1
Pp. 55-74 in Shoshana Grossbard-Shechtman (ed.) Marriage and the Economy: Theory and Evidence from Industrialized Societies. Cambridge: CambridgeUniversity Press. 2003.
The Economics of Divorce
Evelyn L.Lehrer
The far-reaching implications of the dissolution of a marriage for all family members are well known. The wife, who typically retains custody of the children, generally undergoes a significant decline in financial well-being following divorce (Saul Hoffman and Greg Duncan 1988). Growing up in a single-parent household deprives children of important inputs of parental time and money, and also limits their access to other family and community resources, with adverse effects both short term and long term (Sheila Krein and Andrea Beller 1988; Andrea Beller and John Graham 1993; Sara McLanahan and Gary Sandefur l994). It is thus important to understand how various factors influence the probability that a union will be dissolved. The present article reviews the evidence on this issue, focusing primarily on determinants at the micro-level. These include the characteristics of each of the spouses and the quality of their match; the behaviors of each partner during and before the marriage; and the role of new information and events that were unanticipated at the time of the marriage. Differences between first unions and remarriages are also discussed.
The paper is organized as follows. Section 1 outlines the theory on the economic gains from marriage and the process of marital search, which is essential to understanding the determinants of divorce. Based on this framework, Section 2 discusses the empirical evidence on how various factors affect the risk that a union will eventually be dissolved. Section 3 closes the paper by highlighting important areas where additional research is needed.
1. Gains from Marriage and the Process of Marital Search
From an economic perspective, marriage may be viewed as a partnership formed to coordinate and facilitate production and consumption activities, including the production and raising of own children (Gary Becker 1991). Division of labor and specialization within the family increase the couple’s productivity, and represent a key source of gains from marriage. Other sources include (a) the pooling of risks (e.g., one spouse may increase the level of work in the labor force if the other becomes unemployed); (b) economies of scale (e.g., renting a large apartment costs less than renting two small apartments); (c) public goods (e.g., all members of a household can enjoy the beauty of pictures hanging on the walls); and (d) positive externalities (e.g., watching a television program may yield more enjoyment if done with someone else rather than alone, if utility is derived from the partner’s consumption or mere presence).
For all these reasons, marriage may lead to higher levels of production and consumption. The amount of the gain, however, varies across couples, depending on the characteristics of each partner and the quality of their match. Becker (1991) develops implications about the matching of individuals with various characteristics. The optimal sorting is characterized by negative assortative mating for traits that are substitutes, and positive assortative mating for those that are complements.
In practice, many matches are not “ideal” -- in the sense that better matches might have resulted had search continued -- because the search process is costly. The costs include the foregone gains from marriage, as well as various time and out-of-pocket expenses (e.g., expenditures on personal appearance, dating). The best choice of the reservation offer (the minimum level of match quality) is that which equates at the margin the costs of marital search and the present value of the future benefits associated with search (Michael Keeley 1977).
A union between two individuals may seem optimal from the perspective of both partners at a certain point in time. However, this assessment may subsequently change if one or both spouses discover that their expectations about their own characteristics or those of their partner were incorrect, or if major changes that were not anticipated at the time of the marriage take place. If the costs of terminating the union are assessed to be sufficiently small, by one or both spouses, the marriage may be dissolved.
2. Determinants of the Probability of Divorce
The above discussion suggests that marriage dissolution is more likely if the characteristics of the partners or their match, or the behavior of the spouses before and during the union, imply (a) relatively low gains from marriage and / or (b) relatively low costs of divorce. In addition, major unanticipated events that take place during the course of a marriage may be destabilizing. The sections that follow examine the empirical evidence on various factors that are related to these circumstances.
2.1. Characteristics of the Spouses and their Match
Husband’s Permanent Earnings. A high level of husband’s potential earnings is expected to generate a positive income effect that increases marital stability. More resources in the household imply a higher standard of living and may alleviate sources of marital tension associated with economic difficulties. In addition, as Becker (1991) notes, high-income men are likely to gain more from marriage insofar as they can compete more effectively in the marriage market and more easily attract wives with the desired characteristics. Empirically, there is strong evidence in the literature that men with a higher earning potential have more stable unions (Robert Michael 1979, 1988; Yoram Weiss and Robert Willis 1997).
Wife’s Earning Potential. An increase in the wife’s earning capacity would be expected to generate an income effect similar to that discussed above in connection with the husband’s earnings. However, because women’s wages are usually lower than men’s, an increase in the wife’s wage (holding husband’s income constant) may reduce the benefits from the division of labor within marriage, increasing the likelihood of separation. An “independence effect” reinforces this influence: a higher wage rate increases a woman’s ability to support herself and her children outside of marriage. Thus among unhappily married women, those with a higher earning potential can more easily terminate their unions. The empirical findings in the literature differ across studies partly because of differences in the point in time at which female earnings capacity is measured (Evelyn Lehrer 1988). In addition, studies differ widely in the extent to which other variables that are correlated with female wages are controlled-- including education, current hours of work, and years of experience. The weight of the evidence, however, suggests that other factors held constant, an increase in the wife’s earning capacity has a destabilizing effect (Michael 1979; 1988; Marianne Ferber and William Sander 1989; Weiss and Willis 1997).
Interaction between Husband’s and Wife’s Earning Capacity. In Becker’s model (1991) gains from marriage are highest when the spouses have different productivity characteristics and can enjoy the benefits from specialization and division of labor. This theory implies that the optimal sorting is characterized by negative assortative mating on wages, individuals with high wages being matched to mates with low wages who have incentives to specialize in home production. However, as David Lam (1988) has pointed out, joint consumption of public goods is another source of gains from marriage. This represents an offsetting force that generates a tendency for positive assortative mating on wages, due to the returns from the spouses having similar demands for public goods.
The empirical evidence on the relative strength of these forces is mixed. James Smith (1979) finds a correlation of .10 between the husband’s and wife’s earning capacities for white couples, and .41 for black couples. Evelyn Lehrer and Marc Nerlove (1984) report a correlation between .10 and .17 for white couples, and between .11 and .25 for their black counterparts, depending on life-cycle stage. These marriage patterns suggest positive, albeit weak assortative mating in this dimension, implying that complementarities between the husband’s and wife’s earning capacities dominate.1 At the same time, Weiss and Willis (1997) report a very wide gap between the predicted earnings of husbands and wives ($25,005 vs. $11,606, based on data from 1985); they also find that the earning capacities of the spouses interact positively in a divorce equation. The authors interpret these results as evidence that complementarities stemming from the consumption of public goods are not sufficient to offset the substitution effects associated with the division of labor.
Educational Attainment. Holding constant the spouses’ earning capacities, education is a complementary trait within the context of marriage, as it has an impact on a wide range of activities in which husband and wife are jointly involved. Education affects the daily communication and interactions within a marriage; education also influences decisions regarding child-rearing approaches, the level of human capital investments in children, and the allocation of time to home production and leisure activities.
Empirically, studies with various data sets and from different time periods consistently report a high correlation coefficient between the spouses’ schooling levels, between .5 and .6 (e.g., Richard Layard and Antoni Zabalza 1979; Lehrer 1999b). Analyses of the determinants of divorce generally find that high levels of the husband’s and wife’s education have a stabilizing effect (Jessie Tzeng and Robert Mare 1995; Weiss and Willis l997). In addition, the interaction between the spouses’ schooling levels reveals strong complementarities: the higher the husband’s level of schooling, the higher is the contribution of the wife’s education to the stability of the marriage (Weiss and Willis 1997).
Intelligence. Like education, intelligence is a trait for which positive assortative mating is optimal, as it affects virtually all aspects of a relationship between marital partners. Evidence on simple correlations between the spouses’ level of ability supports this view (Arthur Jensen 1978). It also appears that individuals of unusually high ability may have a relatively high divorce rate (Gary Becker, Elizabeth Landes and Robert Michael 1977). Being a genius is a rare trait, and it is difficult for such people to find partners of comparable ability; their tendency to be mismatched along this dimension would be a factor leading to marital instability.
Age. Jensen (1978) cites age as the trait for which positive assortative mating is strongest. In the U.S., a gap of about 2 years between husband and wife is typical, with the husband usually the older partner. Large differences in age, especially when the wife is older than the husband, have been found to be destabilizing. The precise patterns vary across studies, however, and the effects appear to differ by race and marriage order (Lehrer 1996a).
Religious affiliation. Religion is a complementary trait within marriage, which affects many activities beyond participation in religious observances at home and at church. Religion influences the education and upbringing of the children, the allocation of time and money, the cultivation of social relationships, and often even the choice of place of residence. Thus households in which the partners differ in religious affiliation are expected to have reduced efficiency, more conflict, and a higher likelihood of dissolution.
Religious heterogamy has indeed been found to have an adverse impact on marital stability (Michael 1979; Lehrer 1996a). Intermarriage, however, comes in different shades and forms, and not all inter-faith unions are equally unstable. The destabilizing effect of intermarriage is most pronounced in two cases. First, when the spouses are affiliated to religions that have highly dissimilar religious beliefs and practices (e.g., a Jew and a Christian). And second, when the affiliation of one or both partners is exclusivist in nature, with sharply drawn boundaries and membership criteria (e.g., Mormons, fundamentalist Protestants) ( Evelyn Lehrer and Carmel Chiswick 1993).
Other dimensions of religion are also important. Among couples in religiously homogamous unions, it used to be that those affiliated to the Catholic and Jewish faiths were especially stable (Michael 1979; Frances Kobrin Goldscheider1986), but it appears that this is no longer the case. By the late 1980s, homogamous Mormon marriages stood out as the most stable of intra-faith unions (Lehrer and Chiswick 1993). Having some affiliation as opposed to none has a positive effect on marital stability (Lehrer and Chiswick 1993), and higher levels of religious participation also have a favorable influence (Tim Heaton and Edith Pratt 1990).
Race and Ethnicity. It is well known that levels of marital instability are substantially higher among blacks than among whites, primarily for reasons related to pronounced differences in socioeconomic status between the two groups (Steven Ruggles 1997). Another contributing factor is the narrower male-female wage rate in the black population (Paul Secord and Kenneth Ghee 1986), which decreases gains from marriage associated with division of labor. In addition, the ratio of marriageable men to marriageable women is lower in the black population. This difference in the sex ratio implies that compared to their white counterparts, black women have a less advantageous position in the marriage market. The risk of union dissolution is thus higher in the black population, as women generally place a higher priority on marital stability than men (Grossbard-Shechtman 1984; 1985; Marcia Guttentag and Paul Secord 1983). There is also evidence that the way in which various factors influence the probability of divorce varies systematically across these two racial groups (Jay Teachman 1986; Lehrer 1996a).
Much less is known about patterns of marital stability for Hispanics, Asians, and other racial and ethnic groups in the United States. In addition, little research has been done on the implications for marital stability of intermarriage across various racial and ethnic lines.
Non-intact family background. Individuals raised in broken homes may have a higher risk of divorce for several reasons. First, such individuals generally begin their unions with an economic disadvantage, and they tend to do so at an early age. Both of these factors increase the odds of an eventual divorce. Second, gains from marriage may be lower for such individuals, to the extent that they have had fewer opportunities to learn some of the interpersonal skills that are essential to a successful union. Third, for persons brought up in non-intact families, the perceived costs of a marital breakup may be lower, as they have seen that divorce is a viable solution to an unhappy marital situation. Paul Amato (1996) reports evidence suggesting that the first two factors play an important role in explaining the intergenerational transmission of marital instability; he finds little support for the third channel. Research by Sara McLanahan and Larry Bumpass (1988) suggests that the timing of the disruption matters. Compared to individuals who experienced a parental divorce in middle childhood, the adverse effects are stronger for those who experienced it before the age of 5 or during the adolescent years.
2.2. Investments in Marriage-Specific Human Capital
The behaviors of each spouse during the course of their marriage have an important impact on the stability of the union. Becker et al. (1977) emphasize the role of investments in marriage-specific human capital, i.e., those investments that decline substantially in value following the termination of a marriage. Chiswick and Lehrer (1990) refine this notion by distinguishing marriage-specific investments that are transferable among unions from those that are specific to a particular spouse. After the dissolution of a union, the value of the former can be restored through remarriage; in contrast, the decline in the value of the latter is irreversible.
Spouse-Specific Human Capital. Couples who invest in spouse-specific capital enhance the stability of their marriage because, by definition, dissolution of the union would cause an irreversible decline in their value. Examples of spouse-specific investments might include acquiring knowledge about the spouse’s culture, learning about leisure activities the spouse particularly enjoys (e.g., a specific hobby or sport), and learning to get along with members of the spouse’s family. However, by far, children shared with a spouse represent the single most important type of investment in spouse-specific capital. The value of investments in children declines after a divorce, in part because of the tendency for the level of expenditures on children to fall to an inefficiently low level (Weiss and Willis 1985). Under-provision of the couple’s collective good, child expenditures, is likely to occur because of the father’s lack of control over the allocation of resources by the mother, who is typically the custodial parent. To the extent that both parents value their children’s welfare, both would suffer.
Several other adverse changes in connection with the children usually happen after the breakup of a marriage. First, the non-custodial parent has much less contact with the children and the connection between parent and child is weakened (Frank Furstenberg et al. 1983). Second, from the custodial parent’s perspective, the presence of a child from a previous union makes remarriage more difficult and also has an adverse effect on the stability of a future union (Chiswick and Lehrer 1990; Lehrer 1996a). A third consideration is related to the fact that children generate consumption externalities flowing between the parents (Lehrer 1996a). This is one of the important psychic returns from children. For instance, a child’s first steps yield utility to the mother and the father, and each parent derives utility from the other’s enjoyment. This component of utility disappears following divorce.