This is the accepted version of the article: Bueno, Xiana; Vidal Coso, Elena. «Vulnerability of Latin American migrant families headed by women in Spain during the Great Recession : a couple- level analysis». Journal of Family Issues, Vol. 40 Núm. 1 (2019). DOI 10.1177/0192513X18804584 This version is available at https://ddd.uab.cat/record/200913 under the terms of the license Vulnerability of Latin American Migrant Families Headed by Women in Spain during the Great Recession: A Couple-Level Analysis Xiana Bueno xianabuenogarcia@fas.harvard.edu Harvard University Elena Vidal-Coso Elena.Vidal@unige.ch Université de Genève Abstract One of the outcomes of the Great Recession has been the emerging pattern of households maintained exclusively by women. The analysis of intra-couple characteristics is crucial in the context of job segregation by gender and by immigrant origin, such as in Spain. Using the panel version of the Spanish Labor Force Survey from 2008 to 2015, we analyze the transition of dual earner couples to female-earner couples among Latin American and Spanish born households. Our results suggest that migrant vulnerability is not only a consequence of a segregated labor market by gender and origin but is also the result of the partners’ relative occupational and family characteristics. We show that, unlike Spanish-born couples, the risk of Latin American families becoming female-headed is higher for those couples in which the female partner has the weakest position in the occupational scale and for those with children in the household. Keywords : economic recession, Latin American migrants, Spain, gender, female heads of households, migrant families, labor market segregation 1 Introduction Spain, one of the main destinations of Latin American migration, was shaken by the Great Recession starting in 2008. One of the consequences of the financial crisis has been a change in families’ economic arrangements. Hence, households that are economically headed by women have increased significantly. In 2012, a European Commission report noted its concern in this regard, stating that the share of female breadwinner couples had increased to almost 10% during the economic downturn to the detriment of dual-earner couples. This was especially true in those countries that were worst and first hit by the crisis (Bettio et al., 2012: 12). This increase in the number of female-headed households after 2008 has been associated with higher male unemployment in a context of occupational segregation by gender. Has this effect been different for migrant and Spanish-born populations? Although male unemployment during the economic recession has impacted both Spanish and immigrant families, it has affected migrants more than native-born populations (Aysa-Lastra & Cachón, 2015) due to their more vulnerable labor situation. Many scholars have documented a dramatic growth in the unemployment gap between foreign-born and Spanish-born workers in Spain (Aysa-Lastra & Cachón, 2015; Cebolla et al., 2015; Muñoz-Comet, 2012; OECD, 2009; Reyneri, 2004; Reyneri & Fullin, 2011). Scholars have emphasized that the unemployment breach is a male phenomenon that reflects male migrants’ concentration in highly cyclical industries (i.e., construction) and their holding of manual-skilled and low-skilled jobs with less-secure labor conditions. Conversely, immigrant women, particularly Latin American women, remained employed in more stable, although still unskilled, sectors, such as domestic and care services (Gil & Vidal-Coso, 2015). On this basis, women can be considered more resilient to the downturn of the economy than men, even if the crisis has 2 diminished their prospects of upward mobility (Prieto-Rosas, 2016; Vidal-Coso & Vono, 2014). However, we know little about how the recession affected Spanish-born families and immigrant families differently according to couples’ labor force participation. Most previous studies have examined individuals using cross-sectional data. Previous literature has neglected couple-level dynamics in the labor market and its outcomes for immigrant families. In this work, we focus on the less-studied pattern of dual-earner couples that switched during the recession to a female-headed economic model due to the male partner’s job loss. By limiting households to those of couples with or without children, we obtain greater comparability between Spanish-born and migrant couples while avoiding living arrangements in which other adult members might contribute to the economy of the household. We focus on dual-earner couples for two reasons. First, as a consequence of the increase in female labor force participation during the previous decades, the dual-earner model had become the most common employment arrangement within couples in Spain at the onset of the crisis. The same was true for Latin American households due to the prevalence of females with certain human capital who migrated as economic pioneer migrants and who later regrouped with their partners in Spain. This contrasts significantly with the experience of Latin American migrants in the other main destination country, the U.S., where the educational level and labor force participation of Latin American women is lower (Aysa-Lastra & Cachon, 2015). Second, focusing on dual-earner couples allows us to test the effect of the relative resources of the male and female partners. In particular, we explore their educational and occupational advantage to predict their vulnerability and their capacity for resistance during the recession (Vitali & Mendola, 2014). We work with a restricted definition of female heads of household. Although the term female breadwinner in the literature often refers to single mothers or 3 women who out-earn their partners, we limit our definition to females who become the unique economic provider in the couple due to the unemployment of the male partner. Using the panel version of the Spanish Labor Force Survey (SLFS) from 2008 to 2015, we run logistic regression models with random effects to examine the transitions from Latin American and Spanish dual-earner couples to couples in which the male has lost his job and the female has become the sole economic provider. We aim to understand whether the more precarious position in the labor market and the more pronounced gendered pattern of occupational segregation among migrants (compared to the Spanish-born population) is linked to a higher propensity of becoming a female-earner couple. Furthermore, the structural segregation of the labor force by origin and gender may cause the paradoxical situation in which the resilience of Latin American women occurs even when the woman holds lower earnings potential and occupational prestige within the couple. Consequently, we also consider the effect of the partners’ relative positions in the socioeconomic structure on intra-couple employment dynamics. Thus, the following research questions arise. First, how different was the transition from dual- earner to female-headed households among Latin American and Spanish couples when controlling for differences in the partners’ positions in the occupational structure? Second, how is educational and occupational pairing associated with a higher share of female headship? The contribution of this research is to highlight how the dual segregation of the Spanish labor market by gender and origin reveals a major vulnerability of Latin American families during the Great Recession. In addition to this greater vulnerability of migrants compared to Spanish-born, we show that when females become unique providers, Latin American families rely on the income of the partner with the weakest earning power. In contrast, for Spanish couples, female heads of their households are more likely to be the 4 most economically powerful member of the couple. Although female headship mostly emerges for both origins as an unplanned arrangement during the period of economic uncertainty, the key point is that it implies increasing household inequality by origin. In effect, we observe that, unlike Spanish couples, men in Latin American couples suffer a higher vulnerability in the labor market even when they are in hypergamic partnerships in terms of occupation. Finally, we highlight the important role of gender within the migration experience, a perspective that has received less attention in the recent migration scholarship (Donato, et al., 2017). Our findings challenge the discussion of the implications of migration and labor market vicissitudes on the gender-egalitarian dynamics of migrant couples and families by reinforcing the role of women not only as economic providers but also as decision-makers. Latin American Migration during the Economic Expansion and Recession Spain has experienced a dramatic change in its labor market and economy in a very short period of time. The particularities of Latin American migration and the circumstances of its arrival and subsequent years in Spain make it a singular case with which to address our research interests. The labor force demand during the first few years of the 21st century transformed Spain into a country of immigration. The foreign-born population increased from 2.9% in 1998 to 13.1% in 2008. Since the early 2000s, Latin American migration has dominated the incoming flows to Spain, becoming the main continental group at approximately 40% of total migrants in the country. Their shared language and religion made them the favourites (Izquierdo, 2004) and migration regulations were designed accordingly, offering them special treatment with regard to legal status and resulting in a positive change in perceptions and attitudes towards them (Cea D’Ancona & Valles-Martínez, 5 2015). Hence, their cultural proximity made their integration to Spanish society easier compared to other origins (Aysa-Lastra & Cachón, 2015). During the early 2000s, the feminized Latin American flows were a solution to the labor demand for domestic workers and caregivers. This labor demand resulted from the rapid incorporation of Spanish women to the labor market, from 27.9% in 1975 to 82% in 2015 (OECD data), coupled with the familism of the Spanish Welfare Regime. According to the Spanish Labor Force Survey, in 2008, 27% of domestic workers were from Latin American. In subsequent years, economic expansion in the mid-2000s enabled an increase in family reunification processes (Domingo et al., 2012). During those years, Spain saw unprecedented low levels of unemployment and a flourishing economy, primarily in the services sector (Bernardi & Garrido, 2008). That period offered abundant and relatively well-paid manual job opportunities in male-dominated sectors such as construction, agriculture and the industrial sector during the growing period of the housing bubble. Family reunion led to a greater sex composition and occupational balance among Latin American families. By 2007, the year in which Spain registered its maximum immigration flow, many Latin American migrant families had settled and regularized their migrant status, living mainly in nuclear households in dual-earner partnerships (Bueno & De Valk, 2016; Domingo et al., 2012). However, previous studies have emphasized the complementarity of the Spanish workforce, resulting in a segregated labor market by gender and origin (Bernardi, et al., 2011). In general, migrants, especially low-skilled migrants, have a more precarious and marginal position in the labor market (Tilly, 2011). Several studies have explored the occupational mobility of Latin American female migrants in Spain and concluded that a “U”-shaped pattern was observed in their occupational trajectories, mainly from domestic services to other 6 occupations within the secondary segment of the labor market (Aysa-Lastra & Cachón, 2013; Fernández-Macías et al., 2015; Parella et al., 2013; Vidal-Coso & Miret, 2014). From 2008 onwards, Spain was dramatically affected by the Great Recession, resulting in rising levels of unemployment and the deceleration of migratory flows. A simple illustration is that the unemployment rate among migrants was 39.1% in Spain in 2011, which was twice the percentage for Spanish-born individuals (18.4%) (Aysa-Lastra & Cachón, 2012; Colectivo IOE, 2012). Unemployment has dramatically affected traditionally male occupational sectors (e.g., construction and industry) rather than female-dominated sectors (e.g., domestic work and caregiving) (Reyneri & Fullin, 2011). Gil & Vidal-Coso (2015) emphasize the differential impact of the crisis by economic sector, which in turn affected job losses by sex and national origin. Within this context, some authors have confirmed a “feminization” of job stability among foreign women in Spain (Muñoz-Comet, 2012; Zugasti, 2014). Indeed, as reported by Aysa-Lastra & Cachón (2015:5), the activity rate of Latin American women in Spain in 2011 reached 81% (30 points higher than Spanish women). This difference stems from the younger age structure of the migrant population (Cachón, 2009). However, the possible signs of occupational mobility observed in the female migrant population over time were truncated by the arrival of the Great Recession, leaving this population trapped in precarious jobs during the crisis (Vidal-Coso & Vono, 2015). The result is increasing levels of inequality and poverty among those at the lower part of the occupational ladder. Between 2005 and 2013, the proportion of couples under female headship (in which only the female partner works) (Figure 1) increased for Spanish couples from 5% to 11%. Even more significant was the increase for Latin American couples, from 5% to 22%. This outcome has been particularly important for Latin American families (Bueno 7 & Vidal-Coso, 2017) given the circumstances that surround their arrival and settlement process in Spain. -- Figure 1 about here – Little is known about the characteristics of these families that are economically maintained by women. The purpose of this study is to shed light on how both partners’ characteristics are related in these couples by comparing Latin-American and Spanish couples. We now present the theoretical position underlying our analysis. Theoretical Framework Labor Market Segregation by Gender and Origin Previous sociological literature has extensively theorized about inequality emerging from the intersection between race and gender in the labor market (Browne & Misra, 2003; Reskin, 2000). On the one hand, migrant communities have traditionally been subjects of what Portes (1978) defined as “objective vulnerability”, occupying more precarious positions in the social structure of the host countries. Latin American migrants in Spain are not an exception, and the upward occupation mobility for many of them leads them to face a glass door. In the Spanish context, compared to native-born individuals, immigrants suffer from a strong and persistent disadvantage in accessing skilled and stable occupations even after controlling for sociodemographic characteristics (Aysa-Lastra & Cachón, 2015; Bernardi et al., 2011). This situation allowed the emergence of a complementarity process between Spanish-born and immigrant populations (Bernardi & Garrido, 2008). On the other hand, the occupational sex segregation by which women are relegated to limited career paths and low-paid positions in the labor market has also been extensively studied (England, 1984; Maume, 1999; Padavic & Reskin, 2002; Reskin, 8 1984; 1993). Whereas gender occupational segregation affects both native-born populations and migrants, it becomes particularly clear with regard to female migrants (Vidal-Coso & Miret, 2014). Moreover, while human capital theory (Chiswick & Miller, 2002) has been one of the main theories to explain labor market insertion and occupational mobility, it often occupies a secondary role in the labor market performance of migrants due to segmentation dynamics (Piore, 1979). Other studies (Reyneri & Fullin, 2011) have noted how the labor market segregation by origin diminishes the career prospects of migrants, either because of the lack of legal recognition of their studies in their origin countries or simply because of the lack of language proficiency in the destination countries. Economic Recession and Migrant Families Previous studies have focused on how economic recessions impact immigrants, increasing their inequality and vulnerability (Chiswick et al., 2005; Cho & Newhouse, 2012; OECD, 2009; Papademetriou & Terrazas, 2009). In the U.S., Massey (2012) emphasizes that Latinos have been disproportionately affected by the Great Recession and have experienced the largest drop in wealth of any group. Unemployment affects not only individuals but also families’ wealth and wellbeing. Thus, the capacities and strategies of individuals to face and overcome the unexpected vulnerability derived from unemployment should be analyzed at the household level. Consequently, the study of couples’ labor market participation dynamics is key (Harkness, 2013; Harkness & Evans, 2011). In the context of recession, new gender roles may emerge within families. It has been shown that once husbands in male breadwinner families lose their jobs, the female partners may enter the labor force (Humphrey, 1940; Mattingly & Smith, 2010). This is known as the “added worker effect”, and its implicit assumption is that females are “secondary workers” and that their labor supply is a response to the unemployment 9 of their partners. In many contexts, with the prevalence of dual-earner models and the permanent attachment of females to the labor market, this theoretical framework is unable to explain family employment dynamics. Bettio et al. (2012) emphasize that female economic headship is often the result of the onset of an economic crisis within a gendered labor market when unemployment impacts more male-oriented jobs. Economic recessions have led to a significant proportion of families headed by women (Harkness, 2013; Vitali & Mendola, 2014). Behind a female head of household, there is often a male with a more disadvantaged position in the labor market (Drago et al., 2005). In this sense, Chesley (2017) suggests that female breadwinner couples are, in many cases, the result of a non-planned work-family arrangement. Our analysis draws on this perspective. Couples’ Relative Educational and Occupational Statuses Recent studies focusing on female headship take into account females’ relative position in the socio-economic structure and therefore their capacity for resilience in employment. Esteve, García-Román & Permanyer (2012) have shown that the gender- gap reversal of education implies a reduction of the traditional educational hypergamy between partners, which has been balanced by an increase of educational hypogamy and educational homogamy. Patterns of educational and occupational assortative mating are expected to affect the probabilities of female household headship (Klesment & Van Bavel, 2015). For these authors, if the female is more educated than the male, she may have a stronger position in the labor market. However, they also admit that due to the persistence of gendered occupational segregation and to the motherhood penalty, the female educational advantage is not necessarily translated in the reversal of the gender gap in terms of occupational status and earnings. Kramer et al. (2015) found that in the U.S., women in female breadwinner families have a significantly higher level of 10 education than their partners compared to male breadwinner or dual-earner couples. Similarly, Chesley (2011) found that men in female breadwinner families tend to have less education than their partners. In summary, female breadwinner families due to partner unemployment are necessarily dependent on the woman’s position in the labor market and earning potential regardless of whether she is the more economically powerful member of the couple (Drago et al., 2005). Thus, the degree of economic vulnerability within this typology of families is also associated with educational and occupational pairing. Finally, although Klesment and Van Bavel’s study notes major transitions to female- headed households for hypogamous couples, they also state that the proportion of female breadwinners increased after 2008 for couples who were low educated, particularly in countries that were hit badly by the crisis (Klesment & Van Bavel, 2015:29). Similarly, Drago et al. (2005) found that men and women in economic female breadwinner models exhibit lower socio-economic status than those in the equity type. In this regard, female headship is conceived as an unplanned strategy of family resilience during a period of economic uncertainty. Indeed, individuals exercise their resilience to protect or minimize the effects of their vulnerable condition. Their capacity for action depends on a “field of possibilities” delimited by their position in the social structure (Brand & Jax, 2007). Although the recession’s consequences on families have affected both Spanish-born individuals and immigrants, the disadvantaged position in the social structure and lack of social networks among an important part of the migrant population has left them in a situation of relative poverty and material deprivation (Mahía & Arce, 2014). A resilient strategy of immigrant women to compensate for the partner’s job loss may entail accepting or maintaining precarious employment (Papademetriou & Terrazas, 2009). Furthermore, we cannot neglect the possibility that 11 women’s resilience in employment in times of crisis may be associated with employers’ preference for hiring immigrant women, who are considered cheaper, more docile, and flexible workers in low-paid and unskilled jobs (Donato & Bankston, 2008). Hypotheses Our main hypothesis is that within the frame of a structural segmentation of the workforce by origin and gender, Latin American dual-earner couples face a greater vulnerability than dual-earner Spanish-born couples due to more pronounced gendered patterns of labor market segregation for migrant groups. We aim to test three specific hypotheses: H1. The higher likelihood of Latin American women to transition from a dual- to a female-earner model compared to Spanish-born women is the result of the combined lower socio-economic status of both partners. Therefore, it is only partially explained by women’s disadvantaged selectivity in terms of educational level and occupational status. We expect women’s greater probability to be driven by the over-representation of Latin-American men in highly cyclical industries and their greater participation in low-skilled jobs with less-secure labor conditions. H2. A Spanish woman with more education than her partner will increase the odds that she becomes the sole earner of the couple. In contrast, we hypothesize that among Latin American couples, educational assortative mating will have no effect on the probability of transition into female headship. This is a consequence of the segmentation dynamics because immigrants’ labor market performance is not correlated to their human capital to the same extent as it is for Spanish-born individuals. H3. Spanish women in occupational hypogamous unions are significantly more likely to become the unique economic provider than in homogamous or hypergamous 12 partnerships. On the contrary, more pronounced gendered patterns of labor market segregation for migrants lead to a major risk of female headship for hypergamous Latin American couples. This expectation is the result of the pronounced occupational concentration among Latin American women in highly feminized and low-skilled labor positions that are less affected by the crisis, whereas men are more evenly distributed along the occupational scale but in cyclical industries that are rarely affected by unemployment. Consequently, transitions to female headship are not only more common for Latin American than for Spanish families, but the consequences of male unemployment in terms of economic vulnerability are also more accentuated among the former. The safety net in Spanish female-earner couples is stronger because they may rely on the partner who is better positioned in the labor market and who has higher earning potential. On the contrary, the economic disadvantages are greater among Latin American female-headed families depending on the woman’s income when she holds the lower occupational status and earning potential. Data and Methods Data. The data come from the 2008-2015 Spanish Labor Force Survey (SLFS). This survey is a unique source for studying yearly household composition and the characteristics of household members in the inter-census period (Garrido et al., 2000). The SLFS is conducted on quarterly samples of approximately 65,000 households, including information on approximately 200,000 individuals. The SLFS interviews every household during six consecutive trimestral observations. Because the SLFS is a rotating panel, in each wave, 1/6 of the sample is substituted by households of similar characteristics, leaving 5/6 of the sample remaining and resulting in the formation of a 13 new panel after six trimesters. To ensure representativeness, the SLFS demands a specific methodology that takes into account its panel data nature 1. Consequently, we follow women and their partners over 18 consecutive months (six trimesters) to observe changes and intra-couple dynamics in terms of employment status over time. Sample characteristics. The sample is limited to employed married or cohabiting women in heterosexual couples, living with or without children, between the ages of 20 and 60 and whose partners are also aged 20-60 and employed 2. Table 1 describes the characteristics of the sample. The analysis excludes those unions that moved to employment arrangements other than female-earner couples, such as those transitioning to male-earner couples or to couples in which neither partner works. Similarly, the observations of the individuals are truncated at the time at which this transition occurs. Although an individual woman may become a female earner more than once during the observational period, the longitudinal methodology used in the analysis takes into account the panel structure of the data and the fact that we are following individuals’ trajectories during six observations. These steps yield a sample of 76,456 Spanish couples (250,981 observations), of whom 9.2% experienced a transition to female headship during the crisis period of 2008-2015. For Latin American women, the subsample comprises 3,436 couples (8,861 observations); 17% of them became a female-earner couple during the observed period. One aspect to consider in this analysis is that women might experience the transition to female headship more than once during the observational period. The average number 1 Given our interest in the probability of transition to female-earner couples and that there is no attempt to estimate the number of transitions, according to INE (Spanish National Statistical Institute) indications (1989), weightings have not been used. If they were to be used, it would be necessary to use a longitudinal weight for the same individual throughout the observation period, which is not offered by the INE. 2 Note that we only consider partners aged 20-60 to exclude transitions to inactivity due to early retirement, which are very common in Spain for workers aged 61-64. 14 of transitions per person is 1.08 among Spanish-born women and 1.09 among Latin American women. -- Table 1 about here -- Method. We longitudinally explore the economic arrangement within the couples by focusing on the transitions (y=1) of women from a dual-earner couple in wave (t) to a female-earner couple in the following wave (t+1). Transition does not occur (Y=0) for those women who remain in dual-employed couples. We use pooled models that include Spanish-born couples and Latin American couples. To test our hypotheses, logistic regression models with random effects that take into account the multiple waves/observations (t) of each individual (i) are employed. In these models, we assume that the aggregated heterogeneity due to the multiple observations t of every individual i is randomly correlated with the explanatory variables. In Model 1, the main explanatory variable is the woman’s birthplace, including Spanish-born or Latin American-born women. We gradually include control variables in the subsequent models. In Model 2, we include a set of control variables related to the women’s socio-demographic profile: age, age squared, maximum educational level attained, occupational category in t, type of contract and whether the job is in the public or private sector. Model 3 includes family characteristics, the origin of the male partner, and the presence of children in the household. The results in Model 4 are controlled by the couple’s educational and occupational homogamy and the male’s labor market characteristics (occupational category, type of contract (temporary or permanent), and job sector (public or private)) in t. In a second step, to capture differences between Spanish-born and Latin American-born female-earner households at the intra-family level, we run models allowing for several interactions. We first interact women’s birthplace and occupation. Second, we check the interactions between women’s 15 birthplace and family characteristics (children and migrant partner) and then between women’s birthplace and partners’ homogamy (educational and occupational). To ease the interpretation of the resulting interactions, predictive margins are calculated to display the specific effect of these characteristics. The models with interaction terms are not shown, but they are available from the authors upon request. Transitions to Female Headships In this section, we longitudinally explore economic arrangement transitions from households in which the two partners work to households supported exclusively by the female partner. There are three initial economic arrangements from which a couple can become female sustained: dual-income couples, male-breadwinner couples and couples in which neither partner works. We calculated that 66% and 62% of transitions to female-earner couples come from dual-earner couples for Spanish and Latin American couples, respectively. These results reinforce the research interest of focusing exclusively on the movement from families initially maintained by both employed partners. Table 2 presents the likelihood of transitioning from a dual-earner couple to a female- earner couple, which, in other words, refers to the likelihood that the male in the couple loses his job. Model 1 shows that compared with Spanish couples, Latin American couples are almost four times more likely to become female-earner couples. The overall effect of country of birth and the large and highly significant coefficients are in line with the cross-sectional data shown in Figure 1. The differences between Spanish-born and Latin American women decrease considerably after controlling for the female educational and occupation characteristics (Model 2). Women employed in less skilled occupations in temporary jobs and in the private sector exhibit a higher risk of transitioning to a female-headed employment arrangement. While this is the pattern for 16 both Latin American and Spanish women, it is even more evident among the former, and it is especially significant for those in care, cleaning and domestic occupations (Figure 2). Consequently, the results highlight the importance of the concentration of Latin American women in unskilled and highly feminized labor positions, although less affected by the economic recession, in explaining their major resilience in employment. As we will see next, this is related to the occupational pairing of the couple and their higher vulnerability. However, the coefficient in Model 2 remains higher and statistically significant for Latin American couples in relation to Spanish-born couples. Latin American women are almost 2.5 times more likely to become part of female- earner couples. This result reinforces our first hypothesis that it is not only their disadvantaged allocation within the female labor force that explains our outcome. Considering the other covariates, we observe that the probability of transition increases with age. In terms of educational attainment, women with higher education exhibit lower probabilities of becoming the sole earner, in line with the human capital theory and previous studies on female breadwinners (Drago et al, 2005). – Table 2 about here – – Figure 2 about here – Once we add union and family characteristics (Model 3), interesting differences arise. Latin American’s risk of transitioning to a female-headed employment partnership is still being significantly higher to that of Spanish-born women, although differences between Spanish-born and Latin American women decrease. Moreover, the model suggests that women in a union with a migrant male are 2.5 times more likely to become a female head in their households compared to those in a union with a Spanish man. The results also suggest that the presence of children in the household is associated with a lower risk of transitioning to a female-headed couple. At 17 this point, to pursue the goal of testing migrants’ vulnerability, we conducted interactions between women’s birthplace and their family characteristics (children and partner’s origin) for their expected probabilities of female headship. Predictive margins from these interaction effects are displayed in Figure 3. The results indicate that Spanish women coincide with the pattern of showing a higher likelihood for childless women. Conversely, Latin American women are more likely to become female sole earners when they have children in the household (Figure 3A). Moreover, although no differences were found among Spanish-born and Latin American women in partnership with a Spanish-born male, having an immigrant partner clearly increased their probability of becoming the sole earner of the couple. The predictive margins show that while the likelihood for Spanish women is slightly above 6%, it reaches 9% for Latin American women (Figure 3B). This result is consistent with the work of Domingo et al. (2014), who showed that Latin American women in unions with Spanish males belong to couples with higher socioeconomic positions than endogamous couples. –Figure 3 about here – However, we pursue the idea that this result is the effect of the more precarious labor position of Latin American men compared with Spanish men in the Spanish labor market. Indeed, the most important outcome is that when the male characteristics are included in the analysis (Model 4), the differences between Spanish and Latin American couples completely disappear, and statistical significance is lost. Therefore, our first hypothesis that the higher labor-market vulnerability of Latin American men than of Spanish men causes Latin American women to have a resilient role as the unique economic provider is fully corroborated by these results. Because the transition from a dual-earner couple to a female-headed couple implies that the man loses his job, the outcome depends strongly on the former occupation of the male partner. Indeed, the 18 coefficients indicate that a female-headed couple is more probable when the male was employed in manual occupations, particularly less-skilled positions. As expected, a male partner previously working in temporary jobs and in the private sector increases the likelihood of the couple of becoming a female-earner couple. Furthermore, as noted, a higher likelihood is observed for women whose partner is of immigrant origin compared to those with a Spanish-born partner (Figure 3). This occurs even after controlling for the partner’s previous job characteristics, thus reinforcing the idea of a major vulnerability in terms of the unemployment of immigrants compared with Spanish-born couples. Interestingly, once we eliminate the effect of the disadvantaged labor position of Latin American partners in Model 4, the differences in the probabilities of female headship with and without children by origin disappear (Figure 3). These results reinforce our initial expectation regarding the strengthened vulnerability to which Latin-American families are exposed. Model 4 also confirms our second hypothesis that the probability of Spanish women becoming the sole economic provider of their family increases when the woman is more educated than the man. This expected result is in line with the existing literature on the effect of the gender-gap reversal in education (Klesment & Van Bavel, 2015) on women’s less-precarious labor position relative to their partners. However, the predictive margins displayed in Figure 4A contradict our second hypothesis on the lack of effect of female educational advantage for immigrant couples. Indeed, regardless of the women’s origin, female educational hypogamy increases the odds that a woman becomes the only economic provider. –Figure 4 about here – However, as noted, among migrant populations, occupational level is not necessarily a reflection of educational level. Therefore, we have also included in Model 4 the 19 couples’ occupational homogamy. As we observed with the educational level, occupational hypogamy is associated with a greater risk of transitioning toward a female-headed couple. In contrast, as expected, when the male partner has a better labor position than the female partner, the couple is less likely to become female headed. Finally, homogamous couples are the least likely to shift toward female headship. A further revealing check on how this variable interacts with national origin (Figure 4B) shows that, unlike Spanish couples, the likelihood for Latin-American couples of transitioning to a female headship in a hypergamous situation (when the man has a better occupational position than the woman has) is higher compared to hypogamous and homogamous couples. This unexpected finding underpins the more vulnerable position of Latin-American families during the crisis compared to Spanish families. Discussion and Conclusions This study contributes to our understanding of how the economic recession and subsequent male unemployment affected families in Spain differently. It focuses on exploring the transitions from dual-earner households to female-headed households among Spanish and Latin American migrant couples due to the job loss of the male partner. Whereas previous literature has focused on the impact of unemployment on individuals, we frame male unemployment in its family dimension. In effect, unemployment affects not only unemployed individuals but also their family members, and the strategies and capacities for facing it should consequently be analyzed at the household level. In this regard, the focus on dual-earner couples allows us to consider not only the individuals' characteristics but also the effect of the relative resources of the male and female partners, particularly the educational and occupational advantage, in predicting their vulnerability and their capacity for resistance in the context of an economic recession. Using panel data, we are able to observe changes in intra-couple 20 dynamics in terms of employment status over time, allowing us to consider the partners' previous employment profile. Our results suggest an increase in household inequality by origin during the economic crisis. Thus, our aim of understanding how the structural segmentation of the workforce by origin and gender in Spain has translated into greater vulnerability within the families of Latin American immigrants during the recession has been fully addressed. First, we confirm that Latin American females are more likely to become the sole economic providers than Spanish females. Our research has demonstrated the effect of the combined disadvantaged allocation of immigrant men and women in the workforce. After controlling for women's occupational characteristics, our analysis revealed that their higher likelihood is only partially related to their higher concentration in feminized and unskilled labor positions that are less affected by the economic recession. Furthermore, their male partner’s occupational characteristics make Latin American women more likely to become female heads in their households. Therefore, the combination of occupational segmentation by origin and more pronounced gendered patterns of labor market segregation contribute to the resilient role of Latin American women. The high unemployment rates among male migrants and the less negatively affected job sectors among female migrants (i.e., domestic work, cleaning services, and caregiving) explain the more resilient position of women in the labor market relative to their male partners. However, this stable female position is not often synonymous with a non-precarious position. In effect, our results indicate that transitions toward female headship are higher for women employed in low-skilled jobs. What is less obvious from our analysis is to what extent a resilient strategy of immigrant women to compensate the partner’s job loss may entail accepting or maintaining precarious jobs. As Vidal-Coso & Vono (2014) show, the immediate outcome of the recession on female migrants was to 21 leave them trapped in low-scale job sectors. Another question that arises from our analysis is how this additional vulnerability within immigrant families is associated with employers’ preference to hire immigrant women during the crisis because they are considered cheaper, more docile, and flexible workers in low-paid and unskilled jobs. Second, our research revealed the paradoxical situation in which the resilient role of Latin American women occurs even when the woman holds lower earning potential and occupational prestige within the couple. In addition to being more affected by male unemployment compared to Spanish families, Latin American dual-earner families face greater constraints in their capacity to counteract economic uncertainty. Therefore, although the recession has affected both Spanish-born and immigrant individuals, it increased household inequality by origin. Spanish women in female-headed couples are more likely to be the partner with the higher relative resources in terms of education and occupational category. This is an expected result due to the reversal of gender inequality in education and the increase of labor opportunities. In contrast, Latin American female heads of household are more likely to be the partner with the weakest labor market position regardless of their relative educational advantage. This is a consequence of their occupational concentration in highly feminized and low-skilled labor positions, mainly in domestic and care services, whereas Latin American men are more evenly distributed along the occupational scale. It is also important to note the main limitation of this study. Despite using a longitudinal approach, our analysis is limited to couples' transitions during 18 months. The availability of longer employment trajectories would facilitate other analytical approaches, such as survival analysis, which would allow us to follow female-headed couples over time and determine their temporary or persistent nature. Knowledge of the duration of female-headed arrangements would shed light on the understanding of their 22 economic vulnerability in terms of gender roles. Indeed, we speculate that gender roles in households where the woman is the sole employed member might be altered. Implications for Gender Roles and Migrant Families How do the migration experience and the economic crisis affect gender-role dynamics in Latin American families in Spain? While the present analysis is unable to properly answer this question, we highlight it as a concern that motivates our study. We argue that the gender roles of Latin American couples in Spain might follow three different stages. First, Latin American female migrants who migrated before their partners did empowered their role within the household, not only because they were the ones who would later regroup with their male partners but also by having started to build social networks in Spain earlier, which is useful for the labor market, for the housing market, for their knowledge of immigrant legal regulations, and for interacting with the Spanish- born population. We might expect that the longer the period between their arrival and the reunification with their husbands, the stronger their social networks become. Consequently, we assume that during this initial period of sole migration, the hetero- patriarchal gender roles that these couples might have in origin become weaker, although they do not completely disappear (Pedone, 2010). Second, it is possible that once the male partners were regrouped in Spain, a (attempt at) restoration of the former more traditional gender roles might occur for these couples. This would especially be the case for women who regroup with their husbands and children, thereby recovering their role in the reproductive sphere. Finally, the economic crisis implied a new strike to gender roles. The dramatic male-job destruction put women again in a leadership position of becoming the unique economic providers in the household. In this respect, we hypothesize that the experience of Latin American migrants in Spain contributes to greater levels of gender egalitarianism in their households by reinforcing the role of 23 women not only as economic providers but also as decision-makers. However, we also recognize that women may see their role reinforced de facto, but this does not necessary mean that gender dynamics within the household change (Drago et al., 2005). This gender pattern would contrast significantly with the experience of Latin American migrants in other destinations such as the U.S. An additional implication is that this process may bring new patterns of circular/transnational Latin American migration. Unemployed husbands may return to their origin countries while their wives, still employed, remain in Spain. Finally, we would like to draw attention to the risk that the greater vulnerability of Latin American migrant couples might entail for children. Our results suggest that vulnerability is higher for migrant couples with children. This result must not be neglected at the institutional level because it has direct implications for policy-makers. Providing greater levels of labor and social protection for low-income families, especially those with dependent children, represents, from a policy point of view, a challenge for legislators and labor-market institutions. In this sense, a policy change would have implications, first, for reducing vulnerability and inequality among families generated by the segmentation of the labor market by gender and origin; second, it would put the brakes on the downward occupational mobility experienced by migrant families. Policy changes must include changes in migratory policy regulations that facilitate social and labor-market integration and do not contribute to punishing families with members in an unauthorized situation. References Aysa-Lastra, M. & Cachón, L. (2012). Latino immigrant employment during the Great Recession: A comparison of the United States and Spain. Norteamérica 7 (2): 7–42. 24 Aysa-Lastra, M. & Cachón, L. (2013). Movilidad ocupacional segmentada: el caso de los inmigrantes no comunitarios en España. Revista Española de Investigaciones Sociológicas 144: 23-47. Aysa-Lastra, M. & Cachón, L. (Eds.). (2015). Immigrant vulnerability and resilience. Comparative perspectives on Latin American immigrants during the Great Recession (vol. 11). Cham: Springer international publishing. Bettio, F., Corsi, M., D’Ippoliti, C., Lyberaki, A., Lodovici, M. S., & Verashchagina, A. (2012). The impact of the economic crisis on the situation of women and men and on gender equality policies. Brussels: European Commission. Bernardi, F. & Garrido, L. (2008). Is there a new service proletariat? Post-industrial employment growth and social inequality in Spain. European Sociological Review, 24(3), 299-313. Bernardi, F., Garrido, L. & Miyar, M. (2011). The recent fast upsurge of immigrants in Spain and their employment patterns and occupational attainment. International Migration 49: 148-187. Brand, F. S., & Jax, K. (2007). Focusing the meaning(s) of resilience: Resilience as a descriptive concept and a boundary object. Ecology and Society, 12(1), 23. Browne, I., & Misra, J. (2003). The intersection of gender and race in the labor market. Annual Review of Sociology, 29(1), 487–513. Bueno, X. & De Valk, H. (2016). Arreglos familiares de los inmigrantes latinoamericanos en España. ¿Cambios en tiempos de crisis? Notas de Población, 102: 123-148. 25 Bueno, X. & Vidal-Coso, E. (2017). Households economically headed by women in times of expansion and crisis: the case of Latin American migrants in Spain. Revista de Historiografía, 26: 273-296. Cachón, L. (2009). La "España inmigrante": Marco discriminatorio, mercado de trabajo y políticas de integración. [The “Immigrant Spain”: discriminatory frame, labour market and integration policies] Barcelona: Anthropos. Migraciones Proyecto Editorial, 66. Cea D’ancona, M.Á. & Valles-Martínez, M.S. (2015). Changes in the perception of Latin American immigrants in host countries during the Great Recession. In: Aysa- Lastra M. & Cachón L. (eds) Immigrant vulnerability and resilience. International perspectives on migration, vol 11. Cham: Springer. Cebolla-Boado, H., Miyar-Busto, M., & Muñoz-Comet, J. (2015). Is the Spanish recession increasing inequality? Male migrant-native differences in educational returns against unemployment. Journal of ethnic and migration studies, 41(5), 710– 728. Chesley, N. (2011). Stay-at-home fathers and breadwinning mothers: Gender, couple dynamics, and social change. Gender & Society, 25: 642-664. Chesley, N. (2017). What does it mean to be a “breadwinner” mother? Journal of Family Issues, 38(18), 2594–2619. Chiswick, B.R., Lee, Y. & Miller, P. (2005). A longitudinal analysis of immigrant occupational mobility: A test of the immigrant assimilation hypothesis. International Migration Review, 39 (2): 332-353. 26 Chiswick, B.R. & Miller, P. (2002). Immigrants earnings: language skills, linguistic concentrations and the business cycle. Journal of Economy and Society, 48(3):454– 465. Cho, Y. & Newhouse, D. (2012). How did the Great Recession affect different types of workers? Evidence from 17 middle income countries. World Development, 41: 31– 50. Colectivo IOE, (2012). Impactos de la crisis sobre la población inmigrante [The impact of the crisis over migrant population] Organización Internacional de Migraciones. Domingo, A., Bueno, X. & Esteve, A. (2014). El rapto de las latinas: Migración latinoamericana y mercado matrimonial en España, in Cosio-Zavala, M.E. & Rozée Gómez, V. (Coords): Género en movimiento: familias y migraciones [Gender in movement: families and migrations]. México, El Colegio de México. Domingo, A., Treviño, R. & Bueno, X. (2012). Género y estrategias familiares en la migración interna e internacional en Cataluña. Revista de Demografía Histórica, XXIX, II, 2012, segunda época: 59-86. Donato, K. M., & Bankston, C. L. (2008). The origins of employer demand for immigrants in a new destination: The salience of soft skills in a volatile economy. In D. S. Massey (Ed.), New faces in new places: The changing geography of American immigration. New York: Russell Sage Foundation: 99–123. Donato, K. M., Enriquez, L. E., & Llewellyn, C. (2017). Frozen and stalled? Gender and migration scholarship in the 21st century. American Behavioral Scientist, 61(10): 1079–1085. Drago, R., Black, D. & Wooden, M. (2005). Female breadwinner families. Their existence, persistence and sources. Journal of Sociology, 41(4): 343-362 27 England, P. (1984). Wage appreciation and depreciation: A test of neoclassical economic explanations of occupational sex segregation. Social Forces, 62(3), 726- 749. Esteve, A., García-Román, J. & Permanyer, I. (2012). The gender-gap reversal in education and its effect on union formation: The end of hypergamy? Population and Development Review, 38(3), 535-546. Izquierdo, A. (2004). Los preferidos frente a los extranjeros permanentes: La inmigración marroquí en los inicios del siglo XXI. In López, B. & Berriane, M. (Eds.): Atlas de la inmigración marroquí en España [Atlas of the Morroccan migration in Spain]. Madrid: Universidad Autónoma de Madrid: 112–114. Fernández-Macías, E., Grande, R., Poveda, A. del R., & Antón, J.-I. (2015). Employment and occupational mobility among recently arrived immigrants: The Spanish case 1997–2007. Population Research and Policy Review, 34(2), 243–277. Garrido, L., Requena, M. & Toharia, L. (2000). La Encuesta de Población Activa desde la perspectiva de los hogares. Revista de Estadística Española, 42(146): 115-152. Gil, F. & Vidal-Coso, E. (2015). Inmigrantes extranjeros en el mercado de trabajo español: ¿más resilientes o más vulnerables al impacto de la crisis? Migraciones, 37: 97-123. Harkness, S. (2013). Women, families and the 'great recession' in the UK. In: Ramia, G., Farnsworth, K. and Irving, Z., (Eds.) Social Policy Review 25: Analysis and Debate in Social Policy. Bristol: Policy Press. Harkness, S., & Evans, M. (2011). The employment effects of recession on couples in the UK: Women’s and household employment prospects and partners’ job loss. Journal of Social Policy, 40 (4): 675–93. 28 Humphrey, D. (1940). Alleged "additional workers" in the measurement of unemployment. The Journal of Political Economy, 48(3): 412–419. Klesment, M., & Van Bavel, J. (2015). The reversal of the gender gap in education and female breadwinners in Europe. Families and Societies, Working Paper Series, 26. Kramer, K.; Kelly, E.; & McCulloch, J. (2015). Stay-at-home fathers: Definition and characteristics based on 34 years of CPS data. Journal of Family Issues, 36 (12): 1651-1673. Mahía, R. & Arce, R. (2014). Pobreza de la población extranjera en España. In Aja, E., Arango, J. & Oliver-Alonso, J. (Eds.): Inmigración y Emigración: mitos y realidades [Immigration and Emigration: myths and realities], Anuario de Inmigración en España (2014 edition), CIDOB, Barcelona. Mattingly, M. J. & Smith, K. E. (2010). Changes in wives’ employment when husbands stop working: A recession-prosperity comparison. Family Relations, 59 (4): 343–57. Massey, D.S. (2012). Immigration and the great recession. Stanford: Standford Centre on Poverty and Inequality. Maume Jr, D. J. (1999). Glass ceilings and glass escalators: Occupational segregation and race and sex differences in managerial promotions. Work and occupations, 26(4), 483-509. Muñoz-Comet, J. (2012). Evolución del empleo y del paro de las mujeres inmigrantes en el mercado de trabajo español. El impacto de la actual crisis económica. Cuadernos de Relaciones Laborales, 30(1), 115–137. 29 OECD (2009). The labour market integration of immigrants and their children. Key findings from OECD country reviews. High-level policy forum on Migration. Paris, 29–30 June. Padavic, I., & Reskin, B. F. (2002). Women and men at work. Pine Forge Press. Papademetriou, D.G. & Terrazas, A. (2009). Immigrants and the current economic crisis: Research evidence, policy challenges, and implications, Washington, DC: Migration Policy Institute. Parella, S., Petroff, A. & Solé, C. (2013). The upward occupational mobility of immigrant women in Spain. Journal of Ethnic and Migration Studies, 39:9, 1365- 1382. Pedone, C. (2010). “Varones aventureros” vs. “madres que abandonan”: reconstrucción de las relaciones familiares a partir de la migración ecuatoriana. Revista Interdisciplinar da Mobilidade Humana, 16(30). Piore, M. (1979). Birds of passage: migrant labour and industrial societies. Cambridge: Cambridge University Press. Portes, A. (1978). Introduction: Toward a structural analysis of illegal (undocumented) immigration. International Migration Review, 12(4), 469–484. Prieto-Rosas, V. (2016). Las consecuencias de quedarse y de volver en el empleo de los uruguayos que migraron a España. Notas de Población, 43: 149-177. Reskin, B. (1984). Sex segregation in the workplace. Trends, explanations, remedies. National Academy Press, 2101 Constitution Avenue, NW, Washington, DC, 20418. Reskin, B. (1993). Sex segregation in the workplace. Annual review of sociology, 19(1), 241-270. 30 Reskin, B.F. (2000). Getting it right: sex and race inequality in work organizations. Annual Review of Sociology, 26:707–9 Reyneri, E. (2004). Immigrants in a segmented and often undeclared labour market. Journal of Modern Italian Studies, 9(1): 71–93. Reyneri, E. & Fullin, G. (2011). Labour market penalties of new immigrants in new and old receiving west European countries. International Migration 49 (1): 31-57. Tilly, C. (2011). The impact of the economic crisis on international migration: A review. Work, Employment and Society, 25(4): 675–92. Vidal-Coso, E., & Miret-Gamundi, P. (2014). The labour trajectories of immigrant women in Spain: Are there signs of upward social mobility? Demographic Research, 31(13), 337–380. Vidal-Coso, E. & Vono, D. (2014). Equilibrios ocupacionales en los hogares de latinoamericanas en España durante las etapas de expansión y de crisis económica, 1999-2011.In Cosio-Zavala, M.E. & Rozee, V. (Eds.): Género en movimiento: familias y migraciones [Gender in movement: families and migrations], México, Publicaciones El Colegio de México. Vidal Coso, E. & Vono, D. (2015). Entrapped as domestic workers? The effect of economic context on work opportunities. In Domingo, A., Sabater, A. and Verdugo, R. (Ed.) Demographic analysis of Latin American immigrants in Spain, Applied Demography Series 5. Springer International Publishing: 155–80. Vitali, A. & Mendola, D. (2014). Women as main earners in Europe. ESRC Centre for Population Change, Working Paper 56. Zugasti, N. (2014). Incidencia de la crisis en el empleo extranjero. Evidencias a partir de las transiciones laborales de trabajadores indefinidos. Papers, 99(2): 285-306. 31 32 Table 1: Sample characteristics of dual-earner couples included in the analysis at the first time of observation by birthplace Spanish Latin American Age (median) 41,0 37,3 Female's educational level Primary or less 26,4 29,0 Secunary 34,5 44,0 Universitary 39,1 27,1 Female's occupation Managerial and professional 42,0 21,7 Manual skilled 5,8 5,4 Administrative and secretarial 17,0 8,7 Sales and costumers 11,1 10,3 Hotels and restaurants 3,0 8,5 Care, cleanning and domestic 18,3 39,2 Other unskilled 2,9 6,2 Job contract Permanent 85,9 75,0 Temporary 14,1 25,1 Job sector Privat 70,9 92,1 Public 29,1 7,9 Children Not children 54,4 48,9 Children 45,6 51,1 Partner birthplace Spanish born partner 97,0 42,0 Immigrant partner 3,0 58,0 Educational homogamy Same educational level 43,2 49,9 She more educated 35,1 30,9 He more educated 21,7 19,2 Ocupational homogamy Occup. Homogamy 30,5 36,6 Occup. hipogamy 32,2 19,3 Occup. hipergamy 37,3 44,0 Partner's occupation Managerial and professional 41,8 26,85 Manual skilled 38,27 41,54 Administrative and secretarial 6,78 4,22 Sales and costumers 3,95 3,95 Hotels and restaurants 2,38 5,53 Care, cleanning and domestic 2,52 5,17 Other unskilled 4,3 12,74 Partner's job contract Permanent 90,21 74,9 Temporary 9,79 25,1 Partner's job sector Privat 79,43 91,01 Public 20,57 8,99 Observations, not weighted 250.981 8.861 Number of individuals, not weighted 76.456 3.436 33 Source: Spanish LFS, 2008-2015. Table 2: Odds ratio of the likelihood of transitioning from a dual-earner household to a female-earner household. Pooled models for Spanish and Latin Americans. Model 1 Model 2 Model 3 Model 4 Exp (ß). S.E. Exp (ß). S.E. Exp (ß). S.E. Exp (ß). S.E. Female's birthplace (ref: Spain) Latin America 3,85 *** 0,30 2,44 0,19 *** 1,39 0,13 *** 1,1 0,08 ns. Female's age 0,85 0,02 *** 0,86 0,02 *** 0,93 0,02 *** Female's age2 1,00 0,00 *** 1 0 *** 1,00 0,00 *** Female's educational level (ref: Primary or less) Secunary 0,53 0,02 *** 0,53 0,02 *** 0,62 0,03 *** Universitary 0,41 0,03 *** 0,4 0,03 *** 0,55 0,03 *** Female's occupation (ref: Managerial and professional) Manual skilled 1,44 0,12 *** 1,44 0,12 *** 1,19 0,09 ** Administrative and secretarial 1,04 0,06 ns. 1,04 0,06 ns. 0,95 0,05 ns. Sales and costumers 1,55 0,11 *** 1,54 0,11 *** 1,32 0,09 *** Hotels and restaurants 1,58 0,16 *** 1,55 0,15 *** 1,22 0,11 ** Care, cleanning and domestic 2,12 0,13 *** 2,09 0,13 *** 1,29 0,08 *** Other unskilled 2,31 0,22 *** 2,29 0,22 *** 1,38 0,12 *** Temporary job (ref: permanent job) 1,47 0,07 *** 1,47 0,07 *** 1,03 0,04 ns. Public sector (ref: Private sector) 0,88 0,04 ** 0,89 0,04 ** 1,06 0,04 ns. Children (ref: Not children) 0,91 0,04 ** 0,94 0,03 ** Immigrant partner (ref. Spanish born) 2,47 0,2 *** 1,48 0,1 *** Educational homogamy (ref: Same educ. level) She more educated 1,24 0,05 *** He more educated 0,79 0,03 *** Ocupational homogamy (ref: Occup. Homogamy) Hipogamy 1,32 0,06 *** Hipergamy 1,13 0,05 *** Partner's occupation (ref: Managerial and professional) Manual skilled 1,38 0,07 *** Administrative and secretarial 1,2 0,09 *** Sales and costumers 0,91 0,09 ns. Hotels and restaurants 1,82 0,17 *** Care, cleanning and domestic 1,1 0,11 ns. Other unskilled 2,23 0,16 *** Partner's temporary job (ref: permanent job) 13,1 0,47 *** Partner's public sector (ref: Private sector) 0,41 0,02 *** Constant 0,01 0,00 *** 0,34 0,1 *** 0,24 0,1 *** 0,03 0,01 *** Number of observations 259.842 259.842 259.826 Number of individuals 79.892 79.892 79.890 Statistic Significance= "ns" non significant; " * "p < 0.10; " ** " p < 0.05; " *** " p < 0.01. Source: Spanish LFS, 2008-2015. 34 Figure 1: Evolution of couples’ economic arrangements. Women in partnership with or without children and with both partners aged 20-60. Latin American Women 70% 60% 50% 40% 30% 20% 10% 0% Only woman works Both work Only man works Nobody works Spanish Women 70% 60% 50% 40% 30% 20% 10% 0% Only woman works Both work Only man works Nobody works Source: Spanish LFS, 2005-2015. 35 Figure 2: Interaction between women’s occupation and women’s birthplace. Predictive margins of transitioning from a dual-earner household to a female- earner household obtained from pooled models Source: Spanish LFS, 2005-2015. 36 Figure 3: (A) Interaction between children and women’s birthplace. (B) Interaction between immigrant partner and women’s birthplace. Predictive margins of transitioning from a dual-earner household to a female-earner household obtained from pooled models A B Source: Spanish LFS, 2005-2015. Figure 4: (A) Interaction between partners’ educational assortative mating and women’s birthplace. (B) Interaction between partners’ occupational assortative mating and women’s birthplace. Predictive margins of transitioning from a dual- earner household to a female-earner household obtained from pooled models A B Source: Spanish LFS, 2005-2015. 37