New Partners, More Kids: Multiple-Partner Fertility in the United States

Abstract

Declining rates of marriage and overall increases in union instability, combined with high levels of unintended and nonmarital fertility, create the possibility for parents to have children with more than one partner, called multiple-partner fertility, or MPF. The unique characteristics of families with MPF present data and other logistical challenges to researchers studying the phenomenon. Drawing from recent studies and updated data, I present new estimates of MPF that show that about 13 percent of men aged 40 to 44 and 19 percent of women aged 41 to 49 have children with more than one partner, with a higher prevalence among the disadvantaged. Compared to parents with two or more children by only one partner, people with MPF become parents at younger ages, largely with unintended first births, and often do so outside of marriage. This article touches on the implications of MPF for families and concludes by discussing the theoretical difficulties in studying MPF and the challenges it presents to public policy.

Recent changes in union formation and stability, combined with changes in childbearing behaviors, have complicated the ways in which researchers and individuals think about families. In the not-too-distant past, most families consisted of married parents and their shared children. Today, this is less likely to occur, even though a stable partnership with two or more children by only one partner remains the preferred family life path for Americans (Thomson, Winkler-Dworak, and Kennedy 2013). The United States has fairly high rates of teenage, nonmarital, and unintended childbearing in addition to high rates of dissolution among nonresidential unions, cohabitations, and even marriages. As such, men and women are increasingly exposed to, and actually are, having children with more than one partner, a phenomenon known as “multiple-partner fertility,” or MPF.

Although the nomenclature is fairly new, MPF is not exactly a new behavior. In the late eighteenth and nineteenth centuries, it was not uncommon for a spouse and parent who had been widowed or deserted to remarry and have additional children in a new relationship (Degler 1980). In the mid-twentieth century, young unmarried women who became pregnant often put their child up for adoption, then later formed marriages with another man and had additional children; this, too, would be MPF, though we have no estimates of the prevalence of such behavior. Also in the twentieth century, divorce surpassed spousal death as the primary way marriages among men and women in their childbearing years ended, with MPF occurring when a married parent divorced, remarried, and had additional children with his or her new spouse (Logan et al. 2006). By the late twentieth and early twenty-first centuries, however, rising rates of nonmarital childbearing (but lower rates of giving up children from nonmarital births) (Jones 2009; Manlove et al. 2010) and declining rates of remarriage (Brown and Lin 2013) have again shifted the character of MPF, with some or all births across partnerships occurring outside of marriage. Clearly, there is a fair degree of overlap between MPF families and stepfamilies (Sweeney 2010). It is worth noting, though, that not all stepfamilies include a shared child (and thus there is no MPF), and not all MPF families are coresidential, which means they are largely excluded from the stepfamily literature because stepfamilies are generally identified via coresidence. Furthermore, stepfamilies in which all children are born within marriage differ from MPF families in which some or all children are born outside of marriage. The legal privileges and responsibilities are well established for children born within marriages, and there is a large literature addressing custody, visitation, and child support among divorced families (Amato 2010). However, although there is a growing literature focusing on unmarried parents and unmarried stepfamilies (McLanahan and Beck 2010), research gaps remain, and policy lags even further behind in addressing the unique but complicated needs of these parents and their children (Cancian and Meyer 2012).

Consider child support guidelines, which usually mandate that a specific proportion of the noncustodial parent’s income go to his or her child(ren), with the first child commanding a greater proportion than subsequent children. In Illinois, for instance, a noncustodial parent with only one child should pay 20 percent of his or her income, 28 percent for two children, 32 percent for three children, 40 percent for four children, and so on (Illinois Child Support 2014). These guidelines seem to assume that all the children live together with one custodial parent, but what if there were more than one custodial parent? Imagine a noncustodial parent with only one child, paying 20 percent of his or her income in child support, who goes on to have a child with a new partner. If the new relationship ends, and he or she does not have custody of the second child, how should child support be calculated? According to the guidelines, support should total 28 percent of income when there are two noncustodial children—does this mean the second child receives only 8 percent of the noncustodial parent’s income, so that the support paid for the first child does not change, or should the support paid for the first child be reduced to make support more equitable across children? Alternatively, if the court uses its discretion to deviate from the guidelines, should each child get 20 percent of the parent’s income, thus costing the noncustodial parent 40 percent of his or her income? There are no clear answers for such questions. As such, the changing nature of MPF has implications for policy as well as for parents’ and children’s experiences.

At its core, MPF is driven by the demise of a parental relationship; there is no risk of having children by multiple partners if the original relationship is intact. A relationship may end because two individuals never intended to be parents together but experienced an unplanned pregnancy, and sometimes parents who deliberately had children in a partnership they expected to last break up. To understand MPF, then, researchers must understand both the role of childbearing within relationships and patterns of union formation, dissolution, and repartnering. In this article, I review theories about why people have children; situate changing fertility behaviors within changes in union formation and stability; and, after briefly discussing some methodological issues, provide new estimates of MPF. Finally, I discuss the implications of MPF for individual and family wellbeing and directions for future theoretical development and research.

Why Do People Have Children?

Despite the rising opportunity costs of childbearing and a growing number of competing behaviors, most people in the United States still become parents, prompting demographers and family scholars to theorize about why people have children (Morgan and King 2001). Traditionally, economists have viewed childbearing from a rational choice perspective, where people are seen as weighing the costs and benefits of having children and having as many children as they can afford and when the benefits outweigh the costs (Becker 1981; Hotz, Klerman, and Willis 1997). In developed societies, with their emphasis on “high-quality” children requiring substantial investment, the process of weighing costs and benefits tends to translate into small families. Furthermore, economists position children as “relationship-specific capital” in which children signal long-term commitment, discourage dissolution, and ultimately reduce union uncertainty (Friedman, Hechter, and Kanazawa 1994). Under this perspective, having children is selective of the most stable couples (Lillard and Waite 1993; Myers 1997).

Anthropologists and sociologists also note the role that children serve in creating social capital (Coleman 1988), linking individuals to broader social networks and providing parents with social status (Astone et al. 1999), serving as a marker of adulthood (Shanahan 2000), and establishing a couple as a family (Holland and Thomson 2011). And despite declining pronatalism, social and normative pressures to have children still exist (Morgan and Rackin 2010), and American women, by and large, still view motherhood as an important social role (McQuillan et al. 2008). There are also intangible reasons for having children, such as physiological and behavioral dispositions toward reproduction (Morgan and King 2001) and the emotional and psychological rewards of parenting (McMahon 1995). And despite a downward trend in family size, two children remains the norm (Morgan and Rackin 2010), though it could be that second children are “valued primarily as siblings for the first child” (Thomson 2004, 119). In any case, this suggests that parents in developed countries see few reasons to have larger families (three or more children), although there is still some inclination, perhaps weakening, for couples with two sons or two daughters to have a third child in an attempt to have a different sex child (Andersson et al. 2006; Pollard and Morgan 2002).

Theories about why people and couples have children often make several assumptions. First, fertility is largely assumed to be planned; that is, individuals decide to have children because they see some sort of benefit or value and then choose to have children when they are best able to maximize the benefits and minimize the costs. In the United States, however, over a third of all births are unintended (Mosher, Jones, and Abma 2012), occurring either too soon (“mistimed”) or to individuals who do not want any births at all (“unwanted”). High proportions of unintended childbearing suggest that many people are not consciously deciding to signal relationship commitment, present themselves as adults, or follow a normative path. Second, economic theories seem to assume that parents have the number of children they anticipate they can “afford” given the social, relational, and financial commitments childrearing entails. However, since a substantial minority of nonresidential parents are only minimally involved, if at all (Cheadle, Amato, and King 2010), having many children across relationships entails few costs for the subset of parents who do not financially support or spend time with their children. Third, many of these theories place, rightly, the parental union at the center of childbearing and child-rearing. After all, the biological parents are embarking on a multiyear emotional, financial, and legal commitment together. The problem, though, with this assumption is that all parental unions are not the same in terms of stability and strength. A substantial minority of children are conceived in unstable and weak relationships in which there is little explicit or concrete discussion about childbearing or the future of the union (Augustine, Nelson, and Edin 2009; Edin et al. 2007). Couple disagreement on the intendedness of children is not uncommon, either, especially among unmarried parents (Hohmann-Marriott 2009; Korenman, Kaestner, and Joyce 2002). Because few existing theories of fertility simultaneously include birth intentions, parental involvement and investment in children, and relationship variation, our ability to understand increasingly complex fertility patterns is severely hampered.

What do sexual and union behaviors contribute to MPF?

Americans continue to believe that the best environment in which to have and raise children is a stable marriage (Pew Research Center 2010). Although Americans still value marriage, they are delaying it as the perceived social and economic prerequisites for marriage have grown (Cherlin 2010). Delays in marriage have, for some groups, translated into never marrying; race-ethnic minorities and those with lower levels of education are least likely to ever marry (Cohn et al. 2011). The delay in marriage, however, is not accompanied by a delay in sexual debut. Since the 1970s, the median age at first sex has been around 17, and more than 75 percent of Americans have had premarital sex by age 20 (Finer 2007). The implication, then, is that if people prefer to have children within marriage but are engaging in nonmarital sexual activity, there is a lengthy exposure period for nonmarital and unintended childbearing. In itself, this “exposure period” is not problematic if individuals use contraception consistently and correctly, yet we know this is not always the case, especially in the United States (Jones, Mosher, and Daniels 2012).

As individuals have delayed marriage, cohabitation has grown more popular than it once was, but it is less stable than marriage, particularly for the disadvantaged (Lichter, Qian, and Mellott 2006). Of course, marriage itself is no guarantee of union stability, with about half of marriages ending in divorce and higher instability among the disadvantaged (Raley and Bumpass 2003). What is worth emphasizing, however, is that Americans are quite optimistic about relationships and the future (Cherlin 2009), often having children in unions that later dissolve. The instability of both cohabitations and marriages, however, has not seemed to dissuade people from seeking out new partnerships (Cherlin 2009), creating the possibility of having additional children with a new partner.

To understand MPF, we must situate fertility behavior within changes in sexual and union behavior. However, even though there is substantial overlap in the correlates and consequences of teenage, nonmarital, and unintended childbearing, each of these topics is largely studied within a distinct body of research, and it is not uncommon for research to focus on just one of these behaviors—say, marital status—and give only cursory attention (or none at all) to other aspects, such as age or unintendedness. The limited research on MPF—driven by the instability of parental relationships—seldom pays explicit attention to the formation and dissolution of unions. Because research often narrowly focuses on just one aspect of family behaviors, it is difficult to get a cohesive and comprehensive portrait of the interrelated behaviors that ultimately contribute to MPF. Here, I focus on some key fertility behaviors and their linkages to union behavior.

Although teen fertility reached a historic low in 2012 after peaking in the early 1990s (Hamilton, Martin, and Ventura 2013), it remains higher in the United States than in any other industrialized country. The majority of teenage childbearing is unintended and nonmarital, and repeat teen births are common (Mosher, Jones, and Abma 2012). Nonmarital and unintended childbearing, of course, are not solely the domain of teenagers. The proportion of births occurring outside of marriage has steadily increased, reaching 41 percent in 2010, about half of which occur in cohabiting unions (Manlove et al. 2010). A third of all births are unintended, with births to unmarried women particularly likely to be unintended (Mosher, Jones, and Abma 2012). The majority of unintended births arise from using less effective forms of contraception, using contraception inconsistently, or nonuse of contraception (Gold et al. 2009), and these births are concentrated among younger and less advantaged men and women (Finer and Zolna 2011). While accurately understanding contraception and the reproductive process, along with access to affordable family planning services, plays a role in high proportions of unintended fertility among the disadvantaged (Gold et al. 2009; Guzzo and Hayford 2012a), there is also substantial ambiguity about childbearing that influences sex and contraceptive behaviors (Edin et al. 2007). Less-advantaged individuals have fewer opportunity costs related to unintended childbearing and may find greater social and personal rewards in parenthood, even if it is not explicitly planned (Edin and Kefalas 2005). Gender differences in views about relationship commitment and longevity, combined with differences in views on responsibility for contraception (Augustine, Nelson, and Edin 2009; Edin and Kefalas 2005), also influence the chances of having an unintended birth. Relationships in which a teenage, unintended, or nonmarital birth occur are at an elevated risk of union dissolution (Guzzo and Hayford 2012b; McLanahan and Beck 2010), creating a feedback loop between these fertility behaviors and union instability. Relationship instability and teenage, nonmarital, and unintended childbearing are more common among race-ethnic minorities and socioeconomically disadvantaged groups, further widening differences in family behavior (Amato 2010; Smock and Greenland 2010).

Thus, changes in sexual and union behavior have fundamentally altered the contexts in which people have children. The increasing proportion of births to unmarried women is driven, in part, by changes in marital behavior—there are fewer married women, and married women are having fewer marital births (Gray and Stone, forthcoming; Hummer and Hamilton 2010), with declines in marriage among blacks and native-born Hispanics driving the high proportion of nonmarital births in these groups (DeLeone, Lichter, and Strawderman 2009). At the same time, observed changes in fertility behavior are not solely the result of changes in exposure. As the stigma against nonmarital childbearing has declined, the nonmarital birth rate has increased among whites and foreign-born Hispanics (DeLeone, Lichter, and Strawderman 2009), and those who have a first nonmarital birth are decreasingly likely to ever marry (Gibson-Davis 2011), raising the risk of subsequent nonmarital births. These reinforcing mechanisms are driving the increased complexity seen in today’s fertility patterns (as compared to the mid-twentieth century), with exposure to less favorable and stable contexts growing, while Americans are simultaneously becoming more likely to have children in these contexts.

The observed changes in fertility and union behaviors make the conditions in which MPF occurs more common. With fairly high rates of demise among parental unions, parents are forming new unions; these unions are usually cohabitations, only some of which later transition to marriage. For instance, Manning and Cohen (2010) find that by their child’s third birthday two-thirds of teen parents are cohabiting and 15 percent have married, most likely with a new partner. Similarly, within five years of a nonmarital birth, 20 percent of mothers who were cohabiting at birth and 27 percent of those who were romantically involved with their child’s father have had at least one new cohabiting or marital union (McLanahan and Beck 2010). Within a year of filing for divorce, about half of previously married parents are in new relationships they describe as serious (Anderson et al. 2004), and over half remarry within five years, usually after cohabiting with their partner before marriage (Bramlett and Mosher 2002).

Having additional children in these new unions is common. In the best case scenario, shared childbearing with a new partner is planned, functioning as a way to establish legitimacy for the new union (i.e., “commitment value”; Holland and Thomson 2011). This sometimes results in “extra” children that individuals at certain parities would not have had if their original relationship remained intact, as one or both parents revise their ideal number of children to accommodate childbearing in the new union (Iacovou and Tavares 2011). Having a child with a new partner is more likely when the new partner does not have biological children, although couples have to weigh the benefits of a shared child against the costs of an overall higher couple-level family size (Holland and Thomson 2011). However, we also know that individuals with one unintended birth are more likely to have another unintended birth (Guzzo and Hayford 2011), so it is likely that some new-partner fertility is unintended.

Estimates of MPF

Given these dramatic changes in union and fertility behaviors, researchers have increasingly sought to identify, document, and study childbearing across partnerships, but this has proven to be a surprisingly difficult task. One of the difficulties in studying MPF is that it has fairly high data requirements—data explicitly identifying the partners with whom individuals had each child or, at a minimum, a direct question asking individuals about the number of partners with whom they have had children. Unfortunately, traditional fertility data usually collect fertility histories by dates, which are then matched with marriage and cohabitation histories to determine which births occur in a marriage, cohabiting union, or outside of a coresidential union; this tells us little about the context of noncoresidential-union births. And until recently, many surveys did not include men; even the surveys that did likely underestimated their fertility, particularly among disadvantaged men (Joyner et al. 2012). Still, advances in survey methodology, creative (and laborious) attention to household rosters, and access to administrative data have provided opportunities to estimate the prevalence and correlates of MPF, as well as examine how MPF affects individual, family, and child wellbeing. An exhaustive literature search using the terms “multipartnered fertility,” “multiple-partner fertility,” and “stepfamily fertility” produced more than a dozen published or working papers that explicitly measured MPF prevalence, with several other papers examining other aspects of MPF.

In reviewing existing research, several things became apparent. First, the estimates vary widely due to the broad range of data sources, measurement approaches, and study populations. Some of the estimates are derived from relationship matrices in family and household rosters (National Longitudinal Survey of Youth 1979, Survey of Income and Program Participation, Current Population Surveys; all of these are nationally representative), though these sometimes count MPF only within the household (i.e., reported by the household head or individuals included in the household roster). Many estimates are from the Fragile Families and Child Wellbeing Study, which follows a cohort of (largely) nonmarital births in major cities; low response rates among unmarried fathers mean that most estimates of men’s fertility behaviors come from the mothers’ reports. One data source with higher response rates among men, the nationally representative National Survey of Family Growth, allows researchers to obtain estimates directly from questions indexing men’s fertility to specific partnerships. Unfortunately, this survey did not collect partner-specific fertility information for women; this is problematic for women who have more than one birth outside of a coresidential union. The National Longitudinal Study of Adolescent Health, another national dataset, collects fertility information for men and women in the context of relationships, but the sample is fairly young, and the school-based sampling frame may exclude more disadvantaged individuals. Another set of publications uses administrative data from Wisconsin’s welfare and child enforcement systems; these data capture more disadvantaged individuals and rely on formal paternity establishment to identify father’s fertility. Thus, existing data sources are often less than ideal for producing nationally representative and comprehensive estimates of MPF.

Second, the studies often use different subpopulations and units of analysis, ranging from the full population to parents with one or more children, to parents with two or more children (the only group technically able to have MPF). Similarly, many studies present estimates for limited age ranges (sometimes by design, others by necessity) or subgroups such as unmarried parents. Third, the estimates depend on what type of MPF the research is investigating. Some studies measure just the mother’s (or father’s) MPF at a point in time or at the end of childbearing, others examine whether MPF exists within a family or household, and still other studies focus on the child’s experience of their parents’ MPF. These different perspectives, often drawn from different types of data, tend to yield prevalence estimates that vary widely. Many estimates also miss out on overall family or household complexity, as data about only the respondent’s MPF are unable to account for the fact that past or current partner(s) may also have MPF (see Burton, this volume). There is nothing inherently wrong with any of these approaches, but the variability in data and measurement complicates comparisons of estimates and getting a sense of the “big picture.” Finally, nearly all of our estimates are likely to be underestimates for many reasons—many of the sample populations have yet to reach the end of their childbearing years, children outside of the household are excluded, underreporting of nonmarital births is common among men, women’s reports of their partners’ fertility is incomplete, and so on.

With those caveats in mind, I present new estimates of MPF from two nationally representative data sources (wave IV [2007–08] of the National Longitudinal Study of Adolescent Health [Add Health] and cycle 7 [2006–2010] of the National Survey of Family Growth [NSFG]), along with estimates from other surveys, only some of which are nationally representative. Add Health is a school-based survey first conducted in 1995 of adolescents in grades 7 to 12, followed up in 1996, 2001–2002, and 2007–2008. It is one of the few surveys that collects both men’s and women’s fertility histories within relationship contexts, enabling identification of MPF for both genders (an advantage not available in any other survey). Although the respondents in wave IV are aged 25 to 32 (born between 1974 and 1983) and have not yet completed childbearing, many respondents have begun having children by their late 20s. However, fertility during these years—and thus the fertility estimates presented here—are skewed toward the less advantaged, given that well-educated young adults tend to delay childbearing as they complete education and establish careers. The NSFG is a nationally representative survey of individuals aged 15 to 44, and is considered the best source of U.S. fertility data. It contains partnership information for each child of a man, though it did not collect parallel information for women; I draw information from men aged 40 to 44 (born between 1962 and 1970) to isolate those who have largely completed childbearing. Because these estimates are drawn from different sources, they are not directly comparable. Furthermore, Joyner et al.’s (2012) results suggest that the male underreporting in the NSFG may amount to a fifth of all births and is particularly pronounced among nonmarital births. Male underreporting is similarly likely to occur in Add Health; as such, the estimates for men presented here are almost certainly underestimates, and gender comparisons should be made with caution.

Table 1 summarizes the prevalence of MPF in Add Health and the NSFG along with recent estimates from other sources. Estimates are presented for three groups: all adults, all parents, and parents with at least two children. Looking first at overall prevalence among those in their 40s (who have largely completed childbearing), the prevalence of MPF is around 13 percent for men aged 40 to 44 in the 2006–2010 NSFG. Estimates among a similar age group for women (the National Longitudinal Survey of Youth 1979) show that 19 percent of women aged 41 to 49 in 2006 (born between 1958 and 1965) had MPF (Dorius 2011; see Dorius [2012b] for an explanation of how measures were created and why estimates for men are unavailable). Among those in their mid-20s and early 30s in Add Health in 2007–08, about 7 percent of men and 12 percent of women have had children with two or more partners; estimates for men aged 25 to 32 in the NSFG (not shown) are similar but slightly lower, at 6.1 percent, suggesting that Add Health estimates for men are on par with what is seen in other datasets. The lower estimates seen for men relative to women in Add Health likely stem from a combination of greater fertility underreporting among men (Joyner et al. 2012) along with the earlier entrance into parenthood among women. Thus, in the population as a whole, the prevalence of MPF increases with age as exposure to childbearing and union instability increases.

Table 1.

Recent Estimates of Multiple-Partner Fertility, by Author(s) and Source

Authors	Data source(s); estimate year(s)	Overall prevalence	Prevalence among all parents	Prevalence among parents with at least 2 children
Author’s calculation	National Survey of Family Growth (men aged 15–44); 2006–10	13.1% of all men aged 40–44	17.1% of all fathers aged 40–44	22.5% of fathers aged 40–44
Dorius (2011)	National Longitudinal Survey of Youth 1979 (women aged 14–21 in 1979); 2006	18.7% of all women aged 41–49		27.7% of all mothers aged 41–49
Evenhouse and Reilly (2011)	Survey of Income and Program Participation (women aged 15 and older); 2008		7.5% of all mothers 15 and older who have biological children residing in the household
Author’s calculation	National Longitudinal Study of Adolescent Health (enrolled in grades 7–12 in 1995), wave IV; 2007–08	6.9% of all men aged 25–32 12.1% of all women aged 25–32	16.6% of fathers aged 25–32 22.1% of mothers aged 25–32	32.4% of fathers aged 25–32 37.9% of mothers aged 25–32
Scott et al. (2013)	National Longitudinal Survey of Youth 1997 (men aged 12–16 in 1997); 2008		13.7% of all fathers aged 23–27 47.6% of fathers aged 23–27 who were unmarried and non-resident at first birth
Fomby and Osborne (2013)	Fragile Families and Child Wellbeing Study (birth in a major city in 1998–2000); 2007–10 (9-yr follow up)		47.9% of mothers who had a birth in a major city in 1998–2000 57.3% of mothers who had a nonmarital birth in a major city in 1998–2000 18.6% of mothers who had a marital birth in a major city in 1998–2000
Cancian, Meyer, and Cook (2011)	Wisconsin child support enforcement data system and related administrative data as of June 2008		60% of firstborn children to unmarried mothers in 1997 have experienced parental MPF by age 10

Among all parents—which is the estimate produced by most studies—the proportion with children by more than one partner is (not surprisingly) higher than among the population overall, but the estimates vary widely based on the data source, population, and analytical sample. Among fathers aged 40 to 44 in the NSFG, about 17 percent have MPF; there is no analogous estimate for mothers in their 40s from a nationally representative sample. Evenhouse and Reilly (2011), using data from the Survey of Income and Program Participation (SIPP), found that about 8 percent of all mothers 15 and older with coresident biological children have MPF, but the broad age range of their sample, combined with the inability to account for nonresidential children (including grown children), seems to produce very low estimates. For young adults in their 20s and early 30s in Add Health, slightly less than 17 percent of fathers and 22 percent of mothers have MPF; recall that parenthood in this age group is somewhat selective of less advantaged individuals. In a slightly younger group of men aged 23 to 27 in the National Longitudinal Study of Youth 1997, 14 percent of fathers have MPF, rising to 48 percent among the most disadvantaged fathers—those who were unmarried and nonresident when their first child was born (Scott et al. 2013). Estimates are similarly high among samples of other disadvantaged populations— nearly 60 percent of the unmarried mothers surveyed at the baseline of the Fragile Families and Child Wellbeing Study had had a child with a different partner by the nine-year follow-up (Fomby and Osborne 2013), and 60 percent of firstborn children whose mother entered the welfare system in Wisconsin in 1997 had a half-sibling through one or both parents by their 10th birthday (Cancian, Meyer, and Cook 2011).

When we look at the prevalence among the only group that is technically eligible for MPF—those with two or more children—we see even higher proportions. Among those in their 40s with two or more children, using data from the NSFG, about 23 percent of men aged 40 to 44 (author’s calculation) and 28 percent of women aged 41 to 49 have MPF (Dorius 2011). Interestingly, the prevalence is higher at younger ages among those with two or more children. Among those aged 25 to 32, 32 percent of fathers and 38 percent of mothers have two or more children with multiple partners. The more children a person has at younger ages, the more likely it seems that those children are by multiple partners. This is due to the selectivity of childbearing at earlier ages; early childbearing occurs primarily among the disadvantaged, and the relationships in which such births occur are highly unstable (Edin and Tach 2012; Guzzo and Furstenberg 2007b). Conversely, as more individuals become parents with age, they are more likely to have their children within stable unions (Dorius 2012a), essentially diluting the higher proportions seen among young parents with two or more children. Moreover, differences in the rate of MPF at different ages may reflect different union and fertility patterns for those born in different cohorts.

Regardless of the data source and specific estimate, studies have repeatedly shown that MPF is considerably higher among blacks and Hispanics as well as those with lower levels of education (Carlson and Furstenberg 2006; Guzzo and Furstenberg 2007a, 2007b). Similarly, MPF has also been linked to other indicators of social and economic disadvantage, such as parental incarceration (Carlson and Furstenberg 2006) and receiving public assistance (Meyer, Cancian, and Cook 2005); related aggregate factors linked to nonmarital childbearing and union instability, such as a shortage of available mates in an area, unemployment, and incarceration rates, increase the likelihood of MPF as well (Evenhouse and Reilly 2011).

Guzzo and Furstenberg (2007a), using the 2002 NSFG, also found that younger cohorts are transitioning to MPF sooner than older cohorts; updated analyses (not shown) confirm this pattern. For instance, of those aged 30 to 34 in the 2006–2010 NSFG, 9.2 percent of all men, 15.0 percent of all fathers, and 23.0 percent of fathers with two or more children had MPF. Looking at the fertility of those aged 40 to 44 at the time of interview when they were aged 30 to 34, the comparable figures are 7.8 percent, 12.9 percent, and 19.3 percent, respectively. Similarly, of men aged 20 to 24 at the time of interview in 2006–2010, 1.1 percent of all men, 7.3 percent of all fathers, and 29.9 percent of those with two or more children had MPF; the comparable figures for those aged 40 to 44 at the time of interview when they were aged 20 to 24 are 1.0 percent, 5.9 percent, and 19.5 percent, respectively. Although the sample sizes for parents (especially those with two or more children) at younger ages are quite small and the magnitude of the differences is not large, the pattern nevertheless suggests that MPF is likely to increase as younger cohorts age.

The Path to MPF

The circumstances in which people have children, particularly how they enter parenthood, are crucial for whether they go on to have children with new partners. The fertility behaviors discussed earlier—age, intendedness, and relationship context—are all linked to MPF. I explore these key characteristics at first birth and MPF using wave IV (2007–08) of Add Health and cycle 7 (2006–2010) of the NSFG among those with at least two children. The two surveys, populations, and analytical samples are not directly comparable—the young parents in Add Health who have already had at least two children are a highly select group that is more disadvantaged than both their peers without two children and those who will have two children at some later age. The NSFG estimates are limited to men aged 40 to 44, making selection less of an issue (though for both the NSFG and the males in Add Health, underreporting has almost certainly occurred, and readers are cautioned against reading too much into gender differences in Add Health). Still, what can be seen from the data presented below is that although the exact estimates vary, the general patterns are fairly similar across gender and across data sources.

Looking at the top panel of Tables 2 and 3, compared to their peers with two or more children by the same partner (single-partner fertility, or SPF), men and women with MPF have more children, tend to start childbearing earlier, are substantially less likely to have intended their first birth, and are far less likely to have had their first birth in a coresidential union. The young parents in Add Health with MPF (Table 2) started childbearing, on average, over two years earlier than their peers with SPF. Among the older men in the NSFG (Table 3), the difference is greater; the average age at first birth among those with MPF is more than four years earlier than those with children by one partner. Only around a third of those with MPF in Add Health categorized their first birth as wanted compared to about 60 percent of those with SPF. And as seen in past research, relationship differences are substantial. Sixty percent of 25- to 32-year-old mothers with MPF were neither cohabiting nor married at the time of their first birth; in contrast, among those with SPF, 53 percent were married and 27 percent were cohabiting. For young men, the pattern is similar, but the magnitude of differences is smaller, likely reflecting more underreporting of nonunion births among men (Joyner et al. 2012), with just under half of first births occurring outside of a coresidential union among those with MPF compared to 18 percent of those with SPF.

Table 2.

Selected Parental Characteristics by Single vs. Multiple-Partner Fertility among Men and Women Aged 25–32 with Two or More Children in the National Longitudinal Survey of Adolescent Health (Add Health)

	Add Health women (n=2,523)		Add Health men (n=1,433)
	SPF	MPF	SPF	MPF
Fertility characteristics
Average number of births	2.32	2.73	2.26	2.86
Mean age at 1st birth	21.8 yrs	19.1 yrs	22.9 yrs	20.7 yrs
Intended 1st birth	60.0%	30.1%	62.5%	35.7%
Union type at 1st birth
Non-coresidential	20.3%	60.5%	18.0%	48.1%
Cohabiting	27.2%	27.1%	33.8%	34.2%
Married	52.5%	12.4%	48.2%	17.7%
Only marital births	51.8%	4.9%	47.6%	6.2%
Only non-coresidential births	3.9%	19.1%	2.6%	13.1%
Parental well-being
Depressed (CES-D)	24.5%	34.3%	13.6%	20.1%
Happy in role of parent
Strongly agree	79.6%	70.6%	77.3%	58.2%
Agree	18.0%	25.4%	17.1%	24.0%
Neither agree nor disagree	1.4%	2.8%	3.2%	8.2%
Disagree	0.9%	0.8%	1.8%	6.5%
Strong disagree	0.2%	0.6%	0.6%	3.0%
Feel close to child(ren)
Strongly agree	89.0%	81.3%	84.0%	66.6%
Agree	9.8%	17.1%	12.3%	21.7%
Neither agree nor disagree	0.7%	1.2%	1.8%	5.3%
Disagree	0.3%	0.4%	1.7%	4.5%
Strong disagree	0.2%	0.1%	0.3%	1.9%
Child coresidence
Lives with all of their children	97.2%	87.4%	85.1%	13.3%

Table 3.

Selected Parental Characteristics by Single vs. Multiple-Partner Fertility among Men aged 40–44 with Two or More Children in the National Survey of Family Growth

	NSFG men aged 40–44 (n=702)
	SPF	MPF
Fertility characteristics
Average number of births	2.48	3.19
Mean age at 1st birth	27.0 yrs	22.7 yrs
Union type at 1st birth
Non-coresidential	5.6%	40.1%
Cohabiting	18.5%	21.6%
Married	75.8%	38.4%
Only marital births	75.6%	13.8%
Only non-coresidential births	2.6%	4.5%
Parental well-being
Of those with a nonresidential child, how good a job as a father
Very good job	27.7%	11.6%
Good job	36.5%	26.7%
Okay job	21.3%	27.8%
Not very good job	12.2%	19.9%
Bad job	2.3%	14.1%
Child coresidence
Lives with all children	63.7%	6.9%

As shown in Table 3, compared to the young fathers with MPF in Add Health, the older fathers with MPF in the NSFG were more likely to be married when they had their first birth but less likely to be cohabiting. Still, fewer older fathers with MPF were married, and more were neither cohabiting nor married at their first birth than their peers whose children were with the same partner. Finally, although the bottom rows of the top panel of Tables 2 and 3 indicate that it is rare for those with two or more children to have had their children entirely outside of coresidential unions, regardless of MPF, only about 5 percent of the MPF men and women in Add Health and 14 percent of the older men in the NSFG had all their births within marriage. Put differently, nearly all of those with MPF had at least one nonmarital birth.

The implications of MPF

Parental MPF has implications for the well-being of both parents and children, ranging from parental mental health to living arrangements to risky behavior among children and adolescents. Although causality is difficult to ascertain, the lower panel of Tables 2 and 3 reveals clear differences in parental well-being by family complexity. For young parents, a greater proportion of those with children by more than one partner meet the clinical definition of depression than those with children by only one partner. There also seems to be gender differences in how MPF is experienced, and the magnitude of these differences suggests that they cannot be attributed to underreporting and age disparities in the timing of fertility. Fathers are especially less satisfied with their parenting when they have MPF; only 58 percent of young fathers strongly agree that they are happy in their role as a parent, and only 67 percent strongly agree that they are close to their children (the corresponding percentages among fathers with SPF are 77 percent and 84 percent, respectively). Among older fathers with two or more children in the NSFG (Table 3), 14 percent characterize themselves as doing a “bad job” of parenting their nonresidential children compared to only 2 percent among men with SPF. The NSFG data also reveal differences in more formal relationships between older fathers and children (not shown)—of fathers with two or more children and at least one nonmarital birth, only 4.5 percent of men with SPF failed to establish legal paternity for at least one of their children compared to 13.5 percent of men with MPF.

Finally, there are differences in coresidence as well, and again, gender plays a role. The vast majority of young mothers with two or more children, regardless of MPF status, live with all of their children, although it is lower among mothers with MPF (87 percent) than among those with SPF (97 percent). For young fathers, however, the differences are substantial. Eighty-five percent of young fathers with two or more children by the same partner live with all of their children, compared to only 13 percent of those with children by multiple partners; this reflects that most children live with their mothers when their parents are not together. We see a similar pattern across older fathers in the NSFG in Table 3 (though here some of the lack of coresidence is because older children no longer live with either parent). Fewer older fathers live with their children across fertility partner statuses, with 64 percent of fathers with SPF and only 7 percent of those with MPF living with all of their children.

Although not shown in the tables here, prior work also suggests that child wellbeing varies by the presence of SPF versus MPF, though the underlying mechanisms are unclear. Children who have half siblings exhibit more delinquency and externalizing behavior (Fomby and Osborne 2013) and have sex earlier and report more drug use (Dorius and Guzzo 2013) than their counterparts without half siblings. These may be direct effects, if parental MPF reduces the quality of parenting, by leading to parental depression and lower levels of parental engagement (Bronte-Tinkew, Horowitz, and Scott 2009). Parents with MPF have lower levels of social support (Harknett and Knab 2007) and weaker coparental relationships (Carlson and Furstenberg 2007), and men who have children with a new partner often reduce visitation with and child support to prior children (Manning and Smock 1999, 2000). Alternatively, the selectivity of MPF families cannot be ignored; children growing up in families with MPF live in less advantaged households and, it seems, have less involved fathers than in families with SPF. Evidence suggests that maternal repartnering may occur because biological fathers are uninvolved (Kotila and Kamp Dush 2012). It may be also be the case that family structure and instability, rather than MPF itself, affect child wellbeing (Cavanagh 2008), although both studies cited above included family structure/ transition indicators that reduced, but did not eliminate, the association between MPF and child well-being.

There are two important caveats to keep in mind when examining the potential implications of MPF. First, any implications likely vary across mothers and fathers. While mothers usually live with all of their children regardless of MPF, for fathers with MPF, the most common scenario is to have biological children in the household with their current partner and nonresidential biological children with a past partner. As such, their interactions with their children will inevitably differ by coresidence. Second, MPF is not experienced in the same way by all children within a family. A firstborn child whose parents have children with new partners is, in most cases, experiencing the dissolution of his/her biological parents’ union, parental repartnering and thus stepfamily situations (often by both parents), and then the birth(s) of a half sibling. For the youngest child, although they have older half siblings, they may be born into an intact family and may remain in this family structure throughout childhood. Thus, an older child could experience multiple transitions and grow up in a nonintact family whereas a younger child could have a fairly stable family life.

Family theories and complex fertility

To date, research on MPF has been largely atheoretical, focusing on descriptive measures of prevalence and correlates, with more recent work delving into the consequences. In part, this stems from the relatively recent recognition of MPF outside of the stepfamily literature. The bigger reason, though, is that there are multiple underlying mechanisms—changes in union formation, high rates of nonmarital and unintended childbearing, the meaning(s) parents ascribe to shared children within relationships, increasing difficulty in establishing stable relationships, and rising inequality and socioeconomic disadvantage. We might do well, then, to approach MPF from a multitheoretical viewpoint.

Elder’s (1985) life course perspective provides a good starting point. First, the lives of men and women cannot be separated from the historical time period in which they live, and in today’s society, young adults are taking longer to reach adulthood, as structural changes have altered the timing, sequencing, and ease of making key transitions (Settersten and Ray 2010). Accompanying structural changes have been shifts in the normative climate surrounding childbearing and union formation (Lesthaeghe 2010). Together, these produce an environment in which people are more exposed to the possibility of complex family behaviors as the social acceptability of “nontraditional” behaviors grows.

Second, the life course perspective’s emphasis on the timing and sequencing of key transitions underscores how important the entrance into parenthood is— that how people become a parent for the first time seems to set the stage for not only their future childbearing but their future union behaviors as well (Gibson-Davis 2011; Guzzo and Hayford 2011, 2012b). Third, the notion of linked lives can help us to understand not only why MPF happens in the first place but the consequences of MPF as well. Parenthood forever links mothers and fathers, and parents and children, together; even the lack of a coparental relationship or a parent-child relationship is meaningful. For instance, parents seem to respond to their coparents’ level of parental involvement (Kotila and Dush 2012) or new partnerships (Guzzo 2009) by altering their own parental, union, and fertility behaviors.

Within the life-course framework, though, we also need to incorporate elements of theories that are specific to fertility and union behaviors. The prototype/willingness model has emerged as an explanation for unintended and nonmarital childbearing (Barber 2011). Here, even though individuals may not explicitly plan for certain outcomes (like a nonmarital birth) and may subscribe to mainstream ideals (like childbearing within marriage), men and women can find themselves in a situation in which the opportunity to engage in “risky” behavior (like unprotected sex) arises. Some individuals are more or less likely to engage in the behavior without giving thought to the possible consequences or outcomes, and willingness may be more prevalent among those whose lives are otherwise less predictable, ordered, and stable and those for whom the likelihood of following a traditional normative pathway is low.

Another useful explanation comes from Cherlin (2009), who argues that as marriage has become more about individual satisfaction, it has become a “capstone” event, occurring after other markers of adulthood have been achieved. For the less-advantaged segments of our population, this perspective places marriage out of reach, yet childbearing continues to occur while people search for a stable, loving relationship and establish secure employment and economic standing (Edin and Kefalas 2005). Childbearing also seems to confer meaning and status for those unlikely to achieve more tangible markers of success, like higher education and a fulfilling, stable job. And even among those who do marry, viewing marriage as a status almost entirely dependent on personal fulfillment legitimizes marital dissolution and repartnering. Yet within relationships, couples still want a way to symbolize, to themselves and others, that they have a joint and permanent partnership, and children are a common mechanism for doing so, even in higher-order unions and among those who already have children (Holland and Thomson 2011).

Where Do We Go from Here?

Family complexity is unlikely to disappear; if current trends are any indication, we can expect more complexity in the future, concentrated among the least advantaged segments of our population. The increasing proportion of births outside of marriage shows no signs of reversal, and the proportion of births that are unintended has remained stubbornly high over the past decade (Finer and Zolna 2011). Cohabiting unions are increasingly unstable, and there is even some evidence that divorce rates are increasing (Kennedy and Ruggles 2013). Analyses here and elsewhere (Guzzo and Furstenberg 2007a) suggest that MPF is likely to rise over time. The consequences of family complexity, in turn, will also be larger for disadvantaged groups, given lower rates of paternity establishment, child support, and nonresidential father involvement.

The challenges of family complexity are twofold. First, how can we support those who are already experiencing MPF? Existing policies for child support, custody, and visitation, as well as our various social institutions (such as schools and health care systems for which parents must define and establish relationships with and between children), ascribe to an increasingly outdated family model. Revising policies and guidelines is a daunting task, necessitated by the sheer volume of nontraditional families yet complicated by the myriad forms today’s families take. Second, how can we minimize the growth in family complexity? Although many point to cultural and normative changes weakening individuals’ commitments to unions and children, we cannot ignore the role of structural factors, such as labor market restructuring, rising costs of higher education and home ownership, and barriers to family planning services, that make it difficult to achieve a stable life and, by extension, stable families. Unfortunately, there are no simple solutions, and we need a mix of policies and investments to address the structural and normative changes that both lead to, and affect the consequences of, complex families.

Biography

Karen Benjamin Guzzo is an assistant professor of sociology at Bowling Green State University. She is a family demographer whose research focuses primarily on unintended, nonmarital, and multiple-partner fertility, as well as on family formation and stability in the United States.