Demography: Past, Present, and Future
Xie, Yu, Journal of the American Statistical Association
In a classic statement, Hauser and Duncan (1959) defined demography as "the study of the size, territorial distribution, and composition of population, changes therein, and the components of such changes" (p. 2). It was fortunate that Hauser and Duncan explicitly included "composition of population" and "changes therein" in their definition, for their inclusion has broadened demography to encompass two types of demography: formal demography and population studies. Formal demography, whose origin can be traced to John Graunt in 1662, is concerned with fertility, mortality, age structure, and spatial distribution of human populations. Population studies is concerned with population compositions and changes from substantive viewpoints anchored in another discipline, be it sociological, economic, biological, or anthropological; its origin can be traced to Thomas Malthus in 1798. By definition, population studies is interdisciplinary, bordering between formal demography and a substantive discipline that is often, but not necessarily, a social science.
Defined in this way, demography provides the empirical foundation on which other social sciences are built. It is hard to imagine that a social science can advance steadily without first knowing the basic information about the human population that it studies. As a field of inquiry, demography enjoyed a rapid growth in the twentieth century. For example, the membership of the Population Association of America (PAA), the primary association for demographers in the U.S. founded in 1931, grew from fewer than 500 in 1956 to more than 3,000 in 1999. This growth is remarkable given the virtual absence of demography departments at American universities (with a few exceptions, such as the University of California Berkeley). To recognize contributions made by demographers, one only needs to be reminded of factual information about contemporary societies. Much of what we know as "statistical facts" about American society, for instance, has been provided or studied by demographers. Examples include socioeconomic inequa lities by race (Farley 1984) and gender (Bianchi and Spain 1986), residential segregation by race (Duncan 1957; Massey and Denton 1993), intergenerational social mobility (Blau and Duncan 1967; Featherman and Hauser 1978), increasing trends of divorce (Sweet and Bumpass 1987) and cohabitation (Bumpass, Sweet, and Cherlin 1991), consequences of single parenthood for children (McLanahan and Sanderfur 1994), rising income inequality (Danziger and Gottschalk 1995), and increasing economic returns for college education (Mare 1995).
Besides providing factual information, demography has also been fundamental in forecasting future states of human societies. Although demographic forecasting is subject to uncertainty, as any type of forecasting, demographers are able to predict future population sizes by age with a high degree of confidence, utilizing information pertaining to past fertility, regularity in age patterns of mortality, and likely future levels of mortality. A notable example of demographic forecasting is the work by Lee and Tuljapurkar (1994, 1997), who demonstrated how demographic forces (i.e., projected improvements in longevity) dramatically impact future demands on social security.
Formal demography and population studies not only take on different subject matters, but also rely on different methodological approaches. Characteristically, formal demography is built on mathematics and thus is closely tied to mathematical demography. It has a rich arsenal of powerful research tools, such as life tables and stable population theory, the latter of which is usually accredited to Alfred J. Lokta in 1922. Note that mathematical models in formal demography sometimes incorporate stochastic processes. The refinement and formalization of mathematical demography and its successful application to human populations can be found in works by Coale (1972), Keyfitz (1985), Preston and Campbell (1993), Rogers (1975), and Sheps and Menken (1973). …