Printer Friendly

Employment characteristics of older women, 1987.

Early retirement by men and increased labor force participation by women in their mid-fifties have rapidly enlarged the female share of the older paid work force Since the late 1970's, the number of youths in the labor force has fallen by 2.4 million and labor force growth of women in the central age groups (25-44) has slowed considerably. Some employers have found it increasingly difficult to recruit workers and have turned to older persons (age 55 and over) to fill their hiring needs. As the "graying of America" progresses, older workers may become a greatly sought-after resource,

However, while the attractiveness of older persons as a source of labor may have increased, work force participation among those age 55 and over has, until the last few years, declined steadily. Understanding who older workers are, why they leave the labor force or continue to work, and what types of jobs they hold is critical for developing strategies aimed at increasing, or at least maintaining, their employment.1

Most past studies on older workers and retirement have largely ignored women, partly because the female share of the older paid work force has traditionally been small. In 1950, for example, only 2 of 10 workers age 55 and over were women. Since then, early retirement by men and increased labor force participation by women in their mid-fifties have expanded that proportion rapidly. As a result, it is no longer possible to ignore the labor force behavior of women age 55 and over, who now are 4 of every 10 older workers.

This article focuses on women age 55 and over who work. It first presents an overview of the group's work activity, occupational distribution, education, and earnings, and then discusses these characteristics as they vary according to marital status and race. The data for this study come primarily from the Current Population Survey (cps).sup.2 Although it is limited in longitudinal capability, the cps is probably the most comprehensive source of data on employment of older women. Where possible, other sources are used to supplement cps findings.

Work activity

Recent trends. While labor force participation of men age 55 and over has declined dramatically, that of older women as a group has remained remarkably stable. In 1987, 6.2 million women, or 2 of 10 age 55 or older, were in the labor force-about the same proportion as 20 years earlier. However, between 1967 and 1987, the labor force patterns of subgroups of older women varied. Women between the ages of 55 and 59 participated in the general increase in female labor force activity, although to a lesser extent than younger groups of women. (See table 1.) In contrast, the rates for women ages 60 to 61 were largely unchanged. Women in their mid to late sixties, like their male counterparts, responded to changes in both Social Security and private pension benefits by reducing their labor force participation..sup.3

As table 1 illustrates, participation rates decline rapidly with age. This occurs for two reasons. First, as women get older, many leave their jobs and withdraw from the labor force. This is especially common at age 62, when eligibility begins for reduced Social Security benefits, and at age 65, for full benefits. Second, the women in the 55and-over group who were born earliest participated in the labor market to a lesser extent than those born 10 or 15 years later. Thus, some of the age differences observed in the 1987 "snapshot" are not totally a result of aging; they also represent a cohort effect. Cohort effects may include differences in work patterns, attitudes, and education levels between groups of women born at different times.

Women who continue working beyond normal retirement age often reduce their work activity in terms of hours, weeks, or both..sup.4 Chart 1 illustrates a shift in work schedules with age. In 1986, women between ages 55 and 61 worked similar schedules to those of their central-age counterparts-more than half worked full time and year round. That proportion edged down for those between ages 62 and 64, and only a quarter of women age 65 or older had such schedules. Similarly, the proportion of women working the most abbreviated schedules (part time for less than 50 weeks) increased to 1 of 5 after age 62 and 2 of 5 after age 70.

Changing worklife patterns. Women born at different times have had strikingly different worklife patterns. The fact that women at or near retirement ages have dramatically lower participation rates than young women is only partly an aging effect. To a large extent, it reflects differences between cohorts of women. Many over the age of 50 in 1987, for example, first reached theage of labor force entry in the 1940's or 1950's, when women were not usually expected or encouraged to work and when employment opportunities for them were limited (an exception was during World War 11, but most women returned to their prior work patterns after the war).

Changes in social norms regarding women's work force participation -in subsequent decades affected young women. dramatically, and the effect has been lasting. As a result, women in their late thirties today are far more likely to be in the labor force than were their mothers or grandmothers at any age.

Chart 2 illustrates differences in the labor force participation patterns of 5-year cohorts of women who were between ages 30 and 69 in 1987. By linking together the participation rates of these groups as they passed through successive ages (for example, women ages 60 to 64 in 1987 were 45 to 49 in 1972, 50 to 54 in 1977, and so on) a longitudinal perspective is derived from crosssectional cps data.(5) As the chart illustrates, even between cohorts born as few as 5 years apart, differences in labor force participation are striking. At every age, each successive cohort was more likely to work than the preceding one. A comparison of the oldest and youngest groups of women illustrates the cumulative effect of that trend. For example, 7 of 10 women in their thirties worked in 1987, more than twice as many as in 1957.

Not only have participation rates risen with each successive cohort of women, but, as chart 2 shows, fundamental changes in worklife patterns have occurred over time. The older cohorts followed a pattern in which slightly fewer than half of women were in the work force in their early twenties; their participation rates dropped sharply during their childbearing years, and then rose to a peak in their late forties or early fifties. The peak rose higher and higher for each successive cohort, but the pattern remained essentially the same.

In the last decade or so, that pattern has apparently been replaced by one in which participation rates start much higher and continue to rise with no dropoff during childbearing years. The pattern for the two groups of women in their thirties makes it easy to envision future generations having sustained participation rates above 70 percent until they approach retirement age-a pattern reflected in the Bureau of Labor Statistics' projections to the year 2000.(6) In fact, although women's participation rates are lower than men's at each age, the current labor force patterns of young women resemble those of their fathers more closely than those of their mothers. 7

Jobs older women hold

Occupational segregation. Despite media attention focusing on women's movement into nontraditional occupations, a majority of women, both older and younger, work in stereotyped occupations. In 1987, for example, nearly two-thirds of working women age 55 and over and more than half of those between the ages of 25 and 34 were employed in three traditionally female job categories-sales, administrative support (including clerical), and services. (See table 2.)

Some reduction in occupational sex segregation has occurred in recent years, although new opportunities have accrued almost exclusively to young women and new labor force entrants. The limited employment options available in their youth largely determined the type of jobs held by women currently age 55 or older. This explains why only 1 of 10 lawyers and judges between ages 55 and 64 is a woman, compared to 3 of 10 between ages 25 and 34. Young women have also increased their representation in medicine, accounting, engineering, and management. However, this movement into nontraditional occupations has been somewhat overshadowed by strong employment growth in several female-dominated occupations, such as nursing and clerical jobs.(8)

Occupational representation. After age 55, the proportion of women employed in any occupation begins to reflect not only past opportunities, but also other factors. Earnings, retirement income, and opportunities for parttime work all vary between occupations and greatly affect whether women work and what types of jobs they hold. Also, health status determines, to some degree, which jobs older women may be able to perform. The representation index in table 2 provides a measure of the degree to which, among all women, older women are overrepresented or underrepresented in selected occupation and worker categories.(9) An index value of more than I indicates that an age group of women makes up a larger proportion of one occupation than of all occupations combined. A value of 1.16 in sales, for example, indicates that relative to their representation in all occupations, women age 65 and over are overrepresented in sales jobs by 16 percent. It is important to note that the index does not capture the effects of occupational sex segregation; that is, overrepresentation refers only to the unusually large presence of older women among all women in a particular occupation.

Within the 55-and-over group, sharp differences exist in occupational representation. Among other reasons, the jobs held by women in their late fifties are usually full time and year round, while those held by women age 65 or older often reflect a more marginal, near-retirement or postretirement pattern. In general, women in their late fifties are about evenly represented across job groups, while those age 65 and over are substantially overrepresented in some occupations and underrepresented in others.

As seen in table 2, women age 65 and older are overrepresented in both sales and services jobs. While the rigid hours required by many occupations force women to choose between full-time work and no work at all, those in sales and services jobs are often able to work part time for the same employer.(10) Some women switch jobs late in life, moving to these occupations from those that are not as flexible. Apparently, however, job switching is uncommon among older women as most are already working in these occupations." Also, changing jobs becomes less common with increasing age.

Another reason for older women's overrepresentation in sales and services jobs may be low rates of pension coverage in the industries that employ these workers. 1 2 Receipt of a pension other than Social Security greatly affects whether women continue to work beyond normal retirement age. In the Social Security Newly Entitled Beneficiary Survey (NBS), a survey of recipients of retiredworker benefits, it was found that women who did not receive a pension were about 3 times more likely to be working (18 to 30 months after first Social Security receipt) than those who received pension benefits.(13) In this context, it is not surprising that women age 65 and older hold a disproportionate share of both sales and services (especially private household) jobs, as women employed in these occupations may work to later ages than those employed in jobs with higher rates of pension coverage.

Health also plays a role in whether older women continue to work, and is one reason the oldest groups of women are underrepresented in some physically demanding occupations, such as operators, fabricators, and laborers. The relatively small proportion of women who have been employed in such jobs are likely to be in poorer health than are those who have performed less strenuous work. And, even when their health status is comparable to that of other older women, the greater physical demands of these jobs may make continued employment difficult. However, health status is less of a determinant of women's continued work activity than is pension receipt; in 1982, New Beneficiary Survey respondents who reported physical limitations and did not receive pension benefits were two times more likely to be working than persons with no health limitations who received pensions."

Older women are substantially overrepresented among both self-employed and unpaid family workers and are underrepresented among wage and salary workers. Women who are self-employed are often in services industry jobs and may continue workingbecause of low earnings and lack of pension coverage. 1 5Also, these jobs often provide flexible work schedules and make possible continued employment. The predominance of all three age groups of older women (shown in table 2) among unpaid family workers primarily reflects their employment in farming and other family businesses, (16)

Cohort differences not only affect the amount of work older women perform, but also have an impact on the representation of older women in many occupations. The most dramatic example is among private household workers. Limited employment opportunities available to women in the past, especially black women, channeled many into domestic work. In 1987, 8 percent of employed women age 65 or older were in private household jobs (33 percent of black women); this compared to less than I percent of women ages 25 to 34. Cohort effects are also seen in the underrepresentation of older women in executive, administrative, and managerial jobs, and among professionals and technicians. Although many of these jobs remain extremely segregated by sex, employment growth in these fields has created more opportunities for young women. Consequently, there is an overrepresentation of younger women and an underrepresentation of older women in these fields. Education

Education is one of the best predictors of older women's labor force patterns. It dramatically affects not only "whether they participate in the labor force, but also their likelihood of finding employment, of working full or part time, and of holding relatively high- or low-paying jobs.

In general, women age 55 or older have completed fewer years of education than their younger counterparts. As the following tabulation illustrates, in 1987, 4 of 10 women age 55 or older had completed less than 4 years of high school, compared with about I of 10 women between the ages of 25 and 34. Similarly, those in the younger group were more than twice as likely as were older women to have completed 4 years or more of college.

Percent distribution

of female Population

Years of school completed 25-34 55 and over

Less than 4 years of high school . . . . . . . 13.0 41.2

4 years of high school . . . . . . . . . . . . . . . . . . .41.8 38.0

1 -3 years of college . . . . . . . . . . . . . . . . . . . . .22.2 11.8

4 or more years of college . . . . . . . . . . . . . .22.9 9.1

Women who have completed the most years of education are substantially more likely to work than are their less-educated counterparts. Education not only increases the likelihood that one will find work, but it also expands opportunities for jobs which require minimal physical demands and provide high levels of satisfaction, making continued work both possible and desirable for some women.

Differences in labor force participation by years of school completed are illustrated in chart 3. As shown, high school completion dramatically increases the likelihood that older women will be in the labor force. Among women ages 55 to 64, nearly one-half of those who had completed 4 years of high school were in the labor force in 1987, compared with fewer than a third of those with less education. Similarly, about 1 of 10 high school graduates age 65 or older worked, compared to fewer than I of 20 women with less education.

Education is also a good predictor of the amount of work older women perform. Of women who were employed in 1986, those with the most education worked the fullest schedules. Sixty-one percent of college graduates ages 55 to 64 worked full time and year round, compared with only 46 percent of women with 8 years or less of education. Among women 65 and over, 28 percent of the most educated worked these full schedules, compared with a fifth of the least educated.


As women get older, their earnings decline in importance relative to other income sources, For those who continue to work, however, earnings remain a critical source of income. In 1984, for example, 75 percent of nonmarried women ages 62 to 64 and 31 percent of those age 65 or older who worked relied on their earnings for at least half of their total income. 17

Women in their late fifties and early sixties who worked in 1986 had median annual earnings of $11,141, while those age 65 and over earned $5,348. The lower earnings of the oldest group primarily reflected their more marginal work schedules. Earnings of women who worked full time and year round were substantially higher, with those 55 to 64 making $16,066, and those 65 or older, $13,217.

Although earnings depend on a number of factors, the best predictor is probably years of school completed. Table 3 compares the earnings of full-time, year-round workers of different education levels. As shown, older women with the most education earned between two and three times as much as their least-educated counterparts.

The cross-sectional data included in the table reveal different patterns for men and women. Within each educational group, men's earnings are higher for each successive age group-until they peak in the 45- to 54year interval. For women, the pattern is much less clear. In fact, earnings peak at ages 35 to 44 for the group as a whole, and at different ages depending on educational attainment. While men in their late fifties and early sixties earn substantially higher pay than younger men, older women often earn about the same amount or even less than their younger counterparts. This pattern-or lack of one-reflects a combination of factors, including variations between cohorts of women in years of work experience and in occupational mix. Older women, for example, are far more likely than younger women to be employed in low-paying private household work.

Marital status

Marital status has a substantial impact on the work activity of older women. As discussed previously, for women 55 and beyond, participation in the Labor force is partially a function of retirement resources. Married women, who benefit from their husband's income as well as their own, generally have the easiest time retiring or remaining out of the labor force. Many of these women stop working prior to or upon receipt of reduced Social Security benefits at age 62.(18) In contrast, a large proportion of never-married or divorced women rely solely on their own earnings, other income, or both for support. Hence, they must continue working beyond normal retirement age.

Labor force participation rates illustrate differences in the extent of work activity according to marital status. As shown in table 4, older unmarried women have a much stronger attachment to the work force than do married women. At the extremes, 70 percent of divorced women ages 55 to 59 were working in 1987, compared to only 45 percent of those with a husband present. After age 65, divorced women were about three times more likely to be in the labor force than were married women.

Differences in work activity between specific groups of older unmarried women are also striking, and reflect variations in resources available for retirement. For example, divorced women have participation rates far exceeding those of never-married or widowed women. This is not surprising, because divorced women are more likely to rely primarily on their own income for support, while widows generally receive survivor benefits. Also, in contrast to never-married women, many of whom have worked for most of their lives and can retire on their own pensions, divorced women may have started their careers late, accumulating limited pension resources. Data on occupational tenure indicate that half of never-married women age 65 or older have 25 or more years of experience in their current occupation, compared to just 30 percent of their divorced and widowed counterparts.(19) Occupational tenure affects pension eligibility, earnings, and level of Social Security benefits. Hence, divorced women are often worse off than their never-married counterparts and must continue to work to older ages.

It is interesting to note that marital status affects the labor force pattems of men and women in different ways. While being married tends to reduce the work activity of older. women, older married men have higher participation rates than their nonmarried counterparts. Also, except for those over 70, older divorced women are both more likely to be in the labor force and to work full schedules than are divorced men of the same age. (See table 4.)

Differences in the work activity of older women according to marital status have lessened dramatically in recent years. As the following tabulation illustrates, never-married women have reduced their labor force participation dramatically, while married women have increased theirs. Changes in pensions and Social Security benefits increased the ability of older never-married women to stay out of the labor force. Also, changes in cultural norms that encouraged work activity by married women of all ages reduced the differences dramatically.

Labor force

participation rates

1957 1987

Age 55-64:

Never-married............................... 69.2 48.4

Married, spouse present .................... 24.0 37.4

Age 65 and over:

Never-married ............................... 23.7 9.1

Married, spouse present .................... 6.6 6.9

It is important to emphasize again that the work patterns of women age 55 and older not only reflect their current resources, but also depend on past experiences. Many of the differences in work activity between older married and unmarried women should narrow even further for future groups. Marital status often determined the lifetime work experiences of women born in the 1920's and 1930's. For the most part, married women had a more tenuous attachment to the labor force than did their never-married counterparts. Relying on her husband's income, a married woman often remained out of the labor force, worked intermittently, or chose an abbreviated work schedule, while an unmarried woman typically worked full time for many years.

Today, a majority of women work regardless of marital status. The participation rates of never-married and married women ages 25 to 34 in 1957 differed by more than 50 percentage points. By 1987, that gap had narrowed to only 15 points.

Racial differences

Older black women have higher levels of work activity than older white women. Although the labor force participation rate of black women age 55 and older (24 percent) is only slightly higher than that of white women (21 percent), after controlling for educational differences, the greater work attachment of black women is evident. (See table 5.) At every educational level, black women are more likely to be in the labor force than are their white counterparts. Yet, the employment of older black women is concentrated in a relatively narrow range of lower paying occupations.

The more extensive work activity of older black women, in part, reflects differences in available retirement resources. New Beneficiary Survey data indicate that blacks (men and women together) are less likely than whites to receive pension income, to own their own homes, or to own other valuable assets.(20) Also, older black women are much less likely than older whites to be married; hence, black women have fewer resources for retirement and are likely to continue working to support themselves. In 1987, more than half of white women age 55 and older were married with a husband present, compared to only a third of older blacks.

Older black women have completed far fewer years of education than their white counterparts. In 1986, fully two-thirds of black women age 55 or older had completed fewer than 4 years of high school, compared to 39 percent of white women. Of those in the labor force, nearly half of older black women had not completed high school, compared with a fifth of white women. (See table 5.)

Reflecting these differences in education as well as differences in employment opportunities, in 1987, black women age 55 and older were about three times as likely as white women to be employed in service occupations. In fact, about 1 of 3 black women in their late sixties or older was employed as a private household worker; this compared to only 5 percent of white women in the same age group. Also, administrative support (including clerical) occupations employed about a third of older white women, but only 13 percent of older blacks. (See table 6.) In general, workers in services jobs have completed fewer years of school than those in clerical jobs. In fact, service workers are four times more likely than clerical workers to have less than a high school education.

The effect of education is especially clear when comparing the occupational distributions of groups of older black women. More than two-thirds of black women between ages 45 and 64," and nearly all of those age 65 or older who had completed fewer than 4 years of high school worked in the services occupations in 1987, while the majority of those who were college graduates were employed in professional specialties, mostly as teachers or nurses.

Vast differences in both educational and employment opportunities available to black women decades ago and today have resulted in very different occupational employment patterns for today's older and younger black women. The most outstanding difference between the youngest and oldest cohorts is the proportion employed in private household work. Fully 33 percent of black working women age 65 or older worked as cooks, servants, or cleaners (or in related occupations) in 1987, compared with only 1 percent of black women between ages 25 and 34. Young black women, like their white counterparts, were employed primarily in administrative support (including clerical) jobs. Although differences exist even between the youngest groups of black and white women, their relative narrowness is an indication that the occupational distributions of future generations of older women will vary much less by race than in the past.

Black women ages 55 to 64 working year round and full time in 1986 earned about 84 cents for each dollar earned by white women the same age. 21 Median earnings were $13,801 for black women ages 55 to 64 and $16,370 for white women. Sex differences in earnings were greater than race differences, however, as both groups of older women earned less than older black men ($17,556) and older white men ($28,165).

MANY OF THE LABOR FORCE EXPERIENCES of women age 55 and over in 1987 may never be repeated for new groups of older women. In the future, women reaching their late fifties and early sixties will have had substantially more work experience than their mothers or grandmothers. Differences in work history dependent on race and marital status will have lessened, and women's retirement decisions will be more similar to those of men.

However, the employment characteristics of older women in 1987 do provide some insight into what the experiences of later generations will be like. Though future groups of older women will have had many more years of work experience than women currently age 55 or older, their occupation and earnings profiles will undoubtedly continue to affect their work activity at age 55 and beyond.


1Recognizing the importance of issues related to older workers, Secretary of Labor Ann McLaughlin recently convened a Task Force on Older Workers.

2The Current Population Survey (cps) is the monthly household survey (including 59,500 households in 1987) conducted for the Bureau of Labor Statistics by the Bureau of the Census. The cps provides detailed information on the U.S. labor force, with some data going back as many as four decades.

3See Donald Bell and William Marclay, "Trends in retirement eligibility and pension benefits, 1974-83," Monthly Labor Review April 1987, pp. 18-25; and Employee Benefits in Medium and Large Firms, 1986, Bulletin 2281 (Bureau of Labor Statistics, 1987).

4Although from cross-sectional data it is impossible to determine the actual "normal retirement age," it is clear that the range of ages 62 to 65 often represents a turning point in work activity, as Social Security and private pension eligibility lure many workers into retirement.

5For an explanation of this cohort approach, see Philip L. Rones, "Using the cps to track retirement trends among older men," monthly Labor Review, February 1985, pp. 46-48.

6See Howard N Fullerton, Jr., "Labor force projections: 1986 to 2000," Monthly Labor Review, September 1987,pp.19-29.

7For a comparison of male and female participation patterns using cross-sectional labor force data for several years, see Susan E. Shank, "Women and the labor market: the link grows stronger," Monthly Labor Review, March 19 88, pp. 3 - 8.

8See Barbara F. Reskin and Heidi Hartmann, eds. Women s Work, Men's Work: Sex Segregation on the Job (Washington, National Academy Press, 1986), Chapter 2.

9 This technique was used in Philip L. Rones, "Older men-the choice between work and retirement," Monthly Labor Review, November 1978, pp. 3-10.

10For a detailed analysis of the demographic characteristics of older full-time and part-time workers, see William H. Crown, Phyllis H. Mutschler, and Thomas D. Leavitt, Beyond Retirement: Characteristics of Older Workers and the Implications for Employment Pol'cy (Waltham, MA, Heller School, Brandeis University, The Policy Center on Aging, 1987).

11Working paper by Donald C. Snyder, U.S. General Accounting Office, Human Resources Division, "Work After Retirement," p. 5 and table 3.

12For a comparison of pension coverage rates by industry, see Pensions: Worker Coverage and Retirement Income, 1984, Current Population Reports, Household Economic Studies, Series P - 70, No. 12 (Bureau of the Census, 1987), pp. 1 - 13. Also see Donald C. Snyder, "Pension Status of Recently Retired Workers on Their Longest Job: Findings From the New Beneficiary Survey," Social Security Bulletin, August, 1986, pp. 5-21.

13Howard M. lams, "Employment of Retired-Worker Women," Social Security Bulletin, March 1986, pp. 5 - 13.

14Iams, "Employment," p. 8. 15 See Eugene H. Becker, "Self-employed workers: an update to 1983," Monthly Labor Review, July 1984, pp. 14-18. See also Sheldon E. Haber, Enrique J. Lamas, and Jules H. Lichtenstein, "On their own: the self-employed and others in private business," Monthly Labor Review, May 1987, pp. 17-22.

16 See Patricia A. Unpaid family workers: long-term decline continues," Monthly Labor Review October 1982, pp. 3-5

17 See Susan Grad, Income of the Population 55 and Over, 1984, Publication No. 13 - 1 1871 (Social Security Administration, 1985), p. 81.

18 Iams, "Employment," p. 7. 19 Occupational tenure data are for 1987 and are available for three marital status groups: women who are 1) never-married, 2) married (spouse present), or 3) of other marital status. The third category includes both widowed and divorced women as well as married women with a spouse absent.

20 Working paper by Donald C. Snyder, U.S. General Accounting Office, Human Resources Div"Income and Assets of Recently Retired Workers by Race and Hispanic Origin."

21 Occupational data by educational attainment is not tabulated for 55to 64-year-olds specifically, so data for women ages 45 to 64 were used in this case.

22 After controlling for educational attainment, earnings ratios for older black women and white women range between 85 and 99 percent. The overall 84 percent ratio reflects the overrepresentation of black women at the lower end (and white women at the upper end) of the education and earnings scale.
COPYRIGHT 1988 U.S. Bureau of Labor Statistics
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1988 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Herz, Diane E.
Publication:Monthly Labor Review
Date:Sep 1, 1988
Previous Article:Labor Relations: Process and Outcomes.
Next Article:A year's work: labor force activity from a different perspective.

Related Articles
Employment and unemployment in the first half of 1988.
Labor market changes and adjustments: how do the U.S. and Japan compare?
Aging America: implications and impact on vocational rehabilitation.
Women and jobs in recoveries: 1970-93.
New data on multiple jobholding available from the CPS.
Gender differences in occupational employment.
The labor force participation of older women: retired? working? both?
Self-employment among older U.S. workers: the 1990s showed a downward trend in self-employment rates, however, the fact that self-employment rates...
Household survey indicators weaken in 2007.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters