DIVORCE

FIGURE 1

The legal dissolution of marriage has reached historically high levels in most industrial societies. In the United States, divorce rates have been increasing for more than a century. They increased steeply in the 1960s and 1970s and subsequently leveled off, as shown in Figure 1. Slightly less than half of all first marriages in the United States in the early twenty-first century end in divorce. This proportion is roughly the same in Sweden. The figures are somewhat lower (around 40% per marriage) for other Nordic countries, the United Kingdom, Belgium, and most countries of the former Soviet Union. In most other industrial societies, divorce rates are rising, although they are still considerably lower than in the United States.

With the exception of parts of Asia and Latin America, dissolution of marriage is also common, although statistically less well documented, in the developing world. The proportion of women separated or divorced between the ages of 45 and 59 increased on average from 7 percent in the 1980s to 10 percent in the 1990s in Latin America and the Caribbean (the corresponding figures are 5 percent and 9 percent in Europe as a whole and 9 percent and 14 percent in other developed regions). As a combined effect of death, separation, and divorce, in many countries in sub-Saharan Africa, more than one-third of women experience dissolution of marriage before reaching the end of childbearing years.

The rise in divorce in the industrialized countries has been attributed to (1) smaller gains from marriage than in the past; (2) the declining social stigma of divorce; and (3) a relaxation of divorce laws, as represented by the introduction of no-fault divorce. These factors are thought to mutually reinforce each other. For example, increases in divorce reduce social stigma and foster the liberalization of divorce law. Higher expectations of divorce may encourage precautionary behavior (for example, waiting to have children) that itself may reduce the benefits of staying married. In turn, expectation of ease of divorce may encourage less careful deliberation before entering into marriage, which in itself could increase the likelihood of divorce. The plateau in divorce rates in the United States since about 1980 remains largely unexplained.

Although divorce ends individual marriages, most often the individuals who divorce remarry, often to partners who are themselves divorced. Current U.S. remarriage rates imply that about three-quarters of divorced men and two-thirds of divorced women eventually remarry. Divorce and remarriage have, in concert, changed family structure, increasing the number of single-parent headed households and of marriages in which one or both of the spouses was formerly divorced, and increasing the prevalence of step-relationships within families.

The principal impact of divorce that concerns social scientists is its effect on children. Social scientists working with observational data are not able to determine directly whether children of divorce are worse off than those of parents who remain married. Studies that have tried to control for factors other than divorce have yielded mixed results, but some have found that children of divorce do fare worse in various ways in comparison to children brought up in intact marriages. Another consequence of divorce is the economic hardship associated with the dissolution of households, which often falls disproportionately on women.

Higher rates of divorce might be expected to increase the quality of marriage in a society by allowing those in unsatisfactory marriages to choose new partners. On the other hand, it can also be argued that higher divorce rates may have negative consequences for all marriages because couples will be reluctant to make marriage-specific investments, like having children, which might enhance the advantages of marriage. The validity of either of these theories remains unproven.

Marriages most at risk of divorce include those entered into at an early age. There is also a socioeconomic gradient to divorce, such that those with more education and higher income face slightly lower risks of divorce. Part of the effect of socioeconomic status is mediated through age at marriage (those who attend college are less likely to marry young), but multivariate analysis reveals that, even after controlling for age at marriage, those couples who are poor and less educated face slightly higher risks of divorce.

The age pattern of divorce, or more accurately the duration pattern, is such that following the first year, divorce rates decline consistently as the length of the marriage increases. Contrary to popular belief, there is no increase in divorce after seven years of marriage (the seven year itch) apparent in the U.S. pattern. Part of the declining risks of divorce related to duration may be due to the winnowing out of marriages with high risks of divorce. But the advantages of marriage (and associated disadvantages of divorce) are themselves also thought to increase with time. There are more likely to be children, the division of household labor becomes more established, and the amount of other forms of so-called marriage-specific capital increases.

The economic theory of marriage and divorce, pioneered by the economist Gary Becker emphasizes–by analogy with the theory of comparative advantage in international trade–the benefits of marriage that come from the sexual division of labor. According to this theory, the advantages have declined over time as women have entered the labor market and thereby reduced men's comparative advantage as income earners. At the same time, the theory argues that technological advances in the home have reduced women's comparative advantage in housework. In contrast, sociological theories of divorce emphasize the increasingly individualistic and hedonistic orientation of modern societies, and the decline in normative expectations that marriage should be a life-long commitment.

Measures

Divorce, like other demographic rates, can be measured cross-sectionally at a given time (period measure) or along the lifecycle (cohort measure), and at the aggregate or individual level.

The most commonly used aggregate measures are annual rates of divorce. The simplest measure, called the crude divorce rate, is estimated by dividing the total number of divorces in a year by the total mid-year population. The crude divorce rate can be influenced by changes in the population at risk. For example, if age-at-marriage is delayed, the proportion of married people will decrease and the rate will decline even if the risks of divorce to married people remain the same.

A more refined measure, that takes into account the number of marriages at risk, is the general divorcerate, which divides the total number of divorces by the number of married women (a proxy for the number of marriages). The general divorce rate can be thought of as the chance that a randomly chosen couple will divorce in a given year. It is less influenced by changing population composition although it can be influenced by changes in the distribution of marriages by duration. For example, a population with many recent marriages may have a higher general divorce rate because divorce rates are higher for new marriages. The general divorce rate is less often reported than the crude divorce rate because it requires estimation of the number of married women in the population.

In the United States, the crude divorce rate in 1998 was 4.2 per 1000, down from a peak of 5.2 per 1000 in 1980 but nearly double the rate observed in the 1950s. The general divorce rate was 19.5 per 1000 in 1996, slightly more than double the rate observed in the 1950s.

Age-and duration-specific measures of divorce can be used to estimate individual probabilities of divorce. For example, the above-cited estimate that 1 of every 2 first marriages in the United States ends in divorce is calculated by constructing a nuptiality table (analogous to a life table) from marital-duration-specific divorce rates. Estimation of duration-specific rates can be difficult to because it requires counts of marital duration not only for couples that divorce, but also those that remain intact. To calculate such refined rates, data from sample surveys, rather than data from censuses or national registration systems, are often used.

In Europe, a period measure called the total divorce rate, applying to a synthetic (cross-sectional) lifetime, is sometimes reported. This rate, in the same manner as the total fertility rate or the total first marriage rate, sums the age-specific divorce rates (divorces per woman, either married or unmarried) at a given time. This measure is informative but, like the general divorce rate, can be influenced by variations in the population at risk.

In the United States the collection of detailed divorce certificate data was discontinued by the federal government as a cost-savings measure and because the information provided by reporting states was incomplete. Thus, apart from the aggregate rates, most of what is known about divorce differentials and trends is inferred from sample surveys with retrospective marital histories, like the Census Bureau's Current Population Survey or the Survey of Income Program Participation.

As long-term cohabitation increases, which is the case especially in Western Europe, divorce rates alone give an increasingly incomplete picture of actual separation levels.

See also: Cohabitation; Family: Future; Family Demography; Marriage.

bibliography

Becker, Gary. 1991. A Treatise on the Family, enlarged edition. Cambridge, MA: Harvard University Press.

Cherlin, Andrew J. 1992. Marriage, Divorce, Remarriage, 2nd edition. Cambridge, MA: Harvard University Press.

Clarke, Sally C. 1995. "Advance Report of Final Divorce Statistics, 1989 and 1990." Monthly Vital Statistics Report 43(9).

Goldstein, Joshua R. 1999. "The Leveling of Divorce in the United States." Demography 36(3): 409–414.

McLanahan, Sarah, and Gary Sandefur. 1994. Growing Up with a Single Parent. Cambridge, MA: Harvard University Press.

Preston, Samuel H., and John McDonald. 1979. "The Incidence of Divorce within Cohorts of American Marriages Contracted since the Civil War." Demography 16: 1–25.

Schoen, Robert, and Nicola Standish. 2001. "The Retrenchment of Marriage: Results from Marital Status Life Tables for the United States, 1995." Population and Development Review 27(3): 553–563.

United States Census Bureau. 2001. Statistical Abstract of the United States: 2001. Washington, D.C.: US Government Printing Office.

Joshua R. Goldstein