TOWARDS EQUAL? WOMEN IN CENTRAL AMERICA TOWARDS EQUAL? WOMEN IN CENTRAL AMERICA 2018 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The find- ings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dis- semination of its knowledge, this work may be reproduced, in whole or in part, for noncom- mercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Photo cover: (up) Three women plant seeds in a farm in Chimaltenango, Guatemala. Photo: Maria Fleischmann / World Bank (dow) Haiti - Nonouche Rochambert works at the Truitier debris management site. She is one of 176 employees. They work in two shifts sorting the rubble - all that is left of hundreds of thousands of homes and buildings destroyed in the January 12, 2010 earthquake. Photo: Romel Simon / World Bank Contents Acknowledgements 9 Executive summary 11 Introduction 16 Chapter 1 Agency 21 1.1 Legal framework and institutional arrangements 22 1.2 Manifestations of agency 27 Political participation 27 Violence against women 31 Annex 1.1 38 Chapter 2 Endowments 40 2.1 Health 41 Maternal mortality and access to health services 42 Teenage pregnancy 48 2.2 Education 49 Chapter 3 Economic opportunities 57 3.1 Economic participation 57 Obstacles to economic participation 60 Unemployment 62 3.2 Employment outcomes 65 Job quality 67 Labor earnings 72 Entrepreneurship and access to finance 73 Annex 3.1 79 Annex 3.2 81 Chapter 4 Underlying Drivers: Gender Roles and Social Norms 84 4.1 Gender norms, aspirations, and outcomes 85 4.2 The role of households in gender outcomes 93 4.3 Differences between groups 101 Annex 4.1 105 Chapter 5 Taking Stock 107 5.1 Overview of outcomes 107 Agency 107 Endowments 108 Economic opportunities 109 Aspirations and norms 110 Differences in outcomes 111 5.2 Policy discussion 112 References 116 List of Figures Figure 1: Female labor force participation rates (15+) and per capita GDP, LAC region, 2014 17 Figure 2: Poverty reduction in LAC and Central America since 2000 18 Figure 1.1: Levels of gender inequality in social institutions in Latin America, 2014 25 Figure 1.2: Percentage of mayors who are women in Nicaragua, Central America, and Latin America and the Caribbean, 1998-2013 29 Figure 1.3: Global rankings of Central American female political representation 30 Figure 1.4: Percentage of mayors who are women, 2000 and 2014 31 Figure 1.5: Percentage of women ever married or in a union aged 15-49 who reported ever experiencing intimate partner violence by type of violence 33 Figure 1.6: Femicide rate in Central America, per 100,000 women, 2016 34 Figure 2.1: Life expectancy men and women, 1965 and 2015 41 Figure 2.2: Fertility rates 1970-2015 43 Figure 2.3: Maternal Mortality Rates, 1990 and 2015 44 Figure 2.4: Share of births attended by skilled personnel (% of total) 44 Figure 2.5: Pregnant women receiving prenatal care (%) 44 Figure 2.6: Adolescent fertility rates per 1,000 women ages 15-19, 1995, 2005, and 2015 49 Figure 2.7: Adult literacy rate (%), 2000 and 2010 50 Figure 2.8: Gender parity in primary education, 2015 51 Figure 2.9: Secondary education completion rates, 2014 52 Figure 2.10: Gender parity in secondary and tertiary school enrollment 53 Figure 3.1: Female labor force participation in Central America 58 Figure 3.2: Rural female labor force participation in Central America 59 Figure 3.3: Female labor force participation and rural population, 2014 60 Figure 3.4: Probability of women being in the labor force by highest level of schooling completed compared to women with less than a primary education, 2004 and 2014 63 Figure 3.5: Urban unemployment rates, by gender and age, 2014 64 Figure 3.6: Employment rates (18-65), 2014 65 Figure 3.7: Part-time employment rates among women (18-65), 2014 65 Figure 3.8: Unpaid work among rural workers, by gender, 2014 68 Figure 3.9: Unpaid work among young people (18-25), 2014 68 Figure 3.10: Share of wage workers in microenterprises, by gender 2014 69 Figure 3.11: Share of wage workers in public sector, by gender 2014 71 Figure 3.12: Ratio of women’s earnings to men’s earnings, monthly and hourly in 2014 73 Figure 3.13: Change in gender earnings gap, 2004 to 2014 73 Figure 3.14: Gender hourly wage gaps, 2014 74 Figure 3.15: Share of adults with a bank account 77 Figure 4.1: Agreement that wife-beating is justified in certain situations, among women 87 Figure 4.2: Women should only work if the partner does not earn enough 90 Figure 4.3: Correlation between female labor force participation rates and views of women’s work 91 Figure 4.4: Age at first marriage, female, 1980-2010 94 Figure 4.5: Rates of female-headed households 96 Figure 4.6: Rates of child marriage, 2008-2014 97 Figure 4.7: Predicted change in female labor force participation based on number of young children in the household and marital status, 2014 101 Figure 4.8: Probability of being in the labor force by highest level of schooling completed relative to adults with no schooling, 2014 103 Figure 4.9: Gender parity in migration to the US by year of arrival cohort 104 List of Tables Table 1.1: Length of paid maternity and paternity leave 23 Table 2.1: Growth in life expectancy, 1965 - 2015 42 Table 2.2: Share of women ages 15-49 who use modern contraceptives (%) 47 Table 2.3: Area of study in tertiary education, by gender (2015) 53 Table 3.1: Type of employment by gender, 2004 and 2014 67 Table 3.2. Gender segregation by sector of employment, 2014 72 Table 4.1: Average time spent by women in unpaid and paid work 99 List of Boxes Box 1: In focus: Lessons learned from interventions in the region 18 Box 2: Rural barriers or indigenous exclusion? 19 Box 1.1: In focus: Access to land for women in Honduras 26 Box 1.2: Different types of political quota 28 Box 1.3: Data availability and measurement challenges of gender-based violence 32 Box 1.4: Violence in the Northern Triangle 35 Box 1.5: Gender relations within gangs 36 Box 1.6: In focus: Addressing violence against women through community-based interventions: Sasa! in Honduras 37 Box 2.1: Barriers to health services for indigenous women in Honduras and Guatemala 46 Box 2.2: In focus: Lack of sanitation and girls’ attendance in Panama 53 Box 3.1: In focus: Measuring female labor force participation 61 Box 3.2: In focus: Female Ninis in Central America 64 Box 3.3: In focus: Increasing economic opportunities for women through public works and training: The Temporary Income Support Program (PATI) in El Salvador 66 Box 3.4: In focus: Getting women into non-traditional jobs in the roads sector: The Nicaragua rural roads project 70 Box 3.5: In focus: Economic empowerment of indigenous women in Panama: A snapshot of challenges & opportunities 75 Box 3.6: In focus: Guatemala: Women in agriculture: The impact of male out-migration on women’s agency, household welfare, and agricultural productivity 76 Box 4.1: In focus: Roads to agency – barriers and effects of integrating women into non-traditional sectors 88 Box 4.2: Social interactions with leaders matter: Changing households’ investment and aspirations through social interactions: Evidence from a randomized transfer program 92 Box 4.3: In focus: Alliance for education quality in Nicaragua 93 Box 4.4: In focus: The positive effect of targeted productive transfers on intra-household bargaining 100 8 Abstract C entral America lags the rest of Latin America in outcomes related to women. The countries of Central America (Costa Rica, El Salvador, Guate- mala, Honduras, Nicaragua, and Panama) have taken important steps toward increasing the inclusion of women and improving their social and economic out- comes. This report takes stock of this progress as well as continuing challenges faced by women in the re- gion. It assesses women’s legal rights and protections, access to endowments and economic opportunities, and manifestations of agency. A crucial message of this report is that the persistence of gender norms that limit women’s roles and voices remains a key obstacle to increasing economic and social inclusion. World Bank Towards Equal? Women in Central America 9 Acknowledgements T his work was authored by Miriam Muller (So- cial Scientist) and Liliana D. Sousa (Economist) with important inputs from Giselle Del Carmen traczenko (Program Leader Central America) as well as by the peer reviewers, Maria Eugenia Genoni (Se- nior Economist) and Benedicte Leroy De La Brière and Carmen de Paz (Consultants) in the Poverty (Lead Economist). and Eq-uity Global Practice. Andres Cesar The team is particularly grateful to the following (Consultant) pro-vided additional empirical analysis colleagues who kindly contributed with inputs or re- and Fiona Mack-intosh (Consultant) edited the visions on gender interventions led by World Bank study. We would like to thank Oscar Calvo- Group teams: Paula Tavares (Private Sector Devel- Gonzalez (Practice Manager) for his supervision opment Specialist, Pablo Acosta (Senior Economist), and guidance and Seynabou Sakho (Country Director for Central America) for her sup-port to Sevara Melibaeva (Senior Transport Economist), Di- this work. This work builds on gender notes for anna Pizarro (Senior Social Development Specialist), each of the Central America countries, with con- Simone Buechler (Consultant), Chloe Fevre (Senior tributions from Elizaveta Perova (Senior Social Development Specialist), Mary Lisbeth Gon- Economist), Anna Bonfert (Consultant), Sara van zalez (Senior Social Development Specialist), Renos Wie (Consul-tant), and Denisse Wolfenzon Vakis (Lead Economist), and Rita Almeida (Senior (Consultant). Economist). The team would like to gratefully Excellent administrative and logistic support was acknowledge the comments and inputs provided by provided by Ruth Eunice Flores (Senior Program As- Kinnon Scott (Senior Economist), Friederike sistant) and Pamela Gaye Gunio (Program Assistant). Koehler-Geib (Pro-gram Leader Central America), Andrea Guedes (Op-erations Manager Central America) and Tania Dmy- 10 Towards Equal? Women in Central America 11 Executive summary C entral America lags the rest of Latin America in outcomes related to women. The countries of Central America (Costa Rica, El Salvador, Guatemala, sures. In addition, none of the six countries has laws protecting women against being paid less than men for work of equal value, discriminatory family law, Honduras, Nicaragua, and Panama) have taken import- especially regarding early marriage, and institutional ant steps toward increasing the inclusion of women and weaknesses in combating human trafficking are areas improving their social and economic outcomes. They that could and should be improved in the countries of have passed critical legal protections and made import- Central America. ant investments in healthcare and education. Even so, Two manifestations of agency (or the lack thereof ) challenges persist – these are particularly notable in explored in this report illustrate the extent to which women’s weak economic outcomes. A crucial message institutional and legal steps forward have led to im- of this report is that the persistence of social norms that portant advancements for women in some areas but limit women’s roles and their influence on household not in others. These are political representation on the decisions remain a key obstacle to increasing women’s one hand and violence against women on the other. economic inclusion in Central America. In the area of political representation, some Central Overall, the countries of Central America have American countries are high achievers, with Nicara- made significant strides in setting in place the legal gua, in particular, standing out among the best per- frameworks and formal institutions necessary to sup- formers globally in terms of female political partic- port the agency of women and girls. They are signato- ipation. On the other hand, some countries in the ries to various landmark global agreements that pro- region stand out for their high rates of violence against tect women and girls and have, in many cases, passed women. For those countries with available informa- legislation against some forms of gender discrimina- tion, about one in three women reports having been tion. Although these efforts are important, additional exposed to violence by an intimate partner, a minority steps are needed to successfully implement these mea- of whom seek institutional support or protection. 12 Substantial progress has been made in women’s of 2014. In most Central American countries, women health and education outcomes, important endow- are also disproportionately employed in microenter- ments that promote the wellbeing and human capital prises and self-employment, which are overwhelming- of individuals. Most notably, Guatemala, Honduras, ly likely to be informal and low-paying jobs. El Salvador, and Nicaragua have largely closed the gap Differences between men and women in labor between themselves and the rest of the Latin America supply and job outcomes translate into a large disad- and the Caribbean (LAC) region in terms of women’s vantage for women in labor income. This is partially life expectancy, while Panama and Costa Rica contin- explained by the fact that women in Central America ue to outperform the regional average. While Central devote less time to employment activities than men. America trails LAC in terms of general access to ed- Taking into account other factors, including how ucation, girls and women are more likely than boys women self-select into employment and key human and men to be enrolled in secondary and tertiary ed- capital and job characteristics, there is no statistical ucation, while primary education is all but universal evidence of a wage gap between men and women in for both genders in all countries. However, maternal urban areas in Honduras and El Salvador. However, mortality rates remain high, even in richer countries there are sizeable wage gaps in the other four coun- such as Panama, and especially so among the most tries. In those countries, men’s hourly wages are, on vulnerable groups. Rural and indigenous women and average, 11 percent to 14 percent higher than those of those from a poorer socioeconomic background have women with similar characteristics. lower access to basic services. Finally, teenage preg- nancy rates are above the LAC average in all Central There are, of course, significant differences be- American countries except Costa Rica. tween and within countries. Guatemala seems to struggle in essentially all of the dimensions analyzed While Central America has made progress in terms in this report. It is the country with the fewest women of the legal frameworks and access to health and ed- in political positions in Central America, lower than ucation, these improvements have not translated into the LAC average; the only country in the sub-region significant economic opportunities for women. In where girls are disadvantaged in all levels of educa- particular, Central American women have low rates tion; and the country with the lowest female labor of labor force participation, higher rates of unemploy- force participation. Across the region, however, even ment than men, and worse employment outcomes. where progress has been achieved, not all women have Women living in rural areas in particular have low lev- benefitted equally from it. Gender outcomes vary sig- els of participation, ranging from just one in three in nificantly between urban and rural women, between Nicaragua and Guatemala to only half in Panama, the indigenous and non-indigenous women, and among region’s top performer. Fewer than half of all women women from different socioeconomic groups. aged 18 to 65 in Guatemala, Honduras, and Costa Rica were employed in 2014. Among those who were Perceptions and aspirations as well as social norms employed, most worked part time, with fewer than are fundamental drivers of gender outcomes and are one in three women aged 18 to 65 in Guatemala and likely behind the trends observed, although data on Honduras working full time. The best performer by them are scarce. The few existing perception surveys this measure was Panama, where 40 percent of all in the region show that a considerable share of the women in this age group were employed full time as population expects women to have less power with- Towards Equal? Women in Central America 13 in relationships than men. Attitudes and perceptions and taking into account the allocation of labor within about certain aspects of gender equality are strongly households, it is important to consider interventions correlated with specific outcomes, for instance, those aimed at reducing the labor-intensity of household related to women’s role in the labor force or inti- and caregiving tasks, such as access to improved stoves mate-partner violence. Central American countries and to affordable and reliable childcare, which would have among the highest rates of child marriage in the allow more women to work outside the home. Fur- LAC region, which can have significant negative im- thermore, facilitating the process by which property pacts on women’s accumulation of endowments and is registered under the names of both husband and their ability to take advantage of economic opportu- wife would give women more control over the family’s nities. Furthermore, the timing of first marriage and assets, as would extending women’s access to banking first birth has remained almost unchanged over the and credit. In the area of voice and agency, focusing years. Finally, women bear the brunt of household on the support and protection of victims of domestic and caregiving duties, limiting the time which can violence, including consistent legal enforcement are be devoted to economic engagement even while their important and urgent measures to be taken. With overall workload tends to be higher than that of men. regards to political participation, Guatemala could increase rates of women in politics by following the Given these findings, there is a need for two ap- example of other countries in the region and intro- proaches to policy intervention to increase the social ducing gender quotas. and economic inclusion of women in Central Amer- ica. First, there are concrete, short-term entry points In the longer term, increasing the social and eco- that could jumpstart progress in narrowing specif- nomic inclusion of women will require fundamental changes in the prevailing norms about the appropri- ic gender gaps and ensure women and men benefit ate roles women play in society. Policies that promote more equally from services, assets, and opportunities. economic development in general can lead to funda- Second, given the entrenched nature of the inequities mental changes as can increased access to commu- and barriers, there is a need for policies focused on the nications technology. However, norms are persistent long-term horizon. and many structural factors conspire towards inertia. Concrete entry points exist to further progress on Thus, dedicated policies and programs may be need- the three areas necessary for fostering the inclusion of ed to effectively stimulate a change in the underlying women: endowments, economic opportunities, and root causes of the observed gender inequalities. Con- voice and agency. With regards to endowments, en- cretely, as norms favoring inequality may persist due suring that all women have access to effective health- to misinformation, exposure to new role models may care such as skilled birth attendants, especially for help girls and women to update their aspirations and underserved rural and indigenous areas, would help life plans and provide them with the necessary tools to to reduce high maternal mortality rates. Reworking implement those plans. Media can also play a key role curriculums and raising teachers’ awareness on gen- in driving norms change. Exposure to new ideas and der bias could help to promote girls’ interest in math role models through television and soap operas have and science and in occupations beyond those tradi- had quantifiable impact on women’s autonomy, with tionally considered to be appropriate for women. To an increase in their participation in household deci- encourage more economic opportunities for women, sion making and a decrease in the acceptability of do- Executive summary 14 mestic violence. Social norms marketing (an effort in Finally, gender equality is about men and women which messages are disseminated that aim to change and it benefits both: As such, engaging men as partners attitudes and behaviors) can be effective in activating in changing gender norms will be crucial for success. positive social norms and in discouraging negative Experience in the region, such as the Promundo-led attitudes and behaviors, particularly when combined `Program H’ in Brazil, can provide useful insight in with local interventions and engagements. how to engage men in addressing these complex topics. Towards Equal? Women in Central America 15 16 Introduction C entral America lags the rest of Latin America when it comes to outcomes related to women. The countries of Central America (Costa Rica, El Sal- Increasing women’s economic activity is import- ant for empowering women themselves, but it also has a strategic importance for Central America. In- vador, Guatemala, Honduras, Nicaragua, and Pana- creasing women’s empowerment fosters economic de- ma) have taken important steps toward increasing velopment (Agénor and Canuto, 2013), while higher the inclusion of women and improving their social female labor force participation can lead to higher and economic outcomes. They have passed critical le- rates of economic growth and greater macroeconom- gal protections and made important investments in ic stability (Loko and Diouf, 2009; Dollar and Gatti, healthcare and education. Even so, these countries 1999; and Stosky, 2006). Estimates have shown that, if women’s labor income had not grown during the continue trailing most of the other countries in Latin decade of the 2000s, extreme poverty in LAC would American and the Caribbean (LAC) region in crucial have been 30 percent higher in 2010 (World Bank, indicators related to women’s outcomes, including 2012b). This is a particularly important consideration economic participation (see Figure 1). Women’s labor for Central America where poverty reduction, like fe- force participation averages 53 percent in the LAC re- male economic inclusion, have lagged the rest of the gion, but it is only 47 percent in Central America.1 region (Figure 2). Mateo-Diaz and Rodriguez-Cha- While there are significant differences across the per mussy (2016) estimate that increasing female labor capita income levels of the countries of Central Amer- force participation to the level of men would increase ica, they all fall below the regional average when it gross domestic product (GDP) in Honduras by 16.8 comes to women’s economic inclusion as measured by percent and in Costa Rica by 10.4 percent. participation rates. Women’s inclusion also makes long-term contri- 1 Based on unweighted averages of national female labor force butions to development. Evidence shows that, when participation rates in 2014 as reported in the Word Development Indicators (The World Bank). women have a say over household resources and Towards Equal? Women in Central America 17 Figure 1: Female labor force participation rates Figure 2: Poverty reduction in LAC and Central (15+) and per capita GDP, LAC region, 2014 America since 2000 70 50 Female Labor Force Participation 65 PER BOL HTI 40 60 PRY COL Poverty rate (%) BRA URY 55 DOM LAC 30 (15+) NIC SLV CHL 50 ECU PAN HND ARG 45 MEX CRI 20 40 GTM 10 35 LAC Central America 30 0 0 5000 10000 15000 20000 2000 2002 2004 2006 2008 2010 2012 2014 GDP per capita (USD) Source: Authors’ tabulations using World Development Source: LAC Equity Lab tabulations using SEDLAC (CEDLAS and Indicators (The World Bank). the World Bank). Note: Female labor force participation rate is measured as the Note: Poverty is measured at the internationally comparable percentage of all women aged 15 and older who report being income line of $5.50 (2011 PPP) per day per person. employed or actively searching for employment. spending patterns, this can lead to important im- The importance of women’s equality for development provements in the development outcomes of the next is reflected in its inclusion in the UN’s Sustainable De- generation (World Bank 2012b). Maternal education velopment Goals (the achievement of gender equality is associated with reductions in risk factors such as in- and the empowerment of all women and girls). fant or child malnutrition, low birth weight, and fail- The objective of this study is to provide an over- ure to vaccinate children as well as with an increased view of outcomes for women in Central America likelihood of positive outcomes such as using purified across several dimensions, especially economic op- water and seeking healthcare. Improvements in the portunities. The aim of this report is to provide pol- education of women are estimated to have saved the icymakers and civil society in Central America with lives of approximately 2.1 million children under the the information that they need to shape policies, pro- age of 5 between 1990 and 2009 (UNESCO, 2014). grams, and gender-responsive actions to close those In addition, greater gender equality and participation gaps in opportunity between men and women and, can lead to more representative institutions and poli- thus, to enhance growth and sustainability. We com- cy choices to the benefit of the entire society. For in- bine international knowledge and lessons learned stance, in India, giving power to women at the local from local interventions (Box 1) to identify both ar- level resulted in the provision of more public goods, eas of progress in, as well as bottlenecks to, improving such as water and sanitation (Beaman et al, 2011). women’s lives in the region. Introduction 18 Box 1: In focus: Lessons learned from interventions in the region There are several knowledge gaps related to understanding gender equality and challenges for women’s inclusion in Central America. Experiences from other interventions, including interna- tional and national projects, can also shed light on some of the challenges. Throughout this volume, text boxes labeled “ In Focus ” will highlight projects and evalu- ations that have illuminated the challenges to women’s inclusion in Central America. These projects, supported by or in partnership with the World Bank, range from women’s access to land titles in Honduras to women’s engagement in road construction in Nicaragua. While some boxes will highlight the findings of impact evaluations and qualitative analysis of these target- ed interventions, other boxes will highlight how gender-related challenges were reflected in project outcomes. portunities (outcomes pertaining to jobs, production, Framework and technology, and market access).2 organization According to this framework, households are cen- of this report tral to the connection between gender equality and growth. Formal institutions, informal institutions, Gender reflects socially constructed roles and social- and markets affect households while, conversely, ly learned behaviors and expectations for both men household decisions affect the functioning and struc- and women. At times, gender norms lead to the adop- ture of markets and institutions. For example, fami- tion of roles, expectations, and behaviors that put one lies decide when to have children, how many children gender at a disadvantage vis-à-vis the other. For many to have, and how to assign tasks inside and outside Central American men, for example, current gender the household, but these decisions are influenced by norms are associated with lower completion rates of market and institutional factors. In turn, these deci- secondary education than for women and more risky sions also influence prevailing wages and labor sup- behavior such as involvement in gangs and alcohol ply by affecting population growth and women’s time abuse. Yet, as shown in this study, Central American for market activities. The framework also posits that women face quantifiable disadvantages and exclusions agency, endowments, and economic opportunities are in many more dimensions of social and economic life mutually reinforcing in terms of the progress (or lack than their male counterparts. of progress) made in any one of these dimensions. Fi- The study is organized around a framework that nally, it is important to keep in mind that barriers to posits the necessity of progress in three dimensions to gender equality often differ between groups with vary- ing levels of power. In particular, limits on access and achieve inclusion of women: (i) agency (outcomes re- inclusion are often most severe among women who lated to the ability to make choices to achieve desired have multiple disadvantages, such as being a member outcomes, including having a voice in decision-mak- ing); (ii) endowments (outcomes related to education, 2 This framework was introduced by the World Bank’s World De- velopment Report 2012: Gender and Development (World Bank, health, and physical assets); and (iii) economic op- 2012a). Towards Equal? Women in Central America 19 of an ethnic minority, having a disability, or being nificantly worse outcomes than can gender alone can poor (Tas et al., 2013). The intersection of gender, do as we will see in the course of this report (Box 2). age, ethnicity, and place of residence can result in sig- Box 2: Rural barriers or indigenous exclusion? Indicators throughout this report paint a consistent picture of indigenous women experiencing worse outcomes than other women in Central America. Nearly 20 percent of Central America’s population is indigenous, ranging from an estimated 40 percent of Guatemalans to less than 2 percent of El Salvador’s population. In many cases, because of data limitations, it is impossible to determine the extent to which outcomes for this population reflect the exclusion of indige- nous peoples as a whole or of indigenous women in particular rather than other characteristics that are correlated with being indigenous, such as low income or high rates of rural status. The disproportionate rates of rurality among Central America’s indigenous communities is a particular consideration. With the exception of El Salvador, where indigenous peoples are evenly split between rural and urban areas, a majority of Central America’s indigenous popula- tion is rural (World Bank 2015a). Rural populations, in general, have lower access than urban residents to basic goods and services such as running water and improved sanitations and to have lower educational outcomes. To the extent that indigenous women are more likely to live in rural areas than other women in any given country, their less favorable outcomes may be partially explained by rural factors. However, other research has found that even beyond rurality, outcomes for the IP are often worse, reflecting additional barriers. A study of indigenous people across Latin America (World Bank 2015a) found that “indigenous peoples fare worse on most accounts, independently from other factors such as level of education, age, urban or rural location, type of work, and char- acteristics of the household.” Additionally, as noted in this volume, surveys from across the region have found evidence of discriminatory or culturally insensitive practices affecting, for example, health and education outcomes. Source: (World Bank 2015a) Chapters 1, 2, and 3 of this report measure out- While the first two chapters identify real and positive comes in the three dimensions of gender equality change in Central America, the third highlights per- laid out in the framework – agency, endowments, sistent challenges when it comes to economic inclu- and economic opportunities. Chapter 1 analyzes im- sion. Following from this analysis, Chapter 4 explores provements to the legal and institutional frameworks the fundamental reasons for why improvements in governing gender equality in each country, as well as some areas have not translated into increased eco- outcomes in two particularly salient manifestations of nomic opportunities for women. Finally, Chapter 5 agency: violence against women and political partic- ipation of women. Chapter 2 considers women’s ac- offers a brief conclusion bringing together key results cess to two key endowments: education and health. from the first four chapters and proposing a policy Chapter 3 tackles the issue of economic inclusion, approach to tackle the complex mechanisms that de- particularly as relates to employment opportunities. termine women’s outcomes in Central America. Introduction 20 Towards Equal? Women in Central America 21 chapter 1 Agency A gency is the ability of an individual to make choices to achieve desired outcomes. According to Sen (1999), a person with agency is “someone who underlie discriminatory practices that prevent both men and women from having equal access to endow- ments and economic opportunities. Agency plays a acts and brings about change.” (Sen, 1999, p. 19). significant role in individual and social decisions about The idea of agency emphasizes that individuals are not human capital investments and, ultimately, women´s passive recipients but active drivers in their own lives. chances to become active social and economic agents. Agency is also of instrumental value because it serves It is, thus, a key factor in understanding the social and as a catalyst for other development outcomes: economic outcomes of women in Central America. “Instrumentally, agency matters because it has been Equality of opportunity between sexes in all hypothesized, and many times confirmed, that it can spheres of life can only exist on the grounds of equal- serve as a means to other development outcomes. The ity before the law. When legal differences based on agency of women, for instance, has been shown to affect gender are prevalent, women´s capacity to exert deci- positively the wellbeing of all those around them” (Sen sions in all areas of life are constrained, with far-reach- 1999). ing implications throughout the lifecycle. As an ex- Agency can be expressed in many ways—in per- ample, if women´s opportunities in the labor market sonal relationships, in communities (autonomy in de- are not the same as men´s, families may decide not to cision-making, participation in politics, and freedom invest as much in educating girls as in educating boys. of movement), and in an individual’s ability to accu- However, it is not enough to have laws that recognize mulate endowments such as land or property, educa- equality between men and women; these regulations tion, or health. Constraints to agency, for instance, in need to be properly enforced. Often, legal equality co- the form of social norms or institutional biases, often exists with high levels of gender inequality because of 22 poor enforcement or design or because of a lack of on the Prevention, Punishment, and Eradication of capacity.3 Violence against Women (Belém do Pará Conven- tion).5 Panama, Honduras, El Salvador, and Costa This chapter reviews the legal frameworks and in- Rica have also adopted the Beijing Platform for Ac- stitutional arrangements in each country related to tion,6 while Panama and El Salvador participated in gender equality and women’s inclusion, including na- the Quito Consensus7 in 2007 and the Brasilia Con- tional laws and international agreements. The section sensus8 in 2010. also presents a global benchmarking of the countries in Central America, showing how they compare inter- Each of the six countries has passed laws to pro- nationally. The second part of the chapter takes a clos- mote gender equity in employment by providing paid er look at two key dimensions of the manifestation of maternity leave and childcare. In all of the countries, agency—political representation and rates of violence childcare is subsidized or provided free of charge, and against women. the dismissal of pregnant workers is prohibited. Ma- ternal leave, at 100 percent of wages, is offered to for- mal sector workers, with benefits ranging from four 1.1. Legal framework months (120 days) in Costa Rica and 14 weeks (98 and institutional days) in Panama – both above the LAC average of arrangements 88 days – to 12 weeks in Guatemala, Honduras, El Salvador, and Nicaragua (Table 1.1). At the end of The countries of Central America have legal frame- maternity leave, in all countries but Costa Rica, moth- works and national plans that prohibit gender dis- ers are guaranteed an equivalent position as they had crimination and that are designed to reduce gender held before. Although paternity leave is increasingly inequality. Provisions preventing discrimination on the basis of gender have been included in the consti- 5 The Belém do Pará Convention entered into force on March 5, tutions of Honduras, Nicaragua, and Panama over the 1995 and was the world’s first binding international treaty to recognize that violence against women constitutes a violation past two decades. In addition, Costa Rica, El Salva- of human rights. As a legally binding treaty, the Belém do Pará Convention not only condemns violence against women as an dor, Honduras, Nicaragua, and Panama have passed assault on human dignity but also outlines states’ obligations to specific legislation promoting gender equality. Men eliminate it. 6 The Beijing Platform for Action, adopted at the UN’s Fourth World and women are granted equal rights and obligations Conference on Women (Beijing, China, 1995), flagged 12 key ar- within the family, including with regard to parental eas where urgent action was needed to ensure greater equality and opportunities for women and men and girls and boys. It also authority, in all six countries. laid out concrete ways for countries to bring about change. 7 The Consensus followed the Tenth Regional Conference on Wom- Each country is a signatory to the main interna- en in Latin America and the Caribbean held in Quito, Ecuador, in tional laws on gender equality, including the Conven- August 2007. The Consensus focused on two strategic issues: (i) political participation and gender parity in decision-making tion on the Elimination of All Forms of Discrimina- at all levels and (ii) the contribution of women to the economy tion against Women (CEDAW)4 and the Convention and social protection, especially in relation to unpaid work. Representative governments established specific measures to overcome gender discrimination in political participation, employment, education, health, and the economy. 3 World Bank (2015a). 8 Government leaders agreed to facilitate women’s access to new 4 Adopted in 1979 by the UN General Assembly, is often described technologies, promote a democratic and non-discriminatory as an international bill of rights for women. By accepting the media, improve the health and sexual and reproductive rights of Convention, states commit themselves to undertaking a series women, and promote international and regional cooperation for of measures to end discrimination against women in all forms. gender equity. Towards Equal? Women in Central America 23 seen as an important tool to make the distribution of relatively high paying fields such as mining and man- tasks and roles between men and women within the ufacturing (World Bank, 2015b), thus contributing household more equitable,9 only half the countries in to high wage gaps between men and women. In Costa Central America make specific provisions for paterni- Rica, Guatemala, Honduras, and Panama, these re- ty leave and then for only three days in El Salvador, strictions take the form of preventing women from two days in Guatemala, and five days in Nicaragua. working in jobs that are deemed to be either morally or socially inappropriate, hazardous, or arduous. In addition, in Costa Rica, women cannot work after Table 1.1: Length of paid maternity and midnight, while in Nicaragua women cannot engage paternity leave in jobs that require them to lift weights above a cer- tain threshold. Maternity Paternity Costa Rica 120 days 0 Each country has also passed legislation that crimi- Panama 98 days 0 nalizes acts of violence against women, including mar- ital rape and abuse. Femicide, defined as the killing of Nicaragua 84 days 5 days women based on their sex, is a criminal offence in all Guatemala 84 days 2 days countries. Guatemala introduced a law on femicide El Salvador 84 days 3 days in 2008 and was also the first country in the world to Honduras 84 days 0 create specialized courts as a direct judicial response LAC 88 days to femicides.10 Additionally, all countries have special- Source: Women, Business, and the Law (2016) database, ized courts or procedures dealing with cases of inti- World Bank Group. mate partner violence (IPV). While each country in Central America has taken steps to address human trafficking and forced labor, none fully meet the Traf- Even so, legislative barriers to employment for ficking Victim’s Protection Act’s minimum standards women can be found throughout Central Ameri- (USDS, 2017). Among Central American countries, ca. For example, none of the six countries mandates the United States Department of State found that equal remuneration for work of equal value (the ILO Nicaragua was doing the least to meet the minimum standard), and only Honduras mandates non-dis- standards, including by not implementing a dedicated crimination based on gender in hiring. Additionally, anti-trafficking fund. in all countries except El Salvador, women’s work is subject to different legal restrictions that do not apply While laws bar sexual harassment in some spheres to men. This is typical in many countries through- of life, no country in Central America has legislation out the world: of the 173 economies included in the prohibiting sexual harassment on the streets and in World Bank’s Women, Business and the Law database, 100 restrict women from pursuing the same economic 10 Specialized police (or judicial) institutions whose purpose is to increase women’s access to justice have existed for several activities as men. Many of the restricted jobs are in decades in LAC. The first Women’s Police Station (WPS) in the region was inaugurated in São Paulo, Brazil in 1985. More than 400 WPS exist across Brazil, while in Latin America in general, 9 For example, Amin et al. (2016) found a significant positive more than 13 countries have some sort of specialized police relationship between the presence of mandated paternity leave and/or judicial service. The specific role and functions of the in an economy and the share of women workers in a firm. WPS differ by country (Jubb et al., 2008) Chapter 1 Agency 24 transportation. All countries in Central America ex- tional development in the area of gender equality. The cept Guatemala have laws covering sexual harassment OECD’s Social Institutions and Gender Index (SIGI) in employment. In addition, with the exception of categorizes most countries in the world into five cat- Guatemala and Panama, laws protect women and girls egories based on the level of gender discrimination from sexual harassment in schools. However, there are in social institutions: very low, low, medium, high, no laws against sexual harassment in the streets or on and very high (OECD, 2014).11 Panama is the only public transportation. country in Central America, and one of only six LAC countries, included in the very low gender discrimi- Concrete steps have been taken in each country nation category (Figure 1.1). These are the global best to operationalize the promotion of gender equality performers. At the other end of the spectrum, Nica- through either overarching national policies or nation- ragua is the country in LAC with the lowest level of al plans (Annex 1.1). In most cases, these policies or gender equality in social institutions and, along with plans are structured around thematic areas that include Guatemala, is included in the medium category. No broad objectives and/or specific actions to promote country in LAC is in the high or very high category. gender equality in all areas of economic and social life. Costa Rica, El Salvador, and Honduras are in the low Most include employment, health, education, social category. protection, political representation, and gender-based violence. The synthesis report for the 2014 SIGI found that discriminatory family codes –especially topics related In three out of the six countries, the public agen- cy in charge of the promotion of gender equality has to early marriage – were areas in which four coun- been elevated to ministerial rank. This is the case for tries in Central America were lagging behind (OECD, the Institute for Women in both Costa Rica and Nic- 2014).12 Out of 159 countries, Nicaragua ranked 137 aragua and for the Presidential Secretariat for Wom- in discriminatory family codes, placing it in the very en’s Affairs in Guatemala. The National Institute for high category for this dimension. Guatemala, Hon- Women (INAM) in Honduras was at the ministerial duras, and Costa Rica were in the medium category, level until 2014 when it was combined with the Min- ranked 95, 93, and 71 respectively. Among the key istry of Development and Social Inclusion. Panama´s issues identified were high rates of marriage among National Institute for Women (INAMU) is a division women between the ages of 15 and 19, including 30 of the Ministry of Social Development, while the In- percent in Nicaragua and 25 percent in Honduras. An stitute for the Advancement of Women (ISDEMU) important deficit identified in this dimension were in El Salvador reports to a Board of Directors chaired laws that allow for girls and boys to be married under by the Secretary of Social Inclusion and the first lady. the age of 18 with parental consent or judicial autho- These central institutions are responsible for the de- 11 The SIGI classification clusters 108 countries into five levels of discrimination in social institutions: very low, low, medium, high velopment and oversight of the national policies and and very high. These clusters are constructed using a methodol- action plans on gender equality, in some cases through ogy that reduces the variance within classes and maximizes the variance between classes. The ranking is based on indicators in local or sectoral implementation offices or units. the following areas: discriminatory family code, restricted phys- ical integrity, favoring sons over daughters, restricted resources Given the legal framework and advances made on and assets, and restricted civil liberties. the institutional front, Central America fares better 12 In addition to early marriage, this indicator includes information on parental authority during marriage and after divorce and than the global average in terms of legal and institu- inheritance rights. Towards Equal? Women in Central America 25 Figure 1.1: Levels of gender inequality in social institutions in Latin America, 2014 Nicaragua Medium Haiti Jamaica Guatemala Honduras Colombia Peru Paraguay Low Bolivia Costa Rica El Salvador Brazil Ecuador Venezuela Panama Very Low Dominican Republic Trinidad And Tobago Cuba Argentina 0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16 0,18 SIGI value Source: Social Institutions and Gender Index 2014, OECD. http://www.genderindex.org/ranking rization. While Nicaragua recently raised the min- six countries, the SIGI 2014 found that customary imum age of marriage with parental consent to 16, or traditional practices discriminate against women Costa Rica, Guatemala, and Panama recently raised in terms of parental rights in Costa Rica, Guatema- the age requirement for both women and men to 18. la, Nicaragua and inheritance rights in Nicaragua and A key consideration is that making progress in the Honduras. Similarly, although the Law on Equal Op- legal and institutional framework on its own is not portunities for Women in Honduras stipulates that enough as these legal protections must also be success- land can be registered under the name of both spous- fully implemented and enforced. For example, while laws related to parental authority and inheritance es, between 1983 and 2004 only a quarter of all plots grant men and women the same rights in each of the were registered to women (CEDAW, 2006) (Box 1.1). Chapter 1 Agency 26 Box 1.1: In focus: Access to land for women in Honduras Despite laws allowing for land ownership by women, an audit of the Honduras Land Administra- tion Program’s (PATH) first phase revealed that low institutional capacity and lack of knowledge about women’s legal rights in land registries prevented women from claiming their property rights (World Bank 2012a). Initiated in 2003, PATH aimed to formalize property rights, includ- ing those of women. However, the audit showed that the registries lacked adequate legal pro- cedural mechanisms to identify cases where joint titling should be encouraged and/or granted. Focus groups revealed that PATH officials and beneficiaries alike lacked awareness of women’s legal rights. These mechanisms failed to guide women effectively through the joint titling pro- cess. As a result, many of the women who initiated the process failed to complete it and so never obtained the actual physical title. Building on the lessons learned from the first PATH and its audit, the second phase of the PATH project was able to improve outcomes for women in part by increasing the inclusion of women in the project and promoting public awareness of women’s legal rights (World Bank 2017a). The program’s well designed and implemented gender strategy led to the inclusion of women in all phases of the regularization process. Among the concrete actions taken to increase the rates of women filing for joint ownership were: (i) addressing weaknesses in the legal framework; (ii) promoting public awareness of women’s legal rights, both throughout the land regularization process as well as through large information campaigns targeting women; and (iii) improving the indicators used to monitor gender-differentiated targets throughout the life of the project. Initially a woman had to specifically request joint titling, but PATH is current- ly implementing a 2004 property law ( Ley de Propiedad, Decreto Legislativo 82-2004) that facilitates joint titling. Education and training are critical to the proper implementation of property rights. The train- ing and education of officials in regional property registries and municipalities was critical to this process. The project hosted 13 workshops entitled “Ensuring Gender Equity and Equality in Land Access” for 11 target municipalities and La Moskitia, a largely indigenous region. The workshops increased the participants’ understanding of legal rights and raised awareness of PATH’s gender-responsive approach. The project developed a social communication campaign with materials appropriate for men and women of various ages. Female beneficiaries received clear messaging on how a land title could increase their economic opportunities. The campaign targeted radio outlets and developed a script for a radio program on “Gender and the Situation of Women in Indigenous Communities.” The project promoted gender-specific strategies such as outreach activities to promote women’s registration and increasing the participation of in- digenous women in all aspects of intercommunity titling. The results for PATH II achieved every gender-specific indicator, including high satisfaction ratings from female participants and a large increase in land titling for women. Fifty-eight per- cent of the more than 50,000 titles issued by PATH II have a woman as beneficiary (48 percent as individuals and 10 percent via joint tenure in the case of a couple). Further analysis indi- cates that 96 percent of the households that received a land title through PATH II perceived that Towards Equal? Women in Central America 27 they had more secure tenure and a reduced risk of being evicted or suffering a tenure conflict. Compared to the control group, owners with newly titled land felt that the value of their parcel had increased by 30 percent. With this added confidence, a fifth of these owners invested in improving their asset. Interestingly, women who received a title tended to build a new house more often than men. Sources: World Bank (2012a and 2017a) willing to vote for female representatives. In contrast, 1.2. Manifestations violence against women is a negation of basic human of agency rights. It prevents a woman from fully participating in A legal framework promoting gender equality sets the society, and it violates her right to physical integrity, stage for equality of opportunity between women and her right to physical and mental health, her right to men.13 It prepares the playing field on which both liberty and security of the person, and several others.14 women and men can exercise their agency. However, formal laws are not enough to address gender inequal- Political participation ities or reduce discriminatory social norms and prac- The presence of more women in political power has tices. And more specifically, laws on paper do not nec- been shown to yield important policy results. Coun- essarily translate into the lived experiences of women tries with quotas for female legislators systematically and men. Discriminatory social norms and practices have higher levels of spending on social services and as well as unbalanced power relations can significantly welfare (World Bank, 2015b). Female legislators in undermine the application of well-intentioned legal Latin America have had an impact on policy, especial- frameworks. ly in areas such as contraceptive access, gender-based Agency is difficult to measure directly. Instead, violence, and trafficking (Piscopo, 2015). With the we relied on proxies that represent manifestations of aim of increasing women’s levels of representation, agency or the lack thereof. This section focuses on two nearly 40 percent of countries around the world have specific areas: political participation and gender-based implemented gender quotas for representative bod- violence. The levels of political participation of wom- ies. The first of these was established in 1991 in Ar- en indicate the extent to which women access posi- gentina, initially requiring that women constitute 30 tions of power as well as the extent to which society is percent of each party’s nominees for the lower house 13 The IFC/World Bank’s Women, Business, and the Law Project ex- 14 From the 1992 General Recommendation 19 of the CEDAW amined how, by differentiating between men and women, laws Committee: “Gender-based violence, which impairs or nullifies can affect women’s opportunities and incentives to work. The the enjoyment by women of human rights and fundamental regulations covered are working hours and industry restrictions, freedoms under general international law or under human rights parental benefits, retirement ages, and legal rights in the work- conventions, is discrimination within the meaning of article 1 of place. Some of these legal differentiations may help women to the Convention.” (see: http://hrlibrary.umn.edu/gencomm/gen- work, while others may prevent it. http://wbl.worldbank.org/ erl19.htm). Chapter 1 Agency 28 Box 1.2: Different types of political quota Essentially three types of gender quotas are used in politics: (i) reserved seats (constitutional and/or legislative); (ii) legal candidate quotas (constitutional and/or legislative); and (iii) po- litical party quotas (voluntary). While the first type (reserved seats) regulates the number of women elected, the other two (legal candidate quotas and political party quotas) set a minimum share of women that must be on the candidate lists, either as a legal requirement or a measure written into the statutes of individual political parties. While these are the main types of quota, there are many more forms. A particular detail, for instance, is whether or not the rank order of the candidates on the lists is regulated so that women candidates are not just placed at the bottom of the lists. Furthermore, it is also important to look at whether or not there are any sanctions for non-com- pliance in place. Quotas can be mandated in the Constitution, electoral law, or in the statutes of political parties. Source: The Quota Project (International IDEA, Inter-Parliamentary Union and Stockholm University). http:// www.quotaproject.org/aboutQuotas.cfm of the national congress. Since then, all countries in have legal quotas for candidates’ lists in local electoral Latin America, with the exception of Guatemala, have processes (World Bank, 2015b). These quotas have established some form of political quotas (see Box resulted in some countries in Central America – par- 1.2 for the different types of quota). Subsequent re- ticularly Nicaragua, and to a lesser degree, Costa Rica forms to these quota laws have increased female in- and Panama – performing relatively well on a global clusion, for example, by mandating gender-balanced level when it comes to women participating in gov- candidate lists. These appear to result in more equita- ernment. ble representation of women in leadership and deci- Following a 2012 reform, Nicaragua has become sion-making (IDEA International and IADB, 2010). among the best performers in terms of female polit- Quotas also seem to help to change stereotypes and ical participation globally. Nicaragua’s political inclu- attitudes regarding women as leaders and increase sion laws aim to establish gender parity by specifying women’s overall engagement in politics and civic life.15 that men and women must be included in candidates Legislation in each country in Central America lists in such a way that they are evenly distributed except Guatemala establishes quotas for a minimum and alternately ranked for positions of mayor, dep- share of women to be included in candidates´ lists for uty mayor, councilors and deputies (United Nations, elected positions in national elections. The legal quota 2014). What followed the passage of these laws was a is 50 percent in Nicaragua, Panama, and Costa Rica, significant increase in the number of women in both 40 percent in Honduras, and 30 percent in El Salva- national and local positions. For example, in 2011, 14 dor. All countries except Panama and Guatemala also percent of mayors in Nicaragua were women, close to 15 For evidence of the impact of female quotas on a number of the average in Central America. In 2012, this number outcomes, including public service delivery, corruption, and on jumped to 40 percent (Figure 1.2). In 2017, 52.9 per- the aspirations of parents of girls, see Chattopadhyay and Duflo (2004), Beaman et al. (2010), and Iyer et al. (2012). cent of ministerial-level positions were held by wom- Towards Equal? Women in Central America 29 Figure 1.2: Percentage of mayors who are women in Nicaragua, Central America, and Latin America and the Caribbean, 1998-2013 40 % mayors of mayors who are women Nicaragua LAC Central America (without NIC) 30 20 10 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: United Nations (2015) en in Nicaragua – the highest rate in the world and percent and 12.7 percent of parliament seats respec- tied with Bulgaria and France (Figure 1.3). Similarly, tively are held by women. Out of 193 countries, Pan- with 45.7 percent, Nicaragua has the highest propor- ama and Guatemala rank 107th and 139th, in contrast tion of seats in national parliaments held by women with how they were ranked on women in ministerial in Central America in 2017 and ranked fifth globally. positions where both countries were in the top half The presence of women among political appoin- of the ranking. The differences in rankings between tees at the top executive level varies across the other these two indicators show that Panama and Guate- Central American countries but is, in general, high mala have women in positions of power as appointees by global standards. About 30 percent of ministeri- but lag behind in terms of the number of their elected al positions in Costa Rica and Panama were held by female officials. On the other hand, about one-third women in 2017, a level followed closely by Honduras of elected representatives in the Costa Rican and Sal- at 27.3 percent. These three countries rank relatively vadorian parliaments and a quarter of representatives well at the global level in this dimension. Out of 174 in Honduras are women, with these countries ranking countries, they ranked in the top 40. El Salvador and 27th, 36th, and 63rd globally. Guatemala, however, lag behind. As of January 2017, With the notable exception of Nicaragua, the women represented 21.4 and 18.8 percent of all min- countries of Central America have lost ground relative isters respectively. While they lag behind the region, to the LAC region in terms of the share of mayors who both countries were in the top half of the global rank- are women. While in 2000 most Central American ing, meaning that more countries had worse outcomes countries exceeded the rest of LAC in this respect, by than had better outcomes. 2013 only Nicaragua exceeded the regional average Panama and Guatemala have worse outcomes for (Figure 1.4). In fact, Nicaragua is the only country in the levels of female representation in parliament than LAC where women represent more than 30 percent of for political appointees. In these two countries, 18.3 mayors. In 2014, 40 percent of mayors were women Chapter 1 Agency 30 Figure 1.3: Global rankings of Central American female political representation a) Proportion of women in ministerial-level positions, January 2017 60 60 50 50 40 40 Percent 30 Percent 30 20 20 1 30 32 38 72 86 1 30 32 38 72 86 10 10 0 0 Nicaragua Costa Rica Panama Honduras El Salvador Guatemala Nicaragua Costa Rica Panama Honduras El Salvador Guatemala b) Proportion of seats held by women in national parliaments, January 2017 60 60 50 50 40 40 Percent 30 Percent 30 20 20 5 27 36 63 107 139 5 27 36 63 107 139 10 10 0 0 Nicaragua Costa Rica El Salvador Honduras Panama Guatemala Nicaragua Costa Rica El Salvador Honduras Panama Guatemala Source: Inter-Parliamentary Union, www.ipu.org. Note: The rankings are out of 174 countries for panel (a) and 193 countries for panel (b). The numbers next to the bar are the global rankings for each country. – an increase of over 30 percentage points between In El Salvador, almost 11 percent of mayors in 2014 1998 and 2013.16 Costa Rica and El Salvador also per- were women, just below the regional average. form better in this indicator than the other countries On the other hand, Honduras, Panama, and Gua- in Central America. At 12 percent in 2014, women’s temala have much lower levels of female representation political participation at the local level in Costa Rica in local government. Progress in Honduras was nega- was on a par with the LAC average (ECLAC, 2014). tive between 2000 and 2010 when the share of women holding mayoral positions fell from 10 percent to only 3 percent, but it had increased to 6.7 percent by 2014. 16 The next two countries in the region in terms of share of mayors who were women in 2013 were Cuba and Jamaica at 29 percent In Panama, the number of female mayors decreased (UN, 2015) Towards Equal? Women in Central America 31 Figure 1.4: Percentage of mayors who are women, 2000 and 2014 50 40 30 % 20 10 0 2000 2013 Guatemala LAC El Salvador Costa Rica Honduras Nicaragua Panama Source: United Nations (2015) from 13.5 percent in 2000 to 9.3 percent in 2013. As Violence against women with national representation and political appoint- ments, female political representation at the local level Violence against women is widespread in Central is the lowest in Guatemala with 2.1 percent. America.17 There are few sources of high quality data on the prevalence of violence against women, and While they have increased female participation in those that do exist are not recent in many cases. The decision-making, it is unclear what the social impact data that are available in the region show high rates of of gender quotas has been in the region. It is import- physical violence against women (Box 1.3). ant to evaluate whether the quotas that have increased women’s participation in institutions have also ben- While this section focuses on documenting vio- efitted women across the socioeconomic spectrum. lence against women in Central America, this topic is Given the large differences in agency, endowments, also evaluated in Chapter 4 within the larger context and access to economic opportunity among women of social norms. Evidence shows that social norms with different demographic and socioeconomic char- and attitudes that husbands should have authority acteristics in the Central American countries, it is par- over their wives and women relatives are prevalent in ticularly important to assess the extent to which all Central American countries. Such beliefs, as well as groups are effectively represented in institutions. The 17 According to the 1993 United Nations Declaration on the Elimi- existing evidence from Costa Rica suggests that, de- nation of Violence Against Women, “Violence against Women and Girls” refers to any act of gender-based violence that results in, spite the remarkable progress that has been made to- or is likely to result in, physical, sexual, or psychological harm or suffering to women or girls, including threats of such acts, wards gender equality in participation, disadvantaged coercion, or arbitrary deprivation of liberty, whether occurring in groups, particularly indigenous women and women public or in private life. Violence against women and girls is also referred to as violence against women, gender-based violence, with disabilities, remain underrepresented in the po- or sexual and gender-based violence. Both men and women can litical realm (CEDAW, 2011). be victims or perpetrators of violence, but the characteristics of violence commonly committed against women and men differ; in particular, women are more likely to be physically assaulted or murdered by someone they know. Chapter 1 Agency 32 social justifications for wife-beating, create an envi- also women themselves internalize their subordinate ronment in which not only is violence tolerated but roles vis-à-vis men. Box 1.3: Data availability and measurement challenges of gender-based violence More and better data on violence against women (VAW), included data disaggregated across different groups of women, is needed to adequately measure and monitor VAW and the ef- fectiveness of states’ responses to the problem. The recommended approach for collecting prevalence data on VAW is through dedicated population-based surveys specifically designed to gather detailed information on different forms of violence against women (its incidence, prevalence, nature, severity, consequences, and the relation of the victim to the perpetrator). A key disadvantage of these specialized studies is their cost, which presents a challenge for repeating them on a regular basis. One cost-effective way to gather these data may be to add questions or modules concerning VAW to the numerous ongoing surveys. However, figures on violence are highly sensitive to interview methodology, and using the wrong one can result in poor data quality and - even more importantly – concerns about the safety of participants and interviewees. In all types of surveys, even when applying the standards as outlined in international guidelines (WHO, 2003 and United Nations, 2014a), prevalence rates are likely to be underestimated and not internationally comparable because victims can be reluctant to provide sensitive information during interviews. Due to severe underreporting, administrative data cannot be used to estimate the preva- lence of violence against women in the population. Administrative data is collected routinely through public and private agencies such as health centers, police stations, courts, and shel- ters that come into contact with women who have suffered violence. There may be few services available that specifically care for victims or violence against women may be an issue that is largely ignored by police, service staff, or society in general. In these situations, not only will survivors be more reluctant to come forward for support, but staff will also be less inclined to recognize and document these needs. In addition, the reliability and validity of data often varies considerably between service agencies as data collection is not their primary respon- sibility and hence they often do not apply standardized procedures in collecting information. Double counting is another common issue, whereby women seeking repeated services from the same agency or from more than one agency are counted more than once. The optimal case for informed policymaking is achieved when policymakers can base their decisions on both outcome and process indicators. Ideally, this means survey data on preva- lence of VAW combined with administrative data on service usage and protection mechanisms, which together can be used to improve monitoring and evaluation. While surveys are import- ant for measuring the prevalence and trends of gender-based violence, administrative records are essential for monitoring and evaluating the effectiveness of government interventions and services. In order to improve data collection in this field, more and better services for victims and survivors of violence is required in parallel with a reduction in the stigma and discrimina- Towards Equal? Women in Central America 33 tion faced by victims. For example, specialized training for service providers is required to sen- sitize them to the problem and to support general monitoring efforts. In an attempt to address this problem, Costa Rica is harmonizing the surveillance and measurement of gender-based violence through the Unified System for the Statistical Measurement of Gender-based Violence ( Sistema Unificado de Medición Estadística de la Violencia de Género en Costa Rica ). Figure 1.5: Percentage of women ever married or in a union aged 15-49 who reported ever experiencing intimate partner violence by type of violence El Salvador 2008 14,8 9,4 2,1 Guatemala 2014/5 17,3 4,5 5,2 Nicaragua 2006/7 16,3 10,7 2,4 Honduras 2011/12 20,2 5,1 6,5 0 5 10 15 20 25 30 35 Physical but not sexual Physical and sexual Sexual but not physical Source: Bott et al. (2012). Notes: Surveys classified women as “ever married or in a union” if they had ever married or lived with a male sexual partner Reproductive health surveys asked women about violence by any current or former partner in life. All Demographic and health surveys asked about violence by the current or (if no current partner) by the most recent partner only. A partner was defined as a husband or cohabiting male sexual partner. Intimate partner violence is a persistent problem As is the case globally, intimate partner violence is in the Central American countries for which quali- a significant factor in the observed high rates of vio- ty prevalence data are available. In Honduras (in lence against women in Central America. For example, 2011/12), 31.8 percent of women reported having ex- about 60 percent of Honduran women who reported perienced physical, sexual, or both forms of violence being the victims of some form of violence identified from an intimate partner. The total rates of these dif- the main perpetrator being their current spouse/part- ferent forms of violence combined are similar in the ner (35 percent) or their former spouse/partner (24 other countries for which this type of data is avail- percent).18 In El Salvador, 26.3 percent of women re- able. In El Salvador (2008) 26.1 percent of women, in ported being the victim of abuse by their current or Guatemala (2014/2015) 27 percent, and in Nicara- former partners, and 7.7 percent of women reported gua (2006/2007) 29.4 percent reported being exposed during their lives to physical, sexual, or both forms of 18 Estimates based on Honduras’ Demographic and Health Survey violence from an intimate partner (Figure 1.5). (Encuesta Nacional de Demografía y Salud, ENDESA), 2011- 2012. Chapter 1 Agency 34 Figure 1.6: Femicide rate in Central America, per 100,000 women, 2016 11,0 10,2 2,5 1,2 0,9 0,5 El Salvador Honduras Guatemala Nicaragua Panama Costa Rica Source: Economic Commission for Latin America and the Caribbean, CEPALSTAT. http://estadisticas.cepal.org/cepalstat/ WEB_CEPALSTAT/buscador.asp?idioma=i&string_busqueda=femicide Note: LAC average calculated for the countries available in the CEPALSTAT database. recent (in the previous 12 months) physical or sexual counts for about 38.6 percent of all female homicides violence by their intimate partner (Bott et al., 2012).19 globally and 40.5 percent in North and South Ameri- Domestic violence has significant implications for in- ca (Stockl et al., 2013). Stockl et al. (2013) also found tergenerational violence. Girls exposed to violence in that “across all countries where such data are collected, childhood are at higher risk of being victims of vio- women’s main risk of homicide is from an intimate lence later in life. Nearly one in four (24 percent) Sal- partner.” According to the Economic Commission for vadorian women between the ages of 15 and 49 wit- Latin America and the Caribbean (ECLAC) Observa- nessed physical violence perpetrated by a man against a tory of Gender Equality, El Salvador was the country woman within the household, and 31 percent reported in LAC with the highest femicide rate in 2016 at 11.0 being a victim of violence before the age of 18.20 per 100,000 women, closely followed by Honduras The incidence of femicide is high in El Salvador, with 10.2 per 100,000 (Figure 6). Guatemala had the Honduras, and Guatemala (Figure 6). Femicide is fourth highest rate in the region at 2.5 per 100,000 generally understood to involve the murder of women (after the Dominican Republic, not shown). Nicara- because of their sex. Most cases of femicide are com- gua, Costa Rica, and Panama had lower femicide rates, mitted by partners or ex-partners and involve ongoing with Costa Rica reporting the lowest rate in Central abuse in the home, threats or intimidation, sexual vi- America at 0.5 per 100,000 and among the lowest in olence, or situations where women have less power or LAC. However, as with all international comparisons fewer resources than their partner (WHO and PAHO, of crime statistics, these figures should be interpreted 2012). This is known as intimate femicide and ac- with caution given the obstacles to collecting accurate comparative data in this area. 19 Pan-America Health Organization (PAHO) based on on El Salva- dor’s Reproductive Health Survey (Encuesta Nacional de Salud The high rates of femicide observed in El Salvador, Familiar, FESAL), 2008. Honduras, and Guatemala may be explained in part by 20 Estimates based on El Salvador’s Reproductive Health Survey (Encuesta Nacional de Salud Familiar, FESAL), 2008. the high rates of violence observed in these countries Towards Equal? Women in Central America 35 Box 1.4: Violence in the Northern Triangle The countries of the Northern Triangle - El Salvador, Honduras, and Guatemala - have very high homicide rates, due in large part to the presence of strong gangs, drug-trafficking, and orga- nized crime such as human trafficking. In recent years, El Salvador and Honduras have had the highest homicide rates worldwide outside of active conflict zones. While the global average is about 5.3 homicides per 100,000, the official homicide rates were 81.2 per 100,000 in El Sal- vador and 59.0 per 100,000 in Honduras in 2016. Guatemala’s homicide rate is also considered to be very high at 27.3 per 100,000. Although still above the global average, Costa Rica, Pana- ma, and Nicaragua have substantially lower homicide rates (at 11.8, 9.3, and 7.0 per 100,000 people respectively). Note: The global homicide rate is based on 2014 data in the UN Office on Drugs and Crime’s International Homicide Statistics. Country-specific rates were collected by InSight Crime. (Box 1.4). In general, as they are more likely to belong er, punishment and convictions are not the only mea- to gangs and be involved in organized crime, men are sures that should be used to evaluate the efficacy of more likely to both be the perpetrators and victims laws on violence against women; other outcomes like of violent crime than women. Even so, El Salvador protection and service delivery are also important. and Honduras reported the second and third highest The percentage of female victims of violence who rate of female homicides in the world between 2010 seek any type of help, particularly from an institution, and 2015, after Lesotho.21 In particular, El Salvador, is low throughout Central America. Across all coun- Honduras, and Guatemala averaged 13.5, 13.4, and tries for which information is available, the share of 8.2 female homicides per 100,000 women between female victims of violence who sought institutional 2010 and 2015. Panama, Costa Rica, and Nicaragua help was much lower than that of women who turned averaged 2.4, 2.2, and 1.8 per 100,000 women re- to family and friends. The fraction of women who spectively. As gangs have grown in some countries in seek institutional help or turn to family or friends the region, the number of women in gangs has also when they experience IPV is relatively higher in El increased (Box 1.5). Salvador than in other LAC countries, but only 36 Extremely low levels of convictions for all types of percent of women who reported experiencing IPV homicides, including femicides, mean that a majority in the previous 12 months had sought institutional of these crimes remain unpunished. Approximately 98 help (Bott et al., 2012). In Panama, only 5 percent percent of femicide cases in Guatemala end in impuni- of women who experienced physical violence sought ty (OECD, 2014). In El Salvador, during the five-year institutional help, and only 8 percent of the women period between 2005 and 2010, only 5 percent of the turned to family or friends after experiencing a violent cases of female homicides were taken to trial.22 Howev- episode.23 Two-thirds of Guatemalan women received 21 Small Arms Survey Database. While in recent years there has Defense of Rights (PDDH, Procuraduría para la Defensa de los been a concerted effort to improve data collection on this topic, Derechos). the dearth and low quality of available data in many countries 23 Estimates based on Panama’s National Survey of Sexual and makes it very difficult to make comparisons between countries. Reproductive Health (Encuesta Nacional de Salud Sexual y 22 According to El Salvador’s Office of the Prosecutor for the Reproductiva, ENASSER), 2009. Chapter 1 Agency 36 Box 1.5: Gender relations within gangs A 2010 qualitative study of gender relations in the MS-13 and Barrio-18 gangs in El Salvador, Guatemala, and Honduras found that women join gangs for similar reasons as men: the need to belong to a group, search for protection and affection, the need for money, the desire for rec- ognition, and escape from a family environment of conflict and violence (Interpeace, 2013). Female gang members typically become members through relationships with male members, though they may also need to undertake initiation rituals (for example, choosing between be- ing raped or beaten by other gang members for a specific amount of time). Although female gang members are often required to engage in violence, which is the main component of the prevalent male identity within the group, they continue to fulfill traditional female roles within the gang such as being partners to male gang members, being caregivers, doing household tasks, and caring for the sick. Gangs reproduce in an exacerbated or extreme way the existing social norms pertaining to male domination and women´s roles within the group. Men permanently control and dominate women within gangs. This is manifested in different ways - for instance, even when a member is in jail, gangs have systems of control over the imprisoned gang member’s partner(s). Sexual and physical abuse are widespread, but the female body is also used for transport and support in criminal activities where gender stereotypes may help them evade capture. Source: Interpeace (2013) some form of assistance to help them cope with or graphic and Health Surveys (DHS) in 30 countries escape the situation of violence, but only half of these found that only 40 percent of survivors of intimate sought institutional help. Among indigenous women, partner violence (IPV) had ever sought help from any this share was even lower (25 percent).24 In Hondu- formal or even informal source of support (Klugman ras, 29 percent of women who reported experiencing et al., 2014). According to Palermo et al. (2014), violence turned to family or friends after a violent only 14 percent of survivors in LAC made any for- episode, while the percentage of women who sought mal disclosure of their experience of violence. Among institutional help was much lower (19 percent).25 those women who did seek help, the majority turned Sadly, low rates of victims seeking support is a to family and friends, and only a small percentage common phenomenon worldwide due to prevalent sought assistance from the police (United Nations, social norms and a distrust of institutions. Globally, 2015). The main factors explaining the low rates of very few abused women report violence to the police women turning to institutions in all countries include or to dedicated support services. Around the world, a lack of knowledge of available services or barriers to most female victims of violence do not seek institu- accessing them such language barriers, fear of retalia- tional help. World Bank analysis of data from Demo- tion by their families, fear of reliving the experience, embarrassment and wanting to keep the issue private, 24 Estimates based on Guatemala’s Reproductive Health Survey (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2008-09. or feeling that the institutions will not be able to help 25 Estimates based on Honduras’ Demographic and Health Survey them (United Nations, 2015). In Guatemala and Pan- (Encuesta Nacional de Demografía y Salud, ENDESA), 2011- 2012. ama, for instance, rape survivors frequently do not re- Towards Equal? Women in Central America 37 Box 1.6: In focus: Addressing violence against women through community-based interventions: Sasa! in Honduras Sasa!, an integrated and multi-level community-based approach to transforming gender rela- tions, works to change attitudes about tolerance for and the acceptability of IPV and to promote collective action (http://raisingvoices.org/sasa/). First developed by Raising Voices in Uganda to address the root causes of violence against women, Sasa! is an exploration of power—what it is, who has it, how it is used, how it is abused, and how power dynamics between women and men can change for the better. It is organized into four phases designed to ensure that organi- zations can more effectively and systematically facilitate a process of change in the commu- nity. Sasa! means “now” in Kishwahili, but it is also an acronym for the phases of the approach: Start, Awareness, Support, and Action. Evaluations of Sasa! have shown reductions in physical (50 percent) and sexual (33 percent) intimate partner violence (World Bank, 2016a). Adapting Sasa! to the Honduran context has involved revising posters, images, and messag- es to ensure they are culturally appropriate while still aligned with the original methodology. It has also involved providing in-depth training to community leaders and activists in each of the four phases under the guidance of a certified Sasa! trainer. The implementation has started in El Progreso, one of the three municipalities that are beneficiaries of the World Bank’s Safer Municipalities Project, where the local authorities have committed to continuing with the com- plete implementation of Sasa! throughout the four phases. While results are not yet available, this type of community-based approach is an important step towards reducing rates of IPV. Source: World Bank (2016a) port crimes due to social stigma, a lack of confidence in the justice system, and fear of retaliation.26 In Hon- duras, female victims of violence also cited shame, fear of retaliation, not knowing where to go, and not believing that anyone would help them as reasons for not seeking help.27 This hesitation among victims of domestic and sexual assault is likely to be exacerbated by a scarcity of female police officers (Meier and Nich- olson-Crotty, 2006). 26 USDS (2015), and ENASSER 2009. 27 Reproductive Health Survey 2001. Chapter 1 Agency 38 Annex 1.1 Table A1.1: National policies and plans on gender equality Country Name Years Priorities Six strategic objectives: (1) care and social responsibility; National Policy for Gender Equality (2) quality paid job and income generation; (3) education and Equity - Política Nacional de la and health services in favor of equality; (4) effective 2007- Costa Rica Igualdad y Equidad de Género – PIEG. protection of women’s rights and eradication of all forms of 2017 Made operational through two action violence; (5) increased political participation of women and plans (2008-2012 and 2012-2014). achievement of gender parity; and (6) strengthening of the institutional framework for equality and gender equity. Ten global pillars for intervention: (1) equity in socio- political participation; (2) cultural identity; (3) economic National Policy for the Promotion development; (4) the workforce; (5) natural resources, land, and Comprehensive Development 2008- and housing; (6) justice; (7) the eradication of violence, Guatemala of Women (Política Nacional de 2023 discrimination, and racism against women; (8) holistic Promoción y Desarrollo Integral de La health services; (9) education with attention to gender and Mujer) cultural identity; and (10) strengthening of institutional mechanisms for the advancement of women. Aimed at: (1) promotion and protection of the rights of women and adolescents to peace and a life free from violence; (2) promotion and protection of women’s social and political participation rights as well as the right to Honduras, the Second Plan for exercise their citizenship; (3) promotion and protection 2010- Honduras Gender Equality (II Plan de Igualdad y of women’s sexual and reproductive rights; (4) promotion 2022 Equidad de Género en Honduras) and protection of women’s right to education; (5) promotion and protection of the right to employment, access, use, and control of resources; and (6) gender, access, sustainable use and control of the biodiversity, and natural resources management risk. The Plan for Equal Opportunities for Women – (Plan de Igualdad de Seeks to implement structural changes to reduce gender Oportunidades para las Mujeres) equalities in the economic, social, political, and cultural 2015- Panama operationalizes the National Policy for realms, the integration of a gender perspective in public 2019 Equal Opportunities for Women 2012 policies and the inclusion of Panamanian women, youth – (Política Nacional de Igualdad de and girls without discrimination. Oportunidades para las Mujeres)28 El Salvador The National Women’s Policy – 2012- Identifies six priority areas for gender equality: (1) (Política Nacional de la Mujer, PNM) is 2017 economic autonomy; (2) a life free of violence; (3) made operational through the National inclusive education; (4) integral health; (5) care and social Plan of Equality and Equity for the protection; and (6) citizen and political participation. Women of El Salvador (Plan Nacional de Igualdad y Equidad para las Mujeres Salvadoreñas, PNIEMS). 28 Information on Panama’s national policy comes from “Política Pública de Igualdad de Oportunidades para las Mujeres (PPIOM)” by the National Institute of Women (Instituto Nacional de la Mujer). Towards Equal? Women in Central America 39 40 chapter 2 Endowments I nvestments in human capital are critical to enable individuals to benefit from opportunities and ful- fill their potential as active members of society. Dif- household resources; (ii) the acceptance of domestic violence; and (iii) child marriage (World Bank, 2014). Two-thirds had experienced all three deprivations. ferences between men and women in access to, and Women with a secondary education or higher, on the the accumulation of, basic endowments such as health other hand, were far less likely to experience these and education can thus lead to the perpetuation of deprivations, with 18 percent having faced one and unequal gender opportunities throughout the lifecy- only 5 percent having faced three deprivations. cle. In addition, the persistence of such inequalities Education is also an important protective factor entails not only substantial losses for individuals and from violence against women (VAW). In Honduras, families but also economic and social costs for com- Panama, Costa Rica, and Nicaragua, more education munities in the long term. Research also shows that is correlated with lower exposure to violence (Bott et these costs can play a significant role in the inter-gen- al., 2012). Honduran women with higher education- erational transmission of gender inequalities (World al levels reported lower levels of all types of violence; Bank, 2012b). for instance, 6.6 percent of women with a tertiary Globally, access to education is correlated with in- education reported having experienced physical vio- creased agency for women, while the deprivation of lence from their intimate partner in the previous 12 agency is associated with a lack of education. Based months compared to 11 percent among women with on data from 54 countries, a World Bank study found no education. In Panama, 14.1 percent of women that 90 percent of women with a primary education with only a primary education experienced violence or less had experienced at least one of three ways of compared to 4.8 percent of women with some tertia- being deprived of agency: (i) a lack of control over ry education. Towards Equal? Women in Central America 41 Figure 2.1: Life expectancy men and women, 1965 and 2015 85 80 75 70 Life expectancy 65 60 55 50 45 40 1965 2010 1965 2010 Male Female CRI GTM HND NIC PAN SLV LAC Source: World Bank’s World Development Indicators, 2017. This chapter focuses on two types of endowments, years, continue to outperform the regional average for health and education, and the extent to which women both men and women. in Central America have access to these. These are the Women made larger gains in life expectancy than basic building blocks of human capital, with direct men over the past 50 years in five of six countries. The implications for the social and economic inclusion largest gains have been in Guatemala and Nicaragua, of women. The first section of the chapter focuses on where women’s life expectancy grew by more than 51 health outcomes, in particular the region’s high lev- percent – totaling an additional 26 and 26.5 years of els of maternal mortality and teenage pregnancy. The life respectively (Table 2.1). Men in these countries second section considers women’s access to education saw large gains as well, but, at 43.9 and 47.4 percent throughout Central America. respectively, they did not keep up with the gains of women. The only country in which men’s life expec- tancy grew more than that of women was Honduras. 2.1 Health Even there, however, women’s life expectancy had in- Improved health outcomes in the region have led creased by 47.7 percent over 50 years, reaching 76 in to important gains in life expectancy, especially for 2015, while the life expectancy of Honduran men had women, since 1965 (Figure 2.1). These gains have grown by 48.3 percent, reaching 71. been particularly notable in Nicaragua, Guatemala, Gender gaps in mortality are partially explained by Honduras, and El Salvador, all of which in 1965 sig- the different health and lifestyle risks faced by men nificantly trailed behind the LAC average life expec- and women across countries. For example, alcohol use tancy for both men and women. By 2015, this gap and related mortality and morbidity are particularly had largely closed, especially for women. Panama and high for men compared to women in all countries in Costa Rica, where women’s life expectancy is over 80 the sub-region (Monteiro 2007). And even though Chapter 2 Endowments 42 Table 2.1: Growth in life expectancy, 1965 - 2015 Female Male % Years % Years Guatemala 51.6% 26.0 43.9% 21.3 Nicaragua 51.2% 26.5 47.4% 23.2 Honduras 47.7% 24.5 48.3% 23.1 El Salvador 40.4% 22.3 35.7% 18.0 Latin America 29.8% 18.0 28.3% 15.9 Costa Rica 25.5% 16.7 23.7% 14.8 Panama 25.5% 16.4 20.4% 12.7 Source: World Bank’s World Development Indicators, 2017 women have higher obesity rates, men are more like- average. El Salvador and Costa Rica had the lowest ly to use tobacco products, putting them at higher rates (1.9 and 1.8 respectively). Each country has had risk of cardiovascular disease, respiratory problems, important declines in fertility rates, but the drop has strokes, and various cancers. For example, in Panama been largest in Honduras - falling by 4.9 births per men were more likely to have elevated blood pressure women since 1970. levels than women (33 percent and 23 percent respec- Fertility rates are higher among low-income wom- tively) and are more likely to use tobacco products of en and women living in rural communities than for some form (23 percent of men and 4 percent of wom- other women. In Guatemala, women in the lowest en).29 As shown below, maternal mortality, including wealth quintile have, on average, 4.9 children, while the risks associated with teenage pregnancy, continue among the wealthiest, fertility rates are only 1.9 chil- to be a cause for concern in Central America in the dren per woman. Fertility is also higher among in- context of women’s health outcomes. digenous women, who have on average 3.6 children compared to only 2.8 children for non-indigenous Maternal mortality and women. Similarly, women in rural areas have an aver- access to health services age of 3.7 children each compared to only 2.5 for ur- ban women.30 In Nicaragua, fertility is higher in rural In Central America, fertility rates have declined in re- areas than urban areas, 2.9 compared to 2.1, although cent years but remain relatively high in Guatemala and the urban-rural gap has narrowed in recent years (in above the replacement rate in three other countries for 2007, the total fertility rate was 3.5 in rural areas com- which data are available (Figure 2.2). As of 2015, the pared to 2.2 in urban areas).31 fertility rate was highest in Guatemala (3.2), followed by Honduras and Panama (both 2.4) and Nicaragua (2.3). These four countries were above the replace- 30 Estimates based on Guatemala’s Reproductive Health Survey (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2014- ment rate of 2.1 in 2015, which was also the LAC 2015. 31 Estimates based on Nicaragua’s Demographic and Health Sur- vey (Encuesta Nacional de Demografía y Salud, ENDESA), 2007 29 (WHO, 2014). and 2012. Towards Equal? Women in Central America 43 Figure 2.2: Fertility rates 1970-2015 8 Fertily rate (births per woman) 7 6 5 4 3 2 1 0 Costa Rica El Salvador Nicaragua Honduras Panama Guatemala 1970 1980 1990 2000 2010 2015 Source: World Bank’s World Development Indicators, 2017 Note: The solid line represents the replacement rate. Maternal mortality and complications related to ly since 1990, Honduras’ and Guatemala’s still high pregnancy continue to represent significant health mortality rates represent significant gains since 1990. risks for women in Central America. Although ma- The estimated maternal mortality rates in Hondu- ternal mortality rates (MMR) have decreased in all six ras and Guatemala in 1990 were 272 and 205 per countries, in most of Central America, they remain 100,000 live births, the highest in Central America. El high relative to the LAC region and to developed Salvador also made large gains, reducing MMR from economies (where the maternal mortality rate is 69 157 to 54. and 12 respectively) (Figure 2.2 and UNICEF 2014). To reduce maternal and infant mortality rates, the This is especially the case in Nicaragua and Honduras governments in the region have taken important in- with 150 and 129 maternal deaths per 100,000 live stitutional steps to broaden access to maternal health births respectively in 2015. Guatemala and Panama services. Between 2006 and 2013, the government of also have rates above the LAC average, at 94 and 88 Nicaragua increased the number of health and mater- respectively. At the other end of the spectrum, El Sal- nity facilities with health facilities for women increas- vador and Costa Rica both have rates that are below ing from 116 to 143, and the number of dedicated the LAC average (54 and 25 respectively). maternity facilities (Casas Maternas) also increasing While Honduras, Guatemala, and El Salvador (United Nations, 2014b). In 2008, the Honduran have all made significant progress in reducing mater- government initiated a policy called the Accelerat- nal mortality, Nicaragua and Panama did not meet ed Reduction of Maternal Mortality and Childhood the Millennium Development Goal (5a) to reduce 2008-2015 (RAMNI) to continue the development the MMR by three-quarters between 1990 and 2015 of integrated actions to improve maternal and child (WHO, 2015). While Nicaragua’s and Panama’s ma- health. In 2011, Guatemala established the Multi- ternal mortality rates have improved only marginal- sectoral Committee for Safe Motherhood (Comisión Chapter 2 Endowments 44 Figure 2.3: Maternal Mortality Rates, 1990 and 2015 300 250 Maternal mortality ratio 200 150 100 50 0 NIC HND PAN GTM LAC SLV CRI 2015 1990 Source: WHO (2015) Figure 2.4: Share of births attended by Figure 2.5: Pregnant women receiving skilled personnel (% of total) prenatal care (%) Guatemala (2015) 65,5 Guatemala (2015) 91,3 Honduras (2012) 82,8 Panama (2013) 93,4 Nicaragua (2012) 88,0 Nicaragua (2012) 94,7 LAC (2012) 93,3 El Salvador (2014) 96,0 Panama (2014) 93,9 Honduras (2012) 96,6 El Salvador (2014) 98,0 LAC (2012) 96,9 Costa Rica (2014) 99,2 Costa Rica (2011) 98,1 Source: World Bank’s World Development Indicators, 2017; UNICEF’s State of the World’s Children, Childinfo, and Demographic and Health Surveys. Note: The year of the most recent source survey is included in parentheses. Pregnant women receiving prenatal care are the percentage of women who were attended by skilled health personnel at least once during pregnancy for reasons related to the pregnancy. Multisectorial para la Maternidad Saludable) with the mala is the worst performer with only 65.5 percent of objective of monitoring and evaluating a strategy to births being attended by skilled personnel. In Hon- reduce maternal mortality (United Nations 2011). duras, the rate of 82.8 percent in 2012 was a signifi- cant increase (16 percentage points) over the number Three of the six countries in Central America of attended births that occurred in the period 2001- underperform the region in terms of the percentage 2005.32 Nicaragua also trails the region with only 88.0 of births attended by skilled personnel (Figure 2.4). While more recent data are unavailable, data from 2012-2015 show that Guatemala, Honduras, and 32 Estimates based on Guatemala’s Demographic and Health Sur- vey (Encuesta Nacional de Demografía y Salud, ENDESA), 2006 Nicaragua trail behind the rest of the region. Guate- and 2012. Towards Equal? Women in Central America 45 percent of births attended by skilled health staff. This in Panama only 37.6 percent of births among women indicator is almost universal in Costa Rica and El Sal- with no education were assisted by health staff com- vador, with 99.2 percent and 98 percent respectively pared to 99.7 percent of births among women with in 2014. Panama is in line with the regional average some college education.35 In Costa Rica, 92 percent of for LAC, where just over 93 percent of births are at- women with a university education received prenatal tended by medical personnel. care compared with only 77 percent of women with no education. 36 In addition, the share of women at- The countries of the region perform better in terms tended by skilled health staff at birth was 8 percentage of access to prenatal care, with over 90 percent of preg- points lower than the national average for those with nant women in each country benefitting from prenatal only a primary education.37 care (Figure 2.5). Yet only Costa Rica exceeds the re- gional average when it comes to the share of pregnant Evidence suggests that indigenous women and women who receive prenatal care, with 98.1 percent women in rural areas face higher rates of maternal compared to an average of 97 percent in LAC. Hon- mortality. Maternal deaths are higher among poor, duras and El Salvador are on a par with the regional rural women than among their urban counterparts average. In both countries, this represents sizeable in- (USDS, 2016). In Panama, for example, the pro- creases---14 percentage points since 2001 in Honduras portion of births attended by skilled personnel was and 10 percentage points since 2003 in El Salvador. In 93.5 percent in 2011 but just 51.3 percent for the the other three countries, between 5 and 9 percent of Ngäbe-Buglé district or comarca and 44.6 percent for pregnant women do not receive prenatal care. the Embera comarca.38 The combination of limited prenatal care services, lower proportions of skilled Substantial differences in access to these key ma- birth attendance, and, a lack of a reliable rural referral ternal health services exist depending on women’s in- network contributed to rural maternal mortality ra- come and education levels. In Nicaragua, for example, tios that were six times higher (120 per 100,000) than while some departments have nearly universal access those in urban areas (20 per 100,000). In Guatemala, to such services, in the poorer regions (the North and indigenous women account for almost three-fourths South Atlantic Autonomous Regions), considerably of maternal deaths but only 42 percent of the popu- fewer women have access to prenatal checkups (85.6 lation (Sanchez et al., 2016). In Nicaragua, a dispro- percent and 84.9 percent respectively).33 In El Salva- portionately high share (more than half ) of maternal dor, only 65 percent of women in the lowest wealth deaths in the first decade of the 2000s occurred in quintile had their first prenatal care in the first quarter the Caribbean coast and the center-north regions, of pregnancy compared to 88 percent of women in areas that are rural with larger shares of indigenous the highest wealth quintile (Samandari and Speizer, peoples (UNFPA, 2012a). A 2004 study in Hondu- 2010). In Honduras, only 88 percent of women with ras also found that the maternal mortality rate ranged no education received prenatal care compared to 99 percent of those with a tertiary education.34 Similarly, 35 Estimates based on Panama’s National Survey of Sexual and Reproductive Health (Encuesta Nacional de Salud Sexual y Reproductiva, ENASSER), 2009. 33 Estimates based on Nicaragua’s Demographic and Health Sur- vey (Encuesta Nacional de Demografía y Salud, ENDESA), 2012. 36 Statistics published by the Ministry of Health. 2010. Costa Rica. 34 Estimates based on Honduras’ Demographic and Health Survey 37 Statistics published by the Ministry of Health. 2010. Costa Rica. (Encuesta Nacional de Demografía y Salud, ENDESA), 2011- 38 Comarcas are administrative divisions in Panama home to 2012. indigenous peoples. Chapter 2 Endowments 46 Box 2.1: Barriers to health services for indigenous women in Honduras and Guatemala While data specific to indigenous and Afro-descendant women are scarce, there is evidence that cultural barriers reduce the efficacy of health care for some women. In the recent “Poli- cy Agenda for Indigenous and Afro-Honduran Women” (SEDINAFROH, 2012-2013), women re- ported not only that there were an insufficient number and poor quality of health clinics and staff within indigenous communities but also discrimination and a lack of sensitivity among health staff to a world view that values the practice of holistic, traditional medicine. Govern- ment health centers do not recognize, certify, or resource midwives even though they are vital to prenatal care and births for indigenous and Afro-Honduran women. This constitutes a sig- nificant cultural barrier to improving maternal and reproductive health outcomes within these communities. Even where indigenous women have physical access to services, they may refrain from availing themselves of the care that they need because they have experienced and/or antici- pate experiencing discrimination. While this problem can be hard to quantify, there is evidence that substantial cultural and language barriers do exist. For example, indigenous women in Guatemala report that they have been subjected to comments about their perceived excessive fecundity from non-indigenous medical professionals. These discriminatory attitudes com- bined with the unavailability of culturally appropriate and accessible services often discour- age indigenous women and girls from accessing health care services (United Nations, 2011). Human rights advocates have argued that there should be a comprehensive policy or interna- tional plan to address these issues regarding the provision of health care to indigenous peo- ples (UNHRC, 2010). Source: United Nations (2011) between 190 and 255 deaths per 100,000 live births Maternal mortality rates in Honduran departments in the departments that were home to the majority with larger indigenous populations are significantly of indigenous populations (Colón, Copán, Intibucá, above the national average (ECLAC/OPS/UNFPA, Lempira, and La Paz). 2010). Similarly, in Guatemala, there are wide differ- ences by ethnicity with regard to the share of births Access to maternal health services is especially low attended by skilled staff--81.3 percent of non-indig- among rural and indigenous women. In Honduras, enous women compared to only 46.5 of indigenous for instance, women in urban areas are more likely women.40 Not only is access to health facilities much to receive prenatal care (99 percent) than women in lower for those rural populations, but discriminatory rural and indigenous areas (95.9 and 76.9 percent re- practices and lack of cultural sensitivity have also been spectively). Similarly, 99.1 percent of births in urban reported (see Box 2.1). High fertility rates, little access areas were assisted by skilled health personnel, while to and low quality of care, and gender and cultural the share decreased to 83.3 percent in rural areas and to less than half (43.7 percent) in indigenous areas.39 39 Estimates based on Honduras’ Demographic and Health Survey 40 Estimates based on Guatemala’s Reproductive Health Survey (Encuesta Nacional de Demografía y Salud, ENDESA), 2011- (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2014- 2012. 2015. Towards Equal? Women in Central America 47 Table 2.2: Share of women ages 15-49 who use modern contraceptives (%) circa 2000 circa 2015 Guatemala 30.9 (1999) 48.9 (2015) Panama 48.8 (2009) 60.1 (2013) Honduras 50.9 (2001) 63.8 (2012) El Salvador 61.7 (2003) 68.0 (2014) Nicaragua 66.1 (2001) 76.5 (2012) Costa Rica 71.5 (1999) 74.7 (2011) Source: World Bank’s World Development Indicators, 2017 Notes: This indicator reports the percentage of women who are married or in a union and aged 15 to 49 who are currently using, or whose sexual partner is using, at least one modern method of contraception such as female and male sterilization, oral hormonal pills, the intra-uterine device (IUD), the male condom, injectables, the implant (including Norplant), vaginal barrier methods, the female condom, and emergency contraception. Panama’s ENASER 2009 survey reported this indicator at the level of 59.4 percent. http://www.contraloria.gob.pa/inec/Aplicaciones/ENASER/EnasserInformeFinal.pdf barriers to services reduce gains in maternal health for tives, fewer than 10 percent of women in Costa Rica indigenous and Afro-descendant women. and Nicaragua who wanted to use contraceptives have reported not being able to access them (United Na- There has been a noticeable increase in modern tions, 2017). Women in Guatemala, the country with contraceptive uptake across Central America since the the lowest contraceptive uptake, have nearly universal early 2000s. Researchers estimate that, as of 2010, use knowledge of contraceptive methods, and few have of contraceptives had reduced maternal mortality by reported having an unmet need.41 Instead, among almost 44 percent globally. In Central America, the the most common reasons they cited for not using proportion of maternal deaths averted by contracep- contraceptives were because they did not have sexu- tives ranged from 40.6 percent in Guatemala to 67.5 al relations or not very frequently (22 percent), they percent in Costa Rica (Ahmed et al., 2012). However, were breastfeeding (11 percent), they were concerned only a minority of Guatemalan women who are mar- about potential negative side effects (7 percent), or ried or cohabiting have reported using some form of they had moral or religions reasons (7 percent). The modern contraceptive as have fewer than two-thirds of country where women reported the highest rates of Panamanian and Honduran women (Table 2.2). The unmet need for contraceptives was Panama.42 Overall highest rates of uptake are in Nicaragua (with 76.5 one in five (21 percent) women reported having an percent) and Costa Rica (with 74.7 percent). In gen- unmet need for contraceptives. This was particularly eral, rates of contraceptive uptake are higher among the case for young women (including 55 percent of urban women and those with more schooling. Most women in Central America who want to use 41 Estimates based on Guatemala’s Reproductive Health Survey (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2014-15. contraceptives report having access to them. In line 42 Estimates based on Panama’s Multiple Indicator Cluster Survey with the high rates of uptake of modern contracep- (Encuesta de Indicadores Múltiples por Conglomerado, MICS), 2013. Chapter 2 Endowments 48 those aged 15-19 and 33 percent of those aged 20- Teenage pregnancy 24), lower-income women (35 percent of those in the poorest quintile), and nearly half (47 percent) of all Teenage pregnancy and early childbearing is a serious indigenous women. challenge in Latin America, including the countries of Central America. Recent studies have found a strong Abortion is restricted in all countries in Central link between early motherhood, lower educational at- America and is prohibited under all circumstances tainment, and poor employment outcomes for wom- with no exceptions for the health or life of the mother en. Azevedo et al (2012) have shown that teenage in El Salvador, Honduras, and Nicaragua. Interna- pregnancy can also have potential long-term negative tional research suggests that restrictive abortion laws implications for the child, the father, and the fami- are not associated with lower abortion rates; rather ly of the teenage mother. A regional study of LAC, they translate into a greater number of unsafe abor- for instance, found that teenage motherhood had a tions (Sedgh et al, 2012). Unsafe abortions account- clear negative effect on maternal outcomes and on the ed for approximately 13 percent of maternal deaths child, even when controlling for unobservable con- worldwide in 2008 (WHO, 2011) and are considered founding factors. In addition, it provided evidence of a leading cause of maternal mortality by the United a higher risk of maternal mortality, fetal death, infant Nations’ Convention on the Elimination of All Forms mortality, and the suicide of the mothers. of Discrimination Against Women. Despite the legal restrictions that are in place, an estimated 1.1 mil- Teenage pregnancy rates in most of Central Amer- lion abortions were carried out in Central America in ica surpass those of the LAC region. With the excep- 2008 (Sedgh et al., 2012). This suggests that many tion of Costa Rica, in 2015 the countries of Central Central American women are undergoing unsafe il- America had adolescent fertility rates (births per legal abortions. There are also concerns regarding the 1,000 women aged 15-19) that were higher than the way in which these laws are applied. For example, the LAC average (63.7). While rates in El Salvador and concluding observations of the UN’s 2017 CEDAW Honduras were only marginally higher than the re- report on El Salvador noted the “disproportionate gional average, Guatemala, Panama, and in particular criminal penalties applied to women seeking abor- Nicaragua had substantially higher rates. Rates were tion… [and] to women having had a miscarriage”43 highest in Nicaragua (88.1) followed by Guatema- and called for changes in the law to allow abortion la (80) and Panama (73.7). Although these rates are “at least in cases of rape or incest, threats to the life high, they represent significant improvements. Hon- and/or health of the mother, and severe fetal impair- duras has been the most successful in reducing the rate ment” (CEDAW, 2017). Costa Rica is the only Cen- between 1995 and 2015 (from 120.1 to 64.3). Guate- tral American country that permits abortion when the mala and Nicaragua also made substantial gains over mental health, not just physical health, of the mother those 20 years (Figure 2.6). is threatened.44 In all countries in the region, teenage pregnancy rates tend to be higher among rural and indigenous 43 Cases have been reported of women who suffered miscarriages women as well as among those with lower educational being investigated and charged with abortion-related crimes. and income levels (Azevedo et al, 2012). For example, 44 In Panama, abortion is permitted when the health of the mother is at risk and in cases of rape or incest. The legislation in Guate- nearly a quarter (24 percent) of Guatemalan teenage mala allows it in cases where the woman’s life is in danger. Towards Equal? Women in Central America 49 Figure 2.6: Adolescent fertility rates per 1,000 women ages 15-19, 1995, 2005, and 2015 140 133,6 120 108,1 100 88,1 80 86,6 60 66,6 40 56,0 1995 2005 2015 CRI GTM HND NIC PAN SLV LAC Source: World Bank’s Development Indicators 2017 girls in rural areas experienced a pregnancy compared A variety of factors are associated with teenage to 16 percent of those in urban areas.45 This rate was pregnancy. Azevedo et al (2012) found that in LAC similar for indigenous and non-indigenous girls (21.7 inequality seems to be correlated more with the high percent and 19.9 percent respectively). These rates incidence of teenage pregnancy than with per capita were seven times higher for girls with no education GDP. Kearney and Levine (2011) showed that income compared to those with higher education and three inequality across states in the US can explain partially times higher for girls in the lowest income quintile the geographic variation in teenage pregnancy rates than for those in the highest quintile. In Panama, in the country. A study of Demographic and Health teenage birth rates are higher in indigenous areas (32.4 Survey data also showed that household wealth and percent) and rural areas (24 percent), while teenage education are negatively correlated with teenage preg- pregnancy rates for women who have at least started nancy (Azevedo et al, 2012). Furthermore, as dis- secondary education are half of those for women with cussed in more detail in Chapter 4, aspirations, agen- no secondary education. In El Salvador, the highest cy, and social norms are likely to play a significant percentages of teenagers who are pregnant or are al- role as they can lead either to planned pregnancies or ready mothers are women with no education (48.8 to a lack of action to prevent pregnancies. Similarly, percent) or with primary education (39.8 percent). Näslund-Hadley and Binstock (2010) found that ad- Furthermore, women living in rural areas are more olescents with high aspirations in life are more likely likely to have been teenage mothers than women in to have children later in life. Those with fewer aspira- urban areas (27.4 and 18.6 percent respectively).46 tions lack any incentive to prevent pregnancy and may plan to get pregnant. 45 Estimates based on Guatemala’s National Survey of Maternal and Infant Health (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2014-2015. 2.2 Education 46 Estimates based on El Salvador’s Reproductive Health Survey In general, most indicators suggest that girls and (Encuesta Nacional de Salud Familiar, FESAL), 2002-2003 and 2008. women in Central America do not face disproportion- Chapter 2 Endowments 50 Figure 2.7: Adult literacy rate (%), 2000 and 2010 100 90 80 70 60 Literacy rate 50 40 30 20 10 0 2000 2010 2000 2010 2000 2010 2000 2010 2000 2010 2000 2010 PAN HND CRI GTM SLV NIC Women Men Source: World Bank’s World Development Indicators, 2017 Note: Adult literacy rate is the percentage of people ages 15 and above who can both read and write with understanding a short simple statement about their everyday life. ate barriers in accessing education. In fact, according in 2010, a gap double that of El Salvador. Although to many indicators, they seem to outperform boys. literacy rates increased between 2000 and 2010 in In many of these indicators, however, Guatemala is Guatemala, the gender gap remained intact. an outlier where girls and women trail behind their Girls are more likely to finish primary school than male counterparts in rates of completion of primary boys in all countries except Guatemala. Completion and secondary education. These results in Guatemala rates are near universal for both boys and girls in El may in part be driven by indigenous girls having lower Salvador, Costa Rica, and Panama. Yet in line with the rates of access to education than both non-indigenous rates of school repeaters and students who are over-age girls and all boys. and following the LAC trend, boys have higher drop- The countries of Central America have achieved out rates from primary school than girls in both Nica- near universal primary school enrollment for recent ragua and Honduras. In Nicaragua, 89 percent of girls cohorts of students. Adult literacy rates, which reflect completed primary school compared to 82 percent for the accumulation of education over time and hence boys in 2014, and in 2013 in Honduras, 97 percent of can reveal gender biases in earlier cohorts, suggest that girls reached the last grade of primary education com- women in two countries – Guatemala and El Salvador pared to 92 percent of boys. The gap in both countries – had less access to basic education than men (Fig- was larger than the average for the LAC region where ure 2.6). In the other countries, literacy rates are the 93 percent of girls reached the last grade of primary same for men and women. Guatemalan women’s lit- education compared to 89 percent of boys in 2014. eracy rate was 12 percentage points lower than men’s In Guatemala, however, completion rates are slightly Towards Equal? Women in Central America 51 Figure 2.8: Gender parity in primary education, 2015 a) Students who are over-age (% enrollment) b) Primary school repeaters (% enrollment) 35 12 30 Female Male Female Male 10 25 8 20 6 % % 15 10 4 5 2 0 0 El Salvador El Salvador Guatemala Guatemala Costa Rica Costa Rica Nicaragua Nicaragua Honduras Honduras Panama Panama Source: World Bank’s World Development Indicators 2017 Source: UNESCO Institute of Statistics Note: Data are from 2015 except Nicaragua (2010) Note: Data are from 2015 except Nicaragua (2010) and Panama (2014). and Panama (2014). lower for girls than for boys --- 88 percent compared will drop out or not be allowed to continue in school to 85 percent among girls. (Kabay, 2016). Low school quality remains a concern, but girls ap- Lower secondary education completion rates are pear to be faring better in primary school than boys. higher for girls than boys in all countries expect Gua- Gross enrollment rates for primary school are above temala (Figure 2.8). In this case, El Salvador is the 100 percent for both boys and girls in the six coun- country with the highest completion rates for both tries.47 Gross enrollment rates exceeding 100 percent sexes, above the LAC average, and the smallest gap be- are the result of children in primary school who are tween them. The largest gaps to the advantage of girls either under- or over-aged for the grades. With the are again in Nicaragua (13 percentage points) and exception of Panama, more than 10 percent of boys Honduras (11 percentage points). Guatemala is once and girls in primary school in each country is over- more the only country where boys´ rates are higher age (Figure 2.7a). This often reflects children entering than those of girls. This gap in Guatemala is driven by school late or grade repetition. In these measures, girls differences in completion rates in rural areas. While are outperforming boys in primary school; in each in urban areas, the secondary education completion country, both over-age and repetition rates are higher rate is 48 percent among girls compared to only 46 for boys than girls (Figure 2.7b). High repetition rates percent among boys, in rural areas only 24 percent suggest that students have unmet educational needs. of boys and 22 percent of girls complete secondary Repetition also increases the likelihood that a child education.48 48 Estimates based on data from Guatemala’s National Survey of Living Conditions (Encuesta Nacional sobre Condiciones de 47 World Bank’s World Development Indicators 2017 Vida, ENCOVI) 2014. Chapter 2 Endowments 52 Figure 2.9: Secondary education completion rates, 2014 100 Female Male 80 60 % 40 20 0 El Salvador LAC Panama Nicaragua Costa Rica Honduras Guatemala In Source: World Bank’s World Development Indicators 2017 Note: All data are from 2014 except or Nicaragua (2010), Honduras (2013), and LAC (2013). While a quarter (25 percent) of Central American caring for children (12 percent). Among Guatema- youths leave the education system by the age of 15, lan girls ages 13 to 24 who were not enrolled in ed- the reasons for dropping out of school differ between ucation, one-third reported that the main reason was boys and girls. In both lower and upper secondary their household responsibilities (World Bank, 2012b). school, dropout rates in Central America are similar In El Salvador, the need to do household work was between boys and girls (Adelman and Szekely, 2016). cited as a top reason by a high share of girls who had The largest difference is among upper secondary stu- dropped out of school, and over 80 percent of female dents in Nicaragua where boys are 5 percentage points ninis (neither in work nor studying) reported being more likely to drop out. responsible for household work (see Box 3.3). As girls reach secondary school age, the decision to stay at In terms of reasons for dropping out, a lack of in- home is often a result of pregnancy, the decision be- terest was a common reason given by both genders. ing either voluntary or imposed by the school. Other However, boys were more likely than girls to report factors, such as not having running water and proper dropping out for economic reasons while girls were sanitation in schools can also play a role in limiting more likely to cite personal reasons, including need- girls’ school attendance and enrolment (Box 2.2). ing to help with domestic chores at home (Adelman and Szekely, 2016). According to the 2014 Nicaragua Although Central America trails behind the rest of Household Survey, boys between the ages of 13 and LAC in access to secondary and tertiary education, 18 reported dropping out to conduct farm work (46 girls and women are more likely than boys and men percent), because of a lack of interest (25 percent), to continue in school after primary education in all or because of financial constraints (11 percent). Girls, countries except Guatemala. Only Costa Rica’s en- on the other hand, were more likely to mention the rollment rates are above the regional average (Figure importance of leading household work (26 percent), 2.10a). The gender gap in secondary school is most a lack of interest (18 percent), or pregnancy and/or pronounced in Honduras and Nicaragua (where it Towards Equal? Women in Central America 53 Box 2.2: In focus: Lack of sanitation and girls’ attendance in Panama A lack of water and sanitation services in schools can decrease school attendance of girls due to the “menstruation hypothesis.” This is a testable hypothesis: girls’ school attendance (and not that of boys) will be influenced after (and not before) menarche by whether or not schools have water and sanitation services. Using information from the Third Regional Comparative and Explanatory Study (TERCE) on the school attendance of boys and girls in the third grade and in the sixth grade, Almeida and Oosterbeek (2017) confirmed that girls in sixth grade were more likely to be absent from school than both younger girls and sixth grade boys. With schools with inadequate sanitation, the results are large: girls in sixth grade were 6 to 10 percentage points (relative to a base of 0.65) more likely to have missed at least one day of school during the previous six months than both boys in sixth grade and girls in third grade. Figure 2.10: Gender parity in secondary and tertiary school enrollment a) School enrollment, secondary (% gross) b) School enrollment, tertiary (% gross) 150 150 Female Male Female Male 100 100 50 50 0 0 Costa Rica LAC El Salvador Nicaragua Panama Honduras Guatemala Costa Rica LAC Panama El Salvador Honduras Guatemala Source: World Bank’s World Development Indicators 2017 Note: All data are from 2014 except for Nicaragua (2010) and Panama (2013). The gross enrollment rate in tertiary education includes all women and men enrolled at that level (ISCED 5 to 8) regardless of age, expressed as a percentage of the total population of the five-year age group following on from secondary school leaving. amounts to 10 and 9 percentage points respectively). the countries with the highest enrollment rates for It is the smallest in El Salvador where approximately both men and women also have the largest gaps to the 81 percent of both girls and boys are in secondary edu- advantage of women: 17 and 11 percentage points. El cation. The poorest performer in secondary school en- Salvador and Guatemala, the countries with the low- rollment, and the only country in which girls are less est enrollment ratios, are also the two with the lowest likely to be in school than boys, is Guatemala where gender gaps and with a gender parity index closer to 1. only 61.5 percent of girls and 65.5 of boys are en- Tests scores show that, while outperforming boys rolled in secondary schools. Panama and Costa Rica, in language and writing, girls in Central America fall Chapter 2 Endowments 54 behind in math and science. A large body of inter- for women in Guatemala where it accounts for half national work has identified expectations and social of all female tertiary graduates. Education is the top support as important factors in the gender gaps in area of study for women in three out of the five coun- achievements in math and science. Studies have found tries for which data are available for 2015, followed that one’s individual views, as well as those of their closely by business, administration, and law and by mothers and peers on gender equality explain some health and welfare. These three areas account for the of these differences (for example, Leaper et al, 2012). vast majority of female tertiary graduates in all Cen- According to national test data from Honduras, girls tral American countries – ranging from 80 percent in performed better than boys in both Spanish and Math Costa Rica, Guatemala, and Honduras to 64 percent at every grade level in primary school. Data from the in Panama. For men, these three areas account for be- Trends in International Mathematics and Science tween 66 percent (in Guatemala) to 47 percent (in Study (TIMSS) show small gaps in math and science El Salvador) of graduates. In line with global trends, favoring boys in grades 7 through 9, which widen women remain a minority of university students grad- in both subjects for ninth-grade level students. The uating with STEM degrees. Only between 3.3 (El results for Honduras in the Progress in Internation- Salvador) and 8.4 (Panama) percent of female tertiary al Reading Study (PIRLS), which measures reading graduates complete a degree in engineering, science, comprehension, showed girls significantly outper- or mathematics. As has been noted in many studies forming boys in Honduras. Costa Rican trends were of wage and job opportunities in STEM versus other similar, with girls outperforming boys in reading by 15 careers, these differences in subjects studied can have points in the 2015 Program for International Student serious implications for young people’s occupational Assessment (PISA) while boys outperformed girls in choice and future earnings. mathematics and science with a statistical difference The evidence from the region suggests that the of 16 and 18 points respectively. The 2013 Third Re- intersection between gender and indigenous sta- gional Comparative and Explanatory Study (TERCE) tus is associated with lower educational outcomes. results for Costa Rica, Nicaragua, and Guatemala also Throughout the LAC region, indigenous women tend showed boys outperforming girls in mathematics in to have higher levels of illiteracy and school dropout both 3rd and 6th grade, although both genders had rates, which undermine their capacity to take advan- similar results in reading (UNESCO, 2016). In con- tage of economic opportunities (World Bank, 2015a). trast, girls in Panama outperform boys in reading and While this association is only marginal in Nicaragua, have similar results to boys in mathematics. it is significant in Guatemala. As of 2014, the average Differences in test scores in secondary school may non-indigenous woman in Nicaragua had 6.5 years of be contributing to differences in the choices made by schooling, compared to 5.9 years among men, while young men and women about which subjects to study in indigenous communities, women had an average of at the university level. Women represent a dispropor- 6.3 years of schooling compared to 5.7 among men. tionate share of graduates in education and health On the other hand, in Guatemala, school enrollment, and welfare, while men are more likely to graduate retention, and transition rates are consistently lower in information and communication technologies and among indigenous girls than among non-indigenous engineering, manufacturing, and construction (Table girls, indigenous boys, and non-indigenous boys (San- 2.3). The field of education is particularly important chez et al, 2016). Indigenous women had an average Towards Equal? Women in Central America 55 Table 2.3: Area of study in tertiary education, by gender (2015) University area of study Gender CRI SLV GTM HND PAN Business, Administration, Men 37.5 25.2 25.0 29.4 30.1 and Law Women 34.0 29.1 18.3 28.7 27.3 Men 16.2 12.0 32.8 22.9 18.4 Education Women 27.2 19.5 49.4 39.0 28.2 Men 9.8 9.6 8.2 8.0 6.2 Health and Welfare Women 19.3 23.6 11.8 12.5 8.6 Engineering, Manufacturing, Men 12.7 18.5 11.5 18.5 14.9 and Construction Women 3.8 3.3 3.7 5.9 8.4 Information and Communication Men 10.1 18.3 3.0 6.6 7.3 Technologies Women 1.9 6.1 0.5 1.6 4.1 Natural Sciences, Mathematics, Men 1.0 0.9 2.4 0.9 4.7 and Statistics Women 0.7 0.8 1.2 1.1 7.2 Men 12.7 15.6 17.1 13.7 18.5 Other Women 13.1 17.5 15.1 11.3 16.2 Source: Authors’ tabulations using World Bank Education Statistics from UNESCO’s Institute for Statistics. Note: No data are available for Nicaragua. Data for Costa Rica, El Salvador, Guatemala, and Honduras are from 2015, and the latest available data for Panama are from 2011. of 2.8 years of schooling, one year less than indig- enous men (3.8 years), with larger gaps in younger cohorts (Hallman et al, 2006). In Honduras, indige- nous and Afro-Honduran women’s groups cite teacher absenteeism, too few schools, a culturally insensitive curriculum and schedules, and gender norms as sig- nificant barriers to education.49 49 Secretaría de Estado para el Desarollo de los Pueblos Indíge- nas y Afrohondureños (SEDINAFRO), UNCHR, UNWomen, UNDP, Cooperación Suiza. Agenda Política de las Mujeres Indígenas y Afro-hondureñas: Caminando juntas, del diálogo a la propuesta, 2012-2013. Chapter 2 Endowments 56 Towards Equal? Women in Central America 57 chapter 3 Economic opportunities W hen women are not included in economic activities, this leads to substantial losses, not only for women and their families through foregone 3.1 Economic participation income but also in aggregate growth (World Bank, While female labor force participation rates are low 2012a). This chapter examines women’s access to eco- throughout Central America, the size of the gender nomic opportunities through labor market inclusion gap in this key indicator varies significantly across the and entrepreneurship, including access to finance. six countries. For women of prime working age (18- While outcomes vary across the countries and, impor- 65), the labor force participation rate (the percentage tantly, across skill and urban groups, overall it is clear of women who are either employed or searching for that economic opportunities for women in Central America are limited. employment) ranges from 44.7 percent in Guatema- la to over 50 percent of women in each of the other The chapter begins by quantifying the rates of countries, up to 56.6 percent in Panama (Figure 3.1a). female economic participation in each country – in Guatemala’s low female participation rate is particu- particular, labor force participation, unemployment, larly striking as it has the region’s highest male labor and employment. The second section considers differ- ent measures of the quality of economic engagement, force participation at 92.0 percent, meaning that Gua- including hours worked, sectors of employment, and temala’s female participation rate is less than half that wages. It closes by considering entrepreneurship and of men’s (48.6 percent). The region’s best performer in access to capital and credit, which are key indicators this measure is Panama where women’s participation of economic inclusion beyond the labor market. The rate is almost two-thirds that of men’s (63.8 percent) third section concludes. (Figure 3.1b). 58 Figure 3.1: Female labor force participation in Central America a) Labor force participation rate b) Female labor force participation (ages 18-65), 2014 as a percent of male participation rate (2014) Female Male 100 61,2 61,9 63,8 57,2 Labor Force Participation 80 55,7 48,6 60 (%) 40 20 0 GTM HND SLV NIC CRI PAN GTM HND NIC SLV CRI PAN Source: Authors’ tabulations using the Socioeconomic Database for Latin America and the Caribbean or SEDLAC (World Bank and CEDLAS). Women living in rural areas are significantly less participation rates among women increased by 20 and likely to be in the labor force than those living in ur- 27 percent respectively, still significantly higher than ban areas (Figure 3.2a). Just one in three rural women for urban women (14 and 11 percent). On the other in Nicaragua and Guatemala was in the labor force hand, in Nicaragua and Guatemala, female labor force in 2014. Even in Panama, the region’s top performer, participation decreased by 9 and 4 percent respective- only half of all rural women were in the labor force. ly in rural areas. Participation rates for rural women ranged from 40 Could Central America’s limited urbanization percent to 44 percent in El Salvador, Honduras, and combined with the low participation rates of rural Costa Rica. Similarly, in each of the six countries, the women explain the region’s low female participation gap between female and male labor force participation rates compared to the rest of LAC? Figure 3.3 shows rate was significantly higher for rural population; fe- that only two of Central America’s countries have male participation rates ranged from one-third to just higher than average female labor force participation over a half of male participation rates (Figure 3.2b). in urban areas – Nicaragua and Panama.50 The other Rural female labor force participation has in- four countries fall below average in both urban and creased since 2004 in Costa Rica, Honduras, and rural female participation rates and all six countries Panama. Between 2004 and 2014, rural women in have below average female rural labor force partici- Honduras increased their participation rate by 41 per- cent compared to only 5 percent for women in urban 50 Average participation rates are calculated as the unweighted average of the 16 countries in LAC with sufficient data for this areas. In Costa Rica and Panama, rural labor force exercise. Towards Equal? Women in Central America 59 Figure 3.2: Rural female labor force participation in Central America a) Female labor force participation b) Female labor force participation rates (ages 18-65) by urban/rural as a share of male participation rate status, 2014 by urban/rural, 2014 Rural Urban Rural Urban 80 Rural Urban Rural Urban 80 71,9 72,0 69,5 65,6 66,8 71,9 61,0 72,0 69,5 60 65,6 66,8 61,0 53,0 force 60 (%) (%) 46,3 48,4 45,1 53,0 force participation 48,4 labor 40 36,3 36,7 45,1 46,3 participation labor 40 36,3 36,7 Female Female 20 20 0 0 NIC GTM SLV HND CRI PAN NIC GTM SLV HND CRI PAN NIC GTM SLV HND CRI PAN NIC GTM SLV HND CRI PAN c) Female labor force participation rate, urban and rural, 2004 and 2014 70 70 force 50 (%) (%) force 50 participation 30 labor participation 30 labor 10 Female 10 -10 Female -10 Rural Urban Rural Urban Rural Urban Rural Urban Rural Urban Rural Urban CRI Rural SLV Urban Rural GTM Urban Rural HND Urban Rural NIC Urban Rural PAN Urban Rural Urban CRI SLV GTM HND NIC PAN 2004 2014 2004 2014 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). The area of the bar labeled “2014” represents the change between 2004 and 2014. pation. Countries like Bolivia and Paraguay, which Although urban women have higher participation also have sizeable rural populations, have significantly rates than rural women, these are still significantly low- er than male participation rates. Urban female partici- higher levels of female participation in both rural and pation rates vary from 55 percent in Guatemala to 63 urban areas. This suggests that the relatively large rural percent in Nicaragua. Nicaragua and El Salvador have populations can explain only some of the differences the region’s highest gender parity in urban participa- in national female labor force participation rates be- tion rates. In these two countries, female participation tween Central America and the rest of Latin America. rates are 72 percent of male participation rates. Chapter 3 Economic opportunities 60 Figure 3.3: Female labor force participation and rural population, 2014 90 Urban above average and Both above average 80 rural below average Uruguay Peru 70 Colombia Brazil Nicaragua Panama Paraguay Bolivia 60 El Salvador Costa Rica Honduras Ecuador Urban Guatemala Chile 50 Mexico Dominican Republic 40 30 Rural above average and Both below average urban below average 20 20 30 40 50 60 70 80 90 Rural Source: Authors’ tabulations using SEDLAC (CEDLAS and the World Bank). Green horizonal and diagonal lines represent the unweighted average of urban and rural female labor force participation among the 16 LAC countries in the analysis, respectively. A U-shaped relationship between economic devel- between urban and rural married women (Chioda, opment and married women’s labor force participa- 2016). This suggests that the growing shares of female tion has been well documented in the literature (for labor force participation in urban areas observed in example, Boserup 1970; Durand 1975; Psacharopou- Central America not only reflect the greater opportu- los and Tzannatos 1989; Goldin 1995; and Mammen nities available to women in urban areas but probably and Paxson 2000). In other words, when a country also changing norms. is poor, women participate in the labor force out of necessity. As a country grows and develops, econom- Obstacles to economic ic activity moves towards the industrial sector, which participation increases wages, especially for men. At this stage, ac- There are different explanations for the low female la- cording to the theory, women no longer need to work bor force participation in the region. It may partially outside of the home for survival, hence women’s labor be explained by an undermeasurement of female eco- force participation drops. As countries develop even nomic activity (Box 3.1). A second likely factor is the further, education for women increases, fertility de- high labor-intensity of household tasks, especially in creases, and women find more opportunities to work areas with low access to basic services such as running and encounter less stigma about working, particularly water and modern energy sources. This lack of access in the service sector (Goldin, 1995 and Verick, 2014). to these services imply more time is needed to accom- This theory may explain differences in women’s labor plish necessary household tasks typically allocated to force participation within a country, for example, women: for example, having to fetch water and collect Towards Equal? Women in Central America 61 firewood from outside the home, and engage in slower making tortillas as a key daily activity (World Bank cooking methods due to the use of wood fires. A re- 2018). Rural women spend more time than urban cent survey in rural Honduras found that many wom- women and men in household work, mainly because en reported spending hours weekly grinding corn and of a lack of access to basic services (UNDP, 2011). Box 3.1: In focus: Measuring female labor force participation Only about one in five Central American women in rural areas report working in agriculture. How- ever, as rural households grow items for their own consumption in addition to market crops, it may be that household labor allocated to own consumption is underreported in surveys. 51 To the extent that production activities that are directed for household consumption are more likely to be carried out by women, the undercount of these activities can lead to a misleadingly low measurement of female participation. Another element at play is the non-reporting of fam- ily workers in household enterprises, particularly agricultural activities in which many family members contribute. Earlier research suggests that this is indeed happening; Dewalt et al. (1985) and Colverson (1995) found that low labor force participation rates among Honduran women are due, in part, to the underreporting of women’s involvement in agricultural produc- tion and food processing and distribution. This suggests that household surveys are under- measuring female economic production, particularly in rural areas, and that the contribution to women to household income is being underestimated within households. A 2018 qualitative survey (World Bank 2018) in Honduras aimed to better understand the low labor force participation rates among rural women. 52 The interview began with the same questions as those used by the official household survey to measure labor force participation: “during the past week, did you spend an hour or more on any work or activity with payment in cash or in kind, or did you earn any income?”. Most women said they did not work, though fol- low-up questions identified that they often participated in both market and non-market activ- ities in the previous week. When asked about crops and livestock for own consumption, most reported being responsible for these though they did not consider it work. “ We do have hens and orchards. There are pigs but I take care of them for my grandfather. I don’t get paid for that. In the orchard I have plantain, it isn’t much, they haven’t grown. I don’t sell the plantain because it isn’t much. If I had enough, I would even sell hens. ” Woman Participant, Palo Marcado, Honduras While women were often responsible for crops and livestock grown near the home, the men of the households were often responsible for market crops or staples (such as corn) grown further from the home. Even so, women often contributed to the production of the crops grown further from home. 51 New International Labour Organization standards for employment measurement will begin excluding production for auto- consumption from the definition of labor market employment. The surveys included in this analysis predate this change in methodology, including subsistence farming (carried out by men and women) as employment. 52 The study covered four rural communities across the country and interviewed 36 women, 21 men and conducted 6 focal groups with women. Chapter 3 Economic opportunities 62 “I plant squash and beans in a small plot of land in the house. He sows corn, I help him bring water, move corn, shuck and thresh it.” Woman Participant, Las Crucitas, Honduras “I help him. Everything he does, I do it as well. I help him cut the vegetables, peel.” Woman Participant, Palo Marcado, Honduras Source: World Bank (2018) Low access to early childhood education and child- pleted tertiary school are about 29 percent more likely care limit the amount of time women can dedicate to be in the labor force than those with incomplete to market activities outside the home. Indeed, a large primary school. In most countries, the relationship number of women between the ages of 15 and 24 between additional schooling and labor force partic- cite the lack of childcare options as their main reason ipation did not change substantially between 2004 for not being in the labor force in Nicaragua (World and 2014. However, it weakened in Panama and for Bank, 2012a). In Costa Rica, women are still consid- some levels of schooling in Honduras and Nicaragua. ered to be the primary caregiver, and most women On the other hand, the relative likelihood of Guate- face the double burden of employment and domestic malan women with university education participating work (see section 3.1.4). Finally, social norms regard- increased significantly between 2004 and 2014, from ing women’s and men’s roles in society and their di- 21 percent to 34 percent more likely that women with vision of (paid and unpaid) work are drivers behind incomplete primary education. their participation in labor markets. For instance, in four out of the six Central American countries, most Unemployment people believe that women should work only if the When they join the labor force, women are more like- husband does not earn enough money, as discussed ly to be unemployed than men (Figure 3.5). The only in Chapter 4. exception is El Salvador, where female unemployment A key predictor of women’s labor force participa- is lower than male unemployment (3.8 percent versus tion is education. Throughout Central America, wom- 6.4 percent), and Guatemala, where the unemploy- en with higher levels of schooling are significantly more ment rate is 2.7 percent for both men and women. likely to be in the labor force than those with lower At 10.2 percent, Costa Rica has the highest female levels. In Costa Rica, for example, women who com- unemployment rate with a rate higher than that of pleted secondary school and those who completed ter- men (7.3 percent). tiary education were 22 percent and 44 percent more Young women have especially high rates of unem- likely respectively to be in the labor force than women ployment, which suggests that they have particularly who did not complete primary schooling. Panama, the weak job prospects. In Costa Rica, 22.3 percent of country with the highest female participation rate in young women (aged 18-25) in the labor force are un- Central America, has the smallest effect of schooling employed. Young Honduran women face more dis- on participation rates. Even there, women who com- proportionate unemployment rates, with female un- Towards Equal? Women in Central America 63 Figure 3.4: Probability of women being in the labor force by highest level of schooling completed compared to women with less than a primary education, 2004 and 2014 Costa Rica El Salvador 0,5 0,5 Predicted change in participation Predicted change in participation 2004 2014 2004 2014 0,4 0,4 0,3 0,3 0,2 0,2 0,1 0,1 0,0 0,0 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Guatemala Honduras 0,5 0,5 Predicted change in participation Predicted change in participation 2004 2014 2004 2014 0,4 0,4 0,3 0,3 0,2 0,2 0,1 0,1 0,0 0,0 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Nicaragua Panama 0,5 0,5 Predicted change in participation 2004 2014 2004 2014 Predicted change in 0,4 0,4 participation 0,3 0,3 0,2 0,1 0,2 0,0 0,1 -0,1 -0,2 0,0 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Note: These figures report the conditional marginal probability of being in the labor force for women in relation to a similar woman who did not complete primary education. The estimates are conditional on the following characteristics: age and its square, region, urban, current school enrollment, and the number of children in the household under the age of 5 and between the ages of 5 and 14. The sample is limited to women age 18 to 65. See Annex 3.1 for logit results. Chapter 3 Economic opportunities 64 Figure 3.5: Urban unemployment rates, by gender and age, 2014 a) All adults (ages 18-65) b) Young adults (ages 18-25) Women Men 22,3 Women Men 10,2 8,1 18,0 7,3 7,7 7,8 16,0 16,5 15,9 15,4 6,9 6,4 14,4 14,4 5,5 6,1 12,511,6 5,3 11,4 4,2 3,8 9,6 2,7 2,7 7,1 5,7 LAC LAC Panama Panama Honduras Honduras Nicaragua Nicaragua Costa Rica Costa Rica Guatemala Guatemala El Salvador El Salvador Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Note: Unemployment is defined as actively searching for employment and not currently working. This definition varies between countries. LAC unemployment rates are calculated as the unweighted average for 17 countries. Box 3.2: In focus: Female Ninis in Central America One out of every three Central American women aged 15-24 are out of school and not working ( ninis ). Between 2004 and 2014, the share of female ninis increased from 34 to 37 percent, well above the LAC average of 28 percent. The proportion of female ninis varies greatly across the Central American countries, ranging from 23 percent of young women in Costa Rica to 43 percent in Honduras and 47 percent in Guatemala. Moreover, the incidence of young ninis is higher among poorer households, exacerbating existing inequalities and hindering social mo- bility and poverty reduction in the long run. Source: De Hoyos et al (2016) employment at 16.5 percent compared to 9.6 percent lowest (28 percent).53 The long-term negative labor for young men. High unemployment rates, along with market outcomes among youths contributes to the a probable discouraged worker effect, are related to persistence of gender inequalities and low incomes, the nini phenomenon in Central America (Box 3.2). which hinder social mobility and poverty reduction The share of youths who are not in education or em- (Ferreira et al, 2013). ployment (ninis) has historically been higher in Cen- Underemployment is likely to be prevalent across tral America among women than men (44 percent of countries in the sub-region, as hidden or invisible un- women aged 19–24 years, compared with 10 percent deremployment appears to have increased. Underem- of men in 2014). Guatemala has the highest share of female ninis at 54 percent, whereas Costa Rica has the 53 See section 2.2 for further details on ninis in Central America. Towards Equal? Women in Central America 65 Figure 3.6: Employment rates (18-65), Figure 3.7: Part-time employment rates 2014 among women (18-65), 2014 Female Male Female Male 90,8 86,1 86,7 87,2 80,6 80,9 56,4 56,5 46,4 46,6 38,5 42,2 55,3 37,1 38,3 47,9 48,3 51,2 51,5 32,7 44,7 26,6 27,4 29,9 Guatemala Honduras Costa Rica Nicaragua El Salvador Panama Panama El Salvador Costa Rica Nicaragua Guatemala Honduras Source: Authors’ tabulations using the World Bank’s World Development Indicators ployment is defined as the share of employed individ- only half of Central American women of working age uals who would like to work more hours than they being employed. Fewer than half of all adult women are presently working. However, data in this regard in three countries – Guatemala, Honduras, and Costa are only available for Panama and Honduras, and it Rica – are employed (Figure 3.6). In the other three is not clear whether this phenomenon has any gen- countries, only a slight majority are working, from der implications. In Panama, rural male workers have 51.2 percent in Nicaragua to 55.3 percent in Pana- high levels of invisible underemployment, with 20.2 ma. For comparison, employment rates for men of the percent working less than full time at less than min- same age range from a low of almost 81 percent in imum wage in 2013.54 The rates are even higher in Costa Rica and El Salvador to a high of 90.8 percent indigenous areas. Both visible and invisible underem- in Guatemala. ployment has increased in Honduras since the 2009 More than one in four employed women in crisis.55 each country in Central America works part-time, including most employed rural women in Panama 3.2 Employment and Guatemala (Figure 3.7).56 In other words, the already low employment rates hide even deeper dis- outcomes parities in female economic inclusion. Overall, only The combination of low female labor force partici- 27 percent of women in Honduras and Guatemala, pation rates and high unemployment rates results in about 34 percent of women in Costa Rica and Nic- aragua, 37 percent of women in El Salvador and 41 54 Statistics published by the National Institute of Statistics and percent of women in Panama work full time. With Census (Instituto Nacional de Estadísticas y Censo, INEC). the exception of Panama, full-time employment rates 2013. Panama. 55 Estimates based on Honduras’ National Multi-purpose House- hold Survey (Encuesta Permanente de Hogares de Propósitos 56 Part-time employment is defined as working fewer than 35 Múltiples, EPHPM) 2009 and 2013. hours at the primary job. Chapter 3 Economic opportunities 66 for women are less than 90 percent those of men in women worked fewer than 35 hours, including 56.5 all countries. The lowest part-time employment rates percent of employed women in rural areas. Wom- for women are in Panama at 26.6 percent. This dis- en tend to be restricted to occupations with flexible guises a significant disparity between urban and rural or few working hours, which inevitably diminishes women; only 16.8 percent of employed urban wom- their income and limits their work experience and en work part time compared to 56.4 percent in rural human capital accumulation. This is likely to in- areas. The highest part-time employment rates are crease the concentration of women in low-paid jobs in Honduras; 42.2 percent of Honduran employed (Tenjo et al, 2004) (Box 3.3). Box 3.3: In focus: Increasing economic opportunities for women through public works and training: The Temporary Income Support Program (PATI) in El Salvador El Salvador’s Temporary Income Support Program (PATI) was initiated by the authorities in 2010 as a response to the country’s 2009 crisis, which had raised unemployment and poverty, par- ticularly in urban areas that lacked any other form of safety net. PATI was aimed at cities with the highest poverty rates and targeted to groups with the highest unemployment rates, mean- ing young people and female-headed households. As a result, 70 percent of the participants were women. PATI went beyond its name, addressing not only traditional short-term vulnerabil- ity with income support but also long-term employability. Beneficiaries received US$100 per month for a period of six months, conditional on them participating in training and community projects and registering at the employment office. Three lessons learned during the implementation of this project revealed the barriers to employment faced by women in El Salvador: • Occupational segregation : The first cycle of projects focused on traditional urban infra- structure, such as construction work, road rehabilitation, and waste management. These types of projects did not elicit enough interest among unemployed women. The program subsequently responded by including an alternative menu of projects such as urban agri- culture, the rehabilitation of historic buildings, and community events that attracted more women to participate. • Childcare : A lack of childcare services was a critical constraint for women. Once the imple- menting agency (the Social Investment Fund for Local Development, FISDL) realized this, daycare assistance was added as part of the projects funded by PATI. This enabled women not only to participate in the program but also to work in daycare activities as part of the PATI program. • Household work burden : Many women were not looking for fulltime employment but were in- stead interested in flexible home-based work. In response, the program added such courses in, for example, cooking, cosmetology, and tailoring. Towards Equal? Women in Central America 67 Quantitative and qualitative evaluations of the project found that, for women in particu- lar, PATI increased their economic activity and engagement in the labor market. Many women participants said that they felt more prepared to find a job or start a business. Women also increased their participation in community organizations and took on more leadership roles in their communities. However, men were better able to translate their participation in PATI into higher earnings (higher than the income support that they received under PATI), whereas the evaluations found no impact on women’s earnings after their graduation from PATI. Source: FUSADES (2014 and 2015). more precarious than wage work and involves access Job quality to fewer benefits. Young and small firms are typical- In all dimensions of job quality, female employment ly more volatile, gaining and shedding jobs at higher lags behind that of male employment in Central Amer- rates than larger and older firms. Furthermore, the ica. Jobs can be classified into four types: (i) workers quality of jobs is often related to the size of the firm. who are employed by another entity (wage workers); Larger firms are also significantly more likely to be in (ii) the self-employed; (iii) those who employ others the formal sector. Jobs in the formal sector are better (employer); and (iv) workers who are unremunerat- quality as they typically provide access to safety nets, ed. This last category, unpaid workers, often work in such as unemployment insurance and pensions, as family-run enterprises or farms. Self-employment is well as more worker protections. Table 3.1: Type of employment by gender, 2004 and 2014 Female Male Wage Self- Employer Unpaid Wage Self- Employer Unpaid earners employed earners employed Costa Rica 2004 70.5 21.6 3.9 4.0 69.8 19.5 9.4 1.3 2014 78.9 16.9 2.6 1.6 75.2 19.2 4.9 0.7 El Salvador 2004 52.7 38.8 2.5 6.0 69.1 20.8 4.8 5.3 2014 53.9 35.5 3.0 7.6 66.9 22.3 4.6 6.2 Guatemala 2004 40.0 40.8 2.7 16.5 54.2 32.9 6.4 6.4 2014 50.9 34.2 1.8 13.1 67.6 22.5 3.3 6.6 Honduras 2004 48.9 38.7 5.7 6.6 52.7 24.8 15.1 7.4 2014 44.3 40.2 5.7 9.8 54.1 25.5 15.5 4.9 Nicaragua 2004 47.5 37.3 2.7 12.5 54.2 29.7 5.7 10.3 2014 53.8 43.1 2.7 0.4 64.0 27.3 8.0 0.7 Panama 2004 65.5 26.4 1.5 6.7 59.1 33.6 4.0 3.3 2014 66.0 23.6 1.4 9.0 65.5 28.4 3.3 2.8 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Chapter 3 Economic opportunities 68 In El Salvador, Guatemala, Honduras, and Nicara- and Panama respectively in 2014. In that same year, gua, women are less likely than men to be employed this type of employment accounted for 75.2 percent as wage workers and more likely to be self-employed and 65.5 percent of male employment respectively. or unpaid workers (Table 3.1). For both men and Instead, in these two countries, men were more likely women, there has been an increase in the likelihood to be self-employed than women. of being a wage worker and a decrease in the likeli- Unpaid work is particularly common among ru- hood of being self-employed between 2004 and 2014. ral women in Panama, Nicaragua and Guatemala. A little over half of the economically active women Interestingly, throughout most of Central America, in Guatemala, El Salvador, and Nicaragua were wage unpaid work is almost as common for young wom- workers in 2014. This type of employment account- en as it is for young men. The low gender differential ed for about two-thirds of male employment. On the among young workers reflects the importance of fam- other hand, self-employment accounted for one-third ily-owned enterprises for the employment of young of women’s jobs in Guatemala and El Salvador (34.2 people of both genders in Central America. Gender percent and 35.5 percent respectively) and less than differences appear later in life when men are more a quarter of men’s jobs (22.5 percent and 22.3 per- likely to be employed while women are more likely cent). In Honduras and Nicaragua, self-employment to remain in unpaid occupations. In Honduras and accounted for over 40 percent of female employment Panama, even young men are less likely to be unpaid and just about a quarter of male employment. workers, though this gap is significantly lower than In Costa Rica and Panama, on the other hand, among all adults (Figure 3.8 and Figure 3.9). women were more likely than men to be employed as Even among wage workers, in most countries, wage workers. Wage work accounted for 78.9 percent women tend to be employed in microenterprises. and 66.0 percent of female employment in Costa Rica While microenterprises (firms with five employees or Figure 3.8: Unpaid work among rural Figure 3.9: Unpaid work among young workers, by gender, 2014 people (18-25), 2014 25 25 Women Men Women Men 20 20 15 15 % % 10 10 5 5 0 0 Costa Rica El Salvador Guatemala Honduras Nicaragua Panama Costa Rica El Salvador Guatemala Honduras Nicaragua Panama Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS) Towards Equal? Women in Central America 69 Figure 3.10: Share of wage workers in microenterprises, by gender 2014 Women Men 80 60 40 % 20 0 Costa Rica El Salvador Guatemala Honduras Panama Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Information on firm size is unavailable for Nicaragua. less) generate necessary jobs, these workers often lack persistent levels of occupational segregation explained social protection, have irregular and casual contracts, a portion of earnings differentials between men and and are paid lower wages than in larger firms. In Hon- women in Costa Rica. Some key sectors for men, such duras and Panama, an equal share of wage-employed as agriculture, employ few women. Agriculture, which men and women work in microenterprises. However, is particularly important in countries other than Cos- in the rest of the countries, women are more likely ta Rica and Panama, represents between 25.5 percent to work in microentreprises. In Costa Rica, more fe- of all male employment (El Salvador) to 40.6 percent male wage workers than male wage workers work in (Guatemala). Possibly due to measurement issues microenterprises (47 percent versus 39.5 percent). In referenced in Box 3.1, such as cultural definitions of El Salvador, microenterprises account for two-thirds “work” that may not include women’s contributions of all female wage employment and just over half of to agriculture, it accounts for a maximum of 12.1 per- male wage employment (52.4 percent) (Figure 3.10). cent of female employment (in Guatemala) and only No information is available for Nicaragua. 3.7 percent in El Salvador. Instead, the overwhelm- Sector and occupational segregation by gender is ing majority of working women are employed in the high in Central America. Research has found that oc- services sector, with commerce playing a particularly cupational segregation can limit growth: Cuberes and important role. In every country in Central America, Teignier-Baqué (2011) estimate that eliminating bar- commerce is the largest sector of female employment, riers to women working in certain sectors or occupa- ranging from 19.4 percent in Costa Rica to 29.0 per- tions could increase output by as much as 25 percent cent in El Salvador. While commerce is also an im- in some countries. Occupational segregation can also portant sector of employment for men, in every coun- lead to worse wage and employment outcomes for try it accounts for less male employment than female women. For example, Deutsch et al (2005) found that employment. Chapter 3 Economic opportunities 70 Box 3.4: In focus: Getting women into non-traditional jobs in the roads sector: The Nicaragua rural roads project A community-based development approach for the construction of cobblestone roads, known as the Community Modules for Cobblestones ( Módulos Comunitarios de Adoquinado or MCA), has made important strides in including women in what is traditionally seen as “men’s work.” The Nicaraguan government with the support of the World Bank-financed Third Road Rehabili- tation and Maintenance Project first introduced this approach in 2004 by piloting the use of 32 MCAs to pave 28 kilometers with cobblestones (or adoquines ), which are concrete-like paving blocks made of cement, fine aggregates, coarse aggregates, filler, and water. The MCAs were le- gally constituted and accredited by the municipal authorities, who were co-responsible for the execution of the work. Each MCA was selected local participants living in the vicinity of these roads to pave a section of up to 3 kilometers of a rural road. To do this, each had to hire its own local labor for construction work, including women. This labor-intensive method has generated important social benefits, including short-term employment opportunities with over 280 MCAs having been formed and over 9,500 jobs created to date, technology transfer and technical skills building, and community empowerment and ownership of the roads within communities. Since 2007, the proportion of women employees in MCAs has increased from 10 percent to 46 percent. Women have been involved mainly in leadership, technical, and administrative positions, which has increased their income, built their skills, and given them more agency. This work experience has enabled female participants to make effective choices and transform these choices into desired outcomes. Apart from employment, Casabonne et al (2015) found that these experiences have had profound psychological and human development effects in addressing gender gaps, such as increasing women’s self-esteem and autonomy, changing intra-household bargaining and social norms, exposing women to wider social networks, and giving them opportunities to develop leadership skills. While the results have been positive, the study found that women still face some barri- ers and challenges to participating in the MCAs, such as feeling unprepared for these jobs, childcare constraints, and a lack of information. To address some of these, the next iteration of the project (the Rural and Urban Access Improvement Project) is supporting the following activities: (i) technical and gender-sensitization training for MCA staff to diversify tasks and, to encourage women’s agency, giving them the choice of which tasks to do; (ii) a gender-sen- sitive recruitment strategy to provide women with more information about the MCA in various ways such as community radio and the distribution of fliers; and (iii) the provision of childcare to women employed by the MCAs in coordination between Ministry of Transportation, the Min- istry of Family, and the participating municipalities. Women from the former MCAs will also be encouraged to participate in the cooperatives tasked with the routine maintenance of these roads, which will provide them with additional employment opportunities in the longer term. Source: Casabonne et al (2015) Towards Equal? Women in Central America 71 Figure 3.11: Share of wage workers in public sector, by gender 2014 Women Men 25 20 15 % 10 5 Costa Rica El Salvador Guatemala Honduras Panama Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Information on firm type is unavailable for Nicaragua. The public sector is an important source of jobs for services. This reflects the high concentration of wom- women. Public sector employment accounts for more en studying these areas in university as discussed in wage employment among women than men, especial- Chapter 2. These sectors include high rates of public ly in Costa Rica, Honduras, and Panama. Public sec- sector employment. The sectors that, on average, are tor employment usually provides secure earnings with equally important for both female and male employ- employment benefits and protections (Figure 3.11). ment are public administration and defense and man- Nonetheless, Panizza and Qiang (2005) found that ufacturing. women are paid less than men in both the private and Manufacturing is increasingly becoming a sector public sectors, although the private sector gap appears of economic opportunity for women. Potentially due to be larger on average. For Costa Rica and El Salva- to the expansion of maquilas and low-skilled man- dor, the authors found a premium of more than 10 ufacturing, especially in the northern countries of percent associated with working in the public sector, Central America, manufacturing accounts for more although this does not compensate for the wide over- female employment than male employment. The ma- all gender earnings gap. jority of workers in Central American maquilas are Other sectors are also highly gendered – in par- women. In Honduras, manufacturing accounts for ticular, construction work, which remains almost 22.3 percent of female employment and 12.9 percent exclusively male, and domestic work, which remains of male employment. Jobs in maquilas are associated almost exclusively female (Table 3.2). Domestic work with smaller wage gaps. De Hoyos et al (2012) found accounts for 15 percent of all jobs held by women in that the Honduran maquila sector had a gender wage Costa Rica and between 8 to 12 percent in the rest gap that was 16 percentage points smaller than other of the Central American countries. Among other ser- industries. At the same time, it is important to note vices, women are more likely to be involved in teach- that some maquilas have been cited for exploitative ing, social and health services, and other community and dangerous labor conditions. Chapter 3 Economic opportunities 72 Table 3.2: Gender segregation by sector of employment, 2014 CRI SLV GTM HND NIC PAN Construction 0.08 0.03 0.02 0.09 0.04 0.11 Transportation and communication 0.27 0.20 0.11 0.12 0.22 0.28 Agriculture 0.22 0.14 0.30 0.22 0.14 0.39 Electricity, gas, and water 0.30 0.25 0.17 0.29 0.37 0.56 Real estate and business activities 0.81 0.47 0.68 0.70 0.61 0.99 Public administration and defense 1.12 0.52 0.58 1.41 0.93 1.29 Manufacturing industries 0.71 1.31 1.54 1.73 1.11 0.95 Commerce 1.11 1.60 1.68 1.50 1.92 1.39 Financial intermediation 1.18 2.03 1.66 1.62 1.54 2.37 Other community services 1.23 1.65 1.72 2.40 3.76 1.45 Teaching 3.27 2.24 3.71 3.60 3.78 3.42 Social and health services 2.84 2.92 3.31 2.84 3.66 4.69 Hotels and restaurants 1.73 4.83 5.23 4.69 4.10 1.97 Domestic service 14.45 19.70 15.11 7.27 8.07 10.37 Source: Authors’ tabulations using SEDLAC (CEDLAS and the World Bank). Note: This table reports the differences in the distribution of employment by gender across sectors of employment. It is calculated as the ratio between female employment in sector x as a share of all female employment and male employment in sector x as a share of all male employment. A ratio of less than 1 represents sectors that are more important for male employment. A ratio greater than 1 represents sectors that are more important for female employment. and Guatemala where it fell by between 0.07 and Labor earnings 0.08 percentage points. Men have higher labor earnings than women in all The monthly earnings differential between men six Central American countries. Because of the diffi- and women can be partially explained by the fact culty of measuring labor earnings in agriculture, this that men often work more hours and have higher analysis of earnings is limited to urban workers. As rates of full time employment. Average hourly wages of 2014, women’s average monthly earnings ranged for women in El Salvador and Honduras were about from about 70 percent of those of men in Guatema- the same as those of men. In the other countries, the la and Nicaragua to 89 percent in El Salvador (Fig- average wage of women was between 87 percent and ure 3.12). In all countries except for Nicaragua, this 94 percent that of men. With the exception of Cos- gender gap in monthly earnings was an improvement ta Rica and Panama, the gender gap in hourly wages over 2004 (Figure 3.13). The reduction in this gap improved in all Central American countries between was particularly notable in Honduras, El Salvador, 2014 and 2004. In Panama, women earned an average Towards Equal? Women in Central America 73 Figure 3.12: Ratio of women’s earnings to Figure 3.13: Change in gender earnings men’s earnings, monthly and hourly in 2014 gap, 2004 to 2014 Monthly Hourly Monthly Hourly 0,08 1,01 1,00 0,02 0,02 0,89 0,94 0,88 0,87 0,84 0,84 0,89 0,72 0,79 0,70 0,00 -0,03 -0,04 -0,08 -0,07 -0,07 -0,08 -0,08 -0,12 Guatemala Nicaragua Costa Rica Panama Honduras El Salvador Guatemala Honduras El Salvador Nicaragua Costa Rica Panama Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS) Note: The analysis is limited to employed adults ages 18-65 living in urban areas and excludes unpaid workers. of 97 percent of men’s hourly wages in 2004, but this adjustments for individual characteristics that also in- had fallen to 88 percent in 2014. The gains were par- fluence earnings, such as age, education, employment ticularly notable in Honduras where the gap in hourly type, and, to account for intensity of home produc- wages fell by 12 percentage points. tion, the number of children in the household. When these adjustments were made, men in urban areas had However, because of the differences in selection wages that were between 5 percent (in El Salvador) into employment between men and women, the av- and 25 percent (in Nicaragua) higher than those of erage wages for men and women represent a range of women. Adjusting for selection into employment re- different types of work. As seen above, educational duces the wage gap significantly.57 attainment is a more significant factor in explaining labor force participation for women than for men. In addition, the selection into employment is very dif- Entrepreneurship and ferent for men and for women so the two groups of access to finance workers may differ in more than just the characteris- Entrepreneurship often gives women the flexibility to tics that are observable. These differences need to be tend to household chores and childcare duties while taken into consideration when studying earnings gaps. offering an opportunity for professional growth and In fact, our conditional estimates found no urban advancement not always found in the formal sector. gender wage gap in Honduras and El Salvador once In Honduras, half of all owners of microenterprises other differences were taken into consideration (Fig- (defined as those with fewer than 5 employees) are ure 3.14). However, in the remaining four countries, men’s wages were between 11 and 14 percent high- 57 A Heckman selection model was used. Full regression results er than women’s. Specifically, we made econometric are included in Annex 3.2. Chapter 3 Economic opportunities 74 Figure 3.14: Gender hourly wage gaps, 2014 Conditional Selection 25% 18% 19% 19% 17% 13% 14% 12% 11% 5% 5% 2% El Salvador Guatemala Panama Costa Rica Honduras Nicaragua Source: Authors’ tabulations using the World Bank’s World Development Indicators Note: This figure reports the coefficients on being a man using two different econometric specifications of the natural log of wages in urban areas. The first, labeled Conditional, is an OLS model that controls for observable characteristics. The second, labeled Selection, is a Heckman selection model that controls for observable characteristics as well as the differential selection into the labor force of men and women. Each bar represents a coefficient from a separate regression. The gender wage gap is not statistically different from 0 at the 95 percent confidence interval for Honduras and El Salvador. Regression results are reported in Annex 3.2. women compared to 24 percent of small firm owners firms. The available data suggest that barriers to start- (defined as having 5 to 10 employees), and 23 percent ing a business are gender-specific. For instance, men of medium-sized firm owners (defined as having more and women in El Salvador rely on different sources than 10 employees) (World Bank, 2010). Therefore, of funding, with men being more likely to get loans it is clear that women are more likely than men to run from banks, while women are more likely to use re- smaller businesses. mittances to fund their businesses (CONAMYPE and ISDEMU, 2013). This is also the case in El Salvador, where 64 per- cent of the owners of micro businesses are women. For With regard to women as owners and managers of small and medium-sized firms, female ownership falls firms, Nicaragua fares better than other countries in to 24 percent, and among large firms, female own- the region – and even compared to other lower-mid- ership is a low 10 percent. Many female owners of dle-income countries – in terms of women having microenterprises are of necessity entrepreneurs. Pines, access to leadership positions in business.58 Almost Lerner, and Schwartz (2010) noted that barriers to en- one-third (32.3 percent) of firms in Nicaragua have tering the labor market push women into becoming a female top manager, while in the LAC region as a entrepreneurs. This suggests that women’s entrepre- whole and in other lower-middle-income countries, only a fifth (20.8 percent) of companies do. However, neurship is usually a matter of necessity (push) rather than opportunity (pull). Female-owned micro busi- 58 The data in this paragraph come from the World Bank’s Enter- nesses employ fewer people than male-owned firms, prise Survey Database. The Enterprise Surveys collect informa- they focus primarily on neighborhood customers, and tion from a representative sample of the non-agricultural formal private economy. The latest year for which data are available is they have less than half of the sales of male-owned 2010. Towards Equal? Women in Central America 75 there is a clear concentration of female business lead- is only 42.7 percent while the average for lower-mid- ership among small enterprises. While 36.8 percent dle-income countries is only 37.6 percent. of small firms have a female top manager, the share Women’s limited access to capital remains a signif- of female managers drops to 28.3 percent for medi- icant obstacle, but progress has been made. Although um-sized firms and to 8.5 percent for large firms.59 information on entrepreneurship and women in the A similar picture emerges with regard to female par- region is limited, the evidence suggests that a combina- ticipation in the ownership of firms. More than half tion of advancements in women’s rights in the region, (61.9 percent) of Nicaraguan firms has at least one the limited labor options for women in the formal female owner -- 66.5 percent of small firms, 58.7 per- sector, and the need to supplement household income cent of medium-sized firms, and 32.6 percent of large has created an enabling environment for women to companies. The average for the LAC region as a whole become entrepreneurs (Box 3.5). The “feminization of agriculture” in Guatemala is an example of what 59 Small firms are defined as those with an average number of em- positive developments can happen when women have ployees between 1 and 19, medium-sized firms have between access to land and capital (in the form of remittances) 20 and 99 employees, and firms with more than 99 employees are considered large (Business Enterprise Survey Data 2010). (Box 3.6). Box 3.5: In focus: Economic empowerment of indigenous women in Panama: A snapshot of challenges & opportunities “Work we have, but training programs to help us grow we don’t have.” A Ngäbé woman in the Co- marca Ngäbé-Buglé, a member of a women’s handicraft and cultural association, the Cultural and Artisan Organization of Besiko. The economic empowerment of indigenous women in Panama is crucial to combatting the deep poverty in Panama’s indigenous people’s territories where poverty is above 70 percent and extreme poverty above 40 percent (World Bank, 2016b). Indigenous men earn on average 57 percent less than non-indigenous men, while indigenous women earn about 70 percent less than non-indigenous women (World Bank, 2015a). Indigenous women’s primary sources of income in Panama are the conditional cash transfer program, handicrafts, tourism, and some small-scale agriculture. A recent World Bank assessment done in partnership with two indigenous women’s organi- zations in Panama (World Bank, 2016b) has identified key barriers faced by indigenous women for improving their income-generating activities. With the Network of Indigenous Women and Biodiversity (RMIB) and the National Coordinator of Indigenous Women in Panama (CONAMUIP), the study found that the barriers included: (i) a lack of self-esteem among indigenous people; (ii) machismo combined with discrimination against indigenous peoples; (iii) a lack of knowl- edge among indigenous people in the areas of resource management, quality control, business plan development, leadership, and marketing; and (iv) a lack of access to capital, markets, and contacts among indigenous people. Another finding was the importance of involving other key actors who are part of the “ecosystem” critical for supporting women in their productive activities. Chapter 3 Economic opportunities 76 Existing programs for microentrepreneurs in Panama have largely bypassed indigenous women. As of 2017, no programs in Panama have specifically supported or targeted indige- nous women’s economic development. Existing programs have not been designed to take into account indigenous women’s concerns and ignore issues such as their cultural beliefs about living with the land and sharing with community members and their lack of access to modern technology. While a few programs do exist that target women entrepreneurs, one of which is expanding its outreach to indigenous women, the majority of the classes and activities that it offers take place in Panama City or online. Few indigenous women have either electricity or ac- cess to the internet or cell phones. With regard to access to finance, one NGO is supporting the establishment of community banks but only in those communities that are accessible by road. None of these programs have directly addressed the additional barriers faced by indigenous women such as discrimination and the lack of access to markets and capital caused both by high levels of poverty and the remoteness of their communities. Box 3.6: In focus: Guatemala: Women in agriculture: The impact of male out-migration on women’s agency, household welfare, and agricultural productivity Migration is transforming rural economies, landscapes, and, potentially, gender relations. More than 70 percent of Guatemalan migrants are young males. Migration is one of the drivers of the so-called “feminization of agriculture” in Latin America (Deere and León de Leal, 2001). A recent mixed-methods study (World Bank, 2015c) interviewed three groups: • Type 1: Women whose male partners are currently migrants. • Type 2: Women in households where both the male and female heads are present (indepen- dent of any possible migration history). • Type 3: Single female-headed households. Five key findings of interest from the study are as follows: • Contrary to the popular belief held by local officials, policymakers, and researchers alike, the clear majority of households remain in agriculture even when the male head of house- hold migrates. • The continuation of agriculture as a household livelihood strategy is characterized by the transformation and expansion of the role of married women in agricultural production. As men in southeastern Guatemala now migrate for years at a time, their partners face greater responsibilities in agricultural production, both in decision-making and in production itself. • These households, where the male partner has migrated, are more likely than other types of households to employ non-household members or paid workers for agricultural labor. This gap persists even when controlling for the dependency ratio and household size. Towards Equal? Women in Central America 77 • As agriculture is still seen as a traditionally male endeavor, women reported, not only need- ing to take on farming, but also needing to learn how to farm once their husbands migrated. However, extension services and technical assistance generally fail to reach women in rural areas. • Households with a male partner who has migrated have the highest levels of food security and food diversity compared to other households. Given the high level of remittances re- ceived by these households and the fact that remittances tend to go directly to women, this result is in line with literature showing that money controlled by women is more likely to be spent on family nutrition than money controlled by men (Thomas, 1990). Source: Deere and León de Leal (2001), Thomas (1990), and World Bank (2015c) Figure 3.15: Share of adults with a bank account Women Men 60 48 44 41 35 35 24 26 27 23 18 15 14 10 El Salvador Nicaragua Honduras Panama Guatemala LAC Costa Rica (2011) (2014) (2011) (2011) (2014) (2011) (2011) Source: Authors’ tabulations using the World Bank’s World Development Indicators. Year of data is included in parentheses. In all countries in Central America, women are less cent of women and men respectively) in the other five likely to have a bank account than men (Figure 3.15). countries. The percentage of women holding an account at a The percentage of women using their accounts formal financial institution in most countries in Cen- for business purposes is much lower than that among tral America is below the average in the LAC region, men.60 In 2011, only 2 percent of Panamanian and except in Costa Rica. The lowest rate for both wom- Costa Rican women used their account for business en and men is in El Salvador where only 10 percent purposes compared to 4 percent and 7 percent of men of women and 18 percent of men had an account in from each of the two countries, respectively. This gap 2011. In Costa Rica, on the other hand, 60 percent of in business financing is somewhat larger than in sim- men and 41 percent of women had an account. Rates for both sexes fall below LAC averages (35 and 44 per- 60 Unless otherwise stated, statistics about financing are from the World Bank’s Gender Statistics Database for the latest year available. Chapter 3 Economic opportunities 78 ilar countries: the gap is 4 and 7 percent of women and men respectively in LAC, and 3 and 5 percent of women and men respectively in upper-middle-in- come countries. Very few men and women used ac- counts for business purposes in Honduras and El Sal- vador: only 1 percent of Honduran women, 2 percent of Honduran men, 0 percent of Salvadoran women, and 1 percent of Salvadoran men. This is significantly lower than in the LAC region and in lower-middle-in- come countries - 2 and 6 percent for women and men respectively. While data are limited, the available evidence sug- gest that women are more credit-constrained than men in all of the countries. In Panama, for example, 12 percent of men and 9 percent of women held a credit card in 2011. In 2014, around 33 percent of women in Nicaragua and 40 percent in Guatemala reported that it was not possible for them to come up with emergency funds compared to 22 and 29 percent of men respectively, (Demirguc-Kunt et al. 2015). In the same year, 39 percent of women and 46 percent of men borrowed money in Nicaragua. Interestingly, a higher share of women than men made use of a fi- nancial institution rather than resorting to informal sources when borrowing -- 16 percent compared to 13 percent of men. Women also make up the majority of borrowers from microfinance institutions in Nica- ragua -- 55 percent in 2014.61 61 Tabulations from MIX (2014). Towards Equal? Women in Central America 79 Annex 3.1 Table A .1: Estimating the effect of education on female labor force participation Costa Rica El Salvador Guatemala 2004 2014 2004 2014 2004 2014 Complete Primary 0.264*** 0.177*** 0.309*** 0.243*** 0.287*** 0.216*** (0.007) (0.006) (0.005) (0.005) (0.004) (0.003) Incomplete Secondary 0.544*** 0.431*** 0.254*** 0.514*** 0.329*** 0.332*** (0.007) (0.007) (0.008) (0.007) (0.005) (0.003) Complete Secondary 0.753*** 0.841*** 0.613*** 0.406*** 0.802*** 0.798*** (0.008) (0.007) (0.005) (0.005) (0.004) (0.003) Incomplete Superior 0.738*** 0.970*** 0.685*** 0.722*** 1.040*** 1.073*** (0.010) (0.007) (0.009) (0.007) (0.010) (0.007) Complete Superior 1.747*** 2.137*** 1.157*** 1.494*** 0.703*** 1.362*** (0.009) (0.009) (0.008) (0.009) (0.009) (0.008) Age 0.279*** 0.303*** 0.211*** 0.218*** 0.198*** 0.194*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Age-squared -0.004*** -0.004*** -0.003*** -0.003*** -0.003*** -0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Urban 0.320*** 0.317*** 0.557*** 0.644*** 0.601*** 0.734*** (0.005) (0.004) (0.004) (0.004) (0.003) (0.003) Enrolled -0.294*** -0.624*** -1.477*** -1.730*** -0.859*** -0.768*** (0.008) (0.006) (0.009) (0.009) (0.006) (0.006) Married -1.224*** -0.956*** -0.767*** -0.770*** -1.067*** -1.099*** (0.005) (0.004) (0.003) (0.003) (0.003) (0.002) Children under 5 -0.130*** -0.236*** -0.244*** -0.336*** -0.163*** -0.149*** (0.005) (0.004) (0.003) (0.004) (0.003) (0.002) Children 5-14 -0.218*** -0.269*** 0.091*** 0.065*** -0.021*** 0.104*** (0.004) (0.004) (0.003) (0.003) (0.003) (0.002) Regional FE X X X X X X Constant -4.528*** -4.885*** -3.677*** -3.811*** -3.190*** -3.534*** (0.023) (0.021) (0.017) (0.016) (0.012) (0.011) Observations 12,391 12,567 18,885 24,278 16,856 14,996 Pseudo R-squared 0.136 0.152 0.117 0.122 0.0973 0.106 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This table reports the results of logit regression models. The dependent variable is a binary variable indicating whether the individual is in the labor force. The analysis is limited to women between the ages of 18 and 65. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 Chapter 3 Economic opportunities 80 Table A .1: Estimating the effect of education on female labor force participation (continued) Honduras Nicaragua Panama 2004 2014 2004 2014 2004 2014 Complete Primary 0.297*** 0.158*** 0.295*** 0.176*** 0.038*** -0.358*** (0.004) (0.004) (0.006) (0.005) (0.009) (0.009) Incomplete Secondary 0.327*** 0.352*** 0.533*** 0.496*** 0.264*** -0.068*** (0.006) (0.004) (0.006) (0.005) (0.009) (0.008) Complete Secondary 0.932*** 0.627*** 0.699*** 0.557*** 0.646*** 0.308*** (0.006) (0.005) (0.007) (0.006) (0.009) (0.009) Incomplete Superior 0.530*** 0.370*** 1.032*** 0.686*** 1.129*** 0.592*** (0.010) (0.007) (0.009) (0.007) (0.010) (0.009) Complete Superior 1.883*** 1.421*** 1.490*** 1.471*** 1.909*** 1.337*** (0.015) (0.008) (0.010) (0.007) (0.012) (0.010) Age 0.172*** 0.215*** 0.209*** 0.284*** 0.261*** 0.252*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Age-squared -0.002*** -0.003*** -0.003*** -0.003*** -0.003*** -0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Urban 0.631*** 0.298*** 0.497*** 0.697*** 0.224*** 0.060*** (0.004) (0.003) (0.004) (0.004) (0.006) (0.006) Enrolled -0.297*** -0.408*** -0.664*** -0.362*** -0.670*** -0.776*** (0.008) (0.006) (0.007) (0.007) (0.009) (0.008) Married -0.918*** -0.821*** -0.721*** -0.777*** -0.838*** -0.677*** (0.004) (0.003) (0.004) (0.004) (0.005) (0.005) Children under 5 -0.135*** -0.072*** -0.159*** -0.333*** -0.152*** -0.191*** (0.004) (0.003) (0.004) (0.003) (0.005) (0.005) Children 5-14 -0.031*** 0.125*** 0.090*** -0.001 -0.062*** -0.132*** (0.004) (0.003) (0.004) (0.004) (0.005) (0.004) Regional FE X X X X X X Constant -3.221*** -3.869*** -3.729*** -5.127*** -3.620*** -3.559*** (0.017) (0.014) (0.019) (0.018) (0.031) (0.028) Observations 9,180 6,924 9,252 8,779 15,049 12,563 Pseudo R-squared 0.106 0.0806 0.105 0.151 0.133 0.115 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This table reports the results of logit regression models. The dependent variable is a binary variable indicating whether the individual is in the labor force. The analysis is limited to women between the ages of 18 and 65. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 Towards Equal? Women in Central America 81 Annex 3.2 Table A .2: Estimating gender wage gaps, OLS and Heckman specifications Nicaragua Panama El Salvador I II I II I II Male 0.252*** 0.123*** 0.178*** 0.131*** 0.045*** 0.021* (0.016) (0.022) (0.011) (0.016) (0.009) (0.012) Age 0.036*** 0.018*** 0.019*** 0.013*** 0.028*** 0.019*** (0.004) (0.006) (0.003) (0.004) (0.003) (0.003) Age-squared -0.000*** -0.000** -0.000*** -0.000* -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Years of education -0.002 -0.006 -0.031*** -0.033*** -0.032*** -0.033*** (0.006) (0.006) (0.006) (0.006) (0.003) (0.003) Years of education squared 0.004*** 0.003*** 0.005*** 0.005*** 0.005*** 0.005*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Married -0.002 0.004 -0.003 -0.001 0.002 0.004 (0.010) (0.009) (0.003) (0.003) (0.002) (0.002) Enrolled -0.008 -0.011* -0.002 -0.001 0.003** 0.003** (0.006) (0.006) (0.001) (0.001) (0.001) (0.001) Part time 0.417*** 0.507*** 0.426*** 0.475*** 0.321*** 0.332*** (0.022) (0.023) (0.017) (0.019) (0.013) (0.013) Constant 2.042*** 3.268*** 0.169* 0.808*** -0.420*** 0.242*** (0.088) (0.142) (0.092) (0.119) (0.054) (0.083) Observations 9,078 13,305 9,435 13,522 11,605 19,711 Pseudo R-squared 0.180 0.325 0.321 Cens 4228 4087 8186 Rho -0.372 -0.235 -0.211 Sigma 0.748 0.528 0.468 Lambda -0.278 -0.124 -0.0987 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This table reports the results of OLS (column I) and Heckman selection (column II) models. The dependent variable is the natural log of wages. The analysis is limited to urban workers with positive wages between the ages of 18 and 65. The regressions also control for the number of children in the household (ages 0-4 and ages 5-14), as well as regional and employment type fixed effects. The Heckman selection model also includes household head status and school enrollment status. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 Chapter 3 Economic opportunities 82 Table A .2: (continued) Honduras Costa Rica Guatemala I II I II I II Male 0.191*** 0.046 0.190*** 0.135*** 0.172*** 0.119*** (0.024) (0.045) (0.018) (0.027) (0.017) (0.026) Age 0.038*** 0.022** 0.034*** 0.022*** 0.061*** 0.061*** (0.006) (0.009) (0.005) (0.007) (0.004) (0.005) Age-squared -0.000*** -0.000 -0.000*** -0.000** -0.001*** -0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Years of education 0.019** 0.019** 0.002 0.001 0.031*** 0.029*** (0.009) (0.009) (0.009) (0.009) (0.005) (0.005) Years of education squared 0.004*** 0.004*** 0.006*** 0.005*** 0.004*** 0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Married -0.004 -0.002 0.006 0.006 -0.004 -0.001 (0.005) (0.005) (0.004) (0.004) (0.004) (0.004) Enrolled -0.003 -0.002 -0.002 -0.002 -0.001 -0.001 (0.003) (0.003) (0.002) (0.002) (0.002) (0.002) Part time 0.399*** 0.546*** 0.275*** 0.398*** 0.334*** 0.394*** (0.030) (0.032) (0.024) (0.024) (0.022) (0.022) Constant 1.893*** 2.443*** 5.944*** 6.654*** 0.691*** 1.329*** (0.130) (0.240) (0.113) (0.180) (0.084) (0.122) Observations 4,241 6,738 7,390 11,797 8,085 12,010 Pseudo R-squared 0.310 0.329 0.311 Cens 2497 4411 3925 Rho -0.367 -0.260 -0.0856 Sigma 0.746 0.725 0.678 Lambda -0.274 -0.188 -0.0580 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This table reports the results of OLS (column I) and Heckman selection (column II) models. The dependent variable is the natural log of wages. The analysis is limited to urban workers with positive wages between the ages of 18 and 65. The regressions also control for the number of children in the household (ages 0-4 and ages 5-14), as well as regional and employment type fixed effects. The Heckman selection model also includes household head status and school enrollment status. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 Towards Equal? Women in Central America 83 84 chapter 4 Underlying Drivers: Gender Roles and Social Norms T he previous chapters have highlighted key ar- eas of progress in gender equality in the Cen- tral American countries but have also pointed to con- Social norms are “powerful prescriptions of acceptable behavior and they are reflected both in formal struc- tures of society and in its informal rules, beliefs, and tinuing challenges. Progress has been made mainly in attitudes” (Muñoz Boudet et al, 2012). Gender norms terms of improving legal and institutional frameworks can play a determinant role in shaping the legal, insti- as well as in increasing access to education and health. tutional, and policy framework and, what is more, its However, while these are necessary conditions for effective enforcement. In addition, they can also be increasing gender equality, they do not automatical- key drivers of decisions about human capital invest- ly translate into the sphere of employment and jobs. ments and of behavior that favors or deters access to This chapter explores one of the factors that may be economic opportunities for women. underlying this apparent disconnect between progress These norms may not only trigger discrimina- in some areas and stagnation in others -- social norms tion from others but are often internalized by wom- regarding the roles of women. en themselves and negatively affect their agency. In- This chapter argues that a common driver under- dividual decision-making is influenced by context, lies the outcomes described in the previous chapters, networks, and social norms. Together, these factors which is that social norms about the roles played by influence what individuals perceive as desirable and women are evolving only slowly in Central America. what they perceive as possible for them. At the same Towards Equal? Women in Central America 85 time, individual preferences do not develop in a vacu- ties vis-à-vis men’s. This chapter considers the role of um; they are constrained or enabled by broader men- gender norms from three angles. The first assesses the tal models and social structures and norms in society. links between norms, aspirations, and outcomes. The These mental models and social norms can funda- second delves into key household decisions, in partic- mentally restrict individuals’ abilities to choose and ular household formation and labor allocation, which decide freely, leading people to favor the adoption of have enormous implications for gender parity in Cen- “common” or pre-existing behavior. Sociological and tral America. The third section emphasizes that this anthropological literature has shown how people’s story is not monolithic – rather, there are significant attitudes and beliefs are shaped by the social groups differences in progress and outcomes among Central to which they feel they belong (Bourdieu, 1977 and American women. Kleinman, 2006). This can be seen, for instance, when women adapt their aspirations to what they perceive is achievable and possible for them. 4.1 Gender norms, Households are the crucial link between gender aspirations, and norms and outcomes for women. The framework outcomes introduced in the World Development Report 2012 Social norms have an impact on both opportunities (World Bank, 2012b) shows that, while market and and aspirations. By prescribing gender roles regard- formal and informal institutions all influence gender ing the division of power, labor force participation, outcomes in several difeerent ways, the influence of activities, and spaces to occupy within a society, social each is channeled through households. People’s pref- norms may limit the opportunities that are available erences are shaped by social norms, and those, in turn, to women (Muñoz Boudet et al, 2012). Furthermore, influence the decisions of household members. These men’s and women’s choices are framed within their are the decisions that set the stage for the possibilities specific contexts, –meaning that individuals adapt and opportunities available to women in practice in their preferences to the opportunity structure avail- specific societies. Norms shape the division of labor able to them (Elster, 1983). Living in conditions of within the household and assign specific roles to each poverty or social and economic exclusion, for exam- member. In short, social norms and the way in which ple, affects individuals’ attitudes about themselves and they are channeled through the household can have a their aspirations. Poverty can lead to lower aspirations strong impact on the development outcomes of each and lower self-confidence, reducing an individual’s individual member of the household. likelihood and ability to deliberatively plan for the fu- ture (Appadurai, 2004 and World Bank, 2015d). In The prevailing social norms regarding the role of the absence of a real exercise of agency, there is no real women in society vis-à-vis that of men potentially may choice either but merely a “simulacrum of choice”62 be the underlying driver of gender gaps in all areas (Nussbaum, 2000, pg. 53). of economic and social life. Therefore, policymakers need to understand what the predominant social and gender norms are in the countries of Central America, how they are perpetuated, and the ways in which they 62 From the Latin: ‘simulacrum’: “likeness, similarity” – here, em- phasizing the idea of the imitation of a choice but the absence of shape women´s outcomes and access to opportuni- a real choice. Chapter 4 Underlying Drivers: Gender Roles and Social Norms 86 In this way, women and men shape their own mate partner violence and the desirability of women preferences based on the choices that they perceive as working outside of the home. available to them. Because they may adapt to what Based on the existing data, large shares of wom- they deem to be possible for themselves, the absence en in Central America agree with statements that of choice and equality of opportunity may affect the place husbands as authoritative figures over wives. individual’s capacity to aspire (Appadurai, 2004). If The statements reflect the hierarchical power and de- women’s preferences adapt in this way to their con- cision-making structure between the sexes and show text, they may be caught in an “inequality trap (..) that a significant share of women believe that men because of a context that affects their ability to see should be the main decision-makers and hold pow- the pathway to achieve their desired goals” (Muñoz er within their household. As an example, in 2006, Boudet et al, 2012, pg. 24). Agency is important in 43 and 40 percent of Nicaraguan women agreed with this context since it is a necessary requirement when the statements that “a wife should obey her husband” trying to challenge social norms. “Increased agency and “a man needs to show his wife that he is the allows women to move from enduring complete com- boss” respectively.64 Similarly, in El Salvador in 2008, pliance to constraining and unequal gender norms, to 43 percent of women reported that women should questioning those norms in face of potential oppor- obey their husbands, even if they do not agree with tunities, to changing their aspirations, as well as their them.65 In 2006/07, around 50 percent of Guatema- ability to seek and achieve desired outcomes” (Muñoz lan women agreed with the statement that “the man Boudet et al 2012, pg. 14). should demonstrate that he is the one in charge in Norms permeate beliefs and behaviors and are re- the household” and 29 percent of women who were inforced through the expectations of what others will ever married or in a union (15-49) agreed that a wife do (Bicchieri, 2006). An individual’s action depends has an obligation to have sexual intercourse with her both on what he or she believes other people are do- husband, even if she doesn’t want to.66 ing, and on what he or she thinks others expect of her/ These beliefs of a hierarchy between spouses in him. Hence, individual choices affect and are affected some cases translates into support for violence against by those of others. According to Bicchieri, individual women among women themselves. The share of wom- choices are constructed out of beliefs (both normative en who agree with the statement that wife-beating is and empirical beliefs), expectations (both normative justifiable under certain circumstances is particularly and empirical), and material constraints. The result is a feedback loop between gender changes over time often cannot be tracked. The next wave of norms, the aspirations of women, and outcomes that the World Values Survey (WVS wave 7 - 2016-2017) will include can stretch across generations. In this section, two ex- Panama, Costa Rica, and Guatemala and will yield an additional wealth of information on the existing gender norms in other amples of this feedback loop are explored using data areas including political and economic life. from perception surveys63 -- tolerance towards inti- 64 Estimates based on Nicaragua’s Demographic and Health Survey (Encuesta Nacional de Demografía y Salud, ENDESA) 2006/2007. 63 Social norms and beliefs about appropriate behaviors for wom- en and men are reflected in perception surveys and attitudes of 65 Estimates based on El Salvador’s Reproductive Health Survey the general population with regard to different aspects of gender (Encuesta Nacional de Salud Familiar, FESAL), 2008. equality and women’s and men’s roles in society. However, data 66 Estimates based on Guatemala’s Reproductive Health Sur- on these topics are limited. Perception data are not available for vey (Encuesta Nacional de Salud Materno Infantil, ENSMI), all countries, they are unavailable on a number of topics, and 2006/2007. Towards Equal? Women in Central America 87 Figure 4.1: Agreement that wife-beating is justified in certain situations, among women 20,2 15,8 15 13,7 12,4 11 5,3 4,3 2,9 Paraguay Bolivia Panama Nicaragua Honduras Guatemala Perú* Dominican Republic Jamaica* Source: Bott et al. (2012). The data for each country come from the following surveys: Jamaica = 2008/9 Reproductive Health Survey (RHS), Guatemala = 2014/15 Demographic and Health Survey (DHS), Bolivia = 2008 DHS, Nicaragua = 2006/07 RHS, Dominican Republic = 2007 DHS, Peru = 2007/8 DHS, Paraguay = 2008 RHS, Honduras = 2011/12 DHS, and Panama = National Survey of Sexual and Reproductive Health (Encuesta Nacional de Salud Sexual y Reproductiva, ENASSER), 2009. Notes: * Among women ever married or in union aged 15-49. revealing (Figure 4.1).67 In the four Central Ameri- coincides with the reality of high rates of violence can countries with data on this question, more than against women. Figure 4.1 also reflects that women one out of ten women agreed with this statement. In themselves have internalized the belief that wives can Honduras and Nicaragua 12 and 14 percent of wom- “misbehave” and hence IPV can at times be justified. en respectively reported that wife-beating is justified This rationalization may also help to explain why so for at least one reason.68 In Panama, the share of both few women seek out institutional or social support women and men who reported that wife-beating was when they are victims of IPV because even if a woman justifiable reached 15 percent. The share of women herself does not agree with this belief, she knows that who agree, at 11 percent, was lowest in Guatemala. others around her may. These beliefs may also influ- ence the development of formal institutions to com- As seen in Chapter 1, intimate partner violence bat IPV, thus limiting the availability of institutional is a serious issue in Central America. The perception support for victims of IPV. Furthermore, those in that beating one’s wife is justified in certain situations charge of service delivery (if services are available) may 67 According to Bott et al. (2012), Demographic and Health also be guided by such beliefs and may not provide the Surveys and Reproductive and Health surveys asked about dif- ferent but overlapping sets of reasons, including circumstances best possible assistance to victims if they do agree with in which the wife: (i) neglects the children or housework; (ii) statements that justify violence. Indeed, social norms refuses sexual intercourse; (iii) goes out without telling her husband; (iv) is or is suspected of being unfaithful; (v) disobeys shape the design (and enforcement) of laws. This can or argues with husband; or (vi) burns the food. be particularly crucial in the case of gender-based vi- 68 Estimates based on Nicaragua’s Demographic and Health Survey (Encuesta Nacional de Demografía y Salud, ENDESA) olence. For instance, Klugman (2017) shows that in 2006/2007 and Honduras’ Demographic and Health Survey (En- Guatemala, where a law against violence was enact- cuesta Nacional de Demografía y Salud, ENDESA), 2011-2012. Chapter 4 Underlying Drivers: Gender Roles and Social Norms 88 ed in 2008, thousands of cases go to court (only in dez and Fogli, 2006). This research suggests that more 2011, there were more than 20, 000) but less than traditional views negatively correlate with female la- 3% resulted in judgement in 2011 (and often result- bor force participation (and the gender wage gap). ing in small fines). According to Musalo and Bookey For instance, Fortin (2005) analyzing data from the (2014), this has been attributed to gender biases and World Values Survey dataset for a number of OECD stereotypes which prejudice proper investigation and countries shows that past agreement rates with the prosecution. statement “When jobs are scarce, men should have Social norms play an important role in the relation more right to a job than women” are strong predictors between women’s agency and the opportunities avail- of women’s future employment rates. Furthermore, able to them in that they can facilitate or hinder in- the perception of women as homemakers, measured dividuals from taking advantage of available opportu- as agreement with the statement “being a housewife is nities. Specifically, recent research has shown a strong just as fulfilling as working for pay” is also strongly as- link between social norms and female labor force par- sociated with negative female labor market outcomes ticipation (Fortin, 2005; Goldin, 2006; and Fernán- (Box 4.1). Box 4.1: In focus: Roads to agency – barriers and effects of integrating women into non-traditional sectors The Roads to Agency study (Casabonne et al, 2015) looked at factors that enable and constrain women’s participation in road works and ancillary productive activities and at the challenges faced by women during their participation in the program. One of the case studies was on the Nicaragua Fourth Roads Rehabilitation and Maintenance Project, a project effective since May 2006, and the subsequent Rural Roads Infrastructure Improvement Project, referred to as the Fifth Roads Project, approved in December 2011. In each country studied (Argentina, Peru, and Nicaragua), the most cited factors that en- couraged women’s participation in these activities was economic need and their desire to make a contribution to the community. A key barrier to women’s participation was their inability to find childcare. In Nicaragua, women who did not participate often cited a lack of information and their perception of bias in the selection process. Even among those who participated, there were initial concerns about potentially not being able to perform as well as the men. “When I found out about the road work job opportunity, I got nervous about being surround- ed by men, I was afraid of making mistakes of being scolded and telling me ‘this is not how it is.’ I was afraid that they would tell me that it is better for me to leave.” Woman MCA participant, Los Ángeles- Nicaragua, 43 years old Women participants in the road works also mentioned several challenges during program participation. Among these challenges was shouldering the double-burden resulting from working outside the home and bearing the bulk of the household responsibilities, as well as Towards Equal? Women in Central America 89 the hard work conditions. In Nicaragua, women specifically expressed their frustration about the gender-segregation of the road work tasks. Participating in the road works had strong effects on women’s agency. Women that were em- ployed in road works and participated in productive activities related to the newly rehabilitated roads reported not only having more control over their economic resources but also enhanced self-esteem, confidence in themselves, and increased aspirations in terms of their life plans as well. The acquisition of skills and know-how, the pride of making a contribution to the com- munity, and the acquisition of leadership skills were also positive outcomes mentioned by the interviewees. “I liked it very much because I used to bring my own money to the household and I did not depend on him. I could buy what I wanted, that is what I liked the most. Feeling fulfilled as a woman.” Woman MCA participant, Moyogalpa-Nicaragua, 36 years old “This work really strengthened my self-esteem, I can now move on, not only in that job, but also in other tasks.” Woman MCA participant, El Ñámbaro-Nicaragua, 43 years old The projects were also instrumental in broadening women’s networks through their partic- ipation in the community-run road work organizations. The participants’ greater physical mo- bility and resulting social networks gave women increased access to peers, social support, and information. The study showed how barriers limiting women’s agency can (slowly) be deconstructed by increasing women’s participation in the public sphere (work or community participation) and by their taking on non-traditional roles, such as working in road construction and maintenance. While their husbands and community members might have had initial doubts, concerns, or even resistance to these new roles, the projects offered an opportunity for women to step into non-traditional roles and demonstrate the advantages of this to skeptics. Like every Nicaraguan, I was very machista. I thought that the wife should stay at home, even if I met my wife when we worked together. I thought that she should stay at home. Now my viewpoint has changed 180 degrees. Now, I think it is something positive for her to have a job and generate income for the household.” Nicaragua, husband of MCA woman participant Seeing women perform and perform well challenged traditional gender stereotypes. While everyone might not have immediately agreed with these new roles, initiatives such as these are a small step in changing gender norms and increasing gender equality. Women’s husbands reported feeling more respect for their wives and valued them more after they became in- come earners, which indicates that the women’s bargaining power within the household had increased. Chapter 4 Underlying Drivers: Gender Roles and Social Norms 90 At the community level, interviewees reported that the new roads brought many economic and social benefits and a shift in the perception of gender norms. Women’s entry into non-tra- ditional work challenged prevailing social norms on women’s roles, abilities, and participation in the public sphere. The study’s findings provide a powerful rationale for including ways to enhance women’s agency when creating employment opportunities in infrastructure projects and similar eco- nomic empowerment projects. Figure 4.2: Women should only work if the partner does not earn enough a) 2015 b) Change 2008 to 2015 (percentage points) BRA 23,4 PAN CHL 8,0 GTM PER 3,8 MEX URY -0,3 DOM VEN ECU -2,0 HND ARG -2,7 ARG PRY -2,8 SLV COL -4,0 CRI BOL -6,8 BOL CRI -8,3 ECU NIC DOM -8,7 NIC MEX -9,1 PER SLV -10,4 PRY PAN -11,2 CHL HND -13,8 COL GTM -14,0 URY 0 20 40 60 80 100 -16,1 VEN % -19,9 BRA Strongly Agree Agree Disagree More disagree More agree Strongly Disagree Don’t know Source: 2015 data of Latinobarómetro Database, Latinobarómetro Corporation, Santiago, Chile, http://www.latinobarometro.org/latContents.jsp. Note: Panel b reports the change between 2008 and 2015 in the share of respondents who strongly agree or agree that women should only work if their partner does not earn enough Strikingly, in four out of the six Central Ameri- (Figure 4.2a). Yet even in Costa Rica, the country in can countries, most people believe that women should Central America with the lowest share of adults agree- work only if the husband does not earn enough mon- ing, more than two out of five (42 percent) agree. In ey. The exceptions are Costa Rica and Nicaragua Guatemala, at the other extreme, nearly two out of Towards Equal? Women in Central America 91 Figure 4.3: Correlation between female labor force participation rates and views of women’s work 90 Disagree or strongly disagree (%) CHL BRA 80 URY PER ECU 70 ARG PRY 60 DOM CRI COL BOL MEX SLV NIC 50 PAN 40 GTM HND 30 30 40 50 60 70 80 90 Female labor force particiaption rate (15+) Source: World Development Indicators and 2015 data of Latinobarómetro Database Note: The x-axis reports the labor force participation rate is for all women ages 15+ in 2014; The Y-axis reports the share of adults who disagree or strongly disagree that women should only work if the partner does not earn enough. three (63 percent) respondents agree. These beliefs are with yet more challenges in their attempt to enter the indicative of the gender division of roles and responsi- labor market. Not only do they lower women’s aspira- bilities that seems to prevail in the sub-region. tions to work but they also put pressure on husbands to discourage their wives from working as it could re- In most countries in LAC, the share of adults who flect poorly on them. believe that women should only work if the husband does not earn enough has been falling, but this change Attitudes to women’s work in Central America are seems to be slower in Central America (Figure 4.2b). strikingly different to those in South America. The Among the six countries of Central America, the countries of Central America, along with neighbor- change has been most notable in Nicaragua where the ing Mexico and the Dominican Republic, have the share who agreed fell by nearly 9 percentage points highest share of respondents in all LAC countries who between 2008 and 2015. Seven countries in LAC agree that women should only work when their hus- had larger changes during the same period, including bands do not earn enough income. The countries of Brazil in which the share of adults who agreed fell by South America, on the other hand, report large ma- nearly 20 percentage points. However, Panama and jorities disagreeing with that premise, including more Guatemala bucked the regional trend. In 2015, half than 80 percent of respondents in Brazil and Chile. of all Panamanians (51 percent) believed that women Importantly, these attitudes show some relation should work only if her partner does not earn enough; to each country’s female labor force participation this is compared to less than one-third (31 percent) rate (Figure 4.3). The countries with the lowest share in 2008. Such convictions and beliefs present women of adults disagreeing that women should only work Chapter 4 Underlying Drivers: Gender Roles and Social Norms 92 when their husbands do not earn enough are also the complex. A clear causal link between outcomes and countries where labor force participation of women perceptions cannot be established. Gender norms in- is lowest. A simple linear regression analysis shows fluence women’s aspirations and the types of jobs that that 22 percent of the difference in participation rates they seek or even whether they seek employment at across countries can be explained by differences in at- all. Evidence suggests that role models and women’s titudes. While some factors no doubt influence per- leadership can promote the aspirations and agency ceptions and others influence participation and many factors influence both perceptions and participation, of other women around them (see Box 4.2). At the these two indicators support a link between outcome same time, gender norms may reduce opportunities and perception when it comes to women working in for women, for example, by influencing the beliefs Central America. of employers and family about women’s professional As shown in the two examples above, the relation- capabilities, or through legal restrictions on jobs for ship between gender roles and gender outcomes is women, as noted in Chapter 1. Box 4.2: Social interactions with leaders matter: Changing households’ investment and aspirations through social interactions: Evidence from a randomized transfer program Low aspirations can limit households’ investments and contribute to their sustained poverty. Alternately, increased aspirations can lead to investment and upward mobility. Yet how aspira- tions are formed is not always well understood. Macours and Vakis (2014) analyzed the role of social interactions in determining aspirations in the context of a program aimed at increasing households’ investments. Atencion a Crisis was a one-year pilot program implemented between November 2005 and December 2006 by the Ministry of the Family in Nicaragua. It was implemented in the after- math of a severe drought and had two objectives. First, it aimed to serve as a short-run safety net by reducing the impact of the aggregate shock on human and physical capital investments. Second, it was intended to promote long-run upward mobility and reduce poverty by enhancing households’ asset base and their ability to diversify their income. Households were randomly assigned to three different intervention groups within randomly selected treatment communities. The three interventions were: (i) a basic conditional cash transfer (CCT); (ii) the basic CCT plus a scholarship for occupational training; and (iii) the basic CCT plus a grant for productive investment. Community leaders were also randomly allocated to one of these three interventions. The impact evaluation found that social interactions between the leaders and the women who participated in the intervention affected attitudes toward the future and amplified the pro- gram’s positive impact on investments in human capital and productive activities. The empiri- cal evidence indicates that when women interact with motivated and successful leaders, this can give them higher aspirations and cause them to make similar investment decisions. The participants who lived near leaders substantially increased their human capital investment and income diversification and had more hopeful attitudes towards the future. The increased Towards Equal? Women in Central America 93 aspirations of the participants were driven by their increased communication with and motiva- tion from female leaders. The effects were particularly large when participants and leaders par- ticipated in the same intervention. These effects continued two years after the program ended. Sources: Macours and Vakis (2014 and 2016) Box 4.3: In focus: Alliance for education quality in Nicaragua A growing body of evidence suggests that boosting the socioemotional skills of young people can have a positive impact on their education, health, and labor market outcomes in the short and long terms (Gimenez et al, 2015; Cunningham and Villaseñor, 2014). For example, adoles- cent aspirations and expectations about the future can influence the probability of teenage pregnancy (Plotnick 1992, 1993, 2007; Azevedo et al. 2012; Cater and Coleman 2006). Such skills have been proven not only to improve economic outcomes but also to lower the incidence of teenage pregnancy. The Alliance for Education Quality Project supports gender-sensitive classroom practices and the strengthening of students’ socioemotional skills. Promoting gender-sensitive class- room practices can help to ensure that girls and boys benefit from education equally and help them to feel empowered to achieve their aspirations. International experience shows that pro- moting gender sensitivity in classrooms can build more supportive school environments and prepare teachers to challenge gender discrimination, avoid gender stereotypes, and proactive- ly address issues such as gender-based violence, early marriage, or teenage motherhood in the classroom. Source: World Bank (2017b) value placed on each household member’s contribu- 4.2 The role of tion to the well-being of the household as a whole. households in Beyond women’s economic activity, households also gender outcomes make decisions about whether and how to invest in girls’ education and health care. Analyzing and understanding decisions taken within households and how these are influenced by gender Expectations about women’s and men’s roles with- norms is crucial to understanding the progress, or lack in the home and in society more broadly have an im- thereof, of gender equality in Central America. For pact on how individuals see themselves and how they example, decisions about time allocations between construct their life goals. Specifically, beliefs about household work and paid employment are taken with- motherhood can have a strong impact on if and when in the household. The outcomes of those decisions de- women decide to have children. Furthermore, they pend on gender roles, women’s own preferences, the can have consequences for how women fill their role opportunities available in the labor market, and the as mothers. For instance, if motherhood is essential- Chapter 4 Underlying Drivers: Gender Roles and Social Norms 94 Figure 4.4: Age at first marriage, female, 1980-2010 26 1980 1990 2000 2010 24 22 Age 20 18 CRI SLV HND GTM NIC PAN Source: United Nations, Department of Economic and Social Affairs, Population Division. World Marriage Data ly the main role that girls see for themselves in their Timing of marriage adult lives and if motherhood is socially read as the main marker of adulthood for women, then young and of first births women will opt to become mothers earlier in life. On Changes in fertility rates have not been accompanied the other hand, if there are competing roles available by any major changes in the timing of family forma- to women in a given society, if women are expected to tion, including age at first marriage. The average ages have a professional career, to fulfill certain objectives at first marriage and at first births are surprisingly sta- and milestones before becoming a mother, then they ble in all Central American countries. This continuity will first opt to achieve those competing roles and ob- is an indicator of the strong and continuous values jectives before actively choosing motherhood. placed on the family and on the centrality of moth- erhood and family formation for women. As shown At the same time, gender norms and aspirations in Chapter 2, early childbirth is unlikely to be due to can also influence household creation in such a way a lack of access to contraception. Rather, in Central that the household reinforces gender norms. There- American countries, age at first marriage remains be- fore, analyzing the trends in family formation, such low the average for Latin America. In Latin America as age at first marriage and first birth as well as overall as a whole, average age at first marriage for women in- fertility patterns, are indicative of shifts (or the ab- creased from 21 years in the 1970s to over 23 years in sence thereof ) in women’s roles in society. As shown the early 2000s (Chioda, 2016). In Central America, below, in Central America declining fertility rates age at first marriage increased at a much slower pace have not been accompanied by changes in the timing in those countries for which data is available (Figure of marriage or first births. Rather, young marriage, in- 4.4). The average age at first marriage remained almost cluding child marriage, is still common. Age at mar- unchanged in Panama (21.4 in 1980 to 21.6 years in riage has direct implications for women’s educational 2010) and Honduras (20.9 in 1980 to 21.2 in 2010) attainment and economic activity. and was the lowest in Nicaragua (20.9 years in 2010, Towards Equal? Women in Central America 95 up from 19.8 in 1990). Costa Rica and El Salvador devote to housework but not men. Furthermore, hav- experienced the biggest changes, increasing from 22.2 ing children significantly increases the time spent on to 24.2 years and from 20 to 21.9 respectively during care by both men and women but more so for women. the same period. Among rural women, average age at Specifically, the presence of children under 5 years of first marriage remains below 19 years for Honduras age in the household is associated with an additional and Guatemala, the two countries for which data after 1.0 to 2.8 hours of care a day for women for women 2010 are available.69 and between 0.1 to 1.0 hours a day for men depending on the country (Berniell and Sánchez-Páramo, 2011). Those countries in Central America for which data Family formation patterns and the role of women are available have experienced only marginal increases in unpaid work are crucial factors in understanding in age at first birth. For instance, in Guatemala wom- women’s participation in the labor market. As Chioda en are having children a year later in their lives in 2015 (2016) has summarized, the literature essentially lists than in 1995 (21.2 years versus 20.2 years).70 In Hon- five main factors that explain increases in female labor duras, on the other hand, the median age at first birth force participation rates: (i) rising education levels; (ii) remained approximately the same between 2006 and urbanization; (iii) changes in the sectoral structure of 2012 at 20.0 (2005/2006) and 20.4 (20011/2012). the economy; (iv) increases in the use of labor-saving Education and socioeconomic status remain positive- household technology; and (v) changes in marriage ly correlated with age at first birth, with women with patterns and declining fertility. no education being more likely to have a child earlier in life than those with a secondary education or those Even though the average age of marriage and first with a higher socioeconomic status. For instance, in births has not changed much in Central America, a 2015, in Guatemala, the median age at first birth was significant shift has been observed in the composition 19.3 among women with no education compared of households. The share of households headed by with 23.1 years for those with a secondary education women has increased in all countries, ranging from or higher. an increase of 3.3 percentage points in Guatemala to 13.7 percentage points in Costa Rica. In all countries The centrality of marriage, motherhood, and car- except Guatemala, about one in three households are ing domestic roles can essentially frame the place headed by women (Figure 4.5). In earlier literature that women occupy in society and, hence, the way in on poverty in Latin America, female-headed house- which women plan and envision their lives. Economic holds were associated with higher levels of poverty activity and financial independence may not play a and the “feminization of poverty” (Buvinic and Gup- central role in these visions. On the other hand, the ta, 1997 and Pearce, 1978). However, more recent smaller the role that caregiving plays in women’s lives, work has shown that the picture has become more the greater the time available to women to allocate to nuanced, indicating that female-headed households other activities such as education, paid labor, and po- in Latin America do not necessarily fare worse in litical or civic engagement. Research has shown that terms of poverty than male-headed households and marriage significantly increases the time that women that they may actually be a result of the increased 69 Estimates based on Demographic and Health Survey (DHS) data emancipation or economic power of women due to from www.statcompiler.com. their capacity to earn their own income (Liu, Esteve, 70 Data on age at first birth for all countries in this paragraph are DHS data from www.statcompiler.com. and Trevino, 2017). Chapter 4 Underlying Drivers: Gender Roles and Social Norms 96 Figure 4.5: Rates of female-headed households 2000 2015 35,6 35,7 36,2 32,3 33,2 28,4 28,3 30,4 26,1 21,6 22,5 18,3 GTM HND PAN NIC SLV CRI Source: Authors’ tabulations based on SEDLAC (CEDLAS and the World Bank). latest ENDESA (2011/12) showed that sexual initi- Child marriage ation in Nicaragua is happening at a younger age. A The Latin America region has alarmingly high rates of young age of sexual initiation and early marriage are child marriage, and furthermore, it is the only region well-documented correlates of teenage motherhood. in the world where child marriage is not on the de- In all countries in Central America, girls face higher cline. No significant change has been observed in its prevalence over the last 30 years.71 Child marriage is risks that boys of child marriage. defined as a marriage or union taking place before the Poverty, lack of educational and employment op- age of 18. The threshold of 18 years to define child portunities for girls, and prevailing social norms are marriage is used in conventions, treaties, and interna- factors that contribute to this phenomenon. Gender tional agreements, including the Convention on the roles that assign women to the domestic space and Rights of the Child, the Convention on the Elimina- to caregiving roles combined with poverty and lack tion of All Forms of Discrimination Against Women, of opportunities leave few options for girls to choose and the Universal Declaration of Human Rights. from other than becoming a mother and a wife. Not Within the Latin America region, Central Amer- only is the prevalence of early marriage higher among ican countries have some of the highest rates of child the poor, but they may also suffer more from its conse- marriage.72 Figure 4.6 shows that all Central American countries have child marriage rates above 20 percent. quences because of the various constraints facing those Furthermore, Panama, Guatemala, Honduras, and living in poverty (such as barriers to accessing health Nicaragua are in the top six Latin American countries and education services). Gender roles and norms are in terms of the highest prevalence of child marriage. extremely important in the context of child marriage More than one third (43 percent) of all women aged as also reflected in several recent reviews on the topic 20-24 married before the age of 18 in Nicaragua. The (Vogelstein, 2013; UNFPA, 2012b; UNICEF, 2014; 71 Source: UNICEF data reported by www.girlsnotbrides.org. and Klugman et al, 2014; as well as Wodon et al, 2017 72 Source: UNICEF data reported by www.girlsnotbrides.org. for a review of the literature). Towards Equal? Women in Central America 97 Figure 4.6: Rates of child marriage, 2008-2014 45 40 Child marriage (%) Child marriage (%) Percent of women (20-24) 35 (2008-2014)* Married by 15 – (2008-2014)* Married by 18 – 30 25 20 15 10 5 0 Nicaragua Dominican Republic Brazil Honduras Guatemala Panama Cuba Belize El Salvador Uruguay Colombia Guyana Mexico Ecuador (Plurinational State) Costa Rica Peru Paraguay Haiti Trinidad y Tobago Jamaica Saint Lucia Chile Bolivia Source: UNICEF, Global database Note: This figure reports the percentage of women aged 20 to 24 years who were first married or in a union before the ages of 15 and 18. While on the one hand child marriage is a mani- are considered, women who married as children have festation of disadvantage in several dimensions, it can expected earnings (actual or imputed) in adulthood itself contribute to a lack of agency and perpetuate that are lower than women who married after the gender inequality in other dimensions even further. age of 18 (the average loss in earnings is 9 percent). For example, Wodon et al (2017) found that, glob- Furthermore, child marriage is often associated with ally, child marriage is likely to be the cause of at least losses in agency and decision-making for women later 84 percent of births to mothers younger than 18. in life. It may increase the control of their husbands73 Evidence shows that delays in marriage are associat- and in-laws over their lives, thus limiting women’s as- ed with greater education, earnings, and health-seek- pirations and their ability to make decisions, includ- ing behavior (Field and Ambrus, 2008 and Pezzini, ing those about accessing health care during pregnan- 2005). Field and Ambrus (2008), using data from ru- cy and delivery. ral Bangladesh, showed that each additional year that Several studies have also found a link between marriage is delayed is associated with 0.22 addition- child marriage and an increased risk of intimate part- al year of schooling and 5.6 percent higher literacy. ner violence (Clark et al, 2006; Carbone-Lopez, 2006; Child marriage also negatively affects the education of Solotaroff and Pande, 2014; and Hong et al, 2014). In the children born to these mothers because a mother’s Guatemala for instance, 34 percent of women who educational attainment is one of the key factors that 73 A study of 29 countries in Africa and Latin America revealed affect her children’s educational attainment. While that girls who marry at young ages have husbands who are Wodon et al (2017) found that child marriage may five to fourteen years older on average, which can create an insurmountable power differential and preclude autonomous not affect labor force participation once other factors decision-making (Clark et al, 2006). Chapter 4 Underlying Drivers: Gender Roles and Social Norms 98 entered their first relationship before the age of 15 re- the past two decades (Chioda, 2016).74 This is because ported having experienced physical violence at some decisions about labor market participation are inter- point in their life compared to a much lower incidence twined with decisions about marital status and fertil- among women whose first relationship happened later ity. This is partly because of the greater responsibility in life (15.2 percent, for instance, for those who got for domestic and care work placed on women, which married after 25 and 19.7 percent among those mar- limits the amount of time they can devote to working ried between 20 and 24 years old) (Bott et al., 2012). outside the home. Child marriage is a factor in the high rates of wom- Charmes (2015) finds that women in all Central en aged 15-24 who are out of school and not work- American countries tend to bear the brunt of domes- ing (ninis) in Central America who were discussed in tic and care work and have a higher overall workload Chapters 2 and 3. The most important risk factors than men.75 Independent of the exact differences, the that cause young women to exit the labor force and report found a consistent pattern of women devoting drop out of school are marriage before the age of 18 significantly more time to domestic tasks – cleaning, followed by teenage pregnancy (De Hoyos, Rogers, cooking, child care, and care of the elderly – than and Székely, 2016). One of the most common paths men, ranging from 5.3 times as much in El Salvador to becoming a nini is through early school dropout (2010) to 3.7 times as much in Costa Rica (2004) and 2.4 times as much in urban Panama (2011). When followed by unemployment. In Central America, a both paid and unpaid labor is considered, women also substantial share of school dropouts happen in lower work consistently more hours every week than men or upper-secondary school; on average, nearly a quar- (Table 4.1). Therefore, women in Central America ter of youths have left the education system by the carry a “double burden” of employment and unpaid age of 15. Similarly, in El Salvador, female ninis often care work and thus end up with more overall hours mention the need to do household work as their main of work than men. They are not only largely in charge reason for dropping out of school, and 45 percent of of unpaid care or domestic work, but they also spend them are either married or living with a partner com- less time in paid work than men. This is important pared to only 10 percent of male ninis (Calvo-Gonza- since women’s non-remunerated work puts them in a lez and Lopez, 2015). Throughout Central America, position of being economically dependent on others. the nini phenomenon may be keeping the sub-region from fully benefitting from its demographic dividend. Given that women are commonly assigned care responsibilities, a lack of childcare can be a key bar- Households, care work, rier to their labor force participation. Although many different policies will be needed to overcome the bar- and economic outcomes riers that women face to entering the labor market, Family formation patterns affect labor force partici- access to childcare seems to be the policy that has the pation. Changes to family formation (such as having 74 Chioda (2016) estimated that changes in family formation ex- fewer children and delaying the age of first marriage) plain between 18 and 31 percent of the changes in female labor force participation between the 1960s and the 2000s in Latin – as well as increased female education – can explain America. almost two-thirds of the increase in women’s labor 75 The data for Costa Rica and El Salvador come from a dedicated module in the household survey, while the data for Panama force participation in 10 Latin American countries in (urban) come from a dedicated time use survey. Towards Equal? Women in Central America 99 Table 4.1: Average time spent by women in unpaid and paid work (in minutes per day) Costa Rica, 2004 Panama (urban), 2011 El Salvador, 2010 Women Men Women Men Women Men Unpaid work 385 105 286 120 228 43 Paid work 122 352 201 311 192 346 TOTAL 507 457 487 431 420 389 Source: Charmes (2015). most consistent positive effects on female labor force smallest share of users of private services (Mateo Dias participation (Busso and Romero Fonseca, 2015). Ex- and Rodriguez-Chamussy, 2016). isting evidence on access to childcare and female la- Childcare costs, including fees and transportation, bor supply suggests that, as the price of childcare falls, can be a vital factor in women’s decisions to enter the maternal labor force participation increases (Blau and labor force and use non-parental childcare. Among Currie, 2006 and Bick, 2015). Similarly, most litera- households that pay for private childcare, the average ture finds that informal childcare arrangements have out-of-pocket expenditure ranges from 14 percent a positive effect on female labor force participation of household per capita income in Nicaragua to 31 (Posadas and Vidal-Fernandez, 2012; Arpino et al., percent in Honduras. In fact, use of these services by 2010; and Dimova and Wolff, 2011). households in the richest quintile is more than dou- Informal care arrangements are used more fre- ble that of households in the poorest quintile in El quently than formal childcare in Latin America, es- Salvador and Honduras. In Nicaragua, use of for- pecially for younger children. The rate of enrollment mal childcare services is also higher among wealthier in formal childcare programs is below 30 percent for households though the gap between the top and bot- children aged 0-3 in Latin America and ranges from tom quintiles is smaller. Although available informa- 10-75 percent for children aged 3-5. Countries in tion is limited on the frequency and intensity of use of Central America have the lowest rates of formal child- childcare, evidence from Honduras indicates formal care use in the region. Enrollment rates are below 5 arrangements are used 3.6 hours a day, suggesting that percent for children aged 0-3 and below 30 percent mothers who work full-time need to combine formal for children aged 3-5 in Costa Rica, El Salvador, Gua- care with other arrangements (Mateo Dias and Ro- temala, Honduras, and Nicaragua. The majority of driguez-Chamussy, 2016). Similarly, in Honduras and formal childcare users send their children to public Nicaragua, most households that use formal childcare facilities. In Guatemala, Honduras, Nicaragua, and El walk to the facilities. Thus, the distance to the child- Salvador, over 70 percent of children aged 0-5 attend care center appears to be a crucial element in the deci- public childcare centers. Approximately 30 percent sion to enroll kids in these programs. of users in El Salvador, Honduras, and Nicaragua use private childcare services, while Guatemala has the Chapter 4 Underlying Drivers: Gender Roles and Social Norms 100 Box 4.4: In focus: The positive effect of targeted productive transfers on intra-household bargaining A recent World Bank study analyzed a program operated by the Women and Community De- velopment Foundation ( Fundación Mujer y Desarollo Communitario or FUMDEC), a local NGO with multi-year experience in rural Nicaragua on productive transfers and gender. In 2009, the World Bank and FUMDEC partnered to launch a program the objective of which was to design, implement, and evaluate the impact of productive transfers on income generation and wom- en’s empowerment. The intervention was targeted to women in rural communities within the municipality of Santa Maria de Pantasma in the department of Jinotega, located in the north of Nicaragua. The program offered households with at least one adult female member a package of benefits that included: (i) training on community organization and gender awareness; (ii) training in technical or business skills to develop or expand small-scale household enterpris- es, livestock farming, or agricultural activities; (iii) capital transfers in the form of cash, seeds, or livestock; and (iv) follow-up technical assistance. An experimental impact evaluation design was integrated into the project in order to learn from the intervention and to inform any poten- tial scale-up as well as ongoing productive programs targeting poor rural women in Nicaragua. As a result, eligible communities were randomly assigned to treatment and control groups. The experiment was designed to evaluate the impact of the program both on economic em- powerment (including such indicators as income generation, labor market participation, and asset ownership) and on intra-household bargaining empowerment. The relationship between economic empowerment and gender equality more broadly is the question behind this exper- iment: Are initiatives that aim to improve women’s economic empowerment successful in also changing their overall empowerment (within the couple and within the communities or soci- eties in which they live)? In theory, productive interventions targeted to women have the po- tential to generate economic gains and empowerment effects that are mutually reinforcing. Giving more control over resources to women can increase their intra-household bargaining power, which may in turn alter household resource allocations and increase returns to addi- tional household investments. The experiment tested whether a productive transfer intervention providing a package of capital and skills to women in poor rural Nicaraguan communities affected their empowerment. The intervention resulted in increased employment rates and earnings for the main women beneficiaries but no increase in overall household-level employment, income, or consumption. Women’s private goods constituted a larger share of household non-food expenditures after the intervention. The intervention also increased the intra-household bargaining power of the women beneficiaries. Decision-making in the household became more participatory and spou- sal relationships also appear to improve. Source: Hatzimasoura, Premand, and Vakis (2017) Importantly, marital status has a larger impact on additional child in the household between the ages women’s labor force participation than having young of 0 and 4 decreased the mother’s likelihood of be- children. In Central America, women with young ing in the labor force by approximately 5 percent in children in the household are significantly less likely Costa Rica and Panama, and 8 percent in El Salvador to be in the labor force (Figure 4.7). In 2014, each and Nicaragua. The negative correlation was slightly Towards Equal? Women in Central America 101 Figure 4.7: Predicted change in female labor force participation based on number of young children in the household and marital status, 2014 Married Child (under 5) 0,00 - 0,05 - 0,10 Percent - 0,15 - 0,20 - 0,25 - 0,30 Costa Rica El Salvador Guatemala Honduras Nicaragua Panama Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This figure reports the conditional marginal probability of being in the labor force for women relative to a similar woman who is not married or who does not have a child under 5 in the household. The estimates are conditional on the following characteristics: age and its square, educational attainment, region, urban, current school enrollment, and the numbers of children in the household under the age of 5 and between 5 and 14. The sample is limited to adults age 18 to 65. See Annex 3.1 for logit results. smaller in Honduras and Guatemala, where partici- ilar unmarried men. This suggests that, beyond the pation fell by 2 and 4 percent for each young child. “external” barriers to women entering the labor force However, even after taking into account the presence such as the absence of childcare facilities, their labor of children in the household, along with other key fac- force participation is already reduced by the role of be- tors, being married had a larger negative effect on the ing a wife. Marriage itself, to the extent that it assigns labor force participation of women in each country. complementary yet distinct roles and responsibilities Married women in Costa Rica, El Salvador, Hondu- within the household based on gender, can be a barri- ras, and Nicaragua were about 20 percent less likely er to female labor force participation. to be in the labor force than similar women who were not married. The effect was smallest for married wom- en in Panama (16 percent) and highest for married 4.3 Differences women in Guatemala (27 percent). between groups Strict gender roles and familial and household ob- Disaggregating data by geographic, socioeconom- ligations associated with marriage reduce female labor ic, and demographic background wherever possible force participation. In line with the results present- reveals that behind the large overall gaps observed ed in Figure 4.7, Hernandez Ore, Sousa and Lopez between men and women in some crucial areas lie (2016) found that, while a married woman is less like- substantial differences between specific groups, with ly to be in the labor force than an unmarried woman more vulnerable women facing more barriers. Lit- regardless of whether she has children, the reverse is tle information is available on indigenous women, true for men. Married men both in rural and urban as noted throughout the previous chapters, the data areas are more likely to be in the labor force than sim- that are available usually indicate that this group fac- Chapter 4 Underlying Drivers: Gender Roles and Social Norms 102 es substantial and specific barriers that require adapt- viewees thought that wives need to obey their hus- ed responses. More disaggregated data and analysis band compared to 32 percent in urban areas. 77 are needed to fully understand how women from dif- The higher the education and income level of ferent backgrounds are systematically left out or have women, the less likely they are to believe that their their rights bypassed and violated in the countries of husbands should control the household and are jus- the sub-region. This analysis will be crucial to help tified in resorting to violence. In Guatemala, while a policymakers to design effective targeted responses high 73.8 percent of women without any education to improve their situation in the future (see also the agreed that the man should be the one in charge, the “intersectionality” approach put forward in Tas et share among women with tertiary education was only al., 2013). 6.5 percent. A sizeable share of those in the lowest In perception surveys, the share of respondents quintile (73.6 percent) also agreed with that statement who say that they do not experience gender equality compared to only 16.6 percent of women in the high- is much higher among the rural and the poor popu- est quintile. Similarly, 12.1 percent of women in the lation in those countries where data are available. The lowest quintile found wife beating justifiable in cer- available data indicate that traditional gender norms tain situations compared to only 1.1 percent of those may be more common among rural and/or indige- in the wealthiest quintile.78 nous women. Indeed, evidence from across the region It is important to note that change seems possible shows that agreement with the survey statement that – and improvements are happening for some women. “a good wife should obey her husband” and that “a While the trends differ somewhat between countries, man needs to show his wife that he is the boss” was two consistent characteristics are evident. As shown significantly higher among rural women than urban in Chapter 3, in general, women with more schooling women (Bott et al., 2012). The share of indigenous are substantially more likely to be in the labor force, women in Guatemala who supported the statement conditional on individual and household characteris- that “the man should demonstrate that he is the one tics. What is important to note is that this relationship in charge in the household” was almost 20 percent between education and economic participation does higher than among non-indigenous women (50 and not exist for men. In each of the six countries, the par- 60 percent). In Panama, 45.1 percent of indigenous ticipation rates of men with a primary education or women agreed that physical violence was justified for higher differ very little with higher levels of schooling one or more of the reasons provided – compared to even as there is a clear positive relationship between only 15 percent among all women.76 In Nicaragua, education and participation for women. 19.8 percent of rural women agreed that wife beating is justified for certain reasons compared to 9.4 percent In recent years, women and girls from El Salvador in urban areas, and more than half of the rural inter- and Honduras have closed the gender gap in migra- tion. Figure 4.9 shows the changes in the gender com- position of the cohorts arriving in the United States 76 The reasons given in the survey questionnaire were: (i) she burned food; (ii) she argued with her husband; (iii) she left the 77 Estimates based on Nicaragua’s Reproductive Health Survey house without informing the man; (iv) she did not properly care (Encuesta Nicaragüense de Demografía y Salud, ENDESA), for the children; (v) she lacked respect for her husband; (vi) she 2006/2007. was unfaithful to her husband; and (vii) she refused to have sex 78 Estimates based on Guatemala’s Reproductive Health Survey with him. (Encuesta Nacional de Salud Materno Infantil, ENSMI), 2008-09. Towards Equal? Women in Central America 103 Figure 4.8: Probability of being in the labor force by highest level of schooling completed relative to adults with no schooling, 2014 Costa Rica El Salvador 0,5 0,5 Predicted change in participation Predicted change in participation Men Men 0,4 0,4 Women Women 0,3 0,3 0,2 0,2 0,1 0,1 0,0 0,0 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Guatemala Honduras 0,5 0,5 Predicted change in participation Men Men 0,4 0,4 Women Women Predicted change in 0,3 0,3 participation 0,2 0,2 0,1 0,1 0,0 0,0 -0,1 -0,1 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Nicaragua Panama 0,5 0,5 Predicted change in participation Predicted change in participation Men 0,4 Men 0,4 Women 0,3 Women 0,3 0,2 0,2 0,1 0,0 0,1 -0,1 0,0 -0,2 Complete Incomplete Complete Incomplete Complete Complete Incomplete Complete Incomplete Complete Primary Secondary Secondary Superior Superior Primary Secondary Secondary Superior Superior Source: Authors’ tabulations using SEDLAC (The World Bank and CEDLAS). Note: This figure reports the conditional marginal probability of being in the labor force relative to a similar adult with no schooling. The estimates are conditional on the following characteristics: age and its square, region, urban, current school enrollment, and the numbers of children in the household under the age of 5 and between 5 and 14. The sample is limited to adults aged 18 to 65. See Annex 3.1 for logit results for women and Annex 4.1 for logit results for men. Chapter 4 Underlying Drivers: Gender Roles and Social Norms 104 Figure 4.9: Gender parity in migration to the US by year of arrival cohort 75% El Salvador Guatemala Female immigrants (%) 65% Honduras Nicaragua 55% 45% 35% 25% 1974-1980 1986-1990 1998-2000 2007-2009 2010-2012 2013-2015 Source: Author’s calculations using US Census data and the American Community Survey. from the four Central American countries with large country of birth, it is possible to estimate the change migration flows. The United States is the primary des- in the likelihood of employment associated with mi- tination for immigration from the countries of the gration among women with similar characteristics Northern Triangle and the second destination, after (specifically, the same education, age, and marital sta- Costa Rica, for immigration from Nicaragua. While tus). The results for the Northern Triangle show that early cohorts, when migration flows were smaller, had Salvadoran women who migrated are about 6 percent a higher share of women and girls than men and boys, more likely to be employed than those who did not.80 cohorts from around 2000 through to 2010 were Guatemalan women are nearly 3 percent more likely. mainly male from the countries of the Northern Tri- No difference was found between Honduran women angle. After a significant drop in the share of migrants who migrated and those who did not. who were women and girls during the 2000s, recent cohorts of migrants from El Salvador and Honduras have been approaching gender parity. Among Guate- malan emigrants, however, women and girls account for only one out of every three emigrants to the US.79 For some women, migration may mean increased access to economic opportunities. Women who have emigrated from El Salvador and Guatemala to the United States are more likely to be employed than similar women who have not migrated. Using data from household surveys in the United States and the 80 Authors’ estimates based on logit analysis of combined SEDLAC and American Community Survey data. The analysis is based on 79 There is a sizeable flow of female migrants to Mexico from data from 2014, limiting the immigrant sample to those who had Guatemala. immigrated fewer than 10 years ago and immigrated as adults. Towards Equal? Women in Central America 105 Annex 4.1 Table A .1: Estimating the effect of education on male labor force participation Costa Rica El Salvador Guatemala Honduras Nicaragua Panama Complete Primary 0.870*** 0.310*** 0.144*** 0.402*** 0.349*** 0.507*** (0.011) (0.009) (0.008) (0.009) (0.012) (0.017) Incomplete Secondary 1.029*** 0.269*** 0.078*** 0.207*** 0.589*** 0.540*** (0.011) (0.010) (0.008) (0.009) (0.010) (0.017) Complete Secondary 1.015*** 0.387*** -0.243*** 0.516*** 0.943*** 0.566*** (0.011) (0.008) (0.008) (0.011) (0.014) (0.017) Incomplete Superior 0.718*** 0.090*** 0.854*** -0.285*** 0.466*** 0.623*** (0.012) (0.010) (0.012) (0.011) (0.013) (0.019) Complete Superior 1.537*** 0.743*** 0.164*** 0.478*** 1.018*** 0.848*** (0.016) (0.015) (0.013) (0.015) (0.016) (0.020) Age 0.412*** 0.239*** 0.214*** 0.234*** 0.216*** 0.373*** (0.002) (0.001) (0.001) (0.001) (0.002) (0.002) Age-squared -0.005*** -0.003*** -0.003*** -0.003*** -0.003*** -0.005*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Urban -0.112*** -0.214*** -0.550*** -0.489*** -0.681*** -0.646*** (0.008) (0.006) (0.006) (0.007) (0.008) (0.012) Enrolled -1.777*** -2.056*** -2.376*** -1.340*** -2.129*** -2.147*** (0.007) (0.010) (0.007) (0.009) (0.010) (0.010) Married 1.072*** 1.194*** 1.329*** 1.276*** 1.720*** 1.053*** (0.008) (0.007) (0.006) (0.007) (0.008) (0.010) Children under 5 0.438*** 0.196*** 0.092*** 0.380*** -0.113*** 0.347*** (0.010) (0.007) (0.005) (0.007) (0.007) (0.011) Children 5-14 0.063*** -0.073*** 0.081*** -0.026*** 0.018*** 0.049*** (0.007) (0.005) (0.004) (0.006) (0.007) (0.009) Regional FE X X X X X X Constant -5.230*** -2.010*** -0.537*** -2.046*** -1.109*** -2.869*** (0.028) (0.024) (0.019) (0.025) (0.029) (0.056) Observations 11923 20422 13260 5811 7549 11855 Pseudo R-squared 0.272 0.247 0.234 0.229 0.274 0.284 Source: Authors’ tabulations using SEDLAC (World Bank and CEDLAS). Notes: This table reports the results of logit regression models. The dependent variable is a binary variable indicating whether the individual is in the labor force. The analysis is limited to men between the ages of 18 and 65. Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 Chapter 4 Underlying Drivers: Gender Roles and Social Norms 106 Towards Equal? Women in Central America 107 chapter 5 Taking Stock T o understand the challenges that women in Central America face in accessing social and economic opportunities, this report applies the 5.1 Overview of outcomes framework devised by the World Bank’s World De- velopment Report (WDR) 2012: Gender and De- Agency velopment (World Bank, 2012a). This framework In terms of building the legal frameworks necessary posits that progress in the inclusion of women is the for supporting the agency of women, the analysis result of gains in three dimensions—(i) agency (the ability to make choices to achieve desired outcomes, shows that countries in the region have made import- including having a voice in decision-making); (ii) en- ant strides. However, gaps in the legal framework re- dowments (outcomes related to education, health, main. Discriminatory family codes –especially those and physical assets); and (iii) economic opportunities related to early marriage – are causing four countries (outcomes pertaining to jobs, production, technolo- in Central America to lag behind other countries in gy, or market access). As the framework showcases, international comparisons. As discussed in Chapter 4, gender outcomes are not fully determined by external early marriage remains an important obstacle limit- factors such as markets and formal institutions. They ing women’s economic potential in Central America. are also significantly driven by informal institutions Similarly, none of the six countries protects women – or norms – and many are channeled through house- holds. The analysis in this study suggests that norms against being paid less than men for work of equal and aspirations are playing a determinant role in the value, and only one country bans gender discrimina- outcomes of Central American women in the social, tion in hiring. In addition, while human trafficking political, and economic spheres. is an unfortunate reality in Central America, none of 108 the countries are taking the minimum recommended However, levels of violence against women across steps to protect potential victims. Central America remain high. About one in three women reports having been exposed to physical, sex- The six countries vary significantly in terms of the ual, or both forms of violence by an intimate partner steps that they have taken to ensure gender equality in those countries for which quality data are available. and of the outcomes that women in these countries El Salvador and Honduras also have the highest rates have been able to accomplish. Guatemala, in partic- of femicide in LAC at 11.0 per 100,000 women and ular, trails behind its neighbors in several important 10.2 per 100,000 respectively. Extremely low levels of dimensions: it is the only country in Latin America convictions for all types of homicides, including fe- without any type of gender quota in political repre- micides, mean that a majority of these crimes remain sentation even though it has the fewest women in po- unpunished: only 5 percent of female homicide cas- litical positions in Central America and a lower rate es between 2005 and 2010 in El Salvador were taken than the LAC average. While it has internationally to trial. Help and support for victims of violence is recognized comprehensive femicide laws, it is the only limited. Across all countries for which information is country in the region with no laws against sexual ha- available, the share of female victims of violence who rassment in the workplace. Similarly, although Nica- sought institutional help was much lower than that of ragua is a global leader in terms of its share of repre- women who turned to family and friends. sentatives and mayors who are women, it has the most unfavorable family law in Central America (including Endowments a minimum age of marriage with parental consent at While challenges persist and Central America lags 16 years of age, the lowest among all six countries). behind the LAC region in key health and education Overall, the countries of Central America compare outcomes, there have also been undeniable gains in favorably with global outcomes when it comes to in- terms of women’s endowments. Most of the countries clusion of women in the political sphere, including of Central America were lagging behind the rest of the elected office. Following a 2012 reform, Nicaragua region in terms of women’s life expectancy in 1965, has become among the best performers in terms of yet by 2010 they had closed the gap with Panama and female political participation globally. Legislation in Costa Rica continuing to outperform the regional av- each country in Central America except Guatemala erage. Maternal mortality rates in Honduras, Guate- establishes quotas for a minimum share of women to mala, and El Salvador have fallen significantly since be included in candidates´ lists for elected positions 1990 as over 90 percent of pregnant women in each in national elections. However, at local levels, women country now receive some prenatal care. The govern- have not fared as well, with Nicaragua being the top ments in Central America have also taken import- performer when it comes to female mayors at 40 per- ant institutional steps to broaden access to maternal cent. Women have also been included in non-elected health services. positions: over half of ministerial positions in Nicara- In most measures, women in Central America are gua are held by women along with about 30 percent outperforming men in educational attainment. Girls of those in Costa Rica and Panama and 27 percent in and women are more likely to be enrolled in second- Honduras. ary and tertiary education than boys and men in all Towards Equal? Women in Central America 109 countries. Girls also have lower dropout rates in pri- nation. In each of the six countries, teenage pregnancy mary and lower secondary school than boys. The only rates are higher among rural and indigenous women, exception is Guatemala where girls and women are as well as among those with lower educational and in- less likely to finish primary or lower secondary school come levels. than boys and are less likely to be enrolled in second- ary school. Economic opportunities Despite these gains, in some areas the women of Women’s gains in legal and political agency as well as Central America continue to face challenges in acquir- access to endowments have translated into only mod- ing health and education. Four of the six countries est gains in economic inclusion. As a result of low fe- have higher maternal mortality rates than the LAC male labor force participation rates and high female average, including Panama, one of the fastest growing unemployment rates, only half of Central American economies in the region. In addition, teenage preg- women of working age are employed. The employ- nancy is particularly high in Central America, with ment rate for women surpasses 50 percent in only only Costa Rica falling below the regional average. three countries in the region, topping out at 55 per- Challenges in educational outcomes for women cent in Panama. The only country in the region where continue. The reasons for leaving school differ be- unemployment rates are not higher for women than tween men and women, with women being more for men is El Salvador. At 12 percent, female unem- likely to cite household responsibilities as their rea- ployment in Costa Rica is the highest in the region. son for leaving school. Similarly, teenage pregnancy Unemployment rates are particularly high for young leads to many girls dropping out. Some evidence also women, a factor strongly related to the high rates of suggests that lack of running water and sanitation ninis in the region. in schools discourages girls from attending school. More striking still is that more than one in four While outperforming boys in language and writing, employed women in each country works part time. girls in Central America have worse test scores than In other words, women who work full-time repre- boys in math and science. Similarly, female university sent fewer than one in three women in Guatemala students are underrepresented in STEM fields. and Honduras and fewer than two out of five women While there has been substantial progress in terms in Panama. This significant underuse of female labor of access to health and education, some of it has by- represents an important lost opportunity for the econ- passed rural and indigenous women and those from omies of the region. Labor force participation rates the poorest income groups. Indigenous women and are particularly low for women in rural areas and for rural women have worse educational and health out- women with low levels of schooling. This may be a comes than other women, reflecting a lack of access. result of the disproportionate engagement of women in labor-intensive home production activities as well The lower rates of enrollment in school for girls than as the few opportunities for employment available to those for boys in Guatemala, for example, are entirely women. driven by a large deficit in enrollment among rural girls. Among the barriers to access faced by indige- In general, women with more schooling are sub- nous women in access to health services is a lack of stantially more likely to be in the labor force. In Costa culturally appropriate options and perceived discrimi- Rica, for example, women who completed secondary Chapter 5 Taking Stock 110 school and those who completed tertiary education Aspirations and norms were 20 percent and 46 percent more likely respec- tively to be in the labor force than women with no Underlying these outcomes is the central role of so- schooling. However, the increased access to school- cial norms, which affect women’s aspirations, oppor- ing shown in the previous chapter has not necessarily tunities, and agency. Through a feedback loop pro- translated into increased labor force participation. cess, gender norms can affect what aspirations women have and hence the outcomes that they can achieve. In all dimensions of job quality, female employ- Constrained aspirations can hinder individuals from ment lags that of male employment in Central Amer- taking advantage of available opportunities. For in- ica. In El Salvador, Guatemala, Honduras, and Nica- stance, women (and men) believing that men should ragua, women are less likely to be employed as wage have priority rights to a job is negatively correlated workers and more likely to be self-employed or un- with women’s actual participation in the labor force. paid workers than men. Even among wage employed Also, high rates of agreement with men having control workers, in most countries women’s employment is over women and with the justification of wife beat- associated with lower quality microenterprises. Gen- ing are accompanied by high rates of intimate partner der segregation by sector and occupation is also wide- violence against women in Central America. Hence, spread in Central America, with commerce playing a social norms and gender roles play an important role particularly important role, being the largest sector of female employment. While there are positive signs re- in the relation between women’s agency, the opportu- garding female entrepreneurship in the region, wom- nities available to them, and the actual outcomes that en’s lack of access to formal banking and credit is a they are able to achieve. significant obstacle to female entrepreneurship and to Large shares of women in Central America agree the growth of female-owned enterprises. with statements that place husbands as authoritative Women’s monthly labor income is lower than figures over wives and as the main economic earner. men’s in every country in the region, although the gap In 2006, four in ten Nicaraguan women agreed with has been decreasing in most countries. Differences in the statements that “a wife should obey her husband” earnings are mostly explained by men working more and “a man needs to show his wife that he is the boss” hours – hence average hourly wage differences are while similar shares of Salvadoran women in 2008 smaller between men and women than monthly in- agreed that women should obey their husbands, even come differences. Even so, gaps persist. However, once if they do not agree with them. In 2006/07, around other factors are taken into account, including how 50 percent of Guatemalan women agreed with the women self-select into employment and key human statement that “the man should demonstrate that he capital and job characteristics, there is no statistical is the one in charge in the household.” In the four wage gap between men and women in urban areas in Central American countries with data on the justifica- Honduras and El Salvador. However, there are size- tion of wife-beating, more than one out of ten women able wage gaps in the other four countries. In these agreed that it was justified in some cases. On the topic countries, men’s hourly wages are, on average, 11 per- of employment, in four out of the six Central Ameri- cent to 14 percent higher than those of women with can countries, most people believe that women should similar characteristics. work only if the husband does not earn enough mon- Towards Equal? Women in Central America 111 ey. These attitudes correlate with outcomes of violence process cannot be described as “achieved” since legal against women and women’s economic inclusion. initiatives are still needed in several dimensions, for instance, the family law in Nicaragua that allows very Social norms and gender roles are strongly reflect- young marriages referenced earlier. In other cases, the ed in family formation patterns – and in the role that legal framework is either not conducive to women’s women play in the work force. Central America has well-being (for example, a lack of access to legal abor- high rates of child marriage and surprisingly stable tion even in cases where the life of the mother is at ages at first birth. More than a quarter of girls and risk) or is not being implemented and, hence, has not women between the ages of 15 and 19 in Honduras been translated into de facto law. When it comes to and Nicaragua (27 percent and 31 percent respec- the manifestations of agency, Guatemala, in particu- tively) had already married. When family formation lar, lags behind its neighbors in several important di- is initiated at very early ages, especially in childhood, mensions. Specifically, it is the only country in Latin not only are the rights of girls undermined but this America without any type of gender quota and has the also affects other individual-level development out- lowest rate of women in political positions in Central comes such as the accumulation of human endow- America, which is also lower than the LAC average. ments and women’s empowerment to make decisions. However, Nicaragua is a global leader in terms of its Overall, family formation patterns affect women’s share of female political representatives, particularly economic outcomes and, specifically, their labor force at the local level. When it comes to intimate partner participation. Furthermore, the traditional role played violence, rates are stubbornly high among all of the by women within the family mean that their work is countries under review, with roughly one-third of all largely unremunerated, leaving them economically women having suffered from physical or sexual vio- dependent on others. This, in turn, limits their agency lence at some point in their lives. and hence their human capital development. Second there are significant differences within Differences in outcomes countries. In almost all of the indicators analyzed, ru- ral women and indigenous women fare significantly It is important to emphasize that Central American worse than their urban and non-indigenous peers, in- women are not a homogenous group with homoge- cluding in those dimensions of gender equality where neous and static development outcomes. As this re- there has generally been some progress. For instance, port has shown, there are differences in women’s out- access to maternal health services and contraceptive comes and circumstances between countries, within use is especially deficient among indigenous wom- countries, and between different groups of women. en. The same applies to education. In Nicaragua for First, there are differences between countries. For instance, rural women rarely complete primary edu- instance, while Guatemala seems to struggle in essen- cation, whereas the majority of those living in urban tially all dimensions analyzed in this report, some of areas have at least an incomplete secondary education. the other countries are positive outliers – even by re- And in Guatemala, school enrollment and retention gional standards - in specific dimensions. All of the rates are lower for indigenous girls than for non-in- countries have made significant progress in the legal digenous girls and for both indigenous and non-in- and institutional preparation of a level playing field digenous boys. When it comes to female labor force between men and women. At the same time, this participation, rural women are significantly less likely Chapter 5 Taking Stock 112 to be in the labor force than urban women. Just one tral America. In the short term, there are concrete en- in three rural women in Nicaragua and Guatemala try-points across different sectors to narrow specific were in the labor force in 2014. Similarly, in each of gender gaps in those sectors and ensure women and the six countries, the gap between female and male men benefit more equally from services, assets and op- labor force participation rate is significantly higher for portunities. In the longer term, increasing the social the rural population; female participation rates range and economic inclusion of women will require fun- from one-third to just over a half of male participation damental changes in the prevailing norms about the rates. appropriate roles for women to play in society. Policies designed to do this will require an enduring commit- Third, change is happening for some women in ment and appreciation of the difficulty of effecting some parts of these countries. Norms, roles, and aspi- change. rations –key drivers of change - are more progressive among urban and among better educated women. In the short-term, there are immediately identifi- Consistently, two characteristics remain stable across able areas that require attention. For example, it is im- countries: women with more schooling are substan- portant to increase resources for support and protec- tially more likely to be in the labor force, condition- tion of victims of domestic violence, including more al on individual and household characteristics. Also, enforcement actions against perpetrators. In terms traditional gender norms seem to be more common of political participation, Guatemala could increase among less educated and rural women. For instance, rates of women in politics by following the example of the higher the education and income level of women, other countries in the region and introducing gender the less likely they are to believe that their husbands quotas. More resources for maternal health, especially should control the household and are justified in re- for underserved rural and indigenous areas, would re- sorting to violence. duce maternal mortality in the region. Training teach- ers to address gender biases in math and science might lead to higher achievement of girls and women in 5.2 Policy discussion STEM occupations. Simplifying procedures for joint Leveling the playing field between men and women is property rights and access to banking for married cou- not only a matter of social justice and fairness but is ples could lead to women having more access to assets. also smart economic and social policy. When women Passing and enforcing legislation banning gender dis- are excluded from employment opportunities, their crimination in pay and hiring is another area import- human capital is underused and undervalued. Our ant for the region. analysis shows that social norms about women’s roles There are also important knowledge gaps that need are the key driver behind the discrepancy between to be addressed in order to inform better policy. Un- women’s high rates of education but consistently low derstanding the drivers and factors that contribute to labor market outcomes. Most remarkably, in some key gender-based violence is crucial to effectively address social indicators, the role of women seems not to have this problem. More thorough research is needed on changed significantly over the past decades. how different characteristics (such as the legal or insti- The results suggest two levels of policies to increase tutional context and women’s economic conditions) the social and economic inclusion of women in Cen- combine to increase or diminish the vulnerability of Towards Equal? Women in Central America 113 women to violence or their empowerment to report man capital and economic opportunities for women it to inform the design of effective interventions. Bet- and girl can affect private household decisions, even ter understanding of local attitudes and expectations in contexts where social norms overall are changing that may condone or legitimize IPV and other forms slowly. of violence against women is also important to tar- Updating the aspirations of girls and women is get interventions successfully. At the same time, more important so that education can be a catalyst for in- research into discrimination of women in the work- creased gender equality. World Bank (2012b) argues place and in access to credit is needed. The limited that social norms favoring inequality may persist due information available on women in entrepreneurship to misinformation. Following that assumption, pro- suggests that women often engage in informal small viding more information is the key to shifting norms. businesses out of necessity, which may indicate the As a growing share of women in Central America take existence of discrimination or additional constraints. on the same roles and take advantage of the same op- While a lack of access to credit is the key constraint portunities that are available to men, it is important to more female entrepreneurship, in rural areas, a lack to ensure that other women and girls get exposure and of access to land may also be a factor limiting female opportunities to participate in these changed realities. self-employment. Together with targeted efforts to increase women’s en- The second set of policies, those targeted towards dowments, opportunities, and voice, being exposed to the underlying root causes of gender inequality, par- these changed realities can lead women to have greater ticularly gender norms, require long term and multi- aspirations for themselves and are likely to improve faceted approaches. First, it is worth noting that broad outcomes for both men and women beyond those drivers of change, such as economic development or pockets where change has happened up until now. increase in access to communications technology, may Building on work that has already been done (La Fer- themselves impact gender norms. For instance, Segui- rara et al., 2012 and Jensen and Oster, 2009), it will no (2007) finds that, overall, economic development be important going forward to identify what proactive is associated with a shift towards more egalitarian gen- steps can be taken to change gender norms more rap- der norms, although there are exceptions such as in idly in the countries of Central America. high-income Gulf states. That is, continuing to un- Media can also be a very effective entry-point and dertake policies necessary for economic development a driver of change. Social norms marketing is one can itself lead to more egalitarian attitudes towards measure within the scope of the media, an effort in gender. which messages are produced and disseminated that Similarly, exposure to greater economic opportu- carry information or attempt to change perceptions nities can provide women with broader networks and of social norms, of attitudes and behaviors that are enhance their sources of information and support considered acceptable or normal. Social norms mar- (World Bank 2012b, Casabonne et al. 2015). Social keting tries to both activate positive social norms and norms are often shaped in one’s personal context. to discourage certain negative attitudes by communi- Therefore, exposure to alternative ideas and practic- cating the idea that those are not considered typical or es through exposure to role models may be effective. desirable (Paluck and Ball 2010). For instance, Jensen The World Bank (2012b) argues that expanding hu- and Oster (2007) find the introduction of cable tele- Chapter 5 Taking Stock 114 vision in rural India associated with increases in wom- ners in order to change gender norms. For instance, en’s autonomy (such as their ability to go out without in Brazil, The Promundo-led `Program H’ combined permission), with an increase in their participation group education sessions (including role-plays, discus- in household decision making and with a decrease in sion and individual reflection) with a social marketing the acceptability of domestic violence. Beyond those campaign in order to encourage young men to reflect changes in attitude, increases in girls’ school enrol- on how they act as men. In India, a modified version ment as well as decreased fertility were also observed. of the intervention led to a significant reduction in Relatedly, Ferrera et al. (2012) explore the role of tele- support for inequitable gender norms. vision soap operas on fertility patterns in Brazil. The portrayal of a middle and upper-middle class model of the family with none or few children in the soap operas of Rede Globo seems to have stimulated a pref- erence for fewer children and contributed to shifting norms around ideal (smaller) family size. Edutainment is a related approach in which pop- ular characters that the target group can identify with deliver the messages. Paluck and Ball (2010) discuss findings from three evaluations of respective programs (in South Africa, Nicaragua and Brazil) and conclude that awareness messages should be accompanied by strong injunctive norms messages communicating that an influential or relevant social group does not approve of the behavior. Local social engagement is also important. Based on a thorough review of respective evidence, Mar- cus et al. (2014) argue that enhancing mass media programs with discussion groups at the community level can turn into particularly effective transmission channels for new values and norms, especially in ru- ral sites (Marcus et al, 2014). In addition, Bicchieri (2015) suggests that the role of the reference group is a very important driver behind individuals’ motiva- tion to adopt new practices. Following this argument, community based norm change interventions seem promising. Finally, women’s inclusion cannot be treated as only an issue about women. Pulerwitz et al. (2010) emphasize the importance of engaging men as part- Towards Equal? Women in Central America 115 116 References Adelman, M. and M. Szekely. 2016. “School Dropout representation: Evidence from Indian village in Central America: An Overview of Trends, councils.” India Policy Forum 7(1): 159-201. Causes, Consequences, and Promising Inter- Appadurai, A. 2004. “The Capacity to Aspire: Culture ventions.” World Bank Policy Research Work- and the Terms of Recognition.” Pp. 59-84 in ing Paper 7561. Washington D.C.: World Culture and Public Action, edited by V. Rao and Bank. M. Walton. Palo Alto, CA: Stanford University Ahmed, S., Q. Li, L. Liu, and A. Tsui. 2012. “Maternal Press. deaths averted by contraceptive use: an analysis of 172 countries.” The Lancet 380(9837): 111- Arpino, B., C. Pronzato, and L. Tavares. 2010. “All 125. in the Family: Informal Childcare and Moth- ers’ Labour Market Participation.” Institute for Almeida, R. and H. Oosterbeek. 2017. “Water and Economic and Social Research Working Paper sanitation services in Schools and student out- 2010-24. Essex, United Kingdom. comes: Evidence for Panama.” Mimeo. Wash- ington D.C.: World Bank. Azevedo, J. P., M. Favara, S. E. Haddock, L. F. Lo- pez-Calva, M. Muller, and E. Perova. 2012. Amin, M., I. Asif Mohammed, and S. Alena. 2016. “Does paternity leave matter for female em- Teenage pregnancy and opportunities in Latin ployment in developing economies? evidence America and the Caribbean on early child bear- from firm data.” World Bank Policy Research ing, poverty, and economic achievement. Wash- Working Paper 7588. Washington D.C.: ington D.C.: World Bank. World Bank. Berniell, M. I. and C. Sánchez-Páramo. 2011 “Over- Beaman, L., E. Duflo, R. Pande, and P. Topalova. view of Time Use Data Used for the Analysis 2010. “Political reservation and substantive of Gender Differences in Time Use Patterns.” Towards Equal? Women in Central America 117 Background paper for the World Development in Developing Countries?” Economic Develop- Report 2012. Washington D.C.: World Bank.   ment and Cultural Change 45(2), 259–280. Bicchieri, C. 2006. The Grammar of Society: The Na- Calvo-Gonzalez, O. and J. H. Lopez. 2015. El Sal- ture and Dynamics of Social Norms. Cambridge/ vador - Systematic Country Diagnostic: Building New York: Cambridge University Press. on Strengths for a New Generation. Washington, ———. 2015. Norms in the Wild: How to Diagnose, D.C.: World Bank Group.  Measure and Change Social Norms. New York: Carbone-Lopez, K., C. Kruttschnitt, and R. Mac- Cambridge University Press. millan. 2006. “Patterns of Intimate Partner Bick, A. 2015. “The Quantitative Role of Child Care Violence and their Associations with Physical for Female Labor Force Participation and Fer- Health, Psychological Distress, and Substance tility.” Journal of the European Economic Associ- Use.” Public Health Reports 121(4): 382. ation 14(3): 639-668. Casabonne, U., M. Jimenez, F. Bexi, and M. Muller. Blau, D. M., and J. Currie. 2006. “Pre-School, Day 2015. Roads to Agency: Effects of Enhancing Care and After-School Care: Who’s Minding Women’s Participation in Rural Roads Projects the Kids?” Handbook of the Economics of Edu- on Women’s Agency - a Comparative Assessment cation, Vol. 2, edited by E. A. Hanushek and F. of Rural Transport Projects in Argentina, Nicara- Welch. Amsterdam: Netherlands. gua, and Peru. Washington D.C.: World Bank.   Boserup, E. 1970. Woman’s Role in Economic Develop- ment. New York: St. Martin’s Press. Cater, S., and L. Coleman. 2006. “ ‘Planned’ Teenage Pregnancy: Views and Experiences of Young Bott, S., A. Guedes, M. Goodwin, and J. A. Mendoza People from Poor and Disadvantaged Back- (2012). Violence Against Women in Latin Amer- grounds.” Joseph Rowntree Foundation. Bris- ica and the Caribbean: A Comparative Analysis tol: Policy Press. https://www.jrf.org.uk/report/ of Population-based Data from 12 Countries. planned-teenage-pregnancy-views-and-experi- Washington, D.C.: Pan American Health Or- ganization.  ences-young-people-poor-and-disadvantaged. Bourdieu, P. 1977. Outline of a Theory of Practice (Vol. CEDAW. 2006. “Combined Fourth, Fifth and Sixth 16). Cambridge, U.K.: Cambridge University report for States parties: Honduras.” Com- Press. mittee on the Elimination of Discrimination against Women, United Nations. CEDAW/C/ Busso, M., and D. Romero Fonseca. 2015. “Female HON/6. Labor Force Participation in Latin America: Patterns and Explanations.” Working Paper 18, ———. 2011. “Concluding Observations of the Centro de Estudios Distributivos, Laborales y Committee on the Elimination of Discrimina- Sociales. University of la Plata, Argentina. tion against Women: Costa Rica.” Committee Buvinic, M., and G.R. Gupta. 1997. “Female-headed on the Elimination of Discrimination against Households and Female-maintained Families: Women, United Nations. CEDAW/C/CRI/ Are They Worth Targeting to Reduce Poverty CO/5-6. References 118 ———. 2017. “Concluding Observations on the lima/---sro-san_jose/documents/genericdocu- Combined Eighth and Ninth Periodic Reports ment/wcms_224386.pdf of El Salvador.” Committee on the Elimination Cunningham, W., and P. Villaseñor. 2014. “Employer of Discrimination Against Women, United Voices, Employer Demands, and Implications Nations. CEDAW/C/SLV/CO/8-9. for Public Skills Development Policy.” World Charmes, J. 2015.  “Time Use Across the World: Bank Policy Research Working Paper 6853. Findings of a World Compilation of Time Use Washington D.C.: World Bank.   Surveys.” Background Paper UNDP Human De Hoyos, R. E., M. Bussolo, and O. Núñez. 2012. Development Report. New York: United Na- “Exports, Gender Wage Gaps and Poverty in tions. Honduras.” Oxford Development Studies 40(4): Chattopadhyay, R. and E. Duflo. 2004. “Women as 2012-11-06. Policy Makers: Evidence From A Randomized De Hoyos, R., H. Rogers, and M. Székely. 2016. Out Policy Experiment In India.” Econometrica of School and Out of Work: Risk and Opportuni- 72(5): 1409–1443. ties for Latin America’s Ninis. Washington D.C.: Chioda, L. 2016. Work and Family: Latin American World Bank.   and Caribbean Women in Search of a New Bal- Deere, C. D. and M. León de Leal (2001). Empower- ance. Latin American Development Forum Se- ing Women: Land and Property Rights in Latin ries. Washington D.C.: World Bank.   America. Pittsburgh, Pa.: University of Pitts- Clark, S., J. Bruce, and A. Dude. 2006. “Protect- burgh Press. ing young women from HIV/AIDS: The case Demirguc-Kunt, A., L. Klapper, D. Singer, P. Van against child and adolescent marriage.” Inter- Oudheusden. 2015. “The Global Findex Da- national Family Planning Perspectives, 32(2), tabase 2014: Measuring Financial Inclusion 79–88. Around the World.” Policy Research Working Colverson, K. 1995. “Rural Women’s Access to Agri- Paper 7255. Washington, D.C.: World Bank. cultural Information: A Participatory Study of Deutsch, R., A. Morrison, H. Ñopo, and C. Piras. Two Honduran Communities.” Proceedings of 2005. “Working Within Confines: Occupa- the 11th Annual Meeting of the Association for tional Segregation By Sex For Three Latin International Agriculture and Extension Edu- American Countries.” The IUP Journal of Ap- cation, Little Rock, Arkansas.  plied Economics 0(3): 50-59. CONAMYPE and ISDEMU (2013). EFAMEDU: Dewalt, B. R., K. M. Dewalt, and M. A. Fordham. Marco Integrado de Evaluación de un Entorno 1985. “Socioeconomic Constraints to the Pro- Favorable para el Desarrollo de las Mujeres Em- duction, Distribution and Consumption of presarias en El Salvador. Report by the National Sorghum in Southern Honduras.” Lexington: Commission for Micro and Small Firms (CO- University of Kentucky.  NAMYPE) and the Institute for the Advance- ment of Women (ISDEMU). http://www.ilo. Dimova, R., and F. C. Wolff. 2011. “Do Downward org/wcmsp5/groups/public/---americas/---ro- Private Transfers Enhance Maternal Labor Sup- Towards Equal? Women in Central America 119 ply? Evidence from Around Europe.” Journal of FUSADES (Fundación Salvadoreña para el Desarrol- Population Economics 24 (3): 911–33. lo Económico y Social). 2014. “Evaluación de Durand, J. 1975. The Labor Force in Economic Devel- Impacto del Programa de Apoyo Temporal al opment: A Comparison of International Census Ingreso (PATI)”. Data 1946-66. Princeton: Princeton University ———. 2015. “Complemento Cualitativo: Evalu- Press. ación de Impacto del Programa PATI”. ECLAC/OPS/UNFPA. 2010. Salud Materno In- Gimenez. L., S. Van Wie, M. Muller, R. Schutte, M. fantíl de pueblos indígenas y afrodescendientes Rounseville and M. Viveros Mendoza. 2015. de América Latina: Aportes para una re-lectura Enhancing Youth Skills and Economic Opportu- desde el derecho a la integridad cultural. Santia- nities to Reduce Teenage Pregnancy in Colombia. go de Chile: Economic Commission for Latin Washington D.C.: World Bank.   America and the Caribbean. Goldin, C. 1995. “The U-Shaped Female Labor Force Ellsberg, M. and L. Heise. 2005. Researching Violence Function in Economic Development and Eco- Against Women: A Practical Guide for Researchers nomic History.” Pp. 61-90 in Investment in and Activists. Washington D.C.: World Health Women’s Human Capital and Economic Devel- Organization, PATH. opment, edited by T. P. Schultz. Chicago: Uni- Elster, J. 1983. Sour Grapes: Studies in the Subversion of versity of Chicago Press. Rationality. Cambridge: Cambridge University Press. Goldin, C. 2006. “The Quiet Revolution that Trans- formed Women’s Employment, Education, Fernández, R., and A. Fogli. 2009. “Culture: An Em- and Family.” American Economic Review 96 (2): pirical Investigation of Beliefs, Work, and Fer- 1–21. tility.” American Economic Journal: Macroeco- nomics 1(1): 146–77. Grimes, D.A, J. Benson, S. Singh, M. Romero, B. Ga- natra, F. E. Okonofua, and I. H. Shah. 2006. Ferreira, F. H. G., J. Messina, J. Rigolini, L. F. “Unsafe Abortion: The Preventable Pandemic.” López-Calva, M. A. Lugo, and R. Vakis. 2013. The Lancet 368(9550): 1908-1919. Economic Mobility and the Rise of the Latin American Middle Class. Washington D.C.: Hallman, K., S. Peracca, J. Catino, and M.J. Ruiz. World Bank.   2006. “Multiple Disadvantages of Mayan Fe- Field, E. and A. Ambrus. 2008. “Early Marriage, Age males: The effects of Gender, Ethnicity, Pover- of Menarche, and Female Schooling Attain- ty, and Residence on Education in Guatemala.” ment in Bangladesh.” Journal of Political Econo- Research Working Paper No. 211, Population my 116 (5): 881–930.  Council, New York. Fortin, N. M. 2005. “Gender Role Attitudes and the Hatzimasoura, C., P. Premand, and R. Vakis. 2017. Labour Market Outcomes of Women Across “Productive Transfers, Intra-household Bar- OECD Countries.” Oxford Review of Economic gaining and Empowerment Evidence from a Policy 21(3): 416–38. Randomized Trial in Nicaragua.” World Bank References 120 Working Paper, Washington D.C.: World Kabay, S. 2016. “Grade Repetition and Primary Bank.   School Dropout in Uganda.” Harvard Educa- tional Review 86(4): 580-606. Hernandez Ore, M. A., L. D. Sousa, and J. H. Lopez. 2017. Honduras - Unlocking economic potential Kearney, M. and P. Levine. 2011. “Income Inequality for greater opportunities: systematic country diag- and Early, Non-marital Childbearing: An Eco- nostic. Washington D.C.: World Bank.   nomic Exploration of the ‘Culture of Despair,’” NBER Working Paper 17157.  Hong, L. M., T. D. Tran, H. T. Nguyen, and J. Fish- er. 2014. “Early Marriage and Intimate Partner Kleinman, A. 2006. What really matters: living a moral Violence among Adolescents and Young Adults life amidst uncertainty and danger. New York: in Viet Nam.” Journal of Interpersonal Violence Oxford University Press. 29(5): 889-910. Klugman, J. 2017. “Gender Based Violence and the IDEA International and IADB. 2010. Political Parties Law.” Background paper for the World De- and Parity: The Pending Equation. Washington, velopment Report 2017. Washington D.C.: D.C.: Inter-American Development Bank. World Bank.   Interpeace. 2013. Violentas y Violentadas. Relaciones Klugman, J., Hanmer, L., Twigg, S. Hasan, T, Mc- de Género en las Maras de Salvatrucha y Bar- cleary-Sills, J. & Santamaria, J. 2014. Voice and rio 18 en el triángulo Norte de Centro-Améri- Agency: Empowering Women and Girls for Shared ca.  Guatemala City: Interpeace. https://www. Prosperity. Washington D.C.: World Bank.   interpeace.org/latinoamerica/wp-content/up- Kruger, D. and M. Berthelon. 2012. “Household loads/sites/7/2015/08/2013_05_14_ Central_ Choices of Child Labor and Schooling: A sim- Am_Violentas_y_Violentadas_es.pdf ple Model with Application to Brazil.” Journal of Human Resources 47(1): 1-31.  Iyer, L., A. Mani, P. Mishra, and P. Topalova. 2012. “The Power of Political Voice: Women’s Politi- La Ferrara, E., A. Chong, and S. Duryea. 2012. “Soap cal Representation and Crime in India.” Amer- Operas and Fertility: Evidence from Brazil.” ican Economic Journal: Applied Economics 4(4): American Economic Journal 44: 1-31 165–93. Leaper, C., T. Farkas, and C. S. Brown. 2012. “Ad- Jensen, R. and E. Oster. 2009. “The Power of TV: olescent Girls’ Experience and Gender-Relat- Cable Television and Women’s Status in India.” ed Beliefs in Relation to their Motivation in The Quarterly Journal of Economics 124 (3): Math/Science and English.” Journal of Youth 1057-94. and Adolescence 41(3): 268-282. Jubb, N., G. Camacho, A. D’Angelo, G. Yáñez De Liu, C., A. Esteve, and R. Treviño. 2017. “Fe- la Borda, K. Hernández, I. Macassi León, C. male-Headed Households and Living Condi- MacDowell Santos, Y. Molina, and W. Pasina- tions in Latin America.” World Development to. 2008. “Regional Mapping Study of Wom- 90:311-328. en’s Police Stations in Latin America.” Centro Macours, K. and R. Vakis. 2014. “Changing House- de Planificación y Estudios Sociales, Quito. holds’ Investments and Aspirations through Towards Equal? Women in Central America 121 Social Interactions: Evidence from a Random- Muñoz Boudet, A., P. Petesch, C. Turk, and A. Thum- ized Transfer Program.” The Economic Journal ala. 2012. On norms and agency—conversations 124: 607–633.  about gender equality with women and men in 20 Countries. Washington D.C.: World Bank.   Macours, K. and R. Vakis. 2016. “Sustaining Impacts When Transfers End: Women Leaders, Aspi- Musalo, K. and B. Bookey. 2014.  “Crimes Without rations, and Investment in Children.” NBER Punishment: An Update on Violence Against Working Paper No. 22871. Women and Impunity in Guatemala.” Social Justice 40(4): 106-117.  Mammen, K., and C. Paxson. 2000. “Women’s Work and Economic Development.” The Journal of Muzira, S. and D. Hernandez de Diaz. 2013. Re- Economic Perspectives 14(4): 141-64. thinking Infrastructure Delivery: Case Study of a Green Inclusive and Cost-Effective Road Program Marcus, R., E. Page, R. Calder, and C. Foley. 2014. in Nicaragua. Washington D.C.: World Bank.   Drivers of Change in Gender Norms: An Anno- tated Bibliography. London: Overseas Develop- Näslund-Hadley, E. and G. Binstock. 2010. The Mise- ment Institute. ducation of Latin American Girls: Poor Schooling Makes Pregnancy a Rational Choice. Washing- Mateo Díaz, M. and L. Rodriguez-Chamussy. 2016. ton, D.C.: Inter-American Development Bank.  Cashing in on Education: Women, Childcare, and Prosperity in Latin America and the Ca- Nussbaum, M. C. 2000. Women and Human Devel- ribbean. Latin American Development Forum opment. The Capabilities Approach. New York: Series. Washington, D.C.: World Bank and In- Cambridge University Press. ter-American Development Bank.  OECD (Organisation for Economic Co-operation Meier, K. J. and J. Nicholson-Crotty. 2006. “Gender, and Development). 2010. Atlas of Gender and Representative Bureaucracy, and Law Enforce- Development: How Social Norms Affect Gen- ment: The Case of Sexual Assault.” Public Ad- der Equality in Non-OECD Countries.  Paris: ministrative Review 66(6): 850-860.  OECD.  Milosavljevic, V. 2007. “Estadísticas Para la Equi- ———. 2014. Social Institutions and Gender Index dad de Género: Magnitudes y Tendencias en 2014: Synthesis Report. Paris: OECD.  América Latina.” Santiago de Chile: Economic ———. 2015. “Costa Rica Student Performance PISA Commission for Latin America and the Carib- 2015.” Retrieved November 1, 2017 (http:// bean. gpseducation.oecd.org/CountryProfile?prima- MIX. 2014. Microfinance Information Exchange. ryCountry=CRI&treshold=10&topic=PI)  Country Profile Nicaragua. Access online Palermo T., J. Bleck, A. Peterman. 2014. “Tip of the at:https://www.themix.org/mixmarket/coun- Iceberg: Reporting and Gender-based Violence tries-regions/nicaragua in Developing Countries.” American Journal of Monteiro, M. G. 2007. Alcohol and Public Health Epidemiology 179(5):602-12.   in the Americas. A case for action. Washington Paluck, E. L. and E. Ball, with C. Poynton and S. Sied- D.C.: Pan-American Health Organization. loff. 2010. “Social Norms Marketing Aimed at References 122 Gender-Based Violence: A Literature Review Participation.” IZA Discussion Paper 6398, and Critical Assessment.” New York: IRC. Bonn: Institute for the Study of Labor. Panizza, U., and C. Z. Qiang. 2005. “Public–Private Psacharopoulos G. and Z. Tzannatos. 1989. “Female Wage Differential and Gender Gap in Latin Labor Force Participation: An Internation- America: Spoiled Bureaucrats and Exploited al Perspective.” World Bank Research Observer Women?” Journal of Socio-Economics 34: 810- 4(2):187–202. 833. Pulerwitz, J., A. Michaelis, R. Verma and E. Weiss. Pearce, D. 1978. “The Feminization of Poverty: Wom- 2010. “Addressing Gender Dynamics and En- en, Work and Welfare.” The Urban and Social gaging Men in HIV Programs. Lessons Learnt Change Review 11(1-2): 28-36. from Horizons Research.” Public Health Reports 125: 282-292. Pezzini, S. 2005. “The Effect of Women’s Rights on Women’s Welfare: Evidence from a Natural Ex- Quintana, M. E. and M. Chica. 2012. Situación del periment.” Economic Journal 115 (502): C208– Personal de las Casas Maternas para la Detección C227 y Atención de Mujeres y Adolescentes Víctimas de Violencia Sexual. Managua: Ipas Centroamérica.  Pines, A.M, Lerner, M. and Schwartz, D. 2010. “Gen- der differences in Entrepreneurship: Equality, Sagot, M. and L. Guzmán. 2004. Informe Final de Diversity and Inclusion in Times of Global Investigación del Proyecto Encuesta Nacional de Crisis.” Equality, Diversity and Inclusion: An In- Violencia Contra las Mujeres. San Jose, Costa Rica: Centro de Investigación en Estudios de la ternational Journal 29(2):186-198. Mujer, Universidad de Costa Rica. Piscopo, J.M. 2015. “States as Gender Equality Ac- Samandari, G., and I. S. Speizer. 2010. “Adolescent tivists: The Evolution of Quota Laws in Latin Sexual Behavior and Reproductive Outcomes America.” Latin American Politics and Society In Central America: Trends over the Past Two 57(3): 27-49. Decades.” International Perspectives on Sexual Plotnick, R. 1992. “The effects of Attitudes on Teen- and Reproductive Health 36(1): 26–35. age Premarital Pregnancy and its Resolution.” Sanchez, S., K. Scott, and J. H. Lopez. 2016. Gua- American Sociological Review 57(6): 800-811.  temala - Closing gaps to generate more inclusive Plotnick, R. 1993.  “The Effect of Social Policies on growth: systematic country diagnostic. Washing- Teenage Pregnancy and Childbearing Families ton, DC.: World Bank. in Society.”  The Contemporary Journal of Hu- Sedgh, G., S. Singh, I. H. Shah, E. Åhman, S. K. man Services 74(6): 324-328.  Henshaw, and A. Bankole. 2012. “Induced Plotnick, R. 2007. “Adolescent Expectations and De- abortion: incidence and trends worldwide sires about Marriage and Parenthood.” Journal from 1995 to 2008.” The Lancet 379(9816): of Adolescence 30: 943-63. 625-632. Posadas, J., and M. Vidal-Fernández. 2012. “Grand- Secretaría de Estado para el Desarollo de los Pueblos parents’ Childcare and Female Labor Force Indígenas y Afrohondureños (SEDINAFRO), Towards Equal? Women in Central America 123 UNCHR, UNWomen, UNDP, Cooperación UNDP (United Nations Development Programme). Suiza. Agenda Política de las Mujeres Indígenas y 2011. Sustainability and Equity: A Better Future Afro-hondureñas: Caminando juntas, del diálogo for All. New York: United Nations. a la propuesta, 2012-2013. http://americalati- UNESCO (United Nations Educational, Scientific nagenera.org/newsite/images/Agenda_Politi- and Cultural Organization). 2016. Informe de ca_de_las_Mujeres_Indigenas_y_Afro-hondu- Resultados: Tercer Estudio Regional Comparativo renas_FINAL_29ENE2014_copy.pdf y Explicativo. Santiago: OREALC/UNESCO. Seguino, S. 2007. “Plus Ça Change? Evidence on UNFPA (United Nations Population Fund). 2012a. Global Trends in Gender Norms and Stereo- “Final Country Programme Document for types.” Feminist Economics 13(2): 1-28. Nicaragua.” DP/FPA/CPD/NIC/8. New York: Sen, A. 1999. Development as Freedom. New York: Al- United Nations. fred A. Knopf. ——2012b. Marrying too young - End child marriage. Solotaroff, J.L., and R.P. Pande. 2014. Violence against New York, NY: UNFPA. Women and Girls: Lessons from South Asia. UNICEF (United Nations Children’s Fund). 2014. Washington, D.C.: World Bank. Ending child marriage: Progress and prospects. Stockl, H., K. Devries, A. Rotstein, N. Abrahams, J. UNICEF: New York, NY. Campbell, C. Watts, and C. Garcia Moreno. United Nations. 2006. “Ending Violence Against 2013. “The global prevalence of intimate part- Women: From Words to Action.” A/61/122/ ner homicide: A systematic review.” The Lancet Add.1 and Corr.1, Study of the Secretary-Gen- 382(9895): 859-865. eral. Tas, E.O., M. Reimao, and M. Orlando. 2013: “Gen- ——. 2011. “Report of the Special Rapporteur on der, Ethnicity and Cumulative Disadvantage the Right of Everyone to the Enjoyment of the in Education: Evidence from Latin American Highest Attainable Standard of Physical and and African Censuses.” World Bank Policy Mental Health, Mission to Guatemala.” A/ Research Working Paper 6734. Washington HRC/17/25/Add.2, Human Rights Council D.C.: World Bank. 17th Session. Tenjo, J., R.R. Medina, and L.F. Bernat. 2004. “Evo- ——. 2014a. Guidelines for Producing Statistics on lution of Salary Differences between Men and Violence against Women—Statistical Surveys. Women in Six Latin American Countries.” Pp. Department of Economic and Social Affairs 139-185 in Women at Work:  Challenges for Statistics Division. New York: United Nations.  Latin America edited by C. Piras. Washington, ——. 2014b. “National Report Submitted in Accor- D.C.: Inter-American Development Bank. dance with Paragraph 5 of the Annex to Hu- Thomas, D. (1990). “Intra-Household Resource Allo- man Rights Council Resolution 16/21: Nic- cation: An Inferential Approach” The Journal of aragua.” A/HRC/WG.6/19/NIC/1. Human Human Resources 25(4): 635-664. Rights Council 19th Session.   References 124 ——. 2015. The World’s Women 2015 Trends and Sta- Population Division. Geneva: World Health tistics. Department of Economic and Social Af- Organization. fairs. New York: United Nations. ———.2017. “Preventing Unsafe Abortion: Fact USDS (United States Department of State). 2015. Sheet.” Retrieved June 30, 2017 (http://apps. “Guatemala Country Report on Human Rights who.int/mediacentre/factsheets/fs388/en/in- Practices for 2014.” Bureau of Democracy, Hu- dex.html).  man Rights and Labor. U.S. Department of WHO and PAHO (World Health Organization and State, Washington D.C. Pan American Health Organization). ‎2012‎. ———. 2016. “Nicaragua Country Report on Hu- Understanding and Addressing Violence Against man Rights Practices for 2015.” Bureau of De- Women: Femicide. World Health Organization. mocracy, Human Rights and Labor, U.S. De- Wodon, Q., C. Male, A. Nayihouba, A. Onagoru- partment of State, Washington D.C. wa, A. Savadogo, A. Yedan, J. Edmeades, A. ———. 2017. Trafficking in Persons Report 2017. Kes, N. John, L. Murithi, M. Steinhaus and S. Washington, D.C.: U.S. Department of State. Petroni. 2017. Economic impacts of child mar- https://www.state.gov/j/tip/rls/tiprpt/2017/ riage: Global synthesis report. Washington, DC: World Bank. Verick, S. 2014. “Female Labor Force Participation in Developing Countries.” IZA World of Labor World Bank. 2010. Women’s Economic Opportunities 2014: 87 doi: 10.15185/izawol.87.  in the Formal Private Sector in Latin America and the Caribbean: A Focus on Entrepreneurship. Vogelstein, R. 2013. Ending Child Marriage How El- Washington D.C.: World Bank.   evating the Status of Girls Advances U.S. Foreign Policy Objectives. Council on Foreign Relations ———. 2012a. Women’s Economic Empowerment in (CFR). New York Latin America and the Caribbean: Policy Lessons from the World Bank Gender Action Plan. Wash- WHO (World Health Organization). 2003. Putting ington D.C.: World Bank.   Women First: Ethical and Safety Recommenda- tions for Research on Domestic Violence Against ———. 2012b. World Development Report 2012: Women. Geneva: World Health Organization.  Gender Equality and Development. Washington D.C.: World Bank.   ———. 2011. Unsafe Abortion: Global and Regional Estimates of the Incidence of Unsafe Abortion and ———. 2014. Voice and Agency – Empowering Wom- Associated Mortality in 2008. Geneva: World en and Girls for Shared Prosperity. Washington Health Organization.  D.C.: World Bank.   ———.2014. “Panama Health Profile.” Retrieved ———. 2015a. Indigenous Latin America in the November 1, 2017 (http://www.who.int/gho/ Twenty-First Century. Washington D.C.: World countries/pan.pdf?ua=1).  Bank.   ———.2015. Trends in maternal mortality: 1990 to ———. 2015b. Women, Business and the Law 2016: 2015. Estimates by WHO, UNICEF, UNFPA, Getting to Equal. Washington, D.C.: World World Bank Group and the United Nations Bank. Towards Equal? Women in Central America 125 ———. 2015c. Women in Agriculture: The Impact of Male Out-Migration on Women’s Agency, House- hold Welfare, and Agricultural Productivity. Washington D.C.: World Bank.   ———. 2015d. World Development Report 2015: Mind, Society, and Behavior. Washington, D.C.: World Bank. ———. 2016a. Central America Community-Based Approaches to IPV: The Case for Community Mo- bilization Interventions to Prevent Intimate Part- ner Violence. Washington D.C.: World Bank.   ———. 2016b. Locking in success – Panama SCD. Washington D.C.: World Bank.   ———. 2016c. “Program-For-Results Information Document: Improving Access and Quality of Health Care for the Rural Population.” Wash- ington D.C.: World Bank.   ———. 2017a. “Implementation Completion and Results Report: Honduras Second Land Ad- ministration Project.” (IDA – 46410) Wash- ington D.C.: World Bank.   ———. 2017b. “Project Appraisal Document: Alli- ance for Education Quality Project.” Washing- ton D.C.: World Bank.   ———. 2018. “Unseen contributions: Women’s in- visible work in rural Honduras” Washington D.C.: World Bank. References