TRACKING EMPOWERMENT ALONG THE VALUE CHAIN: TESTING A MODIFIED WEAI IN THE FEED THE FUTURE ZONE OF INFLUENCE IN BANGLADESH Akhter U. Ahmed, Hazel Malapit, Audrey Pereira, Agnes Quisumbing, and Deborah Rubin With assistance from Julie Ghostlaw, Md. Latiful Haque, Nusrat Zaitun Hossain, and Salauddin Tauseef International Food Policy Research Institute In collaboration with Data Analysis and Technical Assistance Limited TRACKING EMPOWERMENT ALONG THE VALUE CHAIN: TESTING A MODIFIED WEAI IN THE FEED THE FUTURE ZONE OF INFLUENCE IN BANGLADESH Authors: Akhter U. Ahmed*1, Hazel Malapit*, Audrey Pereira*, Agnes Quisumbing*, and Deborah Rubin** With assistance from Julie Ghostlaw*, Md. Latiful Haque*, Nusrat Zaitun Hossain*, and Salauddin Tauseef* In collaboration with Data Analysis and Technical Assistance Prepared for: United States Agency for International Development Grant Number: EEM-G-00-04-00013-00 Submitted by: International Food Policy Research Institute Policy Research and Strategy Support Program for Food Security and Agricultural Development in Bangladesh House 10A, Road 35, Gulshan 2, Dhaka 1212, Bangladesh Date: August 2018 1Akhter Ahmed (a.ahmed@cgiar.org) is the corresponding author for comments and queries. * International Food Policy Research Institute ** Cultural Practice mailto:a.ahmed@cgiar.org Contents List of Acronyms ................................................................................................................................ i Acknowledgments ............................................................................................................................. ii Executive Summary ........................................................................................................................... iii 1. Background ................................................................................................................................1 1.1 Context of Women’s Empowerment in Agricultural Value Chains in Bangladesh ........................... 1 1.2 WEAI Overview ................................................................................................................................. 1 1.3 Rationale for WEAI extension: Modified WEAI for Value Chains (WEAI4VC) ................................... 3 2. Research Questions ....................................................................................................................3 3. Study Design and Implementation ..............................................................................................3 3.1 Quantitative Survey .......................................................................................................................... 4 3.1.1 Sampling Design .................................................................................................................... 4 3.1.2 Survey Questionnaires .......................................................................................................... 6 3.1.3 Survey implementation and data capture ............................................................................ 8 3.2 Qualitative Research ......................................................................................................................... 9 3.2.1 Sample Selection ................................................................................................................... 9 3.2.2 Qualitative Protocol Design ................................................................................................ 10 3.2.3 Enumeration Team and Training......................................................................................... 11 3.2.4 Fieldwork and Quality Control ............................................................................................ 11 3.2.5 Data Entry and Cleaning ...................................................................................................... 11 4. Findings .................................................................................................................................... 11 4.1 Household Characteristics .............................................................................................................. 11 4.2 Livelihoods ...................................................................................................................................... 15 4.3 Resources ........................................................................................................................................ 24 4.4 Income ............................................................................................................................................ 35 4.5 Leadership ....................................................................................................................................... 37 4.6 Time ................................................................................................................................................ 39 4.7 Intrahousehold Relationships ......................................................................................................... 43 4.8 Other Domains ................................................................................................................................ 45 4.9 WEAI Results ................................................................................................................................... 53 4.10 Key Constraints to Empowerment .................................................................................................. 60 5. Summary and Concluding Remarks ........................................................................................... 66 References ....................................................................................................................................... 71 List of Tables Table 1.1 Domains, indicators, and weights of the WEAI ............................................................................. 2 Table 3.1 List of selected upazilas and districts for the WEAI4VC study ...................................................... 5 Table 3.2 Sample distribution of selected households ................................................................................. 6 Table 3.3 Description of selection of households by PSU ............................................................................. 6 Table 3.4 Household and Individual Questionnaire Modules ....................................................................... 7 Table 3.5 Type and number of interviews conducted ................................................................................ 10 Table 4.1 Household characteristics, by actor and household type ........................................................... 12 Table 4.2 Individual and household characteristics of respondents, by actor and household type .......... 13 Table 4.3 Domains and sub-indicators of the WEAI4VC ............................................................................. 14 Table 4.4 Number of women who reported bringing assets to marriage, by asset type and actor ........... 53 Table 4.5 A-WEAI score and women's empowerment status, by actor and household type .................... 54 Table 4.6 Women’s empowerment status, by actor .................................................................................. 57 Table 4.7 Percent of respondents adequate in modified WEAI sub-indicators, by household type .......... 62 Table 4.8 Key constraints to empowerment (adequacy below 70 percent), by actor and household type ............................................................................................................................................ 63 Table 5.1 Summary of empowerment scores and indicators, by actor and sex ......................................... 69 Table 5.2 Male-female headcount differences in adequacy, by household type ....................................... 70 List of Figures Figure 4.1 Producers: Respondent participation in production activities, by household type .................. 15 Figure 4.2 Entrepreneurs: Respondent participation in entrepreneurship activities, by household type ........................................................................................................................ 16 Figure 4.3 Wage workers: Respondent participation in wage-work activities, by household type ........... 17 Figure 4.4 Percent of respondents who are adequate in input in livelihood activity decisions, by actor and household type ................................................................................................... 18 Figure 4.5 Percent of respondents who are adequate in access to information about livelihood activities, by actor and household type .................................................................................. 19 Figure 4.6 Producers: Percent of respondents who are like the people in the stories .............................. 21 Figure 4.7 Entrepreneurs: Percent of respondents who are like the people in the stories ....................... 21 Figure 4.8 Wage Workers: Percent of respondents who are like the people in the stories ....................... 22 Figure 4.9 Percent of respondents who are adequate in autonomy in livelihood activities, by actor and household type ....................................................................................................... 23 Figure 4.10 Percent of respondents from asset-owning households, by actor and household type ......... 24 Figure 4.11 Percent of respondents from households who solely or jointly own assets owned by their households, by actor and household type ..................................................................... 26 Figure 4.12 Percent of respondents from asset-owning households who can purchase assets, by actor and household type ................................................................................................... 27 Figure 4.13 Percent of respondents from asset-owning households who can rent, sell, give away, or mortgage those assets, by actor and household type ............................................. 28 Figure 4.14 Percent of respondents in households who are adequate in ownership of assets, by actor and household type ....................................................................................................... 30 Figure 4.15 Percent of respondents who are adequate in rights over assets, by actor and household type ........................................................................................................................ 31 Figure 4.16 Percent of respondents from households whose households have access to loans, by actor and household type ................................................................................................... 32 Figure 4.17 Percent of respondents who are adequate in access to and decisions on credit, by actor and household type ....................................................................................................... 33 Figure 4.18 Percent of respondents who solely or jointly have financial accounts, by actor and household type ................................................................................................................. 34 Figure 4.19 Percent of respondents who are adequate in access to a financial account, by actor and household type ....................................................................................................... 35 Figure 4.20 Percent of respondents who are adequate in control over use of income, by actor and household type ....................................................................................................... 36 Figure 4.21 Percent of respondents who are adequate in control over use of agricultural income, by actor and household type ..................................................................................... 37 Figure 4.22 Percent of respondents who are active members of community groups, among groups that are available in the community, by actor and household type ........................... 38 Figure 4.23 Percent of respondents who are adequate in group membership, by actor and household type ........................................................................................................................ 39 Figure 4.24 Average time spent on workload, by actor and household type ............................................. 40 Figure 4.25 Average time spent on workload by age category, by actor and household type ................... 40 Figure 4.26 Percent of respondents who are adequate in workload, by actor and household type ......... 41 Figure 4.27 Average minutes spent on childcare as a secondary activity, by actor and household type .......................................................................................................................................... 42 Figure 4.28 Percent of female respondents in DHHs who are adequate in access to childcare, by actor and household type ................................................................................................... 43 Figure 4.29 Percent of respondents who are adequate in mutual respect among household members, by actor and household type ................................................................................. 44 Figure 4.30 Percent of respondents who believe that a husband is never justified in hitting their wife, by actor and household type ................................................................................. 45 Figure 4.31 Percent of respondents who can visit two or more locations per week, by actor and household type ........................................................................................................................ 46 Figure 4.32 Percent of women respondents who participate in decisions about visiting important locations, by household type .................................................................................................. 47 Figure 4.33 Responses to how households protect women, by actor and household type ...................... 48 Figure 4.34 Responses to whether women (both young and old) are required to cover the head when going out, by actor and household type ........................................................ 48 Figure 4.35 Percent of respondents who have ever heard messages about (in a group, from the media, from an NGO worker) or discussed issues, by actor and household type ............ 49 Figure 4.36 Food insecurity in past 12 months, by actor and household type ........................................... 50 Figure 4.37 Food insecurity in four weeks, by actor and household type .................................................. 51 Figure 4.38 Assets brought to marriage (women respondents only), by actor and household type ......... 52 Figure 4.39 Contribution of each of the six indicators to disempowerment, by sex, actor, and household type ................................................................................................................. 55 Figure 4.40 Percent contribution of each indicator to disempowerment, by actor and household type ........................................................................................................................ 56 Figure 4.41 Top contributors to disempowerment among all women respondents versus only women participants, by actor .......................................................................................... 58 Figure 4.42 Contribution of each of the six indicators to disempowerment, by sex, actor, and household type ................................................................................................................. 59 Figure 4.43 Difference between the percentages of men and women (in dual-adult households) who have adequate achievements in each sub-indicator, by actor and household type ...... 64 Figure 4.44 Difference between the percentages of men and women (in female-adult only households) who have adequate achievements in each sub-indicator, by actor and household type ................................................................................................................. 65 i List of Acronyms 5DE Five domains of empowerment A-WEAI Abbreviated Women’s Empowerment in Agriculture Index BIHS Bangladesh Integrated Household Survey CAPI Computer-assisted personal interviewing DATA Data Analysis and Technical Assistance DHH Dual-headed household FHH Female-headed household FTF Feed the Future GAAP2 Gender, Agriculture, and Assets Project – Second Phase GI Group interview GOB Government of Bangladesh GPI Gender Parity Index IFPRI International Food Policy Research Institute KII Key informant interview OPHI Oxford Poverty & Human Development Initiative PPS Probability proportional to size Pro-WEAI Project Women’s Empowerment in Agriculture Index PRSSP Policy Research and Strategy Support Program PSU Primary sampling unit RAI Relative Autonomy Index USAID U.S. Agency for International Development WEAI Women’s Empowerment in Agriculture Index WEAI4VC Women’s Empowerment in Agriculture Index for Value Chain ZOI Zone of Influence ii Acknowledgments We gratefully acknowledge the United States Agency for International Development (USAID) for funding the Policy Research and Strategy Support Program (PRSSP) in Bangladesh under USAID Grant Number EEM-G-00-04-00013-00. This report is an output of the PRSSP. We also acknowledge the support of the CGIAR Research Program on Policies, Institutions, and Markets (PIM) led by the International Food Policy Research Institute (IFPRI). Data for this report came from the 2017 Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) quantitative survey, which was approved by the Ministry of Agriculture, Government of the People’s Republic of Bangladesh. At IFPRI, we thank the director of the Poverty, Health, and Nutrition Division, Marie Ruel, for her overall guidance. We also thank the qualitative research team for providing in-depth insights, which helped to validate the WEAI4VC quantitative survey, explore men’s and women’s views on empowerment across the value chain, and investigate barriers to entry and growth for women and men in value chains of different commodities. Specifically, we are grateful to the IFPRI field research officers—Shammi Sultana Ferdousi, Tahsin Rahaman, Shuchita Rahman, Waziha Rahman, and Md. Redoy—and the qualitative research coordinator Aklima Parvin, all of whom worked closely with Cultural Practice under the leadership of Deborah Rubin. We thank Jay Willis for his help with the production of this report. The study would not have been possible without the dedication and hard work of the survey enumerators and other staff of the Data Analysis and Technical Assistance (DATA), a Bangladeshi consulting firm that carried out the quantitative survey under IFPRI supervision. iii Executive Summary Upon request of the U.S. Agency for International Development (USAID), the International Food Policy Research Institute (IFPRI) conducted this study to support USAID in assessing the state of empowerment and gender parity of men and women along the agricultural value chain in the Feed the Future (FTF) Zone of Influence (ZOI) in Bangladesh. Specifically, IFPRI’s Policy Research and Strategy Support Program (PRSSP), funded by USAID, piloted the modified Women’s Empowerment in Agriculture Index (WEAI) survey instruments in 10 upazilas (sub-districts) within the FTF ZOI across 1,200 households, which broadly belong to three economic activities of interest: (1) agricultural production, (2) agricultural entrepreneurship, and (3) agricultural sector employment. The quantitative survey was complemented by qualitative research to glean further insights into the facilitators and constraints of empowerment among various actors in the agricultural value chain. The data and analysis generated from this WEAI for Value Chain (WEAI4VC) study may inform USAID’s selection and design of interventions that may, in turn, maximize its programmatic impact on women and men’s empowerment as producers, entrepreneurs, and wage employees. Overall, study results show that empowerment varies based on livelihood activity and gender. Women’s empowerment differs based on the primary economic activity, with greater empowerment among women in producer households than entrepreneur or wage-work households. Qualitative interviews suggest that this may be because female entrepreneurs and wage workers are more susceptible to loss of social respect than female producers since working away from home as a woman deviates from social conventions. Conversely, men’s empowerment is relatively consistent across livelihood activities. The WEAI4VC study finds lack of group membership as a key driver to disempowerment for both men and women, which is consistent with findings from IFPRI’s previous household surveys in the FTF ZOI (Ahmed et al. 2015, IFPRI 2018). Also, group membership and input in livelihood activities were among the top three contributors to disempowerment for both women and men in entrepreneur and wage- work households. The WEAI4VC study suggests two approaches of identifying areas of focus for future programming: (1) the first approach is to identify the indicators with low achievements, which would, therefore, have more potential for substantial improvement; and (2) the second approach is to look at the differences between the male achievements and the female achievements to see which indicators have the largest achievement gaps by gender. Using the first approach, the analysis identified that, workload and physical mobility are key constraints for both men and women in dual-headed households, regardless of livelihood activity. This suggests that interventions will need to consider the time burden required to participate in a particular livelihood activity or to adopt a specific practice or technology. Second, the analysis shows that women remain disadvantaged in terms of mutual respect, attitudes toward domestic violence, and mobility; thus, programming should address these issues. Third, both women and men in entrepreneur and wage-work households have constraints in accessing information, which may represent an opportunity to provide specific livelihood-related training that goes beyond agricultural production. For the second approach—exploring indicators that show gender disparities in achievements— unsurprisingly, we find that men’s empowerment surpasses women’s in most areas. Women in iv entrepreneur households, for instance, have low adequacy scores for autonomy, which may suggest that they do not have many options to choose the type of product, location, and size of their enterprise. There are notable exceptions in which women’s empowerment is greater than men’s: group membership, which favors women across all value chain activities; and workload, which favors women in producer households and in dual-headed entrepreneurial households. Despite challenges in identifying empowerment pathways, the WEAI4VC study has generated evidence on the unique challenges facing value chain actors in Bangladesh, and formulated recommendations for effective targeting of interventions. For producers, increasing participation in groups, decreasing workload, and improving physical mobility are important to close empowerment gaps. It is also important to address the low autonomy that women report in many aspects of agricultural production, and to change attitudes toward domestic violence. For entrepreneurs, increasing autonomy, rights over assets, access to credit, and increasing mutual respect among household members are key to women’s empowerment. For wage workers, increasing autonomy, strengthening rights over assets and control over income, and addressing norms surrounding domestic violence may help close empowerment gaps. Evidence on empowerment generated from high quality data is imperative to guide the design and implementation of gender-sensitive policies and programs. While the WEAI4VC study has assessed empowerment for specific types of livelihood activities, our initial analysis does not fully capture other aspects of livelihoods decisions in diversified households. Further analysis on the multiple roles undertaken by households and individuals to establish a more comprehensive assessment of empowerment that accounts for diversification will sharpen our diagnosis of empowerment gaps along the agricultural value chain in Bangladesh. 1 1. Background 1.1 Context of Women’s Empowerment in Agricultural Value Chains in Bangladesh Despite the recognition of both the agricultural sector as an engine of growth and development and the role of women in rural transformation, it is only recently that robust tools for measuring the impact of agricultural interventions on women’s empowerment have been developed. One of these tools, the Women’s Empowerment in Agriculture Index (WEAI), is a survey-based index developed specifically to measure the empowerment, agency, and inclusion of women in the agricultural sector (Alkire et al. 2012). The WEAI, originally developed as a monitoring indicator for the US Feed the Future (FTF) Initiative, has been applied or modified in more than 49 countries by 69 organizations as of the end of 2017. Bangladesh, one of the countries in which the WEAI was piloted, is also the first to have WEAI data that are representative of the FTF Zone of Influence (ZOI) supported by USAID as well as the entire rural areas of the country. IFPRI’s Policy Research and Strategy Support Program (PRSSP) specifically designed the Bangladesh Integrated Household Survey (BIHS), the most comprehensive, nationally representative household survey conducted to date in Bangladesh, to measure the WEAI. In 2015, Bangladesh became the first country to generate panel data on WEAI. Analysis of the panel data revealed remarkable improvements in women’s empowerment status from the 2011/12 baseline to the 2015 midline in the FTF ZOI. In 2011/12, only 27.4 percent of women were empowered in the FTF ZOI. In 2015, the headcount rate increased 13.9 percentage points to 41.2 percent of all women being empowered. Women who are not yet empowered experienced a 9.1-percentage-point increase in the percentage of domains where they have adequate achievements, from 54.1 percent to 63.2 percent of the domains. At baseline, 40.2 percent of women had gender parity with the primary male in their household; the rate increased 10.5 percentage points to 50.7 percent. The empowerment gap between female and the primary male in their household reduced 10.6 percentage points from 31.6 percent at baseline to 21.0 percent at midline (Ahmed et al. 2015). While these results show positive changes in women’s empowerment in the FTF ZOI over a three-year period, they are only relevant to agricultural production because other areas of empowerment were not measured. It is possible that FTF programming may have had spillover effects on women’s empowerment in agriculture-based rural enterprises as well as the nonfarm sector, but this would not have been captured by the existing WEAI. The need to develop a metric to enable a better understanding of empowerment in other nodes of the value chain, such as entrepreneurship and wage- earning activities, motivated the development of a new index, the WEAI for Value Chains (WEAI4VC). 1.2 WEAI Overview The WEAI was launched in February 2012 and was developed in collaboration between IFPRI, Oxford Poverty and Human Development Initiative (OPHI), and USAID. It is an innovative, survey-based tool for measuring, evaluating, and learning about women’s empowerment and inclusion in the agricultural sector. While originally designed as a monitoring and evaluation tool for USAID’s FTF Initiative, the index can also be used more generally to assess the general state of empowerment and gender parity in agriculture and to identify the key areas where empowerment gaps exist (Alkire et al. 2013). 2 The WEAI is an aggregate index that can be reported at the program-level (as well as other geographic areas) and is composed of two sub-indices: the five domains of empowerment (5DE) and the gender parity index (GPI). The 5DE assesses the degree to which women are empowered in five domains: (1) agricultural production decisions, (2) access to and decision-making power over productive resources, (3) control over use of income, (4) leadership roles within the community, and (5) time allocation. The 5DE is constructed from individual-level empowerment scores, which reflect each person’s achievements in the five domains as measured by 10 indicators with their corresponding weights (Table 1.1). Each indicator measures whether an individual has surpassed a given threshold, or has adequate achievement, with respect to each indicator. A woman or man is defined as empowered if he or she has adequate achievements in four out of the five domains, or has achieved adequacy in 80 percent or more of the weighted indicators. Table 1.1 Domains, indicators, and weights of the WEAI Domain Indicator Definition of Indicator Weight 1. Production 1.1 Input in productive decisions Sole or joint decision making over food and cash- crop farming, livestock, and fisheries 1/10 1.2 Autonomy in production Autonomy in agricultural production reflects the extent to which the respondent’s motivation for decision making reflects own values rather than a desire to please others or avoid harm 1/10 2. Resources 2.1 Ownership of assets Sole or joint ownership of major household assets 1/15 2.2 Purchase, sale, or transfer of assets Whether respondent participates in decision to buy, sell, or transfer assets 1/15 2.3 Access to and decisions about credit Access to and participation in decision making concerning credit 1/15 3. Income 3.1 Control over use of income Sole or joint control over income and expenditures 1/5 4. Leadership 4.1 Group member Whether respondent is an active member in at least one economic or social group 1/10 4.2 Speaking in public Whether the respondent is comfortable speaking in public concerning issues relevant to oneself or one’s community 1/10 5. Time 5.1 Workload Allocation of time to productive and domestic tasks 1/10 5.2 Leisure Satisfaction with time for leisure activities 1/10 Source: Alkire et al. (2013). Unlike other women’s empowerment measures based on interviews of a sole female respondent, WEAI uses survey data from the self-identified primary male and female adult decision makers, aged 18 and over, in the same household. Relative empowerment is captured in GPI, which reflects women’s achievements in the five domains relative to the men in their households. Households are classified as having gender parity if either the woman is empowered (her empowerment score is 80 percent or higher) or her score is greater than or equal to the empowerment score of the male decision maker in her household. 3 All of these indexes have values ranging from 0 to 1, where higher values reflect greater empowerment. The overall WEAI is a weighted average of 5DE and GPI, with weights 0.9 and 0.1, respectively. While the overall WEAI is useful as a headline indicator, similar to how poverty indexes are used to track overall trends in poverty, the WEAI is also decomposable, which allows us to disaggregate the 5DE achievements by domain and by indicator to see which specific areas contribute the most to both women’s and men’s disempowerment. More details about the methodology, piloting, and validation of WEAI are available in Alkire et al. (2012, 2013). 1.3 Rationale for WEAI extension: Modified WEAI for Value Chains (WEAI4VC) The original form of the WEAI has limitations, as recognized by Alkire et al. (2013). These include: • Women who are engaged in decision-making on nonagricultural activities may appear disempowered if they are not also involved in agricultural decisions. • Questions about control over resources and income do not capture many of the nuances behind these domains. • Female-only households are likely to be identified as empowered, even if there are others such as parents, in-laws, or children with whom such women also need to negotiate. • Group membership alone is an inadequate indicator of active participation. • Satisfaction with leisure is subjective and may reflect women’s lower expectations of what is possible in their circumstances. • The focus on agriculture may not capture other domains of empowerment that may be more relevant to specific outcomes. As the rural economies diversify and households become more involved in nonfarm and off-farm economic activities, it is likely that the original WEAI will miss key aspects of empowerment among target beneficiaries who are engaged in rural nonfarm wage earning activities and rural entrepreneurship, which are important livelihood activities in rural Bangladesh. 2. Research Questions In this study, we seek to answer the following research questions: • How empowered are women and men in their roles as producers, wage earners, and entrepreneurs in the FTF ZOI? • What are the sources of disempowerment of women and men as producers, wage earners, and entrepreneurs? What gender-based constraints do they face? • What types of interventions, technologies, or practices can be targeted to women and men producers, wage earners, and entrepreneurs to address sources of disempowerment? 3. Study Design and Implementation The WEAI4VC study combined a quantitative survey and qualitative semi-structured key informant interviews and group interviews. This mixed methods approach to data collection provided opportunities to analyze a rich pool of data that would not have been available with any of these methods on their own. Because of the focus on women’s empowerment and gender equality, sex- disaggregated information was collected covering a wide range of topics. 4 3.1 Quantitative Survey The quantitative data came from a household survey, which was carried out from May–July 2017. This section describes the sampling and fieldwork for the quantitative survey. 3.1.1 Sampling Design We used a sample size of 400 households for each of the three economic activities of interest—(1) agricultural production, (2) agricultural entrepreneurship, and (3) agricultural-sector wage employment—to be able to construct and compare the overall indexes: 5DE, GPI, and A-WEAI—as well as the specialized modules relevant to value chains. We define these economic activities as follows: (1) Agricultural production—A household is classified as a production household if any member has participated in crop farming/fishing/livestock raising in the past 12 months. (2) Agricultural entrepreneurship—A household is classified as an entrepreneur household if any member owns/operates an agriculture-driven business for commercial purposes in the past 12 months. (3) Agricultural-sector wage employment—A household is classified as a wage worker household if any member worked for someone outside the household in exchange of money, food, or goods in the agriculture sector in the past 12 months. This work can be work for agriculture production (crop production, livestock, or fish production), agri-business, or non-agri-business. Livelihoods in rural Bangladesh are diverse: the income source portfolio for rural households is such that many households are likely to be engaged in more than one type of economic activity during the year. It is challenging to identify households that exclusively draw income from one type of economic activity, and is especially difficult to select households that earn their living exclusively from wage employment due to the intermittent nature of the stated activity. Agricultural wage employment is usually short- term, seasonal, and primarily on an as needed basis. Since households are likely to be engaged in more than one kind of economic activity over a 12-month period, instead of categorizing households into only one of the three economic activities before administering the modified WEAI modules, we surveyed 1,200 households in total. This increased our chances of screening and identifying at least 400 households for each category to compare empowerment among the groups. First, from the list of all upazilas (sub-districts) in the FTF ZOI in southeastern Bangladesh, we purposively selected five upazilas for producer and entrepreneur groups, considering diversified agriculture with rice, vegetables, pulses, maize, cut flowers, livestock and poultry, fisheries, and availability of agriculture-based enterprises. Table 3.1 shows the list of selected upazilas and districts, and considerations for their selection in this study. Once five upazilas were selected, four villages were randomly selected with probability proportional to size (PPS) sampling from the list of all villages in the five selected upazilas using village-level population as the basis for size. Thus, 20 villages or primary sampling units (PSUs) were selected, in which a village census was administered using computer assisted personal interviews (CAPI). From the village census lists, we randomly selected 400 producer households (farm households) and 200 agriculture-sector wage-worker households (those who depended mostly on wage earnings). 5 Table 3.1 List of selected upazilas and districts for the WEAI4VC study District Upazila Consideration Barisal Gouronadi Diversified agriculture includes betel leaves and agricultural base enterprises Jessore Jhikargacha Diversified agriculture includes cut flowers and agricultural base enterprises Chuadanga Sadar Diversified agriculture includes cut flowers and agricultural base enterprises Jhenaidah Kaliganj Diversified agriculture includes cut flowers and agricultural base enterprises Satkhira Kaliganj Diversified agriculture includes cut flowers and agricultural base enterprises Source: Constructed by authors. Once the producer and the wage-worker samples of households were selected, a detailed household survey was implemented on the selected households. The individual questionnaire was meant to be administered separately and privately to the primary male and primary female decision makers, usually husband and wife, consistent with the original WEAI protocol, which was possible in all producer households. Less than 4 percent of women in entrepreneur households and wage-worker households were interviewed in the presence of another female adult or children. Less than 3 percent of men in entrepreneur households and wage-worker households were interviewed in the presence of another adult or children. Third, since enterprises are mostly located in urban centers, we decided to use upazila and union centers as PSUs for entrepreneur households and wage employees working for entrepreneur households. We followed the following steps for sampling of entrepreneurs and for agriculture-sector wage employees working for entrepreneurs: using CAPI, we conducted a census of entrepreneurs and agriculture-sector employees working for entrepreneurs in the five selected upazila centers and their union centers. From the census lists, we randomly selected 400 entrepreneur households and 200 wage earner households working for the entrepreneur households. Unlike the producer households, sampling for post-harvest agricultural entrepreneurs was more difficult because these types of entrepreneurs are diverse but not equally prevalent. For example, the field teams identified many irrigation water suppliers and input dealers but not as many agricultural produce transporters or rice/flour mill operators. It was also rare to find female entrepreneurs so any female entrepreneur household identified in the census was automatically selected. In some cases, households with more unusual types of enterprise or wage work activities were also automatically selected, such as those engaged in the cut flower value chain. Highly seasonal activities such as production of GUR (molasses/treacle) are also likely to be missed during the census despite the field team’s best efforts to locate them. Once the entrepreneur and wage employee samples of households were selected, a detailed household survey was administered to the primary male and female respondents in selected households. 6 All three categories of households with both adult male and female accounted for 80 percent of sampled households, whereas households with female adults only accounted for 20 percent of the sample. For agricultural producer/farm households, in which the village is the PSU, there were 16 households with both male and female adults and four households with female adults only in each village. For wage-earner households, there were eight households with both male and female adults and two households with female adults only in each village. For entrepreneur households, since upazila centers were the PSUs, there were 64 households with both male and female adults and 16 households with female adults only in each upazila center. For wage- earner households working for entrepreneurs, there were 32 households with both male and female adults and 8 households with female adults only in each upazila center. Table 3.2 shows the sample distribution of households per economic category by upazila and Table 3.3 shows the number of adult male and female households, as well as the number of female-only households (no adult male present) per PSU. Table 3.2 Sample distribution of selected households Number of households Division District Upazila Producer Producer Labor Entrepreneur Entrepreneur Labor Total Barisal Barisal Guarnadi 80 40 80 40 240 Khulna Chuadanga Chuadanga Sadar 80 40 80 40 240 Khulna Jessore Jhikargachha 80 40 80 40 240 Khulna Jhenaidah Jhenaidah-Kaligan 80 40 80 40 240 Khulna Satkhira Satkhira-Kaliganj 80 40 80 40 240 Total 400 200 400 200 1,200 Source: Constructed by authors. Table 3.3 Description of selection of households by PSU Sample household type Households Upazila Number of PSUs Households per PSU Households per village with adult male and female Female- headed households per village Producers 400 5 20 villages 20 16 4 Wage labor under production 200 5 20 villages 10 6 2 Entrepreneurs 400 5 1 upazila 80 12 4 Wage labor beyond production 200 5 1 upazila 40 6 2 Source: Constructed by authors. 3.1.2 Survey Questionnaires The WEAI4VC survey was composed of a household-level questionnaire administered to the household head or other knowledgeable person in the household, and an individual-level questionnaire 7 administered to the self-identified male and female decision makers regarding the relevant economic activity. The household questionnaire included eight modules on various topics at the household level, including demographics, agricultural production, employment, entrepreneurship, assets, transfers, and shocks. The individual questionnaire included 19 modules covering key dimensions of empowerment such as livelihoods, resources, income, leadership, time use, and intrahousehold relationships and access to information and extension, as well as specific modules for individuals engaged in particular economic activities. Table 3.4 lists the modules of the household and individual questionnaires. Table 3.4 Household and Individual Questionnaire Modules Household-Level Questionnaire Individual Questionnaire Household identification Individual Identification Household listing and demographics Role in household decision making (Producers) Livelihoods and employment Role in household decision making (Entrepreneurs) Dwelling characteristics Role in household decision making (Wage earners) Land and agriculture Access to productive capital Institutional transfers & program operations – cash Access to financial services Institutional transfers & program operations – in kind Time allocation Household shocks Group membership Autonomy in decision making (Producers) Autonomy in decision making (Entrepreneurs) Autonomy in decision making (Wage earners) Intrahousehold relationships Attitudes about domestic violence Physical mobility Parda information Messaging Food insecurity experience scale Wife’s assets that had been brought to marriage Personal information Source: Constructed by authors. The survey instrument used was a modified version of the WEAI called the Women’s Empowerment in Agriculture for Value Chain (WEAI4VC), which was designed to measure the extent of empowerment of women involved in rural agricultural wage employment and entrepreneurship, in addition to agricultural production. These survey modules drew on lessons learned from piloting project-level WEAI (pro-WEAI) under the Gender, Agriculture, and Assets Project–Second Phase (GAAP2), as well as inputs from IFPRI’s ongoing work on women’s empowerment in agricultural value chains and rural nonfarm employment. The IFPRI team designed the WEAI4VC survey to collect data on key dimensions of empowerment across multiple activities in the agricultural value chain. The household and individual questionnaires were conducted using CAPI. Skip patterns and consistency checks were included in the survey program to ensure data quality. 8 3.1.3 Survey implementation and data capture Training For implementing the WEAI4VC household survey, IFPRI contracted Data Analysis and Technical Assistance (DATA), a Bangladeshi consulting firm with expertise in conducting complex surveys and data analysis. DATA worked under the supervision and guidance of senior IFPRI researchers. DATA’s capacity to conduct surveys that collect high quality data was largely built by IFPRI over the past two decades.1 DATA provided experienced survey enumerators and supervisors to administer the household survey. Most of the enumerators and supervisors hold master’s degrees in social science, nutrition, or home economics. From March 20–May 6, 2017, IFPRI researchers and DATA experts trained 40 experienced enumerators (20 females and 20 males), 10 supervisors (5 females and 5 males), and 2 male field monitors. The survey enumerators’ training was approximately fifty days in duration (33 actual training days), and consisted of a formal classroom component as well as closely monitored practice fieldwork. During the formal training, IFPRI researchers and DATA experts briefed the enumerators and supervisors on the objectives and methods of the survey, the sampling design, and the responsibilities of the enumerators. They were trained on how to carry out the interviews using CAPI tablets, Issues related to using tablets and troubleshooting of problems with tablets, line-by-line explanation and interpretation of the questions, the flow and skip-patterns, definitions, and explanations of how to handle unusual cases and when to contact the supervisor for assistance. Field supervisors received additional training related to their supervisory and editing role. In particular, they were trained on the quality control process, cross checking, editing and coding of the questions, and security and confidentiality issues. On April 2, 2017, the questionnaires were field tested in five villages within three unions of Saturia Upazila in Manikganj District. A subsequent field test was conducted on April 20, 2017, in the same set of villages. The field testing determined the appropriate distribution of questionnaire modules among the male and female questionnaires, identified problems with the questionnaires, or additional rules that were needed to address difficult cases. The field testing aimed to approximate the actual implementation of the survey in order to test the full range of survey activities, including questionnaire completion, delivery, and data entry. An additional function of the field testing was to provide practical training to the enumerators in administering the questionnaire. After pre-testing in the field, feedback was incorporated and the survey questionnaire was finalized. 1 DATA carried out all IFPRI surveys in Bangladesh, including more than 50 household surveys and several market, school, and other institutional surveys. In addition, DATA has conducted numerous surveys for various international organizations, such as the World Food Programme (WFP)-Bangladesh, the World Bank, the European Union, the US Department of Agriculture, CARE-Bangladesh, World Vision-Bangladesh, the Population Council– New York, Save the Children (USA), Tufts University School of Nutrition Science and Policy, and the IRIS Center at the University of Maryland. 9 Survey Administration DATA carried out the household survey from May 7–July 16, 2017, under the supervision and guidance of IFPRI researchers in five districts: Jessore, Jhenaidah, Chuadanga, Satkhira, and Barisal, all of which are located within the USAID-supported FTF ZOI in southern Bangladesh. The survey was conducted in two phases: the first phase was conducted from May 7-26, 2017, prior to the fasting month of Ramadan; the second phase was conducted from July 3-11, 2017, after Ramadan. On July 3, 2017, IFPRI and DATA jointly organized a one-day enumerators’ refresher training to ensure the survey team’s retention of knowledge between the two phases. Going into the field, the teams of enumerators were equipped with various materials, such as CAPI tablets, survey manuals, identification cards, and letters of authorization to conduct the survey issued by the Ministry of Agriculture, Government of Bangladesh. The enumerators conducted the interviews one-by-one and face-to-face with the respondents assigned to him or her. The enumerators were supervised by the field supervisors who accompanied them to the village. Each field supervisor was responsible with his/her defined region. All field staff reported their activities to their supervisors using a standard progress report form. Quality Control IFPRI and DATA worked diligently to ensure the quality of the household survey data. In the field, survey supervisors routinely oversaw interviews conducted by enumerators, and verified all data collected by enumerators on a daily basis. If inconsistencies in responses were detected in collected data, then the supervisors visited the relevant respondents to find out the reasons and corrected the responses as needed. IFPRI researchers made frequent field visits to supervise the fieldwork. Data Entry and Cleaning The use of CAPI on programmed tablets minimized data processing time after fieldwork and improved data integrity. Collected data were transferred to the DATA central office in Dhaka on a daily basis for further quality control and validation. After cleaning and labeling by variable and value, DATA delivered the entrepreneurs and wage employees dataset to IFPRI on August 20, 2017, followed by the producer dataset on August 31, 2017. 3.2 Qualitative Research 3.2.1 Sample Selection For the Bangladesh qualitative study (Rubin 2018), the team sought to cover the three categories of respondents that the quantitative survey was focusing on—producers, entrepreneurs, and wage workers—and drew from a subset of the interviewees in the quantitative survey sample described above. The qualitative sample also included interviews with a small set of market traders. Representatives of these categories were interviewed, either in key informant interviews or in group interviews of four to five people. As described above, these categories are not mutually exclusive (Table 3.5). Most of the interviewees in the subsample were also engaged in farming for both home consumption and for sale, even when their main source of income was derived from their occupations 10 as entrepreneurs or traders. In these areas, a smart livelihood strategy is a multifaceted one, and the qualitative interviews illustrated the many ways that households seek to maintain themselves. Interviewees were drawn as much as possible from the list of respondents in the quantitative survey. In some cases, either because that list did not include enough traders or entrepreneurs in the field site, or because the original quantitative respondents were not available, community members were asked to suggest suitable candidates. In total, 102 people were interviewed, including four interviews with government officials or community leaders. Table 3.5 Type and number of interviews conducted Tool Types of respondents Respondents: Minimum number Total Activity (i) Community profile KII w/district or upazila officer, gender focal point, or leading community member 1 person per upazila 1 X 4 = 4 Activity (ii) Group interviews: Local understanding of empowerment Group interviews with: a. Agricultural producers b. Agricultural entrepreneurs c. Wage workers In each upazila, one group of 4-5 men and one group of 4-5 women for each of the three categories 30 X 2 = 60 Activity (iii) Semi-structured interviews Semi-structured interviews with: a. Agricultural Producers b. Agricultural entrepreneurs c. Wage workers In each upazila, for each of the three economic categories, 2 women and 2 men will be chosen by their empowerment status (one empowered; one disempowered. If these data are not available in time, other variables (e.g., age—1 older women and 1 younger woman; 1 older man and 1 younger man) drawn from the quantitative survey list. 4 X 3 X 2= 24 Activity (iv) Key informant interview: Market traders KII with formal-sector traders and with informal-sector traders dealing with main commodities of the community In each upazila, 2 interviewees for each—formal- and informal-sector traders dealing with key commodities in the locality (ideally, 1 man and 1 woman) 4 X 2 X 2 = 16 Total 102 Source: Constructed by authors. 3.2.2 Qualitative Protocol Design Qualitative research methods are particularly useful at exploring perceptions and local understandings of the meanings that people give to their behaviors and beliefs. In this study, the qualitative study sought to clarify respondents’ attitudes toward women’s and men’s involvement in different agricultural value chains and at different nodes along the chain. The study adapted the qualitative protocols developed by IFPRI’s GAAP2 that is constructing another version of the WEAI for use by projects, the pro-WEAI.2 The team reviewed the GAAP2 qualitative data collection instruments and determined which modules would be most useful for exploring the themes of WEAI4VC and that could be done in the time 2 For more information on the pro-WEAI, see http://weai.ifpri.info/. 11 available for the fieldwork. The focus was on collecting information about respondents’ different types of engagement with agricultural value chains and their understanding of concepts of empowerment. 3.2.3 Enumeration Team and Training To learn more about the gender dynamics of agricultural value chains, five IFPRI qualitative Field Officers3 and their qualitative team coordinator4 participated in a training workshop in Dhaka and selected field sites on August 21-29, 2017. The training covered basic concepts related to gender, an overview of gender issues in agricultural value chains, the definition of empowerment and its expression in the Bangladesh context, as well as a range of qualitative data collection and analysis approaches (e.g., coding, categorizing, clustering, and building relationships). The team identified the sample for the qualitative study and practiced techniques of interviewing and analysis. In addition, the group traveled to two different areas to practice interviews and to pilot the interview guides. The revised questions were translated into Bangla. 3.2.4 Fieldwork and Quality Control The Key Informant Interviews (KII) and group interviews were conducted in September and October 2017. Respondents of each category of value chain actor were identified based on the lists of quantitative survey respondents in Gaurnadi Upazila, Barisal District, and Jhikargacha Upazila, Jessore District. The questions used in each interview were tailored to the respondent’s main activity, e.g., as an entrepreneur or an agricultural wage laborer. However, if during the interview, it emerged that the respondent was engaged in more than one income-earning activity, such as farming and daily labor, then the interviewer asked questions about both activities. 3.2.5 Data Entry and Cleaning Following the completion of the fieldwork, the audio recordings were sent to a local firm in Dhaka for transcription. The transcripts were reviewed by the field team multiple times. The final versions were uploaded into NVivo Pro 11 and coded according to a code list prepared by the field officers. Additional analysis was completed in collaboration with the qualitative study team leader. 4. Findings 4.1 Household Characteristics The analysis was conducted on observations with complete data. The quantitative sample consists of 329 dual-headed households (DHH) and 71 female-headed households (FHH) classified as producers, 398 DHH entrepreneur households, and 344 DHH and 56 FHH wage-worker households (Table 4.1). There were only two female-headed entrepreneur households, which were excluded from the analysis due to small sample size.5 Note that respondents—the primary male or female adult in the household— may not always participate in the livelihood activity where their household is assigned, particularly for entrepreneur and wage-work households. The selection criteria for households in each category require that at least one of the respondents participate in production, entrepreneurship, or wage work, respectively. In Bangladesh, households may choose to diversify their livelihood strategies to minimize 3 Waziha Rahman, Shammi Sultana Ferdousi, Shuchita Rahman, Md. Redoy, and S.M. Tahsin Rahaman. 4 Aklima Parvin. 5 The mean age of respondents in the entrepreneur FHH was 32.5 years, and both respondents were secondary school graduates. One of the two women reported that her household was involved in processing. 12 risk, so it is highly unusual to find both male and female respondents engaged in entrepreneurship or doing wage work. Households are equally distributed across the five districts of Barisal, Chuadanga, Jessore, Jhenaidah, and Satkhira. Across all types of actors and household types, most households were Muslim, ranging from 86 percent to 93 percent, followed by approximately 6 percent to 15 percent Hindu. Less than 2 percent of all households were Christian. On average, DHHs were larger than FHHs (3 members vs 2 members). Table 4.1 Household characteristics, by actor and household type Producers Entrepreneurs Wage Workers DHH FHH DHH DHH FHH Number of households Male 329 71 Male 397 Male 344 56 Female 329 Female 398 Female 344 District (%) Barisal 19.45 22.54 20.00 21.22 12.50 Chuadanga 19.76 21.13 20.13 20.64 16.07 Jessore 19.45 22.54 20.13 20.93 14.29 Jhenaidah 20.36 18.31 19.87 20.35 17.86 Satkhira 20.97 15.49 19.87 16.86 39.29 Religion (%) Muslim 87.54 92.96 91.95 91.86 85.71 Hindu 11.55 5.63 8.05 7.27 14.29 Christian 0.91 1.41 0.87 Average household size 3.06 2.07 3.45 2.91 1.84 Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Notes: DHH: Dual-headed households; FHH: Female-only households. FHH among entrepreneurs was excluded due to small sample size (N=2). Male respondents were older than female respondents in DHHs (46 years versus 38 years among producers; 44 years versus 37 years among entrepreneur households; and 41 years versus 34 years among wage-worker households), which is typical of marriage patterns in Bangladesh (Table 4.2). Average age of respondents in female-only wage-worker households (45 years) was higher than that for producer (37 years) and entrepreneur households (33 years). Most respondents had at least some primary school education or were primary school graduates, except male and female respondents in wage worker FHH, who had no schooling. Respondents in entrepreneur households had the highest, and respondents in wage-worker households, the lowest, mean years of schooling. Women in wage-worker FHH had the lowest mean years of schooling at 1.38 years. DHHs owned more land than FHHs for all actors. Entrepreneur DHHs owned the most land (136 decimals6), followed by wage workers (125 decimals) and producers (112 decimals). Producer FHHs owned 50 decimals of land, while wage-worker FHHs owned 40 decimals of land. 6 1 decimal ~ 1/100 acre (40.46m2). 13 Table 4.2 Individual and household characteristics of respondents, by actor and household type Producers Entrepreneurs Wage Workers Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Male (DHH) Female (DHH) Female (FHH) Number of respondents 329 329 71 397 398 344 344 56 Mean age of respondent (years) 46.17 38.28 36.75 44.36 36.58 41.21 34.38 45.07 Education (%) No schooling 29.18 25.84 19.72 16.62 13.60 37.79 25.87 66.07 Some schooling 2.13 0.91 0.76 0.50 2.03 1.16 Some primary school 17.93 14.29 16.90 18.39 15.37 23.55 23.26 21.43 Primary graduate 15.20 19.15 14.08 11.84 13.85 14.83 15.99 5.36 Some secondary school 23.71 30.70 39.44 25.44 38.04 15.70 30.23 7.14 Secondary school graduate 5.78 6.69 7.04 10.58 8.56 4.07 1.74 Completed higher secondary 3.34 1.82 2.82 9.07 6.80 1.74 1.16 College graduate or higher 2.74 0.61 7.30 3.02 0.29 0.58 Madrasa 0.25 Mean years of schooling 4.51 4.71 5.21 6.41 6.24 3.28 4.08 1.38 Mean area of land owned by household (in decimals)† 112.41 50.37 136.02 124.97 40.42 Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Notes: DHH: Dual-headed households; FHH: Female-only households. † Includes agricultural and nonagricultural land. 1 decimal ~ 1/100 acre (40.46m2). Table 4.3 presents the sub-indicators that make up the A-WEAI, as well as potential additional indicators and domains of empowerment that are included in the WEAI4VC, such as intrahousehold relationships, attitudes about domestic violence, and physical mobility. Additional indicators included in the Bangladesh WEAI4VC survey include parda information, awareness of key messages, food insecurity in the household, and women’s assets brought to marriage (female respondents only). The definition of adequacy for each sub-indicator specifies the conditions required for a respondent to be empowered in that sub-indicator. For example, a respondent who participates in at least one community group is adequate in group membership, meaning that s/he is empowered in group membership. 14 Table 4.3 Domains and sub-indicators of the WEAI4VC Domain Sub-indicators Definition of adequacy Livelihoods Input in livelihood activity decisions† Respondent has some input in decisions about livelihood activity or feels they can make decisions in at least two areas of livelihood activities. Access to information Respondent can access information about at least one livelihood activity. Autonomy in livelihood activity Respondent has autonomy (RAI>1)a in at least one livelihood activity. Using income from agricultural and non-agricultural activities Respondent has autonomy (RAI>1) in using income from agricultural and non-agricultural activities Autonomy in Producers Types of crops to grow Respondent has autonomy (RAI>1) in for types of crops to grow. Livestock raising Respondent has autonomy (RAI>1) in livestock raising. Fish production/farming Respondent has autonomy (RAI>1) in fish production/farming. Taking crops/livestock/fish to market Respondent has autonomy (RAI>1) in taking crops/livestock/fish to market. Entrepreneurs Types of products to make and/or sell in the market Respondent has autonomy (RAI>1) in types of products to make and/or sell in the market. Location of the enterprise Respondent has autonomy (RAI>1) in location of the enterprise. Size of the enterprise Respondent has autonomy (RAI>1) in size of the enterprise. Whether to work for someone else for pay Respondent has autonomy (RAI>1) in whether to work for someone else for pay Wage workers Type of work Respondent has autonomy (RAI>1) in type of work Working conditions Respondent has autonomy (RAI>1) in working conditions Resources Ownership of assets† Respondent solely or jointly owns at least one large or two small assets. Rights over assets Respondent solely or jointly has at least one right to at least one agricultural asset that their household owns. Access to and decisions on credit† Respondent solely or jointly makes at least one decision about at least one source of credit that their household used. Access to financial account Respondent has sole or joint access to a financial account. Income Control over use of income† Respondent has at least some input in decisions about income or feels they can make decisions about income, not including minor household purchases. Control over use of agricultural income Respondent has input in decisions related to how to use agricultural income. Leadership Group membership† Respondent participates in at least one community group. Time Workload† Respondent worked less than 10.5 of the last 24 hours. Access to childcare Respondent has someone to care for their child(ren) in their absence. 15 Domain Sub-indicators Definition of adequacy Intrahousehold relationships Mutual respect among household members Respondent has mutual respect with the other respondent in their household, and respondent trusts and is comfortable disagreeing with the other respondent in their household. Attitudes about domestic violence from husband Respondent believes that a husband is never justified in hitting their wife. Mobility Physical mobility Respondent can visit at least two locations once per week. Source: Constructed by authors. † Included in A-WEAI calculation aRAI=relative autonomy index. 4.2 Livelihoods Livelihood activities7 Producers: Most households reported participating in staple grain farming, although participation among FHHs was lower than that of DHHs (Figure 4.1). In producer households, all women (in DHH and FHH) who were interviewed were involved in at least one production activity. A larger proportion of female respondents (in DHHs and FHHs) reported participating in poultry and other small animal raising compared to male respondents, reflecting the common pattern of women’s heavy involvement in livestock production in Bangladesh. Less than 30 percent of households reported participating in fish production. Figure 4.1 Producers: Respondent participation in production activities, by household type 7 Included in A-WEAI calculation. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. 16 Respondents in the qualitative study agreed that production tasks typically differed for men and women as well as by region. Women do not generally work in the rice fields but carry out other agricultural work in home gardens or perform tasks within the confines of their home compounds, including crop sorting, cleaning, and grading, or jute fiber extraction. Entrepreneurs: Approximately 47 percent of male respondents reported participation in retail trading, while 32 percent reported participating in wholesale trading, and 20 percent in food processing (Figure 4.2). However, only 89 out of 398 women in entrepreneur DHHs (22 percent) participated in entrepreneurial activities, consistent with the very low participation rates for female respondents in DHHs. Only 10 percent of women in DHHs participated in retail trading, 7 percent in food processing, and 6 percent in wholesale trading. Figure 4.2 Entrepreneurs: Respondent participation in entrepreneurship activities, by household type Wage workers: Among wage-work households, crop farming was the main reported activity among male respondents in DHHs and female respondents in FHHs (Figure 4.3). Wage work is a relatively uncommon type of livelihood for women in DHHs, with only 37 out of 344 female respondents (11 percent) involved in these activities. A large proportion of female respondents in FHH also reported participating in processing activities, while male respondents reported participating in wholesale service. Female respondents in DHH had very low participation among wage-work activities, with the highest participation (8 percent) in crop farming. This may reflect the social desirability of female seclusion, or purdah, which may constrain women from working outside the homestead for employers who are not family members. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. 17 Figure 4.3 Wage workers: Respondent participation in wage-work activities, by household type The A-WEAI measure for the production domain is the input in productive decisions sub-indicator. In the WEAI4VC, this domain will be called “livelihoods” to reflect the broader set of activities conducted by different types of actors. Additional sub-indicators of access to information about livelihood activities, and autonomy in livelihood activities, are included in the results, but excluded from the A-WEAI calculation. Nearly all respondents had adequate input in livelihood decisions, except for female respondents in entrepreneur and wage worker DHHs, where less than 20 percent had adequacy (Figure 4.4). (In the following bar charts, a # is used to indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level.) If we restrict this indicator to women in DHHs who are themselves participating in entrepreneurship and wage work, adequacy increases to 60 percent for entrepreneurs (n=89) and 95 percent for wage workers (n=37). Table 4.7, presented in Section 4.9 with the A-WEAI results, shows the percent of respondents, by actor and household type, who are adequate in each sub-indicator. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. 18 Figure 4.4 Percent of respondents who are adequate in input in livelihood activity decisions, by actor and household type Most respondents in producer households are also adequate in accessing information about livelihood activities (Figure 4.5). While approximately 20 percent of men in entrepreneur households are adequate, less than 10 percent of female respondents in entrepreneur and wage-work DHHs are adequate in this sub-indicator. If we restrict this sample to women participants only, these estimates do not change. This large gap in adequacy achievements regarding information access across value chain actors indicates gaps in the extension system, which reaches agricultural producers better than entrepreneurs and wage workers. Source: Constructed by authors. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 19 Figure 4.5 Percent of respondents who are adequate in access to information about livelihood activities, by actor and household type Autonomy Autonomy was measured using a short story (or vignette) followed by a series of questions asking respondents if they were similar to hypothetical people who had different motivations for their decisions about the topic. For example, the questions about autonomy in working conditions asked respondents if they were similar to a person who works in unsafe working conditions because s/he has no other choice, because s/he does what s/he is told to do by family members, because s/he does what the family or community expects, or because s/he does what s/he thinks is the best option. Based on these questions, the relative autonomy index (RAI) in working conditions was calculated. Respondents with an RAI greater than one—indicating that their actions were relatively more motivated by their own values than by coercion or fear of others’ disapproval—were considered empowered. Figure 4.6 shows the percent of respondents in producer households who perceive themselves to be like the people described in the stories. Most respondents in production households stated that they are most similar to people in the stories who “do what they think is best” across all the different production decisions, including types of crops to grow, livestock raising, fish production, taking crops or livestock to Source: Constructed by authors. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 20 the market, and how to use income. However, respondents report being least similar to stories of people “doing what they think is best” regarding fish production. Similar to the pattern in production households, most entrepreneur households report being most similar to people in stories that “do what they think is best” across all the different business decisions, including the types of products to make/sell in market, location of enterprise, size of enterprise, and how to use income (Figure 4.7). Respondents in wage-worker households report being most similar to people in stories that “do what they think is best” across all the different employment decisions, including whether to work for someone else for pay, type of work, working conditions, and how to use income. However, women respondents in dual-headed entrepreneur and wage-worker households were less likely to report being similar to people “doing what they think is best” regarding the types of products to make/sell in market, location of enterprise, and size of enterprise in entrepreneur households, and regarding whether to work for someone else for pay, type of work, working conditions in wage-worker households (Figure 4.8). This may reflect the fact that only 22 percent of women in entrepreneur DHHs and 24 percent of women in wage-worker DHHs are themselves engaged in this specific activity. In the qualitative study, most respondents did not view the ability to take their own decisions and to act on them as a positive quality for women. One man, an entrepreneur and a farmer, reflected this widely- held perspective stating “If she takes her own decision without her husband’s consent then other women of this area will not find her to be a good woman even if she is doing good work.…. Men will also not find them good.” Although men were seen as responsible for taking decisions independently and on behalf of the family, the idea that women would act similarly was characterized as disobeying their husbands. 21 Figure 4.6 Producers: Percent of respondents who are like the people in the stories Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71). Figure 4.7 Entrepreneurs: Percent of respondents who are like the people in the stories Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Entrepreneurs – men (N=397); women, dual-adult (N=398). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Types of crops to grow Livestock raising Fish production Taking crops/livestock to market How to use income No other option Does what they tell her/him Wants family/community approval Does what she/he thinks is best 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Male (DHH) Female (DHH) Male (DHH) Female (DHH) Male (DHH) Female (DHH) Male (DHH) Female (DHH) Types of products to make/sell in market Location of enterprise Size of enterprise How to use income No other option Does what they tell her/him Wants family/community approval Does what she/he thinks is best 22 Figure 4.8 Wage Workers: Percent of respondents who are like the people in the stories Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Male (DHH) Female (DHH) Female (FHH) Whether to work for someone else for pay Type of work Working conditions How to use income No other option Does what they tell her/him Wants family/community approval Does what she/he thinks is best 23 Adequacy in autonomy was defined as having a RAI>1 in at least one livelihood activity. Nearly all respondents in producer households achieved adequacy in autonomy, although there was a significant difference by household type (Figure 4.9). Most male respondents in entrepreneur and wage-worker households, and female respondents in FHH in wage-worker households were also adequate. Female respondents in DHH in entrepreneur and wage-worker households were those with the least adequacy in autonomy. Figure 4.9 Percent of respondents who are adequate in autonomy in livelihood activities, by actor and household type Responses in the qualitative study interviews provide insight on issues of autonomy. Most of the interviews stress the importance of a married couple agreeing about the decisions they take as well as in their understanding of who makes which decisions. The interviews reflect that both men and women have areas in which each could legitimately take decisions and act on them either independently or together, for example, where the husband managed the farm and the wife managed the household. However, these gendered areas of decision making were not uniform. In some cases, wives also made decisions around agricultural production or agri-business. Household circumstances influenced these patterns, especially if the husband was working abroad or had died. One informant described this situation as follows, “The woman who has her husband living abroad runs her home according to her wish, takes all the decisions of her children and looks after the family. Such women have no one Source: Constructed by authors. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 24 controlling their movement. They just inform their husbands over the phone.” What was repeatedly noted as important was that the couple in DHHs should agree together on who could decide about what and that women who were heads of households needed to take responsibility for their own decisions. 4.3 Resources Asset ownership8 Across all types of actors, most households own agricultural land or a house or building (Figure 4.10). Agricultural landownership is the lowest for FHHs involved in wage work. Households are also more likely to own poultry, consumer durables, and cell phones than mechanized farming equipment or mechanized transportation. FHHs are less likely to own large or small livestock, or poultry compared to their counterparts. Of respondents who reported that someone in their household owned these assets, more male respondents were likely to solely or jointly own the asset compared to female respondents, except for small livestock, poultry, non-mechanized farming equipment (except female respondents in wage-work DHHs), and consumer durables (Figure 4.11). Rights over assets, including buying and selling the asset, follow similar patterns as sole or joint ownership (Figure 4.12 and Figure 4.13). Female respondents were more likely to have more rights over small livestock, poultry, and consumer durables, but not non- mechanized farming equipment, than male respondents. Among producer households, more female respondents owned storage facilities, and female respondents in FHHs were as likely to own inventory/stock and cell phones compared to men. Among wage workers, female respondents in DHHs were less likely to own any assets compared to at least one of their counterparts. Figure 4.10 Percent of respondents from asset-owning households, by actor and household type 8 Included in A-WEAI calculation. 0% 20% 40% 60% 80% 100% Producers, Men Producers, Women dual-adult Producers, Women, female-only 25 Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 0% 20% 40% 60% 80% 100% Entrepreneurs, Men Entrepreneurs, Women dual-adult 0% 20% 40% 60% 80% 100% Wage Workers, Men Wage Workers, Women dual-adult Wage Workers, Women, female-only 26 Figure 4.11 Percent of respondents from households who solely or jointly own assets owned by their households, by actor and household type Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. 0% 20% 40% 60% 80% 100% Producers, Men Producers, Women dual-adult Producers, Women, female-only 0% 20% 40% 60% 80% 100% Entrepreneurs, Men Entrepreneurs, Women dual-adult 0% 20% 40% 60% 80% 100% Wage Workers, Men Wage Workers, Women dual-adult Wage Workers, Women, female-only 27 Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). Figure 4.12 Percent of respondents from asset-owning households who can purchase assets, by actor and household type 0% 20% 40% 60% 80% 100% Producers, Men Producers, Women dual-adult Producers, Women, female-only 0% 20% 40% 60% 80% 100% Entrepreneurs, Men Entrepreneurs, Women dual-adult 0% 20% 40% 60% 80% 100% Wage Workers, Men Wage Workers, Women dual-adult Wage Workers, Women, female-only 28 Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). Figure 4.13 Percent of respondents from asset-owning households who can rent, sell, give away, or mortgage those assets, by actor and household type 0% 20% 40% 60% 80% 100% Producers, Men Producers, Women dual-adult Producers, Women, female-only 0% 20% 40% 60% 80% 100% Entrepreneurs, Men Entrepreneurs, Women dual-adult 29 Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). Adequacy in asset ownership was defined as owning at least one large asset or two small assets (poultry and small consumer durables). Most respondents achieved adequacy in this category, with female respondents in FHH achieving similar adequacy as male respondents. Female respondents in DHHs achieved the least adequacy (Figure 4.14). 0% 20% 40% 60% 80% 100% Wage Workers, Men Wage Workers, Women dual-adult Wage Workers, Women, female-only 30 Figure 4.14 Percent of respondents in households who are adequate in ownership of assets, by actor and household type A smaller percentage of respondents were adequate in rights over assets (Figure 4.15). Male respondents and female respondents in FHH among producer households achieved similar adequacy rates. However, male respondents in entrepreneur and wage-worker households were more likely to be adequate in the sub-indicator compared to female respondents. Female respondents in FHHs involved in wage work were the least likely to achieve adequacy in rights over assets. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI.Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 31 Figure 4.15 Percent of respondents who are adequate in rights over assets, by actor and household type Access to and decisions on credit9 Most respondents reported that NGOs and banks were the most commonly available formal sources of credit in the community (Figure 4.16). Approximately 20 percent of all respondents said that group- based microfinance was also available. Among informal sources of credit, most respondents reported that they would be able to take a loan or borrow cash/in kind from friends or relatives or informal lenders if they wanted to. Women tended to report higher access to loans from NGOs than men, reflecting the long history of targeting credit to women through NGOs in Bangladesh. 9 Included in A-WEAI calculation. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 32 Figure 4.16 Percent of respondents from households whose households have access to loans, by actor and household type Most respondents were adequate in access to and decisions on credit (Figure 4.17). Female respondents in FHHs were as likely to be adequate as male respondents, and female respondents in DHHs were least likely to be adequate in this sub-indicator. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 0% 20% 40% 60% 80% 100% Men Women, dual-adult Women, female-only Men Women, dual-adult Men Women, dual-adult Women, female-only Producers Entrepreneurs Wage Workers Bank of other finance institution NGO Informal lender Friends or relatives Group based micofinance Informal credit savings groups 33 Figure 4.17 Percent of respondents who are adequate in access to and decisions on credit, by actor and household type Financial accounts Female respondents in DHHs were more likely to solely or jointly have financial accounts with NGOs than male respondents and female respondents in FHHs (Figure 4.18). While female respondents in FHHs were more likely to have financial accounts at banks than males in producer households, the reverse was true for wage-worker households. A larger percentage of female respondents had mobile money financial accounts compared to male respondents, excluding female respondents in FHHs among wage-worker households. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 34 Figure 4.18 Percent of respondents who solely or jointly have financial accounts, by actor and household type Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). Female respondents in producer FHHs, male respondents in entrepreneur households, and female respondents in wage-work DHHs were most likely to achieve adequacy in access to a financial account (Figure 4.19). Among producer households, there were small differences between men and women in DHHs achieving adequacy. In entrepreneur households, however, this difference was much bigger. In wage-work households, women in DHHs were more likely to be adequate in accessing a financial account compared to men, while women in FHHs were the least likely to be adequate in this sub- indicator. 0% 20% 40% 60% 80% 100% Men Women, dual-adult Women, female-only Men Women, dual-adult Men Women, dual-adult Women, female-only Producers Entrepreneurs Wage Workers Bank or other finance institution NGO Mobile money Group-based microfinance 35 Figure 4.19 Percent of respondents who are adequate in access to a financial account, by actor and household type 4.4 Income Control over use of income10 A respondent achieves adequacy in control over use of income if s/he has as at least some input in decisions about income or feels s/he can make decisions about income, not including minor household purchases. Overall, female respondents in DHHs were the least likely to be adequate in control over use of income (Figure 4.20). Most respondents in producer households achieved adequacy in this indicator, although significant differences (p<0.05) exist by respondent type. Male respondents were more likely to be adequate in control over use of income than female respondents in entrepreneur DHHs. Male respondents and female respondents in wage-work FHHs were equally likely to achieve adequacy in this sub-indicator. 10 Included in A-WEAI calculation. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 36 Figure 4.20 Percent of respondents who are adequate in control over use of income, by actor and household type Figure 4.21 depicts the percent of respondents who are adequate in control over use of agricultural income. Female respondents in FHHs were the most likely to achieve adequacy in this sub-indicator, while female respondents in DHHs were the least likely to do so. Across value chain actors, adequacy in control over use of agricultural income was substantially high among producer households, a relatively small gap in adequacy between men and women in DHHs. There were large gender gaps in control over the use of income in entrepreneur and wage-worker, dual-adult households, with much smaller proportions of women in those households achieving adequacy compared to men. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at he 5 percent level. Number of observations: Producers – men (N=329); women, dual-adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 37 Figure 4.21 Percent of respondents who are adequate in control over use of agricultural income, by actor and household type 4.5 Leadership Group membership11 Respondents reported that very few groups were available in the community. Among groups available, female respondents were more likely than male respondents to be active members of credit or microfinance groups and religious groups, reflecting the women-oriented programming of Bangladeshi NGOs (Figure 4.22). Male respondents from entrepreneur households were more likely to be involved in trade associations than female respondents in DHHs. 11 Included in A-WEAI calculation. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which the values for men, women from DHHs, and women from FHHs are significantly different at the 5 percent level. Number of observations: Producers – men (N=329); women, dual- adult (N=329); women, female-only (N=71); Entrepreneurs – men (N=397); women, dual-adult (N=398); Wage workers – men (N=344); women, dual-adult (N=344); women, female-only (N=56). 38 Figure 4.22 Percent of respondents who are active members of community groups, among groups that are available in the community, by actor and household type Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Adequacy in group membership is defined as being an active member of at least one group in the community. Female respondents were more likely than male respondents to achieve adequacy in this sub-indicator (Figure 4.23). Male respondents from wage-work households were the least likely to be adequate in group membership. 0% 10% 20% 30% 40% 50% Men Women, dual- adult Women, female-only Men Women, dual- adult Men Women, dual- adult Women, female-only Producers Entrepreneurs Wage Workers Agricultural/Livestock/Fisheries Water users Forest users Credit or microfinance Mutual help or insurance Trade association Civic group Religious group Other women's group 39 Figure 4.23 Percent of respondents who are adequate in group membership, by actor and household type 4.6 Time Workload12 In this sub-indicator, workload was defined as time spent on cooking and food preparation; caring for children and adults, including the sick and elderly; household chores; shopping or receiving services such as health services; weaving, sewing, and textile care for home use; wage work; work for one’s own business; and work related to farming, processing, trading, and marketing of agricultural products and by-products. In the WEAI4VC, the time domain will also include access to childcare; this new domain will apply only to female respondents. The average total time spent on workload, including childcare, does not vary significantly by respondent type or actor (Figure 4.24 and 12 Included in A-WEAI calculation. Source: Women’s Empowerment in Agriculture Index for Value Chain (WEAI4VC) Quantitative Survey 2017, IFPRI. Note: # indicate categories for which