Bangladesh Social Protection and Labor Review: Towards Smart Social Protection and Jobs for the Poor Bangladesh Development Series Paper No. 33 The World Bank March 2016 www.worldbank.org.bd/bds The findings, interpretations, and conclusions expressed in this paper are entirely those of the author. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. World Bank Office Dhaka Plot – E-32, Agargaon, Sher-e-Bangla Nagar, Dhaka- 1207, Bangladesh Tel: 880 2 5566 7777 Fax: 880 2 5566 7778 www.worldbank.org/bangladesh World Bank 1818 H Street, N.W. Washington DC 20433, USA Tel: 1 202 473 1000 Fax: 1 202 477 6391 www.worldbank.org Standard Disclaimer: The volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominators, and other information shown in this work do not imply any judgement on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission to reproduce portion of the work promptly. 1 Acknowledgements This report reviews the landscape of social protection and labor policy agenda and programs in Bangladesh and provides policy recommendations for the Government of Bangladesh. The review was part of the World Bank’s Programmatic Technical Assistance and was produced under the strategic leadership of Iffath Sharif, Program Leader, and under the guidance of Pablo Gottret, Practice Manager. The team, led by Yoonyoung Cho (Task Team Leader and author), was comprised of Rehan Jamil, Hjalte Sederlof, Mariam Lomaia, Habtamu Fuje, and Abir Hasan. Several colleagues within the World Bank have also contributed to this report. Afrah Chowdhury has provided an analysis on international migration with International Organization for Migration data. The various drafts of the paper benefited from the valuable input and support of Aneeka Rahman, Ashiq Aziz, Cem Mete, Manjula Luthria, Munjir Ahmed, Rebekka Smith, Rubaba Anwar, Shawkat Ali, Shinsaku Nomura, and Simon Davies. Maitreyi Das, Roberta Gatti, Ruslan Yemtsov, and Martin Rama provided valuable comments and guidance through a review process. Special thanks go to Dr. Shamsul Alam and his team from the General Economics Division, Planning Commission of the Government of Bangladesh for their comprehensive feedback. The findings, interpretations, and conclusions expressed here are those of the author and should not be attributed to the World Bank, its management, or its Executive Board of Directors. 1 Foreword Over the past decade, Bangladesh has achieved strong economic growth and impressive poverty reduction. The country’s great strides in human development, reflected in the lowered fertility rate, improved health and epidemiological outcomes, and progress in education, among others, are paving the way for its continued development and economic growth. Despite much progress, however, poverty and vulnerability remains a great challenge. In particular, persistent gaps across different populations exist often due to the labor market conditions experienced by marginalized workers (including unskilled laborers, females, youth, and migrants). A rapid demographic transition, which provides a window of opportunity to utilize young workers before the country faces the issue of population ageing, highlights the need for urgent policy intervention. In light of such dynamic changes in Bangladesh, the role of comprehensive Social Protection and Labor (SPL) policy to facilitate poor and vulnerable households to adequately respond to and manage various risks, is greater than ever. The recently endorsed National Social Security Strategy (NSSS) proceeds in this direction by adopting the lifecycle approach. This Social Protection and Labor Review report complements the NSSS with discussions on specific issues and policy recommendations. The report reviews the landscape of the sector, identifies key challenges to be overcome, and provides options for policy intervention. Overall, the review emphasizes the need for policy direction to shift from individual programs towards an overall strategy for households; from fragmented safety nets towards a comprehensive system; and from (ex-post) disaster reliefs towards (ex-ante) promotion of productive income generating activities. Therefore, the report greatly underscores the importance of strengthening labor market intervention and developing social insurance while addressing implementation challenges related to existing social assistance programs. In contrast to the multitude of various social assistance programs, Bangladesh currently lacks effective measures to promote labor market employment and productivity and adequate social insurance to hedge various risks. In the short term, the report suggests that policy efforts need to be made to introduce active labor market programs to provide greater employment opportunities, particularly for disadvantaged populations. Strategies to develop a pension system that can cover workers regardless of their type of employment are also needed. These measures will serve as building blocks for broader policies for more, better, and inclusive jobs, and a comprehensive social insurance system, which will contribute to the country’s goal of achieving middle income country status by 2021. Xiaoqing Yu Martin Rama Qimiao Fan Acting Senior Director Chief Economist Country Director Social Protection and Labor South Asia Region Bangladesh, Bhutan and Nepal 2 List of Abbreviations ALMP Active Labor Market Program BBS Bangladesh Bureau of Statistics BDT Bangladeshi Taka BES Bangladesh Enterprise Survey BGMEA Bangladesh Garment Manufacturers and Exporters Association BLA Bangladesh Labour Act BMET Bureau of Manpower, Employment, and Training BOESL Bangladesh Overseas Employment Services Limited CCT Conditional Cash Transfer CPIA Country Policy and Institutional Assessment EGPP Employment Generation Program for the Poorest FDI Foreign Direct Investment FFW Food for Works GDP Gross Domestic Product GoB Government of Bangladesh GR Gratuitous Relief HIES Household Income and Expenditure Survey IOM International Organization For Migration ISPP Income Support Program for the Poorest LFS Labor Force Survey MFI Micro Finance Institution MIS Management Information System MoDMR Ministry of Disaster Management and Relief MoE Ministry of Education MoF Ministry of Finance MoFd Ministry of Food MoLGD&C Ministry of Local Government and Rural Development & Cooperatives MoPME Ministry of Primary and Mass Education MoSW Ministry of Social Welfare MRA Microcredit Regulatory Authority NDSS National Sustainable Development Strategy NGO Non-government organization NHD National Household Database NSSS National Social Security Strategy OMS Open Market Sales PES Primary Education Stipend PFDS Public Food Distribution System PIC Project Implementation Committee 3 PMT Proxy Means Test RCT Randomized Controlled Trial RMG Ready Made Garment SNSP Safety Net Systems for the Poorest TFR Total Fertility Rate TR Test Relief TVET Technical and Vocational Education and Training VGD Vulnerable Group Development VGF Vulnerable Group Feeding 4 Table of Contents EXECUTIVE SUMMARY .............................................................................................................................................. 9 INTRODUCTION...................................................................................................................................................... 17 1 SOCIAL PROTECTION AND LABOR ................................................................................................................ 21 1.1 WHAT IS A SOCIAL PROTECTION AND LABOR SYSTEM? ...........................................................................................21 1.2 WHAT IS THE ROLE OF SOCIAL PROTECTION AND LABOR? .......................................................................................21 1.3 WHAT ARE SPL POLICY INSTRUMENTS?...............................................................................................................23 1.4 CONCEPTUAL FRAMEWORKS: A SPL SYSTEM APPROACH .......................................................................................23 2 CONTEXT: BANGLADESH’S SPL POLICY ENVIRONMENT............................................................................... 25 2.1 MACROECONOMIC ENVIRONMENT .....................................................................................................................25 2.2 LABOR MARKET CONTEXT .................................................................................................................................29 2.3 LABOR MOBILITY AND MIGRATION .....................................................................................................................38 2.4 SUMMARY .....................................................................................................................................................41 3 REVIEW OF SPL PROGRAMS, EXPENDITURE, AND PERFORMANCE IN BANGLADESH.................................... 43 3.1 SOCIAL INSURANCE ..........................................................................................................................................44 3.2 SAFETY NETS ..................................................................................................................................................46 3.3 MICROFINANCE ..............................................................................................................................................56 3.4 LABOR MARKET PROGRAMS ...............................................................................................................................61 4 KEY CHALLENGES TO BE ADDRESSED IN SOCIAL PROTECTION AND LABOR .................................................. 71 4.1 CHALLENGE 1: WIDESPREAD POVERTY AND VULNERABILITY ....................................................................................71 4.2 CHALLENGE 2: JOBS CHALLENGES FOR YOUTH ......................................................................................................72 4.3 CHALLENGE 3: GENDER INEQUALITY IN THE LABOR MARKET ....................................................................................75 4.4 CHALLENGE 4: LABOR MIGRATION .....................................................................................................................77 5 POLICY RECOMMENDATIONS ...................................................................................................................... 81 5.1 IMPROVING EXISTING PROGRAMS ......................................................................................................................81 5.2 UTILIZING A WIDER SET OF POLICY INSTRUMENTS TO MANAGE RISKS AND PROMOTE EARNINGS OPPORTUNITIES .............86 6 CONCLUSION ............................................................................................................................................... 99 REFERENCES....................................................................................................................................................... 101 APPENDIX: DATA ............................................................................................................................................... 109 5 List of Figures FIGURE 1. CONCEPTUAL FRAMEWORK ................................................................................................................ 24 FIGURE 2. AVERAGE PER CAPITA HOUSEHOLD CONSUMPTION ................................................................................ 26 FIGURE 3. LORENZ CURVE AND GINI COEFFICIENTS OVER TIME ............................................................................... 26 FIGURE 4. MAJOR BUSINESS CONSTRAINTS .......................................................................................................... 28 FIGURE 5. TOTAL FERTILITY RATES AND POPULATION GROWTH IN SOUTH ASIA ......................................................... 31 FIGURE 6. EMPLOYMENT RATIOS AT A GLANCE..................................................................................................... 32 FIGURE 7. LIFECYCLE PATTERN OF EMPLOYMENT BY REGION AND GENDER ............................................................... 33 FIGURE 8. EXTREME POVERTY RATES BY EMPLOYMENT TYPE .................................................................................. 34 FIGURE 9. TYPE OF OCCUPATIONS AMONG WORKERS OVER TIME: BY GENDER .......................................................... 35 FIGURE 10. EMPLOYMENT SECTOR AMONG WORKERS: BY GENDER ......................................................................... 36 FIGURE 11. SHARE OF FORMAL WORKERS BY GENDER........................................................................................... 36 FIGURE 12. EDUCATION OF THE WORKING AGE POPULATION BY GENDER AND CONSUMPTION DECILES ........................ 37 FIGURE 13. ANNUAL MIGRATION ABROAD AND SIZE OF REMITTANCES .................................................................... 40 FIGURE 14. PERCENTAGE OF HOUSEHOLDS RECEIVING EACH PROGRAM BENEFIT: BY INCOME QUINTILES ...................... 53 FIGURE 15. EMPLOYMENT OUTCOMES OF YOUTH (AGES 15–24) VERSUS ADULTS (AGES 25–64) BY GENDER ............... 72 FIGURE 16. SCHOOL-TO-WORK TRANSITION ........................................................................................................ 73 FIGURE 17. TRANSITION TO THE LABOR MARKET AMONG TVET GRADUATES ............................................................ 74 FIGURE 18. SHARE OF THOSE NEVER MARRIED BY GENDER .................................................................................... 76 FIGURE 19. TIME USE PATTERN BY EMPLOYMENT STATUS AND SECTOR AMONG THE EMPLOYED: BY GENDER ................ 76 FIGURE 20. THE SOURCE OF INFORMATION ON INTERNATIONAL MIGRATION: BY DESTINATION AND EDUCATION ............ 78 FIGURE 21. COMPOSITION OF MIGRATION COSTS: BY FACILITATING AGENCY............................................................ 79 List of Tables TABLE 1. SELECT INDICATORS AND CPIA RATING .................................................................................................. 27 TABLE 2. SOCIAL PROTECTION AND LABOR – THE CONCEPTUAL FRAMEWORK AND ITS APPLICATION TO BANGLADESH ..... 43 TABLE 3. SP BUDGET OVER TIME ....................................................................................................................... 44 TABLE 4. NINE LARGEST SAFETY NET PROGRAMS IN BANGLADESH AS OF FY 2015-16 ............................................... 47 TABLE 5. POVERTY AND SAFETY NET COVERAGE BY DIVISION, 2010 ...................................................................... 51 TABLE 6. ALLOCATION BY LARGEST MINISTRIES .................................................................................................... 55 TABLE 7. FOUR LARGEST MFIS IN BANGLADESH (2012) ........................................................................................ 57 TABLE 8. LARGE-SCALE LABOR PROGRAMS IN BANGLADESH ................................................................................... 61 6 List of Boxes BOX 1. TARGET POPULATIONS OF SPL................................................................................................................. 21 BOX 2. SPL POLICY INSTRUMENTS...................................................................................................................... 23 BOX 3. MACROECONOMIC CONTEXT .................................................................................................................. 25 BOX 4: LABOR MARKET CONTEXT ...................................................................................................................... 29 BOX 5. SLUM CHARACTERISTICS ......................................................................................................................... 39 BOX 6. INTERNAL AND INTERNATIONAL MIGRATION .............................................................................................. 40 BOX 7. EVIDENCE ON THE IMPACTS OF MICROFINANCE.......................................................................................... 60 BOX 8. INSTITUTIONS RELATED TO INTERNATIONAL MIGRATION .............................................................................. 65 BOX 9. BANGLADESH BUREAU OF STATISTICS’ EFFORTS FOR DATA COLLECTION ......................................................... 69 BOX 10. SHOMBHOB AND INCOME SUPPORT PROGRAM FOR THE POOREST (ISPP) .................................................... 83 BOX 11. ONGOING DISCUSSIONS ON SPL PROGRAMS ........................................................................................... 84 BOX 12. PROVIDING SAVING INCENTIVES ............................................................................................................ 88 BOX 13. GRADUATION PROJECTS ....................................................................................................................... 89 BOX 14. JOVENES TYPE PROGRAMS FOR WAGE EMPLOYMENT ............................................................................... 91 BOX 15. DESIGNING ENTREPRENEURSHIP PROGRAMS FOR YOUTH........................................................................... 93 BOX 16. POLICIES FOR SAFE INTERNATIONAL MIGRATION ...................................................................................... 97 7 8 Executive Summary A stronger Social Protection and Labor system is needed in Bangladesh. Social Protection and Labor (SPL) systems are a set of policies and programs that aim to achieve three objectives: (i) resilience by helping people rebound from setbacks from different types of shocks; (ii) equity by reducing poverty and destitution and by promoting equality of opportunity; and (iii) opportunity by building human capital, assets, and access to jobs, which in turn reduces the susceptibility to shocks and builds resilience. Worldwide, SPL is increasingly an important part of each government’s strategy to fight poverty and promote inclusive growth, with a balanced emphasis on the three objectives. A stronger SPL system, which effectively addresses various lifecycle risks and utilizes adequate policy instruments, is required in Bangladesh as recognized by the National Social Security Strategy (NSSS) of Bangladesh in 2015. In particular, Bangladesh should place a greater emphasis on promoting opportunities for better earnings and sustainable livelihoods to complement social safety nets. Over the past decade, Bangladesh has achieved strong economic growth and impressive poverty reduction. Between 2000 and 2014, the average growth rate per year was close to 6 percent. Purchase Power Parity (PPP) adjusted per capita Gross National Income (GNI) has increased from US$1,140 to US$3,330 over the same period. Driven by strong economic performance, the percentage of the population living under the national poverty line declined from 49 percent in 2000 to 32 percent in 2010. Economic growth over the past decade has been slightly progressive; consumption growth of households in the bottom 40 percent of the income distribution is slightly higher than that of average or upper quintile households. Along with economic growth, Bangladesh has made great strides in improving human development and achieving good labor market outcomes. Fertility rates and population growth have steadily declined in Bangladesh; Total Fertility Rate (TFR) reduced from 3.1 to 2.3 between 2000 and 2014. As a result, the dependency ratio has declined while the share of working age population has increased, which is favorable to reaping the demographic dividend. Educational enrollment, particularly at the primary and secondary levels, has significantly improved. Moreover, the gender gap in education is significantly closing. Structural transformation from agriculture to the industry and service sectors is showing slow but steady progress; this is in part due to the continued contribution from the Ready Made Garment (RMG) sector. Moreover, opportunities for overseas employment are expanding and remittances from migrant workers continue to be an important source of income. However, numerous challenges hamper the country’s ability to create more, better, and inclusive jobs, which is critical for the country to benefit from the ongoing favorable demographic transition and progress in human development. The quality of jobs in Bangladesh is a concerning issue for many workers. Many workers (one third of males and over a half of females) are either engaged in unpaid family work or work as day laborers. The majority of workers who are engaged in earnings-generating activities are self- 9 employed. Only a negligible share is employers with one or more paid employees, and the share of workers in wage-employment remains small. A partial explanation for the low quality of jobs is the low level of skills and investment in human capital of workers. Another contributing factor relates to fundamental issues regarding the quality of education and relevance of skills acquired in education systems for labor markets. Given these challenges related to jobs, many workers seek employment and earnings opportunities by (internal and international) migration. Every year, half a million new people arrive in urban areas and the share of urban poor is rising, many of whom often end up in illegal settlements on precarious land with poor living conditions. Similarly, a large number of workers choose international migration for economic opportunities. From 2000–2010, an average of 396,000 Bangladeshis went abroad each year, and this figure increased to about 500,000 between 2011 and 2015. Current SPL programming in Bangladesh is not adequate to effectively address the challenges. Bangladesh’s current SPL system, characterized by a large number of social safety nets with under- developed social insurance and labor market programs, is not adequately equipped to address the formidable challenges in the dynamically changing country. Social insurance (pension) is available only for a very limited number of government employees and those recognized as insolvent Freedom Fighters living in poverty. The overall budget for SPL in FY 2014–15, including the pension and Freedom Fighter’s honorarium programs, was over 2 percent of Gross Domestic Product (GDP); however, excluding the pension and Freedom Fighter’s honorarium programs, the SPL budget is reduced to roughly 1.4 percent of GDP. Labor market programs to promote sustainable means of livelihood and earnings opportunities are very limited in number and budget alike. To the contrary, numerous ministries are currently implementing over 145 safety net programs. The existing safety nets are marred by fragmentation, weak targeting, and inefficiency. Fragmented due to the number of programs which are operating with unclear or duplicative objectives and coverage, and without adequate coordination. For instance, many seemingly analogous social transfers and reliefs are currently being provided under multiple programs. Weak targeting because no scientific or rigorous targeting mechanisms are in place, and it is only recently that GoB has taken steps to construct a household database for the purpose of targeting the people who need these safety nets the most. A considerable portion of social safety net beneficiaries are indeed non-poor as indicated by household surveys. Targeting accuracy tends to suffer because most programs rely on geographical targeting, but do not consider individual household characteristics and poverty status. Finally, implementation inefficiency — most notably beneficiary duplication, leakages, and high administrative costs — hampers the effectiveness of the programs. This is particularly prominent in (but not limited to) food-based programs compared to cash-based ones. Compared to ubiquitous social safety net programs, programs for resilience and opportunities through social insurance or Active Labor Market Policies (ALMPs) are quite limited. Currently in Bangladesh, mechanisms to collectively hedge various risks across the different population groups or to inter- temporally manage the life-cycle risks are limited. The GoB has a long-term plan to introduce a universal health insurance, but there is an urgent need to introduce social and financial programs that facilitate risk 10 management through insurance and savings schemes. Likewise, ALMPs, including skills development to upgrade worker productivity, intervention to facilitate occupational or geographical mobility, provision of market information, business development services for entrepreneurial activities, among others, are not widely available. Recently, the labor discourse in Bangladesh is predominantly focused on labor regulations and international migration. Furthermore, the widespread microfinance sector may serve as a social safety net, but plays a limited role in creating earnings opportunities through promoting business activities. Accessibility to microfinance is high in Bangladesh, particularly in rural areas, and many households often have access to multiple sources of borrowing. One estimate suggested that almost one third of rural households were participating in multiple microcredit programs in 2010–11. In 1998–99, only 8.9 percent of rural households participated in these programs. The high presence of microfinance institutions (MFIs) can be a positive catalyst for expanding financial inclusion and providing earning opportunities through self-employment for the poor. Evaluations of the microcredits suggest that borrowers use the loans to supplement consumption and microfinance has contributed to poverty reduction, but there is room for microcredits to be more closely and effectively linked to business activities. Given the economic and policy environment, four key issues are highlighted where a comprehensive SPL policy intervention is required. The emerging challenges and the landscape of current programming suggest four important gaps that SPL policy attention has to focus upon. First, despite the impressive achievement in growth, poverty and vulnerability remains high due to income shocks and limited access to risk management for the poor. Large segments of the population are routinely exposed to seasonal and natural disasters, making the poor and vulnerable more prone to deprivation. With the majority of the poor engaged in agriculture or informal self-employment, volatility in incomes is high while the ability to cope with risks is limited. Second, with a demographic transition the jobs challenges for youth are high. Youth make up about 30 percent of the working age population, and the large number of young people entering the labor force poses considerable pressure to create jobs. Challenges in the labor market tend to be greater for youth compared to adults. The gap between youth and adults with respect to these indicators is larger for males than females. The school-to-work transition is difficult even among better educated youth. The quality of education in Bangladesh is of concern because skills relevant to the labor market are not taught in school, and training opportunities are very limited. Third, increasing the participation of and improving the outcomes for women in the labor market remains a challenge. Despite strong improvements in female literacy levels, education, and employment in selected sectors, female labor force participation in Bangladesh is still lower than average in the developing world; and a high share of working women is engaged in non-paid family work. Building upon the improved human development outcomes and benefitting from continued economic growth, female labor force participation can potentially rise even further. 11 Fourth, internal and international migration provides both opportunities and substantial risks. With the global market integration and rapid urbanization in Bangladesh, migration provides employment and earnings opportunities to many poor workers. At the same time, many risks pertaining to the migration process, lack of information, and large financial costs associated with recruiting agencies and intermediaries can substantially lower the welfare of migrants and their families. Programs to address the needs of migrants should be formulated as part of the package of broad jobs/employment strategies. Policy efforts to address these challenges require two-fold approaches: (i) improve existing social assistance programs and (ii) promote earnings opportunities for the poor utilizing a wider set of policy instruments. A strategic direction at the policy level is to develop an integrated and comprehensive social protection and labor system, which will facilitate the operationalization of NSSS. Promoting a systems approach to social protection and labor does not mean that all programs should be implemented by a single ministry under a single program. A systems approach emphasizes integrating and harmonizing fragmented programs, and improving implementation efficiency by establishing common administrative systems. Moreover, it promotes a balanced development and utilization of diverse policy instruments to achieve three intertwined objectives (resilience, equity, and opportunity). Therefore, a comprehensive SPL system for Bangladesh will require (i) improving existing social assistance programs; and (ii) utilizing a wider set of policy instruments to promote earnings and jobs opportunities with a strong emphasis on “opportunity.” Immediate priorities in existing social safety nets: The GoB recognizes that the main challenges in implementing social safety nets relate to improving coverage and targeting and reducing leakages. The Government’s NSSS prioritizes a few areas of actions. There is a need to consolidate programs; establish a unified targeting mechanism; improve monitoring and evaluation tools and practice; and use technology to increase administrative efficiency. Several initiatives have already begun to address these issues. Consolidation: A first step to address fragmentation among similar programs and rationalize the delivery of scarce resources would be to consolidate programs with overlapping objectives and functions. Small-scale programs, particularly those with untargeted, universal subsidies, must be carefully reviewed for consolidation. In doing so, risks along the lifecycle must be considered. Unified targeting system: A unified targeting system, which enables various programs to identify the deserving poor, can significantly improve the accuracy of targeting as well as program efficiency. The GoB has taken steps toward fulfilling its commitment to develop a National Household Database, which will contain key household characteristics to determine the poverty status of all households based on the Proxy Means Test (PMT) methodology. Improved efficiency: Weak administrative capacity and systems introduce inefficiencies and increase the scope for fraud and corruption. This is particularly prominent in food-based programs. Thus, it is important to take steps to strengthen the administrative systems that are in place for safety nets while phasing out food-based programs. Technology can be employed for this purpose. As one example, a dynamic 12 Management Information System (MIS) allows record-keeping and consolidation of information on various aspects of program implementation, such as beneficiary lists and numbers, utilization of program expenditures, and grievances and redresses. Monitoring and evaluation: Monitoring programs provide information on the amount of financial resources that have been spent, the number of beneficiaries the program is reaching, and the efficiency with which the program is serving them. Program evaluation is usually an external assessment that determines whether the program has met its defined standard of effectiveness and estimates its net results and impact. To date, the practice of monitoring and evaluating program effectiveness—let alone rigorous impact evaluation—is not very common in Bangladesh, pointing to the need for improvements in this area. Further considerations for safety nets: There is general consensus in the field through international experience that well-targeted cash transfers can be effective as an anti-poverty strategy, that public works can be effective as a crisis response, and that ALMPs can alleviate widespread employment challenges, and this is supported by rigorous evaluations. However, there are several discussions and debates regarding the best practices or design of these programs. Some prominent discussions relevant for Bangladesh need to be considered for future SPL programming include: conditional versus unconditional cash transfers; cash versus in-kind benefits; public works versus public works plus; and rural versus urban. Evidence is still mixed, and more experiments and learning are warranted. Utilizing a wider set of SPL policy instruments: In order to strengthen the overall SPL system to reach beyond social assistance, it is necessary to explore and utilize various policy instruments. Given the four challenges identified above, this review discusses a menu of options and potential programs, although the programs described are not exhaustive. Set out below (Table ES1) are the main challenges and the priority areas for addressing each: 13 Table ES1. Main Challenges and Priority Areas of Policy Intervention Main challenges Priority areas of policy intervention Poverty and - Voluntary insurance/saving schemes should be promoted in the short run vulnerability while efforts to establish a comprehensive pension system or universal health insurance are to be made in the long run. - There is strong evidence that savings improve individuals' ability to manage risks by turning irregular cash flows into lump sums for future expenditures when needed. - Given the market penetration of MFIs in Bangladesh, there is substantial potential to improve financial inclusion and promote more savings for the poor. - Marketing strategies that emphasize the cost of not saving, saving technologies that lower the transaction costs, and frequent nudges and reminders to save, can substantially improve saving outcomes of the poor in a cost-effective way. Youth employment - Lack of skills is considered to be a key constraint to gaining productive employment, and skills training can improve the prospects of youth who are struggling in the labor market. - International evidence, however, shows that training alone would not significantly improve the employability of vulnerable youth. - Comprehensive programs offering practical on-the-job experience and job search assistance in addition to in-class and life skills training can be successful tools. - Since many workers find self-employment as their major source of income, initiatives to promote self-employment and business development among vulnerable youth may be introduced, albeit evidence on global best practices is still scarce. Female participation - Programs must recognize and address extra constraints that women face in in the labor market gaining employment, and should be designed to meet the needs of women. - Vicinity and accessibility to female friendly employment opportunities such as the RMG sector factories could substantially improve female employment, which would lead to delays in marriage and childbearing, and improvements in wellbeing. - Reforms in family law or property rights for women as well as proper implementation and better awareness could also improve women’s access to productive assets and labor market activities. - From the labor supply side, the two notable constraints faced by women are household responsibilities and mobility constraints. Thus, reducing the cost of employment outside the house associated with these two factors can potentially result in increased labor force participation. Risks in migration - The main policy agenda is to make migration more (i) remunerative; (ii) affordable; (iii) inclusive; and (iv) safe, as part of the broad jobs strategy. - Building upon previous experience, the GoB can expand their Government to Government (G2G) programs to other countries that are already in high demand for Bangladeshi labor. 14 - Given the large upfront costs associated with arranging migration and high rates of failure, there is an urgent need to make the migration process safe and transparent for migrant workers. - In line with the overall jobs strategy, facilitating skills development for better paying jobs can substantially improve the development impacts of migration. A longer-term job strategy for the poor is needed as part of a comprehensive SPL system policy. The challenge of poverty reduction and inclusive growth in Bangladesh is not amenable to simple solutions. The country faces an intricate web of challenges and opportunities that lie ahead in the rapidly changing environment. In order to continue to reduce poverty while sustaining continued growth, the key issue is to extend policy attention beyond providing safety nets and coping with risks ex-post toward promoting opportunities and reducing risks ex-ante. To be effective, policies must include early intervention activities that start with maternal health and nutrition and lead to early childhood development, as emphasized by international evidence and also suggested by NSSS. Concurrently, it is necessary to continue developing a balanced set of policy tools available to SPL to maximize the impacts on poverty reduction. A particular emphasis has to center on creating more, better, and inclusive job opportunities. This will require both longer-term strategy and short-term action plans. A longer-term jobs strategy can derive from careful analytical work and strategic thinking. The daunting list of policy agenda items includes, but is not limited to, labor mobility, business environment, structural transformation, infrastructure, climate change, and skills. A jobs strategy may benefit from the use of an in-depth Jobs Diagnostic, which analyzes the current state of both the demand and supply side of jobs and delves into jobs strategies. While the medium- to long-term jobs strategy will strengthen the country’s labor market moving forward, GoB needs short-term action plans specifically targeting the poor and vulnerable so that they can improve their earnings opportunities as well as productivity in order for the country to achieve its goal of becoming a middle income country by 2021. In this review, we discuss the strategic importance of paying policy attention to the poor, youth, women, and migrant workers. Appropriately designed interventions that will address the challenges of these groups could complement existing safety nets and contribute to poverty reduction and inclusive growth. For Bangladesh to harness the benefits of demographic dividends and improved human development, utilize the globalized markets, and seize the opportunities for shared prosperity, immediate actions are urgently needed. 15 16 Introduction Over the past decade, Bangladesh has achieved strong economic growth and impressive poverty reduction. Between 2000 and 2014, the average growth rate per year was close to 6 percent. 1 Purchasing Power Parity (PPP) adjusted per capita gross national income (GNI) has increased from US$1,140 to US$3,330 between 2000 and 2014. Driven by strong economic performance, the percentage of the population living under the national poverty line declined from 49 to 32 percent between 2000 and 2010, and the number of poor people was accordingly reduced by 16 million (from 63 to 47 million) during the same period.2 Economic growth over the past decade has been slightly progressive; the consumption growth of households in the bottom 40 percent of the income distribution was slightly higher than that of average or upper quintile households. Along with economic growth, Bangladesh has made great strides in improving human development and achieving good labor market outcomes. Fertility rates and population growth have steadily declined in Bangladesh. The Total Fertility Rate (TFR) in Bangladesh declined from 3.1 to 2.3 between 2000 and 2014. As a result, the dependency ratio has declined while the share of working age population increased, which is favorable to reaping the demographic dividend. Educational enrollment, particularly at the primary and secondary levels, has significantly improved and the gender gap in education significantly narrowed. Structural transformation from agriculture to the industry and service sectors is progressing slowly but steadily, in part due to the continued contribution from the Ready Made Garment (RMG) sector. Improvement in labor productivity has played an important role in boosting income per capita in the country. Moreover, opportunities for overseas employment are expanding and remittances from migrant workers will likely continue to be an important source of income. Despite these impressive achievements, however, several challenging issues and risks remain that require policy intervention. First, poverty and vulnerability remains high due to income shocks and limited access to risk management for the poor. Large segments of the population are routinely exposed to seasonal and natural disasters, making the poor and vulnerable more prone to deprivation. With the majority of the poor engaged in agriculture or informal self-employment, volatility in incomes is high while the ability to cope with risks is limited. Second, with a demographic transition, the job challenge among young people is high. Youth (15–24 years of age) make up about 30 percent of the working age population, and the large number of young people entering the labor force poses considerable pressure to create jobs. Third, increasing the labor force participation and improving labor market outcomes for women remains a great challenge. Despite strong improvements in female literacy levels, education, and employment in selected sectors, overall labor force participation by women in Bangladesh remains low; a high share of working women engages in non-paid family work. Fourth, a large number of low-skilled migrants and 1 The growth rates are higher in more recent years with the average rate greater than 6 percent per year from 2011-15. 2 Some estimates suggest that the poverty rate fell to 24 percent in 2015. However, given the latest Household Income and Expenditure Survey was conducted in 2010 based upon which the official poverty rate is calculated, the 2010 poverty rate is used as a reference point in this document. 17 aspiring migrant workers are exposed to substantial risks. Given the integration of global markets and rapid urbanization in Bangladesh, migration provides employment and earnings opportunities to many poor workers. However, many risks associated with the migration process, including lack of information, and large financial costs associated with recruiting agencies and intermediaries can substantially compromise the welfare of migrants and their families. Social protection and labor is increasingly an important part of the government’s strategy to fight poverty and promote inclusive growth. For instance, the National Sustainable Development Strategy (NSDS) 2010– 21, emphasized the need for inclusive growth and the role for social protection. In line with the NSDS, the National Social Security Strategy (NSSS) of Bangladesh in 2015 articulates the way forward to improve the social security programs in Bangladesh. The Seventh Five Year Plan (2016-20) outlines the implementation of NSSS as a core goal. Correspondingly, this document reviews the key issues and challenges and presents strategies and implementation plans for social security and risk management programs to reduce poverty. In particular, it emphasizes the need to provide appropriate protection against risks along the lifecycle while paying attention to seasonal shocks and disadvantaged population groups. Recognizing numerous challenges and issues with the current social security programming in Bangladesh, the NSSS also highlights the need for consolidating fragmented programs and introducing modern implementation arrangements to improve efficiency. Complementing the NSSS, this Social Protection and Labor (SPL) review presents the current state of the SPL agenda and programming in Bangladesh from a different angle. While the NSSS focuses on risk management particularly along the lifecycle, this review adopts a framework, which emphasizes the outcomes of an effective SPL programming: resilience, equity, and opportunity.3 This approach promotes a SPL system development with the SPL policy, programs, and administrative tools aligned to achieve the three intertwined objectives of SPL: (i) resilience by helping people rebound from setbacks from different types of shocks; (ii) equity by reducing poverty and destitution and by promoting equality of opportunity; and (iii) opportunity by building human capital, assets, and access to jobs, which in turn reduce the susceptibility to shocks and build resilience. This review describes a diverse set of policy instruments that correspond to each objective. This assessment aims to improve the understanding of the current state of SPL programming in Bangladesh and enhance the analytical underpinning of policy discussion in SPL. Therefore, it complements the NSSS in building a stronger case for strengthening and improving this sector. The review will contribute to raising awareness and developing a consensus among stakeholders so that concerted efforts can be made to assist the GoB in operationalizing the NSSS. Furthermore, this review will help identify areas that require additional development and guide the Bank’s operation in this area in the coming years. This review has five sections. Section 1 briefly discusses the conceptual framework of the review which complements the NSSS. Section 2 presents the context in which the discussion of social protection and labor is taking place and hones in on its macroeconomic context, labor market condition, and global 3 This framework is based on the World Bank 2012–2022 Social Protection and Labor Strategy. See World Bank (2012a). 18 environment associated with international migration. Section 3 describes and analyzes current social protection programming in Bangladesh, and provides comprehensive coverage on the topics of social insurance, social safety nets, microfinance, and labor policy and programs. It also highlights actions that will be required to develop a more efficient social protection system in the future. Section 4 identifies a few key challenges that the current SPL programming has not adequately addressed but require urgent intervention. Finally, Section 5 proposes policy recommendations to address the identified key challenges. The last part concludes the review. 19 20 1 Social Protection and Labor 1.1 What is a Social Protection and Labor System? SPL systems are a set of policies and programs for the poor and Box 1. Target populations of SPL vulnerable and aim to achieve three objectives: resilience, equity, and opportunity. Social safety nets – essentially non- The target populations of SPL systems include: contributory resource transfers in cash or in-kind that facilitate  The extreme poor—typically defined as access to goods and services, and public works – are an those in the bottom 20 percent of the important element in a broader SPL system that may include income distribution, or those under the food poverty line social insurance and labor market policies.  The poor—typically defined as those bottom 40 percent of the income Many developing countries, including Bangladesh, focus on distribution, or all under the poverty line  The vulnerable—typically defined as safety nets as a main tool to reduce poverty and manage risks. those above the poverty line but “at risk” Over 140 countries currently operate more than 450 safety net of becoming poor due to various shocks programs, which cover 345 million people according to a recent  Those in need of social care—typically defined as the individuals at risk, estimate by the World Bank. 4 Over the last decade, these including orphans, mothers and children programs have become part of the central policy agenda to at risk, the elderly, and persons with reduce extreme poverty and promote shared prosperity. disabilities. However, in order to achieve poverty reduction in a more sustainable way, there is a strong need to explore and utilize other SPL instruments while strengthening and complementing safety net programs. In particular, labor market programs that can provide means to support the poor in a more sustainable manner and emphasize the role of the SPL system as a springboard as well as a safety net, deserve more policy attention. The following sections review the role of SPL and the scope of SPL instruments. 1.2 What is the Role of Social Protection and Labor? Individuals and households face various risks — systematic shocks that affect the aggregate performance of an economy or broad regions of an economy, such as natural disasters, conflicts, and economic crises as well as idiosyncratic shocks that affect individuals or individual households, such as illness, layoff, or death. These risks, unless well managed, expose individuals and households to the likelihood of falling into poverty and reduce the chance of escaping poverty. The poor are disproportionately more vulnerable to risks than the non-poor due to their living environment. That is, poorer people tend to be more exposed to risks that can lead to accidents or face living conditions that make them more prone to catching diseases. They have fewer resources, such as savings, to mitigate the consequences of risks, and they are less resilient because their ability to quickly re-enter the labor market is compromised. Moreover, the negative impacts of shocks may last longer for the poor because their main way of coping with these shocks is to reduce their investment in their children’s human capital, or draw down assets that they will subsequently 4The definition and counting of safety net programs may vary across countries. These figures came from The State of Social Safety Nets, 2014. See Gentilini et al. (2014). 21 find difficult to replace. Systematic—or covariate—shocks are risky because they render traditional informal social protection mechanisms useless. If all or most families of a village or a region are affected by the same drought, traditional inter-household lending mechanisms might be less effective or not effective at all. The role of SPL is to help individuals and households escape extreme poverty and manage risks, mitigate the negative consequences of shocks, and make productive investments toward reduced risks and vulnerability in the future. While sustained and equitable economic growth and the creation of well- paying jobs are the main drivers of poverty reduction and economic development, well-designed SPL systems can play an important role in equipping citizens to participate in the benefits of growth. A comprehensive social protection and labor system includes measures that strengthen resilience among the poor and vulnerable by mitigating shocks to their well-being such as unemployment or life-cycle changes including old age; improve equity by alleviating poverty and destitution through appropriate support systems; and promote opportunity by building human capital and facilitating access to higher earnings and better livelihoods. The three objectives need to be balanced in a SPL system given their potential trade-offs and inter-connection. An increasing body of literature assessing the impacts of SPL policies and programs attest that SPL has been a robust tool to fight poverty and reduce inequality. SPL provides income support in cash and in-kind services to prevent the very poor from becoming indigent; and it supports consumption smoothing when poor and vulnerable families are facing the impacts of economic and environmental shocks. SPL programs provide risk-reducing insurance that protects against poverty due to old age and disability and encourages investment in productive high-risk, high-return activities. SPL programs support human capital development by facilitating access to basic social services, often through conditional cash transfers and vouchers; facilitate entry into the labor market through active labor market programs such as skills training, job search assistance, and supportive labor market policies, particularly for youth; and provide the means to facilitate entrepreneurship through micro-business initiatives. In addition, the programs include social care, with a particular focus on assisting the elderly, women, persons with disabilities, and children-at-risk. 22 1.3 What are SPL Policy Instruments? SPL programs conventionally consist of a range of instruments for achieving the three core objectives. A taxonomy of these instruments, grouped according to whether they promote resilience, equity, or opportunity, would be as follows:  Programs to strengthen resilience usually include Box 2. SPL Policy Instruments mandatory or voluntary contributory programs such The objectives and policy instruments of an as social insurance for workers. They protect SPL system include: program participants (and their families) against loss of income due to illness, unemployment, old age,  Resilience for the vulnerable—social insurance disability, or death of main earner. They are financed  Equity for the poor—safety nets (social largely from payroll taxes, but social contributions assistance) and allocations from the general budget are also  Opportunity for all—labor market policy and programs used.  Programs to improve equity — safety nets or social assistance — are noncontributory transfers that are targeted in some manner to the poor or vulnerable. They are meant to help relieve those households that fall into poverty because of some external event or shock, and provide assistance (or sometimes a minimum income) to the more chronically poor. These programs include poverty targeted cash or in-kind transfers, either unconditional or conditional if recipients comply with certain behaviors, usually related to education or health; price subsidies, often for food or energy;5 labor- intensive public works schemes; fee waivers for essential services such as health care, schooling, utilities, or transport; or disaster relief. These programs are usually funded from general budget revenues (or off-budget donor-financing).  Programs to promote opportunity include human development-type interventions, such as active labor market programs that facilitate insertion and re-insertion into the labor market and promote better earnings opportunities by upgrading worker skills and facilitating occupational or geographical mobility; labor policies and regulations that require basic job protection to workers and govern employer-employee relations, or provide income support to the unemployed through unemployment insurance; and social care services and community development initiatives. Programs are usually budget-financed, although services are not necessarily provided by public sector agencies. 1.4 Conceptual Frameworks: A SPL System Approach Most developing countries generally do have SPL programs, but in many cases, they are fragmented and un-harmonized. Often, many parts of governments introduce their own programs at various times, as needed, with little coordination among each other. Each program has its own targeting, identification, enrollment, and service delivery and payment mechanisms. Programs often overlap in their objectives although they may serve different groups, and operate on a small scale without leveraging economies of 5These price subsidies are classified as programs to improve equity. However, in practice, many untargeted subsidies are regressive because they assist those who could afford to purchase and consume. 23 scale. This kind of fragmentation makes it difficult to regularly monitor the adequacy of targeting, coverage, benefits, and expenditures, let alone the impacts of the programs. Therefore, developing a SPL system will reduce inefficiencies and also ensure more equitable and transparent delivery of benefits. Adopting a systems approach in SPL does not mean that all programs and instruments fall under a single national initiative or one implementing agency. Rather, this approach aims to establish a coherent and long-term SPL policy to prioritize areas of intervention, to strategize how to utilize various SP instruments, and to develop a common administrative system to facilitate the process of case management. These activities do not need to be sequential; they can take place concurrently. Initial steps can involve taking advantage of new technologies that facilitate a national identification system, Management Information System (MIS), or the database for socioeconomic status, in order to develop an efficient administrative platform. Harmonizing and streamlining various programs towards consolidating duplicative programs, and strengthening each program is a key pillar for establishing a system. In prioritizing areas of intervention with adequate tools, a systems view as well as a life cycle approach emphasizing the need for a balanced development of policy instruments, provides a useful and complementary framework (Figure 1). Bangladesh’s NSSS adopts a lifecycle approach to reduce poverty, vulnerability and marginalization, and at each stage of the lifecycle, various policy tools can be used to promote equity, resilience, and opportunity, as suggested in the World Bank SPL strategy. Based on these two complementary conceptual frameworks, it is necessary to strengthen the SPL system in Bangladesh to be more comprehensive, which requires (i) improving current social safety net programs; and (ii) exploring and developing other policy instruments, particularly those promoting opportunities for human capital accumulation and access to productive work. Currently, risk management through jobs opportunities during the youth and working age stages in NSSS is relatively muted. Active labor market programs that will facilitate employment and earnings opportunities for the poor, youth, women, and migrant workers deserve more careful policy attention and intervention. Therefore the review emphasizes the need to expand policy attention from equity and resilience towards opportunity. Figure 1. Conceptual Framework Source: The Government of Bangladesh (2015); World Bank (2012a). 24 2 Context: Bangladesh’s SPL policy environment Despite strong economic growth, rates of chronic poverty remain high and millions of people live in extreme poverty or vulnerability. The number of poor individuals decreased from 63 million in 2000 to 47 million in 2010, but close to one-third of the population still lived below the national poverty line in 2010.6 Moreover, risks faced by these poor individuals are substantial and increasing, particularly in the labor market. Agriculture, which still provides jobs for more than half of the active labor force, is exposed to seasonal volatility and weather-related shocks. The industry sector, one of the growth drivers led by the RMG manufacturing, is linked to global export markets and price competition. With increased integration with global markets, the transmission of international economic shocks is stronger and the associated effects on the population are likely beyond the control of Bangladeshi policies.7 Likewise, international migration and remittances, amounting to more than 10 percent of GDP, are open to uncertainties particularly related to receiving countries’ policies and practices. Due to these challenges and issues, SPL will need to play a greater role in Bangladesh in the coming years. The following sections assess the macroeconomic context, the labor market conditions, and changing environment associated with labor mobility. Drawing on analyses of existing data and previous reports, sections 2.1 through 2.3 look at the broad context in Bangladesh, based upon which the discussion on SPL takes place. 2.1 Macroeconomic Environment Bangladesh has been enjoying steady economic Box 3. Macroeconomic Context growth: between 2000 and 2013 the economy grew at a higher pace than it did in the 1990s, averaging 5.9 Some positive and negative trends in recent years in the Bangladeshi economy are as follows: percent annually. PPP adjusted per capita GNI has almost tripled over 13 years from US$1,140 in 2000  Pros: Steady economic growth; poverty (World Development Indicator (WDI), 2014). As a reduction; slight improvement in inequality; continued growth in the industry sector; and consequence, the percentage of the Bangladeshi strong economic fundamentals and favorable population living in poverty fell from 49 to 31 percent global economy. between 2000 and 2010 when poverty is defined in  Cons: A large number of the poor and accordance with the national poverty line.8 vulnerable; uneven benefits from economic growth; political instability; and limited ability to create more and better jobs. 6 World Bank (2013a) 7 The transmission of global shocks and impacts of the more open economy were manifested during the international food, fuel, and financial crisis in 2008–10, which adversely affected the Bangladeshi economy. 8 The national poverty line here is the Upper poverty line described in the 2013 poverty assessment. See World Bank (2013a). 25 Figure 2. Average per Capita Household Consumption Source: PovCalNet. World Bank Growth over the past 10 years has been slightly progressive. The growth rate of monthly consumption of households in the bottom 40 percent of the income distribution was slightly higher than that of average or upper quintile households (See Figure 2). Moreover, the declining Gini coefficients suggest a modest reduction in inequality has been achieved in Bangladesh particularly in urban areas, although the absolute level of inequality is greater in urban than in rural areas (Figure 3). The gap between richer and poorer households has slightly declined. However, differences in living standards between rural and urban households, and between the western and eastern parts of the country, seem persistent (2013 Poverty Assessment). Figure 3. Lorenz Curve and Gini Coefficients over Time 2000 2005 2010 Lorenz curve Lorenz curve Lorenz curve Bangladesh Bangladesh Bangladesh 1 1 1 .1 .2 .3 .4 .5 .6 .7 .8 .9 .1 .2 .3 .4 .5 .6 .7 .8 .9 .1 .2 .3 .4 .5 .6 .7 .8 .9 0 0 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Cumulative population proportion Cumulative population proportion Cumulative population proportion Rural Urban Line of perfect equality Rural Urban Line of perfect equality Rural Urban Line of perfect equality Gini National 0.307 0.310 0.299 Rural 0.271 0.278 0.273 Urban 0.368 0.353 0.330 Source: 2013 Poverty Assessment (World Bank, 2013a) Note: The Lorenz curve shows a cumulative incidence of income/wealth of households reflecting inequality, and the Gini coefficient is obtained from the Lorenz curve. The 45 degree indicates an even distribution of income/wealth across all households and the Gini coefficient in this case is 0. The more severe curvature indicates greater inequality with increases in the Gini coefficient. 26 Strong macroeconomic fundamentals through a prudent management of fiscal and monetary policy, contributing to macroeconomic stability, as well as a recovering global economy and favorable food prices, have helped the country achieve high growth. Select macroeconomic indicators and the Country Policy and Institutional Assessment (CPIA) scores confirm that current macroeconomic conditions and the policy environment are stable although inflation requires close monitoring, and capital formation should be promoted (Table 1). With respect to the business environment, indicators and CPIA ratings that reflect the investment climate suggest that starting and operating a business in Bangladesh is difficult and costly, which is likely to be a deterrent for Foreign Direct Investment (FDI). Table 1. Select Indicators and CPIA Rating Banglad South Developing esh Asia Countries Macro indicators GDP growth (annual, %) 6.0 6.6 5.0 Inflation (consumer price, %) 7.5 6.5 4.1 Public debt stock (% of GNI) 20.6 22.3 22.5 FDI (% of GDP) 1.2 1.4 3.0 Gross capital formation (% of GDP) 26.8 30.1 31.2 Business environment indicators Days to start a business 21.8 16 25.0 Cost of business startup procedures (% of GNI per capita) 18.8 14.7 36.9 Firms expected to give gifts in meetings with tax officials (% 41.0 -- -- of firms) Number of tax payments per year 21 31 31 Ease of doing business index ranking among 181 countries 170 134 -- Start-up procedures to register a business (number) 9 8 7 CPIA ratings (1=low to 6=high) Macro management rating 3.5 3.3 3.6 Debt policy rating 3.5 3.3 3.4 Fiscal policy rating 3.5 2.9 3.3 Business regulatory environment rating 3.5 3.4 3.2 Financial sector rating 2.5 3.1 3.0 Property rights and rule-based governance rating 3.0 2.9 2.9 Source: World Development Indicators, 2014 27 Several challenges affect the country’s ability to create more, better, and inclusive jobs.9 The 2007 and 2013 Bangladesh Enterprise Survey (BES) suggest that political instability, the lack of reliable access to electricity, limited access to finance, and corruption are major challenges to business (Figure 4).10 Due to the change in the questionnaire between the two waves of the Enterprise Surveys, direct comparison is not possible. However, political instability highlighted in the 2013 survey is quite alarming and the persistence of issues such as electricity shortages, corruption, and access to finance calls for attention. These issues pertain not only to small enterprises, but are prevalent across firms of all sizes. Figure 4. Major Business Constraints A. 2007 Enterprise Survey B. 2013 Enterprise Survey Electricity Corruption Access to finance Tax administration Inadequately educated workforce Tax rates Courts system Practices of informal competitors Crime, theft and disorder Customs and trade regulations Business licensing and permits Transportation Labor regulations 0 10 20 30 40 50 60 70 80 90 Bangladesh South Asia All Source: The 2007 Enterprise Survey results from 2013 Poverty Assessment; the 2013 Enterprise Survey results from www.enterprisesurveys.org Political instability such as recurrent hartals (political blockade and strikes) is likely to further weaken the business environment.11 Disruptions in transport, business, and supply chains, due to continued blockades have increased uncertainty and the cost of doing business and hampered the potential for job creation. All sectors in the economy were reportedly affected by the ongoing political turbulence. Although political 9 Better jobs can be defined primarily as those with higher wages for wage workers and lower poverty levels for the self- employed, and secondarily as jobs that reduce the risk of low and uncertain income for the most vulnerable group of workers as defined in World Bank (2012b). 10 See data description in the Appendix for more information on the Enterprise Surveys. 11 According to World Bank estimates, around 1% of GDP might have been lost during the political turmoil occurring in early 2015 (World Bank, 2015c). 28 tension and standstill in Bangladesh is not a new phenomenon, continued gridlocks and business disruptions are likely to take a toll on the economy. 2.2 Labor Market Context In order to understand the labor market context of Bangladesh, we begin with a conventional approach based on a very simple identity. GDP per capita is a product of three parts: output per worker; share of workers out of working age population (employment ratio); and the share of the working age population among the total population, as set out below. = × × (1) where represents aggregate income, is population, is the number of employed workers, and is the size of the working age population. Taking logs and differences of equation (1), we have the following equation that divides output per capita growth into components attributable to changes in the quality of labor (labor productivity), quantity of labor (the employment ratio), and demography (the share of working age population). ∆ = ∆ log + ∆ log + ∆log (2) Growth in per Growth in labor Growth in Growth in the capita output productivity employment share of working ratio age population Between 2003 and 2010, GDPinper Growth capita grew from US$385 to US$535 by almost 40 percent.12 Over the labor Growthfrom productivity ratio changed little same period, the employment in Growth 56.4 to 58.8, and in the the share of working age employment share of working population increased from 56.3 to 60.4. Based on the ratio equations (1) and (2), this translates into an increase age population in labor productivity from US$1,210 to 1,506. That is, about two thirds of per capita GDP growth over this Box 4: Labor Market Context Some noteworthy trends in the labor market context are as follows:  Demography: A steady decline in fertility and slowed population growth; an increasing share of the working age population; a brief window of opportunity to take advantage of the favorable demographic structure before the aging of the population becomes an issue.  Employment: Prevalence of non-wage employment including day labor and self-employment; high inactivity and unpaid work rates among women; a significant gap between urban and rural areas; and slow structural transformation.  Education and skills: Labor productivity depends on various factors, but most notably education and skills. An analysis of education and skills showed that more than 40 percent of the working age population remained illiterate (in 2010) despite significant progress; and there were serious concerns about the quality of education and relevance of skills acquired in education systems for labor markets. 12 Data from WDI. GDP per capita is measured in constant USD 2005. 29 period is explained by growth in labor productivity while the remaining part is explained by favorable demographic change and employment increases. The increase in labor productivity is largely driven by productivity gain (about 76%) rather than inter-sectoral shifts, and the productivity gain is the largest in the industry sector (World Bank, 2013a). We discuss each of these points in more detail in the following sections. 2.2.1 Demography Population growth has considerably slowed down largely due to a decline in fertility rates. The fertility and population growth rates in Bangladesh are currently among the lowest in South Asia (Figure 5). In 2012, the TFR in Bangladesh was 2.2, indicating that the average number of children born to a woman of reproductive age is about two children—close to the replacement rate. 13 Between 2000 and 2008, Bangladesh experienced a steady decline in population growth, although between 2009 and 2012, there was a slight rise in population growth with the fertility rate leveling off and decreased mortality rates. With the reduction in the fertility rate and population growth, the overall dependency ratio has declined while the share of the working age population has increased. 14 Bangladesh’s dependency ratio has gradually declined from 70 percent in 2000 to 53 percent in 2013. 15 A gradually increasing adult population and a declining fertility rate have been important factors in reducing the dependency ratio. Having a larger workforce than dependents helps the economy grow as long as the workforce is engaged in productive income-generating activities. The dependency ratio is projected to continue to fall for another 10 to 20 years (UNFPA, 2015). Therefore, Bangladesh’s main labor market challenge will be to take advantage of the benefits of having a reduced dependency ratio and a higher proportion of the population of working age by ensuring that they have opportunities to engage in income-generating activities. 13 Bangladesh Demographic and Health Survey suggests that the TFR was 2.3 in 2014. 14 The dependency ratio is defined as the number of young and old populations over the number of working age (15 to 64 years old) population. 15 World Development Indicators (WDI). This is equivalent to the change in the share of working age population from 58.8 to 65.3 percent between 2000 and 2012. 30 Figure 5. Total Fertility Rates and Population Growth in South Asia Source: World Development Indicator (WDI) It is important to note that the mere size of the country’s population and the number of workers that need to be absorbed in the labor market poses a formidable challenge. As of 2014, the total population of Bangladesh reached about 158 million, which positioned Bangladesh as the world’s 8th most populous and one of the most densely inhabited countries.16 Almost 2 million people enter the labor market each year. In 2010, the total number of working age individuals, both men and women, was 90 million or, 60 percent of the total population. With the increase in the number of women participating in the labor force along with the increase in working age individuals, the rate of growth in the labor force (about 3.3 percent a year) is indeed greater than overall population growth rate (about 1.2 percent a year). Creating jobs for these individuals will require an increased focus on ensuring continued macroeconomic stability and improvements in the investment climate, notably the ease of doing business. In addition, in the coming years, it will be essential to equip the labor force with better and right skills for improved productivity; improve the labor market matching process to reduce job search and transaction costs; and promote risk taking and entrepreneurial activities. Due to the significant and rapid decline in fertility rates, population aging soon will be critical to address on the policy agenda. However, a current priority is to provide job opportunities and chances for improved productivity to young people and take advantage of the favorable demographic structure. The policy agenda of international migration should be also promoted in this context of utilizing a brief window of opportunity while the country has abundant young workers before facing population aging. 2.2.2 Employment Over the last 10 years, the overall employment ratio has held steady at around 59 percent (59.6 percent in 2005, 58.8 percent in 2010, and 57.1 percent in 2013). 17 However, these overall numbers mask compositional changes in the trend of employment among different population groups. For example, 16 http://www.worldometers.info/world-population/population-by-country/ 17 Employment ratio is defined as the number of the employed over the number of working age population. 31 there is a striking gender difference in the employment ratio. The employment ratios of working age males has fluctuated from 88 to 82 to 81 percent from 2005 to 2010 to 2013; by comparison, that of females changed from 30 to 36 to 33 percent.18 Figure 6. Employment Ratios at a Glance Source: Labor Force Survey (LFS) 2013 The employment outcomes based on 2013 LFS vary widely across different population groups although there is little difference between urban and rural areas (Figure 6).19 It is clear that young male workers are not performing as well as male adults, while the difference between age groups among females is relatively small. The relationship between education and employment outcomes is not monotonous: primary and post-secondary education is associated with high employment ratios, while education levels between the two (secondary and higher secondary levels) as well as no education present lower employment ratios. Being married is associated with higher employment for both genders compared to the single counterparts. Finally, religion is also associated with differential employment ratios, particularly for women, in that Muslim and Hindu women (who are the majority) are less likely to work compared to women of other religions. 18Based on LFS 2005, 2010, 2013. 19While the average male employment ratio from the 2012 Time Use Survey (85%) is similar to those in 2010 and 2013 LFS (82% and 86%, respectively), the average female employment ratio from the 2012 Time Use Survey (27%) is lower than those in 2010 and 2013 LFS (36% and 33%, respectively). Given that the Time Use Survey used the same questionnaire and definition of employment as the LFS, the discrepancy may be due to seasonality and sampling. In Bangladesh, measurement of women’s employment in economic activities in surveys remains a highly inaccurate exercise (UNFPA, 2015) as many women engage in non- paid family work or production for the consumption of households, and the discrepancy may be another example of inaccuracy. 32 Figure 7. Lifecycle Pattern of Employment by Region and Gender Source: LFS 2013. The lifecycle pattern of employment also highlights employment challenges faced by different groups of workers in Bangladesh (Figure 7). Young men predominantly start out in school and transition to paid employment. By age 30, school-to-work transitions are complete for almost all men. As they grow older, the share of the self-employed increases, perhaps because workers tend to wait until they acquire sufficient skills, capital, or social networks to start a business of their own. Wage employment falls rapidly in the late 50s as male workers drop out of employment, but the shift to retirement is more gradual with the self-employed. In line with Gindling and Newhouse (2014), unpaid family work is likely a transient status for men, which almost disappears after they enter their late 30s and early 40s in most cases. Women seem to face more complex challenges. School attendance declines more rapidly for women than men. School-to-work transitions mostly to wage employment observed for men do not take place for women, and most women become inactive or work without pay once they leave school. A large share of women remains in unpaid family work throughout their lives, especially in rural areas, until they leave the 33 labor market in their late 40s. Opportunities for formal wage employment seem severely limited for rural females. An individual’s economic wellbeing and poverty status are associated with the type of labor market activities and occupations they engage in (Figure 8). For instance, the poverty rates among paid employees (salaried workers) are among the lowest, while those among day laborers are the highest. Over time, as poverty rates have decreased overall in the country, the poverty rates of workers in each employment type have declined. Day laborers and agricultural workers have experienced the largest decline in poverty levels. More diversified income sources for day laborers and improved productivity in agriculture and rising prices for agricultural products may be contributing factors to this progress. Figure 8. Extreme Poverty Rates by Employment Type Source: Household Income and Expenditure Survey (HIES) 2000, 2005, 2010 Note: Poverty was estimated using the low poverty line. For many workers in Bangladesh, however, quality of jobs is of concern. Quality of jobs can be measured by the amount and regularity of incomes, but given the data limitation, the sector, status, and formality of employment can proxy the quality of jobs. A substantial proportion of workers are either in unpaid family work or work as day laborers (Figure 9). Although this share has steadily declined for males, it has stayed around a similar level at 60 percent. The majority of workers who are engaged in income generating activities work as self-employed. The share of employers who have one or more paid employees is merely about 0.2 percent among the employed. The share of wage employment has slightly improved over time for men, but remains small for both men and women. This suggests that most workers likely face volatile earnings except for a small number of wage employees. In particular, agricultural 34 workers tend to experience seasonal poverty,20 and day laborers who have limited opportunities for a steady stream of earnings seem to be most vulnerable. Figure 9. Type of Occupations among Workers over Time: by Gender Source: LFS 2003, 2005, 2010, 2013 Although structural transformation has been taking place in Bangladesh, the pace is slow for its strong economic growth (Figure 10). The share of both males and females working in agriculture has declined by close to 10 percentage points from 2003 to 2013. Concurrently, the share of workers in the industry and service sector increased. In particular, the increase in the share of female workers in the industry sector is prominent, which is probably in part due to substantial growth in the garment industry, which absorbs much of the female labor force.21 Despite this gradual shift from agriculture to industry and service, close to half of workers are in the agriculture sector. 20 Bangladeshi agricultural workers used to face severe seasonality with agricultural lean season (called monga) known as months of extreme hunger and poverty with limited work opportunities. However, with the gradual diversification of crops (beyond boro and aman crops) and sources of incomes, the seasonality patterns have changed and severe destitution has declined. 21 From 2004 to 2014, the number of workers employed in the RMG sector has doubled from 2 to 4 million, according to the Bangladesh Garment Manufacturers and Exporters Association (BGMEA) statistics. 35 Figure 10. Employment Sector among Workers: by Gender Source: LFS 2003, 2005, 2010, and 2013 Furthermore, informality is widespread. Depending on various definitions of formality – whether the workplace is a registered entity, whether the individual worker has a written contract, or whether the individual worker has a pension contribution, the share of formal workers varies. Regardless of the definition, however, it is observed that the share of formal workers is small (Figure 11). 22 When it is based on pension contribution, for instance, merely about 5 percent of all workers in the country are currently classified as formal workers. The share of formal workers is slightly greater for men than women except it is reversed when formality is measured based on the existence of written contracts. Figure 11. Share of Formal Workers by Gender Source: LFS 2013 22Registered workplace includes government entities, NGOs, and private enterprises (including household enterprises) registered with the government authority. 36 The low quality of jobs can be attributable to the numbers of workers with a low level of skills and human capital. More educated or skilled workers not only have better job opportunities, but collectively contribute to overall development.23 Workers acquiring more skills create positive spill-over effects that make other workers more productive. With better skilled workers, technology is more easily adapted and adopted, and ultimately new technologies are invented. This process of skills development is critical in engendering technological change and economic diversification, and in enabling the creation of jobs and economic growth. The educational attainment and skills profile of the work force are discussed more in detail in the following section. 2.2.3 Education and Skills In Bangladesh, there has been significant progress in education for both genders over time and this is an important development given the close relationship between education and household income (Figure 12). The share of working age population with no education has substantially declined, and the expansion of primary and secondary education is quite prominent. Despite this impressive improvement, close to 40 percent of the working age population was still illiterate with no education in 2010. The gender gap is also persistent and significant. Moreover, only about 20 percent had a secondary school certificate (SSC) or higher education. Even the formal sector workers present a low level of education; more than a quarter of formal sector workers (28 percent) have primary education or less (World Bank, 2013e). Figure 12. Education of the Working Age Population by Gender and Consumption Deciles Source: 2013 Poverty Assessment Fundamental issues regarding the quality of education are also concerning.24 Many students in primary and secondary schools lack foundational skills that could be useful in the labor market. Pupil assessments have found that those who complete primary school perform on average at a third grade achievement 23 A vast body of literature empirically supports the value of investing in education to develop human capital and acknowledges the contribution of education to growth and development (see, for example, Krueger and Lindahl 2001; Hanushek and Kimko 2000; Hanushek and Woessmann, 2012). 24 See World Bank (2013d). 37 level and lack essential problem-solving skills. 25 In part, due to the low quality of education, school attendance has been quite low and school drop-out rates high, particularly among the poor.26 As a result, inadequate skills of the work force are often cited as a major constraint to firms, after the political instability and infrastructure issues. Access to training opportunities in the labor market is very limited. Currently, more than 3,000 Technical and Vocational Education and Training (TVET) institutions provide skills training in the country, and GoB is making efforts to strengthen the small but growing sector. However, entry to TVET institutions requires completion of grade 8 and those who are unable to complete secondary education, mostly the poor, do not have access to TVET education. Moreover, the TVET system falls short of placing their trainees in jobs with low demand for TVET graduates from the employer side resulting in poor placement rates.27 Thus, TVET graduates despite having mid-level education remain unemployed or engage in low quality jobs. The mismatches between the labor market relevant skills that employers demand and the type of skills trained and fostered through TVET, are considered to be a key issue to poor TVET sector performance. 2.3 Labor Mobility and Migration Migration plays a significant role in the Bangladeshi economy, and poses both opportunities and risks to the migrants and their families. Internal migration is mainly driven by two factors — seasonal motivation reflecting fluctuations in the demand for labor, or longer term economic considerations to find new means of livelihoods to escape poverty. Many agricultural workers join seasonal migration during the lean season, without permanently moving to urban areas but seeking temporary work opportunities. Meanwhile, poor families in rural areas who are at risk of losing their livelihoods due to river erosion, loss of lands, or impoverishment with few work opportunities, often choose to migrate to urban areas. They tend to be more permanent migrant families who likely settle in urban slums. The scale of internal migration – seasonal or permanent -- is difficult to quantify. A survey of 1,600 rural households in the northwest region of Bangladesh indicated that 19 percent of households, and 25 percent of households defined as chronically poor, were engaged in seasonal migration (World Bank, 2015a). Permanent internal migration is highly associated with the urbanization phenomenon. A study in 2007 estimated the number of permanent migrants at 12.7 million in 2004 – about 9 percent of the total population.28 Significant numbers of people continue to migrate to the main urban areas with substantial industrial development, notably in the two largest cities (Dhaka and Chittagong). Bangladesh has been experiencing rapid urbanization with urban population growing at a yearly average rate of 5 percent since the early 1970s, while the national population growth was 2.2 percent. As a result, the share of the urban population has almost doubled from 15 to 30 percent over the past 40 years (between 1973 and 2013). Rapid urbanization is tied to strong economic growth, which has led to a 25 See Tietjen (2003). 26 The primary school survival rate up to the last primary grade, based on administrative data 2008–12, is about 66 percent, indicating that a significant number of students drop out of primary school. See UNICEF (2014). 27 See World Bank (2013e) for a discussion on the findings from an Employer Survey which shows low market demand for TVET graduates. 28 See World Bank (2015a). 38 structural transformation in the economy with the gradual decline in the agricultural sector as a share of GDP. The economic activities of Dhaka and Chittagong dominate Bangladesh’s non-agricultural economy; almost half of national GDP is produced in these two cities.29 This level of economic activity in turn is associated with the thriving RMG sector, which provides over half of formal employment opportunities. With a large influx of labor into urban areas, emerging issues include urban crowding in slums. With a high population density but limited infrastructure and public services, urban slums are expanding especially in government and semi-government land, abandoned buildings/houses, and roadsides. 30 The country’s slum census conducted by the Bangladesh Bureau of Statistics (BBS) describes the characteristics of slums as shown in Box 5. Box 5. Slum Characteristics  Building structures: Small and built from very cheap materials.  High density of living areas: Many members of the household live in a limited space and many structures are closely situated  Water, sanitation, and electricity supply is limited and living conditions are poor.  Socioeconomic status of slum dwellers is very low.  Provision of services such as education is severely limited. Source: BBS (2014c) Every year, half a million new people arrive in urban areas and the share of urban poor is rising. Many poverty-stricken migrants when moved to urban areas often end up in illegal settlements on precarious land with poor living conditions. These areas have few, low-quality services, and rents offered are pricey through middlemen (known as mastaans). Slum evictions are also a constant threat and occur without a clear resettlement plan. UNICEF (2010) reports alarming statistics: the rate of birth registration is about half of the national average in slums; school dropout and repetition rates are higher; and additionally, slums have three times more child labor (13 percent) than the average. Similarly, international migration is also a two-edged sword: promising economic prospects but significant risks. From 2000 through 2010, an average of 396,000 Bangladeshis went abroad each year and this figure increased to about 500,000 between 2011 and 2015(See Figure 13). They are mainly semi-skilled and unskilled workers from the eastern part of the country. Official figures indicate that around 10 percent of the male labor force is currently migrant, and the total number of migrants to a foreign country is about 9.5 million, with remittances amounting to more than 10 percent of GDP – about US$15 billion in 2014 (data from Bureau of Manpower, Employment, and Training). It is likely that these numbers are much higher, because many migrants are believed to have gone unrecorded. Most out-migration is for relatively short-term employment; the Middle East, mainly the Gulf states, and Southeast Asia are the primary destinations. Although the stock of migrants abroad is likely to be relatively stable from one year to the 29 About 10 and 3 percent of Bangladesh’s population lives in the Dhaka and Chittagong, respectivel y. These two cities generate about 36 and 11 percent of GDP. See World Bank (2013b). 30 About 37 percent of the population of the Dhaka metropolitan area resides in slums according to the 2006 urban health survey. Between 1997 and 2014, the population living in slums increased almost three times from 710 to 2,200 (BBS, 2014). 39 next, the official data show migration abroad surged in 2007–2008, possibly due to food and fuel crises and the demand for low cost labor from foreign countries as a result of global recessions. Figure 13. Annual Migration Abroad and Size of Remittances Source: BMET website (www.bmet.gov.bd) and World Bank (2015b). International labor markets currently provide employment opportunities for many Bangladeshi workers, and a large volume of the work force can very likely be absorbed, which will be critical to harness the benefits of the demographic dividend. Although employment opportunities offered through international migration have been options reserved mainly for prime aged males and limited for youth and women (as will be discussed below), there is room to improve in this area. Given that many developed countries experience population aging and the demand for workers is increasing, exploring migration destinations beyond Gulf and other Asian countries could further promote employment opportunities through the international labor market. Migration abroad is an important Box 6. Internal and International Migration source of income for the country. The significance of internal and international migration in reducing The amount of formally received poverty and improving the living standards for households as well as remittances has increased from for the entire economy calls for greater policy attention: approximately 4 to 10 percent of  Prevalence: About 9 percent of the total population is estimated to GDP from 2000 to 2014. When be internal migrants. About 10 percent of the male labor force is remittances through informal estimated to be international migrants. However, the official figures likely underestimate the numbers of migrants because their channels are considered, the role of moves often go unrecorded. remittances is even greater. Of non-  Remittances: Remittance income significantly contributes to the formal remittances, an estimated well-being of the households; in particular, international 40 percent was received through remittances are an important source of national income 31 (amounting to about 10 percent of GDP). the informal hundi system, about 5 percent through friends and relatives, and the remainder through migrant workers on home visits 31A hundi is a traditional financial instrument used as a remittance instrument to transfer money from place to place. It takes the form of an unconditional order in writing made by a person directing another to pay a certain sum of money to a person named in the order. 40 (Siddiqui and Abrar, 2003). In addition to a greater number of migrants, some of the growth in remittances from abroad may reflect the effects of better measurement of flows, greater efficiency in the transfer mechanisms, and exchange rate fluctuations. The impact of remittances on Bangladeshi households is noticeable. This is reflected in a number of findings: in 2012, while poverty incidence at the national level was 31.5 percent, poverty incidence among migrant households was only 10 percent (ILO, 2014); a study (RMMRU, 2014a) noted that both abroad and in-country migrant households enjoyed better living standards than non-migrant households, the former significantly so; and an analysis conducted on the basis of the 2010 Household Income and Expenditure Survey (HIES) concluded that remittance receiving households were more likely to be food secure than households without remittances, with food security measured by access and availability of food (Szabo, 2015). Some of the differences are likely to reflect the effects of remittances, but they are also influenced by the fact that many registered migrants may have been from relatively better-off poor households who could afford to pay the migration costs. The study estimates that about two thirds of remittances are used for consumption, while the remainder is invested, mainly in land. 2.4 Summary Bangladesh’s policy environment highlights the need for a strong and comprehensive SPL programming. Policy measures to reduce poverty and vulnerability to ensure the benefit of strong economic growth to be widely shared; to provide jobs opportunities to reap the demographic dividends; to improve productivity and quality of jobs of those marginalized in the labor market; and to utilize both internal and global labor markets for better opportunities for the poor, while helping with various risks coping, must be considered. A well-developed SPL system can contribute to achieving these goals in Bangladesh. However, as will be seen in the following section, the current SPL programming in Bangladesh may not be well equipped to address the country’s emerging challenges and has much room to be further strengthened. 41 42 3 Review of SPL programs, expenditure, and performance in Bangladesh Many SPL programs have been implemented to help households cope with various risks, but the current SPL programming in Bangladesh is heavily focused on safety net programs. Early efforts in social protection were rolled out as emergency relief measures in response to natural calamities as food aid and public works in response to the seasonal fluctuations of the labor markets in rural areas. Many of these programs are reflected in the current structure, which consists of a large number of social assistance programs. The formal social security system in Bangladesh, which comprises social insurance (essentially contributory old-age, disability, and survivor pensions for public sector workers), approximately 145 social assistance (safety net) programs that transfer cash and in-kind benefits to various population groups, and a small number of active and passive labor market programs, presents the state of unbalanced development of SPL instruments (See Table 2). Table 2. Social Protection and Labor – the Conceptual Framework and Its Application to Bangladesh Instruments – Objective Instruments Bangladesh Social insurance (contributory pensions, Old age pensions for civil servants, by Improving resilience health insurance, unemployment insurance); far the largest program among all SPL (social insurance) severance; short-term benefits (maternity, instruments (BDT 86 billion in FY sick pay). 2014–15) Unconditional poverty-targeted cash and More than 140 programs, including in-kind transfers; conditional cash transfers; social protection and empowerment Improving equity poverty-targeted family allowances; implemented by over 20 ministries (safety nets) workfare/public works; targeted housing, (About BDT 223 billion in FY 2014–15) utility subsidies; disability benefits; non- contributory pensions and social care Very limited except for microfinance. Labor market policies; ALMPs (skills Most microfinance programs are Promoting opportunity development, wage subsidies, job search implemented by NGOs, with the (human capital development) assistance); social funds; and microfinance exception of a few by GoB, with little emphasis on labor market opportunity. The total budget for SP programs has increased over time and so has the number of programs (Table 3).32 The budget has substantially grown at a rate around 16 percent per year from FY 2008–09 to FY 2015–16. So the budget as a percentage of GDP has been around 2 percent. However, this figure includes the budget for pensions for civil servants and honorarium for Freedom Fighters (as will be discussed in more detail), which cover a very small number of beneficiaries but incur a large amount of benefits. If these are excluded, the SP budget as a percentage of GDP is around 1.4 percent, which is much lower than what it appears to be. 32 All social security programs according to the Ministry of Finance’s classification are included as SP programs here. 43 Table 3. SP Budget over Time Budget FY 2008- FY 2009- FY 2010- FY 2011- FY 2012- FY 2013- FY 2014- FY 2015- 09 10 11 12 13 14 15 16 (p) BDT billion 138.3 167.1 208.9 226.4 230.8 267.0 309.0 377.2 US$ billion 1.8 2.2 2.7 2.9 3.0 3.5 4.0 4.9 Number of active 56 77 83 92 91 129 152 145 programs SP budget as % of 1.99 2.13 2.32 2.17 1.94 1.98 2.04 2.20 GDP Source: Ministry of Finance Moreover, SPL programs have been marred with many issues including fragmentation, targeting errors, and inefficiency. The NSSS emphasizes the need to consolidate similar programs along the lifecycle and corresponding risks at each life stage. In line with this, GoB has already started implementing key aspects of the NSSS through its initiatives to improve pro-poor targeting and administrative consolidation of programs. The Safety Nets System for the Poorest project, for instance, has begun to support the five largest programs that consist of three public works and two humanitarian relief programs under the Ministry of Disaster Management and Relief, with a great emphasis on improving pro-poor targeting and synergies between the programs. Also, the Income Support Program for the Poorest project, recently introduced by the Local Government Division of the Ministry of Local Government, Rural Development & Cooperatives (MoLGRD&C), promotes poor households’ investment in children’s health and nutrition by conditional cash transfer, aligned with the NSSS’ emphasis on risk management along the lifecycle. However, limited attention has been paid to promoting labor market programs, which is a crucial element for any SPL program that aims to reduce poverty and vulnerability. This section reviews the landscape of Bangladesh’s SPL programs. It also describes key characteristics of the programs and presents critical issues that must be addressed in order to improve the impact of programs. 3.1 Social Insurance 3.1.1 Program Review Pension for Retired Government Employees and their Families is one of the few public insurance schemes in Bangladesh. The pension program is the largest component of public expenditure among those classified as the Government’s Social Security programs. The expenditure for this program in FY 2012–13 was about BDT 55 billion, equivalent to 24 percent of the total SP budget, and it increased to BDT 86 billion in FY 2014–15, representing about 27 percent of the total SP budget. This expenditure is expected to increase to 116 billion, 31 percent of the total SP budget, in FY 2015–16. Set out below, we describe how this pension as well as the honorarium for insolvent Freedom Fighters is implemented.  Pension for Retired Government Employees and their Families. When an employee retires after serving in the government for a certain period of years, he/she receives a monthly emolument for his/her maintenance or that of his/her family, during the remaining period of his/her life. A civil servant or his/her family members are entitled to one or more of the various types of pensions 44 depending on the circumstances: compensation pension, invalid (disability) pension, superannuation pension, retirement pension, and family pension.33 The current retirement age in government service is 59 years, and an individual is required to have at least 10 years of service to obtain pension rights. The replacement rate varies depending on the length of pensionable service. Generally, a government employee gets 32, 48, 64, or 80 percent of his/her basic salary upon retiring after 10, 15, 20, and 25 years of employment, respectively.  Honorarium for Insolvent Freedom Fighters. Introduced in 2000 under the Ministry of Welfare, and transferred to the Ministry of Liberation War Affairs, this program provides the monthly honorarium to each beneficiary. The amount of monthly honorarium has soared from BDT 300 in 2001 to BDT 500 in 2003, to BDT 900 in 2009, and to BDT 2,000 in 2011. As of FY 2014–15, about 200,000 beneficiaries received benefits in this program, with an annual budget of BDT 12 billion. Although the benefit amount is greater than any other SPL programs in Bangladesh, the number of beneficiaries is quite limited. Eligibility for this program is determined by the Freedom Fighters Allowance Distribution Committees at the districts, Upazillas, and City Corporations level, among those with certificates of Freedom Fighters.34 3.1.2 Key issues with social insurance Issue 1. Limited availability Many countries implement mandatory insurance and savings programs to help individuals manage risks. As discussed above, these programs include insurance for unemployment, disability, illness, longevity, or death. These schemes are often available for wage employees, particularly in the formal sector. In that case, mandatory contributions are levied on salaried employment as part of payroll taxes, which are used to finance the benefits. Typically, this model, based on the payroll system (known as “Bismarckian”), faces severe coverage and distortion issues. The coverage of the program tends to be quite low especially in an economy where a substantial part of the labor force is self-employed or works in subsistence agriculture. Moreover, the mandatory contribution of payroll taxes increases the cost of labor and tax wedges, reducing the incentives to create jobs. In response to these issues, countries often introduce a patch up remedy by providing non-contributory benefits of social insurance to those in the informal sector or the poor for equity. This type of dual system, however, further segregates the labor market and increases firms’ incentives to operate and remain in the informal sector. Currently in Bangladesh, no formal social insurance other than the pension scheme for the Government retirees is available. Mechanisms to collectively hedge various risks across the different population groups 33 The Compensation Pension is provided to officials who have lost jobs because their posts were abolished due to GoB downsizing offices or organizations. The Invalid Pension is provided if an employee becomes physically or mentally disabled, as assessed by a medical board. Superannuation pension takes place upon compulsory retirement at a certain age fixed by GoB (typically 59 years of age, but 65 years for certain occupations such as judges and public university teachers). Retirement Pension is provided when a public servant decides to retire after completing 25 years of service, regardless of age. Family Pension is survivor’s benefits to the family of the public servant who held a pension in case of his or her death. 34 Indeed, little is known about how the status of Freedom Fighters is assessed and determined, and how the benefits are distributed. This process is very political, and it can be controversial to classify this program as social insurance or assistance. 45 or to inter-temporally manage the life cycle risks do not exist. There is an urgent need to introduce social and financial programs that facilitate risk management through insurance and savings schemes. 3.2 Safety Nets 3.2.1 Program Review GoB implements a number of social safety net programs to provide cash and in-kind transfers to the poor, and expenditures on these programs have been growing steadily. The number of active social security programs, of which the majority is safety nets, has grown from 56 in FY 2008–09 to 145 in FY 2015–16. As a result, the budget allocations have increased from BDT 138 billion in FY 2008–09 to BDT 375 billion in FY 2015–16. As discussed above, excluding the two large social insurance programs, the budget as a percentage of GDP, which reflects the safety net budget, is around 1.4 percent. This figure is relatively high in the South Asia region, since expenditures on safety nets has historically been low (e.g., 0.64 percent in India; 0.49 percent in Sri Lanka), but lower than the average of 1.6 percent among developing countries.35 The cross-country comparison of the share of expenditures out of GDP may not be the best measure to assess the adequacy of the volume of safety net programs, due to the variations in the definition and scope of social safety net programs and spending. However, this cross comparison suggests that the allocation of the budget for safety nets is quite small for the daunting poverty challenges facing the country. Although there are many programs, the 9 large safety net programs (in Table 4) constitute about 46 percent of the total budget for safety net programs.36 The size of the 9 largest programs varies from BDT 6 to 16 billion. Unconditional and conditional cash transfers as well as cash-based public works (EGPP) provide cash benefits, but a considerable proportion (about two thirds) is provided in-kind with food. Moreover, while cash transfers have a clear target group and public works rely on (pro-poor) self-targeting, food-based relief programs tend to have less clear target groups. 35See World Bank (2014) for the state of social safety nets. 36The 9 largest safety net programs exclude two social insurance programs (pension for retired government employees and their families and honorarium for insolvent Freedom Fighters), which GoB classifies under social protection programs. 46 Table 4. Nine Largest Safety Net Programs in Bangladesh as of FY 2015-16 Type Program Benefits (in BDT) Number of beneficiaries Eligibility Budget allocation (FY (FY 2015-16) 2015–16) Cash transfers  Old Age Allowance  BDT 400 per month  About 3 million  Annual income less  BDT 14 billion than BDT 3,000; Individuals over the age of 65 (62 for women) Conditional cash  Stipend Program for Primary  BDT 100 (125) per month  About 7.8 million  Categorically poor¹  BDT 9 billion transfers Students (former Food for for one (more than one) students in 487 Education, FFE) pupil Upazilas  Secondary Education Quality  Monthly stipend varying  About 1.5 million  PMT based targeting  BDT 6 billion and Access Enhancement from BDT 100; 125; 160; stipend beneficiaries Project (SEQAE) 180; and 200 from 6 to in 215 Upazilas; 10 grades about 4 million students (when other interventions included) Public works  Food for Works (FFW)  Approximately rice 8kg/  About 0.55 million  Monthly income less  BDT 14 billion programs (cash and  Test Relief (TR) day or equivalent (FFW  About 0.78 million than BDT 4,000 +  BDT 13 billion in-kind transfer) and TR) categorically poor¹  Employment Generation  BDT 200/day for EGPP  About 0.78 million  Categorically poor¹  BDT 15 billion Program for the Poorest (EGPP) Food-based  Open Market Sales (OMS)  BDT 7-10 per kg/rice  About 21 million  Un-targeted  BDT 16 billion emergency/seasonal  Vulnerable Group Feeding  10-20 kg of rice per  About 6.5 million  Disaster affected  BDT 15 billion relief (VGF) month (VGF) households (VGF)  Vulnerable Group  25-30 kg of wheat or rice  About 9.1million  Ultra-poor (extreme  BDT 10 billion Development (VGD) per month (VGD) poor) women (VGD) ¹ The categorical poor include those who lack land ownership and family income and those from certain family background, subject to differ by program. Source: MoF, MoDRM, and MoE administrative data. 47  Old age allowances (Ministry of Social Welfare):37 The Old Age Allowance Program, introduced in 1998, provides monthly benefits of BDT 400 (equivalent to US$5, payable every three months) to poor individuals over the age of 65 (62 for women). Recipient's average annual income must be below BDT 3,000 (US$3 per month). Applicants for the old-age allowance are required to fill out a prescribed form, which shows his/her income level. Then two committees (ward and upazilla) consisting of two representatives of the local Member of Parliament, in coordination with each other, select the list of beneficiaries. When selecting beneficiaries, priority consideration is to be given to the landless and homeless, Freedom Fighters, and widowed females. In FY 2015–16, the total allocation amounted to BDT 14 billion, and the total number of beneficiaries were about 3 million (Ministry of Finance) out of the approximately 8.5 million population above 65 years of age. The cutoff income level, BDT 3,000 per year (or BDT 250 per month), seems very low given that the national poverty line is about BDT 1,600 per month. Compared to this cutoff, the monthly benefit of BDT 400 is generous, underscoring the importance of identifying and selecting the deserving poor in an adequate way. However, it is difficult to assess the targeting performance because typically assets may be more relevant information than incomes to assess the poverty status of the elderly, yet the information on assets is not considered for eligibility.  Primary education stipend program (Ministry of Education): The Food for Education (FFE) program, established in the early 1990s, distributed monthly food grain rations to poor households that sent their children to primary schools. FFE was terminated in 2002 and replaced by the cash-based Primary Education Stipend (PES) program. The current program is designed to provide cash assistance to poor primary school pupils and their families throughout rural Bangladesh. The program objective is to increase the enrollment and attendance rates among primary school-aged children from poor families. Poor students are defined based on five characteristics: children from distressed female-headed households (i.e., destitute widows and divorcees), children of day laborers, children of insolvent artisans/mechanics (i.e., potters, fishermen, blacksmiths, weavers, carpenters, cobblers, etc.), children from landless families (i.e., owning less than 0.5 acres of land), and children of sharecroppers. With these five characteristics as a guideline, a School Management Committee selects students as beneficiaries (Tietjen, 2003). In order to qualify for the stipend, selected students must maintain 85 percent monthly attendance and attain a minimum of 50 percent marks on the annual exam administered for each grade. Households of qualifying pupils receive BDT 100 per month for one pupil and BDT 125 per month for more than one pupil. The amount of the benefits is deemed to be quite small, and little is known about how this program impacts school attendance.  Secondary Education Quality and Access Enhancement (SEQAE) Project (Ministry of Education): The Female Secondary School Assistance Project (FSSAP), initiated in 1993, aims to improve school enrollment among girls. The primary component of FSSAP was the Stipend and Tuition program that provided monthly stipends to female students from Grade 6 to Grade 10 (between 11 and 15 years of age). With its positive results, the project was extended and expanded, and operated until 2008, 37 This is sometimes classified as a non-contributory social pension, captured in pension expenditure. In this report, the old age allowances targeted for the poor elderly are classified as social safety net. 48 contributing to improving girls’ education. Building upon the FSSAPs, the SEQAE launched in 2008. A few notable features were introduced unlike its predecessor. First, given the limited completion rates of secondary schooling for both boys and girls from poor households, this project provides stipends and tuition to both girls and boys for school attendance as well as incentives to students, teachers, and schools based on their performance. The benefit amounts to students for attendance include annual tuition and stipends, which increases with the grade reflecting the opportunity cost of attending school (i.e., BDT 1,380; 1,680; 2,100; 3,100; and 3,450 for grades 6 through 10, respectively). Also, a rigorous pro-poor targeting based on the PMT methodology was introduced to identify the poor.  Food For Work (MoDMR): The Food For Work (FFW) program’s major objectives are to generate employment for the poor during the lean season and to develop and maintain rural infrastructures. MoDMR is the principal authority in charge of this program, and resource allocations are determined based on the population size and poverty rate of the Upazila. Initially, the FFW program involved payments through rice/wheat in exchange for work; however, since 2009, a cash payment method has partially been adopted in an effort to reduce leakage of allocated funds and improve the program’s efficiency. Rather than targeting poor individuals, budget allocations are done at the Upazila level based on its population size and poverty rate. The wage rate is set at 8 kg rice (or equivalent cash) for every 7 hour work/day. The Project Implementation Committee (PIC) in the respective unions implement the project and is also responsible for beneficiary selection. The selection criteria includes household earning less than BDT 4,000 per month, household owning no more than 0.5 acres of land, household head working as day laborer, and household possessing no productive assets. In practice, however, projects are often carried out by independent contractors that are appointed by the PIC; thus, they do not always follow the beneficiary criteria. Moreover, individual participants are often unaware that they are participating in public works because recruitment also takes place through local contractors for insfrastructure projects.  Test Relief (MoDMR): Test Relief (TR) mainly aims to provide food assistance to the rural poor in exchange for their participation in maintaining and developing local institutions, which may include constructing basic infrastructure for rural educational and public welfare institutions. TR is similar to the FFW program. MoDMR is the principal authority in charge. Budget allocations are determined based on population size and the poverty rate of the Upazila. The wage rate and eligibility criteria are similar. The projects are implemented through contractors who often sell the allocated food-grains and independently hire and pay the laborers; as a result of this independence, contractors do not necessarily adhere to the rules regarding participant selection and wages. The major difference between FFW and TR is the type of projects undertaken: FFW is more concerned with developing rural roads and infrastructure, while TR projects are mainly focused on maintaining and developing local educational and religious institutions, including schools, madrasas, orphanages, mosques, temples, among others. 49  Employment Generation Program for the Poor (MoDMR): The main objective of EGPP, which was introduced in 2008 in response to the international food and fuel crisis, is to provide short-term employment to agricultural laborers during the monga/lean periods (from October to December and from March to April). During the lean periods, a large segment of the rural population remains unemployed mainly due to limited farm activities. Through EGPP, each beneficiary works for 40 days in each cycle and is paid BDT 200 for each day’s work.38 Although the rate is deemed to be lower than the market rate, the EGPP wage rate seems to be relatively generous compared to the national poverty line (BDT 1,600 per month) and other benefits discussed above. A few innovations introduced in the EGPP to improve efficiency and equity of the program are different from other public works including FFW and TR. First, pro-poor targeting is enforced with more rigor. Like other public works programs (TR and FFW in this case), participation in EGPP is determined by self targeting through the sub-market wage rate, but the eligibility criteria are simplified and imposed. The criteria state that a household cannot own more than 0.5 acres of land and the head of household must work as a day laborer. Second, payment is entirely in cash and transferred directly to the individual’s bank account. Third, the EGPP program has a quota for at least 33 percent of the total beneficiaries to be females. As a result, female participation in EGPP is higher than in other public works programs.  Open Market Sales (Ministry of Food): Open Market Sales (OMS) is a program through which coarse rice can be purchased at a subsidized price when there is a shortage of food supply. The OMS in its current form was initiatied in 2008 in response to the global food and fuel price crises. The program started with 100 trucks at various spots in Dhaka, and the OMS then expanded the number of dealers and extended its coverage to all divisional cities, 54 district towns, and 379 Upazilas, involving some 2,975 dealers (World Bank, 2013a). The program targets geographical areas where a large number of vulnerable families reside. In theory, families who do not own land or have any permanent source of income are eligible to receive a ‘fair price card,’ with which rice can be purchased. In fact, selection is done by self-targeting as OMS deals with relatively low quality, coarse rice. Each family under this operation can purchase up to 5 kilograms of rice a day or 20 kilograms of rice per month at a subsidized price of BDT 25/kg from the designated OMS outlets. 39 GoB procures rice from both local and international markets and provides it to designated dealers at a price of BDT 23.5/kg, yielding a commission of BDT 1.5/kg for dealers. The OMS, however, has difficulties in adequately regulating dealers from selling poor quality rice or cheating on the weight and amount of rice they distribute.  Vulnerable Group Feeding (MoDMR): The Vulnerable Group Feeding (VGF) program is an in-kind form of humanitarian assistance. The program’s primary objective is to mitigate the consequences of disasters such as floods, cyclones, and other natural calamities for affected families. In addition, VGF also provides rice to poor people during religious festivities (i.e., Eid-ul-Fitr and Eid-ul-Azha) and lean seasons (agricultural lean season as well as fishing lean periods). Poor households generally receive 10 to 20 kg of rice per month under this program. During the fishing lean season, poor households receive 20 to 30 kg of rice per month. Selection criteria were designed to consider numerous factors, 38 The wage rate has been adjusted periodically for inflation over the last couple of years. BDT 200 is as of 2015. 39 The current market price of coarse rice is BDT 32–35/kg (Source: Department of Agricultural Marketing) 50 such as land ownership, assets, income, marital status, and other factors; however, it is not clear how these criteria are verified and enforced. The Gratuitous Relief (GR) program, also under MoDMR, is very similar to VGF and addresses the consequences of disasters and shocks by providing about 10 to 20 kg of rice to each eligible household per month for up to six months following any disaster.  Vulnerable Group Development (Ministry of Women and Children Affairs): This program provides ultra poor women with food grain and a package of development services. Ultra poor women are defined as those from landless households or households owning less than 15 decimals (0.15 acre) of land; women with an irregular income (less than BDT 300 per capita per month) or no household income; women who are daily or casual laborers; and women from households that lack ownership of productive assets. Preference is also given to female-headed households. The program’s main objective is to ensure food and nutrition security of women and children through improved food consumption and improved awareness of the relationship between nutrition and cognitive development. Under this program, poor women receive a ration of 25 to 30 kg of wheat or rice for a period of 24 months, and a package of training in health and nutrition, in their legal and basic rights, and in income-generating activities. 3.2.2 Main issues with safety nets Safety nets play an undoubtedly important role in reducing poverty. The question is whether these programs are cost effective. Rather than analyzing individual programs to address this question, it is worthwhile to look at a few overarching issues that can hamper the cost-effectiveness of the safety net system as a whole. Table 5. Poverty and Safety Net Coverage by Division, Issue 1. Limited pro-poor targeting 2010 Poverty rate (%) Safety Net coverage Most safety net programs rely on a (%) geographical targeting system rather than Barisal 39 34 household or individual level targeting; Rajshahi 36 27 thus, they are pro-poor targeted in a Khulna 32 37 limited way. Although safety net programs Dhaka 31 19 specify that they aim to provide assistance Sylhet 28 24 to the poor or ultra-poor, it is unclear how Chittagong 26 19 Source: HIES 2010 program administrators define and identify Note: Coverage is defined as the share of households that received at least the poor, and verify their poverty status. one safety net benefit among all households in the division. Instead, most programs simply allocate resources according to the prevalence of poverty at the Upazila level. This method likely results in many errors in identifying the deserving poor. Recently, geographical targeting has improved and resource allocation is more closely associated with the prevalence of poverty (See Table 5). However, geographical targeting is unable to ensure whether the funds reach needy and poor households. 51 The majority of safety net programs described in previous sections appear to have targeting criteria in that they consider land ownership, assets, income, and employment status. However, as mentioned previously, there are few mechanisms through which this information can be verified or objectively assessed. More often, beneficiaries are selected by local government officials and the selection process tends to lack transparency and can be subject to biased decisions or patronage and elite capture. Public works programs that employ a self-targeting mechanism are among the few safety net programs with a household/individual level pro-poor targeting. Because public works set a wage rate below the market rate, those who are poor and marginalized in the labor market likely self-select into the program. Despite the sub-market wage rate, however, over-subscription is often the case as more individuals desire to participate than public works can absorb. Although the EGPP applies relatively well-defined eligibility criteria, in many other cases, it is unclear how participants are selected into the program. Even for EGPP, however, a report (Ruthbah, 2013) suggested that the selection process was not impartially conducted as it showed that those who had better connections with local government officials were more likely to participate in the EGPP program. With weak targeting, non-poor participation in safety net programs (inclusion errors) is substantial. When the eight largest programs’ targeting effectiveness was compared (Figure 14), EGPP seems much better targeted than other programs. However, even for EGPP, almost a third of beneficiaries came from non- poor (top 60 percent) households. Moreover, the inclusion errors have increased over time: the overall rate of inclusion among the non-poor increased from 44 percent in 2005 to 60 percent in 2010 (World Bank, 2013a). With respect to coverage, the share of the extremely poor (those in the poorest quintile) that receive at least one benefit is 40 percent. 40 Although this number compares favorably with an average coverage rate of around 30 percent for lower and lower-middle income countries, 41 the Bangladeshi system, with its large numbers of uncoordinated programs, offers significant leeway for enhancing both targeting and coverage. 40 Calculations based on HIES 2010. 41 See World Bank (2014) for the state of social safety nets. 52 Figure 14. Percentage of Households Receiving Each Program Benefit: by Income Quintiles Source: Ahmed (2013), recited from the Project Appraisal Document of the SNSP project. To improve targeting, efforts are underway to construct a National Household Database (NHD) that could serve as a source for any poverty-oriented program to draw beneficiaries, while avoiding duplication. The NHD will take a census approach and contain all households’ information including their family structure, assets owned, living conditions, and locality of residence. The information will be used to assess the poverty status of a household based on a PMT approach.42 Therefore, utilizing the database for various safety net programs should improve the targeting efficiency and transparency of the programs. However, it is recognized that establishing and operationalizing the database is a demanding undertaking. Issue 2. Fragmentation of safety net programs Numerous safety net programs, sometimes duplicative with overlapping coverage and objectives, are provided by multiple ministries. Seven ministries that received the largest budget allocation in FY 2015–16 implement 71 programs, which account for 81 percent of the budget in FY 2015–16.43 The largest six ministries between FY 2008–09 and FY 2012–13 have had little change in budget shares; their programs represent about 85 percent of total SP budget, although this figure has declined between FY 2012–13 and FY 2015–16 (Table 4). The Ministry of Local Government Division (MoLGD) has seen an increase in the number of programs (from 4 programs in FY 2008–09 to 14 programs in FY 2012–13) and in the budget allocation (BDT 2.3 billion in FY 2008–09 and BDT 12.4 billion in FY 2012–13). Also, the Ministry of Health and Family Welfare (MoHFW) portfolio has substantially increased, with new programs and an increased budget. Finally, the number of programs under the Ministry of Social Welfare (MoSW) 42See Sharif (2009) for a discussion of the PMT approach in Bangladesh. 43The seven ministries are the Ministry of Finance, Ministry of Disaster Management and Relief, Ministry of Food, Ministry of Primary and Mass Education, Ministry of Social Welfare, and Ministry of Local Government and Rural Development. 53 continued to grow from 9 to 14 to 24 from FY 2008–09 to FY 2012–13 and to FY 2015–16, while the share of their programs out of GDP remained the same. The sector has become more fragmented over time. The total number of programs increased by over 40 percent from 36 to 52 between FY 2008–09 and FY 2012–13 and by 45 percent from 52 to 132 between FY 2012–13 and FY 2015–16 (Table 6). More resources than before are spread thinly across a larger number of programs. After the largest ministries, the remaining 15 or so other ministries implemented 29 programs in FY 2008–09, 40 programs in FY 2012–13, and 61 programs in FY 2015–16. It appears that the significant increase in programs is due to the introduction of small-scale programs under the ‘social empowerment’ budget defined as “stipends, housing and rehabilitation, microcredit, and other miscellaneous funds.” Unfortunately, these social empowerment programs tend to be universal rather than poverty targeted. Programs that have similar or duplicative objectives could be consolidated or harmonized. For instance, three public works programs under MoDMR – EGPP, FFW, and TR – have quite similar eligibility criteria and objectives, although they differ in wage rates, payment methods, and type of projects. Also, two relief programs under MoDMR – GR and VGF—are analogous in that they currently have no specific target group; rather, they have vague criteria for targeting disaster-affected households. To harmonize programs, the Safety Net System for the Poorest (SNSP) project is currently supporting the five largest programs by MoDMR under one umbrella and making concerted efforts to harmonize the programs within the same management system, improve the payment system, and enhance governance and accountability.44 44 See the SNSP Project Appraisal Document for detailed description. 54 Table 6. Allocation by Largest Ministries FY 2008–09 FY 2012–13 FY 2015–16 No. of BDT % of No. of BDT % of No. of BDT % of programs (billion) total programs (billion) total programs (billion) total budget budget budget MoDMR 9 55 40 MoF 4 67 29 MoF 4 128 34 MoF 4 38 28 MoDMR 12 65 28 MoDMR 12 74 20 MoSW 9 10 7 MoFd 1 18 8 MoSW 24 28 7 MoPME 5 7 5 MoPME 7 17 7 MoHFW 10 24 6 MoFd 1 6 4 MoSW 14 15 7 LGD 17 18 5 MoE 1 3 2 LGD 14 12 5 MoPME 3 17 4 MoFd 1 16 4 Total (6) 36 119 86 Total (6) 52 194 84 Total (7) 71 304 81 Remaining 29 20 14 Remaining 40 37 16 Remaining 61 73 19 ministries ministries ministries Source: Ministry of Finance Notes: MoDMR = Ministry of Disaster Management and Relief; MoF = Ministry of Finance; MoSW = Ministry of Social Welfare; MoPME: Ministry of Primary and Mass Education; MoFd =Ministry of Food; MoE= Ministry of Education; LGD = Local Government Division; MoHFW: Ministry of Health and Family Welfare 55 Issue 3. Efficiency The concern about efficiency is often related to food-based programs and their associated weak administrative capacity, in addition to inefficiencies stemming from targeting and fragmentation issues described in previous sections. Given Bangladesh’s history of, and susceptibility to, catastrophic natural disasters, food-based transfers have traditionally been the main form of income support for the poor. The Public Food Distribution System (PFDS) supplies food grains to various food-based safety nets, which account for around 75 percent of its total PFDS food grain distribution. PFDS maintains food stocks in order to provide emergency relief during natural disasters; alleviate chronic food insecurity through targeted food distribution to poor households; and stabilize the market price of food, especially rice, when necessary. To distribute food, the PFDS operates 15 distribution channels, 8 monetized (sales), and 7 non- monetized channels (relief). PFDS relies on food-based programs, notably the OMS as well as public works (Food for Works and Test Relief) and the food-based relief programs (VGF, GR, and VGD) for its regular turnover of reserve grain supplies. The problem with the food-based social safety net is that it has high transfer costs and susceptibility to leakages. The administrative costs of operating PFDS — procuring, storing, managing, and distributing large stocks of food grain — are high. 45 Moreover, there are substantial discrepancies between what is being distributed and what is being received by poor households. With the growing realization that benefits are not reaching the poor and that PFDS has high operating costs, policy makers are beginning to support transitioning from food to cash. Momentum is also being driven by increasingly liberal trade policies and private investment in the food trade, both elements pressuring GoB to reduce its direct intervention in the food market. At the same time, reduced reliance on rice, with wheat emerging as an alternative to relieve food insecurity, is also reducing the need for PFDS to directly intervene. A complete switch from food- to cash-based programs is, however, unlikely. It is politically challenging as many stakeholders and interest groups have been involved in various stages of procurement for a long time. Also, there are situations such as humanitarian crises or catastrophic natural disasters, when food-based transfers might be a better option. However, it is widely agreed among policy makers that there is a need to modernize the PFDS to improve efficiency. 3.3 Microfinance Home to globally recognized institutions such as the Grameen Bank and BRAC, Bangladesh’s microfinance sector plays a major role in addressing the needs of the poor. Since the 1990s, Bangladesh has experienced a significant expansion of its microfinance sector. Financial products through microfinance institutions including loans, savings, and insurance could in theory reduce poverty and vulnerability as well as promote productive activities and long-term investments, including children’s schooling. With an easy access, microfinance in Bangladesh often serves as a safety net for the poor. 45 See Ahmed et al. (2009) for more detailed discussions. 56 3.3.1 Program Review A large number of MFIs are in practice in Bangladesh. As of 2011, there were more than 576 registered MFIs with more than 35 million active members and over 15 million active borrowers — reflecting an eight-fold increase since 1995. 46 World Bank (2013a) attributes this expansion in the 1990s to three factors: the entry of major nongovernmental organizations (NGOs) in the microfinance sector; the availability of increased donor funds; and the formation of the microfinance apex body, Palli Karma- Sahayak Foundation (PKSF), to supervise and regulate. The variety of microfinance providers in the country includes government-owned banks, private lenders, and NGOs. However, four major MFIs -- Grameen, BRAC, BORO, and Association for Social Advancement (ASA) -- dominate the sector, while many other smaller MFIs coexist, serving smaller clientele including rural communities (Table 7). Along with growth in size and number, the portfolio of MFIs has been increasingly diversified. While the majority of MFIs in Bangladesh primarily focus on offering small microcredit loans, non-lending portfolios have been expanding to include a variety of products including savings, remittances services, and agricultural finance options. Table 7. Four Largest MFIs in Bangladesh (2012) Institution Number of active Loan portfolio Average loan size Saving deposits borrowers (million) (USD million) (per person) (USD million) Grameen  6.7  997  US$ 149 (BDT 11,600)  1,628 BRAC  4.4  750  US$ 170 (BDT 13,300)  311 ASA  4.2  636  US$ 151 (BDT 11,800)  262 BORO  1.0  110  US$ 110 (BDT 8,580)  42 Source: Based on Chen and Rutherford (2013). In part due to the large supply of MFIs, accessibility to microfinance is high in Bangladesh. In 2010–11, for instance, an estimated 69 percent of rural households were microcredit members (World Bank, 2013a). The high level of market penetration by MFIs has meant that in rural areas, many households have access to multiple sources of microcredit. An estimated 31.9 percent of households in rural Bangladesh were members of multiple microcredit programs in 2010–11, which increased from 8.9 percent in 1998–99.47 In particular, the large number of women, especially in rural areas, benefit from easy access to microfinance. Many of Bangladesh’s large MFIs specifically target their loan programs toward low-income women. As a result, almost two thirds of MFI customers in Bangladesh are women. In 2010–11, among the four largest MFIs, women’s share of microcredit lending was the highest for Grameen Bank (89 percent) and the lowest for BRAC (38 percent).48 46 See Khandker and Samad (2013); World Bank (2013a); and Chen and Rutherford (2013). 47 Khandker and Samad (2014). 48 World Bank (2014). 57 International indices to standardize diagnostics that measure market penetration suggest that Bangladesh ranks high with respect to access to microfinance. The first index, developed by the Center for Microfinance, created comparable indicators of microfinance market penetration for more than 80 countries from 2004 to 2010. 49 Bangladesh emerged as the top eighth country with a credit market penetration among the poor of 56 percent.50 Another index, called the Microfinance Index of Market Outreach and Saturation (MIMOSA), utilizes different methodology, capturing financial transactions with both formal financial institutions and MFIs reflected in the Global Findex dataset.51 It created an index from household surveys in 148 countries to construct a market penetration ranking of 1 to 5. In this ranking, Bangladesh stood out as one of the countries with the highest market penetration along with Mongolia and Cambodia; about 27 percent of adults in Bangladesh reported they took out a loan over a 12-month period. 52 It is debatable whether Bangladesh’s microfinance sector is oversaturated and whether it pushes the poor to over-indebtedness with the over-commercialized practices. However, it seems apparent that relatively easy access to microfinance brings both opportunities and risks. 3.3.2 Issues with Microfinance Programs Issue 1. High Transaction Costs, High Interest Rates Rapid growth in the microfinance sector in Bangladesh in the last 20 years and the saturation in loan markets raises questions about the future of microfinance. The high presence of MFIs can be a catalyst to expand financial inclusion for the poor, but over-saturation or over-indebtedness can pose a substantial risk for already impoverished households. Although it is inconclusive whether the sector in Bangladesh is over-saturated or customers are over-indebted, in 2010, responding to concerns about rising interest rates among MFIs, the Microcredit Regulatory Authority (MRA) capped the microfinance interest rates at 27 percent. Many MFIs have been charging quite high interest rates. For instance, the Grameen Bank charged a nominal on-lending rate of 20 percent, 7 percent higher than the commercial bank rate. However, MFIs contend that while MFI interest rates are higher than those of commercial lenders, they remain significantly lower than those of informal money lenders (Khandker and Samad, 2013). High interest rates imposed by MFIs are mainly due to the substantial risks and transaction costs that they face. The risk of default is always present in financial transactions. To mitigate these risks, lenders take various measures. For instance, most commercial banks require collaterals or guarantors, which greatly hinder the poor from obtaining loans. Instead, MFIs can expand poor people’s access to credit while managing the risks by implementing the following measures: group liability; access to larger size loans 49 See Krauss et al. (2012) for the methodology and findings. 50 Countries that rank higher than Bangladesh include Vietnam, Montenegro, Bosnia and Herzegovina, Sri Lanka, Mongolia, Peru, and Tunisia. The poor are defined as those who live below the national poverty line. 51 The simplified model relies on three indicators which act as proxies for market penetration: (1) human development index which the model assumes has a strong relationship with formal sector borrowing; (2) formal sector savings assuming that individuals who are more likely to save at financial institutions tend to also have higher borrowing rates; and (3) semi-formal savings (from moneylenders, employers, and stores) which act as proxy for the intrinsic level of credit demand. 52 See Javoy and Rozas (2013). http://www.cgap.org/blog/introducing-mimosa-microfinance-market-capacity-measurement- tool. 58 upon repayment of outstanding loans; and small frequent installments. In particular, group lending enables the lending agency to reduce risks by having the self-formed group members jointly monitor their activities and create peer pressure for repayment (Bhatt and Tang, 1998). This measure is also designed to reduce transaction costs by dealing with a group instead of multiple individuals, and it transfers monitoring and enforcing costs from the lending agency to the group. The group members’ intimate knowledge of each other or a social norm serves as an implicit instrument for preventing defaults. However, in practice, many hidden transaction costs exist associated with microlending. These include costs associated with reaching out to potential borrowers, screening, appraising the viability of their business activities and facilitating the formation of groups and regular meetings among them; providing borrowers with technical assistance on loan contracts; and providing support for group performance. These transaction costs are incurred regardless of the size of the loans, and adversely affect microlenders’ operation. When MFIs faced less fierce competition and the rhetoric concerning microfinance was almost unanimously positive, MFIs had enough customer base or donor support to compensate for their high transaction costs. As the number of MFIs increases and the sector experiences greater competition, some MFIs’ activities to reduce risks and expand their business can negatively affect the wellbeing of the very populations that they intended to target. For instance, borrowers are often forced or pressured to take more loans to extend the repayment period for their outstanding loans, leading to deeper indebtedness. At the same time, efforts to reduce transaction costs by making less efforts in managing group performance or in appraising viability of proposed activities that the loan would be used for, could significantly reduce the success and repayment rates. Issue 2. Weak linkages between microfinance and productive labor market activities such as entrepreneurship A major goal of microcredit loans to poor borrowers, as initially envisaged, was to provide poor workers (especially women) with the seed capital for small-scale entrepreneurial activities to generate earnings. Proponents of this view argue that microfinance can play an important role in reducing the credit constraints faced by potential entrepreneurs in poor communities and thus contribute to poverty reduction in a more sustainable way. Yunus (2003, 2008), for instance, provides numerous anecdotal evidence and success stories where a group of impoverished women obtained sustainable earnings opportunities through small-scale business with the help of microfinance. Their economic activities, albeit small scale and mostly at subsistence level, were nevertheless considered to be entrepreneurial —and with an injection of minimal funds — they could thrive and enable their workers to improve their livelihoods. These optimistic views and a focus on the productive aspects of their work in large part contributed to the proliferation of microfinance institutions in Bangladesh since the 1990s. However, promotion of business activities through microfinance will have limited success unless support through microfinance is combined with other labor market interventions (See Box 7 for a discussion on the effectiveness of microfinance). Many borrowers use the loans to supplement short-term consumption rather than invest in business activities (Banerjee and Duflo, 2006). Even if business activities are promoted through microfinance, the impacts are weak. Studies suggest that microfinance alone would 59 not be able to promote entrepreneurial activities, because loan amounts are typically small and poor workers face many other constraints such as lack of skills, and limited information and social capital (Cho et al. 2015). Therefore, the role of microfinance programs will be limited in entrepreneurship unless they are combined with other labor market interventions. In Bangladesh, few programs exist to promote business activities for the poor, by providing microfinance combined with other support, which includes training on skills and financial literacy as well as networking and advisory services. Khandker and Samad (2013) find some occupational mobility for microcredit users, gradually transitioning from farm to non-farm activities for both self-employment and wage employment over time. However, extra efforts to facilitate these transitions are not in practice. More recently, major microfinance institutions such as Grameen and BRAC have been developing interventions that help finance social enterprises and community-based collective businesses enabling them to serve the poor and help build more scaled up and sustainable businesses. Box 7. Evidence on the Impacts of Microfinance Despite the initial optimism about the potential causal links between microfinance and poverty alleviation, evidence generated from Bangladesh’s microfinance experience over the past two decades has produced mixed results. Early evaluations tended to emphasize the optimistic view on microfinance. Pitt and Khandker (1998), for instance, found that microcredit helped improve household wellbeing by significantly increasing consumption particularly for women. Another argument is that improved access to credit through microfinance has been instrumental in increasing women’s bargaining power within the household and contributed to women’s empowerment. Hashemi et al. (1996) contributed to the positive views by qualitatively showing that women’s access to microcredit increased the likelihood that women would gain asset holdings in their own names, purchasing power, and political and legal awareness. A series of more recent studies, however, challenged the previous claims and findings on promises and impacts of microfinance. Roodman and Morduch (2009) used the same data set used by Pitt and Khandker (1998), but reached a completely different conclusion. They questioned the estimates and methodology used by the authors, and, based on their own revised estimators, found little measurable evidence about the impacts of microfinance loans on the overall welfare gain among poor households. Moreover, they argued that the impacts on women are not positive. Karim (2011) and Kabeer (2009) also noted that the access to credits does little to improve women’s welfare, partly because loans taken by women are frequently passed on to male members of their families (typically husbands). Studies based on rigorous impact evaluations and randomized controlled trials (RCTs) across various countries also contribute to the mixed findings, if not skepticism, suggesting that microfinance has shown only modest impacts at best on reducing poverty especially with limited results on the outcomes of business and earnings. Banerjee et al. (2015), based on six RCTs — Bosnia and Herzegovina, Ethiopia, India, Mexico, Mongolia, and Morocco — argue that access to microcredit seems to help households cope with shocks (health, climate-related calamities) and contribute to short-term consumption smoothing, but has no transformative impacts. The study emphasized the heterogeneous impacts of MFIs and cautioned against both the panacea-like promises and excessive criticism. Source: based on Pitt and Khandker (1998); Hashemi et al. (1996); Roodman and Morduch (2009); Karim (2011); Kabeer (2009); and Banerjee et al. (2015). 60 3.4 Labor market programs Despite daunting challenges in the labor market discussed above, there are only a small number of programs to promote opportunities through the labor market. The 2013 Poverty Assessment and 2015 Systematic Country Diagnostics (SCD) alike highlighted the importance of jobs to reduce poverty in Bangladesh. Yet, policy intervention in labor is quite limited and most of existing initiatives are driven by international organizations or in pilot stages. Although the importance of reducing poverty and protecting the poor from further deprivation through safety nets has been widely embraced in Bangladesh and many larger scale safety net programs have been created, the role of labor market programs to improve opportunities and promote better earnings has received relatively less attention. 3.4.1 Review of programs Labor market interventions Table 8 presents a few examples of programs that promote income-generating opportunities in Bangladesh. Given that the nature of poverty issues is different between urban and rural areas, programs target either one or the other. Urban programs tend to focus on urban slum dwellers by addressing their constraints to finding productive employment opportunities, while rural programs generally target farmers by helping them improve agricultural technologies, reduce exposure to seasonal risks, and develop agri-business. These programs are still in the early stages of development and tend to be initiatives led by international organizations. Table 8. Large-Scale Labor Programs in Bangladesh Program Organization Target group Budget allocation Empowerment and Livelihood  World Bank and the  Rural poor and vulnerable  US$220 million over Improvement "Nuton Jibon" Social Development households 6 years Project Foundation under the Ministry of Finance Cereal systems initiative for  USAID  Farmers in impoverished  US$411 million over South Asia program and agriculturally 3 years dependent regions Urban partnership for poverty  Ministry of Local  Poor urban communities  US$100 million over reduction Government, Rural and households 5 years Development and Cooperatives (MoLRD), DFID, UNDP, and UN- HABITAT Skills and Training Enhancement  World Bank and the  Low-skill workers including  US$ 79 million over Project (STEP) Canadian International aspirant international 5.5 years Development Agency migrants (CIDA) under the Ministry of Education 61  Nuton Jibon Project: Nuton Jibon means “new life” in Bangla, and the project aims to improve the livelihoods of the rural poor, with a special focus on the poorest segment of the population. The project originated from a pilot in 2003 and built upon the two consecutive social investment programs (Phase I and II) from 2008–15, supports the poor and ultra-poor in more than 20 poorest districts of Bangladesh. The unit of targeting is not an individual household or worker but poor Upazilas, which were identified through the 2014 Bangladesh Poverty Map published by BBS. The project takes a community-driven approach for development through a range of interventions. It focuses on empowering the poor so that they can gain access to income-generating opportunities. For instance, poor households in the catchment villages can be organized into a group of producers, cooperatives, and societies to collectively develop business projects, pursue better linkages to value chains, and strengthen their community institutions. A few noteworthy features of the project include an emphasis on female empowerment by facilitating women’s participation in various community activities; skills training and linkage to potential employers for youth employment; technical assistance for agricultural technology and livestock feeding and rearing through extension services; and collective resource management at the community level.  Cereal Systems Initiative for South Asia - Bangladesh (CSISA-BD): The primary objective of the CSISA- BD project, supported by the United States Agency for International Development (USAID), is to help farmers adopt valid agricultural technologies, improve productivity, and increase earnings. To ensure that farmers have better access to timely and accurate information and technologies, CSISA-BD provides capacity-building training and extension services. Follow-up support for farmers who have adopted new technologies ensues so that they can indeed reap the benefits of adopting the technology. Adaptive research complements these strategies by refining technologies so that they are suitable for local biophysical and socioeconomic circumstances, and it also develops new technologies to improve the scope for future practice.  Urban Partnership for Poverty Reduction (UPPR): In response to various constraints that urban poor and slum dwellers face, this project establishes partnerships with various actors and institutions to improve access to services and promote employment, largely supported by the Department for International Development (DFID). Similar to the Nuton Jibon project which uses a community-driven approach, the UPPR project facilitates the formation of Community Development Committees (CDC) among slum dwellers. The CDCs then collectively identify the needs for support and develop action plans for self-help. The targeting of the poorest households is conducted through the Participatory Identification of the Poor (PIP) process, where responsibilities for defining poverty and identifying those most in need fall on communities. Each CDC discusses and agrees upon relevant social, economic, and physical poverty criteria to identify the poverty status of all households, which are then categorized into the extremely poor, poor, and non-poor. Furthermore, this project highly emphasizes female empowerment and works with women and girls to overcome their disadvantaged social position by encouraging them to take on leadership roles within their communities. The project supports funding for access to basic services such as water, sanitation, and nutrition, and subsidizes households for children’s education and skills training and women’s business development. CDCs 62 manage financial resources while running their own savings and loans schemes. As of 2013, there were more than 2,500 CDCs in 23 target towns and cities.  Skills and Training Enhancement Project (STEP): The demand for a better trained workforce is greater than ever, in response to the need to gain competitiveness in the global market. However, the existing Technical Vocational Education Training (TVET) sector is not well equipped to respond to this need. Trainees graduating from TVET institutes lack skills relevant to the labor market, which has historically been a main reason for poor employment outcomes after they complete their training; the public TVET system is fragmented across numerous ministries and overall intake capacity is very limited;53 and there have been limited opportunities to customize courses to meet the needs of trainees (e.g., language training has not been widely available even for individuals who seek overseas employment). To facilitate public and private skills training institutes to better train the workforce and respond to the market demand for skills, STEP commenced its operation in 2010 under the Ministry of Education in collaboration with the World Bank, and the Government of Canada (through the Canadian International Development Agency, CIDA) began its support in 2013. STEP’s primary objective is to improve the quality of training and offer a wide range of training to workers with post-secondary polytechnic or secondary level education, and other low skilled individuals (with a minimal level of junior secondary education) including aspiring international migrants, through short courses. STEP helps training institutes establish partnerships with industries to ensure that the programs offered are relevant to the needs of the labor market. The program sets performance indicators (e.g., training completion rates, teachers’ training, and student placement rates within six months after completing training) and provides performance grants to institutes which meet the targets. To assist low-income students (in the bottom two quintiles), institutes provide stipends to short course students who are proven poor by the PMT mechanism. Short-term courses receive US$250 (BDT 19,500) per student of which US$10 (BDT 780) per month is paid as a stipend to the student. Labor Regulations Efforts to reform labor regulations are very important initiatives although they are not an intervention directly related to an individual’s or household’s opportunities for employment. In most countries, labor regulations impose rules that employers and employees must comply with. Labor regulations are designed to improve the bargaining power of employees, and to prevent employment practices that are considered abusive and unacceptable. The main aspects of labor regulations include hiring and firing rules, terms of employment contracts, provision of paid and unpaid leave, and minimum wages, and rules on workplace health and safety. Issues related to labor regulations tend to receive little attention especially in low- income countries because a large share of employment is not wage and salary work. In Bangladesh, however, the 2013 high-profile factory collapse in the Rana Plaza, which claimed over 1,000 garment sector workers, triggered a nationwide reform to introduce and enforce more stringent labor compliance laws, particularly in the export-driven RMG sector which employs over 4 million workers. 53The TVET sector has a maximum capacity to train an estimated 200,000 workers annually while almost 2 million people are added to the labor market each year. 63 The most recent labor law, which was enacted in 2006, well before the reform triggered by the factory collapse, extensively consolidated more than 25 labor-related acts and ordinances under the Bangladesh Labour Act (BLA). The BLA introduced improvements in several areas, including regulations on working hours and restrictions in overtime; labor inspections; job protection of union leaders; workers' compensation; and workers' rights to self-defense for disciplinary matters and termination (USAID, 2014). Then the BLA was substantially amended in 2013 extending beyond health and safety regulations, and promoted the overall bargaining power and wellbeing of workers. For instance, the new law requires that factories establish a workers’ welfare fund where they set aside 5 percent of their net profits to improve the living standards of workers. In parallel, GoB significantly increased the minimum wage for garment workers from BDT 3,000 (US$ 38) to BDT 5,300 (US$68) a month.54 However, it is premature to conclude whether these RMG sector-focused reforms are sufficient to properly ensure workers’ wellbeing, or whether the regulation was too restrictive to the extent that it hampers Bangladesh’s competitiveness. Programs for International Migration Similar to other labor programs, appropriate interventions to support migrant workers and their families have been limited. To facilitate the process, GoB has introduced the following useful initiatives:  Enacted the Overseas Employment and Migrants Act 2013 which facilitates a system to prmote safe migration by ensuring the rights and welfare of migrant workers and memebers of their families  Signed bilateral agreements on migration with a number of host countries to formalize the migration process and facilitate safe migration among workers, based on promising approaches with receiving countries such as Malaysia, Korea, and most recently, Saudi Arabia  Established a Migrant Welfare Bank to provide migration loans, handle remittances, and facilitate investment for returning migrants  Streamlined the remittances process through the banking system. Bangladesh’s two largest MFIs, BRAC and Grameen, have both established commercial banks that can transfer remittances through their network at the grassroots level. A pilot project supported by the World Bank (Safe Migration for Bangladeshi Workers) is underway to provide the pre-decision and pre-departure information, orientation, and linkages to skills training for potential migrants via community-based organizations (CBOs) in order to promote safe migration choices and practices.55 54 As the attempt to establish a national minimum wage was not approved by the Parliament in 2007, the national minimum wage apart from the garment sector is not currently implemented in Bangladesh. 55 See World Bank (2013b) for detailed description on the project. 64 Box 8. Institutions related to International Migration An independent ministry, the Ministry of Expatriates’ Welfare and Overseas Employment, administers international migration and overseas employment. There are two operating bodies under the Ministry, which facilitates international migration: the Bureau of Manpower, Employment, and Training (BMET) and the Bangladesh Overseas Employment Services Limited (BOESL). Although government plays a relatively small role in recruiting migrants compared to private agencies or intermediaries, these two government units recruit aspiring migrant workers and facilitate their overseas employment.  BMET: Oversees the District Employment and Manpower Offices (DEMOs) and technical training centers, and aims to create skilled manpower in line with the demand of global labor market. This office is responsible for tracking the formal movement of and ensuring overall welfare of migrant workers.  BOESL: Operates as a state-owned public company which facilitates job matching between Bangladeshi workers and foreign employers. The functional difference between the BMET and BOESL is not always straightforward. For instance, the bilateral government-to-government (G2G) program between Bangladesh and South Korea is under BOESL, while G2G with the government of Malaysia is mostly facilitated by BMET.  Government to Government (G2G) Program: 56 Currently, GoB has G2G programs with three countries: Malaysia, the Republic of Korea, and Saudi Arabia. 57 The discussion of G2G with the Government of Malaysia (GoM) was initiated with the demand for 30,000 Bangladeshi workers in the plantation sector in Malaysia, and a Memorandum of Understanding (MOU) was signed in November 2012. In line with the MOU, GoM provides the Visa with Reference, while BMET facilitates worker recruitment, selection (based on lottery), training, orientation, and medical testing. The process for recruiting and selecting workers is as follows: o BMET advertised in eight national newspapers to register aspiring job seekers in January 2013 o Workers were registered in more than 4,500 Union Digital Centers (formerly, Union Information and Service Centers, UISCs) o Given the plantation sector work, only male workers between the ages of 20 and 45 years with a minimum height of 5 feet and a minimum weight of 50 kg who were able to lift a minimum weight of 20 kg proceeded with the registration (part of which is verified by the photographs of workers sent to the Malaysian Government), and this resulted in about 1.45 million male applicants in the BMET databse o In early 2013, about 12,000 workers were selected through division and union-wise random lotteries for an initial batch of migration o The selected workers received an SMS message notifying them that they were selected and later received a “confirmation card” from BMET o They participated in a 10-day training, medical testing (at any of the nine selected medical colleges across Bangladesh), and final 10-day orientation 56See Box 8 for the institutions related to international migration. 57Given the private sector involvement in the recruiting process to Saudi Arabia, the program is often called as a B2G (Business to Government) instead of G2G. 65 o Workers received a renewable contract with a basic monthly salary of 900 Malaysian Ringgit (US$236; BDT 18,400) for two years o BMET facilitated the purchase of air tickets, emigration clearance, visa stamping, and other items. The total cost from application to departure was about BDT 33,000 (US$415) of which BDT 23,000 was one way airfare. A process evaluation in Bangladesh (Wajid, 2013) suggests that there has been a substantial administrative delay, and only 200 visas among the selected 12,000 workers were received as of April 2013. Although the application process and selection procedures have been transparent and innovative, the lack of private sector involvement severely hampered the smooth flow of workers. As a result, as of October 2015, only 8,500 workers have been sent to Malaysia. In light of this, the two governments have recently amended the MOU to extend the collaboration with private recruiting agencies. Similarly, Bangladesh signed an agreement with the South Korean government on June 4, 2007, under the Korea-Bangladesh Employment Permit System (EPS). The EPS established the principles and rules of managing human resources from foreign countries to South Korea. Bangladesh is one of the 15 countries the Government of Korea has MOUs with in implementing the EPS. 58 About 10,000 Bangladeshi workers have so far received jobs in South Korea through this G2G initiative conducted by BOESL since 2008. The South Korean government facilitates the selection of migrant workers based on their language and technical skills, provides training in the language and culture, and provides support for settlement, job changes, and legal issues particularly with employers. To date, little is known about how the G2G with South Korea has impacted migrant workers in Bangladesh. However, anecdotal evidence suggests that the EPS has been quite successful in reducing migration costs and improving transparency in the process. In February 2015, Bangladesh and Saudi Arabia signed an MOU that focused on the recruiting of domestic workers from Bangladesh to be deployed to Saudi Arabia. This is expected to open a route for female migration, although tensions regarding labor mobility between these two countries still exist, as highlighted in the high-profile domestic workers abuse in Saudi Arabia (Observer, 2014). In contrast to the previous two MOUs with the governments of Malaysia and Korea, the Bangladesh Association of International Recruiting Agencies (BAIRA) has acquired a separate agreement with the Saudi counterpart to facilitate migration and receive commissions from the employers while still utilizing BMET’s infrastructure for registering and tracking workers. This was possible given the lessons learned from the Malaysian G2G case where bypassing private recruiting agencies had negative impacts on the program itself. By this arrangement, BAIRA is expected to conduct background checks and ensure the quality of the workers’ skills. While this administrative arrangement of the Public Private Partnership seems to be welcome, additional cautions are warranted because domestic workers are mostly female migrants, which may have different implications for family’s (especially the children’s) wellbeing as well as the migrant’s welfare. 58The 15 countries include Bangladesh, Cambodia, China, East Timor, Indonesia, Kirgizstan, Myanmar, Mongolia, Nepal, Pakistan, the Philippines, Sri Lanka, Thailand, Uzbekistan, and Vietnam. 66  Safe Migration for Bangladeshi Workers: Many workers fail to migrate, or face severe risks even after they expend significant financial resources, which can be in part due to insufficient information on the migration process. 59 Recognizing that, a Safe Migration for Bangladeshi Workers project was introduced as a pilot initiative. The project, implemented by BRAC and supported by the World Bank, provides potential migrants with better access to timely information and services for safe migration. This is facilitated through CBOs that utilize widely present social networks to reduce exposure of aspirant migrants to potentially abusive and fraudulent intermediaries. For this project, local CBOs deliver: (i) mass awareness campaigns to improve knowledge of migration issues and understanding of the migration process; (ii) pre-migration decision information to inform aspirant migrants about potential costs and benefits of international migration; (iii) basic language and skills training to reduce potential migrants’ barriers to successful migration; (iv) financial management training so that potential migrants become familiarized with the financial instruments and services to manage their financial resources and remittances; and (v) pre-departure orientation for migrants who have already secured their placement to receive practical information about travel to the host country, whom to contact if they encounter problems while abroad, what their rights are in the receiving country, and issues to consider before returning home. The recent baseline study of the pilot offers valuable lessons which have implications for any future intervention in this sector. Some of these lessons include (i) the critical role of middlemen in facilitating overseas migration necessitates more intensive public information campaigns as citizens continue to heavily rely on middlemen in areas where government and NGO actors have not penetrated; (ii) the reliance on middlemen is a lot greater in areas farther away from large cities and where the networks with current and past migrants are weak; and (iii) the lack of a guarantee of a job at the end of training could significantly undermine potential migrants’ motivation and investment in training. 3.4.2 Issues with Labor Programs Issue 1. Limited availability of intervention The foremost issue in labor programs in Bangladesh is limited availability of intervention. Examples are ubiquitous: safety net beneficiaries have very few linkages to pursue for building labor market experience and strengthening employability; an unemployed youth looking for wage-based employment has little support available to assist him/her in the job search; agricultural workers who are using outdated technology and are experiencing diminishing yields have limited access to improve productivity; groups of young workers with plans to operate a small business have limited knowledge in business development even if they can get microcredits from MFIs; and poor rural workers planning seasonal migration do not know where to start other than relying on middlemen. 59 Das et al., (2014) suggests that an estimated one third of aspirant migrants end in failure after they try to migrate. 67 Much of the labor agenda is concentrated on the export-oriented sectors, including the RMG sector, in establishing workers’ right, labor laws, and regulations. Although this agenda is important and under the high-profile scrutiny of international pressure, the fraction of immediate beneficiaries is approximately 7.5 percent (4 million RMG sector workers out of 53 million workers in the active labor force in Bangladesh). For this type of labor regulations and laws to have an impact for a wider range of workers, more opportunities for wage employment need to be developed, to which diverse tools and programs can contribute. Issue 2. Lack of evidence-based policy Evidence-based policy is the practice of establishing relevant policies and programs based on solid analytical underpinning and rigorous evidence. It extends beyond the area of labor and needs to be emphasized in all fronts of SPL. However, the issue of evidence-based policy (or the lack thereof) is more pronounced in the area of labor programs. Major initiatives, such as labor regulations reform or G2G initiatives on international migration, tended to be triggered by accidents or mishaps experienced by workers. The impacts of these initiatives on workers’ welfare have not been rigorously evaluated, and the evidence base for future improvement is weak. This is partly due to limited availability of micro-data required to establish the analytical underpinning of policy intervention. Typically, analyses from the whole spectrum of employment (from both the demand and supply sides) are needed to identify jobs challenges so that more, better, and inclusive jobs are created. From the demand side, investment climate, business environment, firms’ profiles and dynamics, and barriers to trade, among others, are analyzed along with an assessment of general macroeconomic trends. From the supply side, health, education, and skills of the work force, labor force participation decisions from households, employment status and sector, workers’ wages and productivity, and labor market transitions and movement are assessed. Then data on special topics, such as gender, youth, time-use, or manufacturing, provide further insight into the complex issues that individual workers and firms face with jobs challenges. These analyses require a substantial amount of quality data, which are not always available in Bangladesh. In fact, data-collecting activities are very active in Bangladesh. BBS has been leading in this area constructing valuable data on various topics (See Table A1 in Appendix). Various organizations have also contributed to the list of databases. Nonetheless, the caveat is that the constructed data are not always publicly available in a timely manner, and by the time data are released they tend to provide somewhat outdated information. There is great potential to provide useful information and lessons for evidence- based policy if these data can become widely available for various analyses. Recently, efforts to collect and disseminate data on the labor market and poverty to inform policy have been ongoing, and an increasing emphasis has been placed on policy based on analyses and evidence (See Box 9). 68 Box 9. Bangladesh Bureau of Statistics’ Efforts for Data Collection Household survey data that yield the information needed to monitor labor market status and performance as well as poverty in Bangladesh have been collected roughly every five years. The last rounds of the Household HIES and LFS were conducted in 2013. The lack of more frequent household survey data significantly restricts the ability to assess the living conditions of the population and changes in the labor markets, which create important barriers to inform public policies based on factual data. Recently, BBS has decided to increase the frequency of data collection from every five years to every three years for HIES and from every five years to a quarterly scheme (explained below) for LFS surveys in line with their NSDS. Quarterly LFS: Major innovations associated with this activity include not only the frequency of data collection, but also the rotational panel feature. A household will appear two consecutive quarters in data, stay out for two consecutive quarters, and return for another two consecutive quarters. By following this scheme, one is able to observe labor market transitions and dynamics of a household over time with seasonality controlled. The first data collection of the quarterly LFS (QLFS) commenced in July 2015. HIES 2015: Building upon the four rounds of HIES from 1995 to 2010, HIES 2015 will strengthen its function for poverty measurement and monitoring. It will be well aligned with other nationwide data including Economic Census and QLFS, and greatly inform the state of Bangladeshi households in economic, financial, and social domains. Recognizing the need for more frequent data collection and poverty monitoring, a plan to conduct the HIES every three years instead of five years is currently under consideration. Source: author’s articulation. 69 70 4 Key challenges to be addressed in social protection and labor Given the state of existing SPL programs and many associated issues, challenges that are typically within the domain of SPL intervention are still prevalent. This section highlights a few key challenges that SPL policy and policy instruments need to address in the coming years in Bangladesh. These are not an exhaustive list of policy issues related to reducing poverty or creating jobs, but some select challenges that require intervention with SPL measures. 4.1 Challenge 1: Widespread Poverty and Vulnerability Poverty and vulnerability persist particularly due to high exposure to income shocks and limited access to risk management systems. The high share of employment outside the formal sector of wage employment is also correlated with unstable income streams. Day laborers and the self-employed face huge uncertainties in their earnings streams, and those in agriculture are exposed to weather-related shocks and seasonality. The agricultural seasonality, though somewhat mitigated by diversification of income sources, 60 is still of concern and especially pronounced in the Rangpur Division. Natural calamities including floods, earthquakes, and cyclones pose significant income risks. Fluctuations in food prices, especially the price of the main food-grain rice, have been a source of uncertainty as well. The recent experience of a food crisis in Bangladesh in 2008, which pushed many households into poverty and extreme poverty, is an example of the detrimental impact of exogenous shocks on poverty.61 Since food consumption accounts for a large share of total consumption for poor households, this type of crisis has damaging impacts. The poor cope with the risks by reducing the quantity of their food intake; switching to lower quality food; drawing down difficult-to-replace assets; and reducing non-food expenditures (e.g., by pulling their children out of school). Access to social insurance to mitigate the negative impacts of shocks and uncertainties is very limited in Bangladesh. Few workers have crop and livestock insurance, and nearly no one has health insurance. Only a small fraction of the labor force is covered by public pensions, mainly civil servants and their families. Although many households have access to microfinance, availability of saving and insurance is limited and most are concentrated in small amount credits. Large-scale public works programs such as EGPP are available to provide short-term employment opportunities to the poor, particularly during agricultural lean seasons. However, the timing of public works does not always coincide with the lean season due to implementation challenges, although there has been recent progress in this area. Other safety net 60 This includes diversification of crops with Boro rice as well as Aman rice, seasonal migration during lean seasons, and increases in non-farm activities. 61 The food crisis in 2008 was a combination of consequences from multiple factors, which included the damages of the Aman crop that resulted from the July–September 2007 floods and the cyclone in November 2007, and the global food crisis during which food prices almost doubled and when India, the main source of imported rice for Bangladesh, restricted trade in rice. 71 programs that help poor households cope with risks also have some limitations and issues, as discussed above. 4.2 Challenge 2: Jobs Challenges for Youth With the demographic transition towards a larger and younger work force, the job challenge for the large number of young people is high. Young people (15-24 years of age) make up about 30 percent of the working age population (15-64 years of age) and about 39 percent of the active labor force in Bangladesh. With a growing number of young people entering the labor force, the challenges to absorb them in the labor market are formidable. If sufficient jobs can be created to provide earning opportunities for youth, they will contribute to developing a lively and invigorated labor market. Challenges in the labor market tend to be greater for youth compared to adults (See Figure 15). Similar to many other countries, youth are more likely to be unemployed than adults, with a lower employment ratio and labor force participation rate. The gap between youth and adults with respect to these indicators is larger for males than females. The higher share of agricultural and nonpaid family work among youth compared to adults also indicates that youth tend to end up in low productivity jobs. In particular, youth are significantly less likely to be self-employed but more likely to be nonpaid family workers and, for males, day laborers. This suggests that youth are substantially limited in access to quality earnings opportunities. Figure 15. Employment Outcomes of Youth (ages 15–24) versus Adults (ages 25–64) by Gender Source: LFS 2010. School to work transition is difficult. The transition rapidly takes place between the ages of 15 and 20, with only a small proportion of individuals engaged in both working and schooling (See Figure 16). Due to 72 limited data, it is difficult to know how long it typically takes for an individual to transition into full-time employment after he/she completes schooling. Apparently, young workers in Bangladesh, like those in other developing countries, tend to rely on informal networks including family and friends for job searches rather than on formal channels, such as employment services or media (World Bank, 2013e; Matsumoto and Elder, 2010). Therefore, with all things being equal, those who have weak social networks or no role models are more likely to have difficulties finding suitable employment opportunities. Youth from low- income families can be socially less connected than those from higher income families, and the lack of social connections can affect inter-generational correlations of occupations. Figure 16. School-to-Work Transition Even among those better educated who graduated from TVET institutions’ short term (up to six months) training, the transition to the labor market is not without challenges. A tracer study shows that only about a quarter of graduates found wage- or self-employment opportunities, and almost half were out of the labor market and close to one fifth unemployed, six months after the training (Figure 17). The employment outlook does not seem to improve over time, because the survey for another group of graduates whose performance was measured 12 months after completing training presents a similar picture with the share of those working in wage employment less than 20 percent. 73 Figure 17. Transition to the Labor Market among TVET Graduates Source: Adopted from Figure 1 in Nomura et al. 2015. While quantity of education has improved, the quality of education in providing foundational skills is still of concern. An analysis of the National Students Assessment (NSA) showed that only 20 and 13 percent of grade 3 students show grade level competency in Bangla and mathematics, respectively; the equivalent figures for grade 5 students are 25 and 32 percent in Bangla and mathematics, respectively; and at the eighth grade level, the proportion of students who have grade level competencies in Bangla, English, and mathematics are 44, 44, and 35 percent, respectively (World Bank, 2013e). An explanation for the higher proportion of competency as grade levels increase may be due to high dropout rates among the students with poor performance. Furthermore, there is a significant gap across students’ socioeconomic status, and students from poor socioeconomic backgrounds are much more likely to drop out of school. The relevance of skills fostered through the education and training system for the labor market is weak. The lack of basic literacy, numeracy, and non-cognitive skills limits the ability of young people to be gainfully employed by the formal sector.62 The tremendous diversity in the types of skills that matter to different types of employers makes it difficult to define a set of skills as market relevant and prioritize the ones most needed. It is important to underscore the findings of recent empirical analyses, which show that success in the labor market does not just depend on the acquisition of technical skills, but that cognitive and non-cognitive skills in part acquired in early childhood and during basic and secondary schooling, are also important determinants of employment and earnings later in life.63 The literature has also found that the demand for higher-level cognitive skills, relative to demands for manual job-specific skills, tends to increase with technological progress, development, and diversification of the economy.64 62 World Bank (2013d). 63 Studies including Heckman et al. (2006), Carneiro and Heckman (2003), Heckman and Masterov (2007), and Cunha et al. (2006) show that cognitive skills and non-cognitive skills are important determinants of labor market outcomes, and the attainment of these skills begins in early childhood. 64 See Autor, Levy and Murnane (2003), for example. 74 4.3 Challenge 3: Gender Inequality in the Labor Market Despite significant progress in reducing the gender gap in education, women still experience persistent gaps in labor market outcomes. There has been significant progress in reducing the gender gap in education and in substantially increasing female labor force participation. This is attributable in part to growth of the labor-intensive and female- friendly RMG sector, particularly in urban areas.65 For instance, Heath and Mobarak (2014) found that exposure to the RMG sector delayed marriage and childbearing, boosted demand for education, and improved women’s employment outcomes. In rural areas, widespread microfinance, which provided small-scale loans to women, has contributed to increases in women’s labor market activities (Rahman and Islam, 2013). Although identifying causality may not be straightforward, women’s empowerment, improved health, reduced fertility, and increased educational attainment are among the key drivers for increasing female labor force participation. Compared to the impressive reduction of gender gaps in health and education, the labor market gender gap has remained persistently with female labor force participation lower than many other developing countries. This raises questions on the determinants of female labor force participation. Many girls, after leaving school, enter the out-of-the-labor-force status. Among many factors that may determine female labor force participation, data sources point to two key determinants: (i) household responsibilities; and (ii) mobility constraints. The age at first marriage varies significantly by gender: for women, by the age of 20 almost 80 percent are married, in comparison to 20 percent of males (Figure 18). Early marriage contributes to girls dropping out of school early, especially in rural areas. One of the incentives for poor parents to marry their daughters off early is a lower dowry amount than what would be required for an older bride. When young brides drop out of schools and rely heavily on husbands and in-laws it substantially hinders their ability to find employment opportunities outside homes. Early marriage results in early child-bearing, and family responsibilities imposed upon women by social norms severely impede women from seeking out employment opportunities. 65This sector is generally considered as female friendly, not only because they hire large numbers of women but also because the gender wage gap is smaller than for other sectors (Bhattacharya and Rahman, 1999). 75 Figure 18. Share of Those Never Married by Gender Source: LFS 2010. The findings from the Time Use Survey seem to support these arguments. Adult women spend more than twice the amount of time doing household work than men, even when they engage in paid employment in the labor market (Figure 19). Women in self-employment, agriculture, and service spend significantly less time than their male counterparts at work, yet take on more household responsibilities that they spend close to three times more than men in household work. On the other hand, working hours for individuals in regular wage employment and in the industry sector do not differ much by gender: both men and women spend more than 8 hours a day working. Nonetheless, women in regular wage employment spend twice as much time in household work than their male counterparts, at the expense of their leisure and personal care time. This finding suggests that women face greater challenges in maintaining a work-life balance. Figure 19. Time Use Pattern by Employment Status and Sector among the Employed: by Gender 12 12 2.0 2.2 10 1.0 10 3.0 0.9 3.9 1.0 2.9 1.4 1.5 4.2 Time (in hours) per day Time (in hours) per day 8.6 8 8.3 2.0 8.7 8 8.2 8.3 7.0 7.1 6.6 6 6 6.5 5.7 5.1 4.6 4 4 2 2 0 0 Male Female Male Female Male Female Male Female Male Female Male Female Regular Wage Irregular/Day Labor Self-Employed Agriculture Iindustry Service Paid Work Household Work Paid Work Household Work 76 As many previous studies have discussed, although controversial, Purdah, the religious and social practice of secluding women from public spaces and other social norms that restrict female mobility may explain the low female labor market participation rate (World Bank, 2008b). More recently, a study based on the Pathways Survey, conducted as part of a multi-country gender study, argues that women working within the home or economically inactive are far more likely to wear the burka/hijab (when they go out of the house) than those working outside the home. This suggests a strong correlation between religious conservatism (and consequential mobility constraint) and labor market activities. Regardless of whether religion is a dominant factor or not, a large body of literature seems to suggest that the norms, values, and culture affect labor market activities and the scope of mobility (Das, 2012). Issues of violence and public security for women, which have dominated the national discourse on gender, are also a serious impediment to women’s mobility and ability to access new opportunities. Furthermore, poor labor market outcomes and insufficient returns are also likely to discourage women’s labor market activities. As shown above, women are less likely to work in the industry or service sectors but more likely to work in agriculture compared to men; women are far more likely to work as unpaid family workers; time-use surveys also indicate that the average hours of work per day among paid workers are 6.9 hours for men and 5.2 hours for women, which suggests that women lack full time or high intensity employment. With respect to wages, women are paid less than men and the gap is larger for the lower end of the wage distribution than the higher end (Ahmed and Maitra, 2010; 2011). 4.4 Challenge 4: Labor Migration A large number of low-skilled migrants and potential migrants are exposed to substantial risks such as lack of information on the migration process and financial costs associated with recruiting agencies and intermediaries. In stark contrast to the stories of many successful migrants and the statistics that show the positive impacts of remittances on household welfare and the country’s economy, high-profile accidents and issues associated with international migrants, which were widely covered by the international media, highlight migration’s substantial risks. The risks are particularly high for aspirant migrant workers who attempt to cross borders to seek employment (although official statistics regarding accidents and abuse cases may not be readily available). For instance, the mid-sea rescue of migrant workers heading towards Malaysia and destined toward the northern coast of Indonesia illustrated the dire situation of the poor migrants who were seeking opportunities for overseas employment but were misled and misinformed about the high potential risks of this travel (New York Times, 2015). Alerts on human rights violations for migrant workers in the Arab Gulf countries — for instance in Qatar’s World Cup Stadium (Guardian, 2014), the United Arab Emirate’s western museum and college campus construction (Human Rights Watch, 2015), and domestic workers abuse in Saudi Arabia (Observer, 2014) — also emphasized the urgency for the receiving and sending countries’ governments to make concerted efforts to manage migration and to protect migrant workers. 77 Migration is hampered by the high costs generated by inefficiencies in the process, and spotty information about the migration process itself. Data on migrant households conducted by the International Organizations for Migration (IOM) in 2009, highlight the characteristics of (legal) migrants and several challenges (See Table A3 in Appendix).66 The vast majority of these migrants are low- or semi-skilled male workers who are heading to the Gulf countries or, to a lesser extent, to the Southeast Asia region, for employment opportunities (versus other purposes such as studying) without their spouse. 67 Much uncertainty is created in the absence of written contracts between the recruiter and the migrant, insufficient information about the migration process, restrictive visa arrangements, and the absence of mechanisms that could facilitate return and reentry. Workers tend to acquire the information on migration mainly through informal channels and their social networks (Massey and García España, 1987; Munshi, 2003). Those who are not supported by their social networks often acquire the necessary information from overseas middlemen or intermediaries (commonly known as Dalals). Data suggest that those heading to Southeast Asian countries or those with little education tend to acquire information from intermediaries (Figure 20). Only a tiny proportion of workers received information from recruiting agencies. Because workers acquire information from a variety of sources, the information quality each acquires likely varies extensively. Figure 20. The Source of Information on International Migration: by Destination and Education Source: Bangladesh Household Remittance Survey (BHRS, 2009) In line with previous studies, facilitating the migration process is very costly. IOM (2009) indicates that the average cost of migration was about BDT 200,000, while the average monthly earnings among migrants were BDT 21,000, of which BDT 13,000 was saved. Given that the national poverty line, which represents poor households’ income, is about BDT 1,600 per month, the average cost of migration could amount to 66 The analysis presented here is based on the BHRS, which was carried out from January to May 2009 by the IOM). BHRS provides information on a nationally representative sample of 10,673 migrant households from all six divisions of Bangladesh (a response rate was 97.7 percent). 67 Most common destinations among the Gulf countries are Saudi Arabia, United Arab Emirates, and Kuwait. 78 more than 10 years of wage income for low-skilled workers. Total costs incurred for migration vary with the facilitating entities (Figure 21): those who used intermediaries and recruiting agencies tend to pay slightly more for migration than those who used government or informal channels. Regardless of facilitating agency, migrant workers tend to pay significant fees to intermediaries. However, when intermediaries were used workers incurred very few extra costs other than fees directly paid to the intermediaries; whereas, when migrants used government channels, they had to make separate payments to agency and middlemen, and paid more additional fees. This suggests that government and recruiting agencies lack one-stop service, and bring extra hassles. To finance migration costs, more than three quarters of households took out loans and about one quarter of households sold land (See Table A3). These high costs may have been viewed as worth bearing if the migrant successfully migrated, obtained employment and remittance, and welfare gains were obtained for workers and households (Sharma and Zaman, 2009; IOM, 2009). However, for those who were not successful, these upfront costs could lead to severe household indebtedness. Figure 21. Composition of Migration Costs: By Facilitating Agency Source: BHRS (2009) Note: Percentage of workers who report facilitating agency is in brackets. They are not summed to 100% due to multiple answers. Even when migrants successfully migrate to the destination country with legal paperwork, a job offer, and contracts, they are often exposed to exploitation, forced labor, and occupational risks, and deprived of basic human rights. Workers are typically offered different jobs or lower wages than the contract had specified. Among those who received a job arrangement prior to migration, about one fifth later obtained a different type of job, and in more than a third of cases receive 80 percent or less of the original promise. Although many anecdotal cases describe poor working conditions and labor regulations that often fail to protect migrant workers (Human Rights Watch, 2015; ILO, 2014), a significant gap exists between data and knowledge of the various challenges hampering migrant workers’ welfare. Moreover, the cost of remittances, while gradually falling as financial services improve, is still relatively high and their impact is regressive. Surveys indicate that the typical poor migrant sends home about 79 US$200 or less per transaction, with a typical transfer fee of US$18. The fees are regressive, because smaller remittances, usually sent by poorer migrants, cost more per dollar sent (Migration Policy Institute 2007; World Bank 2013c). Substantial uncertainties abound regarding internal migration—including permanent rural-urban migration and seasonal migration during the lean seasons for workers seeking non-farm employment. Like international migration, internal migration involves intermediaries recruiting migrant workers into urban jobs. This process is plagued by uncertain information and an uncertain or absent contractual environment. This holds especially true in the absence of social networks to promote a softer landing for the migrant. Even when migrants arrive to a job at the new location, contractual arrangements shaped by middlemen can be unclearly defined and sometimes contain shaky legal relevance; in addition, migrants may find amenities and social support services hard to access or unavailable. The high prevalence of informality in the labor market and absence of contractual arrangements makes it difficult for GoB to enforce labor laws and weakens labor inspection. A recent focus group discussion with migrants raised awareness about the need for basic amenities, safety, and security for settlements of migrant workers (World Bank, 2014a). These circumstances, combined with uncertain and often short-term informal jobs, do allow increased consumption, but leave little room to accumulate savings for potential investment, even when wages in the city exceed rural wages. Another study identified low wages, irregular payment, long working hours, strenuous working conditions, and discrimination against women as major problems in the workplace (RMMRU, 2014b). 80 5 Policy Recommendations A strategic direction at the policy level is to develop an integrated social protection and labor system, which will help operationalize NSSS. A systems approach emphasizes balanced development and utilization of diverse policy instruments to achieve three intertwined objectives. Specifically, in the context of Bangladesh, a SPL system would require (i) improving existing social assistance programs, embracing more efficient administrative systems and arrangements for SPL programs; and (ii) utilizing a wider set of policy instruments to promote earnings opportunities. 5.1 Improving Existing Programs 5.1.1 Immediate Actions GoB recognizes that the main challenges in implementing social safety nets are improving coverage and targeting and reducing leakages. The Government’s NSSS presents a road map to a more inclusive and adequate social security system and prioritizes a few areas of actions: consolidation; unified targeting mechanism; monitoring and evaluation; and use of technology for administrative efficiency. Several initiatives have already commenced to address these issues. Consolidation: A first step to address fragmentation among similar programs and rationalize the delivery of scarce resources would be to consolidate programs with overlapping objectives and functions. Programs that might benefit from consolidation include those under the same ministry with similar objectives. The NSSS delineates five thematic clusters of ministries for such consolidation and harmonization. In line with this approach, the World Bank supports a Safety Net Systems for the Poorest (SNSP) project which harmonizes five programs under the MoDMR –three public works (EGPP, FFW, TR) and two humanitarian reliefs (GR, VGF). Small-scale programs, particularly untargeted, universal subsidies, also need to be carefully reviewed for consolidation. In consolidating and streamlining existing programs, it is important to consider risks along the lifecycle. Unified targeting system: This type of system enables various programs to identify the poor, and can significantly improve targeting accuracy and program efficiency. As discussed above, the GoB has already made some progress in this area and is committed to developing a NHD, which will contain key household characteristics to determine poverty status based on the PMT methodology.68 Although NHD is being developed under the SNSP project under the MoDMR and BBS, it will be important to establish proper data-sharing protocols and coordination mechanisms across ministries so that the database can be utilized for multiple purposes. In addition, strategies need to be developed to handle appeals and grievances for correcting any targeting errors in a simple and transparent manner and for updating the PMT formula and database on a regular basis. Bangladesh’s move towards adopting an improved poverty targeting system is in line with a regional trend in South Asia, where countries facing similar challenges make efforts to improve SPL systems and establish a robust targeting mechanism. In Pakistan, the flagship safety net program, the Benazir Income Support Programme (BISP), uses PMT-based poverty score cards to select 68 For the detailed discussion regarding the PMT methodology in Bangladesh, see BBS (2014a, b) and Sharif (2009). 81 poor households and their female beneficiaries for a basic income support. Sri Lanka also endeavors to rationalize the institutional framework of all safety net and welfare programs, and to improve the targeting performance of its largest safety net program (the Samurdhi food stamp program). Additionally, the Government of India is in the process of enrolling a billion citizens with biometric identity cards under the Unique Identification Authority of India, to improve public service delivery and to more accurately identify beneficiaries eligible for subsidies. Improved efficiency: Weak administrative capacity and systems introduce inefficiencies and scope for fraud and corruption. This is particularly prominent in food-based programs as discussed above. It is important then to strengthen the administrative system of providing safety nets while phasing out food- based programs. In particular, PFDS needs a modernization reform. Developing a more efficient reserve management system, with an optimal size of food stocks, options for regular monetization, and modern storage technologies, is necessary. In addition, technology can be employed for overall efficiency gain throughout all safety net programs. 69 For instance, a dynamic MIS allows record-keeping and consolidation of information on various aspects of program implementation, such as beneficiary selection and numbers, utilization of program expenditures, and grievances and redresses. With respect to the delivery of services, utilizing effective technology, such as mobile banking or cash cards can lower transaction costs for the beneficiary and reduce leakages. Two safety net programs (ISPP and EGPP) have already begun using electronic debit cards linked to the Bangladeshi postal system to provide cash transfers directly to beneficiaries. Although mobile phone payments are still limited, GoB is moving in this direction to work with the private sector in this area. GoB’s efforts to establish a robust grievances and redresses are also underway. Monitoring and evaluation: Monitoring and evaluation (M&E) are distinct tools used to assess whether programs are on track in achieving their intended results. Monitoring programs provide information on how much financial resources have been spent, how many beneficiaries the program is reaching, and how efficiently the program is serving them. Monitoring is an essential management tool for supplying regular information so that appropriate action can be taken to improve or modify a program’s implementation. Program evaluation is usually an external assessment of a program’s effectiveness to determine whether the program has met the defined standard and estimate its net results and impact. To date, the practice of monitoring and evaluating program effectiveness, let alone rigorous impact evaluation, is not very common; it only takes place on an ad-hoc basis. A strong recommendation is that monitoring and evaluation activities are mainstreamed as a core component of GoB’s social sector programs. A monitoring and evaluation system can consist of a formulation of expected results, selection of output and outcome indicators, data collection from a baseline along key milestones; and a timeline for progress. This can be done by establishing a system for collecting, analyzing and reporting data; monitoring progress; and evaluating the activity to determine its relevance and efficiency. The Income Support Program for the Poorest (ISPP) showcases a well thought-through safety net program (Box 10). The program is based on robust evaluation and evidence, establishes a strong administrative 69 See Muralidharan et al. 2014 for an example. 82 platform and MIS, and builds upon strong leadership of the local government division. All these features are expected to contribute to improved efficiency and strong M&E. Box 10. Shombhob and Income Support Program for the Poorest (ISPP) Background: Although GoB has made significant progress in reducing child and maternal mortality, the country is among the 10 countries with the highest prevalence of malnutrition. The World Health Organization estimates 41 percent children below the age of 5 in Bangladesh are stunted. International evidence suggests that cognitive development in the first five years of a child’s life from conception is critically important and a determining factor in one’s earning capacity. Therefore, investment in children’s health and nutrition is a cost-effective way of reducing inter-generational transmission of poverty. Program (Shombhob): To help address this challenge with greater efficiency, a pilot initiative called ‘Shombhob’ (meaning “possible” in Bangla) was introduced. The pilot took place in two rural areas of the poorest region (Ranpur Division) and one urban slum (Narayanganj City slum). The benefits include BDT 400 for mothers, conditional upon the following: They had to monitor their children’s monthly growth; they had to attend meetings on nutrition (for children up to three years old), and they had to ensure their children attended school more than 80 percent of the time (for children ages 6 to 15). A few noteworthy administrative features were included: (i) a proxy means test based on a poverty scorecard was used to objectively select the deserving poor; (ii) electronic transfers were made through electronic cash cards via the post office system; (iii) a computerized MIS was established to make cash payment and beneficiaries’ compliance traceable; and (iv) for institutional arrangement local government (Union Parishads) played significant roles to improve program accountability at the union level. Impacts (Shombhob): An impact evaluation study found that the program had a great impact on nutrition among young children ages 10 to 22 months, reducing cases of wasting by 40 percent; the impacts were muted among older children and little impact was found for school attendance. The consumption especially of nutritious food significantly increased among the beneficiary families; and mothers’ understanding on nutrition also improved. From the administrative perspective, mobile payment was efficient and transparent, reducing leakages and administrative costs, and the targeting performance was better than other safety net programs. Program (ISPP): In 2014, GoB supported by the World Bank, launched the ISPP project to combat malnutrition and to improve health outcomes among mothers and young children, building upon the success of the Shombhob pilot. Many positive features were carried over with a large scale. The program provides monthly conditional cash transfers to selected 600,000 beneficiaries in 42 of the poorest Upazilas in the country, to help reduce poverty while improving their children's nutrition and cognitive development. The project is implemented by the Local Government Division (LGD) under the Ministry of Local Government, Rural Development and Cooperatives (MoLRD). The cash transfer is conditional. It states that all eligible pregnant women and mothers of children below five years old must (i) attend monthly awareness sessions on nutrition; and (ii) regularly monitor the growth of their children below three years of age, to receive benefits. On fulfilling their responsibilities, the participating mothers receive cash transfers electronically into their own post office accounts using smart cash cards. An important component of the project is establishing common administrative platforms at the union level for safety net beneficiaries. The project team works closely with the local Union Parishads to help implement the cash transfers. The Union Parishads maintain a single beneficiary registry of at least five safety net programs to ensure beneficiaries are accurately monitored, which is an important step in the direction of consolidating data from existing safety net programs. Source: Ferre and Sharif (2014); World Bank (2014b). 83 5.1.2 Considerations for future programming There is a general consensus that well-targeted cash transfers Box 11. Ongoing Discussions on SPL Programs can be effective as an anti-poverty strategy, public works as Some debates and discussions regarding a crisis response, and the ALMPs for employment, among priority for better SPL programs with others, supported by rigorous evaluations. However, there efficiency, impacts on poverty reduction, are several discussions and debates regarding the best and sustainability include: practices or design of these programs. Evidence is still mixed,  Conditional vs. unconditional transfers and more experiments and learning are warranted. Some  Cash vs. in-kind benefits prominent discussions, as will be relevant for future SPL  Public works vs. public works plus  Individual vs. community programming in Bangladesh, are presented here (Box 11 for  Urban vs. rural summary).  Conditional vs unconditional cash transfers. Over the past decade, conditional cash transfer programs (CCTs) have become popular as an alternative, or a complement, to cash transfers without any conditionality. Both are targeted at poor families and provide cash to participating families. While cash transfers regardless of conditionality aim to reduce the current level of poverty in a household by transferring resources, CCTs have an additional objective of influencing future levels of poverty by requiring beneficiaries to make behavioral changes for investments in the human capital of members of the household, usually focusing on children’s health, nutrition, and education. With their emphasis on building for the future, CCTs have spread rapidly in the developing world and overall they have shown positive impacts. 70 The transfer element has increased consumption levels and stabilized income flows of participating households; while the conditionality element has led to significant improvements in school enrollment and attendance, especially in countries where those figures were low to start with. Recently, the role of conditionality is being questioned. A few studies suggest that unconditional cash transfers show equivalently positive or better impacts, 71 and conditionality requires significant administrative capacity, including collaboration across sectors, to be able to provide effective services while monitoring compliance. However, existing evidence still weighs in the importance of conditionality, alternatively co-responsibility, in promoting “desirable behavior” of households and ensuring service delivery of service providers. Therefore, in order to fully achieve the intended objectives of CCTs, administrative capacity – particularly abilities to coordinate between different sectors and operate M&E measures – should be promoted together.  Cash versus in-kind benefits. While safety net programs provide benefits in both cash and in-kind, often in the form of food in the latter case, there has been a gradual shift from food-based towards cash (or near cash – food stamps or vouchers) programs. Each form of benefits has pros and cons. 70See Fiszbein and Shady (2009) for summary. 71The impact evaluation of a small cash transfer in Malawi’s Zomba district illustrates the power of an experimental design to determine a cash transfer’s effects. The Zomba CT, which was designed explicitly for research purposes, tested the impact of conditional and unconditional transfers on educational, marriage, and fertility outcomes for female adolescents and found equivalent or better outcomes from unconditional transfers (See Baird et al. 2011). See also Garcia et al. (2012) for a discussion on emerging trends of unconditional cash transfer in the Africa region. 84 While direct cash payments are obviously a less-costly means of alleviating current poverty by allowing the recipients to allocate the resources according to their needs, cash may not be appropriate in circumstances where the supply of essential goods and services has been disrupted, or the economy faces unstable value of cash (e.g., hyperinflation). To the contrary, food aid can be a preferred option in acute emergencies and conflict situations. However, a common criticism against food-based benefits is that they are logistically costly and complicated, may distort local food production and markets, and do not necessarily match the needs of the beneficiaries. Thus, a general trend is to improve the payment system (e.g., direct payment to a bank account, debit card, or mobile banking) and provide benefits in cash while phasing out food-based programs mainly for emergency relief and reforming food distribution systems.72  Public works versus public works plus. Public works schemes are a means of transferring resources through temporary employment creation, often targeted at workers from poor households,73 and at building local socioeconomic infrastructure. In some instances, they also serve as insurance schemes that guarantee periods of low-wage employment when demanded by workers.74 Public works are useful to relieve a transient shortage of labor demand, especially during times of crises,75 but limited in serving as a pathway to longer-term employment opportunities. Moreover, the cost of transferring a unit of cash is likely to exceed the value of the transferred benefit unless assets and infrastructure are built, repaired, or maintained through public works. Given the limitation, a “public works plus” model has emerged. This model provides the added provision of training or access to credit to improve participants’ chances of obtaining permanent employment in line with the objectives of ALMPs and to better manage community assets and infrastructure collectively (e.g., Ethiopia’s Productive Safety Net Program). The evaluations of the model highlight the potential to create a path out of poverty by linking a public works program to employment and community services, but the program will require more careful planning and design given the complexity of the intervention.  Rural versus Urban. Poverty and vulnerability is generally more prevalent in rural than urban areas. Rural households tend to rely on agriculture and have fewer chances of engaging in non-agricultural employment opportunities, they are exposed to seasonality and extreme shocks (e.g., weather, food price), and have limited access to basic services such as health and education, compared to their urban counterparts. Many SPL programs are indeed geared towards rural households and workers, such as public works during the agricultural lean seasons. In Bangladesh, it is likely that poverty will continue to be a rural phenomenon like in many other developing countries. However, as urbanization is part of mega-trends, more people are migrating to urban and peri-urban centers, and many form urban slum communities and engage in informal market activities, the issue of urban poverty is emerging as an important policy agenda. Worldwide, the urban share of poverty out of total poverty is increasing 72 See Gentilini (2014) for more detailed discussions. 73 Public works schemes are often self-targeted, i.e. wages are set below the market wage for low and unskilled labor, making workfare attractive mainly for the unemployed and underemployed. 74 For instance, India’s Mahatma Gandhi National Rural Employment Guarantee Program. 75 For instance, Argentina’s Jefes y Jefas program during 2002 economic crisis and Korea’s public works during 1998 Asian financial crisis provide good examples. See Galasso and Ravallion (2004) for evaluation of Jefes y Jefas program; See Subbarao (2003) for public works in Asian countries. 85 primarily due to increases in urban population (e.g., Ravallion et al. 2007). A simple comparison of poverty rates between urban and rural areas may underestimate the prevalence of urban poverty given the higher living costs in urban areas. The nature of urban poverty may be different from the challenges and issues in rural areas: poor living conditions, often without a stable supply of water, sanitation, and electricity, are a big concern for urban poor families, particularly children. The availability of short-term employment opportunities as day laborers, for instance, can provide a source of income; however, urban workers face huge uncertainties and risks. Moreover, social stigma, illegality of settlements and businesses, and frequent mobility likely hamper the psychosocial wellbeing of poor households in urban areas along with their economic distress. In light of the situation, policy efforts are needed to adapt SPL programs to urban settings while continuing support to the rural poor. 5.2 Utilizing a Wider Set of Policy Instruments to Manage Risks and Promote Earnings Opportunities To strengthen other SPL programs apart from social assistance, it is necessary to explore and utilize various policy instruments. Given the four issues addressed throughout this paper, the following section discusses different menus of options and potential programs in line with the various objectives. First, we discuss possible programs to address poverty and vulnerability. Then, we discuss potential programs to promote youth employment, followed by programs to promote female empowerment in the labor market; and finally, programs to promote migration support as an employment strategy. 5.2.1 Potential Programs to Address Poverty and Vulnerability Voluntary insurance and savings: Building an adequate but fiscally sustainable pension system is important given the ongoing demographic transition. The NSSS has already delineated a three-pillared pension system that focuses on social pensions for the poor elderly, voluntary pensions for private sector workers, and a mandatory social insurance scheme for the formal sector. While it may be necessary in the long run to develop a formal, mandatory and contributory social insurance such as a universal health insurance or a national pension system, in the short term it is necessary to promote voluntary saving and use of the private insurance market. The primary functions of insurance and savings are to help contributors hedge various risks to avoid destitution and smooth consumption. While insurance focuses more on cross-sectoral risk management, savings enable individuals to manage inter-temporal risks. GoB has traditionally responded to the repeated natural disasters and seasonality of agricultural deprivation through humanitarian aids and reliefs, which were criticized for the lack of clear targeting or lack of timely response. Once established, insurance schemes are less expensive to administer and more timely in the event of shocks, which can lead to a more transparent and sustainable way to cope with agricultural risks. These insurance schemes can be privately developed and implemented, but subsidized by public funds depending on its externalities (See World Bank, 2010, for more detailed discussions on the insurance sector in Bangladesh). 86 Recently, several innovative approaches are being developed and tested to implement incentive- compatible insurance and saving schemes. Examples include weather index-based insurance, community- based health insurance, and extra incentives for saving.76 Weather index insurance is a financial product linked to a weather index highly correlated to local agricultural yields or household incomes. Contracts are written against specific perils or events such as drought, flood, or other natural calamities that a pre- defined regional area may experience. The means for measuring and verifying the incidence or severity of the specific events must be defined prior to launching the financial product. Not only individual farmers, but also relief agencies and NGOs can purchase these insurance products. Community-based health insurance programs, organized and managed in communities, are a voluntary risk pooling mechanism that can provide limited and affordable insurance to the poor. It adopts common insurance rules where participants pay a pre-defined premium and receive financial support in case of health shocks. In many cases, insurance is often based on group solidarity and trust, and organized or facilitated by local NGOs, unions, or local MFIs. Along with insurance, savings improves individuals' ability to manage risk by turning irregular cash flows into lump sums for future expenditures when needed. In the absence of insurance, the importance of savings is even greater for poor people’s welfare. An increasing body of research emphasizes the need for saving for the poor, and investigates why the poor do not save more. The conventional belief has been that the primary reason for the poor not saving is financial exclusion: many do not have access to financial institutions or recognize the need for financial products. In Bangladesh, the Global Findex data (2014) suggests that less than a third of adults have a bank account.77 This is a surprisingly low statistic given the high penetration rate of MFIs in Bangladesh. Even if access to savings is granted or informal savings mechanisms exist, the poor often fail to save or tend to save insufficiently to cope with various risks. Common reasons for undersaving among the poor include credit constraints, myopia, transaction costs, and self-control, among others (Dupas and Robinson, 2013). Increasingly, international research on savings for the poor finds that providing slight nudges such as commitment saving, subsidies, frequent reminders, or visual rewards, can encourage saving (Box 12). The market penetration of financial institutions in Bangladesh yields great potential for improving financial inclusion and promoting more savings for the poor. MFIs, rather than relying solely on traditional microcredits, may be able to expand the scope of their operation and invest in diverse financial products including savings. Studies such as Collins et al. (2010) have emphasized that greater impacts of the availability of MFIs indeed come from savings more than lending. Marketing strategies that emphasize the cost of not saving, saving technologies that lower transaction costs, and frequent nudges and reminders can substantially improve savings outcomes for poor individuals and families, in a very cost- effective way. In addition, combining these tools for encouraging savings with training sessions on how to increase financial capability can enhance the impact of the program. 76There are many other insurance schemes such as crop, livestock, property, and so on. 77See the summary indicators from the Global Findex Data – Bangladesh for more information: http://datatopics.worldbank.org/financialinclusion/country/bangladesh 87 Box 12. Providing Saving Incentives  Commitment saving Rationale: Individuals tend to be myopic and inpatient for near-future trade-offs than for future trade-offs. Combined with the lack of self-control, such myopia leads to withdrawals of saving. Intervention: The project designed and implemented a commitment saving product called a SEED (Save, Earn, Enjoy Deposits) in a rural region in the Philippines. Some distinguishable features of this product are: individuals restrict the rights to withdraw funds until they reached a goal amount or a specific month; each individual was encouraged to set a specific purpose for saving (for tuition, birthdays, or fiestas, for instance); a locked box was provided that enabled frequent deposits but did not allow withdrawals from the deposit. Results: The commitment savings product was effective and attracted those who may face more issues of self- control (maybe due to the demand from other members of the household), and significantly increased the amount of savings.  Frequent reminders for saving Rationale: Inattention to potential future expenditure due to a limited attention span may be part of the reasons why the poor save little. Intervention: The experiments took place in the Philippines, Peru, and Bolivia. Although the financial products and design of treatment slightly varied across these countries, the main intervention was frequent reminders to save through mail (Peru) and text messages (other places). Results: Simple reminders to save increased saving rates by about 6 percent. When the reminders included a target amount of saving or specific future expenditure, the impacts were larger. In addition, visual materials that made the gains from saving (or the losses from not saving) more vivid, concrete, and tangible, made the reminders more effective. Source: Ashraf et al. (2006) and Karlan et al. (2011) A multipronged graduation approach: Establishing links between safety nets, financial support, and employment opportunities can help the poor graduate from social assistance programs and acquire sustainable means of livelihoods. A pioneer ‘graduation approach,’ which combines microfinance with asset transfers (primarily livestock) and skills training for the ultra-poor, was indeed originated from BRAC in Bangladesh. The sequencing of the interventions was carefully planned; with cash stipend as part of safety nets, followed by skills training and transfer of income-generating assets to participants in the program. Compulsory savings was also introduced as part of the program to develop a practice of saving among participants. Regular community meetings were held to build links and social connections between village elites and program participants. Health care advice was provided, and intensive coaching through weekly household visits enabled households to develop a business plan. As a result of the multi-pronged intensive treatment, over two-thirds of participants were able to escape ultra-poverty within two years, and remain on an upward trajectory (See Bandiera et al. 2013). The success of BRAC’s innovative program in Bangladesh inspired a multi-country pilot with a common design to add external validity of the approach. An international research consortium under the Consultative Group to Assist the Poor (CGAP) evaluated the program in a diverse range of contexts 88 including India, Pakistan, Honduras, Peru, Ethiopia, Yemen, and Ghana to understand the impact of this model and its scalability (See Box 13). The findings suggest that the provision of assets, cash transfer, skills training, home visits for counseling, promotion of savings, and basic health services has yielded great impacts on the lives of the rural poor by encouraging self-employment activities. Although such a multipronged approach incurs substantial costs, the results showed that the benefits from the intervention far exceeded the cost. Building upon such experience and linking with existing social safety nets, GoB can implement a program which will help the ultra-poor strengthen their employability and livelihood, and eventually graduate poverty. Box 13. Graduation Projects Background: The needs of the ultra-poor, which are often chronically food insecure and rely on subsistence employment particularly in rural areas, require more than a stand-alone cash transfer or training program. In order to meet the complex needs of the ultra-poor and enable them to build sustainable means of livelihoods, this holistic “graduation” program combined a series of policy instruments with a careful sequencing scheme. Building upon encouraging results from BRAC’s initial program in Bangladesh, an attempt to identify the casual impact of this approach with external validity led to a six country pilot with RCTs. In this pilot, the core design of the program remained similar for each country, while implementation was often accommodated to the context of each country. Program: The program was provided to the poorest of the poor within poor (rural) villages. Once poor villages were selected based on poverty statistics and consultation, ultra-poor households in the selected villages were identified through a participatory procedure where villagers themselves create an economic ranking of all households. The core benefits of the program included (i) one time asset transfers such as live stocks, (ii) regular cash transfers to support consumption, (iii) skills training to manage the productive assets and gain self- employment opportunities; (iv) home visits for coaching and counseling for two years; (v) access to a savings account and encouragement (or requirement) for saving; and (vi) basic health services and life skills training. Impacts: Although there was heterogeneity across countries, overall, significant increases/improvements were found in (i) consumption (per capita consumption and food security); (ii) productive assets; (iii) household income and revenue; (iv) physical and mental health; and (v) political involvement and female empowerment. The positive impacts lasted two and three years after the completion of asset transfer. The program was proven to be cost effective; the internal rate of return ranged from 6.9 percent (Ghana) to 23.4 percent (India) at the discount rate of 5 percent. Source: Banerjee et al. (2015). 5.2.2 Potential programs to promote youth employment Skills development: Lack of skills is considered to be one of the key constraints to productive employment, and skills training may be the best chance to improve the prospects of youth who are struggling in the labor market. However, due to limited availability, costs, and quality, many young individuals do not have access to the TVET opportunities or fully benefit from training. Particularly, vulnerable populations from disadvantaged backgrounds may not fulfill the academic requirements to enter the TVET system nor have the means to pursue further training. Consequently, poor young workers likely engage in the low-skilled informal sector. These issues can be in part addressed by interventions in ALMPs, which could target out- of-school youth for their skills development. 89 International evidence shows that training alone is unable to significantly improve the employability of vulnerable youth, but that comprehensive programs offering practical on-the-job experience and job search assistance in addition to in-class and life skills training can be quite successful (Box 14). Latin America’s Jovenes programs (Argentina, Chile, Dominican Republic, Colombia, Peru, and Panama) replicated in many parts of the world provide a good example of the comprehensive approach described above. For instance, Jovenes en Accion (Youth in Action) project in Colombia, one of the early models of this kind, provided three months of classroom and three months of on-the-job training to unemployed youth whose families were in the lowest two deciles of income distribution in the seven largest cities of the country. Attanasio et al. (2011) found that the intervention significantly improved employment and earnings outcomes particularly among the poor and highlighted the importance of on-the-job training. Building upon this model, supplemental services are provided to further help vulnerable youth find employment opportunities in the labor market. International evidence on ALMPs in the developed countries suggests that job search assistance has yielded positive and significant impacts with relatively low costs (Kluve and Card, 2012). Projects often require training providers to include job search assistance as part of services to improve employment outcomes. More recently, a great emphasis is placed on non- cognitive (socioemotional, behavioral, soft) skills as a critical determinant of employment outcomes. Non- cognitive skills such as self-regulation, self-confidence, motivation, goals, social skills, or emotional stability have been shown to affect employment outcomes (Heckman et al., 2006; Heckman and Kautz, 2012). As a result, an increasing number of youth employment projects are incorporating soft skills training as part of core intervention. In Bangladesh, employment initiatives for vulnerable youth may be introduced starting from urban areas where opportunities for wage employment exist. Such initiatives can focus on specific sectors, especially labor-intensive ones for greater impacts, such as the RMG, construction sectors, and other service sectors. Combinations of job and soft skills training, and mentoring for labor market activities and resource management (e.g., savings) can be considered. Given the low level of education among the poor youth, it may be necessary to utilize training providers who can modify the curriculum and pedagogy for the needs of this group versus relying on the conventional TVET sector. With respect to the employer side, in order to ensure that the program is screening potential employers who are willing to take less skilled and vulnerable youth, it is useful to assess the profiles of small enterprises in their sector, firm size, demand for extra workers, and constraints to additional hiring, etc. Alternatively, public works modified to the context of urban settings for youth, probably combined with training can provide large scale short-term employment that could lead to longer term opportunities. For instance, the Urban Productive Safety Net Program (PSNP) in Ethiopia is introducing public works geared towards urban settings including a range of works such as urban greenery and forestation efforts, solid waste management, construction and maintenance of roads, building drainages and community infrastructure, and development of shelters and workshops for income-generation activities, and others.78 78 Urban Productive Safety Net Program, draft project paper, World Bank. 90 Box 14. Jovenes Type Programs for Wage Employment Background: In the early 1990s, the labor market conditions for youth in Latin American countries were quite pessimistic. The youth unemployment rate was almost three times higher than the adult rate, and the gap between low and high income quintiles was expanding. In response to this situation, governments introduced emergency training programs for vulnerable youth. Dominican Republic’s Juventud y Empleo (youth and employment) and Colombia’s Jovenes en Accion (Youth in Action) were among the new youth employment projects. In the early 2000s, the impacts of these two programs were rigorously evaluated through randomized impact evaluations. Intervention: The Jovenes programs have three distinct features: (i) targeting low-income, out-of-school, vulnerable youth who may not have labor market skills; (ii) providing in-class training (job and life skills); and (iii) providing wage subsidies for internships. Disadvantaged young people are identified using the labor market status, socioeconomic status survey, and poverty mapping, for instance. Training institutions and NGOs provide a package of services including skills training and internship arrangements with local employers to receive funds. In addition to job skills, intensive training is provided on life skills – problem solving, workplace behavior, self-confidence, conflict management, and job search assistance. Impacts: Evaluations of these programs found significant impacts on employment and earnings. Evaluations in Argentina, Chile, Colombia, Dominican Republic (DR), Mexico, Panama, and Peru show heterogeneous impacts in magnitude (0 to 20 percent) but robust results. The impact on employment rate was much higher for women than men (Colombia, Panama, and Peru) and for younger workers (Argentina and Chile). Increases in wages attributable to the program varied widely from insignificant (Argentina and Mexico) to 22 percent (women in Colombia), and 38 percent (Women in Panama). The likelihood of gaining formal employment significantly improved. The key lesson from these evaluations is that a comprehensive support that includes in-class training, workplace training, and supplemental services combined has the greatest impact and is cost effective in improving employment and earnings in the labor market. However, the results are very heterogeneous particularly with the design (e.g., duration and intensity of training, type of supplementary support) and implementation quality as well as demographic groups. Based on Ibarrarán and Shady (2009); World Development Report - Jobs (2013) Entrepreneurship: In the labor market in developing countries, the majority of workers find their employment opportunities from business activities and self-employment rather than wage employment. As discussed above, less than one fifth of young workers in Bangladesh are wage employees, and non- 91 wage employment is the main source of jobs. Therefore, in the context of a strategy to increase the incomes of the poorest workers, it is important to identify interventions to support productive self- employment. These poor self-employed are likely subsistence entrepreneurs with few other options in the labor market, and the constraints they face and the type of support they need differ from higher skilled individuals. Unfortunately, the evidence base of what works to promote the livelihoods of workers through productive self-employment is limited (Box 15). A few programs have been successful, but they tend to be small, and there is limited knowledge available about design and implementation in different contexts and with different types of beneficiaries. It is only recently that the impacts of entrepreneurship intervention for vulnerable workers has begun to be synthesized, and some promising patterns have emerged.79 Many programs use combinations of skills training, financing support (through grants, microcredits, or in-kind investment), counseling and mentoring, or networking services. The key outcomes of interest include labor market activities, earnings and profits, savings and borrowing behaviors, knowledge and business practice, and household income. Studies increasingly emphasize the importance of a holistic approach (e.g., combined training and financing versus stand-alone training) to improve the living standards through self-employment. While enhancing business and employment outcomes is generally challenging, better impacts are observed for youth. In Bangladesh, initiatives to promote self-employment among vulnerable youth may be introduced for business development. Promising models can be adapted to the context of Bangladesh with the lessons learned incorporated in the design. For instance, an evaluation of the Adolescent Girls Initiatives (AGI) in Liberia found that the project providing skills training with complementary services for saving and childcare significantly improved the employment and earnings outcomes of young girls. 80 Uganda’s program provided cash grants so that a group of youth can find the best use of funding, whether this is investment in skills or equipment for business, to operate businesses. 81 In urban Nairobi, Kenya, the International Rescue Committee introduced a microfranchising intervention or cash grants for business for young girls, which provided standardized know-hows and supplies of products for business for profit sharing.82 79 See Cho and Honorati (2014) and Grimm and Paffhausen (2015) for results from meta analyses. 80 See the summary of the results: http://www.worldbank.org/en/programs/adolescent-girls-initiative 81 See Blattman et al. (2014) 82 See Brudevold et al. (2015). 92 Box 15. Designing Entrepreneurship Programs for Youth Constraints: Potential constraints and factors that may affect productivity and earnings through self-employment can be classified in two groups: individual and environmental. Individual constraints include lack of entrepreneurial aptitude and traits, as well as skills and social capital. Environmental constraints include limited access to financial services and products, an adverse business environment, underdeveloped markets for inputs and outputs, and other impeding cultural factors. Even if the market and regulatory failures that create environmental constraints could be removed, the lack of aptitude, skills, and cognitive abilities to understand, discover, and develop business opportunities traps subsistence entrepreneurs in a low-level productivity equilibrium. Technical Component: Entrepreneurship programs provide combinations of (i) training; (ii) financial support; (iii) counseling and mentoring; and (iv) networking and market linkages. Training is a key method for building entrepreneurial capacity and is by far the most commonly provided service. Programs adopt different methods of training (classroom-based training; workplace training, including apprenticeships and on-the-job training; and home/firm visits). There is great heterogeneity across programs in terms of training intensity and duration. Financing support is mostly provided by microcredit or grants, but with an increasing emphasis on financial inclusion, savings are also included as a means to facilitate resource management among the poor as in the case of the Adolescent Girls Initiatives. Counseling and mentoring seems to be an important contributing factor to better results as many small business activities rely on social capital. Finally, networking services include facilitation of economies of scale by horizontally linking related businesses or by vertically linking different players along value chains. Target groups: Common target groups for intervention to promote increased earnings through self-employment have been microfinance clients, social assistance beneficiaries, small business holders, or vulnerable youth. Impacts: The likelihood that small-scale entrepreneurship programs have a positive impact varies by outcomes of interest. Intermediate outcomes such as business practice and knowledge and attitudes present much better impacts than final outcomes such as labor market incomes and activities. However, overall impacts are not strong with the majority showing insignificant impacts, if not negative. Still the evidence base is thin, and more experiments and studies are needed. Source: Based on Cho et al. (2015) and Cho and Honorati (2014). 93 5.2.3 Potential programs to promote women’s labor market activities Programs to promote female labor force participation often target women only. By targeting women and directing the benefits of intervention exclusively for female workers, programs aim to promote female labor market activities and improved incomes. Many youth employment programs mentioned above, for instance Adolescent Girls Initiatives (Afghanistan, Haiti, Jordan, Lao, Liberia, Nepal, and Rwanda), Kenya’s microfranchising, and Jordan New Work Opportunities for Women program, exclusively target young women. Meanwhile, programs targeted for both genders often show differential impacts by gender. Some evaluations — business support mainly by grant and technical assistance in Sri Lanka (De Mel et al. 2009) and apprenticeship for vulnerable youth in Malawi (Cho et al. 2013) — show significantly weaker impacts for women than men. 83 Cho et al. (2013) indeed delved deeper into differential constraints to employment faced by males and females separately. The study suggested that women’s employment tended to be affected more by external factors (vis-a-vis their own decisions) and female trainees had to draw down more savings to support them with accommodation and subsistence even if they received stipends compared to their male counterparts. Programs need to recognize and address extra constraints faced by women to employment to result in stronger impacts. From the labor demand side, vicinity and accessibility to female-friendly employment opportunities such as the RMG sector factories could substantially improve female employment, which leads to delays in marriage and childbearing, and improvements in wellbeing (Heath and Mobarak, 2014). International evidence suggests that reforms in family law or property rights for women could also improve women’s access to productive assets and labor market activities (World Bank, 2011). From the supply side, the two notable constraints faced by women, as emphasized in the analyses above, are household responsibilities (mainly associated with early marriage and childbearing) and mobility constraints (mainly associated with cultural norm and physical safety and security). Thus, reducing the cost of employment outside the house associated with these two factors can potentially be effective to increases in labor force participation. Therefore, the design of programs to promote women’s labor market activities would be similar to that of youth employment programs with added measures to reduce female-specific constraints. Evaluations of the Adolescent Girls Initiatives suggest that the provision of child care substantially enhanced the programs’ effectiveness. A project in Cambodia, which provided cleaner, fuel-efficient cook stoves, enabled women to substantially reduce the time spent on collecting firewood and improve labor market activities (World Bank, 2012c). Measures to relieve time for household responsibilities can help women’s labor market participation. Similarly, mobility constraints can be addressed by innovative intervention as well. The provision of mobile training in Pakistan facilitating women’s training at or near home greatly encouraged women’s participation in training (Cheema et al. 2012). On a slightly different context, an intervention in India distributing free bicycles to secondary school enrollees significantly increased the school attendance rates by improving female mobility outside the village (Muralidharan and Niehaus, 2013). Likewise, providing safe transport exclusively available for women can reduce mobility constraints 83 To the contrary, the impacts of Latin America’s Jovenes programs were generally stronger for females. 94 while ensuring their physical safety. 5.2.4 Potential programs for migration as an employment strategy The main policy agenda is to make migration more (i) remunerative; (ii) affordable; (iii) inclusive; and (iv) safe (World Bank, 2015b), as part of an employment strategy. The reason that remittance per worker is low in Bangladesh and migration is not as remunerative as it can be is because international migration has been concentrated on low-skilled workers’ migration. There is a need to improve the skills level of migrant workers and expand the types of jobs available from overseas employment. Given the high out-of-pocket cost of migration, the recruitment process must be made more transparent and efficient, in order to reduce the cost of migration and facilitate international migration among the poor. Finally, the entire migration process and involved agencies must be better managed and regulated for safer migration. To achieve these objectives, programs to promote successful international migration in Bangladesh should focus on three intertwined areas: (i) expanding and strengthening G2G programs; (ii) facilitating safe migration; and (iii) providing skills development for potential migrant workers. Expanding and strengthening G2G programs Building upon previous experience, the GoB can expand G2G programs to other countries that are already in high demand for Bangladeshi labor. Based on the experience with the governments of Korea and Malaysia, it would be useful to conduct comprehensive evaluations and assessments from the perspectives of sending country, receiving country (and employers), and migrant workers. Other countries’ experiences, such as Australia’s or New Zealand’s seasonal worker programs for Pacific Islanders, or the Philippines’ bilateral arrangements with roughly 25 countries, can also be useful to review in strengthening G2G programs (See Gibson and McKenzie, 2008; McKenzie and Yang, 2015). In essence, the G2G program should identify a meaningful number of migrant workers; visa specifications; employment arrangements; and clear rules and regulations for worker protection while in the country.84 Doing this requires close coordination and collaboration with the governments of receiving countries as this is often the policy area of the receiving country and not within the control of the sending country. However, it would be important for the GoB to establish a clear set of principles and policies in developing and strengthening G2G programs. Facilitating safe migration Given the large upfront costs associated with arranging international migration and high rates of failure, GoB urgently needs to make the migration process safe for migrant workers. There are multiple stages where GoB can intervene through partnership with NGOs and other community-based organizations: pre- decision; pre-departure; during migration; and post-return (see Box 16). The need to have more extensive public outreach campaigns to improve awareness on available services, and to reduce reliance on middlemen is a key learning from recent implementation experience. Given the prevalence of intermediaries in facilitating migration and transmission of remittances in Bangladesh, it is critical to 84 See World Bank (2008a) for a discussion on some other measures to ensure migrant workers’ welfare. 95 manage and formalize the activities of intermediaries so that workers are less vulnerable to abuse, human trafficking, or fraud. Upgrading skills Employment for low skilled workers in the mostly menial sectors (e.g., Malaysia’s plantation sector; Gulf countries’ construction sector) has been dominating for international migration among Bangladeshis. It is critical to develop worker skills to position them for better paying jobs. Doing this can substantially improve the development impacts of international migration. Concurrently, soft skills should be fostered among potential migrants so that they can better adjust to new working environments in foreign countries and more effectively work with employers. Depending on the destination country, training in the language and culture can be a useful supplement to skills training.85 85For instance, Chiswick and Miller (1995) suggest that level of earnings and level of proficiency in the language of the receiving country are positively correlated. 96 Box 16. Policies for Safe International Migration Pre-decision Intervention: For potential migrant workers to make informed decisions on international migration, it is important to make relevant information widely available and accessible. As discussed in Beam et al. (2014), two types of activities can be envisaged: (i) information about overseas employment opportunities, migration process, financing options, passport processing, and so on; and (ii) assistance on job search by establishing a job matching website or holding a job fair. The evidence based on the impacts of this type of intervention is quite weak, whether the information and assistance can induce more workers to decide to migrate or change the composition of workers who apply for the jobs. However, this intervention can improve knowledge of overseas jobs. Pre-departure Intervention: For those who have already decided to migrate, a few training and orientation programs can be useful. These programs include (i) life skills and basic language training to facilitate the transition to life abroad, (ii) orientation on measures to prevent abuse and protect workers ’ rights; and (iii) training on financial literacy and procedures and options for making remittances. Given the importance of remittances on the welfare of migrant workers, their families, and receiving communities ’ economy, an increasing emphasis is being placed on financial literacy training with respect to financial planning, savings, debt management, remittances transactions, and so on. Studies generally show positive impacts of financial training and stressed the importance of training provision to both migrant workers and the remaining family members. During-migration policies: While varying by destination country, a set of basic rights – such as equal treatment with nationals with respect to access to labor regulations including minimum wages, health and safety rules, the ability to form associations and unions, the right to change employment, among others – should be accessible and communicated to migrant workers. The GoB should facilitate networking through their embassies to the migrant workers so that they exchange information and provide support to each other. In addition, a series of interventions should take place in destination countries for migrant workers to (i) access financial institutions to facilitate the transmission of remittances, (ii) reduce the costs and fees of sending remittances, and (iii) improve financial management including savings access. With respect to remittances, MFIs with their footprint all over the country are uniquely positioned to play a role in facilitating the transactions. However, regulatory constraints currently prevent most MFIs/NGOs from offering international remittance services or credit products to families of migrants. Consequently, Bangladesh’s two largest MFIs (BRAC and Grameen) have established commercial banks that can transfer remittance through their wide network and cater their services to non-resident Bangladeshis who wish to remit money to the country. Given the significance of remittances, tailored financial and saving products that address the needs of the migrant’s families would be an important area of development. Post migration: The return migrants can bring accumulated skills, experience, and financial resources back to their communities. The policy for returning migrants is to focus on reintegration and readjustment to the communities and labor markets. For instance, the Overseas Foreign Worker program in the Philippines provides training and psychosocial support for returning migrants. However, examples of how return migrants are utilizing their new skills and experience when they return are limited, and very little is known about the best practices to effectively utilize the returned work force. Source: Elaborated based on McKenzie and Yang (2015) and World Bank (2013d) 97 98 6 Conclusion From Social Transfers to Jobs for the Poor The challenge of poverty reduction and inclusive growth in Bangladesh is not amenable to simple solutions. The country faces an intricate web of challenges and opportunities that lie ahead in the rapidly changing environment. For Bangladesh to continue to reduce poverty while sustaining continued growth, it will be critical to extend policy attention beyond providing safety nets and coping with risks ex-post towards promoting opportunities and reducing risks ex-ante. Key policies include early intervention that starts from maternal health and nutrition that leads to early childhood intervention, as recognized and suggested by the NSSS. Concurrently, a balanced development of policy tools must be available for SPL to maximize the impacts on poverty reduction. In particular, policies and programs to promote more, better, and inclusive job opportunities deserve more attention. A Longer Term Job Strategy Needed A jobs agenda requires both longer term strategy and short-term action plans. A longer term jobs strategy must be based on careful analytical work and strategic thinking. A daunting list of policy agenda items is available as follows:  Consider the global trend in labor mobility and international trades and formulate policies to tap into global markets  Review the business environment for job creation and the environment for foreign investment  Examine the progress in structural transformation in the country and identify promising engines of growth  Put into place infrastructure, including logistics, transportation, and Information and Communication Technology (ICT) which facilitate many economic activities  Consider climate change and its consequences and propose environment-friendly jobs policies  Provide additional policy attention to urbanization and special development  Ensure the provision of fundamentals through the education and health systems which contribute to the production of an adequately skilled work force. This type of strategy may benefit from an in-depth Jobs Diagnostic, which analyzes the current state of both the demand and supply sides of jobs and delves into jobs strategies. Such strategies can be a great complement to the Government’s NSSS in improving the wellbeing of the poor by building sustainable means of livelihoods. Short Term Action Plans for the Poor and Vulnerable While the medium- to long-term jobs strategy will help the country move forward in achieving its goal as a middle income country by 2021,86 short-term action plans specifically targeting the poor and vulnerable to improve their earnings opportunities must be in place, particularly for economic growth to be more inclusive. The review has discussed the strategic importance of paying policy attention to the poor, youth, women, and migrant workers. Appropriately designed interventions to address the challenges of these 86 See Government of Bangladesh (2011) for a plan 99 groups will complement existing safety nets and further contribute to poverty reduction and inclusive growth. International evidence on various models provides a solid ground on which Bangladesh can design and implement its own initiatives. For Bangladesh to harness the benefits of demographic dividends, utilize the globalized markets, and seize the benefits of shared prosperity, immediate actions are urgently needed. 100 References Ahmed, S., and Maitra, P. (2010). “Gender wage discrimination in rural and urban labor markets of Bangladesh.” Oxford Development Studies, 38(1). Ahmed, S., and Maitra, P. (2011). “A distributional analysis of the gender wage gap in Bangladesh.” Monash University. (mimeo) Ahmed, A., Quisumbing, A., Nasreen, M., Hoddinott, J. and Bryan, E. (2009). “Comparing food and cash transfers to the ultra-poor in Bangladesh.” International Food Policy Research Institute, Research Monograph 163. Washington, DC. Ahmed, A. (2013). “Profile of the feed-the-future-zone in the South and other regions of Bangladesh.” Presentation at a workshop. International Food Policy Research Institute, Dhaka, January 16, 2013. Ashraf, N., Karlan, D. and Yin, W. (2006). “Tying odysseus to the mast: Evidence from a commitment savings product in the Philippines.” Quarterly Journal of Economics, 121(2). Attanasio, O., A. Kugler, and C. Meghir (2011). "Subsidizing vocational training for disadvantaged youth in developing countries: Evidence from a randomized trial," American Economic Journal: Applied Economics 3:188-220. Autor, D., Levy, F., and Murnane, R. (2003). “The skill content of recent technological change: an empirical exploration.” Quarterly Journal of Economics, 118(4). Baird, S, McIntosh, C. and Ozler, B. (2011). “Cash or condition: Evidence from a cash transfer experiment.” Quarterly Journal of Economics, 126(4). Bandiera, O., Burgess, R., Das, N. Gulesci, S. Rasul, I., Shams, R. and Sulaiman, M. (2013). “Can basic entrepreneurship transform the economic lives of the poor?” London School of Economics. UK. (mimeo) Banerjee, A. and Duflo, E. (2006). “The economic lives of the poor.” Journal of Economic Perspectives, 21(1). Banerjee, A., Duflo, E., Glennerster, R., and Kinnan, C.G. (2013). “The miracle of microfinance? Evidence from a randomized evaluation.” National Bureau of Economic Research. Working Paper No. 18950. Banerjee, A., Duflo, E., Goldberg, N, Karlan, D. Osei, R., Pariente, W., Shapiro, J., Thuysbaert, B., Udry, C. (2015). “A multifaceted program causes lasting progress for the very poor: Evidence from six countries.” Science, 348 (6236). 101 BBS. (2014a). Design and Implementation of Bangladesh Poverty Database. Bangladesh Bureau of Statistics. Dhaka, Bangladesh. BBS. (2014b). Proxy Means Test formula for Bangladesh. Bangladesh Bureau of Statistics. Dhaka, Bangladesh. BBS. (2014c). Slum Census and Floating Population Count 2014. Bangladesh Bureau of Statistics. Dhaka, Bangladesh. Beam, E., McKenzie, D. and Yang, D. (2014). “Unilateral Facilitation Does Not Raise International Labor Migration from the Philippines.” Economic Development and Cultural Change, forthcoming. Bhatt, N. and Tang, S. (1998). “The problem of transaction costs in group-based microlending: An Institutional Perspective.” World Development, 26(4). Bhattacharya, D. and Rahman, M. (1999). “Female employment under export propelled industrialization: Prospects for internalizing global opportunities in the Bangladesh apparel sector.” Occasional Paper No. 10. The United Nations Research Institute for Social Development. Geneva. Blackden, M. and Wodon, Q. (2006). “Gender, time-use and poverty in Sub-Saharan Africa.” World Bank Working Paper No. 73. Washington, DC: World Bank. Blattman, C., Fiala, N., and Martinez, S. (2014). “Generating skilled self-employment in developing countries: Experimental evidence from Uganda.” Quarterly Journal of Economics, 129(2). Brudevold, A., Honorati, M., Jakiela, P., and Ozier, O. (2015). “A firm of one’s own: The impacts of microfranchising on young women in Nairobi.” University of Maryland. (mimeo). Carneiro, P., and J. Heckman. (2003). “Human capital policy.” In Inequality in America: What role for human capital policies? ed. James J. Heckman, Alan B. Krueger, and Benjamin M. Friedman. Cambridge, MA: MIT Press Cheema, A., Khwaja, A., Naseer, F., Shapiro, J., Lodhi, A., Sheikh, S., Siddiqui, S., Tourek, G., Niazi, M. and Shoaib, A. (2012). “Punjab economic opportunities program: The skills for employability evaluation report.” Center for Economic Research in Pakistan. (mimeo). Chen, G. and Rutherford, S. (2013). “A microcredit crisis averted: The case of Bangladesh.” Focus Note. No. 87. Consultative Group to Assist the Poor (CGAP). Chiswick, Barry and Miller, P. (1995). “The endogeneity between language and earnings: International analyses.” Journal of Labor Economics, 13(2). Cho, Y., Kalomba, D., Mobarak, A. M., Orozco, V., and Wolfson, D. (2013). “Gender differences in the effects of vocational training: Constraints on women and drop-out behavior.” IZA discussion paper, No.7408. Cho, Y and Honorati, M. (2014). “Entrepreneurship programs in developing countries: A meta-regression approach,” Labour Economics. 28 (C). 102 Cho, Y., Robalino, D., and Watson, S. (2015). “Supporting self-employment and small-scale entrepreneurship: Potential programs to improve livelihoods for vulnerable workers.” IZA Journal of Labor Policy. forthcoming. Collins, D., Murdoch, J., Rutherford, S., and Ruthven, O. (2010). Portfolios of the Poor: How the World’s Poor Live on $2 a Day. Princeton: Princeton University Press. Cunha, F., Heckman, J., Lochner, L., and Masterov, D. (2006). “Interpreting the evidence on life cycle skill formation,” In Handbook of the economics of education, ed. Eric A. Hanushek and Frank Welch. Amsterdam: North-Holland Das, M. (2012). “Exclusion and discrimination in the labor market.” Background Paper for the World Development Report on Jobs 2013. Das, N., de Janvry, A., Mahmood, S., and Sadoulet, E. (2014). “Migration as a risky enterprise: A diagnostic for Bangladesh.” University of California at Berkeley. (mimeo). de Mel, S., McKenzie, D. and Woodruff, C. (2009). “Are women more credit constrained? Experimental evidence on gender and microenterprise returns,” American Economic Journal: Applied Economics, 1(3):1-32. Dupas, P. and Robinson, J. (2013). “Why don’t the poor save more? Evidence from health savings experiments.” American Economic Review, 103(4). Ferre, C. and Sharif, I. (2014). “Can conditional cash transfers improve education and nutrition outcomes for poor children in Bangladesh? Evidence from a pilot project.” World Bank Policy Research Working Paper, No. 7077. Fiszbein, A. and Shady, N. (2009). Conditional cash transfers: Reducing current and future poverty. Washington, DC: World Bank. Gallaso, E. and Ravallion, M. (2004). “Social protection in a crisis: Argentina’s Plan Jefes y Jefas.” World Bank Economic Review, 18(3). Garcia, M., Charity, M., and Moore, T. (2012). “The cash dividend: The rise of cash transfer programs in Sub-Saharan Africa.” Directions in Human Development, No. 67208. Washington, DC. Gentilini, U. (2014). “Our daily bread: What is the evidence on comparing cash versus food transfers?” Social Protection and Labor Discussion Paper No. 1420. World Bank. Gentilini, U., Honorati, M., and Yemtsov, R. (2014). The state of social safety nets 2014. Washington, DC: World Bank. Gibson, J. and McKenzie, D. (2008). “Australia’s Pacific seasonal worker pilot scheme: Development impacts in the first two years.” World Bank. (mimeo). 103 Gindling, T. and Newhouse, D. (2014). “Self-employed in the developing world,” World Development. 26(C). Government of Bangladesh. (2015). National Social Security Strategy of Bangladesh, General Economic Division, Planning Commission, Dhaka: Bangladesh. Grimm, M. and Paffhausen, A. (2015), “Do interventions targeted at micro-entrepreneurs and small and medium-sized firms create jobs? A systematic review of the evidence for low and middle income countries,” Labour Economics, 32(1). Hanushek, E., and Kimko, D. (2000). “Schooling, labor force quality, and the growth of nations.” American Economic Review, 90(5). Hanushek., E. and Woessmann., L. (2012) “Do better schools lead to more growth? Cognitive skills, economic outcomes, and causation,” Journal of Economic Growth, 17(4). Hashemi, S., Schuler, R., Riley, A. (1996). “Rural credit programs and women's empowerment in Bangladesh.” World Development, 24(4). Heath, R. and Mobarak, M. (2014). “Manufacturing growth and the lives of Bangladeshi women.” National Bureau of Economic Research Working Paper No. 20383. Heckman, J.A., Stixrud, J., and Urzua, S. (2006). The effects of cognitive and noncognitive abilities on labor market outcomes and social behavior. Journal of Labor Economics, 24(3). Heckman, J. A. and Kautz, T. (2012). “Hard evidence on soft skills.” IZA Discussion Paper 6580. Human Rights Watch. (2015). “Migrant workers’ rights on Saadiyat Island in the United Arab Emirates.” Human Rights Watch. Ibarrarán, P. and Rosas-Shady, D. (2009). “Evaluating the impact of job training programmes in Latin America: Evidence from IDB funded operations.” Journal of Development Effectiveness, 1(2). ILO (2014). In the corridors of remittances – Cost and use of remittances in Bangladesh. International Labour Organization. Country Office in Bangladesh. IOM (2009). The Bangladesh Household Remittance Survey 2009: Summary Report. International Organization for Migration. Javoy, E. and Rozas, D. (2013). Introducing MIMOSA: Microfinance market capacity measurement tool. http://www.cgap.org/blog/introducing-mimosa-microfinance-market-capacity-measurement- tool. Consultative Group to Assist the Poor. Kabeer, N. (2009). “Conflicts over credit: Re-evaluating the empowerment potential of loans to women in rural Bangladesh.” World Development, 29(1). 104 Karim, L. (2011). Microfinance and its discontents: Women in debt in Bangladesh. London: Minnesota. University of Minnesota Press. Karlan, D., McConnell, M. Mullainathan, S., and Zinman, J. (2011). “Getting to the top of mind: How reminders increase saving.” National Bureau of Economic Research Working Paper No.16205. Khandker, S.R., and Samad, H.A. (2013). “Are microcredit participants in Bangladesh trapped in poverty and debt?” World Bank Policy Research Working Paper No. 6404. Washington, DC: World Bank Khandker, S.R., and Samad, H.A. (2014). “Dynamic effects of microcredit in Bangladesh.” World Bank Policy Research Working Paper No. 6821. Washington, DC: World Bank. Krauss, A. L., Lontzek, J.M., and Frommelt, M. (2012). “Lack of access or crowded markets? Towards a better understanding of microfinance market penetration.” University of Zurich. (mimeo). Krueger, A., and Lindahl, M. (2001). “Education for growth: Why and for whom?” Journal of Economic Literature, 39(4). Massey, D. and García España, F. (1987). “The social process of international migration,” Science, 237(4816). Matsumoto, M. and Elder, S. (2010). “Characterizing school-to-work transitions of young men and women: Evidence from the ILO school-to-work transition surveys.” International Labour Office. (mimeo). McKenzie, D. and Yang, D. (2015). “Evidence on policies to Increase the development impacts of international migration.” World Bank Research Observer. 30(2). Migration Policy Institute. (2007). “Leveraging remittances for development.” Policy Brief. Munshi, K. (2003). “Networks in the modern economy: Mexican migrants in the U.S. labor market.” Quarterly Journal of Economics, 118(2). Muralidharan, K., Niehaus, P., and Sukhtankar, S. (2014). “Payments infrastructure and the performance of public programs: Evidence from biometric smartcards in India.” National Bureau of Economic Research Working Paper No. 19999. Cambridge. Muralidharan, K. and Niehaus, P. (2013). “Cycling to school: Increasing secondary school enrollment for girls in India.” National Bureau of Economic Research Working Paper No. 19305. Cambridge. Nomura, S., Rahman, M., Goyal, S., Nakata, S., and Al-Zayed, S. (2015). “How does the short-term training program contribute to skills development in Bangladesh: A tracer study of the short-term training graduates.” South Asia Education Development Series. Washington DC: World Bank. Pitt, M. and Khandker, S. (1998). “The impact of group-based credit programs on poor households in Bangladesh: Does the gender of participants matter?” The Journal of Political Economy, 106(5). 105 Rahman, R. and Islam, R. (2013). “Female labour force participation in Bangladesh: Trends, drivers and barriers.” ILO Asia-Pacific Working Paper Series. Ravallion, M., Chen, S. and Sangraula, P. (2007). “New evidence on the urbanization of global poverty.” World Bank Policy Research Working Paper No. 4199. Washington, DC. Roodman, D. and Morduch, J. (2009). “The impact of microcredit on the poor in Bangladesh: Revisiting the evidence.” Center for Global Development. Working Paper 174. RMMRU. (2014a). “Impact of migration on poverty and development.” Policy Brief No. 14. Refugee and Migratory Movements Research Unit, University of Dhaka. RMMRU. (2014b). “Internal migrant workers and the construction sector in Bangladesh: Tackling informality and exploitative labour practices.” Policy Brief No. 11. Refugee and Migratory Movements Research Unit, University of Dhaka. Ruthbah, U. (2013). “The employment generation for the poor: a process evaluation.” University of Dhaka. (mimeo). Sharif, I. (2009). “Building a targeting system for Bangladesh based on Proxy Means Testing.” Social Protection Discussion Paper No. 0914. World Bank. Sharma, M. and Zaman, H. (2009). “Who migrates overseas and is it worth their while? An assessment of household survey data from Bangladesh.” World Bank Policy Research Working Paper No. 5018. Siddiqui, T., and Abrar, C.R. (2003). “Migrant worker remittances and microfinance in Bangladesh.” Working Paper 38, Social Finance Program, ILO, Geneva. Subbarao, K. (2003). “Systemic shocks and social protection: Role and effectiveness of public works programs.” Social Protection Discussion Paper Series No. 0302. World Bank. Tietjen, K. (2003). “The Bangladesh primary education stipend project: A descriptive analysis.” World Bank. (mimeo). UNFPA. (2015). “The impact of the demographic transition on socioeconomic development in Bangladesh: Future prospects and implications for public policy.” United Nations Population Fund. Bangladesh Country Office. UNICEF (2010). Understanding Urban Inequalities in Bangladesh: A Prerequisite for Achieving Vision 2021. United Nations Children’s Fund. Dhaka, Bangladesh. UNICEF (2014). The State of the World’s Children 2015: Reimagine the Future, Innovation for Every Child. United Nations Children’s Fund. New York. USAID (2014). Bangladesh Labor Assessment. Dhaka: Bangladesh. 106 World Bank. (2008a). “Protecting temporary workers: Migrant welfare funds from developing countries.” Migration and Development Brief No. 7. Pu World Bank. (2008b). “Whispers to voices: Gender and social transformation in Bangladesh.” Bangladesh Development Series, Paper no. 22. Washington, DC: World Bank. World Bank (2010). “Agricultural Insurance in Bangladesh: Promoting Access to Small and Marginal Farmers” Report No. 53081-BD. Dhaka: World Bank. World Bank (2011). Women, Business and the Law 2012: Removing Barriers to Economic Inclusion. Washington, DC: World Bank. World Bank (2012a). The World Bank 2012-2022 Social Protection and Labor Strategy: Resilience, Equity, and Opportunity. Washington, DC: World Bank. World Bank (2012b). More and Better Jobs in South Asia. Washington, DC: World Bank. World Bank. (2012c). “Building on tradition as the way to women’s empowerment in Cambodia.” East Asia and Pacific Region Social Development Notes. Washington, DC: World Bank. World Bank (2013a). “Bangladesh poverty assessment: Assessing a decade of progress in reducing poverty (2000–2010).” Bangladesh Development Series, Paper No. 31. Washington, DC: World Bank. World Bank (2013b). “Safe migration for Bangladeshi workers.” Project Information Document, Washington, DC: World Bank. World Bank (2013c). Bangladesh: The Path to Middle-income Status from an Urban Perspective. Washington, DC: World Bank. World Bank (2013d). “Barriers to access to payment systems in sending countries and proposed solutions.” Special Purpose Note. Washington, DC: World Bank. World Bank (2013e). Education Sector Review: Seeding Fertile Ground: Education That Works for Bangladesh. Washington, DC: World Bank. World Bank (2014a). “Employability, mobility and skills for the poor in South Asia: A qualitative study in Bangladesh, Bhutan and India.” Praxis – Institute for Participatory Practices. World Bank (2014b). “Income support program for the poorest.” Project Appraisal Document, Washington, DC: World Bank. 107 World Bank (2015a). “Addressing inequality in South Asia” in South Asia Development Matters. Washington, DC: World Bank. World Bank (2015b). “More and better jobs to accelerate shared growth and end extreme poverty” Bangladesh Systematic Country Diagnostic. Washington, DC: World Bank. World Bank (2015c). Bangladesh Development Update, The World Bank Office, Dhaka. Yunus. M. (2003). Banker to the Poor: Micro-Lending and the Battle Against World Poverty. Public Affairs. Yunus. M. (2008). A World Without Poverty: Social Business and the Future of Capitalism. Public Affairs. 108 Appendix: Data A wide range of data sets is used or referred to in this review. The two key data sets include the LFS and HIES. The LFS provides information on the nationally representative sample of individuals above the age of 15 on their labor market activities and performance. The HIES has been instrumental in assessing the country’s poverty rates and performance of social programs for the poor by providing information on household consumption and characteristics for the nationally representative sample of households. Multiple rounds of these two data sets are available and provide key information to facilitate the understanding of economic activities of the work force and their poverty profiles; supplementary data are used to inform the labor demand and supply sides as well as the state of SPL programs. These include enterprise surveys and financial inclusion database (demand side), skills surveys, time use survey, demographic and health survey and migration survey (supply side), and household survey for poverty monitoring and SPL coverage (Table A1). The detailed descriptions of these data are as follows. Table A1. List of key microdata used or discussed in the report Main Data Used/Discussed Years Availability of raw data 2003, 2005, 2010 Fully available Labor Force Survey 2013 Key indicators available Household Income and Expenditure Survey (HIES) 2000, 2005, 2010 Fully available Enterprise Survey 2007, 2013 Fully available - Enterprise Survey Innovation Follow up 2013 Fully available Entreprise Skills Survey (ESS) 2012 Fully available Bangladesh Time Use Survey 2012 Fully available Bangladesh Household Remittance Survey 2009 Fully available (BHRS) *Key indicators were available for LFS 2013 as of October 2015. 1. Labor Force Surveys: Conducted by the BBS, this survey provides key information on the labor market activities of nationally representative adults (age 15 and above) in the country. Labor force surveys conducted in 2003, 2006, and 2010 have provided key labor market indicators of the nationally representative sample of various populations groups over time. The most recent LFS was conducted in 2013, the final public release is still on hold. Due to an increased emphasis on the importance of obtaining reliable and frequent labor market data, GoB has decided to conduct the survey every quarter starting July 2015 with a rotational panel scheme incorporated. For this purpose, a total sample size of about 31,000 households, intends to deliver reliable quarterly estimates of unemployment and other relevant labor force indicators for 24 analytic domains: 10 City Corporations, and the rural and urban portions of the country’s seven divisions. For the analysis above, we have used 2010 LFS extensively and key indicators from 2013 LFS. 2. HIES: Conducted by BBS, this survey provides key information that can analyze the wellbeing of nationally representative households in the country, including the characteristics, dwelling conditions, 109 economic activities and incomes, consumption, and assets. HIES has been conducted every five years since 1995; the most recent HIES was conducted in 2010, and BBS is preparing for its 2015 HIES. As the importance of reliable and more frequent income data mainly for poverty monitoring has been increasingly emphasized, BBS is considering conducting HIES every three years. 3. Enterprise Survey: Conducted by the World Bank, this survey focuses on the many factors that shape the business environment by providing information on infrastructure, trade, finance, regulations, taxes and business licensing, corruption, crime and informality, and perceptions about obstacles to doing business. It is a firm-level survey of a representative sample of an economy’s private sector. Enterprise surveys include both qualitative and quantitative data collected through the surveys to connect a country’s business environment characteristics with firm productivity and performance. Bangladesh’s most recent enterprise survey was conducted in 2013 for a representative sample of firms in the non- agricultural formal private sector. Business owners and top managers in 1,442 firms, among which the largest sectors include garments and other manufacturing, were interviewed. The majority of the interviewed firms pointed out political instability, electricity, and access to finance as top business constraints. Given the importance of innovation in firms’ growth and to improve the measurement of innovation at the firm level, a follow-up survey was conducted in 2013 for a subset of firms interviewed for the Enterprise Survey. The total number of firms re-interviewed was 990 out of 1,442. 4. Enterprise Skills Survey: In order to inform the skills mismatches and gaps in the formal sector, the World Bank Education Team conducted an employer-employee matching skills survey in 2012. The employer survey interviews with business owners or top managers, while the employee survey interviews with the randomly selected employees from the roster of employees in each establishment. The survey measured cognitive, non-cognitive, and technical skills. Based on business registry 2009, which contains the registered firms with more than 10 employees, 500 firms in manufacturing, commerce, financial and insurance, public administration, and education, and 7,254 employees were interviewed. Apart from generally low education and the gender gap, a few striking findings emerge from the survey: worker turnover from the formal sector is so low that almost 90 percent of workers responded that they have never changed jobs; the majority of Bangladeshi workers lack basic literacy and numeracy skills; both employers and employees assess that non-cognitive skills are more important than cognitive skills. 5. Bangladesh Time Use Survey: Conducted by BBS, this survey provides information on how Bangladeshi adults spend their time. The TUS was conducted in 2012, and the sample included nationally representative 3,780 households from all 7 divisions. Of 3,780 households, 2,380 households (63 percent) are from rural areas and the remainder is from urban areas. All household members above 15 years of age were surveyed for the TUS, which resulted in 10,002 individuals. For household information and characteristics, a standard household survey (similar to the questionnaire of the Labor Force Survey) was used, while a time diary was implemented for recording their time use every 30 minutes for one week day and one weekend during the survey week. There are broadly two different ways to record time use. One is a time diary method where each individual records his/her use of time for certain activities every 30 (or other pre-designated) minutes. The other 110 method is based on face-to-face recall interviews. Studies have found that the diary method yields better quality data and also captures simultaneous activities compared to the retrospective interviews (Blackden and Wodon, 2006). The 2012 TUS primarily used the diary method, but the high level of illiteracy underscored the difficulty of employing this method as diary was recorded by someone else in the household or supplemented by enumerators by face-to-face recalls. Recorded activities, as per International Classification of Activities for Time Use Statistics, were classified into 15 different types, then three System of National Accounting categories, and five different work classifications (Table A2). Table A2. Time Use Survey: activity categories ICATUS SNA Category Work Activity Description 01 Corporations/quasi-corporations, non-profit and government 02 Household enterprises in primary production 03 SNA Economic activities Household enterprises in non-primary production 04 Household enterprises in construction 05 Household enterprises providing services for income 06 Unpaid domestic services for own final use within the Household work household 07 Non-SNA Unpaid care-giving services to household members 08 Community services and help to other households Other work 09 Learning/education 10 Socializing and community participation 11 Attending cultural, entertainment, and sports events Non-productive 12 Leisure Engaging in hobbies, games, and other leisure time activities 13 Indoor and outdoor sports participation 14 Use of mass media 15 Personal Personal care and maintenance ICATUS: International Classification of Activities for Time Use Statistics SNA: System of National Accounting 111 6. Bangladesh Household Remittance Survey (BHRS): Conducted from January to May in 2009 by the IOM, this survey provides data from a nationally representative sample of 10,673 migrant households from all six divisions of Bangladesh and had a response rate of 97.7 percent. The advantage of this survey is that it does not only ask remittance related questions, but also includes detailed information on migrants and migrant sending households and hence is a good data source for analyzing Bangladeshi migrant’s remittance behavior. The survey utilized a stratified sampling method using the district level concentration of migrant households as the defining characteristic of strata (based on 2001 Census). Three divisions — Chittagong, Dhaka, and Sylhet — which contain higher concentrations of migrant households than other divisions participated in this survey; two strata from each of these divisions were selected. The remaining divisions also participated in the survey but only one stratum was selected in each of these divisions. For the selected nine strata, clusters were defined and selected using probability proportional to the number of migrant households. Then, a census approach was used for all clusters to identify all migrant households. A migrant household was defined as a household that has at least one of its members living or working abroad. All the identified migrant households from the household listing of the selected clusters were included in the sample. The initial sample size selected following the above-mentioned procedure was 10,926 households, out of which 10,673 households were successfully interviewed.87 About 17 percent of the surveyed migrant households had more than one migrant. Among all the migrants included in the survey, 24 percent migrated in the last year of the survey implementation. A data limitation is that migrant information was not reported by the migrants themselves but by a member of the migrant- sending households. Thus, information about migrant’s income or savings in the country of destination or even information about the migration process may not be absolutely reliable. Nonetheless, this information provides a window to observe and understand migrants’ experiences and behavior from the perspective of a migrant-sending household. The characteristics of migrant workers are presented in Table A3 by destination region. Table A3. Characteristics of Migrant Workers by the Destination Region Gulf countries Asia Others * Current age (years) 32.1 29.9 33.1 Female (%) 1.8 0.5 5.5 Years of education (%) No schooling 10.8 9.6 0.9 1 to 5 29.2 24.6 7.9 6 to SSC 50.6 53.4 34.0 HSC 5.8 8.5 15.4 At least graduate 2.3 3.3 39.2 Household head/spouse of the head (%) 31.6 26.5 8.3 Spouse accompanied when married (%) 2.7 5.1 52.3 87 For a detailed description of the sampling, see IOM (2009, 2010). 112 Duration of migration (%) Less than 2 years 35.2 63.1 27.0 3 to 5 years 22.2 13.5 25.1 6 to 10 years 23.3 9.4 21.6 More than 10 years 19.2 13.9 26.3 Took a loan/dissaving to finance mig. (%) 75.2 78.3 44.0 Sold land to finance mig. (%) 24.3 26.7 15.5 Facilitator of migration (%) Government/formal agency 8.1 12.5 24.5 Intermediaries 38.0 53.0 18.5 Relatives/friends 44.0 23.5 37.7 Own efforts 9.8 11.1 19.2 No. of observations 9,423 2,838 680 Source: Bangladesh Household Remittance Survey, conducted by IOM in 2009 * Others generally include the developed world such as North America, Europe, and Australia and New Zealand. 113