Preventing intimate partner violence through paid parental leave policies.

Paid parental leave policies have the potential to strengthen economic supports, reduce family discord, and provide opportunities to empower women (Basile et al., 2016; Niolon et al., 2017). In this article, we present a theory of change and evidence to suggest how paid parental leave may impact intimate partner violence (IPV). In doing so, we present three mechanisms of change (i.e., reduction in financial stress, increase in egalitarian parenting practices, and promotion of child/parent bonding) through which paid parental leave could reduce rates of IPV. We also describe limitations of the current state of knowledge in this area, as well as opportunities for future research. Ultimately, our goal is to facilitate the identification and implementation of approaches that have the potential to reduce violence at the population level. Paid parental leave embodies the potential of policies to change societal-level factors and serve as an important prevention strategy for IPV.


Introduction
Intimate partner violence (IPV) is a significant public health issue, with 37.3% of women and 30.9% of men in the United States experiencing contact sexual violence, physical violence, or stalking by an intimate partner in their lifetime . Global estimates suggest that the lifetime prevalence of physical and/or sexual IPV against women is approximately 30% (World Health Organization, 2013). However, we currently have few effective strategies to prevent the onset of violence (i.e., primary prevention) or reduce violence that is already ongoing (i.e., secondary/tertiary prevention) in intimate relationships Whitaker, Murphy, Eckhardt, Hodges, & Cowart, 2013). Moreover, those few that are effective focus on individual-or relationship-level factors and have limited population impact due to inability to scale up these strategies (Frieden, 2010;Spivak et al., 2014;Whitaker, Hall, & Coker, 2009;Whitaker et al., 2013).
In this vein, policy-based prevention approaches have the potential to change the outer layers of the social ecology (i.e., community and societal factors; Bronfenbrenner, 1979), altering social inequalities and ultimately changing norms that support the use of violence (Dahlberg & Krug, 2002). Given that the impacts of such policies can be broad, reaching communities and/or society at large, it seems that policy approaches may be ideally suited to modify those societal factors that contribute to rates of violence in communities. In support of this effort to identify potential policy approaches to prevent violence, it may be useful to evaluate whether the effects of current policies designed or enacted for other purposes, extend beyond their original purpose to affect rates of violence. Policies from various sectors (e.g., education, economic, criminal justice) designed to affect health inequities may serve as effective violence prevention strategies. For example, Kearns, Reidy, and Valle (2015) summarized the literature examining alcohol-related policies and their association with IPV. The authors reported an association between alcohol outlet density and rates of IPV, which suggest that policies that regulate the number of alcohol outlets in a given community may be an effective method to curb IPV in those communities. In similar fashion, D'Inverno, Kearns, and Reidy (2016) argued that policies designed to increase girls' and women's enrollment in science, technology, engineering, and math (STEM) fields may be an effective primary prevention strategy for teen dating violence (TDV) and IPV in great part due to effects on strengthening household financial security and reducing financial stress and its impact on relationship discord (Matjasko, Niolon, Valle, 2013;Niolon et al., 2017). In addition, supporting girls' and women's enrollment in STEM fields may also lead to more distal effects of promoting attitudes and beliefs about women as equals thereby increasing gender equity (Glick & Fiske, 2001). Indeed, given the links among economic deprivation, gender, health disparities, and IPV, policies that reduce familial financial stresses and increase gender parity may likely be effective tools to prevent IPV (D'Inverno et al., 2016;Niolon et al., 2017).
Paid parental leave represents one policy-based approach that has potential to strengthen economic supports, reduce family discord, and provide opportunities to empower women all of which have the potential to affect rates of IPV (Basile et al., 2016;Niolon et al., 2017). Paid parental leave 1 supports new parents by providing job-protected, paid time off to care and bond with a new child without interruptions to household income or conflict between work and family responsibilities. This bonding period may be invaluable in fostering positive parenting skills and promoting healthy family relationships and lifestyles (Chatterji & Markowitz, 2012;Goodman, 2012;Huang & Yang, 2015;Johansson, Wennberg, & Hammarström, 2014;Månsdotter, Lindholm, Lundberg, Öhman, & Winkvist, 2006;Månsdotter & Lundin, 2010;Saade, Barbour, & Salameh, 2010;Whitehouse, Romaniuk, Lucas, & Nicholson, 2013). In this sense, paid parental leave simultaneously supports the family as a whole while also strengthening support for mothers individually. But beyond the multitude of social, mental, and physical health benefits proffered by paid leave practices, paid parental leave policies may be an effective strategy to prevent future instances of violence in intimate relationships.
In this article, we outline the potential for paid parental leave to influence IPV indirectly through its purported influence on risk and protective factors associated with IPV. We present a rationale behind paid parental leave as a promising prevention approach for IPV, including a theoretical model based on empirical evidence of the various pathways by which paid parental leave may influence rates of IPV. We also describe limitations of the current state of knowledge in this area, as well as opportunities for future research. Our goal is to facilitate the identification of evidence-based, societal-level approaches for preventing violence (such as policy-based approaches) in order to achieve greater population impact. Ultimately, this article is a call for researchers, practitioners, and stakeholders across disciplines to collaborate in the implementation and evaluation of innovative strategies to prevent IPV.

Theoretical Model Describing Paid Parental Leave and its Impact on Intimate Partner Violence
There is a dearth of research that directly examines the relation between paid parental leave and IPV. For many states, implementation is still in the early stages; thus, there has been limited opportunity to examine the relation between paid parental leave and IPV. However, there are several theoretical reasons to expect that paid parental leave may affect rates of IPV. We propose three processes or, mechanisms of change, through which paid parental leave may potentially prevent or decrease IPV (illustrated in Figure 1).

•
Path 1 -paid leave maintains household income preventing financial stressors and associated relationship discord that can incite instances of relationship violence; • Path 2 -paid leave increases egalitarian parenting practices and decreases the impact of work interruptions on women's advancement in the workplace, thereby increasing gender equity, which is associated with lower rates of IPV against women; and • Path 3 -paid leave provides new parents a period of time to bond with a child free of conflict between work and family demands, which facilities IPV/TDV protective factors and reduces risk factors in youth (e.g., healthy parenting practices, healthy relationships, good parental mental health, etc).
Collectively, the proposed paths may work together in additive and multiplicative fashion to attenuate risk factors and increase protective factors, with the shared objective of preventing or reducing IPV. Below we present empirical evidence supporting the argument for each of these mechanisms to prevent IPV.

Path 1
The economic benefits of paid parental leave may be likely to impact the frequency of IPV in a relationship by reducing financial stress and worry about insufficient household income that can serve as precipitant stressors for violence. Poverty and stress related to financial strain have been linked to negative outcomes, including relationship dissatisfaction and conflict, which are risk factors for IPV (Byun, 2012;Capaldi, Knoble, Shortt, & Kim, 2012;Davis & Mantler, 2004;Dew, 2008;Fox & Chauncey, 1998;Neff, Holamon, & Schluter, 1995;Slep, Foran, Heyman, & Snarr, 2010). For the most economically disadvantaged, paid leave may proffer reduction in the number of violent events given that financial stressors such as food insecurity, eviction, disconnected phone service, and being unable to pay utilities are significant predictors of physical IPV perpetration among men and women (Schwab-Reese, Peek-Asa, & Parker, 2016). For example, a qualitative study of women who had experienced IPV during or shortly after giving birth found IPV often existed in conjunction with other stressful life events, including financial and housing difficulties (Bacchus, Mezey, & Bewley, 2003). Similarly, Breiding, Basile, Klevens, and Smith (2017) found robust associations between food and housing insecurity in the preceding 12 months and rates of IPV and sexual violence victimization. Notably, when the state of California implemented a paid leave policy, the most economically disadvantaged families showed the greatest increase in leave-taking (Bartel, Baum, Rossin-Slater, Ruhm, & Waldfogel, 2014). Thus, it seems providing paid parental leave could mitigate relationship stress about finances among the most at risk families during this critical and already stressful period.
Only one study has directly assessed the association between paid leave and IPV. Gartland and colleagues (2011) surveyed 1,507 Australian women during pregnancy and three, six, and twelve months postpartum about their experiences with physical and emotional IPV.
Women were also asked about employment status and eligibility for paid maternity leave. The authors identified three groups, women that: (1) worked during pregnancy and qualified for paid maternity leave; (2) worked during pregnancy but did not qualify for paid maternity leave; and (3) did not work during pregnancy, thus did not qualify for paid maternity leave. After controlling for maternal age at birth, relationship status, income, and education level, women who worked during pregnancy and qualified for paid maternity leave reported 58% lower odds of IPV in the first twelve months postpartum compared to women who did not have access to paid maternity leave (i.e., the combination of women that worked during pregnancy but did not qualify for paid maternity leave and women that did not work during pregnancy and therefore did not qualify for paid maternity leave, see Aitken et al., 2015).
Unfortunately, the authors did not test differences in rates of IPV between working mothers with access compared to working mothers without access to paid leave. Additionally, interpretation of the results is limited because the authors were unable to determine whether the women with access to paid maternity leave actually used their leave. It is possible that other factors, such as the perception of support in the workplace, may have played a role in decreasing violence against women. Nevertheless, there is evidence to suggest a trend between access to financial resources and reduced violence in intimate relationships (Ellsberg et al., 2015;Kim et al., 2007;Matjasko et al., 2013). Hence, it is possible that even partial wage replacement during parental leave may mitigate the stress associated with household finances, thereby reducing relationship problems, and consequently reducing the frequency of violent events in the relationship.

Path 2
Paid parental leave also has potential to influence rates of IPV by promoting more egalitarian parenting practices, which in turn, generalize to promote less traditional gender norms and ultimately reduce gender inequality. This is pertinent because traditional (i.e., patriarchal) gender norms and gender inequality are risk factors for violence against girls and women (Gressard, Swahn, & Teten, 2015; World Health Organization [WHO]/London School of Hygiene and Tropical Medicine, 2010). For example, at the individual level, endorsement of patriarchal gender role attitudes has been linked to physical and sexual violence against an intimate partner (Parrott & Zeichner, 2003;Reidy Berke, Gentile, & Zeichner, 2014;Smith-Hunter, Parrot, Swartout, & Teten-Tharp, 2015). Likewise, at the societal level, indices of gender inequality are strongly associated with the rates of girls' (but not boys') physical dating violence victimization (Gressard et al., 2015). Accordingly, it seems altering patriarchal gender norms and consequent gender inequality may be fruitful in the prevention of IPV.
In the U.S., working mothers spend approximately twice as much time as working fathers engaged in domestic work and this difference is due largely to primary childcare duties such as feeding, changing diapers, bathing, taking care of children when they are sick, and managing children's schedules and activities (Allard & Jane, 2008;Bureau of Labor Statistics, 2015;Pew Research Center, 2015). Evidence suggests policies that provide and encourage fathers to take paid leave increase their participation in these childcare duties (Haas & Hwang, 1999;OECD, 2016;Tanaka & Waldfogel, 2007). Moreover, fathers who participate early in childcare duties tend to stay involved throughout a child's life (OECD, 2016). This participation is notable because fathers who are more involved in direct physical and emotional care of children hold more gender-equitable attitudes (Bonney, Kelley, & Levant, 1999;Bulanda, 2004;Craig, 2006). In fact, involved fathers who attend prenatal visits, take paternity leave, and help their children with homework, etc., are less likely to perpetrate IPV (Chan, Emery, Fulu, Tolman, & Ip, 2017). Thus, as fathers' use of parental leave becomes more commonplace, the stigma of assisting with childcare (and other domestic work) may abate potentially altering traditional, hegemonic, masculine ideologies that are associated with gender inequality and ultimately IPV against women (Farmer & Tiefenthaler, 2003;Gressard et al., 2015;McCauley et al., 2013;Murshid & Critelli, 2017;Reidy, Shirk, Sloan, & Zeichner, 2009;Smith-Hunter et al., 2015).
In addition to supporting increased involvement from fathers, paid parental leave would also have potential benefits for mothers. Research has shown that women in the United States who have access to job-protected maternity leave are more likely to return to their previous employers after childbirth and experience positive wage benefits, even when controlling for employer characteristics (Waldfogel, 1998). 2 Both in the United States and internationally, one consistently documented source of gender inequality relates to the wage gap between male and female workers. Median weekly earnings for women in the United States represented 82% of median weekly earnings for men in 2016 (Bureau of Labor Statistics, 2017), and research suggests that the gender wage gap grows with age, becoming even more pronounced for women with children (Budig & England, 2001;Goldin, 2014;Slaughter, 2015). Gangl and Ziefle (2009) found that when controlling for work experience, working mothers in the United States experience a 4-7% wage penalty per child. This penalty was largely accounted for by work interruptions for childcare, changes in employer at reentry into the labor market, and other economic responses to motherhood. In the United States, the gender wage gap may also play a role in the slow uptake in fathers utilizing paid leave, since the unavailability of paid leave incentivizes the parent making the most money to keep working. In other countries where paid leave is offered, but only at a percentage of the parent's earned salary, the gender wage gap also motivates the parent making the most money to continue working, yet another reason for the slow uptake in fathers utilizing paid leave. In fact, some countries have implemented successful strategies such as "bonus periods" and non-transferrable parental leave to increase parental leave in men (Haas & Rostgaard, 2011). A couple may receive extra weeks of paid leave if the father uses a certain amount of paid parental leave, providing a "bonus period." Non-transferrable parental leave provides each parent with their own paid leave period, which cannot be used by the other parent. Non-transferrable parental leave has doubled the number of parental leave days taken by men in Iceland and Sweden (Organisation for Economic Co-operation and Development, 2016). Encouraging fathers to take parental leave is critical to combat traditional patriarchal gender roles being reinforced when mothers exclusively stay home to care for a new child. In this way, paid parental leave represents an opportunity to not only encourage fathers to participate more frequently in childcare duties, but also to support mothers in returning to the workforce following the birth of a child. This has the potential to decrease the impact of work interruptions on future earning potential and ultimately advance gender equality in the long-term.

Path 3
Paid parental leave policies may also support prevention of IPV through the prevention of TDV, a risk factor for IPV (Exner-Cortens, Eckenrode, Bunge, & Rothman, 2017). Specifically, paid parental leave has demonstrated positive impacts on parental involvement and positive parenting practices. For example, this dedicated time encourages new parents to learn and become interested in child development, increases involvement in child caretaking responsibilities, offers them the opportunity to become more attentive to the infant's needs, and increases the probability and duration of exclusive breastfeeding (Feldman, Sussman, & Zigler, 2004;Nepomnyaschy and Waldfogel, 2007;Galtry & Callister, 2005;Roe, Whittington, Fein, & Teisl, 1999). These parenting behaviors, in turn, contribute to improved child and family physical, behavioral, and mental health including decreasing the risk of externalizing disorders, depression, substance use, and risky sexual behavior (Oddy et al., 2010;Cookston & Finlay, 2006;Deptula, Henry, & Schoeny, 2010). Pertinently, a recent review of the literature on risk and protective factors for TDV identified parenting-related factors (e.g., low parental monitoring, harsh parenting practices, and negative parent-child interactions) as increasing the risk of TDV perpetration (Vagi et al., 2013). Likewise, the aforementioned physical, behavioral, and mental health outcomes are risk factors that exacerbate TDV (Vagi et al., 2013). Clearly, longitudinal research is necessary to claim longterm impacts of paid parental leave on child and adolescent outcomes, including perpetration of TDV (and ultimately IPV). However, it seems possible that utilizing parental leave may improve parenting practices and family bonding, thereby reducing adverse mental and behavioral health disorders in adolescence, which in turn, may reduce risk for TDV perpetration.
In a related vein, some parenting practices influenced by paid leave, such as duration of breastfeeding, have also been linked to lower risk for child abuse and neglect (Klevens, Luo, Xu, Peterson, & Latzman, 2016;Strathearn, Mamun, Najmun, & O'Callaghan, 2009). To the extent that paid parental leave prevents a child from being a victim of child maltreatment or a witness to IPV, the intergenerational transmission of IPV may also be interrupted, as these forms of family violence are also predictors of IPV (Ireland & Smith, 2009;Linder & Collins, 2005). Thus, paid parental leave may buffer against known risk factors for perpetration against intimate partners by promoting parent-child bonding and healthy parenting practices, which in turn may decrease risk for child maltreatment and promote the healthy development of youth.

Caveats & Conclusions
Despite widespread recognition of the significant public health implications of IPV, a body of accumulated research, a vast field of dedicated practitioners, and the resources that have been devoted to IPV/TDV over the past decades, we still have significant progress to make in preventing such violence Whitaker et al., 2013). This is surely, in part, due to the multifaceted nature of IPV/TDV (Reidy & Niolon, 2012), wherein there is no singular cause for any one person, and no unified set of causes across persons. Consequently, truly effective interventions will likely necessitate comprehensive strategies that incorporate multiple causal mechanisms at multiple levels of the social ecology. Herein, we have laid out three potential mechanisms of change whereby paid parental leave may influence the perpetration of IPV/TDV. Of course, these paths at this point are primarily theoretical, albeit based on empirical links. Accordingly, we believe these to be three fruitful areas of exploration for prevention researchers in investigating the relationship between paid leave and IPV outcomes.
However, there are a number of critical questions that future research on this topic should likely consider. For example, the ideal length of paid leave has not yet been determined. Some studies suggest that too much leave could be harmful to a woman's career in the form of lower wages, lower labor market attachment, and workplace discrimination (Hegewisch & Gornick, 2011;Morgan & Zippel, 2003;Ray, Gornick, & Schmitt, 2009), with some evidence suggesting that even short interruptions from work (i.e., less than four months) can increase a woman's risk for downward mobility and decrease chances for upward mobility (Aisenbrey, Evertsson, & Grunow, 2009). Conversely, it will be important to establish the minimum time necessary to achieve the positive outcomes related to paid leave. In addition, policies of this nature may produce different results in the United States relative to other countries. For instance, the Nordic countries, which are known to have high levels of gender equality, paradoxically report high levels of IPV (Gracia & Merlo, 2016;World Economic Forum, 2014). One potential explanation for this unexpected association is backlash from males who respond to an increase in female independence with violence because it challenges the norm of male dominance and female dependence (Macmillan & Gartner, 1999). Many policies attempt to curtail a behavior and change power roles, so that behavior may increase in response to the policy before deeply rooted ideologies change and the behavior declines. It is also possible that reports of a behavior will increase once a policy raises awareness of the problem behavior. In the "Nordic paradox" above, while not a specific policy per se, IPV may be elevated because of higher levels of disclosure, as incidents of violence against women are more likely to be openly addressed and challenged in societies with greater equality. It is also worth noting that research examining gender equality and IPV can present a complicated picture. Using data from 30 states in the U.S., Yllo and Straus (1990) found a curvilinear trend such that states with the lowest and highest levels of gender equality had the highest rate of wife assaults, but once the study was updated to include a larger sample size and more recent data from all 50 states, Straus (1994) found that states with higher levels of gender equality also reported lower rates of wife assaults. Future research should continue to carefully examine policies, their impact on gender equality, and the effect this has on IPV.
The way in which policies are analyzed can also have implications on the findings. Research at the individual-level and policy-level may produce different results. A systematic review that examined both individual-level and policy-level comparisons of paid leave found the individual-level showed positive maternal health benefits (e.g., lower psychological distress and reduced odds of poor physical health), whereas the studies at the policy-level showed negative or null effects of paid leave (e.g., no differences in depression; less life satisfaction and poorer general health; Aitken et al., 2015). The authors concluded that the studies conducted at a policy-level aggregated the effects for women who do and do not take leave, thus accounting for the null findings (Aitken et al., 2015). Studies at the policy-level may also limit understanding of the impacts of paid leave on people of differing marital status, sexual orientation, race, socioeconomic status, and job roles (Aitken et al., 2015). This is particularly problematic for learning more about subpopulations in which gender role norms may differ from typical patriarchal gender norms (e.g., cisgender individuals and those in same sex relationships).
Despite these unknowns, there are clear benefits of paid parental leave (Chatterji & Markowitz, 2012;Goodman, 2012;Huang & Yang, 2015;Johansson et al., 2014;Månsdotter et al., 2006;Månsdotter & Lundin, 2010;Saade et al., 2010;Whitehouse et al., 2013). As such, exploring additional outcomes (i.e., IPV prevention) seems only logical. In addition, we should point out that the three mechanisms of change we present here are not necessarily exhaustive. It is entirely possible that paid parental leave policies may have preventive effects on IPV through additional paths. For example, the period of time following the birth of a new child represents a period of heightened risk for IPV victimization, especially for younger, lower income mothers (Agrawal, Ickovics, Lewis, Magriples, & Kershaw, 2014;Harrykissoon, Rickert, & Wiemann, 2002). For some women, it is possible paid leave could reduce financial dependence on their abuser enough to allow them to escape a potentially escalating abusive situation, even if only temporarily. This guaranteed income, combined with time off, might empower a woman to leave her abuser without having to immediately return to work or fear losing her job. The potential of paid parental leave to prevent IPV in the first place or reduce IPV in relationships where it is already occurring merits further empirical investigation.
Paid parental leave is only one of many policies that may prove to be an important violence prevention tool. This policy can offer families a multitude of benefits, but other related policies, in conjunction with paid leave, could be considered as well. For example, equal pay for equal work, subsidized childcare, or policies which put value on unpaid childcare at home through basic income or participation income where parents can combine paid and unpaid work may impact IPV through similar pathways. These policies have previously been recognized as potential approaches to empower and support women by increasing economic stability and decreasing gender inequality (Basile et al., 2016). Paid parental leave is a concrete approach that highlights the potential of policies to change societal-level factors and serve as an important prevention strategy that prevents multiple, connected forms of violence.