Testing the Economic Independence Hypothesis: Union Formation Among Single Mothers in Chile

Abstract

Despite the increasing proportion of single-mother families, the literature on union formation among unpartnered parents in Latin America is scant. Using a sample from a Chilean longitudinal survey (N = 3,318), we estimated regression models to test associations between single mothers’ economic resources in 2010 and coresidence with the biological father of their children in 2012. We also examined whether these associations differed in subgroups with higher rates of labor market participation. We found that mothers’ paid work was not associated with union formation. Receiving government benefits and living with a parent were associated with lower probability of coresidence with the biological father, even among the most advantaged subgroups. We conclude that in contexts in which most mothers of young children are not doing paid work, economic resources coming from government and extended family may be more consequential than earnings’ potential in influencing single mothers’ union formation behavior.

Keywords

single parents independence hypothesis economic issues family policy Latin America

Introduction

The economic independence hypothesis posits that an increase in a woman’s economic resources reduces the benefits of partnering or staying in a coresidential union. While evidence from the United States and European countries supports the hypothesis (Cancian & Meyer, 2014; Carlson et al., 2004; Lyngstad & Jalovaara, 2010), little is known about the role of women’s economic resources in their union formation in Latin America, and whether the economic independence hypothesis holds in this region. A number of factors limit the application of prior research to the Latin American context. Economic resources coming from the labor market—a key predictor of women’s union formation in the United States and Europe—may be less significant among Latin American women, whose average labor market participation is relatively lower with respect to the United States and Europe (ECLAC & ILO, 2019). Financial support coming from extended family—a persistent feature of the Latin American family system—or government programs—in some cases more generous than in the United States—may play a more crucial role in the determination of coresidential unions. These factors may be particularly relevant to the increasing population of single mothers (Institute for Family Studies & Wheatley Institution, 2019), who may be at higher risk of poverty in the absence of these resources.

The fact that biological mothers establish a coresidential union with their children’s father is of great interest to policymakers and social scientists for a number of reasons. While nonmarital childbearing has increased dramatically, weakening the connection between union formation and fertility, and contributing to the growth of single-mother families worldwide (Laplante et al., 2015; Liu et al., 2017), children living with both parents are less likely to live in poverty (Cuesta et al., 2018; Salinas, 2011) and more likely to perform better academically and socially (Hofferth, 2006; Mariani et al., 2017) than children in single-parent families. Notably, living with a social father (i.e., a mother’s new partner) does not provide the same benefits for children as does living with their biological father (Hofferth, 2006).

A few studies have examined union formation among single mothers in affluent nations, particularly the United States (Bzostek et al., 2012; Cancian & Meyer, 2014; Carlson, McLanahan, et al., 2004). However, women who have had a nonmarital birth in the United States usually enter new partnerships (Bzostek et al., 2012). Data from Chile shows single mothers with low incomes rarely partner with social fathers. To the best of our knowledge, there is only one study that looks at changes in partnership status among unmarried mothers in Chile, but these mothers were already cohabitating (Salinas, 2016). We know of no studies in Latin America that examine the predictors of union formation among biological parents living apart. Extant research is also limited in that it only included women from the capital of Chile, sample design was non-probabilistic, and economic predictors of union formation were not included in the analyses.

Our study addresses these limitations and extends this literature by investigating the predictors of union formation among biological parents of young children in Chile, and the extent to which the economic independence hypothesis holds. Using nationally representative, longitudinal data from the Chilean Encuesta Longitudinal de Primera Infancia (ELPI), we regress mother’s union status in 2012 on both economic resources obtained by the mother herself and economic resources obtained through extended family and government in 2010, net of other factors associated with union formation. Furthermore, we examine whether these associations vary between subgroups of women with large differences in their labor market participation in Chile. Investigating how single mothers’ economic resources are associated with their partnering behavior in Chile will improve our current understanding of father’s absence in Latin America, and provide insight on how to improve the well-being of children born to unmarried parents, who are substantially worse off than children in other family arrangements (Salinas, 2011) and represent the majority of births in Chile (INE, 2017) and many other Latin American countries (Institute for Family Studies & Wheatley Institution, 2019).

The Chilean Context

The decline of marriage (OECD, 2018), the rise of cohabitation (Binstock et al., 2016), and the increase in nonmarital childbearing (INE, 2002, 2017) have transformed Chilean families more rapidly than families in other Latin American countries (Ramm, 2016; Salinas, 2016). Within just a few decades, the crude marriage rate dropped almost half (from 6.6 in 1992 to 3.4 in 2015) and the proportion of children born to unmarried parents increased from 49% in 2000 to 72% in 2015 (INE, 2002, 2017; OECD, 2019). Yet, little is known about the formation of unions by biological parents living apart, which is rather common among new mothers in Chile: approximately half of first-time mothers were not coresiding with the child’s biological father in 2010 and about one in seven of these families experienced biological fathers’ movement into the household within two years (Reynolds et al., 2018). The factors associated with union formation among single mothers in Chile, specifically whether mothers’ access to economic resources is potentially discouraging coresidential unions, has not yet been examined.

Relatively low levels of Chilean women’s labor market participation suggest earnings may not be as determinant of women’s union formation behavior. While the proportion of women aged 15–64 doing paid work increased approximately 8 percentage points over the last decade, the national rate is lower (49.2%) than the average observed overall in Latin America (56.6%) and a sample of European countries and the United States (70.7%) (ECLAC & ILO, 2019). However, there are some differences within the population. The percentage of women employed among those with college (58.7%) or graduate education (87%) is much higher than those with elementary (31.2%) or high school education (45.6%) (INE, 2015). Likewise, the employment rate among women under 25 years is less than half (25%) that of women aged 25–54 (over 60%) (INE, 2015). Finally, urban–rural gap estimates show a 41% participation in paid work among urban women while only 27.7% of women in rural areas are employed (ECLAC & ILO, 2016).

For women not doing paid work, other resources may be more consequential in their decision to coreside with the biological father of their children. Extended families may provide basic needs such as housing and food, which in turn may reduce single mothers’ economic insecurity. While this type of support is not unique to Latin America, the share of single mothers living in extended families is much higher than in the United States or Europe; in Chile, 81.8% single mothers aged 25–29 were living in extended family households at the beginning of the 21st century (Esteve et al., 2012). Government benefits may also reduce single mothers’ economic insecurity. Poverty alleviation programs in Chile are allocated by household income, which is lower for single mothers, and, in some cases such as the housing subsidy, are preferentially assigned to single mothers (Ramm, 2016). Unlike the United States, single mothers in Chile are not required to work to receive cash welfare and those receiving public assistance are allowed to keep all child support paid on behalf of their children.

Theoretical Framework and Prior Research

Theoretical Framework

The role of economic resources has been central to theoretical explanations of union formation. Becker posits that entering any co-residential union is contingent on whether the benefits of the relationship exceed its costs (Becker, 1973). The economic independence hypothesis emerges from this framework: an increase in women’s income—particularly earnings—reduces their benefits of coresidential unions. While this theory primarily focuses on the role of labor income, it can be broadened to include nonlabor resources. For example, it may be parents’ resources—rather than mothers’ earnings—what matter for women’s union formation behavior. Additionally, relying on government transfers may allow women to avoid partnerships that are primarily motivated by their fear of experiencing economic hardship. We review empirical research on union formation among single mothers in light of this theoretical framework.

Mothers’ economic resources

Studies in the United States have found a positive association between single mothers’ higher educational achievement—a key predictor of their earnings’ potential—and the formation of coresidential unions (Cancian & Meyer, 2014; Carlson et al., 2004). While no prior research has examined union formation among single mothers in Chile, one study looking at changes in cohabitation in the capital of Chile suggests the behavior of Chilean single mothers may be aligned with Becker’s predictions: higher educational attainment does not necessarily encourage marriage (Salinas, 2016). The United States literature finds that single mothers with any earnings are as likely to enter a coresidential union as those who are not doing paid work (Carlson et al., 2004). Prior research on union formation in Chile has not examined the role of earnings, though women’s lower participation in the labor market suggests paid work may not be as consequential in single mothers’ decision to coreside with their children’s biological father. Furthermore, social stigma around single motherhood and prevalence of traditional gender roles may weaken the effect of women’s earnings on union formation. Qualitative research finds Chilean women prioritize motherhood and domestic work over paid work and males prefer that their partners do not participate in the labor market (Ramm, 2016).

Having access to economic resources provided by the government may also influence union formation among single mothers. Findings from research in the United States are mixed. Early studies found a negative impact of welfare benefits on union formation (Garfinkel & McLanahan, 1986; Moffitt, 1998) while more recent literature finds no negative effect (Bitler et al., 2004; Carlson et al., 2004; Gennetian & Knox, 2003). Separations of couples with children increased when Brazil’s cash transfer program for mothers expanded (Litwin et al., 2019) but in Mexico, there was an increase in separations among those already partnered and an increase in unions among poor, young women (Bobonis, 2011). Although quantitative studies about union formation in Chile have not included measures of public assistance (e.g., Salinas, 2016), one qualitative study including single mothers finds they prefer to remain single in order to access housing subsidies (Ramm, 2016). Hence, in our study, we expect a negative association between receiving government subsidies and entry into a coresidential union.

The presence of a woman’s parents in the household may increase her economic independence (at least from the child’s biological father) by increasing her family income (Mutchler & Baker, 2009) or facilitating paid work (Posadas & Vidal-Fernandez, 2013). Because Chile has a high prevalence of extended family arrangements (Binstock et al., 2016) and the vast majority of single mothers live in extended family households (Esteve et al., 2012), we expect the presence of a parent will deter union formation.

Fathers’ economic resources

Studies in the United States generally find that fathers with higher economic capacities are more likely to enter a coresidential union than those with lower educational attainment or earnings (Carlson et al., 2004; Landale & Forste, 1991). However, one study finds that Hispanic fathers with limited earning potential are much more likely to marry their child’s mother than their African American counterparts (Harknett & McLanahan, 2004). This finding may be explained by less social acceptance of single motherhood among Latinos and the belief that, regardless of fathers’ circumstances, children and their mothers are better if fathers are around (Ramm, 2016). A random-assignment experiment in the United States finds child support transfers are not associated with union formation with a biological father but single mothers’ receiving these transfers are less likely to cohabit with a social father (Cancian & Meyer, 2014). Qualitative research from Chile suggests fathers who have higher economic capacities may be more likely to play the breadwinner role (Ramm, 2016). We expect fathers’ higher education, earnings, and child support payments will encourage coresidential unions.

Noneconomic factors

Other characteristics of single mothers’ families and the context in which they are raising their children may also influence their union decisions. Consistent with evidence for the United States (Carlson et al., 2004), prior research for the capital of Chile finds a negative association between children’s age and entry into a marital union (Salinas, 2016). The birth of a son has a strong positive effect on the probability of partnering with the child’s biological father in the United States (Lundberg & Rose, 2003). We anticipate a similar finding for Chile because son preference has been also documented in Latin America (Reynolds, 2018). Child’s poor health may deter union formation in the United States (Reichman et al., 2004), but Chile offers universal health care. Additionally, with a larger portion of mothers caring for their children at home in Chile, the stress of scheduling doctor’s visits may be reduced. Thus, we expect children’s health to have a smaller impact in Chile than in the United States. Cultural factors such as social acceptance of single motherhood may also influence single mothers’ partnering behavior (Carlson et al., 2004), and we expect these mechanisms to work similarly in Chile as in the United States. Other factors that may influence union formation include mother’s age, household structure, and mother’s relationship with the biological father. While there is evidence of a positive association between women’s age and union formation in the United States (Lichter & Graefe, 2001), evidence for the capital of Chile suggests there is no association between mother’s age at first birth and entry into a marital union (Salinas, 2016). Single mothers living in households with presence of more children may have more financial and childcare needs than those with fewer children around, which may deter or encourage biological father’s movement into the household. Finally, we expect fathers who maintain contact with children and or lived with the mother in the past will be more likely to partner.

Current Study

We address gaps in the literature on union formation among single mothers of young children in Latin America, particularly how economic resources are related to their decision to coreside with their children’s biological father. Our two research questions are: 1) how are single mothers’ economic resources associated with union formation with her children’s biological father in Chile? and 2) do these associations between mothers’ economic resources and union formation differ for subgroups of single mothers who have higher rates of participation in the labor market in Chile? We aim to test three hypotheses.

Hypothesis 1: Given lower participation in the labor market among Chilean women, we do not expect an association between single mothers’ earnings or earnings potential (education) and union formation with their children’s biological father.

Hypothesis 2: Receiving a government subsidy and coresiding with a parent will be associated with a lower probability of union formation between the single mother and her children’s biological father.

Hypothesis 3: Mothers’ economic resources, including those obtained from the job market, will have a negative association with union formation among older, urban, and highly educated single mothers.

Method

Data

We used data from the Chilean ELPI 2010 and 2012 waves. This nationally representative survey provides longitudinal information about children’s development and includes information on child, parent, and family characteristics. As a focal child survey, children in the national birth registry born between January 1, 2006 and August 31, 2009 were selected using a cluster stratified sampling technique. The clusters (counties) were stratified by region, income per capita, and population of children. The 2010 survey included 15,175 households, but there was a 15% attrition rate in the 2012 follow-up (Centro Microdatos, 2010).

Sample

In 2010, 30% of focal children lived in the same household as their biological mother but not their biological father (N = 4,654). For our analysis on the biological father joining the focal child’s household, we excluded families with focal children whose mothers had already re-partnered with a social father (N = 217), whose fathers were deceased (N = 51), traveling (N = 21), or in prison (N = 92) at baseline. Of these households, 85.3% were re-surveyed in 2012 (N = 3,644). We then excluded families in which the focal child’s biological mother was absent from the household roster in 2012 (N = 62), and families in which the focal child’s biological father was deceased (N = 24), traveling (N = 6) or was in prison (N = 28) in 2012. We excluded families where the focal child was not the youngest child in the household (N = 207), and we also dropped one family in which the time between survey rounds was not available. After these exclusions, our analytic sample included 3,318 single-mother families at baseline (see Table 1 for descriptive statistics). Our analytic sample is relatively similar to the single-mother households excluded from the analyses (results available upon request).

Table 1.

Sample Descriptive Statistics in 2010 by Union Status in 2012.

	Full Sample		No Change		Union Formed		Diff.	% Missing
	Mean or %	SD	Mean or %	SD	Mean or %	SD	Diff.	% Missing
Mother’s economic resources
Education (base: completed higher ed.)	7.70		7.70		7.70			0.9
Elementary school or less	12.70		12.80		12.60			0.9
High school completed	62.20		61.90		64.20			0.9
Some higher education	17.30		17.60		15.60			0.9
Math score (standardized)	0.03	0.02	0.03	0.02	0.02	0.04		5.0
Vocab score (standardized)	−0.03	0.02	−0.04	0.02	0.03	0.04		5.0
Work income (of earners only)^{a b}	352	342	352	352	346	274
Has work contract	30.80		30.80		30.50
Other household income^b	702	1,292	702	1,298	704	1,260		4.1
Government subsidy	33.80		34.90		27.80		**
Parent in household	74.10		74.90		69.50		*
Housing tenure status (base: rented/borrowed)	54.20		33.10		39.00		*	0.2
Fully paid home	11.80		54.90		50.20		+	0.2
Mortgaged home	34.00		12.00		10.80			0.2
Father’s economic resources
Education (base: completed higher ed.)	8.10		8.00		8.30
Unknown	18.80		19.90		13.20		***
Elementary school or less	10.70		11.00		8.70
High school completed	50.50		49.20		57.50		**
Some higher education	11.90		11.90		12.40
Supports child economically	66.20		63.10		83.70		***	0.5
Child’s characteristics
Has male child	61.50		61.60		60.40
Age of youngest child	1.95	0.02	1.97	0.02	1.82	0.05	**
Low birth weight	3.80		3.70		3.90			8.0
Ever hospitalized or had surgery	25.90		25.70		27.00
Gastrointestinal problems	8.90		8.80		9.80
Respiratory problems	46.90		47.30		44.30
Other physical problems	14.80		15.20		12.40		+
Psychological problems	5.80		5.70		6.70
Child development score (standardized)	0.02	0.02	0.01	0.02	0.09	0.04		5.7
Cultural context
Prevalence of single motherhood	40.60	0.00	40.60	0.00	40.90	0.01
Urban	91.30		91.10		92.30	0.01
Other factors
Mother’s age	26.31	0.12	26.41	0.13	25.78	0.29	+
Mother had first child at age 21 or younger	61.20		60.90		62.80
Num. of bio. children	1.51	0.02	1.51	0.016	1.52	0.039
Num. of additional children in household	0.64	0.02	0.62	0.02	0.72	0.05	*
Num. of adults (excludes mother and parents)	0.86	0.02	0.85	0.02	0.92	0.05
Contact with bio. father (number of days)	10.47	0.21	9.35	0.22	16.64	0.54	***
Never lived with father	61.40		63.40		50.20		***
Months between survey rounds	26.08	0.03	26.08	0.03	26.08	0.08
N (single mothers)	3318		2810		508

Notes. Diff. column indicates statistically significant difference in estimates between those who formed union and those who did not form union using a t-test.

N = 1,765 full sample, N = 1,509 no change, N = 256 union.

Monthly income in U.S. $.

Num. = Number; bio. = biological; ed. = education.

p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.

Measures

Union formation

Our measure of union formation between a single mother and the biological father of her children was an indicator of whether the focal child’s biological father was living in the household at the time of the second survey wave (2012).¹ Twenty percent of those who formed a union were formally married but we do not distinguish between formal and informal union.

Mother’s economic resources

We distinguished between economic resources that were generated by the mother and those provided by institutions such as government and family. The former included education, cognitive abilities, and labor market indicators. We chose to include both education and cognitive abilities because education measures achievement while cognitive abilities are indicative of skill in two distinct realms (mathematics and vocabulary); all could influence the mother’s ability to acquire and maintain employment. We measured mother’s education with five dummy variables: elementary school or less, high school completed, some higher education, higher education completed (base category), and unknown. We used standardized scores of the mother’s mathematics and vocabulary skills on the Wechsler Adult Intelligence Scale (Apfelbeck & Hermosilla, 2000) as measures of cognitive abilities. We used the natural log of mother’s earnings in the past month, with a value of 0.0001 for the 47% of the sample who were not working. In order to capture mother’s access to employment benefits we included a dummy variable of whether the mother has a work contract. We included government support with an indicator of whether the mother’s household received public assistance from the conditional cash transfer program Subsidio Unico Familiar (Unique Family Subsidy) in the past month. We also included several covariates to measure family support: to measure household income (excluding mother’s income), we used the log of other income in the past month, with a value of 0.0001 for those who did not have other sources of income. We also included an indicator of whether the focal child has a grandparent in the household, which based on prior research we assume to be the mother’s parent (Reynolds et al., 2018). Finally, we used housing tenure status to measure mother’s assets. We included three dummy variables: fully paid, mortgaged, or rented/borrowed (base category).

Father’s economic resources

We used mother’s reports of father’s education and economic support for the child. We measured father’s education using the same categories created for mothers. We also included a dichotomous measure of whether the child received economic support from the father as the exact amount was not recorded.

Noneconomic factors

We included noneconomic factors that may influence mothers’ union behavior as discussed in the prior research section. We included an indicator of whether any of the mother’s children is male. We measured age of the youngest child as a continuous measure of child’s years. We used information available for the focal child to measure child’s characteristics that may influence parents’ decision to partner: an indicator for low birth weight, dummy variables for hospitalization or surgery, respiratory, gastrointestinal or other physical illnesses (i.e., kidney, growth, visual, audio, and skin), and psychological challenges (i.e., learning disorder or delays, mental health, trauma, and neurological). We also included a measure of child development. We created this index from factor analyses of age-standardized scores. For children ages 24 months and older, we generated this index using the cognitive, language, and motor subscales of the Tepsi evaluation of child development (Haussler & Marchant, 1980) and, for children younger than 24 months, we generated this index using the personal–social, adaptive, motor, communicative, and cognitive subscales of the Battelle test (Newborg, 2005). Though each of these tests has its own scoring of overall ability, we performed factor analysis to ensure similar procedures across tests. To account for the role of cultural factors² we included a dummy variable for urban and also generated a measure of the prevalence of single motherhood where the mother was residing at the time of the first wave. We also included other factors that may influence union formation such as single mother’s age (and age squared), measured as continuous variables of her self-reported age in years, and whether the mother had her first child at age 21 or younger, calculated as the difference between the mother’s age and focal child’s age or the focal child’s oldest sibling’s age. We measured the number of biological children by the number of siblings of the focal child plus the focal child. The number of adults included any household member age 18 and older who is not the biological mother or the focal child’s grandparents. Similarly, we measured number of additional children as all children age 17 and younger who are not the focal child’s siblings. We measured child’s contact with the biological father using the mother’s reports of father–child contact. The possible answers were re-scaled to indicate days per month (daily = 30, 5–6 days a week = 22, 3–4 days a week = 14, 1–2 days a week = 6, biweekly = 2, monthly = 1, never = 0). A dummy variable indicates that the mother never lived with the biological father. We included months between survey rounds, which was measured by the difference in the focal child’s age between surveys.

All predictor variables were measured at the first survey wave (2010). A few variables had a very small percentage of missing data; see Table 1 for details. We imputed the mean value if the variable was continuous and the modal value if the variable was discrete.

Analytic Plan

Main analyses

To examine associations between mothers’ economic resources and union formation we estimated a series of logit regression models predicting coresidence with the biological father in 2012. In Model 1, our most basic model, we estimated the associations between different measures of mothers’ economic resources and coresidence with the biological father in 2012. In Models 2 to 5, we progressively adjusted for other factors associated with union formation by adding fathers’ economic resources (Model 2), child characteristics (Model 3), proxies for cultural context (Model 4), and other factors (Model 5); we followed this strategy to examine how findings from the basic model changed when other determinants of union formation are considered. We estimated marginal effects at the mean using robust standard errors and included months between rounds. All analyses were performed in Stata 14.

Sensitivity analyses

One analytic challenge that pervades the literature on the economic independence hypothesis is the possibility that women’s economic behavior may be shaped by the anticipation of whether or not she will be living with the biological father of her children in the near future. For instance, if a single mother is not planning to marry or cohabit with the biological father of her children, she may be more likely to live with extended family members or apply for government benefits. While we cannot rule out reverse causality between mothers’ economic resources and union formation, we address this issue by estimating Model 5 for families in which the youngest child was 2 years or younger in 2010. In this subset of families, marriage or cohabitation is more of an open question because we can be sure a romantic relationship occurred relatively recently. Therefore, we may expect the direction of the association between the independent and dependent variable to go from single mothers’ economic resources to union formation and not the other way around. We also examined whether our main findings changed when we excluded cases that increased statistical precision, but may have potentially biased our estimates: families in which the parents coresided prior to 2010, families in which the reason for father absence in 2010 is unknown, and families in which the mother partnered with other men by 2012.

Subgroup analyses

To determine whether the associations between single mothers’ economic resources and union formation differ for subgroups of mothers who have higher labor market participation, we estimated three subgroup analyses using Model 5: we divided the sample by whether the mother had a child at a young age (21 years old or less), whether the family lived in an urban area, and whether the mother had some higher education. We estimated the p-value of a Wald test equality of each of the marginal effects (at the mean of each subsample) and we also used the Wald test to estimate the joint significance for the equality of all the marginal effects.

Results

Table 1 presents sample descriptive statistics in 2010 by union formation in 2012. In total, 53% of single mothers in our analytical sample are employed and most have completed high school. Most strikingly, 74% coreside with a parent. Although the mothers are not that young, (26 years of age on average), over 60% of them had their first child at age 21 years or younger.

Predictors of Union Formation

Overall, 15% of single mothers of young children partnered with their children’s biological father within two years. Our main results of the predictors of this event are presented in Table 2. Model 1 shows the associations between mothers’ economic resources in 2010 and coresidence with the biological father of her children in 2012 before adjusting for other factors associated with union formation. Several of the associations are statistically significant: single mothers’ earnings, receiving a government subsidy, and living with a parent are all characteristics negatively associated with the probability of coresiding with the biological father. At a trend level (p < .10), single mothers living at homes that have been fully paid or have a mortgage have a lower probability of union formation with the biological father. After adding fathers’ economic resources in Model 2, most of these associations remained statistically significant; however, mothers’ earnings are associated with lower probability of union formation only at a trend level (p < .10). Model 2 also shows that child support payments from the nonresident father are positively associated with union formation. After adding child characteristics in Model 3, single mothers’ earnings are no longer statistically significant, suggesting resources from the labor market are confounded with other factors. Consistent with prior research (Carlson et al., 2004; Salinas, 2016), we find that mothers with younger children are more likely to form unions with the biological father than those who have older children. At a trend level (p < .10), we found support for the negative association between child’s physical ailments and union formation also found in the United States (Reichman et al., 2004).

Table 2.

Predictors of Union Formation.

	Model 1		Model 2		Model 3		Model 4		Model 5
	ME	SE	ME	SE	ME	SE	ME	SE	ME	SE
Mother’s economic resources
Education (base: completed higher ed.)
Elementary school or less	0.013	(0.031)	0.032	(0.031)	0.033	(0.030)	0.033	(0.031)	0.029	(0.030)
High school completed	0.016	(0.025)	0.019	(0.025)	0.019	(0.025)	0.018	(0.025)	0.011	(0.024)
Some higher education	−0.015	(0.029)	−0.013	(0.027)	−0.013	(0.027)	−0.015	(0.027)	−0.019	(0.027)
Math score (standardized)	−0.004	(0.007)	−0.005	(0.007)	−0.006	(0.007)	−0.006	(0.007)	−0.006	(0.006)
Vocab score (standardized)	0.014⁺	(0.008)	0.011	(0.007)	0.011	(0.007)	0.011	(0.007)	0.012⁺	(0.007)
Work income (natural log)	−0.003*	(0.001)	−0.002⁺	(0.001)	−0.002	(0.001)	−0.002	(0.001)	−0.002	(0.001)
Has work contract	0.012	(0.018)	0.012	(0.017)	0.012	(0.017)	0.012	(0.017)	0.013	(0.017)
Other household incomes (natural log)	0.00	(0.002)	−0.001	(0.002)	−0.001	(0.002)	−0.001	(0.002)	−0.002	(0.001)
Government subsidy	−0.049**	(0.014)	−0.037**	(0.014)	−0.037**	(0.013)	−0.036**	(0.013)	−0.034**	(0.013)
Parent in household	−0.035*	(0.015)	−0.026⁺	(0.014)	−0.030*	(0.014)	−0.031*	(0.015)	−0.043**	(0.015)
Housing tenure status (base: rented/borrowed)
Fully paid home	−0.025⁺	(0.014)	−0.024⁺	(0.013)	−0.022⁺	(0.013)	−0.021⁺	(0.013)	−0.015	(0.013)
Mortgaged home	−0.037⁺	(0.021)	−0.035⁺	(0.020)	−0.034⁺	(0.020)	−0.034⁺	(0.020)	−0.024	(0.019)
Father’s economic resources
Education (base: completed higher ed.)
Unknown			−0.023	(0.027)	−0.024	(0.027)	−0.023	(0.027)	−0.001	(0.027)
Elementary school or less			−0.022	(0.030)	−0.025	(0.030)	−0.025	(0.030)	−0.02	(0.029)
High school completed			0.018	(0.023)	0.016	(0.023)	0.015	(0.023)	0.011	(0.023)
Some higher education			0.009	(0.027)	0.006	(0.027)	0.005	(0.027)	0.004	(0.026)
Supports child economically			0.123**	(0.015)	0.122**	(0.015)	0.123**	(0.015)	0.071**	(0.015)
Child characteristics
Has male child					−0.009	(0.012)	−0.009	(0.012)	−0.009	(0.012)
Age of youngest child					−0.011*	(0.005)	−0.011*	(0.005)	−0.009	(0.006)
Low birth weight					0.005	(0.030)	0.006	(0.030)	0.005	(0.029)
Ever hospitalized or had surgery					0.015	(0.014)	0.015	(0.014)	0.011	(0.013)
Gastrointestinal problems					0.021	(0.020)	0.021	(0.020)	0.019	(0.019)
Respiratory problems					−0.019	(0.012)	−0.018	(0.012)	−0.017	(0.011)
Other physical problems					−0.033⁺	(0.018)	−0.033⁺	(0.018)	−0.022	(0.017)
Psychological problems					0.031	(0.025)	0.031	(0.025)	0.028	(0.023)
Child development score (standardized)					0.009	(0.006)	0.009	(0.006)	0.007	(0.006)
Cultural context
Prevalence of single motherhood							0.011	(0.034)	−0.055	(0.063)
Urban							0.02	(0.022)	0.000	(0.022)
Other factors
Mother’s age									0.011	(0.011)
Mother’s age²									0.000	(0.000)
Mother had first child at age 21 or younger									0.000	(0.016)
Num. of bio. children									0.006	(0.010)
Num. of additional children in household									0.015*	(0.006)
Num. of adults (excludes mother and parents)									0.009	(0.005)
Contact with bio. father (number of days)									0.004**	(0.000)
Never lived with father									−0.055**	(0.012)
Months between survey rounds	0.001	(0.004)	0.001	(0.003)	0	(0.003)	0	(0.003)	0.000	(0.003)
N (single mothers)	3318		3318		3318		3318		3318

Notes. Marginal effects (and standard errors) presented. The base probability of union formation is 15.3% controlling for time between rounds. ME = Marginal effects; SE = Standard errors; Num. = Number; bio. = biological; ed. = education.

p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.

The significant associations between mothers’ economic resources and union formation established in Model 1 and that were still strong in Model 3 remained statistically significant and consistent in magnitude throughout Models 4 and 5, as was the marginal effect for fathers’ child support. Though the coefficient on parent in the household seemed to increase with more control variables, the coefficient in Model 5 is not statistically distinct from the coefficient in Model 1, suggesting an interaction with another variable is unlikely. Our proxies for cultural factors added in Model 4 were not associated with union formation. Child contact with the father was positively associated with union formation (Model 5). This variable slightly reduced the magnitude of the marginal effect of child support, but not by so much that it became statistically insignificant.

Our findings provide partial support to the economic independence hypothesis. Single mothers’ economic resources were associated with a lower probability of union formation in Chile; however, this association was only observed for resources provided by the government and extended families; the initial finding of earnings being associated with union formation was explained by other variables. Similarly, mother’s education and intellectual ability, other proxies for human capital and ability to access economic resources, were not associated with union formation. Father’s education did not predict formation, but his economic support and interactions with the child were statistically significant factors.

Sensitivity Analyses

Our main analyses included single mothers with a focal child approximately age 0 to 5 when parents were first observed in 2010. Because the role of single mothers’ economic resources in their union formation may be more challenging to predict among parents with older children—who may have applied to government benefits after deciding against union formation and not the other way around—we estimated Model 5 with a subset of families in which the youngest child was 2 years or younger in 2010. Results are consistent with findings from our main models: economic resources provided by government or extended family have a statistically significant negative association with union formation (see Table 3, columns 1 and 2). Our main findings remained robust when we excluded cases that increased precision of our estimates but may have potentially biased our results (see Table 3, columns 3–8).

Table 3.

Sensitivity Analyses of Predictors of Union Formation.

	Families in Which Youngest Child was 2 Years or Younger in 2010		Mothers Had Never Lived with Father Prior to 2010		Excludes Fathers Not in the hh. in 2010 for “Other” Reason		Mothers Who Had Not Partnered with Other Men by 2012
	ME	SE	ME	SE	ME	SE	ME	SE
Mother’s economic resources
Education (base: completed higher ed.)
Elementary school or less	0.059	(0.038)	0.028	(0.032)	0.009	(0.030)	0.03	(0.032)
High school completed	0.009	(0.032)	−0.006	(0.026)	−0.003	(0.024)	0.011	(0.026)
Some higher education	−0.026	(0.035)	−0.052⁺	(0.028)	−0.028	(0.026)	−0.025	(0.028)
Math score (standardized)	−0.008	(0.008)	0.000	(0.007)	−0.003	(0.006)	−0.006	(0.007)
Vocab score (standardized)	0.013	(0.009)	0.004	(0.008)	0.008	(0.007)	0.011	(0.008)
Work income (natural log)	−0.002	(0.001)	−0.001	(0.001)	0	(0.001)	−0.002	(0.001)
Has work contract	0.01	(0.022)	0.007	(0.018)	−0.002	(0.017)	0.016	(0.018)
Other household incomes (natural log)	−0.003	(0.002)	−0.002	(0.002)	−0.002	(0.001)	−0.001	(0.002)
Government subsidy	−0.055**	(0.016)	−0.035*	(0.014)	−0.030*	(0.013)	−0.037**	(0.014)
Parent in household	−0.061**	(0.019)	−0.046**	(0.017)	−0.046**	(0.016)	−0.052**	(0.017)
Housing tenure status (base: rented/borrowed)
Fully paid home	−0.004	(0.016)	−0.003	(0.013)	−0.018	(0.013)	−0.017	(0.014)
Mortgaged home	−0.046⁺	(0.026)	−0.008	(0.020)	−0.023	(0.020)	−0.025	(0.021)
Father’s economic resources
Education (base: completed higher ed.)
Unknown	0.004	(0.036)	−0.042	(0.027)	0.001	(0.027)	−0.006	(0.028)
Elementary school or less	−0.019	(0.038)	−0.047	(0.033)	−0.02	(0.030)	−0.026	(0.032)
High school completed	0.026	(0.030)	−0.007	(0.024)	0.01	(0.023)	0.009	(0.025)
Some higher education	0.026	(0.035)	−0.019	(0.027)	−0.003	(0.027)	−0.003	(0.028)
Supports child economically	0.075**	(0.019)	0.056**	(0.017)	0.057**	(0.015)	0.075**	(0.016)
N (single mothers)	2205		2037		2963		3098
Probability of union formation ^a	0.163		0.125		0.142		0.164

Notes. Marginal effects (and standard errors) presented.

Controlling for time between rounds. Models adjust for child characteristics, cultural context, and other factors.

ME = Marginal effects; SE = Standard errors; Num. = Number; bio. = biological; ed. = education; hh. = household.

p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.

Subgroup Analyses

We tested if the economic independence hypothesis would be more likely among women with higher rates of labor market participation. The Wald Test for jointly testing the differences of the marginal effects showed significant differences in the models contrasting younger mothers with older mothers and mothers living in urban areas with mothers living in rural areas (see Table 4). On the other hand, contrasting mothers with some higher education or more to mothers with secondary education or less were not separable according to this test, although the marginal effects of some variables were statistically distinct when tested individually.

Table 4.

Subgroup Analyses of Predictors of Union Formation.

	Young Mother ^a		Older Mother		Diff ^b	Urban		Rural		Diff ^b	Some higher ed. or more		Secondary or less		Diff ^b
	ME	SE	ME	SE	Diff ^b	ME	SE	ME	SE	Diff ^b	ME	SE	ME	SE	Diff ^b
Mother’s economic resources
Education (base: completed higher ed.)
Elementary school or less	−0.012	(0.051)	0.038	(0.042)		0.023	(0.032)	0.061	(0.086)
High school completed	−0.029	(0.047)	0.025	(0.027)		0.012	(0.025)	0.02	(0.080)
Some higher education	−0.063	(0.049)	0.004	(0.030)		−0.021	(0.027)	0.06	(0.092)
Math score (standardized)	−0.013	(0.008)	0.005	(0.010)		−0.004	(0.007)	−0.027	(0.019)		−0.014	(0.012)	−0.003	(0.008)
Vocab score (standardized)	0.008	(0.009)	0.015	(0.010)		0.011	(0.007)	0.031	(0.025)		0.011	(0.012)	0.01	(0.008)
Work income (natural log)	−0.002	(0.001)	−0.001	(0.002)		−0.002	(0.001)	0.000	(0.003)		0.002	(0.002)	−0.003*	(0.001)	+
Has work contract	0.024	(0.022)	−0.006	(0.024)		0.017	(0.017)	−0.048	(0.054)		−0.017	(0.031)	0.022	(0.019)
Other household incomes (natural log)	−0.003	(0.002)	−0.001	(0.002)		−0.002	(0.001)	0.007	(0.008)		0.001	(0.003)	−0.003⁺	(0.002)
Government subsidy	−0.025	(0.016)	−0.052*	(0.023)		−0.038**	(0.014)	−0.026	(0.036)		−0.05	(0.035)	−0.030*	(0.014)
Parent in household	−0.038⁺	(0.020)	−0.045*	(0.022)		−0.051**	(0.016)	−0.003	(0.045)		−0.041	(0.031)	−0.047**	(0.017)
Housing tenure status (base: rented/borr.)
Fully paid home	−0.025	(0.016)	−0.005	(0.019)		−0.02	(0.013)	0.042	(0.034)	+	0.031	(0.024)	−0.031*	(0.014)	*
Mortgaged home	−0.038	(0.026)	−0.008	(0.027)		−0.035⁺	(0.020)	0.157*	(0.070)	**	−0.027	(0.034)	−0.02	(0.023)
Father’s economic resources
Education (base: completed higher ed.)
Unknown	0.089*	(0.042)	−0.076*	(0.035)	**	−0.011	(0.028)	0.094	(0.071)		−0.008	(0.039)	0.006	(0.037)
Elementary school or less	0.078⁺	(0.044)	−0.122*	(0.048)	**	−0.036	(0.032)	0.113	(0.073)	⁺	−0.146	(0.108)	−0.001	(0.038)
High school completed	0.074⁺	(0.039)	−0.015	(0.028)	⁺	0.013	(0.024)	0.014	(0.064)		0.019	(0.028)	0.016	(0.033)
Some higher education	0.073⁺	(0.043)	−0.037	(0.033)	*	0.002	(0.028)	0.071	(0.083)		−0.015	(0.031)	0.032	(0.040)
Supports child economically	0.081**	(0.020)	0.047*	(0.022)		0.068**	(0.016)	0.087⁺	(0.044)		0.068*	(0.030)	0.067**	(0.017)
N (single mothers)	2030		1288			3028		290			830		2488
Probability of union formation ^c	0.157		0.147		*	0.155		0.134		**	0.142		0.157

Notes. Marginal effects (and standard errors) presented.

Young mothers had their first-born child at age 21 or younger; ^bp-values from testing that the coefficients are equal; ^ccontrolling for time between rounds. Models adjust for child characteristics, cultural context, and other factors.

ME = Marginal effects; SE = Standard errors; Num. = Number; bio. = biological; ed. = education; borr = borrowed.

p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.

Yet even though some of the models differ by subgroup overall, marginal effects on government subsidy and parent in the household are not statistically distinct across subgroups. Differences emerged for other key variables. For younger mothers, having lower levels of fathers’ education is associated with union formation while for older mothers the opposite is the case. For urban mothers, a home that is owned is associated with independence, whereas for rural mothers, it is associated with union formation. This relationship is similar for the less educated mothers, who are less likely to form a union if they live in a home that is owned.

Discussion

We examined the role of economic resources in single mothers’ decision to coreside with biological fathers of their children in Chile. Our results provide support to the economic independence hypothesis, which predicts that an increase in women’s economic resources reduces their benefits of coresidential unions. However, these findings are distinct from Becker’s theoretical predictions. Instead of relying on their own economic capabilities to remain single, single mothers rely on government and parents, as receiving a government subsidy and coresiding with a parent are both associated with less likelihood of union formation. Single mothers also are more likely to form a union with a partner who is providing child support. Our findings indicate that external economic resources (e.g., family, government, and partner) rather than own economic potential shape union formation behavior of Chilean single mothers.

In contrast to the traditional conceptualization of the economic independence hypothesis, mothers’ education and earnings were not statistically significant predictors of union formation. This finding may be partly explained by traditional gender-related norms in Chile. While earnings could be critical resources for the well-being of the mother and her young children, societal expectation that mothers should prioritize motherhood and domestic work over paid work may encourage single mothers to look for economic support outside the labor market. Whether mother’s earnings become determinant of union formation when children reach school-age years is an important question for future research.

With Chile’s leap in economic growth in the past 30 years, government and families may have more resources to support young, unpartnered mothers. The government subsidy is not particularly large (U.S. $15 per month), so we would not expect this income to be a sole source of support for single mothers. However, being registered in the social services system may provide access to other types of assistance such as housing subsidies (Ramm, 2016). While the variable household income excluding mother’s income was not significant, descriptive statistics show more detail on how parents may be providing financial support to single mothers. In 92% of households in which young mothers did not coreside with their child’s biological father but coresided with at least one parent (specifically, the child’s grandparent), a parent was indicated as the household head. These parents provided labor income or pensions in 70% of households with single mothers. When the parent was present, the home was owned and fully paid in 60% of households. In contrast, in households with single mothers but not parents, this figure was 36%.

We found differences in the association between fathers’ education and union formation when contrasting the behavior of younger and older women. Women who became mothers when older than age 21 are more likely to partner if the fathers have higher levels of education, as expected. In contrast to Becker’s theory suggesting father’s ability to support the family is an important factor in union formation, women who became mothers at younger ages are more likely to partner with fathers who have not completed higher education. It may be the father’s preference rather than the mother’s that explains this result: while mothers may prefer fathers with higher education, fathers with college degrees may prefer highly educated mothers as well, and women who had children at young ages are unlikely to have completed their college education.

Aligned with Becker’s theory of union formation, we found father’s economic support to his children is strongly associated with union formation, with the likelihood of formation increasing by 12% when the father supports children economically. This reduces to 7% when controlling for time spent with the child, which suggests the financial support is correlated with father involvement. That economic ability to support children remains important beyond the time investment, however, highlights the importance of the father’s economic commitment. This may be even more important than economic earning potential, since education was not associated with union formation. However, further research should seek to understand which aspect of fathers’ economic capacities encourage union formation in order to shed light on policies designed to promote child and family well-being.

Our findings should be interpreted in light of the following limitations. First, data on absent fathers is reported by the mother and not as detailed as data on coresiding fathers. However, since most couples have relationships prior to becoming pregnant, we can assume basic information such as education is accurate. The question on whether the father supports the child economically is vague, without a time frame or monetary value. Still, analyzing child support receipt is an important first step, and our findings support that future surveys add more detail on this topic. Finally, we do not have information on the fathers’ paternity of additional children outside the household in 2010, which is associated with fathers’ ability to partner.

That the data come from a focal child survey is a strength regarding the detailed information on the focal child but does come with some limitations. We do not have information on siblings’ health and child development scores, but these are typically correlated across children. Furthermore, we designed our analytical sample to ensure that the focal child was the youngest child, whose challenges may be most pertinent. Another limitation from the focal child survey structure is that the parent variable was determined by the relationship to the focal child, not the single mother. However, the Chilean culture is matrilocal, so the child’s grandparent is unlikely to be paternal, especially since the father is absent from the household.

Despite these limitations, our study has a number of strengths. This is the first study that examines predictors of union formation among unpartnered, biological parents of young children in Latin America. The sample size is relatively large for the literature and a relatively high proportion of unions are formed. We include a large number of control variables. Our sensitivity analyses confirm the robustness of the results and our subgroup analyses show the importance of family and government economic support across various populations.

Our findings provide insight for child and family policy. This research ameliorates concerns about women’s labor market participation disrupting families as predicted by the economic independence hypothesis. Instead, promoting mothers’ paid work is unlikely to result in changes in union formation in Chile, but may improve the economic situation of the family. Programs that encourage nonresident father involvement may also improve future union formation as well. However, focusing policy only on union formation is simplistic. While there is empirical evidence that children benefit from being raised by both biological parents, partnering behavior that is entirely determined by economic hardship may be more harmful to single mothers and their children than father absence. In these cases, engaging extended family and social services providers in supporting the mother and child can be helpful. The other cases are where partnering with the biological father is only inhibited because family economic support is more stable than the young father’s resources. Encouraging the stability of the young couple in spite of still having to rely on external financial support is important. In order for this family organization to work, the father’s role may have to shift from sole breadwinner toward increased caregiving responsibilities.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received financial support for the research, authorship, and/or publication of this article: This work was supported by the Berkeley Population Center andthe Eunice Kennedy Shriver National Institute for Child Health and Human Development (P2CHD07396).

ORCID iD

Laura Cuesta

Notes

References

Apfelbeck

E.-M.

Hermosilla

(2000). Manual de Administración y Tabulación del Test de WAIS. Santiago, Chile.

Becker

(1973). A theory of marriage: Part I. Journal of Political Economy, 81(4), 813–846.

Binstock

Cabella

Salinas

López-Colás

(2016). The rise of cohabitation in the Southern cone. In Esteve

Lesthaeghe

R. J.

(Eds.), Cohabitation and marriage in the Americas: Geo-historical legacies and new trends (pp. 247–268). Springer. https://link-springer-com-443.web.bisu.edu.cn/content/pdf/10.1007%2F978-3-319-31442-6.pdf

Bitler

M. P.

Gelbach

J. B.

Hoynes

H. W.

Zavodny

(2004). The impact of welfare reform on marriage and divorce. Demography, 41(2), 213–236. https://doi.org/10.1353/dem.2004.0011

Bobonis

G. J.

(2011). The impact of conditional cash transfers on marriage and divorce. Economic Development and Cultural Change, 59(2), 281–312. https://doi.org/10.1086/657123

Bzostek

S. H.

McLanahan

S. S.

Carlson

M. J.

(2012). Mothers’ repartnering after a nonmarital birth. Social Forces, 90(3), 817–841. https://doi.org/10.1093/sf/sos005

Cancian

Meyer

D. R.

(2014). Testing the economic independence hypothesis: The effect of an exogenous increase in child support on subsequent marriage and cohabitation. Demography, 51(3), 857–880. https://doi.org/10.1007/s13524-014-0295-8

Carlson

Garfinkel

McLanahan

Mincy

Primus

(2004). The effects of welfare and child support policies on union formation. Population Research and Policy Review, 23(5), 513–542. https://doi.org/10.1007/s11113-004-2707-8

Carlson

McLanahan

England

(2004). Union formation in fragile families. Demography, 41(2), 237–261.

10.

Centro Microdatos. (2013). Presentación: Segunda Ronda Encuesta Longitudinal de la Primera Infancia. Departamento de Economia, Universidad de Chile.

11.

Cuesta

Hakovirta

Jokela

(2018). The antipoverty effectiveness of child support: Empirical evidence for Latin American countries. Social Policy & Administration, 52, 1233–1251. https://doi.org/10.1111/spol.12437

12.

ECLAC, & ILO. (2019). Employment situation in Latin America and the Caribbean: Evolution of and prospects for women’s labour participation in Latin America (No. 21; Employment Situation in Latin America and the Caribbean). https://www.cepal.org/en/publications/44917-employment-situation-latin-america-and-caribbean-evolution-and-prospects-womens

13.

Esteve

García-Román

Lesthaeghe

(2012). The family context of cohabitation and single motherhood in Latin America. Population and Development Review, 38(4), 707–727. https://doi.org/10.1111/j.1728-4457.2012.00533.x

14.

Gennetian

L. A.

Knox

V. W.

(2003). Staying single: The effects of welfare reform policies on marriage and cohabitation (No. 13; The Next Generation Working Paper Series). MDRC. https://www.mdrc.org/publication/staying-single

15.

Haussler

I. M.

Marchant

(1980). TEPSI test de desarrollo psicomotor 2-5 años (TEPSI test of psychomotor development 2-5 years) (8th ed.). Ediciones Universidad Catolica.

16.

Hofferth

S. L.

(2006). Residential father family type and child well-being: Investment versus selection. Demography, 43(1), 53–77.

17.

INE. (2002). Anuario de estadísticas vitales 2000. Instituto Nacional de Estadísticas.

18.

INE. (2015). Mujeres en Chile y mercado del trabajo. Participación laboral femenina y brechas salariales. Instituto Nacional de Estadísticas.

19.

INE. (2017). Anuario de estadísticas vitales 2015. Instituto Nacional de Estadísticas.

20.

Institute for Family Studies, & Wheatley Institution. (2019). World family map 2019. Mapping family change and child well-being outcomes. https://ifstudies.org/ifs-admin/resources/reports/worldfamilymap-2019-051819.pdf

21.

Laplante

Castro-Martín

Cortina

Martín-García

. (2015). Childbearing within marriage and consensual union in Latin America, 1980–2010. Population and Development Review, 41(1), 85–108. https://doi.org/10.1111/j.1728-4457.2015.00027.x

22.

Lichter

D. T.

Graefe

D. R.

(2001). Finding a mater? The marital and cohabitation histories of unwed mothers. In Wu

L. L.

Wolfe

(Eds.), Out of wedlock: Causes and consequences of nonmarital fertility (pp. 317–343). Russell Sage Foundation.

23.

Litwin

Perova

Reynolds

S. A.

(2019). A conditional cash transfer and women’s empowerment: Does Bolsa Familia influence intimate partner violence? Social Science & Medicine, 238, 112462. https://doi.org/10.1016/j.socscimed.2019.112462

24.

Liu

Esteve

Treviño

(2017). Female-headed households and living conditions in Latin America. World Development, 90, 311–328. https://doi.org/10.1016/j.worlddev.2016.10.008

25.

Lundberg

Rose

(2003). Child gender and the transition to marriage. Demography, 40(2), 333–349. https://doi.org/10.2307/3180804

26.

Lyngstad

Jalovaara

(2010). A review of the antecedents of union dissolution. Demographic Research, 23(10), 257–292.

27.

Mariani

Özcan

Goisis

(2017). Family trajectories and well-being of children born to lone mothers in the UK. European Journal of Population, 33(2), 185–215. https://doi.org/10.1007/s10680-017-9420-x

28.

Moffitt

R. A.

(1998). The effect of welfare on marriage and fertility. In Moffitt

R. A.

(Ed.), Welfare, the family, and reproductive behavior. National Academy Press.

29.

Mutchler

J. E.

Baker

L. A.

(2009). The implications of grandparent coresidence for economic hardship among children in mother-only families. Journal of Family Issues, 30(11), 1576–1597. https://doi.org/10.1177/0192513X09340527

30.

Newborg

(2005). Battelle Developmental Inventory (2nd ed.). Riverside.

31.

OECD. (2018). Family Database. SF3.1: Marriage and divorce rates. http://www.oecd.org/els/family/SF_3_1_Marriage_and_divorce_rates.pdf

32.

OECD. (2019). OECD. Stat. https://stats.oecd.org/

33.

Posadas

Vidal-Fernandez

(2013). Grandparents’ childcare and female labor force participation. IZA Journal of Labor Policy, 2(14), 1–20.

34.

Ramm

(2016). Changing patterns of kinship: Cohabitation, patriarchy and social policy in Chile. Journal of Latin American Studies, 48(4), 769–796. https://doi.org/10.1017/S0022216X16000365

35.

Reichman

N. E.

Corman

Noonan

(2004). Effects of child health on parents’ relationship status. Demography, 41(3), 569–584. https://doi.org/10.1353/dem.2004.0026

36.

Reynolds

(2018). Sex of children and family structure in Brazil: Father and grandmother bias? Revista Brasileira De Estudos De População, 35(1), 1–29. https://doi.org/10.20947/s102-3098a0069

37.

Reynolds

Fernald

Lia Deardorff

Behrman

(2018). Family structure and child development in Chile: A longitudinal analysis of household transitions involving fathers and grandparents. Demographic Research, 38(58), 1777–1814.

38.

Salinas

(2011). Socioeconomic differences according to family arrangements in Chile. Population Research and Policy Review, 30(5), 677–699.

39.

Salinas

(2016). Changes in cohabitation after the birth of the first child in Chile. Population Research and Policy Review, 35(3), 351–375. https://doi.org/10.1007/s11113-015-9378-5