Demography is the study of human populations. It is an important part of sociology and the other social sciences because all persisting social aggregates—societies, states, communities, racial or ethnic groups, professions, formal organizations, kinship groups, and so on—are also populations. The size of the population, its growth or decline, the location and spatial movement of its people, and their changing characteristics are important features of an aggregate whether one sees it as a culture, an economy, a polity, or a society. As a result some anthropologists, economists, historians, political scientists, and sociologists are also demographers, and most demographers are members of one of the traditional social science disciplines.
A central question for each of the social sciences is: How does the community, society, or whatever, seen as a culture, an economy, a polity, or whatever, produce and renew itself over the years? Formal demography answers this question for aggregates seen as populations. This formal part of demography is fairly independent of the traditional social sciences and has a lengthy history in mathematics and statistics (Smith and Keyfitz 1977; Stone 1997; Desrosieres 1998). It depends on a definition of age and on the relationship of age to fertility and mortality. Those relationships certainly are socially conditioned, but their major outlines are constrained by biology.
Beyond pursuing formal demography, the task of most social scientist-demographers is detailing the relationships between demographic change and other aspects of social change. Working with concepts, methods, and questions arising from the traditions of each of the social science disciplines as well as those of demography per se, scholars have investigated the relationship between demographic changes and such social changes as those in the nature of families (Davis 1985; Sweet and Bumpass 1987; Bumpass 1990; Waite 1995), levels of economic growth (Johnson and Lee 1987; Nerlove and Raut 1997), the development of colonialism (McNeill 1990), changes in kinship structures (Dyke and Morrill 1980), nationalism and interethnic strife (Tietelbaum and Winter 1998), and the development of the nation-state (Watkins 1991).
FORMAL DEMOGRAPHY
At the heart of demography is a body of strong and useful mathematical theory about how populations renew themselves (Keyfitz 1968, 1985; Coale 1972; Bourgeois-Pichat 1994). The theory envisions a succession of female birth cohorts living out their lives subject to a schedule of age-specific mortality chances and age-specific chances of having a female baby. In the simplest form of the model the age-specific rates are presumed constant from year to year.
Each new annual birth cohort is created because the age-specific fertility rates affect women in earlier birth cohorts who have come to a specific age in the year in question. Thus, the mothers of a new cohort of babies are spread among previous cohorts. The size of the new cohort is a weighted average of the age-specific fertility rates. The sizes of preceding cohorts, survived to the year in question, are the weights.
As a cohort of women age through their fertile period, they die and have children in successive years according to the age-specific rates appropriate to those years. Thus, the children of a single birth cohort of women are spread over a sequence of succeeding birth cohorts.
The number of girls ever born to a birth cohort, taken as a ratio to the initial size of the cohort, is implicit in the age-specific fertility and mortality rates. This ratio, called the net reproduction rate, describes the growth rate over a generation that is implicit in the age-specific rates. The length of this generation is also implicit in the age-specific rates as the average age of mothers at the birth of the second-generation daughters. With a rate of increase over a generation and a length of the generation, it is clear that an annual rate of increase is intrinsic to the age-specific rates.
The distribution of the children of a birth cohort over a series of succeeding cohorts has an important effect. If an unusually small or large birth cohort is created, the effects of its largeness or smallness will be distributed among a number of succeeding cohorts. In each of those succeeding cohorts, the effect of the unusual cohort is averaged with that of other birth cohorts to create the new cohort’s size. Those new cohorts’ ”inherited” smallness or largeness, now diminished by averaging, will also be spread over succeeding cohorts. In a few generations the smallness or largeness will have averaged out and no reflection of the initial disturbance will be apparent.
Thus, without regard to peculiarities in the initial age distribution, the eventual age distribution of a population experiencing fixed age-specific fertility and mortality will become proportionately constant. As this happens, the population will take on a fixed aggregate birth and death rate and, consequently, a fixed rate of increase. The population so created is called a stable population and its rates, called intrinsic rates, are those implicit in the net reproduction rate and the length of a generation. Such rates, as well as the net reproduction rates, are frequently calculated for the age-specific fertility and mortality rates occurring for a single year as a kind of descriptive, ”what if’ summary.
This theory is elaborated in a number of ways. In one variant, age-specific rates are not constant but change in a fixed way (Lopez 1961). In another elaboration the population is divided into a number of states with fixed age-specific migration or mobility among the states (Land and Rogers 1982; Schoen 1988). States may be geographic regions, marital circumstances, educational levels, or whatever.
In part, the value of this theory is in the light it sheds on how populations work. For example, it explains how a population can outlive all of its contemporary members and yet retain its median age, percent in each race, and its regional distribution.
The fruit of the theory lies in its utility for estimation and forecasting. Using aspects of the theory, demographers are able to elaborate rather modest bits and pieces of information about a population to a fairly full description of its trajectory (Coale and Demeny 1983). Combined with this mathematical theory is a body of practical forecasting techniques, statistical estimation procedures, and data-collection wisdom that makes up a core area in demography that is sometimes called formal demography (United Nations 1983; Shryock and Siegel 1976; Pollard, Yusuf, and Pollard 1990; Namboodiri 1991; Hinde 1998).
DEMOGRAPHIC DATA
Generating the various rates and probabilities used in formal demography requires two different kinds of data. On the one hand are data that count the number of events occurring in the population in a given period of time. How many births, deaths, marriages, divorces, and so on have occurred in the past year? These kinds of data are usually collected through a vital registration system (National Research Council 1981; United Nations 1985, 1991). On the other hand are data that count the number of persons in a given circumstance at a given time. How many never-married women age twenty to twenty-four were there on July 1? These kinds of data are usually collected through a population census or large-scale demographic survey (United Nations 1992; Anderson 1988; Anderson and Fienberg 1999). From a vital registration system one gets, for example, the number of births to black women age twenty. From a census one collects the number of black women at age twenty. The division of the number of events by the population exposed to the risk of having the event occur to them yields the demographic rate, that is, the fertility rate for black women age twenty. These two data collection systems—vital statistics and census—are remarkably different in their character. To be effective, a vital statistics system must be ever alert to see that an event is recorded promptly and accurately. A census is more of an emergency. Most countries conduct a census every ten years, trying to enumerate all of the population in a brief time.
If a vital registration system had existed for a long time, were very accurate, and there were no uncounted migrations, one could use past births and deaths to tally up the current population by age. To the degree that such a tallying up does not match a census, one or more of the data collections systems is faulty.
SOCIAL DEMOGRAPHY
One of the standard definitions of demography is that given by Hauser and Duncan: ”Demography is the study of the size, territorial distribution and composition of populations, changes therein, and the components of such change, which may be identified as natality, mortality, territorial movement (migration) and social mobility (change of status)” (1959, p. 2). Each of these parts—size, territorial distribution, and composition—is a major arena in which the relationships between demographic change and social change are investigated by social scientist-demographers. Each part has a somewhat separate literature, tradition, method, and body of substantive theory.
Population Size. The issue of population size and change in size is dominated by the shadow of Malthus (Malthus 1959), who held that while food production can grow only arithmetically because of diminishing returns to investment in land and other resources, population can grow geometrically and will do so, given the chance. Writing in a time of limited information about birth control and considerable disapproval of its use, and holding little hope that many people would abstain from sexual relations, Malthus believed that populations would naturally grow to the point at which starvation and other deprivations would curtail future expansion. At that point, the average level of living would be barely above the starvation level. Transitory improvements in the supply of food would only lead to increased births and subsequent deaths as the population returned to its equilibrium size. Any permanent improvement in food supplies due to technological advances would, in Malthus’s theory, simply lead to a larger population surviving at the previous level of misery.
Although there is good evidence that Malthus understood his contemporary world quite well (Lee 1980; Wrigley 1983), he missed the beginnings of the birth control movement, which were contemporaneous (McLaren 1978). The ability to limit births, albeit at some cost, without limiting sexual activity requires important modifications to the Malthusian model.
Questions of the relationship among population growth, economic growth, and resources persist into the contemporary period. The bulk of the literature is in economics. A good summary of that literature can be found in T. P. Schultz (1981), in Rosensweig and Stark (1997), and in a National Research Council report on population policy (1985). A more polemical treatment, but one that may be more accessible to the noneconomist, is offered by the World Bank (1985).
Sharing the study of population size and its change with Malthusian issues is a body of substantive and empirical work on the demographic revolution or transition (Notestein 1945). The model for this transition is the course of fertility and mortality in Europe during the Industrial Revolution. The transition is thought to occur concurrently with ”modernization” in many countries (Coale and Watkins 1986) and to be still in process in many less developed parts of the world (United Nations 1990). This transition is a change from (1) a condition of high and stable birthrates combined with high and fluctuating death rates, through (2) a period of initially lowering death rates and subsequently lowering birthrates, to (3) a period of low and fairly constant death rates combined with low and fluctuating birthrates. In the course of part (2) of this transition, the population grows very considerably because the rate of increase, absent migration, is the difference between the birthrate and the death rate.
In large measure because of anxiety that fertility might not fall rapidly enough in developing countries, a good deal of research has focused on the fertility part of the transition. One branch of this research has been a detailed historical investigation of what actually happened in Europe, since that is the base for the analogy about what is thought likely to happen elsewhere (Coale and Watkins 1986). A second branch was the World Fertility Survey, perhaps the largest international social science research project ever undertaken. This project conducted carefully designed, comparable surveys with 341,300 women in seventy-one countries to investigate the circumstances of contemporary fertility decline (Cleland and Hobcraft 1985). In 1983, a continuation of this project was undertaken under the name Demographic and Health Surveys. To date, this project has provided technical assistance for more than 100 surveys in Africa, Asia, the Near East, Latin America, and the Caribbean. For more complete information about current activities see their web site at http:// www.macroint.com/dhs/.
Scholars analyzing these projects come to a fairly similar assessment of the roots of historical and contemporary fertility decline as centering in an increased secular rationality and growing norms of individual responsibility.
Detailed investigation of the mortality side of the European demographic transition has primarily been conducted by historians. A particularly useful collection of papers is available in Schofield, Reher, and Bideau (1991).
The utility of the idea of a demographic transition to understand population change in the developing world has, of course, been a matter of considerable debate. A good summary of this literature is found in Jones et al. (1997).
Territorial Distribution. Research on the territorial distribution of populations is conducted in sociology, geography, and economics. The history of population distribution appears to be one of population dispersion at the macroscopic level of continents, nations, or regions, and one of population concentration at the more microscopic level of larger towns and cities.
The diffusion of the human population over the globe, begun perhaps as the ice shields retreated in the late Pleistocene, continues to the present (Barraclough 1978; Bairoch 1988). More newly inhabited continents fill up and less habitable land becomes occupied as technological and social change makes it possible to live in previously remote areas. Transportation lines, whether caravan, rail, or superhighway, extend across remote areas to connect distant population centers. Stops along the way become villages and towns specializing in servicing the travelers and the goods in transit. These places are no longer ”remote.” Exploitation of resources proximate to these places may now become viable because of the access to transportation as well as services newly available in the stopover towns.
As the increasing efficiency of agriculture has released larger and larger fractions of the population from the need to till the soil, it has become possible to sustain increasing numbers of urban people. There is a kind of urban transition more or less concomitant with the demographic transition during which a population’s distribution by city size shifts to larger and larger sizes (Kelley and Williamson 1984; Wrigley 1987). Of course, the continuing urban transition presents continuing problems for the developing world (United Nations 1998).
Population Composition. The characteristics included as compositional ones in demographic work are not predefined theoretically, with the exception, perhaps, of age and sex. In general, compositional characteristics are those characteristics inquired about on censuses, demographic surveys, and vital registration forms. Such items vary over time as social, economic, and political concerns change. Nonetheless, it is possible to classify most compositional items into one of three classes. First are those items that are close to the reproductive core of a society. They include age, sex, family relationships, and household living arrangements. The second group of items are those characteristics that identify the major, usually fairly endoga-mous, social groups in the population. They include race, ethnicity, religion, and language. Finally there is a set of socioeconomic characteristics such as education, occupation, industry, earnings, and labor force participation.
Within the first category of characteristics, contemporary research interest has focused on families and households because the period since 1970 has seen such dramatic changes in developed countries (Van de Kaa 1987; Davis 1985; Farley 1996, ch. 2). Divorce, previously uncommon, has become a common event. Many couples now live together without record of a civil or religious ceremony having occurred. Generally these unions are not initially for the purposes of procreation, although children are sometimes born into them. Sometimes they appear to be trial marriages and are succeeded by legal marriages. A small but increasing fraction of women in the more developed countries seem to feel that a husband is not a necessary ornament to motherhood. Because of these changes, older models of how marriage comes about and how marriage relates to fertility (Coale and McNeil 1972; Coale and Trussell 1974) are in need of review as demographers work toward a new demography of the family (Bongaarts 1983; Keyfitz 1987).
Those characteristics that indicate membership in one or another of the major social groupings within a population vary from place to place. Since such social groupings are the basis of inequality, political division, or cultural separation, their demography becomes of interest to social scientists and to policy makers (Harrison and Bennett 1995). To the degree that endogamy holds, it is often useful to analyze a social group as a separate population, as is done for the black population in the United States (Farley 1970). Fertility and mortality rates, as well as marital and family arrangements, for blacks in the United States are different from those of the majority population. The relationship between these facts, their implications, and the socioeconomic discrimination and residential segregation experienced by this population is a matter of historic and continuing scholarly work (Farley and Allen 1987; Lieberson 1980; Farley 1996, ch. 6).
For other groups, such as religious or ethnic groups in the United States, the issue of endogamy becomes central in determining the continuing importance of the characteristic for the social life of the larger population (Johnson 1980; Kalmijn 1991). Unlike the black population, ethnic and religious groups seem to be of decreasing appropriateness to analyze as separate populations within the United States, since membership may be a matter of changeable opinion.
The socioeconomic characteristics of a population are analyzed widely by sociologists, economists, and policy-oriented researchers. Other than being involved in the data production for much of this research, the uniquely demographic contributions come in two ways. First is the consideration of the relationship between demographic change and change in these characteristics. The relationship between female labor force participation and changing fertility patterns has been a main topic of the ”new home economics” within economic demography (Becker 1960; T. W. Schultz 1974; T. P. Schultz 1981; Bergstrom 1997). The relationship between cohort size and earnings is a topic treated by both economists and sociologists (Winsborough 1975; Welch 1979). A system of relationships between cohort size, economic well-being, and fertility has been proposed by Easterlin (1980) in an effort to explain both fertility cycles and long swings in the business cycle.
A second uniquely demographic contribution to the study of socioeconomic characteristics appears to be the notion of a cohort moving through its socioeconomic life course (Duncan and Hodge 1963; Hauser and Featherman 1977; Mare 1980). The process of leaving school, getting a first job, then subsequent jobs, each of which yields income, was initially modeled as a sequence of recursive equations and subsequently in more detailed ways. Early in the history of this project it was pointed out that the process could be modeled as a multistate population (Matras 1967), but early data collected in the project did not lend itself to such modeling and the idea was not pursued.
POPULATION POLICY
The consideration and analysis of various population policies is often seen as a part of demography. Population policy has two important parts. First is policy related to the population of one’s own nation. The United Nations routinely conducts inquiries about the population policies of member nations (United Nations 1995). Most responding governments claim to have official positions about a number of demographic issues, and many have policies to deal with them. It is interesting to note that the odds a developed country that states its fertility is too low has a policy to raise the rate is about seven to one, while the odds that a less developed country that states that its fertility is too high has a policy to lower the rate is about 4.6 to one. The prospect of declining population in the United States has already begun to generate policy proposals (Teitelbaum and Winter 1985).
The second part of population policy is the policy a nation has about the population of other countries. For example, should a country insist on family-planning efforts in a developing country prior to providing economic aid? A selection of opinions and some recommendations for policy in the United States are provided in Menken (1986) and in the National Research Council (1985).
IMPORTANT CURRENT RESEARCH AREAS
Three areas of demographic research are of especially current interest. Each is, in a sense, a sequela of the demographic transition. They are:
1. Institutional arrangements for the production of children,
2. Immigration, emigration, ethnicity, and nationalism,
3. Population aging, morbidity, and mortality.
Institutional arrangements for the production of children. The posttransition developed world has seen dramatic changes in the institutional arrangements for the production of children. Tracking these changes has become a major preoccupation of contemporary demography. Modern contraception has broken the biological link between sexual intercourse and pregnancy. Decisions about sexual relations can be made with less concern about pregnancy. The decision to have a child is less colored by the need for sexual gratification. People no longer must choose between the celibate single life or the succession of pregnancies and children of a married one. Women are employed before, during, and after marriage. Their earning power makes marriage more an option and a context for child raising than an alternative employment. Women marry later, are older when they have children, and are having fewer children than in the past. Divorce has moved from an uncommon event to a common one. Cohabitation is a new, and somewhat inchoate, lifestyle that influences the circumstances for childbearing and child rearing. Although only a modest number of children are currently born to cohabiting couples, most children will spend some time living with a parent in such a relationship before they are adults (Bumpass and Lu 1999). Changes such as these have occurred in most of Europe and North America and appear to be increasing in those parts of Asia where the demographic transitions occurred first. The changes have, of course, been met with considerable resistance in many places. Will they continue to occur in other countries as the demographic transitions proceed around the world? How serious will the resistance be in countries more patriarchal than those of Europe? These issues await future demographic research.
Immigration, emigration, ethnicity, and nationalism. The difference between immigration and emigration is, of course, net migration as a component of population growth. In general, the direct impact of net migration on growth has been modest. Certainly emigration provided some outlet for population growth in Europe during the transition (Curtin 1989) and played a role in the population of the Americas, Australia, and New Zealand. Most of the effect of immigration to the New World was indirect, however. It was the children of immigrants who peopled the continents rather than the immigrants themselves. Contemporary interest in international migration has three sources. First is an abiding concern for refugees and displaced persons. Second is interest in a highly mobile international labor force in the construction industry. Third, and perhaps most important, is the migration of workers from less developed countries to developed ones in order to satisfy the demand for labor in a population in real or incipient decline (Teitelbaum and Winter 1998). In France, Germany, and England immigrants who come and stay change the ethnic mix of the population. In the United States, legal and illegal immigration from Mexico has received considerable attention (Chiswick and Sullivan 1995). The immigration of skilled Asians raises issues of another kind. In all of these countries, the permanent settlement of new immigrants arouses powerful nationalistic concerns. The situation provides ample opportunity for the investigation of the development of ethnicity and the place of endogamy in its maintenance. As with changes for child rearing, these changes seem consequences of the demographic transition. Below-replacement fertility seems to generate the need for imported workers and their consequent inclusion in the society. If this speculation is true, how will countries more recently passing through the transition react to this opportunity and challenge? This is another arena for future demographic investigation.
Population aging, morbidity, and mortality.
Accompanying the reduction of fertility and of population growth is the aging of the population (Treas and Torrecilha 1995). Much of the developed world is experiencing this aging and its concomitants. Countries that made a rapid demographic transition have especially dramatic imbalances in their age distribution. Concerns about retirement, health, and other programs for the elderly generate interest in demographic research in these areas. Will increasing life expectancy add years to the working life so that retirement can begin later? Or will the added years all be spent in nursing homes? These questions are currently of great policy interest and pose interesting and difficult problems for demographic research (Manton and Singer 1994). Traditional demographic modeling has assumed that frailty, the likelihood of death, varies according to measurable traits such as age, sex, race, education, wealth, and so on, but are constant within joint values of these variables. But what happens if frailty is a random variable? In survival models, randomness in frailty is not helpful, or at least benign, as it is in so many other statistical circumstances. Rather, it can have quite dramatic effects (Vaupel and Yashin 1985). How one models frailty appears to make a considerable amount of difference in the answers one gets to important policy questions surrounding population aging.
DEMOGRAPHY AS A PROFESSION
Most demographers in the United States are trained in sociology. Many others have their highest degree in economics, history, or public health. A few are anthropologists, statisticians, or political scientists. Graduate training of demographers in the United States and in much of the rest of the world now occurs primarily in centers. Demography centers are often quasi-departmental organizations that serve the research and training needs of scholars in several departments. In the United States there are about twenty such centers, twelve of which have National Institutes of Health grants. The Ford Foundation has supported similar demography centers at universities in the less developed parts of the world.
Today, then, most new demographers, without regard for their disciplinary leanings, are trained at a relatively few universities. Most will work as faculty or researchers within universities or at demography centers. Another main source of employment for demographers is government agencies. Census bureaus and vital statistics agencies both provide much of the raw material for demographic work and employ many demographers around the world. There is a small but rapidly growing demand for demographers in the private sector in marketing and strategic planning.
Support for research and training in demography began in the United States in the 1920s with the interest of the Rockefeller Foundation in issues related to population problems. Its support led to the first demography center, the Office of Population Research at Princeton University. The Population Council in New York was established as a separate foundation by the Rockefeller brothers in the 1930s. Substantial additional foundation support for the field has come from the Ford Foundation, the Scripps Foundation, and, more recently, the Hewlett Foundation. Demography was the first of the social sciences to be supported by the newly founded National Science Foundation in the immediate post-World War II era. In the mid-1960s the National Institute of Child Health and Human Development undertook support of demographic research.
