Age Structure and Dependency

Updated About content Print Article Share Article
views updated


Every individual is some particular age. Populations are collections of individuals, and rather than being some particular age, populations are characterized by the frequency distribution of the ages of the individuals who constitute them. This is called the population age distribution or age structure. The age structure can be summarized in various ways, for example, by the average or median age of the population. A population with a low median age is called "young," and one with a high median age is called "old." One can also define life cycle stages, such as youth, working age, and old age, and describe the age structure by the percentage of the population in each of these categories, for which various age boundaries are used. Thus one can speak of the proportions of the population below age 15 or 18 or 20; proportions between one of these ages and ages up to 60 or 65; and proportions above ages 60 or 65. With reference to these same life cycle stages, so-called dependency ratios may be calculated. The ratio of the elderly to the working age population is the old age dependency ratio, the ratio of youth to the working age population is the child or youth dependency ratio, and the sum of these two is the total dependency ratio.

Some of these measures are shown in Table 1 for the years 1950 to 2000, along with United Nations projections to the year 2050, for the more developed countries (DC) and the less developed countries (LDC), the countries being classified according to their economic status in 2000. In every year, the DCs have a higher median age than the LDCs, with a smaller proportion of children and a higher proportion of the elderly. Evidently, aging has already affected the LDCs as well as the DCs; the phenomenon is not restricted to the industrial nations. The aging of the DC populations is shown by the projected increase in their median age by about 18 years from 1950 to 2050, and by the corresponding increase for the LDC populations by about 14 years.

Stable Population States

The age structure of a population is shaped by the past history of births, the past age distributions of deaths, and the age characteristics of net migrations. Consider first the case in which net migration is always zero at all ages (a closed population, on net), and age-specific fertility and mortality rates have been unchanging for a long time–about a century or more. In this case, it can be shown that the population will converge to a so-called stable state, in which the percentage age distribution is constant over time, and the population and every age group grow at the same constant exponential rate. Furthermore, this stable population age distribution is independent of the population age distribution that existed a sufficiently long time ago–again, about a century. That is, the particular features and shape of the initial age distribution tend to be forgotten as time passes, and the eventual age distribution depends only on the constant age-specific fertility and mortality rates. Depending on those rates, a stable population can have a constant growth rate within a wide band of particular values, and the rate can be positive, negative, or zero. A stationary population is a stable population with a zero growth rate.

Members of a population who are now age x were born x years ago and have survived for years. In a stable population, the number of births grows at the exponential population growth rate. In a growing population, the number of births x years ago will be smaller than the number of births today, and the higher the population growth rate, the smaller will be the generation born x years ago relative to the generation born in the current year. The opposite will be true if the growth rate is negative and the population is shrinking. Indeed, the rate of population growth is the most important determinant of the age distribution of a closed, stable population. The age distribution, however, is also affected by mortality, which determines the proportions of


births that survive to each age x. The lower the mortality rate, the higher will be the proportions surviving from birth to older ages.

Nonstable and Irregular Population Distributions

For a given level of mortality, higher fertility will always be associated with faster population growth and therefore with a younger stable age distribution in a closed population. Mortality differences, however, have two contradictory effects. On the one hand, lower mortality makes a stable population older by increasing the proportions surviving from birth to older ages. On the other hand, lower mortality tends to make a stable population younger, because it raises the population growth rate (for a given level of fertility). When the initial level of mortality is high, the net outcome of lower mortality is to make a population younger. When the initial level of mortality is low, lower mortality tends to make a population older. For intermediate initial levels, the effects of lower mortality are mixed, sometimes leading to higher proportions of both youth and of elderly and sometimes hardly changing the age distribution at all. These different effects of mortality decline are observed in real-world situations as well as in the hypothetical stable populations. For example, Table 1 shows that the LDC population in 1975 had a younger median age than it did in 1950, as well as a higher proportion of children and lower proportion of elderly. Mortality declined rapidly from 1950 to 1975, illustrating how falling mortality can make a population younger.

Many actual population age distributions are highly irregular rather than smooth and geometric like those of stable populations. Irregular distributions can come about in several major ways. The populations of a number of industrial countries, for example, experienced a baby boom from the late 1940s through the mid-1960s, followed by subsequent baby busts. These changes in fertility created large bulges and hollows in the population age distributions as the affected birth cohorts reached higher ages. The changing relative sizes of cohorts had important consequences for average wages, unemployment rates, and prospects for promotion, and they eventually will exert differential fiscal pressures through public pension and health-care systems. Population age distributions can also be heavily marked by traumatic events such as major wars or, for example, China's disastrous famine resulting from the Great Leap Forward (an economic plan launched in the late 1950s). Such crises cause heavy mortality that is sometimes concentrated at certain age and sex groups, and they also lead to sharp reductions in fertility and therefore in the size of generations born during and immediately after the crisis. When a population age distribution is strongly distorted by influences such as these, the distortions simply age with the population, moving up from younger to older ages as time passes. For example, the effects of both World War I and World War II are still clearly apparent in the age pyramids (as the conventional graphic representation of age distributions are labeled) of many European countries. A third cause of irregular age structure is age-focused patterns of immigration and emigration. These are more frequently seen in a sharp differential at the local rather than the national level. Often such patterns occur in towns with universities, prisons, army bases, or retirement communities. A fourth cause is emigration of the younger population from some rural areas, which leaves behind an elderly population. Some characteristic age pyramids are shown in Figure 1.

Not only do distorted age structures tend to persist over time as the population ages; they also can be transmitted to the stream of new births through the processes of reproduction, as echoes. If some generations are unusually large, because of an earlier baby boom, for example, then when the members of those generations enter their peak reproductive ages they will themselves generate an unusually large number of births, given typical levels of fertility. In this way they create another bulge in the age distribution, albeit a somewhat smaller one than the first. Formal analysis shows that populations with nonstable age distributions but which are subject to constant age-specific fertility and mortality will tend to move in cycles about one generation (25 to 30 years) long as they converge to stability. This result can be generalized to populations that are constantly subjected to random perturbations. Historical time series of baptisms often show evidence of such cycles. Sometimes, however, there is negative feedback in the renewal process, so that large generations of young adults experience adverse economic conditions and consequently have lower fertility and give birth to smaller, rather than larger, generations. In this way cycles longer than one generation may be generated; these are known as Easterlin cycles.

Consequences of Population Age Distributions

In many contemporary societies, there is particular interest and concern about the process of population aging and rising old age dependency ratios, because these factors will affect the cost per worker of supporting the elderly retired populations. Some analysts suggest that governments should seek to raise fertility in order to reduce and postpone population aging. Others propose to alleviate population aging through increased immigration, because immigrants are typically younger than natives and have higher fertility. But analysis shows that any gains from such a policy would be short lived and smaller than most people expect, because immigrants grow old themselves and require support. Only constantly accelerating rates of immigration achieve much effect, and such policies cannot be sustained for long. As an example, the U.S. Bureau of the Census reported the old age dependency ratio and the median age in 1995 to be .21 and 34.3 years, respectively. The Census Bureau projected these quantities to the year 2050 under low and high immigration assumptions that differed by more than a million immigrants per year. With low immigration, the Census Bureau projected that, by 2050, the old age dependency ratio would rise to .38 and the median age to 38.8 years. With a million more immigrants per year, these figures were projected to be only slightly lower at .35 and 37.6 years. The additional 55 million immigrants would have a big effect on population size but only a small effect on population aging.


Population age distributions have a range of socioeconomic consequences, because people's behaviors, abilities, and entitlements all vary with age. These variations reflect biological changes over the life cycle, but in addition they reflect somewhat arbitrary institutional age categories and individual choices in response to various preferences and incentives. On the biological side, it appears that health and vitality at the older ages have been increasing over time, so that working life could be extended to older ages. This option, however, is apparently not commonly viewed as desirable, because actual ages at retirement have declined by five to ten years over the twentieth century in industrial countries. These declines are due in part to the desire for more leisure as incomes rise, pensions becoming more common, and financial institutions making saving easier. It is also clear, however, that the structures of both public and private pensions provide strong incentives for early retirement, and that this has contributed to the decline. This trend slowed or slightly reversed in the 1990s in many countries.

The boundary age for dependency in youth also reflects a number of factors, most notably the length of time spent in formal education, that influence the age at which the workforce is entered. These age boundaries for youth and old age correspond roughly to directions of flows of intergenerational transfers, through the family and through the public sector. The public sector in industrial countries provides pensions and health care for the elderly and education for youth. The size of these public transfer programs for the young and the old swamps the transfers to those of working age. Private transfers in most industrial countries consist mainly of parental support of children and assistance from the elderly to their adult children and grandchildren through transfers made prior to death and through bequests.

As the population age distribution changes, pressure on those who make these transfers is relaxed or intensified. Population aging often goes with reduced fertility, resulting in not only a reduced need for public and private transfers to children but also a greatly increased need for transfers to the elderly for health care and pensions. A generalized fiscal dependency ratio can be calculated as follows. The numerator is determined by weighting each population age group by the costliness of public transfers it receives, with the denominator equal to the level of taxes that age group pays. Holding these weights fixed, one can then see how the fiscal dependency ratio changes over time for demographic reasons alone. For the United States, for example, the federal fiscal support ratio has been projected to increase by 56 percent from 2010 to 2075. This means that in order to provide the same set of age-specific benefits, age-specific tax rates financing intergenerational transfers would have to be raised by 56 percent. Alternatively, if age-specific tax rates financing transfers were held constant, then benefits received as transfers would have to be scaled back by 36 percent.

See also: Aging of Population; Cycles, Population; Generational Accounting; Intergenerational Transfers; Oldest Old.


Coale, Ansley J. 1972. The Growth and Structure of Human Populations: A Mathematical Investigation. Princeton: Princeton University Press.

Lee, Ronald. 1994. "The Formal Demography of Population Aging, Transfers, and the Economic Life Cycle." In The Demography of Aging, ed. Linda Martin and Samuel Preston. Washington, D.C.: National Academy Press.

United Nations Department of Economic and Social Affairs, Population Division. 2002. World Population Aging, 1950–2050. New York: United Nations.

Ronald Lee