DIREC TIONS IN DE VELOPMENT Countries and Regions Getting to Work Unlocking Women’s Potential in Sri Lanka’s Labor Force Jennifer L. Solotaroff, George Joseph, and Anne T. Kuriakose Overview DIREC TIONS IN DE VELOPMENT Countries and Regions Overview Getting to Work Unlocking Women’s Potential in Sri Lanka’s Labor Force Jennifer L. Solotaroff, George Joseph, and Anne T. Kuriakose This booklet contains an Overview of Getting to Work: Unlocking Women’s Potential in Sri Lanka’s Labor Force doi: 10.1596/978-1-4648-1067-1. A PDF of the final, full-length book, once published, will be available at https://openknowledge.worldbank.org/ and print copies can be ordered at http://Amazon​ .com. Please use the final version of the book for citation, reproduction, and adaptation purposes. © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank with external contributions. The findings, interpreta- tions, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Solotaroff, Jennifer L., George Joseph, and Anne T. Kuriakose. 2017. “Getting to Work: Unlocking Women’s Potential in Sri Lanka’s Labor Force.” Overview booklet. World Bank, Washington, DC. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content con- tained within the work. The World Bank therefore does not warrant that the use of any third-party- owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. Cover design: Naylor Design, Inc. Cover image: © Luxshmanan Nadaraja / World Bank. Further permission required for reuse. Getting to Work (Overview) 80°E 81°E IN DIA SRI LANKA 10°N 10°N SELECTED CITIES AND TOWNS PROVINCE CAPITALS it a Point Pedro NATIONAL CAPITAL r St RIVERS k Jaffna MAIN ROADS Pal Delft Elephant Pass RAILROADS Island PROVINCE BOUNDARIES Palk Bay Killinochchi INTERNATIONAL BOUNDARIES Iranamadu Tank Mullaittivu Ferr y Ad Talaimannar Manakulam am 's B ridge 9°N Mannar Island Mannar NORTHERN Pulmoddai Aruvi A Vavuniya ru SRI LANKA Trincomalee Gulf of Karaitivu NORTH CENTRAL Island Rambewa Mutur Mannar Anuradhapura Yan Oya Galkulama Kalpitiya Ka Kaud la O ulla Oya ya Bay of Bengal Puttalan Habarane 8°N 8°N NORTH Madura Oya WESTERN Maho Oya Batticaloa uru ed Kattankudi Mahaweli Ganga EASTERN D Chilaw Madura Oya Kurunegala Reservoir Kalmunai CENTRAL h a Oy a Ma Ampara Negombo Kandy Gal Oya Kegalla Victoria Falls Reservoir U VA Senanayake WESTERN Pidurutalagala (2,524 m) Samudra 7°N ga Badulla 7°N Kela Gan COLOMBO ni Sri Jayewardenepura Kotte Pottuvil Monaragala Moratuwa Ratnapura lu Ganga Wellawaya Ka SABARAGAMUWA Kirin Kalutara di Oya Kumana Laccadive Wala Kataragama INDIAN e w Ga Sea OCEAN ng a SOUTHERN Hambantota Galle Tangalla 0 20 40 60 Kilometers 6°N Matara IBRD 42743 | FEBRUARY 2017 Dondra Head 0 10 20 30 40 Miles 81°E 82°E Contents of the Overview Acknowledgments vii About the Authors ix Executive Summary xi Abbreviations xiii Overview 1 I. Introduction 1 II. Summary of Descriptive Data: Demographic Changes over Time 5 III. Hypothesis Testing: All Explanations for Women’s Poor Outcomes Are Still Supported 18 IV. Conclusion and Recommendations 38 Annex OA  Summary of Recommended Interventions (Cross Sectoral) 46 Annex OB  Sectoral Recommendations: Findings from Five Private Sector Industries 47 Notes 60 References 63 Figures O.1 Labor Force Participation, by Country 2 O.2 Female Labor Force Participation, by Select Country, Economic Status, and Region,1993–2016 2 O.3 Labor Force Participation, by Age and Gender, 2009 and 2015 5 O.4 Labor Force Participation, by Gender and Residential Sector, 2006–15 6 O.5 Labor Force Participation, by Gender and Ethnicity, 2009 and 2015 7 O.6 Labor Force Participation, by Education and Gender, 2009 and 2015 14 O.7 Unemployment, by Age and Gender, 2015 15 O.8 Unemployment, by Education Level and Gender, 2015 16 O.9 Reasons for Not Working Last Week 21 Getting to Work (Overview)   v   vi Contents of the Overview O.10 Social Acceptability of Long-Distance Commuting and Migration for Unmarried Men and Women 22 O.11 Social Acceptability of Long-Distance Commuting and Migration of Married Men and Women 22 O.12 Gender Differences in Skill Level, by Education, 2015 25 O.13 Gender Differences in Skill Level, by Education, 2009 26 O.14 Workers’ Perceptions: Most Important Characteristics Employers Seek in New Hires 29 O.15 Employers’ Expectations: Most Important Characteristics of Male and Female Workers 29 O.16 Vocational Education and Apprenticeship 30 O.17 Perceptions of Gender-Based Discrimination in the Job Market Despite Similar Levels of Education 31 O.18 Employers’ Preference in Hiring at Different Levels 35 O.19 Labor Market Tightness and Hiring of Female Workers 36 Maps O.1 Labor Force Participation Rate, by District 7 O.2 Poverty Head Count Ratio and FLFP, by District 10 O.3 FLFP and Domestic Household Remittances 11 O.4 FLFP and International Household Remittances 12 Tables O.1 Labor Force Participation, by Consumption Decile, 2009–10 8 O.2 Labor Force Participation, by Consumption Decile, 2012–13 9 Getting to Work (Overview) Acknowledgments This report was prepared by a core team led by Jennifer L. Solotaroff, Senior Social Development Specialist, South Asia Social Development Unit, Social, Urban Rural and Resilience Global Practice (GSU06). David Warren (Practice Manager, GSU06) provided managerial guidance and support. Idah Z. Pswarayi-Riddihough (Country Director, Sri Lanka and Maldives), Françoise Clottes (Director, Strategy and Operations, and former Country Director, Sri Lanka and Maldives), Rolande Simone Pryce (Manager, World Bank Indonesia Program, and former Operations Advisor, Sri Lanka and Maldives), and Valerie Marie Helene Layrol (Senior Operations Officer, Sri Lanka and Maldives) provided overall guidance. The core team members included Jennifer L. Solotaroff (GSU06), George Joseph (GWAGP), Anne T. Kuriakose (GCCCI), Jayati Sethi (GSU06), and Mohammed Ghani Razaak (GSU03). Maria Isabel Larenas Gonzalez provided quantitative data analysis and maps, and Yukari Shibuya provided support on select background research and logistics. Dilinika Peiris, Joe Qian, and Yann Doignon (SAREC) guided the team on communications. Special thanks go to Kamani Jinadasa and Bandita Sijapati (GSU06) for liaising with client counter- parts and general coordination. For excellent editorial and publication support, thanks go to Aziz Gökdemir (GSPDM), with assistance from Jewel McFadden (DECSO). Kerima Thilakasena and Niluka Nirmalie Karunaratne Sriskanthan provided administrative support. Thanks also are due to Rohanthi Perera and her staff at the Sri Lanka Business Development Centre (SLBDC) for data collection. The team extends heartfelt thanks for the care and attention that the Honorable Ms. K.D.M. Chandrani Bandara (Minister, Ministry of Women and Child Affairs, Government of Sri Lanka), Ms. Chandrani Senaratna (Secretary, Ministry of Women and Child Affairs), and Mrs. Ashoka Alawatte (Additional Secretary, Ministry of Women and Child Affairs) gave to the final report draft through written comments and constructive suggestions. The report peer reviewers were Gladys-Lopez Acevedo of the South Asia Region Chief Economist’s Office (SARCE), Nistha Sinha (GPV03), Varun Singh (GSU06), Emcet Oktay Tas (GSU06), and Dileni Gunewardena (Professor of Economics, University of Peradeniya, Sri Lanka). Getting to Work (Overview)   vii   viii Acknowledgments Discussions on framing the issues have benefited greatly from the views of Harini Amrasekera (Open University), Luis Andres (World Bank GWAGP), Nisha Arunatilake (Institute of Policy Studies of Sri Lanka), Harsha Aturupane (World Bank GED06), Juergen Depta and his colleagues at Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), Halil Dundar (World Bank GED13), Nilan Fernando (Asia Foundation), Priyanthi Fernando (Center for Poverty Analysis, Sri Lanka), Nelan Gunasekera (Asian Development Bank), Graeme Harris (IFC, CASSB), Subangi Herath (University of Colombo), Swarna Jayaweera (Center for Women’s Research—CENWOR), Seenithambi Manoharan (World Bank GFA06), Carmen Niethammer (IFC, CASWB), Rosanna Nitti (World Bank GSU09), David Newhouse (World Bank GPV06), and Nihal Somaweera (former Additional Secretary [Regional Development], Government of Sri Lanka). This report has been made possible by Trust Fund support—from a Department of Foreign Affairs and Trade (DFAT), Government of Australia grant through the Partnership for South Asia (PFSA) South Asia Gender (SAGE) Initiative window; and an allocation from the World Bank Group’s Umbrella Facility for Gender Equality (UFGE), a multi-donor trust fund— and by Bank Budget. Getting to Work (Overview) About the Authors JENNIFER LYNN SOLOTAROFF is a Senior Social Development Specialist in the Social, Urban, Rural and Resilience Global Practice and the Gender Coordinator for the World Bank Group’s South Asia Region. Her research inter- ests include gender and labor markets, gender-based violence, and social stratifi- cation in South Asia and East Asia. She has a PhD in sociology, MA in economics, and MA in East Asian studies from Stanford University. GEORGE JOSEPH is a Senior Economist with the Water Global Practice of the World Bank, Washington, DC. His research interests are centered on develop- ment economics and behavioral and applied microeconomics. He received his PhD in economics from Rutgers, the State University of New Jersey, and an MA in economics from Jawaharlal Nehru University, New Delhi, India. ANNE T. KURIAKOSE is a Senior Social Development Specialist at the Climate Investment Funds at the World Bank in Washington, DC. Her research interests include gender and labor, social protection, climate adaptation, and rural liveli- hoods. Anne holds a PhD in development studies from the University of Wisconsin-Madison, an MA in gender and development from the University of Sussex, and BA in political science from McGill University. Getting to Work (Overview)   ix   Executive Summary This report is an overview of a new World Bank study (Solotaroff, Joseph, and Kuriakose, forthcoming) that updates and expands upon the 2013 World Bank policy study, Getting In and Staying In: Increasing Women’s Labor Force Participation in Sri Lanka. Both studies are intended to provide a better under- standing of the puzzle of women’s persistently low labor force participation (LFP) rates and other poor labor market outcomes in the country. The earlier study focused on the years leading up to the end of the Sri Lankan Civil War (2006–09); the current study compares the earlier findings to data from the years following the civil war (2010–15). Using nationally representative secondary survey data, as well as primary qualitative and quantitative research, both studies test three hypotheses that would explain gender gaps in labor market outcomes: (1) household roles and responsibilities, which fall disproportionately on women; (2) a human capital mismatch, whereby women are not acquiring the proper skills demanded by job markets; and (3) gender discrimination in job search, hir- ing, and promotion processes. The current study finds that not only are all three hypotheses supported, as they were with the earlier report, but also the social norms governing women’s responsibilities for child care, elder care, and housework—and that inhibit women from joining labor markets, obtaining employment, and closing gender wage gaps—have become more entrenched since the end of the civil war. Having young children in the household is now associated with even lower odds of LFP, lower chances of becoming a paid employee, and lower earnings for women compared with before 2010, and compared with those of men for all three out- comes. The disparity between marriage’s association with men’s versus women’s odds of LFP is the only gender gap associated with household roles that appears to be shrinking over time; however, marriage still penalizes women in labor mar- kets (lowering their odds of LFP by 4.4 percentage points), whereas for men it provides an 11 percentage point premium in their odds of LFP.1 Gender norms that restrict women’s mobility more than men’s—especially lack of social sup- port for women commuting to work—and that prevent women from accessing safe and comfortable transportation to work, as well as parents’ greater encour- agement of sons’ rather than daughters’ pursuit of careers (especially in the pri- vate sector) are other supply-side factors undermining women in labor markets. Getting to Work (Overview)   xi   xii Executive Summary The analysis also suggests that since 2009, women find it even more challeng- ing than men to translate their educational attainment into high-skill and higher- paying jobs. This is true of women with even university education or higher, who still queue for public sector jobs in spite of limited openings, pushing up their rates of unemployment among young women. Another worrying trend is that poorer and less educated women are falling further behind more educated and wealthier women in chances of LFP and other employment and wage-related outcomes. The good news for women is that raw gender wage gaps are shrinking every year; moreover, the explained portions of these wage gaps are increasing over time. In other words, gender discrimination appears to play less and less of a role in these gaps in earnings; discrimination also appears to determine gender gaps in LFP rates to a diminishing degree over time. The primary data bolster these findings: employers, on average, report that they look for the same skills and experience in men and women, actively discriminating by gender to a much smaller degree than employees suspect. Employers in some industries studied in the primary research—such as the garment industry and tea estate sector—express a preference for hiring women workers because they believe them to be more reliable and hard working than men. Yet, persistent occupa- tional segregation across industries suggests that these preferences may not hold for promotions—especially into high-skill and management jobs, which men continue to dominate. The report concludes with four priority areas (summarized in annex OA) for addressing the multiple supply- and demand-side factors to improve women’s LFP rates and reduce other gender gaps in labor market outcomes. It also offers specific recommendations for improving women’s participation in the five private sector industries studied for the primary data collection: information and com- munication technology (ICT), tea estate work, tourism, garments, and commer- cial agriculture (see annex OB). Common recommendations across the five industries include the provision of care services to ease women’s time poverty, and improvements in providing safe, comfortable transportation to worksites or near- worksite accommodations for women so that they are at lower risk of the gender- based violence that is highly prevalent on public transportation and in public spaces. Together, these recommendations are intended to help the government, the private sector, and other stakeholders in Sri Lanka collaborate and harmonize efforts in getting women to work. Note 1. A full discussion of the study’s methodology, as well as tables of all results from the analyses of primary and secondary data (including results from multivariate analysis of nationally representative secondary data) can be found in the full version of the report (Solotaroff, Joseph, and Kuriakose, forthcoming) at http://www.worldbank.org​ /en/programs/world​-bank-south-asia-region-gender-innovation-lab. Getting to Work (Overview) Abbreviations A-level General Certificate of Education Advanced Level DCS Department of Census and Statistics EPZ export processing zone FGD focus group discussion FHH female-headed household FLFP female labor force participation GDP gross domestic product GoSL government of Sri Lanka HIES Household Income and Expenditure Survey ICT information and communication technology ISCO International Standard Classification of Occupation IT information technology LFP labor force participation LFS Labour Force Survey LKR Sri Lankan rupee NVQ National Vocational Qualification O-level General Certificate of Education Ordinary Level SGBV sexual and gender-based violence SLBFE Sri Lanka Bureau of Foreign Employment STEM science, technology, engineering, and mathematics TVET technical and vocational education and training Getting to Work (Overview)   xiii   Overview I. Introduction Sri Lanka has the 14th-largest gender gap in labor force participation (LFP) globally (WEF 2016). This large gap is surprising given the country’s long-standing achievements in human development outcomes, such as high levels of female education (including gender parity at most levels) and low total fertility rates, as well as its status as a lower-middle-income country with overall improvements in economic growth of more than 6 percent annually over the past decade (World Bank 2015a, 2016). LFP rates among Sri Lankan women age 15 years and older were 36 percent for 2015 and 2016, versus 75 percent for same-age men for both years (DCS 2015, 2016b). In contrast, the 2015 LFP rates for women age 15 and older in Thailand and Malaysia—which are upper-middle- income countries—were 63 percent (compared with 80 percent for same-age men) and 49 percent (compared with 78 percent for same-age men), respec- tively (World Bank 2015c). Sri Lanka’s LFP gender gap is even greater than that of several other South Asian countries (figure O.1), despite Sri Lanka’s serving as a model for the region in many other gender outcomes. Sri Lanka shows remarkable persistence in low LFP rates for women over the past three decades—with even a slight decline as the economy has expanded (figure O.2). This presents significant challenges to the country’s growth and equity goals, such as the current government of Sri Lanka’s aim of creating 1 million jobs, fostering investment in the private sector, and enhancing social inclusion outcomes,1 as it strives to become an upper-middle-income country. Recent economic policy statements have emphasized the need to create an enabling environment for women’s participation in the economy to achieve the government’s goal of inclusive and balanced development. The government envisages that 40 percent of the jobs created by 2020 will employ women, and it seeks to encourage women’s greater involvement and leadership in small and medium enterprises. The most potent route to growing Sri Lanka’s overall work- force will come from increased numbers of women working. Raising the rate of women’s LFP by 15 percentage points over current rates will add more than 1 million workers to the labor market each year (Sinha 2012). Getting to Work (Overview)   1   2 Overview Figure O.1  Labor Force Participation, by Country 90 Percent of male and female population 80 70 60 age 15 and older 50 40 30 20 10 0 11 6 14 16 4 11 5 Q2 −1 01 −1 20 20 5− 0− 16 s2 15 14 01 01 an an 20 20 20 ve a2 l2 ist ut di ka sh n pa Bh an di al ta an de M In Ne kis gh iL la Pa Af Sr ng Ba Male Female Source: National estimates, various years. Figure O.2  Female Labor Force Participation, by Select Country, Economic Status, and Region,1993–2016 90 Percent of females age 15 and older 80 70 60 50 40 30 20 10 19 3 94 19 5 96 97 98 99 00 01 02 03 04 20 5 06 20 7 08 09 10 11 12 20 3 14 15 16 9 9 0 0 1 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Year Nepal Middle income South Asia Afghanistan Sub-Saharan Africa Malaysia India Upper middle income Sri Lanka Pakistan Source: World Bank datacenter, modeled ILO estimate. Note: Some ILO data may predate the 2016 LFS data noted in the text. ILO = International Labour Organization; LFS = Labour Force Survey. Getting to Work (Overview) Overview 3 Nonetheless, women’s experience in Sri Lanka’s labor market remains char- acterized by low LFP; high unemployment, especially for women younger than age 30; and persistent, though shrinking, wage disparities between the sexes. As this study shows, determinants of these poor gender outcomes include household roles and responsibilities, a mismatch between skills acquired in school and those demanded in the labor market, and gender discrimination in labor supply as well as labor demand dynamics. This report is intended for policy makers and employment program practi- tioners in the Sri Lankan government, the private sector, and the donor and nongovernmental organization communities. It also targets academia and other research institutions, in part to call upon them to undertake additional studies that can continue to identify the most effective means of engaging and sustaining more women in the workforce—particularly in the private sector. Finally, this study is intended to reach any others who have a stake in helping Sri Lanka’s economy grow by taking advantage of this relatively untapped population of potential labor, innovation, and productivity: women. By examining gender norms about work as well as the typical economic fac- tors in analyses of gender and labor dynamics, this study explores why, compared with men, women continue to be well educated but less commonly working for pay in Sri Lanka. It identifies means of promoting women’s entry into and con- tinued employment in the labor market, which will grow the economy. Improved female LFP (FLFP) will also be critical to helping the country cope with its now-rising inverse dependency ratio: the demographic transition now under way suggests that the population older than age 60 will double in the next quarter-century, whereas the younger working-age population will continue to decrease because of lower total fertility rates (World Bank 2016). This study provides an overview of a new World Bank study (Solotaroff, Joseph, and Kuriakose, forthcoming) that updates and expands upon the 2013 World Bank policy study, Getting In and Staying In: Increasing Women’s Labor Force Participation in Sri Lanka. Both studies are intended to provide a better understanding of the puzzle of women’s persistently low LFP rates and other poor labor market outcomes in the country as a whole, rather than in a particular city or province. Quantitative analyses are nationally representative. Most of the earlier economic analyses have attributed gender gaps in these outcomes (that is, LFP, employment, and earnings) that were unexplained by household time constraints or human capital factors to the “black box” of social factors, including gender-based discrimination. Using primary data as well as existing national-level survey data from the Sri Lanka 2006–10 Labour Force Survey (LFS) and the 2009–10 Household Income Expenditure Survey (HIES), the earlier report sought to unpack the social processes underlying these gender differences. It posed three hypotheses that would explain gender gaps in labor market outcomes: (1) household roles and responsibilities, which fall disproportionately on women; (2) human capital mis- match, whereby women are not acquiring the proper skills demanded by job mar- kets; and (3) gender discrimination in job search, hiring, and promotion processes. Multivariate analysis of the secondary data found support for all three hypotheses. Getting to Work (Overview) 4 Overview The earlier analysis of secondary data reflected gender-biased labor market dynamics in the final years of Sri Lanka’s civil war, which ended in 2009. Although the analysis summarized these dynamics in Sri Lanka, national-level surveys tended to exclude the more conflict-affected areas2 (for example, dis- tricts in the Northern Province and sometimes the Eastern Province) until 2011. The primary research for the current report was conducted in 2012 in select industry sectors in the Badulla, Gampaha, and Trincomalee districts in lower-central, western, and eastern Sri Lanka, respectively. The researchers used quantitative and qualitative methods to ask questions of different groups of workers, household members, and employers about their labor market expe- riences and attitudes toward work. Industry sectors were selected to include a mix of “new” and traditional drivers of the economy: information and com- munication technology (ICT), tea estates, tourism, the garment sector, and commercial agriculture. The primary research (particularly the qualitative data) helps explain employer preferences and incentives as well as gender dif- ferences in educational choices, job preferences, occupational aspirations, job search channels, household decision making, and time use patterns—among other labor-related factors—between men and women of different ages, educa- tion and income levels, ethnicities, and employment types. The analysis of secondary and primary data together adds value to existing labor studies of Sri Lanka by using the additional lenses of gender norms, identity, and agency (the ability to make decisions as well as take advantage of opportunities).3 One of the most recent existing studies (Gunewardena 2015) analyzes the 2012 World Bank STEP Skills Measurement Program survey data to explore why Sri Lankan women’s educational gains are not translating into workforce advantages. The findings contribute an unprecedented, nuanced understanding of Sri Lankans’ perceptions of their own skills and how they are linked to labor market advantages. The findings also provide sharper definition to the 2013 policy report’s mixed-methods exploration of a mismatch between women’s educational attainments, on the one hand, and skills sought by employers— especially those in the private sector—on the other. This updated report analyzes more recent national survey data (2011–15 LFS and 2012–13 HIES) to shed light on whether and how labor force patterns have changed for women over the past decade, with particular attention to the years since the end of the civil war. Any quantitative analysis using national survey data (that is, LFS and HIES) was conducted twice for each survey year—first using the full sample (all provinces) from that year and then using a sample that dropped the districts and provinces not included in surveys from years before 2011 to allow for comparability across years.4 This study also presents findings from the qualita- tive and quantitative primary data more comprehensively than the previous study. Finally, this update identifies ways to promote women’s entry into and sustained employment in the labor market. Recommendations at the end of this report (and summarized in annex OA) are tailored to different stakeholders, with annex OB providing special focus on certain industries in the private sector that have strong potential to absorb and sustain Sri Lankan women in paid decent work. Getting to Work (Overview) Overview 5 II. Summary of Descriptive Data: Demographic Changes over Time In recent years, broader macroeconomic changes in Sri Lanka have reflected an ongoing transition from agriculture to the industry and services sectors. Construction, transport, domestic trade and banking, and insurance and real estate have become important contributors to economic growth. Together, these sectors accounted for 50 percent of the total increase in gross domestic product (GDP) between 2009 and 2014 (World Bank 2015a). Manufacturing and ser- vices constituted the second-largest share of GDP growth at 44.7 percent. In comparison, the agriculture sector’s contribution was only 5.3 percent. Understandably, overall employment trends are analogous. Data from National LFSs from 2000 through 2015 show employment rates expanding for the indus- try and services sectors but contracting for agriculture. Gender Gaps in LFP by Residential Sector, Age, District, and Ethnicity Throughout these macroeconomic transitions, the FLFP rate has been declining and remains lowest for urban women. The 36 percent FLFP rate for 2015 (and 2016) indicates a distinct decline since 2010, when it was 41 percent versus 82 percent for men of the same age group (15 years and older).5 In 2006, the FLFP rate was even higher, at 46 percent. Over nearly a decade, therefore, FLFP rates have dropped by roughly 10 percentage points. Like the population at large, the Sri Lankan workforce is aging—particularly for women—because of the demo- graphic transition under way (figure O.3, panels a and b). Although urban women continue to participate the least in labor markets (with LFP hovering around 30 percent), LFP rates have fallen most among Figure O.3  Labor Force Participation, by Age and Gender, 2009 and 2015 a. 2009 b. 2015 1.0 1.0 0.9 0.9 0.8 0.8 0.7 0.7 0.6 0.6 Percent Percent 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 15 20 25 30 35 40 45 50 55 60 65 70 75 80 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Age Age Male Female Source: World Bank calculation based on 2009 and 2015 Sri Lanka Labour Force Surveys. Note: Population age 15 and older. Data from the Northern Province were excluded to maintain comparability over time. The 2009 weight factor was adjusted by World Bank projection of total population from the 2012 census. Getting to Work (Overview) 6 Overview Figure O.4  Labor Force Participation, by Gender and Residential Sector, 2006–15 80 Percent of population age 15 and older 70 60 50 40 30 20 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year Rural, male Estate, male Urban, male Rural, female Estate, female Urban, female Source: World Bank calculation based on Labour Force Surveys for 2006 through 2015. Note:The Northern and Eastern Provinces were excluded to keep comparability among years.The weight factor was adjusted by World Bank projection of total population from the 2012 census, except for 2014 and 2015. women who work in the estate sector (that is, tea plantation and other planta- tion estates on the island) by nearly 10 percentage points between 2006 and 2015 (figure O.4). The LFP rate of women in estates still far surpasses that of rural and urban women, however. In all residential sectors, men’s participation is considerably and consistently higher than women’s participation, though the LFP rate for urban men has declined by about 5 percentage points since 2006. Map O.1 geographically depicts the stark disparity between men’s and women’s LFP rates by district. Women’s participation rates were lowest (less than 20 percent) in Kilinochchi District and peak in the range of 40–60 percent in eight districts—all in the middle of the island (except Kandy) in 2015. Not surprisingly, many of these districts are in the estate sector, particularly those with heavy tea cultivation, such as Badulla and Nuwara Eliya, and rubber culti- vation, such as Ratnapura. Men’s participation rates, on the other hand, are consistently 60–80 percent. FLFP rates also vary considerably by ethnicity6 (figure O.5), income level, and education: Indian Tamil women and women with the greatest wealth and educational attainment exhibit the highest participation rates in the most recent years. Despite falling FLFP rates in the estate sector, Indian Tamil women there continue to attain the highest FLFP rates (58 percent in 2009 and 50 ­percent in 2015) of all ethnic groups in Sri Lanka, including the Sinhala majority. Sri Lankan Moor women tend to have the lowest rates (17 percent in 2009 and 15 percent in 2015). Men’s LFP rates are much more uniform—hovering between 65 and 77 percent for the six main ethnic groups in 2015—which sug- gests greater cultural constraints on women’s participation than on men’s in the Getting to Work (Overview) Overview 7 Map O.1  Labor Force Participation Rate, by District a. Female b. Male Ja na Ja na < 20% < 20% 20−40% 20−40% Kilinochchi 40−60% Kilinochchi 40−60% Mullaitivu 60−80% Mullaitivu 60−80% > 80% > 80% Mannar Mannar Vavuniya Vavuniya Trincomalee Trincomalee Anuradhapura Anuradhapura Puttalam Puttalam Polonnaruwa Polonnaruwa Batticaloa Batticaloa Kurunegala Kurunegala Matale Matale Kandy Ampara Kandy Ampara Gampaha Kegalle Nuwara Eliya Gampaha Kegalle Nuwara Eliya Badulla Badulla Colombo Moneragala Colombo Moneragala Ratnapura Ratnapura Kalutara Kalutara Hambantota Hambantota Galle Matara Galle Matara Source: World Bank calculation based on the 2015 Labour Force Survey. Note: Population age 15 and older. Figure O.5  Labor Force Participation, by Gender and Ethnicity, 2009 and 2015 80 Percent of population age 15 and older 60 40 20 0 Male Female Male Female 2009 2015 Sinhala Indian Tamil Malay Other Sri Lankan Tamil Sri Lankan Moor Burgher Source: World Bank calculation based on the 2009 and 2015 Labour Force Surveys. Note: Data from the Northern Province were excluded to maintain comparability over time. The 2009 weight factor was adjusted by World Bank projection of total population from the 2012 census. Getting to Work (Overview) 8 Overview same ethnic group. The lower comparative LFP rates for both men and women in the Sri Lankan Tamil, Moor, Malay, and Burger ethnic groups (in comparison with Sinhalese and Indian Tamils) suggests sustained socioeconomic exclusion of these four groups. Gender Gaps in LFP by Household Income Level, Poverty, and Migration The U-shaped curves that characterized women’s LFP by income and schooling in 2009–10 no longer hold; the curves are “flatter” for those with lower levels of wealth and education, suggesting that perhaps the poorest and least-educated women in Sri Lanka may be bearing the brunt of deterio- rating labor outcomes for women. Before the end of the civil war, women from households with the lowest incomes (consumption deciles 1–2) and the highest incomes (deciles 8–10) participated more in labor markets than women from middle-income households (table O.1), according to HIES 2009–2010 data. This U-shaped pattern is evident for national averages of FLFP and is especially pronounced for women in the urban and rural sectors, whereas urban and rural men’s FLFP rates remain similar across income deciles, dipping only at the highest levels of wealth. The relatively high rates of estate sector women’s LFP—compared with those of other women—are fairly consistent across all but the highest income deciles. According to 2012–13 HIES data for the country as a whole, however, women from the poorest households participate the least (30–31 percent) and the wealthiest the most (36–39 percent), with rates for middle-income women at 32–34 percent (table O.2). This monotonically increasing pattern is strongest in the rural sector, where FLFP rates exhibit an inverted-U shape, with middle-income women achieving the highest rates. Also important to note is the overall decline in the estate sector’s LFP rates—for men and women alike—between 2009 and 2013.7 The relative disadvantage of poor women in the Table O.1  Labor Force Participation, by Consumption Decile, 2009–10 percent Consumption National Urban Rural Estate decile All Male Female Male Female Male Female Male Female 1 60.2 83.1 39.3 80.4 38.2 83.6 37.9 80.5 60.1 2 60.1 83.3 39.6 78.3 35.4 83.5 38.3 87.3 59.0 3 59.3 84.0 37.4 80.6 29.4 84.3 36.2 85.8 61.3 4 58.7 83.4 36.6 80.0 35.0 83.9 34.9 83.4 61.7 5 58.8 83.0 37.8 80.8 30.1 82.9 37.2 88.7 58.7 6 58.2 84.5 34.8 83.9 28.9 84.3 34.0 89.3 59.5 7 58.2 83.0 36.8 77.0 34.9 83.6 35.9 91.5 59.9 8 58.7 81.4 39.1 79.1 39.2 81.9 38.5 82.2 56.8 9 58.8 81.0 39.6 79.8 36.8 81.3 39.7 81.4 65.0 10 59.2 78.4 42.5 76.4 40.5 79.0 43.2 85.2 47.0 Source: Data from Household Income and Expenditure Survey 2009–10. Note: Includes persons age 15 and older. Getting to Work (Overview) Overview 9 Table O.2  Labor Force Participation, by Consumption Decile, 2012–13 percent Consumption National Urban Rural Estate decile All Male Female Male Female Male Female Male Female 1 51.3 76.0 30.2 72.9 25.1 76.6 29.8 71.8 40.5 2 51.8 75.9 30.6 74.2 24.1 76.4 29.5 73.2 51.5 3 53.6 78.8 31.1 75.1 22.9 79.6 30.9 76.5 50.3 4 54.5 78.2 33.9 75.0 25.7 78.7 33.6 78.7 56.0 5 53.8 78.7 32.0 77.6 23.1 78.8 32.3 79.8 52.4 6 53.0 77.4 32.7 73.2 21.1 78.4 34.5 77.8 52.2 7 53.5 77.7 33.8 72.9 27.0 78.8 34.6 80.6 54.2 8 52.9 77.2 32.9 69.8 27.2 79.5 34.2 82.1 51.2 9 53.3 74.8 35.6 68.4 30.2 76.9 37.3 90.1 47.3 10 55.2 74.5 39.3 69.1 34.8 77.0 41.4 84.4 48.6 Source: Data from Household Income and Expenditure Survey 2012–13. Note: Includes persons age 15 and older. rural and estate sectors appears to be more pronounced than before, as is—by extension—the need to prioritize these groups with interventions to improve access to labor markets. If poverty were the sole driver of women’s labor market participation in Sri Lanka, women’s participation rates would be highest in the most impover- ished districts, and lower in districts with lower poverty rates. The data do not bear out this pattern, however. Using the district-level poverty head count ratio8 from HIES 2012–13, map O.2 displays each district’s poverty head count ratio by FLFP rates in that district. The map shows much area in which women’s participation is low but poverty is high (all the eastern and northern coastal districts—Trincomalee, Batticaloa, Mullaitivu, Mannar, Jaffna, and Kilinochchi) and in which women’s participation is high but poverty is low (Kurunegala, Matale, Kegalle, Kalutara, Nuwara Eliya, and Hambantota). Women continue to make up a significant share of the overseas Sri Lankan labor force, though their share—relative to men’s—has decreased since 1997 and since 2012 has dropped to less than 50 percent. In 2003, women made up 53 percent of the total overseas workforce (World Bank 2007) and as much as 62 percent in the later years of that decade (Arunatilake et al. 2010). Of the 282,331 Sri Lankans working abroad in 2012, 138,547 (49 percent) were women and 86 percent of these women were housemaids in domestic service (Jayasura and Opeskin 2015). Using the number of departures for foreign employment, disaggregated by sex, to investigate changes over time in shares of men and women migrating overseas, the Sri Lanka Bureau of Foreign Employment (SLBFE) reports that, even though the total number of departures has been increasing over the past 30 years (from 14,456 in 1986 to 300,703 in 2014), the share of women has fluctuated widely. This share was 24 percent in 1986, peaked at slightly more Getting to Work (Overview) 10 Overview Map O.2  Poverty Head Count Ratio and FLFP, by District Ja na FLFP−HCR High−high High−low Kilinochchi Low−high Low−low Mullaitivu Mannar Vavuniya Anuradhapura Trincomalee Puttalam Polonnaruwa Batticaloa Kurunegala Matale Kandy Ampara Kegalle Gampaha Nuwara Eliya Badulla Colombo Moneragala Kalutara Ratnapura Galle Hambantota Matara Source: World Bank calculation based on HIES 2012–13. Note: FLFP = female labor force participation. Population age 15 and older. Getting to Work (Overview) Overview 11 Map O.3  FLFP and Domestic Household Remittances Ja na FLFP−local High−high High−low Kilinochchi Low−high Low−low Mullaitivu Mannar Vavuniya Anuradhapura Trincomalee Puttalam Polonnaruwa Batticaloa Kurunegala Matale Kandy Ampara Kegalle Gampaha Nuwara Eliya Badulla Colombo Moneragala Kalutara Ratnapura Galle Hambantota Matara Source: World Bank calculation based on HIES 2012–13. Note: FLFP = female labor force participation. Population age 15 and older. Getting to Work (Overview) 12 Overview Map O.4  FLFP and International Household Remittances Ja na FLFP−abroad High−high High−low Kilinochchi Low−high Low−low Mullaitivu Mannar Vavuniya Anuradhapura Trincomalee Puttalam Polonnaruwa Batticaloa Kurunegala Matale Kandy Ampara Kegalle Gampaha Nuwara Eliya Badulla Colombo Moneragala Kalutara Ratnapura Galle Hambantota Matara Source: World Bank calculation based on HIES 2012–13. Note: FLFP = female labor force participation. Population age 15 and older. Getting to Work (Overview) Overview 13 than 75 percent in 1997, and declined thereafter to 52 percent in 2009 and 37 percent in 2014. SLBFE (2015) attributes the sharp decline of women among overseas workers to the rapid increase in men’s outmigration along with several SLBFE policy changes occurring in late 2013, including raising the minimum age for women who emigrate for domestic work. Female overseas workers employed in domestic service and in the health sec- tor contribute greatly to national and household incomes. Remittances from migrating household members may partly explain some of the spatial variation in FLFP rates.9 HIES data limitations do not allow identification of who in the household—especially whether male or female—is migrating for work and send- ing remittances. Still, it is worth noting that districts in northern, eastern, and western Sri Lanka have both low FLFP rates and low receipts of domestic house- hold remittances (map O.3); FLFP and domestic (local) remittances are both high throughout the center of the country. Only Kurunegala, Matale, and Kalutara have both high remittances and high FLFP rates; most districts with high international receipts tend to have low FLFP rates—Kandy; Jaffna in the far north; Ampara, Polonnaruwa, and Batticoloa in the east; and the west coast dis- tricts except for Galle and Kalutara (map O.4). The low levels of both domestic and international migration from the conflict-affected Northern and Eastern Provinces are surprising. Given that these areas have among the highest poverty rates in the country, the relative lack of out-­ migrating women suggests that mobility and other sociocultural constraints may be acting on women in the largely Tamil Muslim local population. Other explanations could be that these women lack the social networks that facilitate migration, or that they simply lack the funds required for travel to the destination. Gender Gaps in LFP Are Rising At All but the Highest Education Levels Women’s participation rates with respect to education also resembled a U-shaped curve before the end of the conflict (panel a of figure O.6), with LFP at its lowest for those who stopped schooling after completing grade 10 (General Certificate of Education Ordinary Levels, or O-levels)—though the slope on the right side of the U-shaped curve was much steeper than on the left (that is, education beyond O-levels was associated with much higher FLFP rates than education completed before O-levels). More recent data (panel b of figure O.6) show a similar skewed-U-shaped curve for FLFP with respect to education; however, the lowest levels of educa- tion (education below grade 6 or no education) are associated with slightly lower FLFP than before, whereas the middle-upper range of educational attainment is associated with higher FLFP rates than before. Although O-level education is still associated with the lowest FLFP rates, the 2015 rate is about 35 percent—a few percentage points higher than in 2009. In 2015, women’s participation rates also were higher for those who continued beyond O-levels and A-levels (General Certificate of Education Advanced Levels) to university education, where the FLFP rate increases sharply to more than 85 percent for women with university education, as opposed to less than 80 percent in 2009. Getting to Work (Overview) 14 Overview Figure O.6  Labor Force Participation, by Education and Gender, 2009 and 2015 a. 2009 b. 2015 100 100 90 90 80 80 70 70 Percent Percent 60 60 50 50 40 40 30 30 n 6 8 ed ed y n 6 8 ed ed y sit sit 6− 6− io io e e et et et et ad ad at at er er e de pl pl pl pl uc uc iv iv gr gr ad m m m m a Un Un ed ed w w Gr Gr co co co co lo lo No No ls ls ls ls Be Be ve ve ve ve le le le le O- A- O- A- Educational attainment Educational attainment Male Female Source: World Bank calculation based on 2009 and 2015 Labour Force Surveys. Note: Persons age 15−64. Data from the Northern Province was excluded to maintain comparability over time. The 2009 weight factor was adjusted by World Bank projection of total population from the 2012 census. FLFP rates for those with less than grade 6 education fell between 2009 and 2015, from roughly 46 percent to about 42 percent. For women with no edu- cation, LFP rates dropped slightly to about 41 percent in 2015. LFP rates have also declined for men with no schooling through grade 6 education, risen for those completing education between grade 6 and O-levels, and held steady for those with A-levels and above. The gender gap in the LFP “payoff” to men’s and women’s investments in education has narrowed, but only at higher levels of education. One large factor is the rising returns to schooling for women with more advanced educational attainment. The gender differential in participation rates is shrinking for those who attend university (from a 10 percentage point gap in 2009 to a 2015 gap of less than 5 percentage points). The new trend showing LFP rates falling more in 2015 for men who continue on to complete their A-levels is a concern. It may be reflecting the higher numbers of boys dropping out of secondary education compared with girls—in part because they are losing confidence that higher levels of education can equip them with the skills they need for the job market and in part because they are under increasing pressure to earn income for their families (World Bank, forthcoming). Gender Gaps in Unemployment, Wages, and Employment Type The gender gap in unemployment appears to be shrinking slightly as well, especially in the rural sector. This gap has dropped steadily each year, from ­ Getting to Work (Overview) Overview 15 5 percentage points in 2006 (when women’s unemployment rate was 9.7 percent compared with men’s rate of 4.7 percent) to 2.9 percentage points in 2015 (DCS 2016a). Rural women continue to have the highest unemployment rates of all, though these rates have declined from nearly 11 percent in 2006 to 8 percent in 2015—thus the shrinking gender gap in rural unemployment (rural men’s unemployment rates range from 2.9 percent to 5 percent over the same period). Estate sector women tend to have the lowest unemployment rates among all women; since 2008, their rates also have stayed within 2 percentage points of those of estate men, who consistently have the lowest unemployment rates in all sectors, with a rate of 2.7 percent in 2015. Men’s unemployment is highest in urban areas, at 3.5 percent, while urban women’s unemployment is nearly twice that, at 6.7 percent (DCS 2015). Young women still have the highest rates of unemployment in Sri Lanka (figure O.7), and the gender gap in youth unemployment rates is expanding. In 2015, unemployment appeared to be most entrenched for women ages 20–37 or so. Female unemployment rates are consistently higher than those for same- aged males—except around age 18, at which point they converge. Unlike in previous years, when female unemployment peaked at higher ages, its apex now occurs for girls age 15–16, drops to about 22 percent around age 18, and shoots up again to 35 percent for 20-year-old women. Female unemployment rates do not fall below 5 percent until age 37. Male unemployment peaks at age 17, falls below 7 percent by age 25, and remains under 5 percent for men ages 27 and older. Even if this shift primarily reflects the addition of the many poor in the conflict-affected Northern Province—who have no choice but to seek work even though job opportunities are scarce compared with most other parts of the Figure O.7  Unemployment, by Age and Gender, 2015 35 30 25 Percent 20 15 10 5 0 15 20 25 30 35 40 45 50 55 60 65 Age Male Female Source: World Bank calculation based on 2015 Labour Force Survey. Note: Persons age 15−64. Getting to Work (Overview) 16 Overview country—young women are generally facing increasing barriers to employment, compared with young men, in Sri Lanka as a whole. The new trend of high female unemployment rates at the young ages of 15 and 16 may signal an uptick in dropout rates among girls around O-levels (grade 10), perhaps because of an increasing need among the poorest households for as many family members as possible to generate income. Unemployment by education level also differs for men and women: whereas unemployment tends to rise with education for both groups, it rises at an increasing rate for women who complete their educations between grade 6 and grade 12 (A-levels) (figure O.8). The gender gap in unemployment rates is larg- est for those who stop education after A-levels. It shrinks only slightly for univer- sity attendance, remaining above 10 percent for women but dropping to about 3 percent for men. If Sri Lanka is to improve its unemployment as well as LFP rates, it will need to ensure that (1) more girls and boys complete their educa- tions at least through grade 12 and (2) skills acquired at higher levels of educa- tion are better aligned with jobs—in particular, for women who graduate from high school but do not continue to university. Women also remain on the losing end of gender wage gaps, although the average raw earnings gap has narrowed over time. According to the 2015 data, the raw earnings gap (calculated by multiplying hourly wage by hours worked in the month before the survey, plus all other earnings, including ben- efits) averaged across all provinces was 15.9 percent, with women’s average Figure O.8  Unemployment, by Education Level and Gender, 2015 12 10 8 Percent 6 4 2 0 n 6 8 ed ed y sit 6− tio de et et er a a e pl pl uc iv gr ad m om Un ed w Gr co c lo No ls ls Be ve ve le le O- A- Educational attainment Male Female Source: World Bank calculation based on 2015 Labour Force Survey. Note: All provinces. Persons age 15−64. Getting to Work (Overview) Overview 17 monthly wage being 17,729.5 Sri Lankan rupees (LKR) and men’s being 20,839.6 LKR. This is smaller than the 2011 gap of 18.7 percent. In fact, the raw earnings gap has been steadily shrinking even since before the end of the civil war. Dropping the Northern Province from the samples to allow for com- parison across years, the gender gap was 19.9 percent in 2009 compared with 15.8 percent in 2015. Women may thus be benefiting from the higher labor earnings that helped drive the country’s impressive poverty reduction between 2002 and 2012–13.10 The gender wage gap is highest in the estate sector and lowest in the urban sector. Although women earn less than men in all sectors, urban women, on average, earn more than both men and women in the rural and estate sectors. As for changes over time in the type of work women are doing, this investiga- tion finds only slight improvements in their employment status. The share of women among own-account workers increased slightly during the past decade— from 26 percent in 2006 to 27 percent in 2015—as did the share of women among employers (from 9.4 percent in 2006 to 12.8 percent in 2015), with the shares of men decreasing by the same margin on both counts. However, the share of women among those with contributing-family-worker status, which is unpaid, rose from 72 percent to 78 percent over the same period. To provide a better sense of scale, with the denominator being all employed individuals, in 2015, contributing family workers accounted for 18.8 percent of all employed females, though this share is very low compared with other countries in South Asia (DCS 2016a). Still, it is far higher than for employed males, among whom only 2.8 percent engage in unpaid work. More than 36 percent of working men are engaged in own-account work and 57 percent of men work as paid employees— compared with 25 percent and 55 percent of working women, respectively (DCS 2016a). Only 1.1 percent of working women are employers, compared with 4.2 percent of working men. These numbers have not changed much over the years, although the mild increase in female employers (from 0.8 percent of employed women in 2006, as opposed to the decrease in male employers, from 4.4 in 2006) and in women own-account workers is promising (DCS, various years). Yet women clearly continue to be vastly overrepresented among unpaid workers. They also are overrepresented among public sector employees (19 percent of all employed women versus 13 percent of employed men) and underrepre- sented among private sector employees (36 percent of employed women versus 44 percent of employed men). Women have indeed shifted from private to public sector employment more than men since 2006, when 15.6 percent of employed women worked in the public sector (compared with 12 percent of all employed men) and 39 percent worked in the private sector (compared with 44 percent of employed men, the same as in 2015). Among paid employees in Sri Lanka’s private sec- tor, slightly more than 50 percent are formal workers; the remaining persons are informal workers in informal (19 percent) or formal (31 percent) enter- prises (DCS 2016a). More than one-third of working women are employed in the informal sector, which is characterized by low wages, lack of social Getting to Work (Overview) 18 Overview protection, and the failure to reward workers’ skills at levels comparable to those in the formal sector (ADB 2008; DCS 2016a). Men are more likely to be employed in informal jobs than are women—56 percent of men versus 37 percent of women (DCS 2016a; World Bank 2012). More than one-third (36 percent) of all informal workers are employed in microenterprises. Private sector formal employment accounts for just one-fourth of those who are working (DCS 2016a). Although women are moving out of agriculture at a pace similar to that of men, compared with men they are moving more into the services sector and less into the manufacturing sector. Among women participating in the labor force in 2011, 37.5 percent were in agriculture, 24.2 percent in manufactur- ing, and 38.3 percent in services (compared with 30.6 percent, 23.4 percent, and 45.9 percent of men distributed across agriculture, manufacturing, and services in 2011). As of 2015, the share of working men in agriculture dropped by 4 percentage points (to 26.5 percent) and increased by 2 percent- age points in manufacturing (to 25.4 percent) and by more than 2 percentage points in services (to 48.1 percent). Women, on the other hand, saw a 5 per- centage point drop in agriculture (to 32.5 percent of working women), but a 4 percentage point increase in services and only a 0.5 percentage point increase in manufacturing (to 24.7 percent). Occupational segregation has grown within the services sector as well: between 2011 and 2015, the share of working men in traditional services dropped by 9 percentage points, but by less than 0.5 percentage point for women; in intermediate services, the share among working women also fell, whereas the share rose for men by nearly 9 percentage points. The good news is that women are moving more into mod- ern services than men are.11 III. Hypothesis Testing: All Explanations for Women’s Poor Outcomes Are Still Supported All three explanations for low FLFP in Sri Lanka—household roles and related mobility constraints, human capital mismatch, and gender discrimina- tion—continue to play a significant role in labor outcome gender gaps, though to different degrees than before. The multivariate analysis of more recent data suggests that women’s roles in the marital household still signifi- cantly lower women’s odds of joining the labor force—and even more so than before the end of the civil war. Qualitative research confirms the increasingly powerful influence that these gender norms have on family roles and on women’s safety and mobility—all of which obstruct women from getting to work, diminishing both labor supply and labor demand for women workers: employers as well as families and communities expect women either to not enter or to leave the workforce if they are married and have children. Traveling to work becomes even less acceptable once women move into these family roles. Getting to Work (Overview) Overview 19 Household and Family Roles and Mobility Constraints Still Penalize Women in Labor Markets, Especially Women with Young Children As found in the previous report (World Bank 2013) and Gunewardena (2015), gender norms around household roles—such that housework, elder care, and child care responsibilities typically fall to women—continue to create time poverty and lower social support for women’s LFP and employ- ­ ment. The negative effects of some household roles have been dampened in recent years, whereas for other roles the negative effects have intensified. As always, these roles and responsibilities create significant hurdles for women in entering the labor market and continuing employment, especially after marriage and childbirth. Marriage penalizes women’s participation in labor markets, but the penalty has shrunk over the past decade. In 2015, odds of participation ­ were 4.4 percentage points lower for married women than for unmarried women—whereas marriage provided an 11 percentage point premium for men nationwide. Compared with the years before the conflict’s end, the penalty on married women appears to have shrunk (from 8 percentage points in 2006 and 6 percentage points in 2009), as has the premium for married men (14 percentage points in both 2006 and 2009). In other words, the gender gap in marriage’s effect on LFP appears to have narrowed over the past decade, when considering residential areas in aggregate. When looking at the urban population alone, the gender disparity contin- ues to be more pronounced than in other residential sectors: in 2015, mar- riage lowered urban women’s LFP odds by 11 percentage points (compared with those of unmarried urban women), but raised them for urban men by 10.3 percentage points. This is an improvement since 2006, however, when marriage was associated with a 17 percentage point penalty for urban women (and a 15 percentage point boost for urban men). Marriage now even more drastically lowers women’s odds—by 26 percent- age points—of becoming a paid employee, whereas for men it slightly increases the odds, by 2.5 percentage points. In 2009, marriage was associated with 18 percentage point lower odds for women. Similarly, the presence of children younger than age five in the household now has the greatest-ever association with lower chances of LFP for Sri Lankan women, whereas it has never had any significant effect for men—at least over the decade covered in this study. Considering all residential areas together, hav- ing a child under age five in the household makes women 7.4 percentage points less likely to join the labor force than women without young children, but it has no effect for men. This negative association is larger than it was in 2009, 2011, and 2013, when childrearing was associated with lower LFP odds for women of 7, 5, and 6 percentage points, respectively. In 2006 this negative association was more pronounced for women in the rural and estate sectors (8.4 percentage points) than for urban women (3.6 percentage points), but in 2015 it was about the same for urban women and non-urban women. For women with young Getting to Work (Overview) 20 Overview children in the household, living in an estate or rural area is associated with 7.7 percentage point lower odds of LFP (compared with those of women without children), and by 6.2 percentage points if living in an urban area. Yet, ­ there is no significant effect of having young children on men’s prospects, regardless of where they live. The need for child care services may be rising for women in urban areas, and it remains high for women in the rural and estate sectors. This rising need is likely coupled with slight improvements in attitudes about the acceptability of women’s paid work, which would encourage more women to enter the labor force. On the other hand, being married and having young children is associated with lower odds of unemployment for most women and men in 2015, who have 2–3 percentage point lower odds of unemployment than those who are unmarried. The only group for whom marriage and unemployment have no significant relationship is urban women—possibly reflecting their relatively lower odds of LFP; those who do not join the labor market have no chance of being unemployed. The significance and magnitude of marriage and children’s negative effects on unemployment in 2015 were virtually identical to those of earlier years. Qualitative results regarding the social desirability of women’s LFP and employment—and thus unemployment—tend to vary by location, which may have some relationship to women’s presence (or lack thereof) in the dominant local industry. In Badulla district, a young man employed in commercial agricul- ture remarks, “[Culturally] women being employed cannot be approved. The term ‘housewife’ means that the woman of the household should attend to all the household chores and be responsible for the upbringing of the children and not be someone who should go out and earn money.”12 His views reflect those of many men and women in this conservative rural Tamil village. In Gampaha district, where ICT and garments are prominent industries, a man says, “Local customs demand that women bring up their children and look after the home. The man of the house should work and provide for the family. Although these customs are changing now, generally parents consider that their male children should be employed full-time.”13 Women in the tea estate focus group discus- sions (FGDs) provide some more nuanced views, saying that, because men on the estate cannot be relied upon to earn income for the household, estate women must enter the workforce. This finding implies that the social norm that women should not work unless as a secondary income earner as a household necessity has been overcome in the estate sector, primarily because of economic need. The presence of elderly household members has the potential to offset some of young mothers’ child care responsibilities; however, it could also add to their overall care burden. In 2009, the presence of those over age 64 had a positive, significant association with odds of LFP for both men and women in urban and non-urban areas, with the greatest positive association with LFP of urban women. This positive association appears to have shrunk in 2015 and shifted entirely to the rural and estate sectors: both men and women in these sectors who live with elderly household members have just 1 percentage point higher Getting to Work (Overview) Overview 21 Figure O.9  Reasons for Not Working Last Week 100 Percent of population age 15 and older 80 60 40 20 0 Men without Women without Men with Women with children children children children Family (housework or child care) Still didn’t find a job Off season Student Family needs (general) Others Source: World Bank calculations based on Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. odds of joining the labor force compared with people in households without elderly members. These results suggest that the help elderly in the home pro- vide to offset women’s home care responsibilities—to the extent that they can participate in the labor force—is declining and has vanished entirely for urban women. The loss of a positive association between LFP and the presence of elderly household support is likely another explanation for young women’s declining LFP rates, signaling an even more urgent need for child care support from outside the household. The primary research further substantiates the hypothesis that being a wife and mother of small children restricts women’s workforce participation. Among women from households with children, 40 percent indicate that housework and child care lead to their nonparticipation. Among women without children in the household, 29 percent state that burdens surrounding the household gender division of labor keep women out of the labor force (figure O.9). As in the existing literature, the primary research points to substantial gender constraints on social and physical mobility, particularly after marriage, that restrict Sri Lankan women’s employment. These constraints exist on a contin- uum from, for example, the most conservative preferences that women’s paid work (if any) be limited to the physical homestead environs (such as through petty trade or small enterprise work), to preferences for shorter commutes, to restrictions on or lack of support for long-distance commuting and domestic or overseas migration. Figures O.10 and O.11 present responses from the household survey, disaggregated by gender. Respondents were asked about both married and Getting to Work (Overview) 22 Overview Figure O.10  Social Acceptability of Long-Distance Commuting and Migration for Unmarried Men and Women 100 Percent of population age 15 and older 80 60 40 20 0 Male Female Male Female Commute long distances for work Migrate Male respondents Female respondents Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. Figure O.11  Social Acceptability of Long-Distance Commuting and Migration of Married Men and Women Percent of population age 15 and older 80 60 40 20 0 Male Female Male Female Commute long distances for work Migrate Male respondents Female respondents Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. Getting to Work (Overview) Overview 23 unmarried men and women commuting long distances for work or migrating. Three valuable insights emerged to illuminate levels of support for these prac- tices: First, regardless of marital status, men hold more social leeway in commut- ing long distances for work and migrating away from the household among both male and female respondents. Second, as the FGDs reveal, married women’s mobility is more restricted than that of unmarried women. Finally, for all women, migration is more acceptable than long distance com- muting. Other women, such as grandmothers, may be relied upon to carry out household care roles and act as substitutes for migrating mothers, which is more acceptable than for commuting mothers. This substitution appears to be less common for long-distance commuting, where women are still expected to fulfill household roles of preparing meals and caring for children (for example, “dinner on the table and school homework”). As one male FGD member states, “It is better for a woman to carry on with a business or enterprise from her home, rather than being employed away from home. Then she can [fulfill her] dual roles of housewife and income-earner.”14 Another man says, “For women, prox- imity to home is the most important criterion in selecting a job, because they will spend less time on the road, [and have] more time to attend to household chores. Apart from saving on transportation to and from work, there will be fewer wor- ries about their security.”15 Any male or female who lives in a household that is headed by a woman has a significantly lower chance of participating in the labor force than someone living in a male-headed household, except for the female house- hold head herself. This seems a bit paradoxical, except when considering the intense pressure to earn income that the female heads of household may experience: in 2015, their odds of working or seeking work were 12 percent- age points higher than for women who were not household heads—and 13 percentage points higher for women in the rural and estate sectors, but 10 percentage points higher for urban women. These positive associations between LFP and being a female household head were slightly smaller than in 2013 (15 and 11 percentage points for non-urban and urban women, respectively), but appear to be greater than those for female household heads in earlier years. The positive association was smallest in 2006 (about 8 per- cent across all residential sectors), and grew slightly for non-urban women in 2009 (9 percentage points) and then again in 2011 for all women (9 and 11 percentage points for urban and non-urban women, respectively). The more recent years of LFS data (2011–15) evidently are picking up the grow- ing number of female household heads who, since the end of the civil con- flict, have little choice but to enter the labor force to provide for their vulnerable households. Unfortunately, greater odds of LFP do not translate into enhanced likelihood of employment or increased earnings for these women. In 2015, being a female household head was associated with higher unemployment rates, lower odds of becoming a paid employee, and lower earnings than for all other groups on the island. These findings suggest a deterioration in labor market conditions for female household heads. Getting to Work (Overview) 24 Overview For earlier rounds of LFS data, being a female head had no significant associa- tion with odds of unemployment, but now women who head households are having to wait longer—or indefinitely—for employment, compared with others in the labor force. On a more positive note, the extreme economic pressure on women heads of household is easing, at least based on the size of the penalty on securing work as a paid employee. Being a female head lowered the chances of becoming a paid employee by about 10 percentage points in 2015, which is less than the 12 percentage point penalty in 2013 and the 17 percentage point pen- alty in 2011, though it had no significant effect on these chances in 2009. Analysis of 2006 data suggests a 12 percentage point penalty, although the 2006 data did not cover the Eastern and Northern Provinces. The findings for more recent years could indicate that women who head households are increasingly looking for opportunities for self-employment to earn income. Multivariate analysis results suggest that those who belong to a female- headed household (FHH) are as vulnerable as the female heads themselves, with their vulnerabilities appearing to increase over time. For LFP, this is true for both men and women, though more so for urban women than for any other group; belonging to an FHH lowers urban women’s chances of joining the labor market by more than 8 percentage points. More than any other household or individual characteristic (except for being married), being a member of an FHH substantially lowers a woman’s odds of becoming a paid employee—by 12 percentage points in 2015, 30 percentage points in 2011, and 29 percentage points in 2009. This negative relationship between becoming a paid employee and belonging to an FHH is true for men as well, though to a lesser degree, and it does not become statistically significant until 2011 and later. Finally, men who live in FHHs have higher rates of unemployment than men who do not, especially urban men in FHHs, whose odds of unemployment were 4 percentage points higher than those for other men in 2015. The effect for women is not significant. Together, these results provide a rough sketch of labor-related opportunities and behaviors of those who live in FHHs. According to the 2012–13 HIES, about one-third (or 40,000) of the country’s 1.2 million FHHs lie in the Northern Province, with 20,000 in Jaffna District alone; this is in large part because of the high prevalence of war widows in the conflict-affected northeast, where an esti- mated 90,000 widows reside (United Nations, Sri Lanka 2015). Income-earning opportunities are relatively scarce in this region, which is still being rehabilitated after decades of conflict. Because of the region’s high incidence of extreme pov- erty (World Bank 2015a, 2015b), however, men in FHHs must at least try to obtain work. The high numbers of job seekers relative to available jobs elevates their rates of unemployment. Poor urban men in this region might be especially prone to unemployment because the infrastructure and markets for nonagricul- tural work are not developing as quickly as those for agriculture. The finding that women in FHHs are not prone to unemployment may be capturing the relatively strong sociocultural constraints on women’s job-seeking among the ethnic groups that make up majority populations in the northern (Sri Lankan Tamils) and east- ern (Sri Lankan Moor) parts of the island. This sketch would have to be Getting to Work (Overview) Overview 25 verified—and, even better, definitive details revealed—by targeted research in northeast Sri Lanka. Fortunately, the World Bank is conducting a strategic social assessment of the conflict-affected Northern and Eastern Provinces, the results of which should become available in 2018. Skills Mismatch and Occupational Segregation Educational and occupational streaming continue to create a skills mismatch between women’s human capital attainment and market demand for their paid labor. Men have retained their advantage in obtaining high-skill jobs even though more women are now attending university (and more women are attend- ing than men). The gender gap for those with any university-level education who secure highly skilled employment (for example, as managers, professionals, or legislators) favors men by more than 2 percentage points. University-educated women who beat the relatively high odds of unemployment tend to work in medium-skill jobs instead—more than 5 percentage points more than men who attend university, as of 2015 (figure O.12). These “gender skills gaps”—that is, the gender differences in the way in which education levels translate into jobs of various skill levels—have not changed much. In 2009 (figure O.13), university-educated men had a slightly greater Figure O.12  Gender Differences in Skill Level, by Education, 2015 8 6 4 Percentage point di erence WOMEN 2 0 −2 −4 MEN −6 −8 University A-levels O-levels Grade 6–8 Below No education completed completed grade 6 High skill Medium skill Low skill Source: World Bank calculation based on 2009 Labour Force Survey. Note: All provinces; data from the Northern Province were not collected this year. Employed age 15–64. Occupation was classified according to International Standard Classification of Occupation 2008 (ISCO - 88) definition. Occupations are defined as High skill (managers, legislators, professionals); Medium skill (armed forces, technicians, clerks, service workers, skilled agricultural and fishery workers); Low skill (craft and related trade workers, plant and machine operators, elementary occupations). Getting to Work (Overview) 26 Overview Figure O.13  Gender Differences in Skill Level, by Education, 2009 8 6 4 Percentage point di erence WOMEN 2 0 −2 −4 MEN −6 −8 University A-levels O-levels Grade 6–8 Below No education completed completed grade 6 High skill Medium skill Low skill Source: World Bank calculation based on 2009 Labour Force Survey. Note: All provinces; data from the Northern Province were not collected this year. Employed age 15–64. Occupation was classified according to International Standard Classification of Occupation 2008 (ISCO - 88) definition. Occupations are defined as High skill (managers, legislators, professionals); Medium skill (armed forces, technicians, clerks, service workers, skilled agricultural and fishery workers); Low skill (craft and related trade workers, plant and machine operators, elementary occupations). The 2009 weight factor was adjusted by World Bank projection of total population from the 2012 census. advantage in landing high-skill jobs than in 2015, and university-educated women had about the same advantage as they did in 2015 in obtaining medium- skill jobs. Men with university education also are more likely than same-educated women to work in low-skill employment, and slightly more so in 2015 than in 2009. The only education level at which women have an advantage over men in obtaining high-skill jobs is the completion of A-level exams (leaving school after grade 12), which may reflect their relative practicality. University education is not necessarily required for high-skill jobs in the private sector; moreover, university- educated women prefer jobs in the public sector even though they tend not to occupy the highest-level positions there. Unfortunately, even female A-level graduates’ advantage over men in obtaining high-skill employment has dissipated by about five percentage points (down to less than 2) since 2009. In 2015, female A-level completers’ advantage appears to have shifted to obtaining medium-skill jobs; these women obtain medium-skill jobs at a rate of nearly 5 percentage points higher than male A-level completers, more than doubling their advantage over their male counterparts since 2009. The only reversal in gender gaps regard- ing high-skill employment has occurred among O-level completers: in 2009, women leaving school at this point had about a 1 percentage point edge over men Getting to Work (Overview) Overview 27 with the same educational attainment, whereas these men had a slightly greater edge over comparably schooled women as of 2015. Female O-level completers secure more low-skill jobs than males—even more so in 2015 than before. The gender skills gaps at lower levels of education in 2009 have all but disappeared. Among those with no education, however, small gender disparities that resulted in more women than men in medium- and high-skill jobs in 2009 flipped to men’s advantage by 2015. More uneducated women than men had low-skill jobs in 2015, whereas the reverse was true in 2009. In sum, women at all levels of educational attainment have lost ground to men in securing high-skill jobs. Women with A-level and higher education have shifted instead to medium-skill jobs, while women with O-level and lower schooling are increasingly concentrated in low-skill jobs. Even though the odds of college-educated women joining labor markets have improved over time, these women have seen a concomitant loss in their ability (or willingness) to obtain low-skill jobs—compared with college-educated men. The findings sug- gest a strong aversion to “settling” for low-skill jobs among women who attend university or complete their schooling after A-levels—as opposed to higher- educated men and women with less education, who are more amenable to low- skill employment. Gender skills gaps among those who complete their schooling at O-levels or less seem to be disappearing, except for those with no schooling at all, among whom we see women shifting over time toward low-skill jobs and men toward high- and especially medium-skill jobs. When university-educated women secure jobs at any level, they still receive lower pay and lower returns to higher education than men do. Compared with having no education, university education is significantly associated with wom- en’s much higher odds of LFP (by 21 and 26 percentage points in 2009 and 2015, respectively) but not with men’s (whose LFP odds have not been signifi- cantly associated with university education since 2013, when it improved men’s chances of LFP by 20 percentage points, and women’s by 34 percentage points) according to multivariate analysis of LFS data. Women from the rural and estate sectors who attend university appear to enjoy the greatest boost in LFP odds for all years in the analysis (27 percentage points in 2015 and ranging from 22 to 33 percentage points from 2009 to 2014). Compared with lower levels of educa- tion, university education is associated with the greatest increase in earnings and greater odds of obtaining a job as a paid employee for both men and women, with gender disparities in both shrinking between 2009 and 2015. However, university education is also associated with higher odds of unemployment for women than for men, though the unemployment gender gap is not nearly as pronounced as it is for those who stop schooling after A-levels (Gunatilaka 2013). All lower levels of educational attainment in 2015—even completion of A-levels—are associated with significantly lower LFP rates than those for women with no schooling at all. For men, LFP odds are significantly increased—by 5 percentage points—only for education below grade 6 and are not significantly associated with other education levels other than university attendance. The least-educated women had better LFP chances in 2009, when those with some Getting to Work (Overview) 28 Overview education below grade 6 were no more likely to participate in labor markets than uneducated women—and when those who ended their educations somewhere between grade 7 and A-level completion suffered less of a penalty on LFP chances than they did in 2015. The primary research delves into what might underpin gender gaps in labor market outcomes related to education, skills, and experience. Well-educated women’s losses over time in obtaining high-skill jobs (as opposed to educated men’s gains) is very likely related to educational and occupational specialization among women, who—according to FGDs—exhibit a preference for humanities and arts in their educational training, rather than in technical skills that are better suited to jobs in industry. This trend is also capturing the vast numbers of female university graduates who seek jobs in the public sector, especially government jobs. The final section of this report suggests policy and other interventions to reverse this trend. Young women also are encouraged by parents to study humanities, arts, and biological sciences and—though parents are less invested in the career choices of daughters than sons—to aspire to jobs as doctors and especially as teachers and government workers, given that public sector jobs in teaching and government tend to have regular work hours, maternity leave, and decent benefits packages. The reality of the labor market for public sector jobs, however, is that there are far fewer openings than those seeking them, such that high proportions of female university graduates queue for years awaiting such jobs—thus contributing to the very high rates of unemployment among young (especially educated) women. Young men, on the other hand, are encouraged to pursue majors and careers in engineering, computer science, and medicine, as well as in government. Girls’ and women’s preferences about the level and type of education they acquire is one way in which their aspirations are misaligned with employers’ needs for skilled workers. Another point of misalignment is the difference between the behaviors and “soft” skills that women seek to cultivate, on the one hand, and what employers value in new hires, on the other. Female workers (34 percent) are more likely to believe that work ethics and honesty are valued by employers in new hires, whereas male workers (37 ­ percent) cite industry experience as the most important characteristic that they believe employers seek (figure O.14). What employers, in fact, value most is educational qualifications (45 percent) of new hires (figure O.15). This mismatch between what employers seek in new hires and what workers think is valued by employers could lead to reduced educational investment on the part of workers if they believe that the market will not reward such invest- ment. Surveys from the primary fieldwork also reveal that there is no gender difference in the characteristics (such as education, experience, technical, and soft skills) that employers seek in new hires. In other words, employers report that they expect female workers to have the same educational qualifications for the job that male workers have. Overall, young women also receive less “moral support” and career guid- ance (including parental expectations that they will work and choose a Getting to Work (Overview) Overview 29 Figure O.14  Workers’ Perceptions: Most Important Characteristics Employers Seek in New Hires Accounting skills Age Communication, marketing skills Educational qualification Industry experience Nothing special Physical strength Teamwork Work ethics, honest attitude 0 10 20 30 40 Percent of population age 15 and older Male respondents Female respondents Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. Figure O.15  Employers’ Expectations: Most Important Characteristics of Male and Female Workers Family/personal contact Proximity to workplace Training program qualifications Industriousness Industry experience Educational qualification 0 10 20 30 40 50 Percent of employers Male workers Female workers Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. career) from information providers such as parents, compared with young men. Parents are more willing to leave daughters’ career decisions up to the girls themselves (20 percent and 12 percent for fathers and mothers, respec- tively). Conversely, for sons, only 8 percent of fathers and 9 percent of moth- ers are willing to leave the decision to the boys, reflecting the heavier Getting to Work (Overview) 30 Overview importance that is placed on sons’ career success in the Sri Lankan context. The strength of parental roles in occupational choice does vary by sector and industry: workers in the estate sector and in commercial agriculture do not consider the parental role in occupational choice to be as prominent as do workers in the tourism sector (96 percent of male and 94 percent of female tourism workers). It is not surprising, then, that women also are vastly underrepresented in technical and vocational education and training (TVET), including apprentice- ships, despite having achieved gender parity in formal education years ago. The sample from the primary survey confirms relative parity in formal schooling, but a significant disparity in vocational education and apprenticeship among men and women (figure O.16). Among surveyed households, vocational education for women is much lower compared with that for men in the same age cohort (24 percent vs. 15 percent). Among workers surveyed, 33 percent of men have had apprenticeships compared with just 20 percent of women. The entrenched occupational segregation in Sri Lanka encourages girls to aspire to occupations that do not require a firm technical grounding; thus, they do not enroll in vocational education. This decision later places women at a greater disadvantage compared with men if they aim to advance to higher-skilled and more technical jobs, away from unskilled or clerical positions. The eschewing of vocational training may be more characteristic of urban women, however, given that rural participants in FGDs—both male and female—complain of poor access to market-oriented TVET facilities and job guidance centers, which are scarce in rural areas. Figure O.16  Vocational Education and Apprenticeship 40 Percent of population age 15 and older 30 20 10 0 Vocational education Apprenticeship (workers only) Men Women Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. Getting to Work (Overview) Overview 31 The advantage that TVET provides in improving women’s and men’s employ- ment chances, as well as earnings, however, will be negligible unless that training directly provides the skills that employers need in their workers. The finding in Gunewardena (2015) that TVET, training, and apprenticeships provide no advantage beyond that of general schooling suggests that such programs are not sufficiently aligned with employer demands. This lack of alignment—along with the lack of availability of TVET programs—is addressed in the Conclusion and Recommendations section, below. Analysis of the primary data does not provide any indication that parental perceptions are biased against female children with regard to scholastic ability, aptitude, and interest in the fields of science and engineering. In fact, a large majority of parents believe that children of both genders perform equally well in school and have similar aptitudes in science, technology, engineering, and math- ematics (STEM) subjects. However, nearly half of the respondents surveyed believe that girls with education levels similar to those of boys are not offered the same job opportunities, confirming strong perceptions about gender-based labor market discrimination (figure O.17). It is plausible, then, that occupational streaming could be emanating, at least partly, from parental perceptions of labor market discrimination. Parental expectations about the job market may be trans- ferred to children, resulting in corresponding educational and career choices by boys and girls. Figure O.17  Perceptions of Gender-Based Discrimination in the Job Market Despite Similar Levels of Education Responses to the statement: Girls and boys with similar education are offered similar job opportunities a. All respondents b. Educational level Don’t know or no opinion, 60 Percent of population age 15 and older 14 Agree, 40 38 20 0 n 6 9 el l ity ve 6− io e v s le le ad at er e O- A- uc iv gr ad Un ed w Gr lo No Be Disagree, 48 Agree Disagree Don’t know or no opinion Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers Getting to Work (Overview) 32 Overview Gender Discrimination Using Oaxaca-Blinder decomposition models to explore select labor market outcomes—which consider the influence of individual, household, and other characteristics on gender inequities in these outcomes—confirms the strong presence of gender discrimination that impedes women’s entry into the labor force and reduces their wages. Both active and institutional forms of gender discrimination—which manifest in gender wage gaps, discrimina- tory workplaces, and weaker job networks for women than for men—are associated with reduced FLFP and higher female unemployment. Given such a poor forecast for their performance in the labor market, some women choose never to participate or are induced to withdraw from the market entirely. Labor Force Participation Although the overall gender gap in LFP was 39 percent in 2015, just 6.8 percent of this gap is explained by differences in endowments (for example, education, ethnicity, family characteristics, age and experience, and so forth), while more than 93 percent of the gap remains unexplained.16 It is the unexplained share that can be attributed to bias (discrimination). Family characteristics and the presence of children younger than age five in the household tended to increase the explained portion of the gap in 2015. The presence of those older than age 64 in the household, educational attainment, ethnicity, and location tended to reduce it. The explained portion of the LFP gender gap appears to be shrinking over time—with the unexplained portion increasing—which suggests progressive entrenchment of gender discrimination in determining the gap. The explained portions range from 9.0 percent in 2006 and 7.8 percent in 2009 to 7.4 percent in 2011, 7.7 percent in 2013, and just 6.8 percent in 2015.17 The unexplained portion thus has increased from 91 percent in 2006 to 93.2 percent in 2015. Family characteristics and the presence of children younger than age five in the household consistently increase the explained portion, both before the end of the civil conflict (the 2006 and 2009 LFS samples) and after it (the 2011, 2013, and 2015 samples). The finding that education significantly reduces the explained part of LFP gender gaps only in 2013 and 2015 is indicative of the recent, though growing, influence of gender bias in education as it relates to gender disparities in LFP. It also provides further support to the skills mismatch hypothesis: overall, women’s educational attainments are increasing, but the type of education they acquire is not being rewarded in labor markets. The presence of elderly in the household and location—but not ethnicity— consistently reduce the explained portion over time. It is not until 2011 that ethnicity becomes significantly and negatively associated with LFP gender gaps (lowering the explained portion). This negative association with ethnicity contin- ues in 2015, suggesting that ethnicity-related discrimination has emerged in these gaps in the years following the civil conflict. Another explanation for this emergence could be the inclusion of the Northern and Eastern Provinces in the Getting to Work (Overview) Overview 33 2011, 2013, and 2015 LFS samples; both provinces were excluded in 2006 and the Northern was excluded in 2009. These two explanations are not mutually exclusive. Earnings Oaxaca-Blinder decomposition of log earnings suggests that, controlling for type of employment, women in the labor market are better endowed with characteristics that employers reward, such as education and age or experi- ence, than men but still face lower wages because of gender discrimination. The decomposition, however, finds that in 2015 the gender gap in earnings was about 26 percent; of this, 18 percent is explained by endowments (characteris- tics), while the remaining 82 percent is unexplained. Although having young children has no significant association with the gender wage gap in 2015, other family characteristics, ethnicity, and location significantly increase the gap. For example, in conflict-affected northeastern Sri Lanka—where fewer jobs are available, on average, than in other parts of the country—women who do earn money may have to settle for the lowest-paying jobs. Education, age and expe- rience, and employment category (that is, public sector, private formal sector, or private informal sector) appear to significantly reduce the gap in 2015. Considering changes in the gender earnings gap over time, the good news for women is that the gap itself is shrinking; moreover, the gap is increasingly attributable to endowments (that is, bias in determining these gaps is dimin- ishing over time). Not only is the raw wage gap consistently narrowing, as discussed earlier, but the gap estimated when controlling for men’s and women’s characteristics is narrowing as well—from 33 percent in 2006 and the 30 percent in 2009, to 28 percent in 2013 and 26 percent in 2015. Family characteristics, location, and ethnicity tend to increase the explained portion of the gap over time, whereas education, employment category, and age and experience reduce it. The explained portion was only 2 percent in 2009, and jumped to 4 percent, then 8 percent, and finally 18 percent in 2011, 2013, and 2015, respectively. The type of industry and occupation that women choose tends to increase the explained portion of the wage gap. Further decomposition of 2015 log earnings by employment category finds gender wage disparities to be greatest in the pri- vate informal sector (a 61 percent gap) and least in the public sector (7 percent), with the private formal sector gap (40 percent) in between. In 2011 and 2013, the earnings gap itself was again largest in the private informal sector (58 percent and 56 percent in 2011 and 2013, respectively) and smallest in the public sector (14.5 percent and 15 percent, respectively), with the size of the private formal sector gap (37.5 percent and 38 percent) falling in between. In 2006 and 2009, the gaps across employment categories followed the same pattern, although the greatest range occurred in 2006, which shows an 11 percent gap in the public sector and a 62 percent gap in the private informal sector. Despite smaller gender wage gaps in the public sector than in the formal and informal private sectors, the explained portions of these gaps are consistently Getting to Work (Overview) 34 Overview smallest in public sector employment. Its explained portion was as small as 3 percent in 2011 and 14 percent in 2009, though it increased to 22 percent in 2013. The private formal sector consistently has the largest explained portions of gender wage gaps, from a high of 52 percent in 2006 to its lowest share yet (41 percent) in 2015. Gender wage gaps thus appear to be more attributable to gender discrimina- tion in the public sector than in the private sector, even though the raw gaps are smaller in the public sector—which is likely another reason why women are attracted to public sector jobs, as shown by the primary research and existing literature. Even though women are increasing their educational attainments over time—and are thus becoming better qualified than men for public sector jobs— women’s wages in the public sector are not reflecting their higher attainments and qualifications. If they were, the gender wage gap would most likely disap- pear in the public sector, or even reverse itself to favor women. If Sri Lankan women were to look more to the private sector—especially the formal private sector—for employment, the chance that their endowments, rather than gender bias, would determine their earnings would improve. Endowments such as fam- ily characteristics, location, age, and years of experience tend to significantly increase the explained portion of the gender wage gap in the private sector. In the public sector, however, age and experience tend to reduce the explained portion of the wage gap, whereas ethnicity tends to increase it. These findings shed further light on the causes of women’s preference for public sector jobs, in spite of the drawbacks associated with seeking to acquire these jobs. Although the educated women who queue for public sector jobs risk unemployment, they are making rational calculations in prioritizing these jobs—not only because of the women-friendly benefits provided by the public sector, but also because gender wage gaps there technically are smaller than in the private sector. However, ultimately they would fare better by seeking jobs in the private sector—especially the formal private sector—in both earnings potential and chances of employment when joining the labor market. If they were to acquire the “right” kinds of education and skills—that is, those that employers value and those that tend to be acquired by men—these endow- ments would be rewarded with higher earnings for women, reducing the gender wage gap in the private sector. Multifaceted interventions are needed to redirect them toward private sector jobs and, hopefully, to break through the boundaries of occupational segregation in the economy. Role of Employers Occupational segregation results not only from “push factors” in labor supply, such as social norms that value women’s roles as wives and mothers over their role in the workplace, parents’ higher career aspirations for sons than for daugh- ters, preferences for jobs that provide maternity leave but relatively low pay, and perceptions that employers discriminate between men and women with the same educational attainments (figure O.17); it is driven also by “pull factors” of labor demand. Getting to Work (Overview) Overview 35 The FGD findings from the primary field research uncover gender-biased values and norms held by employers that play out in their decisions about hir- ing, promotions, and salary. Explicit forms of active gender discrimination are less detectable in the primary survey results. However, several norms and beliefs prevailing among workers and employers regarding hiring, promotion, and salary conspire against women workers by affecting the career growth of already- employed female workers, on the one hand, and by discouraging new female entrants on the other. Survey results indicate employers’ clear preferences for men over women in hiring at all levels. This difference is particularly pronounced at the managerial and skilled-worker levels, which confirms the existence of a glass ceiling for women workers in the country (figure O.18). Employer preferences for hiring men are not necessarily perceived by employees. About one-third of managers believe that employers prefer men, while the majority of workers at all levels think that their employers do not prefer one sex over the other in hiring, promotion, or salary increases. Because the formal pri- vate sector is more regulated than the informal private sector, the former tends to provide benefits of particular interest to women, such as maternity leave, flexible working hours, and transportation facilities. Gender discrimination may also be detectable from labor market tight- ness; nearly 60 percent of employers report difficulty in finding new workers, reflecting excess demand for labor (figure O.19). In situations of labor mar- ket tightness, economic theory predicts that employers will hire new workers from the reserve of unemployed or inactive workers—if employers believe that potential workers have the necessary skills for the job—to avoid rising wage costs. As the employer responses demonstrate, even in the context of Figure O.18  Employers’ Preference in Hiring at Different Levels Hiring Managerial workers Promotion Salary increase Hiring workers Skilled Promotion Salary increase Hiring Unskilled workers Promotion Salary increase 0 20 40 60 80 100 Percent of population age 15 and older Men Women Either sex No response Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. Getting to Work (Overview) 36 Overview Figure O.19  Labor Market Tightness and Hiring of Female Workers a. Employer difficulty in b. Employers planning to increase the hiring new workers number of female workers Some difficulty, Much Yes, 41 difficulty, 46 19 No, 53 Not much difficulty, No response, 40 1 Source: Primary data collected in 2012. Note: The data were collected in Gampaha, Badulla, and Trincomalee districts. Sample size: 556 households and 157 employers. labor market tightness, a significant proportion of employers (53 percent) are not willing to hire women (figure O.19). Employers’ perceptions about whether women or men are better workers differ by sector. Employers in the tea and tourism sectors deem women to be better workers than men, whereas the reverse is true in the ICT and com- mercial agriculture sectors. Employers’ views on the performance of workers by gender generally mirror whether they plan to hire more women. For example, tea sector employers report planning to increase the number of female workers, whereas those in ICT and commercial agriculture sectors do not. The latter two sectors, which are two of Sri Lanka’s most promising economic drivers and also have the potential for high earnings, appear to be biased against the hiring of women. Ethnicity As with any discussion of discrimination and—more broadly—social inclu- sion, this study must consider the intersection of gender with multiple axes of exclusion. In previous sections, the discussion centers on how gender inter- sects with age, education, marriage, and other individual as well as household- level characteristics. Another important consideration is the relationship between gender and ethnicity, and how it plays out in labor markets. The Oaxaca-Blinder decomposition models suggest that ethnicity18 has had Getting to Work (Overview) Overview 37 a statistically significant—though small—association with the LFP gender gap in that it has tended to reduce the explained part of the gap since 2011, though not before then. As for earnings gaps, however, ethnicity tends to increase the explained part of the gap (with a significant association in 2009, 2011, 2015). In other words, after the civil conflict ended, more ethnic minorities in conflict-affected areas (for example, Sri Lankan Tamils) entered the workforce—or already had been in the workforce but were included in national survey rounds only after the conflict. Because of the scarcity of well- paying jobs in these areas, however, the (mostly impoverished) ethnic minor- ities inhabiting them must accept low-paying jobs. Women in the labor market there thus face the double burden of being female and belonging to an ethnic minority. Among the different ethnic groups of women in Sri Lanka, Indian Tamil women tend to fare best in labor market outcomes, except for earnings. Across years (2006–15)19 and compared with women from all other ethnic groups (including the Sinhalese majority), Indian Tamil women have the greatest odds of participating in the labor force and the best chances of securing a job as a paid employee according to the multivariate analysis. (For the latter outcome, however, in more recent years they have not retained the great advantage they had in 2006 and especially 2009, when they were 72 percent more likely than Sinhala women to become a paid employee, and more than twice as likely if they lived in the urban sector. As of 2015, they were 37 percent more likely than Sinhala women to become a paid employee). Indian Tamil women also are less likely to be unemployed than other women. By contrast, women in Sri Lanka’s Moor ethnic group fare the worst in labor markets. They are the least likely to participate, with 23 percentage point lower odds of participation than Sinhala women and nearly 30 percentage point lower odds than Indian Tamil women. (It is interesting to note that Malay women tended to fare almost as badly in LFP as Moor women up until 2013, after which Malay women pulled ahead.) Since 2011, Moor women have been the least likely to find work as paid employees; moreover, they have tended to have the highest rates of unemployment, except in 2015 when Burgher women’s unemployment rates surpassed those of all other women. Indian Tamil women’s high rates of LFP and working as paid employees are not surprising, given that Indian Tamils are predominant in tea and other estates. Unfortunately, their extreme disadvantage in earnings—relative to women from other ethnic groups—also is not surprising, given the poor work- ing conditions in the estate sector. Indian Tamil women now earn significantly lower wages than any other group of women, even when considering only those women who live in the estate and rural sectors. This suggests that Indian Tamil women’s relative wages have deteriorated in recent years; in 2006 and 2009 their wages were not significantly different from those of other women. That Tamil men also tend to have significantly lower earnings and lower LFP odds than other men (though not the case in 2006 for LFP) underscores the dire situation for estate sector residents in general. These findings are in line Getting to Work (Overview) 38 Overview with the estate sector’s recent identification (World Bank 2016) as one of Sri Lanka’s remaining pockets of poverty. IV. Conclusion and Recommendations A range of factors contribute to women’s low LFP and persistent gender wage gaps in Sri Lanka. Among them are prevailing gender roles developed from early childhood that steer women toward primarily domestic identities within the household rather than toward well-remunerated jobs. Social pro- hibitions against women working after marriage—and especially after the birth of children—remain strong. Commuting to jobs, as well as employment in certain sectors (such as the garment industry), retains a stigma, especially for married women. As is common elsewhere in the world, Sri Lanka exhibits considerable educa- tional and occupational streaming by gender, resulting in women’s low rates of acquiring the skills valued in labor markets. The human capital mismatch between women’s education and training, on the one hand, and skills that employers seek in growth sectors of the economy, on the other, disadvantages women even more now than before the end of the civil war. Although higher education is even more strongly associated with increased chances of FLFP than it was a decade ago, it also is associated with a decline in women’s acquisition of medium-skill and especially high-skill jobs. In other words, the gender gap in education as an avenue to obtaining higher-skill jobs expanded between 2009 and 2015 to increasingly favor men. Women face the added disadvantage that their perceptions of what employers seek—in soft skills and behavioral charac- teristics, as well as in technical skills—are quite different from employers’ actual expectations. Men appear to be gaining on the earnings front as well. Although more highly educated women in 2009 enjoyed greater earnings returns to education (com- pared with men who completed A-levels or attended university), in 2015 men with university schooling had pulled ahead of university-educated women in their earnings potential. Women’s university attendance is associated with a boost in earnings potential of about 50 percent (compared with women with no education) whereas men’s university attendance is associated with a 65 percent boost. This gender gap has flipped in favor of men since 2011, when university education improved earnings for women by more than 65 percent and for men by 50 percent. Finally, gender discrimination continues to lower women’s odds of participa- tion as well as their earnings relative to men’s. The role of gender discrimina- tion in LFP gender gaps is increasing over time, but for gender wage gaps it is slightly declining, especially in the private sector. Primary data confirm the persistent influence of gendered social norms, beliefs, and behaviors on labor market outcomes for women, expanding LFP gaps. These norms—which per- meate the attitudes and beliefs of employers as well as workers, their families, and their communities—play out in various forms of statistical and institutional Getting to Work (Overview) Overview 39 discrimination. They constrain employers’ views as to who makes a good employee and manager, with the prevailing image of a manager being a man from the ethnic majority. Statistical discrimination on the part of employers not only remains a barrier to the hiring of women, but it also combines with institutional discrimination to undermine women’s safety and security in work- places and on the forms of transportation that get them to work. Overcoming gender gaps in labor market outcomes will require attention to the following: (1) reducing tradeoffs between women’s household and market roles; (2) addressing occupational segregation, particularly through investment in human capital and skills valued by labor markets; and (3) establishing an enabling environment for gender equality in the workplace—particularly in the formal private sector—to help draw women into private sector jobs and away from public sector positions that are increasingly scarce relative to the number of women seeking them. Considering changes over time, which particular groups are gaining and which are losing ground in Sri Lanka’s labor markets? The study’s findings underscore the need to differentiate between women from diverse age, educa- tional, ethnic, regional, and family backgrounds. Disaggregated analysis is impera- tive so that program and policy responses can be customized to effectively address the different needs of different groups. Men have maintained their advantage over women in labor market outcomes since the civil war—though most gender gaps have narrowed, except for higher education’s rising boost to men’s earnings and acquisition of high-skill jobs, with each year. Although urban women have seen a slight bump (1 to 2 percentage points) in LFP rates in the past decade, the LFP gender gap remains the greatest in urban areas when considering the polarizing effect of marriage: getting married is associated with 11 percentage point lower odds of LFP for urban women and 10.3 percentage point higher odds for men—a gap of more than 21 percentage points. Urban women also have lost their relative advantage over rural and estate women with regard to household variables: Having young children now depresses their chances of LFP to about the same degree as it does for other women. Moreover, the boost that elderly in the household used to provide to urban women’s LFP chances has now disappeared. The poorest and least-educated women, regardless of residential status, are falling further behind higher-educated and wealthier women in LFP chances and access to higher-skill, higher-wage jobs according to the quantitative analy- sis. School dropouts—including women who have completed only O-levels— need special attention. They require the concerted efforts of general education and TVET programs, employers, and their own families to help them succeed in the school-to-work transition. The fact that girls age 15–16 years have the highest unemployment rates in Sri Lanka (according to 2015 LFS data) is unprecedented and is cause for great concern—primarily because it signals climbing dropout rates among girls around O-levels (and, it follows, declining rates of high school completion among girls). This concerning trend may reflect an increasing need among the poorest households for as many family members as possible to Getting to Work (Overview) 40 Overview generate income—even at the expense of their education. Also concerning is the stubborn, decade-old trend that women who leave school after completing O-levels have significantly lower LFP odds than women of all other levels of education—including women with no schooling at all. O-level-educated women also are significantly less likely than all other women to secure jobs as paid employees. Young men just a couple years older (age 18) who complete their educations around grade 12 consistently have the highest unemployment and the lowest LFP rates of all men, although as of 2015 uneducated men have the low- est participation rates. These young men, too, need enhanced support in the school-to-work transition. The need to target women from already-vulnerable groups seems increasingly urgent. Women who live in the estate sector, in the most conflict-affected areas in Northern and Eastern Provinces, and in other “pockets of poverty”20 in Sri Lanka should be prioritized in decisions about intervention locations because their positions in labor markets are continuing to deteriorate. Not only do these locations need a greater presence of financial institutions (ideally, formal financial institutions, as opposed to just microfinance institutions) and TVET providers, but the TVET programs themselves also need to be much better aligned with employers’ needs for skills training, especially in local job markets, and they need to provide—or be linked to—providers of business development training. On the demand side of labor, these job markets must be developed to better tap into local comparative advantages, such as the availability of arable land and water sources for agriculture and livestock production or nonfarm rural enterprises; proximity to coastal hubs for creation or expansion of trade routes and connec- tivity to trade partners in other South Asian countries and beyond; and, in the case of the estate sector, a large pool of women with experience in agricultural production. The economic opportunities for women in rural or otherwise remote regions of Sri Lanka would improve vastly if their mobility were enhanced. The easing of mobility constraints is a key factor in “getting women to work.” Providing women with safe and comfortable transportation (see annex OB) to jobs or job training programs would help increase social acceptance of their travel outside the household to access job opportunities. Women in the estate sector, despite having lower unemployment and higher LFP than other women, have been losing ground in LFP: their LFP rates have declined by nearly 10 percentage points since 2006. They face not only the great- est gender wage gaps across residential areas, but significant hurdles in upward job mobility as well—both toward higher-skill and managerial jobs in estate work and, even more important, in making the transition to more remunerative employment outside the estates. Their growing disadvantage in earnings—­ compared with other women and with men in the estate sector—is reflected in the deteriorating state of relative wages for women of the Indian Tamil ethnic group, who make up the majority population in the estates. Some estates have been experimenting with new middle-management models that place women in team leader positions over groups of workers, granting these women higher wages and greater decision-making responsibility. Given the ground that tea estates are Getting to Work (Overview) Overview 41 losing in Sri Lanka’s economy, however, interventions will need to offer alterna- tive, sustainable livelihoods to these women. Such interventions could take the form of program responses—for example, expansion of already-successful rural livelihoods projects that organize women into savings groups, producer groups, and trader groups and provide them with skills and business development training—and policy responses, such as reform of laws and practices around land ­ and other asset inheritance and ownership, which women could use for growing crops and livestock or as collateral for loans to create and grow enterprises. The strikingly poor labor market outcomes of Sri Lankan Moor women— even when compared with women of other ethnic minorities—can be explained by Moors’ predominance in the eastern provinces (some of which were sites of armed conflict) only up to a point. Moor women’s low LFP rates and poor odds of working as paid employees might indicate their own preferences (or family pressure) to opt out of labor markets—if it weren’t for their high unemployment rates in rural and estate sectors. Those who actively seek work are hard-pressed to obtain it. That this is also the case for Moor men, who have high LFP but high unemployment rates and low odds of wage work, suggests a pattern of excluding Moors from labor markets. Oaxaca-Blinder decomposition confirms the increas- ing degree—in the years since the conflict—to which discrimination explains variations in LFP rates across ethnic groups, along with the obvious determinants related to endowments (or the lack thereof) such as poverty and the associated low levels of education. To the extent that sociocultural norms restrict Moor women from working outside the homestead, interventions to increase their income-earning possibilities should concentrate on opportunities within the homestead. Such opportunities could include providing microloans and skills training for farming of high-value crops on family land, aquaculture in backyard pools, or (in the case of very small loans) for the purchase and fattening of live- stock for future sale. ICT is another sector that could offer myriad possibilities for in-house paid work (see annex OB for details). War widows and female heads of household, many of whom reside in the conflict-affected northeast, appear to have become increasingly vulnerable in labor markets since the civil war. Their high LFP rates are not translating into sustainable economic opportunities: Though not the case in 2009, since 2011 they have become significantly less likely than other women to become paid employees. Moreover, as of 2015 they had significantly lower earnings than all other women, which was not the case in previous years. Targeted, multifaceted efforts are needed to improve labor market opportunities for women household heads and their family members, who also face growing insecurity in their chances of LFP and obtaining wage employment; men from FHHs have been experiencing significant drops in both since 2011. Land resettlement policies and practices should ensure that the names of female household heads are on land titles. Livelihoods programs should complement women’s enhanced land inheri- tance and ownership with microcredit—or in the most destitute circumstances, cash grants—and training in financial literacy and business development so that they can use their land for productive, remunerative purposes. Getting to Work (Overview) 42 Overview Female overseas migrant returnees, like war widows and women heads of household, are another target group who would greatly benefit from efforts to enhance their use of microfinance, complementary business development train- ing, and land titling reforms to help build women’s assets for use as collateral or other economic leverage. A potential employment sector for returnees—many of whom have worked as domestic servants providing child care in households abroad—could be the sorely needed, high-quality care economy that has yet to develop in Sri Lanka, as in many South Asian countries: less than one-quarter of children in South Asia even benefit from preschool provision (Niethammer 2017). The SLBFE’s institution of more stringent requirements on women seek- ing overseas employment in 2013 was meant to protect migrating women by endowing them with advanced skills that would preclude their need to settle for low-skill jobs such as housekeeping work; however, these policy changes leave a growing population of poor Sri Lankan women in need of domestic, alternative livelihoods. It should be noted that many overseas female migrant returnees come from estate backgrounds, meaning that there may be some overlap in these two target groups. Finally, women working or aiming to work in the public sector also consti- tute a vulnerable group. Although their earnings tend to be higher than those of women working in the informal private sector—and their maternity leave and other benefits are more assured than those for women in the informal and formal private sector—women seeking public sector jobs far outnumber the jobs themselves, which contributes to high unemployment rates among well- educated women. Moreover, it is in the public sector that gender wage gaps are most attributable to bias rather than to worker endowments. For the ben- efit of women and Sri Lanka’s macroeconomic growth, policy makers need to improve incentives for women to seek employment in the private formal sector (see the priority areas in the General Recommendations section, below, for a more detailed discussion of these incentives) and for employers there to hire them. The informal private sector—with its lack of regulation, lack of benefits, and lowest wages among all employment types—is a precarious work environment for women and men alike. This report gives detail and direction to several policy areas that the govern- ment of Sri Lanka is already considering to expand female LFP, such as enhanced skills attainment, greater participation in the private sector, parity between the public and private sectors for maternity leave coverage, provision of child care through early childhood education programs, and transportation and housing services for female workers. Primary research conducted in Badulla, Gampaha, and Trincomalee reveals intervention areas through which constrain- ing factors on women’s LFP, acquisition of market-valued skills, and wage gaps can be strategically addressed. Recommended interventions to improve wom- en’s labor market outcomes in general can be found below. Detailed recommen- dations regarding each private sector industry studied in the three districts can be found in annex OB. Getting to Work (Overview) Overview 43 General Recommendations (Cross Sectoral) Policy recommendations derived from this research are summarized in four pri- ority areas below. They are presented in matrix form in annex OA. PRIORITY AREA ONE: Reduce barriers to women’s participation in paid work, particularly lack of child care services and socio-physical constraints on women’s mobility. Support in this area will require shifting some child and elder care responsibilities away from women who seek outside work, ensuring wom- en’s safe travel to and from places of work, promoting more home-based income-generation opportunities for women, and, in the longer term, helping change attitudes toward women working outside the household, especially after marriage. Priority area one might comprise the following interventions: • Expansion of opportunities for women to access part-time work and maternity leave • Expansion of opportunities for women to generate income in their homes by providing some combination of technical and vocational training, financial lit- eracy and business development training, access to credit or other financial assistance, and links to markets • Expansion of the care economy through (1) public-private partnerships that support a quality-focused accredited system, (2) on-site day care at workplaces (some ICT companies based in Colombo already have incorporated day care centers for workers’ young children, with promising results; moreover, a recent report by the International Finance Corporation (IFC 2017) presents the com- pelling business case, supported by extensive cross-country research, for child care that is supported by employers), or (3) more localized provision of care services that absorb women returning from employment abroad as domestic workers • Improved access to child care services through short-term measures (for example, using Sri Lanka’s expanding program of early childhood develop- ment centers, which can provide child care without the stigma associated with purely child care facilities) and long-term campaigns to remove the stigma (that is, expanding the market for, and acceptability of, child care services through, for example, media tie-ins to promote images of working women and children in day care) • Improved public transportation safety for women, and partnerships with private employers to encourage firm-specific transportation for female workers • Expanded housing stock for firms’ female workers (through provision of firm incentives), as well as housing in the vicinity of worksites—for female internal migrant workers and other working women—that is leased at ­affordable rates Getting to Work (Overview) 44 Overview PRIORITY AREA TWO: Strengthen girls’ early orientation to career development and to acquiring the types of education and skills that prepare them for labor markets. One of the most effective ways to improve social acceptance of women’s employment (especially outside the home) is to start early in the life cycle and reinforce messages at each stage of girls’ education and development. Related interventions could include the following: • Provide support for girls’ career identification and development—in the early years at home and in the community—through community campaigns and outreach to parents and teachers regarding girls’ skills acquisition and employment • Incorporate in campaigns an emphasis on nontraditional fields for girls and women (with particular attention to STEM courses and a special emphasis on computer science) in the general education system and raise awareness about the benefits of TVET courses that build skills valued in the growth sectors of Sri Lanka’s economy • Provide in-school mentoring of girls and young women in general education (well before O-levels, to help mitigate dropout around grade 10), higher edu- cation, and TVET programs to help them identify and refine their knowledge of career paths in the private sector • Link adolescent girls and young women to successful adult role models in desired careers who can provide examples of balancing work with marriage and raising children PRIORITY AREA THREE: (1) Improve the jobs orientation of education and skills providers and (2) expand provision of job-matching services and TVET that target women and respond to employers’ needs. Improving the jobs focus and the provision of vocational and employment ser- vices in educational and community settings would help reduce educational and occupational streaming by gender. Such efforts will also bolster the general rec- ommendation to expand active labor market policies in Sri Lanka focused on women through skills development as well as enhancement of job information services and job matching. Specific interventions could include the following: • Provide mentoring, peer group support, and internships • Expand industry-linked internships and school-based business incubator and exposure programs for female students at the lower-secondary school level • Improve the jobs orientation of educational institutions and curricula; the ­ government might even consider making computer education mandatory for girls and boys beginning in middle school, as this would ensure that market- able skills are acquired by all youth in Sri Lanka Getting to Work (Overview) Overview 45 • Enhance job information and job placement services in school and community settings, with an increased emphasis on job centers in rural areas • Expand TVET programs and strengthen the National Vocational Qualification system, with island-wide accreditation of providers (the government’s efforts to expand the National Vocational Qualification system and link to rural-based providers should ease some of the constraints experienced by rural women in accessing TVET) • Ensure the presence of accommodation facilities for girls and women at voca- tional and technical training colleges and other TVET providers • Provide one-stop-shop job centers that coordinate job vacancy and TVET information and training services, while providing post-training assistance such as peer group counseling for adult job seekers • Strengthen the relevant skill sets of women who are interested in obtaining high-skill, highly paid foreign employment by providing language skills and scholarships along with skills training PRIORITY AREA FOUR: (1) Ensure gender equity in labor legislation and nondiscriminatory workplace environments and (2) undertake affirmative action and ethical branding initiatives to expand women’s share of employment and firm ownership in emerging sectors. Improving women’s LFP in Sri Lanka also requires enhanced review and enforce- ment of labor laws, antiharassment measures, and maternity leave—especially related to leave provisions in the private sector. Recommended activities include the following: • Review labor laws for clauses that restrict women’s rights to paid employment • Expand application of maternity leave legislation to include enforcement in the private sector • Enhance safety regulation and labor monitoring audits of the workplace and sector-specific approaches such as promotion of child-safe tourism Direct support for women’s expanded employment and firm ownership in emerging sectors could include the following areas: • Provide subsidies and tax incentives to firms that support women’s employ- ment (for example, that hire and retain new female graduates or that partici- pate in vocational training programs for women) • Institute preferential government procurement for goods and services offered by women-owned firms or those with female-majority management • Establish public-private partnerships that contribute to verified certification and branding of firms as gender-responsive • Improve business development and financial services for women’s enterprises Getting to Work (Overview) 46 Overview Annex OA  Summary of Recommended Interventions (Cross Sectoral) Key constraints Range of interventions Household roles and responsibilities Child care burden Child care and job sharing • Ensure provision of accessible and affordable child care (at workplace; in community) • Improve social acceptance of paid child care, for example, through media campaigns and awareness-raising in communities about the benefits of child care • Promote part-time work and job sharing to encourage participation of young mothers in workforce Social constraints on Transportation and housing female mobility • Safe and affordable transportation for women • Housing for female workers at or near workplaces • Business development and financial services for self-employment and small enterprises run by women, especially in rural areas Parents’ and educators’ Life planning and identity formation campaigns roles in girls’ • Community campaigns and outreach to parents and teachers; educational and customized approaches for communities with lower female labor force occupational choices participation (such as certain ethnic minority groups) Human capital Educational and Improved jobs orientation in school and community settings occupational • Provision of mentoring and peer group support for female students, streaming by gender particularly in science, computer science, and other technology, engineering, mathematics, and business-related subjects; compulsory computer education for all secondary school students • Expansion of industry-linked internships; school-based business incubator programs • Support career counseling and placement function of high schools and universities, with trained counselors Low participation by Technical and vocational education and training (TVET) women in employer- • Scholarships for girls to attend TVET in nontraditional fields informed vocational • Coordination with private sector on training and job placement training • Expansion of NVQ-accredited TVET providers in lagging regions, especially private sector providers • Public-private coordination with industry in designing new TVET courses and outreach policy focused on women’s employment Information barriers in Job matching matching jobs and • Job information services via local government; one-stop job centers workers Discrimination Legal discrimination Antidiscrimination legislation • Amend legislation that restricts women’s rights to worka • Legal prosecution of firms that violate women’s civil rights • Support maternity leave; parity for public and private sector guarantees Employer bias against Affirmative action in hiring and procurement hiring and promoting • Outreach to employers from TVET institutes and colleges female workers • Preferential government procurement from female-owned firms Poor workplace facilities Labor monitoring and conditions • Safety regulations, including measures taken to mitigate sexual harassment, and workplace monitoring by Labor Department • Gender certification and ethical branding for firms Source: World Bank. Note: NVQ = National Vocational Qualification. a. As examples, “protective” legislation currently limits the number of night shifts per month and puts constraints on overseas work by mothers of small children. Getting to Work (Overview) Overview 47 Annex OB  Sectoral Recommendations: Findings from Five Private Sector Industries This section draws from secondary sources and the primary research—both qualitative and quantitative—conducted in Gampaha, Badulla, and Trincomalee districts. Surveys, FGDs, and other qualitative methods were used to gather information from household members, employers, and workers in these districts. The research focused on five private sector industries that were strongly repre- sented in the districts and regarded as a mix of emerging and traditional drivers of the economy: ICT, tea estates, tourism, the garment sector, and commercial agriculture. The following discussion summarizes basic information about wom- en’s involvement in these industries—as well as barriers to their entry—and then offers recommendations to increase this involvement. Many of these recommen- dations can be applied to other industries, with the overall intent of achieving much greater participation of women in Sri Lanka’s private sector. Common across the industry recommendations is the need to focus on women’s safety in transportation, which is a crucial factor in raising FLFP rates and actually “getting women to work.” Information and Communication Technology A fast-growing and relatively new industry in Sri Lanka, ICT has not yet been fully gender-typed as “men’s work” and thus, compared with other industries, may have potentially lower barriers to entry for skilled women. With higher education institutions consistently providing less than half the number of ICT graduates needed in the industry, information technology (IT) companies face the challenge of finding suitably trained professionals to fill vacancies (LIRNEasia 2006; Senewiratne 2011, quoted in Morgan 2012). Despite their potential for meeting this demand, women ICT workers represent a meager 0.28 percent of the female workforce. ICT employers, however, do not appear to be implement- ing any formal strategies for attracting or recruiting female employees to the sector (Morgan 2012), nor do they express any intent or need to do so, according to the primary fieldwork. Unfortunately, women’s entry into the sector is restricted by the narrow range of study and training courses considered to be gender appropriate, perceptions of parents and employers (as discussed earlier in this report), and limited English language skills (Jayaweera et al. 2007). LFS data confirm this pattern: those few women working in ICT are already becoming locked into jobs associated with lower-skill and lower-paying “women’s work”; these jobs also generally lack decision-making authority. In 2009 and 2015, men outnumbered women in nearly all ICT occupations, with the greatest gender gaps occurring among sci- ence and engineering professionals and technicians. Although some of these women are in lower-skilled jobs because of their limited technical education, the majority occupy lower-level positions than their male counterparts despite com- parable education and skills attainment (Morgan 2012). The presumed value addition occurring within the industry is concentrated at occupational tiers Getting to Work (Overview) 48 Overview within which women are not represented. Consequently, studies have found gender differences across all salary bands among ICT professionals in Sri Lanka, with a larger percentage of women than men in the lower pay bands (Jayaweera and Wanasundera 2006, quoted in Jayaweera et al. 2007; Gamage 2004, quoted in Morgan 2012). Women are also starkly absent in ICT decision-making struc- tures, including boards, senior management levels of private firms, and ministries responsible for ICT policy and regulatory institutions (DCS 2016a; Jayaweera and Wanasundera 2006). Of the five sectors studied, ICT and commercial agriculture offer the fewest worker benefits that support women. In fact, several studies have identified that ICT employers prefer male applicants because they are reluctant to invest in “extra facilities” for women employees to accommodate social norms around caretaking, mobility constraints, and the like (Morgan 2012; Ranasinghe 2004). According to the primary qualitative research, the workflow of the software industry, which requires long hours of work, makes it especially hard for women with young children to seek or retain employment in the sector. One young woman from Gampaha in the ICT study site states, “Women who are mothers of small children face a difficult situation in having to choose between employment and caring for their child. Often home responsibilities take precedence.”21 These concerning trends in the ICT industry reflect increasing barriers to women’s entry and promotion, but can be reversed by taking a number of simultaneous actions. • Create an enabling policy environment for increasing women’s participation in the sector. Encourage the government of Sri Lanka and IT companies to acknowledge the low share of females in the ICT sector, and introduce a for- mal strategy for actively recruiting women. Collection of disaggregated data should be a requirement for the planning and policy-making process, and the consideration of gender should be increased throughout all national programs and projects that seek to promote use of IT. Measures should be put in place to ensure adequate representation of women in decision-­ making bodies tasked with developing IT policy and strategy. Organizations with a good understand- ing of the opportunities, needs, and constraints experienced by women in the sector, as well as private sector stakeholders, should be involved in policy and strategy formulation, implementation, monitoring, and evaluation. • Make hiring criteria (education, skills, and experience) for ICT employment clear and transparent. Government and industry leaders need to work together to determine and publicize objective hiring and advancement stan- dards that ensure deserving female applicants and employees are not over- looked in employment and promotion opportunities. Making ICT employers accountable for gender-neutral hiring standards will help minimize gender discrimination. • Introduce, formalize, and publicize worker benefits that support women, such as on-site child care and flexible, part-time work options. The lack of a Getting to Work (Overview) Overview 49 gender-focused policy among ICT employers increases their reliance on informal or piecemeal mechanisms that may not be sufficient to transform male-biased employment dynamics. FLFP in the industry would likely increase were ICT employers to introduce set schedules (as opposed to the current norm of fluctuating shift work), on-site child care (which, according to research, contributes to reduced rates of sick leave, absenteeism, and turn- over among employees [IFC 2017]), and flexible work hours, which would give women the option of working when children are at school—on a part- time basis—or in the evenings, as long as safe transportation to and from the work site is ensured. These interventions would allow women who want to work to plan their employment around their domestic and other responsibilities. • Increase female enrollment in ICT technical education by scaling up pro- motion of STEM courses to girls and their parents. In fact, introducing computer education as a compulsory subject in secondary school, or even earlier, may be the fastest route to ensuring that girls as well as boys are acquiring these important technical skills for their futures. This is a mea- sure that a range of governments—including in India and several European and South American countries—have either considered instituting or fully instituted (Reddy 2015). Short of taking this measure, governments have less costly options as well. Female-targeted awareness-raising campaigns could be embedded at the basic education level. Providing STEM mentor- ing and peer support would also encourage girls to pursue careers in ICT over traditional “female” occupational streams. A critical companion to school-based ICT awareness campaigns would be widespread efforts to dis- card the gendered presentation of roles, duties, and employment as female or male. Students should learn about all existing options in the job market, and be encouraged by teachers, school administrators, and educational poli- cies to pursue further studies and choose careers according to their inter- ests, abilities, and labor market demand. Vocational training programs and universities, in particular, will need to improve their marketing of ICT- related courses to female trainees and students. • Expand industry-linked internships and school-based business incubator and exposure programs for female students at the secondary school level. After- school and extracurricular activities can be brought to bear to enhance girls’ familiarity with skills development and jobs.22 Expanding internships and work experience opportunities for secondary students may also deepen students’ on-the-job experience, clarify career goals, and make graduates more attractive to future employers in ICT and other growth industries. This effort could include attention to a range of technical opportunities and exposure to the government’s “sectors of focus,” such as tourism and construction.23 Getting to Work (Overview) 50 Overview • Provide career counseling and job market information in higher education. A number of school-based interventions can be provided to promote increased FLFP in ICT, including internship placements with ICT employ- ers and mentoring relationships with higher-skilled workers and managers (especially female ones). Mentoring and peer group support for girls in STEM areas as well as in business can reduce gender streaming in education and occupational choice, so that women increasingly pursue traditional “male” careers such as engineering, ICT, and business.24 At the university level, Sri Lanka has experimented with supporting arts graduates in setting up their own business enterprises as entrepreneurs—and helping students plan for related coursework—while they are still enrolled. Success rates (including persistence of new firms over time) should be evaluated; and, if warranted, the intervention should be replicated, particularly for students with O-level or A-level education. Schools can consider partnerships with nongovernmental organizations or private sector TVET providers to increase career exposure for youth (especially young women and school leavers) to expose them to a wider range of employment opportunities in a number of sectors. Such enhanced recruitment and retention of female technology professionals in ICT would help the sector reach its growth targets (GoSL 2012). • Establish special on-the-job training and mentoring opportunities for female ICT workers to enable their rise into leadership roles. Formalized leadership and technical training as well as mentorship programs with managers (espe- cially female ones) can encourage female ICT employees to stay in the indus- try and seek advancement while providing them with clearer paths to vertical mobility. This approach is especially important given evidence of employers “guiding” women into particular job roles on the “softer” side of computing, such as technical writing, quality assurance, and training, even when those women express an interest and desire to work in more technical areas such as programming (Morgan 2012). Such factors also contribute to wage differences between men and women in ICT industries. Increasing women’s access to training in more technical aspects of ICT work can help remedy female employees’ overwhelming presence in low-skill jobs and low pay bands (Ranasinghe 2004). • Encourage education and vocational training institutions to invite input from ICT employers when developing education and training curricula. This input can equip young women and men with a better sense of both the technical and soft skills sought by employers as well as help prepare them for working in the often coeducational environment of the ICT workplace. Getting to Work (Overview) Overview 51 Tea Estates Although women in the estates have higher odds of working than women else- where, the poor quality of their work, low wages, and declining interest in this work among the younger generation signal the need for intervention either to improve their working conditions or to provide them with alternative liveli- hoods outside of estate work. In collaboration with the United Nations Development Programme, the Ministry of Hill Country New Villages, Infrastructure and Community Development has developed a Ten Year National Plan of Action for the Social Development of the Plantation Community 2016–2025. The document presents a comprehensive strategy for improving working condi- tions and living standards for women as well as men in Sri Lanka’s estate com- munity. With its prioritization of development interventions and detailed action plans that address successive life-cycle stages starting from early childhood development, the plan has little need for improvement. This discussion shares a few observations from the qualitative field research that are relevant to improve- ments in women’s work and living conditions in the sector—particularly those observations concerning the quality of their work, wages, and opportunities for advancement. The qualitative research finds that women in the estate sector tend to hold more nuanced views than other women (and men) regarding the social desir- ability of women’s LFP. Respondents in the primary research point out that women have no choice but to enter the workforce because men on the estate cannot be relied upon to earn for the household. Thus, the social ideal that women should not work still holds in the estate sector but may not be adhered to because of economic necessity. • Build upon the strong presence of women in the plantation sector to obtain “quick wins” through promotions of women into management positions and use of role models. The fact that social biases against women’s work have broken down is a solid starting point for improving labor market out- comes for women in the estate sector, as is the sector’s tendency to offer more flexible working conditions than most other work. Tea sector employ- ers’ reported plans to increase the number of female workers could serve as an effective entry point for advancing better terms of employment for women. Regular jobs in the tea sector, of all sectors studied, also offer the highest number of employer-provided benefits (such as maternity leave and child care) and facilities (dedicated transportation and separate toilets) for women. • Adopt new practices that promote women’s technical training and access to higher-skill work, group leader positions, and even middle- to upper-level management jobs that are associated with greater pay and enhanced ­decision-making responsibilities. Where the tea estates fail women at work is in the types of jobs to which most are confined: low skill and low pay, with Getting to Work (Overview) 52 Overview little to no decision-making responsibility and scarce opportunities for advancement. The primary fieldwork suggests that women’s representation in management is lowest in the tea sector (3 percent) among all sectors studied. The Ten Year National Plan of Action for the Social Development of the Plantation Community 2016–2025 covers these issues, recognizing that women need to be better represented at all levels of decision-making author- ity. Technical skills upgrading is aligned with the government’s objective to modernize and improve the efficiency, competitiveness, profitability, employment creation, and social, economic, and environmental sustainabil- ity of the tea plantation estate sector. • Increase awareness of sexual and gender-based violence (SGBV) redress mechanisms and improve the ability of these mechanisms to respond to women workers in the estate sector. Collaborative efforts between the National Committee of Women, the Sri Lankan police, and nongovernmental organiza- tions have made substantial progress in developing programs to address and respond to SGBV in Sri Lanka, including engaging men and boys in prevention campaigns as well as providing timely response, referral, and legal assistance through 24-hour helplines. As acknowledged in the 2016–2020 Policy Framework and National Plan of Action to Address Sexual and Gender-based Violence in Sri Lanka, expanding the outreach of these ongoing initiatives and increasing awareness of response services will be key to the well-being of estate sector workers, who experience relatively higher rates of SGBV compared with national averages (according to the 10-year plan for the plantation community). • Encourage women living in tea estates to pursue work outside the estate. These women may work in the tea sector, combine this work with other employment outside the estate, or not be engaged in paid employment at all, according to the primary research. The estates function as self-con- tained social enclaves in many ways, and residents view the estates as their home, even if they migrate out for work (see also Gunetileke, Kuruppu, and Goonasekera 2008). Tensions arise, however, in the social stigma that estate residents face; and, according to the FGDs, many young residents put off estate work because they fear such experience will prejudice their chances of employment outside the estate. Residents are increasingly seek- ing non-estate opportunities, including in small-scale horticulture as well as wage employment. In particular, women from the estates need support for livelihoods diversification. Such support could include approaches used by successful rural livelihoods programs that (1) organize women into savings groups (such as the World Bank–supported rural livelihoods projects across South Asia that have so successfully helped women collectively save and use funds), producer groups, and trade groups for marketing and sale of horticulture or other products and (2) provide them with technical and business-development skills training—especially if these women have Getting to Work (Overview) Overview 53 access to credit programs. Ideally, programmatic interventions would be accompanied by policy interventions that improve women’s ability to inherit, own, purchase, rent out, and sell land—or use it to generate income in other ways. Tourism Unlike the tea estate sector, where representation of women is higher than aver- age, the share of female employees is much lower than average in the tourism industry, a promising growth sector in Sri Lanka. Primary among these con- straints is the perceived—and real—vulnerability of tourism to potential labor abuses and criminal activity, including the sexual abuse and trafficking of women and children. Sadly, research by the United Nations Children’s Fund and the International Labour Organization has identified Sri Lanka as one of the five countries in the world with the highest prevalence of child prostitution (Iaccino 2014). The survey results in the primary research confirm this finding: women themselves, as well as their parents and husbands, fear for their safety in tourism, which tends to inhibit them from considering this sector for work opportunities. Among the employers interviewed for the primary research, those from the tour- ism sector (85 percent) are most likely to identify marriage as the reason for women’s exit from the workforce. The industry will need to provide greater incentives to potential female employees to help overcome such stigma. Stigmatization of tourism jobs for women may be one factor in the notable lack of female trainees in related job skills programs, where such programs exist. Another factor is the lack of such programs in the country. In Trincomalee district, FGD participants note that, despite demand from the tourism sector for workers, there is a dearth of local skilled workers because of higher school dropout rates and an accompanying lack of TVET options for the local populace. These spatial disparities in TVET service provision particularly affect women, who are less likely than men to enroll in and travel to programs located far from home—either because they are relatively time poor and cash poor or because of safety concerns and social stigma associated with young women traveling long distances. A young female member of the FGD in Trincomalee comments that “Most of the big establish- ments, like tourist hotels, have been established by outsiders, who [bring] skilled workers from outside. Because training facilities are not available locally in Trincomalee, those who wish to acquire skills have to go to Colombo, which costs a lot of money.”25 In spite of high social barriers to women’s participation in the tourism industry—many of which appear to emanate from the labor supply side—the sector has relatively low demand-side barriers to entry (at least in the unskilled and semi-skilled job tiers) and promotion of women. In fact, more than any of the five sectors studied, the tourism industry is seeking a greater share of women as employees and is willing to make efforts to accommodate them. First, among all employers surveyed for this study’s primary research, tourism Getting to Work (Overview) 54 Overview sector employers are most likely to report that women are better, more reliable workers than men. Second, among all five sectors, tourism has the highest rep- resentation of women in management (31 percent). Third, the tourism s ­ector is more likely to offer maternity leave (74 percent) and separate toilets for women (91 percent), compared with other sectors on which respondents reported. Fourth, tourism employers are the most willing to make contributions for child care. In the tourism sector, employers contribute 39 percent of costs compared with other sectors studied, where average employer contributions tend to be 10 percent or less. Tourism industry leaders, as well as policy makers and practitioners, need to take advantage of employers’ relatively strong, posi- tive attitude toward employing women by tackling supply-side constraints on female participation in the sector as well as by scaling up the attractive attri- butes on the demand side. • Address the supply-side constraints that create stigma and fear of the tour- ism industry as a potential employer for women, and publicize these improve- ments. Ensuring greater safety in tourism industry workplaces would require collective and sustained engagement between hotels, catering, and tourism enterprises and their owners to agree on the uniform application of equal opportunity policies and practices (Baum 2013). Firms within the sector’s supply chain, whether a large hotel or a small handicrafts industry supplier, should be required to adopt a zero-tolerance policy against child labor, sexual harassment, and other forms of SGBV. Adopting this measure would include required gender-sensitivity training for all employees and management. Similar training for banking sector employees in Pakistan—through the International Finance Corporation’s Banking on Women Program—has shown measurable improvements in the attitudes and behaviors of the men and women working together. Industry leaders also need to work with law enforce- ment and the legal sector to strengthen safeguards for women employed in the sector (UN Women 2011). Such legal protections would include the establishment and enforcement of laws pertaining to minimum wages, maxi- mum work hours, and equal pay for equal work, as well as safe working condi- tions (Kamal and Woodbury 2016; UN Women 2011). Well-functioning workplace grievance redress mechanisms should be in place and publicized so that women know where and how to report discrimination or harassment (Kamal and Woodbury 2016). Once these protections are in place, leaders also should work with the concerned government agencies and private sector media to increase public knowledge of these improvements and to promote a positive image of women in the tourism industry. These steps could be cou- pled with messages to enhance awareness of women’s important economic role in the tourism sector (UN Women 2011). Awareness-raising campaigns using radio and television, social media, and even community-based in-person transactions can be quite effective in reaching large populations and changing their perceptions. Getting to Work (Overview) Overview 55 • Recognize and meet the transportation needs of women. Law enforcement should be included when addressing another fear of the tourism sector for women: the distances required to travel to the industry’s worksites, especially on public transportation. More than 80 percent of women and girls in Sri Lanka experience sexual harassment while using public transportation. Perera, Gunawardane, and Jayasuriya (2011), in their review of evidence on SGBV in Sri Lanka, find that sexual harassment on buses triggers anxiety among women regarding the use of public transportation. Employers need to ensure that women who commute daily to jobs in tourism are provided with secure and comfortable transportation. Transportation providers, moreover, also will need to be held accountable to a zero-tolerance policy for SGBV of any sort on the transport itself and at transport stops and stations. They also will need to receive training on how to recognize and confront harassment and other SGBV that is being perpetrated by others. If public transportation on an individual basis is not an option, the tourism sector can help organize collective transpor- tation for workers and build residences for women with no means of transpor- tation from home or those who live far from their places of work (Baum 2013; Karkkainen 2011). • Take advantage of tourism employers’ acceptance of the need for child care services and other benefits and their willingness to help pay for them. Policy makers and practitioners can build on employers’ amenability by helping orga- nize and certify child care providers to operate in or near sites of tourism activity, or in sites near workers’ residences. Introducing tax breaks for tourism enter- prises that take more responsibility for the welfare of their employees—such as by providing free or low-cost child care—is one means of providing incentives to hotels and other firms operating in the tourism industry to hire more women (Karkkainen 2011; Ferguson 2011; Baum 2013). Similarly, implementation of maternity leave requirements, work-from-home options, and flexible workplace hours would also help draw more women into tourism jobs. • Encourage entrepreneurship of women in the tourism sector. Entrepreneurship may be another attractive option for women with child care responsibilities— self-employment may allow them to work out of their homes or at least close to home. Facilitating mechanisms include the improvement of women’s access to land, other property, and credit as well as training in how to responsibly manage funds and create (and sustain) enterprises with them (UN Women 2011). Ecotourism, in particular, can be a very hospitable subsector for women working in the industry (ADB 2013). • Create greater access to training in skills that are marketable in tourism. If tourism is to absorb large numbers of female workers, as envisaged by national human resources strategies like Sri Lanka’s National Human Resources and Employment Policy, training programs linked to tourism jobs will need to be greatly expanded (so that women in rural and remote areas Getting to Work (Overview) 56 Overview have greater access), as will people’s perceptions of what is acceptable work for women in this sector (ADB 2013). Women often get trapped in jobs that reflect traditional gender roles—such as cooking, cleaning, and handicrafts— and that tend to be highly labor intensive, yet yield products for which mar- kets are saturated. Women need to be provided with training in areas beyond these gender-stereotyped occupations, such as through work-based training programs in hotels. Requiring quotas for women in training would promote incentives for enhanced hotel outreach to women (Karkkainen 2011). UN Women (2011) also suggests creating tertiary-level education and vocational training opportunities oriented toward careers in tourism, while also focusing on improving the education level of women who are already employed in dif- ferent parts of the sector. • Develop a strategy, and involve women in every part of its planning and exe- cution. Multiple actors will need to come together and make concerted efforts to encourage women to fill employment gender gaps in the tourism sector. Chief among these are the concerned ministries in the central government; private sector leaders of the hotel, catering, and tourism industries and other enterprises operating in the tourism sector; those tasked with designing TVET curricula and certifying providers; and those who run job search mechanisms, whether based in social or traditional media or in-person job placement cen- ters, so that trainees can be linked with jobs requiring the skills they have acquired, and employers’ needs can be met through the administration of industry-informed training curricula. Together, these actors will need to ensure that women are represented in all groups and in all stages of the strategy, including development of any tourism legislation and policies and, moreover, in development of action plans that involve targets for women’s participation across all subsectors within the tourism industry (UN Women 2011). Garments As with tourism, garment sector work also appears to carry a stigma for female employees (Ariyarathne et al. 2012). According to the primary research, both tourism and the garment industry are maligned even more by men than women, which implies that fathers and husbands may steer their daughters and wives away from such work; moreover, unmarried women may be less inclined to seek these jobs for fear of jeopardizing their marriage prospects. In spite of the stigma against it, however, more than 70 percent of apparel workers in Sri Lanka are women as of 2015, according to LFS data. With the garment industry rebounding and expanding since the end of the civil war, even more workers are needed to fill labor shortages—and thus there is ample space for an influx of female as well as male workers. Prasanna and Kuruppuge (2013) identified about 30,000 vacancies for women in the garment sector, especially in the export processing zones (EPZs). To attract more women, employers and other stakeholders in the industry will need to address the stigma-related barriers to their entry, as well as improve Getting to Work (Overview) Overview 57 working conditions for women by addressing gender-related skills, pay, and promo- tion gaps in this highly occupationally segregated industry. • Have women who are more highly placed in the garment industry act as role models. A large proportion of garment employment occurs in EPZs, contrib- uting about 4.4 percent of total employment in the country. Typically hired right out of secondary school (whether having graduated or dropped out), women in EPZs are the majority among worker cadres, whereas men hold most supervisory and technical as well as top administrative positions. Although the garment sector employs far more women than men, men his- torically have tended to obtain the higher-paying jobs and rise more quickly to positions with decision-making authority. Until 2009, men occupied the majority of management positions; however, 2015 LFS data reveal a shift in this pattern, with greater representation of women in production and special- ized services management positions. The industry needs to build on these successes in women’s advancement, though they are still far from the norm, by engaging women who have risen to higher positions as role models and as mentors to other female employees and in outreach efforts that target poten- tial female applicants. Industry leaders also should take advantage of the rela- tively high education levels of their female employees (80 percent of garment workers have A-level or O-level educations) to provide them with on-the-job training in higher-skill work as well as greater opportunities for promotion into management positions. • Improve workplace conditions for women and raise awareness about these improvements. The garment sector has previously launched efforts to attract more women employees—including a 55 million rupee (US$490,000) cam- paign in 2011—but these efforts have not yielded the anticipated boost in women workers (Samath 2011). Perhaps this and similar campaigns did not sufficiently reflect improvements made to the industry that would provide women with better incentives to consider working there. In addition to ensur- ing that workplaces are comfortable (that is, having separate restrooms for women) and safe (for example, requiring gender-sensitive training for all man- agement and staff as well as mandating a zero-tolerance policy for sexual harassment), the garment sector can better market itself to potential female employees by advertising its relatively generous maternity leave benefits. According to garment employees in Gampaha, firms allow maternity leave of up to 84 days for their employees, as required by law. Unlike many private sec- tor firms, garment sector firms also tend to comply with other requirements, including provisions for pregnant women (slightly shorter working hours and freedom from hazardous work) and for nursing mothers (permitting daily vis- its and travel time with nursing infants). A Gampaha firm worker says, “Some women opt to [remain] employed after availing their maternity leave, because of the nutrition they can obtain from the factory-supplied meals, which they cannot afford at home.”26 These are formal sector firms, however; Getting to Work (Overview) 58 Overview the vast majority of employment in Sri Lanka remains located in the informal sector, where poorer working conditions can constrain women’s entry and per- sistence in the workforce. Similar requirements should be made of informal sector garment firms. In addition, incentives should be given to garment sector firms—both formal and informal—to share the costs of child care services with workers and to provide flexible work hours for women. Increasing workers’ awareness of their rights would also contribute to holding garment industry employers accountable for providing proper worker benefits and pay and decent working conditions. Finally, improving wages—even marginally—for women workers would further encourage women to seek jobs in the sector. A 1 percent boost in Sri Lanka’s expected garment industry wage would increase the odds of women’s LFP by 89 percent (Lopez-Acevedo and Robertson 2016). • Enforce transparent labor recruitment standards to reduce discrimination against married and older female workers. The participation of married women in the garment sector workforce is reported to be as low as 10 percent (Meyer and Scott 2011, quoted in Ariyarathne et al. 2012). This low rate is partly due to self-selection out of the sector as a result of the stigma as well as married women’s increased domestic and caretaking responsibilities. However, research also reveals management preferences for young unmarried workers in the recruitment process, due in large part to a desire to avoid worker “chal- lenges” related to pregnancy and child care (Prasanna and Kuruppuge 2013). The enforcement of labor standards that minimize such bias and allow women to balance their care-giving and other domestic responsibilities is needed to provide equal work opportunities for married and older women who wish to sustain their employment in the sector. • Establish garment factories in remote areas of Sri Lanka in collaboration with local and foreign investors. Given the need for expanded job opportunities in rural and remote regions, such as the conflict-affected areas, the garment industry needs greater incentives to move into these regions. Rural areas would be conducive to increasing garment sector work in grassroots-level job mar- kets, where women would be able to work from their homesteads and better balance work and household responsibilities. • Finally, as in most sectors, providing women with dedicated, safe transporta- tion to and from garment sector jobs—or providing secure and comfortable living accommodations to those who cannot make the daily commute—is imperative for attracting and retaining more women in the industry. Providing support for child care is important as well: the global Sri Lankan textile and apparel manufacturer Mas Holdings reported that once it provided on-site child care in one of its factories in Jordan, the incidence of sick leave fell by 9 percent after a mere 8 months and production lines were more stabilized (Niethammer 2017; IFC 2017). Investments in female worker housing and Getting to Work (Overview) Overview 59 transportation would address some of the safety concerns of women and their families and lower the cultural barriers to women’s work in the sector (Ariyarathne et al. 2012). These services would not only open up new employ- ment options for women but also promote the long-term growth of the indus- try given the large number of workers the sector is looking to recruit to meet increasing international demand. Commercial Agriculture Even though respondents to the primary survey tend to feel that commercial agriculture presents the fewest social barriers for work-seeking women, employ- ers from this sector were most likely (80 percent) to report that they do not have plans to hire women. They also are among the employers most likely to express the opinion that men are better workers than women. Policy makers, practitio- ners, and development partners need to make concerted efforts to improve opportunities for women in agricultural value chains; once these opportunities come to fruition, the attitudes of those working in the sector may start to change. • Agricultural development programs that emphasize diversification to higher- value crops should work closely with the government to develop and pilot schemes that help women practicing subsistence farming make the transition to more profitable crops and markets. Investments in modern machinery that women farmers can use at various stages of production, along with technical training to improve efficiency, yield higher-quality products. Greater efficiency also eases the time poverty of women, allowing them more time for child care, and higher-quality products fetch higher prices in product markets. At the same time, addressing female farmers’ consistent disadvantage in access to markets may be remedied to a large extent by expanding women-friendly and affordable transport links and increasing their representation in farmer, pro- ducer, and marketing groups. • Commercial agriculture is a valuable entry point for reaching some of the most vulnerable groups of women, such as those who are heads of households (many of whom are war widows in Sri Lanka); other women living in the impoverished, conflict-affected areas of northern and eastern Sri Lanka; and women who aspire to supplement or replace their income from their low- reward work in estate sector jobs. Agricultural development programs should incorporate explicit provisions to address constraints on the participation of women who are resource poor or otherwise marginalized because of, for example, household status, ethnicity, or lack of education. Such provisions could include intensive outreach and training for these marginalized groups as well as efforts to address the lack of child care or other household responsibili- ties that tend to keep women from engaging in paid work. The use of female extension trainers to work with women in more conservative communities— especially those that value restrictions on women’s mobility—can ensure the success of programs in these communities. Also important is ensuring that Getting to Work (Overview) 60 Overview professional mobilizers and trainers have the resources required for targeted outreach and training. • The private sector needs to be provided with greater incentives to invest in rural industries. Expansion and promotion of agro-based and off-farming industries at the rural level could focus on local opportunities in the garment industry (for example, as suppliers of materials to larger actors who are further along in value chains), other cottage industries, food and beverage production, light engineering, and eco-tourism. • Finally, interventions to improve women’s participation in and remunera- tion from commercial agriculture must take into account its lack of worker benefits that support women as well as the lack of protections for women. Any improvements that can be made to ease the time poverty of women (for example, through providing child and elder care support, greater access to electricity and water, and reliable transportation to and from mar- kets) and lower the risk of harassment and other violence against women working in agriculture (for instance, through community-level awareness raising and provision of safe and dedicated transportation) are likely to yield a substantial increase in women’s participation in the sector. Fortunately (and somewhat surprisingly), according to the employer and household surveys, commercial agriculture, compared with other sectors, is reported to have the most formal procedural responses in place to address workplace sexual harassment. Thus, compared with other sectors, com- mercial agriculture provides a stronger foundation on which to address the threat of SGBV. Notes 1. As articulated by the Honorable Prime Minister in his Economic Policy Statement of November 5, 2015 (found at http://www.news.lk/fetures/item/10674-economic​ -­policy-statement-made-by-prime-minister-ranil-wickremesinghe-in-parliament). More recently, the Prime Minister has said of the government’s new (2017) policy document, “Powerful Sri Lanka,” that the government aims to “establish an economy providing equal opportunities to one and all…not an economy that will yield benefits to a few….” (http://mnpea.gov.lk/web/index.php?option=com_content&view=article​ &id=170:our-aim-is-to-enhance-the-economic-condition-of-all-sri-lankans&catid=9& Itemid=112&lang=en). 2. Conflict-affected areas were excluded because of the absence of a sampling frame. It had not been possible to conduct a census in these areas from 1981 until after the end of the conflict. 3. The 2012 World Development Report on Gender and Development understands agency as “the process through which women and men use their endowments and take advantage of economic opportunities to achieve desired outcomes” (World Bank 2011, 150). Getting to Work (Overview) Overview 61 4. For example, the same descriptive and multivariate analysis of 2015 LFS data has been conducted using the full 2015 LFS samples as well as the 2015 LFS without Northern and Eastern Provinces. Estimates from both samples are virtually identical, with the same directionality, same levels of statistical significance, and, only in a few cases, slight differences in magnitude. 5. In this study, LFP rates for all years and both sexes are measured as the percentage of that population group, age 15 and older, that is employed or actively seeking employment. 6. Sri Lanka’s Department of Census and Statistics (DCS) identifies the presence of the following ethnic groups in its Labour Force Survey and other relevant surveys, includ- ing the 2012 Census of Population and Housing: Sinhalese, Sri Lankan Tamil, Indian Tamil, Sri Lankan Moor, Burgher, Malay, and Other (the 2012 census also identifies the Sri Lanka Chetty and Bharatha ethnicities, but their share of Sri Lanka’s popula- tion is reported to be 0 percent). According to the 2012 census, the shares of population among ethnic groups, in declining order, are 74.9 percent Sinhalese, 11.2 percent Sri Lanka Tamil, 9.3 percent Sri Lanka Moor, 4.1 percent Indian Tamil, 0.2 percent Burgher, 0.2 percent Malay, and 0.1 percent Other. 7. To ensure that the changes in LFP rates by income level over time are not simply a function of different samples (some districts in the Northern Province were omitted from the 2009–10 HIES survey, while all districts in all provinces were included in the 2012–13 survey), the authors also calculated LFP estimates using a sample of the 2012–13 HIES data that excluded the same Northern Province districts as in 2009– 10. The results show no meaningful difference. There is absolutely no change in sign or magnitude of the coefficients, and any differentials in magnitude are minimal— ranging between 0 and 1 percent. This outcome is also true for all descriptive and regression results calculated using data from Sri Lanka’s Labour Force Surveys. Tables of the quantitative analysis results using different samples of these secondary data sources are available upon request. 8. The poverty head count ratio is calculated as the share of the population living below the national poverty line. The national poverty head count ratio for Sri Lanka in 2013 was 6.7 percent; in 2010 it was 8.9 percent. (http://data.worldbank.org/indicator​ SI​ /­ .POV.NAHC). 9. Remittances may lead to low female labor force participation for several reasons. Remittances may lead to labor substitution within the household to compensate for the migrant household member’s absence, could have positive education effects by increasing enrollment of female children in school, and can lead to labor market inac- tivity because of the work-leisure tradeoff as incomes rise from remittances. 10. It is important to note that the estimated decline in head count poverty from 22.7 percent to 6.1 percent during this period does not include data from districts in the Northern and Eastern Provinces (World Bank 2016). 11. Based on World Bank calculations from data in LFS 2011, LFS 2013, and LFS 2015 for population age 15–64. Traditional services include retail and wholesale trade, trans- port and storage, public administration, and defense (Eichengreen and Gupta 2011). Intermediate services are a hybrid of traditional and modern services, consumed mainly by households—education, health and social work, hotels and restaurants, and other community, social, and personal services. Modern services include financial intermediation, computer services, business services, communications, and legal and technical services. Getting to Work (Overview) 62 Overview 12. Unemployed male, 28 years old, O-level pass, unmarried—Konthahela village, Badulla district. 13. Unemployed male, 24 years old, unmarried—Henegama village, Gampaha district. 14. Unemployed male, 26 years old, grade 10 pass, married—Henegama village, Gampaha district. 15. Unemployed male, 26 years old, grade 10 pass, married—Henegama village, Gampaha district. 16. According to results from the Oaxaca-Blinder decomposition of labor force participa- tion, the total gap in 2015 was 39 percentage points (that is, the coefficient is 0.390 with p < 0.01). The coefficient for the explained portion of the gap is 0.0267 (p < 0.01), or 2.67 percentage points. The gender differences in endowments (all variables) thus explain 6.8 percent of the gap, since (2.67/39) × 100 = 6.8. 17. Samples from the three rounds include population age 15 years and older in all dis- tricts except those in the Northern and Eastern Provinces, which are excluded to allow for comparability across years. 18. The Oaxaca-Blinder analysis uses LFS data and thus defines ethnicity according to LFS and other surveys used by Sri Lanka’s Department of Census and Statistics (see note 6 for a discussion on ethnicities in Sri Lanka and their relative shares among the national population). 19. This holds even when omitting from the 2009 and 2015 samples both the Eastern and Northern Provinces, which were the two provinces not represented in the 2006 LFS. 20. The Sri Lanka Systematic Country Diagnostic (World Bank 2015a) cites pockets of high poverty rates in Monaragala district, the estate sector, and the north and east of Sri Lanka. 21. Unemployed female, 28 years old, A-level pass, married—Yaithena village, Gampaha district. 22. The Junior Achievement program in the United States, for example, orients students to business and entrepreneurship activities while enhancing financial literacy. The Girl Scouts and Girl Guides programs have related activities in the area of girls’ business and career acumen that could be further supported in the Sri Lankan context (CGO 2012). 23. The Tertiary and Vocational Education Commission is the Ministry of Education agency that oversees accreditation of training providers. Training requirements outpace available spaces in such industry areas as hotels and tourism (1,960 train- ing places for 28,000 jobs to be filled), building and construction (4,538 places for 25,000 jobs), and metal and light engineering (4,300 places for 16,000 positions) (GoSL 2009). 24. Authors in Sri Lanka have cited gender constraints faced by female students at the university level, particularly in regard to mentoring, professional development, and educational streaming (Gunawardena 2003; Gunawardena et al. 2006). More gener- ally, we note that there appear not to have been studies in Sri Lanka regarding pres- ence of gender-regressive practices in the classroom setting—for example, teachers exhibiting favoritism toward male students—despite research elsewhere pointing to such subtle or overt practices. 25. Female hotel worker, 29 years old, O-level pass, married—Kumburupitty village, Trincomalee district. Getting to Work (Overview) Overview 63 26. Male machine operator in garment factory, 26 years old, O-level pass—Henegama village, Gampaha district. References ADB (Asian Development Bank). 2008. Country Gender Assessment: Sri Lanka. Manila: Asian Development Bank. ———. 2013. Gender Equality in the Labor Market in the Philippines. Manila: Asian Development Bank. Ariyarathne, Tehani, Gayathri Lokuge, Nadhiya Najab, and Priyanthi Fernando. 2012. “The Vulnerability of Women in the Economy.” Economic Review, People’s Bank, Colombo. Vol 38. http://www.cepa.lk/uploads/4bae1bc28cb6cbe73a08a2bf084b8fa3-120830​ -ER-The-vulnerability-of-women-in-the-economy---CEPA.pdf. Arunatilake, Nisha, Suwendrani Jayaratne, Priyanka Jayawardena, Roshini Jayaweera, and Dushni Weerakoon. 2010. Impact of Remittances on the Household of the Emigrant and on the Economy of the Migrant’s Country: Sri Lanka. Institute of Policy Studies of Sri Lanka. Baum, T. 2013. International Perspectives on Women and Work in Hotels, Catering and Tourism. Geneva: International Labour Organization. Carlin, Bruce Ian, and David T. Robinson. 2012. “What Does Financial Literacy Training Teach Us?” Journal of Economic Education 43 (3): 235–47. CGO (Center for Gender in Organizations). 2012. “Dreaming Big: What’s Gender Got to Do with It? The Impact of Gender Stereotypes on Career Aspirations of Middle Schoolers.” CGO Briefing Note 35, Center for Gender in Organizations, Simmons College, Boston, MA. http://www.simmons.edu/som/docs/insights_35​ _­v2_%282%29.pdf. DCS (Department of Census and Statistics). 2015. Sri Lanka Labour Force Survey 2014 Annual Report. Colombo: DCS. http://www.statistics.gov.lk/samplesurvey/LFS​ _­Annual%20Report_2014.pdf. ———. 2016a. Sri Lanka Labour Force Survey 2015 Annual Report. Colombo: DCS. http://www.statistics.gov.lk/samplesurvey/LFS_Annual%20Report_2015.pdf. ———. 2016b. Sri Lanka Labour Force Survey 2016—Quarterly Bulletin. Colombo: DCS. http://www.statistics.gov.lk/samplesurvey/LFS_Q2_Bulletin_WEB_2016_final.pdf. Eichengreen, B., and P. Gupta. 2011. “The Service Sector as India’s Road to Economic Growth.” NBER Working Paper No. 16757, National Bureau of Economic Research, Cambridge, MA. Ferguson, L. 2011. “Promoting Gender Equality and Empowering Women? Tourism and the Third Millennium Development Goal.” Current Issues in Tourism 14 (3): 235–49. Gamage, G. 2004. “Gender and ICT: A Sri Lankan Perspective.” Paper presented at the 9th National Convention on Women’s Studies, March, Colombo. GoSL (Government of Sri Lanka). 2009. Labour and Social Trends in Sri Lanka. Ministry of Labour Relations and Manpower, Central Bank of Sri Lanka and Department of Census and Statistics. Colombo: GoSL. ———. 2012. The National Human Resources and Employment Policy for Sri Lanka. Secretariat for Senior Ministers. Colombo: GoSL. Getting to Work (Overview) 64 Overview Gunatilaka, Ramani. 2013. “Factors Holding Women Back from Market Work in Sri Lanka.” International Labour Office, ILO DWT for South Asia and Country Office for India, New Delhi. Gunetileke, Neranjana, Sanjana Kuruppu, and Susrutha Goonasekera. 2008. The Estate Workers’ Dilemma: Tensions and Changes in the Tea and Rubber Plantations in Sri Lanka. Study Series 4. Colombo: CEPA. Gunawardena, Chandra. 2003. “Gender Equity in Higher Education in Sri Lanka: A Mismatch between Access and Outcomes.” McGill Journal of Education 38 (3): 437–50. Gunewardena, Chandra, Yoga Rasanayagam, Tressir Leitan, Kanchana Bulumulle, and Asha Abeyasekera-Van Dort. 2006. “Quantitative and Qualitative Dimensions of Gender Equity in Sri Lankan Higher Education.” Women’s Studies International Forum 29: 562–71. Gunewardena, Dileni. 2015. “Why Aren’t Sri Lankan Women Translating Their Educational Gains into Workforce Advantages?” The 2015 Echidna Global Scholars Working Paper, Center for Universal Education, Brookings Institution, Washington, DC. Iaccino, Ludovica. 2014. “Top Five Countries with Highest Rates of Child Prostitution.” International Business Times, February 6. http://www.ibtimes.co.uk/top-five​ -countries-highest-rates-child-prostitution-1435448. ILO (International Labour Organization). 2014. “Key Indicators of the Labour Market.” ILO, Geneva. Retrieved from World Bank Data Center: http://data.worldbank.org​ /­indicator/SL.TLF.CACT.FE.NE.ZS?locations=TH-MY. International Finance Corporation (IFC). 2017. Tackling Childcare: The Business Case for Employer-Supported Child Care. Washington, DC: IFC. Jayasura, R., and B. Opeskin. 2015. “The Migration of Women Domestic Workers from Sri Lanka: Protecting the Rights of Children Left Behind.” Cornell International Law Journal 48 (3): 579–638. Jayaweera, S., and L. Wanasundera. 2006. “Gender and ICT in Sri Lanka: An Overview.” Paper presented at the 10th National Convention on Women’s Studies, April, Colombo. ———, Hiranthi Wijemanne, Leelangi Wanasundera, and Kamini Meedeniya Vitarana. 2007. Gender Dimensions of the Millennium Development Goals in Sri Lanka. Colombo: UNDP and Centre for Women’s Research. http://www.undp.org/content/dam​ /­srilanka/docs/mdg/Gender_Dimensions%20of%20Sri%20Lanka.pdf. Kamal, A., and L. Woodbury. 2016. Emerging Opportunities for Women in Khyber Pakhtunkhwa: Industry (Manufacturing and Construction), Mining, Tourism, and Agriculture. London: International Growth Center, London School of Economics and Political Science. Karkkainen, O. 2011. Women and Work in Tunisia: Tourism and ICT Sectors: A Case Study. Turin: European Training Foundation. LIRNEasia. 2006. “A Baseline Sector Analysis of the Business Process Outsourcing Industry in Sri Lanka.” Information and Communication Technology Agency of Sri Lanka, Colombo. Lopez-Acevedo, G., and R. Robertson. 2016. Stitches to Riches? Apparel Employment, Trade, and Economic Development in South Asia. Washington, DC: World Bank. Meyer, R., and H. Scott. 2011. “Women in the Global Factory—The Garment Industry in Sri Lanka.” Accessed August 23, 2012. http://www.utexas.edu/cola/insts​southasia​ /­ /_files. Getting to Work (Overview) Overview 65 Morgan, Sharon. 2012. “Women’s ICT Sector Employment in Developing Countries: Dualism of Rhetoric vs. Reality in the Case of Sri Lanka.” Development Informatics Working Paper No. 49, Centre for Development Informatics, Institute for Development Policy and Management, SED, University of Manchester. Niethammer, Carmen. 2017. “The Business Case for Tackling Childcare: Good for Employees and Employers, Good for Sri Lanka.” Financial Times, Sri Lanka. October 3. http://www.ft.lk/columns/The-business-case-for-tackling-childcare--good​ - for​ -employees​-and-employers--good-for-Sri-Lanka/4-640772. Perera, Jennifer, Nalika Gunawardane, and Vathsala Jayasuriya, eds. 2011. Review of Research Evidence on Gender Based Violence (GBV) in Sri Lanka. Colombo: Sri Lanka Medical Association. Prasanna, R. P. I. R., and R. H. Kuruppuge. 2013. “General Wage Situation of Apparel Industry Workers in Sri Lanka.” Asia Floor Wage Alliance. http://asia.floorwage.org​ /­r esources/wage-reports/general-wage-situation-of-apparel-industry-workers​ -in-sri-lanka. Ranasinghe, Athula Lakshman. 2004. “ICT and Decent Work: An Assessment of the Labour Market Impacts of the Advent of Information and Communication Technology in Sri Lanka.” Department of Economics, University of Colombo, Sri Lanka. Prepared for and presented at the JILPT/ILO Networking of National Institutes for Labour Studies: Meeting on Joint Investigative Studies for “Determining the Impact of Information and Communications Technology on Decent Work in the Asia and Pacific Region,” Bangkok, Thailand, December 2–3. Reddy, R. Ravikanth. 2015. “Computer Education May Be Made Mandatory.” The Hindu. January 20. http://www.thehindu.com/news/cities/Hyderabad/computer-education​ -may-be-made-mandatory/article6800774.ece. Samath, Feizal. 2011. “Sri Lanka: Garment Industry Woos Women Workers.” Inter Press Service News Agency. February 10. http://www.ipsnews.net/2011/02/sri-lanka​ -garment-industry-woos-women-workers/. Senewiratne, H. H. 2011. “‘Lanka Needs More Graduates for IT Industry Growth’ –Madu Ratnayake.” The Island Newspaper, Colombo. Sinha, Nistha. 2012. “Demographic Transition and Female Labor Force Participation in Sri Lanka.” South Asia Human Development Sector Discussion Paper 41, World Bank, Washington, DC. SLBFE (Sri Lanka Bureau of Foreign Employment). 2015. Annual Statistical Report of Foreign Employment—2015. Colombo: SLBF. http://www.slbfe.lk/page.php?LID​ =1&MID=213#. Solotaroff, J. L., G. Joseph, and A. T. Kuriakose. Forthcoming. Getting to Work: Unlocking Women’s Potential in Sri Lanka’s Labor Force. Washington, DC: World Bank. United Nations, Sri Lanka. 2015. Mapping of Socio-Economic Support Services to Female Headed Households in the Northern Province of Sri Lanka. Colombo: United Nations, Sri Lanka. UN Women. 2011. Global Report on Women in Tourism 2010. New York: United Nations. WEF (World Economic Forum). 2016. Global Gender Gap Report 2016. Geneva: World Economic Forum. http://www3.weforum.org/docs/GGGR16/WEF_Global_Gender​ _Gap_Report_2016.pdf. World Bank. 2007. Sri Lanka Poverty Assessment. Engendering Growth with Equity: Opportunities and Challenges. Washington, DC: World Bank. Getting to Work (Overview) 66 Overview ———. 2011. World Development Report 2012: Gender Equality and Development. Washington, DC: World Bank. ———. 2012. “Sri Lanka—Gender and Economic Opportunities: Securing Jobs for Women and Men.” Unpublished, World Bank, Washington, DC. ———. 2013. Getting In and Staying In: Improving Women’s Labor Force Participation in Sri Lanka. Washington, DC: World Bank. ———. 2015a. Sri Lanka—Ending Poverty and Promoting Shared Prosperity: A Systematic Country Diagnostic. Washington, DC: World Bank. ———. 2015b. Sri Lanka—Poverty and Welfare: Recent Progress and Remaining Challenges. Washington, DC: World Bank. ———. 2015c. World Development Indicators Database. World Bank, Washington, DC. http://data.worldbank.org/indicator/SL.TLF.CACT.FE.NE.ZS?locations=TH-MY. ———. 2016. International Bank for Reconstruction and Development, International Development Association, International Finance Corporation, Multilateral Investment Guarantee Agency Country Partnership Framework for the Democratic Socialist Republic of Sri Lanka for the Period FY17–FY20. Report 104606-LK. Washington, DC: World Bank. ———. Forthcoming. Sri Lanka’s Underperforming Boys: The Gender Dimensions of Educational Access and Achievement in Sri Lanka. A World Bank Study. Washington, DC: World Bank. Getting to Work (Overview) Sri Lanka has shown remarkable persistence in low female labor force participation rates—at 36 percent in the past two years, compared with 75 percent for same-aged men—despite overall economic growth and poverty reduction over the past decade. The trend stands in contrast to the country’s achievements in human capital development that favor women, such as high levels of female education and low total fertility rates, as well as its status as a lower-middle-income country. This study intends to better understand the puzzle of women’s poor labor market outcomes in Sri Lanka. Using nationally representative secondary survey data—as well as primary qualitative and quantitative research—it tests three hypotheses that would explain gender gaps in labor market outcomes: (1) household roles and responsibilities, which fall disproportionately on women, and the associated sociophysical constraints on women’s mobility; (2) a human capital mismatch, whereby women are not acquiring the proper skills demanded by job markets; and (3) gender discrimination in job search, hiring, and promotion processes. Further, the analysis provides a comparison of women’s experience of the labor market between the years leading up to the end of Sri Lanka’s civil war (2006–09) and the years following the civil war (2010–15). The study recommends priority areas for addressing the multiple supply- and demand-side factors to improve women’s labor force participation rates and reduce other gender gaps in labor market outcomes. It also offers specific recommendations for improving women’s participation in the five private sector industries covered by the primary research: commercial agriculture, garments, tourism, information and communications technology, and tea estate work. The findings are intended to influence policy makers, educators, and employment program practitioners with a stake in helping Sri Lanka achieve its vision of inclusive and sustainable job creation and economic growth. The study also aims to contribute to the work of research institutions and civil society in identifying the most effective means of engaging more women—and their untapped potential for labor, innovation, and productivity—in Sri Lanka’s future. SKU 33125