3 Family Dynamics and Living Arrangement of Elderly Persons
It is very clear that Chinese population aging will be extremely rapid and the size of the elderly population will be exceptionally large in the first half of the next century. Population aging is accompanied by changes in family household structure (see e.g., Wolf 1994, for a review). Elderly persons depend upon spouses and children for emotional and physical support, and financial aid as well, especially in rural areas in China. Past research has established that family care is an important part of long-term care, and has substantial impact on caregiving arrangements for the elderly (e.g. Soldo et al. 1990). In particular, the use of institutional long-term care has been shown to vary by family status . Cohorts who will become the elderly of the 21st century have been on the leading edge of the family changes, with rapid decreases in fertility and quick rises in divorce, and changes in attitudes about co-residence between parents and married children . Long-term care costs in the U.S. have doubled during each decade since 1970, reaching an annual level of $106.5 billion in 1995. Home health care costs grew 90.7% from 1990 to 1995, in contrast to 33.4% for institutional care costs . Thus, the mix of home based and institutional care has been rapidly shifting towards home health care, especially for the oldest elderly . Clearly, changes in family structure strongly affect caregiving needs, the long term care health service system, and health-related policy-making (, ).
How may demographic and socio-economic changes alter the family households and living arrangements of the Chinese elderly population? How many elderly persons live alone, or with a spouse only, or with children or grandchildren or other relatives, or in institution? How many elderly live with a son or daughter and his or her spouse? What are the gender and rural-urban differences? What are changes of family structure and living arrangements in the recent past, and what are the future perspectives? Mainly based on the data derived from the one-per-thousand micro data files of the 1990 and 1982 censuses, as well as family household projection, this section will address questions such as these, which are important for elderly caregiving and health-related policy-making.
3.1 Family structure and living arrangement of Elderly Persons in 1990
Table 3 and Figure 2 present detailed and simplified percentage distributions of family structure and living arrangements of elderly persons in Mainland China, comparing rural and urban areas, based on the 1990 census data. We classify the elderly into three age groups: the moderately old (persons age 65-79), very old (persons age 80-89), and extremely old (persons age 90+). The broader age category "oldest old" is a combination of the very old (age 80-89) and extremely old (age 90+). We will employ this terminology in the following discussions.
3.1.1 Co-Residence with Children
A large majority of the moderately old women (73.9%) and men (68.7%) live with their children (children include grandchildren hereafter, unless otherwise specified). A higher proportion (79.9% of females and 69.1% of males) of the very old live with their children. The corresponding figures for the extremely old are the highest: 82.4% for women and 76.7% for men (see Table 3). It is very clear that a large majority of the Chinese elderly live with their children, and the higher the age, the higher the proportion living with their children. Female elderly persons of all age groups are more likely to live with their children, because elderly women are more likely to be economically dependent and widowed.
In the Chinese censuses, householders' children and children's spouse were coded as one category "children", so that it is impossible to distinguish between married sons and married daughters who live with their parents. Thus, we have to rely on other data sources to examine the living arrangements with sons versus daughters. According to the 1998 healthy longevity survey that sampled elders age 80+ in 22 provinces [Note 6], the percent living with a daughter (among those living with children) is 17.3, 17.7, and 17.7 for men age 80-89, 90-99, and 100+, respectively. The corresponding figures for women age 80-89, 90-99, and 100+ are 15.4, 18.8, and 18.1, respectively. These figures show that, on one hand, most of the oldest old live with adult sons, and on the other hand, a considerable portion of them live with adult daughters.
Based on 1990 census data, among the moderately old who live with offspring, a majority (66.8 and 79.6 percent of men and women) lives with both children and grandchildren. A larger majority (82.6% for men and 83.2% for women) of the very old who live with offspring lives with both children and grandchildren. The corresponding figures for the extremely old men and women are 79.5% and 77.1%, respectively (see Table 3). In the cultural context of Chinese society, multi-generation family households are one of the main living arrangements for the elderly, especially for the oldest old.
Slightly more than 2 percent of the moderately old men and women live with grandchildren without son or daughter present. The corresponding figures are 2.1 and 1.5 percent for very old males and females, and 2.7 and 1.3 percent for extremely old men and women (see Table 3). This kind of two-generation household, consisting of grandparents and grandchildren, indicates that in Chinese society grandchildren may care for their grandparents when the middle generation is not available (perhaps due to reasons of job location or death), or another way around - not very old grandparents take care of young grandchildren.
3.1.2 Living with a Spouse
Among moderately old men and women living with children, 36.8 and 69.7 percent do not have a spouse present. Among the very old who live with children, a large majority (69.5% for men and 94.6% for women) does not live with a spouse. The corresponding figures are 80.7 and 98.9 percent for extremely old men and women (see Table 3). The proportion of elderly men who live with spouse only is 21.6, 16.0, and 8.0 percent at ages 65-69, 80-89, and 90+, respectively, in contrast to 15.2, 4.0, and 0.5 percent for their female counterparts (see Table 3). The proportion of not living with a spouse increases tremendously with age, due to high rates of widowhood at old ages (divorce rate in China is extremely low). Many more elderly women are widowed than men because of the gender differential in mortality at advanced ages.
3.1.3 Living Alone or with Other Relatives or Non-Relatives
The proportion of moderately old men and women living alone is 8.0 and 10.2 percent, in contrast to 13.2 and 15.2 for the oldest old males and females, respectively (differences in the proportions living alone between ages groups 80-89 and 90+ are very small, see Table 3). Note that elderly women are much more likely to be widowed and thus live alone. On the other hand, elderly women are economically more dependent. Therefore, the disadvantages of women in marital life and living arrangements are substantially more serious than those of men at advanced ages.
A very small proportion (0.6 and 0.3 percent for moderately old men and women; 0.4 and 0.2 percent for oldest old men and women, respectively) of the Chinese elderly lives with other relatives or non-relatives, without children present (see Table 3). This fact demonstrates that Chinese elderly rarely live with other people than offspring and spouse.
3.1.4 Living in Institutional Household
The proportions of moderately old, very old, and extremely old men who live in institutional households in 1990 were 1.1, 1.2, and 1.9 percent, respectively [Note 7]. The corresponding figures for moderately old, very old, and extremely old women are 0.3, 0.7 and 1.1 percent, respectively (see Table 3). Given the extremely limited facilities available for institutional care for elderly, perhaps the major cause of institutionalization of elderly persons in China in 1990 is childlessness (or absent of children). Therefore, the percent of the institutionalized elderly was extremely low, as compared to developed countries where the most important reason for an elderly person to move into an institution is disability. Note that the Chinese elderly women's lower social and economic status made them less likely be able to access long-term care facilities. This is another social disadvantage faced by elderly women in Chinese society, which merits attention from the society and government.
3.2 Rural-urban Differentials in 1990
Table 3 and Figure 2 present detailed and simplified percentage distributions of family structure and living arrangements of elderly persons in 1990, with a comparison between rural and urban areas. The proportion of the moderately old men who live with children in rural and urban areas in 1990 was 69.7 and 65.9, respectively, and the corresponding figures for women were 74.7 and 71.8, respectively (see Table 3). Thus, the rural moderately old persons are more likely to live with their children than their urban counterparts. However, the proportions of very old and extremely old persons who live with their children in the rural areas is slightly lower than in the urban areas, except for men age 80-89. More studies are needed to explain this phenomenon.
A very interesting finding is that many more urban oldest old persons live with daughters than their rural counterparts. Among the oldest old males and females age 80+ who live with children in urban areas, 26.6 and 28.8 percent live with a daughter, while the corresponding figures are 11.8 and 9.5 percent in the rural areas, based on the 1998 healthy longevity survey data. Such striking rural-urban differentials of proportions living with daughters among the oldest old also exist at age groups 80-89, 90-99, and 100+. This demonstrates that the traditional idea of relying on sons for old age care is less popular in urban areas, and is changing with modernization. More and more old people in urban areas accept or even prefer to live with a daughter if possible, since daughters are more likely to provide better care to old parents than sons. This gives us a hope that the traditional son preference in China may be reversed if urbanization is accompanied with appropriate social programs which aim at increasing women's status and encouraging old persons to live with their daughters.
The proportions of those moderately old, very old, and extremely old men and women who live with both children and grandchildren are all higher in rural than in urban areas (see Table 3). This confirms that multi-generation family household is more popular in rural areas than in urban areas in contemporary China.
The proportion of the moderately old, very old, and extremely old who live with grandchildren without son or daughter present is about 2-5 times higher in urban areas than rural areas (see Table 3). This suggests that job location (rather than death) of the middle generation is the main reason for these special two-generation households consisting of grandparents and grandchildren in China.
Slightly more moderately old men and women live with only a spouse in urban areas than rural areas. However, there are no clear pattern of rural-urban differences in living with spouse only among the very old and extremely old (see Figure 2 and Table 3).
There is no clear pattern of rural-urban differences of proportions of the moderately old who live alone, but it is evident that the proportions of the very old and extremely old who live alone are higher in rural areas than urban areas (see Figure 2 and Table 3). People generally speculate that the urban elderly are more likely to prefer privacy and independent family structure and living arrangement, and thus are more likely to live alone than their rural counterparts. But the Chinese census data do not support this hypothesis. Perhaps some other factors such as higher widowhood rates, lower remarriage rates, and fewer long-term care facilities in rural areas than urban areas offset the effects of rural-urban attitudes differences. One may also speculate that the preference for privacy and independent family structure and living arrangements, even among Chinese elderly in the urban areas are still not strong. More in-depth studies are needed.
The proportions of institutionalized moderately old, very old and extremely old men and women were 2 - 4 times among urban than rural residents in 1990 (see Figure 2 and Table 3). This is not surprising. In the rural areas, very few institutional facilities for the elderly are available. Those rural oldest old living in the institutional households are there because of childlessness, rather than disability. Chinese policies in rural areas allow only those elderly who have no close relatives to stay in the government subsided institutional households. In urban areas, the facility limitation and institutional policy are relatively less restrictive, permitting more elderly to live in institutional households.
3.3 Changes between 1982 and 1990
Table 4 and Figure 3 present detailed and simplified percentage distributions of family structure and living arrangements of elderly persons (rural and urban areas combined [Note 8]), with a comparison between 1990 and 1982, based on the censuses data collected in these two years. The proportion of the elderly living with children decreased slightly from 69.6 percent in 1982 to 68.7 percent in 1990 for the moderately old men, and remained virtually unchanged for the moderately old women (74.1% in 1982 and 73.9% in 1990). The proportions of the very old, and extremely old men and women who live with children increased by 2-4 percentage points between 1982 and 1990 (see Table 4). The proportion of elderly men and women age 65-79, 80-89, and 90+ who live with both children and grandchildren also increased between 1982 and 1990 (see Table 4). Furthermore, the proportion living alone declined considerably at all ages 65-79, 80-89, and 90+ (see Figure 3 and Table 4). Were the family structure and living arrangements of Chinese elderly in 1990 more traditional than 1982? This seems unlikely, given the rapid socio-economic development and opening door to the outside world that occurred in China during the 1980s. How then, can we interpret this phenomenon that is contradictory to the general and theoretical expectations? We believe that the following factors might contribute to this phenomenon. First, elimination of the previous food ration system and changes in the function of household register booklets around 1990 might play a role. In the 1970s and early 1980s, very low efficiency in the collective agriculture production system resulted in severe food shortages. In addition to the main food ration, other foodstuffs such as meat, fish, eggs etc. were primarily supplied on the basis of the household register booklet. Each household could periodically purchase certain amounts of low-priced subsidiary food, and the household register booklet was used as ID. This led some young or old people, who actually lived with their family members, to register as a separate household. The number of one-person households including elderly persons living alone, was over-enumerated, and the elderly living with children was under-enumerated. Such biases resulted in that the State Statistical Bureau adjusted the urban average family household size enumerated in the 1982 census from 3.84 to 3.95 through a post-census sample check. The rural household size was not adjusted, but a similar bias (that might be smaller) existed in the rural areas in the 1982 census. Such biases were much less serious in 1990 because the food ration system was basically eliminated and the incentives for having more household register booklets were greatly reduced. Second, increases in remarriage rates and decreases in death rates at old ages may also partly explain why the proportion of the elderly who lived alone declined between 1982 and 1990. The increase in remarriage rates among elderly persons was a result of the social reform and the progress of mate-made services in the late 1980s. The reform aimed at protecting elders' rights, including the right to remarry that were often violated by the intervention of children and other family members in traditional Chinese society. Rapid economic development, accompanied by the substantial improvements in the standard of living might decrease death rates at old ages. Although these explanations are speculative due to the lack of empirical data, we believe that the family structure and living arrangements of the Chinese elderly in the 1980s were not a return to tradition. We also believe that the traditional norm that Chinese elderly should live with their children changed little, if at all, between 1982 and 1990.
The proportion of the elderly living with spouses (including with spouse only and with spouse and children) increased considerably between 1982 and 1990, among the moderately old and very old men and women, but remained almost unchanged among women age 90+ and decreased by 3 percentage points among men age 90+ (see Table 3). Similar changes in the proportions of the elderly living with spouse only between 1982 and 1990 among the moderately old, very old and extremely old persons were also observed (see Figure 3 and Table 4). We believe that the possible explanatory factors for this interesting phenomenon are increases in remarriage rates among the elderly age 65-89 and decreased death rates in the 1980s, as discussed above. Another possible candidate of the explanations is decreases in divorce rates among the moderately old persons. But this factor can be ruled out, since divorce rates among elderly persons are extremely low in China, and there is no evidence of further decline. In fact, the general divorce rates for the whole population have substantially increased in China since 1980 .
As compared to 1982, the proportion of the elderly living in institutional households increased by 1990 among women age 80-89 and men and women age 90+, remained unchanged among men age 80-89 and women age 65-79, but decreased among men age 65-79 (see Figure 3 and Table 4). Socioeconomic development in the 1980s may have contributed to improvement of long-term care facilities, which facilitated the increases in the proportion of those living in institutional households, although it is still very low. Commercial nursing home facilities for the elderly have been growing quickly since 1990 due to rapid economic development, and the percents of the institutionalized oldest old in 1998 are 6.1 for men and 6.7 for women age 80+, based on the 1998 survey data conducted in the 22 provinces. These figures are still low, as compared with the Western countries, but have tremendously increased since1990. Modernization in China may further increase the proportion of the elderly living in long-term care institutions in the future.
3.4 Future Perspectives
Note that, on average, elderly persons enumerated in the 1990 and 1982 censuses and analyzed above had about 6 children per couple. Average number of children and other demographic conditions of the Chinese elderly a few decades later in the 21st century will dramatically differ from the elderly today. What would be the future perspectives of the family structure and living arrangement of elderly in China? This section intends to address this question.
According to an application to the multidimensional family household projection model developed by Zeng, Vaupel and Wang (, ), the family household structure and living arrangements of the Chinese elderly will change dramatically during the first half of the 21th century. Base population and other required data for the Chinese family household projection are derived from the 1% data file of the 1990 census, the 1988 two-per-thousand Fertility and Contraceptive Survey, and the 1985 In-Depth Fertility Survey (see , for more detailed resources, estimates and discussions). We present two major scenarios with the same medium fertility, but different (medium and low) mortality assumptions, which are described in the endnotes 1 and 2 of this paper. The assumptions about first marriage, divorce, remarriage, and leaving the parental home for the two scenarios are the same. The rural and urban propensities of divorce are assumed to increase from 6 and 10 percent in 1990 to 15 and 25 percent in 2050, respectively [Note 9]. By the middle of next century, the propensities of remarriage after divorce and widowhood are assumed to decline by 20 to 50 percent in rural and urban areas. The rural and urban proportions of parents who have at least one married child but do not live with any of them are assumed to increase by 30 to 45 percent. The proportion of population residing in urban areas is assumed to increase from 26.5 percent in 1990 to 70 percent in 2050. It should be emphasised that the nature of the above assumptions in our scenarios is purely for policy analysis to answer "what if" questions, and we do not intend for precise predictions.
Figure 4 and 5 present the projected percent of total population that are elderly by types of living arrangement in the rural and urban China from 1990 to 2050. Percent of elderly age 65+ living alone among the total population in 2030 would be 5 and 6 times as large as that in 1990 in the rural areas and urban areas, respectively. In 2050, the percent of the elders age 65+ living alone would be 11 and 12 times as large as that in 1990 in the rural and urban areas, respectively. The increase of the percent of the oldest old age 85+ living alone in 2030 and 2050 would be even much more dramatic. As compared with 1990, and under the medium and low mortality scenarios, percent of elders aged 85+ living alone would be 11-15 times and 10-14 times larger in 2030 in the rural and urban areas, 60-85 times and 46-65 times larger in 2050 in the rural and urban areas, respectively. Similarly, percent of the elders living with spouse only will also increase dramatically. The percent of elderly living with children, and in the institution will also increase due to very large increase in the total percent of the elderly population, but to a much less extent (see Figures 4 and 5). Moreover, there would be a total of 89 million or 96 million elderly age 65+ living alone in the year 2050, under the medium and low mortality assumption, respectively. Some 22 or 38 million of this group would be aged 85 and older.
Note that such dramatic increases of percent of the elderly, especially the oldest old, living alone or living with spouse only among the total population are largely due to dramatic increase of the total percentage share of the elderly. For example, the total percent of the oldest old age 85+ in 2050 would be 21-34 times as large as in 1990 in the rural areas, 14-22 times as large as in 1990 in the urban areas, which is a result of low fertility (average number of children of elderly who reach age 65 after 2010 would be less than one third of that of those elderly in 1990), and increasing longevity. It is also due to decrease in co-residence between parents and married children, and migration away of children, while increase in divorce rates may have very small effects on percent of elderly living alone [56, p77].
Although future developments in technologies of communication, transportation, housing and social services will change what it means to "living alone" today by the year 2050, 89-96 million elderly living alone and 22-38 million among them are over age 85 are large burdens for the society. Clearly, the middle of the next century will be a difficult time for the country due to serious problems of population ageing.
Family Dynamics of 63 Million (in 1990) to more than 330 Million (in 2050) Elders in China
Zeng Yi, Linda George
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871