Paid paternity leave-taking in the United States

Abstract

Surprisingly few studies have focused on paternity leave-taking in the U.S. This study utilizes data from three national datasets to provide a comprehensive examination of the attitudes, practices, and predictors of paid paternity leave-taking in the U.S. Specifically, this study focuses on (a) describing attitudes towards fathers receiving a share of paid parental leave, (b) describing rates and lengths of paid paternity leave-taking, and (c) analyzing the extent to which economic capital, cultural capital, social capital, and father identities predict paternity leave-taking practices. The results indicate that most people support fathers receiving a share of paid parental leave in the U.S. Yet, rates of paid paternity leave-taking are relatively low and the majority of fathers who take paid leave take only one week or less. Economic capital, cultural capital, social capital, and father identities that prioritize engaged fathering are positively associated with taking paid leave and taking longer periods of leave. Overall, the results emphasize that the current structure of U.S. paternity leave policies seems to limit access to paid paternity leave and contribute to patterns of inequality due to more advantaged fathers having greater access and ability to take paid paternity leave than less advantaged fathers.

There has been an increased interest in paternity leave policies in recent years among policymakers, the media, and the general public. Rising interest in expanding paternity leave policies is due, in large part, to the potential benefits of such policies. For example, greater access to paternity leave may encourage men to become more invested in their families and engaged in their children’s lives (Huerta et al., 2014; Neponmyaschy & Waldfogel, 2007; Pragg & Knoester, 2017). Greater access to paternity leave may also help to alleviate rising levels of work-family conflict among fathers and help to reduce stress within families (Aumann, Galinsky, & Matos, 2011; Harrington, Van Deusen, & Humberd, 2011). Perhaps most notably, increased access to paternity leave may help to reduce gender inequality. By encouraging men to become invested in their child’s life from birth, fathers may gain parenting mastery and take some of the burden of childcare away from mothers (Rehel, 2014). As a result, mothers may experience reduced stress and be able to re-enter the labor force sooner (Gault et al., 2014; Johansson, 2010; Sejourne et al., 2012).

Despite the increased interest and potential benefits of paternity leave, there has been limited empirical research on paternity leave in the U.S. Results from the few studies on paternity leave suggest that most American fathers take a short period of leave when their child is born, and advantaged fathers (e.g., higher SES, white, access to more comprehensive leave policies) are more likely to take paternity leave and take longer periods of leave (Han, Ruhm, & Waldfogel, 2009; Harrington et al., 2014; Neponmyaschy & Waldfogel, 2007; Pragg & Knoester, 2017). While these studies have provided a glimpse into the patterns and predictors of paternity leave-taking in the U.S., research has yet to distinguish how the patterns and predictors of paid leave may differ from any (i.e., paid or unpaid) paternity leave. Such a focus is important as paid leave is relatively uncommon in the U.S., fathers are unlikely to take leave unless it is paid, and paid leave may provide more advantages to fathers and their families relative to unpaid leave (Harrington et al., 2014; Klerman, Daley, & Pozniak, 2012). Such knowledge will be helpful to scholars and policymakers who are interested in assessing whether the current structure of leave policies in the U.S. is adequate.

The current study attempts to fill this gap in the literature by utilizing three national datasets to provide a comprehensive examination of the attitudes, practices, and predictors of paid paternity leave-taking in the U.S. In particular, the purpose of this study is to: (a) present an overview of attitudes about paid paternity leave-taking in the U.S., (b) describe rates of paid paternity leave-taking, and (c) analyze the extent to which economic capital, cultural capital, social capital, and father identities predict the likelihood of taking paid paternity leave and lengths of paid leave-taking.

Background

Current policies on paternity leave, cultural and economic practices, and gendered structures influence attitudes toward paternity leave and create institutional constraints in regard to who is able to take paternity leave and for how long. For example, most countries have policies that provide paid parental leave to their citizens. Recent analyses show that 96% (178 of 186 countries studied) of countries provide paid maternity leave. In fact, all high-income countries have national policies mandating paid maternity leave except the United States (Heymann & McNeill, 2013). Although paid paternity leave policies are less widespread, 44% of countries have policies that allow fathers to receive some form of paid parental leave (i.e., leave shared by mothers and fathers), and the vast majority of these countries offer paid paternity leave (Heymann & McNeill, 2013; ILO, 2014).

The only U.S. policy that includes provisions for parental leave is the Family and Medical Leave Act (FMLA), which allows employees to take up to 12 weeks of unpaid leave after childbirth or for other eligible family or medical reasons (Cantor et al., 2001; Heymann & McNeill, 2013). Employees are elgible to take leave under FMLA if they work for an employer with 50 or more employees and if they have worked for the employer for at least 1,250 hours in the previous year (Han & Waldfogel, 2003; Melamed, 2014). Because of these restrictions, 40–50% of all U.S. employees are not eligible for leave under FMLA (Melamed, 2014; Winston, 2014).

In addition to FMLA, some U.S. states provide other leave benefits. Specifically, five states (California, Rhode Island, New Jersey, New York, and Hawaii) offer temporary disability insurance (TDI) with partial wage replacement to mothers due to their temporary “disability” from pregnancy and/or childbirth (Rossin-Slater, Ruhm, & Waldfogel, 2013; Winston, 2014). This benefit does not extend to fathers. Additionally, three states (California, Rhode Island, and New Jersey) expand on their TDI coverage to provide paid family leave with wage replacement to mothers and fathers who have a child (Winston, 2014). Similar policies will be implemented in New York in 2018, Washington in 2020, and Washington, D.C. in 2020. These policies vary in their level of wage replacement (ranging from 50–66%), amount of time offered (4–12 weeks for paid leave), and job protection (RI and NY offer job protection, CA and NJ do not) (Winston, 2014). There is evidence suggesting that these policies have led to dramatic increases in rates and length of maternity leave-taking and there is also some evidence suggesting that the California paid family leave policy has led to an increase in paternity leave-taking (Bartel et al., 2017; Baum & Ruhm, 2016; Milkman & Appelbaum, 2013; Rossin-Slater et al., 2013).

Due to the lack of a national paid leave policy, fathers’ access to leave is often dependent on whether their employer offers paternity leave. Decisions about employer-based work-family policies are often negotiated within the labor market, and the lack of a centralized collective bargaining system in the U.S. combined with the cultural ideal of a good employee being one that is always at work and invests long hours at work results in relatively few U.S. employers offering paid parental leave to their employees (Albiston & O’Connor, 2016; Berg et al., 2013). Specifically, only 13% of private workers and 14% of civilian workers have access to paid family leave (Desilver, 2017). One report suggests that only 9% of workers are employed at companies that offer paid paternity leave to all male employees (20% of workers work for companies that provide paid paternity leave to some male employees) (Klerman et al., 2012). As a result, many men may use other ways of taking time off (e.g., vacation or sick time, personal days, or unpaid leave) when they have a child (Harrington et al., 2014). In contrast, 35% of workers have access to paid maternity leave at their workplace (Klerman et al., 2012).

Conceptual framework

In the midst of this milieu, our conceptual framework draws on understandings of gender as social structure, the implications of socioeconomic inequalities, “diverging destinies,” Bourdieu’s forms of capital (1986), and the relevance of father identities. We consider how these concepts shape attitudes, opportunities for, and practices of paid paternity leave-taking.

First, we recognize that gender operates as structure on an individual, interactional, and institutional level (Risman, 1998; 2004). At the individual level, socialization processes encourage males to value the breadwinner role and females to value the caregiver role. As such, fathers may be less likely to want to take paternity leave (and mothers may prefer this arrangement as well). At the interactional level, gendered meanings and cultural practices may lead individuals to expect mothers to take leave from paid labor but not fathers, and stigmatize fathers who want to do the same. At the institutional level, the organization and practices of institutions are gendered; workplaces often follow the ideal worker norm, which assumes that paid work is the only (or primary) responsibility of employees (Acker, 1990; Williams, 2000). Yet, we know that women are disproportionately responsible for domestic labor and men are disproportionately responsible for paid labor (Hochschild, 1989; Milkman & Appelbaum, 2013).. Thus, U.S. institutions rarely develop or prioritize family leave policies, provide for the needs of working parents, or encourage paternity leave-taking.

Second, we emphasize socioeconomic inequalities, acknowledging the trends of “diverging destinies” and how access to various forms of capital maintains and exacerbates these inequalities (Bourdieu, 1986; McLanahan, 2004). Bourdieu (1986) argues that socioeconomic inequalities are rooted in the accumulation of economic capital (i.e., monetary resources), and are maintained and reproduced through the possession of cultural capital (i.e., knowledge of the dominant culture) and social capital (i.e., social networks that help to maintain advantage). Furthermore, the inequalities that result from differing levels of capital have increased over time (McLanahan, 2004), likely resulting in markedly more options for family leave-taking for advantaged families. Moreover, the ability to take leave may further contribute to “diverging destinies” among families with unequal access to various forms of capital.

Finally, we consider the influence of father identities. Father identities are the culmination of the meanings and importance of fathering roles that are developed through social interactions, cultural messages, and fathers’ participation in their children’s lives. The salience, commitments, and nature of these identities reflect and shape subsequent attitudes and behaviors (Pasley, Petren, & Fish, 2014; Pragg & Knoester, 2015; Stryker, 1968). New fatherhood ideals are increasingly urging fathers to be more nurturing and engaged in their children’s lives, and men seem to be increasingly developing father identities that prioritize involvement with children (Marsiglio & Roy, 2012; Pasley et al., 2014). Such identities may encourage men to support the establishment of paternity leave opportunities and utilize such opportunities when available.

U.S. attitudes about paternity leave

Attitudes about paternity leave are important because they indicate the extent to which support for paternity leave may exist and have the potential to influence cultural practices (Cotter, Hermsen, & Vanneman, 2011). Attitudes about paternity leave may be influenced by cultural changes which expect fathers to become more involved in their children’s lives yet still meet the demands of traditional fatherhood (Aumann et al., 2011; Marsiglio & Roy, 2012). In balancing these competing demands, individuals who embrace “new fatherhood” ideals may be more likely to support paid paternity leave than individuals who embrace traditional gender ideologies (Lammi-Taskula, 2008).

Research on attitudes toward paternity leave seems to reflect these competing demands. For example, one survey of relatively advantaged fathers indicates that a large majority of fathers believe that companies should offer paid paternity leave, with most fathers advocating for between 2–4 weeks of paid leave. The vast majority of fathers also stated that they would not take leave unless most of their salary was compensated (Harrington et al., 2014).

These attitudes are consistent with international patterns of paternity leave-taking, as fathers are more likely to take leave when it is paid and targeted towards them (O’Brien, 2009). Even in countries with statutory paid leave policies, many fathers do not take leave (or do not take the full leave period allowed) due to the financial and social pressures to fulfill provider roles (Brandth & Kvande, 2016; Duvander, 2014; Malin, 1994).

As such, gender as social structure may lead men to prioritize breadwinning, believing that paternity leave-taking is inappropriate—or at least less appropriate than maternity leave-taking. In fact, the social structure of gender may operate to disproportionately encourage women to endorse paternity leave-taking. This may reflect the dissatisfaction that many women have over taking on a “second shift” of domestic responsibilities (Hochschild, 1989).

Paternity leave-taking

Beyond the institutional constraints imposed by U.S. policies, gendered attitudes, socioeconomic inequalities, and father identities are expected to influence paternity leave-taking practices. Indeed, previous research has examined paternity leave-taking practices, but has not distinguished between paid and unpaid paternity leave. Consistent with larger patterns of inequality (Bourdieu, 1986; McLanahan, 2004), fathers with greater levels of economic, cultural, and social capital may be especially more likely to have access to paid paternity leave, and be better able to take leave, than fathers who have less capital. Father identities may also shape the willingness of men to take paid paternity leave.

The lack of paid paternity leave policies and gendered structure of society might suggest that relatively few U.S. fathers take paternity leave. Yet, previous studies suggest that most fathers take short periods of paternity leave (Harrington et al., 2011; Nepomnyaschy & Waldfogel, 2007; Pragg & Knoester, 2017). However, only a small minority of fathers have access to paid paternity leave, suggesting that many fathers may be using other forms of leave such as vacation time when they have a child (Klerman et al., 2012; SHRM, 2015). Thus, given the structural barriers to paid paternity leave-taking, rates of paid leave-taking are likely to be low in the U.S.

Nonetheless, we anticipate that Bourdieu’s (1986) forms of capital influence leave-taking practices. Economic capital may strongly influence fathers’ access to, and likelihood of taking, paternity leave. Research consistently shows that fathers with greater economic capital are more likely to take, and take longer periods of, paternity leave (Brandth & Kvande, 2002; Huerta et al., 2014; Nepomnyaschy & Waldfogel, 2007). Specifically, fathers with higher incomes may be more able to afford leave than low-income fathers (especially if wage replacement is less than 100%), but are also more likely to have access to paid leave from their employers (Klerman et al., 2012; SHRM, 2015; Winston, 2014).

In addition, cultural capital may influence fathers’ access to, and willingness to take, paternity leave. For example, education may be linked to access to leave, but also may provide men with the knowledge needed to navigate the fragmented structure of paternity leave in the U.S. Specifically, more educated fathers may better understand the available leave options and be better able to take advantage of such policies (Klerman et al., 2012; Winston, 2014).

Occupations may also be a source of cultural capital that influences patterns of leave-taking. This is particularly important because leave-taking is often stigmatized; the expectation that fathers should be the main financial provider in a family may lead fathers who take leave to feel that they are not fulfilling their responsibilities as fathers (Malin, 1994; Marsiglio & Roy, 2012). Indeed, many men fear that taking leave will damage their career (Brandth & Kuvande, 2002), and numerous studies suggest that men encounter a flexibility stigma, which is the bias against workers who take caregiving leave (Bornstein, 2013). Men who request leave encounter resistance to their requests, receive lower performance ratings, are more likely to be viewed as poor workers, and are viewed as less masculine (Coltrane et al., 2013; Rudman & Mescher, 2013). Leave-taking may also negatively impact men’s future earnings (Rege & Solli, 2013).

The likelihood of encountering the flexibility stigma is partially dependent on cultural capital, as professional and more-educated workers are more likely to be trusted and are less likely to have rigid schedules. In contrast, low-wage and less-educated workers are less likely to have flexibility, and more likely to be penalized when requesting leave (Williams, Blair-Loy, & Berdahl, 2013). Thus, fathers with greater cultural capital may be more likely to take leave than fathers with less cultural capital (Bygren & Duvander, 2006; Dahl, Loken, & Mogstad, 2014).

Fathers’ social capital may also structure patterns of leave-taking. For example, white fathers may be more likely to work for employers that offer paid leave and less likely to be penalized for taking leave due to their advantage within society. Indeed, white fathers are more likely to take paternity leave, and longer periods of leave, than black or Hispanic fathers (Huerta et al., 2014; Nepomnyaschy & Waldfogel, 2007; Pragg & Knoester, 2017). Black fathers are also more likely to be penalized for requesting leave than white fathers (Rudman & Mescher, 2013).

In addition, marriage may provide an advantage, relative to other family structures, that increases the likelihood and length of leave-taking. Specifically, fathers who demonstrate a greater commitment to birth mothers through marriage may be more likely to have salient father identities (Pasley et al., 2014; Townsend, 2002). This commitment works as a form of social capital that provides married residential fathers with a wage premium that is not experienced by unmarried fathers (Killewald, 2013). The social capital advantage that married fathers hold may allow them to take paid leave, and longer periods of leave, than cohabiting and nonresident fathers (Huerta et al., 2014; Pragg & Knoester, 2017).

Paternity leave-taking may also be dependent on the salience of, and commitments to, father identities. That is, viewing fatherhood as a key aspect of one’s identity and being committed to enacting fathering roles may increase the likelihood that fathers take paternity leave as one way of fulfilling fathering expectations (Pasley et al., 2014; Stryker, 1968). For example, fathers who have more positive attitudes about fathering may be more likely to take leave, and longer periods of leave, than men who are less interested in fathering or who emphasize the provider role (Duvander, 2014; Pragg & Knoester, 2017; Romero-Balsas, 2012).

In addition, father identities may be more salient for men making their initial transition to fatherhood, which may lead new fathers to be more likely to take paternity leave, and take longer periods of leave, than men with other children (Harrington et al., 2014; Nepomnyaschy & Waldfogel, 2007). A more salient father identity may also lead fathers to engage in fathering behaviors prior to the birth of a child, and prenatal involvement increases the likelihood of leave-taking (Huerta et al., 2014). Religious commitments may also motivate fathers to take paternity leave due to the religious emphasis on the importance of family life (Petts, 2007).

Hypotheses

• Hypothesis 1: Attitudes about paternity leave-taking will indicate substantial support for paid paternity leave opportunities. Female and more egalitarian adults are expected to be especially likely to support paid paternity leave opportunities.

• Hypothesis 2: Fewer fathers will take paid paternity leave, as compared to the number of fathers who endorse them.

• Hypothesis 3: Fathers with higher levels of economic, cultural, and social capital, and fathers with especially salient and committed father identities will be more likely to take paid paternity leave than less advantaged fathers and fathers with less salient father identities.

• Hypothesis 4: Fathers with higher levels of economic, cultural, and social capital, and fathers with more salient and committed father identities will take longer periods of paid paternity leave than less advantaged fathers and fathers with less salient father identities.

Data and Methods

Sample

Data is taken from three national studies. The 2012 General Social Survey (GSS) contains reports on attitudes about family (and paternity) leave. In addition, two national longitudinal studies offer information on paid paternity leave-taking: the Fragile Families and Child Wellbeing Study (FFCW) and the National Longitudinal Study of Youth 1997 (NLSY97). Overall, the complementary nature and uniqueness of each dataset (i.e., attitudes about paid paternity leave from a nationally representative sample in the GSS, and leave-taking behaviors from a relatively disadvantaged group of fathers in the FFCW and a relatively young group of fathers in the NLSY97) allows for a comprehensive assessment of attitudes and behaviors about paid paternity leave in the U.S.

The GSS is an annual (or biennial) nationally representative survey of adult Americans. The 2012 GSS (N = 1,974) contained a special module on family and changing gender roles, which included questions on attitudes towards family leave. The FFCW is a longitudinal birth cohort study that follows 4,898 children born between 1998 and 2000 and their parents. Fragile families are defined as unmarried parents and their children, and these data focus on an urban sample from 20 large cities that contains high percentages of low-income, minority, and unmarried parents. Parents were interviewed at the hospital shortly after birth (W1) and when children were approximately one (W2), three (W3), five (W4) and nine years old (W5). Data from W1 and W2 of the FFCW were used. The NLSY97 contains a nationally representative sample of approximately 9,000 youths who were 12–16 years old as of December 31, 1996. Youths were first interviewed in 1997, and have been re-interviewed 15 times through 2013. All available waves of data from the NLSY97 were used.

To analyze attitudes toward paternity leave, the sample was restricted to the random subset of respondents who were asked questions about paid family leave and provided valid responses to these items (N = 1,099).¹ To analyze paternity leave-taking behaviors, the samples were restricted to fathers who were employed at the time of their child’s birth (to be eligible to take paternity leave) and who answered the questions about paternity leave-taking. Resident single fathers were also excluded due to insufficient cases (N < 15), and the NLSY97 was further restricted to focus on only one focal child for each father (fathers were asked questions about leave for any child born, but the sample is restricted to only the first instance in which fathers meet all of the eligibility criteria). These restrictions result in final sample sizes of 2,154 fathers in the FFCW and 1,720 in the NLSY97.

Attitudes about paternity leave

In the 2012 GSS, respondents are presented with a scenario involving a couple who both work full-time, have a newborn child, and one of the parents stops working to care for the child. Respondents are then asked if they think paid leave should be available for this couple and how leave should be divided between the mother and father. Support for fathers receiving a share of paid parental leave indicates respondents who indicate that: (a) paid leave should be available for this couple and (b) the father should take at least some of the paid leave (1 = yes).

Paid paternity leave-taking

In the NLSY97, respondents were asked at each wave whether they took any paid leave from work since the previous wave because of a pregnancy or birth of a child and how long the period of leave was. Respondents were specifically asked whether there were any periods in which they took a full week or more of paid leave, but a sizeable number of fathers (20%) reported taking less than a week of leave (based on the dates in which leave began and ended). As such, some short leaves are likely not reported in the NLSY97 (Baum & Ruhm, 2016). In the FFCW, fathers reported on whether they took any time off of work due to the birth of the focal child, how many weeks of leave they took, and how many paid weeks they received while on leave in the W2 survey.²

Information from these responses was used to construct two indicators. Paid paternity leave-taking is a dichotomous variable indicating whether or not fathers took any paid paternity leave (1 = yes). Length of paid paternity leave indicates whether fathers took (a) no leave (used as reference category), (b) one week or less of paternity leave, (c) more than one, but no more than two weeks of paternity leave, or (d) more than two weeks of paternity leave.

Predictor variables

All predictor variables are taken from the 2012 GSS, W1 in the FFCW, and the survey prior to the child’s birth in the NLSY97.

Economic capital

Measures of economic capital include household income (1 = less than $5,000 to 7 = more than $250,000)³ and whether the birth mother (or a spouse/partner in the GSS) is employed (1 = yes). Hours worked is coded as (a) part-time (less than 35 hours a week), (b) full-time (between 35 and 54 hours a week, used as reference category), and (c) overtime (55 hours a week or more).

Cultural capital

Education ranges from 1 = did not complete high school to 4 = college degree. Occupation was categorized as (a) professional, (b) labor (used as reference category), (c) sales, (d), service, or (e) other. An additional indicator was included in the GSS for fathers who were not working (used as the reference category for GSS analyses).⁴

Social capital

Race/ethnicity is coded as (a) White (used as reference category), (b) Black, (c) Latino, or (d) other race/ethnicity.⁵ Relationship status in analyses predicting paternity leave practices is coded as (a) married to the birth mother (used as reference category), (b) cohabiting with the birth mother, or (c) nonresident father.⁶ Corresponding measures in the GSS (which are not in reference to the birth mother) are (a) married, (b) cohabiting, and (c) single (used as reference category).

Father identities

In analyses using the GSS and FFCW, we incorporate two variables that indicate father attitudes. Unfortunately, indicators of fathering attitudes are unavailable in the NLSY97. Traditional father attitudes is a dichotomous variable indicating whether one agrees (1 = yes) that it is more important for fathers to focus on providing while mothers care for the family. New fatherhood attitudes is a dichotomous variable indicating whether one agrees (1 = yes) that it is important for fathers to be involved and spend time with his family. In the GSS, this is indicated by respondents’ reactions to the statement that family life suffers when men work too much. Two additional variables are used in analyses using the FFCW. Positive father attitudes is represented by fathers’ level of agreement (1 = strongly disagree to 4 = strongly agree) to three statements (being a father is one of the most fulfilling experiences for a man, I want people to know that I have a new child, and not being a part of child’s life would be one of the worst things), and the mean is used (α = .72). Prenatal involvement indicates whether fathers (a) gave the birth mother money or bought things for the baby and (b) helped in other ways like providing transportation or doing chores (1 = yes) prior to the birth of the child. Additionally, whether the father is a first time father (1 = yes) is included in all analyses predicting paternity leave practices. Finally, religious participation (0 = never to 4 = at least once a week) is included in all analyses.

Other controls

Father’s age is coded in years. Child’s gender (1 = male) is included in analyses predicting paternity leave practices. Additional controls from the GSS to predict attitudes about paternity leave include gender (1 = male), number of children, number of resident children, and having at least one resident male child (1 = yes).

Analytic strategy

We first analyze descriptive statistics to ascertain patterns in attitudes and practices surrounding paid paternity leave. This allows for an initial consideration of hypotheses 1 and 2. To test the remaining hypotheses, we use logistic and ordinary least squares (OLS) regression models to predict attitudes about paid paternity leave and leave-taking practices. Specifically, we first use logistic regression models to further test Hypothesis 1 and assess whether the three types of capital and father identities predict support for fathers receiving a share of paid parental leave using the full GSS sample and then a subsample of men only (since only men can take paternity leave). To test the third hypothesis, we use logistic regression models to predict the likelihood of taking paid paternity leave. To test the fourth hypothesis, we use OLS regression models to predict length of paid paternity leave.⁷ In all analyses, multiple imputation is used to account for missing data. Less than 5% of cases have missing data on variables of interest (with the exception of father’s income in the FFCW, in which 17% of cases are missing). Missing values were imputed using all variables included in the analyses. Combined results from ten imputed models were used.

Results

Mean values for all variables are included in Table 1. Consistent with the first hypothesis, results show modest support for fathers receiving a share of paid parental leave: 54% of all GSS respondents (including 47% of male respondents) favor fathers receiving a share of paid parental leave (in contrast, 80% of all GSS respondents in our sample believe that paid leave should be available to the couple, results not shown). Nonetheless, consistent with the second hypothesis, rates of paid paternity leave-taking are comparatively low. Only 24% of fathers take paid paternity leave in the NLSY97 and 42% take paid leave in the FFCW.

Table 1.

Mean Values for all Variables

	GSS	NLSY97	FFCW
Paid Leave-Taking
Support for fathers receiving a share of paid parental leave	0.54	-	-
Father took paid paternity leave	-	0.24	0.42
Length of Paid Paternity Leave
No Paid Leave*	-	0.76	0.58
One week or less	-	0.09	0.27
1–2 weeks	-	0.09	0.11
More than 2 weeks	-	0.06	0.04
Economic Capital
Income	3.30	3.57	6.09
Works Part Time	0.14	0.20	0.11
Works Full Time*	0.36	0.52	0.71
Works Overtime	0.11	0.28	0.18
Not Working	0.39	-	-
Employed Mother/Partner	0.35	0.70	0.65
Cultural Capital
Education	2.68	2.19	2.31
Occupation: Professional	0.23	0.18	0.17
Occupation: Labor*	0.11	0.47	0.49
Occupation: Sales	0.06	0.08	0.08
Occupation: Service	0.21	0.25	0.24
Occupation: Other	0.02	0.02	0.02
Occupation: None	0.38	-	-
Social Capital
White*	0.66	0.51	0.27
Black	0.15	0.22	0.42
Latino	0.14	0.27	0.26
Other Race	0.05	-	0.05
Married*	0.45	0.56	0.34
Cohabiting	0.10	0.44	0.40
Singlea	0.45	-	-
Nonresident Father	-	-	0.26
Father Identities
Positive Father Attitudes	-	-	3.76
Traditional Gender Attitudes	0.23	-	0.39
New Fatherhood Attitudes	0.54	-	0.77
Prenatal Involvement	-	-	0.93
First Time Father	-	0.72	0.42
Religious Participation	0.91	1.94	1.91
Other Controls
Age	46.98	25.77	28.26
Male	0.46	-	-
Male Child	0.21	0.52	0.52
Number of Children	1.82	-	-
Number of Resident Children	0.56	-	-
N	1099	1720	2154

^*Used as reference category;

^adifferent reference category used in GSS analyses.

The difference in leave-taking between the two samples is likely attributable to a few factors. First, because the initial question about leave-taking in the NLSY97 focused on leave periods lasting a week or more, short leaves are likely underreported. Second, the NLSY97 sample is significantly younger (fathers are 18–34 years old, with a mean age of 25.77) than the FFCW sample (fathers are 18–57 years old, with a mean age of 28.26). Younger fathers may be less secure in their careers, and have less access to (or are less willing to take) leave than older fathers. Third, there is some evidence from previous research than urban fathers take longer periods of leave than fathers who do not live in urban areas (Neponmyaschy & Waldfogel, 2007). Given that the FFCW is an entirely urban sample, this may contribute to higher rates of leave-taking and longer leaves compared to the NLSY97 sample.

Among fathers who take paid leave, most take short durations—typically, one week or less. Thus, very few fathers – less than 5% of all fathers (and 14% of fathers who take leave) – take more than even two weeks of paternity leave. Specifically, 64% of the fathers who take leave in the FFCW take one week or less. In the NLSY97, nearly 38% of the fathers who take paid leave take one week or less.

Attitudes about fathers receiving a share of paid parental leave

Results in Table 2 provide additional support for the first hypothesis in showing that female and more egalitarian adults are more likely to support fathers receiving a share of paid parental leave. Specifically, men are 43% less likely to support paid paternity leave opportunities than women (Exp B = .57, p < .001), and individuals who support traditional gender ideologies are 31% less likely to support fathers receiving a share of paid parental leave compared to individuals with more egalitarian gender ideologies (Exp B = .69, p < .05). In addition, we find that education is positively related to support for fathers receiving a share of paid parental leave (Exp B = 1.19, p < .05). There is also some evidence that age is negatively associated with support for fathers receiving a share of paid parental leave (Exp B = .98, p < .001).

Table 2.

Results from Logistic Regression Models Predicting Support for Fathers Receiving a Share of Paid Parental Leave

	Full Sample (GSS)		Men Only (GSS)
Variable	Exp B	SE B	Exp B	SE B
Economic Capital
Income	1.05	0.06	1.04	0.08
Works Full Time	1.07	0.22	1.03	0.32
Works Overtime	0.80	0.21	0.62	0.23
Significant Other Employed	1.23	0.27	1.31	0.39
Cultural Capital
Education	1.19	0.09*	1.26	0.14*
Occupation: Professional	1.00	0.24	1.08	0.40
Occupation: Labor	0.85	0.24	1.06	0.40
Occupation: Service	1.13	0.26	1.66	0.59
Occupation: Sales	0.72	0.23	1.11	0.56
Occupation: Other	2.13	1.13	2.33	1.53
Social Capital
Black	1.15	0.22	1.12	0.35
Latino	0.97	0.19	0.98	0.27
Other Race	1.22	0.37	0.86	0.34
Married	0.75	0.15	0.57	0.17
Cohabiting	0.83	0.23	0.51	0.19
Father Identities
Traditional Gender Attitudes	0.69	0.11*	0.88	0.21
New Fatherhood Attitudes	1.11	0.15	1.24	0.25
Religious Participation	0.99	0.07	1.13	0.12
Other Controls
Age	0.98	0.00***	0.99	0.01
Male	0.57	0.08***	-	-
Number of Children	1.03	0.05	1.04	0.07
Number of Resident Children	1.07	0.12	1.07	0.21
Having a Male Child	0.97	0.25	1.42	0.61
R2	0.06		0.06

Note: Results Reported as Odds Ratios.

^*p < .05;

^**p < .01;

^***p < .001

Paid paternity leave-taking

Next, we analyze whether the forms of capital and father identities are associated with taking paid paternity leave (Hypothesis 3). Results are presented in Table 3. First, we find relatively strong and consistent evidence that economic capital is positively associated with paid paternity leave-taking, as hypothesized. Specifically, fathers with higher incomes are more likely to take paid leave in both the NLSY97 (Exp B = 1.63, p < .001) and FFCW (Exp B = 1.20, p < .002). Moreover, compared to full-time workers, part-time workers are less than half as likely to take paid leave in the FFCW (Exp B = .42, p < .001).

Table 3.

Results from Logistic Regression Models Predicting Paid Paternity Leave

	NLSY97		FFCW
Variable	Exp B	SE B	Exp B	SE B
Economic Capital
Income	1.63	0.11***	1.20	0.04***
Father Worked Part Time	1.01	0.17	0.42	0.22***
Father Worked Overtime	0.96	0.15	0.80	0.17
Mother Employed	0.85	0.14	0.77	0.14
Cultural Capital
Education	1.14	0.10	1.28	0.08**
Occupation: Professional	1.87	0.36**	1.29	0.22
Occupation: Sales	1.90	0.44**	0.94	0.25
Occupation: Service	1.66	0.28**	1.35	0.16
Occupation: Other	1.32	0.56	1.06	0.51
Social Capital
Black	0.72	0.14	0.47	0.19***
Latino	1.37	0.21*	0.71	0.21
Other Race	-	-	0.35	0.33**
Cohabiting with Birth Mother	0.66	0.10**	0.58	0.18**
Nonresident Father	-	-	0.32	0.20***
Father Identities
Positive Father Attitudes	-	-	1.21	0.16
Traditional Gender Attitudes	-	-	0.97	0.14
New Fatherhood Attitudes	-	-	1.08	0.16
Prenatal Involvement	-	-	3.57	0.28***
First Time Father	1.77	0.30**	1.44	0.14*
Religious Participation	1.04	0.04	1.00	0.05
Other Controls
Father Age	1.06	0.02*	0.99	0.01
Child is Male	1.01	0.13	1.00	0.13
R2	0.18		0.05

Note: Results Reported as Odds Ratios.

^*p < .05;

^**p < .01;

^***p < .001

Consistent with the third hypothesis, cultural capital appears to increase the likelihood of taking paid leave. Specifically, education is positively associated with taking paid leave (Exp B = 1.28, p < .01) in the FFCW. Also, in the NLSY97, fathers who are professionals (Exp B = 1.87, p < .01), in sales (Exp B = 1.90, p < .01), or in a service occupation (Exp B = 1.66, p < .01) are markedly more likely to take paid paternity leave than fathers who are laborers.

Results in Table 3 also provide evidence that social capital is associated with the likelihood of paid leave-taking. Fathers who identify as black (Exp B = .47, p < .001) or other race/ethnicity (Exp B = .35, p < .01) are less likely to take paid leave than white fathers in the FFCW, as expected. Unexpectedly, there is evidence that Latinos are more likely to take paid paternity leave compared to whites in the NLSY97 (Exp B = 1.37, p < .05). As one may anticipate, however, this is only the case after taking into account differences in economic and cultural capital. In addition, cohabiting fathers are less likely to take paid leave than married fathers in the NLSY97 (Exp B = .66, p < .01). Similarly, cohabiting (Exp B = .58, p < .01) and nonresident fathers (Exp B = .32, p < .001) are less likely to take paid leave than married fathers in the FFCW.

We also find evidence that father identities also matter in predicting paid leave-taking. First-time fathers are more likely to take paid leave than fathers who already have a child in the NLSY97 (Exp B = 1.77, p < .01) and FFCW (Exp B = 1.44, p < .05). Fathers who are involved in preparing for their new child are also more likely to take paid leave (Exp B = 3.57, p < .001). Overall, consistent with the third hypothesis, each type of capital (economic, cultural, and social) and father identities all appear to influence the likelihood that fathers take paid paternity leave.

Length of paid paternity leave

Finally, we analyze whether the forms of capital and father identities are associated with length of paid paternity leave (Hypothesis 4). Results are shown in Table 4. Consistent with our hypothesis, economic capital is associated with length of paternity leave. Specifically, income is positively associated with length of paid paternity leave in the NLSY97 (B = .10, p < .001) and FFCW (B = .05, p < .001). In addition, fathers who work part-time take shorter leaves than full-time workers in the FFCW (B = −.20, p < .001).

Table 4.

Results from OLS Regression Models Predicting Length of Paid Paternity Leave

	NLSY97		FFCW
Variable	B	SE B	B	SE B
Economic Capital
Income	0.10	0.02***	0.05	0.01***
Father Worked Part Time	−0.02	0.05	−0.20	0.05***
Father Worked Overtime	−0.01	0.05	−0.07	0.04
Mother Employed	−0.02	0.05	−0.05	0.03
Cultural Capital
Education	0.08	0.03**	0.10	0.02***
Occupation: Professional	0.21	0.06**	0.24	0.05***
Occupation: Sales	0.17	0.08*	−0.05	0.06
Occupation: Service	0.18	0.05***	0.16	0.04***
Occupation: Other	0.06	0.14	0.23	0.13
Social Capital
Black	−0.08	0.05	−0.05	0.05
Latino	0.12	0.05*	0.04	0.05
Other Race	-	-	−0.12	0.08
Cohabiting with Birth Mother	−0.11	0.05*	−0.19	0.05***
Nonresident Father	-	-	−0.29	0.05***
Father Identities
Positive Father Attitudes	-	-	0.09	0.04*
Traditional Gender Attitudes	-	-	−0.05	0.04
New Fatherhood Attitudes	-	-	0.06	0.04
Prenatal Involvement	-	-	0.17	0.07*
First Time Father	0.15	0.05**	0.08	0.04*
Religious Participation	0.00	0.01	0.00	0.01
Other Controls
Father Age	0.02	0.01***	−0.00	0.00
Child is Male	−0.01	0.04	−0.01	0.03
R2	0.14		0.16

^*p < .05;

^**p < .01;

^***p < .001

Consistent with the fourth hypothesis, we also find that cultural capital is associated with length of paternity leave. Specifically, education is positively associated with length of paid paternity leave in the NLSY97 (B = .08, p < .01) and FFCW (B = .10, p < .001). Also, professionals (B = .21, p < .01), sales workers (B = .17, p < .05), and service workers (B = .18, p < .001) are more likely to take longer periods of paid leave than laborers in the NLSY97. Similarly, professionals (B = .24, p < .001) and service workers (B = .16, p < .001) take longer periods of paid leave than laborers in the FFCW.

Also consistent with the fourth hypothesis, some social capital indicators are associated with length of paternity leave. Specifically, cohabiting fathers take shorter paid leaves in the NLSY97 (B = −.11, p < .05), and FFCW (B = −.19, p < .001) than married fathers. Nonresident fathers (B = −.29, p < .001) also take shorter periods of paid leave than married fathers. However, similar to the estimates of predicting any paid leave-taking, Latino fathers (B = .12, p < .05) take longer paid paternity leaves than white fathers in the NLSY97.

Also consistent with the fourth hypothesis, father identity indicators are associated with length of paternity leave. In particular, first-time fathers take longer paid leaves in the NLSY97 (B = .15, p < .01) and FFCW (B = .08, p < .05) than fathers who already have a child. Moreover, positive father attitudes (B = .09, p < .05) and prenatal involvement (B = .17, p < .05) are positively associated with length of paid leave-taking in the FFCW.

Discussion

Despite increased interest in creating more extensive family leave policies, there is little research on the attitudes and practices regarding paid paternity leave in the U.S. Paternity leave is especially important to study because it represents a version of family leave-taking that is rare in comparison to maternity leave, yet it offers substantial possibilities for alleviating work-family conflicts and encouraging increased father involvement, mothers’ well-being, and gender equity in the divisions of domestic and paid labor responsibilities (Gault et al., 2014; Pragg & Knoester, 2017). In the present study, we utilize three national data sets to provide a comprehensive examination of the attitudes, patterns, and predictors of paid paternity leave-taking in the U.S. As such, this study expands prior research on paternity leave as well as confirms previous findings.

In regard to attitudes towards paid leave, we find support for fathers receiving a share of paid parental leave, as expected. In the GSS, over half of all respondents favored the creation of paid parental leave in which fathers receive at least some paid leave. In supplementary analyses, we found that women under the age of 40 were especially likely to support fathers receiving a share of paid parental leave (69% in favor), yet a slight majority of men in this age range were also supportive (54% in favor). Although most respondents support paid paternity leave, support for leave was less widespread in the GSS than in previous studies (Harrington et al., 2014). This discrepancy may be due to the wording of the questions on leave in the GSS. When respondents are asked whether they support paternity leave, the vast majority of men (89%) say they do (Harrington et al., 2014). However, when asked how to divide up parental leave between mothers and fathers, far fewer men (47%) believe fathers should take leave (Harrington et al. 2014). Although men are supportive of leave, they appear to be less supportive if this comes at the expense of mothers’ leave time. Such beliefs reflect the gendered structure of U.S. society that emphasizes breadwinning roles for men and caretaking roles for women (Risman, 1998; 2004).

We also find that fewer fathers take paid leave when compared to those that support paid leave-taking opportunities. Although previous research suggests that the vast majority of employed fathers take paternity leave when they have a new child (Harrington et al., 2011; Nepomnyaschy & Waldfogel, 2007; Pragg & Knoester, 2017), results from this study suggests that far fewer fathers take paid paternity leave. In addition, when fathers do take paternity leave, these leaves are likely to be short in duration. Less than 5% of fathers take more than two weeks of paid leave; among fathers who take leave, the majority (56%) takes one week or less. These findings echo results from other studies that focus on leave more generally (Harrington et al., 2011; Nepomnyaschy & Waldfogel, 2007; Pragg & Knoester, 2017). In contrast, in countries that offer statutory paid leave, the vast majority of fathers take at least some paid leave and many take substantially longer leaves (averaging 2–3 weeks or longer in the Nordic region) (Akerstrom, 2015; Lero, 2015; Schulze & Gergoric, 2015; van Belle, 2016).

The low rates of paid paternity leave-taking, as we anticipated, certainly seem to reflect the lack of paid paternity leave options that exist in U.S. workplaces (Klerman et al., 2012; SHRM, 2015). Thus, the institutional structure of gender may block access to paid leave for most men. Moreover, despite the rise in the new fatherhood ideal, there are often professional and relationship challenges that emerge if men do not meet breadwinning expectations (Coltrane et al., 2013; Marsiglio & Roy, 2012). Thus, the gendered structure of society at the interactional and individual levels may deter fathers from taking paternity leave. Furthermore, even among fathers who can take paid paternity leave, most men may choose to take short periods of leave to maintain a balance between spending at least some time with their new child while avoiding extended time away from work to minimize the risk of stigmas and other penalties.

We also find support for our third and fourth hypotheses that economic capital, cultural capital, social capital, and father identities lead to different patterns of paid paternity leave-taking. Economic capital consistently predicts paid paternity leave-taking. Higher levels of income are associated with an increased likelihood of taking paid leave and longer periods of leave. Similarly, fathers who work part-time are less likely to take paid leave and take shorter leaves than full-time workers. Fathers with high incomes are more likely to be in a financial position that allows them to take time off work when a new child is born due to greater economic security. This privileged position may suggest that the opportunity cost of taking leave is higher for fathers with high incomes than for fathers with lower incomes (especially if wage replacement is less than 100%). Yet, higher levels of economic capital may also be linked to greater benefits and autonomy in one’s career (which may include being employed in workplaces that offer more generous leave options), allowing these fathers to avoid stigmas associated with taking leave (Williams et al., 2013). In contrast, fathers who work part-time may not have the option to take paid leave, or may be more likely to be penalized at work by taking leave (Williams et al., 2013).

Cultural capital also predicts patterns of leave-taking. Higher education is associated with an increased likelihood of taking paid paternity leave and longer periods of leave. Given that leave options in the U.S. vary by workplace, many individuals may be unaware of the leave options available to them or how to utilize such options (Klerman et al., 2012). More educated fathers may be better able to take advantage of available leave policies. Relatedly, fathers in professional occupations are more likely to work for companies that offer paid leave (Bygren & Duvander, 2006; Harrington et al., 2014; Nepomnyaschy & Waldfogel, 2007). In contrast, fathers who work in labor occupations may not have the option to take paid leave or may be stigmatized if they request leave (Williams et al., 2013). Indeed, we find that laborers are especially less likely to take paid leave (and longer periods of leave) than other workers.

Results from this study also indicate that family structure appears to be a key source of social capital that influences patterns of paid leave-taking. In particular, married fathers are consistently more likely to take paid paternity leave, and longer periods of leave, than cohabiting and nonresident fathers. Marriage may require greater commitments from fathers, leading them to be more likely to invest in their family life in the form of paternity leave-taking (Pragg & Knoester, 2017). This finding may also reflect evidence of a fatherhood premium; research suggests that married, resident, and biological fathers experience a wage premium (Killewald, 2013). This premium may extend to greater benefits and flexibility for married fathers.

However, in contrast to our hypothesis, racial/ethnic differences in patterns of paid leave-taking are mixed. In the FFCW, black fathers were less likely to take paid leave than white fathers as expected. Minority fathers are more likely to work in occupations that do not provide paternity leave, and are more likely to be penalized for requesting leave than white fathers (Melamed, 2014; Rudman & Mescher, 2013). Yet, all else equal, Latino fathers are more likely than white fathers to take paid leave, and to take longer lengths of paid leave, in the NLSY97. The younger fathers in this sample may be embracing the cultural emphasis on familism, or placing importance on one’s family, within Latino communities (Christerson, Edwards, & Flory, 2010). The emphasis on familism may lead young Latino fathers to place more emphasis on the new fatherhood ideal than fathers from other racial/ethnic groups.

Finally, salient and committed father identities that prioritize engaged fathering may encourage paternity leave-taking and longer periods of leave-taking. In particular, fathers who engaged in parenting activities prior to the child’s birth were more likely to take paid leave, and longer periods of leave, than fathers who were not involved prenatally. These fathers may have embraced caretaking father identities early, and sought out ways to enact fathering roles (Pasley et al., 2014; Stryker, 1968). Father identities may also be more salient for men making the transition to fatherhood as they actively seek ways to develop fathering skills (Harrington et al., 2014; Nepomnyaschy & Waldfogel, 2007). Indeed, we find that first-time fathers are more likely to take paid leave, and longer periods of leave, than fathers with other children. Finally, positive attitudes towards fathering are associated with longer periods of paid leave-taking.

Utilizing two national datasets to assess variations in paid leave-taking increases our confidence in these findings and suggests that the patterns uncovered in this study persist across various subsets of the U.S. population. Regardless of whether we look at a relatively young sample (NLSY97), or a relatively disadvantaged sample (FFCW), economic capital (i.e., income) consistently predicts a greater likelihood of paid leave-taking and longer leaves. Cultural capital gained from education consistently increases the likelihood of taking longer periods of paid leave. Greater social capital due to marriage is also consistently associated with being more likely to take paid leave, and longer periods of leave, compared to unmarried fathers. Furthermore, salient fathering identities as evidenced by prenatal involvement and being a first-time father are consistent predictors of taking paid leave and longer periods of leave.

There are some limitations. First, there is no information about what types of paternity leave programs respondents have access to, if any, and the amount of wage replacement received. Fathers may be utilizing workplace paternity or parental leave programs or using other forms of leave (e.g., vacation or sick days). There is also no information on whether, and to what extent, fathers may have been expected to work (e.g., responding to email) while on leave. Short leaves (i.e., less than one week) are likely also underreported in the NLSY97 (Baum & Ruhm, 2016). Similarly, questions from the GSS indicate support for paid parental leave (and how this leave should be divided) as opposed to support for paid paternity leave specifically. Knowing the types of leave opportunities that Americans support and that fathers have available—and are using—is essential to getting an accurate assessment of the determinants of, and potential barriers to, leave-taking. This information is also important for policymakers to scrutinize as they consider how statutory paid leave may be beneficial to families.

Second, although the use of multiple datasets offers complementary analyses of paternity leave-taking that encompasses diverse groups of fathers, this approach also presents challenges. Specifically, father identity measures are unavailable in the NLSY97 and somewhat limited in the GSS, which prevents us from examining whether fathers’ attitudes and behaviors regarding leave are due to more salient and committed father identities. In addition, we are only able to examine the influence of resident status on leave-taking within the FFCW. Having more consistent measures in each dataset would have allowed us to more fully assess whether the influence of father identities on leave-taking is consistent among various groups of fathers.

Finally, there are potential selection and social desirability bias issues that may influence the results, especially in the FFCW where data is only collected after children are born. It is unclear how these men felt about fatherhood prior to having a child, which may influence their patterns of leave-taking. Having data collected prior to the child’s birth (as in the NLSY97) would be helpful in minimizing these potential issues.

Despite these limitations, this study represents a substantial step forward in improving our understanding of the attitudes, patterns, and predictors of paid paternity leave-taking in the U.S. Overall, results indicate that most Americans believe that paid leave should be available to working parents who have a child, and a slight majority believe that fathers should receive at least a share of this leave. These attitudes are consistent with parental leave programs in many European countries (Heymann & McNeill, 2013). Yet, given that paid leave programs in the U.S. are largely workplace-based (and provided to individual employees, not both parents as is the case in many European countries), few fathers seem to be able to access paid leave opportunities. Socioeconomic inequalities seem to shape opportunities for paternity leave-taking; fathers with higher levels of economic, cultural, and social capital are more likely to take paid paternity leave, and to take longer periods of paternity leave, than less advantaged fathers. Finally, more salient and committed father identities that endorse nurturing seem to encourage more extensive paternity leave-taking practices. Thus, to the extent that paternity leave-taking offers positive consequences for families, there are additional reasons to encourage new fatherhood ideals.

Because the U.S. does not offer statutory paid paternity leave, access to leave is overwhelmingly dependent on workplace policies. This structure exacerbates the diverging destinies of families because more advantaged fathers have greater access and abilities to take paid leave than less advantaged fathers. Future research should further assess the consequences of these disparities and examine the ways in which paternity leave-taking may influence families. Such knowledge will be important as policymakers seek to determine how changes in family leave policies may benefit American families.