Birth and Death Rates

views updated

BIRTH AND DEATH RATES

Much of the birth and death information published by governments is in absolute numbers. These raw data are difficult to interpret. For example, a comparison of the 42,087 births in Utah with the 189,392 in Florida in 1996 reveals nothing about the relative levels of fertility because Florida has a larger population (Ventura et al. 1998, p. 42).

To control for the effect of population size, analyses of fertility and mortality usually use rates. A rate measures the number of times an event such as birth occurs in a given period of time divided by the population at risk to that event. The period is usually a year, and the rate is usually expressed per 1,000 people in the population to eliminate the decimal point. Dividing Florida's births by the state's population and multiplying by 1,000 yields a birth rate of 13 per 1,000. A similar calculation for Utah yields 21 per 1,000, evidence that fertility makes a greater contribution to population growth in the state with the large Mormon population.


BIRTH RATES

The crude birth rate calculated in the preceding example, is the most common measure of fertility because it requires the least amount of data and measures the impact of fertility on population growth. Crude birth rates at the end of the twentieth century range from over 40 per 1,000 in many African countries and a few Asian countries such as Yemen and Afghanistan to less than 12 per 1,000 in the slow-growing or declining countries of Europe and Japan (Population Reference Bureau 1998).

The crude birth rate is aptly named when used to compare childbearing levels between populations. Its estimate of the population at risk to giving birth includes men, children, and postmenopausal women. If women of childbearing age compose different proportions in the populations under consideration or within the same population in a longitudinal analysis, the crude birth rate is an unreliable indicator of the relative level of childbearing. A portion of Utah's 61 percent higher crude birth rate is due to the state's higher proportion of childbearing-age women, 23 percent, versus 20 percent in Florida (U.S. Bureau of the Census 1994). The proportion of childbearing-age women varies more widely between nations, making the crude birth rate a poor choice for international comparisons.

Other rates that more precisely specify the population at risk are better comparative measures of childbearing, although only the crude birth rate measures the impact of fertility on population growth. If the number of women in childbearing ages is known, general fertility rates can be calculated:


This measure reveals that a thousand women in Utah of childbearing age produce more births in a year than the same number in Florida, 89 versus 64 births per 1,000 women between ages fifteen and forty-four (Ventura et al. 1998, p. 42). The 39 percent difference in the two states' general fertility rates is substantially less than that indicated by their crude birth rates which are confounded by age-composition differences.

Although the general fertility rate is a more accurate measure of the relative levels of fertility between populations, it remains sensitive to the distribution of population across women of childbearing ages. When women are heavily concentrated in the younger, more fecund ages, such as in developing countries today and in the United States in 1980, rather than the less fecund older ages, such as in the United States and other developed countries today, the general fertility rate is not the best choice for fertility analysis. It inflates the relative level of fertility in the former populations and deflates the estimates in the latter populations.

Age-specific fertility rates eliminate potential distortions from age compositions. These rates are calculated for five-year age groups beginning with ages fifteen to nineteen and ending with ages forty-five to forty-nine:


Age-specific fertility rates also provide a rudimentary measure of the tempo of childbearing. The four countries in Figure 1 have distinct patterns. Zimbabwe, a less-developed country with high fertility, has higher rates at all ages. At the other extreme, Japan's low fertility is highly concentrated between the ages of twenty-five and thirty-four, even though the Japanese rates are well below those in Zimbabwe. In contrast, teenage women in the United States continue to have much higher fertility than teenage women in Japan and other industrialized countries, despite declines in the 1990s. The younger pattern of American fertility also is evident in the moderately high rates for women in their early twenties. At the other extreme, Irish women have an older pattern of fertility.

More detailed analyses of the tempo of childbearing require extensive information about live birth order to make fertility rates for each age group specific for first births, second births, third births, and so forth (Shryock and Siegel 1976, p. 280). A comparison of these age-order-specific rates between 1975 and 1996 reveals an on-going shift toward later childbearing in the United States (Ventura et al. 1998, p. 6). The first birth rate increased for women over thirty while it decreased for women ages twenty to twenty-four. As a result, 22 percent of all first births in 1996 occurred to women age thirty and over, compared to only 5 percent in 1975.

When the tempo of fertility is not of interest, the advantages of age-specific fertility rates are outweighed by the cumbersome task comparing many rates between populations. As an alternative, each population's age-specific rates can be condensed in an age–sex adjusted birth rate (Shryock and Siegel 1976, pp. 284–288). The most frequently used age–sex adjusted rate is calculated:


if the age-specific fertility rates are for five-year age groups. Single-year age-specific rates are summed without the five-year adjustment. When expressed per single woman, as in equation (4), the total fertility rate can be interpreted as the average number of births that a hypothetical group would have at the end of their reproduction if they experienced the age-specific fertility observed in a particular year over the course of their childbearing years.

Age-specific rates in real populations that consciously control fertility can be volatile. For example, the fertility rate of American women ages thirty to thirty-four fell to 71 per 1,000 in the middle of the Great Depression and climbed back to 119 during the postwar Baby Boom, only to fall again to 53 in 1975 and rebound to 84 in 1996 (U.S. Bureau of the Census 1975, p. 50; Ventura et al. 1998, p. 32). Consequently, a total fertility rate calculated from one year's observed age-specific rates is not a good estimate of the eventual completed fertility of childbearing-age women. It is, however, an excellent index of the level of fertility observed in a year that is unaffected by age composition.

The total fertility rate also can be interpreted as an estimate of the reproductivity of a population. Reproductivity is the extent to which a generation exactly replaces its eventual deaths. While the total fertility measures the replacement of both sexes, other measures of reproductivity focus only on the replacement of females in the population (Shryock and Siegel 1976, pp. 314–316). The gross reproduction rate is similar to the total fertility rate except that only female births are included in the calculation of the age-specific rates. It is often approximated by multiplying the total fertility rate by the ratio of female births to all births.

Theoretically, women need to average two births, one female and one male, and female newborns need to live long enough to have their own female births at the same ages that their mothers gave birth to them to maintain a constant population size. In real populations some female newborns die before their mothers' ages at their births and the tempo of fertility fluctuates. As a result, both the total fertility rate and the gross reproduction rate overestimate reproductivity. The net reproduction rate adjusts for mortality, although it remains sensitive to shifts in the tempo of childbearing. This measure of reproductivity is calculated by multiplying the age-specific fertility rates for female births by the corresponding life-table survival rates that measure the probability of female children surviving from birth to the age of their mothers, and summing across childbearing ages. If the tempo of fertility is constant, a net reproduction rate of greater than one indicates population growth; less than one indicates decline; and one indicates a stationary population.

Because of the impact of female mortality, the total fertility rate must exceed two for a generation to replace all of its deaths. In industrialized countries with a low risk of dying before age fifty, the total fertility rate needs to be about 2.1 for replacement. Developing countries with higher mortality need a higher total fertility rate for replacement. Malawi, for example, with an infant mortality rate of 140 per 1,000 live births and a life expectancy of only thirty-six needs a rate of about three births per woman for replacement.

The fertility of most industrialized countries has fallen below replacement (Population Reference Bureau 1998). Some, including the United States (Figure 2), Ireland, Iceland, and New Zealand, are barely below replacement. Others have declined to unprecedented low rates. Spain (Figure 2), Portugal, Italy, Greece, Germany, most Eastern European countries, and Japan have total fertility rates of 1.1 to 1.4 births per woman. Some of these populations are already declining. Without constant, substantial net in-migration, all will decline unless their fertility rates rebound to replacement levels or above.

In contrast, the total fertility rates of most developing countries whose economies are still dependent primarily on agriculture exceed replacement (Population Reference Bureau 1998). The highest rates are found in Africa where most countries have rates greater than five births per woman. A few, including Ethiopia, Somalia, Niger, and Angola, equal or exceed seven births per woman. As a result, sub-Saharan Africa is growing at about 2.6 percent per year. If unchanged, Africa's population would double in twenty-seven years. Change, however, appears underway in most African countries. The largest fertility declines since 1980 have occurred in North African countries like Egypt (Figure 2) and in Kenya, Zimbabwe, and South Africa. Nevertheless, fertility in all African countries remains well above replacement. The continent's 1998 total fertility rate was 5.6 births per woman.

Fertility has declined to lower levels in other developing regions. Led by three decades of decline in Mexico (Figure 2), Brazil, Ecuador, Peru, and Venezuela, fertility rates for Central and South America have declined to 3.4 and 2.8 births per woman, respectively. Most Caribbean countries have near replacement-level fertility. Haiti, the Dominican Republic, and Jamaica are the major exceptions.

Rapid fertility declines also have occurred throughout much of Asia. China's aggressive birth control policy and nascent economic growth reduced its fertility to 1.8 births per woman (Figure 2), about the same as its Korean and Taiwanese neighbors. A number of other Asian countries, including the large populations of Bangladesh, Iran, Thailand, Vietnam, and Turkey, experienced more than a 50 percent decline in the last two decades of the twentieth century. The even larger populations of India and Indonesia continued their previous slow downward trend, yielding 1998 rates of 3.4 and 2.7 births per woman, respectively. Only a few Asian countries, other than traditional Moslem societies such as Afghanistan, Iraq, and Pakistan, continue to have more than five births per woman.

The fertility measures discussed up to this point are period rates. They are based on data for a particular year and represent the behavior of a cross-section of age groups in the population in that year. Fertility also can be measured over the lifetime of birth cohorts. Cumulative fertility rates can be calculated for each birth cohort of women by summing the age-specific fertility rates that prevailed as they passed through each age (Shryock and Siegel 1976, p. 289). This calculation yields a completed fertility rate for birth cohorts that have reached the end of their reproductive years. It is the cohort equivalent of the period total fertility rate.


DEATH RATES

The measurement of mortality raises many of the same issues discussed with fertility. Rates are more informative than absolute numbers, and those rates that more precisely define the population at risk to dying are more accurate. Unlike fertility, however, the entire population is at risk to dying, and this universal experience happens only once to an individual.

The impact of mortality on population growth can be calculated with a crude death rate:


Crude death rates vary from over 20 per 1,000 in some African countries to as low as 2 per 1,000 (Population Reference Bureau 1998). The lowest crude death rates are not in the developed countries of Europe, North America, and Oceania, which have rates between 7 and 14 per 1,000. Instead, the lowest crude death rates are found in oil-rich Kuwait, Qatar, and United Arab Emirates, where guest workers inflate the proportion of young adults, and in the young populations of developing countries experiencing declining mortality. Developed countries that underwent industrialization and mortality decline before 1950 have old-age compositions. The proportion of people age sixty-five and over ranges between 11 and 17 percent in these countries compared to less than 5 percent in most African, Asian, and Latin American populations. Although there is a risk of dying at every age, the risk rises with age after childhood. Consequently, older populations have higher crude death rates.

To control for the strong influence of age on mortality, age-specific rates can be calculated for five-year age groups:


Before age five, the age-specific mortality rate usually is subdivided to capture the higher risk of dying immediately after birth. The rate for one- to four-year olds, like other age-specific rates, is based on the midyear estimate of this population. The conventional infant mortality rate, however, is based on the number of live births:

The infant mortality rate is often disaggregated into the neonatal mortality rate for the first month of life and the postneonatal rate for the rest of the year.


Infant mortality varies widely throughout the world. Iraq, Afghanistan, Cambodia, and many African countries have 1998 rates that still exceed 100 per 1,000 live births, although most have declined (Population Reference Bureau 1998). Infant mortality rates in most other African, Asian, and all other Latin American countries have declined as well. They now range in developing countries from 7 in Cuba to 195 in Sierra Leone. In contrast, developed countries have rates at or below 10, led by Japan with under 4 infant deaths per 1,000 live births.

The declining American infant mortality rate, 7 per 1,000 live births in 1998, continues to lag behind Japan, Australia, New Zealand, Canada, and most Northern and Western European countries due to a higher prevalence of prematurity and low-birth-weight babies which are major causes of infant death (Peters et al. 1998, pp. 12–13). The prematurity and low-birth-weight rates for the country's largest minority, African Americans, are double those of non-Hispanic whites (Ventura 1998, pp. 57–58). Not all minority mothers have higher rates than non-Hispanic whites. Low-birth-weight rates are lower for Americans of Chinese, Mexican, Central and South American origin, about the same for Native Americans and those of Cuban origin, and higher for Puerto Rican, Filipino, and Japanese Americans.

Table 1. Death Rates per 100,000 for the Leading Causes of Death in the Population 45–64 Years of Age, by Sex and Selected Race and Ethnic Categories: United States, 1997.
Death Rates per 100,000 for the Fourteen Leading Causes of Death in the Population 45-64
Years of Age, by Sex and Selected Race and Ethnic Categories: United States, 1997
 malesfemales
 non-hispanic non-hispanic 
 whitesblackshispanicwhitesblackshispanic
source:hoyert, d. deaths: final data for 1997.
heart diseases2544691609323169
cancers254431147219285129
strokes259132195820
respiratory diseases2631 24227
accidents      
motor vehicle1930219119
other2349268146
pneumonia and flu1132118156
diabetes206033165429
hiv67225 185
suicide25 128  
liver disease and cirrhosis264453101514
kidney disease     4
septicemia   414 
homicide and legal intervention 2812   


Infant mortality also varies by sex. The rate for male infants born to white mothers in the United States is 6.7 per 1,000 live births, compared to 5.4 for female infants (Peters 1998, p. 80). Similarly, the rate for male infants born to African-American mothers is 16 per 1,000 live births, compared to 13 for female infants.

Like race and ethnic differences, the sex difference in mortality is evident at all ages. The greatest gap is among young adults; American males are two and one-half to three times more likely to die between the ages of fifteen and twenty-nine than females due largely to behavioral causes. These include motor vehicle and other accident fatalities, homicides, and suicides. Higher male death rates from congenital anomalies at younger ages and from heart disease in middle age (Table 1) suggest that biological factors also contribute to the sex difference in mortality. Behavioral causes continue to play a significant role in the sex difference in mortality in middle age with chronic liver disease and cirrhosis and human immunodeficiency virus (HIV) adding premature male mortality. Note that cause-specific death rates are calculated per 100,000 population in a specified group, rather than per 1,000 to avoid working with decimals.

Age-specific mortality rates usually are specific for sex and race or ethnicity because of the large differences evident in Table 1. Using five-year age categories results in thirty-eight rates for each racial or ethnic group. Analyses of such data can be unwieldy. When the age pattern of mortality is not of interest, an age-adjusted composite measure is preferable. The most readily available of these for international comparisons is life expectancy.

Life expectancy is the average number of years that members of an age group would live if they were to experience the age-specific death rates prevailing in a given year. It is calculated for each age group in a life table (Shryock and Siegel 1976, pp. 249–268). Life expectancy has increased since the nineteenth century in developed countries like the United States in tandem with economic growth and public health reforms. Japan has the longest life expectancy at birth, eighty years, due to extremely low infant mortality. With the exception of Russia and some of the other former Soviet-bloc countries of Eastern Europe that have had declines in life expectancy, all other developed countries have life expectancies at birth that equal or exceed seventy years (Population Reference Bureau 1998).

Figure 3 presents trends in life expectancy since World War II in selected countries. Some of the most rapid increases have been in Asia and Latin America. China, Sri Lanka, Malaysia, Mexico, and Venezuela, for example, have 1998 life expectancies at birth of over seventy years. Increases in life expectancy in African countries, in contrast, where there is a relatively high prevalence of poor nutrition, unsanitary conditions, and infectious diseases, generally have lagged behind Asia and Latin America. Most have 1998 life expectancies in the forties or fifties, similar to those of developed countries at the end of the nineteenth century. The future trend in life expectancy in many African countries is uncertain. Progress made in some is being undermined by the spread of HIV and political conflict. The drug therapies that have reduced mortality from HIV in developed countries have so far proven too costly for widespread use in Africa and political unrest continues.

Declines in infant and child mortality from infectious diseases have been largely responsible for the historical increase in life expectancy around the world. As a result, life expectancy has increased far more at birth than at age sixty (Figure 4). Heart disease, cancer, and cerebrovascular diseases have become the major causes of death in the United States (Peters et al. 1998, p. 7), other developed countries, and some developing countries. Bringing these chronic diseases under control has proven to be difficult. Research suggests that both genetic composition and unhealthy behaviors affect the development of such chronic diseases over time. In the United States where health education campaigns have stressed the link between these diseases and obesity, use of tobacco, and lack of sufficient physical activity, mortality has declined from heart disease since 1950 and from cancer since 1990. If these trends continue, gains in life expectancy at older ages will accelerate.


references

Hoyert, D., Kenneth Kochanek, and Sherry Murphy 1999 "Deaths: Final Data for 1997." National Vital Statistics Reports, vol. 47, no. 19. Hyattsville, Md.: National Center for Health Statistics.

Population Reference Bureau 1998 1998 World Population Data Sheet. Washington, D.C.: Population Reference Bureau.

——1980 1980 World Population Data Sheet. Washington, D.C.: Population Reference Bureau. Shryock, Henry, and Jacob Siegel 1976 The Methods and Materials of Demography. San Francisco: Academic Press.

United Nations 1998 Demographic Yearbook, 1996. New York: United Nations.

——1998 World Population Prospects, 1996. New York: United Nations.

U.S. Bureau of the Census 1994 "Population Projections for States by Age, Sex, Race, and Hispanic Origin: 1993 to 2020." Current Population Reports, series P-25, no. 1111. Washington, D.C.: Government Printing Office.

——1975 Historical Statistics of the United States, Colonial Times to1970. Washington D.C.: Government Printing Office.

Ventura, Stephanie, Joyce Martin, Sally Curtin, and T. J. Mathews 1998 "Report of Final Natality Statistics, 1996." Monthly Vital Statistics Report, vol. 46, no. 11, Supp. Hyattsville, Md.: National Center for Health Statistics.


Deborah A. Sullivan
Department of Sociology
Arizona State University