Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Sociodemographic Aspects of Future Unpaid Productive Roles George C. Myers, Kenneth G. Manton, and Helena BacelIar In some respects the appraisal of forecasts puts a greater bur- den on the policymaker than the original task of forecasting itself. The accuracy of current forecasts is of course yet to be determined. Evaluation of the methodology of various forecasts may require technical sophistication at least as great as, and perhaps greater than, that of the specialist in forecasting. Yet the policymaker is rarely a specialist in forecasting techniques, nor usually an authority on the phenomena being projected (Ascher, 1978: 1-21. The purpose of this paper is to provide a background for consid- eration of the sociodemographic factors relating to unpaid pro- ductive roles in an aging society. The prospective nature of this task requires that use be made of projections or other analytic procedures that attempt to gauge the nature of the population structure in the decades ahead. This is a challenging task in itself, but it is made even more difficult by the task of determin- ing the relevant aspects of a concept as diffuse as unpaid produc- tive roles. The National Institute on Aging's Report of the National Research on Aging Planning Pane! (1982) identified four types of George C. Myers and Kenneth G. Manton are affiliated with the Center for Demographic Studies, Duke University, Durham, North Carolina. Helena Bacellar is a member of the Department of Sociology, Duke University, Durham, North Carolina. 110
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 111 activities that do not involve direct wage remuneration: (1) work activities contributed without payment to a family farm or busi- ness, or work undertaken on a do-it-yourself basis; (2) involve- ment in voluntary organizations civic, church, social, and so forth; (3) mutual help provided to (or by) family, friends, and neighbors; and (4) self-help that encompasses care of one's own person or one's immediate living space. The report emphasized that relatively little is known about the nature and value of such roles for individuals and for society, about the mechanisms that promote such activities, and about the obstacles and constraints that prevent such activities from being pursued on a larger scale. Although considerable research has been devoted to paid employment among older persons (for example, studies on the impact of relaxing mandatory retirement laws, delaying early retirement decisions, the effect of earnings tests), relatively little attention has been given to unpaid activities and the design of interventions that might alter the conditions affecting such activities. Thus, emphasis has been placed almost totally on the monetary benefits of activities among older persons. Virtually ignored has been the impact of such behavior on the general well- being of older persons and the impact it might have on the society at large (Rosow, 19761. Adopting a broadened perspective, these varied notions can be encapsulated under the general concept of "vintage capital," by which we mean anything that enhances a person's power to engage in useful activities (i.e., producing goods and services). The different types of unpaid activities reflect on the interplay between formal and informal roles and the relative status that may or may not adhere to such roles. Older age is generally viewed as a stage in the life course in which formal and perhaps informal role deficits occur. The most noticeable role change involves formal work roles, but other roles also may be lost, dropped, or modified. Unpaid activities can be viewed as alterna- tives to paid work and as adaptations to changing life conditions. The dimensions of such activities, however, are even more com- plex in nature. For example, in addition to the formal or informal nature of such roles, there is variation in the types of activities that might be pursued, for whom they are intended, the level of activity required, the reward structures (which may sometimes include partial remuneration), and the nature of the environ- ments in which the activities take place. So, too, attention must
112 GEORGE C. MYERS ET AL. be given to both suppliers (providers) and consumers (users) of such activities. (In the case of self-help, of course, these are the same person.) The activities also may include younger persons as well as older persons, although the main emphasis of this paper is on older persons. Finally, it should be noted that there are role requirements attached to being a user as well as to being a pro- vider. While it is often assumed that persons in need of care are willing to accept such attention, this may sometimes not be the case. Nor, for that matter, can it be assumed that persons fully capable of engaging in activities necessarily are motivated to do so. We have dwelled at some length on various conceptual issues because they are relevant in important ways to the main concern of this paper. The general demographic situation provides a con- text for determining the extent to which such activities are of societal importance (i.e., demand) as for example in determin- ing potential suppliers and consumers. The sociodemographic characteristics of these population aggregates obviously com- mand attention, especially in terms of changes that may occur over time. But the selection of which characteristics might fruit- fully be examined is difficult because so little is known empiri- cally about the phenomenon in question. Even in areas that have been researched, such as participation in voluntary associations, philanthropy, political involvement the correlates of activity and sociodemographic characteristics are sketchy and inconclu- sive, especially when the subject pertains to older persons. More- over, it is likely to be the case that the greater the empirical knowledge about these relationships, the more complex they will become with respect to the specificity of the activities in ques- tion, the interaction among the characteristics of providers and users, and the likelihood of intercohort patterns changing over time. ~ provide a first slice of this potentially rich pie, we have elected in this paper to present some information about sociode- mographic characteristics of the population that could possibly influence the levels of nonpaid productive activities in the future. We have relied upon existing national population projections and selected analytical studies in this effort. In so doing, issues are raised about both the relevance, technical suitability, and relia - bility of such projective exercises. The last section of this paper
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 113 devotes specific attention to a review of the current state of these activities. The task before us, then, is not to test any hypotheses but rather to examine a limited set of sociodemographic characteris- tics of the total population and older segments of the population. We suspect that such characteristics of the population as the age structure, sex composition, labor force participation, education, household structure and marital status, kinship structure, and health may be related to providers and users. In a real sense, we are actually only considering potential pools of such persons. However, by examining these sociodemographic characteristics of populations projected into the future, we learn not only about possible outcomes but about past trends, because most forecasts are based on known trends that are "projected" into the future. We can project certain features of the older population with some assurance, mainly because they are dependent on only one or a few parameters (numerical counts, age, sex composition); other features are more difficult to forecast (marital status, labor force participation, health). Thus, there is a conceptual leap being made in such an effort that exemplifies Ascher's comment at the opening of the paper. The difficult tasks lie not only in the act of forecasting but in interpreting the results and drawing issues of policy concern from them. PROJECTIONS OF MAJOR POPULATION CHARACTERISTICS Table 1, which is a detailed accounting of the age and sex distribution of the population, is included here to provide the most recent (1982) Bureau of the Census projections. While the projections are intended mainly for reference, they do enable us to see how the U.S. population will continue to grow in size over the period to 2050 and then stabilize. In addition, the figures reveal shifts in population structure. Females predominate over males, both in the aggregate and at ages over 25 currently, but at later ages in subsequent years. This change reflects on assump- tions that have been made about changes in mortality differen- tials by sex, and possibly migration, which rest in the technical details of the forecast. Finally, and most relevant for our pur- poses, the figures for the older population, which we refer to
114 GEORGE C. MYERS ET AL. generally in this paper as all persons 65 years of age and over, show the rapid growth of this subpopulation in numbers and as a proportion of the total population. From 11.4 percent of the population in 1980, it is expected to increase to 17.3 percent by the year 2020. The most recent projections reveal considerably greater growth in both the numbers and the proportions of older persons in the population than was true for the last "official" projections in 1977. This reflects mainly the modified forecasts of mortality reductions at all ages, including the older ages. Table 2 provides an overview of some relevant features of these projections of the older population to the year 2080. The number of older persons will increase at a fairly modest pace for the rest of this century and then will increase steadily until the baby-boom generation reaches age 65 in the period 2012 to 2025. By 2020 the aged population is projected to be 51.4 million, or about 17 percent of the total population. The age distribution of the aged population fluctuates, as might-be expected, depending mainly on the size of entering cohorts. These structural dynamics have direct relevance for the issues raised in this paper and probably have greater importance than the size of the aggregate older population per se. What is particularly note- worthy is the increased size of the very old group, 85 years of age and over, which by the year 2000 will constitute 14.1 percent of the total aged population. The sex ratio of this population group reveals an extremely high proportion of females over two females for each male although these ratios will become less extreme over time. A simple means of assessing structural shifts in population is by the use of ratios relating one age grouping to another; these are sometimes referred to as dependency ratios. The ratios of the so-called active population (20 to 64 years of age) to the total aged population decline sharply to the year 2020, decrease to the year 2080. continuing to Another ratio that is of interest relates the younger older per- sons to the very old (that is, persons aged 65 to 74 to those 85 and over). This ratio drops sharply to the year 2000; it then rises, but drops again to less than two younger older persons to a single very old person in 2050 and 2080. Finally, a ratio that is some- times called a familial aged dependency ratio, which relates older persons to the population 45 to 49 years of age, can be
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 115 calculated. This group might be thought of as the cohort of chil- dren related to persons reaching ages 65 to 69. These ratios also decline through the year 2000, rise sharply by 2020, and slowly continue to increase thereafter. The characteristics of the elderly population portrayed in Table 2 are commonly noted in overviews of the aged population, but their importance cannot be overemphasized. In considering both demand and supply issues relating to unpaid productive roles, these characteristics clearly show the profound aging process that is under way, the full impact of which will not be felt until the baby-boom generation reaches old age. But the san- guine attitude that sometimes prevails regarding trends for the next few decades up to that point of explosion in the next century encourages a vision of a "breathing spell" that is probably unwarranted. In fact, the aged population will grow steadily in the rest of this century while net additions to the older popula- tion slowly decline in number, although they are still positive, to the year 2000. After that point the net additions increase at ever higher levels to the year 2012, when the growth further acceler ates. The aged population will continue to become older itself during the next three decades. The proportion of females is expected to increase for the total aged population and among the very old. To the extent that providers, in our terms, tend to be younger elderly women, the trends may suggest a potential increase in their supply. On the other hand, the preponderance of women at extremely old ages will probably lead to an increase on the demand side at least for the next 30 years. Finally, we should note one major characteristic of the older population the high rate of turnover of individuals within the population. This turn- over means that only 40 percent of the persons who are members of the older population at any point in time (say, 1980) would have been members of the population 10 years earlier. This high turnover not only affects the numbers in the population but may markedly alter its social and economic structure, due to changing characteristics of new entrants (cohorts), selective survival of the earlier older population, and some modifications in status through behavioral changes that may occur over time. The population projections prepared by the Bureau of the Cen- sus have over time come to include an increasingly large set of alternative projections that reflect different assumptions made
6 GEORGE C. MYERS ET AL. TABLE 1 Estimates and Projections of the Population of the United States Including Armed Forces Overseas, by Age and Sex, 1980 to 2080 (Middle Series Projections) (in thousands) 1980 1990 2000 Age Total Male Female Total Male Female Total Male Female 227,658 110,834 116,824 249,657 121,518 128,139 267,955 130,491 137,464 0-416,4488,4138,03619,1989,8279,37117,6269,0228,604 5-916,5958,4828,10818,5919,5119,08018,7589,5999,159 10-1418,2279,3188,91916,7738,5868,20719,5199,9867,532 15-1921,12310,75810,36516,9688,6708,29918,9439,6819,262 20-2421,60510,90010,70518,5809,4339,13717,1458,7238,422 25-2919,7639,8769,88721,52210,87810,64517,3968,8048,592 30-3417,8248,8458,97922,00711,81410,99219,0199,5809,439 35-3914,1266,9669,16220,0019,93310,06821,75310,92510,828 40-4411,7525,7605,99217,8468,7999,04821,99010,94111,049 45-4911,0475,3725,67513,9806,8317,14819,7639,73910,024 50-5411,6845,6125,07511,4225,5195,90317,3568,4578,899 55-5911,6195,4836,13610,4334,9545,47913,2806,3636,917 60-6410,1344,6915,44310,6184,9175,70110,4874,9095,578 65-698,8053,9144,8919,9964,4585,5389,0964,1084,989 70-746,8432,8733,9708,0393,4054,6348,5813,6654,196 75-794,8151,8562,9596,2602,4363,8257,2952,8694,426 80-842,9721,0301,9424,0891,4202,6695,0231,7713,252 85-891,5414831,0582,1576421,5153,0259072,118 90-945751644118492116381,3553351,020 95-9913034962535519843890348 100+268185411421081890 0-47.2%7.6%6.9%7.7%8.1%7.3%6.6%6.9%6.3 5-97.37.66.97.47.87.17.07.46.7 10-148.08.47.66.77.16.47.37.76.9 15-199.39.78.96.87.16.57.17.46.7 20-249.59.89.27.47.87.16.46.76.1 25-298.78.98.58.69.08.36.56.76.3 30-347.88.07.78.89.18.67.17.36.9 35-396.26.36.18.08.27.98.18.47.9 40-445.25.25.17.17.27.18.28.48.0 45-494.84.84.85.65.65.67.47.57.3 50-545.15.15.24.64.54.66.56.56.5 55-595.14.95.24.24.14.35.04.95.0 60-644.44.24.64.34.04.43.93.84.1 65-693.93.54.24.03.74.33.43.13.6 70-743.02.63.43.22.83.63.22.83.6 75-792.11.72.52.52.03.02.72.23.2 80-841.30.91.71.61.22.11.91.42.4 85-890.70.40.90.90.51.21.10.71.5 90-940.30.10.40.30.20.50.50.30.7 95-990.10.00.10.11.40.20.20.00.3 100 +0.00.00.00.00.00.00.00.00.1 Total100.0 100.0100.0 100.0100.0 100.0100.0 100.0 100.0 SOURCE: U.S. Bureau of the Census. 1984. "Projections of the Population of the United States, by Age, Sex, and Race: 1983 to 2080." In Current Population Reports. Series P-25,
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 117 2020 2050 2080 Age Metal Male Female Ibtal Male Female lUtal Male Female 296,597 144,457 152,140 309,488 149,419 160,070 310,762 149,901 160,862 0-418,357 9,397 8,960 17,665 9,043 8,621 17,202 8,808 8,395 5-918,590 9,513 9,077 18,051 9,220 8,796 17,471 8,942 8,529 10-1418,306 9,366 8,939 18,217 8,322 8,895 17,747 9,083 8,644 15-1917,958 9,181 8,778 18,251 9,331 8,920 17,940 9,174 8,766 20-2418,308 9,324 8,984 18,381 9,362 9,019 18,103 9,222 8,881 25-2919,533 9,898 9,635 18,892 9,574 9,318 18,418 9,335 8,083 30-3420,301 10,252 10,049 18,491 9,844 9,647 18,819 9,506 9,313 35-3919,644 8,890 9,754 19,658 9,903 8,756 19,106 9,626 9,480 40-4417,699 8,874 8,826 18,186 9,635 9,552 19,116 9,604 9,513 45-4917,559 8,767 8,792 18,553 9,280 9,274 18,866 9,443 9,423 50-5418,621 9,230 9,391 18,439 9,169 9,270 18,568 9,243 9,325 55-5920,507 10,059 10,449 18,824 9,275 9,550 18,344 9,054 9,290 60-6419,791 9,495 10,296 18,503 8,985 9,518 17,970 8,754 9,216 65-6916,080 70-7413,325 75-798,824 80-845,662 85-893,582 90-942,158 95-99975 100 +361 7,7218,899 5,8537,381 3,6175,207 2,0783,585 1,1212,466 5551,063 206769 60301 16,6197,872 13,4956,133 11,4784,891 9,7853,795 7,8252,649 4,9151,405 2,261541 1,029191 8,747 16,914 7,363 14,984 6,587 12,659 5,990 10,305 5,179 7,977 3,510 5,433 1,720 2,946 838 1,870 8,059 8,855 6,880 8,105 5,486 7,172 4,091 6,213 2,800 5,178 1,650 3,783 766 2,181 372 1,498 0-46.2% 6.5% 5.9% 5.7~o 6.1% 5.4% 5.5% 5.9% 5.2 5-96.3 6.6 6.0 5.8 6.2 5.5 5.6 6.0 5.3 10-146.2 6.5 5.9 5.9 6.2 5.6 5.7 6.1 5.4 15-196.1 6.4 5.8 5.9 6.2 5.6 5.8 6.1 5.4 20-246.2 6.5 5.9 5.9 6.3 5.6 5.8 6.2 5.5 25-296.6 6.9 6.3 6.1 6.4 5.8 5.9 6.2 5.6 30-346.8 7.1 6.6 6.3 6.6 6.0 6.1 6.3 5.8 35-396.6 6.8 6.4 6.4 6.6 6.1 6.1 6.4 5.9 40-446.0 6.1 5.8 6.2 6.4 6.0 6.2 6.4 5.9 45-495.9 6.1 5.8 6.0 6.2 5.8 6.1 6.3 5.9 50-546.3 6.4 6.2 6.0 6.1 5.8 6.0 6.2 5.8 55-596.9 7.0 6.9 6.1 6.2 6.0 5.9 6.0 5.8 60-646.7 6.6 6.8 6.0 6.0 5.9 5.8 5.8 5.7 65-695.6 5.3 5.8 5.4 5.3 5.5 5.4 5.4 5.5 70-744.5 4.1 4.9 4.4 4.1 4.6 4.8 4.6 5.0 75-793.0 2.5 3.4 3.7 3.3 4.1 4.1 3.7 4.5 80-841.9 1.4 2.4 3.2 2.5 3.7 3.3 2.7 3.9 85-891.2 0.8 1.6 2.5 1.8 3.2 2.6 1.9 3.2 90-940.7 0.4 1.1 1.6 0.9 2.2 1.7 1.1 2.4 95-990.3 0.1 0.5 0.7 0.4 1.1 0.9 0.5 1.4 100+0.1 0.0 0.2 0.3 0.1 0.5 0.6 0.2 0.9 Metal100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 No.952. Washington, D.C.: U.S. Government Printing Office. Unpublished data were used for ages 85-100+ in 1980.
118 GEORGE C. MYERS ET AL. TABLE 2 Selected Statistics on Population 65 Years of Age and Over, United States 1980-2080 (Middle Series Projections) Year 1980 2000 2020 2050 2080 Ibtal population (in thousands)25,70834,91251,42267,407 73,089 Percentage of aged11.313.017.321.8 23.5 Ages-number (in thousands) 65-698,8059,09616,62016,619 16,914 70-746,8438,58113,23513,495 14,984 75-794,8157,2958,82411,478 12,659 80-842,9725,0235,6629,785 10,306 85-891,5413,0253,5877,825 7,977 90-945751,3552,1584,915 5,433 95-991304389752,261 2,946 100 +261083611,029 1,870 Ages-percentage100.0100.0100.0100.0 100.0 65-6934.226.032.324.7 23.1 70-7426.624.625.720.0 20.5 75-7918.720.917.217.0 17.3 80-8411.614.411.014.5 ~14.1 85-896.08.77.011.6 10.9 90-942.23.94.27.3 7.4 95-990.51.21.93.4 4.0 100+0.10.30.71.5 2.6 Sex ratio67.565.070.268.8 70.0 65-6980.082.386.890.0 91.0 70-7472.474.679.383.3 84.9 75-7962.764.869.574.3 76.5 80-8453.054.558.063.4 65.9 85-8945.742.845.551.1 54.1 90-9439.932.834.640.0 43.6 95-9935.125.926.831.5 35.1 100+44.420.019.922.8 24.8 Percent nonwhite of total aged9.310.913.819.1 23.5 Population structure ratios 20-64/201.792.112.352.36 2.38 20-64/65 +5.044.533.342.52 2.29 20-64/20-65 +1.321.441.381.22 1.17 65-74/85 +6.883.594.221.88 1.75 65-79/45-491.851.262.202.24 2.36 SOURCE: U.S. Bureau of the Census. 1984. "Projections of the Population of the United States, by Age, Sex, and Race: 1983 to 2080." In Current Population Reports. Series P-25, No. 952. Washington, D.C.: U.S. Government Printing Office. Unpub- lished data were used for ages 85-100 + in 1980.
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 119 about future levels of fertility, mortality, and migration. Whereas mortality assumptions may affect the numbers and percentage of the aged population over both the short and long terms, vari- ations in fertility and migration mainly operate in the long term on the numbers of older persons and on both time frames with respect to the relative proportions of the older population. For the 1984 projections, we can illustrate this effect on the projected numbers of older persons by examining 3 out of the 30 series produced: the lowest, middle (which has been presented earlier), and highest series. The lowest series reflects a low fertility assumption of 1.6 births per woman, a low net migration of 250,000 persons, and high mortality. The middle series reflects middle assumptions on fertility with 1.9 births per woman, a net migration of 450,000 persons per year, and middle mortality. Finally, the highest series reflects high fertility assumptions of 2.3 births per woman, a high net migration of 750,000 persons, and low mortality. On the other hand, with middle assumptions on fertility (1.9 births per woman) and net migration (450,000 persons), high mortality assumptions reflect a life expectancy at birth of 77.4 years in the year 2080. For the middle and Tow mortality assumptions, life expectancy at birth in the year 2080 rises from 81.0 to 85.9 years. Table 3 shows how these assump- tions affect the older population. In the middle and high series, there is continuous growth of the population 65 years of age and over through the 100-year period. In the lowest series the size of the aged population increases and then declines after 2040. If we consider the extreme values as providing certain levels of confidence about the middle series, then the range of possible error increases over time, with the difference between the low and high series reaching over 60 mil- lion by the year 2080. In terms of the impact of these extreme series on the proportion of older persons, the low series would produce 13.1 percent of the population aged in 2000 and 25.6 percent in 2080, while comparable figures for the high series would be 12.9 and 20.7 percent. Although we may fee] fairly confident that the middle series represents a set of reasonable (if perhaps somewhat Tow on the life expectancy improvement) assumptions, we also must recog- nize that the possibilities of being in error increase over time. Nonetheless, even under the least favorable assumptions, the numbers of older persons will more than double in size in the
120 GEORGE C. MYERS ET AL. TABLE 3 Variations in Alternative Projections of Population 65 Years of Age and Older, United States, 2000-2080 (in thousands) Number of Persons Year 65 and Over 2000 2020 2040 2060 2080 Lowest series Middle series Highest series Difference (high-low) Percentage of difference (low-middle) Percentage of difference (high-middle) 33,621 34,921 36,246 2,625 -3.7 47,139 51,422 56,332 9,193 58,116 66,988 78,558 20,442 -8.3 - 13.2 54,871 70,081 90,808 35,937 49,035 73,089 109,895 60,860 -21.7 -32.9 +3.8 +9.5 + 17.3 +29.6 +50.4 SOURCE: U.S. Bureau of the Census. 1984. "Projections of the Population of the United States, by Age, Sex, and Race: 1983 to 2080." In Current Population Reports. Series P-25, No. 952. Washington, D.C.: U.S. Government Printing Office. next 50 years, and the proportion of the aged population will closely approach 20 percent. The presentation of these figures does give ample evidence about the fragility of efforts to project populations and the importance of the basic assumptions that enter into their derivation. GEOGRAPHIC DISTRIBUTION In addition to the size of the older population and its composi- tion by age and sex, it is also important to assess current and future changes in its spatial distribution. Table 4 provides projec- tions to the year 2000, which unfortunately is as far as the cur- rent projections extend in time. In terms of total population, the projections reveal an increasing proportion of the U.S. population residing in the South and West and a declining share in the Northeast and in the North Central states. The trends for the proportion of older persons follow comparable patterns, although the Northeast will continue to have a proportionately greater share of older persons and the West will continue to be somewhat underrepresented by older persons. More than a third of the older population will continue to be found in the South. The lower panel of Table 4 provides greater age detail for the regions. The figures emphasize the considerable aging of the
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 121 aged population itself to the year 2000. The larger percentage gains in the population 85 years and over in the South and West are particularly noteworthy. The Northeast, which has had the highest proportions of these oldest persons, also will continue to show increasing numbers of the very old. These changing fea- tures of geographic distribution for different age groups provide a strong argument for greater attention to the distribution of programs and service support among the major regions of the country and undoubtedly among selected states within these regions. These dynamics of redistribution are often neglected in considering changing patterns of demand for the U.S. population as a whole. LABOR FORCE PARTICIPATION Forecasts of labor force participation up to the year 2000 have recently been prepared by the Bureau of Labor Statistics as part of general economic projections made to 1995. These labor force participation rates are then simply applied to projected census figures to determine the size of the labor force at future dates. The figures for 1982 in Table 5 provide an indication of current patterns in labor force participation. Rates for males, which are higher than those for females at all ages, peak at ages 35 to 44 and then decline sharply after age 55. In earlier estimates for 1982, it was shown that only slightly more than a quarter of men were in the labor force between ages 65 to 69, a level that has shown a precipitous decline in the past few decades. The partici- pation rates of females tend to follow the same pattern as men but always at lower levels. In fact, rates for women are less than half the levels for men at the older ages, which reflects earlier cohort differences in participation for the most part. The projections to the years 1990 and 1995 show that slight declines are expected in the rates of participation for men at most ages. However, a continuation of the rapid change for women toward greater involvement in the labor force up to age 60 is viewed as likely. Rates for males are still above those for women at all ages in 1990 and ~ 995. At ages over 60 the participation rates are expected to decline for males and to remain fairly stable for females. Thus, the work patterns that have emerged in the past few decades are forecasted to continue, with proportionately more women participating throughout the life course than was
122 on on on Cal or CQ o . - b~ o Cal be ¢ o En Cal - o Cal o A o a, o 5- v o C: sir o ~ o .- =~ o 8 l Go 8 o l o o o o 8 Cal o o Go o o (R CON Ct U) o o .- o ._ C~ _I o oo ~ CD ~ C~ 0 C~ ~ O ~ ~ C~ ~ ~ C~ C~ ~ . . . . . . . . . . . . . . . c~ c9 0 ~ r~ ~ co 1 co ~ cc ~ ~ ca ~ c~ c~ C~ ~ 00 ~ C~ C~ ~ ~ C~ C~ O CD . . . . . . . . . . . . . . . . ~ ~ 0 c~ oo 0 ~ ~ ~ r~ 0 ~ CD ~ C~ 00 1 1 ~ ~ ~ 1 ~ ~ ~ 1 O ~ ~ C~ C~ ~ 00 ~ ~ CD ~ O ~ 00 . . . . . . . . . . . . . . . . O ~ C~ ~ C~ ~ C~ ~ 1 ~/ c~ c~ ~ ~ cc e~ c~ ~ ~ c~ cc u: O ~ CO O ~ O 0 ~ c ~O O ~ C~ ~ C~ O O ~ C~ O O ~ C~ _I ~ C~ ~ ~ O O ~ ~ C~ CO . . . .. . . . . . N CC C~O ~ ~ O =' ~4 ~1 C~ C~ C~ O C~ r~ ~ ~ ~ 0 ~ ~ (D oo . . . .. . . . . . 00 0 ~ ~ ~ C~ O CO ~ O ~ C~ ~ ~ ~ O ~ O ~ ~O ~ C~ C~ 00 0 ~ ~ 00 U~ 00 O ~ CD ~ ~ O C~ ~ CO C~ O ~ CO a: ~ oo O C~ C~ ~ ~O C ~ O CO C~ ~1 C~ ~ ~ 00 O ~ C~ 14 ~ °O CO ~ 00 ~ d4 u: ~ co C~ C~ ~ ~ ~ O O C~ CO ~ C~ C~ ~ C~ U: ~ a) o^ ~ c~^ CO 0 U: C~ CO 00 ~ ~14 ~ O C~ C~ It~ CO ~ C~ oo ~ ~ oo ~ co c~ 0 00 ~ ~ ~O ~ ~ C~ ~ C~ ~ C~ CD ~ ~ 00 r~ C ~ C~ C~ O ~ ~ ~ (D 0 ~ O O C~ ~ O ~ C~ CO CC) ~ 00 ~ O O C~ O CD ~ ~ ~ ~ 0 00 CC) ~ C~ C~ ~ r~ ~ ~ ~ ~ c~ ~ ~ oo ~ ~ ~ c~ ~ O O CO ~ ~ ~ ~ ~ ~ 00 ~ ~ 00 It: ~1 00 C~ ~1 ~ O CC) ~ C~ 1= t~ 1- ~ a) C~ CO ~ 00 ~ ~U: ~00 ~U) 00 ~C ~C~ C~ ~ ~ ~ ~C ~C~ C~ O _ ~ ,= O S~ V] zo CO o ._ C~ + Ct O C) C~ O =0 0 ~ ~ Z + + CO ~ ~ CO UO: ~ ~^ o ._ Ct o ~ ~ ~ l ~ + v:) ~ o u: o u: . cs r~ ~ 0 ~
123 ~ co ~ ~ 0 ~ 0 ~ 0 cow ~ cc ~ cam CS O ~ rem rem ~ d4 ~ O Cal . . . . . . . . . . . . . . . . . . . . . . . . c~ ~ cc c~ ~ oo ca ~ r~ ~ ~ co 0 ~ ~ ~ ~ ~ ~ 0 co 0 ~ ~ c~ co ~ ~ ~ c~ co oo co c~ ~ r~ 0 ~ u: c~ ~ oo ~ co 1 1 oo ~ oo ~ 0 0 0 ~ c~ ~ c~ co oo r" c~ ~ r~ . . . . . . . . . . . . . . . . . . . . . . . . L~ ~ c~ ~ ~ co c~ co ~ ~ ~ 0 co ~ ~ c~ r~ ~ ~ c~ ~ ~ 1 ~ 1 ~ c ~ 1 ~ C ~ 1 ~ C~ C~ CD ~ 1 ~ c~ c~ ca o ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ C ~C~ o o · ~t · · · . ~ C~ . . . . . . . . . . . . . . · · ~ C~ C~ ~ · · oo ~ oo CC ~ ~ ~ oo ~ U: U) ~ C~ ~ C~ ~ ~ ~ ~ C~ ~ ~ ~ ~ ~ C~ ~ CO C~ C~ ~ ~ ~ ~ C~ C~ o o ~ U: . . . . . . o U: ~ o ~ U: o C~ o ~ ~ C~ oo . . . . . . o ~ ~ ~ C~ o CO C~ ~ ~ ~ o o ~ oo o o . . . . . . o ~ ~ oo o C~ C~ ~ U: oo o C~ ~ C~ o ~ o ~ 0 oo C~ CO ~ ~ o e~ ~ ~ ~ o c~ ~ ~ ~ co . . . . . . O ~ ~ O d. o c~ ~ ~ o 'c ~ co . . . . . . o o ~ ~ ~ ~ o ~ c~ ~ ~ ~ 0 c~ 0 0 c~ r" . . . . . . o ~ cs ~ c~ ~ o ~ c~ ~ ~ o co c~ oo oo co . . . . . . o cD u) o c~ ~ ~ o oo ~ ~ c~ o . . . . . . o ~ u: ~ c~ o o ~ c~ ~ ~ ~ o ~ u: oo ~ oo . . . . . . O u~ r~ oo 0 o co c~ ~ ~ r~ ~ ~ 0 ~ c~ cs ~ ~ ~ ~ o ~ 0 ~ o ~ ~ ~ cO ~ ~t lD CD ~ 00 00 00 ~ ~ ~ ~ ~ ~- ~ l - - - - - - o co c~ ~ o cD . . . . . . o co O N C~ C~ ~i O ~ U: C~ . . . . . . O C~ U: ~ C~ O O C~ C~ ~ ~1 O C~ ~ C~ C~ . . . . . . O ~ CD 00 O CO C~ ~ ~ ~ ~ O C~ ~ O 00 C~ CO C~ ~ O CO 00 ~ ~ ~ O c~ 0 ~ ~ ~ ~ c~ ~ CO ~ ~ ~ ~ C~ O ~ co r~ ~ co c~ 0 C~ 0 O 1O d4 ~ ~ ~ O O C~ C~ CS) CO C ~C~ C5) ~ C~ ~ ~1 1- C~ 00 0 <9 C~ 0 ~4 0 a) ~4 CC\ /53 C~ ~ ~1 0 00 ~4 ~1- cc c~ ~ ~r c~ ~ ~ ~ c~ ~ ~u: ~ ~ ~ C~ 00 U: ~ ~ C~ ~ O ~ 0 0 ~ ~ O O ~ ~ ~ 0 ~ ~ ~ C~ O C~ CD ~ C~ C~ -/ a) C~ ~ CO _l ~ 0 0 e~ ~ _~ ~ ~ ~ c~ 0 0 0 0 0 ~ ~ r~ ~co c ~0 co ~ ~ c~ ~ ~ co c~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~_ _ _ _ CC C~ ~ ~CO C~ ~ ~0 C~ C~ CO - CQ ~ ~S: ~14 ~ ~ _ C~ ~ ~ 0 1 1 1 1 + ~, ~ O ~ O ~ O ~ ~ ~ 0 0 CO - Ct s" V ~ ~ ~ ~ _4 co r~ r~ 0 1 1 1 1 + ~, ~ O U) O L~ O tS) 1!~ ~ 0 0 ~+ CD ~5) ~ ~ ~4 ~ CD ~ ~ ~ =- I I 1 01 + ~ Ca I 1 01 + U: O U) O ~CR U~ O ~ O ~ o CD ~ ~ 0 0 ~ CO r~ ~ oo 0 U] ~ 00 o0 .. X U2 cd b~ ct ~ u) o ~ bC ~ · - ~o ~ ·-o ~ ~ o ~ - o v: ~ .. ~ v ~o ~ o o ·~ ~} - ~ (R . .5 o o &m . 0 ~ 3, CQ U) V ~ ~: ~o C: a, V] s~ .. V ~ =^. o o V2 C~
124 GEORGE C. MYERS ET AL. TABLE 5 Labor Force Participation Rates by Age and Sex, United States, 1982-1995 (Middle Series Projections) Age 1982 1990 1995 Males 16-19 56.37 62.3 62.9 20-24 84.9 84.4 84.1 25-29 ~94 7 92.9 92.3 30-34 ~· 94.5 93.9 35-39 ~96.2 95.8 40-44 ~95.3 94.8 45-49 t 912 93.2 50-54 S · 88.7 88.4 55-59 ~702 78.1 76.7 60-64 ~· 52.8 50.5 65-69 ~23.3 21.2 70-74 ~17.8 15.0 13.8 75 + J 6.9 6.3 Total: 16-75+ 76.6 76.5 76.1 Females 16-19 51.4 56.8 58.2 20-24 69.8 78.1 82.0 25-29 t 68 0 78.5 82.2 30-34 S · 77.8 81.3 35-391 68 0 78.0 82.0 40-44 S · 79.2 83.5 45-49 t 61 6 69.2 71.3 50-54 ~· 64.5 67.3 55-59 ~41 8 51.1 51.9 60-64 ~· 32.1 32.1 65-69) 15.1 15.2 70-74 ~7.9 7.2 7.1 75 + J 2.2 2.0 Total: 16 -75 + 52.6 58.3 60.3 SOURCE: Bureau of Labor Statistics. 1984. Employment Project tions for 1995. Bulletin 2197, Appendix A-1. Washington, D.C.: U.S. Government Printing Office. true previously. Proportionately fewer older mates and about the same proportion of older females are expected to be in the labor force. The National Institute on Aging (1984) has developed longer- term projections to the year 2055 from its macroeconomic- demographic model, projections that permit an examination of the size of the labor force in broad age categories (Table 6~.
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 125 Although the labor force as a whole is expected to increase by over 28 million workers by the year 2000, the number of older persons in the labor force is projected to decline, notably among men. Although the number of females in the labor force will continue at fairly stable levels until the middle of the next cen- tury, the number of older males after 2000 is expected to increase. This is due largely to the increase in their total num- bers, not to the proportions employed. What is striking about these projections, which of course represent only one possible scenario, is the very modest impact that employment of older persons has upon the labor force as a whole. From a supply point of view, the pool of women younger than 60 years of age and potentially available for service to others is expected to decline to the year 2000. On the other hand, both TABLE 6 Projected Labor Force Participation by Age and Sex 1980-2055 (in millions) 1980 2000 2020 2050 2055 Male 16-24 12.1 11.1 11.2 12.8 13.4 25-64 44.4 55.9 56.4 57.0 57.3 65 + 1.9 1.8 22.9 3.4 3.9 Total 58.4 68.8 90.5 73.2 74.6 Female 16-24 9.7 10.8 11.5 12.6 13.2 25-64 28.2 45.1 52.9 57.2 58.2 65+ 1.1 1.1 1.1 1.2 1.4 Total 39.0 57.0 65.5 71.0 72.8 Rate of Growth (average annual percentage) 1980-2020 2020-2055 1980-2055 Male 16-24 -0.06 0.36 0.15 25-64 0.60 0.04 0.34 65 + 1.06 0.85 0.96 Total 0.49 0.14 0.32 Female 16-24 0.42 1.39 0.41 25-64 1.57 0.27 0.97 65 + 0.00 0.69 0.32 Total 1.30 0.30 0.83 SOURCE: National Institute on Aging. 1984. Macroeconomic-Demographic Mode! Table 1-7. Washington, D.C.: U.S. Government Printing Office.
126 GEORGE C. MYERS ET AL. women and men out of the labor force at older ages will be numerically and proportionately greater. The effects of various changes in retirement behavior, coupled with and generated by changes in federal legislation (especially related to Social Security) on work levels, have been examined in projection studies by the Urban Institute (Wertheimer and Zediewski, 1980; Hendricks and Storey, 1981) and ICE Incorpo- rated (Anderson and McNaught, 1982~. They permit us to see how such enactments, some of which are already in place, might influence labor force participation. Projections made by the Urban Institute (Wertheimer and Zedlewski, 1980) utilized the microsimulation mode! DYNASIM to examine 1990 labor force participation rates under a high and a Tow scenario. Essentially, the two series differ in terms of early or late departure from the labor force. In Table 7, work status and annual hours worked are examined for males and females in different age groups. Under the Tow series, which relates closely to present conditions, there is a clear decline in labor force partic- ipation and corresponding decreases in mean hours worked for males at all ages and for females at age intervals 65 years and over. The high series effects mainly occur at ages 62 to 64 for mates, especially in the amount of full-time employment. At 65 years and over the levels are only slightly elevated for mean hours worked by workers, but the proportions of individuals who have retired are roughly comparable to levels in 1977. The same impact is found for females. Another application of the microsimulation mode! by Hen- dricks and Storey (1981) has examined different assumptions regarding the effect of mandatory retirement policies on labor force participation rates of men in the year 2000. Inasmuch as some of these changes have already been implemented, the results are interesting to examine in terms of their effect on working levels through the ages 60 to 70. Under the existing change that raised the minimum mandatory retirement age to 70, the increase in mate labor force participation was projected as minimal in the year 2000 for the 60 to 61, 62 to 64, and 68 to 70 age groups, but the change had a somewhat larger effect at ages 65 to 67. Eliminating mandatory retirement altogether barely increases any of the rates up to age 67, but it has a modest effect at ages 6S to 70. The expected net effect of most of the recent changes in policies
127 on rat - Cal ¢ - U: JO Cal ¢ Cal AIR V) ~ _ o o 5- ._ ~ · 5-1 0 cn ~ ~ .O 8 ~ sly o Cal ._ ~ _' JO ~ rat ~ _ c53 3 rat ox o 3 U' o ~4 o ~ CO Go CD . . . . . ~ lo d. N Cal 00 Cal rat Do 0 0 rat 0 O ~1 00~ 1 1 1 1 1 'D 1 co 1 1 1 1 1en 1 0 ~ rat rat ~ cc ~0 CDrat 0 Cal ~ ~ ~ 0 ~a:) 1 1 1 1 1 ~1 CO Cal ~ ~ Cal ~O ~ CS) ~ ~ ~ ~U: ~ . ~ ~ ~=4 As ~ Ed ~ O ~Go CC Cal ~ 00 ~ ~CO Cal ~ Can Cal ~ 00 Cal of (~) ~ Cal Cal O O ~ 00 ~ ~ Cal O ~ 00 00 0 CO Lo00 0 Cal o (3,) ~ Cal Cal O O ~oo Cal O ~ It: CO a) ~1> Go ~ _ 0 ~ =4_ or Co 00 ~ 00 ~64 Cal ~ ~ ~ ~ ~ ~Cal O Cal I CC Cal CO 1~ ~ Cal ~ of) CCHI Cal (31 Cal ~4 iO C70 _ ~4 0 ~ Co =4 ~ ~0 Cal ~ U: ~ ~ ~CO CS ~ Cal ~ CD Cal ~Lo Ct ~ (c 0 Ct ~ t£) Cal -0 0 433 ~4 Co ~4 Cal ~ -A0 ~0 0 0 U: ~ ~ Cal ~ ~ ~ ~ ~ CS ~O Cal m ~CD co ~ d4 CS ~ O ~ ~Cal ~ ~ CS CC ~ Od4 ~ (53 0 ~Cal U: ~ 0 0 _ ~ Cal Cal ~ CO en ~Cal ~ 0 _ CC CC_1 O ~ 0 ~ ~ ~O ~C,~ ~ To ~ g ~ ~, or ~ ~ ~so o lo lo o or ~- ~of o ~ A - C o 3 o o cat me ~ o i`: 3 o:> ~ ~ O O Sly ~ 0 3 ~ S ~ ~ O O ~ ~ ~ ~ ~ _ ._ CO 0 _ O · ,~3 sit _3 . .~ sit TV .~ O 0 U) 3 - U: - ~i - 5_,- ~ ._ s=- .. - V ~ ~ lo. ~ V, U: ~
128 so - o -v Cal rib Cal o AIR o - - by o a) Cal Ct A ¢ s; -v o - v ¢ sit - C~ CQ o s~ C~ o o C~ ~_ - E-°' P: b~ C) - C: C: Ct 3 o C~ o Ct P: o o C~ o oo o o C~ o o N o C~ o o , - o o C~ o C~ o C~ o 00 C~ ~ ~ . . . C~ 00 U: . . . ~ O O C~ ~ 00 ~ ~ CO . . . C~ ~ C~ L~ . . . O ~ 00 a) 6g C C~ . . . O CD 1~ C~ 6g C~^ O . . . C~ 00 00 0O 6g,. CD . . . O O C~ _i ~ ^ _ ~ S~ O U) ·: V O Q _b ~0 .~ t4 .6 O sQ~ $ b~ ·_. O U' _ ~ C' ~ ~ E E ~ o o ~
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 129 on projected male labor force participation shows only modest increases in participation, which would of course reverse the trend toward earlier retirement. Similar trends are found for females as well. Of greater importance has been the total retire- ment income prospects of individuals, which include growing pri- vate pension support. Projections of total income by persons over 66 years of age have been estimated by the National Institute on Aging model to the year 2060. Table ~ shows a projected increasing level of total income for older persons and an increasing proportion of this income compared with total GNP income for the country. This is particularly true up to the year 2020, with a lowered rate of growth thereafter to the middle of the next century. In terms of income per capita for older persons, there is an increasing level throughout the projection period but a progressively lower rate of increase. Judging from the previous analysis of levels of employ- ment, it is not expected that this income will come from wages but from other sources. Nonetheless, this evidence suggests an increasingly more affluent older population and the importance of its income as a share of the income share of overall GNP. Both demographic and economic factors appear to be operating in these projections to gradually retard growth in the relative finan- cial well-being of older persons, but the outlook is still generally favorable. This examination of projected labor force participation and income expectations lends weight to the view that paid work among older persons is not likely to increase significantly in the future, thereby supporting the idea that other unpaid activities can be an alternative condition for sizable numbers in the grow- ing older population. To the extent that higher proportions of women stay in the labor force until retirement, then the poten- tial pool is diminished; but for those turning 66 the prospects are not diminished. Of course, recent changes in Social Security laws raising the entitlement age for future cohorts will have a long- term effect when and if that reform is introduced. Over the short term, however, the effect is likely to be minimal. EDUCATION Educational attainment has been widely viewed as a personal characteristic that is positively related to high levels of formal
130 U: Cal be ¢ A= . ¢ o . - Cal .~ CO Y ~ + Ct be - o V _ o no V] be CQ sit C: o U] 5- C: a' us: CQ Go U' o i' l co 0 rib ~ ~ d ~ rib . . . . . . . . . . Cal ~ Cal ~ Cal o ~ Go 0 0 o ~ CO CO Cal Cal o . . . . . . . . . . 0 ~ U: ~ ~ Cal o ~ ~ 0 . . . . . . . . . . lo ~ ~ O ~ 0 CD Lo U: ~ o 0 Cab ~ ~ ~ o CO o U: . . . . . . . . . . Cal CO ~ ~ Cal ~ ~ 0 ~ Cal Cal ~ ~ ~ ~ ~ ~ ~ CO 0 CO ~ 0 CS ~ ~ CO . . . . . . . . . . ~ red ~ ~ con car Cal ~ CO Cal Cal CO ~ Cal Cal ~ ~ O ~ ~ ~ 0 ~ CS 0 CO . . . . . . . . . . 00 ~ Cal ~ ~ ~ CD ~ (D cat 0 ~ ~ cs ~ co . . . . . . . . . . cat ~ ~ rib cr: 0 ~ ~ ~ cat Cal Cal Cal ~ ~ U: ~ 0 0 0 Cat . . . . . . . . . . CO 0 ~ ~ ~ ~ ~ Cat ~ ~ Cal Cat ~ ~ ~ ~ ~ CO CO ~ ~1 1 1 1 1 1 1 1 1 + _ 0 U: 0 ~ 0 ~ 0 U: Cal ~ CO ~ ~ U: ~ CC Cat CO ~ ~ ~ Cat Cal ~ ~ ~ . . . . . . . . . . 0 Go 0 ~ Cal ~ Cal . . . . . . . . . . CO ~ Cat Cal Cat ~ CO CO 0 . . . . . . . . . . Cal ~ 0 CO ~ ~ ~ ~ ~ ~ ~ 0 Cat Cal ~ 0 ~ 0 CS Cal . . . . . . . . . . 0 ~ Cal CO ~ ~ ~ ~ 0 0 U: ~ ~ UP ~ ~ 0 ~ Cal Cat . . . . . . . . . . red us ~ co oo Cal Cal rib rat ~ cat cat 0 us 0 . . . . . . . . . . cat ~ ~ ~ rib co ~ 0 cat co ~ ~ U: 0 ~ ~ ~ ~ ~ ~ . . . . . . . . . . Cal CO N ~ ~ ~ Cat ~ O ~ Cat Cal U: CO ~ 1~ ~ ~ O Cal CO ~ Cat Cal r ~ ~ U: Up es By ~ cat1 1 1 1 1 1 1 1 1 + O ~ O ~ O ~ O ~ O CO CO so ^. 0 C5) .. CQ C) =: =; ~ O U) by .= .- ~ ~ ._ - · - ~ En ¢ ·: CrS V O ~ O ~5 Eel ·' - (Q 0 ~ c';i . ~ cO ,-) us ~ ~ of ~ o o ~ ~ to ce ~ m ~- V V) ~ O REV ~ 0 0 U: ~
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 131 participation in voluntary and service organizations. No doubt it is also important in explaining differences in self-help and activ- ity levels generally. There is virtual neglect in forecasting educational attainment in official population projections; the last Bureau of the Census effort at projecting education dates back to 1972. In contrast, projections of school enrollment at various levels of schooling, especially over the short term, are very common. This undoubt- edly relates to the direct relevance that this information has for service and manpower planning. The 1972 projections, which were made to the year 1990, were found subsequently to be rather poor indicators of educational attainment trends for both males and females in that they overestimated the male trends and underestimated female patterns. This point is being noted because it emphasizes that projections based on previous trends or even current developments may often be distorted by cohort differences attributable to rather strong period effects. The situa- tion in the 1960s and early 1970s was unusual in that large numbers of mates remained in school to avoid the draft, thus inflating enrollments. The enrollments also were inflated by large-scale entry of veterans of military service who were enti- tled to educational benefits. Instead of presenting these dated projections, a perspective on prevailing trends can be gained from the last educational attain- ment figures that were published from the 1979 Current Popula- tion Survey (U.S. Bureau of the Census, 19801. These figures give us a view of what future developments may be as these cohorts reach older ages. (This assumes that no differential mortality by educational status exists, no doubt a questionable assumption.) The most noteworthy feature of Table 9 is the large increase in attainment for successively younger cohorts. Older persons at the turn of the century (beginning with the cohort aged 45 to 49) will have much higher levels of formal education and sharply lower proportions of persons who left school at the elementary or even high school level. After the turn of the century, a leveling off in median years of attainment can be anticipated for older per- sons and therefore a narrowing of the gap between older persons and younger cohorts. These results are in accord with recent projections of Shapiro and Easterlin (1981) on median school years completed by comparable age groups up to the year 2000. Median education levels of males are higher than females for
132 GEORGE C. MYERS ET AL. ages up to age 50, but a crossover takes place at that age. This does not reflect a change in college completion levels, for higher proportions of males complete college and do postgraduate work more often than females at all ages. Rather, it reflects a higher rate of completion of high school and some college for a larger share of women. We can conclude from this rather naive approach to forecasting that the possibility of sharply increased proportions of more edu- cated older persons could be realized by the year 2000. Thereaf- ter, older persons on the whole will have profiles of educational attainment that are not markedly different from younger cohorts. To the extent that formal educational attainment is posi- tively related to more productive roles, then the next few decades should witness a great improvement in this regard. Of course, at the same time we should balance this against the likelihood of higher labor force participation rates, especially for better edu- cated women, for cohorts up to the time they reach the older ages. HOUSEHOLDS AND MARITAL STATUS The two characteristics households and marital status are examined together inasmuch as the types of households that are found are closely related to the marital status of persons at speci- f~ed ages. The household or living arrangements in which older persons participate are clearly of major importance in determin- ing the life conditions, social relationships, and needs that are important in assessing productive roles. The Bureau of the Cen- sus prepares projections of households based on survey informa- tion and base population projections. The latest series up to the year 1995-was issued in 1979; therefore, it was based on the 1977 projections of population. Although somewhat outdated, they provide an indication of the likely trends that may occur in the years ahead for persons at different ages. Nearly two-thirds of households with householders aged 55 to 64 are intact husband-wife households (Table 101. This level falls sharply with age-to about a third for families with householders 75 years of age and over. Nonfamily households, which may con- sist of persons living alone or with nonrelatives, are more com- mon for females than for males, even at ages 55 to 64. The propor- tions in nonfamily households rise moderately with age for males but markedly for females, shifting from 16.S percent to 45.0 per
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 133 TABLE 10 Percentage Distribution of Households, by Types of Household and Age of Householder, 1980 and 1995 (Series C) 1980 1995 Ages 75 Ages and Over 55-64 Type of Household Ages Ages 55-64 65-74 Ages Ages 75 65-74 and Over Family household Husband-wife 65.7 49.9 34.6 65.0 49.5 34.6 Female-no spouse 8.0 7.2 7.3 7.6 5.4 4.1 Male-no spouse 2.1 1.4 2.0 2.3 1.1 1.1 Nonfamily Male 7.4 8.7 11.1 9.6 10.0 9.5 Female 16.8 32.7 45.0 15.6 33.9 50.7 Total 100.0 100.0 100.0 100.0 100.0 100.0 Number (in thousands) 12,494 9,917 6,121 12,138 12,055 9,302 SOURCE: U.S. Bureau of the Census. 1979. "Projections of the Number of House- holds and Families: 1979 to 1995." In Current Population Reports. Series P-25, No. 805, 1bble 2. Washington, D.C.: U.S. Government Printing Office. cent at 75 years and above. Female nonfamily households, there- fore, contain many very old persons a group generally viewed as a prime target population for special assistance. In 1980 it was estimated that there were 2.S million such households headed by females 75 years of age and over, the vast majority of whom were living alone. The projections reveal that little change is expected by 1995 in the proportions of family households made up of husbands and wives at the older ages. However, family households in which no spouse is present are projected to decline. NonfamiTy households will increase for males under 75 but will decline for those 75 and over. Female nonfamily households will be proportionately fewer in the ages 55 to 64 but will continue to increase proportionately for women 75 years of age and older. An interesting feature brought out by the projections is the decline in the number of households expected with householders aged 55 to 64. In con- trast, for older householders the increase is about 22 percent for the age category 65 to 74 and 52 percent for those 75 years and older. Once again the potential consuming population in terms of . . . . . . services Is Increasing at a more rapid pace. These trends are largely dependent on differential patterns of
134 Lo Cal on I be ¢ o . - Cal o o U] Cal U: Cal . - Ct By o o . - ,Q · ~ ·_' be Cal _ ~ V i,;; · _1 so a, _' o ~ l U: l l CO l 1 + CO l U: Cal l CO l ~7 U) Cal U: - ._ sin U: Cal =^ Go 0 Do o U: Go U: ~ lo ~0 0 Cat ~ i; -^ lo Cal ~Cal ~ O ~ C ~oo 0 C ~ O ~oO CO oO 0 CD o0 C~ U: d4 =4 ~ 0 CO C ~C~ 0 ~ CD C ~- CO C ~(D C ~_, ~ CO ~= C~ C ~CC C~ ~4 0 ~ CO CD C ~C ~CS) 0 ~O o O O ~ ._ ~ .a) 3 o U3 _) o . o . . ~Q - ._ Ct ~ O _ o ~ . - o _ ~ ~o ~ ~ u: c- o v u) ~ o ~ c) o b~ ~o ~ ~ .- - u:) t :~ co ,-l ~ - r~ ~ ~ o ~ oo ~ o ~o ~ s" ~ .o v) v) ~ o v ~ ~ ~: o o u] ~
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 135 mortality and survival, which have tended to keep families intact at younger ages during this period of life and produced increased widowhood at the extreme older ages. The data pre- sented in Able 11 clearly show the impact of these factors on marital status at older ages. The proportions of both males and females who are single are small and expected to decline by 1995. Over 80 percent of males between the ages 55 to 64 are in intact marriages compared with only 67 percent for females. With increasing age the male proportion declines somewhat, but the proportion of married females with a spouse living drops sharply to a level of 17.3 percent for females 75 years of age and older. Correspondingly, the proportions of widowed, separated, and divorced (the large majority of whom are widowed at these older ages) increase moderately for males and increase greatly for females. At ages 75 and older, three-quarters of females are with- out a spouse. The trend by age reflected in these projections is for males to experience intact marriages later in life. For example, a gain of 7 percentage points for males 75 and older may be noted within the category defined as married with spouse present. For females, there is a slight tendency for an increase. The proportion of wid- owed, separated, or divorced mates is projected to increase slightly at ages up to 75, but a large decline is expected in the oldest age category. Female proportions will tend to decrease for all age groups to 1995. Although these results may seem anoma- lous, they are confirmed in recent cross-national research (includ- ing the United States) in which the proportions widowed appear to be declining and intact marital situations increasing, particu- larly for males (Myers, 1982; also see Masnick and Bane, 19801. Part of the explanation for this development lies in improved survival at younger ages for those already considered to be old, but it also reflects on the large open-ended terminal category that is often used. If more detailed older age categories are used, then the proportions of persons in intact marriages do decline for both sexes. Longer-term projections to the year 2060 have been proposed by the Social Security Office of Actuaries for the aged population as a whole (Table 121. Some of the fluctuations in the proportions of older persons in each marital status reflect the age composi- tions of the aged population. This is true for married and wid- owed categories, in particular, in which there are declines, then
136 GEORGE C. MYERS ET AL. TABLE 12 Marital Status of Population 65 Years and Over in the Social Security Area, by Sex, United States, 1980-2060 Year Total Single Married Widowed Divorced NUMBER Male 1980 10,397600 7,686 1,688424 2000 14,232716 10,128 2,1181,270 2020 21,5671,236 15,730 2,7431,858 2040 26,8411,951 18,798 3,7942,299 2060 28,1602,196 19,848 3,6022,514 Female 1980 15,544940 5,989 8,024590 2000 21,767980 8,092 9,9622,734 2020 30,9041,597 12,606 11,5135,188 2040 39,7882,289 15,473 16,1845,842 2060 41,6182,548 16,185 16,6936,192 Total 1980 25,9411,540 13,675 9,7121,014 2000 35,9991,696 18,220 12,0804,004 2020 52,4712,833 28,336 14,2567,046 2040 66,6294,240 34,271 19,9788,141 2060 69,7784,744 36,033 20,2958,706 PERCENTAGE Male 1980 100.0 5.8 73.9 16.2 4.1 2000 100.0 5.0 71.2 14.9 8.9 2020 100.0 5.7 72.9 12.7 8.6 2040 100.0 7.3 70.0 14.1 8.6 2060 100.0 7.8 70.5 12.8 8.9 Female 1980 100.0 6.0 38.5 51.6 3.8 2000 100.0 4.5 37.2 45.8 12.6 2020 100.0 5.2 40.8 37.2 16.8 2040 100.0 5.8 38.9 40.7 14.7 2060 100.0 6.14 38.9 40.1 14.9 Total 1980 100.0 5.9 52.7 37.4 3.9 2000 100.0 4.7 50.6 33.6 11.1 2020 100.0 5.4 54.0 27.2 13.4 2040 100.0 6.4 51.4 30.0 12.2 2060 100.0 6.8 51.6 29.1 12.5 SOURCE: U.S. Department of Health and Human Services. 1980. United States Population Projection by Marital Status for OASDI Cost Estimates, 1980. Actuarial Study No. 84. Washington, D.C.: U.S. Government Printing Office.
PRODUCTIVE ROLES: SOCIODEMOGRAPNIC ASPECTS 137 . increases, then declines. The general pattern, however, is toward Tower proportions of widowed persons, especially females; some increase in single persons; and large gains in the proportions of both male and female divorced persons. While these percentage figures do not reveal any major changes in the marital categories (e.g., single, widowed, divorced) that might be expected to lead toward greater dependence, it should be kept in mind that large numerical gains are characteristics of all of the categories as the overall aged population increases. In terms of service require- ments, it is the number of persons requiring assistance that is of paramount importance. These findings about trends in household patterns, particu- larly for females living alone and in nonfamilial circumstances, suggest that the demand for services may well increase, espe- cially at the oldest ages, but that persons in intact marriages may be both relatively and numerically more numerous and, therefore, be in positions for supplying services. KINSHIP RELATIONS The extent to which unpaid productive roles relate to mutual aid for other family members must necessarily be influenced by the potential kinship networks that exist. Demographic trends pointing to steadily declining mortality at older ages, which increases survival, and continued Tow fertility levels have modi- fied family structures by altering the number of siblings, chil- dren, and multigenerational families. Of particular importance for the topic is the increased likelihood that persons entering the older age categories will themselves have parents still living. Ceteris paribus, this creates new and rather different needs for mutual aid to parents, at the same time limiting the pool of persons available to provide aid to other nonfamiTy individuals · , . Or organizations. Several modeling approaches may be taken to derive estimates of available kin, and various analyses already have been made to capture those dynamics (Hammer et al., 1981; Pullum, 19821. Wolf (1983) has made the most recent effort to mode] the process, and some results from his projection study are presented in Table 13. The data are presented only for males inasmuch as the results for females are quite similar. The probability of a triage being 66 to 69 years of age and
138 to CO Lo o --o o-.~.oo o v of Cal r a C) o a Ct C - O. U2 o . - - .= Cat ·_1 o :^ - · ~. o CC d. 1 O O Cal a 1 CO U: Cal 0 O Cal U: Cal ~ O Cal 0 O Cal to Cal Cal Cal ~ O Cal s~ ~ .~ ° ~ ~ ~ co O rig 11 11 so ~ 0 Ce ~ do U) Cal Car Cal O en . . . . Do us rig Go ~ Cal U: 0 O C~ C~ ~ . . . . CO C~ C~ C~ O ~ C~ . . . . r~ c~ r~ O . . . . CO O C~ ~ U: ~ O ~ ~ 00 . . . . O C~ ~ ~ ~ C~ O O 0 0 . . . . m . ~ =o Ct 5- - ~, ~.> =~ ~0 ~ ~ C~ s~ E~ . . V o ._ CQ 5 CC s~ ~3 o CO s~ ~: ~4 ·_} .S CO ·_, 00 ¢ U) _ C~ _ b~ O ~ O 5- Ct 5- U] o s~ . . _ ·_ O ~ U] ~
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 139 having a father still alive was .014 in 1975; it increases to .06 by the year 2015. The probability of having a mother alive at these ages is much higher, as expected, and it increases in 2015 to a level in which a quarter of the males would be in this situation. The probabilities of having any living parent alive is naturally higher. Having at least one sibling alive is more sensitive to prior fertility (note the dip in 1921-1925 and later cohorts), but the levels are very high, reaching 88 percent in 2015. Finally, having a living child rises to 91 percent in the year 2000 for males aged 65 to 69 but then declines gradually to 84 percent in 2015. This reflects the present "birth dearth," and assumes low fertility lev- els for the near future. It is likely, then, that there will be an enlarged pool of family members for whom mutual aid may be necessary. In turn, youn- ger family members (at ages 65 to 69) are also available who could provide assistance if someone was in need. The term "potential" must be emphasized, inasmuch as the family support system depends on many other factors as well. These figures suggest that mortality conditions play a somewhat greater role than fertility in the structure of family relations and touch upon a whole range of issues relating to living arrangements, migra- tory patterns, and mutual aid and assistance. HEALTH STATUS There have been relatively few attempts to forecast the health status of the older population. One recent effort by the National Center for Health Statistics uses fairly conventional procedures of ratio estimation for different health dimensions applied to pro- jections by age and sex. As will be noted subsequently, alterna- tive conceptualizations of the issue are also possible. Table 14 provides the main conclusions from this study using two different mortality assumptions in the projections for contrast. Even if we consider the projections under the unrealistic assumption of constant mortality, by the year 2003 the impact of the aging population may be clearly noted on all of the health dimensions examined. Declining mortality shows large increases over the 25-year period in terms of persons with limitations of activity, hospital short-stay visits, and especially nursing home residency (the number of such residents more than doubles under the assumed ratios). Physician visits are the least likely to
140 Ct Cat of - o Cal ¢ o U: so LO ca o 5- ,~o, CQ Cal ·_4 ·_4 Cal U: - cd E-°. Cat o + CO .. be Cal CO o _ Ct E° by ¢ _ ¢ o _ E-°' .= ~ . - U) 5_, CC o o g v ~ ~ Go Cat be ~ ·_ - _ ._ Ct - C~ o ° o V ~ ~ oo r~ b~ ~ . _ - · - ~ Ct:) o ° . - Ct: ~ o ° v ~ ~o oo C5) 0 ~r~c~oo ca~ ..... .. C~o C~ C~ C~ ... ~CSoo CO ~C~ ~C9 ... ~CO CO ~CO CDCO 0U:) C~ 00~ 8 ~c~ C~ C~ 0 o co~ 0r~ .. .. ~CO ~CD 0C~ C~0 0C~ ~ .. .. COC~ ~C~ 0C~ C~0 CO0 ~ .. .. ~c~ ~r~ 0~ COC~ U: C~ ~ oCO 0o ~<=, ~ C~ C~ ... .. o ~0 oC~ o ~C~ ~CD0 ~CO 0co _.... .. oooCO ~ ~o oo~ oC~ 0 `' ~C~U, C~ - o o ~0 0 c~7c, N 1O CO - ~U: ~O ~ ~C~ C ~CO C ~O_ ~- ^ C~ . ~C~ U) C ~O C-C`4 ~CD ~C~/= z) o ~~ E-. E ~ U] Z ,; E ~, ,, ~ a ~ ,, ~2 ~ ~ ~ 2 z; ~ ~ ~ z; v ~ ~ & - ·: ~ v ct 0 = 0 c~ ~ _ ~ ._ . c~ s~ ~ Z _ c~ ~ ~v =._ ~ .~ _) . _ ~ b~ b~ v - - t3 0 ~ 45, · CQ ·_ _ CQ · - u: ~ . ~ g - c`;l ~ 0 - ° v ~ . _ ~ ~ ·- ~ O ~ o.= ~ . - · = - v O x u:
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 141 increase markedly. Nonetheless, the relative burden of health care requirements for the older population would appear to be a major cause of concern for the health care system. Many of these dimensions also would lead to increased demand for informal care for example, for the nearly 40 percent of the population with limitations of activity represented by aged persons. In terms of health expenditures these projections show the growth in fiscal liability generated by the aging population. For these three services the 1978 expenditures of $124.7 billion increase to $167.3 and $185.7 billion in the year 2003 under the two mortality assumptions. The aged share of these expenditures increases from 32.6 percent to 35.6 and 40.7 percent, respectively, in 2003. The cost gains are particularly large in terms of nursing home care, exceeding even the aged expenditure increases for hospital care. Thus, both the demand for services and expendi- tures are highly sensitive to demographic changes generated by an aging population. Figures such as these, which are undoubt- edly underestimates, are the impetus for increased calls for attention to alternative types of community and family care that might complement the formal health care system. Consideration of health status is quite different from the other characteristics discussed earlier in that the outcome variable is itself part of the input to a projective model. It has been argued that appropriate demographic projections will have to relate changes in the age structure of the over-65 population (which are, to a large part, a function of mortality at those ages) to the associ- ated morbidity and disability changes. Thus, such projections must forecast mortality, morbidity, and disability (Ioss of func- tional capacity) as correlated phenomena, each driven by the underlying processes of physiological aging. Perhaps the simplest way to appreciate the age implications of the biomedical correlation of mortality, morbidity, and disability is to consider Figure 1. In this figure, the horizontal axis repre- sents age and the vertical axis represents the proportion of a birth cohort that would survive to a given age without experienc- ing a particular type of health status change. The three curves in the figure represent the trajectory of change in the age-specif~c probability of the event. Specifically, the first curve (A) repre- sents the simple probability of survival to age x. A second curve (B) represents the probability that a person will survive to age x without suffering serious limitations of activity. Curve (C) indi
142 GEORGE C. MYERS ET AL. eO and e60 are the number of years expected to be lived autonomously at birth and at age 60, respectively. M50, M25, and M10 are the ages to which 50, 25, and 10 percent of individuals could expect to survive in an autonomous state. 0.8 0.6 0.4 o ~_; ·;~_` ~: Mortality (A) Morbidity (C) ` ~ j :\ art" , 1 1 J~` ~ _ 0 10 20 30 40 50 .. ~ __ _- 60 ]01 801 1 ~100 1 10 AGE eOM50e6o M25 Mao FIGURE 1 Mortality, morbidity, and disability survival curves. cates survival without a morbid condition. The figure thus clearly represents (1) the changing age correlation of mortality and disability and (2) the age dynamics that cause the prevalence of disease and disability to rise in conjunction with a rise in the risk of mortality. The figure also defines the question: Given changes in survival (A), how does the prevalence of disease and disability in the population change? Clearly, this latter question is of critical importance in determining both the demand for vari- ous types of volunteer services among the elderly and the poten- tial pool of elderly who are healthy and able to provide such services. Although the need for projections giving the age change in the relation of mortality to other health status is clear, such projec- tions rarely have been performed. This is rather surprising because the data requirements for such projections are not as great as they first appear. One example is provided in a simula- tion study of population aging in Japanese society conducted by Nihon University (19821. In an analysis by Koizumi (1982), data from standard vital statistics life tables were combined with information from three nationally representative surveys of health characteristics, health service utilization, and welfare.
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 143 The results from such analyses were used to interrelate morbid- ity, health status changes, perception of subjective health status, health service utilization, and mortality over age. As such they are relevant to both supply and demand issues. These relations can be examined in an example using U.S. data, which is presented in Figure 2. In this figure, data on heart and hypertensive disease and limitations are combined with the survival probabilities derived from a standard U.S. life table. As we can see, the age-specific prevalence of morbidity rises initially and is followed secondarily by a rapid increase in disability at later ages. The use of projections that interrelate various population health states in this way does not resolve all of the issues in assessing health, health service utilization, and the implications of health for the supply of and demand for volunteer services. However, it does provide information on the basic parameters of such behavior for the system. The fact that relatively little effort has been applied to resolving the nature of such associations on a population level let alone forecasting such relations into the future- indicates some serious gaps in the information base needed to plan for the requirements of an aging U.S. population. ~ 0.8 I In ~ 0.6 I ~ 0.4 I o ~0.2 Heart Disease/Hypertension ~ I\ & Limitation `` \ \ · Heart Disease/Hypertension `` \ \ oL I I I I I I I ~ I 0 1 0 20 30 40 AGE \\ \\ At' 50 60 70 80 85 FIGURE 2 Survival curves for the 1978 U.S. population and for heart disease and hypertension, with and without limitation.
144 GEORGE C. MYERS ET AL. DISCUSSION Clearly, population projections are needed in order to create an awareness of the social changes that are occurring in a society and that are likely to occur in the future. Although demogra- phers and other statistical scientists began to make such projec- tions in the nineteenth century, it is only in the past 50 years that systematic and periodic projections have become common- place. For example, the first systematic projections of the U.S. population were made in 1934 by the U.S. National Resources Board and prepared by Thompson and Whelpton. And it was not until 1947 that the Bureau of the Census began to prepare projec- tions as a regular activity. In this half century there has been a growing acknowledgment that population projections are an integral part of the planning process for government agencies on all levels and for business enterprises and the like. With this increased demand has emerged a continuing need for more disag- gregated projections that relate more proximally to the specific components of the population of interest, that is, the conditions of the aged population itself. The projections from which data have been drawn for this review are basically general projections of the total population that have been made with little specific attention to particular dynamics of change at the older stages of life. For example, the issues raised about health status indicate the complex and per- haps rapidly changing nature of disease prevalence and disabil- ity at older ages. These factors directly affect mortality rates, which play such an important role in projections of older persons. The rather poor record that has been achieved in projecting the numbers of aged persons in the past reflects on the inadequacy of mortality forecasts that failed to capture the dramatic declines in mortality at older ages during the past decade. The changes in labor force participation at points of retirement demand careful single-year-of-age determinations, but these issues have only recently been pursued and no current data are available in pub- lished form. In short there appears to be a need for projections that devote more attention to the dynamic properties of the older population. Many of the projections reported earlier derive from the so- called official projections of the U.S. Bureau of the Census in the sense that they apply ratios of the subject matter to the projected
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 145 age-sex-race base population figures Tong, 19811. This is true of projections of educational attainment, households, marital sta- tus, labor force, and the like. In spite of the apparent linkage between these projections, the different efforts that go into them are generally uncoordinated. We find the projections beginning at different years and being carried forward for different time periods, the use of varying age categories, and release at differ- ent points in time. This may be disconcerting to the users, but it also reflects on the lack of systematic interplay of crucial dimen- sions within the projections themselves. For example, labor force projections are made without consideration of such factors as educational attainment and marital status. Mortality or fertility trends are developed without consideration of labor force or mari- tal status. The components, therefore, are not interrelated on the conceptual level, and this is carried through to the technical level. It is not surprising, then, that organizations outside the federal government have developed their own projection models. The microsimulation model DYNASIM created by the Urban Insti- tute, the macroeconomic-demographic model used by {CF Incor- porated, and the macro mode} used by the Joint Center for Urban Studies of MIT and Harvard University are cases in point. But although studies using these more sophisticated models are often of great interest, there is an even greater diversity in the modes of presenting results from these studies than there is in the feder- ally produced series. This circumstance makes reviewing the findings difficult and interpreting discrepancies virtually impos- sible. To summarize, this brief review has pointed to a number of conceptual, technical, and organizational factors that limit the use of current projections for planning and policy formulation purposes. While we are not suggesting that a single set of projec- tions should be developed and strictly adhered to in the federal government, we do think that greater coordination would be desirable. There also is a clear need to examine more sophisti- cated models of both inputs in projections (i.e., mortality, fertility, migration) as well as outcomes (e.g., educational attainment, health status). This may be accomplished by using more structur- ally dynamic analytic models, such as multistate mathematical applications that include increment-decrement life table approaches, and biomedical models of the sickness-death process.
146 GEORGE C. MYERS ET AL. Moreover, adequate assessment of the accuracy of population pro jections should be an ongoing activity. Current development sug gests that the time may be appropriate for providing information about confidence limits at the time projections are issued (Key f~tz, 1981, 1982; Stoto, 19831. In short, we are suggesting that projections be given a higher research priority than has previ ously been the case. The National Academy of Sciences could well play an important role as a catalyst in this development. REFERENCES Anderson, J. M., and W. McNaught. 1982. Projecting Alternative Futures for the Retirement Income System. Final report to the National Institute on Aging. Washington, D.C.: ICE, Inc. Ascher, W. 1978. Forecasting: An Appraisal for Policymakers and Planners. Baltimore, Md.: Johns Hopkins University Press. Hammel, E. A., K. W. Wachter, and C. K. McDaniel. 1981. "The Kin of the Aged in A.D. 2000: The Chickens Come Home to Roost." PP. 11-39 in S. B. Kiesler, N. Morgan, and V. K. Oppenheimer, eds. Aging: Social Change. New York: Academic Press. Hendricks, G., and J. R. Storey. 1981. The Long-Run Effects of Alternative Mandatory Retirement Policies. Washington, D.C.: The Urban Institute. Keyfitz, N. 1981. "The Limits of Population Forecasting." Population and Development Review 8(4):579-593. Keyfitz, N. 1982. "Can Knowledge Improve Forecasts?" Population and Development Review 8 (4):729-751. Koizumi, A. 1982. "Toward a Healthy Life in the 21st Century." Chapter 6 (pp. 6-1 and 6-19) in Population Aging in Japan. Problems and Policy Issues in the 21st Century. International Symposium on an Aging Society: Strategies for 21st Century Japan, November 24-27, Nihon University. Long, J.F. 1981. "Survey of Federally Produced National Level Demographic Projections," Review of Public Data Use 9:309-329. Masnick, G., and M. J. Bane. 1980. The Nations Families: 1960-1990. An Outlook Report of the Joint Center for Urban Studies of MIT and Harvard University. Myers, G. C. 1982. "Cross-national Variations in Marital Status Among The Elderly." Paper presented at Gerontological Society of America Meetings, Boston, Massachusetts. National Center for Health Statistics. 1983. "Changing Mortality Patterns, Health Services Utilization and Health Care Expenditures: United States, 1978-2003." In Analytical and Epidemiological Studies. Series 3, No. 23. Washington, D.C.: U.S. Government Printing Office. National Institute on Aging. 1982. A National Plan for Research on Aging: Report of the National Research on Aging Planning Panel. Washington, D.C.: U.S. Government Printing Office. National Institute on Aging. 1984. Macroeconomi~Demographic Model. Washington, D.C.: U.S. Government Printing Office. Nihon University. 1982. "Population-Aging in Japan: Problems and Policy Issues in
PRODUCTIVE ROLES: SOCIODEMOGRAPHIC ASPECTS 147 the 21st Century." Paper presented at the International Symposium on an Aging Society: Strategies for 21st Century Japan, November 24-27. Pullum, T. W. 1982. "The Eventual Frequencies of Kin in a Stable Population." Demography 19:549-565. Rosow, I. 1976. "Status and Role Change Through the Life Span." In R. H. Binstock and E. Shanas, eds. Handbook of Aging and the Social Sciences. New York: Van Nostrand Reinhold. Shapiro, M. O., and R. A. Easterlin. 1981. "Educational Attainment by Sex and Age, 1980-2000." Review of Public Data Use (9):323-329. Stoto, M. A. 1983. "The Accuracy of Population Projections." Journal of the American Statistical Association (18):13-20. U.S. Bureau of Labor Statistics. 1982. Economic Projections to 1990. Bulletin 2121. Washington, D.C.: U.S. Government Printing Office. U.S. Bureau of Labor Statistics. 1984. Employment Projections for 1995. Bulletin 2197, Appendix A-1. Washington, D.C.: U.S. Government Printing Office. U.S. Bureau of the Census. 1979. "Projections of the Number of Households and Families: 1979 to 1995." In Current Population Reports. Series P-25, No. 805. Washington, D.C.: U.S. Government Printing Office. U.S. Bureau of the Census. 1980. "Educational Attainment in the United States: March 1979 and 1978." In Current Population Reports. Series P-20, No. 356. Washington, D.C.: U.S. Government Printing Office. U.S. Bureau of the Census. 1983. "Provisional Projections of the Population of the United States by Age and Sex: 1980 to 2000." In Current Population Reports. Series P-25, No.917. Washington, D.C.: U.S. Government Printing Office. U.S. Bureau of the Census. 1984. "Projections of the Population of the United States, by Age, Sex, and Race: 1983 to 2080." In Current Population Reports. Series P-25, No. 952. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Health and Human Services, Social Security Administration, Office of the Actuary. 1980. United States Population Projection by Marital Status for OASDI Cost Estimates, 1980. Actuarial Study No. 84. Washington, D.C.: U.S. Government Printing Office. Wertheimer, R. F., and S. R. Zedlewski. 1980. The Aging of America A Portrait of the Elderly in 1990. Washington, D.C.: The Urban Institute. Wolf, D. A. 1983. Kinship and the Living Arrangements of Older Americans. Washington, D.C.: The Urban Institute.
