How to Improve Indicators on Women’s Economic Empowerment Strengthening Guidance Note on Gender-Sensitive Gender Survey Design and Implementation Statistics How to Improve Indicators on Women’s Economic Empowerment Guidance Note on Gender-Sensitive Survey Design and Implementation Published March 2022 Strengthening Gender Statistics This guidance note was developed by Anna Bonfert (lead and corresponding author), Heather Moylan, Miriam Muller, and Talip Kilic under the World Bank Strengthening Gender Statistics Project with support from the Bill and Melinda Gates Foundation. The team is grateful for the valuable comments and contributions from Kathleen Beegle, Maria Eugenia Genoni, Daniel Halim, Rose Mungai, and Amparo Palacios- Lopez. Additionally, we would like to thank Georgina Maku Cobla and Kiran Correa for research assistance and Honora Ann Mara for editorial support. Contents Introduction................................................................................................................5 Overview of gender data gaps ................................................................................................... 5 Efforts to narrow economic gender data gaps ....................................................................... 6 Why this guidance note? ............................................................................................................. 8 PART I Questionnaire Design for Better Gender Statistics..............................................11 (A) Asset ownership and control.............................................................................................. 12 Recommendations at a glance................................................................................................... 12 Sample module ............................................................................................................................ 12 Related SDG indicators and UNSD minimum indicators ...................................................... 13 Recommendations explained..................................................................................................... 13 (B) Work and employment ........................................................................................................ 17 Recommendations at a glance................................................................................................... 17 Sample module ............................................................................................................................ 19 Related SDG indicators and UNSD minimum indicators ...................................................... 19 Recommendations explained..................................................................................................... 20 (C) Entrepreneurship.................................................................................................................. 25 Recommendations at a glance................................................................................................... 25 Sample module ............................................................................................................................ 26 Related indicators ........................................................................................................................ 26 Recommendations explained..................................................................................................... 26 (D) Spotlight on respondent selection and importance of self-reporting......................... 28 PART II Data Collection Process for Better Gender Statistics...........................................31 Recommendations at a glance.................................................................................................. 32 Sample field protocols............................................................................................................... 32 Recommendations explained.................................................................................................... 32 References and Additional Resources...................................................................36 Appendix A Why Focus on Assets, Work and Employment, and Entrepreneurship?...........................................................................................39 Appendix B About the Strengthening Gender Statistics Project.........................41 Appendix C Relevant Indicators in the Economic Domain....................................45 Introduction This guidance note provides recommendations on how to operationalize methodological advances for indicators related to women’s economic empowerment in survey design and implementation, focusing on three key dimensions: asset ownership and control, work and employment, and entrepreneurship. It seeks to facilitate the transfer of internationally recommended best practices into country-level, household survey– specific advice. Applying the targeted recommendations presented here will allow countries to improve their ability to collect meaningful gender statistics that serve as a critical input to designing policies to improve economic opportunities for all. Overview of gender data gaps Globally, increased demand for gender data has not been matched by increased supply. Gender data are important for understanding gendered differences in opportunities, access to services, and outcomes across all domains and over the entire life cycle. They are also a prerequisite to designing policies to address inequities. The shift from the Millennium Development Goal (MDG) to the Sustainable Development Goal (SDG) era has increased demand for sex-disaggregated and nuanced data.1 Not only the SDGs but also national development strategies and World Bank priorities, including the World Bank Gender Strategy and the Data for Policy (D4P) package,2 call for increased emphasis on more and better gender- disaggregated data. Currently, however, only 22 percent of the 54 gender-specific indicators in the SDGs are produced regularly worldwide (UN Women 2018). Similarly, countries have made unequal progress in full coverage of the United Nations (UN) Statistical Division’s Minimum List of Gender Indicators (“UNSD minimum indicators”). Consequently, meeting the steadily rising demand for gender statistics requires more concerted efforts across data producers, funders, and users to ensure that the promise of gender data is fulfilled. 1 See for example a mapping of gender data gaps by Data2X. 2 The D4P package targets a core set of economic, social, and sustainability statistics. In order to inform the design and monitoring of public policies, it will help countries improve the availability of multi-disaggregated statistics, for example, by sex and disability. It will seek to increase individual-level, non-proxy respondent data in official (or national) household surveys on physical and financial assets, employment/labor markets outcomes, and (selected) human capital indicators. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 5 Not all gender data gaps are created equal, and they differ significantly by topic. Whereas health and education data tend to be available, data are much scarcer in some other domains. Even within sectors, availability exists along a broad spectrum—ranging from routinely published indicators to partially available indicators to others that are not available at all. Indicators on women’s economic empowerment are particularly affected by availability and quality issues, despite the importance of these indicators for effective policy making (see appendix A). Although definitions of women’s economic empowerment vary, research highlights that it has three key ingredients: access to more and better jobs, entrepreneurship, and ownership of productive assets such as mobile phones, land, and financial accounts. A priority in addressing gender data gaps is therefore to increase availability of the indicators related to economic status, including those related to these three areas.3 Measuring women’s economic empowerment is not straightforward. Many countries face challenges in producing and disseminating indicators related to women’s economic empowerment. For example, only 5 of the 15 sex-disaggregated SDG indicators in the economic realm are classified as Tier 1.4 Globally, close to 80 percent of countries regularly produce sex-disaggregated statistics on mortality, labor force participation, and education and training; but less than a third of countries disaggregate statistics by sex on informal employment, entrepreneurship (ownership and management of a firm or business), gender-based violence, and unpaid work (Buvinic, Furst-Nichols, and Koolwal 2014). Efforts to narrow economic gender data gaps Significant investments have been made to address methodological concerns and practical challenges related to economic gender data. The SDGs require sex-disaggregated data on asset ownership, labor, time use, and roles in family enterprises. To meet these new ambitious data requirements, significant effort has gone into developing new guidelines and standards. In recent years, new 3 Although the focus of this guidance note is on closing gaps related to women’s economic empowerment, the Strengthening Gender Statistics project will leverage the work of other key agencies such as UN Women, Data2X, and the International Food Policy Research Institute, which have worked on assessing and filling gender data gaps in other critical areas. For example, the UN Women’s Women Count program aims to address gender data gaps in violence against women and in time use. 4 A Tier 1 Indicator is conceptually clear, has an internationally established methodology, and relies on data that are regularly produced by countries for at least 50 percent of countries and of the population in every region where the indicator is relevant. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 6 international guidelines and recommendations for improving questionnaire design and data collection have been released, building on years of applied research in countries across the globe. Specifically, the efforts of the following initiatives and partnerships have significantly enhanced the global evidence base on women’s economic empowerment indicators: • The Gender Asset Gap Project was created in 2009 to generate evidence on the importance and feasibility of collecting individual-level data on wom- en’s and men’s access to and ownership of property.5 • The UN Evidence and Data for Gender Equality (EDGE) initiative6 launched in 2013 to develop international guidelines on asset ownership and en- trepreneurship, including field-testing collection methods in six countries. Findings are encapsulated in the Guidelines for Producing Statistics on As- set Ownership from a Gender Perspective (United Nations 2019). A related document is the guidance document by FAO, World Bank, and UN-Habitat (2019), Measuring Individuals’ Rights to Land, providing recommendations on survey data collection for SDG indicators 1.4.2 and 5.a.1, related to land tenure security and agricultural land ownership. • The Methodological Experiment on Measuring Asset Ownership from a Gender Perspective7 (MEXA), a randomized household survey experiment, was implemented as part of the EDGE initiative in Uganda in 2014 to test the relative effects of different approaches to respondent selection and questionnaire design on estimates of ownership of and rights to physical and financial assets. It also provided inputs for the UN guidelines men- tioned previously. • The Women’s Work and Employment Partnership (WWEP), a collaborative initiative of the World Bank, the International Labor Organization (ILO), and Data2X, has conducted studies to inform updated and improved guidelines for measuring women’s participation in different forms of work—paid and unpaid—to guide policy and program development. • The Living Standards Measurement Study – Plus (LSMS+) program was established in 2016 to enhance the availability and quality of intra-house- 5 The project conducted field surveys in Ecuador, Ghana, and the state of Karnataka in India. Measures of gender asset and wealth gaps have been calculated for the three countries. 6 EDGE was created as part of a collaboration between the US government, the UN Statistics Division, and UN Women, along with key regional and international agencies including the Organisation for Economic Co-operation and Development and the World Bank. 7 MEXA was a collaborative effort of LSMS, EDGE, and the Uganda Bureau of Statistics. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 7 hold, self-reported, individual-disaggregated survey data collected in low- and middle-income countries on key dimensions of men’s and women’s economic opportunities and welfare.8 Why this guidance note? Despite these efforts, the methodological advances in measuring women’s economic empowerment have not translated seamlessly into survey operations. The findings from the new guidelines are often dense, hard to digest, and scattered across different documents, making it difficult for survey practitioners to consider the extensive recommendations during questionnaire review and ultimately resulting in limited adoption in survey practice. To overcome the “know-do gap,” this guidance note distills the main recommendations from these initiatives to support their operationalization in ongoing and future survey operations. Box 1 World Bank Strengthening Gender Statistics Project Launched in October 2020 with support from the Bill and Melinda Gates Foundation, the Strengthening Gender Statistics project catalyzes ongoing statistical capacity projects in International Development Association partner countries by providing technical assistance to national statistical offices on survey design and postsurvey analysis and dissemination. Premised on a demand-driven approach, the project seeks to reduce gender data gaps in the economic domain—specifically in asset ownership, employment, and entrepreneurship—in 10 partner countries by the end of 2023 and to create spillover influence to increase focus on gender gaps in broader World Bank statistical capacity operations. The project is a joint initiative of the Gender Group, Poverty and Inequality Global Practice, and Development Data Group’s Living Standards Measurement Study (LSMS) team. Harnessing available evidence and guidelines can help to bridge current gaps in data production. Although data production challenges are often recognized in global assessments of gender data gaps, addressing these areas requires a concerted effort to work at the country level. The World Bank’s Strengthening Gender Statistics (SGS) project9 (see box 1) supports the scale-up of the production 8 As of July 2021, nationally representative surveys directly supported by LSMS+ include the 2016 Malawi Integrated Household Panel Survey, 2019/20 Tanzania National Panel Survey – Extended Panel, 2018/19 Ethiopia Socio-Economic Survey, 2019/20 Cambodia Socio-Economic Survey, and 2021 Sudan Labor Market Panel Survey. 9 For additional information on the SGS project, please see appendix B. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 8 and use of gender data by leveraging ongoing statistical capacity operations. The project does not attempt to tackle the systemic challenges surrounding the production of gender data but instead follows a pragmatic approach focusing on specific and targeted adjustments to survey design and implementation. The proposed adjustments presented in this guidance note are firmly grounded in technical international guidelines and standards. Equally important, the adjustments are within reach of NSOs and can have an outsized effect on their capacity to generate more and better gender statistics. This guidance note provides targeted and strategic advice on how to adapt existing data collection tools and processes. To help scale up the production and use of gender data, it synthesizes the main findings from past investments in survey methodology that have shed light on how to generate more accurate and relevant data on women’s work and their access to productive assets. As such, it is intended as a resource for NSOs and other survey practitioners, both inside and outside the World Bank, who are pressed for time and looking for streamlined suggestions on how to make their questionnaires more effective in collecting information on assets, employment, and entrepreneurship. Importantly, the practical advice provided herein applies specifically to multitopic household surveys. Other types of data such as administrative data and enterprise data are covered in separate guidance documents produced under the SGS project. Part I of this guidance note provides specific recommendations on how to incorporate a gender perspective in questionnaire design. Cognizant that all gender-informed questionnaires also need to be properly implemented to generate meaningful gender statistics, part II of the guidance note then discusses recommendations to ensure gender-sensitive data collection to avoid common pitfalls in the field. Relevant topics to explore gendered economic status and impacts are grouped into three thematic areas: asset ownership and control, work and employment, and entrepreneurship. Key recommendations10 to overcome well-known challenges in indicator production are summarized at the beginning of each section in part I. Each recommendation highlights which data gap it helps to address and is linked to one or more paragraphs in the subsection titled “Recommendations explained,” which provides a detailed explanation of why this indicator or broader subject matters 10 Please note that this list is not all-encompassing and therefore is understood as a starting point when creating or reviewing a household survey questionnaire with a gender lens. Additional considerations may emerge in a specific country context. To capture the required granularity and nuance, it is recommended to validate the proposed questions as part of the regular stakeholder and expert consultations. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 9 from a gender perspective. To facilitate a quick overview of which data gaps specifically can be filled with the help of this guidance note, each section includes relevant SDG indicators and UNSD minimum indicators. For each economic domain, a sample module is provided containing relevant questions to generate these priority indicators. Filling these data gaps will accelerate progress toward complete measurement of SDG indicators and UNSD minimum indicators. Given global policy commitments to achieve the SDGs and World Bank operational requirements linked to increasing the availability of UNSD indicators at the country level (see box 2), these indicators present a logical entry point. Depending on country context and priorities, additional indicators should be considered. Box 2 UNSD Minimum Indicators and the Gender Tag in StatCap Operations The United Nations Statistical Division (UNSD) minimum list of gender indicators was prepared by the Inter-Agency and Expert Group on Gender Statistics as a guide for the national production and international compilation of gender statistics. Revised in 2019, the list contains 52 quantitative indicators and 11 qualitative indicators related to national norms. World Bank statistical capacity operations aiming to receive the gender tag can consider including the indicator “Number of UNSD Minimum List gender indicators available using data collected within the past 5 years” in their results framework. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 10 PART I Questionnaire Design for Better Gender Statistics (A) Asset ownership and control Recommendations at a glance Table 1 summarizes the key recommendations on asset ownership and control, highlights which data needs will be addressed by taking them into consideration, and cross-links to the detailed explanations in the subsection titled “Recommendations explained.” Table 1 Summary of Key Recommendations on Asset Ownership and Control Corresponding # Recommendation paragraph Define relevant asset categories, and ensure the inclusion of priority assets in the questionnaire. • To address the data needs for SDG indicator 8.10.2/UNSD minimum indicator I.11, (A1) (A2) 1 consider including questions on bank account ownership. • To address the data needs for SDG indicator 5.b.1/UNSD minimum indicator I.18, consider including questions on mobile phone ownership. Distinguish between core assets and other assets. For core assets, ask about different 2 (A3) (A4) types of ownership. 3 Allow for multiple owners to be recorded for the same asset. (A5) Land is a core asset, and all types of land (agricultural and nonagricultural) should be 4 (A6) considered. For land, be sure to ask questions on the bundle of rights. 5 • This can help satisfy the data requirement for SDG indicator 5.a.1/UNSD minimum (A7) indicator I.12 and SDG indicator 1.4.2 on secure land ownership. For all questions on personal ownership of and rights to physical and financial assets, encourage self-reporting instead of proxy reporting. 6 See section (D) • For multitopic surveys consider collecting data on assets immediately after modules on education and health. Source: World Bank. Note: SDG = Sustainable Development Goal; UNSD = United Nations Statistical Division. Sample module • A sample module covering the different types of asset ownership as well as different rights to an asset can be found in Annex 1 of the LSMS+ oper- ational guidance. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 12 • On land specifically, see Questionnaire Module Version 4 of FAO/WB/UN Habitat Guidance. Related SDG indicators and UNSD minimum indicators Table 2 provides the SDG indicators and UNSD minimum indicators related to asset ownership, including the indicator description and link to the indicator’s metadata file. Table 2 SDG and UNSD Minimum Indicators Related to Asset Ownership UNSD SDG Indicator description minimum indicator (with hyperlink to metadata file) indicator Proportion of adults (15 years and older) with an account at 8.10.2 I.11 a bank or other financial institution or with a mobile-money service provider 5.b.1 I.12 Proportion of individuals who own a mobile telephone, by sex (a) Proportion of total agricultural population with ownership or secure rights over agricultural land, by sex; and (b) share of 5.a.1 I.18 women among owners or rights-bearers of agricultural land, by type of tenure Proportion of total adult population with secure tenure rights to 1.4.2 land, who (a) have legally recognized documentation and (b) per- ceive their rights to land as secure, by sex and type of tenure Source: World Bank. Note: SDG = Sustainable Development Goal; UNSD = United Nations Statistical Division. Recommendations explained (A1) What asset categories should be considered? Assets vary in nature and can include physical items such as housing, land, livestock, household durables, nonfarm business assets, and financial assets. The UN EDGE guidelines recommend that countries collect, at a minimum, information on a “core” set of assets that has been found to make up most of personal wealth. This core set of assets comprises principal dwelling, agricultural land, and other real estate. Additionally, the UN EDGE guidelines advise countries to determine additional assets on which to collect data based on policy needs and prevalence of assets within a given country. Asset categories to consider include nonagricultural enterprise assets, livestock, How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 13 large and small agricultural equipment, financial assets and liabilities, valuables, and consumer durables such as vehicles. Under the LSMS+ program, for example, the three asset classes considered in all six countries were land (both agricultural and nonagricultural), financial accounts, and mobile phones. Box 3 Ownership Definitions Reported ownership refers to the persons who consider themselves to be owners of the asset in question, irrespective of whether they possess legal, or documented, ownership of the asset. Economic ownership refers to the persons who would have control over the proceeds from the sale of an asset. Documented ownership refers to the existence of any document recognized by the government that an individual can use to claim ownership rights in law over an asset by virtue of the individual’s name being listed as an owner on the document. Bundle of rights to an asset includes the right to sell, right to bequeath, right to rent out, right to use as collateral, and right to make improvements/invest. Source: LSMS+ (A2) Defining relevant asset classes and categories requires country- specific knowledge to ensure proper phrasing of questions. Asset-related questions need to be informed by legal context, and relevant asset types may differ by country. The UN EDGE guidelines therefore recommend reviewing the following when designing a survey questionnaire: • The legal framework and customary norms that govern property rights, in- cluding those related to marital and inheritance regimes, across different areas of the country • The link between asset ownership and other development issues, includ- ing poverty, livelihoods, entrepreneurship, agriculture, women’s empower- ment, and gender equality • Government programs and policies regarding core assets, including the distribution and titling of land and housing • Existing quantitative studies providing information on the prevalence of asset ownership, at both individual and household levels, and wealth dis- tribution across different population groups; and studies related to land tenure systems across the country How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 14 (A3) For core assets, the questionnaire should distinguish between different types of asset ownership. Specifically, ownership can be defined in terms of who states that the asset is theirs, who gets to use the asset for market transactions, and who has their name on the ownership title (see box 3). The distinction between reported, economic, and documented ownership is not merely semantic but carries important implications for gendered patterns of intergenerational wealth accumulation (Doss, Grown, and Deere 2008). (A4) A comprehensive analysis of assets extends beyond ownership and considers individuals’ rights to and control over assets. Apart from ownership, the rights to an asset can also differ. The bundle of rights that should be collected for core assets—that is, land and real estate—includes the right to sell, right to bequeath, right to rent out, right to use as collateral, and right to make improvements or invest. The questionnaire should capture the capacity to exercise rights independently by asking permission/consent provider(s). Finally, there is also a difference between ownership and control of assets, which can overlap but do not need to. This distinction is especially relevant for plot ownership and management, which often do not coincide and should properly be accounted for in the questionnaire. Table 3 summarizes the scope of recommended data collection according to the asset class. Table 3 Scope of Recommended Asset Data Collection Asset Reported Economic Documented Bundle of classes ownership ownership ownership rights Land X X X X Financial accounts X X Mobile phones X Livestock X Durables X Source: World Bank. (A5) Assets can be jointly or individually owned. Collecting information on exclusive or mutual ownership is particularly important to understand marital and inheritance regimes. It is recommended to ask who owns, manages, and inherits key assets, allowing for the identification of at least three individuals within the How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 15 household for each question (Hasanbasri et al. 2021a). The questionnaire should allow for a unique identification of any joint owners or right holders both inside and outside the household. To capture individual-disaggregated microdata, specific questions in each module should be linked back to the household roster for the unique identification of individuals who are associated with specific phenomena, such as ownership of an asset or management of an agricultural plot. Box 4 discusses how to address conflicting answers on asset ownership. (A6) Given its importance for wealth storage and accumulation, land is considered a core asset and as such deserves special attention in the questionnaire. To collect land data well, it is suggested to start with a listing of all the land owned by the household. The recommendation is to collect data on all land—both agricultural and nonagricultural—to address the data needs for SDG indicators 1.4.2 and 5.a.1. This recommendation covers all land owned or accessed via use rights and follows recent recommendations (see FAO, World Bank, and UN- Habitat 2019). Importantly, it also includes the land/parcel on which the dwelling is located, meaning that ownership/rights questions should be shifted from the housing module. (A7) If possible, ownership and rights questions should be asked separately to all adult individuals in the household for all parcels that they own. The land roster would include all parcels used, owned, or accessed via use rights by any household member, irrespective of use (that is, including agricultural, residential, pastoral, forest, and commercial parcels). This roster should be created with participation from as many household members as possible to ensure its completeness. The recommendation for land ownership is to provide identification of the (a) respondent, (b) reported owners (at least two individuals), (c) documented owners (at least two individuals) along with information on formal document, (d) bearers of right to sell (at least two individuals), and (e) bearers of right to bequeath (at least two individuals).11 11 If this is not possible, it is encouraged that the parcel roster questions be designed as in Questionnaire Module Version 4 of the linked document—at the parcel level, proxy respondent approach. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 16 Box 4 How to Deal with Conflicting Answers on Asset Ownership The United Nations’ Evidence and Data for Gender Equality (UN EDGE) guidelines recommend collecting rosters of assets (except financial assets) at the household level to eliminate the complication of reconciling assets reported by multiple respondents. Discrepancies are still possible, however, if multiple individuals within the household report ownership and rights indicators. Although the Living Standards Measurement Study – Plus modules found a relatively high degree of intrahousehold agreement (but cases of disagreement were not negligible), the UN EDGE guidelines recognize that agreement can vary significantly by context. They recommend two approaches for reconciling discrepancies: (a) ignore discrepancies and assign ownership according to the ownership status reported by each individual respondent, and (b) override any discrepancies in the ownership reported by individual respondents on exclusive or joint ownership so that persons will be considered exclusive owners if they are the only persons reporting owning an asset in the same household. If other household members also report owning the same specific asset, these persons, along with all others reporting owning the asset, will be considered joint owners. Source: United Nations 2019. (B) Work and employment Recommendations at a glance Table 4 summarizes the key recommendations on work and employment, highlights which data needs will be addressed by taking them into consideration, and cross- links to the detailed explanations in the subsection titled “Recommendations explained.” Table 4 Summary of Key Recommendations on Work and Employment Corresponding # Recommendation paragraph Follow standards for definition of work and employment adopted by the 19th International Conference of Labour Statisticians (ICLS-19). • To address the data needs for UNSD minimum indicator I.2 on average number of hours spent on total work (total work burden) 1 (B1) (B2) • To address the data needs for UNSD minimum indicator I.3 on labor force participation rate for persons aged 15-24 and 15+ • To address the data needs for UNSD minimum indicator I.15 on prime-age employ- ment-to-population ratio by sex, household type, and presence of children Record the number of hours spent on specific activities such as preparing meals, childcare, and cleaning. 2 (B3) • To address the data needs for SDG indicator 5.4.1/UNSD minimum indicator I.1 on average number of hours spent on unpaid domestic and care work How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 17 Corresponding # Recommendation paragraph Collect information on primary and secondary wage jobs over the last seven days. • To address the data needs for SDG indicator 8.5.1/UNSD minimum indicator I.13 on gender wage gap • To address the data needs for SDG indicator 8.8.1 on occupational injuries • To address the data needs for SDG indicator 5.5.2 on proportion of women in manage- 3 (B4) rial positions • To address the data needs for UNSD minimum indicator I.8 on percentage distribution of employed population by sector • To address the data needs for SDG indicator 8.3.1/UNSD minimum indicator I.9 on informal employment in non-agriculture employment Add a seven-day employment screening question for different activities at the beginning of the employment module, and include recovery questions to capture work and employment 4 information. (B5) (B6) • To address the data needs for UNSD minimum indicator I.14 on proportion of employed working part-time Explicitly ask about own use production and contributing family workers. • To address the data needs for UNSD minimum indicator I.4 on proportion of employed 5 who are own-account workers (B7) • To address the data needs for UNSD minimum indicator I.5 on proportion of employed who are contributing family workers Include a question on intended destination of agricultural production to distinguish between market and own use. 6 (B8) • To address the data needs for SDG indicator 2.3.2 on average income of small-scale food producers 7 Ask about the control over earnings. (B9) Include questions on unemployment and job searching over the last 30 days, and collect information on whether individuals are searching for more or different work. • To address the data needs for SDG indicator 8.5.2/UNSD minimum indicator I.10 on 8 (B10) unemployment • To address the data needs for SDG indicator 8.6.1/UNSD minimum indicator I.7 on proportion of youth not in education, employment or training Encourage self-reporting rather than proxy reporting for the labor module. • To improve questionnaire flow, consider collating preexisting individual-level survey (B11) and see 9 modules, such as education and health, with those that report on work and employ- section (D) ment, nonfarm enterprises, and asset ownership. Source: World Bank. Note: SDG = Sustainable Development Goal; UNSD = United Nations Statistical Division. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 18 Sample module • A recommended sample labor questionnaire module is in Annex I of the LSMS Guidebook on Employment and Own-Use Production in Household Surveys. Related SDG indicators and UNSD minimum indicators Table 5 provides the SDG indicators and UNSD minimum indicators related to employment and work, including the indicator description and link to the indicator’s metadata file. Table 5 SDG Indicators and UNSD Minimum Indicators Related to Employment and Work UNSD SDG Indicator description minimum indicator (with hyperlink to metadata file) indicator 5.4.1 I.1 Proportion of time spent on unpaid domestic and care work, by sex, age, and location I.2 Average number of hours spent on total work (total work burden), by sex I.3 Labor force participation rate for persons aged 15–24 and 15+, by sex I.4 Proportion of employed who are own-account workers, by sex I.5 Proportion of employed who are contributing family workers, by sex Proportion of youth (aged 15–24 years) not in education, employment, or training, by sex and 8.6.1 I.7 age Percentage distribution of employed population by sector, each sex (sectors here refer to I.8 agriculture, industry, and services) 8.3.1 I.9 Proportion of informal employment in nonagricultural employment, by sex 8.5.2 I.10 Unemployment rate, by sex, age, and persons with disabilities Average hourly earnings of female and male employees, by occupation, age, and persons 8.5.1 I.13 with disabilities I.14 Proportion of employed working part-time, by sex I.15 Prime-age employment-to-population ratio by sex, household type, and presence of children 2.3.2 Average income of small-scale food producers, by sex and indigenous status How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 19 UNSD SDG Indicator description minimum indicator (with hyperlink to metadata file) indicator 5.5.2 Proportion of women in managerial positions 8.7.1 Proportion and number of children aged 5–17 years engaged in child labor, by sex and age 8.8.1 Frequency rates of fatal and nonfatal occupational injuries, by sex and migrant status Source: World Bank. Note: SDG = Sustainable Development Goal; UNSD = United Nations Statistical Division. Recommendations explained (B1) Consistent adoption of the 19th International Conference of Labour Statisticians (ICLS-19) definitions of work and employment will enable a better picture of gendered differences in employment. The new international definitions for work and employment (see box 5 and figure 1) require survey instruments to be updated with a skilled understanding of the nuances of new concepts, without which countries will lag in producing new measures accurately. In some questionnaires, “work” and “employment” continue to be used interchangeably, even though the definitions differ. Employment includes wage and salaried work and any for-profit work in household enterprises or self-employment. Key gender-relevant changes in the new framework are the classification of own-use production of goods as work (not employment) and the inclusion of own-use provision of services as work. Box 5 Definitions of Work and Employment Work refers to any paid or unpaid activity to produce goods and services (that is, any productive activity). This includes employment as well as time spent providing or creating services and goods for own use. This “own-use production” includes both goods (for example, farming, fishing, preserving foods for storage, sewing your own clothes) and services (for example, childcare, eldercare, food and meal preparation for immediate consumption, other household chores). Employment is a specific subset of work, defined as work performed for pay or profit. Source: ICLS 2013. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 20 Figure 1. The ICLS-19 Definition of Work and the SNA 2008 Intended destination of For own final use For use by others product Production work… Volunteer work… Employment Unpaid Other …in market …in households Forms of work …of (work for pay trainee work and non- producing… …of goods or profit) work activities services market units …goods …services Relation to Activities within the SNA Production Boundary 2008 SNA Activities within the SNA General Production Boundary Source: ICLS-19 2013. Note: SNA = System of National Accounts. Box 6 Underreporting Women’s Economic Activity: Evidence on Reasons and Implications from Honduras A study assessing the discrepancy between women’s reported and actual engagement in economic activities in rural Honduras found that women underreport their engagement in economic activities, including production for own consumption, production of market goods, and remunerated services and commerce. Consequently, Honduras’s rural female labor force participation rate is potentially underestimated by up to 23 percentage points. The study also shed light on the reasons for underreporting of women’s economic activity. On the one hand, women regard themselves (and are identified) primarily as housewives, and the concepts of housework and employment are taken as mutually exclusive. On the other hand, women define “employment” on the basis of a set of necessary characteristics that exclude many of their own activities. Specifically, work needs to (a) be conducted physically outside the home, (b) be in exchange for money, and (c) entail sufficient time commitment. Applying the standards from the 19th International Conference of Labour Statisticians would hence dramatically change official statistics. Source: Muller and Sousa 2020. (B2) The updated ICLS-19 standards offer several opportunities to better capture women’s work and employment. First, the ICLS-19 standards recognize that individuals may be engaged in different forms of work simultaneously during a particular reference period and the labor module should attempt to capture this concurrence as well as possible. This recognition is particularly relevant for women who often engage in more than one form of paid and/or unpaid work. Second, the revised labor standards also clearly define the boundaries between market and own-use production. Importantly, distribution of individuals across the different forms of work is not gender-neutral. Broadly, women tend to perform more activities that fall under production of own-use services than do men, who tend to provide more own-use goods. From a gender perspective, it is important How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 21 to ensure that people with small-scale or informal jobs or who help in family businesses or on farms are captured by the surveys as employed (if the work is for pay or profit) or engaged in the own-production of goods (if the work is intended for own use). Currently, employment figures in many countries still do not reflect this updated classification, thereby undercounting women’s work and employment, and ultimately their contribution to society. (B3) Given women’s outsized role in the provision of domestic and care work, the inclusion of own-use provision of services as work has significant implications. Unpaid domestic and care work is often inadequately accounted for in traditional survey design, which prevents a better understanding of gendered labor allocation patterns. This type of work can absorb many hours daily or weekly, particularly for women. Persons who are employed, unemployed, and outside the labor force can engage in domestic and care work. Collecting these data can be achieved by asking survey participants to record the number of hours spent on specific activities such as preparing meals, childcare, and cleaning, or by including time diaries in household surveys or conducting specialized time-use surveys. We recommend that countries implement specialized time-use surveys or at least incorporating a time use module into their standard multi-topic surveys, however if this is not possible then adding stylized questions should be considered as a last option.12 (B4) A thorough understanding of the primary and secondary wage jobs is indispensable for meaningful gender analysis. The international standards establish that the share of employed population, share of employment in different sectors, and employment in informality are all derived from information collected on the main and secondary jobs in the last seven days. The questionnaire should contain information on primary employment—including characteristics by industry, sector, working time, earnings, job attachment, and level of informality— to enable a nuanced subsequent analysis. If needed, for entrepreneurship and other indicators, a module on jobs held over the last 12 months can be added with questions that cannot be addressed with the seven-day module, for example when there is high seasonality. (B5) Prevailing social norms and gender biases on what constitutes work can skew labor participation figures and estimates of hours worked by gender. One issue is that female employment often looks different than male employment. For example, women, who make up a large proportion of contributing family workers (that is, household members whose work supports other members engaged in self- 12 The World Bank is currently involved in methodological work to produce updated guidance on measuring time use. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 22 employment, specifically in running a farm or nonfarm enterprise), may not self- identify as having employment. According to the new classifications, however, they are in fact employed because their work implies indirect compensation in the form of household income if the enterprise operates for profit. Similarly, women engaging in small-scale or occasional work may not consider this work. As a result, work is underreported unless the questionnaire explicitly addresses these issues (see box 6). (B6) Employment screening questions and recovery questions can help to identify all forms of work undertaken by all household members. To fully account for differences in men’s and women’s labor contributions, it is therefore important to include in the questionnaire a broad range of activities reflecting different types of work. The LSMS+ guidance suggests adding a seven-day employment screening question for different activities at the beginning of the employment module to elicit a more complete answer because many activities may not be regarded as work. It is recommended that the questionnaire list specific economic activities for respondents to choose from, rather than asking for a “yes or no” answer. Similarly, recent ILO and World Bank research shows that recovery questions to capture work and employment information for individuals who may not identify their activities as such can significantly improve the accuracy of employment and work classifications for both men and women, but especially the latter (see box 7). Recovery questions are very context-specific and as such need to be carefully worded. When crafted thoughtfully, they are an important tool in correctly classifying people’s labor contributions. Box 7 Underreporting Women’s Economic Activity: Evidence on Improving Accuracy of Responses from Sri Lanka A pilot study conducted by the World Bank and International Labour Organization in partnership with the Sri Lanka national statistical office generated important lessons for measuring women’s and men’s work through household surveys operationalizing the standards from the 19th International Conference of Labour Statisticians. Among other important findings, the study highlighted the importance of recovery questions. It found that some household members, especially contributing family workers, reported their work activity only if they were asked the relevant recovery question. The inclusion of questions for contributing family workers in agriculture and of recovery questions on small or casual jobs can lead to a more accurate and consistent capture of women’s participation in all forms of work and is therefore recommended best practice. Source: ILO and World Bank 2021. (B7) It is important to be mindful of double standards for men and women in indicator definition and calculation. Although employment patterns for females diverge from those of their male peers in many areas, in some instances females How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 23 engage in the exact same activity and yet are classified differently. Women working in the family business, farm, or shop are often counted automatically as contributing family workers even when they are working on an equal footing with their husbands. In such instances, international classifications advise that they should be recognized as employers (where there is at least one employee) or as own-account workers (where no employee exists) like their male counterparts. Including specific questions in the module can help avoid applying biased perceptions when defining employment categories for males and females. (B8) Agricultural production has a particularly large employment gap by gender. Underreporting in agriculture is common because women are less likely than men to define their activities as work and because women tend to work longer hours than men (Francavilla 2019). LSMS+ guidelines recommend that, to properly capture women’s involvement in agriculture, surveys should include a question on intended destination of agricultural production (market or own use). (B9) In addition to identifying all sources of income, investigate who gets to control the earnings. Labor income is an important determinant of women’s economic empowerment. However, a woman’s ability to earn income does not ensure that she has control over how it is spent, saved, or invested. The LSMS recommendation for labor/employment is to ask for identification of (a) respondent and (b) household member(s) (a minimum of two individuals) controlling the use of earnings—asked separately for each job listed for the individual during the reference period of interest. This recommendation extends to other sources of income such as agricultural income from crop sales, livestock production, or fishery and also should be applied to nonlabor income such as social transfers and remittances. (B10) Measuring labor underutilization can help make more visible women’s unmet need for employment and constraints to working more hours. Underutilized labor, defined as the difference between the desire of individuals to work and their ability to do so, is included as a new category in the ICLS-19 standards. This category is particularly relevant when exploring gender differences in access to more and better jobs. For example, in ILO Labor Force Survey pilot studies, among household members who reported an interest in working or working more, women cited family responsibilities more than five times as often as men as a reason for not searching for employment and almost four times as often as a reason for not being available to accept employment (Benes and Walsh 2018). Similarly, among those employed, women are significantly more likely than men to mention family responsibilities as a reason for working fewer hours than a typical workweek. The LSMS+ program therefore encouraged countries to include How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 24 questions on unemployment and job searching over the last 30 days and to ask for information on whether individuals are searching for more or different work. (B11) Self-reporting is particularly important for the labor module. Per the WWEP and LSMS guidelines, every effort should be made to collect information directly from each household member regarding their own labor activities. Only when this is not possible may another adult household member be asked to provide the information on behalf of the family member in question. For complete and accurate information on work and employment for all members of the household, it is important not only to collect information at the individual level but also to give all age-eligible13 members of the household the opportunity to report their own labor activities. Proxy respondents may not provide accurate information on the income and labor inputs of others, including on that of their spouse. (C) Entrepreneurship Recommendations at a glance Table 6 summarizes the key recommendations, highlights which data needs will be addressed by taking them into consideration, and cross-links to the detailed explanations in the subsection titled “Recommendations explained.” Table 6 Summary of Key Recommendations on Entrepreneurship Corresponding # Recommendation paragraph 1 Adopt a clear definition of entrepreneurship. (C1) Include screening questions to identify potential entrepreneurship-related activities, and 2 (C2) include data checks with labor module. 3 Reference period for entrepreneurship questions should be the last 12 months. (C2) 4 Separate business ownership from management. (C3) 5 Distinguish between ownership of business and ownership of assets. (C4) 6 Collect information from the associated manager for each enterprise. (C4) Opt for self-reporting instead of proxy reporting for all questions regarding 7 see section (D) entrepreneurship. Source: World Bank. 13 Labor data must be collected for all working-age members of the household. The working age is to be defined at the survey or country level (Durazo et al. 2021). How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 25 Sample module • A recommended sample is the EDGE sample survey instrument (see Annex 1 of EDGE Entrepreneurship Report). Related indicators Table 7 provides the UNSD minimum indicators related to entrepreneurship, including the indicator description. Table 7 Indicators Related to Entrepreneurship UNSD SDG minimum Indicator description indicator indicator 1.6 Proportion of employed who are employer, by sex See note Percentage of adult population who are entrepreneurs, by sex Source: World Bank. Note: This entrepreneurship indicator was previously a UNSD minimum indicator but was removed in the most recent round. Please note that the recommendations outlined for this indicator are in line with the latest international guidance, however there is still ongoing methodological work giv- en the complexity of this definition. SDG = Sustainable Development Goal; UNSD = United Nations Statistical Division. Recommendations explained (C1) A clear methodological definition can help to generate internationally comparable information on entrepreneurship. The EDGE project defined an entrepreneur as a person who has direct control over the activities of an enterprise they own alone or with other individuals. Even though this definition is still broad, it allowed for the concept to be operationalized across the six participating pilot countries. (C2) Selecting the right survey type is important to fully capture women’s entrepreneurship.14 Household surveys are a great source of information on 14 For developing gender indicators on entrepreneurship, two data sources are particularly relevant: (a) data from household surveys, which mostly collect information on micro and small businesses; and (b) firm-level data from registers, enterprise surveys, and economic censuses with information on business owners. Agricultural surveys may tilt toward How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 26 female entrepreneurship. EDGE focused only on household surveys, and its pilot studies revealed that this source captures mainly women and men entrepreneurs operating micro or small home-based enterprises, which primarily operate in trade and service industries. Therefore, these entrepreneurs may not be adequately covered in an enterprise survey or business register. Most women entrepreneurs are concentrated in micro or small enterprises, which are often household-based. Generally, it is recommended to work through existing national data collection instruments to ensure sustainability of data collection. Adding a few questions or appending a module on entrepreneurship to an existing labor force or other household survey can help keep costs manageable. To capture the irregular nature of certain entrepreneurial activities, the survey reference period for entrepreneurship questions is often the last 12 months. (C3) Asset ownership in the context of nonagricultural enterprises merits special attention in questionnaire design. Doss, Grown, and Deere (2008) support taking account of both the ownership and management of the business separately, by gender, without assuming that ownership implies management or vice versa. It is also important not to infer from survey questions that business assets are owned by the owner of the business. It is advised that individual and joint ownership of both the business and its assets be probed because owners of the business may not necessarily own some of the business’s assets. Furthermore, computer-assisted personal interviewing applications should have embedded data quality controls to flag any internal inconsistencies between this module and the individual-level module on labor regarding household members who are enterprise laborers. (C4) Selecting the right respondent is key to assess gendered aspects of entrepreneurship. The respondent for each enterprise should be the associated manager (or one of the available managers in the case of jointly managed enterprises). Self-reporting in gathering entrepreneurship data is key because proxy reporters may have difficulty recalling other entrepreneurs in their households and may not have full information on entrepreneurial motivations and aspirations, entrepreneurial resource and constraints, and entrepreneurial performance of other household members on whom they report (UN 2018). The LSMS guidance on household enterprises is to identify (a) respondent, (b) reported owners (a minimum of two individuals), (c) managers (a minimum of two commercial/larger plots; thus, they disproportionately leave out women in agriculture who farm smaller plots. Formal firm surveys may miss the full picture of the gendered firm distribution as women-owned firms are more likely to be informal and smaller. Similarly, labor force surveys are often predominantly focused on formal labor arrangements. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 27 individuals), and (d) household laborers (a minimum of four individuals)—along with information on labor inputs for each laborer. (D) Spotlight on respondent selection and importance of self-reporting (D1) Collecting individual-level data is a prerequisite for meaningful sex- disaggregation. Many missed opportunities to fill gender data gaps arise because data are not collected at the individual level. Data on assets and income activities are often collected only at the household level, even when these assets and activities are individually attributable. The current standard of practice risks omitting important intrahousehold disparities. For example, some countries might ask whether the household owns a mobile phone, without collecting information on who in the household has ownership. As a result, accurate sex-disaggregated data about mobile phone ownership cannot be produced. Even when household survey data are collected at the individual level, the information is often collected from a single respondent, typically the self-identified “most knowledgeable” household member (see box 8). Box 8 Analytical Shortcomings of Household Headship Concept Many surveys collect data by interviewing proxies such as “household heads,” even though the reliability and significance of estimates of households differentiated by “headship” continue to be the subject of debate. Several shortcomings have been pointed out: (a) ambiguities in how headship is defined and understood by both enumerators and respondents; (b) assumptions in many countries that the household head is male, with data likely subject to gender-biased perceptions of enumerators and respondents; (c) limited usefulness of a category such as “female-headed household” because it fails to reflect diverse economic realities of these households; and (d) outdatedness of the headship concept, long discontinued in nationally representative surveys in countries such as Canada, the United Kingdom, and the United States. A better approach is to consider household composition in the analysis of survey data and to look at the demographic and earnings profiles of household members. Source: Munoz Boudet et al. 2018; UN Women 2019. (D2) Ideally, each individual aged 18 years and older should self-report personal information on asset ownership and control, work and employment, and entrepreneurship. Studies have shown that a proxy respondent may provide a different response than the actual individual, which can lead to bias in the analysis (Bardasi et al. 2011; Kilic and Moylan 2016). For instance, compared with the international best practice of privately interviewing adults about their personal How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 28 asset ownership and rights, the business-as-usual approach of interviewing the most knowledgeable household member(s) on adult household members’ ownership of and rights to assets has been shown to generate (a) higher rates of exclusive reported and economic ownership of agricultural land among men, and (b) lower rates of joint reported and economic ownership among women (Kilic et al. 2020). The sex of the proxy, as well as their relationship to the target respondent, could also introduce additional bias. Proxy reporting aggravates issues of inaccuracy—particularly in low-income settings, which are usually associated with low rates of documented land ownership; widespread seasonal and informal employment, which are complex to measure; and social norms or regulatory barriers that can influence how proxy respondents provide information on other household members’ economic roles and ownership of and rights to assets (Hasanbasri et al. 2021b). Box 9 summarizes the experience of LSMS+ countries moving from proxy to direct responses and box 10 discusses the cost implications of conducting individual-level interviews. Box 9 Moving from Proxy to Direct Responses: Experience of LSMS+ Countries Response rates among eligible adults were high across the four completed Living Standards Measurement Study – Plus (LSMS+) countries. In Malawi, 82 percent of all eligible adults were successfully interviewed. In Cambodia, Ethiopia, and Tanzania, Ethiopia and Cambodia, the shares of eligible adults who were interviewed were 92, 96, and 80 percent, respectively. Across specific asset modules, for households with nondwelling land, 95 percent of respondents in Ethiopia were interviewed; these shares were 83 percent in Malawi and 79 percent in Tanzania. The interview rates were somewhat higher for dwelling land in Ethiopia and Tanzania. In Ethiopia, about 96 percent of eligible respondents were interviewed for the mobile phone module, whereas these shares were 82 percent in Malawi and 80 percent in Tanzania. Overall, the urbanrural differences in interview rates were not very large, but gender differences were substantial—particularly in Malawi and Tanzania, where eligible female respondents were much more likely than eligible men to be available for interviews. Source: Hasanbasri et al. 2021b. (D3) Despite its potential drawbacks, proxy reporting can be a valid option when budget and time constraints do not allow for individual interviews with all household members. When individual interviews are not feasible, the use of proxy respondents is considered preferable to missing data. Irrespective of self- or proxy reporting, it is essential to ensure proper respondent identification. As such, it is important to collect the roster ID of the actual respondent providing data for each person, to capture whether individual-level information is being reported by proxies or by the individuals themselves. A core problem in past surveys has How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 29 been that the inability to attribute an answer to a certain individual obfuscates any meaningful gender differences in responses. The questionnaire should provide specific instructions on who is supposed to answer the questions and include the respondent ID. At the beginning of each new module, the questionnaire should clarify (a) whether the respondent is reporting for him/herself and (b) who is responding on behalf of (name) to make sure each answer can be coded correctly. Box 10 The Additional Cost of Implementing Individual Interviews Collecting individual data is more time-intensive than using a proxy respondent approach, but what is the exact differential? In Malawi, two surveys identical in scope were conducted at the same time, with the only difference being that one (IHPS) used the individual interview approach whereas the other (IHS4) was carried out with the business-asusual approach of interviewing the “most knowledgeable” household member(s) regarding asset ownership. A side-by-side comparison of the approaches reveals that, on average, field teams spent approximately 3.4 days in an IHS4 enumeration area, with one enumerator visiting each household. The same field teams spent an average of 4.5 days in an IHPS enumeration area, which involved interviewing, simultaneously whenever possible, up to four adults in each household. Moving from one to up to four respondents hence increased the interview time by 1.1 days, or 32 percent above the standard interview approach. Similarly, paradata from the Tanzania National Panel Survey 2019/20 shows that enumerators spend, on average, 270 minutes at a household, with 79 of those minutes—approximately 30 percent of their time—devoted to interviews with adult individuals in that household. At the national level, an interview with an adult household member takes about 30 minutes, on average. Source: Hasanbasri et al. 2021b. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 30 PART II Data Collection Process for Better Gender Statistics Recommendations at a glance Table 8 summarizes the key recommendations to better collect gender data and cross-links to the detailed explanations in the subsection titled “Recommendations explained”. Table 8 Summary of Key Recommendations on Gender-Sensitive Data Collection Corresponding # Recommendation paragraph 1 Conduct a survey-specific assessment. (E1) (E2) Raise awareness of gender concepts among interviewers—in 2 (E3) (E4) training, instructions, written guidance, and supervision. 3 Ensure privacy in the interview process. (E5) Aim for a gender match between interviewer and interviewee, if 4 (E6) appropriate in country context. 5 Conduct a gender-sensitive pilot before taking the survey to the field. (E7) Source: World Bank. Sample field protocols • A sample enumerator manual for asset modules can be found in Annex II of the LSMS+ Program overview document. • A sample Interviewer Training Manual is available in Annex 2 of the LSMS Guidebook on Employment and Own-Use Production in Household Surveys. Recommendations explained (E1) Integrating a gender perspective into surveys goes well beyond questionnaire design. Gender considerations also need to be embedded into survey implementation.15 This section highlights relevant aspects to consider when 15 This section focuses on field implementation. Two other critical components of the data collection process to ensure balanced gender data are sample design and data coding, which are outside the scope of this guidance note. The sample design of the survey should also ensure that the figures by gender are reliable and that they can be disaggregated by other characteristics as needed for statistically meaningful gender analysis (for example, region, education, age groups, marital status, and so on). Additionally, training and other protocols may need to be put in place to ensure that biases are not present at the postcoding, editing, and imputation phases. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 32 taking a survey to the field that is supposed to generate meaningful gender data. Although most recommendations presented here are considered best practice for data collection efforts more broadly, some considerations are particularly relevant to gender statistics. Beyond adhering to best practice in data collection, generating reliable gender statistics depends on thoughtful consideration of gender imbalances in survey implementation. Generally, good data collection principles can be considered a necessary, but not sufficient, condition for reliable gender statistics. Common pitfalls for collection of gender data specifically include insufficient awareness of gender concepts among interviewers, lack of privacy, or gender mismatch between interviewers and interviewees. Disregarding these possible problems can introduce bias to the data collected. (E2) Cost and accuracy trade-offs need to be carefully balanced. Conducting a survey-specific assessment can help alleviate these concerns. Based on the findings from such assessment, the setup for fieldwork implementation and scheduling of interviews may be slightly tweaked to allow interviewers to spend their time more efficiently and to better target individuals and decrease the reliance on proxy respondents. Given the context-specific nature of these trade- offs, recommendations are not generalizable but LSMS+ work has shown that interviewing individuals is feasible within different types of fieldwork setups— mobile teams moving to different enumeration areas as well as in contexts with residential enumerators. (E3) Successful field implementation relies on detailed instruction manuals for both supervisors and interviewers. The language of the manual should be straightforward and easy to understand and should be a useful guide during training, as well as a reference document for use during field operations. The fieldwork manual needs to cover all aspects of the survey. Its language should be free of gender-based biases or other stereotypes, including attention to the examples given to minimize reinforcing gender stereotypes. Moreover, additional information required to aid field supervisors in their supervisory responsibilities include the following: procedures for organizing and controlling the flow of materials to and from the field, fieldwork monitoring approaches, advantages of adherence to timetables, procedures for quality control and on-field review of completed questionnaires, and steps to take when serious errors are discovered. (E4) Interviewer manuals and enumerator training need to pay special attention to new concepts. Some of the recommended updates to questionnaires on asset ownership, work and employment, and entrepreneurship need to be explained in detail in the interviewer manuals to ensure that enumerators can How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 33 elicit correct responses and respond thoughtfully to any clarifying questions that may arise in the interview process. For example, responses to questions on ownership and rights to assets are not necessarily consistent, and interviewers should feel confident to deal with these ambiguities as they record answers. Similarly, enumerators can make the case for individual interviews instead of using a proxy respondent more convincingly if they understand why this is the preferred approach. Training should be organized to better explain gender-related concepts to mitigate the challenges associated with collecting questions related to assets or employment from a gender angle. For instance, it may be helpful to draw on both female and male trainers, have women and men trained to interview persons of the same and of the opposite sex, and ensure that materials and vignettes include examples illustrating men and women to minimize reinforcing stereotypes. (E5) Gathering information on assets, work and employment, and entrepreneurship requires a mindful data collection approach. Given the sensitive nature of these topics, disclosing personal information will require an environment of trust and confidentiality to ensure that accurate and truthful answers can be recorded. It is possible that household members, particularly women, will have hidden assets that other people in the household know nothing about. A key recommendation is therefore to conduct private interviews with all adult household members on their personal ownership of and rights to assets. The LSMS+ experience furthermore suggests that interviews be conducted, if possible, simultaneously across different household members. The international guidelines for the collection of required data for SDG indicator 5.a.1 also recommend interviewing in private either all adult household members or a randomly selected household member. It may be necessary for the field staff to carefully explain the rationale for individual and private interviews to survey respondents. (E6) Ensure as much as possible a gender match between the interviewers and the respondents. In selecting interviewers for the survey, it is crucial to determine to whom interviewees will be more comfortable disclosing sensitive or confidential information (as per the survey context). LSMS+ work showed that respondents may be more comfortable disclosing such personal information— hidden assets, financial accounts, and so on—to enumerators of the same sex. The EDGE guidelines acknowledge, however, that in some country contexts it may be less safe for female enumerators to be part of the field staff and travel alone. Additionally, survey practitioners should weigh the costs and benefits of gender matchup based on the country context and also consider the fieldwork setup and the feasibility of assigning at least one male and one female enumerator How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 34 to a household given the logistics involved and the potential time and cost implications of doing so. To increase the likelihood of securing one-on-one time with selected respondents, it may be worthwhile to identify the preferred days/ times when scheduling visits, interviewing at the residence or in another place, and interviewing different members at the same time. In addition, recording who is also present during the interview could help assess the quality of the interview in situations where privacy is not possible. (E7) Piloting is particularly important to the collection of gender statistics given the conceptual complexity associated with data on assets, work and employment, and entrepreneurship. Survey pilots can help identify any issues with regard to respondent selection, interview flow, and topic coverage. Pilots act as an early warning system, because the experience from the field and the analysis of pilot data can raise any red flags that may impede successful implementation of the survey and offer the survey team the opportunity to address any content- specific, technical, operational, or financial challenges. The scope of the pilot should include determining what works in terms of the gender composition of the field team and in eliciting responses to matters that might be new or sensitive to respondents. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 35 References and Additional Resources Guidelines and Factbooks Beegle, K., N. Buehren, D. Halim, and G. Roberts. 2021. Tip Sheet for Engendering Data in COVID-19 Response Household Surveys. World Bank Operational Note, February 5, World Bank, Washington, DC. Durazo, J., V. Costa, A. Palacios-Lopez, and I. Gaddis. 2021. LSMS Guidebook: Employment and Own-Use Production in Household Surveys: A Practical Guide for Measuring Labor. Washington DC: World Bank. FAO (Food and Agriculture Organization of the United Nations), World Bank, and UN- Habitat (United Nations Human Settlement Programme). 2019. Measuring Individuals’ Rights to Land: An Integrated Approach to Data Collection for SDG Indicators 1.4.2 and 5.a.1. Washington, DC: World Bank. Hasanbasri, A., T. Kilic, G. Koolwal, and H. Moylan. 2021b. Overview and recommendations for improving individualdisaggregated data on asset ownership and labor outcomes. World Bank, Washington DC. Oseni, G., A. Palacios-Lopez, H. K. Mugera, and J. Durazo. 2021. Capturing What Matters: Essential Guidelines for Designing Household Surveys. Washington DC: World Bank. UN (United Nations). 2018. Measuring Entrepreneurship from a Gender Perspective: Lessons Learned from the EDGE Project. Technical Report. United Nations Department of Economic and Social Affairs, Statistics Division, New York. UN (United Nations). 2019. Guidelines for Producing Statistics on Asset Ownership from a Gender Perspective. Studies in Methods, Series F No. 119. ST/ESA/ STAT/SER.F/119, United Nations Department of Economic and Social Affairs, Statistical Division, New York. Research and Policy Papers Bardasi, E., Beegle, K., Dillon, A. and Serneels, P., 2011. Do labor statistics depend on how and to whom the questions are asked? Results from a survey experiment in Tanzania. The World Bank Economic Review, 25(3), pp.418- 447. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 36 Benes, E., and K. Walsh. 2018. Measuring Employment in Labour Force Surveys: Main Findings from the ILO LFS Pilot Studies. Statistical Methodology Series 4, International Labour Office Department of Statistics, Geneva. Desiere, S., and V. Costa. 2019. Employment Data in Household Surveys Taking Stock, Looking Ahead. Policy Research Working Paper 8882, World Bank, Washington DC. Discenza, A., I. Gaddis, A. Palacios-Lopez,and K. Walsh. 2021. Measuring Women and Men’s Work Main Findings from a Joint ILO and World Bank Study in Sri Lanka. Washington, DC: World Bank. Doss, C., C. Grown, and C. Deere. 2008. Gender and Asset Ownership: A Guide to Collecting Individual-Level Data. Policy Research Working Paper 4704, World Bank, Washington DC. Gaddis, I., G. Oseni, A. Palacios-Lopez, and J. Pieters. 2020. Who Is Employed? Evidence from Sub-Saharan Africa on Redefining Employment. Policy Research Working Paper 9370, World Bank, Washington DC. Hasanbasri, A., T. Kilic, G. Koolwal, and H. Moylan. 2021a. LSMS+ Program in Sub- Saharan Africa: Findings on Individual-Level Data Collection on Labor, and Asset Ownership and Rights. World Bank, Washington DC. Kilic, T., and H. Moylan. 2016. Methodological Experiment on Measuring Asset Ownership from a Gender Perspective. World Bank, Washington, DC. Kilic, T., G. Van den Broeck, G. Koolwal, and H. Moylan. 2020. Are You Being Asked? Impacts of Respondent Selection on Measuring Employment. Policy Research Working Paper 9152, World Bank, Washington DC. Muller, M., and L. Sousa. 2020. She Helps Me All the Time—Underestimating Women’s Economic Engagement in Rural Honduras. Policy Research Working Paper 9217, World Bank, Washington DC. Munoz Boudet, A., P. Buitrago, B. Leroy de la Briere, D. Newhouse, E. Rubiano, K. Scott, and P. Suarez-Becerra. 2018. Gender Differences in Poverty and Household Composition through the Life-Cycle—A Global Perspective. Policy Research Working Paper 8360, World Bank, Washington DC. Other Resources Buvinic, Mayra, Rebecca Furst-Nichols, and Gayatri Koolwal. 2014. Mapping Gender Data Gaps. https://data2x.org/wp-content/uploads/2019/05/Data2X_ MappingGenderDataGaps_FullReport.pdf. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 37 Francavilla, F. 2019. The Landscape of Gender Data. Mapping Players and Initiatives by Themes. Report commissioned by the William and Flora Hewlett Foundation with cooperation from the Bill & Melinda Gates Foundation. Grantham, K. 2020. Mapping Gender Data Gaps: An SDG Era Update. Data2X, Washington, DC. ICLS (International Conference of Labour Statisticians). 2013. Resolution Concerning Statistics of Work, Employment, and Labour Underutilization. Nineteenth International Conference of Labour Statisticians, October. UN Women (United Nations Entity for Gender Equality and the Empowerment of Women). 2018. Turning Promises into Action: Gender Equality in the 2030 Agenda for Sustainable Development. UN Women, New York. UN Women (United Nations Entity for Gender Equality and the Empowerment of Women). 2019. Progress of the World’s Women 2019–2020: Families in a Changing World. UN Women, New York. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 38 Appendix A Why Focus on Assets, Work and Employment, and Entrepreneurship? This guidance note focuses on how to improve indicators on women’s economic empowerment. Below we highlight the benefits of considering the recommendations in this guidance note, by economic domain. Collecting data on assets, work and employment, and entrepreneurship is key to measuring gender inequalities, monitoring progress toward gender-related sustainable development goals and for conducting research for the design and evaluation of policies and interventions that aim to eliminate gender inequalities. Why focus on assets? Research has shown that ownership and control over assets—such as land and financial accounts—can have many benefits. It can ease access to credit and help boost productivity and income while also improving bargaining power and decision- making within households. Therefore, accurate information on intrahousehold asset ownership and control can play an important role in policy making, including the design of land reforms, on economic empowerment. Additionally, monitoring international goals such as the Sustainable Development Goals on land ownership and rights relies fundamentally on the quality of underlying data, which, in the context of surveys, are directly affected by how respondents are selected. These issues are particularly relevant for agriculture in developing-country contexts where a clearer understanding of individual land ownership and rights can help raise productivity and secure property rights among farmers. Access to formal financial services such as savings, insurance, payments, credit, and remittances is essential to the ability of people to manage their lives, build their futures, and grow their businesses. Why focus on work and employment? On the macroeconomic side, women’s inclusion in the labor market is a key driver of growth and economic development. Considering microeconomic effects, income generated through market-based work can improve women’s bargaining power and decision-making within households. Going beyond market-based work, it is important to capture all types of work to account for gender differences in care work How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 39 and household responsibilities that may have a limiting effect on an individual’s time to be dedicated to employment activities. Accurate data on employment are needed to design labor policies and training programs that account for the realities of local labor markets and the gendered characteristics of the workforce. Methodologically sound employment data are needed to monitor international goals such as the SDGs and national development plans. Why focus on entrepreneurship? Entrepreneurship is a subcategory of employment and is similarly relevant to boosting economic inclusion and growth. Research shows that women’s entrepreneurship supports the diversification of business, stimulating innovation and diversification in management, in production and in marketing practices, as well as in products and services. Male and female entrepreneurs differ significantly. For example, female entrepreneurs tend to provide different solutions to management, organizational, and business problems than men do. To create an environment conducive of entrepreneurship it is therefore important that policy makers understand business barriers and opportunities as perceived by both women and men. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 40 Appendix B About the Strengthening Gender Statistics Project Why gender data? This year marks the 10-year anniversary of the World Development Report 2012 on gender and development. Gender data is essential as we revisit remaining obstacles to the promise of productive and self-determined lives for women and girls globally: to measure and report on progress achieved and to design meaningful policies moving forward. Gender data go well beyond sex-disaggregation. Contrary to popular belief, gender data is not only on women or women’s issues but encompass the full range of the life cycle intersecting with all sectors from getting a proper ID at birth to accessing quality education to getting a decent job to using public transport and feeling safe at home and in public spaces. Gender data also pays specific attention to women and men’s different realities and how those interact with other forms of individual characteristics. The foundation for designing effective policies that benefit women and men, girls and boys, gender data is often incomplete, methodologically flawed, or completely lacking. Consequences of inadequate gender data are dire: social and economic policies are less impactful, opportunities to improve project design are missed, and a blind eye is turned to stubborn inequalities in gendered social and economic realities. How to tackle gender data gaps? To address this issue in our partner countries and generate lessons for a global engagement on gender statistics, the World Bank, with support from the Bill and Melinda Gates Foundation, is implementing the Strengthening Gender Statistics (SGS) project. SGS draws on expertise from the Gender Group, the Poverty and Equity Global Practice, and the Development Data Group’s Living Standards Measurement Study (LSMS) program, to provide technical assistance to select IDA countries on improving gender data production and dissemination of gender statistics, with a focus on the economic domain. In doing so, the project is leveraging our long-standing engagements with national statistical offices, and our concerted efforts in the recent past, through the How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 41 LSMS+ Initiative and the Women’s Work and Employment Partnership, to improve the availability and quality of individual-disaggregated survey data on economic outcomes. Through a demand-driven process, SGS is currently supporting twelve IDA countries with ongoing or planned statistical operations to improve the scope and quality of their gender data and statistics. Homing in on women and girls’ economic opportunities and outcomes To make the most of available resources the SGS project targets gender data gaps specifically in the areas of employment, entrepreneurship, and asset ownership and control. Women’s economic empowerment is instrumental to ending extreme poverty and promoting shared prosperity and expanding survey data collection on these topics is a priority for the World Bank. Additionally, this focus is complementary to other partners’ efforts in improving gender data such as the UN Women Count project and the multi-donor 50x2030 initiative. Narrowing existing gaps within economic statistics has also been identified as a policy priority among many partner countries. Leveraging existing engagements with National Statistical Offices globally The SGS project is collaborating with National Statistical Offices from a diverse set of countries and regions ranging from Bangladesh and Lao PDR in Asia to several Western African countries including Benin, Burkina Faso, Cameroon, Ghana, Mali, and the Republic of Congo as well as Tanzania, Somalia and Madagascar in Eastern Africa, and Djibouti representing the MENA region. While each country is facing unique data challenges, they are united in their desire to improve the availability and quality of their gender data. Tackling the various sources of gender data gaps – one at a time The SGS project is designed to mainstream the adoption of international best practices in the production of economic gender data. There are several entry points to effectively address these gaps. Our approach recognizes that data impediments can emerge at different stages of the data life cycle. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 42 Data gaps arise due to various reasons… Data no collected Collecte but methodologically weak Metodologically strong ... but not processed & analyzed Processed & analyzed ... but not processed & analyzed • Lack of surveys • Inappropiate or • No calculation of • Lack of timely and Problems overall incomplete relevant indicators gender statistical • Lack of special topic questionnaire design • Limited tabulation and factbooks surveys • Flawed data collection • No disaggregation by • Insu cient approaches sex and other • No communication and administrative data characteristics dissemination strategy Proposed solution Outside of scope of SGS SGS Component 1 SGS Component 2 SGS Component 3 Data production Data analysis Data dissemination Strengthening Gender Reduced gender data gaps in the economic domain Statistics … and require a holistic approach to be narrowed and closed Broadly, problems can be grouped into three categories: 1. Data production: Significant investments have been made to address meth- odological concerns and practical challenges related to economic gender data. New evidence has emerged from projects such as LSMS+ and the Women’s Work and Employment Partnership on the implications of re- spondent selection, updated guidance on employment indicators, and the urgency of sex-disaggregation. These findings demonstrate the need and value of adopting a gender lens in the survey design process. However, advances in measuring women’s economic status are often not translat- ed into survey operations. The SGS project provides targeted fixes to survey design and implementation protocols within each country’s specific survey context. 2. Data analysis: Even where data are being disaggregated by sex, in-depth analysis of that data is not always undertaken. Too often, gender data use is hindered by limited tabulation and dissemination of existing data. As a result, the uptake and use of gender data are challenged in the transla- tion and communication of statistics to less technical audiences and lever- aged in policy dialogue. The SGS project offers country-tailored data analysis training on generating basic gender statistics calculated using internationally recognized methodology. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 43 3. Data dissemination: Efforts to collect high-quality gender data are only helpful if the data is used. Disseminating collected data is not going the extra mile, but a fundamental step in the data production cycle to ensure uptake and use of data for policymaking. The SGS project supports the pro- duction of gender abstracts and factbooks and works with partner countries to promote the dissemination of data sets and associated reports. Our vision: Starting a new chapter for economic gender data Specific activities within these three categories are defined based on the needs of the partner country and depending on the country context. Results are expected by the end of 2023 and implementation will be shared throughout the duration of the SGS project. Equipped with new and better gender data, partner countries will be in a better position to shape the economic lives and livelihoods of a new generation. Producing, leveraging, and sharing gender data is critical to AccelerateEquality in the next decade. For access to more than 900 gender-related indicators, please visit the World Bank Gender Data Portal, the World Bank Group’s comprehensive source for the latest sex-disaggregated data and gender statistics. How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 44 Appendix C Relevant Indicators in the Economic Domain To identify relevant indicators for the economic domain this Guidance draws on two sources, the UN Minimum Set of Gender Indicators and data2X Gender- Relevant SDG Economic Opportunity Indicators. The table below shows the gender indicators that the SGS project (see details in appendix A) is designed to address. Each indicator is mapped to one of the three focus areas of the SGS project: assets, employment, and entrepreneurship. This list can serve as a starting point for NSOs wishing to identify key gender data gaps that the country is experiencing, and other indicators can be added as needed. To assess the availability of indicators, the SGS project will use a scale that classifies indicators based on three characteristics: (a) whether the indicator is available in the public domain, (b) whether it is calculated according to the internationally recommended definition, and (c) whether the data to calculate the indicator are collected on a regular basis (defined as at least every three years). The assessment scale has six levels, allowing for a detailed investigation of the current indicator status: 1. Available—Meets all aspects of indicator definition and regularly collected 2. Available—Meets all aspects of indicator definition but not regularly collected 3. Partially available—Some aspects of indicator definition missing but collect- ed regularly 4. Partially available—Some aspects of indicator definition missing and not col- lected regularly 5. Unavailable—Data collected but indicator not calculated or published 6. Unavailable—Data not collected How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 45 Table C.1 SGS Project Focus Indicators UNSD Custodian SDG # Indicator Description Tier minimum agencies indicator indicator Assets Proportion of adults (15 years and older) with an account at 1 a bank or other financial institution or with a mobile-money 1 WB 8.10.2 I.11 service provider*+ Proportion of individuals who own a mobile telephone, by 2 2 ITU 5.b.1 I.18 sex*+ (a) Proportion of total agricultural population with ownership or secure rights over agricultural land, by sex; and (b) share 3 2 FAO 5.a.1 I.12 of women among owners or rights-bearers of agricultural land, by type of tenure*+ World Bank Proportion of total adult population with secure tenure rights 4a 2 and UN 1.4.2 to land, (a) with legally recognized documentation+ Habitat Proportion of total adult population with secure tenure rights World Bank 4b to land (b) who perceive their rights to land as secure, by sex 2 and UN 1.4.2 and type of tenure+ Habitat Employment Average number of hours spent on unpaid domestic and care UNSD/UN 5 work, by sex, age and location Note: separate domestic work 2 5.4.1 I.1 Women and care work, if possible)* + Average number of hours spent on total work (total work 6 2 UNSD I.2 burden), by sex* Labor force participation rate for persons aged 15-24 and 7 1 ILO I.3 15+, by sex* Proportion of employed who are own-account workers, by 8 1 ILO I.4 sex* Proportion of employed who are contributing family workers, 9 1 ILO I.5 by sex* Proportion of youth (aged 15-24 years) not in education, 10 1 ILO 8.6.1 I.7 employment, or training, by sex and age*+ Percentage distribution of employed population by sector, 11 each sex (sectors here refer to Agriculture; Industry; Ser- 1 ILO I.8 vices)* Proportion of informal employment in nonagriculture em- 12 2 ILO 8.3.1 I.9 ployment, by sex*+ How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 46 UNSD Custodian SDG # Indicator Description Tier minimum agencies indicator indicator Unemployment rate, by sex, age, and persons with disabili- 13 1 ILO 8.5.2 I.10 ties*+ Average hourly earnings of female and male employees, by occupation, age, and persons with disabilities+ 14 2 ILO 8.5.1 I.13 Note: the corresponding UNSD indicator is Gender gap in wages, by occupation, age and persons with disabilities* 15 Proportion of employed working part-time, by sex* 2 ILO I.14 Prime-age employment-to-population ratio by sex, household 16 1 ILO I.15 type and presence of children* Average income of small-scale food producers, by sex and 17 2 FAO 2.3.2 indigenous status+ 18 Proportion of women in managerial positions+ 1 ILO 5.5.2 Proportion and number of children aged 5–17 years engaged 19 2 ILO/UNICEF 8.7.1 in child labor, by sex and age+ Frequency rates of nonfatal occupational injuries, by sex and 20a 2 ILO 8.8.1 migrant status+ Frequency rates of fatal occupational injuries, by sex and 20b 2 ILO 8.8.1 migrant status+ Entrepreneurship 21 Proportion of employed who are employer, by sex* 1 ILO I.6 Percentage of adult population who are entrepreneurs, by sex  22 3 Note: this indicator was previously on the UNSD minimum indicator list but was removed in the most recent round. *UN Minimum Indicators List Indicator +Data2X Gender-Relevant SDG Economic Opportunity Indicator Sources: IAEG-GS Advisory Group on Gender Indicators and https://data2x.org/wp-content/up- loads/2020/03/MappingGenderDataGaps_Economic.pdf How to Improve Indicators on Guidance Note on Gender-Sensitive Women’s Economic Empowerment Survey Design and Implementation 47 Strengthening Gender Statistics