Can Crossnational Theories of Nonmarital Childbearing Explain Nonmarital Childbearing Patterns Across Observed Across U.S. States?
Working Paper/PCS, November 10, 2009
Celeste Benson
Can Crossnational Theories of Nonmarital Childbearing Explain Nonmarital Childbearing Patterns Across Observed Across U.S. States?
Perhaps the biggest trend in fertility behavior in Western industrialized countries[1] over the last several decades is the rise in non-marital childbearing. Up until the mid-1960s, marriage and childbearing were closely linked across these countries. Marriage occurred in early adulthood, and nonmarital childbearing was rare. But by the close of the twentieth century this was increasingly not the case. Today, in a handful of Western European countries, a child is more likely to be born to unmarried than to married parents (Ventura 2009). In the United States, this proportion is slightly less, with one in every 2.5 children born outside of marriage[2] (Hamilton et al. 2009).
While in Europe unmarried parenthood has become progressively accepted as an alternative family form, in the United States nonmarital childbearing has sparked great controversy and attention has overwhelmingly centered on discouraging its growth, particularly through the promotion of marriage for unmarried parents (Kiernan 2004). Behind this emphasis is the argument that nonmarital childbearing is associated with higher family instability and long-term disadvantages for children than marital childbearing—due to a greater likelihood of relationship dissolution for cohabiting parents as well as a lower likelihood of single non-cohabiting birth mothers establishing long term stable unions (Heuveline 2003; Bumpass and Lu 2000; Graefe and Lichter 2002; Bennett et al. 1995; Bennett et al. 1989). However, with respect to cohabitation, studies have also shown that its meaning for childbearing is not fixed and that the stability of cohabiting unions varies notably across countries and social contexts (Heuveline 2003, 2004; Thomson 2005; Sigle Rushton and McLanahan 2002).
Crossnational researchers point out that across Western countries, cohabitation has taken on different meanings, especially with respect to family formation—reflecting an alternative or equivalent to marriage and an acceptable setting for childbearing in some countries, while acting largely as a childless prelude to marriage in others (Heuveline et al. 2004; Thomson 2005; Jensen and Clausen 2003). Such research indicates that the United States does not follow an identifiable pattern and shows little signs of convergence with other Western countries in any cohabitation and childbearing trend (Raley 2001; Heuveline et al. 2004). Using national longitudinal survey data from the late 1980s and mid-1990s, Raley (2001) concluded that the U.S. did not appear to be following the same path as European countries with respect to its patterns of childbearing within cohabitation due to the high level of instability for cohabiting unions in the U.S.[3] as well as persistently high fertility rates for single mothers who did not cohabit following a nonmarital conception (Raley 2001). This resonates with other U.S. research by Lichter et al. (2006) which tracked the union status of cohabiting couples over the period of 1979 to 2000 using national survey data and found that half of all cohabiting unions in the United States ended within one year and over 90% ended by the fifth year—the large majority ending is dissolution rather than marriage. In a crossnational study using U.S. data from the early 1990s, Heuveline and Timberlake made a similar judgment to Raley (2001), basing their reasoning on the comparatively short median duration of cohabitation in the U.S., as well as the high likelihood of cohabiting union dissolution in the U.S. relative to other developed countries (Heuveline and Timberlake 2004). Based on their findings they suggested that the best characterization of cohabitation with respect to family formation in the United States is “an alternative to being single” a term which captures the low level of commitment and stability it implies on average for U.S. couples (Heuveline and Timberlake 2004). This conclusion is consistent with the findings of a separate study by Heuveline, Timberlake and Furstenburg (2003) found that 76% of all children born to cohabiting parents in the United States experienced parental separation by the age of 15.
Such studies suggest that due to the strong likelihood of separation for cohabiting parents in the United States (Heuveline et al 2003), children born to cohabiting unions may differ little from those born to single mothers. This interpretation is largely consistent with traditional economic models of marriage[4], which understand all childbearing outside of marriage as equivalent, whether a child is born to a single mother or to cohabiting parents (Becker 1973, 1974, 1981, 1991). However, research also indicates that in the United States, important diversity in family and childbearing patterns exists across socioeconomic groups, across U.S. states and regions, between racial/ethnic groups, and with respect to individual cohabiting couples (Manning 2001; Musick et al. 2007; Lesthaeghe and Neidert 2006; Goldin and Katz 2002; Sigle-Rushton and McLanahan 2002). Survey data reveals that over the last several decades in the United States, social acceptance of cohabitation has increased significantly, that the proportion of women who have ever cohabitated continues to rise, and that the duration of cohabitation has been increasing as well (Thornton and Young-DeMarco 2001; Bumpass and Lu 2000; Kennedy and Bumpass 2008). Evidence also suggests that cohabiting couples in the United States have become somewhat more likely to plan or intend pregnancy and childbearing outside of marriage over time, and also indicates that single and cohabating mothers differ significantly in their childbearing intentions (Manning 2001; Musick 2007; Finer and Henshaw 2006). For example, in 2001, Finer and Henshaw found that while rates of unintended pregnancy for cohabiting women were quite high, rates of intended childbearing for cohabiting women were much more comparable to married than to single women (Finer and Henshaw 2006). In addition, using mid-1990s data, Musick (2002) found that planned childbearing was significantly more likely for cohabiting than for single women, at least for non-Hispanic whites and Hispanics. Longitudinal data also shows that as childbearing within cohabitation has become more common, the strength of its association with low socioeconomic status has weakened as well. While it is still quite rare among women with 4-year college degrees, it has grown noticeably for working class women with some college education in recent decades (Sassler and Cunningham 2008; Sassler et al. 2009). Between the early 1980s and 2001, the proportion of all births among women who reported themselves as having “some college”, and also cohabiting rose from 3 to 15 percent (Kennedy and Lu 2000; Kennedy and Bumpass 2008). These findings imply that cohabitation may increasingly serve as a stable alternative to marriage for some couples.
While these trends point to increasing institutionalization of cohabitation as an alternative family form in the U.S., important confounds remain. Despite the rising share of unmarried births to cohabiting mothers, births to single noncohabiting mothers in the U.S. remain strikingly high. While having a child outside any union is uncommon across most of continental and northern Europe[5] (Kiernan 1999; 2002; Andersson 2002), in the United States half of all nonmarital births are to non-cohabiting mothers (Kennedy and Bumpass 2008)[6]. In addition, by international standards, Americans also have an unusually high incidence of births resulting from unintended pregnancy (Jones et al. 1989; Ranjit et al 2001; Frejka and Kinkade 2003; Quesnel-Vallée and Morgan 2003; Musick et al. 2007) and data indicates that in recent years in the U.S., single mothers have increasingly cohabited as a response to pregnancy (Raley 2001; Kefalas et al 2005; Reed 2006)[7]. The stronger prevalence of unintended pregnancy and childbearing in the U.S. might indicate that relative to Europe, cohabitation in the U.S. may be host to a less stable element due to these couples’ lower intent and preparation for parenthood. In addition, the United States also has significantly higher poverty levels than other developed countries (OECD 2008), another factor associated with lower union formation and higher relationship instability, particularly when measured in terms of male wages (Goldscheider and Waite 1986; Teachman et al. 1987; Oppenheimer et al. 1997; Smock and Manning 1997; Sassler and Schoen 1999; Oppenheimer 2000; Sweeney 2002; Xie et al. 2003 Clarkberg 1999; Smock and Manning 1997). Other findings suggest that measures of equality rather than individual wages may play an important role in shaping nonmarital childbearing patterns. Thomson for example found that across developed countries, measures of gender equity were positively associated with higher childbearing within cohabitation, but had no relationship to single motherhood. However, she found that levels of single motherhood were consistently associated with levels of economic inequality across countries once a threshold of nonmarital childbearing was reached. While economic or social inequality are rarely explored as a predictor of unmarried childbearing patterns in U.S. research, as Blossfeld and Timm point out, because inequality uniquely shapes marriage markets, and because marriage in return shapes inequality, the two should be understood as closely interconnected (Blossfeld and Timm 2003). Compared with other developed countries, the implications of inequality may be particularly important for the United States, as levels of inequality in the United States are unusually high (OECD 2008).
Theories of Nonmarital Childbearing in the United States
In the United States, non-marital childbearing and unmarried parent families have received significant scholarly attention. However, despite this attention, researchers admit that the reasons for their rise and rapid growth in recent decades remain little understood (Jencks and Ellwood 2002). Since the 1970s, explanations have been largely dominated by traditional economic models of marriage, derived from Becker’s theory of marriage (Becker 1973, 1974, 1981). Becker’s theory of marriage argues that marriage is a matter of individual utility maximization whereby individuals compete for partners in what Becker refers to as a “marriage market” and marry only when the expected gains from marriage exceed the expected costs of not marrying. Because the process of marriage is essentially a process of exchange, Becker posits that a higher inequality in role specialization for the sexes predicts a higher likelihood of marriage, as it ensures greater “gains from trade” for each partner. This conclusion is also built on the assumption that the chief purpose of marriage is the production of biological children, which are viewed as the primary form of marriage specific capital. Due to these two factors, Becker claims that traditional marriage, with its reliance on a sharp role division between husband and wife, whereby a woman specializes in reproduction and household labor in exchange for a man’s wage earnings in the labor market, represents an optimal market exchange, and thus the most stable formula for marriage. Any imbalance to this equilibrium in the form of increased wages or contributions from either partner that is not able to be offset by the other partner decreases the likelihood of marriage formation and has the potential to induce marital instability for married couples (Becker 1973, 1974, 1977). While Becker recognized nonmarital childbearing as lying outside the bounds of his formal marriage model, in laying out his theory of marital instability, he suggested that as women’s labor market earnings rose and more women entered the labor market, prospects for marital instability would rise not only due to women’s individual earnings offsetting their gains to marriage, but also as a byproduct of changes in the marriage market caused by the increased availability of new partners due to increased divorce (Becker 1977). The overall impact of this shifting context, Becker suggested, could be a rise in nonmarital births, particularly for economically advantaged women who would be more able to opt out of marriage under low returns and unstable conditions (Becker 1977).
Becker’s marriage model understood marriage and childbearing as indistinguishable, making no allowances for nonmarital births, prompting Becker to note its non-suitability for explaining nonmarital births (Becker 1977). However, despite the fact that Becker did not see his theory as appropriate for explaining nonmarital childbearing, economists have subsequently extended and employed it in various ways to do so, spawning an extensive research literature. The majority of this attention has centered on two areas, wages and welfare. Early applications of these models posited that nonmarital childbearing was a byproduct of women’s increased labor market position resulting in women’s declining gains to marriage; while later extensions explained it as either a rational response to government welfare (see Moffitt 1998 and Rosenzweig 1995) or a rationale response to low male wages and constrained marriage markets (Willis 1996, 1999). As Ellwood and Jencks (2004) note, while traditional economic models have received some support in the area of welfare[8] and especially male wages and marriage markets (see Harknett and McLanahan 2004 for review), overall they have performed rather poorly in explaining nonmarital childbearing patterns in the United States. This has prompted researchers in the United States to turn to non-economic explanations such as the role of social norms and values, gender role conflict, personal and interpersonal efficacy and the importance of effective contraception and abortion (Ellwood and Jencks 2004). Today, while these areas stand at the forefront of research on nonmarital childbearing, the influence of each is not well understood.
The Question of Single and Cohabiting Births in the United States: How Much Difference?
As Wu points out, nonmarital fertility in the United States and Europe differs in several important respects, the most important of which is arguably the high prevalence of births to single non-cohabiting mothers in the United States (Wu 2001). In the United States, unmarried childbearing has been largely framed as equivalent to single mother childbearing, while in Europe unmarried childbearing, as it is more likely to occur within cohabitation, is considered broadly equivalent to marriage (Kiernan 2004). The American way of understanding unmarried childbearing is also consistent with traditional economic models of marriage, which consider all nonmarital births as equivalent, and has focused U.S. research on the unmarried childbearing patterns of teen and non-white mothers, which have both predominantly been to single rather than cohabiting mothers (Wu 2001). However, the rate of childbearing to teens has declined significantly in the United States since the early 1990s[9] while adult women have experienced a strong general trend towards higher nonmarital childbearing overall (Ventura and Bachrach 2000; Finer and Henshaw 2006; Martin et al. 2009; Kennedy and Bumpass 2008). Moreover, this trend has been strongly concentrated within cohabitation and the trend towards childbearing within cohabitation has occurred not just for whites but across racial groups (Kennedy and Bumpass 2008). Data suggests that today in the United States, single and cohabiting mothers differ little with respect to one another in terms of socioeconomic background, but are both much more likely to be come from lower socioeconomic backgrounds than married mothers (Kennedy and Bumpass 2008). The primary difference between single and cohabiting mothers remains age and race, with single mothers being somewhat more likely to be adolescents as well as black[10], while cohabiting mothers are more likely to be older and non-Hispanic white or Hispanic (Kennedy and Bumpass 2008). In 2001, 59% of women under the age of 20 gave birth as single mothers compared with 44-46% of women over the age of 20; in addition, 60% of black women gave birth as single mothers compared with 45% for non-Hispanic whites and 38% for Hispanics (Kennedy and Bumpass 2008).