FERTILITY DECLINE AND CHANGING LIVING ARRANGEMENTS
Paulo Saad, United Nations (UN) Economic Commission for Latin America and the Caribbean, offered his thoughts on the relevance of the study of living arrangements of older people in Latin America and the Caribbean (LAC). He illustrated his considerations with information from a previous UN study on the living arrangements of older people around the world and presented preliminary results from a just-begun study on recent trends of living arrangements of older people in the region.
The initial question he posed is why the study of living arrangements of older people is particularly relevant in LAC. In the first place, he said, informal support plays a very important role in the well-being of older people wherever the provision of adequate formal support is absent or deficient, as in most countries of the region. And coresidence is usually a very important element in this intrafamily support transfer system. A second point he made is the increasing need for support of older people because of the aging of the population; not only do more people survive to older age, but those who survive tend to live longer.
At the same time, he said, several factors tend to constrain the ability of families to provide support to older members. One is the increasing labor force participation of women. Women have traditionally been the main providers of support for the older family members, but many women now
have less time for doing this because they are participating much more in the labor force. A second point is lower fertility and the decreasing number of children, which decreases the options for coresidence for older people. There is also increasing mobility of the younger generation, such that children are moving away from parents and older relatives, which also decreases the possibility of coresidence.
In studying living arrangements, it is important to consider three main elements, according to Saad. One is the consequence of different living arrangements for the well-being of older people. Some arrangements may provide older people more assistance with activities of daily living (ADL), for instance. Other arrangements might increase the flow of direct support (money, goods, services) to older people and/or give them more satisfaction with life in general. The second element is the determinants of these living arrangements, especially in order to inform public policy to better address the consequence of different living arrangements. The third element has to do with classification: how to choose a small set of meaningful categories of living arrangements from a potentially large set of choices.
Saad explained that a given living arrangement is the result of a balance between preferences, not only of older people—whether they prefer to live alone or prefer to live with their children—but also the preferences of other family members. There are also cultural influences on these preferences. An older person living alone, for instance, may be seen as a sign of abandonment by children. These preferences are constrained by other factors, including kin size and composition, physical feasibility of living independently, and financial feasibility. He then noted the importance, for classification purposes, of understanding the relationship between an older person and other members of a family and/or household. Often, a researcher only has information from general surveys and censuses that show an older person’s relationship to the head of household. If the older person is the head of a household, classification is straightforward. But for those older people who are not heads of household, there will be uncertainty as to the real living arrangement of this person. Ideally, information about the relationship of an older person to each member of a given household would be available, but this information must come from a specialized survey.
Saad presented data from a comprehensive UN study of more than 100 countries worldwide, based largely on census data (United Nations Department of Economic and Social Affairs, 2005). He showed that there were large differences in living arrangement patterns between developed and developing regions. For example, three-quarters of older (age 60+) people in
developing regions lived with a child(ren) or a grandchild, compared with one-quarter of older people in more developed regions. Coresidence of older people with children is lower in Latin America than in Asia and Africa, but much higher than in Europe. In all regions of the world, older women are more likely than older men to live alone. But if looking only at unmarried older people, men are more likely than women to live alone, because most older men are still married and most older women are unmarried. In all regions of the world, there is an increased likelihood over time for older people to live alone.
Looking at economic and social differentials for the entire set of countries in the UN study, Saad noted that in countries with very low levels of development, coresidence with children is associated with higher social and economic status of the older person. Among countries at moderate levels of development, this association tends to disappear or even reverse; living with children starts being associated with lower social and economic status. One conclusion is that, in the poorest countries, older people living alone tend to be an especially disadvantaged group. A second finding, Saad said, was that among unmarried older people, although lower levels of socioeconomic and physical vulnerability increased the probability of living alone, coresidence substantially increased the probability of receiving support with ADL for those in need of such support. Among married older people, neither coresidence nor the number of children affect the probability of receiving assistance with ADL, which means that the spouse is the primary provider of support in the case of married older people.
Saad concluded by mentioning a new study that seeks to update the earlier UN effort using 2010-round census data and focusing primarily on LAC, which will afford a look at trends over time. He also mentioned several preliminary results from a multivariate analysis: higher age first increases and then decreases the probability of independent living for older people; for unmarried older people, the probability of living alone is higher among men; for married older people, the probability of living with spouse only is higher among women; and higher educational level increases the probability of independent living (i.e., alone or with spouse only).
SOCIAL MOBILITY ACROSS GENERATIONS IN LATIN AMERICA
Florencia Torche, New York University, tackled the subject of intergenerational mobility, which is defined as the association between parents’
well-being and adult children’s well-being. A high association means very little mobility, a situation determined by the accidents of birth and by what parents had or did not have. In this light, the concept of mobility is said to capture equality of opportunity. An immobile society is a society that does not give equal opportunity to its members, she stated. As such, mobility is likely related to inequality, and there are several channels that induce this association.
First, Torche said, inequality implies high returns to schooling, so that people with high education have higher incomes. High inequality also implies less progressive investment in the human capital of the next generation, preventing mobility. Inequality strengthens the political influence of the elite, the wealthy, thereby potentially reducing the scope for redistribution. Inequality also induces residential segregation. Based on these pathways (and there may be many others), she said one might hypothesize that high inequality will result in very limited mobility opportunity. If that is the case, noted Torche, then Latin America, which is highly unequal, should feature very little mobility.
To address this hypothesis, Torche presented a series of stylized facts. The first is that Latin America is the most unequal region of the world and has been since records have been kept from the mid-20th century. On average, Latin America has much higher Gini indices than Asia, the developed nations, Africa, and Eastern Europe and Central Asia (Gasparini and Lustig, 2011). Secondly, Latin America is particular not only because of the level but also the pattern of inequality. The top 10 percent of the income distribution receives about half of the income in the region, compared with about one-third in the rest of the world (Bourguignon and Morrisson, 2002). Inequality by definition is related to concentration, but the Latin American case is extreme. She said a third and more positive fact is that inequality, again based on the Gini index, has declined slightly in the recent past as a result of educational expansion and progressive government transfers such as Opportunidades in Mexico and Bolsa Familia in Brazil (Lopez-Calva and Lustig, 2010).
Torche explained that researchers who study mobility essentially have used three sources of data. The first is the longitudinal panel survey that includes information on more than one generation. The second is administrative data, administrative records that match parents with their adult children. The third is cross-sectional surveys of adult respondents in which those adults report retrospective information about their parents. Unfortunately for analysis in the LAC region, only the third option is presently
available, and acceptable retrospective data generally are limited to Argentina, Brazil, Chile, and Mexico.
Returning to the question of the level of intergenerational mobility in Latin America, Torche said that the answer depends on the measures used. Using occupational class, the intergenerational association in Latin American countries is seen to be as low or lower than in advanced industrial countries, meaning that there is no less and perhaps more mobility. Looking at economic mobility, however, the association between parent’s earnings and adult children’s earnings appears much greater in Brazil and Chile than in the more industrialized world. The same is true with regard to educational mobility—looking at the association between parents’ schooling and adult children’s schooling, the intergenerational association (the opposite of mobility) is much higher in Latin America than in the rest of the world (Hertz et al., 2007). In terms of the pattern of mobility, she presented transition matrices by earnings quintiles for Brazil and Chile, which suggest that the pattern of economic mobility is similar to the pattern of inequality—income concentration at the top, as well as strong intergenerational immobility in the top income quintile.
Torche then raised the issue of how people understand mobility. She said that scholarly research tends to look at the association between parents and children net of any economic growth across generations. In other words, if everyone is doing better or everyone is doing worse from one generation to the next, this is removed from the picture. But what people perceive in their lives is change that includes the economic cycle. She presented data from a study in Mexico that indicate that even among people whose relative economic standing deteriorated compared to their parents’ relative standing, 40 percent said they had moved upwards and 50 percent said they were the same, because they benefitted from overall economic growth.
To summarize, Torche noted that available data suggest there is much less economic mobility, educational mobility, and equality of opportunity in Latin America than elsewhere, but that the number of LAC countries with good mobility data is small. She explained that every Latin American country now has at least one household survey that is nationally representative or almost so. Therefore, it would be possible to add a small module with retrospective questions about parents in those surveys, she suggested. A modest set of questions could enable the construction of indices of economic well-being for the parental and current generations that would enhance the analysis of mobility throughout the region. She said it will also be important to move from bivariate intergenerational associations to
the more important question of how mobility works; that is, the processes involved when socioeconomic well-being is transmitted or not transmitted across generations.
Finally, she stressed that education in Latin America is by far, and more so than in other parts of the world, the most relevant factor explaining intergenerational mobility (Torche, 2014). This is both good and bad news, according to Torche. The good news is that education is highly amenable to policy intervention; there is widespread support for increasing and equalizing educational opportunity. The bad news is that advantage is reproduced across generations in Latin America through education. Wealthy parents can afford more and better education for their children, and education in turn pays off in the labor market. The strong mediating role of education creates a situation of “inherited meritocracy,” intergenerational persistence that is legitimized and naturalized by educational attainment, she said. This situation emerges from the strong barriers that disadvantaged families in Latin America face with regard to accessing both quantity and quality of education. It is a meritocracy because people do well based on their education, but it is inherited because education depends on the family of origin. This has been documented in Latin America, and highlights the importance of moving from a bivariate perspective to one that better explains how socioeconomic advantage is replicated across generations.
RAPID SOCIAL CHANGES AND IMPLICATIONS FOR AGING
Cassio Turra, Federal University of Minas Gerais, focused his remarks on the empowerment of older people in Brazil during the preceding decades. He began with some institutional and demographic context for Brazil, noting that Brazil is a very large country with a large population, a large economy, and high levels of inequality that have persisted for many years. There is great racial and regional diversity. Brazil is characterized by very rapid demographic transition and other transitions—educational transition, labor market transition, and the expansion and consolidation of social welfare programs—which are under way at the same time.
He presented data on two of the most important indicators of demographic transition, infant mortality and fertility. The infant mortality rate declined by a factor of about 10 over the last 50 years, faster than in China and Mexico. The decline in infant mortality was followed by a decline in the total fertility rate, which was also very fast, the rate plunging from more than 6 children per women in 1960 to 1.9 in 2010, and the decline
continues. These trends have consequences for population aging, family size, and family structure. At the same time, there was a large increase in the proportion of women in the labor force. In terms of education, he said that Brazil was late in making investments in education; in 1960, half of the adult female population (ages 30–69) had no schooling.
In the late 1980s, Brazil was ending a military government and preparing a new constitution. Part of the motivation for the new constitution was to build a large social welfare program, and Brazil undertook simultaneous investments in education, in social security, and in health care. There were sharp increases in some program coverage rates between 1988 and 1991, particularly in noncontributory programs. This was a challenge, Turra said, because Brazil was spending money on the old and the young at the same time. This was in contrast to many developing countries that invested first in children, then in economic expansion and improved productivity, and then consolidation of social security programs. As to why Brazil expanded social security protection to all older people at once, Turra explained that at least part of the rationale was a “social inequality” argument. Looking at the 1930 birth cohort (i.e., people who were turning age 60 when the new constitution was written), 70 percent had lived in childhood poverty, the female labor force participation rate for this cohort was about 20 percent, and the average educational level was about 4 years of schooling. Therefore, there was a high probability of being poor at older ages in 1988, and the decision was made to provide a minimum benefit to all elders.
With regard to outcomes stemming from policy changes, the expansion of social security coverage to all older people had a large period effect on poverty rates. Turra explained that there were sharp declines in poverty after 1988 for six birth cohorts of older Brazilians (born between 1901 and 1932). There were also declines in poverty rates among children, but these were based more on cohort and intergenerational effects. In terms of income inequality, the demographic transition and expansion of benefits among older people resulted in a greater concentration of older Brazilians in higher income deciles. Echoing Saad’s earlier remarks, Turra noted an increasing prevalence of older people living alone or in one-generation households between 1970 and 2010, which is related to the expansion of social security and probably indicates that older people have increasing autonomy within households. He also reported that there has been increasing financial independence among older people living in multigenerational households; those who live in multigenerational households now hold a larger share of household income. There is a need, said Turra, to do analyses
on the consumption side of the equation, which is important because older people spend a good deal on health services.
In conclusion, he reiterated that Brazil is in the midst of dramatic demographic and institutional change, and the challenge is how to keep social progress occurring with a growing proportion of the population aged 60 and older. Old-age support systems (social security, health, and long-term care) must take into account population aging and smaller family size. The definition of “aged” needs to change over time; in his opinion, part of Brazil’s fiscal solution involves having a minimum age of retirement that is flexible and can be increased in response to longevity increases. He noted that this has been very difficult to implement in Brazil, because Brazilians retire at a relatively early age. He mentioned the importance of rethinking issues surrounding fertility levels and international migration. Consideration needs to be given to the level of public transfers to older people, which may be too high to encourage savings. To better assess these and other issues, Brazil needs more longitudinal data, he said, and an integration of different source of data: administrative records, other linked data, longitudinal studies, repeated cross-sections, and census data.
Timothy Miller, United Nations Economic Commission for Latin America and the Caribbean, added a different twist to the workshop discussion, shifting the focus from aging populations to aging economies. He noted that economies can also grow old, and defined an aged economy as one in which consumption by older people exceeds that of children, meaning that more resources are going to elders than to children. Using this metric, he pointed out that there were no aged economies in the world until the late 1990s, when Japan set the pace. By 2010, 18 other countries were in this category, and the number is expected to reach 77 aged economies by 2040. This is a new phenomenon that will eventually come to dominate the world, and one about which there is little historical information.
One of the best inventions that economists created in the last century, said Miller, was the construct called national accounts, which produce estimates of gross domestic product (GDP) and other indicators. This concept has been broadened to what are called national transfer accounts (NTA). NTA do two things. The first is to measure the age distribution of economic activity, for example, labor earnings by age. The second is to measure the transfer of resources between population age groups within a national
economy, and whether that transfer is occurring through the government, through financial markets, or through families. Thus, while national accounts address relationships between economic sectors, NTA address relationships between population groups (young and old, men and women, etc.). The NTA concept has been expanded to include (1) national time transfer accounts, because when looking at the distribution of resources and economic activity, time must be taken into account and not just spending; and (2) national inequality accounts (e.g., a focus on rich and poor).
To recap, Miller said that national accounts focused on a very important problem of the previous century, the problem of growth. NTA can also be used to look at growth, but one of the main motivations behind NTA is to look at distributions. The main themes are population aging, inequality, and sustainability. NTA are based on household surveys, administrative data, and the national accounts themselves. More than just an analytical framework, he explained, there is also an NTA network of researchers with more than 100 investigators in 46 countries.1 NTA methodology and results may be found in a United Nations manual (United Nations Department of Economic and Social Affairs, 2013).
Miller then presented summary data for economic lifecycle by age. Figure 6-1 shows information for an average of 23 countries. The dashed line is consumption, which rises during the younger ages and is rather flat after age 20. The solid line represents the average labor earnings profile. The “problem” for societies is identified by the two gray zones, the two periods of economic dependency. The core problem is how economies and countries finance consumption during these two periods of early and late life. A period of dependency in young life and a period in old life are characteristic of human societies, whether industrialized or agrarian. Societies need to solve the problem using a balancing equation, which can be thought of as a simple accounting framework. On the one side is consumption: goods consumed, savings, taxes paid, and family transfers provided. On the other side of the balancing equation are labor earnings, asset income, public transfers received, and family transfers received. Support of consumption can only occur through these four channels. Each society will adopt a different combination of these four channels to satisfy consumption needs.
To illustrate, Miller contrasted the roles of labor income, family transfers, public transfers, and private assets in supporting consumption in Japan, Mexico, and Sweden. In Mexico, families are the primary supporter
_________________________
1The NTA project website is http://www.ntaccounts.org/web/nta/show/ [August 2015].
FIGURE 6-1 Lifecycle deficit in youth and old age.
NOTE: Average of data for 23 national transfer account countries, circa 2000.
SOURCE: Lee and Mason (2011).
of youth (ages 0–19) consumption, with some role for public transfers and some role for labor income coming from older youth. At the other end of the spectrum, older people on average give more resources to family than they receive from family. This is a common pattern around the world, seen also in India, Indonesia, Philippines, and South Africa. In Mexico, the dominant form of family transfers is assets, and older Mexicans are using up their assets in retirement. In Japan, the government takes a more active role in supporting youth, and there is greater reliance on public transfers and less on assets. This also is common to Chile, Costa Rica, Germany, and Peru. In Sweden, the government plays as large a role as does the family in terms of support to children. Raising children in Sweden is a very cooperative endeavor, and older age consumption in Sweden in characterized by the dominance of public transfers. Older Swedes use relatively few of their assets in retirement.
Miller noted that one global pattern is that as net public transfers to older people rise, net public transfers to youth also tend to rise. There are, of course, numerous outliers, such as Brazil in the 1990s with very high transfers to older ages (as Turra discussed earlier). Miller also noted that consumption differences within countries are as great as differences between countries, and that intranational analyses are important. He concluded with some data on rising health care expenditures related to population aging and
growing incomes. Contrasting rich and poor countries, he noted that patterns of health expenditure at younger ages were quite similar, but that there was an enormous difference at older ages. Over time, health expenditures in poorer countries are likely to increase because there will be more people in older age groups and countries will become wealthier. Miller suggested that the world will see a continuing transition away from economies in which 20–30 percent of GDP is devoted to food to economies in which 20–30 percent of GDP is devoted to health care.
This page intentionally left blank.
