Do Gender Norms Become Less Traditional with Displacement?

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.


Policy Research Working Paper 9822
Conflict-induced displacement is associated with loss of human and physical capital and psychological trauma. Households and social structures that produce and reproduce gender norms are disrupted, providing opportunities for change. This paper operationalizes a definition of gender norms that brings together the behaviors and attitudes of displaced and non-displaced women using household survey data for Colombia. The results of a two-step estimation involving kernel-based propensity score matching and multilevel linear regression models show that gender norms condoning violence against women relax with displacement, while those that limit women's economic opportunities become more rigid. The findings also reveal a misalignment between attitudes and behaviors in other domains. Displaced women have less rigid patriarchal attitudes, but their ability to decide about contraception and their own earnings decreases following displacement. This paper is a product of the Gender Global Theme. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The author may be contacted at erubiano@worldbank.org.

Introduction
Women can be disproportionately affected by the negative effects of conflict-induced displacement. Access to essential services such as reproductive health care can be disrupted. Displacement can result in higher levels of gender-based violence (GBV) (Annan & Brier, 2010;Callaway & Martin, 2011;Cohen et al., 2013;Vu et al., 2014;Wirtz et al., 2014). However, it can also provide opportunities to challenge gender norms that limit women's access to opportunities and ability to make decisions.
Following displacement, women might adopt new roles that would not have been possible before (Justino et al., 2012;Meertens & Stoller, 2001;Pirtskhalava, 2015). For example, in the absence of men, displaced Nuer women in South Sudan took on male responsibilities as income providers and assumed roles perceived as male, including negotiating dowries. To date, however, partly due to data limitations, few quantitative studies take account of the changes in gender relations among women and men in situations of displacement.
This paper aims to bridge this knowledge gap by building on the work by Heise and Cislaghi (2020), who propose a definition of gender norms that brings together two streams of theory and practice around gender equality. The first stream is the work on social norms, which emerged from studies in social psychology and evolved with behavioral economics (Bicchieri, 2005;Mackie et al., 2015). The second stream is the study of gender norms advanced by feminist scholars (Badgett & Folbre, 1999;Connell & Pearce, 2014;Connell & Pearse, 2015). The definition is operationalized by measuring behaviors and attitudes (Alesina et al., 2013;Harper et al., 2020).
The analysis focuses on Colombia, a country with a long-standing conflict and home to the second largest internally displaced population in the world. It uses three rounds of Demographic and Health Surveys (DHS) data for the 2005-2015 period to examine the extent to which gender norms that limit women's access to reproductive health, economic opportunities, and mobility, and norms that tolerate violence against women, and endorse patriarchy become less traditional among 3 women in situations of displacement. The empirical approach involves a two-step estimation using kernel-based propensity score matching (PSM) and multilevel linear regression models to estimate the effect of displacement behaviors and attitudes on the matched sample of women.
The findings show mixed evidence regarding norm change. Gender norms that tolerate violence against women become less traditional with displacement, while those that limit women's economic opportunities become more rigid. Results also reveal a misalignment between attitudes and behaviors in specific domains of gender norms. In particular, displacement reduces the likelihood of agreeing with patriarchal statements such as 'families with men have less problems' or 'a good wife obeys her husband,' but women's ability to decide about contraception and their own earnings, proxies for behaviors, decreases with displacement. This is consistent with previous studies showing that displaced women in Colombia report increased controlling behaviors when they pursue employment while their husbands are unemployed. Furthermore, gender norm transgression might lead to a backlash of more restrictive gender norm attitudes. Displaced men's controlling behaviors might be exacerbated by psychological trauma, stress and loss of financial stability (Hynes et al., 2016;Wirtz et al., 2014).
The remainder of this paper is organized as follows. The next section presents a review of the literature. Section 3 describes the Colombian context. Section 4 presents the theoretical framework. Section 5 describes the data, followed by Section 6, which presents the empirical approach. Section 7 discusses the results before concluding in Section 8.

Literature Review
Displacement disrupts social and community relations, alters the structure and size of households, and is associated with changes in gender roles (Gururaja, 2000;Ibáñez & Velásquez, 2009;Ibáñez & Vélez, 2008;Levine et al., 2019;Vélez & Bello, 2010). In some cases, displacement might also provide opportunities to renegotiate gender roles (Aysa-Lastra, 2011). Women often take on the role of providers and protectors of families when their male partners die, disappear, are recruited by armed groups, or migrate in search of economic opportunities (Ayssa & Massey, 2004;El-Bushra, 2000). For example, Meertens and Stoller (2001) and Meertens and Segura-Escobar 4 (1996) document changes in gender roles among internally displaced populations (IDP) in Colombia. The authors argue that most displaced rural women were raised in a context of patriarchal traditions characterized by rigid gender roles around domestic chores and participation in agricultural activities close to the home. When fleeing to urban settings, displaced women often become the main breadwinners for the first time in their lives.
Gender norms that assign women to the domestic sphere have significant economic consequences in the context of displacement. As women enter paid employment, these gender norms are reproduced by occupational segregation (Badgett & Folbre, 1999). Displaced women tend to be employed as domestic workers or they are often engaged in petty trade (Bouta et al., 2005). They also maintain their roles as primary caregivers, creating a double burden compounded by poor security, limited infrastructure, and gender norms (Culcasi, 2019;Petesch, 2017;Pirtskhalava, 2015). For example, Culcasi (2019) reports that Syrian refugee women in Jordan have become breadwinners for their households, but their responsibility as the caretakers for their families has not diminished. Similar dynamics have been reported for IDP widows in Nepal (Ramnarain, 2016), Chechen refugees in the Czech Republic (Szczepanikova, 2005), and IDP women in Darfur (De La Puente, 2011).
Notwithstanding displaced women's increasing participation in the labor market, the evidence on the effect of displacement on intra-household bargaining power is mixed. In a study on the impact of conflict on women's activities, Justino et al. (2012) find that women in Colombia participate more actively in labor markets during and immediately after conflict. Moreover, greater engagement in paid work is accompanied by improvements in women's economic empowerment within households. In another study for Colombia, Calderón et al. (2011) confirm that IDP women work more hours per week than non-IDP women in rural areas. However, their ability to participate in important household decisions remains unaltered. Further, IDP women often report increased domestic violence when they pursue employment or education while their husbands were unemployed . In Turkey, Gulesci (2018) finds that displaced men were more likely than their non-displaced counterparts to display controlling behaviors, either by limiting their wives' movements or social interactions. At the same time, displaced women were more inclined to believe that domestic violence is acceptable, compared to non-displaced women.

5
Women's ability to perform activities deemed unsuitable pre-displacement often depends on the contestation of gender norms at the community level. In the above mentioned study in Darfur, De La Puente (2011) shows that IDP women were involved in health-related activities at the community level, but they did not participate in decisions related to the infrastructure or management of the camp, which were perceived as male fields. In contrast, Ramnarain (2016) finds that in the aftermath of the conflict in Nepal, widows engaged in employment outside the home and some of them even crossed over into male-dominated fields, such as construction labor or transport.

Colombian Context
The number of IDPs in Colombia constitutes 12 percent of the 45.7 million people who had been forced to flee worldwide by the end of 2019, resulting in the second largest displacement in the world, behind only the Syrian Arab Republic (IDMC, 2020). Displacement in Colombia is directly linked to violence, but the underlying causes are as complex as the protracted conflict itself.
Similar to other situations of conflict in the world, the attacks on civilians are the main triggers for displacement in Colombia (Bohra-Mishra & Massey, 2011;Czaika & Kis-Katos, 2009;Engel & Ibáñez, 2007;Lischer, 2007;Shultz et al., 2014). The violence is directly linked to land disputes and drug trafficking, but it is also used as a strategy of armed groups to exert greater power and destroy social networks (Ibáñez, 2008;Kay, 2001;Perez, 2002). Official figures indicate that the main groups responsible for the displacement are guerilla and paramilitary groups (Ibáñez, 2009).
The nature of the conflict implies that displacement is not confined to specific areas of the country.
Between 1997 and 2018, more than 6 million people were forced to flee their homes from 90 percent of the country's 1,123 municipalities ( Figure 1) and its incidence has been more intense in rural areas characterized by weak institutional presence (Angrist & Kugler, 2008;Ibáñez, 2009).
Nonetheless, cities hosting large populations of IDPs, such as Medellín, Cali and Bogotá, have also experienced violence by parties of the armed conflict and criminal groups, which has led to the phenomenon of intra-urban displacement (Atehortúa et al., 2013;Jacobsen, 2011). The decades-long armed conflict in Colombia has affected men and women in a number of ways.
Men have been more likely than women to be kidnapped, killed, injured, and forcibly recruited by armed actors (Oficina del Alto Comisionado para la Paz, 2020). Women and girls, on the other hand, are more likely to be victims of sexual violence and forced labor. They often assume the role of household heads and tend to be the caregivers for disabled family members (Bouvier, 2016).
Rural women are particularly vulnerable, as they have limited access to land and other productive assets (World Bank, 2019). This situation is aggravated by the exposure to sexual violence in situations of displacement (ABColombia & US Office in Colombia, 2013).
The Colombian state has established a solid normative framework for gender equality. On paper, the legislation recognizes women's rights, penalizes GBV, and mandates a 30 percent quota for women in electoral lists. The peace accord between the government and the Revolutionary Armed Forces of Colombia (FARC for its Spanish acronym) is often referred to as a model for gender inclusion, as it recognizes gender inequities in multiple areas; guarantees the rights of rural women; and addresses the rights of the victims (PRIO Centre on Gender, Peace and Security, 2016;  Navarro, 2019). Furthermore, the Victims and Land Restitution Law, which enables victims of the conflict to receive assistance and reparation, established preferential treatment for IDP women and reparations for survivors of sexual violence (Bouvier, 2016;Valcarcel & Samudio, 2017).
Notwithstanding, the legislative framework has not yet translated into conditions of gender equality. Gender discrimination prevents applicants from receiving property rights after their husbands have died or disappeared (Garcia-Godos & Wiig, 2014;Meertens, 2010). Access to justice also remains a challenge for victims of conflict-related sexual violence, despite an increase in the number of formal complaints (United Nations Security Council, 2020). One of the challenges to achieve gender equality in Colombia lies in deeply rooted gender norms. Women are expected to take on the bulk of domestic responsibilities, whereas men are seen as the head and main breadwinners for their families (Chant, 2002). Affordable, good-quality childcare services are lacking, and no legal provision exists for paid parental leave to be shared between both mother and father (World Bank, 2019Bank, , 2020. In fact, women do nearly four times as much as much unpaid domestic and care work as men do. 1 According to van der Gaag et al. (2019), alongside outliers such as Bangladesh and Algeria, Colombia is one of the countries where both the laws and gender norms around caregiving are relatively resistant to gender equality. These structural barriers are reflected in women's lack of access to economic opportunities.

Theoretical Framework
In general, norms specify rules, conventions and institutions that dictate what should or should not be done (Harper et al., 2020). Gender norms, in particular, are defined by Cislaghi and Heise (2020) as: Social norms 2 defining acceptable and appropriate actions for women and men in a given group or society. They are embedded in formal and informal institutions, nested in the mind, and produced and reproduced through social interaction. They play a role in shaping women and men's (often 8 unequal) access to resources and freedoms, thus affecting their voice, power and sense of self. 3 (p.

415).
Following Cislaghi and Heise (2017) and using elements outlined in Marcus and Harper (2015) in relation to gender norm change, this section describes the main individual, social, material, and structural factors that could shift gender norms in situations of displacement. Depending on the context, they can either promote a positive change, that is, gender norms become less traditional and new practices emerge, or a negative change, which entails more discriminatory practices.
Gender norms are learned early in life. They are adopted and endorsed by women and men through their behaviors and attitudes (Cislaghi & Heise, 2020;Harper et al., 2020;Lundgren et al., 2019).
Gender norms are reinforced or contested in school, at the workplace, by the media, and other social institutions (Tenenbaum & Leaper, 2002). Aspirations and skills acquired over time contribute to the reproduction or change of gender norms. For instance, gender norms become manifest in educational materials that portray characters in stereotypical roles in the household and at the workplace (Blumberg, 2008;Islam & Asadullah, 2018;Mahmood & Kausar, 2019;Miroiu, 2004). Providing people with access to unbiased education materials and curriculum can thus contribute to norm change. For instance, a semester-long course on gender equity as part of preservice training in Turkey developed more favorable attitudes toward gender equality among aspiring teachers (Erden, 2009). Improved access to education can also foster more liberal attitudes and a break in the intergenerational transmission of gender norms (Marcus & Harper, 2015).
Mass media can reflect and sustain gender norms over time, but it can also foster positive change.
By moving from remote to more densely populated areas, displaced women and men may access both factual and overt messaging about gender equality. These messages can also be transmitted through social interaction and popular entertainment programs that present an alternative vision of gender relations. For example, evidence suggests that soap operas played an important role in the reduction of fertility rates in Brazil and in shifting gender norms around domestic violence in Nigeria (Banerjee et al., 2019;Ferrara et al., 2012).
Migration can reduce exposure to the structures that tend to reinforce gender norms, such as traditional and religious leaders (Muñoz-Boudet et al., 2013). Behavioral studies suggest that what other people do has a greater influence than what they say, particularly for behaviors that are visible, such as child marriage (Palluck & Ball, 2010). For behaviors that are less visible, norms are more likely to be spread by people talking about endorsing them (Bursztyn et al., 2020).
Gender norms are influenced by material conditions and the environment in which individuals are born and live. Gender norms underpin inheritance laws, ownership and control over assets, and intra-household dynamics (Agarwal, 1997;Connell & Pearce, 2014). For example, land privatization via government redistribution programs has often disadvantaged women by placing land in the hands of male relatives (Whitehead & Tsikata, 2003). On the other hand, the loss of assets by displaced families represents an overall loss of wealth for the household, but it might also 'level the playing field' for women. This could foster changes around decision making, resource allocation, and patriarchal notions around men as the main breadwinners.
Policies, regulations, and institutional biases reinforce gender norms. In the labor market, for instance, gender norms influence recruitment, the work environment, wage differentials, and career progression. In Colombia, displaced men often face large spells of unemployment as their agricultural skills are less relevant in urban settings. Displaced women, on the other hand, are frequently employed as domestic workers who are poorly remunerated (Meertens & Stoller, 2001).
These dynamics can challenge patriarchal gender norms whereby men are no longer the breadwinners and increase intra-household tensions (Calderón et al., 2011;Meertens & Segura-Escobar, 1996). On the other hand, the recognition of women's economic contribution might call patriarchal norms into question (Gutmann & Viveros, 2005;Jensen, 2012).

Data and Descriptive Statistics
This paper uses data from the Colombian DHS for 2005, 2010, and 2015. 4 Surveys are representative of the female population ages 13-49 at the national, urban, and rural levels. They collect information on health outcomes and socio-economic characteristics. The most recent waves include attitudes towards gender equality, women's role in society, GBV, and intra-household decision-making. The DHS employs two-stage sampling designs. Primary sampling units ( The surveys allow for the direct identification of displaced household members. Questions ask whether the respondent lived in one or more places in the last 5 years, the date of migration, and the reason for migrating. 5 Displaced households are defined as those that had at least one member who was forced to flee due to conflict. This is a reasonable assumption for Colombia, where nearly 91 percent of IDPs migrate with all household members (Ibáñez, 2008). Furthermore, to facilitate the provision of reparations and other entitlements, as the IDP status is attached to a household, it is transmitted across generations (Sarzin, 2017;Shultz et al., 2014). IDPs were oversampled in the three survey waves and they represent 7 percent of the individuals who migrated internally each survey-year.

Measurement of gender norms
Gender norms relate to multiple spheres of life. This study focuses on norms around reproductive health, economic opportunity, mobility, violence against women, and patriarchy. The analysis of gender norms in other spheres, although important, remains beyond the scope of this paper because of data limitations. For the purposes of operationalization, the analysis combines the definition proposed by Cislaghi and Heise (2020) with elements of the social norms theory (Bicchieri, 2005(Bicchieri, , 2017. 6 Specifically, it pays attention to the role of a reference network, which refers to the group of people whose actions and beliefs individuals care about when they act. Depending on the context and the sphere of life that gender norms refer to, this group can be given by neighborhoods, villages, or people on the street. Hence, as people move from one place to another, they move across reference groups and may knowingly change their behavior to comply with the norms in place in the new setting (Choe et al., 2014).
The analysis focuses on behaviors or actions and attitudes or empirical expectations (Alesina et al., 2013). 7 These two components are measured and analyzed separately, rather than combined in an index. Following previous studies, survey clusters are used as a proxy for reference networks (Storey & Kaggwa, 2009;Uthman et al., 2011;Vyas & Heise, 2016). The prevalence of behaviors and attitudes at the reference network level is inferred by aggregating (non-self) reported values across individuals in the same cluster, as it is reasonable to think that people residing in the same cluster might have direct contact with each other. This approach is consistent with the feminist literature which has theorized gender norms as having blurry boundaries, rather than focusing on a particular group with similar demographic characteristics (Oakley, 2015). Also, following gender norms theory, which focuses on the alignment between the norm and personal attitudes, the analysis in this paper assumes that gender norms become less (more) traditional when both attitudes and behaviors become less (more) rigid.
The analysis examines 17 items (10 attitudes and 7 behaviors) classified into five domains.
Reproductive health, is measured using one indicator for attitudes and one indicator for behaviors, as shown in Table 1. The question on attitudes refers to women's approval of contraception to prevent pregnancy. The proxy for behavior measures the ability of women to decide upon the use of contraception. The norms that limit women's access to economic opportunities, are measured by one indicator for attitudes and two indicators for behaviors. These items deal with the intersection of family and work as well as the ability of women to decide on their earnings. The norms that limit women's mobility, are proxied by one indicator for attitudes and one indicator for 12 behaviors, which capture women's ability to make decisions around their own mobility. There are four indicators to measure tolerance towards violence against women and two proxies for violent behaviors. Patriarchal gender norms include six indicators measuring the disagreement with statements around male dominance within the household. Four of the indicators on attitudes are combined into a single measure that differentiates women who disagree with all statements compared to those who agree with at least one of them. All indicators are assigned a 1 for a less traditional attitude/behavior and 0 otherwise. Only the more recent waves include questions about attitudes towards gender equality.  Table 2 shows statistics comparing IDP and non-IDP women interviewed in the three rounds of DHS. On average, IDP women are younger and less educated than their non-IDP counterparts.

Descriptive statistics
They are less likely to be married but more they likely to be widowed, or not be in union. IDP households have one more member than non-IDP households; they have slightly more children and adult members, but fewer elderly. The average time in displacement is 2.5 years. 14 In terms of reproductive health, most women approve the use of contraceptives, but less than 1 in 5 women use any modern method and are the sole decision makers regarding contraceptive use ( Figure 2). When it comes to economic opportunities, the proportion of women disapproving their relegation to the domestic sphere is lower than the share disagreeing with traditional attitudes in the other four dimensions. Also, few women decide on their earnings. Regarding mobility, IDP and non-IDP women are equally likely to endorse nontraditional attitudes and behaviors. When it comes to violence against women, few women disagree with the statements 'when men are mad it is better not to tempt them' or 'women stay in abusive relations because they like it.' In contrast, most of them reject the statement 'men have the last word in household decisions' and would call out a friend who abuses another woman. In terms of patriarchy, few women disagree with all four statements around men's superiority in the household. In addition, IDP women are less likely than non-IDP women to have a say in important household decisions.

Empirical Strategy
Following Ho et al. (2007), the empirical approach involves a two-step estimation. In the first stage, the analysis employs kernel-based propensity score matching (PSM) to pre-process the data and control group for the displaced (treatment) before applying the parametric analysis in the second stage. 8 The treatment group includes all women living in a household where at least one member was displaced due to violence, while the control group is defined as individuals who were not displaced. Voluntary migrants are excluded from the analysis. The main assumption is the exogeneity of the treatment, that is, armed groups attack civilians, seize their property and force them to flee. Therefore, displacement is not a voluntary decision to improve economic conditions (Ceriani & Verme, 2018;Ruiz & Vargas-Silva, 2015). In Colombia, in nearly 9 of every 10 cases, displacement is a reaction to being a victim of violent attacks (Ibáñez & Vélez, 2008). or if it moved to another department where there were massacres in any other department two years prior to their migration; and 0 otherwise. For non-displaced households, the variable assigns a 1 to households that live in a municipality where there were massacres with a two-year lag.
The matching is done for each survey wave separately. The balancing property is fulfilled, that is, the mean propensity score is the same for individuals in the treatment and control groups. There is also a high degree of overlap between the two distributions, indicating that the common support assumption is satisfied (see balance tests in the Appendix).
In the second stage, the analysis employs a multilevel linear regression model to estimate the effect of displacement on behaviors and attitudes on the matched sample. The model is given by equation (1): To examine whether a multilevel model is appropriate, intraclass correlations were computed from the empty model with only the random error allowed to be free. The intraclass correlation captures 17 the proportion of variance that lies between level-2 units, which ranges from 0.15 for attitudes and behaviors around contraception to 0.9 for variables around violence against women. These amounts of variation are moderate to high, as Snijders and Bosker (2012) note that intraclass correlations with values between 0.05 and 0.2 are common. Table 3 shows the effect of displacement on attitudes and behaviors around reproductive health.

Results
Columns (1)-(5) indicate that displacement does not alter attitudes towards the use of contraception. This is explained by the fact that most women in the sample agree with the use of contraception. When it comes to behaviors, columns (6) Women's lack of decision-making power around contraception could also be explained by less 'gender-equitable' practices within the household, which tend to the transmitted through the family and passed down to the next generation (Fernández & Fogli, 2009). Given the approach to determine whether gender norms become less traditional (or not) adopted in this paper, these findings do not provide evidence to suggest that displacement relaxes gender norms around reproductive health. Turning to economic opportunities, Columns (2)-(5) in Table 4 show that displacement is associated with more traditional attitudes around women in the domestic sphere. Depending on the specification, displacement reduces the probability of disagreeing with the statement 'a woman's main role is family caregiving and cooking' by 6-8 percentage points. This might be explained by the fact that social expectations that assign women to the domestic sphere can make them 'protective' of the unpaid care space, attaching value to their leadership of it and being resistant to   Ramnarain, 2016;Szczepanikova, 2005), which reveal that women can work for pay outside the home, but there is an expectation that they will be the main caregivers in the household. In the case of Colombia, Calderon et al. (2011) show that IDP women work more hours than non-IDP women in rural areas, but greater engagement in the labor market does not translate into improved bargaining power. Indeed, gender norms about paid and unpaid work are intertwined, but they can also move in different directions. According to Harper et al. (2020), this is the case when norms stretch to encompass women doing paid work, without any corresponding shifts in male responsibilities.    is also associated with less traditional behaviors. These results, however, do not provide enough evidence to say that displacement is associated with less traditional gender norms around mobility.

20
In terms of the effect of displacement on gender norms around violence against women, columns (1) and (5) in Table 6 show that displacement reduces the likelihood of disagreeing with the statement 'it is better not to tempt men when they are mad,' but increases the probability of supporting the statement 'women stay in abusive relations because they like it.' In terms of behaviors, on the other hand, the effect of displacement is significant and stable across specifications for one of the two proxies. IDP women are more likely than their non-IDP counterparts to state that they would call out a friend who abuses a woman.
These findings might reflect the fact that the acceptability of violence spans a continuum. Some women believe that violence is justified under certain circumstances, but they do not accept it completely (Harper et al., 2020). Furthermore, attitudes towards violence against women might be slow to change because of sticky norms that reflect patriarchy (Harper et al., 2020), but results in Table 6 provide some evidence to suggest that some attitudes and behaviors around violence against women appear to change with displacement.
Strong legislative frameworks that support the rights of displaced women and condemn different forms GBV can also contest traditional gender norms. Some of these laws shape values and norms, which in turn, can influence individual attitudes and behaviors (Klugman, 2017;Nadler, 2017).
For example, in a study of 12 Sub-Saharan African countries Maswikwa et al. (2015) found that the prevalence of child marriage was 40 percent lower in countries with laws against this practice compared with countries with no legislation.    -squared 4,196 4,196 4,152 4,152 2,358 4,196 4,196 4,152 4,152 2,358 4,195 4,195 4,151 4,151 2,358 4,191 4,191 4,147  Finally, Table 7 shows the effect of displacement on patriarchal norms. IDP women are significantly less likely than non-IDP women with similar characteristics to disagree with individual statements such as 'families with men have less problems' and 'a good wife obeys her husband,' but there is no significant difference when asked about men as heads of household and men's last word in household decisions. Also, displaced women are significantly more likely than non-displaced women to disagree with all patriarchal statements at the same time. Both groups, however, are relatively small in the sample. In terms of behaviors, columns (18)-(22) indicate that displacement does not alter women's say in important household decisions. While these findings do not suggest changes in patriarchal gender norms, they reveal important changes around women's attitudes which might indicate slow shifts in intra-household dynamics.

Conclusions
The findings in this paper show mixed evidence regarding norm change. Specifically, gender norms that tolerate violence against women become less traditional with displacement, while those that limit women's economic opportunities become more rigid. These patterns are not necessarily surprising, as the evidence in this area is mixed. For example, Culcasi's study (2019) of Syrian refugees in Jordan reveals that women can work for pay outside the home, but they continue to be the main caregivers in the household. Similar experiences have also been reported for IDP widows in Nepal, Chechen refugees in the Czech Republic and IDP women in Darfur (De La Puente, 2011;Ramnarain, 2016;Szczepanikova, 2005).
The study also revealed a misalignment between attitudes and behaviors in specific domains of gender norms. For example, displacement is associated with less traditional patriarchal attitudes such as 'families with men have less problems' or 'a good wife obeys her husband,' but women's ability to decide about contraception and earnings decreases following displacement. These findings shed light on the complexity of gender norm change, which does not operate in a vacuum.
Moreover, change can be contradictory and improvements in one area do not imply that all others will automatically follow.
The findings of this study have implications for policy making. Greater access to reproductive health for displaced women could be ensured through universal health coverage schemes, for example, by including contraception in basic packages. Social assistance programs for IDPs could also address barriers to contraception by providing information to increase uptake and men's support for modern methods (Khan et al., 2016). While the expansion of reproductive health services might not guarantee a shift in gender norms, it might provide opportunities for women to overcome some of the challenges imposed by those norms (Malhotra et al., 2019).
In terms of economic opportunities, the findings show that, in situations of displacement, paid work does not necessarily translate into increased decision-making power. Hence, providing access to economic opportunities is not a guarantee that gender gaps will be reduced if men have full control of the gains, as determined by patriarchal norms. Economic empowerment programs 30 for displaced women, in particular, should have built-in guidelines for the protection of women and should engage men in promoting more gender-equitable relationships (Heilman & Barker, 2018;van der Gaag et al., 2019). Given the long-term nature of displacement in Colombia, it is important to build capacity for both displaced women and men to access economic opportunities, which can eventually replace social assistance. Program interventions should identify occupations and sectors where they could work given their skills and include support services, such as flexible working hours and childcare to address constraints related to domestic responsibilities.
Finally, the risk of sexual and GBV in situations of displacement has devastating impacts on the individual and on the household. This is a complex issue and requires a battery of interventions for prevention and response. Interventions that have worked in non-displaced settings such as strategies for addressing norms that condone violence against women, designing effective facilities and services for survivors, and engaging men and boys in prevention and response could be piloted in situations of displacement (see Jewkes et al., [2015] and Ellsberg et al., [2015]).
This study has limitations. Large-scale household surveys rarely include questions to identify and measure a norm. Moreover, the choice of proxy indicators for attitudes and behaviors is driven by data availability and some of them could arguably be mapped onto multiple spheres or domains of gender norms. Given data limitations, mixed methods studies can offer more insights as the economics literature on the gender-differentiated effects of conflict-induced displacement evolves.
The limited time span covered by the migration question might pose a challenge to capturing changes in gender norms. Nonetheless, the theoretical framework argues that displacement can accelerate change by, for instance, opening up economic opportunities for women in urban areas and reducing exposure to traditional structures that reinforce gender norms (Cislaghi & Heise, 2020;Harper et al., 2020;Marcus & Harper, 2015;Muñoz Boudet et al., 2013). The lack of data for men is another limitation. Gender norms are produced and reproduced by women and men; hence, the analysis omits part of the story. On the other hand, focusing on women's views and their own decision-making power within the household provides an overview of intra-household dynamics that could hint at men's and other household members' gender norms. Year (2015) 0