The economy is always changing, renewing, and reformulating itself. New products and new ways of producing goods and services are always being developed. Thus, in many routine ways we have a “new economy” every year. The more profound question is whether we are making or have made a transition to a permanently better economy, i.e., a more efficient economy leading to increased living standards for working families.

In the post-1979 period, economic policy has moved decisively toward creating a more laissez-faire, deregulated economy. Industries such as transportation (trucking, intercity buses, railroads, airlines) and communications have been deregulated. Management has actively pursued the weakening of union protections, the right to organize, and the right to collectively bargain. Social protections, such as safety, health, and environmental regulations, the minimum wage, government cash assistance (e.g., welfare), and the unemployment insurance system, have been weakened. Increased globalization, including greater international capital mobility and international trade, has given greater scope to managerial discretion. Taxes on capital and the average and marginal tax rates for high-income families and business have been reduced. We have pursued the anti-inf
lationary policies preferred by investors, Wall Street, and the bond market. In sum, there has been a conscious, decided shift of national policy designed to unleash market forces and empower management decision makers.

The promise of all of these policies was to raise living standards and to generate more overall income growth. As with all policies and economic transformations, there were expected to be, and there have been, losers, as the large redistribution of income, wealth, and wages since 1979 attests. The question is, was there an overall improvement in the economy that would justify all of the social costs? In economists’ terms, did the benefits outweigh the costs? Or simply, was the gain worth the pain? Is there reason to believe we are making a transition to a better economy?

The issue of whether we have a “new” and “better” economy requires that we specify when the “new economy” originated. Some analysts have identified 1992 as the starting point, while others focus on 1994 or 1996. But regardless of the starting point, it is inherently difficult, given the cyclicality of the economy, to distinguish whether positive trends in any recovery year are due to the cyclical recovery or to the onset of a “new economy.” Therefore, we analyze this issue in three ways. First, we examine whether the economy’s performance in this recovery and business cycle was better than in earlier ones. The merit of such an approach is that it employs an “apples to apples” comparison that looks beyond the data for a few, selected years. The second part of the analysis addresses whether the solid wage and productivity performance in 1996 and 1997 are indicative of our having ascended to a new information technology era. The third part examines how “knowledge workers,” presumably the cutting edge and clear beneficiaries of the new economy, have fared in recent years.

Business cycle comparisons
In comparing one business cycle to another, three indicators are key: the growth of productivity (output per worker), growth in real hourly compensation, and growth in real income per capita. The first, productivity growth (the growth in private nonfarm output per hour worked), allows us to judge the rate at which the private sector is becoming more efficient. Central to the “new economy” idea is that growing inequalities and insecurity have been necessary (or the price paid) for generating a significant acceleration in productivity growth. But productivity growth in the current business cycle (1990-98) has been no better than that of the business cycles of the 1970s and 1980s – between 1.0% and 1.2% annually (Table D) – and it pales in comparison to the growth over the five business cycles from 1948 to 1973, when productivity grew from 1.9% to 3.3% annually.

Table D Growth in nonfarm business productivity, hourly compensation, and income per capita over business cycles, 1948-98

Recent productivity growth is similarly unspectacular in the shorter period that includes only the recovery. Regardless of the time period used to date the current recovery (starting in early 1991, as per the National Bureau of Economic Research, or in mid-1992, when unemployment started falling), its productivity performance has been no better than the recoveries of the 1970s and 1980s and significantly below those of the earlier postwar period. So, in terms of the key indicator of efficiency – productivity – there is nothing new in the new economy.

Some have argued that productivity is mismeasured or understated (particularly in services) and that the payoff will come as we learn to exploit microelectronic/computer technologies. Productivity may or may not be mismeasured, but the only relevant issue here is whether there has been a greater understatement of productivity growth in the 1990s than in the 1980s, 1970s, or earlier periods (otherwise, our comparison of time periods would remain correct). No analysis has shown a growth in mismeasurement. For instance, any errors in measuring productivity in particular service sector industries or in the service sector as a whole have been present for decades, and the expansion of the most difficult-to-measure services, such as finances and health care, has not been sufficient to substantially affect overall trends. The payoff may come eventually, but when? There have been claims ever since the 1950s that computerization will soon transform productivity, and we are still waiting.

The second indicator in Table D is the growth of average real hourly compensation (which includes payroll taxes, pensions, health insurance, wages, and salaries for all nonfarm business sector employment). This measure captures the compensation growth of all workers, from executives to day laborers, and therefore does not reflect trends experienced by “typical workers.” Average hourly compensation growth does, however, provide a measure for how fast the compensation pie is growing. Far from posting a better performance, the current business cycle has seen slower compensation growth (0.6%) than during the stagflationary 1970s (1.1%) and the earlier postwar periods (when it ranged from 2.0% to 3.3%). Compensation growth in the current recovery also does not suggest a better performance for the 1990s.

The final measure, income per capita (gross domestic product per person) reflects growth in all types of income and growth in the percent of the population employed. As with compensation growth, per capita income is an average across the population and therefore does not take into account changes in distribution – in other words, it does not reflect trends for a typical family. By this measure, income growth in the current cycle has been no better than that of the 1970s and 1980s and far slower than that of the 1948-73 period. The 1.8% or 1.9% per capita income growth in the current recovery, however, is decidedly inferior to the recoveries of the 1970s or 1980s and just half that of the earlier postwar recoveries with the exception of the mid-1950s.

Thus, if we are in a “new economy” it is not necessarily a better economy in terms of productivity, wage, or income growth.

The last two years
Although productivity and other measures of economic performance have not shown strong or even better-than-average growth over this business cycle, productivity and compensation growth in the last two years has been strong. For instance, productivity grew 1.9% and 1.7%, respectively, in 1996 and 1997. Can or should these trends be interpreted as evidence of the economy ascending to a new, higher rate of productivity growth?

Caution is called for in interpreting short-run gains as evidence of a new economic regime. Simply put, two good years no more proves the existence of a new economy than does the dismal performance of the three prior years (productivity growth of 0.1%, 0.4%, and 0.2%) prove a permanent, lower productivity growth regime. The convention of examining complete recoveries and business cycles remains the most enlightening type of analysis. But the question of why productivity picked up so much in 1996 and 1997 is still an important one. It seems unlikely that the long-awaited gains from the computer revolution, the tax cuts of the 1980s, deregulation, or globalization suddenly took hold in 1996. Rather, a more plausible explanation may be that persistent low unemployment and a spurt in demand pushed productivity up. Business did not expect nor plan for an economy with unemployment below 5.5%, so as demand grew faster than expected there was greater-than-expected capacity utilization (primarily in services) and resulting productivity gains. That is, businesses were faced with the opportunity to sell more goods and services than expected and managed to produce more even though they could not add workers as quickly. This conventional macroeconomics sto
ry fits the facts more reasonably, in our view, than does a sudden ascent to a new economy.

‘Knowledge’ workers
One prominent formulation of the new economy story holds that we are enjoying an information-technology-fueled ascent to a new era. However, this story does not correspond to trends in the labor market. For instance, a transition to a technology economy would presumably be associated with an increased need for highly skilled and educated workers. Yet, our analysis of wage and employment trends for the 1990s up through 1997 shows that this decade has not been a good one for white-collar and college-educated workers compared to the 1980s. For instance, the decades-long shift into white-collar occupations actually ground to a halt in the early 1990s, and it has been slower over the 1989-97 period than in prior decades. Moreover, projections to 2005 show a further slowdown in the shift to white-collar employment. Corresponding to this slow growth are historically high rates of white-collar job displacements from downsizing and greater job instability and insecurity among college-educated men.

Perhaps most disconcerting to a new economy interpretation is that wage trends for white-collar and college-educated workers have not been especially favorable in the 1990s. This is especially true for men over the 1989-97 period: wages for nearly every white-collar occupation group were stagnant or fell; health insurance coverage did not expand; wages among the college educated rose just 1.2%; and the college/high-school wage premium has been flat over the 1992-97 recovery. Remarkably, the entry-level wages earned by new college graduates, male or female, were 7% less in 1997 than in 1989.

Even so-called information technology workers have not done all that well. For example, newly hired engineers and scientists are earning 11% and 8% less in 1997 than their counterparts did in 1989, despite good wage growth over the 1996-97 period. Young college graduates in computer science and mathematics occupations were earning only 5% more in 1997 than in 1989, with all of the gains occurring in 1997 following seven years of wage stagnation.

These trends do not easily fit with a story in which information technology is transforming the workplace, allowing those equipped to participate to enjoy prosperity while those lacking skills lag behind. Rather, it seems that white-collar workers’ experiences in the 1990s – wage losses, displacement, and job instability – mirrors the unpleasant experiences of blue-collar workers in the 1980s. This phenomenon might be described as the “blue-collarization” of white-collar worklife in the 1990s. How can a new information-age economy be expected to lift all of our wages when it cannot even do so for white-collar workers and young college graduates working in technical occupations, presumably the best-educated, most computer literate, and most flexible segment of the workforce?

In sum, looking at the current recovery compared to the past, the dramatic changes of the last two years, and the labor market status of key “knowledge” workers, there seems to be little if any evidence that changes in the economy have led to greater long-term capacity to raise incomes or produce goods and services. Thus, the factors causing the pain of greater dislocation, economic vulnerability, and the long-term erosion of wages have not made the economy into a “better” economy, nor has the large-scale redistribution of income and wealth documented throughout this book been associated with improved economic efficiency, compensation, or income growth. For the vast majority, the slogan for the last two decades might be “all pain, no gain.”


As the 1990s recovery proceeds, unemployment has fallen sharply and real wage growth has become strong and broad based. These are very favorable developments, coming on the heels of long-term wage losses and stagnating living standards for most working families.

And yet, despite consistent expansion of the economy since 1991, the wage of the median worker remains below its prerecession level, and the income of the median family took longer to recover in this business cycle than in any prior postwar recovery. Income and wealth have become more concentrated, and jobs have become more insecure. Even groups whose wages grew in the 1980s, such as middle-wage women, white-collar workers, and young college graduates, have faced real wage losses in the 1990s.

This critical viewpoint of the 1990s economy has little in common with the conventional wisdom of much contemporary economic analysis, most of which proclaims that we are enjoying the “best economy in 30 years” (or “in 50 years,” or “ever”). But it is unclear in such proclamations what the yardstick is for judging the economy, and for whom the economy is considered “best.” With real median family income and median compensation still below their pre-recession levels, it cannot be said that this is best economy for working families. And while declining unemployment, accelerating productivity, and rising real wages are evidence of an economy getting better, it is difficult, in light of the longer-term negative trends examined in this book, to call it the “best.”

Economic discourse as well as economic policy needs to be explicit about who benefits from economic growth, and which yardstick is used to judge improvements. America’s conversation about the economy, and, more pointedly, about the living standards of working families, cannot be limited to short-term analysis with a sole focus on the recent gains in unemployment, productivity, and wages. Nor can the conversation convincingly reference a “new economic order” when the evidence is so starkly lacking. Instead, the economic discourse must recognize and address the scope of wage and income problems that over the last two decades have diminished the quality of life of American families.