August 9, 2008
Mate Availability and Unmarried Parent Relationships*
By Harknett, Kristen
Theoretically, a shortage of males in a local marriage market may influence the formation, quality, and trajectory of unmarried parent relationships. To test these hypotheses, I combine city-level sex ratio data from the U.S. census with microdata on unmarried couples who recently had a child from the Fragile Families and Child Wellbeing study. A shortage of men in a marriage market is associated with lower relationship quality for unmarried parents. Male shortages are associated with lower rates of marriage following a nonmarital birth, in part because of the mediating influence of relationship quality. A shortage of men is not significantly related to the economic quality of male, nonmarital childbearing partners. In many U.S. cities, single women in their 20s and 30s outnumber men. These sex ratio imbalances come about because of differential migration by sex and higher rates of mortality for males than females (Case and Paxson 2005). Sex ratio imbalances have been exacerbated in recent decades by an exponential increase in incarceration, which has disproportionately affected young and African American men (Pettit and Western 2004). Sex ratios for different racial/ethnic groups vary widely across localities, offering the opportunity to examine the influence of sex ratio imbalances on social demographic outcomes.Research has shown that a shortage of men relative to women in a local marriage market is associated with lower rates of ever marrying as well as higher rates of nonmarital childbearing and divorce (Fossett and Kiecolt 1993; Lichter et al. 1992; Lichter, LeClere, and McLaughlin 1991; South and Lloyd 1992; South, Trent, and Shen 2001). This article considers some potential consequences of male shortages that have not been examined previously: do women have children with less economically desirable men in marriage markets where men are in short supply? Do unmarried parents have lower quality relationships in contexts where men are in short supply? Connecting these analyses with prior research on ever marrying, I also consider whether male shortages inhibit marriage following the birth of a child because they are associated with lower quality childbearing partners (in economic terms) and lower quality relationships.
If marriage markets affect the economic quality of childbearing partners, the quality of parent relationships, and marital transitions after a birth, these processes have implications for child well-being. In recent years, over one-third of children were born outside of marriage in the United States (Sutton and Matthews 2004). If women have children with economically undesirable men in places where men are in short supply, children are more likely to grow up in economically insecure households. If parent relationships are low quality when men are in short supply, fathers in these contexts may be less involved with their children (Erel and Burman 1995), and child well-being may be compromised (Howes and Markman 1989). If male shortages inhibit transitions to marriage, children born in these contexts will be less likely to be raised by two parents. Therefore, in marriage markets that are unfavorable for women, children born to unmarried parents may be particularly prone to the disadvantages associated with single-parent families and father disengagement (Amato and Gilbreth 1999; McLanahan and Sandefur 1994).
To examine the influence of marriage markets on partner economic quality, relationship quality, and transitions to marriage, I combine marriage market data from the U.S. census with data from the Fragile Families and Child Wellbeing study on unmarried parents. The study includes 20 cities that vary widely in their marriage market characteristics, offering the opportunity to test theoretical predictions about how marriage markets influence mate selection and union formation processes.
THEORY AND PRIOR RESEARCH
Individuals search for romantic partners in particular marriage market contexts that may be relatively more or less favorable. Research typically assumes that marriage markets are segmented by racial/ethnic group, age group, and geographic area, and that individuals search for partners within these boundaries. Consequently, some men and women will face favorable marriage markets in which they have a relatively large number of partners from which to choose and with relatively little competition for these partners, but others will face unfavorable markets in which partners are in short supply and competition for these partners is stiff.
Marriage markets and partner quality. Many authors have drawn parallels between job search and marital search processes (Becker 1981; England and Farkas 1986; Hutchens 1979; Oppenheimer 1988). For example, the strength of the labor market influences the job that one can get and one's minimum standards for a job. Along these lines, marriage markets can be predicted to influence the quality of the partner that one can attract and the minimum quality partner that one will accept. Similar to the search for a job in a particular labor market, the search for a romantic partner involves gathering information about the distribution of opportunities and then choosing the best available opportunity, given one's own qualifications and attractiveness. In an unfavorable marriage market, an individual may be unable to attract a high quality partner and thus may lower his or her standards required for a partner. The process of lowering standards when faced with an unfavorable marriage market is similar to lowering one's reservation wage in an unfavorable labor market. In theory, women can be expected to accept less economically desirable childbearing partners when faced with a shortage of potential partners and to find more economically desirable partners when potential partners are in relative abundance. In any marriage market context, one's own characteristics also affect the quality of partner that one can attract.
In the case of a job, salary, hours, and working conditions all contribute to quality. Although the quality of a childbearing partner is more subjective, prior research gives some indication of the attributes that are in demand on the marriage market. Research has consistently shown that men's earnings and earning potential are positively related to marriage and marital stability (Lichter et al. 1991; Sweeney 2002; Xie et al. 2003). All else equal, men who are employed and have more earning power and education are expected to be more desirable partners. These are by no means the only measures of childbearing partner quality, but these attributes can be used to test theoretical predictions about marriage markets and partner quality.
The formation of a relationship requires the cooperation of both parties. Notably, taking the male or female perspective yields the same predictions about marriage market influences on partner quality. In an unfavorable marriage market, women may be more willing to have a child with a man who is unemployed or earns a low wage, and they may have a harder time attracting an economically desirable man because of the stiff competition for this type of partner. Meanwhile, in this context, men may perceive that many opportunities for romantic partners are available to them; consequently, they may be less willing to commit to fatherhood or marriage.
To summarize, theory predicts that a shortage of males may lead to partnerships between women and economically unattractive male partners. Alternatively, the local supply of partners may have no effect on the economic quality of male childbearing partners. Women may prolong their search and postpone childbearing until they find a partner who meets their standards (Oppenheimer 1988). Some women may expand their search by going outside their local marriage market to find a suitable partner. Other women may forgo childbearing if they never find an adequate partner (Heaton, Jacobson, and Holland 1999). Therefore, an alternative hypothesis is the null hypothesis that marriage market conditions will not influence the economic desirability of male childbearing partners.
Marriage markets and relationship quality. In addition to influencing the economic quality of male partners, marriage markets may influence the quality of couple relationships. Theoretically, the sex in short supply will have more bargaining power in their relationships because alternative romantic partners are relatively easy to come by (Guttentag and Secord 1983). Household bargaining models suggest that the pool of alternative partners influences relationship quality. If one partner perceives that his or her utility would be higher outside the relationship given other available partners, this can lead to a noncooperative equilibrium within the relationship, reflected by lower relationship quality (Lundberg and Pollak 1996). England and Folbre (2002) argued that children tip the bargaining scales in men's favor. Women's greater attachment to children and their custodial responsibility for children decreases women's utility outside of marriage and makes women's investments in their relationship relatively insensitive to marriage market conditions. Qualitative research supports the asymmetry in bargaining power in relationships when children are involved. Anderson's (1989) ethnographic account of young, inner- city adults described women's aspirations for a middle class, nuclear family juxtaposed against men's desires for casual sex and to avoid responsibility for children. Therefore, one can predict that relationship quality will be better in marriage markets that favor women. Relationship quality is multifaceted and has been measured in many ways in prior research (Norton 1983). Recent research using the Fragile Families and Child Wellbeing study derives a supportiveness scale based on the perceptions that a partner is fair, loving, supportive, and not often critical (Carlson and McLanahan 2006; Carlson, McLanahan, and England 2004). This same research has operationalized low relationship quality as frequent conflicts in common areas of disagreement for couples: time, money, sex, substance abuse, pregnancy, and faithfulness. In this article, a father's presence at the hospital during or after baby's birth is also taken as an indicator of some amount of commitment to the mother and child. A final, indirect indicator of relationship quality is the father's multipartnered fertility. Prior quantitative research suggests that a father's children from prior relationships detract from the time and resources that he can devote to his new family (Cooksey and Craig 1998; Manning and Smock 1999, 2000). Prior qualitative research suggests that a father's having children from prior relationships creates tension in his relationship with the mother of his most recent child (Monte 2007). Marriage markets that favor women are expected to be associated with better relationship quality on each of these measures.
Marriage markets and transitions to marriage. Marriage markets in which men are in short supply are theoretically expected to inhibit marriage. The literature on male shortages and rates of marriage generally focuses on two explanations. First, a numeric shortage of men will constrain women's ability to marry because there are too few men to go around (Wilson 1987). This mechanism is not relevant for the current article because all women in the sample have found men with whom to have children. Instead, this article focuses on the quality and trajectory of the parent relationships. Second, shortages of men are expected to affect marriage by influencing bargaining power and commitment in relationships ( Guttentag and Secord 1983; Lundberg and Pollak 1996). Men are expected to be less committed to their female partners and less likely to marry when they perceive that available alternative dating opportunities are plentiful (Willis 1999; Wilson 1996). I examine this mechanism later in the analysis of marriage markets and relationship quality. A third possibility that has received less attention in the literature is that women are more likely to form relationships with low quality men in unfavorable marriage markets. In unfavorable marriage markets, women may have children with men whom they do not deem marriageable (Edin and Kefalas 2005) and with men who are unwilling to commit to marriage (Anderson 1989). I examine this third mechanism later in the analyses of the economic quality of male partners and of transitions to marriage.
Prior Empirical Research
An extensive body of research has examined the relationship between marriage markets and marriage rates (Fossett and Kiecolt 1993; Kiecolt and Fossett 1995; Lichter et al. 1992; Lichter et al. 1991; South 1996; South and Lloyd 1992). In general, this research suggests that shortages of males are associated with lower rates of ever marrying. This research has typically measured marriage markets using crude sex ratios defined by age, race/ethnicity, and locality. Although these assumed marriage market boundaries are an oversimplification, prior research has demonstrated that crude sex ratios are more strongly predictive of social demographic outcomes than more refined marriage market measures, such as availability ratios (Fossett and Kiecolt 1991; Goldman, Westoff, and Hammerslough 1984; Lampard 1993). Therefore, crude sex ratio measures of marriage markets continue to prevail in marriage market research.
Most prior research on marriage markets and marriage rates has focused on rates of ever marrying in the aggregate or on individual- level transitions to marriage without taking children into account. In contrast, this article focuses in particular on parental relationships and transitions to marriage following a birth. One prior study found evidence that un favorable marriage markets inhibited transitions to marriage following a nonmarital birth (Harknett and McLanahan 2004). Although this prior study was primarily focused on explaining racial and ethnic differences in marriage, the authors presented some evidence that marriage markets are correlated with partner and relationship quality. The current article builds upon this earlier work by (1) analyzing these correlations in a multivariate context, controlling for a range of mother characteristics; (2) using an appropriate, multilevel modeling approach; and (3) examining the mediating role of partner and relationship quality in the relationship between marriage markets and marriage transitions.
Research linking marriage market characteristics to the quality of male partners or the quality of relationships is scarce. The existing research focuses on marital relationships that may or may not include children. Trent and South (2003) combined data from the National Survey of Families and Households with census data and found no relationship between marriage market conditions and reported marital happiness. Lichter, Anderson, and Hayward (1995) combined data from the National Longitudinal Survey of Youth with census data and found mixed evidence on the relationship between marriage markets and the quality of marital partners. In unfavorable marriage markets, women were more likely to marry men with relatively low economic status. However, the strength of the marriage market was not related to educational assortative mating; unfavorable marriage markets did not lead more women to marry down in terms of education.
By focusing on unmarried parents who recently had a child, this article provides evidence on how marriage markets affect the quality of children's fathers. A recent book by Edin and Kefalas (2005) analyzed qualitative interviews with 162 unmarried mothers, shedding light on women's marriage and childbearing decisions in an unfavorable marriage market. These authors found that low-income women place a high value on children and are unwilling to forgo childbearing if they do not find a marriageable man. Quantitative research has also suggested that low-income women are the most likely to have children outside of marriage and that voluntary childlessness is far more common among women with higher incomes (Abma and Martinez 2006; Ellwood and Jencks 2004). Qualitative evidence from the male perspective has suggested that men are reluctant to commit to a relationship when alternative partners are relatively numerous (Wilson 1996). Together, this research has suggested that women-at least, low-income women-may settle for low quality childbearing partners when faced with an unfavorable marriage market and that male partners in these contexts may be reluctant to marry. This article focuses on unmarried parent couples in urban areas, a population that tends to be relatively economically disadvantaged.
Based on prior research, this article tests the following set of hypotheses. First, unfavorable marriage markets for women will be associated with less economically desirable male, nonmarital childbearing partners. Second, unfavorable marriage markets for women will be associated with lower relationship quality among unmarried parents. Third, in unfavorable marriage markets, unmarried parents will be less likely to marry following a nonmarital birth in part because male partners are less economically desirable and relationships are lower quality.
DATA AND METHODS
This article uses recent data from the longitudinal Fragile Families and Child Wellbeing study (hereafter referred to as the Fragile Families study). The study oversampled unmarried parents, and the analysis in this article is based exclusively on the unmarried sample. I use data from baseline interviews with unmarried mothers and fathers who had just had a baby and from follow-up interviews with mothers both 1 year and 30 months after the birth. The baseline interviews were conducted between 1998 and 2000 in 20 large U.S. cities. Response rates to the baseline survey were 87% for unmarried mothers and 75% for unmarried fathers. Of those unmarried mothers included in the baseline survey, 90% responded to the 1-year follow-up survey, and 88% responded to the 30-month follow-up survey. Cities were randomly selected for inclusion in the study from among U.S. cities with populations of over 200,000 and were stratified by policy and labor market context. The Fragile Families unmarried sample is representative of nonmarital births in large U.S. cities. For more details on the design of the Fragile Families study, see Reichman et al. (2001).
One city in the study, Norfolk, VA, is an outlier on the main independent variable-the crude sex ratio-because a mostly male military base is located in that city. Therefore, couples in Norfolk (n = 74) are excluded from the analysis sample. Parents who were not African American, Hispanic, white, or Asian are excluded from the sample because of small cell sizes or because sex ratio data could not be determined for other ethnicities (n = 311). Cases missing data on any dependent variable are excluded (n = 944). After these sample restrictions, the analysis is based on 2,382 unmarried couples.
Measures of partner quality. The fathers' economic quality is measured at baseline by variables indicating that the father was employed in the week before the child's birth (0/1), that the father completed some college (0/1), and the father's predicted hourly wage. The father's predicted hourly wage is based on fathers' self- reports of their earnings at their current job or the last time they worked. About one-third of fathers were missing wage data because they were nonrespondents to the baseline survey, did not volunteer wage information, or had never worked. For all fathers, predicted hourly wages were estimated based on a regression of fathers' reported hourly wages (for those with nonmissing wage data) on fathers' characteristics, such as age, education, race/ethnicity, and city of residence. In this analysis, predicted wages are used for all fathers, even those who had nonmissing wage data. The predicted wage measure should be interpreted as an estimate of the father's earning potential rather than as a measure of his actual wages. Father quality is also measured relative to the mother's characteristics at baseline. The two measures of relative quality are the ratio of father's to mother's predicted hourly wages, and whether the father has as much or more education than the mother (0/ 1). For the wage ratio measure, mothers' predicted hourly wages are estimated in the same way as for fathers (as explained earlier). The relative education variable is based on four levels of educational attainment: less than a high school diploma, high school diploma only, some college, and college degree or higher. If a father and mother fell into the same educational category or if the father had a higher level of educational attainment than the mother, the father was coded as having as much or more education as the mother. Table 1 presents descriptive information on the economic quality of unmarried fathers. Most fathers are employed and have as much or more education than their female partners. Only about one-fourth of fathers have some college education. The average father in the sample could earn about $11 per hour if employed, and father's predicted wages exceeded mother's by about one-third.
Measures of relationship quality. Following Carlson et al. (2004), relationship quality is measured using a scale that combines mother and father reports of whether the other parent is fair, loving, helpful, and critical based on the following questions:
Thinking about your relationship with [BABY'S FATHER/MOTHER], how often would you say that:
S/he is fair and willing to compromise when you have a disagreement?
S/he expresses affection or love for you?
S/he encourages or helps you to do things that are important to you?
S/he insults or criticizes you or your ideas?
Parents answered these questions on scale of 1 to 3 representing "never,""sometimes," or "often," respectively. The scale for each parent represents the average response to these four questions, with responses to the insults or criticizes question reverse-coded. Therefore, higher values on the scale represent higher relationship quality. In this article, relationship quality is treated as a characteristic of the relationship rather than as a characteristic of an individual mother or father; therefore, mother's and father's reports of relationship supportiveness are averaged together. When fathers did not respond to surveys, mothers' reports of relationship supportiveness are used in place of their average report. The scale has a Cronbach's alpha reliability of .65. Although mother and father reports of relationship quality do not always agree, analyzing the mother's and father's reports of relationship quality separately yields results consistent with the average reports I present.
I include an additional scale measuring conflict in relationships. This scale sums the number of areas in which the parents have frequent disagreements among the following areas: time, money, faithfulness, the pregnancy, sex, and drugs. Mother and father reports are averaged together. The alpha reliability for this scale is .77.
As shown in Table 1, reported relationship supportiveness is high (2.6 out of 3 on the scale), and relationship conflict is low (0.7 out of 6 on the scale). These measures of relationship quality were collected soon after the baby's birth, when parents may have been unusually positive about their relationships. The interest in this article is not in the level of relationship quality but rather in the comparison of relationship quality across marriage markets. The unique time of data collection may have minimized variation in reports of relationship quality. If so, this analysis may underestimate the difference in relationship quality across marriage markets.
A third measure of relationship quality is a dummy variable indicating whether the father visited the mother in the hospital during or after the birth of their child. The Fragile Families sample was drawn from births in hospitals, so all mothers gave birth in a hospital. Table 1 shows that 83% of fathers visited the mother and baby in the hospital.
A final measure of relationship quality is a variable indicating that the father does not have a child with a previous partner (0/ 1). The measure of the father's children from previous partners is based on data collected in the 1-year follow-up survey. For fathers who were nonrespondents to the survey, mothers' reports of father characteristics were used. Table 1 shows that 58% of fathers did not have any children with previous partners, meaning that a large minority of fathers (42%) did have children from prior relationships.
Transitions to marriage. Marriage is measured in the mother's 30- month follow-up survey and is defined as marriage to the baby's father. Only 2% of mothers were married to someone other than the father at the 30-month follow-up, and they are coded as not married to baby's father. Table 1 shows that 15% of unmarried parents were married by the 30-month follow-up interview.
Measuring marriage markets. I use crude sex ratios to measure the supply of males relative to females in a local marriage market. Consistent with prior research, I focus on heterosexual relationships and make the simplifying assumption that marriage markets are segmented by age, city, and race/ethnicity (Fossett and Kiecolt 1991). The presence of gays and lesbians in the localities I study is a source of imprecision in my measure of the heterosexual marriage market. Crude sex ratios are defined as the ratio of 18- to 34-year-old men to women by city and racial/ethnic group. In this sample, 92% of mothers and fathers fell in this age range at the time of their baby's birth. The analysis includes 19 cities and 4 racial/ethnic groups, yielding 76 (19 x 4) possible marriage markets. Because some cities did not have any Hispanic or Asian parents in this sample, the analysis includes 64 marriage markets after these empty cells are excluded.
Crude sex ratio data come from the 2000 U.S. census. As shown in Table 1, sex ratios vary widely across cities in the analysis sample: from 0.73 (Milwaukee, WI) to 1.12 (San Jose, CA) for African Americans; from 0.96 (Corpus Christi, TX) to 1.65 (Indianapolis, IN) for Hispanics; from 0.92 (Oakland, CA) to 1.14 (Newark, NJ) for whites; and from 0.86 (Richmond, VA) to 1.25 (Austin, TX) for Asians.
All marriage market research in the United States that relies on census estimates of mate availability faces the problem of the differential undercount. African American men are more likely to be undercounted in the census than men in other racial/ethnic groups or than African American women (Robinson et al. 1993). Some authors argue that those not counted in census enumeration are likely to be excluded from the marriage market because they belong to marginalized subpopulations, such as the homeless (South and Lloyd 1992; Spanier and Glick 1980). The census remains the best source of data on mate availability, but potential bias from the differential undercount should be kept in mind when interpreting results.
For mixed race/ethnicity couples (11% of the sample), the marriage market definition varies for the mother and father. The assumption that marriage markets are segmented by race/ethnicity does not apply to these couples. For the purposes of the analysis, I make two assumptions for mixed race/ethnicity couples. First, the supply of partners from the woman's perspective will be related to the economic quality of her childbearing partner. If a woman has many alternative partners, she will choose and be able to attract a male partner who is employed, has relatively more earning power, and has relatively more education. Second, the supply of partners from the man's perspective will be related to relationship quality and to marriage transitions following a nonmarital birth. If a man has many alternative partners, he will be less committed in his relationship with the mother, relationship quality will suffer, and the couple will be less likely to marry. The sex ratio variable for mixed race/ ethnicity couples is defined accordingly in the upcoming analyses. In particular, marriage markets are defined according to mother's race/ethnicity in analyses of father's economic characteristics and according to father's race/ethnicity in analyses of relationship quality and transitions to marriage. The choice of mother's or father's racial/ethnic group in defining the marriage market for mixed race/ethnicity couples does not alter the pattern of results.
In this analysis, most incarcerated men are excluded, de facto, from crude sex ratios because the sample is exclusively urban and because prisons are almost exclusively located in nonurban areas. Explicitly excluding incarcerated men and women has very little effect on the sex ratio for this sample. In separate analyses, I tested several alternative measures of the marriage market. Substituting the nonincarcerated male/female ratio, the employed male/female ratio, or the unmarried male/female ratio for the crude sex ratio yielded the same pattern of results presented. Using different age cutpoints (20-34, 20-29, or 20-39 years) for the sex ratio definition also yielded the same pattern of results. Control Variables
In most models, I control for the following characteristics of mothers measured at baseline: age in years, race/ethnicity (African American; Hispanic; with white, non-Hispanic and Asian, non- Hispanic as the reference cell),1 immigrant (0/1), mother and father have a different race/ethnicity (0/1), predicted hourly wage ($), education (high school diploma, some college, with less than high school as the reference cell), received welfare or food stamps in the prior year (0/1), and self-reported fair or poor health (0/1). I also control for the presence of mother's children from prior partners (0/1) measured at the 1-year followup survey. In models predicting the relative quality of male partners (father's predicted wages or level of education relative to the mother's), I omit controls for mother's wages and education, which are encompassed by the relative quality dependent variables. Regression estimates were used to predict missing values on control variables, and a dummy variable is included in regression models to indicate that data were imputed. No more than 4% of cases were missing on any one variable. Table 1 presents the mean values for individuallevel control variables.
As a test of robustness, I include controls for contextual variables that may be correlated with sex ratios and outcomes: the local divorce rate, nonmarital childbearing rate, unemployment rate, and an estimate of the census undercount in the city. These data were derived from the census, the Bureau of Labor Statistics, and vital statistics. Table 1 summarizes these contextual variables for the 19 cities included in the analysis. Including these contextual control variables does not alter the pattern of results presented. The results from these robustness checks are discussed in the text but are not presented in tables.
Multilevel Modeling Approach
The analytic models that follow combine data measured at the marriage market and the individual levels; therefore, I use multilevel regression models. I use the Stata xtmixed procedure to estimate multilevel linear models for continuous outcomes, and the Stata xtmelogit procedure to estimate multilevel logistic regression models for dichotomous outcomes. I use the Stata xtmepoisson command for the relationship conflict variable because this is a left- skewed count variable. Each of these models is specified with a random intercept and fixed slopes.
An advantage of multilevel models is that they provide estimates of the level-2 variance in dependent variables: in this case, the variance across marriage markets. The null, randomintercept model shows that all of the dependent variables vary significantly across marriage markets. When the crude sex ratio is added as a predictor, the variation in relationship quality across marriage markets is no longer significant. Crude sex ratios also account for most of the variation across marriage markets in transitions to marriage after a nonmarital birth. In contrast, the father's economic characteristics continue to vary significantly across marriage markets even after controlling for crude sex ratios. The subsequent tables and results focus on the relationships between crude sex ratios and dependent variables.
Table 2 summarizes the relationship between sex ratios and five measures of the economic desirability of male, nonmarital childbearing partners. Sex ratios are expected to be positively associated with each of these measures. For each of the five measures of the economic desirability of male partners, sex ratios have the expected positive sign. However, for only one outcome measure is the relationship statistically significant.
In favorable marriage markets, women are more likely to have a child with a male partner who has at least as much education as they do, and women are less likely to partner down in terms of education. In a favorable marriage market, a woman is no more likely to have a child with a man who is employed or who has higher predicted wages, higher predicted wages relative to her own, or some college education than she is in an unfavorable marriage market. Separate analyses show that the local unemployment rate is also not related to the economic or educational characteristics of fathers.
The economic quality of a male partner has much more to do with mother's own characteristics than it does with the marriage market or local economy. Mothers' predicted hourly wage and education are positively correlated with fathers' employment, predicted wages, and education. Older and immigrant mothers also tend to have partners who are more economically desirable. In contrast, African American mothers as well as mothers who crossed racial/ethnic lines in their relationships have partners with lower predicted wages and who are less likely to be employed, and Hispanic mothers have partners with lower predicted wages and lower levels of education. Mothers' welfare receipt is also associated with less economically desirable partners. Perhaps surprisingly, mothers who had children with previous partners are more likely to partner up in terms of predicted wages and to partner with a man with as much or more education. This finding is consistent with research from the Fragile Families study showing that women who repartner in the study do so with higher quality men (Bzostek, Carlson, and McLanahan 2007).
Table 3 shows the relationship between sex ratios and four measures of relationship quality. Favorable marriage markets are associated with better quality relationships by all four measures. Marriage markets that favor women are associated with significantly more supportive relationships and less relationship conflict. Additionally, in marriage markets that favor women, fathers are more likely to demonstrate some level of commitment to the mother and baby by visiting the hospital during or after the baby's birth. Finally, in marriage markets that favor women, fathers are less likely to have children with a previous partner.
In addition to marriage markets, some mothers' characteristics are also predictive of relationship quality. Immigrant mothers have more supportive relationships and less relationship conflict than nonimmigrant mothers. Older mothers, mixed-race couples, and mothers in poor health have more relationship conflict than their counterparts. As reported in prior research, older mothers and African American mothers are more likely to have children with men who have children from previous partners (Carlson and Furstenberg 2006). Also, mothers with children from a previous relationship are more likely to partner with fathers who also have children from a previous relationship. African American mothers are less likely than other mothers to be visited in the hospital by fathers at the time of the birth.
Table 3 focuses on local sex ratios, but other features of the local context-such as the unemployment rate or the prevalence of divorce and nonmarital childbearing-may be correlated with sex ratios and the quality of unmarried parents' relationships. Depressed labor markets may lead to male shortages and to low relationship quality. However, in separate analyses (not shown), I find that controlling for the city unemployment rate does not alter the reported relationship between sex ratios and relationship quality. High rates of divorce and nonmarital childbearing may lead to a normative climate of low commitment and less investment in relationships. Controlling for the local divorce rate and the nonmarital birth rate does not alter the positive and significant relationship between sex ratios and better quality relationships. Controlling for the estimated census undercount in a city, which may bias estimated sex ratios, has no effect on the relationship between sex ratios and relationship quality.
Table 4 examines transitions to marriage after a nonmarital birth. Model 1 estimates the relationship between sex ratios and marriage after I control for mother characteristics. Higher sex ratios are positively associated with marriage following a nonmarital birth. Holding mother characteristics constant, moving from a context in which men are in large shortage (sex ratio = .80) to a context in which men are in large surplus (sex ratio = 1.20) increases the predicted probability of marriage from .12 to .20 based on the Model 1 regression estimates. Mother characteristics are related to transitions to marriage in the expected manner: African American mothers, couples of mixed race/ethnicity, and welfare recipients are less likely to marry, whereas mothers with at least a high school education and mothers who are immigrants are more likely to marry.
Model 2 shows that relationship quality and father's economic characteristics are associated with marriage. The father's predicted hourly wage is positively associated with marriage, but his employment status, education, and wages and education relative to the mother are not. Couples with more supportive relationships are more likely to marry, but the extent of relationship conflict is not related to marriage. Couples are more likely to marry if fathers visited the hospital at the time of their baby's birth and if fathers had no children with a previous partner.
Model 3 includes sex ratios, father quality, and relationship quality measures as predictors. The relationship between sex ratios and marriage is still positive and statistically significant but smaller in magnitude after I control for the quality of partners and the quality of relationships. Controlling for other characteristics of localities-divorce, nonmarital childbearing, and unemployment- and for estimates of the census undercount does not alter the reported relationship between sex ratios and marriage. Returning to the hypotheses proposed earlier in this article, the results from Tables 2-4 can be summarized as follows: (1) marriage markets are not associated with the economic desirability of male, nonmarital childbearing partners; (2) marriage markets that favor women are associated with better relationship quality among unmarried parents; (3) unfavorable marriage markets are associated with a lower probability of marriage following a nonmarital birth in part because of the lower quality of unmarried parents' relationships in these contexts.
Marriage markets have been linked in prior research to marriage, divorce, and nonmarital childbearing. Marriage markets may also influence unmarried parent relationships in a number of ways. This article considers whether local marriage markets affect the economic quality of male childbearing partners, the quality of unmarried parent relationships, and the chance that unmarried parents will marry after a birth. The results in this article are generalizable to unmarried parents in large cities.
In marriage markets that favor women, unmarried parents' relationships seem to be of better quality. As predicted by Guttentag and Secord (1983), parents' relationship quality seems to suffer when men have many alternative partners but seems to improve when women have many alternative partners. This gender asymmetry in marriage market dynamics suggests that men and women exercise their marriage market advantage very differently: men invest less in the relationship with the mother, but women strengthen the relationship with the father. Norms of maternal attachment and maternal custody may make investments in the parental relationship a higher priority for mothers than for fathers (England and Folbre 2002). These findings are also consistent with research showing that the presence of children weakens women's position in the remarriage market (Qian, Lichter, and Mellott 2005). Because my analysis measures relationship quality at one point in time, I cannot determine whether unmarried parent relationships are lower quality from their inception in the context of male shortages (a selection effect) or whether unmarried parent relationships deteriorate over time in these contexts (a household bargaining effect).
The measures of relationship quality in this article have important implications for families and children but are not comprehensive. This article measures relationship quality in terms of relationship supportiveness, relationship conflict, the father visiting the mother in the hospital, and the father having no children from previous relationships. These measures of relationship quality could be broadened to include direct measures of relationship satisfaction. Also, relationship quality is not static, and incorporating repeated measures of relationship quality in future research would be useful.
This article finds that marriage markets are not related to the economic quality of male childbearing partners. The article measures the quality of male, nonmarital childbearing partners in terms of fathers' predicted wages, employment, and education as well as fathers' predicted wages and education relative to mothers. In future research, the quality of male partners could be broadened beyond economic status to include measures such as how helpful men are with housework and child care.
This research on unmarried parents finds a different pattern of results from research on married couples. Whereas Lichter et al. (1995) found that marriage markets that favor women led to marriages with more economically desirable men, I find that marriage markets are not associated with the economic characteristics of unmarried fathers. Whereas Trent and South (2003) found that marriage markets were not associated with marital happiness, I find that marriage markets are associated with the quality of unmarried parent relationships. A key difference between this article and these two studies is that I analyze unmarried parents, and those studies analyzed married couples who may or may not have had children. The differing results suggest that marital status, parental status, or both are important moderators of marriage market effects.
This article also examines whether the economic quality of male partners or relationship quality may be the mechanisms through which marriage markets inhibit transitions to marriage. I find evidence consistent with a theoretical model in which unfavorable marriage markets lead to lower quality relationships, which, in turn, inhibit marriage. Nevertheless, only 15% of unmarried parents in this sample were married at the 30-month follow-up in spite of high average reports of relationship quality. The low rates of marriage in the overall sample cannot be understood simply in terms of marriage markets or the parent and relationship characteristics measured in this article. Unfavorable marriage markets and low relationship quality explain only a small portion of nonmarriage in this study.
Because of my focus on new parents, the findings have direct implications for children's living arrangements and childrearing. My findings show that marriage markets appear to influence the characteristics of unmarried parent relationships and the transitions to marriage following a nonmarital birth. From these observations, one can speculate that children with unmarried parents will face disadvantages in unfavorable marriage markets: their parents' relationships are lower quality, and their parents are less likely to get married after their birth. Based on this research, one can predict that children born to unmarried parents in marriage markets that favor women will fare relatively better than their counterparts in marriage markets that favor men, which is a proposition that can be empirically tested in future research.
1. White and Asian groups were combined in the reference cell because the Asian sample size is small and white and Asian groups are similar in their means on outcomes and predictors.
Abma, J.C. and G.M. Martinez. 2006. "Childlessness Among Older Women in the United States: Trends and Profiles." Journal of Marriage and Family 68:1045-56.
Amato, P.R. and J.G. Gilbreth. 1999. "Nonresident Fathers and Children's Well-being: A Metaanalysis." Journal of Marriage and the Family 61:557-73.
Anderson, E. 1989. "Sex Codes and Family Life Among Poor Inner- City Youths." Annals of the American Academy of Political and Social Science 501:59-78.
Becker, G. 1981. A Treatise on the Family. Cambridge, MA: Harvard University Press.
Bzostek, S., M.J. Carlson, and S. McLanahan. 2007. "Repartnering After a Nonmarital Birth: Does Mother Know Best?" CRCW Working Paper 2006-27-FF. Center for Research on Child Wellbeing, Princeton University.
Carlson, M.J. and F.F. Furstenberg. 2006. "The Prevalence and Correlates of Multipartnered Fertility in the United States." Journal of Marriage and Family 68:718-32.
Carlson, M.J. and S.S. McLanahan. 2006. "Strengthening Unmarried Families: Could Enhancing Couple Relationships Also Improve Parenting?" Social Service Review 80:297-321.
Carlson, M.J., S. McLanahan, and P. England. 2004. "Union Formation in Fragile Families." Demography 41:237-61.
Case, A. and C. Paxson. 2005. "Sex Differences in Morbidity and Mortality." Demography 42: 189-214.
Cooksey, E.C. and P.H. Craig. 1998. "Parenting From a Distance: The Effects of Paternal Characteristics on Contact Between Nonresidential Fathers and Their Children." Demography 35:187-200.
Edin, K. and M.J. Kefalas. 2005. Promises I Can Keep: Why Poor Women Put Motherhood Before Marriage. Berkeley and Los Angeles, CA: University of California Press.
Ellwood, D. and C. Jencks. 2004. "The Spread of Single-Parent Families in the United States Since 1960." Pp. 25-65 in The Future of the Family, edited by D.P. Moynihan, T.M. Smeeding, and L. Rainwater. New York: Russell Sage.
England, P. and G. Farkas. 1986. Households, Employment, and Gender: A Social, Economic, and Demographic View. New York: Aldine Publishing Company.
England, P. and N. Folbre. 2002. "Involving Dads: Parental Bargaining and Family Well-being." Pp. 387-408 in Handbook of Father Involvement: Multidisciplinary Perspectives, edited by C.S. Tamis- LeMonda and N. Cabrera. Mahwah, NJ: Lawrence Erlbaum.
Erel, O. and B. Burman. 1995. "Interrelatedness of Marital Relations and Parent-Child Relations: A Meta-analytic Review." Psychological Bulletin 118:108-32.
Fossett, M.A. and K.J. Kiecolt. 1991. "A Methodological Review of the Sex Ratio: Alternatives for Comparative Research." Journal of Marriage and the Family 53:941-57.
_____. 1993. "Mate Availability and Family Structure Among African Americans in U.S. Metropolitan-Areas." Journal of Marriage and the Family 55:288-302.
Goldman, N., C.F. Westoff, and C. Hammerslough. 1984. "Demography of the Marriage Market in the United States." Population Index 50:5- 25.
Guttentag, M. and P.F. Secord. 1983. Too Many Women? The Sex Ratio Question. Beverly Hills: Sage Publication.
Harknett, K. and S.S. McLanahan. 2004. "Racial and Ethnic Differences in Marriage After the Birth of a Child." American Sociological Review 69:790-811.
Heaton, T.B., C.K. Jacobson, and K. Holland. 1999. "Persistence and Change in Decisions to Remain Childless." Journal of Marriage and the Family 61:531-39.
Howes, P. and H.J. Markman. 1989. "Marital Quality and Child Functioning: A Longitudinal Investigation." Child Development 60:1044-51.
Hutchens, R.M. 1979. "Welfare, Remarriage, and Marital Search." American Economic Review 69:369-79.
Kiecolt, J.K. and M.A. Fossett. 1995. "Mate Availability and Marriage Among African Americans: Aggregate and Individual-Level Analysis." Pp. 121-35 in The Decline in Marriage Among African Americans: Causes, Consequences, and Policy Implications, edited by B.M. Tucker and C. Mitchell-Kernan. New York: Russell Sage Foundation. Lampard, R. 1993. "Availability of Marriage Partners in England and Wales: A Comparison of Three Measures." Journal of Biosocial Science 25:333-50.
Lichter, D.T., R.N. Anderson, and M.D. Hayward. 1995. "Marriage Markets and Marital Choice." Journal of Family Issues 16:412-31.
Lichter, D.T., G. Kephart, D.K. McLaughlin, and D.J. Landry. 1992. "Race and the Retreat From Marriage: A Shortage of Marriageable Men." American Sociological Review 57:781-99.
Lichter, D.T., F.B. LeClere, and D.K. McLaughlin. 1991. "Local Marriage Markets and the Marital Behavior of Black and White Women." American Journal of Sociology 96:843-67.
Lundberg, S. and R.A. Pollak. 1996. "Bargaining and Distribution in Marriage." Journal of Economic Perspectives 10:139-58.
Manning, W.D. and P.J. Smock. 1999. "New Families and Nonresident Father-Child Visitation." Social Forces 78:87-116.
_____. 2000. "'Swapping' Families: Serial Parenting and Economic Support for Children." Journal of Marriage and the Family 62:111- 22.
McLanahan, S. and G. Sandefur. 1994. Growing Up With a Single Parent: What Hurts, What Helps. Cambridge, MA: Harvard University Press.
Monte, L.M. 2007. "Blended But Not the Bradys: Navigating Unmarried Multiple Partner Fertility." Pp. 183-203 in Unmarried Couples With Children, edited by P. England and K. Edin. New York: Russell Sage.
Norton, R. 1983. "Measuring Marital Quality: A Critical Look at the Dependent Variable." Journal of Marriage and the Family 45:141- 51.
Oppenheimer, V.K. 1988. "A Theory of Marriage Timing." American Journal of Sociology 94: 563-91.
Pettit, B. and B. Western. 2004. "Mass Imprisonment and the Life Course: Race and Class Inequality in U.S. Incarceration." American Sociological Review 69:151-69.
Qian, Z.C., D.T. Lichter, and L.M. Mellott. 2005. "Out-of- Wedlock Childbearing, Marital Prospects and Mate Selection." Social Forces 84:473-91.
Reichman, N.E., J.O. Teitler, I. Garfinkel, and S.S. McLanahan. 2001. "Fragile Families: Sample and Design." Children and Youth Services Review 23:303-26.
Robinson, J.G., B. Ahmed, P. Das Gupta, and K.A. Woodrow. 1993. "Estimation of Population Coverage in the 1990 United States Census Based on Demographic-Analysis." Journal of the American Statistical Association 88:1061-71.
South, S.J. 1996. "Mate Availability and the Transition to Unwed Motherhood: A Paradox of Population Structure." Journal of Marriage and the Family 58:265-79.
South, S.J. and K.M. Lloyd. 1992. "Marriage Markets and Nonmarital Fertility in the United States." Demography 29:247-64.
South, S.J., K. Trent, and Y. Shen. 2001. "Changing Partners: Toward a Macro-structural Opportunity Theory of Marital Dissolution." Journal of Marriage and Family 63:743-54.
Spanier, G.B. and P.C. Glick. 1980. "Mate Selection Differentials Between Whites and Blacks in the United States." Social Forces 58:707-25.
Sutton, P.D. and T.J. Matthews. 2004. "Trends in Characteristics of Births by State: United States, 1990, 1995, and 2000-2002." National Vital Statistics Reports 2(19).
Sweeney, M.M. 2002. "Two Decades of Family Change: The Shifting Economic Foundations of Marriage." American Sociological Review 67:132-47.
Trent, K. and S.J. South. 2003. "Spousal Alternatives and Marital Relations." Journal of Family Issues 24:787-810.
Willis, R.J. 1999. "A Theory of Out-of-Wedlock Childbearing." Journal of Political Economy 107: S33-S64.
Wilson, W.J. 1987. The Truly Disadvantaged: the Inner-City, the Underclass, and Public Policy. Chicago: University of Chicago Press.
_____. 1996. When Work Disappears: The World of the New Urban Poor. New York: Alfred A. Knopf Press.
Xie, Y., J.M. Raymo, K. Goyette, and A. Thornton. 2003. "Economic Potential and Entry Into Marriage and Cohabitation." Demography 40:351-67.
*Kristen Harknett, Department of Sociology, 3718 Locust Walk, 271 McNeil Building, Philadelphia, PA 19104; E-mail: email@example.com. The author acknowledges the support of the National Institutes of Health-National Institute on Aging Grant No. P30 AG12836, B.J. Soldo, P.I.; the National Institutes of Health- National Institute of Child Health and Development, Grant No. R24 HD044964, H.L. Smith, P.I.; and the Boettner Center for Pensions and Retirement Security-all at the University of Pennsylvania. This research was also supported by a grant from the University of Pennsylvania Research Foundation. The author gratefully acknowledges comments from Kathryn Edin, Frank Furstenberg, Jean Knab, Hugh Louch, Sara McLanahan, Tod Mijanovich, Julien Teitler, and Jane Waldfogel.
Copyright Population Association of America Aug 2008
(c) 2008 Demography. Provided by ProQuest Information and Learning. All rights Reserved.