MALE OUTMIGRATION AND WOMEN’S WORK AND EMPOWERMENT IN AGRICULTURE The Case of Nepal and Senegal June 2018 MALE OUTMIGRATION AND WOMEN’S WORK AND EMPOWERMENT IN AGRICULTURE The Case of Nepal and Senegal June 2018 © 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 findings, 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 dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial 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. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons. org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work,including for commercial purposes, under the following conditions: Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. Images: Smallholder female farmer in Nepal (Left): Poverty Alleviation Fund II Project, Government of Nepal. Small holder female farmers Ramata Niass and Faty Penda Niasse (Right): Daniella Van Leggelo-Padilla / World Bank CONTENTS Acknowledgments  v Abbreviations and Acronyms vii Forewordix Executive Summary xi Chapter One: Introduction 1 Chapter Two: Country Context: Nepal and Senegal 5 Nepal’s agriculture sector and migration background 5 Senegal’s agriculture sector and migration background 6 Chapter Three: Survey Methodology 9 Survey locations 9 Survey instruments  9 Samples  12 Chapter Four: Characteristics of Migration in the Surveyed Areas 15 Characteristics of Nepali migrants 15 Characteristics of Senegalese migrants  17 Chapter Five: Individual Characteristics of Women Who Stay Behind 19 Nepal19 Senegal19 Country comparison 20 Chapter Six: Employment Characteristics of Women Who Stay Behind 21 Nepal21 Senegal22 Chapter Seven: Characteristics of Women’s Empowerment  25 Chapter Eight: Food Security Characteristics 27 Chapter Nine: Empirical Strategy 29 Chapter Ten: Results 33 Linkages between male outmigration and women’s employment  33 Associations with women’s empowerment  34 Chapter Eleven: Conclusions  37 Chapter Twelve: Policy Recommendations 39 Generalized policy recommendation  39 Country-specific policy recommendations 40 ANNEXES The Abbreviated Women’s Empowerment in Agriculture Index Annex A:  (A-WEAI) Used in Nepal and Senegal Surveys 43 Annex B: Descriptive Analysis of Key Variables 47 Annex C: Association Between Food Insecurity Experience Scale and Migration Status 53 Annex D: Regression of Interest (Employment Outcomes) 55 Annex E: Regression of Interest (Empowerment Outcomes) 59 Male Outmigration and Women’s Work and Empowerment in Agriculture iii Annex F: Regression of Interest (Addressing Endogeneity) 63 References67 TABLES Table 1: C  haracteristics of international migrants versus nonmigrants, working-age individuals (age 16+), Nepal 16 Table 2: C  haracteristics of international and internal migrants versus nonmigrants, working-age individuals (age 16+), Senegal17 Table A1: D  omains and indicators from the Abbreviated Women’s Empowerment in Agriculture Index (A-WEAI) used in Nepal and Senegal surveys 43 Table A2: Empowerment outcomes by sex in Nepal 44 Table A3: Empowerment outcomes by sex in Senegal 45 Table B1: Characteristics of female family members, Nepal 47 Table B2: Characteristics of female family members, Senegal 48 Table B3: E  mployment characteristics by international migration experience for all working-age adults and for working-age women only, Nepal 49 Table B4: E  mployment characteristics by migration status for all working-age adults and for working-age women, Senegal50 Table B5: Women’s empowerment outcomes by migration status, Nepal 51 Table B6: Women’s empowerment outcomes by migration status, Senegal 52 Table C1: T  he correlation between migration status, remittances, and household food insecurity, Nepal 53 Table C2: T  he correlation between migration status, remittances, and household food insecurity, Senegal 54 Table D1: The impact of migration on employment outcomes for women, Nepal 56 Table D2: The impact of migration on employment outcomes for women, Senegal 57 Table E1: T  he association between migration (with and without remittances) and the empowerment of women, Nepal, OLS 60 Table E2: The impacts of migration and remittances on the empowerment of women, Senegal, OLS 61 Table F1: The impact of migration on types of work for women, Nepal, 2SLS 64 Table F2: The impact of migration on types of work for women, Senegal, 2SLS 65 FIGURES Figure 1: R  easons for migrating abroad, Nepal (Left: international, Right: internal) 2 Figure 2: Map of Nepal with the sampled districts 10 Figure 3: Map of Senegal with the sampled regions 11 Figure 4: Use of remittances, Nepal 16 Figure 5: Use of remittances, Senegal 18 Figure 6: W  omen’s employment outcomes by household migration status, Nepal 22 Figure 7: W  omen’s employment outcomes by household migration status, Senegal 23 Figure 8: The prevalence of food insecurity based on FIES, Nepal 27 Figure 9: The prevalence of food insecurity based on FIES, Senegal 28 iv The Case of Nepal and Senegal ACKNOWLEDGMENTS This Discussion Paper was produced jointly by the The team thanks Flore Martinant de Preneuf and World Bank Group and the Food and Agriculture Tammy Mehdi for their assistance with the communi- Organization of the United Nations (FAO). The cations aspects, and to Venkat Ramachandran, Pawan discussion paper was led by Anuja Kar (World Bank Sachdeva, Hien Minh Vu and Beulah Noble for their Group, Team Leader) in collaboration with Vanya administrative support. The team thanks Amy Gau- Slavchevska (FAO) and Susan Kaaria (FAO) under tam for excellent support with editing this paper. The the excellent guidance of Louise Scura (Practice Senegal Map is sourced from the World Bank Agri- Manager, World Bank Group). The team greatly culture Observatory, prepared by Caroline Sar-torato thanks Susan Kaaria for her guidance, supervision Silva Franca. The team is extremely grateful to Nepa and management of FAO-related activities for this School of Social Science and Humanities and a team discussion paper. The team received very helpful led by Dr. Sudhindra Sharma at the Inter Disciplinary inputs from Erdgin Mane (FAO), Riccardo Ciacci Analysts for their diligent and highly efficient work (FAO), and Yurie Tanimichi Hoberg. in implementing Nepal survey and National Statistics Office of Senegal for implementing Senegal Survey. The team is grateful to the peer reviewers: Holger A. Kray, Dilip Ratha, Mio Takada, and Aphichoke Koti- The team is thankful to colleagues Sanjiva Cook, kula, for their tremendous guidance. Dorte Verner, Michael Morris, Mellissa Williams, Natasha Hayward, Nichola Dyer, Izabela Leao, Aira The team extends thanks to Sanna Lisa Taivalmaa for Maria Htenas, Ioannis Vasileiou, and Patricia Van de laying out the foundation of this paper. Velde for their tremendous support and inputs to future areas of research. Special thanks to team of advisers: Robert Townsend, Victoria Stanley, Agnes Quisumbing (IFPRI) for their We wish to extend our gratitude to senior manage- guidance which greatly improved the manuscript. ment from the World Bank Agriculture Global The team also gratefully acknowledges Madhur Gau- tam, Federica Marzo, Dhushyanth Raju for extremely Practice for their guidance and support during this pro- helpful suggestions. cess, including Juergen Voegele, Simeon K. Ehui, Mar- tien van Nieuwkoop, Louise Scura, Kathryn Hol-lifield, The team is grateful for the tremendous support Marianne Grosclaude, Preeti Ahuja, and Rob Townsend. and encouragement from the colleagues from Nepal Country Team (Mio Takada, Karishma Wasti, Also, sincere thanks and acknowledgement to senior Purna Bahadur Chhetri, Chris Jackson, Sanjay Sriv- management in the Gender GSG for their tremen- astava, and Omar Lyasse,) and Senegal Country dous support, especially Caren Grown and Lucia Team (Aifa Fatimata Niane Ndoye, El Hadj Adama Hanmer. This paper is dedicated to the thousands of Toure, Feder-ica Marzo, Paolo B. Zacchia, and women farmers in Nepal and Senegal, who served as Sophie Naudeau). the central motivation. Authors: Anuja Kar (World Bank Group), Vanya Slavchevska* (FAO), Susan Kaaria (FAO), Sanna Lisa Taivalmaa, Erdgin Mane (FAO), Riccardo Ciacci (FAO), Yurie Tanimichi Hoberg (World Bank Group), Robert Townsend (World Bank Group), and Victoria Stanley (World Bank Group). * The author is currently with the International Center for Tropical Agriculture (CIAT). Male Outmigration and Women’s Work and Empowerment in Agriculture v ABBREVIATIONS AND ACRONYMS A-WEAI Abbreviated Women’s Empowerment in M&E Monitoring and Evaluation Agriculture Index NGO(s) Non-Governmental Organization(s) FAO Food and Agriculture Organization of the Nepal Nepal Living Standards Measurement United Nations LSMS Survey FIES Food Insecurity Experience Scale OLS Ordinary Least Squares GDP Gross domestic product TLU Tropical Livestock Units IFPRI International Food Policy Research WEAI Women’s Empowerment in Agriculture Institute Index Male Outmigration and Women’s Work and Empowerment in Agriculture vii FOREWORD The advent of internal and international migration of 2. Empowerment: Male outmigration is linked to empow- people is not new but migration and its consequences erment in some domains and disempowerment in have turned into a pressing item on the development others. In Nepal, receipt of remittances is positively agenda in recent years. The number of international associated with increased female decision-making on migrants reached 266 million globally in 2017, driven the farm, greater group membership, and their holding both by economic and non-economic factors. a financial account. However, in the absence of remit- tances, spouses of international migrants are worse off International and internal migration is predomi- in several domains of empowerment, including deci- nantly male which raises questions on what happens sion making on productive activities and agricultural to the women who stay behind. Rural women have income, and access to information. always worked, but the additional roles they assume 3. Food security: Migration of household members who increases their paid and unpaid work and caring do not send remittances is likely to increase household roles. In this context, understanding the impact of food insecurity. The evidence is stronger and significant migration on labor market outcomes for women, in the case of Senegal, where both international and empowerment of women, and food security will be internal migration are positively associated with food important to guide domestic policy. Data from two insecurity. comparable surveys for Nepal and Senegal collected between August and November 2017, were used to Recommendations that emerge from this study study these three effects. include reducing remittances costs, supporting women’s engagement in higher-earning activities, The analysis shows that: and providing tailored extension services to female farmers. 1. Labor market outcomes: Male outmigration is asso- ciated with significant changes in women’s roles in Our hope is that this initial study on this topic will agriculture, where for example in the case of Nepal, pave the way for further work and policy dialogue so women move from contributing family workers to self- that the women who stay in rural areas become posi- employed workers on the farm. The employment out- tive agents of change who can lead their families and comes become stronger if accompanied by remittances. larger communities toward great development gains. Juergen Voegele Caren Grown Senior Director Senior Director Food and Agriculture Global Practice Gender Group The World Bank Group The World Bank Group Male Outmigration and Women’s Work and Empowerment in Agriculture ix EXECUTIVE SUMMARY Migration is important in the development agenda1 In the absence of their migrant husbands, women and is closely connected with agriculture in many coun- may increase their roles in decision-making around tries. Limited available evidence suggests that across a range of household and farm activities, partly the globe the migration originating from rural areas because remote monitoring of rural households and is predominantly male (Mueller et al. 2015), which agriculture activities can only be done imperfectly. could potentially lead to significant socioeconomic At the same time, the migration of spouses may changes in rural areas, including changes in tradi- lead to higher work burden and stress, which may tional gender norms. Yet limited rigorous evidence disempower women. These consequences of migra- exists on the direct impact of male outmigration on tion have only been explored in small-scale, mostly women’s work within and outside of agriculture, with qualitative, studies. To the authors’ knowledge, the even less evidence on its consequences on intrahouse- only study that provides a detailed account, includ- hold decision-making and women’s empowerment. ing quantitative analyses, establishing the linkages This is due to the fact that most existing survey data between migration and women’s empowerment in include information on either migration or women’s agriculture is the work done by Stanley (2015) for empowerment but rarely on both aspects together. Guatemala. Stanley (2015) points out that despite migration, women who stay behind continue to farm Migration affects women’s work and empower- even though farming is traditionally seen as men’s ment mainly through the loss of migrants’ labor and work in Guatemala. Women have to overcome vari- through the flow of remittances. In response to the ous constraints, including the challenge of hiring and absent migrant labor, women may be required to managing male labor, but they do see an improve- increase their labor allocation on the family farm to ment in their decision-making power. keep agricultural production at the same level. (Alter- natively, migrant households may change or reduce agricultural production.) Remittances have a separate OBJECTIVES OF THIS REPORT effect on women’s labor supply: they may raise wom- The objective of this study is to examine the linkages en’s reservation wages, resulting in reduced time in between migration and women’s work and empow- remunerated employment; or they may relax growth erment in agriculture in Nepal and Senegal. In par- constraints for family farming, making family farming ticular, this analysis seeks to understand: (i) how more attractive than other paid or unpaid activities. outmigration influences women’s work in agriculture; These hypotheses have been tested in various studies, (ii) the consequences of male-dominated migration however, there has been little attention to the types of on gender roles and women’s empowerment; and paid and unpaid work performed by women. (iii) whether and how outmigration impacts house- hold food security. The fact that migration may alter intrahousehold decision-making processes has been understudied. The study tested several hypotheses: 1 An initial identification carried out by the United Nations shows all 1. Employment: whether women in households with a Sustainable Development Goals (SDGs) and targets are directly relevant migrant reduce participation in income-generating to migrants and migration: http://www.un.org/en/development/desa/ activities, controlling for the individual characteristics population/migration/events/coordination/14/documents/back- of the women, household characteristics, and regional grounddocs/GMPA_14CM.pdf dummies. Male Outmigration and Women’s Work and Empowerment in Agriculture xi 2. Types of Employment: whether the migration of a (male) family member is linked to changes in the types EMPIRICAL FINDINGS of work women do—for example, whether women increase employment in nonfarm activities and reduce WOMEN’S EMPLOYMENT participation in farm activities. The study finds that in Nepal male outmigration 3. Empowerment: whether women in migrant households from rural, primarily agricultural areas is not linked are significantly more likely than women in nonmigrant to a decrease in women’s employment, but it is associ- households to experience improvements in empower- ated with significant changes in women’s roles in agri- ment, as measured by several indicators based on data culture. The study finds no evidence that living in a collected through the Abbreviated Women’s Empower- migrant-sending household causes women to reduce ment in Agriculture Index (A-WEAI). 4. Remittances: whether the effects differ if the migrant overall participation in income-generating activities. households receive remittances or not. In Nepal, male outmigration from rural, primarily 5. Food Insecurity: whether migration is associated with agricultural areas is strongly and significantly linked changes in the food insecurity status of the household, to changes in women’s roles in agriculture—women where food insecurity is measured with the Food Inse- shift from being contributing family members to curity Experience Scale (FIES), and whether the link being self-employed on the farm. These changes are between migration and food insecurity is mediated by stronger when migration is accompanied by remit- the receipt of remittances. tances. Contrary to some previous studies, the report DATA AND RESEARCH does not find evidence that women in households with a family member who is currently abroad reduce METHODS their engagement in off-farm wage employment and Using data from two comparable surveys for Nepal off-farm self-employment. On the other hand, in and Senegal collected between August and Novem- Senegal male-dominated outmigration is not associ- ber 2017, this study assesses the effects of male out- ated with changes in women’s roles in agriculture. migration from rural, primarily agricultural areas This is because most rural women in Senegal live in on women’s work and empowerment in agriculture large extended families in which other members may and in the household. These innovative surveys take on the roles and responsibilities of the migrant were designed to capture detailed individual-level spouse (Marzo and Atuesta 2018). information on both nonmigrant members of rural households and all current and return emigrants. WOMEN’S EMPOWERMENT They also included comprehensive modules on The study reveals that male-dominated outmigration crop production, livestock rearing, social protec- is not always associated with women’s empowerment. tion, and employment outcomes of all household Based on evidence from the A-WEAI, male outmi- members. In addition to this household question- gration is linked to empowerment in some domains naire, which was administered to the most knowl- and disempowerment in others. These results differ edgeable person in the household, one individual substantially by country. In Nepal, direct interviews from each household (either the spouse of the with spouses of migrants reveal that the receipt of migrant or the man or woman from the primary remittances is positively associated with increased couple) was separately interviewed about his or decision-making on the farm, group membership, her own empowerment status using the Abbrevi- and holding a financial account. In Senegal, with the ated Women’s Empowerment in Agriculture Index exception of decisions regarding credit, there is no (A-WEAI) questionnaire (Malapit et al. 2015; Alkire evidence that male outmigration leads to women’s et al. 2013). The surveys also collected information empowerment. Moreover, in the absence of remit- on the food security status of the households using tances, spouses of international migrants are worse the Food Insecurity Experience Scale (FIES) devel- off in several domains of empowerment, including oped by FAO’s Voices of the Hungry Project (Bal- decision-making on productive activities and agricul- lard, Kepple, and Cafiero 2013). tural income, and access to information. xii The Case of Nepal and Senegal HOUSEHOLD FOOD SECURITY remittances can influence significant changes in wom- The consequences of migration on household food en’s roles in agriculture and are positively associated with women’s empowerment in several domains (such security are country-specific and mediated by the as decisions on farm, group membership, and holding receipt of remittances. The study finds that migra- a financial account for Nepal, and access to decisions tion of household members that is not followed by about credit for Senegal). One way to facilitate remit- remittance transfers is likely to increase household tance transfers would be to reduce the cost of sending food insecurity. The evidence is stronger and signifi- remittances. Sustainable Development Goal (SDG) 10 cant in the case of Senegal, where both international aims to reduce the cost of remittances to three per- cent by 2030 and eliminate remittance corridors with and internal migration are positively associated with costs higher than five percent. This will be an avenue to food insecurity. In Nepal, no significant correlation formalize remittances channels. One key constraint in exists between migration and food security, but the Nepal, especially in the mountain and hill areas, is the lack of significant results may be due to the rather lack of access to financial services. small survey sample size. iii. Enact policies to support women’s engagement in higher- earning activities. A smaller share of women in Senegal GENERALIZED POLICY than in Nepal report being economically active. There is a need to better understand women’s low participation RECOMMENDATION in the labor market in Senegal, but besides that, women who are economically active are largely concentrated in A more generalized and priority policy action emerg- the production end of agricultural value chains. Very ing out of the analysis suggests the importance of rec- few women in either Nepal or Senegal engage in pro- ognizing the changing roles of women in agriculture, cessing or trade of agricultural products. and providing targeted interventions to support their roles. General policy actions are to: COUNTRY-SPECIFIC POLICY i. Encourage greater availability of gender-relevant, sex- RECOMMENDATIONS disaggregated data to monitor the effects of male out- A set of policy recommendations was derived for migration on women’s work and empowerment. The current practice of collecting and disseminating sex- each country. Each set addresses the country-specific disaggregated data is done in a scattered manner across challenges identified in this study. different agencies. To identify tailored knowledge gaps and policies targeted specifically to women left behind after the outmigration of a male spouse, it is extremely NEPAL important to improve the availability of evidence-based, The following approaches appear promising in targeted surveys and to centralize the survey packages addressing the problems identified by the study: for future research and policy dialogues. It is also impor- tant to build national capacity to collect and analyze sex-desegregated data covering migrant-sending and Adapting Agricultural Extension nonmigrant households in agriculture. This is a system- i. Provide tailored extension services to female farmers. atic pathway of providing policy makers with sufficient The study finds that as a result of male outmigration in baseline information to institute favorable changes to Nepal, the on-farm responsibilities and decision-making existing policies, which currently affect women and of the women left behind increase. In Nepal, all migra- men differently in migrant households. This will also tion is linked to a change in women’s roles in agricul- form the basis of institutionalizing such rigorous evi- ture from being a contributing family worker to being dence to strengthen existing and future World Bank self-employed in agriculture, and the effect is larger operations or multi-stakeholder programs that are tar- for women who live in households with international geted at women engaged in on-farm activities, where migrants who send remittances. This clearly indicates monitoring and evaluation (M&E) systems are often the need for improving female farmers’ access to exten- less comprehensive in terms of capturing progress on sion services to increase productivity on their farms and women’s empowerment in different domains. ensure the sustainability of agricultural production. ii. Facilitate the flow of international and internal remit- ii. Strengthen women’s access to higher-earning activities tances. Evidence from the case studies indicates that in agricultural value chains. The study shows very low Male Outmigration and Women’s Work and Empowerment in Agriculture xiii engagement in higher value chain activities such as pro- empowerment in Senegal. That said, the impor- cessing and trading, which can be linked to women’s tant role of remittances in mediating the effects of low skills, lack of access to market information, and migration on women’s empowerment is evident in transportation and time constraints. Extension services Senegal as well. for women should go beyond the traditional focus on production and should provide technical assistance, training, and access to resources that can scale up wom- The following approaches appear promising in en’s involvement beyond subsistence agriculture and in addressing the problems identified by the study: the higher-value nodes of the supply chains. iii. Ensure that a gender-sensitive approach is adopted for Reducing the Cost of Remittances the provision of agricultural extension services, includ- i. Reduce the cost of remittances to positively affect dis- ing through hiring more female agricultural exten- posable household income and improve incentives to sion agents. Studies have shown positive experiences remit more (World Bank 2005). The cost of sending with hiring female extension agents to better support remittances through formal channels is very high in female farmers (Acharya and Bennet 1983; World Bank Senegal, a situation accompanied by a high gender dis- 2010) and the importance of local groups for mobiliz- parity in the receipt of remittances: male-headed house- ing public awareness to mainstream gender balance holds receive higher remittances than female-headed in agriculture extension. A concerted involvement of ones (Orozco et al. 2010). Positive remittances will also decentralized government bodies, nongovernmental help mitigate the negative effects from the lost labor of organizations (NGOs), private agencies, and individu- migrants and therefore will help mitigate the negative als can create an enabling environment. effects on women’s empowerment. ii. Conduct more research to understand the factors behind the low economic activity status of women in Senegal. Addressing Labor Shortages A very small share of women in Senegal report hav- i. Promote small-scale rural mechanization to reduce ing engaged in any work activity in the last 12 months. women’s time burden and improve diversification of Although women in migrant households have even income-generating activities in Nepal (Biggs and Justice lower employment rates than women in nonmigrant 2015). As suggested by the results, women in migrant households, the analysis suggests that the lower employ- households in Nepal are more overworked and time- ment probability is not attributed to migration but to constrained compared to both men and women in non- other factors, which may also be correlated with migra- migrant households. This may be due to the scarcity of tion, including household demographics. The presence agricultural labor and low access to labor-saving tech- of larger extended families may facilitate migration but nologies for Nepalese women. may also mediate the potential transformative effects of migration on spouses who stay behind. Therefore, in an environment with low employment rates for both men Improving Enabling Environment for and women and large extended families, the migration Productive Use of Remittances by Female of male family members is less likely to lead to significant Farmers changes in women’s employment and empowerment, as i. Reduce the cost of remittances to create an enabling other family members can step in to do the work of the environment for women to mobilize remittances for migrant man or to make decisions in his absence. productive purposes, including more investments in agriculture or small businesses and savings through IMPLICATIONS FOR FUTURE development of money management skills (Dhakal and Maharjan 2018). In certain areas of Nepal the cost of RESEARCH remittances is quite high. Currently, at least some of The outmigration linkages for rural women left the remittances are used for the purchase of food, but behind to participate in agriculture can vary widely a non-negligible amount is also invested in agriculture. across countries, depending on the socioeconomic environment, cultural norms, migration type, and SENEGAL the influence of cross-cutting areas such as climate The study finds no significant association between change and fragility. For example, migration can male outmigration and women’s employment and be caused by economic as well as crisis factors. This xiv The Case of Nepal and Senegal study mainly highlights the association between It is essential to understand all of the dimensions male outmigration due to economic reasons and discussed above to identify the observed and unob- women’s employment and empowerment in rural served factors that impact employment and empow- areas. The area of crisis-led migration (e.g., migra- erment outcomes. This is beyond the scope of this tion caused by political upheaval, disaster, security, stand-alone quantitative research and must be com- or other push factors) requires expanding country plemented with qualitative research (such as focus coverage. Similarly, issues of migration status and group discussions with survey respondents) to bet- spell duration play a critical role in affecting the out- ter understand the results of data analysis and the comes of employment and empowerment. Also, the narrative of their behaviors. That can be the key characteristics of international and internal migra- ingredients to the provision of robust policy recom- tion differ in many ways, which deserve additional mendations. Future research, including research field research and analysis. using the data collected for this study, will have to address these dimensions. The issue of women’s empowerment requires explo- ration beyond the A-WEAI, which remains heavily The linkages between migration, agriculture, wom- focused on agriculture. Future research needs to en’s empowerment and food security are very com- expand on additional dimensions that are impor- plex and deserve more attention. Male outmigration tant to understand the situation of women as well as is associated with changes in women’s roles in agricul- migration dynamics. Similarly, the overarching and ture (in some contexts) and it is also likely associated complex notion of related social norms and custom- with changes in the agricultural sector overall. Future ary and legal frameworks may dictate employment research should continue in-depth exploration of the as well as empowerment outcomes in developing effects of male outmigration on agricultural produc- countries (e.g., forthcoming research by Marzo and tion, productivity, and food security and how the Atuesta (2018) outlines some implications for labor effects are mediated by the changes in women’s roles market outcomes and productivity). in agriculture. Male Outmigration and Women’s Work and Empowerment in Agriculture xv CHAPTER ONE INTRODUCTION Attention to the implications of rural outmigration is growing, but little evidence exists on its association with women in agriculture. In 2017, there were 266 mil- lion2 international migrants, up from 220 million in 2010 and 173 million in 2000 (UN DESA 2017). Internal or domestic migration, generally from rural to urban and peri-urban areas, is an even larger phenomenon—in 2005, there were 763 million internal migrants worldwide (UN DESA 2013). Most migra- tion flows originate from rural areas, which raises concerns about their conse- quences on rural communities. The limited available evidence suggests that across the globe, migration originating from rural areas is predominantly male (Mueller et al. 2015).3 Hence, this type of migration could lead to significant socioeconomic changes in rural areas, including changes in traditional gender norms. While in a great number of developing countries women’s share of the agricultural labor force (relative to that of men) increased significantly over the past few decades, including in response to male outmigration (Slavchevska, Kaaria, and Taivalmaa 2016), there is limited rigorous evidence on the direct impacts of male outmigration on women’s work in and outside of agriculture, and even less evidence on its consequences for intrahousehold decision-making and women’s empowerment. These gaps in the literature are largely attributed to limited data, as most existing surveys focus on either migration or women’s empowerment but rarely on both issues (with the exception of Stanley’s 2015 small-scale study of migration and women’s agency in Guatemala). Migration affects women’s work and empowerment mainly through the loss of migrants’ labor and through the flow of remittances. In response to the absent migrant labor, women may increase their labor allocation to the family farm to keep agricultural production at the same level. (Alternatively, migrant 2 KNOMAD database https://www.knomad.org/data/migration/immigration 3 The sex composition of migration varies significantly by region, and even by country within the same region. The composition is also expected to change over time, with initially male-dominated patterns followed by more gender-balanced emigration trends later on. However, data and statistics on internal migration, particu- larly on rural outmigration, are extremely scant. Male Outmigration and Women’s Work and Empowerment in Agriculture 1 FIGURE 1. R  EASONS FOR MIGRATING ABROAD, NEPAL (LEFT: INTERNATIONAL, RIGHT: INTERNAL) Educ tion Educ tion To look for b tt r job To look for b tt r job Assi nm nt/ Assi nm nt/ mplo m nt opprtunit mplo m nt opprtunit Joinin spous /m rri Joinin spous /m rri D th of spous /p rtn r D th of spous /p rtn r F mil probl ms F mil probl ms Joinin oth r m mb rs Joinin oth r m mb rs of th hous hold of th hous hold R turn to pr vious r sid nc R turn to pr vious r sid nc Non-poss ssion or Non-poss ssion or in d qu t cultiv bl l nd in d qu t cultiv bl l nd Poor/d r d d l nd Poor/d r d d l nd H lth probl ms H lth probl ms In d qu t cc ss to In d qu t cc ss to soci l prot ction b n fits soci l prot ction b n fits Educ tion of childr n Educ tion of childr n 0 100 200 300 400 0 20 40 60 R son to mi r t bro d (counts) R son to mi r t insid N p l (counts) households may change or reduce agricultural pro- In the absence of their migrant husbands, women duction.) Remittances have a separate effect on may increase their roles in decision-making around women’s labor supply—they may raise women’s res- a range of household and farm activities, partly ervation wages, resulting in reduced time in remu- because remote monitoring of rural household and nerated employment; or they may relax growth agriculture activities can only be done imperfectly. constraints for family farming, making family farm- ing more attractive than other paid or unpaid activi- The fact that migration may alter women’s intra- ties. These hypotheses have been tested in various household decision-making processes has received studies, though with little attention to the types of limited coverage and attention. The only study that paid and unpaid work performed by women.4 Much provides a detailed account of the linkages between less attention has been paid to the fact that migration migration and women’s empowerment in agriculture also alters intrahousehold decision-making processes. is the work done by Stanley (2015) for Guatemala. The 2015 study pointed out that despite migration, women who stay behind continue to farm even 4 See Funkhouser (1992) for Nicaragua; Rodriguez and Tiongson (2001) though farming is traditionally seen as men’s work for the Philippines; Amuedo-Dorantes and Pozo (2006) for Mexico; in Guatemala. Women must overcome various con- Binzel and Assaad (2011) for Egypt; Mu and van de Walle (2011) for China; Mendola and Carletto (2012) for Albania; and Lokshin and Glin- straints, including the challenge of hiring and man- skaya (2009) and Phadera (2016) for Nepal. Using data from the 2010-11 aging male labor, but they do see an improvement in Nepal Living Standard Survey, Phadera (2016) examined the effects of their decision-making power. migration on the participation and hours spent in self-employment and wage employment of both men and women who stay behind. The study It is also important to distinguish between the vari- found that migration led to women relocating time from wage employ- ment to self-employment, where self-employment largely consisted of ous aspects of empowerment. Autonomy in deci- subsistence farming. Similar studies for Senegal could not be identified. sion-making is only one aspect of empowerment. In 2 The Case of Nepal and Senegal their study of migration and women’s autonomy in that collected detailed information on all types of Mozambique, based on data for 2000-2006, Yabiku, outmigration from rural areas in Nepal and Senegal. Agadjanian, and Sevoyan (2010) found that both suc- Detailed information was also collected on women’s cessful and unsuccessful cases of male outmigration5 and men’s work in sending communities and wom- are linked to significantly higher autonomy for wives en’s empowerment in agriculture using the Abbre- who stay behind, and the gains in autonomy persist viated Women’s Empowerment in Agriculture Index after husbands return. At the same time, although (A-WEAI). In addition, the survey inquired about unsuccessful migration increases women’s autonomy, households’ food insecurity using the Food and Agri- it may have disempowering effects on women. Unsuc- culture Organization’s (FAO) recently developed cessful migration itself can be a strain on women’s Food Insecurity Experience Scale (FIES). time, as they have to assume the work of their migrant husbands and deal with the financial difficulties that The objective of this study is to examine the linkages accompany unsuccessful migration experiences. between migration and women’s work and empower- ment in agriculture in Nepal and Senegal. In particular, The complex issue of rural outmigration also has this analysis seeks to understand: (i) how outmigration implications for household food security. First, fam- influences women’s work in agriculture; (ii) the conse- ily members who stay behind may struggle to com- quences of male-dominated migration on gender roles pensate for the lost income from the migrant labor. and women’s empowerment; and (iii) whether and Second, remittances may have a separate effect on how outmigration impacts household food security. household food security. Empirical studies generally The study tested several hypotheses: find a positive relationship between migration and food security, largely attributed to remittances (Zezza i. Employment: whether women in households with a et al. 2011). Third, several studies raise the issue of migrant reduce participation in income-generating changing agricultural practices, which may nega- activities, controlling for the individual characteristics tively affect food security. Small-scale studies from of the women, household characteristics, and regional Nepal suggest that at least in some regions women indicators. who stay behind and take over the farm management ii. Types of Employment: whether the migration of a (male) family member is linked to changes in the types adopt less labor-intensive crops, shorten cropping of work women do—for example, whether women cycles, reduce the diversity of crops they grow, and increase employment in nonfarm activities and reduce even abandon agricultural land (Paudel, Tamang, participation in farm activities. and Shrestha 2014; Tamang, Paudel, and Shrestha iii. Empowerment: whether women in migrant households 2014). A more standardized approach and a compa- are significantly more likely than women in nonmigrant rable indicator are required, which is applicable in a households to experience improvements in empower- cross-country analysis. ment, as measured by several indicators based on data collected through the A-WEAI. iv. Remittances: whether the effects differ if migrant To address these existing knowledge gaps in a frame- households receive remittances or not. work that combines gender, migration, and food secu- v. Food Insecurity: whether migration is associated with rity, this study exploits a rich, comprehensive survey changes in the food insecurity status of the household, where food insecurity is measured with the Food Inse- 5 In this study, migration of a family member that is not accompanied by curity Experience Scale (FIES), and whether the link the receipt of remittances is considered unsuccessful. In turn, migrants between migration and food insecurity is mediated by who send remittances back home are deemed successful. the receipt of remittances. Male Outmigration and Women’s Work and Empowerment in Agriculture 3 CHAPTER TWO COUNTRY CONTEXT: NEPAL AND SENEGAL Primary data were collected in Nepal and Senegal to explore the linkages between migration and (i) changes in women’s work and empowerment in agri- culture, and (ii) household food security. These two countries were selected for several reasons. First, in both countries outmigration from rural areas is high and dominated by men, leading to potentially significant changes in intra- household labor allocations and decision-making. The consequences on the women who stay behind will be affected by the specific drivers of migration and whether the migration is successful (i.e., whether there are remittance transfers). However, as mentioned earlier, the limited data preclude rigorous exploration of the issue. Second, a review of the literature provided evidence of women’s high and growing participation in agriculture in these countries, especially relative to that of men (Slavchevska, Kaaria, and Taivalmaa 2016). Yet robust evidence linking the change to migration is limited. Third, the available statistics only capture women’s growing visibility in the agriculture sector but do not provide enough details about the types of activities women engage in, whether changes are linked to women’s higher economic empowerment, and whether any adverse effects on household food security are incurred. Finally, in both countries, rural areas continue to be heavily dependent on agriculture, which directs attention to the consequences of migration and the potential intrahousehold changes in labor and decision-making on agricultural produc- tion and food security. NEPAL’S AGRICULTURE SECTOR AND MIGRATION BACKGROUND Agriculture is the main sector of employment for most Nepali men and women, but it has become much more important for women. Agriculture is the back- bone of Nepal’s economy. Agricultural work is the primary activity for almost 66 percent of working-age women (over 15 years old) compared to 53 per- cent of working-age men. The inability of subsistence agriculture to provide for basic household needs (Maharjan, Bauer, and Knerr 2012) has pushed many households in Nepal to diversify their income-generating activities into Male Outmigration and Women’s Work and Empowerment in Agriculture 5 off-farm employment, including engaging in interna- security. In the past, economic reasons generated tional migration. According to World Bank (2015), significant internal migration, but today most Nepali at 29.2 percent Nepal has one of the highest shares migrants search for better economic opportunities in of remittances in gross domestic product (GDP). international destinations, rather than in urban cent- Remittances from international migration have also ers at home. India remains an important destination been linked to huge gains in poverty reduction in the for migrants (35 percent of international migrants country. Almost one-fifth of the country’s poverty from Nepal go to India) but has been surpassed by reduction between 1995 and 2004 is attributable to Malaysia and the Gulf countries, which receive more migrant remittances (Lokshin, Bontch‐Osmolovski, than 60 percent of international Nepali migrants. In and Glinskaya 2010). the study areas, there are no clear patterns of current hills–Terai migration; internal migration today is pri- Men dominate international migration. Ninety-seven marily to Kathmandu. While the top motivations for percent of Nepali migrants are men aged 15-44 (Lok- internal migration are education, jobs, and employ- shin and Glinskaya 2009) who leave women behind to ment, family reasons, such as marriage and joining take care of the household (Gartaula, Niehof, and Vis- family members, also play a role (Figure 1). ser 2010). Male outmigration, scarce off-farm employ- ment opportunities—especially for women—and Both international and internal migration costs in biased gender norms are largely behind the growing Nepal are principally financed through savings. For role and visibility of women in agriculture in Nepal around 56 percent of international migrants and 75 (Allendorf 2007; Gartaula, Niehof, and Visser 2010; percent of internal migrants, savings are the most Lokshin and Glinskaya 2009; Maharjan, Bauer, and important source for paying migration costs. Loans Knerr 2012; Tamang, Paudel, and Shrestha 2014). from lenders are the second most important source of financing migration (for around 20 percent of Patterns of migration have evolved over the years. both internal and international migrants). A few Historically, the majority of internal migration (80 migrants also list contributions or loans from rela- percent) was from the hills toward the Terai,6 a trend tives as important financial sources. reportedly started largely after the 1950s. The Terai has played an important role as a receiving region, but this role is now being challenged by the ever-increas- SENEGAL’S AGRICULTURE ing outmigration from the Terai. Since the 1990s, SECTOR AND MIGRATION outmigration from both the hills and the Terai has BACKGROUND exhibited an increasing trend, and today the Terai is a In Senegal as in Nepal, agriculture is more important major migrant-sending area. Unlike historical migra- for women than for men. In 2017, according to the tion where whole families would relocate in search World Bank, 59 percent of women’s employment is in of better economic opportunities, as was the case for agriculture compared to 49 percent of men’s employ- hills–Terai migration, the current migration is largely ment. As in Nepal, the agriculture sector in Senegal characterized by individual migration, whereby one is a key employer for most of the population—even or more family members migrate to urban centers or though agriculture constitutes only around 17 percent abroad for a few years and then return. of the country’s GDP. In some higher-value agricul- ture sectors such as horticulture, women dominate the The main drivers of migration are unemployment labor share (Maertens and Swinnen 2009). Although and low agricultural income, as subsistence agricul- women seem to be concentrated in low-skill, labor- ture is often unable to ensure households’ financial intensive tasks, with the few managerial positions typi- cally filled by men, some positive effects on women’s 6 Lowland region in southern Nepal (Shrestha and Bhandari 2007; Gar- empowerment arise from their higher involvement in taula and Niehof 2013). paid wage employment (Maertens and Swinnen 2012). 6 The Case of Nepal and Senegal Migration in Senegal is an important livelihood diver- The key destinations of Senegalese economic sification strategy. Both men and women participate migrants are France, certain Francophone African in migration but are often motivated by different rea- countries, and Dakar. Within Europe, France is the sons. Chort, De Vreyer, and Zuber (2017) analyzed most prevalent destination for both current and gendered patterns of internal migration in Senegal past international migrants. Gabon, Mauritania, The using panel data collected in 2006–2007 and 2010– Gambia, and the Democratic Republic of the Congo 2012. They concluded that women often move shorter are the main destinations of Senegalese international distances and tend to migrate from one rural area to migrants to other African countries. For internal another. Moreover, women’s migration is often driven migrants, Dakar is the most prevalent destination by marriage or family reasons, while men are signifi- (for about 50 percent of both current and past inter- cantly more likely to migrate for economic incentives. nal migrants). Male Outmigration and Women’s Work and Empowerment in Agriculture 7 CHAPTER THREE SURVEY METHODOLOGY This study draws on two unique household surveys from Nepal and Senegal. The two survey questionnaires are essentially the same, but some modules are adapted to the local context (e.g., the lists of livestock and crops in the two countries are different). SURVEY LOCATIONS The survey sample from Nepal consists of 1,002 households from five districts (Achham, Rolpa, Nawalparasi, Makwanpur, and Jhapa). These districts were purposefully selected for the study based on two main criteria: (i) high emigra- tion rates, and (ii) wide geographic coverage. Because of limited resources, a nationwide survey could not be carried out, but the selected districts are distrib- uted across two ecological zones (the hills and the Terai; the mountains were excluded because of extremely low population densities) and the five former7 developmental regions (Figure 2). The survey sample was drawn from rural areas of the selected five districts and is, therefore, representative of their rural areas. The survey sample from Senegal includes 999 households8 from two regions (Matam and Kaolack) (Figure 3). As in Nepal, the two regions were purpose- fully selected because of their high rates of internal and international migra- tion. The sample is representative of rural areas in the two regions where the survey was implemented. As in Nepal, only rural areas were surveyed. SURVEY INSTRUMENTS It is important to highlight that each overall survey consisted of three separate instruments: a household questionnaire, the Abbreviated Women’s Empower- ment in Agriculture (A-WEAI) questionnaire, and the Food Insecurity Experi- ence Scale (FIES). The household questionnaire was completed by the most 7 This is the administrative division before the new constitution in 2015 in Nepal. 8 The operational definition of the household includes all wives and all children of both the household head and of the household head’s spouse(s). Male Outmigration and Women’s Work and Empowerment in Agriculture 9 FIGURE 2. MAP OF NEPAL WITH THE SAMPLED DISTRICTS Source: “Technical Report on Survey of Migration and Women’s Empowerment in Agriculture” prepared by Nepa School of Social Sciences and Humanities, September 2, 2017. knowledgeable person in the household. The A-WEAI data (de Brauw and Carletto 2012). As the focus of questionnaire was completed by the migrant’s spouse the study is migration out of rural areas, the house- or a member of the primary couple, and only collected hold questionnaire collected information on the information on the respondent. Thus, the A-WEAI was determinants of current and past international and used to collect self-reported information regarding internal outmigration, employment characteristics various domains of empowerment (see below) from a of migrants before and after the migration episode, subset of the whole individual sample. These data are family migration history, cyclical and seasonal migra- therefore not representative of all adult rural women, tion episodes, remittances, and migration financ- unlike data from the household questionnaire, which ing. Throughout the household questionnaire, collected information about employment and other but particularly throughout the migration module, characteristics for all adult rural women. individual-level, gender-relevant questions related to migration were included, such as: who made the The household questionnaire was designed to cap- decision to migrate, who in the household receives ture detailed, sex-disaggregated, and gender-relevant remittances, and how much control do migrants and information on migration as well as on agriculture, recipients have over the use of remittances. employment, and other characteristics of rural house- holds. Its migration modules built and improved on The modules from the abbreviated version of the Wom- existing surveys and closely followed recent guide- en’s Empowerment in Agriculture Index (A-WEAI) were lines and recommendations for collecting migration also included in the survey (Table A1 in Annex A). The 10 The Case of Nepal and Senegal FIGURE 3. MAP OF SENEGAL WITH THE SAMPLED REGIONS Source: World Bank. A-WEAI focuses on the same five domains of empow- When the migrant did not have a spouse, or if the erment as the WEAI—input into decisions about agri- spouse was unavailable, the A-WEAI instrument was cultural production, access to and decision-making administered to another woman in the household who about resources (including ownership of assets and was randomly selected. In nonmigrant households, access to and decisions about credit), control over use the A-WEAI instrument was administered either to the of income, group membership, and time use (Alkire man or woman of the primary couple.10 et al. 2013), but excludes some of the modules that were difficult to implement (Malapit et al. 2015). The decision-making and control over income may increase for women who A-WEAI was used to keep the multitopic questionnaire stay behind, but higher workload and time poverty may move the index in at a reasonable length to minimize interview fatigue as the opposite direction. Therefore, to study the linkages between migration well as costs. In migrant households, the A-WEAI mod- and women’s and (men’s) empowerment, it is essential to focus on the components of the index rather than on the composite index. ules were administered to the spouse of the migrant.9 10 In a nuclear household, there is only one couple. In multigenerational households, the primary couple is largely defined on the basis of age 9 The A-WEAI modules were administered to one individual per household as the couple in prime working age. In households where there were for several reasons. First, it is impossible to interview the man or woman of multiple primary couples, the A-WEAI was administered to any of the the primary couple in households where one of the partners is a migrant. prime working-age couples. The objective was to avoid administering the Second, it is costly and time-consuming to interview two people per house- A-WEAI to elderly couples, which might have occurred if the focus was hold for the A-WEAI. Third, the components of the index rather than the household head. The underlying hypothesis is that migration plays the index itself are of primary interest for the study. Male outmigration is a more transformative role in changing gender roles and perceptions unlikely to influence all aspects of empowerment in the same direction— among the younger generation rather than for the elderly. Male Outmigration and Women’s Work and Empowerment in Agriculture 11 Another innovative feature of the overall survey was for 5,227 (migrant and nonmigrant) family members its module on household food security status. This was collected. Since the analysis focuses primarily on module solicited information for the Food Insecurity work and empowerment outcomes, the sample was Experience Scale (FIES) developed by FAO’s Voices restricted to those 16 years and older, which left 3,544 of the Hungry Project (Ballard, Kepple, and Cafiero individuals. Furthermore, individuals who were not 2013). The FIES is an experience-based metric of the in the household at the time of the survey (interna- prevalence of food insecurity that relies on direct tional and internal migrants) and those who were yes/no responses to eight questions regarding access residing in the household were distinguished.12 At to food. FAO recently developed the FIES to estimate the time of the survey, 530 adults lived abroad (inter- two indicators – the prevalence of moderate or severe national migrants) and another 92 adults resided food insecurity (FImod+sev) and the prevalence of in Nepal but not in the locality of their household severe food insecurity (FIsev). The FIES is compara- (internal migrants). In the final sample, 12 individu- ble across different countries and cultures. Moreover, als were excluded because of missing information on FImod+sev was selected, together with the prevalence some of the variables included in the final model. of undernourishment, as an indicator to monitor Sus- The remaining 2,910 adult individuals from the sam- tainable Development Goal target 2.1: By 2030, end ple of working-age adults who resided in rural areas hunger and ensure access by all people, in particular, the at the time of the survey were the main subjects of poor and people in vulnerable situations, including infants, this study. These individuals belonged to one of three to safe, nutritious and sufficient food all year round. different types of households: (i) households with an international migrant (1,181 individuals from In both Nepal and Senegal, the FIES module was 443 households); (ii) households with an internal administered at the household level rather than the migrant but no international migrants (133 individu- individual level. The individual-level version inquired als from 55 households); and (iii) households with directly about interviewed individuals’ perception of neither internal nor international migrants (1,596 food insecurity. As these data were collected at the individuals from 504 households). household level, this study can only examine the relationship between migration and food insecurity In Senegal, individual-level information was collected of the whole household, not evidence of any differ- from 999 rural households for 10,380 migrant and ences in food insecurity at the individual level (such nonmigrant family members. There were 6,350 indi- as between women and men). viduals 16 years and older. Excluding migrant mem- bers left a sample of 5,125 adult individuals (from 997 SAMPLES households13). Some 154 individuals (and 9 house- holds) were excluded from the analysis because of While most of the literature focuses on migrant house- holds in destination areas, the current analysis studies periods of time to work, receive education, or visit relatives (current the consequences of outmigration in sending commu- migrants). Thus, the household membership selection criteria stipulated nities. Therefore, references to migrant households the inclusion of all children of the man and woman of the primary cou- always imply households in sending communities. ple (working-age) provided that the member (i) did not have another family; and (ii) shared food from a common source with other household In Nepal, 1,002 rural households were sampled for the members when present. 12 Individuals over 16 years old who migrated in the 12-month period prior survey. Individual-level information was collected for to the survey and were back home at the time of the survey were included all household members, including current migrants in the migrant group, regardless of their intention to stay or go back to the who were absent.11 Thus, individual-level information migration destination. In both Senegal and Nepal, these individuals con- stituted a very small number. Thus, it is unlikely that their classification as migrant or past/return migrant influenced the overall results. Because of the focus of the study on migration, the definition of the 11 13 In two households, the current location of some members was not pro- household was extended to include all people who belong to this house- vided, making it impossible to determine their migration status. These hold and do not have another family, even if they may be away for long households were excluded from the sample. 12 The Case of Nepal and Senegal missing information for the variables of interest. The migrant; 1,694 individuals in 354 households with an final sample included 4,971 individuals from 988 internal migrant but no international migrant; and households, distributed as follows: 1,428 individu- 1,849 individuals in 368 households with no current als in 273 households with at least one international migrants. Male Outmigration and Women’s Work and Empowerment in Agriculture 13 CHAPTER FOUR CHARACTERISTICS OF MIGRATION IN THE SURVEYED AREAS CHARACTERISTICS OF NEPALI MIGRANTS As expected, rural outmigration in Nepal in the survey sample is heavily domi- nated by men—more than 93 percent of reported migrants are men (Table 1).14 Working-age migrants are relatively younger than the overall working-age population in Nepal15—31 years on average compared to 38 years for non- migrants. Migrants are also better educated: only 9 percent of international migrants have no education compared to 33 percent of rural people who stay behind; 24 percent have a primary education compared to 18 percent of non- migrants; and 67 percent have a secondary education compared to 48 percent of the nonmigrant population. Like nonmigrants, almost three-quarters of migrants are married. Around 43 percent of surveyed households in Nepal receive remittances; the median amount received is more than double the per capita GDP. In Nepal, 87 percent of households with international migrants receive remittances (only 13 percent do not receive remittances); 65 percent of households with only inter- nal migrants receive remittances; and only 6 percent of households with no migrants received remittances, perhaps from relatives or friends abroad. Most often remittances are sent every three months and 86 percent of the house- holds who receive remittances get them twice per year or more frequently. The median amount of remittances sent by all migrants in the 12-month About three-quarters of households have a member who lived in the household at the time of the survey, but 14 who was a migrant or lived somewhere else a year earlier. Nearly 80 percent of these individuals are women, and the major reason for moving to the current location is family reasons, such as joining the husband’s household. This type of migration is not included in this analysis, which focuses on economic migration. 15 In Nepal, the sample of households with only internal migrants is very small (92 respondents out of 3,544). Therefore, most of the discussion that follows focuses on the differences between international migrant households and all other households (i.e., nonmigrant households and households with internal migrants are combined). Male Outmigration and Women’s Work and Empowerment in Agriculture 15 TABLE 1. C  HARACTERISTICS OF INTERNATIONAL MIGRANTS VERSUS NONMIGRANTS, WORKING-AGE INDIVIDUALS (AGE 16+), NEPAL   (1) INTERNATIONAL MIGRANTS (2) NONMIGRANTS   MEAN STD. ERR. MEAN STD. ERR. P-VALUE Individual characteristics Age (years) 31.19 0.47 37.93 0.41 *** Female† 0.07 0.01 0.57 0.01 *** Never married† 0.24 0.02 0.20 0.01 ** Married† 0.75 0.02 0.73 0.01 Cohabiting† 0.00 0.00 0.00 *** Widowed/divorced† 0.01 0.00 0.07 0.01 *** No education† 0.09 0.01 0.33 0.01 *** Primary education† 0.24 0.02 0.18 0.01 ** Secondary education† 0.67 0.02 0.48 0.01 *** High caste 0.41 0.03 0.43 0.01 Low caste 0.21 0.02 0.12 0.01 *** Muslim 0.04 0.01 0.02 0.00 * # observations 530   2910     Note: * the difference is significant at the 10% level; ** – at the 5%: *** – at the 1% level. period prior to the survey was 160,000 Nepali rupees FIGURE 4. USE OF REMITTANCES, NEPAL (approximately US$1,555). International migrants sent more—the median amount sent was 200,000 Clothin nd footw r Dw llin construction/ Nepali rupees (approximately US$1,944). This is r p irs a significant amount in a country where GDP per Educ tion capita in 2016 was only US$729. Almost two-thirds Food of remittance senders indicate how the remittances H lth should be used. Although other family members Hous hold’s f rmin ctiviti s may also participate in the decision about the use of Oth rs sp cif remittances, the decision-making process remains P m nt of d bts heavily dominated by men, since most migrants are Purch s of dw llin men. In 61.4 percent of the households that receive remittances, the only decision-makers about the use Purch s of plot l nd of remittances are men; in 22.2 percent the only S vin S rvic s ( l ctricit , decision-makers are women; and in 16.4 percent of w t r, phon ) households, both men and women make decisions 0 100 200 300 regarding the use of remittances. Us of r mitt nc s (Counts) Note: Respondents were allowed to choose as many categories as needed. Remittances are predominantly used to purchase food (Figure 4). In addition, remittances are used for clothing, education fees, payment of debts, and use remittances for household farming activities, health care costs. Around 30 percent of households including for the purchase of land. This to some 16 The Case of Nepal and Senegal extent validates the hypothesis that households use International and internal migrants appear to have the capital obtained from international migration distinct characteristics. International migrants are mainly to overcome liquidity constraints for subsist- slightly older than the nonmigrant working-age pop- ence production. ulation by about three years, while internal migrants are significantly younger by around five years. Almost CHARACTERISTICS OF 80 percent of international migrants are married, mostly monogamously, compared to 73 percent of SENEGALESE MIGRANTS working-age nonmigrants. Internal migrants are least Unlike Nepal, internal migration dominates inter- likely to be married—only about one-half report national migration in the Senegalese sample. About being married. About 17 percent of international 13 percent of working-age individuals in the two migrants and 21 percent of nonmigrants are in polyg- study regions were internal migrants at the time of amous marriages. the survey or had migrated within Senegal in the 12-month period prior to it. The incidence of inter- Low education levels are characteristic of the whole national migration was about one-half that of inter- working-age population in Senegal. Compared to nal migration – about 6.5 percent of the working-age Nepal where one-third of nonmigrant adults have population resided abroad or had lived abroad in the no education, in Senegal three-quarters of nonmi- 12-month period prior to the survey. grants have no education (Table 2). A similar share of international migrants has no education. Inter- Men dominate both internal and international migra- nal migrants appear to be slightly better off in this tion. Around 17 percent of internal migrants and respect—64 percent have no education, but the rest nine percent of international migrants are women. have at least some primary or even some secondary TABLE 2.  CHARACTERISTICS OF INTERNATIONAL AND INTERNAL MIGRANTS VERSUS NONMIGRANTS, WORKING-AGE INDIVIDUALS (AGE 16+), SENEGAL   (1) INTERNATIONAL (2) INTERNAL DIFF (1) MIGRANTS MIGRANTS (3) NONMIGRANTS VS (3) MEAN STD. ERR. MEAN STD. ERR. MEAN STD. ERR. P-VALUE Individual characteristics Age (years) 38.74 2.35 29.71 1.55 35.36 3.54 * Female† 0.09 0.02 0.17 0.02 0.53 0.10 *** Never married† 0.20 0.04 0.47 0.05 0.27 0.08 ** Married monogamous† 0.63 0.03 0.44 0.02 0.44 0.05 *** Married polygamous† 0.17 0.05 0.07 0.03 0.21 0.03 Widowed/divorced† 0.01 0.01 0.02 0.01 0.08 0.01 *** No education† 0.78 0.04 0.64 0.03 0.78 0.02 Primary education† 0.07 0.01 0.08 0.01 0.07 0.00 Secondary education† 0.15 0.03 0.28 0.03 0.15 0.03 Ethnicity: Pular† 0.80 0.03 0.48 0.02 0.53 0.01 *** Ethnicity: Sirer† 0.06 0.03 0.12 0.02 0.20 0.00 *** Ethnicity: Wolof/Libou† 0.12 0.02 0.37 0.03 0.24 0.00 *** # observations 412   813   4971     Note: * the difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. Male Outmigration and Women’s Work and Empowerment in Agriculture 17 FIGURE 5. USE OF REMITTANCES, SENEGAL The Wolof/Libou, on the other hand, are more likely to migrate internally. The Wolof/Libou account for Clothin nd footw r 24 percent of the adult sample, but for only about 12 Educ tion percent of all international migrants and 37 percent Food of all internal migrants. The third most populous Fun r l/W ddin / C r moni s/R li ious ethnicity in the sample is the Sirer, but their share fr stiv l among all migrants is significantly smaller than their H lth Oth r (to b sp cifi d) share in the whole population. P m nt of d bts Purch s of p rc l In Senegal, about 30 percent of surveyed households of f rml nd receive remittances. About 56 percent of households Purch s of ricultur l input with at least one international migrant and 42 per- Purch s of liv stock cent of households with internal migrants (but no S vin international migrants) receive remittances. Very S rvic s ( l ctricit , few (about three percent) households without any w t r, phon ) 0 200 400 600 800 1,000 international or internal migrants receive any remit- Us of r mitt nc s (Counts) tances. The median amount of remittances sent in Note: Respondents were allowed to choose as many categories as needed. the 12-month period prior to the survey was 50,000 CFA francs (approximately US$95) from internal migrants and 55,000 CFA francs (approximately education. The relatively higher educational achieve- US$105) from international migrants, roughly 11 ment among internal migrants could imply that some percent of Senegal’s per capita GDP. In 40 percent of the reasons for migration are the pursuit of higher of households, the senders indicate how the money education, such as adolescents migrating for educa- should be spent. tion purposes. Food is by far the most often stated use of remit- The data also suggest that individuals from certain tances (Figure 5). As in Nepal, clothing, education ethnic groups are significantly overrepresented fees, payment of debts, and health care costs comprise among migrants. For example, about 53 percent of an important share of use of remittances. Unlike in all nonmigrants in the sample are Pular, but they Nepal, farming activities are rarely listed as a use of comprise 80 percent of the international migrants. remittances. 18 The Case of Nepal and Senegal CHAPTER FIVE INDIVIDUAL CHARACTERISTICS OF WOMEN WHO STAY BEHIND NEPAL Some noticeable differences arise in the individual characteristics of women in international migrant households versus those in nonmigrant households in Nepal. For example, compared to women in nonmigrant households, women in migrant households are more likely to be married (Table 1). In terms of household characteristics, migrant households have more young children (under five years old) compared to nonmigrant households, and significantly more adult women and men, suggesting that migration may be facilitated by the presence of extended families, since other adults can take over the tasks of the migrant or help with the care of very young children. In addition, migrant households are more likely to belong to a low caste compared to nonmigrant households. SENEGAL In Senegal, the individual characteristics of women in households with migrants are very similar to those of women in nonmigrant households, except for their ethnicities. For example, Pulars are more represented among international migrant households than Sirers and Wolof/Libous (Table 2). And women in households with international migrants are slightly better educated than women in households with no migrants. However, significant differences arise between the household characteristics of migrant and nonmigrant households. Households with an international or internal migrant have fewer very young children (under age 10) than house- holds with no migrants. In addition, migrant households have significantly more adult women and men than nonmigrant households. As mentioned in the case of Nepal, the presence of more adults in migrant households may be a key factor facilitating the decision to migrate. Male Outmigration and Women’s Work and Empowerment in Agriculture 19 Households of international migrants also differ in terms of household wealth, as suggested by the char- COUNTRY COMPARISON acteristics of their dwellings. Nearly 80 percent of Compared to Nepal, the differences between migrant women in households with international migrants and nonmigrant households in Senegal are signifi- live in houses with cement walls compared to 65 cantly more pronounced. It is difficult to draw con- percent of women in nonmigrant households. That clusions whether returns to migration are higher in international migrant households are better off than Senegal, however, or whether only better-off house- nonmigrant households or even households with an holds can afford to send a family member abroad internal migrant is also shown by the higher quality given the high costs of migration. of the roof and the floor of the dwelling, the access to better toilet facilities and piped water, and a source of drinking water in the house. 20 The Case of Nepal and Senegal CHAPTER SIX EMPLOYMENT CHARACTERISTICS OF WOMEN WHO STAY BEHIND In both countries, almost all economically active men and women report farm- ing as one of their economic activities. However, differences exist between the two countries in terms of labor force participation rates, employment rates, and reported engagement in types of agricultural work by individuals who stay behind. Engagement in other income-generating activities outside the family farm (including working as laborers in or outside of agriculture, in processing, or in trade of agricultural products) is rare for both men and women in the two countries. The discussion below highlights a few employment-related charac- teristics of the rural women who stay behind. NEPAL In Nepal, women and men in international migrant households are just as likely to be economically active as those in nonmigrant households. Nearly 90 percent of all adult men and women, regardless of the migration status of their family, participated in at least one employment activity in the 12 months prior to the survey.16 There are no significant differences in the probability of employment between women in migrant-sending households and women in nonmigrant households. Therefore, the data do not support the notion of women dropping out of the labor force or reducing employment in response to the migration of their partners or other family members. 16 In the survey, respondents were asked whether they are engaged in seven broad types of activities: (i) self- employed, employer, or contributing family member; (ii) agricultural worker; (iii) processing of agricultural products; (iv) trader/seller of agricultural products; (v) nonagricultural worker, nonagricultural artisan, or worker engaged in commerce; (vi) professional (private and public sector); and (vii) other. A detailed list of activities/professions was included in each category so that enumerators could easily classify the economic activity of the surveyed individuals. For each activity, respondents were further asked whether it is done as self-employment or as an employee, whether it is market-oriented, the number of months performed in the last 12 months, number of days per month, and average number of hours per day. In addition, earnings infor- mation was collected as well as information on whether the activity is regular employment or not. Male Outmigration and Women’s Work and Empowerment in Agriculture 21 FIGURE 6. WOMEN’S EMPLOYMENT in migrant households and seven percent of women OUTCOMES BY HOUSEHOLD in nonmigrant households engage in agricultural MIGRATION STATUS, NEPAL wage labor. Most of the wage work is on small farms with fewer than five workers. Only 3.6 percent of all Non- ricultur l work adults employed as wage workers report that they are Tr din ( ricultur l products) employed on a regular, full-time basis for the whole Proc ssin year; 22 percent are part-time employees; and the ( ricultur l products) rest are classified as seasonal, short-term, or casual A ricultur l l bor r employees (statistics not included in the table). These F rm contributin characteristics of agricultural wage employment sug- f mil work r gest that it is not a major source of employment. It is F rm s lf- mplo d more likely to be a livelihood diversification strategy Workin , as family farming may not be sufficient for house- n occup tion hold food and financial security. Furthermore, only 0% 20% 40% 60% 80% 100% five to seven percent of adult women engage in the P rc nt of wom n n d in th ctivit processing of agricultural products and even fewer HH with No Mi r nt HH with n Int rn tion l Mi r nt women are engaged in the trade of agricultural products. Finally, less than five percent of women in nonmigrant and migrant households are engaged in In Nepal, women in migrant households are signifi- nonagricultural activities and the difference is not cantly more likely to be identified as self-employed17 statistically significant. in agriculture compared to women in nonmigrant households (Figure 6). About 32 percent of women in migrant households are classified as self-employed SENEGAL in agriculture compared to 20 percent of women in The Senegalese data show low labor participation in nonmigrant households. In addition, less than 60 general. A significantly smaller share of the population percent of women in migrant households are classi- in Senegal was economically active in the 12 months fied as contributing family workers compared to 72 prior to the survey compared to Nepal. Within Sen- percent of women in nonmigrant households. Sub- egal, differences arise in economic activities by sex sistence farming appears to dominate family farming; and by migration status of the household. Only 40 approximately 55 percent of self-employed adults percent of both men and women in households with report that less than 50 percent of their agricultural an international migrant participated in at least one production is intended for the market. economic activity in the 12 months prior to the survey compared to around 60 percent of men and women In Nepal, very few women (and men) engage in agri- in households with internal or no migrants. cultural wage work. Only five percent of all women Women in households with an international migrant report some of the lowest employment rates. Only 17 Self-employment includes jobs “whose remuneration depends directly 26 percent of women in international migrant house- on the (expectation of) profits derived from the goods and services holds report having worked in the 12 months prior to produced” and “engage one or more persons to work for them as the survey compared to about 50 percent of women ‘employees’ on a continuous basis” (http://www.ilo.org/global/statis- tics-and-databases/statistics-overview-and-topics/status-in-employment/ in households with internal or no migrants (Figure current-guidelines/lang--en/index.htm). In this study, the definition of 7). Thus, unlike Nepal, where no clear relationship self-employment is expanded to include own-account workers, who are is found between migration and the employment also self-employed individuals but do not hire employees on a continuous status of family members who stay behind, in Sen- basis. Contributing family workers are those who “hold self-employment jobs in an establishment operated by a related person, with a too-limited egal, a glaring negative relationship arises between degree of involvement in its operation to be considered a partner” (ibid). migration and the probability of having worked in 22 The Case of Nepal and Senegal FIGURE 7. WOMEN’S EMPLOYMENT about one-quarter of men not working were actively OUTCOMES BY HOUSEHOLD searching for a job. MIGRATION STATUS, SENEGAL Engagement in farming activities is significantly Non- ricultur l work lower among women in households with interna- Tr din tional migrants compared to women in nonmigrant ( ricultur l products) Proc ssin households. Agriculture is the most important sector ( ricultur l products) of employment for most rural women (and men) in A ricultur l l bor r Senegal. Most women working in agriculture, how- F rm contributin ever, are classified as contributing family workers f mil work r rather than self-employed workers or employers (Fig- F rm s lf- mplo d ure 7). Only two percent of women in households Workin , with international migrants are self-employed com- n occup tion pared to four percent in households with an inter- 0% 10% 20% 30% 40% 50% 60% nal migrant, and six percent in households with no P rc nt of wom n n d in th ctivit migrants. Thus, if women are employed, they are HH with No Mi r nt HH with n Int rn l Mi r nt HH with n Int rn tion l Mi r nt most likely to be contributing family workers. Working outside of the family farm, even as an agri- the last year. Women’s main reason for not work- cultural laborer on other farms, is rare in Senegal. ing in the past year is that they were doing domes- Less than one percent of all adult men and women tic work without pay, which means that these women combine work as agricultural wage laborers on other were not actively looking for a job and were therefore people’s farms or process agricultural products, excluded from the labor force. Men offered different and less than five percent trade agricultural prod- reasons for not working in the past year, such as stud- ucts. These statistics are even lower considering only ying (almost 40 percent). And in contrast to women, women (Figure 7). Male Outmigration and Women’s Work and Empowerment in Agriculture 23 CHAPTER SEVEN CHARACTERISTICS OF WOMEN’S EMPOWERMENT18 As mentioned earlier, detailed information on various indicators of empower- ment as specified in the A-WEAI was collected for only a subset of women (and men) in both Nepal and Senegal. Therefore, the results related to empower- ment are not valid for all women in Nepal or Senegal but are rather intended to capture the empowerment of the woman most directly linked to the migrant, such as a spouse or a mother. Nepalese women in nonmigrant households have a more diverse income-gener- ating portfolio than do spouses of migrants, as captured by the A-WEAI (Table B5 in Annex B). In Nepal, women in nonmigrant households participate in a greater number of productive activities than women in migrant households. Women in nonmigrant households are slightly more likely than women in migrant households to engage in off-farm and self-employment in addition to working on the family farm. Women in nonmigrant households are also slightly more likely than women in migrant households to be engaged in poultry rear- ing. These statistics suggest that women in nonmigrant households have a more diversified portfolio of income-generating activities, perhaps because they can- not rely on remittances to cushion the negative effects of poor harvests. Women in Senegal are characterized by a low level of economic activity in gen- eral (Table B6 in Annex B). As seen earlier, a large share of women (but also 18 As mentioned earlier, the questionnaire used to collect information on empowerment builds upon the Abbre- viated Women’s Empowerment in Agriculture Index Questionnaire (A-WEAI) developed by the International Food Policy Research Institute (IFPRI). However, it was modified to include additional questions about decision- making and control of income from nonagricultural livelihoods. This was necessary because the original A-WEAI only collected information about rural women’s agricultural activities, thus potentially leading to misleading estimates of the empowerment status of women whose livelihoods were not based on agriculture. At the center of the A-WEAI is the definition of empowerment as “the expansion of people’s ability to make strategic life choices, particularly in contexts where this ability had been denied to them” (Alkire et al. 2013). The five domains of empowerment of the A-WEAI include indicators that focus on respondents’ capacities to make decisions. See Annex A for the exact set of indicators used to understand women’s empowerment in the various domains. Male Outmigration and Women’s Work and Empowerment in Agriculture 25 men) did not work in the last 12 months, not even compared to 47 percent of women in nonmigrant on the family farm, according to the data collected households are active members of at least one group. through the household questionnaire. The lower level of economic activity in rural Senegal compared Women in migrant households in Nepal are more over- to that in rural Nepal is also reflected in the responses worked than women in nonmigrant households. More to the A-WEAI modules. According to the data col- than one-half of all women in Nepal report working lected through the A-WEAI, few women engage in more than 10.5 hours a day—a figure that does not economic activities regardless of whether they are in account for the fact that women’s work activities may agriculture or not. Primary-age women interviewed also overlap with child care. Yet only about 21 percent using the A-WEAI module in Senegal report doing of men report working more than 10.5 hours a day less than one economic activity on average, while (Table A2). Forty-eight percent of women in house- women in Nepal engage in nearly three different eco- holds with an international migrant work fewer than nomic activities, most of which are within agriculture. 10.5 hours a day compared to 56 percent of women In Senegal, only about 17 percent of all women who in nonmigrant households, suggesting a potential dis- responded to the A-WEAI module reported working empowering effect of migration on women who stay in staple grain farming compared to 95 percent of behind through higher work burden. women in Nepal. More than 60 percent of women in In both countries, regardless of the immigration sta- Nepal keep livestock compared to about 6 percent of tus of the household, significant gender gaps arise in women in Senegal. Small livestock and poultry rear- access to resources, information, and decision-mak- ing is also not as common in Senegal as it is in Nepal. ing in various domains (Table A2 and Table A3). Most striking is the gender gap in ownership of land, a key Control over agricultural income is dependent pri- agricultural asset for agriculture-based livelihoods. In marily on the level of engagement in the income-gen- Nepal, only one-third of women own land versus about erating activity itself, regardless of the household’s two-thirds of men (based on the responses to A-WEAI migration status. In Nepal, women have high control module). In Senegal, 88 percent of men but only 56 over agricultural income, regardless of the house- percent of women own any land solely or jointly. hold’s migration status, perhaps because of their high engagement in agriculture. In addition, women Gender gaps in access to information about agri- in nonmigrant households have higher control over cultural production are also noticeable. Despite nonagricultural income, since they are, on average, reportedly high levels of access to information about more likely to engage in off-farm work as well. In agricultural production in Nepal, women are still sig- Senegal, no such differences arise, in part because nificantly disadvantaged in that respect compared to women in all households have similarly low levels men. In Senegal, access to information is rather low of participation in agricultural and nonagricultural for all, but is significantly lower for women—only 26 income generation. percent of women report being able to access infor- mation about agricultural production compared to Participation in local groups is higher among women 41 percent of men. in migrant households than among women in non- migrant households in Nepal. About one-half of all A comparison between Nepal and Senegal clearly women in Nepal and one-third of all women in Sen- shows that the gender gaps are even more striking in egal are active members of at least one agricultural, Senegal. Women in Senegal are disadvantaged rela- financial, social, or religious group. In Senegal, no tive to men in almost all domains: they have lower statistically significant differences occur in group decision-making power for agricultural activities, membership by migration status of the household, lower ownership of land, lower access to credit and but in Nepal, pronounced differences are found. decision on credit, and lower control of income from About 56 percent of women in migrant households both agricultural and nonagricultural sources. 26 The Case of Nepal and Senegal CHAPTER EIGHT FOOD SECURITY CHARACTERISTICS On average, only about one in ten households in the Nepalese sample reported severe or moderate food insecurity. The FIES-based estimates of food insecu- rity in Nepal are presented in Figure 8. The average prevalence rate of severe or moderate food insecurity (FImod+sev) in the five study districts is about nine percent. However, the results vary considerably across districts. Achham is by far the worst off, with FImod+sev equal to 26.4 percent. Although not sta- tistically significant, food insecurity is higher for households with at least one migrant abroad than for households with no migrants. However, the results reverse when distinguishing between households that receive remittances and those that do not. The prevalence of food insecurity for households that do not receive remittances is about one percentage point higher than the prevalence of food insecurity among households that do receive remittances (whether FIGURE 8. THE PREVALENCE OF FOOD INSECURITY BASED ON FIES, NEPAL 30% 25% 20% 15% 10% 5% 0% M kw npur All 5 Districts N w lp r si Achh m Mi r nt Non-Mi r nt R mitt nc No-R mitt nc Jh p Rolp Districts Mi r tion Abro d R mitt nc FIES S v r FIES Mod r t to S v r Male Outmigration and Women’s Work and Empowerment in Agriculture 27 FIGURE 9. THE PREVALENCE OF FOOD INSECURITY BASED ON FIES, SENEGAL 60% 50% 40% 30% 20% 10% 0% K ol ck ions M t m Mi r nt Non-Mi r nt Mi r nt Non-Mi r nt R mitt nc No-R mitt nc Both R R ions Mi r tion Abro d Int rn l Mi r tion R mitt nc FIES S v r FIES Mod r t to S v r they have a migrant abroad or not), highlighting the severe food insecurity is considered; the prevalence importance of successful migration to household rate is almost double in households without remit- well-being and food security. tances (15.6 percent versus 8.7 percent). Close to one-half of all households in the Senega- The impact of migration on household food secu- lese dataset reported severe or modest food inse- rity depends on whether the migration is successful curity. The FIES-based estimates of food insecurity or not. In Nepal, no significant correlation exists in Senegal are presented in Figure 9. The average between migration and food insecurity (see Annex prevalence rate of severe or moderate food insecu- C). The prevalence of food insecurity in the five rity (FImod+sev) in the two study regions is 44.3 per- districts in Nepal is much lower than the national cent, with similar levels in Kaolack (43.8 percent) average, as estimated by Voices of Hungry using and Matam (45.0 percent). However, the prevalence Gallup data (FAO 2016). However, the signs of the of severe food insecurity is much larger in Matam coefficients provide suggestive evidence that it is (20.1 percent) than in Kaolack (10.9 percent). Inter- not migration per se that is associated with lower estingly, food insecurity is lower for households with household food insecurity, but rather the receipt of at least one migrant abroad compared to households remittances from migrants. Migration of household with no migrants, but the results reverse when inter- members not followed by remittance transfers is nal migration is considered, in which case FImod+sev likely to increase household food insecurity. This is is equal to 51.2 percent and 42.6 percent, respectively, clearer in Senegal. Both international and internal for households with and without internal migrants. migration are positively associated with food insecu- The prevalence of moderate or severe food insecu- rity (though only the coefficient on internal migra- rity is 46 percent for households that do not receive tion is statistically significant), but the receipt of remittances and 36 percent for those that receive remittances is linked to lower food insecurity (Table remittances. The difference is even starker when C2, column 2). 28 The Case of Nepal and Senegal CHAPTER NINE EMPIRICAL STRATEGY This study models the labor allocation and empowerment of women as a func- tion of whether they live in a household with an international migrant, M1h, and of individual, household, and community characteristics, Xih: 1. Yih = α + βM1h + γXih + εi (Nepal) where Yih is a set of different indicators for women’s work in and outside of agriculture. In Nepal, because there are too few households with an internal migrant, the model simply differentiates between households with an inter- national migrant and all other households, combining households with no migrants or only internal migrants into the base category.19 In Senegal, both internal and international migration are significant, so controls for both types of migration are included: 2. Yih = α + β1M1h + β2M2h + γXih + εi (Senegal) In this case, M1h indicates a household with at least one international migrant and M2h indicates a household with at least one internal migrant but no inter- national migrants.20 The base category (comparison group) in Senegal includes households with no current or recent (in the last 12 months) internal or inter- national migrants. The same model is employed to study the linkages between male-dominated migration and women’s empowerment in and outside of agriculture. The 19 The models tested if the empirical results changed depending on whether households with current domestic migrants were included (i) in the base category, (ii) separately as a control, or (iii) completely dropped from the analysis. The estimates were not at all sensitive to how domestic migrants were included in the model. 20 Some households had several family members who emigrated. If at least one family member emigrated abroad, then the household was classified as a household with an international migrant. The assumption was that international migration would have a stronger effect on women’s work and empowerment for various reasons, including higher potential returns and initial costs, the difficulty of the migrant to return home frequently, and the exposure to foreign social and cultural norms. Male Outmigration and Women’s Work and Empowerment in Agriculture 29 indicators of empowerment are based on the five and religious background); household demographic domains of the A-WEAI based on whether the characteristics; household wealth and asset char- respondent: (i) is adequately empowered in deci- acteristics (quality of the construction materials of sions about agricultural production; (ii) has adequate the dwelling, quality of sanitary facilities, source of control and access to resources; (iii) has control of drinking water, access to electricity, household own- income; (iv) is overworked (based on a 24-hour time- ership of land, land area owned and cultivated, and use recall module); and (v) is a member of an active livestock ownership expressed in Tropical Livestock group in the community. εi is the error term in all Units (TLU)); and a dummy variable for whether the three equations. household received any social assistance. The model for Nepal includes district fixed effects; for Senegal, To separate the labor effect of migration and the department fixed effects are included. income effect from the receipt of remittances, model 1 and model 2 are re-estimated with the following The key problem for studies on the impacts of migra- indicators: (i) M1R1h is an indicator for whether the tion is that migration is a selective process—migrants household has an international migrant who has sent are likely to be significantly different from nonmi- any positive remittances in the last year; (ii) M1R0h grants in both observable and unobservable ways. is an indicator equal to one if the household has an The decision to migrate may be based on the same international migrant but has not received any remit- factors that affect the employment and empower- tances in the past year; and (iii) M2h is an indicator ment outcomes of interest—this is the classic omitted equal to one if the household has at least one internal variable problem. Moreover, reverse causality may migrant (and no international migrants), regardless play a role. Migration may change intrahousehold of whether the internal migrant has sent remittances. dynamics and women’s decision-making power, but if The base category includes women in households women and men value migration differently, women with no international or internal migrants and no who are more empowered may exert a higher influ- remittances. A very small share of households in ence on the husband’s migration decision. Using both countries receives remittances without hav- longitudinal data from Mexico, Nobles and McKel- ing any migrants. These households are too few to vey (2015) showed that an exogenous positive shock get an accurate picture of their characteristics and to women’s empowerment, proxied by the decision- to understand what differentiates them from other making over household resources, leads to a lower migrant-sending and nonmigrant households. For probability that the husband migrates. that reason and for greater clarity in interpreting the results, they are excluded from this model. To help solve the endogeneity problem, an instru- mental variable approach is employed. The ideal 3. Yih = α + β1M1R1h + β2M1R0h + β3M2h + γXih + εi instrument must be correlated with the decision to migrate and uncorrelated with the error term; it Rather than using information on the amount of remit- should affect the outcome of interest only through tances received, the study uses an indicator variable for its effect on migration. Therefore, drawing on the remittance receipts. An indicator variable is potentially migration literature and taking into consideration less subject to measurement or reporting errors as it is the available data, the study uses two different varia- likely that the respondent remembers whether some- bles as instruments for the migration decision: (i) the one in the household received remittances in the past share of households in the community21 with at least year but may not remember or may not know the exact one migrant; and (ii) the family migration history. amount received over the whole year. Vector X includes: individual characteristics (age, age squared, marital status, education, and ethnic 21 The community is the ward in Nepal and the village in Senegal. 30 The Case of Nepal and Senegal The first instrument is a proxy for the current migra- another country. Similar to migration networks, this tion network22 at the place of origin; it is constructed instrument is expected to influence the migration from the listing data collected before the survey. decision through increased information regarding Both the current migration network in the commu- migration experiences, and through reduced costs nity (Acosta 2006; Binzel and Assaad 2011) and the related to undertaking the trip and finding a job. historical migration network (Lokshin and Glinskaya 2009; Mendola and Carletto 2012) have been used in In both Nepal and Senegal, a listing of the households the literature. In general, the extent of the migration in the study areas was carried out prior to implemen- network should influence the decision to migrate by tation of the survey. Slightly different information reducing costs and improving information regarding was collected during the listings in the two coun- migration. Using the historical migration network in tries. In Nepal, enumerators only recorded whether the community as an instrument is potentially a bet- there were current or recent migrants in the house- ter solution for the reverse causality problem, but no hold, regardless of the type of migration; in Senegal, data are available on historical migration networks. more detailed information about the destination of As migration networks take a long time to develop, migrants was collected. Therefore, for Nepal having the current network is likely a result of many years of a current international migrant in the household migration flows rather than a recent phenomenon, (M1h) is instrumented with the current migration and should thus be a valid exogenous instrument.23 network at the ward level and with family migra- tion history. For Senegal, two potential endogenous The second instrument is the family migration his- regressors are used —both international (M1h) and tory. This indicator equals one if the parents or par- internal migration (M2h)—and three instruments: ents-in-law of the household head have ever lived in international migration network, internal migration network, and family migration history.24 22 The 2010 Nepal Living Standards Measurement Survey (Nepal LSMS) is potentially a good source for constructing a measure of the historical While there are good instruments for having an inter- migration network in a ward or a village. Due to time constraints, it was not national (M1h) or an internal migrant in the house- possible to explore how best to match the information in the Nepal LSMS hold (M2h), no exogenous instruments exist for the with the information in this survey. This exercise will be conducted later. decision to send remittances. Therefore, model 3 is 23 Migration networks variables and the migration history are theoretically good instruments as they are correlated with the endogenous variable estimated using ordinary least squares (OLS) and the (migration) and conceptualized to have an effect on the employment findings are interpreted as associations. outcomes of women only through their effect on the migration status of the household. Given F-Statistics of larger than 10, it can be reason- ably argued that the instrument is not weak, and the Sargan-Hansen test confirms that the instruments are exogenous. The results were also run with the endogenous regressor (without correcting for endogeneity); the 24 Yet findings for Senegal should be interpreted more cautiously, since results were qualitatively the same as the results from the two-stage least migration history can be tracked to (at least) the previous generation in squares (2SLS) model. around 60 percent of cases. Male Outmigration and Women’s Work and Empowerment in Agriculture 31 CHAPTER TEN RESULTS LINKAGES BETWEEN MALE OUTMIGRATION AND WOMEN’S EMPLOYMENT There is no evidence that rural women in sending communities reduce their employment in response to the migration of male family members (see Annex D). In both Nepal and Senegal, a negative relationship is found between wom- en’s employment in any activity and the presence of international migrants in the household, but these coefficients are not statistically significant in any of the specifications. Likewise, the receipt of remittances does not influence women’s employment (Panel B, Table D1 and D2). Women in households with international migrants do not seem to reduce over- all employment, but depending on the social and cultural contexts in which migration takes place, women may experience changes in their roles and responsibilities on the family farm. Compared to women in households with no current migrants, women in households with international migrants in Nepal are significantly more likely to report being self-employed on the farm (with or without employees) and less likely to report being contributing family work- ers. This implies that their responsibilities and decision-making on the farm increase with the outmigration of male family members. Panel A of Table D1 shows that women in households with international migrants are 17 percent- age points more likely to report being self-employed on the family farm rather than a contributing family worker. The coefficients are even larger when cor- recting for the endogeneity of migration; the results in Table F1 provide strong evidence that these are not merely associations between migration of family members and women’s changing roles on the farm, but that the changes are in fact attributed to the migration of the male family member. There is no evidence that women relocate labor in other activities, including outside of agriculture. This may be due to the need for labor in agriculture, or opera- tion of household farms remaining a dominant economic activity in rural areas (McCullough 2015), or limited employment opportunities for women in rural areas outside the family farm. Male Outmigration and Women’s Work and Empowerment in Agriculture 33 However, in Senegal, there is no such strong evidence last two years (2015 or afterwards), whether it was that the outmigration of male family members is between three and five years ago (between 2012 and associated with changes in women’s work on and off 2014), or if it was before 2012. In Nepal, no signifi- the farm. The lack of significant changes in women’s cant association is found between how much time roles in Senegal could be linked to the prevailing has passed since the current migrants first migrated social and cultural norms in the country. In particu- and the employment outcomes of women who stay lar, women’s roles in Senegal are prescribed to the behind. The results are robust to changes in the cut- domestic sphere and women are expected to be sup- offs of the variable and to the use of a continuous ported by their husbands. In the absence of their variable for time since first migration. In Senegal, husbands, they do not automatically become house- there is some evidence in households in which the hold heads. Household decision-making falls into the current migrant left in the last two years that women hands of the migrant’s extended family (Mondain et who stay behind increase their employment. How- al. 2011). Using qualitative methods, Mondain et al. ever, this result is only marginally significant; it is not (2011) looked into the linkages between male outmi- significant when a continuous variable is used. This gration and women’s roles in Senegal and concluded question deserves more attention in future work. that migration reinforces men’s status as primary earners and does not directly challenge existing gen- ASSOCIATIONS WITH der norms. WOMEN’S EMPOWERMENT Women’s growing self-employment in agriculture The tables in Annex E focus on the linkages between (that is, their growing role as primary farmers) is male outmigration and women’s empowerment in linked to both the migration of the spouse and the agriculture in several domains. As mentioned earlier, receipt of remittances. In Nepal, all migration is the abbreviated version of the Women’s Empower- linked to a change in women’s roles in agriculture ment in Agriculture Index (A-WEAI) was adminis- from contributing family workers to self-employment tered to one person per household: either the spouse in agriculture. Yet the effect is larger for women who of the migrant25 or the man or woman from the pri- live in households with international migrants who mary couple in nonmigrant households. Thus the send remittances compared to women who live in empowerment-related estimates based on the A-WEAI households with international migrants who do not are valid for only a subsample of women in the whole send any remittances. Remittances are strongly asso- population, unlike the employment-based estimates, ciated with women taking on more responsibilities which are valid for all working-age women. Although on the farm; this may be linked to the fact that in the sample on which the employment-related out- Nepal almost one-third of households invest some of comes discussed in the previous section are based the remittances on the farm. There is no evidence on is larger than the sample for the A-WEAI-related that male outmigration leads to changes in women’s estimates, its disadvantage is that the information probability of engaging in off-farm employment, for all individuals in the households was provided by which may be linked to the scant nonagricultural a single respondent, while for A-WEAI the selected employment opportunities in Nepal’s rural areas. women (and men) reported directly only about the activities and decisions that pertained to them. The linkages between migration and women’s employ- ment are not strongly dependent on the migration The evidence shows that male outmigration is not duration. Information about the timing of the first always positively linked to the empowerment of migration episode of the current migrants was used women who stay behind. Moreover, the results differ to create a proxy for the duration of migration. A categorical variable was constructed to differentiate 25 If the migrants were not married, the survey was administered to whether the first migration episode was within the another female member of the household. 34 The Case of Nepal and Senegal significantly by country. In Nepal, the spouses of In addition, the receipt of remittances is positively international migrants revealed that they reduced associated with increased decision-making on the the total number of productive activities that they farm, active participation in community groups, and participated in, which may be linked to the loss of access to a financial account. These are positive con- the male labor and the need for women to take over sequences of migration on women’s empowerment some of the tasks previously done by men. The results in Nepal, but they are restricted to women in house- from the A-WEAI provide suggestive evidence that in holds where the migrant sends back remittances. response to the migration of their spouses abroad, In fact, migration without remittances is associated women decrease participation in nonfarm activities; with negative effects, though mostly not statistically related to that, they also decrease decision-making significant, on almost all empowerment indicators of regarding nonagricultural income (Table E1). This women. is not necessarily disempowering if it is a choice and does not affect the welfare of the respondents. It is dis- The important role of remittances in mediating the empowering if it is not done out of choice, but out of effects of migration on women’s empowerment is necessity because of labor and time constraints, and evident in Senegal as well. With the exception of deci- if it reduces the welfare of the household through sions regarding credit, there is no evidence that male reduced diversification of livelihoods and potentially outmigration leads to Senegalese women’s empow- lower food security. erment. The analysis shows that in the absence of remittances, spouses of international migrants are These effects on empowerment are strongly medi- disempowered in several domains, including par- ated by the receipt of remittances, however. If the ticipation in productive activities in and outside agri- migration is accompanied by remittances, there is no culture, decision-making on productive activities, evidence of a reduction in the number of agricultural decision-making on the use of agricultural income, activities in which the spouses of migrants participate. and access to information about agriculture. Male Outmigration and Women’s Work and Empowerment in Agriculture 35 CHAPTER ELEVEN CONCLUSIONS This paper adds to the scarce evidence on rural outmigration and its interlink- ages with women’s employment and empowerment in agriculture. The poten- tial of migration to be a transformative factor for gender equality and women’s empowerment has attracted attention, but empirical research on the issue is limited. This study explores the linkages between rural outmigration and wom- en’s work, empowerment status, and food security from unique data collected specifically for this purpose. Migration from rural areas is increasing as more people (predominantly men) seek better opportunities to earn money. How- ever, it is important to note that, as with any economic action taken to improve household welfare, risks are involved with migration (i.e., migrants may not find lucrative job opportunities at their migration destination). The Nepal- ese data showed that only 45 percent of households with migrants reported receiving remittances, and the share is even lower in Senegal (30 percent). The remittance amount is found to be quite high in Nepal (more than double the average per capita GDP of the country), which would likely make the risk worth taking for many households. The situation is quite different in Senegal, where the average remittance amount for the 30 percent of households that reported receiving remittance is only about US$100 per year, making migration a risky move with less likelihood of “success.” The study finds that male outmigration from rural, primarily agricultural areas is not linked to a decrease in women’s employment, but it is associated with significant changes in women’s roles in agriculture. The study finds no evidence that living in a migrant-sending household causes women to decrease their overall participation in income-generating activities. In Nepal, male outmigra- tion from rural, primarily agricultural areas is strongly and significantly linked to changes in women’s roles in agriculture—women shift from being contrib- uting family members to being self-employed on the farm. These changes are stronger when migration is accompanied by remittances. Contrary to some pre- vious studies, the report does not find evidence that women in households with a family member who is currently abroad reduce their engagement in off-farm Male Outmigration and Women’s Work and Empowerment in Agriculture 37 wage employment and self-employment. On the there is no evidence that male outmigration leads to other hand, in Senegal male-dominated outmigra- women’s empowerment. Moreover, in the absence tion is not associated with changes in women’s roles of remittances, spouses of international migrants in agriculture. This is because most rural women are worse off in several domains of empowerment, in Senegal live in extended families in which other including the number of productive activities in members may take on the roles and responsibilities which they participate, decision-making on produc- of the migrant spouse. tive activities and agricultural income, and access to information. The study reveals that male-dominated outmigration may not always be associated with women’s empower- The consequences of migration on household food ment. Based on evidence from the A-WEAI that was security are country-specific and mediated by the administered to either the spouse of the migrant or receipt of remittances. The study finds that migra- the man or woman from the primary couple, male tion of household members that is not followed by outmigration is linked to empowerment in some remittance transfers is likely to increase household domains and disempowerment in others. The results food insecurity. The evidence is stronger and signifi- differ substantially by country. In Nepal, direct inter- cant in the case of Senegal, where both international views with spouses of migrants revealed that the and internal migration are positively associated with receipt of remittances is positively associated with food insecurity. In Nepal, no significant correlation increased decision-making on the farm, group mem- exists between migration and food security, but the bership, and holding a financial account. In Senegal, lack of significant results may be due to the rather with the exception of decisions regarding credit, small survey sample size. 38 The Case of Nepal and Senegal CHAPTER TWELVE POLICY RECOMMENDATIONS GENERALIZED POLICY RECOMMENDATION A more generalized and priority policy action emerging out of the analysis suggests the importance of recognizing the changing roles of women in agri- culture, and providing targeted interventions to support their roles. General policy actions are to: i. Encourage greater availability of gender-relevant, sex-disaggregated data to monitor the effects of male outmigration on women’s work and empowerment. The current practice of collecting and disseminating sex-disaggregated data is done in a scattered manner across different agencies. To identify tailored knowledge gaps and policies targeted specifically to women left behind after the outmigration of a male spouse, it is extremely important to improve the availability of evidence-based, targeted sur- veys and to centralize the survey packages for future research and policy dialogue. It is also important to build national capacity to collect and analyze sex-desegregated data covering migrant-sending and nonmigrant households in agriculture. This is a systematic pathway of providing policy makers with sufficient baseline information to institute favorable changes to existing policies, which currently affect women and men differently in migrant households. This will also form the basis of institutional- izing such rigorous evidence to strengthen existing and future World Bank opera- tions or multi-stakeholder programs that are targeted at women engaged in on-farm activities, where M&E systems are often less comprehensive in terms of progress on the various dimensions of women’s empowerment. ii. Facilitate the flow of international and internal remittances. Evidence from these studies indicates that remittances can influence significant changes in women’s roles in agriculture and are positively associated with women’s empowerment in several domains (such as decisions about farm, group membership, and holding a financial account for Nepal and access to decisions about credit for Senegal). One way to facilitate remittance transfers would be to reduce the cost of sending remittances. Sustainable Development Goal (SDG) 10 aims to reduce the cost of remittances to three percent by 2030 and eliminate remittance corridors with costs higher than five percent. This will be an avenue to formalize remittances channels. One key constraint in Nepal, especially in the mountain and hill areas, is the lack of access to financial services. iii. Enact policies to support women’s engagement in higher-earning activities. A smaller share of women in Senegal than in Nepal report being economically active. There Male Outmigration and Women’s Work and Empowerment in Agriculture 39 is a need to better understand women’s low participa- extension. A concerted involvement of decentralized tion in the labor market in Senegal, but apart from that, government bodies, NGOs, private agencies, and indi- women who are economically active are largely con- viduals can create an enabling environment. centrated in the production end of agricultural value chains. Very few women in either Nepal or Senegal engage in processing or trade of agricultural products. Addressing Labor Shortages i. Promote small-scale rural mechanization to reduce COUNTRY-SPECIFIC POLICY women’s time burden and improve diversification of income-generating activities in Nepal (Biggs and Justice RECOMMENDATIONS 2015). As suggested by the results, women in migrant households in Nepal are more overworked and time- A set of policy recommendations was derived for constrained compared to both men and women in non- each country. Each set addresses the country-specific migrant households. This may be due to the scarcity of challenges identified in this study. agricultural labor and low access to labor-saving tech- nologies for Nepalese women. NEPAL The following approaches appear promising in Improving Enabling Environment for addressing the problems identified by the study: Productive Use of Remittances by Female Farmers Adapting Agricultural Extension i. Reduce the cost of remittances to create an enabling i. Provide tailored extension services to female farmers. environment for women to mobilize remittances for The study finds that as a result of male outmigration productive purposes, including more investments in in Nepal, the on-farm responsibilities and decision- agriculture or small businesses and savings through making of the women left behind increase. In Nepal, development of money management skills (Dhakal and all migration is linked to a change in women’s roles in Maharjan 2018). In certain areas of Nepal the cost of agriculture from contributing family workers to self- remittances is quite high. Currently, at least some of employment in agriculture, and the effect is larger the remittances are used for the purchase of food, but for women who live in households with international a non-negligible amount is also invested in agriculture. migrants who send remittances. This clearly entails the need for improving female farmers’ access to extension services to improve the productivity on their farms and SENEGAL ensure the sustainability of agricultural production. The study finds no significant association between male ii. Strengthen women’s access to higher-earning activities outmigration and women’s employment and empow- in agricultural value chains. The study shows very low erment in Senegal. That said, the important role of engagement in higher value chain activities such as pro- remittances in mediating the effects of migration on cessing and trading, which can be linked to women’s women’s empowerment is evident in Senegal as well. low skills, lack of access to market information, and transportation and time constraints. Extension services The following approaches appear promising in for women should go beyond the traditional focus on production and should provide technical assistance, addressing the problems identified by the study: training, and access to resources that can scale up wom- en’s involvement beyond subsistence agriculture and in the higher-value nodes of the supply chains. Reducing the Cost of Remittances iii. Ensure that a gender-sensitive approach is adopted for i. Reduce the cost of remittances to positively affect the provision of agricultural extension services, includ- disposable household income and improve incentives ing through hiring more female agricultural extension to remit more (World Bank 2005). The cost of send- agents. Studies have shown positive experiences with ing remittances through formal channels is very high hiring female extension agents to better support female in Senegal, a situation accompanied by a high gender farmers (Acharya and Bennet 1983; World Bank 2010) disparity in the receipt of remittances—male-headed and the importance of local groups for mobilizing public households receive higher remittances than female- awareness to mainstream gender balance in agriculture headed ones (Orozco et al. 2010). Positive remittances 40 The Case of Nepal and Senegal will also help mitigate the negative effects from the lost other factors, which may also be correlated with migra- labor of migrants and therefore will help mitigate the tion, including household demographics. The presence negative effects on women’s empowerment. of larger extended families may facilitate migration but ii. Conduct more research to understand the factors behind may also mediate the potential transformative effects of the low economic activity status of women in Senegal. migration on spouses who stay behind. Therefore, in an A very small share of women in Senegal report hav- environment with low employment rates for both men ing engaged in any work activity in the last 12 months. and women and large extended families, the migration Although women in migrant households have even of male family members is less likely to lead to significant lower employment rates than women in nonmigrant changes in women’s employment and empowerment, as households, the analysis suggests that the lower employ- other family members can step in to do the work of the ment probability is not attributed to migration but to migrant man or to make decisions in his absence. Male Outmigration and Women’s Work and Empowerment in Agriculture 41 ANNEX A: THE ABBREVIATED WOMEN’S EMPOWERMENT IN AGRICULTURE INDEX (A-WEAI) USED IN NEPAL AND SENEGAL SURVEYS TABLE A1. D  OMAINS AND INDICATORS FROM THE ABBREVIATED WOMEN’S EMPOWERMENT IN AGRICULTURE INDEX (A-WEAI) USED IN NEPAL AND SENEGAL SURVEYS DOMAIN INDICATOR DEFINITION OF INDICATOR 1. Production Input in productive 1.1  • Number of agricultural and nonagricultural activities in which an indi- decisions vidual participates • Number of agricultural production activities in which an individual participates • Whether respondent has sole or joint decision-making over food and cash-crop farming, livestock, and fisheries • Whether respondent makes decisions about what to plant on ANY land 1.2 Access to information • Whether respondent has access to information for at least ONE agricul- tural activity 2. Resources 2.1 Ownership of assets • Whether respondent solely or jointly owns AT LEAST two small assets • Whether respondent owns land solely or jointly Access to and decisions • Whether respondent has access to and participates in decision-making 2.2  about credit concerning credit • Whether respondent has access to a financial account 3. Income Control over the use of • Whether respondent decides about the use of agricultural income 3.1  income • Whether respondent decides about the use of nonagricultural income 4. Leadership 4.1 Group member • Whether respondent is an active member in at least one economic or social group 5. Time 5.1 Workload • Minutes spent on work • Whether respondent worked less than 10.5 hours in the previous 24 hours Source: While the domains of empowerment are the same as in Alkire et al. (2013), the selected indicators for the analysis may differ because the A-WEAI was implemented, rather than the WEAI, and some additional indicators were added. Male Outmigration and Women’s Work and Empowerment in Agriculture 43 TABLE A2. EMPOWERMENT OUTCOMES BY SEX IN NEPAL A-WEAI SAMPLE WOMEN MEN   N MEAN SE N MEAN SE P-VALUE Production # of work activities 724 2.83 0.05 271 2.78 0.04 # of agriculture activities 724 2.67 0.05 271 2.41 0.03 *** Input in decision-making in AT LEAST TWO 697 0.96 0.01 260 0.98 0.01 productive domains† Decision-making, solely or jointly, land† 692 0.85 0.02 260 0.92 0.02 Access to agriculture information† 696 0.93 0.01 259 0.98 0.01 ** Resources Respondent owns assets, solely or jointly† 724 0.99 0.00 271 1.00 0.00 Respondent owns land, solely or jointly† 692 0.32 0.02 260 0.64 0.04 *** Decision-making on credit† 724 0.48 0.02 271 0.50 0.04 Has a bank account† 724 0.52 0.02 271 0.48 0.04 Income Decision-making: agricultural income† 724 0.93 0.01 271 0.95 0.01 Decision-making: nonagricultural income† 724 0.15 0.02 271 0.35 0.04 Leadership Membership (any group)† 724 0.52 0.02 271 0.41 0.04 *** Time use # minutes work 724 589.66 6.69 271 454.62 15.27 *** Respondent worked <10.5hrs in previous 24hrs† 724 0.51 0.02 271 0.79 0.03 *** * the difference is significant at the 10% level; ** – at the 5%: *** – at the 1% level. † A dummy variable. ‡ An active member of that group. SE = standard error. 44 The Case of Nepal and Senegal TABLE A3. EMPOWERMENT OUTCOMES BY SEX IN SENEGAL A-WEAI SAMPLE WOMEN MEN   N MEAN SE N MEAN SE P-VALUE Production # of work activities 534 0.56 0.05 375 0.94 0.04 *** # of agriculture activities 534 0.50 0.05 375 0.84 0.04 *** Input in decision-making in AT LEAST 534 0.30 0.03 375 0.47 0.02 *** TWO productive domains† Decision-making, solely or jointly, land† 353 0.36 0.07 303 0.94 0.01 *** Access to agriculture information† 532 0.26 0.02 372 0.41 0.02 *** Resources Respondent owns assets, solely or jointly† 534 0.86 0.01 375 0.86 0.02 Respondent owns land, solely or jointly† 352 0.56 0.03 303 0.88 0.03 *** Decision-making on credit† 534 0.14 0.02 375 0.24 0.01 *** Has a bank account† 534 0.03 0.01 375 0.06 0.01 ** Income Decision-making: agricultural income† 534 0.26 0.02 375 0.39 0.02 *** Decision-making: nonagricultural income† 534 0.03 0.01 375 0.08 0.00 *** Leadership Membership (any group)† 534 0.33 0.02 375 0.32 0.02 * the difference is significant at the 10% level; ** – at the 5%: *** – at the 1% level. † A dummy variable. ‡ An active member of that group. SE = standard error. Male Outmigration and Women’s Work and Empowerment in Agriculture 45 ANNEX B: DESCRIPTIVE ANALYSIS OF KEY VARIABLES TABLE B1. CHARACTERISTICS OF FEMALE FAMILY MEMBERS, NEPAL (1) WOMEN IN HOUSEHOLD WITH (2) WOMEN FROM ALL INTERNATIONAL MIGRANTS OTHER HOUSEHOLDS VARIABLE N MEAN SE N MEAN SE P-VALUE Individual Characteristics Age (years) 763 36.61 0.69 904 37.38 0.75 Female† 763 1.00 904 1.00 Married† 763 0.78 0.02 904 0.73 0.02 * Never married† 763 0.14 0.01 904 0.16 0.02 Cohabiting† 763 0.00 0.00 904 0.00 0.00 Widowed/separated† 763 0.08 0.01 904 0.10 0.01 No education† 763 0.44 0.02 904 0.44 0.02 Primary education† 763 0.12 0.01 904 0.16 0.02 Secondary education† 763 0.44 0.02 904 0.40 0.02 High caste† 763 0.43 0.02 904 0.42 0.02 Low caste† 763 0.17 0.02 904 0.10 0.01 *** Other caste† 763 0.37 0.02 904 0.47 0.02 *** Muslim† 763 0.03 0.01 904 0.01 0.00 ** Household Characteristics # children <5 years 763 0.49 0.03 904 0.38 0.02 *** # children 5-10 years 763 0.56 0.03 904 0.53 0.03 # males 11-14 years 763 0.20 0.02 904 0.22 0.02 # females 11-14 years 763 0.18 0.02 904 0.18 0.02 # males 15-17 years 763 0.19 0.02 904 0.17 0.02 # females 15-17 years 763 0.19 0.02 904 0.24 0.02 # female adults 763 2.11 0.04 904 1.81 0.03 *** # male adults 763 2.15 0.04 904 1.65 0.03 *** Note: * the difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. Male Outmigration and Women’s Work and Empowerment in Agriculture 47 TABLE B2. CHARACTERISTICS OF FEMALE FAMILY MEMBERS, SENEGAL (1) WOMEN IN HOUSEHOLD (2) WOMEN IN (3) WOMEN IN WITH HOUSEHOLD HOUSEHOLD INTERNATIONAL WITH INTERNAL WITH NO MIGRANTS MIGRANTS MIGRANTS (1) VS (3) N MEAN SE N MEAN SE N MEAN SE P-VALUE Individual Characteristics Age (years) 894 35.71 2.34 981 35.29 2.02 951 34.95 1.84 Female† 894 1.00 981 1.00 951 1.00 Never married† 894 0.14 0.03 981 0.16 0.04 951 0.16 0.06 Married monogamous† 894 0.48 0.03 981 0.43 0.05 951 0.44 0.07 Married polygamous† 894 0.25 0.05 981 0.28 0.03 951 0.27 0.04 Widowed/divorced† 894 0.14 0.04 981 0.13 0.03 951 0.13 0.03 No education† 894 0.77 0.03 981 0.79 0.04 951 0.81 0.03 ** Primary education† 894 0.07 0.01 981 0.06 0.01 951 0.06 0.00 Secondary education† 894 0.15 0.03 981 0.15 0.03 951 0.13 0.03 * Ethnicity: Pular† 894 0.77 0.02 981 0.49 0.02 951 0.50 0.02 *** Ethnicity: Sirer† 894 0.07 0.02 981 0.13 0.02 951 0.22 0.01 *** Ethnicity: Wolof/Libou† 894 0.14 0.01 981 0.36 0.02 951 0.24 0.01 *** Household Characteristics # children <5 years 894 1.12 0.08 981 1.26 0.10 951 1.39 0.08 ** # children 5-10 years 894 1.85 0.10 981 2.05 0.13 951 2.07 0.09 *** # males 11-14 years 894 0.53 0.03 981 0.63 0.07 951 0.54 0.03 # females 11-14 years 894 0.45 0.04 981 0.45 0.03 951 0.54 0.03 *** # males 15-17 years 894 0.50 0.04 981 0.48 0.03 951 0.42 0.02 ** # females 15-17 years 894 0.51 0.04 981 0.71 0.06 951 0.45 0.04 # female adults 894 4.62 0.26 981 3.73 0.24 951 3.16 0.27 *** # male adults 894 4.14 0.17 981 3.63 0.17 951 2.52 0.14 *** Note: * the difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. 48 The Case of Nepal and Senegal TABLE B3.  EMPLOYMENT CHARACTERISTICS BY INTERNATIONAL MIGRATION EXPERIENCE FOR ALL WORKING-AGE ADULTS AND FOR WORKING-AGE WOMEN ONLY, NEPAL 1) HOUSEHOLD 2) HOUSEHOLD WITH A CURRENT WITH NO CURRENT INTERNATIONAL INTERNATIONAL MIGRANT MIGRANT   MEAN SE MEAN SE P-VALUE A. All working-age adults Employment, any occupation† 0.884 0.010 0.898 0.009 Farm self-employed† 0.361 0.017 0.334 0.015 Farm contributing family worker† 0.552 0.017 0.584 0.015 Agricultural laborer† 0.050 0.007 0.087 0.008 *** Processing (agricultural products)† 0.048 0.008 0.061 0.010 Trading (agricultural products)† 0.010 0.005 0.008 0.002 Nonagricultural employment† 0.075 0.009 0.166 0.012 *** Observations 1181 1726 B. Working-age women only Employment, any occupation† 0.892 0.013 0.893 0.013 Farm self-employed† 0.323 0.020 0.197 0.018 *** Farm contributing family worker† 0.594 0.021 0.722 0.019 *** Agricultural laborer† 0.049 0.008 0.069 0.010 Processing (agricultural products)† 0.046 0.010 0.068 0.014 Trading (agricultural products)† 0.003 0.003 0.001 0.001 Nonagricultural employment† 0.037 0.008 0.054 0.009 Observations 763 904 Note: * the difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. Male Outmigration and Women’s Work and Empowerment in Agriculture 49 TABLE B4. E  MPLOYMENT CHARACTERISTICS BY MIGRATION STATUS FOR ALL WORKING- AGE ADULTS AND FOR WORKING-AGE WOMEN, SENEGAL (2) HOUSEHOLD (3) (1) HOUSEHOLD WITH A HOUSEHOLD WITH A CURRENT CURRENT WITH NO INTERNATIONAL INTERNAL CURRENT MIGRANT MIGRANT MIGRANT (1) VS (3)     MEAN SE MEAN SE MEAN SE P-VALUE A. All working-age adults Working, any occupation† 0.392 0.015 0.604 0.013 0.612 0.016 *** Farm self-employed† 0.047 0.005 0.097 0.009 0.136 0.015 *** Farm contributing family worker† 0.245 0.014 0.425 0.013 0.396 0.013 *** Agricultural laborer† 0.002 0.001 0.010 0.003 0.007 0.002 *** Processing (agricultural products)† 0.002 0.001 0.005 0.002 0.010 0.002 *** Trading (agricultural products)† 0.018 0.003 0.041 0.005 0.032 0.003 *** Nonagricultural laborer† 0.091 0.008 0.095 0.007 0.098 0.010 Observations 1428 1694 1849 B. All working-age women Working, any occupation† 0.255 0.015 0.477 0.014 0.467 0.025 *** Farm self-employed† 0.017 0.004 0.043 0.005 0.062 0.008 *** Farm contributing family worker† 0.171 0.015 0.362 0.014 0.342 0.021 *** Agricultural laborer† 0.001 0.001 0.003 0.002 0.006 0.002 ** Processing (agricultural products)† 0.001 0.001 0.009 0.003 0.011 0.004 ** Trading (agricultural products)† 0.016 0.003 0.033 0.006 0.026 0.004 * Nonagricultural laborer† 0.041 0.006 0.046 0.005 0.045 0.009 Observations 894 981 951 Note: * the difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. 50 The Case of Nepal and Senegal TABLE B5. WOMEN’S EMPOWERMENT OUTCOMES BY MIGRATION STATUS, NEPAL WOMEN ONLY, A-WEAI SAMPLE 1) HOUSEHOLD WITH A CURRENT (2) HOUSEHOLD INTERNATIONAL WITH NO CURRENT MIGRANT MIGRANT   N MEAN SE N MEAN SE P-VALUE Production # of work activities 421 2.76 0.07 303 2.92 0.04 ** # of agriculture activities 421 2.64 0.05 303 2.72 0.04 ** Input in decision-making in AT LEAST TWO 408 0.96 0.01 289 0.96 0.01 productive domains† Decision-making, solely or jointly, land† 405 0.86 0.07 287 0.85 0.08 Access to agriculture information† 407 0.91 0.04 289 0.95 0.01 Resources Respondent owns assets, solely or jointly† 421 0.99 0.00 303 1.00 0.00 Respondent owns land, solely or jointly† 405 0.33 0.03 287 0.31 0.05 Decision-making on credit† 421 0.46 0.03 303 0.51 0.03 * Has a bank account† 421 0.55 0.01 303 0.48 0.06 Income Decision-making: agricultural income† 421 0.93 0.03 303 0.94 0.01 Decision-making: nonagricultural income† 421 0.11 0.01 303 0.19 0.02 ** Leadership Membership (any group)† 421 0.56 0.03 303 0.47 0.00 ** Time use # minutes worked 421 593.01 16.12 303 585.43 2.80 Respondent worked <10.5 hours in previous 24 421 0.48 0.05 303 0.55 0.03 * hours† Note: * The difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. Male Outmigration and Women’s Work and Empowerment in Agriculture 51 TABLE B6. WOMEN’S EMPOWERMENT OUTCOMES BY MIGRATION STATUS, SENEGAL WOMEN ONLY, A-WEAI SAMPLE (1) HOUSEHOLD (2) HOUSEHOLD WITH A WITH A (3) HOUSEHOLD CURRENT CURRENT WITH NO INTERNATIONAL INTERNAL CURRENT (1) vs   MIGRANT MIGRANT MIGRANT (3)     N MEAN SE N MEAN SE N MEAN SE P-VALUE Productive activities # of work activities 153 0.51 0.08 181 0.78 0.05 200 0.52 0.06 # of agriculture activities 153 0.44 0.08 181 0.71 0.05 200 0.47 0.06 Input in decision-making in AT LEAST 153 0.26 0.04 181 0.35 0.03 200 0.29 0.03 TWO productive domains† Decision-making, solely or jointly, land† 96 0.37 0.10 111 0.34 0.09 146 0.36 0.07 Access to agriculture information† 153 0.20 0.03 181 0.31 0.03 198 0.25 0.03 Asset ownership Respondent owns assets, solely or jointly† 153 0.78 0.03 181 0.88 0.02 200 0.87 0.01 ** Respondent owns land, solely or jointly† 96 0.56 0.05 110 0.55 0.05 146 0.56 0.03 Decision-making on credit† 153 0.13 0.02 181 0.27 0.03 200 0.11 0.02 Has a bank account† 153 0.05 0.02 181 0.02 0.01 200 0.03 0.01 * Decision-making: land and income Decision-making: agricultural income† 154 0.21 0.03 181 0.32 0.02 200 0.25 0.03 Decision-making: nonagricultural 154 0.06 0.02 181 0.05 0.02 200 0.02 0.01 ** income† Group membership Membership (any group)† 154 0.26 0.06 181 0.35 0.02 200 0.33 0.02 Note: * The difference is significant at the 10% level; ** -- at the 5%: *** -- at the 1% level. † A dummy variable. SE = standard error. 52 The Case of Nepal and Senegal ANNEX C: ASSOCIATION BETWEEN FOOD INSECURITY EXPERIENCE SCALE AND MIGRATION STATUS TABLE C1.  THE CORRELATION BETWEEN MIGRATION STATUS, REMITTANCES, AND HOUSEHOLD FOOD INSECURITY, NEPAL FImod+sev FImod+sev FImod+sev FImod+sev FImod+sev     (1) (2) (3) (4) (5) International migrant in household –0.00689 0.00387 0.00375 0.00767 0.00944 (0.0188) (0.0277) (0.0215) (0.0252) (0.0225) Remittances –0.0146 (0.0293) Total remittances in US$, outliers removed –5.84e–06 –1.16e–05 1.04e–05 (4.34e–06) (1.63e–05) (1.72e–05) Total remittances SQUARED in US$, outliers 9.62e–10 removed (2.28e–09) Total remittances in US$ INTERACTED with –1.77e–05 ABROAD migration, outliers removed (1.73e–05) Observations 994 994 994 994 994 Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 All models also include the following controls: Household head age and age squared; marital status; sex; education; whether the head belongs to high or low caste; whether the head is a Muslim; the maximum education achieved by anyone in the household; household demographic structure (the number of children under 5, children 5-10 years old, male and female children 11-14 years old, males and females 15-17 years old, number of adult men and adult women in the household); wealth variables (including material of walls, roof, and floor, the type of toilet, access to electricity, access to piped water, whether the drinking water source is on the household grounds, whether the household owns land and area of land owned, livestock ownership measured in TLU); whether the respondent is a woman; and district-level dummies. Male Outmigration and Women’s Work and Empowerment in Agriculture 53 TABLE C2.  THE CORRELATION BETWEEN MIGRATION STATUS, REMITTANCES, AND HOUSEHOLD FOOD INSECURITY, SENEGAL FImod+sev FImod+sev FImod+sev FImod+sev FImod+sev     (1) (2) (3) (4) (5) International migrant in –0.0149 0.0450 –0.00858 0.0280 0.0239 household† (0.0522) (0.0575) (0.0544) (0.0523) (0.0533) 0.0509 0.0920** 0.0527 0.0738* 0.0640 Internal migrant in household† (0.0406) (0.0450) (0.0411) (0.0403) (0.0418) –0.128** Remittances† (0.0534) Total remittances in US$, outliers –4.72e–05 –0.00069*** 0.000204* removed (0.000109) (0.00016) (0.000111) Total remittances SQUARED in 3.81e–07*** US$, outliers removed (1.05e–07) Total remittances in US$ –0.00044*** INTERACTED with ABROAD migration, outliers removed (0.00013) Total remittances in US$ –0.000346** INTERACTED with INTERNAL migration, outliers removed (0.000147) Observations 976 976 976 976 976 Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 All models also include the following controls: Household head age and age squared; marital status; sex; education; whether the head belongs to Pular, Sirer, or Wolof/Libou ethnic groups; the maximum education achieved by anyone in the household; household demographic structure (the number of children under 5, children 5–10 years old, male and female children 11–14 years old, males and females 15-17 years old, number of adult men and adult women in the household); wealth variables (including material of walls, roof, and floor, the type of toilet, access to electricity, access to piped water, whether the drinking water source is on the household grounds, whether the household owns land and area of land owned, livestock ownership measured in TLU); whether the respondent is a woman; and whether the household receives social transfers. 54 The Case of Nepal and Senegal ANNEX D: REGRESSION OF INTEREST (EMPLOYMENT OUTCOMES) Male Outmigration and Women’s Work and Empowerment in Agriculture 55 56 TABLE D1. THE IMPACT OF MIGRATION ON EMPLOYMENT OUTCOMES FOR WOMEN, NEPAL Farm Agricultural Processing Trading Employed Farm self- contributing (wage) (agricultural (agricultural Nonagricultural (any) employed family workers laborers products) products) workers Professional VARIABLES (1) (2) (3) (4) (5) (6) (7) (8) A. Base model - no controls for remittances (N=1667) , OLS International migrant in –0.00508 0.167*** –0.177*** 0.00199 –0.0332** 0.00309 –0.00604 0.00298 household (0.0174) (0.0241) (0.0274) (0.0118) (0.0168) (0.00382) (0.0124) (0.00952) B. Controlling for migration and remittances (N=1618‡), OLS Household with an 6.71e–05 0.214*** –0.218*** –0.00104 –0.0400** 0.00311 0.000198 0.00227 international migrant, with remittances (0.0186) (0.0252) (0.0291) (0.0134) (0.0188) (0.00419) (0.0130) (0.0103) Household with an –0.0419 0.0745* –0.135*** –0.0326 –0.00817 0.00372 0.00239 –0.00703 international migrant, no remittances (0.0427) (0.0425) (0.0512) (0.0230) (0.0268) (0.00289) (0.0203) (0.00940) Internal migrant in –0.0234 0.190*** –0.252*** –0.0320* –0.0403 0.000553 0.0248 –0.00813 household (0.0382) (0.0499) (0.0589) (0.0190) (0.0337) (0.00174) (0.0278) (0.00774) Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 ‡ For greater clarity, women in households that receive remittances but do not have an international migrant are excluded from the estimation in Panel B (these women constitute around 3 percent of the finale female sample). In Panel B the base category includes households with no internal or international migrants that do not receive remittances either. All models also include the following controls: age; age squared; marital status; educational attainment; whether the woman is high caste or low caste; whether she is Muslim; household demographic structure (the number of children under 5, children 5–10 years old, male and female children 11–14 years old, males and females 15–17 years old, number of adult men and adult women in the household); wealth variables (including material of walls, roof, and floor, the type of toilet, access to electricity, access to piped water, whether the drinking water source is on the household grounds, whether the household owns land and area of land owned, livestock ownership measured in TLU); and district dummies. The Case of Nepal and Senegal TABLE D2. THE IMPACT OF MIGRATION ON EMPLOYMENT OUTCOMES FOR WOMEN, SENEGAL Farm contributing Agricultural Processing Trading Employed Farm self- family (wage) (agricultural (agricultural Nonagricultural (any) employed workers laborers products) products) work Professional Other VARIABLES (1) (2) (3) (4) (5) (6) (7) (8) (9) A. Base model - no controls for remittances (N=2826), OLS International migrant –0.0188 –0.00442 –0.0122 –0.00325 –0.000912 –0.00610 –0.00512 0.00782 0.0119* in household (0.0248) (0.00970) (0.0213) (0.00253) (0.00355) (0.00805) (0.0125) (0.00497) (0.00661) Internal migrant in 0.0214 –0.0121 0.0279 –0.000696 –0.00388 0.00544 0.00286 0.00616 –0.00427 household (0.0236) (0.0121) (0.0222) (0.00370) (0.00543) (0.00823) (0.0112) (0.00487) (0.00805) B. Controlling for migration and remittances (N=2795‡), OLS Household with an –0.0396 –0.00372 –0.0243 –0.00455* –0.00321 –0.0110 0.00723 0.00347 0.00505 international migrant, with remittances (0.0285) (0.0117) (0.0247) (0.00237) (0.00463) (0.00889) (0.0145) (0.00627) (0.00738) Household with an –0.0173 –0.0163 –0.0113 –0.00225 0.00133 –0.00179 –0.0203 0.0131** 0.0198** international migrant, no remittances (0.0325) (0.0105) (0.0280) (0.00415) (0.00288) (0.0100) (0.0150) (0.00657) (0.00960) Male Outmigration and Women’s Work and Empowerment in Agriculture Internal migrant in 0.00421 –0.0172 0.0159 –0.000933 –0.00457 0.00358 0.00462 0.00559 –0.00607 household (0.0243) (0.0124) (0.0229) (0.00376) (0.00578) (0.00826) (0.0115) (0.00506) (0.00885) Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 ‡ For greater clarity, women in households that receive remittances but do not have any migrants are excluded from the model. About one percent of women in the final sample belong to households that report receiving remittances although they do not have an international migrant. In Panel B the base category includes households with no internal or international migrants that do not receive remittances. All models also include the following controls: age; age squared; marital status; educational attainment; whether the women is Pular, Sirer, or Wolof/Libou ethnicity; household demographic structure (the number of children under 5, children 5–10 years old, male and female children 11–14 years old, males and females 15–17 years old, number of adult men and adult women in the household); wealth variables (including material of walls, roof, and floor, the type of toilet, access to electricity, access to piped water, whether the drinking water source is on the household grounds, whether the household owns land and area of land owned, livestock ownership measured in TLU); whether the household receives social transfers; and department dummies. 57 ANNEX E: REGRESSION OF INTEREST (EMPOWERMENT OUTCOMES) Male Outmigration and Women’s Work and Empowerment in Agriculture 59 60 THE ASSOCIATION BETWEEN MIGRATION (WITH AND WITHOUT REMITTANCES) AND THE EMPOWERMENT TABLE E1.  OF WOMEN, NEPAL, OLS participates which individual # of activities in participates in which individual # of AG activities AT LEAST 2 domains Input in decisions in least 1 AG activity Access info for at assets AT LEAST two small Solely or jointly owns about credit Makes decisions account Access to a financial on ANY land about what to plant Makes decisions jointly owns land Resp. solely or use of AG income Decides about the income use of non-AG Decides about the community group Member of at least 1 work Minutes spent on in previous 24 hours less than 10.5 hours Respondent worked (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) A. Base model - no controls for remittances, OLS International –0.185* –0.0933 0.0013 –0.0077 2.16e–05 –0.0249 0.0386 0.070** 0.0636 0.0136 –0.089*** 0.120*** –4.851 –0.0315 migrant in household (0.097) (0.096) (0.018) (0.021) (0.005) (0.049) (0.042) (0.031) (0.043) (0.022) (0.034) (0.046) (12.59) (0.045) Observations 726 726 699 698 726 726 726 694 694 726 726 726 726 726 B. Controlling for migration and remittances‡, OLS Household with –0.223** –0.104 –0.0131 0.00165 0.00458 0.00820 0.0839* 0.0842** 0.0679 0.0305 –0.116*** 0.153*** 6.120 –0.0660 an international migrant, with remittances (0.107) (0.107) (0.0152) (0.0216) (0.00606) (0.0541) (0.0469) (0.0335) (0.0463) (0.0252) (0.0391) (0.0518) (13.94) (0.0500) Household with –0.547** –0.418* –0.0207 –0.141** –0.0221 –0.0185 –0.0626 0.0743 –0.0625 –0.0834 –0.109* –0.0719 –26.68 0.0558 an international migrant, no remittances (0.230) (0.218) (0.0406) (0.0695) (0.0351) (0.0868) (0.0787) (0.0735) (0.0698) (0.0595) (0.0631) (0.0843) (26.36) (0.0906) Internal migrant in –0.271 –0.118 –0.0479 –0.0341 0.0129 0.0864 0.0803 0.0589 –0.0935 0.0158 –0.145** 0.0862 19.62 –0.0854 household (0.179) (0.179) (0.0460) (0.0557) (0.0105) (0.0928) (0.0997) (0.0485) (0.0671) (0.0501) (0.0588) (0.0814) (25.20) (0.0913) Observations 706 706 680 679 706 706 706 675 675 706 706 706 706 706 Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. OLS = ordinary least squares. ‡ For greater clarity, women in households that receive remittances but do not have an international migrant are excluded from the estimation in Panel B. In Panel B the base category includes women in households with no internal or international migrants that do not receive remittances. All models include the same controls as in Table B1. The Case of Nepal and Senegal TABLE E2. THE IMPACTS OF MIGRATION AND REMITTANCES ON THE EMPOWERMENT OF WOMEN, SENEGAL, OLS participates which individual # of activities in participates in which individual # of AG activities domains in AT LEAST 2 Input in decisions least 1 AG activity Access info for at two small assets owns AT LEAST Solely or jointly about credit Makes decisions financial account Access to a plant on ANY land about what to Makes decisions jointly owns land Resp. solely or use of AG income Decides about the income use of non-AG Decides about the group least 1 community Member of at   (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Female Respondents International migrant in –0.182 –0.167 –0.122* –0.141** –0.0434 0.0792** 0.0301 –0.0141 –0.109 –0.119* 0.00515 –0.0153 household (0.137) (0.132) (0.0670) (0.0636) (0.0492) (0.0400) (0.0267) (0.0837) (0.0814) (0.0658) (0.0286) (0.0594) Internal migrant in 0.139 0.110 –0.00132 –0.00121 0.00139 0.159*** –0.000650 0.0219 –0.0343 0.00777 0.0340* –0.0317 household (0.126) (0.122) (0.0594) (0.0569) (0.0399) (0.0452) (0.0186) (0.0725) (0.0709) (0.0580) (0.0181) (0.0527) Observations 534 534 534 532 534 534 534 353 352 535 535 535 Household with an –0.0730 –0.0654 –0.0834 –0.124* –0.0746 0.0566 0.0517 –0.0151 –0.114 –0.0844 0.00742 –0.0195 international migrant, Male Outmigration and Women’s Work and Empowerment in Agriculture with remittances (0.163) (0.155) (0.0774) (0.0736) (0.0587) (0.0437) (0.0328) (0.109) (0.103) (0.0753) (0.0336) (0.0701) Household with an –0.334** –0.313** –0.175** –0.164** 0.00850 0.118** –0.00232 –0.00406 –0.105 –0.163** 0.00293 –0.0191 international migrant, no remittances (0.150) (0.141) (0.0790) (0.0745) (0.0618) (0.0570) (0.0344) (0.0923) (0.100) (0.0752) (0.0369) (0.0750) Internal migrant in 0.141 0.112 –0.000157 0.000437 0.00602 0.157*** –0.000874 0.0271 –0.0392 0.00833 0.0345* –0.0350 household (0.129) (0.125) (0.0604) (0.0578) (0.0410) (0.0454) (0.0191) (0.0739) (0.0727) (0.0592) (0.0182) (0.0538) Observations 529 529 529 527 529 529 529 349 348 530 530 530 Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. OLS = ordinary least squares. ‡ For greater clarity, women in households that receive remittances but do not have an international migrant are excluded from the estimation in Panel B. In Panel B the base category includes women in households with no internal or international migrants that do not receive remittances. All models include the same controls as in Table B2. 61 ANNEX F: REGRESSION OF INTEREST (ADDRESSING ENDOGENEITY) Male Outmigration and Women’s Work and Empowerment in Agriculture 63 64 TABLE F1. THE IMPACT OF MIGRATION ON TYPES OF WORK FOR WOMEN, NEPAL, 2SLS Farm Agricultural Processing Trading Employed Farm self- contributing (wage) (agricultural (agricultural Nonagricultural (any) employed family workers laborers products) products) workers Professional     (1) (2) (3) (4) (5) (6) (7) (8) B. Women (obs. 1,667) International migrant in household –0.136 0.253* –0.427*** 0.0596 0.108 0.0132 0.119 –0.0989 (0.0924) (0.135) (0.151) (0.0789) (0.0863) (0.0190) (0.0734) (0.0623) F-test 20.90 20.90 20.90 20.90 20.90 20.90 20.90 20.90 Sargan-Hansen (p value) 0.9147 0.368 0.0540 0.246 0.00303 0.251 0.904 0.257 Note: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 2SLS = two-stage least squares. All models include the same controls as in Table B1. The Case of Nepal and Senegal TABLE F2. THE IMPACT OF MIGRATION ON TYPES OF WORK FOR WOMEN, SENEGAL, 2SLS Farm contributing Agricultural Processing Trading Employed Farm self- family (wage) (agricultural (agricultural Nonagricultural (any) employed workers laborers products) products) workers Professional Other VARIABLES  (1) (2) (3) (4) (5) (6) (7) (8) (9) B. Women (2,637) International migrant in 0.269 0.144* 0.242 0.013 0.0104 –0.083 0.0706 0.0306 0.0347 household (0.249) (0.0834) (0.205) (0.0244) (0.0333) (0.103) (0.143) (0.0193) (0.0777) Internal migrant in 0.00697 –0.0898 0.307 0.063 –0.0369 –0.0132 0.0544 0.0406 0.292* household (0.25) (0.144) (0.258) (0.0477) (0.0476) (0.102) (0.125) (0.0285) (0.162) Observations 2,637 2,637 2,637 2,637 2,637 2,637 2,637 2,637 2,637 F-test 10.38 11.55 10.38 10.38 10.38 10.38 10.38 10.38 10.38 Sargan-Hansen (p value) 0.752 0.0589 0.473 0.269 0.118 0.921 0.567 0.59 0.175 Note: Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 2SLS = two-stage least squares. All models include the same controls as in Table B2. Male Outmigration and Women’s Work and Empowerment in Agriculture 65 REFERENCES Acosta, Pablo. 2006. “Labor Supply, School Attend- by the Women Left Behind.” Labour Economics 18: ance, and Remittances from International Migra- S98–114. tion: The Case of El Salvador.” World Bank Policy Research Working Paper No. 3903. Washington, DC: Brauw, Alan de, and Calogero Carletto. 2012. “Improv- World Bank. ing the Measurement and Policy Relevance of Migra- tion Information in Multi-Topic Household Surveys.” Alkire, Sabina, Ruth Meinzen-Dick, Amber Peter- Living Standards Measurement Study Working Paper 14. man, Agnes Quisumbing, Greg Seymour, and Ana Vaz. 2013. “The Women’s Empowerment in Agricul- Chort, Isabelle, Philippe De Vreyer, and Thomas ture Index.” World Development 52: 71–91. Zuber. 2017. “Gendered Internal Migration Pat- terns in Senegal.” < https://hal.archives-ouvertes.fr/ Allendorf, Keera. 2007. “Do Women’s Land Rights hal-01497824/document> Promote Empowerment and Child Health in Nepal?” Dhakal, N.H., and A. Maharjan. 2018. “Approaches World Development 35 (11): 1975–88. to the Productive Use of Remittances in Nepal.” ICI- Acharya, M., and L. Bennet. 1983. “Women and MOD Working Paper 2018/1. Kathmandu: ICIMOD. the Subsistence Sector: Economic Participation and FAO. 2016. Methods for estimating comparable rates Household Decision-Making in Nepal.” DEDPH Staff of food insecurity experienced by adults throughout Working Paper No. SWP 526. Washington, DC: World the world. Rome: FAO. Bank. Funkhouser, Edward, 1992. “Migration from Nica- Amuedo-Dorantes, Catalina, and Susan Pozo. 2006. ragua: some recent evidence,” World Development, “Migration, Remittances, and Male and Female Elsevier, vol. 20(8), pages 1209–1218, August. Employment Patterns.” American Economic Review, 96 (2): 222–226. Gartaula, Hom Nath, and Anke Niehof. 2013. “Migra- tion to and from the Terai: Shifting Movements and Ballard, Terri J., Anne W. Kepple, and Carlo Cafiero. Motives.” The South Asianist 2(2): 28–50. 2013. “The Food Insecurity Experience Scale: Devel- opment of a Global Standard for Monitoring Hunger Gartaula, Hom Nath, Anke Niehof, and Leontine Vis- Worldwide.” Technical Paper. Food and Agriculture ser. 2010. “Feminisation of Agriculture as an Effect Organization of the United Nations, Rome. http:// of Male Out-Migration: Unexpected Outcomes from www.fao.org/3/a-as583e.pdf. Jhapa District, Eastern Nepal.” International Journal of Interdisciplinary Social Sciences 5(2). Biggs, S., and S. Justice. 2015. Rural and Agricultural Mechanization: A History of the Spread of Small Lokshin, Michael, Mikhail Bontch-Osmolovski, and Engines in Selected Asian Countries. IFPRI Discussion Elena Glinskaya. 2010. “Work-Related Migration and paper 01443, Development Strategy and Governance Poverty Reduction in Nepal.” Review of Development Division. International Food Policy Research Institute. Economics 14(2): 323–32. Binzel, Christine, and Ragui Assaad. 2011. “Egyptian Lokshin, Michael, and Elena Glinskaya. 2009. “The Men Working Abroad: Labour Supply Responses Effect of Male Migration on Employment Patterns Male Outmigration and Women’s Work and Empowerment in Agriculture 67 of Women in Nepal.” The World Bank Economic Review Social Relationships?” Available at http://www.nai. 23(3): 481–507. uu.se/ecas-4/panels/81-100/panel-95/ [visitado: 29/03/2012]. Maertens, Miet, and Jo Swinnen. 2009. “Are Modern Supply Chains Bearers of Gender Inequality.” Paper Mueller, V., C. Kovarik, K. Sproule, and A. Quisumb- presented in Rome, Italy: 31 March–2 April. ing. 2015. “Migration, Gender, and Farming Systems in Asia: Evidence, Data, and Knowledge Gaps.” IFPRI Maertens, Miet, and Jo Swinnen. 2012. “Gender and Discussion Paper 01458, IFPRI, Washington, DC. Modern Supply Chains in Developing Countries.” Journal of Development Studies 48: 1412–1430. Mu, Ren, Dominique van de Walle. 2009. “Left behind to farm? Women’s labor re-allocation in rural China.” Maharjan, Amina, Siegfried Bauer, and Beatrice Policy Research working paper no. WPS 5107. Wash- Knerr. 2012. “Do Rural Women Who Stay Behind ington, DC: World Bank. Benefit from Male Out-Migration? A Case Study in the Hills of Nepal.” Gender, Technology and Development Nepa School of Social Sciences and Humanities. 16(1): 95–123. 2017. “Technical Report on Survey of Migration and Women’s Empowerment in Agriculture.” Malapit, Hazel J., Chiara Kovarik, Kathryn Sproule, Ruth S Meinzen-Dick, and Agnes R. Quisumbing. Nobles, Jenna, and Christopher McKelvey. 2015. 2015. “Instructional Guide on the Abbreviated Wom- “Gender, Power, and Emigration from Mexico.” en’s Empowerment in Agriculture Index (A-WEAI).” Demography 52(5): 1573–1600. International Food Policy Research Institute (IFPRI), Washington, DC. Paudel, Krishna P., Sujata Tamang, and Krishna K. Shrestha. 2014. “Transforming Land and Livelihood: Marzo, F., and B. Atuesta. 2018 forthcoming. “Break- Analysis of Agricultural Land Abandonment in the ing Out of the Productivity Trap: How Gender Ine- Mid Hills of Nepal.” Journal of Forest and Livelihood qualities Lock Senegal’s Women into Lifetimes of 12(1): 11–19. Lower Income.” Washington, DC: World Bank. Phadera, Lokendra. 2016. “International Migration McCullough, Ellen B.. 2015. Labor productivity and Its Effect on Labor Supply of the Left-Behind and employment gaps in Sub-Saharan Africa. Policy Household Members: Evidence from Nepal.” Paper Research working paper, no. WPS 7234. Washington, prepared for the 2016 Annual Meeting of the Agri- DC: World Bank Group. cultural and Applied Economics Association, Boston, July 31–August 2. Mendola, Mariapia, and Calogero Carletto. 2012. “Migration and Gender Differences in the Home Rodriguez, Edgard R., and Erwin R. Tiongson. 2006. Labour Market: Evidence from Albania.” Labour Eco- Temporary Migration Overseas and Household nomics 19 (6): 870–80. Labor Supply: Evidence from Urban Philippines. International Migration Review. 35. 709–725. Mohapatra, Sanket, and Dilip Ratha. 2011. “Remit- tance Markets in Africa.” Directions in Development Shrestha, Sundar S., and Prem Bhandari. 2007. “Envi- and Finance. Washington, DC: World Bank. ronmental Security and Labor Migration in Nepal.” Population and Environment 29(1): 25–38. Mondain, Nathalie, Sara Randall, Alioune Diagne, and Alice Elliot. 2011. “Consequences of Male Inter- Slavchevska, Vanya, Susan Kaaria, and Sanna-Liisa national Migration for Women’s Position in Sen- Taivalmaa. 2016. “Feminization of Agriculture egal: Reinforcement or Weakening of Traditional in the Context of Rural Transformations: What 68 The Case of Nepal and Senegal Is the Evidence?” Washington, DC: World Bank. Migration.” Global Economic Prospects and the https://openknowledge.worldbank.org/han- Developing Countries (GEP). Washington, DC: dle/10986/25099 License: CC BY 3.0 IGO. World Bank. Stanley, V. 2015. “Migration and Women’s Agency in World Bank. 2010. “Gender and Governance in Rural Agriculture – Women in Agriculture: The Impact of Services: Insights from India, Ghana, and Ethiopia.” Male Out-Migration on Women’s Agency, Household Washington, DC: World Bank. Welfare and Agricultural Productivity.” Washington, DC: World Bank. World Bank. 2015. “World Development Indica- tors.” Retrieved from http://data.worldbank.org/ Tamang, Sujata, Krishna P. Paudel, and Krishna K. indicator. Shrestha. 2014. “Feminization of Agriculture and Its Implications for Food Security in Rural Nepal.” Jour- World Bank. 2018 forthcoming. “Food Insecurity nal of Forest and Livelihood 12(1): 20–32. Experience Scale Exploration Paper: GAFSP Opera- tionalization and Target Setting.” Washington, DC: UN DESA. 2013. “Cross-National Comparisons of World Bank. Internal Migration: An Update on Global Patterns and Trends. Technical Paper.” Technical Paper No. World Bank, Food and Agriculture Organization, 2013/1. United Nations, Department of Economic and International Fund for Agricultural Develop- and Social Affairs, Population Division. http://www. ment. 2009. “Gender in Agriculture Sourcebook.” un.org/en/development/desa/population/publica- Agriculture and Rural Development, Washington, tions/pdf/technical/TP2013-1.pdf. DC: World Bank. UN DESA. 2017. “International Migration Report 2017: Yabiku, Scott T., Victor Agadjanian, and Arusyak Highlights.” New York: United Nations, Department Sevoyan. 2010. “Husbands’ Labour Migration and of Economic and Social Affairs, Population Division. Wives’ Autonomy, Mozambique 2000–2006.” Popula- http://www.un.org/en/development/desa/popula- tion Studies 64(3): 293–306. tion/migration/publications/migrationreport/docs/ MigrationReport2017_Highlights.pdf. Zezza, Alberto, Calogero Carletto, Benjamin Davis, and Paul Winters. 2011. “Assessing the Impact of World Bank. 2005. “Global Economic Prospects Migration on Food and Nutrition Security.” Food Pol- 2006: Economic Implications of Remittances and icy 36(1): 1–6. Male Outmigration and Women’s Work and Empowerment in Agriculture 69 1818 H Street, NW Washington, DC 20433