SOCIOECONOMIC POSITION

Social Gradient and Health

An examination of a social gradient for health is important for the older worker for at least five reasons. First, the nature of the work may play an important part directly in generating social inequalities in health. Second, for most people, before retirement the job is the principal determinant of income level and general standard of living. Third, the job is an important shaper of self-identity and a means through which personal growth and development are realized, or not. Fourth, occupation is an important criterion of social stratification. These are all potentially important for health. Fifth, the other side of having a job is not having one, or having insecure employment. This is relevant for health because of direct effects of unemployment and job insecurity and because these will be related to the wider aspects listed above.

There is by now a general agreement that health follows a social gradient. Studies from both the United Kingdom and the United States illustrate that the social gradient in health is not confined to poor health for those at the bottom of the social hierarchy and good health for those above a threshold of absolute deprivation. This is true for morbidity as for mortality. Figure 4-1 shows mortality rates from the original Whitehall Study of British Civil Servants, with men classified according to their employment grade, which is a precise guide to their position in the occupational hierarchy (Marmot and Shipley, 1996; van Rossum et al., 2000). Even at the oldest age, there is a social gradient in mortality.

FIGURE 4-1

All-cause mortality by grade of employment: Whitehall men, 25-year follow-up. SOURCE: Marmot and Shipley, 1996.

Figure 4-1 also shows that the occupational classification continues to predict mortality long after these men left the workforce. The relative differences in mortality are slightly less at ages 70–89, but the absolute differences are greater because overall mortality is higher. In addition, there is a substantial social gradient in morbidity that continues well into retirement. Figure 4-2 presents results from a follow-up of the original Whitehall cohort, 29 years after the original baseline examination. At the time of resurvey, two-thirds of participants were over age 75. Employment grade continues to predict mental and physical health and disability after retirement (Breeze et al., 2001). The Whitehall II study, in a cohort studied 20 years after the original Whitehall cohort, documented persisting social gradients in morbidity (Marmot et al., 1991, 1997).

FIGURE 4-2

Morbidity (percent) at resurvey by baseline grade, median age 77 (range 67–97): Whitehall men 1997–1998. SOURCE: Breeze et al., 2001.

Investigations as longstanding and robust as the Whitehall studies are not available from the United States. However, there is evidence to suggest a similar social gradient effect on health exists in the United States. Figure 4-3 presents the mortality gradient for older adults according to household income. The data are from the Panel Study of Income Dynamics (PSID), a household survey (McDonough et al., 1997). Those with poorest house-hold income had highest mortality rates, but for people at each level of income, mortality rates were higher than for those above them in the hierarchy. The PSID graph also shows that even after adjustment for education the relation between income and health remains.

FIGURE 4-3

Mortality PSID: United States, ages 45–64 years. SOURCE: Mcdonough et al.; AJPH, 1997.

A National Center for Health Statistics report (1988) with a special focus on health disparities shows that a gradient in life expectancy according to income exists at age 45 and at age 65. The gradient is steeper for men than for women. As a result, the male-female difference in life expectancy is greater for low-income people than it is for high-income people. It is true also for blacks and whites. There is also a racial disparity: at any given income category blacks have lower life expectancy than whites.

In addition, inequalities in mortality by measures of socioeconomic status (SES) have been increasing in the United States. One study used an area-based classification of socioeconomic level. It showed clear social gradients in all-cause mortality. The slope of the gradient increased between 1969 to 1970 and 1997 to 1998. For example, in 1969 the odds ratio between the bottom quintile of low SES and the top quintile for men was 1.4. From 1997 to 1998 this odds ratio had increased to 1.7. For women the odds ratio had increased from 1.3 to above 1.5 (Singh and Siahpush, 2002). Another study reports a comparison of two American Cancer Society cohorts, 1959–1996, and shows increasing mortality according to education for a range of diseases (Steenland, Henley, and Thun, 2002). Using the 1986 National Mortality Followback Survey and the 1986 National Health Interview Survey, Pappas et al. (1993) showed that disparity in mortality rates according to income and education had increased over a 26-year period.

The main markers that have been used to measure socioeconomic position—education, income, and occupation—are likely to relate to health and to one another in complex ways. For example, if education, income, and occupation are entered into a multivariate equation with health as the outcome and education drops out of the model as a predictor, this does not mean that education is unimportant. Indeed the lack of education may lead to ill health because people with low education are likely to end up in low-paying jobs, and hence to have low occupational status and low income. Conversely if occupation stays in and the others drop out, it does not mean that it is the job itself that is the cause of social inequalities in health. For the reasons set out above, occupation may be a measure of soecioeconomic position, in addition to being a guide to the nature of work. For economists, the unit of analysis is the individual. When economists speak of income inequalities they commonly use a metric such as the Gini coefficient, which measures the degree of unevenness of income among individuals. When public health people speak of health inequalities or health disparities, they most usually refer to inequalities between social groups, or racial/ethnic groups, which are probably closely related. These are quite different questions; the determinants of individual differences in illness or disorder may be different from the determinants of the differences between social groups and, more particularly the social gradient in health and illness or disorder.

In the Whitehall II study, low control at work and imbalance between efforts and rewards are both related to incidence of coronary heart disease and mental illness or disorder (Bosma et al., 1997, 1998; Stansfeld et al., 1999; North et al., 1996). These psychosocial work characteristics, particularly low control, make an important contribution to explaining the social gradient in these two illnesses or disorders (Marmot et al., 1997).

As people move from work to retirement the relative importance for health, and specifically for health inequalities, of work and nonwork factors will change. As noted above, resurvey of men from the original Whitehall study 29 years later showed that employment grade continues to predict mental and physical health and disability after retirement (Breeze et al., 2001). It is less likely that this is a direct effect of work than it is a result of the correlation between employment grade and other aspects of socioeconomic position.

The results of the Whitehall studies related to worker health are likely to be generalizable to other populations. For example, in a general population case-control study in the Czech Republic, low control in the work-place was strongly related to myocardial infarction risk and appeared to be an important mediator of the social gradient in chronic heart disease (Bobak et al., 1998).

The Whitehall studies related to retirement may be less generalizable, because they will relate to the circumstances of the British Civil Service and the nature of its pension scheme. Three sets of factors were related to early retirement. First, men and women in high grades were more likely to retire early. Qualitative interviews suggest that they had more options for an interesting postwork life. Second, those with worse health were more likely to retire early. Third, those with lower job satisfaction were more likely to leave early. In multivariate analysis, these three factors were independently related to early retirement (Mein et al., 2000). Material problems tended to keep people working.

The Role of Work

In the Whitehall studies, employees were classified according to their position in an occupational hierarchy. The Panel Study of Income Dynamics makes clear that classifying people by income or education would also reveal a social gradient in mortality. This raises the question of how important work is. Do men and women at different points in the occupational hierarchy have differing patterns of health and illness or disorder because of their occupation or because of other factors associated with their social position?

Work is a major source of income and life chances. This is of particular relevance to older workers. When people leave work, incomes generally decline, and so does access to social participation that is associated with the world of work. But this scenario is likely to differ according to occupational level. A high-status worker who leaves formal employment is more likely to pursue a portfolio career of paid and unpaid work than is a low-status worker. When one is employed, one's income is strongly related to position in the hierarchy. Illustrating this point, Table 4-1 shows that financial difficulties are clearly related to grade of employment (lower positions are denoted by higher category numbers): the lower the position, the greater the difficulty.

TABLE 4-1

Characteristics of Men and Women (Aged 35–55) by Grade of Employment in the Whitehall II Study (Age-Adjusted).

Work is a major definer of social and personal roles. The acquisition of appropriate education and development of appropriate skills is of great relevance to younger workers. For older workers the application of these skills may define the extent to which the workers see themselves as playing an important social role and are seen to be doing so by others. For an older worker, lack of work or change to tasks with less responsibility may influence mental and physical health. As Table 4-1 shows, people of higher status have more active social networks beyond the family. They are more likely to be married, to have hobbies, and to be physically active in leisure time. They are also less likely to experience stressful life events.

Occupation is one definer of social status. Research on health inequalities suggests that relative position in the hierarchy may be important for health in addition to the material conditions of life that go with that relative position.

Work is a major source of pain and pleasure, frustration and fulfillment, demands and rewards. These are in addition to any positive or negative effects of physical exposures in the workplace. Table 4-1 also illustrates differences by grade of employment in psychosocial work characteristics.

In sorting out the importance of work for the health of the older worker, we have to take into account both the wider set of influences to which people in different occupations are subject and differences that people bring with them to work. Table 4-1 shows that smoking is strongly related to position in the hierarchy. Similarly, there is a social gradient in hostility. Running through each of the ways that the wider aspects of work can affect health is the crucial distinction for older workers of working versus not working.

As discussed in Chapter 3, the nature of work has been changing. In 1900, about 1 in 6 of the U.S. workforce were in professional, managerial, clerical, and service occupations. In 1980 it was 6 in 10 (National Research Council, 1999). The changing nature of work has led to increasing polarization in work according to education. People with higher education tend to be in jobs requiring a higher level of skills compared to those with less education. There are cohort effects, such that older workers are somewhat less likely to have higher educational qualifications. Cross-sectional data from the Current Population Survey 2000 show that older people have had less experience of higher education. This means there will be some tendency for older workers to be underrepresented in jobs requiring higher education. This has implications for their experience of work. The Gantz Wiley Research WorkTrends™ Survey (National Research Council, 1999) examines attitudes toward work of a national sample of workers. In general, the higher the status of the job in which they are employed the more likely are workers to report high satisfaction, good use of skills and abilities, greater participation in decision making which affect work, opportunities to improve skills, and trust in management.

The MIDUS survey (Americans at Midlife) provided an opportunity to confirm these trends and examine interactions with age (Marmot et al., 1998; Grzywacz, personal communication to M.G. Marmot, 2002). MIDUS did indeed confirm the link between education and experience of work. The higher the level of education, the less likely people were to be classified as having low autonomy at work and low use of skills or variety. However, the relation between education and these characteristics of work did not differ according to age, though, interestingly, for a given level of education women had less autonomy on the job than men.

Another important way that age and social position can be related to work is in risk of unemployment. Job displacements are closely linked to educational and occupational status. The higher the education the lower the risk of job displacement defined as loss of job through plant closure, cessation of trading, and layoffs (Newman and Attewell, 1999). The age group at highest risk of job displacement is the youngest—new entrants to work—followed by the oldest workers. Four year retention rates by age are 29 percent (at age 16–24), 57 percent (age 25–39), 67 percent (age 40–54), and 45 percent (age 55 and older) (National Research Council, 1999).

GENDER

There have been significant changes in the gender composition of the older workforce (see Chapter 2). Among women between 55 and 64 years of age, labor force participation rates have increased steadily from 42 percent in the mid-1980s to 52 percent in 2000. According to the Bureau of Labor Statistics (BLS), this participation rate is expected to increase to 61 percent by 2015 (GAO, 2001). Similar increases are anticipated among women age 65 and older, whose labor participation rate was about 7 percent in the mid-1980s, 9 percent in 2000, and is expected to grow to 10 percent by 2015. The labor force participation rate of men over 55 is similarly expected to increase in the future, although this rate has remained relatively stable in recent years. In 2000, 67 percent of men aged 55 to 64 were in the labor force, and this rate is expected to increase to approximately 69 percent by 2015. Among men over 65, labor force participation rates in 2000 were at 17 percent, expected to rise to 20 percent by 2015. As these figures illustrate, despite the rapidly increasing labor force participation of older women, men are more likely than women to be in the labor force in their later years.

These patterns underscore the importance of taking gender into account as an important social determinant of the work experiences of and related health outcomes for older workers. Men and women differ in their earlier lifetime experiences and in the broader contexts of social and historical change that have shaped gender roles (Moen, 1996). It is necessary not only to analyze work experiences and health outcomes for gender, but also to understand gender as an organizing structural force, that is, to examine the role of gender as an independent variable influencing the work experiences and related health outcomes of older workers, as well as a factor in itself that needs examination (Moen, Robison, and Dempster-McClain, 1995; Umberson, Wortman, and Kessler, 1992; Walsh, Sorensen, and Leonard, 1995). Gender is important for the health and safety concerns of older workers for several reasons, including its influence on the types of jobs men and women hold, the resulting work-related exposures, the patterns of work over the life course, consequent income differentials, and differences in men's and women's experiences of retirement. Each of these is discussed below.

Gender influences the nature of work experiences. The sex segregation of the labor market has decreased in recent years but continues to structure the nature of work for men and women alike, although it is most notable in the types of jobs women hold. In 1980 about 80 percent of women were employed in the 20 leading occupations for women. By 1996, this proportion was reduced by about half, but those 20 occupations remain a considerable force influencing women's work experiences (Walstedt, 2000). As a consequence of the segregation of the labor market, men and women are exposed to different types of demands, strains, and hazards. Men are more likely than women to report hazardous work exposures. For example, approximately 39 percent of working men report that they have been exposed to substances at work that they believe were harmful if breathed or placed on the skin, compared to 23 percent of working women. Similar proportions of workers aged 45 to 64 reported these exposures (22 percent of women and 33 percent of men), although reported exposures were less common among workers over age 65 (Centers for Disease Control and Prevention, 1997). Women's occupations may, nonetheless, have unnoticed adverse health effects (Messing, 2000). An example is prolonged standing, common among such female-dominated occupations as bank tellers, grocery cashiers, restaurant workers, and sales clerks; prolonged standing may cause back, leg, and foot pain (Seifert, Messing, and Dumais, 1997). Women are reported to have higher rates of occupational musculoskeletal disorders than men; a large proportion of this difference may be attributable to ergonomic exposures (Punnett and Herbert, 1999; Zahm, 2000). Female-dominated jobs are also more likely than male-dominated occupations to be characterized by low pay, low levels of autonomy, low levels of authority and power, low levels of complexity and high levels of routinization, and responsibility for providing care and support for others (Bulan, Erickson, and Wharton, 1997; Marshall, 1997; Pugliesi, 1995; Ross and Mirowsky, 1992; Starrels, 1994; Wright et al., 1995).

Gender and age have intersecting influences on patterns of workforce participation. Although the percentage of employed persons working part-time increases with age, the proportion of women working part-time is consistently higher, across the working years, than that for men. Among workers over 55 years of age in 1997, 19.4 percent of men compared with 35.4 percent of women worked part-time (National Center for Health Statistics, 1998). Whether women are choosing to work part-time or are unable to find full-time work is unclear (Hill, 2002). Working part-time may have the salubrious result of reduced potential for hazardous exposures, or alternatively may place workers at risk of lower income and fewer work benefits.

Gender influences income, and hence the standard of living, as well as access to resources that may permit the choice of retirement or the necessity of working later in life. Women earn less than men. In 2000, women earned on average 76 percent of what men earned (U.S. Department of Labor, 2001). This earning differential is more pronounced among older workers. Among workers 55 to 64 years of age, the female-to-male earning ratio was 68.5 percent. In contrast, among those 25 to 34 years of age, women earned 81.9 percent of what men earned. Men earn more than women despite the fact that women have higher educational levels in similar occupations (Marini, 1980; McGuire and Reskin, 1993). Women are 70 percent more likely to spend their retirement in poverty than are men (Parsons, 1995). Women also are more likely than men to work at jobs that lack pension coverage (Richardson, 1999). Socioeconomic disadvantage and irregular career trajectories experienced by women in their middle years influence the availability of pensions and savings in later life (Moss, 2000).

Financial strain is a particular concern for older women. With age, women become increasingly at risk for poverty, reflecting the fact that women tend to earn less than men and that women are more likely than men to work at jobs without pensions (Richardson, 1999). Systematic inequalities in retirement policies further place women at a disadvantage; for example, women who divorce prior to 10 years of marriage are not eligible for dual Social Security entitlements. Women who receive Social Security benefits based on their own work records average $151 per month less than men (Logue, 1990; Richardson, 1999). While women comprise about 58 percent of the population over 65, they constitute about 75 percent of the elderly poor (U.S. Bureau of the Census, 1991). Among those aged 56–65, 27 percent of women and 17 percent of men have been poor at least once, and for women aged 66–75, that number rises to 35 percent (Duncan, 1996). Women from ethnic and racial minority groups are especially likely to be poor. For example, older African American women are twice as likely as older white women and five times as likely as older white men to be poor (Richardson, 1999). Risk of economic strain is also heightened among women who live alone or are widowed. Evidence indicates that widowhood is associated with approximately an 18 percent reduction in women's standard of living (Bound et al., 1991). Umberson et al. (1992) reported that financial strain was the primary variable accounting for higher levels of depression among widowed women relative to married women.

Not surprisingly, gender also plays an important role in retirement patterns. While the average age of retirement has declined in recent years among men, it has shown less of a decline in women. The traditional conceptualization of retirement is based on the presumption of a linear work path that results in a point of retirement representing the cessation of paid employment (Richardson, 1999). For many women who work intermittently or part-time in order to balance work and home responsibilities or who have few retirement benefits on which to rely, it may be necessary to either continue working or return to work. Indeed, retirement does not always signal the end of employment; one-third of older workers become reemployed after retirement, and this is especially true among women who have been intermittently employed before retirement (Han and Moen, 1998; Marshall and Clarke, 1998). Nonetheless, several researchers have concluded that although poverty may be a real threat, the personal and non-economic aspects of women's lives, such as family situations and previous labor force attachment, may be even more influential than economic factors in determining whether older women worked or not (Haider and Loughran, 2001; Hill, 2002; Honig, 1985). For women, there is a strong correlation between labor force participation early and later in life (Pienta, Burr, and Mutchler, 1994).

Men's and women's differing experiences of retirement are further shaped by socioeconomic position and race. For example, for low-income African Americans who lack private pensions and other sources of income during retirement, it may be necessary to work periodically in their later years (Gibson, 1991). Studies of retirement must therefore consider the qualitatively different experiences of differing subgroups as they face retirement. Understanding differences by gender and social class in the experience of retirement requires that we examine differing work histories, patterns of movement in and out of the workforce, shifting family responsibilities, occupation, and available financial resources for retirement (Moen et al., 1995).

The situation and experiences of older men and women today do not necessarily predict what future generations of older workers will experience. For example, many women in the current cohort of older workers either did not work outside the home or had discontinuous work patterns that excluded them from pensions or other retirement benefits. In recent years, women have experienced rapid changes in their status and roles. These changes may influence the very definition of work. As increasing numbers of women move into the labor force, work that has traditionally been in the private sphere without financial compensation, such as house-work, is increasingly being done for pay (Messing et al., 2000). Likewise, changes in the extent to which men and women share responsibilities for childcare and work in the home will have consequences for future generations of workers of both genders.

Future research on the health of older workers needs to control for gender and also needs to examine the roles that gender may play in the health of older workers. Men and women often work in different types of jobs, both in their later years and throughout their working lives, and consequently they experience differing exposures to work-related hazards. Their patterns of work differ across the life course, in part because women generally have greater household and caregiving responsibilities than men. Women consequently earn less money and are more likely to be economically disadvantaged than men in their later years. Understanding these variations in the work experiences of men and women across the life course provides an important departure point for planning future research and informing social policy.

RACE AND ETHNICITY

In contemporary American society, race is a key determinant of social identity and access to resources. Many minority older workers have been exposed to adverse social circumstances throughout their life courses, including deficits in education and health care during childhood and experiences of poverty, discrimination, and other forms of exclusion during adulthood. The research literature suggests that many of these challenges persist into old age and shape the opportunities and outcomes for minority elders. Although there has been little research specifically on occupational health and safety concerns for older minority workers, an understanding of factors that influence the general health of the minority elderly provides a useful point of departure and raises important questions for future research.

Traditional interest in minority elders has been dominated by studies of black and white differences. However, there is growing recognition that the minority elderly are racially and ethnically diverse, and that there are important intergroup and intragroup differences within these populations. In recent decades, there has been dramatic growth in both the number and proportion of older persons. The number of ethnic minorities is increasing at a faster rate than the white population. In the year 2000 non-Hispanic whites were the largest percentage of the total population of older persons, representing 84 percent of those older than 64 and 78 percent of those between ages 45 and 64. Their percentages are projected to decline by the year 2050 to 64 percent for those over age 64 and 55 percent for those between the ages of 45 and 64 (U.S. Bureau of the Census, 2000; Federal Interagency Forum, 2000). In 2000, 8 percent of those over the age of 64 were non-Hispanic black, 6 percent were Hispanic, 2 percent were Asian and Pacific Islander (API), and 0.4 percent were American Indian/Alaska Native. The corresponding percentages were higher for those in the 45–64 age range: 10 percent non-Hispanic blacks, 8 percent Hispanics, 4 percent APIs, and 1 percent American Indians.

The population of Hispanic elderly is growing particularly rapidly and is estimated to increase by the year 2050 to 16 percent of those over age 64 (21 percent of those between ages 45 and 64). In the year 2050, 12 percent of those over the age of 64 are projected to be non-Hispanic blacks, 6.5 percent APIs, and six-tenths of a percent American Indians. The percentages are predicted to be higher for those in the 45–64 age range: 15 percent non-Hispanic blacks, 10 percent APIs, and 1 percent American Indians. By 2050, over one-third of those over the age of 64 (and close to half of those between the ages of 45 and 64) will be black, Hispanic, or Asian.

There is considerable racial variation among minority elderly in years of formal education. For example, among persons aged 65 and older, almost six out of every ten blacks and seven out of every ten Hispanics have not completed high school; whites 65 years and older have rates of high school graduation that are more than twice that of Hispanic elders and 1.7 times that of blacks. Compared to whites, Asian American elders are overrepresented at both extremes of the educational distribution. The API elderly are more likely than whites to not have completed 12 years of education or to have a bachelor's degree or more (U.S. Bureau of the Census, 1996). The pattern for Hispanics reflects the impact of immigration, with large numbers of Latinos being raised outside of the United States in the context of lower educational opportunities compared to their U.S. born counterparts. The black-white differentials reflect the unequal educational opportunities and lack of investment in education for blacks that characterized U.S. society during the time period when today's black seniors were growing up.

Patterns of poverty also differ by race and ethnicity. During the latter half of the 20th century, there was a steady decline in the poverty rates among the aged of all races. At the same time, rates of poverty have remained relatively high among the elderly. One-fourth of all black elders, one-fifth of Latino elders, one-tenth of white elders, and one-eighth of API elders reside in households that fall below the federal poverty line (U.S. Bureau of the Census, 2001). The level of poverty for American Indian elders resembles that of blacks (John, 1996). Data on poverty tell only a part of the story of economic vulnerability, however, given the large number of persons who are only slightly above the poverty level. Data from the 2000 census show that combining the poor (annual income below the poverty threshold) and the near-poor (annual income above the poverty threshold but less than twice the poverty level), 30 percent of the American elderly are economically vulnerable (12 percent are below the poverty level). Among those over the age of 64, 35 percent of non-Hispanic whites, 56 percent of blacks, and 56 percent of Hispanics fall into this vulnerable group; among those between the ages of 45 and 64, the corresponding percentages are 14 for non-Hispanic whites, 35 for blacks, and 37 for Hispanics (U.S. Bureau of the Census, 2001).

Race and socioeconomic position are related but nonequivalent concepts. For example, although the rate of poverty is three times as high for the black compared to the white elderly, two-thirds of the black elderly are not poor, and two-thirds of all poor elderly are white. There are important variations within these categories. For example, although the overall rate of poverty among Hispanic elders was 22.5 percent in 1990, the rate for Puerto Ricans was 31.7 percent (Chen, 1995).

Beyond the issue of poverty per se, other large racial differences are apparent in income across elderly groups. The 1998 median income for elderly whites ($22,442) was 1.6 times that of elderly blacks ($13,936) (U.S. Bureau of the Census, 1999). There are also striking differences in the sources of income by race and ethnicity. In 1998, income from Social Security provided at least half of the total income for 63 percent of the beneficiaries (Social Security Administration, 2000). Minority elders depend more heavily on Social Security than their majority peers. For example, 33 percent of black and Hispanic and 30 percent of American Indian elders, compared to 16 percent of whites, depend on Social Security for all of their income (Hendley and Bilimoria, 1999).

Research is needed to understand the role of several key factors likely to influence the health of older minority workers. It is important to explore the role of acculturation and length of residence in the United States. Across a broad range of health status indicators, research suggests that foreign-born Hispanics have a better health profile than their counterparts born in the United States; for example, rates of cancer, high blood pressure, and psychiatric disorders increase with residence in the United States (Vega and Amaro, 1994). It is also important to clarify the intersections between race and socioeconomic position, given that the minority elderly are overrepresented among lower income groups. Research is needed to examine the effects of racism. This concern is especially notable for the African American elderly population. Although many groups have suffered and continue to experience prejudice and discrimination in the United States, blacks have always been at the bottom of the racial hierarchy and the social stigma associated with this group is probably greatest (Massey and Denton, 1993; Lieberson, 1980). It is also important to understand the role of specific work experiences. For persons over the age of 45, a higher proportion of white males and females participate in the labor force than their black and Hispanic counterparts, with one exception: the labor force participation rate for white men does not exceed that for Hispanic men in the 45 to 64 age range (Siegel, 1996; Fullerton, 1999). Research reveals that even after adjusting for education and work experience, employed blacks are more likely than their white counterparts to be exposed to occupational hazards and carcinogens (Williams and Collins, 1995). Additional research is needed to better understand how lifelong job-related exposures combine with specific work experiences in later life to affect the health and well-being of minority elders.

AGE DISCRIMINATION

Legal protections are provided to older workers to prevent discrimination on the basis of age (see Chapter 7 for a detailed discussion of laws pertaining to age discrimination). Nonetheless, a recent survey of older workers found that over two-thirds of workers over 45 years of age were concerned that age discrimination was a barrier to their advancement and well-being at work (American Association of Retired Persons, 2002). Evidence of age discrimination may be found in the length of time it takes to find employment, the wage loss experienced by many on reemployment, and the size of award as a result of reported discrimination (American Association of Retired Persons, 2002).

Inequalities in work opportunities, experiences, and health outcomes may be the consequence of discrimination on the basis of age, gender, race/ ethnicity, social class, sexual orientation, or disability (Krieger, 2000; Minkler and Estes, 1999). For many older workers, the accumulated effects of discrimination related to race or gender, for example, have persistent influences on work experiences, retirement patterns, and health outcomes in the later years (Dressel et al., 1997). Between-group comparisons that focus solely on aging, ignoring the intersections of other social determinants, are likely to mask the important roles of other factors (Dressel, 1988). For example, focusing on the effects of age discrimination for older African American women who have experienced poverty, racism, and sexism throughout their lives must account for the long-term consequences of social inequalities resulting from multiple forms of discrimination (Dressel et al., 1997; Dressel and Barnhill, 1994; Hill, 2002). Social inequalities on the basis of race, social class, gender, and age represent interlocking systems of inequality.

The theory of political economy of aging, which has been applied to the field of gerontology as “critical gerontology,” provides a useful lens for understanding age-based discrimination (Estes, 1999; Minkler and Estes, 1999). This perspective describes the experiences of older persons as socially and structurally produced through the distribution of material and political resources, as defined by social policy. Public policy reinforces the life chances of individuals based on social class, gender, race/ethnicity, as well as age. Accordingly, the status and resources of older persons are conditioned by their location within the social structure. The political economy approach has been criticized for its emphasis on structural disadvantage at the expense of a focus on individual agency and for a lack of attention to cultural change (Bury, 1995). Nonetheless, the perspectives offered by the theory of political economy highlight the important roles of social structure in shaping older workers' experiences, which might be balanced with the dynamic approaches provided by life course perspectives.

Discrimination is likely to be of particular concern for minority older workers due to the consequences of lifelong differences in opportunity. Among persons over 65 years of age in 1996, for example, 31 percent of whites, 57 percent of blacks, and 70 percent of Hispanics had less than a high school education (U.S. Bureau of the Census, 1996). These differentials reflect the historically unequal educational opportunities and lack of investment in education for blacks that was prevalent in the United States when these older persons were growing up (Williams and Wilson, 2001). The consequences of racism thus persist throughout a lifetime, and have clear implications for the work opportunities and retirement possibilities for older African Americans. Reflecting the dual discrimination of ageism and racism, a recent survey of older workers found that African Americans were more likely than other older workers to view ageism as a problem for older workers and were also most likely to report that their employers treated them worse than other workers because of their race (American Association of Retired Persons, 2002).

Discrimination may constrict work opportunities, influence overall economic well-being, and ultimately influence health outcomes. Research on the health effects of discrimination is a new but growing field of study. Krieger (2000) outlines five potential pathways whereby discrimination may influence health outcomes: economic and social deprivation, including residential and occupational segregation; increased exposures to toxic substances and hazardous conditions, resulting from residential or occupational environments; socially inflicted trauma, with consequent physiologic responses; targeted marketing of legal and illegal psychoactive substances, including marketing of pharmaceuticals to older persons; and inadequate health care. Although much research on these pathways has focused on effects of racial discrimination, age discrimination may follow similar pathways to influence the health of older workers and must be studied as well.

THE NATURE OF WORK

The preceding sections suggest that work, or its lack, plays an important part in people's lives and may have a profound effect on health. The nature of the work itself may also be crucial (see Chapter 6). As advanced industrial economies move increasingly away from heavy industry toward the service sector, the nature of work hazards changes. For office workers, some physical hazards of work may be less important, in terms of population-attributable risk of ill health, than are psychosocial hazards. Even for blue-collar workers, these loom large.

There is no strong suggestion from the literature that the relation between psychosocial working conditions and ill health differs by age. It has been reported from Sweden that control at work tends to increase with age, because of increasing seniority. This peaks at age 55. Thereafter degree of control may decline.

UNPAID WORK ROLES

The health and safety needs of older workers arise not only from their paid employment, but also from their unpaid work roles, including participation in volunteer work, caregiving responsibilities, and other household responsibilities. Volunteer work may provide important health benefits for older men and women. Moen, Dempster-McClain, and Williams (1992) found that volunteer work on and off through adulthood was positively associated with health; memberships in clubs or organizations were associated with women's longevity (Moen, Dempster-McClain, and Williams 1989; Moen et al., 1992). Preference, choice, and level of autonomy may be important elements in linking these roles to health (Moen et al., 1992). Volunteer work may likewise provide increased social ties, recognition, reduction in anxiety and self-preoccupation, and social support (Moen et al., 1992). Of course, it is possible that older workers who enjoy good health are more likely to engage in volunteer work in the first place, but even for already-healthy workers, volunteerism seems to hold potential for benefits.

Housework and caregiving responsibilities tend to be structured by the overall division of labor by gender. Multiple roles may take a toll on health while also offering potential health benefits. There are health risks associated with multiple roles. Employed women spend about 50 percent more time than men on domestic tasks (Canadian Advisory Council on the Status of Women, 1994). Several studies show that although the total amount of time spent on paid and unpaid work is comparable for men and women, women do more of the unpaid work characterized by low schedule control that is associated with psychological distress (Barnett and Baruch, 1987; Barnett and Shen, 1997). Psychological and physical health problems may additionally result from efforts to balance work and family (Moss, 2000). Women are more likely to experience role strain and overload as a consequence of family responsibilities in combination with work-related stress (Arber, 1991). These stressors may be either compounded or alleviated by material well-being (Arber, 1991).

Women, including daughters and daughters-in-law, are primarily responsible for providing care to elderly family members (Starrels et al., 1997; Walker, Pratt, and Eddy, 1995). About 55 percent of women between the ages of 45 and 59 with one parent living can expect to provide some level of care to a parent in the next 25 years, and with increases in life expectancy, this percentage has been estimated to increase to as high as 74 percent (Himes, 1994). Caregiving is associated with higher rates of depression and lower levels of self-rated health (Moen et al., 1992; Schulz, Visintainer, and Williamson, 1990). Caregiving may also pose particular strains for employed women. Among caregivers, women are more likely than men to miss work due to responsibilities in caring for an older family member (Anastas, Gibeau, and Larson, 1990). Among female caregivers, there is no difference in the amount of care provided between those employed and not employed (Stone and Short, 1990). To cope with elder care, working women may rely on rearranging schedules, job flexibility, and leave (Bird, 1997). Such flexibility and the resources available to provide this care clearly differ by socioeconomic position. For example, poor- and working-class women are more likely than their middle- and upper-class counterparts to provide hands-on care and less likely to function as a care manager (Archibold, 1983). Additionally, the cost of leaving the workforce to care for a family member is highest for those in low paying jobs with few fringe benefits, for whom the loss of a job signals further reductions in income and pension benefits (Sidel, 1996).

On the flip side, despite the risks posed by multiple roles, participation in a range of social roles may also provide resources that have been associated with older workers' health (Marshall, 2001). For example, in one study women in their 50s and 60s who were currently caregiving reported a higher sense of mastery than women not currently caregivers (Moen et al., 1995). Men and women with more roles tend to be in better health (Hopflinger, 1999), although the healthy worker selection effect may partially explain these findings. Women who successfully manage multiple roles over their life course seem to benefit in terms of increased confidence and self-esteem later in life (Moen et al., 1992; Thoits, 1995). By contrast, men who avoided household responsibilities in their younger years may experience reductions in instrumental activities of daily living. For example, a Swiss study of men and women over age 75 found that 5 percent of older women and 29 percent of older men were not able to prepare a meal (Stuck et al., 1995). Compared to older men, older women also have more social contacts and are more involved in neighborhood activities and family networks (Hopflinger, 1999). Occupying multiple social roles augments an individual's social network, power, prestige, resources, and emotional gratifications (Moen et al., 1992). Compared to men, women have more intimate relationships and receive more support from these relationships (Turner and Marino, 1994; Umberson et al., 1996). Social support provides an important buffer against the negative health effects of stress (Cohen, 1988).

WORK AND THE WIDER CONTEXT

The issue of the health of older workers has to be set in a socioeconomic context. Both the nature of work and the wider implications of work will be important for the health of older workers. The balance of gains and losses associated with work versus retirement will be influenced by wider social and economic forces. The number of lifetime hours in paid employment has been diminishing as the number of discretionary hours has been increasing; monetary income from paid work probably represents a minority of total benefits; and the egalitarian challenge for the future is equalization of spiritual resources (Fogel, 2000). These resources include self-fulfillment, family ties, social cohesion, and control over life circumstances. It is important to put work in this context. For some older people, work will be a source of these spiritual and psychosocial opportunities. For others the reverse will be the case. The health of older workers will be influenced by where the balance lies.