Skip to page content
Office of the Secretary
Bookmark and Share

U.S. Department of Labor Futurework
  Trends and Challenges for Work in the 21st Century
---division bar --

Opinions and views in these papers are those expressed by the author(s). They are not to be taken as expressions of support for particular positions by the Department of Labor. Please do not cite these papers without prior permission of the author(s).

AN OVERVIEW OF ECONOMIC, SOCIAL,
AND DEMOGRAPHIC TRENDS
AFFECTING THE US LABOR MARKET

Robert I. Lerman
Stefanie R. Schmidt

The Urban Institute
Washington, D.C.

Final Report
August 1999

This report was prepared at the Urban Institute for US Department of Labor, Office of the Assistant Secretary for Policy, under DOL Contract No. J-9-M-0048, #23. The views expressed are those of the authors and should not be attributed to the Department of Labor, the Urban Institute, its trustees, or its funders.

Abstract

This paper offers an assessment of broad social, economic, and demographic trends affecting the US labor force now—in this time of strong economic growth—and in the future. The focus is on demographic trends, work and family issues, health and pension patterns, technical change, adjustment to low unemployment, globalization, and the plight of low-skilled workers. The paper identifies several important trends and patterns, including: 1) the largest demographic shift relevant to the job market is the impending decline in the share of prime-age workers; 2) over 60 percent of workers do not have their own children in their home, but an increasing share of workers care for elderly relatives; 3) the impact of the substantial shift from defined-benefit (DB) to defined-contribution (DC) pension plans on workers is unclear, but some estimates suggest that the typical worker will gain financially; 4) while investment in computers is spurring technical change, the impacts on productivity in firms vary a great deal because of the varying organizational responses to technology; 5) the labor market has adjusted surprisingly well to low unemployment, partly because college-educated workers have accounted for over 90 percent of the net growth in employed adult workers during the 1992-99 expansion; 6) globalization of production is unlikely to have weakened the position of US workers because overall foreign investment in the US has exceeded US investment abroad and foreign direct investment has been nearly as high as US direct investment; 7) while the economic expansion greatly reduced unemployment and expanded job opportunities for low-skill workers, many less-educated men who left the labor force in earlier years have not reentered the job market.

Brief Description

This paper offers an assessment of broad social, economic, and demographic trends affecting the US labor force now—in this time of strong economic growth—and in the future. The focus is on demographic trends, work and family issues, health and pension patterns, technical change, adjustment to low unemployment, globalization, and the plight of low-skilled workers.

Keywords:Labor Market, Jobs, Economic, Social, Demographic

Table of Contents

Introduction

I. Demographic Change and the Future Workforce

II. Trends in Work and Family, Health Insurance, Pensions

III. Trends in Employer-Provided Health and Pension Benefits and Families

IV. Technology and Work Organization

V. Adapting to Tight Labor Markets

VI. Globalization

VII. The Low-Skilled Labor Market

Tables

Table 1: The Changing Mix of the US Labor Force by Age, Ethnicity, and Sex: 1976-2006

Table 2: Participation in Job-Related Education and Training by Age Group in the US and Other Selected Countries

Table 3: Percent of Women in the Labor Force in Various Types of Families: Second Quarter of 1998

Table 4: Percent of men in the Labor Force in Various Types of Families: Second Quarter of 1998

Table 5: Annual Work Hours of Husbands and Wives with at Least One Child

Table 6: Gains in Employment-Population Ratios and Unemployment Rate Reductions by Age, Ethnicity, and Education: 1992-1998

Table 7: Distribution of Net Employment Growth of Population, Ages 25 and Over, by Educational Status: 1992-1998, 1st Half

Table 8: Relationship Between Changes in Wage Rates and Weekly Earnings and State Labor market Conditions: 1995:I to 1998:I

Figures

Figure 1: Relationship Between Unemployment Rate, Consumer Price Index (CPI), and Employment Cost Index (ECI): 1980-1998

Figure 2: Trends in Unemployment Rates and Labor Force Participation Rates: 1970-1998

Figure 3: Distribution of Unemployment Rates by State: 1998:I

Figure 4: Average of Exports Plus Imports as a Share of Gross Domestic Product: 1959-1998

The recent performance of the US job market has proved surprisingly strong Unemployment rates are at a 30 year low and far below what most macroeconomists predicted could be reached without substantial increases in inflation. Job growth has been strong. Employers have expanded their recruitment to reach large numbers of youth, low-skilled workers, mothers heading families, and other groups generally not favored in the labor market. Even wages, which had been rising only slowly, have been increasing more rapidly.

In the context of today’s good times, it makes sense to step back and assess the broader trends affecting the job market of today and the future. One rationale is to put in place policies that can help sustain the economic expansion without spurring a new round of inflation. A second rationale is to improve our understanding of the interactions between the job market, emerging social, economic, and demographic trends, and public policies. What forces will policymakers have to confront in the future? Is wage inequality likely to increase in the future? How can we best meet employer demands while giving priority to the needs of low-skilled workers and their families?

The purpose of this paper is to contribute to the dialogue about these and related questions. We review seven broad social, economic, and demographic trends affecting the US labor force now and in the future. The seven topics deal in turn with demographic trends, work and family issues, health and pension patterns, technical change, adjustment to low unemployment, globalization, and low-skilled workers. The purpose is first, to set the context for new research by bringing together existing knowledge and second, to provide some initial ideas relevant to public policy. Clearly, government policies have a considerable influence on the job market. If new policies are derived informed by the latest research, better policies may materialize.

I. Demographic Change and the Future Workforce

Important demographic trends will take place in the workforce over the next 10-15 years. The emerging patterns are the result of ups and downs in birth rates (low in the late 1920s and early 1930s, high in the late 1940s through the early 1960s, and modest growth in the late 1970s through the early 1990s). The population and labor force will continue to diversify, as immigration continues to account for a sizable part of population growth. Projections suggest that the Hispanic and Asian shares of the population will rise from 14 percent in 1995 to 19 percent in 2020.

BLS projections imply that over the next decade, 40 million people will enter the work force, about 25 million will leave the work force, and 109 million will remain. Although only a modest reduction will take place in the overall growth in the workforce (from 1.3 percent per year to 1.1 percent per year), the composition of growth will generate rising shares of young (under 25) and older (45 and over) workers and a decline in the share of middle-age workers.

These trends constitute a sharp reversal of the last decades. Consider the trends in the youth labor force (16-24 year-olds). After declining by 9 percent from 1986-96 and not growing between 1976 and 1986, the youth labor force will keep pace with the overall labor force with an expected 15 percent increase over the next decade. More dramatic are the changing patterns of growth among prime age workers and older workers. The prime age group of 25-54 year-olds accounted for virtually all the workers added to the labor force over the last two decades. Between 1976 and 1996, 38 million prime age workers and 1.7 million workers 55 and over joined the labor force, while reductions in the youth labor force amounted to about 2.1 million. Over the next decade, instead of having nearly all increases in employment coming from the 25-54 year-old age group, fewer than one in three (31 percent) of the added workers will be in this category. Nearly half of the additional workers will come from the 55 and older category, while about one in five will come from the youth labor force.

The overall reversal in the prime age category of workers masks a major change within the group. Note in Table 1 that the most experienced workers (45-54 year-olds) will expand rapidly enough to raise their share of the labor force. At the same time, declines will take place in the absolute numbers of 25-34 year-olds and of 35-44 year-olds. As a result, the proportion of 25-44 year-olds in the labor force will decline from 52.6 percent in 1996 to 44.5 percent in 2006. Workers in the 45+ categories will raise their demographic share from 32 percent to 39 percent. These are large and dramatic changes for a decade. The labor force share is increasing among older workers (from 28.8 percent to 36.2 percent among 45-64 year-olds) and younger workers (from 15.8 percent to 16.4 percent among 16-24 year-olds). However, the proportion of workers beyond the normal retirement age of 65 will remain below 3 percent.

The specific trends in the age composition of the work force vary with future time periods and are subject to uncertainty related to labor force participation rates. The aging of the population is largely the result of boom in births during the 1946-64 period. Over the coming decade (through 2005), substantial growth will occur among 45-64 year-olds, but the number over age 65 will increase only modestly (by five percent). However, between 2005 and 2010, the population of 65-69 year-olds will rise by 17 percent and then explode by another 37 percent in the 2010-2020 period. The jump in the 70 and over population will occur between 2010 and

Table 1: The Changing Mix of the US Labor Force by Age, Ethnicity, and Sex:

1976-2006

  1976 1986 1996 2006

Total Labor Force

96,158

117,834

133,943

148,847

Percentage of the Labor Force

Age

16-19

9.4

6.7

5.8

6.0

20-24

14.9

13.1

10.0

10.4

25-34

25.2

29.4

25.3

20.7

35-44

18.0

23.1

27.3

23.8

45-54

17.7

15.1

19.7

23.6

55-64

11.9

10.1

9.1

12.6

65-74

2.6

2.2

2.4

2.2

75+

0.4

0.4

0.5

0.6

Ethnicity

White, Non-Hispanic

N/A

79.8

75.3

72.7

Black

N/A

10.7

11.3

11.6

Hispanic

N/A

6.9

9.5

11.7

Asian

N/A

2.9

4.3

5.4

Age within Sex

Males

59.5

55.5

53.8

52.6

16-24

22.3

18.7

15.5

16.4

25-39

35.3

42.6

39.7

32.4

40+

42.4

38.7

44.8

51.2

Females

40.5

44.5

46.2

47.4

16-24

27.2

21.2

16.2

16.4

25-39

33.1

42.1

38.9

31.7

40+

39.7

36.7

44.8

52.0

Source: Howard Fullerton, ALabor Force 2006: Slowing Down and Changing Composition, Monthly Labor Review, November 1997, p. 23-38.

2020, rising by 38 percent from 24.6 to 31.8 million people.

How these figures translate into labor force participation is a major question mark. Among older workers, there are a variety of relevant factors. Private retirement pensions will

cover an increasing share of workers 65 and over and thus should encourage retirement. However, older workers will increasingly be drawn from those with college degrees or some years at college and fewer will be high school dropouts. In 1997, 27 percent of 60-69 year-olds lacked a high school degree or equivalent and only 18 percent had a BA degree. In one decade, only 17 percent of 60-69 year-olds will be without a high school degree and 27 percent will have earned a BA degree. Since educated workers participate in the work force at substantially higher rates than less educated, the country could see a reversal of past trends in labor force participation rates. In one such scenario, the Hudson Institute predicts substantial increases in the labor force participation rates of the 55 and older population. Participation rates of male 55-64 year-olds would return to 1970 levels and about half of 65-70 year-old men would work. While these upward shifts appear unlikely, even moderate growth in participation by older workers would significantly raise the growth of the overall labor force beyond what is currently projected. Poll data supports this claim, indicating that most baby boomers expect to work beyond age 65.

Outside the US, other OECD countries also exhibit the trend toward an older labor force. Between 1995 and 2030 the proportion of the labor force made up of 45-59 year-olds is projected to increase from 25.6 percent to 31.8 percent in the OECD as a whole, while the share of workers ages 60 and over is projected to rise from 4.7 percent to 7.8 percent. Increases in the educational attainment of older workers, together with the increasing demand for skilled workers, may raise the share of older workers further by stimulating higher labor force participation rates. In OECD countries as whole, the proportion of 45-59 year-olds with more than a high school degree will rise from about 21 percent in 1995 to 30 percent in 2015. Although the US will see minimal educational gains in this age group, the educational attainment of US workers over age 60 will rise significantly. Policy measures, such as raising the official retirement age and lessening financial disincentives to work, may encourage delays in retirement.

Shifts in the ethnic composition of the workforce will continue the patterns of recent decades. Immigrants will account for as much as half of net population growth over the next decades. Between 1996 and 2006 white non-Hispanic entrants will make up 49 percent of new labor force entrants, up from 43 percent during the previous decade, but well below the 1995 level of 76 percent. As a result, the share of non-Hispanic whites will fall to 73 percent in 2005. Of the nearly 15 million worker increase in the 1996-2006 period, about 7 million will be Hispanic or Asians. Hispanic-Americans will raise their share of new workers slightly from 29 percent to 31 percent, as will Asian-Americans whose share will grow from 14.5 percent to 15.7 percent. By 2020, white non-Hispanic workers will make up only 68 percent of the work force.

One concern about the changing ethnicity is the potential impact on the educational structure of the work force.Hispanic workers have the lowest educational attainment of any major ethnic group; only 55 percent of the Hispanic population over age 25 had completed high school as of 1997, well below the 85 percent completion levels among non-Hispanics. Thus, unless Hispanic youth and immigrants raise their educational attainment, their growing presence in the job market will lower the educational base of the labor force at the very time when the demand for skills is continuing to increase. The expanding share of Asians in the labor force will moderate this trend, since their educational attainment is higher than the rest of the work force. As of 1997, 42 percent of Asians over 25 had at least a BA degree, well above the 23 percent rate for the overall population. As a whole, immigrants have an educational profile that embodies higher proportions lacking a high school diploma, but the same share of college graduates as non-immigrants.

While the last two decades witnessed significant increases in the share of women in the work force (rising from 40.5 percent in 1976 to 46.2 percent in 1996), the female share will barely increase over the next decade. Still, by 2006 women will account for nearly half (47 percent) of the work force. However, given the age composition shift away from the 25-44 year-olds, a declining share of women workers will be mothers with young children.

Changing marital and living arrangements could have significant implications for the work force. Labor force participation rates are much higher and unemployment rates much lower among married than among unmarried men and women. Even in today=s tight job market (1998:1), unemployment rates are high among individuals who are in the never-married category. Never-married men experienced an 8.2 percent unemployment rate, far below the 2.1 percent rate among men who are married and living with a spouse. In addition, labor market outcomes are better among men living with at least one of their own children than among men with no children. The unemployment rate of never-married men is only 5.7 percent among those with children, but over 8 percent among those without children. Projections indicate that the share of American households consisting of families with children will decline from 48 percent to about 41 percent and that married couple families with children will make up less than one-third of families by 2010. To some extent, it is changes in employment opportunities that cause changes in marriage and family formation patterns and not the other way around. However, some of the marital and family changes have other causes and may well lead to worse job market outcomes.

Demographic trends will vary substantially by region of the country. Population growth will be much more rapid in the West and South than in the Northeast and North Central regions of the country. Projections suggest that California alone will add about 10 million people by 2015, or 22 percent of the nation=s total population growth. Texas and Florida will add close to another 10 million people. These three states, which currently account for about 25 percent of the US population, will absorb 44 percent of the nation=s population growth. The growth of these states can be attributed to the fact that immigrants will continue to make up a large share of net population growth and they tend to concentrate in a few states. In 1996, for example, over two of three immigrants declared their intended state of residents as California, New York, Texas, Florida, New Jersey, or Illinois. These states absorb an even higher share (perhaps 85 percent) of the illegal immigration.

Selected Implications of the Changing Demographics

The declining proportion of middle age individuals in the work force has a number of implications. First, rising shares of workers will have over 25 years of experience or less than seven years of labor market experience. Fewer will be in their early careers. The age shifts in the labor force should exert little or no impact on the aggregate unemployment rate. Given today=s unemployment rates within age categories, the overall unemployment rate in 2006 will be identical to today=s average rate. Changes in the age distribution of the work force will neither raise nor lower the overall unemployment rate.

Second, the declining numbers of 25-34 year-olds, together with their changing ethnic mix, may portend shortfalls in key professional areas. The number of earned BA degrees will remain constant over the next decade (at about 1 million per year) despite the rising demand for skilled workers and the increasing size of the labor force. As a result, new BA=s will decline as a proportion of all new entrants to the labor force from 32 percent in the 1986-96 period to 30 percent over the following decade.

Third, demographic trends raising the percentage of older workers and potential workers have implications for individual, firm-based, and government training. According to a recent OECD report, the US is distinctive in that training peaks in the 45-54 year-old years and drops off only moderately among the 55-64 year-olds. Table 2 shows that while older US workers are more likely to obtain training than older workers in other countries, US firms are less likely to finance training for younger workers than firms in other countries. Still, to the extent that the US attempts to raise participation rates of older workers, the current moderate amounts of training provide a base on which to build. Labor markets are generating jobs with higher skill requirements, but taking advantage of these opportunities requires expanded training opportunities, especially among older workers trying to avoid the effects of obsolescence. Since firms generally do not train less educated workers, the growing number of older, less educated workers are likely to place an added strain on the public training system.

The decline in labor force participation with age is also highest among less educated workers.Part of the reason is that their limited skills leave them with only low-wage options. Another explanation is that their Old Age Insurance under social security provides a higher than average replacement rate. Still, concern for the plight of this group causes many people to oppose raising the retirement age. Thus, effective training for the less educated could have a large payoff

Table 2: Participation in Job-Related Education and Training and Professional Training by Age Group in the US and Other Selected Countries

(Text Only)

 

% Participating in job-related continuing education and training

% Participating in professional and career upgrading training

 

Total

Paid by the employer

Total

Paid by the employer

United States

All Workers

44.5

31.7

29.7

24.1

15-24 year-olds

42.4

11.7

9.4

6.2

25-44 year-olds

46.4

36.2

32.5

26.9

45-54 year-olds

45.7

36.2

35.8

29.4

55-64 year-olds

36.8

25.8

28.4

21.4

United Kingdom

All Workers

50.2

39.2

15.4

12.9

15-24 year-olds

55.6

30.3

14.4

12.1

25-44 year-olds

55.2

46.3

18.1

15.0

45-54 year-olds

42.8

35.3

13.1

11.1

55-64 year-olds

32.1

27.2

8.6

7.4

Switzerland

All Workers

31.9

20.1

26.0

16.8

15-24 year-olds

34.5

23.5

21.0

12.6

25-44 year-olds

33.9

20.9

28.7

17.8

45-54 year-olds

29.7

18.9

25.6

18.8

55-64 year-olds

25.2

15.9

20.7

12.9

10-Country Average

All Workers

34.2

22.7

20.6

14.9

15-24 year-olds

38.8

16.0

15.2

9.1

25-44 year-olds

35.6

25.7

22.8

16.8

45-54 year-olds

30.8

22.6

21.1

15.8

55-64 year-olds

23.3

17.4

15.4

11.5

Source: Employment Outlook: June 1998, OECD, 1998, p.140, Based on International Literacy Survey.


first because of the enhanced capabilities of older trainees and second because their improved job accessibility may allay concerns over changing incentives under social security for all older workers.

Public training programs such as JTPA are likely to face rising shares of older workers among eligibles seeking services. The majority of older workers calling on JTPA services have utilized the displaced worker program and not the standard training programs. However, the share of young people is growing as well, especially among groups traditionally eligible for programs for the disadvantaged. Thus, JTPA will simultaneously see increases in the job displacement problems of older workers and in the initial training requirements for young workers.

From the standpoint of employers, there are advantages and disadvantages in hiring older workers. Their health care costs are disproportionately high and, since many will have more seniority than younger workers, they may receive higher pay and qualify for longer vacations. On the other hand, older workers are less likely to move and less likely to have an accident at work (though it takes longer for them to recover).

Third, a declining share of workers will have very young children. Women in the 25-44 age category will make up 21.1 percent of the work force of 2006, down from 24.2 percent in 1996. These figures incorporate an expected rise in the labor force participation rates of 25-44 year-old women from 76 percent to 79 percent. On the other hand, more women and men will have to care for elderly parents.

Fourth, the work force will increasingly become more heterogeneous by educational status and by gender. The proportions with BA degrees are especially variable by ethnic status among younger workers. As of March, 1997, a striking 51 percent of Asian 25-29 year-olds had earned a BA, as compared to 29 percent of whites, 14 percent of blacks, and 11 percent of Hispanics. Except for Hispanics, rates of high school completion were similar across groups, at about 85 percent. Another recent phenomenon is the emerging gender differences among black and Hispanic workers. Among 25-29 year-olds in the labor force, 20 percent of black women but only 13 percent of black men had earned BA degrees; among Hispanics, 17 percent of women but only nine percent of men had graduated college. These educational patterns are indicative of broader trends indicating that minority worker problems are becoming more concentrated among men.

II. Trends in Work and Family, Health Insurance, Pensions

The complications embedded in efforts to combine work and family are not a new phenomenon. Throughout the first half of the century, it was mainly low income and/or black women who faced the biggest struggle balancing work and family. These women often remained in the labor force after the birth of a child because they could not afford to quit working. Women married to middle- and upper-income men typically did not work for pay while their children were young, but waited to reenter the labor force at least until their children were school aged (Klerman and Leibowitz, 1994). Few children lived in single parent families.

Since the 1960s, several demographic trends have changed the structure of the American family and the way that parents balance work and family responsibilities. Women married to men at all income levels have increased rapidly their participation in the labor force, with the most rapid growth among college educated women married to relatively high income men. The majority of women now return to paid work within a year of the birth of a child. The rising divorce rate and the growing prevalence of children born to unmarried mothers means that many children live in single parent families.

Working parents, especially working mothers, report a great deal more stress in their lives than other workers. Journalists and academics argue that the very structure of the work place contributes to that stress; few jobs allow workers the flexibility of dealing with family responsibilities during normal business hours. Some employers have adopted policies to make the work place more "family-friendly" including flex-time, job sharing, generous parental leave following the birth of a child, and on-site child care. Evidence on the success of these programs in reducing stress is mixed. A rising share of workers are choosing self-employment, consulting, temporary work, or other forms of contingent work, which give them more flexibility in balancing work and family responsibilities. Moreover, an increasing share of men and women are providing assistance to elderly relatives. Women provide the majority of eldercare, and many providing eldercare are part of the so-called "sandwich generation," caring for children and elderly relatives at the same time.

Most Workers Do Not Have Children

In order to put a discussion of work and family in perspective, it is important to note that most workers presently do not live in households with their own children under 18. Although most women do have children at some point in their lives, only 40 percent of the female labor force and 36.2 of the male labor force live in a household with their own children under age 18 (See Tables 3 and 4). In addition, a relatively small share of workers live with their own small children. 16.4 percent of female workers and 16.5 percent of male workers live with at least one child under age six, and 9.1 percent of female workers and 9.8 percent of male workers live with at least one child under age three.

Table 3: Percent of Women in the Labor Force in Various Types of Families: Second Quarter of 1998


(Text Only)

Living Arrangements

Percent of female labor force

Percent of female full-time workers

Own children under age 18

40.0

39.6

Own Children under age 6

16.4

15.1

Own Children under age 3

9.1

8.1

Married, spouse present with children under age 18

28.1

27.2

Divorced, separated or widowed and own Children under 18

7.2

8.1

Never married and own children under 18

4.7

4.3

Source: US Bureau of Labor Statistics, unpublished tabulations.


Table 4: Percent of Men in the Labor Force in Various Types of Families: Second Quarter of 1998

(Text Only)

Living Arrangements

Percent of male labor force

Percent of male full-time workers

Own Children under age 18

36.2

38.6

Own Children under age 6

16.5

17.5

Own Children under age 3

9.8

10.4

Married, spouse present with children under age 18

34.3

36.6

Divorced, separated, or widowed and own children under 18

1.9

2.0

Never married and own children under 18

1.0

0.99

Source: US Bureau of Labor Statistics, unpublished tabulations.


Tables 3 and 4 also show that single parents comprise a relatively small share of the labor force; 7.2 percent of working women were ever married (divorced, widowed, or separated) and had children under 18, and 4.7 percent of working women were never married and had children under 18. For men, single parenthood was much less common. 1.9 percent of male workers were ever married with children under 18, and one percent were never married with children under 18.

Growth in Married Women’s Labor Force Participation

Part of the growing concern over balancing work and families is driven by the growth of dual-career couples with children. Women married to high wage, college educated men have witnessed the largest increase in labor force participation since the late 1960s. The labor force participation rate of the wives of men in the highest wage quintile increased 16.6 percent between 1969 and 1989. The wives of men in the middle wage quintile increased their labor force participation rates by 11.7 percent, and wives of men in the bottom wage quintile increased their labor force participation rate by 6.9 percent.

Women’s labor force participation has grown in part because women are taking less time out of the labor force after the birth of a child. Since the 1970s, mothers of infants have rapidly increased their labor force participation. By 1995, 55 percent of women who had had a child in the last year were in the labor force, almost double the 1976 participation rate of 31 percent (US Census Bureau, 1998). Today, mothers face the tradeoff between work and family even when their children are infants. In the 1970s and 1980s, most women did not return to the labor market until their children were in preschool; in the 1950s and 1960s, most mothers waited until their children reached elementary school to return to the labor force (Klerman and Leibowitz, 1994). In addition, the hours of work for employed women with children under the age of six have grown; the average employed white woman with at least one child under the age of six worked 1,487 hours in 1970 and 1,638 hours in 1988. College-educated women accounted for nearly all of the increase in the annual hours of work for mothers of young children conditional on employment (Coleman and Pencavel, 1994).

Working mothers with infants face difficulties that working mothers of older children do not. First, breast feeding has become more prevalent, but few workplaces provide facilities for breast pumping and storage of breast milk. Second, infant care is expensive and difficult to find. Third, the sleep deprivation that comes with having a baby can make work more difficult.

Growth in Single Parent Families and Single Mothers’ Labor Force Participation

Single parents face special challenges in balancing work and family. Between 1970 and 1997, single female headed families increased from 17 percent to 27 percent of all families with children, and single father headed families increased from one to five percent of families (US Census Bureau, 1998).

This section will address issues related to single mothers, since they are the vast majority of single-parent families. Women can become single mothers because of separation, divorce, or because they were unmarried when they gave birth to a child. The divorce rate more than doubled between 1966 and 1977 (from 10 to 21 divorces per 1,000 married women per year) and has remained relatively stable since. Over 40 percent of all existing marriages are expected to end in divorce (National Center for Health Statistics, 1998). Non-marital births as a percent of all births have grown quite rapidly in since the 1960s, rising from 5.3 percent in 1960 to 30.1 percent in 1992. The groups that experienced the largest increases in non-marital birth rates were 15 to 24 year-old women and black women.

Since 1992, the labor force participation rate of never-married mothers has grown enormously. Historically, never-married mothers participated in the labor force at low levels and often relied on AFDC. But since 1996, approximately 1 million never-married mothers have left the welfare rolls and entered the labor market. The labor force participation rate of never-married mothers rose from 53 percent in 1992 to 60 percent in 1996 and reached 70 percent in the second quarter of 1998. These increases in labor force participation were due to the combination of federal welfare reform legislation, which established time limits for the receipt of benefits for the vast majority of recipients, and a booming labor market (Bishop, 1998).

Historically, divorced and separated mothers have participated in the labor market at much higher rates than mothers with a spouse present and never married mothers. Recently, the gap has narrowed. As of the second quarter of 1998, 77 percent of divorced and separated mothers were in the labor force, compared with 70 percent of mothers with spouse present and 70 percent of never-married mothers.

Unmarried mothers living in poverty often face particular difficulties managing their work and family responsibilities. Because of the lack of affordable child care, these women often must place their children in poor quality care. States do provide some subsidized child care for former TANF recipients and other poor or near-poor women, but the programs in several large states have long waiting lists and cannot provide subsidies to all who apply (Urban Institute, 1998). In addition, women who rely on public transportation often face long and logistically difficult trips getting from home to child care and work.

Are Americans working longer hours than they did in the past?

One debate about work and family issues concerns whether work hours have grown in recent decades. Juliet Schor, in her 1991 best-selling book, The Overworked American, declared that workers were working longer hours than they had in the past, and that long work hours were robbing workers of satisfying lives. In 1997, Robinson and Godbey argued that Americans were actually working less than they did in the 1960s. Because individuals were also doing less housework and childcare, Robinson and Godbey found a net increase in leisure.

The increase in leisure that Robinson and Godbey describe is due to the rise in early retirement for workers in their 50s, and the delay of childbearing and marriage for workers in their 20s and 30s. The vast majority of married couples with children are spending more total time in paid work than they did in 1979 or 1989. Husbands worked an average of 2096 hours in 1979 and 2159 hours in 1994. Wives worked an average of 581 hours in 1979 and 1168 hours in 1994 (Economic Policy Institute, 1998).

Table 5 shows the work hours of married men and women according to the percentile of work hours of the husband. With the exception of husbands in the lowest and third fifth of work hours, husbands worked more annual hours in 1994 than they did in 1979. Men in the top five percent of work hours witnessed the largest increase in work hours, from 2490 in 1979 to 2615 in 1994. With the exception of the wives of men in the top five percent of work hours, wives worked substantially more hours in 1994 than they did in 1979. The husbands in the top of the work hours distribution are typically college-educated men married to college-educated women (Coleman and Pencavel, 1994). Therefore, it is college-educated dual career couples who are facing the biggest decline in nonwork hours

Table 5: Annual Work Hours of Husbands and Wives with at Least One Child

Percentile of Husbands’ Work Hours
(Text Only)

 

Bottom fifth

Second fifth

Middle fifth

Fourth fifth

80-95 percentile

Top 5 percent

Entire sample

Husbands’ Average Annual Work Hours

1979

1678

2074

2155

2226

2300

2490

2096

1989

1706

2138

2212

2274

2364

2554

2148

1994

1607

2145

2257

2340

2387

2615

2159

Wives’ Average Annual Work Hours

1979

478

733

876

1086

1164

828

581

1989

653

1009

1190

1332

1407

1146

1105

1994

634

1062

1290

1419

1605

929

1168

Source: Economic Policy Institute DataZone, 1998. Sample includes married couples with at least one child and one spouse between the ages of 25 and 54. Families with zero earnings were excluded.

Husbands in the lowest fifth of annual work hours actually experienced a decline in work hours, from 1,678 working in 1979 and 1,607 working in 1994. Men in the bottom of the work hours distribution are likely low-skilled men, who have decreased their labor force participation rates and their work hours in response to the marked decline in their wages (Juhn and Murphy, 1997).

When paid work and unpaid work are added together, on average, men and women spend equal amounts of time on work. Men spend more hours working for pay than women do, and women spend more time doing housework and child care (Barnett and Shen, 1997).

The struggle to juggle work and family likely takes a toll on the emotional health of parents—particularly mothers. In a recent survey of workers, 38 percent of mothers and 19 percent of fathers said they felt nervous and stressed out often or very often. Similarly, 48 percent of mothers and 35 percent of fathers responded that they were often or very often tired when getting up in the morning. Thirty percent of mothers and 23 percent of fathers registered the same responses when asked if they felt emotionally drained from their work (Galinsky et al, 1998).

Shortage of Quality Child Care

As women’s labor force participation has grown, the demand for child care has also grown. A recent study found that 56 percent of mothers with children under age 5 said that finding affordable child care was a serious problem for them (US Department of Labor, Women’s Bureau, 1998). The shortage of quality is particularly acute for infants, many of whom are placed in care that is physically dangerous to them. Many parents are forced to miss work when a child becomes sick because they lack backup care.

There is also a severe shortage of child care available on the evenings and weekends. Nearly one-fifth of workers worked nonstandard hours in 1991, and women comprised one-third of those working nonstandard shifts. Service sector jobs requiring these nonstandard work hours are among the fastest growing (US Department of Labor Women’s Bureau, 1998).

Recent research has shown that poor quality child care could ultimately take a toll on children. Research on the brain development of children shows that the first three years of life are key in developing a child’s full intellectual and emotional potential. Poor quality care reduces a child’s future cognitive abilities and emotional health (Galinsky et al, 1998).

Eldercare

The projected growth in the share of the population older than age 65 means that a growing share of adults will spend time caring for their relatives. Americans are now living longer lives than ever before, but many of their later years are spent in poor health, with the need of assistance from family, friends, or health care workers. Many working adults must also care for elderly relatives. In 1997, one-quarter of workers had provided special assistance to someone 65 years or older within the last year, while 13 percent had done so in the last month. One-fifth of workers had cared for both elderly relatives and children within the last year. Those who do care for elderly relatives spend an average of 11 hours per week doing so. Women do more eldercare than men, and studies show that women often reduce their work hours in response to taking on the responsibility of eldercare.

Can Firms Help Their Employees Balance Work and Family Responsibilities?

Many firms offer policies that aim to help workers balance work and family responsibilities. In a 1992 survey, 90 percent of workers said they had access to family sick leave, 57 percent have access to job sharing, 47 percent can take extended lunch breaks, 44 percent have the ability to work more hours one day and fewer the next day, 29 percent can choose flextime, and 24 percent have the ability to work at home on a regular basis (Galinsky et al., 1996). While few small firms offer assistance with child care and eldercare, many Fortune 1000 firms do; 55 percent of Fortune 1000 firms offer child care resources and referral, 21 percent offer eldercare resource and referral, and 13 percent offer on-site child care.

Some family friendly policies, such as flexible time, parental leave, and dependent care assistance, have little impact on parental stress (Galinsky et al, 1998). The take-up rates of policies that reduce work hours, such as work sharing and flextime tend to be quite low because workers fear that reducing their work hours will hurt their careers. Anecdotal evidence indicates that workers who opt for flexible or reduced work hours often are promoted at much lower rates than their colleagues (National Public Radio, 1998). The programs that have the highest take-up rates are those which enable employees to increase their work hours: on-site child care, child and elder care referral services, and emergency backup child care services (Stone, 1997).

The job attributes that were most correlated with low parental stress were not policies explicitly targeted towards families, but attributes that improve overall job satisfaction. Employed parents experienced the least conflict between their work and family responsibilities when they were in jobs that had greater autonomy, that were less demanding and hectic, and that offered greater job security (Galinsky et al, 1998).

Anecdotal evidence indicates that some employers are finding that workers that lead balanced lives are more productive. Those firms are aiming to improve employee satisfaction by increasing communication between workers and bosses, improving scheduling flexibility, and recognizing workers’ needs to balance work and family (Wall Street Journal, July 1998). Such policies are not aimed at a small subset of workers who choose flextime, but are transforming their entire work culture. Xerox reported a 10 percent productivity increase when implementing flexible scheduling and Johnson & Johnson reported a 50 percent decline in absenteeism among employees who used flexible work options and family leave policies (Price Waterhouse, 1998).

Several large consulting and accounting firms are using more flexible work scheduling as a method of recruiting and retaining workers in a tight labor market (National Public Radio, 1998; Price Waterhouse, 1998). At the same time, consulting and accounting firms are trying to change their workaholic culture to make flexible scheduling more feasible.

Two policies, flextime and the four-day compressed work week, appear to be effective in increasing the amount of time parents spend with children. Some researchers have also argued that increasing the amount of paid paternal leave would increase fathers’ involvement with their children. The liberality of companies’ paternity leave policies is a predictor of the length of parental leave. Fathers’ involvement with older children is positively correlated with the amount of time they spent with their children as infants. Researchers argue that if fathers spend more time with their newly born children, they are more likely to bond with their children, and spend more time with children as they grow older. Many companies do not offer paid paternity leave, although the 1996 Federal Medical and Family Leave Act does require all firms with more than 50 employees to offer two weeks of unpaid maternity or paternity leave. The vast majority (75 to 91 percent) of fathers take time off following the birth of child, but the average length of leave averages only five days.

Because relatively few jobs provide the autonomy and flexibility that allow workers to balance work and family effectively, some workers, especially women, have likely turned to contingent employment and self-employment. Contingent workers include workers who are employed by temporary agencies and other workers whose implicit or explicit contracts with their employers define their jobs as temporary. Women who are self-employed or contingent workers typically do not work full-time full-year, but vary their schedule over the year and over the week to meet family demands. The majority of women who are self-employed, on-call, or independent contractors enter into these situations voluntarily, and many state that they work in such arrangements for family reasons (Economic Policy Institute, 1997). The female self-employment rate has grown from 4.1 percent in 1990 to 6.7 percent in 1994 (Devine, 1994). Nearly three percent of the female work force is in temporary, on-call, or contract company jobs (Economic Policy Institute, 1997).

III. Trends in Employer-provided Health and Pension Benefits and Families

The Decline in Employer-Provided Health Insurance

Employer-provided health insurance coverage declined sharply in the last decade. In 1988, 67 percent of the population had such coverage, in 1993, 61 percent did. The rate of coverage declined for every income group, with the sharpest decline (7 percentage points) for individuals with incomes between 100 and 200 percent of poverty and the smallest decline among individuals with incomes below the poverty line (Blumberg and Liska, 1996).

Most of the decline in employer-provided health insurance coverage is due to declining coverage for dependents, with the largest decline for children. About 18 percent of workers have no employer-provided health coverage, and an additional five percent have health insurance coverage for themselves, but not for their families. Moreover, 30 percent of firms now require individuals to share the cost of family health insurance coverage, and many workers opt not to cover dependents when the price is too high. About 20 percent of firms also require individuals to share in the cost of their coverage, but employee costs for their own coverage are typically modest compared to the cost of family coverage (General Accounting Office, 1997).

For poor women and children, Medicaid has made up for some of the decline in private health insurance coverage. Since the late 1980s, the Medicaid program has been expanded to cover poor pregnant women and children who are not recipients of AFDC. As a result, between 1988 and 1993, Medicaid coverage expanded from 8.5 percent to 12.4 percent of the nonelderly population. Because of the Medicaid expansions, poor children are much more likely to be insured than poor adults (Blumberg and Liska, 1996). Some analysts have argued that the expansion of Medicaid eligibility actually contributed to the decline in employer-provided health coverage (Cutler and Gruber, 1996). But Medicaid only fills the private insurance gap for about 50 percent of individuals below the poverty line.

How will the decline in employer-provided health insurance affect the health status of the population? Individuals who are uninsured have much less access to health care than those who do not, even adjusting for income and health status. Uninsured individuals are likely to delay or forgo medical care, even in the case of serious medical problems. They tend not to receive preventative care, so health problems are often much worse when they finally visit the doctor. In addition, when persons without health insurance coverage do receive care, doctors tend to perform fewer procedures. The uninsured receive better care in communities with public hospitals and clinics as these institutions provide free medical care.

Does the lack of access to health care result in a decline in the health of uninsured individuals? A lack of health insurance coverage is correlated with a poor health status and a decline in life expectancy. However, it is difficult to identify a causal relationship because health insurance coverage is endogenous; persons in poor health are less likely than healthy workers to find good jobs that provide health insurance coverage.

For many workers, the absence of health insurance coverage is only temporary. The median spell without health insurance is seven months, and 48 percent of spells of non-insurance end within five months. A significant minority, 19 percent, last beyond two years (Blumberg and Liska, 1996). An interesting unanswered research question is whether short lapses in health insurance coverage affect access to health care as much as long spells.

The Change in the Pension System

Over the last twenty years, the fraction of workers covered by defined contribution (DC) pension plans has increased rapidly, while the fraction covered by defined benefit (DB) plans has declined. In 1975, 13 percent of workers with pensions had DC plans as their primary pension, compared to 33 percent in 1988 and 42 percent in 1993 (Samwick and Skinner, 1998).

Some analysts are concerned that the switch from DB to DC plans will reduce workers’ retirement savings. Under DC plans, workers bear much of the responsibility for accumulating adequate assets for retirement for three reasons. First, contributions to DC plans are often voluntary. In a sample of workers from several large firms, only about 70 percent chose to participate in voluntary DC plans. On average, workers contribute about six percent of their salaries (Employee Benefits Research Institute, 1995).

Second, workers can control the allocation of assets in their DC portfolio. Many workers choose far too conservative investments, which dramatically reduce their value of their assets at retirement. About one-sixth of workers invest none of their portfolio in equities (Samwick and Skinner, 1998) and many of these very conservative investors are relatively young. In one dimension, workers’ investment strategies are sensible: the percent of assets in equities declines as workers get closer to retirement (Employee Benefits Research Institute, 1995).

Third, workers can choose to consume, rather that reinvest, the DC lump sum distributions they receive when they leave a job. Only half of those who receive lump sum distributions reinvest all or part of it into an IRA or their new employers’ retirement plan (Employee Benefits Research Institute, 1997).

The lump sum distributions of DC plans have the advantage of making the pensions portable. If workers choose to reinvest their lump sum distribution, they suffer no penalty in retirement income from changing jobs. When workers change jobs, they suffer a substantial decline in the expected value of their DB plan income.

A recent paper (Samwick and Skinner, 1998) examined the retirement income yield for average pension recipients under representative DB and DC plans. The Samwick-Skinner model takes account of the reality that many workers will consume the lump sum distributions from DC plans when they change jobs and that many workers will continue to be conservative in their portfolio choice. Still, they find, somewhat surprisingly, that the average worker will have a higher retirement income under a DC plan than under a DB plan. However, workers at the lowest ten percent of the retirement income distribution are projected to earn higher retirement income under DB than under DC plans. Although many DC plans do not adequately provide for retirement, neither did many DB plans in the past.

IV. Technology and Work Organization

Business investment in computers has grown very rapidly over the last 25 years. In 1973, business computer purchases accounted for .005 percent of GDP, and in 1996, they accounted for 3.1 percent of GDP. Computers now comprise about one-fifth of businesses’ new capital investment. This investment in technology has meant that one-half of American workers are likely to use a computer in their work, compared with one-fourth in 1987 (Madrick, 1998).

On average, technology has increased the demand for workers who have good math, communication, and teamwork skills. However, the impact of technology has not been uniform; it varies by occupation, industry, and among firms within the same industry. Technology increases the skill demands of many manufacturing jobs, but may deskill some clerical work. Industries vary in their investment in information technology. Evidence shows that industries that invest the most in technology have the highest demand for skilled workers, and pay a higher wage premium to workers than other industries. Manufacturing industries that invest the most in technology have a higher demand for non-production workers (managers, secretaries, and janitors) relative to production workers (line workers).

The impact of technology varies considerably among firms within the same industry because firms differ widely in their organizational responses to technology. Technology is not simply an input into the production function. It changes the production function by affecting work organization and the expanding the types of possible outputs. In the long run, a firm’s type of product, its work organization, and its level of investment in technology are complementary. Although firms choose a myriad of organizational responses to technology, it is heuristically useful to think of two types of strategies, customization and more traditional work practices (Piore, 1998). With customization, firms produce products in which quality and/or control of the production product are key. They invest heavily in information technology and employ a cluster of work organization practices that both expand workers’ decision making power and flexibility and compensate workers for their performance. Firms that follow the customization strategy require the most skills from their workers, and also experience the largest productivity gains from adopting computers.

Firms that choose more traditional work organization produce products that are less customized and they invest less in technology. Such firms do not fully transform their work organization in response to technology, rather their method of work organization is closer to mass production. Firms that follow the second strategy do not demand as many skills from their workers as firms that follow customization, and do not experience as large a boost in productivity from adopting new technology. They may adopt some new work place organization practices, but do not fully transform their workplace. The choice of strategies varies widely among firms within the same industry.

The future impact of technology on the labor force will depend on whether more firms adopt the customization strategy in response to new technologies. Competitive pressures may push more firms to change their workplace organizations. However, many firms will find it difficult to transform themselves because of the resistance of workers, lower to mid-level managers, and/or labor union leaders.

Recent Changes in Work Organization

Accompanying the growth in computer investment has been a movement away from the Ford/Taylor model of mass production, once a major form of work organization in the US Under the mass production model, workers performed a narrow set of tasks, and had little incentive or need to understand how their work fit into the larger picture. Monitoring quality was the responsibility of supervisors.

No single form of work organization has emerged as dominant, but most firms have adopted flexible forms of production. Workers often perform a broad range of tasks thereby gaining the ability to respond quickly to changes in the production process and assuming some responsibility for the quality of the product (Piore, 1998).

Information technology was not the only factor contributing to movement away from the mass production model. During the 1970s, several changes in the macroeconomic environment led to increased levels of uncertainty and volatility in the demand for products. The macroeconomic changes included the opening of the economy to trade and deregulation. Additionally, the change from fixed to fluctuating exchange rates led to wide swings in the value of the dollar. The 1970s also witnessed more rapid shifts in consumer demand, which led to quicker product cycles, more products, and quicker supplier responses to changes in consumer tastes (Piore, 1998). Consumer demand also became more fragmented; the number of new food and household products introduced each year has increased 15 or 20 fold since 1970.

Heuristically, it is again useful to think of two types of strategies that were responses to the changing economic environment: customization and more traditional work arrangements. The nomenclature is based on Piore’s (1998) analysis of the case study literature on manufacturing firms; Bresnahan, Brynjolfsson, and Hitt (1998); and quantitative analysis of data from 303 Fortune 1000 firms in both manufacturing and services.

First, firms that chose the strategy of customization responded to rapid changes in consumer demand by producing products in which service was an important component that can be easily customized by controlling the production process. These firms have been heavy investors in information technology and have transformed their work organization practices. They organize work into self-directed teams, which gives workers more decision-making authority (particularly over pace and method of work). They also rotate workers among jobs, improving their understanding of the production process. In addition, these firms implement incentive systems to reward and encourage performance. Such firms also require workers to perform highly skilled tasks, such as reprogramming equipment for a change in the product line, so firms typically invest heavily in worker training and screening (Piore, 1998). According to one estimate, nearly 40 percent of firms follow the customization model (Osterman, 1994). The positive relationship between information technology and new work organization practices cannot be explained by industry type alone; the relationship exists both within and between industries (Bresnahan, Brynjolfsson, and Hitt, 1998).

Second, some firms choose to produce products in which service and producer control are not as important. These firms tend to use fairly traditional forms of work organization. Most firms have implemented at least one but not all of the nontraditional human resource practices that exist in the customization strategy (Osterman, 1994). Appelbaum and Batt (1994) find that the fraction of Fortune 100 firms with at least one employee involvement practice rose from 70 percent in 1987 to 85 percent in 1992.

Firms that follow the customization strategy experience greater productivity gains from the adoption of information technology than other firms—even firms within their own industry (Beede and Young, 1996; Brynjolfsson, and Hitt, 1998). The customization strategy also results in large improvements in productivity when new organization practices deeply penetrate the firm (Cappelli et al (1998), Ichniowski et al (1997). However, firms that implement only some of the new work organization practices experience only marginal improvements in productivity.

Will More Firms Adapt Their Work Organization to Fully Exploit Technology?

Technology’s future impact on the demand for skills and on productivity will likely depend on how many additional firms adopt customization as a form of work organization. David (1990) argues that a new technology can take decades to have its full impact on productivity because it takes some time for firms to adopt work organization to the technology. Such was the case with the adoption of electricity in the early twentieth century.

Although competitive pressures may push additional firms to adopt customization, many firms may find it difficult to reform their work organization. Piecemeal changes to work organization are easier for firms with a tradition of mass production to implement than customization (Piore, 1998). Management and union leaders may oppose such changes because they threaten their power and workers may have difficulty conforming to new modes of work organizations because old work norms have become ingrained (Cappelli et al., 1998; Bresnahan, Brynjolfsson, and Hitt, 1998). Also, decades of animosity between management and labor may prevent the cooperation necessary to reform work. Because they lack an institutional memory, greenfield (newly built) plants are more likely to adopt customization than other plants (Ichniowski et al., 1995).

Does Technology Change the Demand and Returns to Skills?

The growth in information technology and changes in work organization have likely contributed to the rising demand and returns to higher-order cognitive skills. Anecdotal evidence shows that new technologies and workplace practices often require workers to have good writing and verbal skills, good math skills, and good problem solving skills. In addition, given the increased autonomy and responsibilities in many environments, there is evidence that so-called "soft skills," motivation, work habits, etc., have become more important. Although there has been an economy-wide increase in the demand for skills, some jobs may have been deskilled by the changes in work organization and growth in information technology. In particular, some clerical work has become less skilled now that office technology, such as word processing software, can perform many of the most skilled clerical tasks.

Because of other sweeping economy wide changes that were occurring at the same time as technological change and the change in work organization (including globalization, deregulation, the decline of unions), it is difficult to empirically identify the magnitude of technology’s effect on the rising skill premium.

Technology has played a role in the rising demand and return to skill. Three papers (Bound, Berman, and Griliches, 1994; Allen, 1997; Haskel and Slaughter, 1998) find that industries that invest the most in technology have the highest demand for skilled workers, and pay the highest relative wages to skilled workers. When the three studies are taken together, the result is robust across the manufacturing and service sectors, across different measures of technology, and across different proxies for unskilled and skilled workers.

The shortcoming of the industry-level studies is that they cannot say whether the adoption of computers per se increases the demand for skilled workers, or whether the adoption of computers is simply a proxy for other organizational changes in the firm. Recent evidence from worker-level data suggests that it is not the use of the computer per se that raises the relative earnings of skilled workers; the use of a computer is simply a proxy for whether individual sits and thinks at his or her job (DiNardo and Pischke, 1997).

Mishel and Bernstein (1997) argue that technological change cannot explain the rising demand and returns to skill. They believe that in order for the relationship to hold true, technological change should have accelerated between the 1980s and the 1970s, because the wage gap between skilled and unskilled workers grew more quickly during the 1980s than during the 1970s. They find no evidence of an acceleration in investment in technology in the 1980s. The discrepancy in timing between the introduction of technology and wage inequality may be explained by the fact that the new work organization practices that exploited technology became more widely adopted in the 1980s. Bernard and Jensen (1994) also argue that technological change cannot account for much of wage inequality, because they show that the within-industry changes in the employment shares and wage shares of skilled workers were larger than the between-industry changes. They assume that all firms within the same industry would have similar technologies, which has been shown to be empirically false.

Many papers addressing the issue of skill biased technical change have simply assumed the change was the residual that could not be explained other factors, such as changes in demographics, changes in the composition of industries, or trade.

Given that whether and how firms adopt technology depends on their response to changes in the external environment (e.g. trade, deregulation), a research program that attempts to allocate contributions to wage inequality may nevertheless fail to achieve a consensus. For policy purposes it is more important to first determine what skills are valued in the economy, then to go about training workers in those particular skills.

V. Adapting to Tight Labor Markets

With the US economy reaching the lowest unemployment rates in 30 years and the employed share of the adult population at an all time high, today’s primary concerns are labor shortages and inflationary pressures resulting from tight labor markets. According to many predictions, the 1995-1996 unemployment rates of 5.5 percent should have already led to excessive wage growth. In a recent estimate, Akerlof, Dickens, and Perry (1996) concluded that the rate of unemployment consistent with no increases in the inflation rate was in the 5.5-6.0 percent range. The US experience of 1997 and 1998 cast doubt on these and similar projections. Even after reaching 4.5 percent unemployment rates, the US economy has yet to experience inflationary wage pressures.

How have these pressures been averted? Are they about to arise shortly? How are employers coping with the tight job markets? Is rapid wage growth taking place in the lowest unemployment rate areas? To what extent have new workers been drawn into the job market to mitigate shortages and wage pressures? What mechanisms other than wage increases are employers using to recruit and retain workers? Are employers turning to low turnover strategies with job ladders and extensive training? To what extent are employers able to lower their formal job qualifications in response to a shortage of workers?

The National Trends

The impact of the 1990s expansion on the labor market is unusual in two respects. First, the absence of any observable wage pressure in the context of a 4.5 percent unemployment rate is unexpected. Figure 1 graphs the trend in unemployment with the trend in the Employment Cost Index (ECI), which best captures the potential inflationary pressures emerging from the labor market, and the Consumer Price Index (CPI). Note that from 1986 to 1989, when unemployment

rates fell from 7.0 percent to 5.3 percent, the ECI rose from a growth rate of 0.7 percent per year to 1.2 percent per year and the CPI doubled from 2 percent to 4 percent per year. Yet, in the 1990s, even larger reductions in the unemployment rate has induced no increase at all in the ECI or CPI. Second, the dramatic reduction in unemployment rates has stimulated only a modest

{short description of image}

impact on participation in the job market. Note in Figure 2 falling unemployment in the late 1970s and mid to late 1980s attracted many new workers into the market, raising participation rates 1.5 to 2 percentage points. In contrast, the decline in the 1990s to 4.5 percent unemployment rates has only led to a 0.7 point increase in participation. Thus, large increases in labor supply cannot explain the limited impact on wages and prices.

One feature of the current expansion that follows past patterns is that the expansion has raised employment most among the more disadvantaged groups. Table 6 reveals that the percentage point gains in employment-population ratios and declines in unemployment rates were substantially higher among minorities, teenagers, and less educated workers than among prime-age males. For example, the unemployment rate among black men, ages 20 and over, fell an extraordinary seven percentage points, from almost 16 percent to about 9 percent. White men, ages 20 and over, also experienced sizable reductions in unemployment rates, but virtually no movement in employment. Similarly, the unemployment rate of college graduates declined from 3.2 percent in 1992 to 1.7 percent in 1998, while their employment-population ratio of college graduates remained constant at 78 percent.

With the economy apparently running out of skilled workers since nearly all were already employed earlier in the business cycle, employers must turn to less-qualified workers to fill the new job. These pressures are good for the disadvantaged---firms are more willing to take a inexperienced, less educated workers, to expand training, and to lower hiring standards. But shortages of high-skilled workers could lead to inflationary wage increases while adding low-skilled workers could lower productivity and raises costs.

Firms might have to alter their production approaches to the extent that the mix of skills

{short description of image}

Table 6: Gains in Employment-Population Ratios and Unemployment Rate Reductions by Age, Ethnicity, and Education: 1992-1998


(Text Only)

 

1992

1998 (1st Half )

Percentage Point Change

% Change

Employment-Population Ratios

White Males, 20+

73.0

74.7

1.7

2.3%

White Females, 20+

54.9

57.7

2.8

5.0%

Black Teens

22.8

29.6

6.8

26.1%

All Teens

41.7

45.3

3.6

8.3%

Black Males, 20+

64.3

67.3

3.0

4.6%

Black Females, 20+

53.6

59.4

5.8

10.3%

Hispanic Workers

59.1

63.6

4.5

7.3%

HS Dropout

36.5

39.6

3.1

8.2%

HS Graduate

61.8

62.7

0.9

1.4%

Some College

71.0

72.3

1.3

1.8%

BA or Higher Degree

78.8

78.9

0.1

0.1%

Unemployment Rates

White Males, 20+

6.4

3.2

-3.2

-69.3%

White Females, 20+

5.5

3.5

-2.0

-45.2%

Black Teenagers

39.8

27.5

-12.3

-37.0%

All Teenagers

20.1

14.3

-5.8

-34.0%

All Black Males

15.7

8.7

-7.0

-59.0%

All Black Females

13.2

9.4

-3.8

-34.0%

Hispanic Workers

11.6

6.9

-4.7

-51.9%

HS Dropout

11.5

7.1

-4.4

-48.2%

HS Graduate

6.8

4.0

-2.8

-53.1%

Some College

5.7

3.1

-2.6

-60.9%

BA or Higher Degree

3.2

1.8

-1.4

-57.5%

Source: US Bureau of Labor Statistics, unpublished tabulations.

among new workers differs sharply from the mix among existing workers. Employers would seem to face serious problems integrating large numbers of less skilled workers into their organizations, particularly at a time when the demand for skill is increasing on a long-run basis.

Surprisingly, a close look at the data provides an entirely different picture of recent job market trends. As Table 7 reveals, fully 94 percent of the 11.7 million newly employed adult workers (ages 25 and over) had at least some college or a BA degree and over half of the came from the highest educational category. Demographics and educational distributions by age allows us to reconcile the substantial improvement in the position of less-educated workers with the high levels of education among the newly employed. The normal tendency at peaks in the business cycle for employers to draw on less educated workers has been offset by the long-term increase in the educational status of the population.

While the typical dropout had a much easier time finding a job in 1998 than in 1992, dropouts did not account for any of the growth in employment. The reason was that the high school dropout population declined by 2.8 million, as the number of young high school dropouts becoming adults was smaller than the number of older dropouts dying. At the same time, while the typical adult college graduate was no more likely to be employed in 1998, the population of college graduates ages 25 and over increased by about 7.5 million, or 20 percent, well above the 7 percent growth in the total adult population. As a result, college graduates constituted 65 percent of the 11.5 million increase in the 25 and over population.

Thus, the striking reality is that employers have been able to draw on a growing pool of highly educated workers, even over the 1992-98 expansion. Despite the fact that nearly all college graduates looking for work had jobs in 1992, the 20 percent increase in the population of

Table 7: Distribution of Net Employment Growth of Population,
Ages 25 and Over, by Educational Status: 1992 to 1998, 1st Half

(Text Only)

Educational Group

1992 Employment

1998 (1st Half) Employment

Change in Employment

Percent of Net Growth in Employment

High School Dropouts

11,845

11,754

-91

-0.8

High School Graduates, No College

35,305

36,090

785

6.7

Some College

25,523

30,480

4,957

42.5

BA or Higher

27,273

32,487

6,005

51.5

Total

99,946

111,601

11,655

100

Source: Tabulations by Urban Institute based on data from the US Bureau of Labor Statistics.

college graduates, ages 25 and over, provided a substantial pool of new educated workers. While the 25 and over population has increased by seven percent since 1992, the numbers with at least some college jumped by 18 percent.

These surprising figures put to rest a "worker mix" explanation for limited wage growth. Had the mix of workers become less educated, this compositional factor would have exerted a downward impact on average wages. Put another way, increased wages among existing workers could have been offset by a rising share of low-wage workers. In fact, the opposite took place. The average educational level of the work force increased considerably over the period, thereby increasing average wage growth above the growth in wages among individual groups.

The rapid expansion in the supply of college-educated workers may explain why worker shortages have not become so widespread as to stimulate wage inflation. The long-term trend toward a rising demand for skilled and educated workers continued over the current expansion. In recent years, the structure of occupations has shifted dramatically toward high skill positions. Professional, managerial, and technical jobs account for nearly two-thirds of the net growth in employment, far above the 28 percent of jobs in these occupations in 1988. Despite the shift toward high skill occupations and the increased demand for skill within occupations, the substantial growth in the supply of the college-educated population apparently provided enough of an inflow to prevent the types of shortages one would expect at this stage of the business cycle.

Area Variation in Unemployment Rates and Wage Pressures

The national picture captures the average market conditions for the nation, but does not show the variation across labor markets in the degree of tightness and any induced pressures on wages. In Figure 3, we can see the wide variation in unemployment rates, ranging from about three percent in the ten states with the lowest unemployment to almost six percent for the ten with the highest unemployment. Fully half (25) the states had unemployment rates at four percent or below and only five states at about six percent or higher. Growth in employment is a second indicator of labor market tightness in an area. Figure 3 displays percentage employment change between the first quarter of 1995 and the first quarter of 1998, when employment in the nation rose by about 5.5 percent. It is striking that high employment gains did not necessarily go with low unemployment rates. In fact, the correlation between unemployment rates and percent growth in employment was slightly positive at .05.

The wide variations across states might provide indications as to whether inflationary wage pressures are finally emerging from tight labor markets. After all, fifteen states now are

{short description of image}

experiencing unemployment rates at about 3.5 percent or below. If migration is limited, one would expect to observe faster wage growth in these low unemployment rate areas than in high unemployment areas.

To test this possibility, we compiled data on nominal wage growth between 1995:I and 1998:I by state and calculated growth in average weekly and average hourly wages and salaries from the outgoing rotation samples of the relevant Current Population Surveys. Next we tabulated growth in earnings by states ranked on the basis of 1998 unemployment rates, 1995-1998 percentage change in employment, and 1995-1998 percent change in unemployment rates. The calculations grouped areas by quartiles of each labor market indicator. The unemployment rate groupings were those at 3.7 percent or below, 3.7-4.7 percent, 4.7-5.1 percent, and over 5.1 percent.

The well-known Phillips curve relates wage growth to levels of unemployment rates. Looking at this relationship across areas, one finds that wage growth was no higher in the lowest unemployment areas than in other areas. The numbers in Table 8 show no apparent relationship between rates of wage growth and area employment conditions. For example, the mean wage rates and mean weekly earnings were virtually identical across all four quartiles of unemployment rates.

Employer Responses to Tight Job Markets

How have employers kept wages in check in the face of these extremely tight labor markets? The answer is unclear. However, two sets of strategies appear to have emerged. The first is common to expansions and involves widening recruitment, expanding training, upgrading

Table 8: Relationship Between Changes in Wage Rates and Weekly Earnings and State Labor Market Conditions: 1995 to 1998

Quartiles of Unemployment Rates, 1998

Wage Levels

Below 3.7%

3.7% to 4.7%

4.7% to 5.1%

Above 5.1%

Percent Change in Hourly Wage Rates: 1995-1998

Mean

7

8

7

7

Median

13

13

13

9

25th percentile

10

8

8

8

75th percentile

8

8

8

8

Percent Change in Weekly Earnings: 1995 -1998

Mean

11

11

12

11

Median

11

13

11

7

25th percentile

8

13

15

9

75th percentile

11

9

9

8

Quartiles of Percentage Change in Employment: 1995 - 1998

 

Above 7%

5% to 7%

4% to 5%

Less than 4%

Percent Change in Hourly Wage Rates: 1995-1998

Mean

10

8

5

8

Median

13

13

9

11

25th percentile

8

11

8

13

75th percentile

9

8

6

7

Percent Change in Weekly Earnings: 1995 -1998

Mean

11

12

7

14

Median

10

10

6

14

25th percentile

10

13

11

15

75th percentile

12

5

5

13


Source: Urban Institute tabulations from Current Population Surveys, January - March 1995 and January - March 1998.

existing workers, and/or lowering normal hiring standards. The second involves the use of signing bonuses, stock options, profit-sharing, and other forms of non-wage compensation.

Types of Employer Responses

Employers engage in a number of strategies in response to a shortage of workers. Among the most common are:

  • To increase recruiting, by advertising more, turning more to employment agencies, reaching a wider geographic area, and even paying recruiting bonuses to employees;
  • To increase overtime work and turn part-time positions into full-time jobs;
  • To reduce education and other requirements for new hires;
  • To restructure work in ways that adapt to the available work force;
  • To substitute capital for labor;
  • To expand the supply of qualified workers by conducting additional training;
  • To improve working conditions;
  • To offer bonuses, stock options, and other forms of non-wage compensation to new and/or existing employees; and
  • To improve wages and fringe benefits.

Although employers generally turn to increases in wages and fringe benefits only as a last resort, a significant impact on wages unusually emerges by this point in the business cycle. A natural explanation is that employers are choosing to emphasize responses other than wage increases over the current expansion. While the evidence concerning non-wage responses is spotty, individual cases and limited data suggest employers are indeed adopting the strategy of emphasizing non-wage approaches.

Training

Employers are apparently increasing their training in a number of ways. Unfortunately, there are few consistent data sets documenting training practices over time. The 1994 National Employer Survey (NES) of over 4,000 employers, conducted by the US Bureau of the Census on behalf of the National Center on the Educational Quality of the Workforce, University of Pennsylvania, gives a detailed picture of training patterns and expectations for growth over time. One striking finding is that over two-thirds of employers reported that the skills required to perform production or support jobs increased over the prior three years. Over three out of four employers said they had increased training outlays over the prior three years, while only about three percent or less had decreased their amounts of training. Employers reporting rising skill requirements on production and support jobs were especially likely to have increased training; 85 percent of this group increased training as compared to 58 percent of employers who said skill requirements had not increased. In addition, the majority of employers projected a further increase in training.

Of the employers reporting an increase in training, over 80 percent cited changes in the work process, such as changes in technology or changes in the structure of work. Over 60 percent attributed the increases to product changes and 90 percent saw expanded training as a way of upgrading quality. In addition, nearly two-thirds of employers indicated that increased training was motivated by the fact that new hires did not have the necessary skills.

The 1994 NES reveals that several types of training are offered by large proportions of employers (see Table 6). Note that over three in four employers reimburse workers for tuition at colleges and training institute and that over 70 percent provide training in production equipment, in computer literacy, cross training, and teamwork. Most of these training areas are expected to grow over the next three years.

A Department of Labor survey (Frazis, et al., 1998) undertaken in 1995 showed that 70 percent of workers received some formal training in 1995 and virtually all (96 percent) spent time in informal training. Formal training is training that is planned in advance and has a structured format and a defined curriculum. Much of the formal, employer-sponsored training involved only a modest number of hours. Employees reported averaging only about 13 hours of formal training during a six month period and about 31 hours of informal training. Reports by employers showed an even lower number of hours. Still, the costs of training, counting wages and salaries paid to trainees, tuition reimbursements, wages of trainers, and payments to outside trainers, amounted to over $50 billion per year. The youngest (under age 25) and oldest workers (over age 54) experienced the least amount of training. Smaller firms provided somewhat less training, though few differences were observed between medium size establishments (100-499 employees) and large establishments (500 and over employees). Firms implementing four or more new workplace practices, such as pay for skills, employee involvement in technology decisions, job redesign, quality circles, and self-directed work teams, reported almost twice as much formal training as other firms. Formal training varies significantly among types of workers. More training reaches the high-paid, well-educated, full-time workers, workers in establishments with medium or low turnover, and workers with long tenure at the firm. For example, 90 percent of workers with a BA or more received formal training but only 60 percent of those with a high school degree or less did so. At the same time, average hours of training were higher among workers in the production, construction, and material handling occupations than among mangers.

These BLS data contrast sharply with data reported by the OECD from the International Literacy Survey. Their report suggests only about 23 percent of workers received any job or career-related training paid for by employers.

In any event, there is little indication that firms are providing depth in their formal training sufficient to raise significantly the capacities of less skilled workers. Formal training averages less than one week per year. Despite these limitations, some firms are increasingly emphasizing training, not only to improve the productivity of existing employeers but also to increase the supply of qualified workers in various fields.

Many companies have begun working with high schools to develop a new work force.

Charles Schwab is a good example of a company making an effort to shift from recruiting only workers with a BA degree to developing their own work force through a combination of work-based learning, work experience, and school-based learning. CISCO Systems is working with high schools to help young people qualify for jobs as computer network administrators. The auto industry has promoted a variety of programs to upgrade the quality of training for future auto mechanics. Other industries are working closely with career academies, which focus on industry or occupational fields in such areas as finance, travel, health, and computers.

Apparently, the involvement of employers goes beyond a few individual cases. According to the 1997 follow-up of the National Employer Survey of about 7,000 employers, an extraordinary 74 percent said they were participating in school-to-work partnerships and nearly 24 percent reported providing internships.

It will take time before many of these initiatives generate a substantial increase in the labor supply of skilled workers. However, in the interim, hiring and training young people as low-wage interns may allow many firms to limit the costs of expanding their work force.

Several important questions arise about training responses. First, to what extent do firms perceive special barriers that limit the amounts of training they sponsor? Some firms may hold back on extensive training efforts because of their concern that workers will leave the firm once they undergo training. Second, can we learn anything about the nature of worker shortages from the training undertaken by firms?

Non-wage Forms of Compensation

Employers are apparently expanding their use of special forms of compensation that do not involve direct salary increases. Despite the thriving economy, firms see themselves in a highly competitive environment, one in which raising prices can mean large losses of sales.

Anecdotal evidence suggest firms are turning to bonuses and variable compensation as a way of attracting workers in today’s tight labor market. According to Louis Uchitelle (1998), signing bonuses are proliferating and reaching well beyond upper level managers and skilled technicians. He cites examples such as the Labor Department’s offer of a $4,000 bonus to attract young economists and Price Waterhouse hiring bonus of $10,000 provided to newly hired management. Reportedly, the 1998 class of Cornell MBAs received an average bonus of $17,500, nearly double the 1996 average of $9,400. Increasingly, employers are willing to extend signing bonuses to other college graduates, including public school teachers.

A second expanding source of compensation is employee stock ownership and stock options. According the 1994 National Employer Survey, 35 percent of employers were offering stock options to their workers. By 1999, the figure is no doubt considerably higher. The National Center for Employee Ownership estimates that nearly 8 million workers participate in Employee Stock Ownership Plans or stock bonus plans and at least 6 million are in broad stock option plans. Changes over time in the compensation provided through these and other stock option or stock purchase plans are unknown. The Bureau of Labor Statistics does not try to measure the implicit income provided to workers who receive stock options.

One related trend in compensation is the area of nonproduction bonuses. According to Schwenk (1997), the proportion of compensation going to nonproduction bonuses nearly doubled, rising from about 0.7 percent in 1986 to 1.3 percent in 1996. As expected, year-to-year percentage changes in nonproduction bonuses are highly variable and presumably highly sensitive to the profits of firms paying such bonuses (Walker and Bergman, 1998). For example, since the 1991-92 period, percent changes in bonuses were 16 percent, 21 percent, 7 percent, 14 percent, and 5 percent.

The share of wages and salaries in total compensation seems to have declined only slightly from 73 percent in 1986 to 71.9 percent in 1996, if we take account of bonuses and benefit costs, including paid vacations, sick leave, health insurance, retirement plans, and social security and workers’ compensation. However, these data ignore the value of stock options and possibly other types of special payments.

Whatever the facts, should policymakers should promote compensation schemes linked to a firm’s performance? Might the apparent increases in bonuses and stock options help firms remain cost-conscious in today’s highly competitive environment? In the early 1980s, Martin Weitzman (1984) proposed "the share economy," one in which compensation would be based less on fixed wage contracts and based more on arrangements in which what workers received depended on the firm’s revenues or profits. Through the 1990s, Weitzman has continued to make the case for moving away from fixed wages and his work has stimulated a lively debate in the economics profession. The debate deals with issues involving microeconomics (relating to the impact on individual firm hiring and performance) and macroeconomics (relating to aggregate unemployment and inflation).

Research on the role of pay incentives to firm performance certainly predates the work by Weitzman. A substantial literature has developed around these issues and many of the findings suggest positive impacts on productivity and profitability from shifting compensation away from fixed wage contracts.

The focus here is on whether variable pay helps firms respond to booming macroeconomic conditions. Although Weitzman’s primary argument was that variable pay could induce hiring and reduce the 1980s problem of stagflation, he (1988) subsequently argued that the shift to variable pay could lower NAIRU, the unemployment rate consistent with non-increasing inflation. Two key elements of the argument are 1) that incentive pay arrangements encourage firms to increase hiring, and 2) that increased flexibility in pay will permit reduced variability in employment. In the case of pure wage contracts, firms hire to the point where the marginal cost (the wage) of the extra worker is equal to the marginal revenue obtained from the extra worker’s contribution to sales. In this case, if sales or prices fall and thus reduce marginal revenue below the wage costs, then the profit-making firm is likely to lay off workers until it reaches a new equilibrium. On the other hand, if workers were to receive a percentage of sales, then the firm would have no incentive to lay off marginal workers so long as they contributed some amount to revenues (even if it were less than the wage or the average revenue per worker).

Overall revenue to the firm would have declined but the firm would continue to offer employment to existing workers, since the revenues generated by the marginal workers would benefit the firm. However, average pay to the remaining workers would be lower with a no layoff policy than with the types of layoffs arising in firms with wage contracts.

One obvious implication is that workers might be able to gain increased employment stability but at the cost of increased variability of compensation. Workers confident in their ability to keep a job might well choose to avoid the risks of variable pay. On the other hand, workers vulnerable to layoff and workers optimistic about their firm’s future are likely to prefer compensation linked to their firm’s performance.

Notwithstanding the logic of the share economy, the advantages of an increased emphasis on profit-sharing or revenue-sharing might be illusory (John, 1991). In a profit-sharing firm, adverse shocks reduce profits and thus lower compensation per worker. Although firms have no direct incentive to lay off workers in this situation, they may be reluctant to allow average worker compensation to fall significantly. They may lose their best workers to competitors or may find that their jobs do not pay enough to deter workers from shirking. To avoid this scenario, firms may lay off workers even though the marginal workers do not generate any direct costs. Another possibility is that reduced compensation will lower the supply of labor and again induce reductions in employment in response to the shock to profits. Thus, while some labor market conditions suggest little gain from the shift away from fixed wage contracts, other scenarios indicate more employment stability.

In today’s context, the more important question may well be the impact of profit-sharing schemes on NAIRU. Here, Weitzman (1988) sees some potential gains for the share economy but concedes that some causes of a high natural rate of unemployment would apply as much in a profit-sharing context and as in a fixed wage context. Neither would have any advantage if high NAIRUs resulted from high social benefits, excessive bargaining power by unions, or efficiency wages. However, if the problem was that firm insiders were so powerful that external unemployment did little to affect wages, then bias toward expanding employment in profit-sharing firms would help speed up reductions in unemployment when the economy expanded. A world emphasizing profit-sharing might also reduce frictional and structural unemployment.

Surprisingly, Weitzman says little about the potential impact on inflation or on how uncertainty on the part of the firm will affect employment and prices in a low unemployment economy. Firms concerned about raising prices and uncertain about future sales may wish to avoid building high wages into their permanent cost structure. At the same time, they must recruit workers in the context of a tight labor market. The response by many firms is to try to make one-time payments, especially to newly hired workers, and to offer stock options and bonuses which will ultimately be linked to profits. To the extent firms can attract and retain the same quality of workers using these compensation schemes---perhaps because good workers are willing to take risks in return for high potential rewards---the package may make financial sense as a hedge for firms. Since payments to workers will be directly correlated with the firm’s performance, the losses due to weak firm performance will be mitigated by the lower compensation to workers while gains due to strong firm performance will be moderated by the higher compensation going to workers.

Already, the government encourages profit-sharing in a number of ways. There are special rules that provide tax advantages to workers and firms who create Employee Stock Ownership Plans (ESOPs). The tax treatment of stock options is also favorable. A worker receiving an option to buy his company’s stock at the current market price obtains something of value even when the option price is the current market price. Modern financial economics can place a value on these options. Yet, workers will not pay any tax on the options until and unless they exercise the options and sell the stock at a price higher than the option price. A key question is whether the government should do more to encourage these and other profit-sharing arrangements.