Household Structure and Older Persons

Abstract This research explores the life circumstances of older persons (aged 60 years and above), focusing on the sociodemographic and socioeconomic conditions of those who live alone. We situate the living arrangements of older persons within the global context of changing household structures in 76 countries from all regions of the world. Older persons who live alone are among those most likely to need governmental and other forms of social support. The analysis presented here is crucial for supporting policy responses to the needs of older persons, including the special attention they require during the current COVID-19 crisis. It also supports the operationalization of the Madrid International Plan of Action on Ageing (MIPAA)(United Nations, 2002), the realization of United Nations Principles for Older Persons (United Nations, 1991), and the broader framework of the Programme of Action of the International Conference on Population and Development(ICPA-POA).


Introduction
This booklet explores the life circumstances of older persons (aged 60 years and above), focusing on those who live alone, in the context of changing household structures in 76 countries from all regions of the world. The analysis is based on harmonized international census data obtained from the Integrated Public Use Microdata Series (IPUMS) at the University of Minnesota. The analysis presented here is crucial for supporting policy responses to the needs of older persons, including the special attention they require during the current COVID-19 crisis, as they are disproportionately more likely to be infected and die from COVID-19 related complications than younger persons. It also supports the operationalization of the Madrid International Plan of Action on Ageing (MIPAA) (United Nations, 2002) and the realization of United Nations Principles for Older Persons (United Nations, 1991), and within the broader framework of the Programme of Action of the International Conference on Population and Development (ICPD-POA).
The focus on older persons is due to all regions of the world experiencing population ageing, albeit at different rates. Globally, projections show that the proportion of older persons aged 60 and above will increase from 13.5% to 21.4% between in 2020 and 2050. Regionally, Latin America and the Caribbean and Asia and will experience the fastest growth in the proportion of older persons, from 13% to 24.9% and 13.1% to 24.1%, respectively. Africa will experience slowest growth, from 5.5% to 8.7% whilst in North America and Europe the proportions of older persons will increase from 23.1% to 28.5% and 25.7% to 35.0%, respectively. 1 Older people living alone receive special attention in this booklet because they can require targeted social and policy support (United Nations, 2005). "Older persons that live alone are more likely to need outside assistance in the case of illness or disability, are at greater risk of social isolation and, even in countries with well-developed systems of social security, are disproportionately likely-especially older women-to be poor" (ibid., p. 16).
Alongside global population ageing, other megatrends include changing household structures, urbanization, globalization, increasing incomes and rising education levels, all which are reshaping the contexts in which older persons live (United Nations, 2017), and the amount and nature of the support they enjoy.
Household structures and relationships within households have become increasingly important for policy analysis because they have implications for support and welfare (Mberu, 2007). The United Nations defines a household as "a group of persons who make common provision of food, shelter and other essentials for living" (United Nations Population Division, 2017). A household is, therefore, a fundamental socio-economic unit in human societies. While there are critiques of this definition (see Kriel andothers, 2014, andothers, 2015, for example), this "housekeeping" conceptualization of a household is widely used in censuses (United Nations, n.d.) and international large-scale household surveys such as the Demographic and Health Survey. This makes it crucial to situate analyses of socioeconomic, demographic and cultural phenomena in the context of the household. Further, the household is the nucleus of demographic, socioeconomic and cultural processes, where decisions are made about childbearing, division of labour, education, labour force participation, migration, health care, child-rearing, and the teaching of core values and norms.
The same processes in turn shape the structure of households. The combination of fertility, mortality, nuptiality, and migration levels and trends, for example, affect the age structure of a population, and can define household structure, composition and size (Jamba andMmatli, 2016, andBongaarts, 2010). Cultural factors also determine the structure and composition of households, including the formation of new ones, within the context of life-course events such as marriage, divorce, widowhood, orphanhood, employment, etc. (Dommaraju and Tan, 2014).
The rest of this booklet is organized as follows. Section 2 presents a brief overview of concepts used to define household structures. It provides evidence of changing household structures across the world. Section 3 examines trends in the living arrangements of older people. Section 4 examines the sociodemographic characteristics of older persons who live alone and some of the correlating major socioeconomic characteristics. Section 5 offers brief conclusions. The analysis presented here is crucial for supporting policy responses to ageing, including under the Madrid International Plan of Action on Ageing (United Nations, 2002) and the United Nations Principles for Older Persons (United Nations, 1991), and within the broader framework of the ICPD-POA.

Defining household structureoverview of concepts
The widely used typology of household structure characterizes families by type. This is not surprising because families and households are separate but related entities that describe how a population is organized within communities. Multiple criteria to define household structure include the relationships of household members to the head of household, characteristics of the household head (child-headed households, grandparent-headed households and female-headed households, for example); household size (single-person and multi-person households); or simply type of family household (non-family and family household).
The demographic transition theory postulates that, as a corollary to birth and death rates moving from higher to lower levels as societies modernize, populations also experience a shift from predominantly complex extended-family households to simpler nuclear and conjugal ones (Lesthaeghe, 2014). The latter are characterized by smaller household sizes and married couples living with their children, with fertility levels at around replacement level. This transition has been observed in most parts of Europe and North America. Since the 1970s, many countries in these two regions have, however, progressed to below replacement fertility. They have seen an emergence of new and diverse household structures that are not tied to familial or conjugal relationships, a disconnection between marriage and child bearing, and increasing proportions of one-person as well as single-parent households. This has been described as the second demographic transition.

Evidence of changing household structures
Figures 1 and 2 show evidence of a worldwide transition in household structure. The figures map regional trends in the proportion of single-person households and households with three or more people. Most African and Asian countries analyzed did not experience such rapid increases. Exceptions are Botswana and South Africa (Africa) and Israel (Asia), which have reached levels comparable to those in some countries in Europe and North America. Note: Each dot represents a country at a point in time. The dots are colour coded by region. The trend linesalso colour coded by regionrepresent unweighted regional averages of the proportions of the two types of households at a point in time, calculated using available data for the countries included in the analysis.  Figure 2 shows general declines in the proportion of large households (with three persons and above) in all regions between 1960 and 2010. Declines are again more pronounced in Europe and North America, and not so much in Asia and Africa. This could indicate an incomplete fertility transition and the preponderance of extended-family households in the latter.
Figure 2. Regional trends in the proportion (percentage) of households with three persons and above based on 76 countries with available data, 1960-2015 Note: Each dot represents a country at a point in time. The dots are colour coded by region. The trend linesalso colour coded by regionrepresent unweighted regional averages of the proportions of the two types of households at a point in time, calculated using available data for the countries included in the analysis.

Living Arrangements of Older Persons
Alongside the overall increase in the proportion of single-person households between 1960 and 2010, there was a corresponding rise in the share of older single-person households in Europe, Latin America and the Caribbean, and North America ( Figure 3).
In Africa and Asia, the proportion did not change, remaining below 5 per cent for most countries included in the analysis. Israel is an exception in Asia, having experienced an increase from about 5 per cent in the 1970s to 10 per cent in the mid-1990s. Note: Each dot represents a country at a point in time. The dots are colour coded by region. The trend linesalso colour coded by regionrepresent unweighted regional averages of the proportions of the two types of households at a point in time, calculated using available data for the countries included in the analysis.  Figure 4 shows that in Africa, Asia, and Latin America and the Caribbean, older persons predominantly live in households made up of three or more members, regardless of age group (60 to 69 years, 70 to 79 years, or 80 years and above). This reflects a high preponderance of extended family households in these regions.
In North America and Europe, on the other hand, older persons are more likely to live in two-person households (most likely with their spouses). In both regions, the proportion of older persons living in single person households is significant, ranging from about 2 in every 10 older persons (18 per cent) in the age group from 60 to 69 years, to nearly 4 in every 10 older persons (43 per cent) among older persons aged 80 years and above. The variation is widest in Europe and Latin America and the Caribbean. Among people aged 80 years and above, for example, the proportion ranges from a low of 30 per cent in Spain to a high of 61 per cent in Poland (Europe), and from a low of 7 per cent in Nicaragua to a high of 32 per cent in Uruguay (Latin America and the Caribbean). See also Figure A-1 in the annex.
The regional cross-sectional aggregates of the distribution of older persons living alone by age group (Figure 4) hide country-level variations in the likelihood of living alone by age cohort. Figure 6 shows that in Europe and North America, the proportion of older persons living alone more than doubles as people progress from their 60s to their 80s.

Percentage living alone
Percentage living in two-person households Percentage living in households with three or more people Figure 6. Age-specific trends in the living arrangements of older persons by type of household and region, 1980-2010

With whom do older persons live?
Among older persons who do not live alone, most Asians live with their children: 52 per cent live with one child, 15 per cent live with two or more children, and 32 per cent live with people other than their own children (Figure 7). Older women in Asia are most likely to live with their children. Seven in every 10 women aged 60 years and above live with one or more own children in the region. In contrast, North American women are the least likely to live with their children; only 30 per cent do so. Among men, co-residence with own children is highest in Africa, where 71% men aged 60 years and above live with their children, and lowest in North America where only 21% do so. Further sex disaggregation shows that in all regions, except Africa, women are more likely to live with their children than men. Patterns of living arrangements among older persons are likely associated with socioeconomic factors such as traditionalism, culture and attitudes towards older persons.

Characteristics of Older People Living Alone
Older women are more likely to be single-person householders than older men. Worldwide, 17 per cent of women aged 60 years and above live alone, compared to only 9 per cent of men of the same age ( Figure 8).
The sex differential in the likelihood of older people living alone cuts across all regions, but the gap is widest in Africa and Europe. In both regions, the proportion of older women living alone is more than double the share of older men.  Figure 9 shows that while there are normally more females than males at older ages, given that women tend to live longer, older women are more likely to live alone for reasons that go beyond longevity. In Asia, for example, among older people living alone, for every 100 females aged 60 to 69, 70 to 79 and 80-plus, there are 61, 48 and 42 males, respectively. These sex ratios are significantly different from those observed for the general population of older people, where per 100 females aged 60 to 69, 70 to 79 and 80-plus, there are 99, 91 and 74 males. Among African countries included in the analysis, older men live alone at roughly half the rate of older women. For Europe and North America, male-female sex ratios among older persons living alone decrease rapidly as age increases.  Analysis of data from a few countries with censuses after 2000 show that while the proportion of older persons living alone is low in Africa and Asia compared to other world regions, older people are more likely to live alone in rural areas than in urban ones ( Figure 10).  The size of the bubbles reflects the relative size of older persons aged 60 and above in each country.  Rural/urban ratio above 1 means more than 50% of older persons living alone are in rural areas and vice versa Older persons who live alone are more likely to be either divorced (particularly in North America) or widowed (particularly in Asia), as shown in Figure 11. This pattern is true for all regions and both sexes except older males in Europe, who are more likely to live alone even when married. Among them, 48 per cent are married. Never-married people account for the smallest share of older persons living alone. Never-married males in Latin America and Caribbean are a relatively higher proportion (22 per cent) of older males who live alone, compared to other regions. There was a general increase in the proportion of divorced older persons living alone in Europe, Latin America and the Caribbean, and North America.   In all regions of the world, the proportion of older women living alone is higher among those who never had children compared to those who have children.
Note: the bars do not add up to 100 because they do not represent the same universe of women. The green bars show percentages among women who have no children and the grey bars percentages among women who have children.   Figure 15. Percentage distribution of older women aged 60-plus who live alone by age at their first marriage, countries with available data Among countries that have collected data on age at first marriage, early marriage (before age 20) is associated with a higher likelihood of living alone at older ages among older women in Asia and Europe. In Africa, late marriage is associated with a higher likelihood of living alone ( Figure 15). Annex Figure A-3 presents data for Indonesia, and compares, by age at first marriage, the age-specific distribution of older women who live alone and who co-reside with other people. The graph shows that proportionately more women who married before age 18 are found among older women living alone, compared to those not living alone. Further research is needed to unravel the underlying causes for the relationship between age at marriage and living arrangements in old age, across regions. Table 1 shows that even after controlling for parity (i.e. number of children ever born), women who married before age 18 are generally more likely to live alone after age 60 compared to women who married at the ages of 18 to 24, 25 to 34 and 35 years and above, in all regions except Africa, and regardless of the number of children they have.
The combination of early marriage and childlessness increases the likelihood that women will live alone in older ages.

Conclusion
Population ageing is a global phenomenon. Shares of persons aged 60 and above are increasing in countries in all regions, albeit at different rates. Alongside age structure changes, all regions are also experiencing household structure changes. More people are living alone, and household sizes are shrinking.
The proportion of older single-person households has climbed in Europe, Latin America and the Caribbean, and North America. The predictors of living alone in old age include age at marriage, the number of children ever born, child survival, current marital status and the share of people aged 60 years and above in the total population.
Although based on relatively dated data, the analysis presented here is crucial for supporting policy responses to the needs of older persons, including the special attention they require during the current COVID-19 crisis, as they are disproportionately more likely to be infected, and die from, COVID-19 related complications than younger persons. The greater vulnerability of older persons to the disease is related to their fragile health and suppressed immune system, on the one hand, and their living arrangements on the other. Available data show that countries and regions within countries where the number of older persons is high are high-risk areas for COVID-19 cases and fatalities, and should be among the regions that receive priority in crisis response. Older persons that live alone may not receive the support they need especially if they get ill; living in institutional care facilities they might get necessary care, but they also face a higher risk of falling ill. However, in the case of COVID-19, living with or in proximity to children and younger adults also poses health risks to older persons, as COVID-19 is less likely to show symptoms in children and younger adults.
The analysis presented in this booklet also supports the operationalization of the Madrid International Plan of Action on Ageing (MIPAA) (United Nations, 2002) and the realization of United Nations Principles for Older Persons (United Nations, 1991), and within the broader framework of the Programme of Action of the International Conference on Population and Development (ICPA-POA).
This booklet does not in any way suggest that lack of co-residence among older members of society is bad or automatically translates to a lack of support or care. While it is possible that a substantial proportion of older single-person households may be vulnerable, and therefore require some form of targeted social protection and support, especially in the context of COVID-19 as shown above. It is also possible that many older people live on their own out of choice. They may enjoy supportive interactions with family members even if they do not co-reside with them. In this light, the overall aim was not to portray living alone among older persons as an undesirable outcome, but to show the changing life circumstances and characteristics of older single-person householders as an increasingly important demographic group everywhere in the world.