Bangladesh Enabling faster and more equal income growth in rural Bangladesh Rural Income Diagnostic Maria Eugenia Genoni Mansur Ahmed Pablo Antonio Tillan Madhur Gautam Poverty and Equity Global Practice Agriculture and Food Global Practice © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staȂ of The World Bank with e3ternal contributions. The ‫ޔ‬ndings, inter- pretations, and conclusions e3pressed in this work do not necessarily re‫ޕ‬ect the views of The World Bank, its Board of E3ecutive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are speci‫ޔ‬cally reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Genoni M.E., M.M. Ahmed, P.A. Tillan, and M. Gautam (2021). Bangladesh Rural Income Diagnostic: Enabling faster and more equal income growth in rural Bangladesh. Washington DC. © World Bank. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: %&01/+0)1&,+40+,1 /"1"!6%",/)!+(+!0%,2)!+,1" ,+0&!"/"!+,ƛ&- cial World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by the World Bank. Third-party content—The World Bank does not necessarily own each component of the content con- tained within the work. The World Bank therefore does not warrant that the use of any third-party- owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. E3amples of components can include, but are not limited to, tables, ‫ޔ‬gures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. Contents Acknowledgments V Abbreviations VI Executive summary 7 Introduction and objectives of this review 17 Data sources and levels of analysis 23 Main data sources 23 Levels of analysis 25 ey de‫ޔ‬nitions 2ч Section 1. Assessing context and heterogeneity - What do we know about rural incomes and poverty in Bangladesh? 31 Strong rural poverty reduction but with increased spatial disparities 32 The role of diȂerent economic sectors in poverty reduction and East-West disparities 35 Understanding rural incomes 41 Rapid income transformations—that remain vulnerable 41 Three factors behind heterogeneity in rural incomes 53 Rural income levers: household assets and markets 71 Education and earning power: literacy gaps hamper progress in the rural West 71 Average land holdings are small, and land markets could work better 75 Key agricultural inputs: irrigation, fertilizers, seeds, and mechanization 78 Electricity access has fueled consumption growth and poverty reduction, but gaps persist 81 Connectivity challenges go beyond infrastructure 83 Commercialization and ‫ޔ‬nancial markets 88 BANGLADESH RURAL INCOME DIAGNOSTICS Section 2. Analyzing opportunities and constraints for rural income growth 99 Raising incomes in agriculture 99 Raising non-farm and migration incomes 104 Raising womenҁs incomes 10ч onstraints to increasing rural incomes 10ч Section 3. Setting priorities for action: Which constraints should be tackled ‫ޔ‬rst? 121 Policy focus and regulations 122 Investment gaps 123 Market-related constraints 123 Knowledge and capacity constraints 123 Gender norms 123 Section 4. Identifying policy solutions 129 Closing remarks 135 Appendix 1 – Summary statistics by quintile 139 Appendix 2 – Summary statistics by division 139 Appendix 3 – Opportunities and constraints in the livestock and ‫ޔ‬sheries sectors 140 References 143 IV Acknowledgments This report has been prepared jointly by the Poverty and Equity Global Practice and the Agriculture and Food Global Practice of the World Bank Group. The World Bank Group team was comprised by Maria Eugenia Genoni (co-TTL), Md. Mansur Ahmed (co-TTL), Madhur Gautam, and Pablo Antonio Tillan. Outstanding inputs were pro- vided by oaquin Endara, Claudia N. Berg, Thea Hilhorst, and Ale3ander Irwin. The authors are grateful to Ruth Hill, Michael Morris, Gayatri Acharya, and Willem Janssen for peer reviewing this report. The authors are also grateful for comments and discussions provided at diȂerent stages of the preparation of the report by Akhter Ahmed, Faizuddin Ahmed, Muneeza Mehmood Alam, Sylvia Maria Cesar, Johannes Hoogeveen, Afsana I. Khan, Nandini Krishnan, Gladys Lopez-Acevedo, Aissatou Ouedraogo, Nethra Palaniswamy, Zillur Rahman, Juan Segnana, Binayak Sen, Wameq Raza, and Samina Yasmin. The authors bene‫ޔ‬ted from valuable insights and guidance provided by colleagues from the Bangladesh Institute of Development Studies (BIDS), General Economics Division (GED), Bangladesh Rural Advancement Committee (BRAC), International Food Policy Research Institute (IFPRI), United States Agency for International Development (USAID), Food and Agriculture Organization (FAO), and Palli Karma Sahayak Foundation (PKSF) who participated in discussions and bilateral meetings in Dhaka. The team is especially thankful to Imran Matin and Narayan Das from the BRAC Institute of Governance and Development for their valuable comments. Finally, the discussions held at the Bangladesh Development Forum 2020 were very helpful in informing the analysis for this study. The report has been prepared under the direction, guidance, and support of Carolina Sanchez-Paramo (Global Director, Poverty and Equity Global Practice), Mercy Tembon (Country Director, Bangladesh and Bhutan), Yutaka Yoshino (Program Leader), Benu Bidani (Practice Manager, Global Unit, Poverty and Equity Global Practice), Andrew Dabalen (Practice Manager, Poverty and Equity Practice, South Asia), and Loraine Ronchi (Practice Manager, Agriculture and Food Global Practice, South Asia). Funding for this work was provided by the Bill and Melinda Gates Foundation, and Alan Rennison provided intellectual guidance from the foundation as the diagnos- tic method and report were developed. Cover illustration and design by Reyes Work. V Abbreviations ASA Association for Social HIES Household Income and Advancement E3penditure Surveys HYV High-yielding variety BBS Bangladesh Bureau of ICT Information and communica- Statistics tions technology BDP2100 Bangladesh Delta Plan 2100 IFPRI International Food Policy BIDS Bangladesh Institute of Research Institute Development Studies kcal Kilocalorie BIHS Bangladesh Integrated kg Kilogram Household Survey km Kilometer BPDB Bangladesh Power kWh Kilowatt hour Development Board LFS Labor Force Survey BRAC Bangladesh Rural LTS Lal Teer Seed Advancement Committee MSME Micro, small, and medium BSA Bangladesh Seed Association enterprises CBN Cost of Basic Needs MT Metric tonne CPI Consumer Price Inde3 NGO Nongovernmental DAE Department of Agricultural organization E3tension PKSF Palli Karma Sahayak DAM Department of Agricultural Foundation Marketing QLFS Quarterly Labor Force Survey DLS Department of Livestock R&D Research and development Services RID Rural Income Diagnostic EBA Enabling the Business of RMG Ready-made garment Agriculture SME Small and medium FAO Food and Agriculture enterprises Organization TFP Total factor productivity FAOSTAT Food and Agriculture TTL Task Team Leader Organization Statistical US United States of America Database USAID United States Agency for FDI Foreign Direct Investment International Development FY Fiscal Year UTC Universal Time Coordinated GCI Global Connectivity Inde3 WDI World Development GDP Gross Domestic Product Indicators ha Hectare WSA Water spread areas VI Executive summary This Rural Income Diagnostic (RID) aims to answer the question: “What are the main opportunities and constraints to faster, sustained income growth for poor and vulnerable households in rural Bangladesh?” This analysis is moti- vated by recent evidence highlighting the centrality of rural areas for poverty reduction in Bangladesh and the need to update our understanding of rural income dynamics to better inform policy solutions.1 The objective of the analy- sis is to inform the World Bank Systematic Country Diagnostic and government planning. The analysis focuses on areas where progress can be made in the next ‫ޔ‬ve years, consistent with the country’s long-term development path. The focus on short-term priorities to accelerate rural income growth needs to be implemented in a manner that is consistent with, and does not distract from, long-run goals and investments that will have very high future returns, especially for the poor. These include investments in child nutrition, health, and education. The RID takes four steps to identify and prioritize opportunities and constraints for rural income growth: • Step 1: Evaluate context and key dimensions of heterogeneity. What do we know about rural incomes and poverty in Bangladesh? How do rural households currently allocate time and assets across activities to ma3imize income and reduce variability? • Step 2: Identify opportunities for income growth. What are the opportuni- ties for agricultural growth and rural non-farm growth, and how may migra- tion contribute? • Step 3: Prioritize constraints to achieving growth. What are the most important constraints preventing poor households from taking advantage of opportunities to raise their incomes? • Step 4: Identify policy solutions. What are feasible policy actions that would help poor households overcome these constraints and take advan- tage of opportunities for income growth? 1 Hill and Genoni (2019). 7 BANGLADESH RURAL INCOME DIAGNOSTICS Understanding context and income heterogeneity Bangladesh’s gains against poverty in rural areas have been remarkable. Although poverty remains high, and may have risen with COVID-19, Bangladesh has achieved strong reductions in poverty levels in the past decades. Bangladesh reduced its rural poverty from 52 percent in 2000 to 27 percent in 2016/17, accord- ing to the latest oȂicial statistics. About 80 percent of the countryҁs poor live in rural areas.2 Economic transformation has allowed many rural households to diversify income sources and increase income and consumption, but gains remain frag- ile. All sectors of Bangladeshҁs economy have contributed to poverty reduction in the past, re‫ޕ‬ecting a rapidly transforming and interconnected economy, in which many households derive labor income from multiple sources. For all rural areas, on average 39 percent of households derive income from diȂerent sectors (i.e., agriculture, industry, and services). Importantly, these patterns do not diȂer substantially between poor and richer households in rural areas. However, many households transition in and out of poverty. About 4 in 10 rural Bangladeshi house- holds belonged to the bottom income quintile at some point between 2000 and 2014. This highlights the importance for sustained rural development of focusing not only on the poor, but also on vulnerable non-poor households. Recurrent movement in and out of poverty partly re‫ޕ‬ects the low quality of jobs and weather-related risks, combined with limited ability to cope with income shocks. Informal self-employment and daily wage activities largely dominate rural employment. Outside agriculture, the poor are more engaged in small-scale trad- ing and lower-quality service jobs. A workerҁs type of employment (daily laborer, self-employed, or salaried) appears to better predict their household poverty than does their main economic sector of work. On the demand side, a large share of employment is based on informal and subsistence microenterprises.3 In addition, Bangladeshҁs location renders it especially vulnerable to the adverse impacts of weather and climate change. The vulnerability in rural incomes is accompanied by limited ability to cope with income shocks. Negative shocks e3perienced by rural households are largely managed with the householdҁs own resources. Due to lack of options, the poor rely more on reducing food consumption when facing a shock. 2 Poverty estimates based on the oȂicial upper poverty line. See Ahmed et al. (2019); Hill and Genoni (2019). 3 Farole and Cho (2017). 8 EXECUTIVE SUMMARY Even before the COVID-19 crisis, studies indicated that some of Bangladesh’s traditional growth drivers were losing steam. The contribution of agricultural growth to poverty reduction has diminished since 2010, garment job creation has stagnated, and poverty in the informal service sector in urban areas has increased. In addition, recent poverty reduction has been uneven across regions, and spatial disparities have re-emerged, with urban poverty stagnating and Rangpur and Rajshahi divisions recently seeing rural poverty rise.4 The COVID-19 crisis highlights challenges linked to Bangladesh’s current pat- tern of growth and poses signi‫ޔ‬cant risks for the ‫ޔ‬ght to eradicate poverty. Reliance on the garment sector for employment creation and growth has turned into a source of vulnerability for Bangladesh. Remittances were central to support- ing incomes but have now come under strain, with important localized eȂects. In addition, the large share of incomes derived from vulnerable and informal jobs means that many Bangladeshi households remain at risk of falling into poverty. Importantly, the economic impact of the COVID-19 crisis is not limited to major cities, as a large share of non-farm employment in Bangladeshҁs rural areas is also in the sectors hardest hit by the pandemic.5 Three features of rural incomes are key for identifying future opportunities: • First, rice dominates rural farm incomes, but the sector has been under- performing, limiting agriculture’s contribution to poverty reduction. Rice is by far the dominant crop for rural Bangladeshi households. However, growth in rice production has slowed signi‫ޔ‬cantly, a key factor in the over- all slowdown of agricultural growth. While many crops oȂer farmers higher average returns than rice, the majority of farmers continue to grow rice, in part because cereal production involves lower risks. Public policy support in the form of input subsidies, procurement support, and research and devel- opment for food security contribute to the lower risk of producing cereals. • Second, there is still a deep gender divide in the rural labor market, and women’s earning potential remains largely untapped. Employment and earning sources diȂer sharply between men and women in rural areas. Female labor force participation in rural areas has increased steadily, but remains much lower than menҁs. Womenҁs share in agricultural employment has grown over time, re‫ޕ‬ecting an increasing ҂feminization of agriculture.҃ However, about 3 in 10 women who report working are engaged in livestock 4 Hill and Genoni (2019); Farole and Cho (2017). 5 Genoni et al. (2020). 9 BANGLADESH RURAL INCOME DIAGNOSTICS and 11 percent in unpaid farm activities. In contrast, only 6 percent of men report deriving income from livestock, and only 2 percent are unpaid. • Third, location in‫ޕ‬uences rural income inequality, as income opportu- nities vary across areas. Greater engagement in agriculture in Western Bangladesh has tied poverty reduction there more closely to the perfor- mance of the agricultural sector.6 On the other hand, Western areas and poorer parts of the East have higher agricultural diversi‫ޔ‬cation potential, which could boost agricultural incomes in these regions in the future.7 To date, faster income gains among households engaged in garments and other manufacturing have increased spatial disparities, as these activities are concentrated in the Eastern divisions around the Dhaka-Chittagong corridor. Finally, migration incomes are important for poverty reduction, but the concentration of migrants in Dhaka and Chittagong adds to spatial diȂerences in rural incomes. Strategic household assets and services (education, land, seeds, irrigation, and access to electricity) and key markets (farm-product, land, credit, and insur- ance) oȂer levers for improving rural incomes. This report highlights substan- tial progress in household assets and access to services but also persistent gaps. For instance, half of the rural adult population is illiterate, especially women and people in Western areas. Land is fragmented into small farms with limited space for e3pansion, and land markets could function more eȂiciently to circumvent these challenges. Agricultural productivity faces both longstanding and emerging challenges related to irrigation, fertilizers, mechanization, and seeds. Two in ten rural households do not have access to electricity, and power cuts are frequent. Connectivity between the countryҁs periphery and Dhaka-Chittagong—the dom- inant economic center—is compromised by inadequate transport infrastructure and limited broadband connectivity. Substantial improvements are needed in the commercialization of agricultural products, access to credit, and insurance markets. Key opportunities for inclusive rural income growth Agriculture needs to become more productive, not only to boost farm earnings but also to spur rural non-farm incomes. Despite the countryҁs economic trans- formation, a large share of incomes in rural areas still depend on agriculture. The recent slowdown in agricultural growth is an important reason why poverty has 6 Hill and Genoni (2019). 7 Ahmed et al. (2021). 10 EXECUTIVE SUMMARY increased or stagnated in Western areas, so the performance of agriculture is cru- cial for reducing spatial welfare disparities. Boosting growth in agriculture is also important because of its multiplier eȂect on growth in rural non-farm incomes.8 The spatial distribution of ‫ޔ‬rms is highly correlated with non-farm incomes, indi- cating that agricultureҁs multiplier role can contribute to reducing income dis- parities across the country, especially in stimulating non-farm incomes in areas remote from large urban centers. While potential for diversi‫ޔ‬cation through arable land expansion is limited, diversi‫ޔ‬cation could be achieved through higher production intensity and sub- stitution of crops. Closing yield gaps in rice remains important. Analysis indicates that there is room to close remaining yield gaps in aman and aus rice. This can support freeing up suitable boro rice areas for a diverse range of high-value and more sustainable crops, while accelerating the development of the livestock, poultry, and ‫ޔ‬sheries subsectors, which have growth potential. The e3panding domestic market for high-value agricultural products provides an opportunity to increase incomes by diversifying agricultural output into higher value-added crops, livestock, and ‫ޔ‬shing. Past trends and projections indicate that domestic consumption of rice is declining, while that of other, typically higher-value com- modities is rising. This means there is room to e3pand output of other agricultural products, without compromising food security objectives. Supporting the growth of urban areas outside Dhaka and Chittagong can boost rural incomes and distribute the bene‫ޔ‬ts of internal migration more evenly. Bangladeshҁs urbanization and economic-activity development have been largely concentrated around the Dhaka and Chittagong corridor. Manufacturing has grown in peri-urban areas of Dhaka, but job growth in secondary cities and other urban areas has thus far been limited. Recent improvements in access to electric- ity and education have substantially narrowed gaps in education and infrastruc- ture between Dhaka and other urban areas. This could open fresh opportunities for job creation and higher productivity in those areas, if other investment gaps are addressed. As internal migration is driven by the pursuit of economic oppor- tunities, a more balanced urbanization that creates new economic openings in smaller cities can spread the bene‫ޔ‬ts of internal migration across the country. Given the current size of Dhaka, locally led ‫ޔ‬rm growth, and not the reallocation of ‫ޔ‬rms, is more likely to drive growth in secondary cities and other urban centers. 8 Shilpi and Emran (2016). 11 BANGLADESH RURAL INCOME DIAGNOSTICS The existing spatial distribution of ‫ޔ‬rms suggests that sectors that serve the internal market, in particular agro-processing, have more potential for short- term non-farm job creation in urban areas outside Dhaka and Chittagong. The current spatial distribution of agro-processing activities is more even across the country than is the case for other non-farm activities, highlighting the potential of agro-processing to create non-farm income opportunities in a more decentral- ized way.9 Secondary and other urban centers are well placed to host value-added processing of agricultural products and can be a critical link to boost agricultural productivity and rural incomes.10 Developing this sector should be seen as com- plementing broader eȂorts to e3pand quality job growth in urban areas along with other non-farm activities, including greater e3port orientation of non-RMG manufacturing activities. Finally, supporting greater labor force participation and higher income produc- tivity for women is another important avenue for rural income growth. Womenҁs participation can ease pressures in the agricultural labor market and support crop diversi‫ޔ‬cation, as the availability of agricultural labor is central for this process.11 The changing gender composition of the agricultural workforce suggests that even larger bene‫ޔ‬ts can be achieved by further reducing barriers to womenҁs par- ticipation. High unemployment among young women underscores the need to e3pand job opportunities for women in rural areas. Constraints to rural income growth and policy levers to liȅ them The RID identi‫ޔ‬es a series of constraints that, if addressed, could help Bangladesh take advantage of the rural income opportunities just described. The RID identi‫ޔ‬es and prioritizes constraints based on two main criteria: (i) strong evidence that current performance in the area is falling short; and (ii) strong evi- dence that tackling the constraint can substantially improve rural income growth. The evidence used comes from two sources: the descriptive analysis undertaken for this study and an e3tensive literature review. The report organizes the con- straints under ‫ޔ‬ve areas for policy intervention: (i) policy focus and regulations; (ii) investment gaps; (iii) market-related constraints; (iv) knowledge and capacity constraints; and, (v) gender norms. 9 Muzzini and Aparicio (2013) show that agro-processing and te3tiles contributed signi‫ޔ‬cantly to the growth of non-farm activities in rural areas between 2001 and 2009. Over that period, the share of agro-processing jobs located in rural areas increased from 42 to 55 percent, and the share of te3tiles jobs rose from 38 to 50 percent. 10 Muzzini and Aparicio (2013); Farole and Cho (2017). 11 Hoque and Ahmed (2020). 12 EXECUTIVE SUMMARY Following the identi‫ޔ‬cation of key constraints, the analysis identi‫ޔ‬es policy levers that can help liȅ constraints and trigger opportunities. The RID recom- mends policy solutions based on two features: (i) they are supported by a strong evidence base, and (ii) they are feasible within ‫ޔ‬ve years. In the current state of evidence, some policy areas are clearer and more speci‫ޔ‬c than others. On some fronts, such as tackling gender norms, continued e3perimentation can be import- ant to tailor approaches to the country conte3t and for scalability.12 This report identi‫ޔ‬es three main constraints related to policy focus and regu- lations. First, as income growth and urbanization continue, agricultural policy will need to support a more diversi‫ޔ‬ed agriculture. The report highlights the importance of levelling the playing ‫ޔ‬eld across products by addressing bias against some high-value non-rice production in agricultural support. The evi- dence indicates that, with changing demand patterns and higher incomes, this can be done without compromising the countryҁs food-security goals. Second, streamlining current regulations can create a more robust seed sector, with even higher private-sector participation and higher-quality seeds. Third, improving the coordination of transport policies and road safety enforcement can contrib- ute to reducing transaction costs associated with congestion and the quality of the transport network. Policy levers related to policy focus and regulations include: Rebalance agricultural policies. Some actions comprise: (i) Review and reform input subsidies, especially on fertilizers, and leverage e3tension services to make fertilizer use more eȂicient; (ii) Rebalance the agricultural research and develop- ment (R&D) budget toward non-rice products; (iii) Revisit regulations preventing private-sector actors from breeding and producing seeds for noti‫ޔ‬ed crops;13 (iv) Increase public-private coordination on quality seed supply for noti‫ޔ‬ed crops; (v) Increase market transparency on the quality and eȂectiveness of seeds. 12 Evidence of e3isting successful approaches in Bangladesh was collected from evaluation documents, recent high-quality assessments, and through stakeholder consultations. Results were cross-refer- enced with e3tensive literature reviews on the eȂectiveness of e3isting policies and potential alter- natives. The analysis highlights areas where evidence suggests that the policies now being used to improve rural incomes in Bangladesh can be enhanced to achieve greater impact. 13 Noti‫ޔ‬ed crops are rice, wheat, potato, jute, and sugarcane, which are considered important in terms of food and national security. Seed segments for these crops are highly regulated. 13 BANGLADESH RURAL INCOME DIAGNOSTICS Review policy actions and regulations that can lower transport and logistics costs.14 Policy areas that can be tackled in the ne3t ‫ޔ‬ve years include: (i) Assess roles to improve coordination across public transport ministries and agencies; (ii) Review outdated policies and regulations to move from development of infrastruc- ture to also focus on the quality of services; (iii) Fully implement the Competition Act, which aims to prevent collusion in the logistic services market and price ‫ޔ‬3ing by industry players; (iv) Strengthen enforcement of e3isting regulations to prevent road accidents and overloading of trucks. The RID analysis also prioritizes water management, electricity, and trans- port-related investments. Given the e3tent of ‫ޕ‬ooding in the country and the need for reliable water supply in oȂ-monsoon seasons, better irrigation and drain- age remain crucial for productivity and diversi‫ޔ‬cation in agriculture. Priorities to close investment gaps include: (i) Invest in high-eȂiciency irrigation coverage and ‫ޕ‬ood protection, particularly in hilly and highland divisions; (ii) Invest in reliable drainage, particularly in low-lying areas where production of non-paddy crops is especially diȂicult; (iii) Continue promoting the shiȅ away from boro paddy, as this crop is rapidly depleting groundwater storage. Renewed eȂorts in terms of energy and transport remain central for both farm and non-farm income growth. Actions to close electricity gaps in terms of coverage and reliability of the supply should continue. Closing transport infrastructure gaps is also central, with a stron- ger emphasis on infrastructure maintenance. It is important to make the land market work better. With average land size very small and limited room for e3pansion, land markets need to be more eȂicient to help create economies of scale for investment. Land rental markets help landless farm households to access cultivable land. However, typical land rental tenure is limited to three years and largely informal. A speci‫ޔ‬c policy area related to this point is reviewing the agricultural land rental market to identify speci‫ޔ‬c market failures that hinder longer-term rental of agricultural land. The report identi‫ޔ‬es knowledge and capacity constraints related to low skills and weak extension services. Despite recent gains, the human-capital base of Bangladeshҁs rural adult population remains low, constraining opportunities to improve agricultural productivity and access productive non-farm opportunities. Fully overcoming this key barrier will be a slow process, however, skills-develop- ment eȂorts should be pursued. Policy makers can also bolster e3tension systems to deliver more timely and relevant information to farmers. 14 These recommendations build on the e3tensive analysis undertaken by Herrera Dappe et al. (2019). 14 EXECUTIVE SUMMARY Policy areas that can improve extension services include: (i) Revise the budget and service-delivery capacities of the Department of Agricultural E3tension (DAE); (ii) Provide training to DAE e3tension staȂ on new agricultural knowledge and techniques for delivery; (iii) Increase support for private-sector-led training pro- grams (e.g., PRAN); (iv) Support closer integration between DAE and the private sector, which could enhance impact despite limited resources.15 The report highlights gender norms that constrain women’s ability to engage in more productive, paid farm and non-farm activities. The prevalence of early marriage is still high in Bangladesh compared to other South Asian countries, increasing the likelihood that younger Bangladeshi women will e3it school or the labor force due to household responsibilities. Mobility constraints also limit wom- enҁs ability to engage in non-farm activities and commercialization of farm prod- ucts. Restrictions on womenҁs ability to inherit property or obtain credit constrain their labor market participation.16 Policy areas to tackle gender norms: There is some recent evidence of progress in lowering barriers to income growth among rural women, particularly for younger cohorts. However, resolving constraints will require a comprehensive approach that tackles multiple barriers at once. In the short run, policy priorities include the following: (i) Invest in boosting productivity in womenҁs livestock work; (ii) Engage men and communities to promote gender equality; (iii) Institute legal reform for womenҁs inheritance of land, and address social norms that oppose womenҁs land ownership and registration; and (iv) Target interventions to increase womenҁs control of ‫ޔ‬nancial resources. World Bank (2020a). 15 16 Cameron, Dowling, and Worswick (2001); Gonzales et al. (2015); Kabeer (1990); Amin (1997); World Bank (2012); Klugman et al. (2014); Srivastava and Srivastava (2010); NeȂ, Sen, and Kling (2012), SolotaroȂ et al. (2019). 15 Introduction and objectives of this review In the past two decades, Bangladesh has achieved remarkable welfare gains for its people, both in monetary and non-monetary terms, registering rates of poverty reduction that rank among the most impressive in the world. Since 2010, however, the country’s progress in poverty reduction has slowed, and spatial welfare disparities have widened. Bangladeshҁs earlier welfare progress was accompanied by robust economic growth, a sweeping structural transfor- mation of the countryҁs economy, and rapid urbanization. Aȅer 2010, however, while the countryҁs economic growth accelerated, poverty reduction slowed. Even before the COVID-19 crisis, studies indicated that some of Bangladeshҁs tradi- tional growth drivers were starting to lose steam. The contribution of agricultural growth to poverty reduction has diminished, garment job creation has stagnated, and poverty in the informal service sector in urban areas has increased. In addi- tion, recent poverty reduction has been uneven across regions, and spatial dispar- ities have re-emerged, with urban poverty stagnating and Rangpur and Rajshahi divisions seeing rising rural poverty.17 The COVID-19 crisis highlights the challenges linked to Bangladesh’s current growth pattern and poses signi‫ޔ‬cant risks in the ‫ޔ‬ght to eradicate poverty. Reliance on the garment sector for non-farm employment creation and growth is now a source of vulnerability for Bangladesh. Remittances, central to supporting incomes in the past two decades, are now under strain, with important localized eȂects. In addition, the large share of incomes derived from informal and more vul- nerable jobs puts many Bangladeshi households at risk of falling into poverty. The economic impact of the COVID-19 crisis is not limited to major cities, as a large share of non-farm employment in impacted sectors is in Bangladeshҁs rural areas.18 Today, renewed eȂorts will be needed to continue Bangladesh’s remarkable achievements in poverty reduction and advance the country’s ambitious development goals. This report marshals evidence and analysis for strategies to 17 Hill and Genoni (2019); Farole and Cho (2017). 18 Genoni et al. (2020). 17 BANGLADESH RURAL INCOME DIAGNOSTICS increase income growth in rural areas of Bangladesh, with a particular focus on the poor and those vulnerable to becoming poor. The recently completed World Bank Poverty Assessment19 highlights the centrality of rural areas in the poverty reduction story in Bangladesh before the pandemic. By the latest oȂicial ‫ޔ‬gures, from 2016/17, 1 in 4 Bangladeshi people lived in poverty, and about 8 in 10 of the poor were in rural areas. Shedding light on ways to reduce poverty in the rural space is not straightforward, however. It requires a comprehensive understand- ing of the income transformation process underway in rural Bangladesh, the links between rural incomes and changing conditions in urban areas, and how speci‫ޔ‬c enabling factors can support rural income gains. This Rural Income Diagnostic (RID) aims to answer the question: “What are the main opportunities and constraints to faster, sustained income growth for poor and vulnerable households in rural areas of Bangladesh?” The objective of the RID is to better understand how people who currently live in rural areas could have higher incomes, but it does not delve into the broader challenge of how to raise rural GDP. The focus of the analysis is on identifying areas where progress can be made in the ne3t ‫ޔ‬ve years, while remaining consistent with the countryҁs long-term devel- opment path. This means that short-term eȂorts to accelerate rural income growth should not distract from long-run priorities and investments that have very high long-run returns. This applies especially to investments that will disproportionately raise future welfare in poor communities, notably investments in child nutrition, health, and education.20 This report follows a common framework used across diȂerent countries under- taking rural income diagnostics.21 The framework provides a structure and set of questions that draw on the literature on structural transformation, agricultural growth, and rural non-farm growth.22 Building on the organizing and theoretical frameworks set out in Timmer (2008), Foster and Rosenzweig (2004), and McCulloch et al. (2007), the RID identi‫ޔ‬es three broad sources of labor income growth for rural households: growth in agricultural incomes, growth in non-farm incomes, and growth generated through rural to urban migration. Within these categories, it considers more speci‫ޔ‬c sources of income growth that are presented in Figure 0.1. 19 Hill and Genoni (2019). 20 Hill (2018). 21 Hill (2018). 22 Including Christiaensen and Martin (2018); Barrett et al (2017); Dercon and Gollin (2014); Haggblade, Hazell and Reardon (2007); Timmer and Akkus (2008); Lanjouw and Feder (2001); and some of the references therein. 18 Labor income growth forINTRODUCTION AND OBJECTIVES OF THIS REVIEW rural households Agricultural growth Non-farm growth Migration Hâ°ÝÝÐçÝÀÄô÷ç»ü°ÑãÑã¼ôÄ°÷ÄÀñôçķü°»ÑÝÑüē Growth in non-tradable Temporary migration of from existing goods, through favorable goods and services (wage or family members to urban changes in relative prices, more stable prices, self-employment) areas or internationally yield growth, reduced yield volatilty, area expansion, more cost-effective production Growth in tradable goods Permanent migration of through mining, tourism, or individuals or families Smallholders achieve greater market partici- rural manufacturing (wage or pation with existing goods, through commer- self-employment Commuting cializing a larger share of production, selling further up the value chain, engaging in contract farming or outgrower schemes Smallholders transition into higher-value outputs from both crop and non-crop agricul- ture, including livestock The diagnostic takes four steps to identify opportunities and prioritize constraints within Growth in large-scale commercial farming increases demand for agricultural wage labor and across these types of income growth: 1 2 3 4 Assess context and Analyze opportunities for Prioritize constraints to Identify feasible policy heterogeneity: How do income growth: What are growth: What are the solutions: What are households currently the opportunities for most important feasible policy actions allocate time and assets agricultural growth and constraints preventing that would help poor across activities to rural non-farm growth? poor households from households overcome the maximize income and What opportunties could taking advantage of constraints and take reduce variability? How is emerge from migration? growth opportunities? advantage of the oppor- this different for poor tunities for income households? How does growth? this vary across the rural 19 space? Source: Hill (2018). BANGLADESH RURAL INCOME DIAGNOSTICS Households are oȅen engaged in more than one type of activity in order to ma3imize incomes and reduce the volatility of incomes to e3ternal shocks. Growth of income from one source has impacts on the opportunities and constraints for growth from other sources. This report applies the framework to identify opportunities, challenges, and policy options for boosting income growth among poor households in rural Bangladesh. The diagnostic incorporates results from a series of recent techni- cal assessments, complementing these studies with descriptive analysis and benchmarking. Based on the available evidence, it proposes a set of policy rec- ommendations to continue reducing rural poverty. The reportҁs main contribu- tions include systematic comparisons between poor and non-poor households and across the countryҁs divisions and districts. This approach is motivated by recent policy discussions in Bangladesh that have highlighted the need to better diȂerentiate opportunities and options geographically and across income levels. The report’s ‫ޔ‬rst section provides a descriptive analysis of the main sources of rural labor income and shows how assets and markets shape the capac- ity of households to earn income. Using the framework, the analysis distin- guishes three broad types of earned labor income that have been the main income sources for rural households and key drivers of rural poverty reduction: income from agriculture, income from non-agricultural activities, and income from migration, i.e. remittances. The relative importance of these types of rural income varies across regions, in‫ޕ‬uenced by agroecology, connectivity, human capital endowments, and other factors. Accordingly, the analysis emphasizes comparisons across divisions in Bangladesh. It also e3plores patterns across consumption quintiles to identify policy measures most likely to deliver strong bene‫ޔ‬ts for the poor or near poor. The second section identi‫ޔ‬es key opportunities and constraints for sustainable agricultural, rural non-farm, and migration income growth for rural households, especially the poor. Some constraints are speci‫ޔ‬c to particular types of income growth, while others—such as the presence of macro-stability and the quality of institutions—broadly aȂect growth of all types. While these broad factors are crit- ical, the analysis of macro-constraints and institutional and governance aspects is beyond the scope of this report. The third section prioritizes the main barriers to reducing rural poverty in Bangladesh. These include market and policy failures that constrain opportu- nities for rural income growth. Constraints are prioritized in the analysis when 20 INTRODUCTION AND OBJECTIVES OF THIS REVIEW strong evidence shows that: (i) they limit rural income growth; (ii) they impact a large number of households, particularly poorer households; and (iii) they con- strain multiple types of income growth important for reducing rural poverty. A constraint is also given more weight in the analysis when it operates as a bottle- neck: that is, when progress in addressing several other barriers can be acceler- ated if this constraint is tackled ‫ޔ‬rst. The report’s ‫ޔ‬nal section identi‫ޔ‬es policy levers that can help liȅ some of the prioritized constraints. Evidence of e3isting successful approaches in Bangladesh was collected from evaluation documents and through stakeholder consulta- tions. Results were cross-referenced with literature reviews on the eȂectiveness of e3isting policies and potential alternatives. The analysis highlights areas where evidence suggests that the policies now being used to improve rural incomes in Bangladesh can be enhanced to achieve greater impact. It is important to note that the rural income diagnostic focuses on income that is earned by a rural household. Therefore, shortfalls in human capital will be identi‫ޔ‬ed where relevant, but the diagnostic does not speci‫ޔ‬cally analyze con- straints aȂecting fertility reduction or health, nutrition, and education outcomes. Similarly, macroeconomic factors and governance challenges are not e3amined in detail unless they aȂect speci‫ޔ‬c pro3imal incentives for rural income growth (e.g., fertilizer subsidies). To set the stage for the core analysis, the following pages describe data sources used for the RID, specify levels of analysis, and clarify the de‫ޔ‬nitions of select key terms. 21 Data sources and levels of analysis Main data sources The descriptive analysis presented in this report relies primarily on three household surveys: the Bangladesh Integrated Household Survey (BIHS), the Household Income and Expenditure Survey (HIES), and the 62-village panel surveys (62-village panel). These data combined make it possible to compare patterns across areas and across household per capita consumption quintiles, a feature central for the analysis. The BIHS is a rural representative household panel survey that collects detailed data on household-level demographics and assets, income and con- sumption, and plot-level agricultural production and practices.23 This survey is representative of rural areas of each of Bangladeshҁs seven administrative divisions. The ‫ޔ‬rst round of the BIHS was conducted from November 2011 to March 2012 and is used as a reference point to measure progress through repeat surveys. The second BIHS round was conducted from January to June 2015. A third BIHS round was completed in 2018. The main analyses presented here rely on the 2018 round. We complement the analysis with the HIES, which is a comprehensive, nation- ally representative survey used to measure monetary poverty in Bangladesh. The HIES 2016/17 is the fourth round in the HIES cross-sectional series conducted by the Bangladesh Bureau of Statistics (BBS). Previous rounds were conducted in 2000, 2005, and 2010. For the years 2000, 2005, and 2010, the survey was representative at the division level and by rural and urban areas. For 2016/17, the HIES is also repre- sentative of the countryҁs 64 districts.24 The longer time period covered by the HIES allows us to better portray changes since the 2000s. It is important to note that the 2016/17 survey has some limitations in terms of the quality of the income informa- tion collected,25 and also lacks important details regarding agricultural practices and incomes, as they are not part of the core objectives of the survey. 23 Ahmed (2016). 24 Ahmed et al. (2017). 25 Hill and Endara (2019b). 23 BANGLADESH RURAL INCOME DIAGNOSTICS The 62-village panel allows for a deeper analysis of dynamics and drivers of rural income growth and income poverty across time. The panel surveys were undertaken to assess changes in rural poverty and livelihoods and identify the role of diȂerent factors in driving these changes, including technological prog- ress, prices, and others. The panel data is a nationally representative sample of rural households in Bangladesh. The baseline of the survey was carried out in 1988, when 1,260 households from 62 villages in 57 districts of the country were visited. These households were revisited in 2000, 2004, 2008, and 2014. Besides revisiting the original households and their oȂshoots, researchers have adjusted the sample size in repeat surveys in each round to make the sample representa- tive of the rural population for the survey year. The analysis in this diagnostic used the surveys in 2000, 2008, and 2014. Each of these three surveys brings its own advantages and limitations to the analysis. Whenever possible, the analysis was repeated for all three surveys, and results are presented when ‫ޔ‬ndings seem to be consistent. The descrip- tive analysis in the ne3t section will present ‫ޔ‬gures using the best data for the particular purpose at hand. Speci‫ޔ‬cally, we use the BIHS data to describe householdsҁ income sources, as this is a survey speci‫ޔ‬cally designed to capture rural incomes. We use the HIES surveys to describe recent poverty trends across Bangladesh, migration patterns at the subnational level, analysis across time on urbanization and poverty, and for the analysis of household assets, such as edu- cation and electricity. We use the 62-village panel to better understand changes in income sources and their links to poverty and vulnerability by looking at the same households across time. Note that the BIHS data could have been used for the same purpose, however, the ‫ޔ‬nal round became available aȅer most of the analyses for the RID were completed. Therefore, the dynamic analysis using BIHS was not included. The report supplements ‫ޔ‬ndings from the three main surveys with infor- mation from other surveys conducted by the BBS, including the Economic Census 2013, the Agriculture Censuses, the 2018 Agriculture and Rural Survey, and Labor Force Surveys (LFS). The 2018 Agriculture and Rural Survey was used to provide more updated district-level statistics that cannot be obtained using HIES. LFS was used to analyze unemployment and wage gaps by gender. These sources are complemented with price data from the Bangladesh Department of Agriculture Marketing (DAM) and statistics from the World Development Indicators (WDI), FAOSTAT, and other sources for benchmarking. 24 DATA SOURCES AND LEVELS OF ANALYSIS Levels of analysis The descriptive work undertaken here explores whether there are substan- tial distinct patterns across two dimensions: household per capita consump- tion and geography. The Bangladesh literature in general provides limited comparisons across the consumption distribution (i.e., between poor and non-poor households) or across regions in Bangladesh. The recent World Bank Bangladesh Poverty Assessment noted a reemergence of the countryҁs East- West economic divide, with poverty in Western divisions stagnating or increas- ing, even as it has continued to fall in Eastern areas.26 This highlights the need to better understand regional dynamics, as constraints and enablers of growth may diȂer geographically. To explore diȂerences across households with diȂerent levels of resources, we construct per capita household consumption quintiles. In HIES, the BBS oȂicial consumption aggregate is used to obtain these quintiles, de‫ޔ‬ned only for house- holds in rural areas. A similar approach is adopted for the BIHS. We opt not to compare just poor and non-poor households but also compare across the con- sumption distribution to provide a more nuanced pro‫ޔ‬le. When comparing poor and non-poor households, we use the upper poverty line determined by the BBS.27 For the 62-village panel analysis, we rely on household income as the wel- fare aggregate. Appendi3 1 includes summary statistics presenting the variation in consumption and income across rural consumption quintiles to help the reader understand diȂerences across these groups. To shed light on geographic diȂerences, the analysis incorporates divi- sion-level comparisons. In Bangladesh, divisions are the ‫ޔ‬rst-level adminis- trative geographical partitions and are redrawn periodically. As of 2016, the country has eight divisions: Barisal, Chittagong, Dhaka, Khulna, Mymensingh, Rajshahi, Rangpur, and Sylhet. For this analysis, seven divisions are adopted (treating Mymensingh and Dhaka as one), to be comparable across time. 26 Hill and Genoni (2019). 27 Total household consumption is calculated as the sum of all food and non-food e3penditures e3cept for ta3es and fees, wedding e3penses, and interest and insurance e3penses. Rent is included and imputed when missing. A household is considered poor if its per capita consumption is lower than the poverty line for the survey stratum in which it lives. Poverty lines vary across geographic strata to account for spatial diȂerences in the cost of purchasing basic needs. The national upper poverty line is set at the cost of consuming 2,122 Kcal per person per day and an allowance for non-food e3penditure. Poverty lines are updated between survey rounds using a food price inde3 that captures food in‫ޕ‬ation in each stratum and the non-food consumer price inde3 (CPI). See Ahmed et al. (2019). 25 BANGLADESH RURAL INCOME DIAGNOSTICS Analysis below the division level is limited by the representativity of the HIES and BIHS surveys. However, when possible, district-level comparisons are applied. In that case, seven divisions are subsequently divided into 64 districts, or zilas. The division comparison aligns relatively well with diȂerences in agro- ecology, population density, and connectivity to main urban centers (Bo3 1). Appendi3 2 includes summary statistics presenting the variation in consump- tion and income across the rural areas of the divisions. Rangpur division is located to the west of the Jamuna river and has low access to the large markets in Dhaka, as well as high rain intensity throughout the year. Rajshahi division is also west of the Jamuna river but has relatively easier access to large urban markets due to its pro3imity to the Jamuna bridge. This division has both rainy and dry seasons and is highly irrigated. Khulna and Barisal divi- sions, also west of the Jamuna, are less connected to Dhaka than other west- ern areas. These two divisions have rainy and dry seasons (largely vulnerable to cyclones). Chittagong division is located in the southeast and is the second most urbanized region. Access to markets across Chittagong also depends on pro3imity to Dhaka. Those areas of Chittagong closer to the capital have better road networks. Chittagong division has some mountainous and forest areas. It also includes both areas with high rain intensity throughout the year and others with distinct dry and rainy seasons. Home to the capital, Dhaka is Bangladeshҁs most urbanized division and largest market. The southern portion of Dhaka still lacks good connectivity to the north, however. Dhaka has rainy and dry seasons. Sylhet division is located to the northeast and e3periences high rain intensity throughout the year. Box 1. Profile of Bangladesh’s divisions Temperature is stable around Bangladesh. Almost all of the country is at sea level or below (e3cept for the south of Chittagong). Virtually the whole country is ‫ޔ‬t to plant rice or similar crops (again e3cept the south of Chittagong). The Jamuna river has historically been identi‫ޔ‬ed as a natural boundary for the dif- ferent regional patterns of economic growth seen in Bangladesh (Figure B 1). 26 DATA SOURCES AND LEVELS OF ANALYSIS Figure B 1. Spatial variation of key features in Bangladesh a. Divisions b. Population Rangpur • Sylhet • RAJSHAHI SYLHET Rajshahi • DHAKA Dhaka • KHULNA Khulna Barisal • • CHITTAGONG BARISAL Chittagong • km km 0 50 100 200 0 50 100 200 Population Density, 2015 People per square km High: 90,000 Low: 22 c. Farming systems d. Land cover km km 0 50 100 200 0 50 100 200 Rice Developed Water Coastal artisanal fishing Rainfed Cropland and Pasture Wetland Rice-wheat Irrigated Cropland Barren Highland mixed Grassland/Shrubland Tundra Savanna Other Forest Source: Maps b-d are sourced from the United Nations Food and Agriculture Organization (FAO). (ŚƩƉ͗ͬͬǁǁǁ͘ĨĂŽ͘ŽƌŐͬĐŽƵŶƚƌLJƉƌŽĮůĞƐͬŵĂƉƐͬĞŶ͍ͬŝƐŽ3с') 27 BANGLADESH RURAL INCOME DIAGNOSTICS Key definitions This report uses Bangladesh’s oȂicial classi‫ޔ‬cation of urban areas. According to this de‫ޔ‬nition, urban areas are those with more than 40,000 individuals within a mouza.28 It is important to note that, due to this threshold and the fact that many urban areas are very dense, many households oȂicially classi‫ޔ‬ed as rural in Bangladesh could be considered urban in other conte3ts. When using agglom- eration indices rather than oȂicial de‫ޔ‬nitions of urban areas, the proportion of people living in urban areas in Bangladesh is as much as 20-36 percentage points higher than oȂicial estimates.29 Figure 0.2 presents the share or households living in rural areas across districts according to the HIES 2016/17. 30 Figure 0.2. Rural population share, by district, 2016 0.74 – 0.91 0.57 – 0.74 0.41 – 0.57 Bay of Bengal 0.24 – 0.41 0.07 – 0.24 Source: Authors’ calculations using HIES 2016/17. 28 Mouza, or mauza, is the lowest administrative unit with a separate jurisdiction and a well-demar- cated cadastral map. 29 Uchida et al. (2010) suggest an urban population share that is 20 percentage points higher, while Robert et al. (2017) show that agglomeration indices predict an urban population share 30-36 percent- age points higher than oȂicial estimates. Other methods would suggest an even larger divergence. 30 According to the WDI, the share of rural population in Bangladesh is similar to other comparator countries. While the country has 62 percent of its population living in rural areas, Vietnam, Pakistan, India, and the average for the South Asia region have 63, 63, 65 and 65 percent, respectively. 28 DATA SOURCES AND LEVELS OF ANALYSIS The HIES analysis applies some corrections to the classi‫ޔ‬cation of rural and urban areas to ensure comparability across time and re‫ޕ‬ect observed urbanization rates (Table 0.1).31 The analysis using BIHS and the 62-village panel follows the same households across time, thus the information is representative of the rural popula- tion at baseline. Household-level attrition is relatively small in both panel surveys. Table 0.1. Urban population share in Bangladesh, 2000-2016 Urban share of population from census data (year) / projections Urban share (HIES) 2000 23.8 (2001) 20.1 2005 24.7 2010 28.0 (2011) 26.3 2016 29.1 Source: BBS Population and Household Census 2001 and 2011 and HIES 2000, 2005, 2010, and 2016. For calculation details concerning the urban share in HIES 2016, see Hill and Genoni (2019). De‫ޔ‬nition of agriculture. Agriculture is de‫ޔ‬ned as the sector consisting of crops, livestock, agroforestry, and ‫ޔ‬sheries. In this report, the terms ҂agriculture҃ and ҂farm҃ are used interchangeably. High-value agriculture refers to ‫ޔ‬sh, livestock products, fruits, vegetables, and spices that have a higher market value than tradi- tional food grains. Non-farm activities include all rural economic activities outside farm agriculture, such as self-employment, wage employment (full-time, formal, informal, and seasonal), and non-farm production. 31 Ahmed et al. (2019). 29 BANGLADESH RURAL INCOME DIAGNOSTICS 30 SECTION 1 Assessing context and heterogeneity What do we know about rural incomes and poverty in Bangladesh? This section synthesizes evidence to inform the analysis and policy recommen- dations presented later in the report. It characterizes the current state of rural incomes in Bangladesh, shows trends, and identi‫ޔ‬es key dimensions of rural income heterogeneity, for e3ample across geographical regions and gender. The section also seeks to understand migration and labor patterns in rural Bangladesh. It draws on this evidence to address a series of questions that are important for policy decisions, in particular: Who and where are Bangladeshҁs rural poor, and how has poverty in rural areas been changing? How do rural Bangladeshis earn their incomes, and what risks do they face? What are the main factors shaping het- erogeneity in rural incomes? How do key household assets and access to markets translate into rural income gains? The section starts by summarizing recent trends in rural poverty and identifying the factors behind the poverty performance seen, particularly over the period 2010-2016. We emphasize the 2010-2016 period because the oȂicial poverty sta- tistics for those years are the latest available. Some of the descriptive analysis reaches the year 2018. Macroeconomic conditions and sectoral growth trends suggest that the situation did not change substantially enough between 2016 and the advent of the COVID-19 crisis to aȂect the relevance of the ‫ޔ‬ndings. Lack of representative post-COVID-19 data in rural areas limits our ability to analyze poverty conditions following the crisis. Nonetheless, the features highlighted in this section are important in order to recognize vulnerabilities that e3isted before COVID-19 and so understand the pandemicҁs implications for the rural poor. 31 BANGLADESH RURAL INCOME DIAGNOSTICS Strong rural poverty reduction but with increased spatial disparities In the past two decades Bangladesh has achieved substantial reductions in pov- erty levels. Between 2000 and 2016, Bangladesh reduced its poverty rate from 49 percent to 24.5 percent. Rural poverty fell from 52.3 percent in 2000 to 26.7 percent in 2016. Over the same period, urban poverty declined from 35 percent to 19.3 percent (Figure 1.1). This represents about 25 million Bangladeshi e3iting poverty between 2000 and 2016. Despite improvements, poverty remains high, and the pace of poverty reduction has recently slowed in urban areas. The latest oȂicial statistics estimate that, as of 2016-17, about 1 in 4 Bangladeshi lived in poverty. About 80 percent of the countryҁs poor lived in rural areas.32 The rural poor numbered around 31.7 million, out of the total of 39.6 million poor people in Bangladesh.33 Recent trends show that poverty reduction remained strong in rural areas, while poverty reduction in urban areas slowed down. Rural poverty fell from 35.2 percent in 2010 to 26.7 percent in 2016. This contrasts with trends in urban areas, where poverty rates declined by only 2 percentage points, with no reduction in urban e3treme poverty during this period.34 Despite reductions in poverty, there was an increase in inequality in rural areas over the 2010-2016 period. The Gini coeȂicient increased from 0.27 to 0.29. This increase in inequality came because the bottom 10 percent of households in rural areas did not fare well, while richer rural households e3perienced higher consumption growth. Poverty reduction was largely e3plained by growth and not redistribution of consumption.35 The increase in inequality was accompanied by an uneven poverty reduction across geographical areas, with poverty recently increasing in the Western divi- sions. Between 2010 and 2016, poverty declined in the Eastern divisions, stag- nated in Khulna and Rajshahi, and increased in Rangpur. The stronger rate of pov- erty reduction in the Eastern regions widened a welfare gap between Eastern and Western Bangladesh that had previously been narrowed between 2005 and 2010.36 32 According to the 2016 HIES, about 71 percent of the population was rural. 33 Poverty estimates based on the oȂicial upper poverty line. See Ahmed et al. (2019) and Hill and Genoni (2019). Note that poverty estimates are calculated using diȂerent poverty lines across the country to account for diȂerences in the cost of purchasing a similar food basket. 34 Hill and Genoni (2019). 35 Hill and Genoni (2019). 36 Hill and Genoni (2019). 32 ASSESSING CONTEXT AND HETEROGENEITY The re-emerging divergence between the East and West occurred largely in rural, instead of urban areas. Rural poverty increased by about 3.7 percentage points in Rangpur and about 1.6 percentage points in Rajshahi between 2010 and 2016 (Figure 1.2). Since 2010, district-level spatial disparities have also increased. The poverty rate rose by more than 10 percentage points in 11 districts of the country and decreased by more than 10 points in 25 districts (Figure 1.3). Districts with poverty increases or stagnation are largely concentrated in the Western part of the country and in Chittagong, while the remaining districts in the center, north, and east of the country performed better. Figure 1.1. Poverty rates across time, 2000-2016 ǽ 1&,+)-,3"/16 ǽ/++!/2/)-,3"/16/1"0 60% 60% 52,3% 50% 50% 48,9% 43,8% 40% 40% 40% 35,1% 35,2% 34,3% 31,5% 28,4% 26,7% 30% 30% 25,1% 24,5% 21,3% 19,3% 20% 20% 17,6% 13% 10% 10% 0% 0% 2000 2005 2010 2016 2000 2005 2010 2016 Poverty Rural 51/"*"-,3"/16 Urban Source: Authors’ calculations using HIES 2000-2016. Notes:,3"/16%"! ,2+1/1"0/" ) 2)1"!0"!,+1%",ƛ& &)2--"/-,3"/16)&+"20&+$1%" Cost of Basic Needs (CBN). 33 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.2. Rural poverty rates by division, 2000-2016 70% 60% 50% 40% 30% 20% 10% 0% 2000 2005 2010 2016 Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Source: Authors’ calculations using HIES 2000-2016. Figure 1.3. District poverty rates and recent changes a. Poverty rates in 2016 b. Change in poverty 2010-2016 60-72% Bay of Bengal Bay of Bengal 45-60% 30-45% Increased 15-30% No change 0-15% Decrease Source: Authors’ calculations using HIES 2016 and 2010 poverty maps. Notes: In panel b, changes in poverty are indicated in orange (increased) or brown (decreased), if they are statistically significant. Districts without statistically significant change are displayed in beige. 34 ASSESSING CONTEXT AND HETEROGENEITY Rural poverty is spatially concentrated, with 11 districts out of the total 64 account- ing for some 40 percent of the rural poor. Rural poverty is relatively higher in the Western half of the country, with the e3ception of three hilly districts (Bandarban, Rangamati, and Khagrachhari) in the Eastern division of Chittagong (Figure 1.4a). The district of Dinajpur in Rangpur is home to the largest share of the rural poor found in any single district (6 percent), followed by Kurigram, also in Rangpur, and Kishoreganj in Dhaka (with 5 percent each). Rangpur and Gaibandha in Rangpur division account for another 8 percent of the countryҁs rural poor (Figure 1. 4b). A large share of Bangladeshҁs rural poor is located in northern parts of the country. Figure 1.4. Poverty rates and distribution of the poor across districts, 2016 a. Rural poverty rate by district b. Percentage of rural poor across districts 60-74% Bay of Bengal Bay of Bengal 45-60% 30-45% 4-6% 15-30% 2-4% 0-15% 0-2% Source: Authors’ calculations using HIES 2016. Note: Panel (a) displays the district upper poverty rates. Panel (b) displays the distribution of the poor population across districts. The role of different economic sectors in poverty reduction and East-West disparities A decomposition analysis shows that improvements in education, reductions in household size, and expansions in access to electricity and other assets were sizeable enough to explain about 70 percent of all rural consumption growth between 2010 and 2016. This was a larger contribution from these sources than 35 BANGLADESH RURAL INCOME DIAGNOSTICS in the period 2005-2010, when they e3plained about 53 percent of consumption growth. The increased importance of these factors in 2010-16 was largely linked to the faster e3pansion of electricity coverage in some areas of the country, as later sections of the report will show (Figure 1.5). However, reductions in family size and gains in adults’ education were slower in the Western divisions, partly explaining the West’s slower progress in poverty reduction. Compared to the period 2005-2010, the contribution of gains in assets to consumption growth between 2010 and 2016 was proportionally much higher in the Eastern divisions (Figure 1.6). Hill and Endara (2019b) show that slower progress in reducing family sizes and in average education of adult members in the rural Western divisions is associated with the limited progress in poverty reduction in the rural West and the reemergence of the East-West welfare divide. Figure 1.5. Estimated contribution of demographic and asset changes to consumption growth in rural Bangladesh, 2005-2010 and 2010-2016 0,05 Predicted change in log consumption 0,02 0,02 0,01 0,01 0,01 0,02 0,02 (0,01) (0,03) 2005-2010 2010-2016 Household size % members younger than 18 Land size Access to electricity Average years of education Source: Authors’ calculations using HIES 2005, 2010, and 2016. Note: Results from an Oaxaca-blinder decomposition of changes in log consumption between two years. Y-axis measures the predicted consumption per capita growth over the reference period from changes in household demographics and assets (i.e., education of adult members, household demographics, access to electricity, and size of land ownership). 36 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.6. Estimated contribution of demographic and asset changes to consumption growth in rural Bangladesh: results by division in two periods 0,16 0,14 0,12 0,10 0,08 0,06 0,04 0,02 0 al na i a g r et l ah Al pu on ak ris lh ul jsh ng Dh ag Sy Ba Kh Ra Ra itt Ch 2005-2016 2010-2016 Source: Authors’ calculations using HIES 2005, 2010, and 2016. Note: Results from an Oaxaca-blinder decomposition of changes in log consumption between two years. The bars measure the predicted consumption per capita growth over the reference period from changes in household demographics and assets (i.e., education of adult members, household demographics, access to electricity, and size of land ownership). Another element behind the diȂerential performance across areas was the smaller contribution of agriculture to poverty reduction compared to previous periods. Between 2010 and 2016, poverty reduction among households primar- ily engaged in agriculture was still positive, but considerably slower than in the past.37 Even though about 47 percent of rural households were primarily engaged in agriculture, they contributed only 27 percent of poverty reduction. About 36 percent of the poverty decline in rural areas was led by households engaged in services, and 23 percent stemmed from households in industry. This contrasts with the period 2005-2010, when agriculture was the most important sector reduc- ing poverty (Figure 1.7). During the period from 2005 to 2010, 69 percent of rural poverty reduction was among households primarily engaged in agriculture. 38 37 Hill and Genoni (2019). 38 An income growth decomposition based on the 62-village panel shows that multiple sources of income contributed to faster income growth and poverty reduction between 2000 and 2014. For house- holds that were poor at baseline, all sources of income (farm, non-farm, and remittance) contributed equally to households’ moving out of poverty aȅer 2000, indicating diversi‫ޔ‬ed and multiple pathways 37 BANGLADESH RURAL INCOME DIAGNOSTICS The declining role of agriculture in poverty reduction is associated with the slow per- formance of the sector since 2010, particularly the crop subsector, which accounts for half of agricultural GDP. While the growth rates of agricultural output were above 5 percent between 2005 and 2010, they have been declining consistently since 2010 and remained much lower compared to the trend growth rates in the industry and service sectors (Figure 1.8). Total factor productivity (TFP) growth in agriculture has decelerated sharply in the same period. The TFP growth for the sector declined from an average 2.4 percent annually in 2001-10 to 1.0 percent in 2011-16. Figure 1.7. Contribution of economic sectors to rural poverty reduction across time 70% 60% 50% 40% 30% 20% 10% 0% Agriculture Services Industry Sectoral shiȅ 2000-2005 2005-2010 2010-2016 Source: Authors’ calculations using HIES 2000, 2005, 2010, and 2016. Note: Results obtained from Ravallion and Huppi (1991) decompose changes in poverty over time into intra-sectoral eȂects, a component due to population shiȅs across sectors, and an interaction (not displayed). Sector of employment defined based on reported hours of work in each sector. Sectoral shiȅ is relatively small as it is defined at the household level. out of poverty. Remittance incomes were the largest contributor of income gains for non-poor house- holds. In more recent years, 2008 and 2014, non-agriculture income, particularly from businesses and services, gained importance in rural income growth and was the major driver of income growth for poor households. Within agriculture, between 2000 and 2008, income from all sources (rice, non-rice cereals, non-crop agriculture, and labor) made signi‫ޔ‬cant contributions to poor households’ income. However, the role of non-crop agriculture has become much more prominent aȅer 2008, while the role of crop agriculture (both rice and non-rice) for poor households has become modest in the same period (Ahmed and Gautam 2020). 38 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.8. Annual growth rates in total and sectoral GDP, 2000-2019 12% 10% 8% 6% 4% 2% 0% 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 GDP Agricultural Services Industry Source: Authors’ calculations using WDI. With a larger share of households engaged in agriculture in the Western regions, slower growth in agriculture contributed to the limited progress in poverty in those areas. As subsequent sections will highlight, households in the Western divisions (Khulna, Rajshahi, and Rangpur) are more likely than households in the Eastern divisions to report their main sector of work as agriculture. As agri- cultural growth was slower and less poverty reducing39 between 2010 and 2016, diȂerences in sectors of work between households in West and East seem to have played an important role in driving the increased disparities across areas.40 The slowing performance of agriculture is rooted in falling real prices since 2010. While the upward changes in real prices of agricultural commodities accounted for more than half of agricultural growth between 2005 and 2010, the real prices of agricultural commodities have declined aȅer peaking in 2008. Prices rebounded brie‫ޕ‬y in 2012 and 2013 but have since fallen sharply, turning negative in 2016 (Figure 1.9). 39 Agriculture was also less poverty reducing, as the elasticity between growth and poverty also declined between 2010-2016, compared to previous periods. 40 Hill and Genoni (2019). 39 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.9. Decomposition of aggregate crop revenue growth, 1996-201841 12 10 8 6 4 2 0 -2 -4 -6 15 16 17 18 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 20 20 20 20 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Yield Area Diversification Price Interaction Revenue Source: Authors' calculations using FAOSTAT data. Despite the reduced contribution of agriculture to poverty reduction in recent years, faster progress in industry and services allowed rural poverty reduction to continue at a strong pace. Faster consumption growth among non-agricultural households supported a strong pace of poverty reduction of about 1.4 percentage points per year, though this was slower than the 1.7 percentage points per year observed between 2005-2010. Between 2005 and 2010, about 30 percent of poverty reduction was driven by industry and services. In contrast, between 2010 and 2016, households in these two sectors accounted for 59 percent of poverty reduction. Stronger poverty gains among households engaged in garments and manu- facturing also help explain increased spatial disparities, as these activities are concentrated in the Eastern divisions around the Dhaka-Chittagong corridor. Between 2010-2016, households in garments, other manufacturing, and transport showed much stronger contributions to poverty reduction than in the past (Figure 1.10). While garments and other manufacturing contributed about 3 percent of rural poverty reduction between 2005 and 2010, households in these sectors contributed 18 percent of poverty-reduction gains between 2010 and 2016. For households in transport, the contribution was about 7 percentage points higher in 2010-2016. More households in Eastern areas were primarily engaged in gar- ments and manufacturing, and so bene‫ޔ‬ted more from the gains in those sectors. 41 For details on decomposition methodology, see Minot et al. (2006). 40 ASSESSING CONTEXT AND HETEROGENEITY For instance, in 2016, about 16 percent of rural households in Dhaka division were primarily engaged in garments and manufacturing, compared to 8 percent in Khulna and 9 percent in Rangpur divisions. Bene‫ޔ‬ts in the transport sector were more widespread, as employment shares in this sector are more evenly distrib- uted across areas. Figure 1.10. Contribution of sectors within industry and services to rural poverty reduction across time ǗǕʢ ǖǛʢ ǖǗʢ ǝʢ Ǚʢ Ǖʢ 1 %" &) 1/6 + 1 /&+ / ,/ "+ 12 1%" ,1 1 1&, 20 0- /0 $ ! /" /*  +! 1/2 + !"ȡ +  / / +0  %" /  , 2# 1 + ǗǕǕǚȒǗǕǖǕ ǗǕǖǕȒǗǕǖǛ  Source:1ƛ ) 2)1&,+020&+$ ǗǕǕǚǾǗǕǖǕǾ+!ǗǕǖǛǽ Note:"02)10,1&+"!#/,*3))&,++! 2--&țǖǞǞǖȜ!" ,*-,0" %+$"0&+-,3"/16,3"/1&*"&+1, &+1/Ȓ0" 1,/) "ƛ" 10Ǿ  ,*-,+"+1 !2" 1, -,-2)1&,+ 0%&ƞ0  /,00 0" 1,/0 ț+,1 !&0-)6"!ȜǾ +! + &+1"/ 1&,+ț+,1!&0-)6"!Ȝǽ" 1,/,#"*-),6*"+1!"#&+"!0"!,+/"-,/1"!%,2/0,#4,/(&+" % 0" 1,/ǽ Understanding rural incomes Having examined rural poverty trends, their underlying drivers, and key factors in‫ޕ‬uencing East-West disparities in poverty reduction, we have a foundation for characterizing rural income dynamics more broadly. Accordingly, we now turn to look at how Bangladesh’s rapid development has reshaped rural income patterns, enabling swiȅ consumption gains for many households, but leaving many rural people still exposed to substantial risks. Progress in rural incomes has not been uniform, and identifying main factors shaping that heterogeneity is important for setting policy priorities. Rapid income transformations—that remain vulnerable The fact that all sectors of Bangladesh’s economy have contributed to poverty reduction in the past re‫ޕ‬ects a country that has been transforming rapidly. Since 2000, there have been large shiȅs in the sectoral distribution and geographical focus 41 BANGLADESH RURAL INCOME DIAGNOSTICS of economic activity in Bangladesh. Bangladesh has moved at a faster pace than most other developing countries in this process of structural transformation.42 Many Bangladeshi households derive labor income from multiple sources. Considering all rural areas, on average 39 percent of households derive income from diȂerent sectors, while 39 percent derive income only from agriculture, and the remaining households only from industry (3 percent) or only from ser- vices (20 percent). Comparing across income quintiles shows that these patterns do not diȂer substantially between poor and richer households in rural areas (Figure 1.11), although diȂerences are important in absolute terms. Median household per capita labor income was 2,255 Takas in 2018 and varied from 700 Takas for the poorest quintile to 5,000 Takas for the richest quintile.43 Figure 1.11. Percentage of total labor income derived from agriculture, industry, and services Rural areas, by quintiles, 2018 100% 90% 80% 36% 39% 41% 41% 37% 39% 70% 60% 14% 50% 15% 20% 4% 18% 22% 28% 3% 40% 3% 3% 3% 3% 30% 20% 46% 43% 38% 33% 32% 39% 10% 0% 1 2 3 4 5 All Consumption per capita quintile Only agriculture Only industry Only services /,*!&ƛ"/"+10,2/ "0 Source: Authors’ calculations using BIHS 2018. Note: The figures denote the average share of sectoral sources of income across consumption quintiles. 42 Hill and Genoni (2019) show that, between 2010 and 2016, the share of employment in agriculture in Bangladesh fell about 24 percentage points, compared to 14 points in the South Asia region as a whole. 43 Figures estimated using BIHS 2018. Household labor income is calculated by adding all the incomes of all activities performed by the individuals within the household. 42 ASSESSING CONTEXT AND HETEROGENEITY The movement to oȂ-farm activities has also been observed for the rural poor. Despite their stronger connection to agriculture, households in the poorest quin- tile have also moved out of agriculture as an exclusive source of income. Between 2012 and 2018, the percentage of households engaged exclusively in agriculture fell 5 percentage points. Sources of income by sector vary between poor and non- poor households (Figure 1.12), but not dramatically. Comparing the poorest and richest quintiles, the average share of income from agriculture declines from 59 percent to 44 percent. Figure 1.12. Percentage of total labor income obtained from agriculture and non-agriculture activities Rural areas, by quintiles, 2018 100% 90% 80% 44% 52% 48% 70% 59% 58% 60% 50% 40% 30% 56% 48% 52% 20% 41% 42% 10% 0% 1 2 3 4 5 Consumption per capita quintile Non-Ag Agriculture Source: Authors’ calculations using BIHS 2018. Note: Bars show the average percentages by consumption quintile. This process of transformation has allowed many households to improve their income and consumption. However, a large share of households in rural areas still have consumption levels very close to the poverty line. In 2016, about 55 percent of rural households were not poor, but remained vulnerable, with con- sumption levels between the poverty line and twice the poverty line (Figure 1.13). When faced with income shocks, these households are at risk of falling into pov- erty. Median household per capita consumption in rural areas was 3,000 Takas, while the average upper poverty line for rural areas was about 2,300 Takas.44 44 Figures estimated using HIES 2016.These patterns are robust if we use the 2018 BIHS consumption aggregate. 43 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.13. Poverty and vulnerability by area (percentage of the population in 2016) 100% 90% 22% 19% 30% 80% 70% 60% 55% 50% 54% 51% 40% 30% 20% 25% 27% 10% 19% 0% National Rural /+ Headcount 2)+"/)" Middle Class Source: Authors’ calculations using HIES 2016/17. Note: ,3"/16 !"#&+"! 20&+$ 1%" ,ƛ& &) 2--"/ -,3"/16 /1"ǽ 2)+"/)" %,20"%,)!0 /" %,20"%,)!0 4&1%-"/ -&1 ,+02*-1&,+"14""+1%",ƛ& &)2--"/-,3"/16)&+"+!14& "1%"2--"/-,3"/16)&+"ǽ &!!)"Ȓ )00%,20"%,)!0/"1%,0"4&1%-"/ -&1 ,+02*-1&,+,3"14& "1%"2--"/-,3"/16)&+"ǽ Panel data con‫ޔ‬rm substantial transitions in and out of income-based poverty. Household surveys tracking the same rural households over time indicate that 42 percent of households were in the bottom income quintile at some point between 2000 and 2014 (Figure 1.14a). Households appeared to have experienced signi‫ޔ‬cant mobility (both upward and downward) in the same period. The majority of those who were extremely poor or poor in 2000 were able to escape poverty by 2014 (almost 9 out of 10), while a substantial share of vulnerable non-poor households improved their economic status (Figure 1.14b). Among the households that were moderately poor in 2000, 58 percent moved out of poverty by 2014, while 11 percent of this group descended into poverty during the same period. Among the non-poor, 39 percent of households classi‫ޔ‬ed as vulnerable non-poor in 2000 were able to move up to the rich non-poor category by 2014, but a substantial 35 percent of the vulnerable non-poor fell back into poverty by 2014. Almost 2 out of 5 rich non-poor households experienced downward mobility to a lower economic status (Figure 1. 14b). This highlights the importance for sustained rural development of focusing not only on the poor but also on vulnerable non-poor households. 44 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.14. Household poverty transitions, 2000-2014 a. Percentage of households moving b. Percentage of households moving in and out of the poorest income quintile in and out of poverty 100 80 3 60 87 58 39 40 Percentage 59 42 20 % 31 26 13 55 0 -11 -20 -35 -41 -40 -60 Po me Po ate Po le or n- ab Po or or or tre er No lner n- od Ex No M Vu ch Ri Always in bottom 20% Poverty Status 2000 Always in top 80% At some point in bottom 20% Static Up Down Source: Authors’ calculations using 62-village panel. Note: For details on methodology, see Ahmed and Gautam (2020). Poverty is defined using the income aggregate and the oȂicial poverty lines from BBS. The substantial movement in and out of poverty partly re‫ޕ‬ects low quality of jobs in rural areas, largely dominated by self and daily-wage employment. A worker’s type of employment (daily laborer, self-employed, or salaried) appears to be a better pre- dictor of whether or not they belong to a poor household than is the main economic sector they work in. The share of workers in agriculture declines with consumption levels, but the drop is not substantial (Figure 1.15a). About 56 percent of bottom-quin- tile workers are engaged in agriculture, compared to 40 percent of the richest quintile. The largest factor that separates the poor from the non-poor is their type of employ- ment: whether they are day laborers, self-employed, or salaried. About 41 percent of workers in the bottom quintile are day laborers, compared to 12 percent in the rich- est quintile (Figure 1.15b). In addition, about 7 in 10 rural workers are self-employed, either in farm or non-farm activities. Higher-quality salaried jobs only comprise about 7 percent of employment and are more likely for higher consumption quintiles. 45 BANGLADESH RURAL INCOME DIAGNOSTICS When looking at younger cohorts, a substantial share of persons are working as day laborers. In 2018, 34 percent of young male rural workers (ages 15-29) were day laborers, and 15 percent were unpaid workers. Figure 1.15. Sector and type of employment by quintiles, 2018 a. Sector of employment b. Type of employment 100% 3% 3% 5% 4% 3% 4% 90% 80% 44% 47% 51% 52% 70% 58% 60% 54% 63% 65% 66% 60% 71% 75% 50% 40% 4% 5% 30% 56% 7% 7% 53% 49% 42% 48% 41% 8% 20% 40% 12% 31% 26% 26% 10% 19% 12% 0% 1 2 3 4 5 All 1 2 3 4 5 All Agriculture Non-agriculture Daily wage Salaried Self-employed Unpaid Source: Authors’ calculations from BIHS 2018. Note: Working-age adults (15-64 years old) grouped according to their household consumption quintile. There is large variation in employment types between workers in poor and non-poor households. Within agriculture, 1 in 3 workers are daily workers. Day laborer status is much more prevalent among the poor (about 52 percent). Outside agriculture, the poor are more engaged in small-scale trading and low- er-quality service jobs. Trade comprises 32 percent of rural employment out- side agriculture, followed by transport (17 percent), construction (13 percent), and other services (12 percent). The poorest quintile is less likely to engage in trade and services (Figure 1.16). When bottom-quintile workers do participate in trade, they are more likely to be engaged in smaller-scale trading (e.g., road- side stands) or low-productivity jobs, such as rickshaw driving, maid services, and begging (Figure 1. 17). In contrast, the richest quintile is largely involved in medium and large-scale trading, as well as in sales within private enterprises. Within industry, the poorest workers are more likely to be in masonry, tailoring, construction, and carpentry. 46 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.16. Types of non-farm employment by quintile, 2018 100% 15 14 9 15 90% 21 16 80% 8 12 11 17 12 70% 9 12 10 Percentage 23 20 17 60% 22 10 8 50% 5 13 15 13 8 6 40% 4 9 19 30% 12 10 20% 11 41 44 28 32 10% 25 18 0% 1 2 3 4 5 All Trade Education Manufacturing Construction Transport Other Services Health Other Source: Authors’ calculations using BIHS 2018. Figure 1.17. Main non-farm occupations reported by poorest and richest quintiles, 2018 a. Share of workers in industry b. Share of workers in services Mason Medium trader Retail/sales Tailor/ Miscellaneous seamstress jobs Carpenter Large trader Teacher Construction labor Small trader Factory worker Driver Government Handicraȅs Rickshaw/van pulling Goldsmith/ Hair cutter silversmith Apprentice Small industry Beggar 0% 10% 20% 30% 40% 0% 5% 10% 15% 20% 25% 30% Richest quintile Poorest quintile Source: Authors’ calculations using BIHS 2018. Note: A residual category, not displayed, includes activities involving fewer than 2 percent of workers. On the demand side, a large share of employment is based on informal and subsistence microenterprises. According to Bangladesh’s latest Economic Census, from 2013, 98 percent of ‫ޔ‬rms in the country had less than 10 employ- ees. Household-based enterprises are more prevalent in rural areas (42 percent of ‫ޔ‬rms) than in urban areas (36 percent). Microenterprises account for about 55 percent of jobs in Bangladesh, which is higher than comparator countries. 47 BANGLADESH RURAL INCOME DIAGNOSTICS Tanzania has a similar share of microenterprises, but those ‫ޔ‬rms comprise less than 40 percent of Tanzania’s employment. In Vietnam, while microenterprises represent about 70 percent of ‫ޔ‬rms, they account for less than 10 percent of employment. Most rural microenterprises are engaged in relatively unsophisticated activ- ities and are largely informal.45 A detailed breakdown of rural enterprise jobs (4-digit level) using the latest Economic Census shows that 12 activities account for about 6 in 10 rural enterprise jobs (Figure 1.18). While retail grocery stores comprise many rural enterprise jobs, some manufacturing activities such as ready-made garments (RMG), handloom, brick, and furniture manufacturing also play an important role supporting rural enterprise jobs. However, most of the enterprise jobs are in areas with limited space for growth. This includes gro- cery stores, tea stalls, and similar businesses. In addition, Gautam and Faruqee (2016) ‫ޔ‬nd that there has been very little change in the sophistication of non- farm activities in rural areas over the decade 2003-2013, even though these activities are more likely to bring higher returns to households. Rural nonfarm enterprises’ limited capacity to upgrade or bene‫ޔ‬t from improved connectivity suggests that, on average, these ‫ޔ‬rms are not well positioned to expand and oȂer a major source of job growth. Changing demographics and rapid urbanization will have a signi‫ޔ‬cant impact on labor market developments in the coming decades. Bangladesh’s working-age population is expected to increase from about 67 percent to 70 percent of the total population between 2018 and 2030. This trend is already putting pressure on the labor market to absorb new cohorts of workers.46 Rapid urbanization implies that, by 2030, about 46 percent of the population will be living in urban centers. Rising rural unemployment, particularly among youth and women, is a concern. Since 2010, rural unemployment rates have increased, driven to a large extent by younger cohorts. Female unemployment rates are dou- ble those of males, and increases in unemployment have been greater for females. Figure 1.19 shows that female youth unemployment has increased about 300 percent, compared to a 50 percent increase for males across all rural areas. The spikes in unemployment are signi‫ޔ‬cant in Western divisions, including Barisal. In those divisions, unemployment rates also increased for older cohorts, although they remain at a very low level. 45 Farole and Cho (2017). 46 Farole and Cho (2017). 48 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.18. Leading rural enterprise activities, by number of jobs, 2013 Manufacture of rice/rice milling Manufacture of wooden furniture Retail sale of pharmaceutical/ cosmetic/toiletries Manufacture of handloom textile Manufacture of brick/ block, tiles Tailoring services Retail sale of food Tea stalls Activities of religious organizations Education Manufacturing of wearing apparel (RMG) Retail grocery stores 0 200 400 600 800 1000 Number of jobs (thousands) Source: Economic Census 2013. The larger increases in the share of women looking for jobs may also indicate an increased likelihood of new, better-educated female cohorts’ seeking to engage in the labor market. This is consistent with recent research.47 For better-educated workers, a recent rise in unemployment rates signals chal- lenges in access to good jobs. The unemployment rate among youth who have com- pleted secondary school and above is over 10 percent, and higher for females than for males; the unemployment rate among postsecondary-educated youth is close to 20 percent. Part of the increase in youth unemployment observed in recent years might be explained by a change in the educational composition of new labor-market entrants and the inability of the market to absorb those entrants. Better-educated youth who enter the labor market may have a high reservation wage, while high-pay- ing jobs are unavailable, given the slowdown in the labor market in recent years.48 47 SolotaroȂ (2019). 48 Farole and Cho (2017). 49 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.19. Rural youth unemployment rates, 2010 and 2016, by division a. Males 18-34 years old b. Females 18-34 years old 25% 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% g a na i al et sh r ah pu g a na i al et sh r on ak ah ris pu lh on ak de ul ris lh jsh de ng ul Dh ag Sy jsh Ba Kh ng Dh la ag Sy Ba Kh la Ra Ra itt ng Ra Ra itt ng Ch Ba Ch Ba 2010 2016 Source: Authors’ calculations using LFS 2010 and 2016. As employment in agriculture falls, it is important to ensure that new sources of non-farm income are of high quality. The recent slowdown in job creation in manufacturing, particularly in the RMG sector, is a source of concern, as this can push new workers to engage in lower-quality service jobs. Evidence for urban areas indicates that this is already happening. The limited job creation in manufacturing observed since 2013 has reduced productivity among the urban self-employed in services, in turn signi‫ޔ‬cantly slowing poverty reduction in urban areas.49 Attention to the quality of new employment is vital to ensure that structural transformation does not generate greater vulnerability, associated with low-quality or subsistence non-farm self-employment. Weather and climate-change risks are another important factor behind the sub- stantial movement in and out of poverty. Bangladesh’s location at the lower end of the Ganges-Brahmaputra-Meghna delta renders it especially vulnerable to the adverse impacts of weather and climate change. Two-thirds of the country is less than ‫ޔ‬ve meters above sea level, hence susceptible both to sea level rise and tidal 49 Hill and Genoni (2019). 50 ASSESSING CONTEXT AND HETEROGENEITY ‫ޕ‬ooding during storms. The Global Climate Risk Index ranks Bangladesh as the world’s seventh most aȂected country over the period 2000-2019. Agricultural incomes are particularly vulnerable to extreme weather events. Agricultural production in the country has already been adversely aȂected by climate change, including through temperature increase, sea level rise, saltwater intrusion, rainfall variability, and increased extreme weather events. These impacts are partic- ularly marked in coastal areas and arid and semi-arid areas.50 Climate change will further exacerbate the vulnerability and risks facing agriculture. Increased soil and water salinity due to sea level rise is projected to result in a 15.6 percent yield reduc- tion in high-yielding rice varieties by 2050.51 Climate change is also adversely aȂect- ing the ‫ޔ‬sheries sector, and about 15 percent of ‫ޔ‬shermen have already changed their occupation due to a decrease in ‫ޔ‬sh collection rates in open water. More than 15 percent of livestock die of diȂerent diseases every year, and many producers are going out of business due to disease tolls that result in large economic losses. This vulnerability is exacerbated by limited ability of households to cope with income shocks. Negative shocks experienced by rural households are largely man- aged with own resources.52 According to HIES 2016, about 17 percent of rural house- holds reported a negative event or income shock in the year preceding the interview, 81 percent of which were related to weather events (such as ‫ޕ‬oods, droughts, or landslides). The main coping strategies reported were reducing food consumption (58 percent), using own savings or selling assets (66 percent), and getting help from friends (53 percent). About 33 percent of households reported obtaining govern- ment help and 20 percent reported using credit. According to FINDEX (2017), about 63 percent of all rural households reported that they could obtain emergency funds if needed. However, only 50 percent of households in the two poorest quintiles believed they could do so. Due to lack of options, the poor rely more on reducing food consumption when coping with a shock. This is consistent with large transitions in and out of pov- erty. The likelihood of reporting negative shocks varies little between poor and bet- ter-oȂ households or across divisions. However, poorer and better-oȂ households use diȂerent coping responses. Households in the poorest quintile are more likely to adjust food consumption (77 percent), get help from friends (63 percent), and receive government help (52 percent). Across divisions, Rangpur households are 50 Alam et al. (2018). 51 World Bank (2019c). 52 Genoni et al. (2020). 51 BANGLADESH RURAL INCOME DIAGNOSTICS more likely to adjust food consumption and obtain government help. Using credit is reported more oȅen in Barisal and Sylhet. Migration is also a means of coping with shocks, for instance in Chittagong (29 percent) (Table 1.1). Table 1.1. Coping responses to shocks reported in the past year, percentage of households Looked Reduced for Own savings Help from food more Government or assets friends consumption work Migrated help Credit Consumption quintile 1 44% 63% 77% 25% 12% 52% 13% 2 63% 54% 66% 23% 16% 28% 20% 3 78% 49% 51% 19% 14% 28% 21% 4 74% 44% 46% 17% 14% 29% 27% 5 78% 50% 43% 10% 14% 22% 25% Division Barisal 84% 64% 33% 3% 7% 15% 49% Chittagong 67% 51% 36% 7% 29% 33% 28% Dhaka 82% 54% 21% 17% 11% 33% 21% Khulna 88% 37% 84% 37% 5% 7% 21% Rajshahi 90% 22% 35% 24% 22% 28% 31% Rangpur 25% 79% 84% 15% 7% 64% 5% Sylhet 71% 51% 46% 19% 25% 9% 42% Rural 66% 53% 58% 19% 14% 33% 21% Bangladesh Source: Author’s calculations using HIES 2016 for rural areas. 52 ASSESSING CONTEXT AND HETEROGENEITY Three factors behind heterogeneity in rural incomes When thinking about ways to increase rural income growth in an inclusive man- ner, three features or rural incomes are important to highlight. First, incomes within agriculture are still largely centered around rice production. Second, rural Bangladeshi women are still largely outside the labor market or performing unpaid work. Third, geographic location largely shapes inequalities in income due to diȂer- ential engagement in agriculture and manufacturing, urbanization, and migration. Rice is central for rural incomes, but the sector has been underperforming Despite the growing importance of non-agricultural incomes in rural areas, agricultural incomes remain a large component of total rural incomes. On aver- age, about 54 percent of household labor income still comes from agriculture.53 Rice production contributes a large share of agricultural incomes. Rice accounts for 71 percent of the country’s cultivated area and 59 percent of total crop value.54 Rice is by far the dominant crop for all rural households in agriculture. BIHS data indicate that about 95 percent of farmers across consumption quintiles grow rice (either alone or in combination with other crops). About 52 percent of farmers grow exclusively rice (which may include diȂerent varieties), and 15 percent grow exclusively one variety of rice. Farmers in the poorest quintiles are slightly more likely to produce exclusively rice than are farmers in the richest quintile (53 per- cent versus 47 percent) (Figure 1.20). The share of Bangladeshi households across all consumption quintiles growing rice (including in combination with other crops) is high compared to similar countries such as Vietnam. In Vietnam, rice is grown by 85 percent of households in the poorest quintile but only 49 percent of farmers in the richest quintile.55 Bangladesh has made many agricultural policy reforms since the 19ч0s with a focus on ensuring food security. These policy reforms were broad based and have contributed to the sustained high growth, including TFP in agriculture, from 1990s to about 2010. However, within the broadly admirable policy framework, food secu- rity concerns have also prioritized certain policies with a preferential treatment for rice which have tilted farmer incentives in favor of rice. These include a heavy focus on rice in R&D, which has raised rice productivity and lowered production risks. 53 BIHS 2018. 54 Ahmed et al. (2021). 55 CazzuȂi et al. (2017). 53 BANGLADESH RURAL INCOME DIAGNOSTICS In addition, public procurement for strategic reserves is exclusively focused on rice, which in competitive markets create a bias against other crops. Finally, strong returns to fertilizers with improved rice varieties and sustained fertilizer subsidies also tilt the bene‫ޔ‬t-cost ratios in favor of rice, compared to other crops. These pol- icies have contributed to the country’s success in achieving self-suȂiciency in rice production but are now creating headwinds constraining the emergence of a more diversi‫ޔ‬ed production system. Figure 1.20. Percentage of farmers growing only rice, across consumption quintiles, 2018 43% 41% 44% 42% 43% 46% 53% 56% 51% 53% 47% 52% 4% 4% 5% 5% 7% 5% 1 2 3 4 5 All Only non rice crops Only rice Rice and other crops Source: Authors’ calculations using BIHS 2018. Despite the large share of farm households engaged in rice, the rice sector has not been performing well recently, and this has limited its contribution to income growth and poverty reduction. GDP growth in the rice subsector slowed from 5.1 percent per year in 2005-2010 to 1.2 percent in 2010-2016, contributing substantially to a slowdown in overall agricultural growth. This was mainly driven by a decline in real prices along with a slowdown in TFP growth. In contrast, the ‫ޔ‬shing and forestry sectors have performed well (Figure 1.21). As previously discussed, between 2010 and 2016, the contribution of agriculture to poverty reduction declined, partly due to the slowdown in agricultural growth substantially attributable to the crop sector.56 56 Hill and Genoni (2019). 54 ASSESSING CONTEXT AND HETEROGENEITY Analysis tracking the same households over the period 2000-2014 shows that income from agriculture accounted for about 40 percent of households’ annual per capita income growth, but that this was not driven by the crop sector. Non-crop agri- culture was the main source of income gains in agriculture (2 percent), followed by agricultural labor (0.1 percent). Income from crop agriculture (both rice and non- rice) registered negative growth over the same time period.57 Figure 1.21. Sectoral growth rate of real GDP, 2005-2019 10 9 8 7 6 5 4 3 2 1 0 6 7 8 9 0 1 2 3 4 5 6 7 8 9* -0 -0 -0 -0 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 05 06 07 08 09 10 11 12 13 14 15 16 17 18 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Total agriculture and forestry Crops & horticulture Animal Farmings Forest and related services Fishing Source: Authors’ calculations using Ministry of Finance (2019). Note: Real GDP is calculated with base year 2005-6. *2018-2019 provisional While many crops oȂer farmers higher returns than rice, the majority of farm- ers continue to grow rice, in part because cereal production involves lower price risks. Bangladesh has seen some diversi‫ޔ‬cation toward non-paddy crops in recent years, but progress remains slow.58 While potential gross margin is an important driver for farmers’ cropping choices, risks also play a critical role in their decision making. While Bangladeshi farmers have signi‫ޔ‬cant know-how for rice production, rice also ensures farmers’ own food security, as it is their main staple. Public policy further contributes to farmers’ perception that rice carries less risk. Public policy support in the form of input subsidies, procurement support, and 57 Ahmed and Gautam (2020). 58 Ahmed et al. (2020). 55 BANGLADESH RURAL INCOME DIAGNOSTICS R&D to ensure food security contributes to the lower risks of producing cereals.59 In particular, the production of rice and wheat carries much lower price risks than other high-value agricultural commodities, as price volatilities for cereals are much lower than those of non-cereal agricultural commodities.60 Both price vola- tilities across time and space are higher for non-cereal agricultural commodities. The government maintains a regular rice procurement program to stabilize falling paddy prices following harvest seasons and also manages the public food distri- bution system to keep rice prices stabilized throughout the year.61 Rural labor markets still show a deep gender divide Employment and income sources diȂer sharply between men and women in rural areas. Female labor force participation has been increasing steadily, to about 36 percent. However, this is considerably lower than the 82 percent labor force par- ticipation of males. Bangladesh’s female labor force participation outperforms the South Asia regional average (23.5 percent) but falls far short of ‫ޔ‬gures in other comparator countries. For comparison, Kenya, with a similar per capita GDP, has 72 percent of females in the labor force. In Vietnam, female labor force participation is around 73 percent.62 Bangladesh’s female labor force participation is higher in rural areas than in urban areas. In 2016, women’s labor force participation was 37 percent in rural and 31 percent in urban areas. Among divisions, female labor force partici- pation is highest in Rajshahi (50 percent all areas and 54 percent in rural areas). It is lowest in Sylhet (19 percent all areas, 21 percent in rural areas) (Table 1.2). Of women who work, about 60 percent are in agriculture, with a large pro- portion taking care of livestock or engaged in unpaid agricultural activities. Women’s share in agricultural employment has been growing over time, re‫ޕ‬ecting an increasing ҂feminization of agriculture.҃ Women workers accounted for 44 per- cent of total agricultural employment in Bangladesh in 2016, up from 34 percent in 2006. The number of women workers in agriculture increased from 7.7 million in 2006 to 11.2 million in 2016 (LFS surveys). However, 28 percent of women working 59 World Bank (2020a). 60 Hoque and Ahmed (2020). 61 Considering the importance of this staple in the Bangladeshi diet and the focus of the government on ensuring food security, rice imports are an important mechanism for price stabilization in the local mar- ket. While Bangladesh maintains a 28 percent import duty on rice, it oȅen reduces the duty or imposes a supplemental duty depending on the trend in domestic price. For example, the government cut import duty on rice to 2 percent in 2017 to address production shortages due to two consecutive ‫ޕ‬oods; how- ever, it raised the duty to 28 percent in 2018 following a normal rice production year and increased it further to 55 percent in 2019, following record domestic production of rice in 2019. 62 WDI 2019. 56 ASSESSING CONTEXT AND HETEROGENEITY in agriculture are engaged in livestock and 11 percent in unpaid farm activities. In contrast, only 6 percent of working men report deriving income from livestock, and only 2 percent are unpaid. Women’s role in livestock tends to be primarily home-based activities such as feeding and milking cows, as well as raising small ruminants and backyard poultry. Women’s involvement in marketing livestock is limited. Men are more likely to participate in the non-farm sector (Figure 1.22), though women have been increasing their participation. Table 1.2. Men’s and women’s labor force participation in Bangladesh, 2016 National Rural Areas Urban Areas Male Female Male Female Male Female Chittagong 80% 32% 79% 34% 81% 27% Dhaka 83% 33% 82% 33% 85% 34% Khulna 84% 34% 84% 36% 83% 27% Rajshahi 84% 50% 85% 54% 83% 36% Barisal 81% 26% 81% 27% 82% 24% Sylhet 82% 19% 83% 21% 81% 14% Rangpur 85% 39% 85% 41% 84% 32% Bangladesh 83% 35% 83% 37% 84% 31% Source: Authors’ calculations based on active-age population using LFS 2016. Rural women engage in very few occupations within industry, while there is more variation in their occupations within services. Women who have com- pleted secondary school or more are more likely to be employed in the industry or service sectors.63 Women in oȂ-farm activities within industry derive income mainly from two activities, seamstress and craȅs (Figure 1.23a). Among women in the poorest quintile, the dominant job is seamstress (48 percent) followed by craȅswoman (43 percent). In the richest quintile, seamstress is the most import- ant activity (69 percent). Women’s employment in the service sector is more var- ied (Figure 1.23b). Comparisons of occupations across time indicate very little change in the oȂ-farm occupational pro‫ޔ‬le of women in rural areas. 63 Multivariate regression models using LFS 2016/17 for rural areas show that unskilled women are more likely to participate in the labor market, re‫ޕ‬ecting the large presence of women in unpaid agriculture and livestock. When assessing the probability of participating in nonagricultural sectors, married women and those living in larger households are less likely to participate. Women who have completed secondary school or more are more likely to be employed in the industry or service sectors. 57 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.22. Types of employment for rural males and females, 2018 Percentage of workers by type of employment, by gender, 2018 20% 35% 16% 6% 5% 10% 15% 16% 11% 6% 28% 22% Male Female Non-farm salaried Farm daily labor Livestock Crop Non-farm daily labor Farm unpaid Fishing Non-farm unpaid Non-farm self-employed Source: Authors’ calculations using BIHS 2018. Figure 1.23. Main off-farm occupations among women in rural areas, 2018 a. Percentage of female rural workers b. Percentage of female rural workers in industry in services Teacher Tailor/seamstress Sales in private enterprise Miscellaneous jobs Medium trader Handicraȅs Government Housemaid NGO worker Small trader Factory worker Beggar 0% 20% 40% 60% 80% 0% 10% 20% 30% 40% Richest quintile Poorest quintile Source: Authors’ calculations using BIHS 2018. Note: Bars present the percentage of women in diȂerent occupations conditional on being in industry (panel a) or services (panel b). Occupations with small or undefined sample shares are not included. 58 ASSESSING CONTEXT AND HETEROGENEITY Women are engaged in diȂerent sectors than men, but once in the same sector, gender wage gaps appear to be small. The raw gender wage gap has decreased dramatically in the past years, from a 43 percent gap in 2013 to a 24 percent gap in 2016. In fact, Bangladesh has the smallest gender wage gap in the region.64 Conditional wage gaps within sectors indicate no large diȂerence between men and women, particularly in agriculture and services (Figure 1.24). This suggests that gender productivity gaps may be linked to women’s engagement into diȂer- ent occupations and paid employment. Once those barriers are considered, and men and women are compared within the same sector, wage disparities are small. Figure 1.24. Gender wage gap by sector and division, 2016 a. Gap in rural areas by sector b. Gap in agriculture by division -0.1 -0.05 0 0.05 -0.2 -0.1 0 0.1 0.2 Agriculture Industry Services Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet c. Gap in industry by division d. Gap in services by division -0.15 -0.1 -0.05 0 0.05 0.1 -0.1 -0.05 0 0.05 0.1 Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Source: Authors’ calculation using LFS 2016. Note: Dot indicates the point estimate of the coeȂicient for gender (femaleۙ1) in a log-hourly earnings regression controlling for district fixed eȂects, household characteristics, age, education, and type of employment (wage, daily, self). The lines indicate the 95 percent confidence interval. 64 SolotaroȂ et al (2019). 59 BANGLADESH RURAL INCOME DIAGNOSTICS Location shapes rural income inequalities Incomes in Western divisions still depend more on agriculture The importance of agricultural income varies more across space than between poor and richer households. Labor incomes in Rangpur, Khulna, and Rajshahi still rely heavily on agriculture, compared to the national average. About 60 percent of average labor incomes in Rangpur, Khulna, and Rajshahi are derived from agri- culture (Figure 1.25). This contrasts with the Eastern divisions, where Chittagong marks the other end of the spectrum, with on average just 37 percent of rural labor income coming from agriculture. Greater engagement in agriculture in Western Bangladesh has tied poverty reduction there more closely to the performance of the agricultural sector than in the country’s Eastern divisions.65 Figure 1.25. Agricultural income as a share of total rural household labor income, averages by division, 2018 100% 90% 80% 39% 41% 43% 49% 52% 70% 57% 63% 60% 50% 40% 30% 61% 59% 57% 51% 48% 20% 43% 37% 10% 0% Rangpur Khulna Rajshahi Dhaka Barisal Sylhet Chittagong Agriculture Non Ag Source: Authors’ calculations using BIHS 2018. Western areas and poorer areas in the East have high diversi‫ޔ‬cation potential that could open a promising path to raise agricultural incomes in these areas. Diversi‫ޔ‬cation potential into non-rice crops varies across divisions, being more limited in Sylhet, northern Dhaka (Mymensingh), and Barisal divisions. In these areas, the proportion of low-lying wetland (haor), which remains under water for 65 Hill and Genoni (2019). 60 ASSESSING CONTEXT AND HETEROGENEITY much of a year, is relatively high. With favorable agro-ecological conditions and good irrigation coverage, districts in Western areas have higher agricultural diver- si‫ޔ‬cation potential. An area-based diversi‫ޔ‬cation measure, known as the Simpson Index,66 shows that the diversi‫ޔ‬cation of agricultural production systems is high among districts in the Western divisions, along with Dhaka division from the East region (Figure 1.26). While Rangpur division has even greater availability of high- lands than Rajshahi and similar (or lower) exposure to droughts, it is less diverse compared to Rajshahi, primarily due to lower irrigation coverage (Figure 1.27). The low degree of diversi‫ޔ‬cation among districts in Sylhet, Mymensingh, and Barisal divisions re‫ޕ‬ect their large shares of lowland areas, where cropping intensity is also low. Sylhet consistently shows the highest share of farmers growing only rice (Figure 1.27). Most districts experienced positive changes in crop diversi‫ޔ‬cation between 2014 and 2018. The shiȅ toward a more diversi‫ޔ‬ed production system was most prominent among coastal districts of Khulna and Barisal.67 Figure 1.26. Simpson Diversification Index by districts, 2018 (0.65-0.78] (0.50-0.65] (0.37-0.50] (0.28-0.37] (0.08-0.28] Source: Ahmed et al (2021). Note: Darker shading indicates greater diversification. 66 The Simpson diversi‫ޔ‬cation index is calculated as SIDۙ1 – , where Pi is the proportionate area (or production value for a value-based index) of the ith crop in gross cropped area (total production value). 67 Ahmed et al (2021). 61 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.27. Percentage of farmers growing rice and other crops, across divisions, 2018 14% 37% 39% 49% 51% 47% 59% 85% 54% 57% 46% 45% 49% 38% 10% 4% 5% 3% 4% 4% 1% Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Only non rice crops Only rice Rice and other crops Source: Authors’ calculations using BIHS 2018. The spatial concentration of oȂ-farm manufacturing activities around Dhaka and Chittagong in‫ޕ‬uences rural income opportunities country-wide Economic transformation in Bangladesh has been accompanied by rapid urban- ization.68 Since 2000, urban population growth has exceeded 3 percent per year while rural population growth has been negative since 2014. The growth in the urban population has been faster than in most other countries in the South Asia region, with an average growth rate of 3.9 percent annually in Bangladesh since 2000, compared to 2.7 percent for South Asia.69 As the country urbanizes, an increasing share of households living in rural areas report working in urban centers. In 2016, about 20 percent of workers liv- ing in rural areas reported working in an urban area (19 percent for the poorest quintile and about 26 percent for the richest quintile). Between 2010 and 2016, 68 Note that there is no sharp separation between urban and rural areas. This increase could also re‫ޕ‬ect more urbanization and not necessarily a change in the location of employment sources. 69 United Nations (2018). 62 ASSESSING CONTEXT AND HETEROGENEITY the share of rural workers commuting increased from 10 to 19 percent.70 Workers who commute get higher wages, indicating that there is a higher return for those working in urban jobs. The median wage of a commuter is 9,000 Taka, compared to 6,000 Taka for the non-commuter. Conditional on other worker characteristics, those who commute receive about 33 percent higher wages. Among workers who commuted in 2016, only 8.8 percent were women. A simple regression analysis controlling for individual and household characteristics suggests that women are 38 percent less likely than men to commute for work. Mobility constraints aȂect- ing women may explain part of these patterns. More educated workers are also more likely to commute, conditional on other characteristics. The rapid increases in rural-urban commuting since 2010 suggest that this pat- tern is not only due to fuzzy urban-rural distinctions but may re‫ޕ‬ect increased oȂ-farm labor opportunities in the urban periphery.71 A decomposition analysis of the sectors driving these changes indicates that about 37 percent of the rise is due to increased employment in manufacturing (garments and textiles) and another 31 percent to various types of services. These trends in commuting are consistent with evidence of a recent shiȅ of ‫ޔ‬rms toward the urban periphery, which could re‫ޕ‬ect responses to the relative advantage of locating in rural or urban areas.72 Commuting has been increasing across the country, again suggesting growing opportunities in oȂ-farm activities linked to urban areas. However, rural work- ers in Dhaka and Chittagong divisions are more likely to bene‫ޔ‬t from this process than those in other divisions (Figure 1.28). Rural workers in close proximity to the large metropolitan areas of Dhaka and Chittagong comprise about 50 percent of all commuters. While in Dhaka about 1 in 3 rural workers have jobs in urban areas, this is true for about 1 in 10 rural workers in Khulna and Sylhet. More generally, varying degrees of urbanization across districts strongly shape employment opportunities and poverty reduction. District-level panel data for the period 2000-2016 show a positive correlation between districts that have urbanized more rapidly and faster consumption growth. This correlation holds even when controlling for initial district-level conditions, as well as changes over time in district-level factors such as access to services, connectivity, and population.73 70 HIES 2016. 71 Asfar (1999); Rahman (2014); World Bank (2015a). 72 Farole and Cho (2017). 73 Authors' background analysis can be shared upon request. 63 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.28. Share of rural residents working in urban areas a. By division in 2010 and 2016 b. By district in 2016 35% 30% 25% 20% 15% 10% 5% 0% al g a na i et r ah pu on ak ris lh Bay of Bengal ul jsh ng Dh ag Sy Ba Kh Ra Ra 0.63 – 0.78 itt Ch 0.48 – 0.63 0.32 – 0.48 2010 2016 0.17 – 0.32 0.02 – 0.17 Source: Authors’ calculations based on HIES 2010 and 2016. Note: “Commuter” refers to workers who reside in a rural area but report working in an urban area. Bangladesh’s urbanization and economic activity development have been largely concentrated around Dhaka and Chittagong. The Dhaka-Chittagong cor- ridor is the backbone of the export-oriented manufacturing sector in Bangladesh. As urbanization continues, a natural next step is reallocation of economic activity, particularly manufacturing, to peri-urban areas of large cities and then to sec- ondary cities and adjacent rural areas.74 However, compared to other urbanizing countries, Bangladesh’s secondary cities remain disproportionally small com- pared to Dhaka and Chittagong.75 Sectors that serve the internal market, such as food processing, show a more even distribution of ‫ޔ‬rms across the country. Figure 1.29 contrasts the distribution of RMG and food processing non-microenterprises across the country’s upazilas. The location of RMG ‫ޔ‬rms highlights the concentration of activity around Dhaka. This is also observed for other export-oriented ‫ޔ‬rms in textiles and light manufacturing. 74 Bird et al. (2018). 75 Muzzini and Aparicio (2013). 64 ASSESSING CONTEXT AND HETEROGENEITY In contrast, the distribution of food processing ‫ޔ‬rms is widespread, with a high concentration of ‫ޔ‬rms in the Western divisions. The location re‫ޕ‬ects the proximity to agricultural production. The largest concentration of non-microenterprises in food processing is outside the city of Bogra (in Rajshahi division), and the largest concentration of microenterprises is outside of Rajshahi.76 Figure 1.29. Distribution of non-microenterprises by upazila, 2013 a. RMG b. Food procesing Apparel firms w/ 10+ emp Food firms w/ 10+ emp Established & Permanent Established & Permanent < 50 1 - 100 51 - 100 101 - 200 101 - 500 201 - 500 501 - 3968 501 - 1000 Kurigram nodata 1001 - 2331 Kurigram nodata Source: Farole and Cho (2017) using the Economic Census 2013. Note: RMG = readymade garments. Migration patterns add to spatial disparities in rural incomes The COVID-19 crisis underscores how, in Bangladesh, rural income dynam- ics are tied to income patterns in urban areas. Domestic migration is one important link in this association. At the onset of the COVID-19 crisis, the lockdown propelled a large movement of people out of Dhaka, including many economic migrants returning to their origin households (Figure 1.30). 76 Farole and Cho (2017). 65 BANGLADESH RURAL INCOME DIAGNOSTICS Results from a representative phone survey in poor urban areas of Dhaka and Chittagong indicate that about 10 percent of respondents changed their place of residence following the start of the crisis and the lockdown that began in late March 2020.77 Figure 1.30. COVID-19 urban-to-rural population movements across Bangladesh, March to May 2020 Percentage change in FB users -75+ 0 75+ Source: Analysis of data based on Facebook Data for Good. Note: he map shows the percentage change in mean (anonymized) Facebook users per 600m x 600m grid cell, compared to the 45 day mean prior to March 23, 2020. Users’ locations are based on the mean recorded location of each user activating Facebook from 1600-2400 UTC on the selected day. 77 Genoni et al., (2020). 66 ASSESSING CONTEXT AND HETEROGENEITY A large share of rural households receive income from remittances (domestic and international). In 2016, about 6 percent of rural Bangladeshi households reported receiving international remittances, while 14 percent received domestic remittances (Table 1.3). Based on household reporting, median per capita remit- tance income per month was 1,389 Takas in 2018.78 In terms of international remit- tances, Chittagong and Sylhet show the highest share of recipient households. For domestic remittances, the share of recipient households does not vary sig- ni‫ޔ‬cantly across divisions, except in Sylhet. Overall, Western divisions are signi‫ޔ‬- cantly less likely to receive international remittances. Compared to countries like India or Vietnam, Bangladesh has a higher share of the rural population receiv- ing domestic remittances. According to FINDEX, in 2017, 22 percent of the rural Bangladeshi population older than 15 years received domestic remittances, while in India and Vietnam this share was 17 percent. Poor households are less likely to receive remittances, but income from migration is pivotal for poverty reduction. At ‫ޔ‬rst glance, Table 1.3 suggests that poor households are less likely to receive international remittances. In addition, remittance amounts are signi‫ޔ‬cantly lower for the poorest quintiles (In 2018, median per capita monthly remittance income was 667 Takas for the poorest quintile compared to 4,375 takas for the richest quintile). However, a decomposition of annual per capita income growth (4.2 percent) between 2000 and 2014, based on the 62-village panel survey, shows that remittances contributed signi‫ޔ‬cantly to the rise in incomes (2.2 percent). This implies that, while poor households may have lower odds of receiving remittances, receipt of remittances increases the odds of moving out of poverty. In addition, remit- tances support incomes and poverty reduction indirectly by boosting local economies. Spending on local goods and services with remittance income boosts local demand, and wages can increase as local labor supply is reduced. From 2000 to 2016, poverty reduction was faster in districts where international migration was higher: for each additional 0.1 percent of a district’s population migrating internationally, poverty in that district fell by 1.7 percent.79 Using BIHS panel data and matching/IV methods, Tillan et al. (forthcoming) ‫ޔ‬nd that rural households that had a member migrating experienced faster consumption and income growth. Sen et al. (2014) ‫ޔ‬nd that diȂerential rates of urbanization and international migration can help explain the spatial pattern of poverty reduction across districts in Bangladesh. 78 BIHS 2018. 79 Hill and Endara (2019a). 67 BANGLADESH RURAL INCOME DIAGNOSTICS Table 1.3. Percentage of households receiving international or domestic remittances, 2016 International remittances Domestic remittances All Poorest 40% All Poorest 40% Barisal 3% 2% 16% 18% Chittagong 10% 7% 16% 11% Dhaka 7% 3% 14% 13% Khulna 3% 3% 13% 11% Rajshahi 2% 0% 17% 14% Rangpur 1% 1% 13% 12% Sylhet 11% 4% 6% 5% Rural Bangladesh 6% 3% 14% 12% Source: Authors’ calculations using HIES 2016. The geographic pattern of migration constrains its potential to reduce coun- try-wide income disparities. Dhaka alone absorbs 60 percent of all recent migrants in Bangladesh and has the largest share of all types of migrant ‫ޕ‬ows. Chittagong absorbs another 16 percent of migrants.80 While Rangpur com- prises about 12 percent of the rural population, only 4 percent of migrants come from that division (Figure 1.31). The majority of migration is concentrated in the Eastern areas of the country around Dhaka and Chittagong (Figure 1.32), and this has changed little over time. Greater Dhaka is the main destination for migrants from Barisal, Rangpur, Rajshahi, and Dhaka divisions (Figure 1.33). Chittagong and Khulna divisions show slightly greater variation in migration des- tinations, but greater Dhaka still concentrates about 50 percent of migrants. In Sylhet, most migration is within the division. The temporal persistence of these geographic patterns is consistent with the literature showing that migration is more likely in places where the stock of migrants is already high, as prospective migrants can rely on existing migrant networks.81 80 HIES (2016). 81 Hanson (2010); Litch‫ޔ‬eld et al. (2015). 68 ASSESSING CONTEXT AND HETEROGENEITY Lower overall migration in poorer areas of the country may be linked to higher risks and costs of undertaking this opportunity. The literature on internal migra- tion for Bangladesh indicates that internal migration largely re‫ޕ‬ects the pursuit of better economic opportunities and is strongly linked to the country’s urbaniza- tion.82 In some areas, seasonal migration is also important as a way to cope with weather shocks. According to HIES 2016, 14 percent of households experiencing an income shock (largely due to weather) used migration as a coping strategy. Experimental evidence in poor areas of Western Bangladesh shows that migration is pro‫ޔ‬table and may be an underutilized strategy to smooth incomes for poor households. Qualitative information suggests that incomes close to subsistence and lack of knowledge about returns to migration contribute to low participa- tion in migration. Credit constraints may be another important factor preventing households from migrating.83 Figure 1.31. Spatial pattern of rural out-migration, by division, 2010 and 2016 10 11 12 4 5 6 13 11 26 26 43 22 7 6 % of total population % of total rural out-migrants Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Source: Authors’ calculations using HIES 2016. Note: Percentage of rural households reporting a recent domestic or international migrant, by the division where the household was interviewed. “Population” bar indicates the share relative to the total population. “Migrants” bar reports the areas’ respective shares of the total migrant population. 82 UNDP (2013). 83 Bryan, Chowdhury, and Mobarak (2014). 69 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.32. Share of households sending migrants, by district, 2016 a. Domestic b. International 0.14 – 0.17 0.31 – 0.38 0.10 – 0.14 Bay of Bengal 0.23 – 0.31 Bay of Bengal 0.07 – 0.10 0.15 – 0.23 0.03 – 0.07 0.08 – 0.15 0.00 – 0.03 0.00 – 0.08 Source: Authors’ calculations using HIES 2016 for both urban and rural areas. Figure 1.33. Main destination district of migrants by division of origin, 2016 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Barisal Rangpur Dhaka Rajshahi Chittagong Khulna Sylhet Greater Dhaka Barisal Chittagong Comilla Khulna Mymensingh Rajshahi Rangpur Sylhet Habiganj Maulvibazar Sunamganj Others Source: Authors’ calculations using HIES 2016. Note: Greater Dhaka includes Dhaka, Gazipur, Manikgabj, Munshiganj, Narayanganj, Narsingdi. 70 ASSESSING CONTEXT AND HETEROGENEITY Rural income levers: household assets and markets This section examines assets and markets whose availability and quality aȂect rural households’ earning capacities. Key household assets that have been shown to in‫ޕ‬uence earnings in rural areas include education and land. Markets important for rural incomes include farm-product, credit, and insurance markets. In the following pages, analyzing these and other strategic productive assets and markets gives a clearer picture of how people in rural Bangladesh earn their incomes, why rural incomes diȂer across regions and population groups, and what impedes rural households’ eȂorts to improve their earnings. Results will inform the assessment and prioritization of opportunities to boost rural incomes through policy action—the heart of the RID. A stylized assets framework guides our analysis of assets and markets in rural Bangladesh.84 Education and earning power: literacy gaps hamper progress in the rural West Despite progress, the rural adult population remains poorly educated. In 2000, about 64 percent of the rural adult populaltion was illiterate, de‫ޔ‬ned as not being able to write a letter. In 2016, this was true for 50 percent of rural adults. When focusing on the poor, literacy rates increased by 14 percentage points between 2000 and 2016. However, in 2016, 6 out of 10 Bangladeshi adults living in poor rural households remained iliterate. Illiteracy rates for male farmers are particu- larly high, ranging between 70 and 75 percent.85 Educational disparities across areas remain large, and this partly determines income opportunities. Sylhet, Rajshasi, and Rangpur have literacy rates below the rural average, consistent with the higher poverty levels previously noted for districts in these divisions. Progress in the accumulation of education acceler- ated aȅer 2010 across divisons and for the poor. However, as discussed above, progress in educational attainment has been slower in the rural West, slowing income growth in those areas. While urban literacy rates increased by 15 per- centage points nationally between 2010 and 2016, literacy rates increased by only 2 percentage points in Rangpur, 5 percentage points in Rajshahi, and 8 per- centage points in Khulna. Sylhet also showed limited progress, with a rise of just 7 percentage points (Figure 1.34). In 2018, about a quarter of the rural pop- ulation older than six years had no formal education. The districts of Jamalpur (38 percent) and Gaibandha, Sherpur, and Kurigram (33 percent each) were the 84 Lopez-Calva and Bussolo (2014). 85 World Bank (2020a). 71 BANGLADESH RURAL INCOME DIAGNOSTICS weakest performers. Their situation contrasts with Patuakhali and Barguna, where only 7 percent of the population older than six years had no formal edu- cation (Figure 1.35). Figure 1.34. Literacy rates in rural areas, 2000-2016, by division a. All rural areas b. Rural poor 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 2000 2005 2010 2016 2000 2005 2010 2016 Barisal Chittagong Khulna Dhaka Rajshahi Rangpur Sylhet Notes: Authors’ calculations using HIES 2000-2016. Sample are adults aged 15 and older. Literate is defined as being able to write a letter. Figure 1.35. Percentage of the rural population older than 6 years with no education, 2018 a. Map b. Ranking 40% 35% 30% 25% 20% 15% 10% 5% Bay of Bengal 0% 0.32 – 0.38 Jamalpur Kurigram Bandarban Kishoreganj Sirajganj Rangpur Chuadanga Natore Khagrachhari Habiganj Brahmanbaria Naogaon Faridpur Sunamganj Gazipur Shariatpur Sylhet Feni Narayanganj Chittagong Pirojpur Barguna 0.25 – 0.32 0.19 – 0.25 0.13 – 0.19 0.07 – 0.13 Source: Authors’ calculations using BBS (2018). 72 ASSESSING CONTEXT AND HETEROGENEITY Adult women remain less educated than men. About 54 percent of adult women have not completed primary education, compared to 49 percent of men. In addi- tion, women are less likely to pursue higher levels of education. Only 11 percent of adult women in rural areas have completed secondary school or above, com- pared to 18 percent of males (Figure 1.36). Figure 1.36. Educational attainment by gender (percentage of adults older than 18), 2016 100% 11% 90% 18% 80% 23% 70% 20% 60% 12% 13% 50% 40% 30% 49% 54% 20% 10% 0% Male Female Less than primary complete Primary complete Some secondary Secondary complete or more Source: Authors' calculations using HIES 2016. ounger cohorts are signi‫ޔ‬cantly more educated, but the quality of education remains low. Figure 1.37 shows that only 6 percent of those 18-25 years old had no education in 2016, compared to 30 percent of those 25-45 and 60 percent of those 45-64 years old. This showcases the substantial improvement in education over time. However, the bulk of the youth workforce is still composed of individ- uals with relatively low skill levels: close to 1 in 3 youth aged 18-25 in rural areas are primary school completers or less. Recent analysis ‫ޔ‬nds that, by age 16, close to half of young individuals are already out of school. Most females drop out of the labor force as they exit formal schooling. While a child is expected to complete 73 BANGLADESH RURAL INCOME DIAGNOSTICS 10.2 years of school by her 18th birthday, when factoring in what children learn, the average child can expect to receive just 6 years of school.86 Access to skills training beyond formal education is limited. The percentage of the Bangladeshi workforce that has taken training is low overall, with about 2 percent of the overall workforce and of youth reporting that they have received training in the past 12 months. Data suggest that access to skills development is limited for individuals from lower-income households, youth (ages 15–24), and less-educated workers, regardless of their desire for training opportunities. Training opportunities tend to be concentrated on better-educated urban males.87 Apart from technical and vocational training, few other active labor market pro- grams exist in Bangladesh to facilitate access to jobs for disadvantaged popula- tions. Support instruments oȅen targeted to unemployed youth in other parts of the world include wage subsidies, apprenticeship and internship, counseling or mentoring, job search assistance, and employment services. These tools are largely unavailable in Bangladesh. Interventions to promote entrepreneurship and self-employment, which may be more applicable in an environment where labor demand is limited, are also lacking.88 Figure 1.37. Educational attainment by age group in rural Bangladesh, 2018 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 18-25 25-45 45-64 Older than 65 No education Less than primary Education Primary Complete Secondary Incomplete Secondary Complete Tertiary Education Source: Authors’ calculations using BIHS 2018. 86 World Bank, Human Capital Index 2020. 87 World Bank (2019d). 88 Farole and Cho (2017). 74 ASSESSING CONTEXT AND HETEROGENEITY Average land holdings are small, and land markets could work better In addition to education, land is another key asset for rural households in Bangladesh. About 63 percent of rural households own at least 0.05 acres of culti- vable land. The rest either own less or have neither cultivable nor operated land. On average, a rural household owns 0.80 acres and operates 0.82 acres of land.89 With 70 percent of agricultural land, Bangladesh holds the highest share of agricul- tural land as a percentage of total land area among comparators and is well above the 56 percent of the South Asian region. Indeed, Vietnam, Pakistan, and India have 39 percent, 47 percent, and 60 percent of agricultural land, respectively.90 There is a clear correlation between consumption levels and size of land owned, although overall land sizes are very small. Considering only households that report productive land, 79 percent owned that land, and 21 percent were land- less.91 About 63 percent of households own marginal or small areas, 7 percent medium-sized areas, and 8 percent large properties. In the poorest quintile, 4 in 10 households are landless, compared to 7 percent for the richest quintile (Figure 1.38). These patterns of land ownership have changed little in the past decade. Overall farm sizes in Bangladesh are among the smallest in the world on average, and there is limited room for expansion. 92 Figure 1.38. Land ownership by land size and consumption quintile, 2018 1% 3% 4% 3% 6% 9% 12% 6% 7% 7% 16% 9% 21% 12% 23% 24% 27% 41% 35% 41% 40% 39% 37% 35% 39% 29% 26% 24% 18% 7% 1 2 3 4 5 All Landless Marginal Small Medium Large Source: Authors’ calculations using BIHS 2018. Notes: Landless are those with land size=0; Marginal: land size>0 and <0.20 hectares; Small: land size>=0.2 and <=0.6 hectares; Medium: land size>=0.61 and <1.01; Large: land size>= 1.01 hectares. 89 BBS (2019). 90 WDI (2020). 91 Productive land excludes homestead land and other land not usable for production (BIHS 2018). 92 FAO (2020). 75 BANGLADESH RURAL INCOME DIAGNOSTICS The relationship between land ownership and consumption has become weaker in the Eastern divisions, signaling a faster structural transformation of those areas. Since 2010, the correlation (conditional on other factors) between land holdings and household per capita consumption has declined signi‫ޔ‬cantly for most households in the Eastern divisions (Figure 1.39a). This contrasts with Western areas, where this correlation has not changed across time (Figure 1.39b), re‫ޕ‬ecting the persistently strong engagement in agriculture in the West. As shown in Figure 1.25, Western divisions have a higher share of household income coming from agriculture than Eastern divisions. Figure 1.39. Looser ties between consumption and landholding —but only in the East a. Eastern Bangladesh b. Western Bangladesh 0.5 0.5 0.4 0.4 +! ,)!&+$,"ƛ& &"+10 +! ,)!&+$,"ƛ& &"+10 0.3 0.3 0.3 0.3 0.1 0.1 0 2 4 6 8 0 2 4 6 8 2005 2010 2016 Source: Hill and Endara (2019b). Note: Figures indicate how closely consumption levels were correlated with land ownership in 2005 (green), 2010 (blue), and 2016 (red). In Eastern Bangladesh, consumption became less tied to land holding over this period. In Western areas, this trend was much less pronounced. With Bangladesh’s rural population still growing, rural households appear to be working around the challenge of land fragmentation by establishing land leasing and tenancy relationships. Half of landless households rent land, mak- ing poor households relatively more engaged in land rental markets (Figure 1.40a). This re‫ޕ‬ects a longer-term upward trend in the share of cultivated land under tenancy and a noticeable rise in the share of landless tenancy.93 93 In 1988, only 23.4 percent of cultivated land was under tenancy in rural Bangladesh; the matched share increased to 32.8 percent in 2000, rising further to 39.8 percent in 2008 and 47.5 percent in 2014. While 43.6 percent of households rented land from others in 1988, the matched share rose to 54.2 percent in 2000 and 58.3 percent in 2008. In 1988, only 6.7 percent of total cultivated land was under landless tenancy; the matched share increased to 12.4 percent in 2000, rising further to 23.1 percent in 2014 (Hossain and Bayes 2009, 2018). 76 ASSESSING CONTEXT AND HETEROGENEITY About 1 in 4 farmers rent land using sharecropping agreements (Figure 1.41a). The rate is higher for the lowest quintile (35 percent of those renting in) and in Rangpur and Sylhet (about 1 in 3 households renting). Figure 1.40. Percentage of households renting land in and out, by land size and consumption quintile, 2018 a. Renting in b. Renting out 3% 3% 8% 4% 6% 8% 10% 13% 6% 10% 17% 19% 14% 14% 14% 19% 13% 14% 14% 33% 26% 14% 15% 36% 40% 37% 36% 35% 39% 43% 39% 37% 38% 38% 58% 49% 45% 46% 46% 39% 40% 33% 32% 33% 24% 27% 1 2 3 4 5 All 1 2 3 4 5 All Landless Marginal Small Medium Large Source: Authors’ calculations using BIHS 2018. Figure 1.41. Percentage of households renting with sharecropping agreements, 2018 a. By quintile b. By division 40% 40% 35% 35% 30% 30% 25% 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% 1 2 3 4 5 All al g Kh a Ra na Ra hi et r pu on ak ris a lh ul jsh ng Dh ag Sy Ba itt Ch Source: Author’s calculations using BIHS 2018. 77 BANGLADESH RURAL INCOME DIAGNOSTICS Land markets could work more eȂiciently to support higher-value invest- ments and economies of scale. Two important ‫ޔ‬ndings related to land are that rented land is signi‫ޔ‬cantly and positively associated with higher eȂiciency, and that land fragmentation is signi‫ޔ‬cantly and negatively associated with farming eȂiciency.94 Together, these results suggest that although land markets are func- tional, they remain ineȂicient. Better-functioning land and lease markets could allow more-eȂicient households to rent in land (or, perhaps equally important, allow ineȂicient households or absentee households to rent out land without fear of losing it), thereby reducing the eȂiciency premium currently associated with rented land. About 31 percent of households report not possessing legal docu- ments for agricultural land.95 It is likely that these households have little incentive to make investments that might enhance productive eȂiciency. Gender gaps in ownership and rights over land and other household assets per- sist. Among rural women, 13 percent solely or jointly own agricultural land and 7 percent own nonagricultural land, compared to 70 percent and 86 percent of rural men, respectively.96 Survey data also indicate that 96 percent of household land in rural areas is still owned by husbands alone.97 The gender gap in land ownership is primarily attributable to male-biased practices in inheritance, the main channel through which land is acquired. Women inheriting land face obstacles in register- ing and legally owning their inheritance. Barriers include procedural obstacles; women’s limited mobility to travel to government oȂices; and high fees for regis- tering land. To cover registration fees, women may have to request money from their husbands.98 Key agricultural inputs: irrigation, fertilizers, seeds, and mechanization Investment and reforms in the 19ч0s encouraged use of better inputs, promoted mechanization, and enabled improvements in irrigation. EȂorts at that time gave a large share of agricultural households access to irrigation, mechanization, and quality inputs such as seeds. These policies had a strong, positive impact on labor and land productivity in agriculture.99 94 Gautam and Faruqee (2016). 95 BBS (2018). 96 BIHS 2015. 97 Quisumbing, Kumar, and Behrman (2018). 98 SolotaroȂ et al. (2019). 99 Mandal (2014); Gautam and Faruqee (2016). 78 ASSESSING CONTEXT AND HETEROGENEITY Irrigation Today more than ч0 percent of agricultural land has access to water for irriga- tion on the edge of the ‫ޔ‬elds.100 Using a comparable de‫ޔ‬nition of irrigation, in 2016, 60 percent of Bangladesh’s agricultural land was irrigated, compared to 50 percent in Pakistan and 37 percent in India (last estimate 2013). Irrigation access is strongly associated with the level of diversi‫ޔ‬cation among marginal, small, and medium farmers.101 However, irrigation coverage is very low in hilly and highland divisions, such as Rangpur, due to constraints in availability and accessibility to water. While Rangpur division has even higher availability of highlands than Rajshahi and similar (or lesser) exposure to droughts, it is less diverse compared to Rajshahi primarily due to lower irrigation coverage (32 against 90 percent). In 2016, over 95 percent of 7.6 million hectares of irrigated lands were irrigated with tube wells and power pumps.102 Groundwater accounts for 75 percent of irrigation. However, groundwater storage is declining rapidly due to expansion of boro paddy. One estimate suggests a 32 percent decline in groundwater storage between 2003 and 2013.103 Several districts in the northwest region and around Dhaka city have reached critical points for groundwater depletion.104 Fertilizer use About 95 percent of farmers use fertilizers. However, technical analysis suggest that farmers are overusing them.105 Fertilizer use has increased steadily over the past three decades. From 2003 to 2016, fertilizer consumption jumped from 160 to 289 kg per hectare of arable land, which implies an average annual growth rate of 5 percent. This compares with a South Asia average of 160 kg in 2016. During the same 2003-2016 period, India, Vietnam, and Pakistan had a growth rate in fertilizer use of 4 percent, 2 percent and 3 percent, respectively.106 Signi‫ޔ‬cant subsidies dedicated to fertilizer use explain these excessive consumption rates. The strong focus of past policies on nitrogenous fertilizers led to an imbalance in the use of nitrogen relative to other nutrients. Subsidies were increased for potassium and phosphorous macronutrients in recent years to address concerns about resource degradation resulting from an overuse of nitrogen. 100 Gautam and Faruqee (2016). 101 Haque and Ahmed (2020). 102 Pumping using electricity or solar power (not diesel). 103 Khaki et al. (2018). 104 BPC (2018). 105 Gautam and Faruqee (2019) and World Bank (2020a). 106 WDI (2020). 79 BANGLADESH RURAL INCOME DIAGNOSTICS Evidence suggests that almost half of farmers are overusing fertilizer with appar- ently no additional output. Conversely, the overuse of fertilizer could have neg- ative environmental, health, and even farm productivity consequences through resource degradation.107 Seeds Improved seeds have driven past productivity, but quality seeds are now in short supply. Seeds embody technological change and are at the core of the pro- ductivity agenda. Improved varieties have driven past success in raising produc- tivity and achieving food security.108 However, the formal seeds sector supplies only 20 percent of total demand.109 The remaining share is covered by farmers’ own seeds (88 percent) and smuggled seeds from India (12 percent).110 Supply gaps for formal seeds vary from crop to crop. For most crops, seed de‫ޔ‬cits are not caused by limited production capacity but rather dissemination constraints. In the potato seed supply chain, for instance, inadequate basic infrastructure (roads, electricity) and ‫ޕ‬awed logistics system prevent timely dissemination of seeds.111 On the contrary, horticulture seed supply is constrained mainly due to farmers’ limited access to ‫ޔ‬nance, extension, and markets.112 At present, both public and private sectors are active in the seed industry. The public seed sector’s focus is on noti‫ޔ‬ed crops, whereas the private sector cov- ers above 90 percent of hybrid seeds for two noti‫ޔ‬ed crops (paddy and maize) as well as improved and hybrid varieties of vegetables (54 percent), pulses, and legumes.113 As of 2013, there were around 200 seed companies with 17,000 regis- tered and 50,000 mobile seed vendors with market value of over US$125 million. Most of these companies enjoy good access to international seed technologies through partnership with major international seed companies and seed exports/ imports. However, only two companies - BRAC and Lal Teer Seed (LTS) - have com- plete seed value chains from research and breeding to extension. Private com- panies joined together to form the Bangladesh Seed Association (BSA), through which they collaborate with the government on seed policy and regulations. 107 Gautam and Faruqee (2016). 108 Gautam and Faruqee (2016). 109 World Bank (2020a). 110 BPC (2018); USAID (2014); BRAC (2019); and LTS (2019). 111 World Bank (2020a). 112 LTS (2019). 113 USAID (2014). 80 ASSESSING CONTEXT AND HETEROGENEITY Mechanization Reforms in the late 19ч0s supported the expansion of mechanization,114 exerting a strong, positive impact on labor and land productivity in agriculture. This allowed adopter farmers to release labor for other (mostly nonfarm) activities and diversify their incomes, while becoming more productive and eȂicient and achieving higher gross farm margins.115 About 85 percent of land preparation is done mainly by power tillers and to some extent by tractors.116 The shortage of family labor and access to electricity have been important drivers of farm mechanization, along with proximity to infrastructure and services.117 Among comparator countries, Bangladesh has had the fastest growth of mechanization, measured by the stock of farm machinery in 40-CV tractor equivalents. In Bangladesh, mechanization grew at an average annual rate of 22 percent between 1980 and 2015, while in Vietnam, India, and Pakistan, the average growth rates were 12, 9, and 5 percent, respectively. 118 Today most farmers use some type of mechanization. About 96 percent of rural farmers use some from mechanization (BIHS), and there has been a rapid con- vergence in mechanization across areas, with Barisal and Sylhet catching up to other divisions. Today, the high use of mechanization can also be seen across all districts. However, room for improvement remains. Agricultural mechanization is still used in an unbalanced way, largely for planting, threshing (rice), and process- ing.119 The use of mechanized operations in non-crop agriculture (e.g., horticul- ture and livestock) remains very limited. This imbalance disproportionally aȂects female farmers, who dominate in post-harvest processing of crops.120 Electricity access has fueled consumption growth and poverty reduction, but gaps persist Recently, rural electricity access in Bangladesh has expanded substantially, with Khulna showing an especially impressive performance. Between 2000 and 2016, the share of rural households with access to electricity increased sharply, and gains have been faster since 2010 (Figure 1.42). Khulna, for instance, showed an impressive gain in coverage for its whole population, including the poor. 114 Mandal (2014). 115 Gautam and Faruqee (2016). 116 Tiwari et al. (2017). 117 Ahmed and Gautam (2015). 118 According to data from Fuglie (2012 and 2015). 119 Hossen (2019). 120 Alam and Khan (2017). 81 BANGLADESH RURAL INCOME DIAGNOSTICS In 2000, only 7 percent of Khulna’s poor households had electricity, but by 2016 that share had climbed to about 71 percent, well above the national average for the rural poor. Dhaka accounts for about 25 percent of rural households without electricity, followed by Rangpur with 21 percent. More recent data from the 2018 rural agricultural survey point to continued expansion in electricity coverage, with 82 percent of rural households using electricity as their main source of lighting. Kerosene and solar power are other sources of energy, with 10 and 8 percent of rural households relying on those sources, respectively. These gains in electricity access were associated with faster income and therefore consumption growth between 2010 and 2016. This indicates that closing remaining electricity gaps can support faster income growth. Figure 1.42. Rural access to electricity, 2000-2016, by division a. All rural areas b. Rural poor 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 2000 2005 2010 2016 2000 2005 2010 2016 Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Source: Authors’ calculations using HIES 2000-2016. Access to electricity still varies geographically, with Barisal and Rangpur show- ing the largest gaps. Barisal has a much lower share of electricity use (65 percent) compared to other divisions but complements largely with solar energy (26 percent) and kerosene (9 percent). In contrast, Rangpur also reports a comparatively low share of households using electricity (73 percent) but largely relies on kerosene to compensate (21 percent), with only 6 percent of households using solar. Figure 1.43 shows the variation across districts in the use of electricity and other power sources. 82 ASSESSING CONTEXT AND HETEROGENEITY Poorer performers, with high rates of kerosene use at the expense of electricity or solar, are Khagrachhari, Thakurgaon, and Bandarban districts, where about 28 percent of households rely on kerosene. Use of solar power for agriculture is lim- ited. As noted, spatial patterns of access to electricity are correlated with higher incomes and lower poverty. Figure 1.43. Energy sources used by rural households, 2018 a. Percentage of households using b. Energy sources across districts kerosene 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Kushtia Comilla Joypurhat Jessore Chittagong Sylhet Jhalokati Natore Bagerhat Dinajpur Cox's Bazar Lalmonirhat Thakurgaon Pirojpur Barguna Khagrachhari Bay of Bengal 0.24 – 0.29 0.18 – 0.24 0.12 – 0.18 0.06 – 0.12 0.01 – 0.06 Electricity Solar Kerosene Source: Authors’ calculations using BBS (2018). Connectivity challenges go beyond infrastructure There is a strong link between better accessibility to markets and urban centers and higher incomes. Figure 1.44 shows metrics of market accessibility and clearly highlights the disconnect between the country’s periphery and its dominant eco- nomic centers (the Dhaka-Chittagong corridor). These maps correlate strongly with the location of a large share of the poor. Evidence based on regression anal- ysis also points to a strong link between accessibility to urban areas and higher incomes and lower poverty.121 121 Gautam and Faruqee (2016); Blankespoor and Yoshida (2010). 83 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 1.44. Market accessibility across Bangladesh in 2021 km 0 100 Major markets (cities) Market access levels by 2021 pop Admins Highest market access 121,439 - 250,000 Divisions 2nd quintile 250,000 - 500,000 Districts (Zilas) 3rd quintile 500,000 - 3,000,000 4th quintile 3,000,000 - 9,799,517 Lowest market access Source: Robert Steven Banick’s calculations using road network and population projections data. Note: The map shows each union’s cumulative access to all markets, represented by Bangladesh’s 43 recognized City and Municipal Corporations. Each market’s accessibility to a given union is measured by weighting its projected population size in 2021 by the inverse of travel times to that market from the union (e.g., a gravity model). These 43 measures are summarized per union for a cumulative index of market accessibility. Past improvements in physical connectivity in Bangladesh have been linked to faster income growth, higher productivity in agriculture, and increased oȂ-farm income opportunities.Forinstance,theJamunabridgehascontributedtointegratingthecoun- try’s northwestern areas by promoting diversi‫ޔ‬cation into high-value crops, improving access to credit, and increasing market participation for both tenants and owners.122 122 Bayes (2007). 84 ASSESSING CONTEXT AND HETEROGENEITY The construction of the bridge facilitated a farm to non-farm employment shiȅ, along with lower household unemployment, in two adjacent districts connected by the bridge.123 There is also evidence of a shiȅ toward non-agricultural activities, especially in trade and transport.124 The Jamuna bridge has reduced transport costs and improved access to key consumption centers, as well as access to basic services such as natural gas, electricity, and telecommunications.125 Poor inland connectivity remains an important barrier to rural income growth.126 Road density in Bangladesh is low: a mere 0.13 kilometers per 1,000 population for highways and district roads, and 1.9 kilometers per 1,000 population if rural roads are included. These ‫ޔ‬gures compare poorly with road density in Pakistan (1.5 kilometers per 1,000 population), India (3.5 kilometers per 1,000 population), Sri Lanka (5.5 kilometers per 1,000 population) and Bhutan (9.7 kilometers per 1,000 population).127 Bangladesh is exposed to recurrent ‫ޕ‬oods that can cause severe disruptions in the transport network. According to Multi-Hazard Risk Assessment reports by the Bangladesh Ministry of Disaster and Relief, more than 50 percent of all road types are exposed to diȂerent levels of ‫ޕ‬ooding (Figure 1.45) increasing the relevance of road maintenance. In addition to the poor quality of the road network, important challenges include limited to nonexistent multimodality in the transport system and an uncompetitive logistic service market.128 129 123 Mahmud and Sawada (2014). 124 Bayes (2007). 125 Hossain, Sen, and Yasuyuki (2013). 126 Only 27 percent of the country’s rural roads and 40 percent of main roads are paved, and only half of paved roads are in good condition. Many parts of the regional road network are narrow and oper- ating at or near maximum capacity. Many sections are incapable of handling more freight vehicles. Congestion on roads doubles standard trucking costs, while road crashes alone account for about 11 percent of truck operating costs in Bangladesh (Herrera Dappe et al. 2020). 127 ŶĚƌĞƐĞƚĂů͘;2013Ϳ͘ 128 Herrera Dappe et al. (2020). The core infrastructure for all modes of transport is in place, but intermodal facilities are lacking. About 91 percent of shippers surveyed reported few to virtually no instances in which logistics service providers use multimodal transport to reduce logistics costs. Shippers typically work with more than one service provider to complete each shipment. Trucking and inland water shipping operations are provided by thousands of agents, many of them using very small ‫ޕ‬eets or single units. Drivers’ unions and associations set prices and control access to cargo, inter- fering with the market mechanism and preventing full competition. This weakens incentives among service providers, as they are not rewarded for the quality of their services. 129 Road transport rates in Bangladesh are high compared to other countries (Herrera Dappe et al. 2020). For instance, facilitation payments, fuel, insurance, registration, taxes, and tolls cost 2 percent- age points more in Bangladesh than in Vietnam. Repair, maintenance, and tires cost 5 percentage points more in Bangladesh than in Vietnam. Capital costs, administrative costs, and employment costs represent 4, 3, and 2 percentage points more in Vietnam than in Bangladesh. 85 BANGLADESH RURAL INCOME DIAGNOSTICS Connectivity costs particularly complicate fresh-product value chains and increase transaction costs that may disincentivize farmers’ diversi‫ޔ‬cation out of cereals. Transportation diȂiculties limit supply chains geographically.130 Direct and indirect logistics costs are high in most sectors and are particularly large for horticulture. Transportation costs comprise the majority of direct logistic costs for the dairy sector. Low connectivity in the Western areas of Bangladesh limits the ability of farmers to diversify, even if diversi‫ޔ‬cation potential in the region is higher. Figure 1.45. Flood exposure of Bangladesh’s road networks km 0 50 100 200 Not flood prone Moderate tidal surge Low flash flooding Severe flash flooding Low river flooding Severe river flooding Moderate flash flooding Severe tidal surge Moderate river flooding Source: TU Delȅ (2018). Small physical distances suggest that reducing congestion and logistic costs can have a signi‫ޔ‬cant impact on connectivity, thereby increasing rural farm incomes and expanding access to oȂ-farm labor opportunities. Despite challenges in con- nectivity, local markets are still physically close for most Bangladeshi farmers. 130 Gautam and Faruqee (2016). 86 ASSESSING CONTEXT AND HETEROGENEITY About 36.5 percent of rural households live within 2 kilometers of the hat-bazar (an open-air market), another 37.5 percent are between 2 and 4 kilometers, and 15 per- cent are between 4 and 5 kilometers (Figure 1.46). Although this metric does not re‫ޕ‬ect the actual extent of connectivity, it shows that physical distances are not large and suggests that transport costs are high primarily because of congestion and the suboptimal quality of infrastructure. It is estimated that reducing congestion on the national highways would cut logistics costs as a share of sales by 0.5 percentage points, increase exports by 3.7 percent, lead workers to migrate to and around Dhaka, thus concentrating employment there, and increase real wages in all districts.131 Figure 1.46. Physical distance to hat-bazar a. Percentage of households where b. Average distances by district hat-bazar is further than 5km 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Brahmanbaria Noakhali Tangail Kushtia Pirojpur Khulna Madaripur Bagerhat Netrakona Lakshmipur Dhaka Cox's Bazar Narail Munshiganj Bandarban Panchagarh Shariatpur Maulvibazar Rangamati Natore Pabna Bogra Bay of Bengal 0.37 – 0.46 0.28 – 0.37 0.18 – 0.28 0.09 – 0.18 0.00 – 0.09 Below 2km 2-4 km 4-5 km 5km or more Source: Authors’ calculations using BBS (2018). 131 These estimates are based on a spatial equilibrium model for Bangladesh used to simulate the poten- tial impacts that logistics interventions would have on prices, wages, and employment. The model has three building blocks—geography, economic activity, and workers—and is calibrated using survey and geo-coded transport data. Prices and wages adjust to clear the market following an intervention. In the present example, the introduction of a 40 km per hour minimum travel speed on national highways would reduce congestion, thus decreasing logistics costs (Herrera Dappe et al. 2020). 87 BANGLADESH RURAL INCOME DIAGNOSTICS Mobile phone coverage has expanded rapidly in rural Bangladesh, providing an alternative way to enhance connectivity, but access to broadband connectivity is still very limited. In 2000, only 6 percent of rural households reported owning a mobile phone, and 0.5 percent had internet and email access. By 2016, these ‫ޔ‬gures had reached 90 percent and 5 percent, respectively. This dramatic increase in access was made possible by signi‫ޔ‬cant declines in user costs. Since 2010, the cost of inter- net-enabled handsets has dropped signi‫ޔ‬cantly, and competition among inter- net-service providers has driven prices down, making it more aȂordable to connect. However, though 3G networks now cover 93 percent of the Bangladeshi population, mobile internet uptake is still low, at 21 percent in 2017.132 A combination of demand and supply factors, including low literacy and access, help explain these patterns.133 Citizens’ broadband and smartphone penetration rates also remain relatively low, with only 20 percent of Bangladeshis subscribed to mobile internet and 31 percent of total connections being made using smartphones.134 Comparing the number of secure internet servers per million people, Bangladesh is below the regional aver- age and other comparator countries. Indeed, while in Bangladesh there are only 100 secure internet servers per million people, Vietnam, India, and the South Asia region as a whole have 2,596, 389, and 313 servers respectively (WDI). Gender gaps in mobility and mobile phone use also constrain women’s oppor- tunities to earn higher incomes. As will be discussed shortly, social norms related to women’s mobility increase their likelihood of working near home, reduce their participation in the commercialization of agriculture, and limit their access to oȂ- farm labor opportunities. Lower use of mobile phones by women may also con- strain their ability to access mobile banking and other digital resources. Commercialization and ‫ޔ‬nancial markets In this sub-section, we describe the current situation in rural Bangladesh in terms of commercialization of farm products and ‫ޔ‬nancial markets. Evidence on the functioning of these key markets, in addition to land markets, is central for bet- ter understanding the reasons behind the income patterns seen across space and gender, as well as potential constraints to income growth. 132 GSMA (2018). 133 On the Global Connectivity Index (GCI), Bangladesh improved its ranking from 76th among 79 countries in 2018 to 73rd in 2019. This progress re‫ޕ‬ects signi‫ޔ‬cant growth in smartphone demand and mobile broadband subscriptions over the years. User experience has also improved, with more aȂord- able mobile and ‫ޔ‬xed broadband, better cybersecurity, and cloud services. However, slow progress has been reported in overall ICT investment, which is essential to build capacity. 134 World Bank (2019a). 88 ASSESSING CONTEXT AND HETEROGENEITY ost farmers commerciali5e their products through eƙicient but informal markets Commercialization of farm production is high, with about 7 in 10 households selling part of their produce (Figure 1.47). For those households that sell, about 41 percent of the production is sold. For those households not selling their pro- duction, the main use is own consumption (38 percent), followed by storage (31 percent), animal feed (15 percent), and sharecropping (10 percent). There is very little diȂerence across consumption quintiles in terms of the percentage of households selling some production. However, the share of the production sold increases with the consumption quintile (35 percent for the poorest quintile and 46 percent in the richest quintile). Across divisions, Rajshahi and Rangpur have the highest shares of households selling, while Sylhet has the lowest. Figure 1.47. Percentage of households selling their farm production, 2018 a. Across consumption quintiles b. Across divisions 100% 100% 90% 90% 17% 19% 28% 27% 24% 27% 29% 24% 25% 27% 80% 80% 47% 41% 51% 70% 70% 60% 60% 50% 50% 40% 40% 83% 81% 72% 73% 76% 73% 71% 76% 75% 73% 30% 30% 53% 59% 49% 20% 20% 10% 10% 0% 0% 1 2 3 4 5 itt al Dh g Kh a Ra lna Ra ahi et l Sy r Al pu on ak Ch aris lh jsh u ng ag B Sell Do not sell Source: Authors’ calculations using BIHS 2018. Farmers largely sell in the traditional village markets instead of at the farm- gate (Figure 1.4ч). Given the fact that farm sizes are small, production is usually not large enough to justify marketing it directly to wholesalers. The large majority of farmers (74 percent) sell their products at the village market (hat-bazar). 135 Hat-bazar market access varies across districts, however. Some districts have a relatively high share of agricultural households selling at home. 135 A Hat bazar is an open-air village market that serves as a trading venue for local people in rural areas. 89 BANGLADESH RURAL INCOME DIAGNOSTICS More than 50 percent of households in Gopalganj, Kishoreganj, Sunamganj, Meherpur, and Nilphamari sell their products at home, for example. In Nilphamari district, 2 out of 3 households sell at home. Figure 1.48. Where farmers sell their produce a. Share of farmers selling at home b. Comparison across districts 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Faridpur Rajbari Narail Jhalokati Sherpur Netrakona Natore Jamalpur Satkhira Maulvibazar Magura Jhenaidah Rajshahi Lalmonirhat Sylhet Munshiganj Jessore Cox's Bazar Brahmanbaria Kushtia Kishoreganj Nilphamari Bay of Bengal 0.41 – 0.51 0.32 – 0.41 0.22 – 0.32 0.12 – 0.22 0.03 – 0.12 At hat-bazar At home Other markets Source: Authors’ calculations using BBS (2018). Commercialization of rice is lower than for other products, in particular for poorer farmers exclusively producing rice. The share of farmers only produc- ing rice who sell their products is lower than the corresponding share among households that are producing both rice and other crops (62 and 96 percent, respectively). More importantly, while more diversi‫ޔ‬ed households across the whole consumption distribution are equally likely to sell their products, among households that only produce rice, those in the lowest quintile have a far lower probability of participating in markets than those at the high end of the con- sumption distribution (46 versus 64 percent). This re‫ޕ‬ects the fact that poor farm households are more likely to self-consume or store part of their rice and engage in sharecropping arrangements. 90 ASSESSING CONTEXT AND HETEROGENEITY Agricultural exports still account for a low share of the country’s total exports. In 2018, over 90 percent of Bangladesh’s exports were RMG. The main agricul- ture-related exports were very small in comparison: frozen foods (1.3 percent), agricultural products (2 percent), leather and leather products (3 percent), and jute and jute goods (2 percent).136 Exports of rice are very limited. Despite near self-suȂiciency in rice production in recent years, export of non-fragrant rice was banned for over a decade, likely depressing local prices relative to exportable price and reducing incentives to improve eȂiciency. Export of fragrant rice was only allowed on a limited scale. The government liȅed the export ban in May 2019 amid a drastic fall in paddy prices, following a record domestic production of 37.2 million metric tons (MTs) of rice. However, since the ban was liȅed, export of non-fragrant rice has not risen as expected, given the low rice price in international markets. On the other hand, export of fragrant rice has been slowly increasing in recent years, though it remains miniscule relative to the total volume of the country’s exports. Export destinations for fragrant rice are limited to niche markets in settings with a sub- stantial Bangladeshi diaspora. Bangladesh exported more than 22,400 MTs of rice worth US$ 17.7 million in ‫ޔ‬scal year 2019, of which 100 percent was fragrant rice. The export in ‫ޔ‬scal year 2018 had been 11,000 MTs, worth US$ 7.13 million. Bangladesh’s rice markets appear to be well integrated. Analysis using daily price data from the Department of Agricultural Marketing (DAM) suggests that prices of rice move in step across the country and that price disparities compared to Dhaka are not very large. Figure 1.49a shows monthly average retail prices of boro rice across divisions, relative to Dhaka. Overall changes in prices across months are correlated across markets, and prices move within 20 percent of the Dhaka average retail price. Figure 1.50a compares wholesale and retail prices for boro rice across markets and months and shows that rice markups are below 20 percent. This indicates that rice markets are integrated. For more perishable products, such as brinjal (eggplant), prices and mark- ups vary more across the country. Figure 1.49b shows price variations com- pared to Dhaka for brinjal. In this case, retail prices vary between 50 and 120 percent of the Dhaka average prices. For this product, markups vary more than for rice, though in many cases they are below 40 percent (Figure 1.50b). 136 More than 50 percent of Bangladesh’s food exports consist of ‫ޔ‬sh and crustaceans, whether live, frozen, or prepared (World Bank 2020). 91 BANGLADESH RURAL INCOME DIAGNOSTICS Higher variation is expected due to the perishable nature of the product and the more local nature of commercialization. Price variations right aȅer the COVID-19 lockdown also suggest that markets are reasonably able to smooth shocks in demand and supply (Figure 1.49 and Figure 1.50). Aȅer the March 25, 2020, lockdown, prices of rice across markets became more similar, re‫ޕ‬ecting the government response to protect rice sup- ply and prices. Markups did not vary compared to previous months. Variations in prices and markups increased for brinjal, though not dramatically. Food in‫ޕ‬a- tion rates aȅer COVID-19 have remained low. Recent studies also show that price information ‫ޕ‬ows eȂiciently from traders to farmers, indicating that traders are unlikely to enjoy market power on the basis of asymmetric information.137 Figure 1.49. Prices of boro rice and brinjal across divisions, compared with Dhaka average, 2019 a. Boro retail prices b. Brinjal retail prices 1.5 Ratio over Dhaka price of Brinjal ‫ ە‬Superior Quality 1.2 Ratio over Dhaka price of Boro‫ە‬Medium 0.1 1 0.8 0.5 0.6 0.4 0 2019m1 2019m7 2020m1 2020m7 2019m1 2019m7 2020m1 2020m7 Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Mymensingh Source: Authors’ calculation using DAM (2019). However, the distribution and marketing of produce is still largely informal and localized. As Bangladesh’s economy transitions toward an upper-middle-in- come economy, formal markets (e.g., supermarkets and large-scale processors) 137 Gautam and Faruquee (2016). 92 ASSESSING CONTEXT AND HETEROGENEITY will be needed to increase the value of commercialization. Almost 98 percent of agricultural goods transit via rural wholesale markets, larger wholesale markets around large cities, and wet markets and corner shops in cities. These markets face challenges of inadequate infrastructure, inconsistent quality, and lack of enforcement of food safety standards. Retail via supermarkets remains at a very limited scale (about 1 percent of retail sales of fruits and vegetables), making it harder to justify investments along the value chain. In some other countries, urbanization and eȂorts to improve standards and management systems have altered such dynamics. Approaches aiming to boost quality and safety at the pri- mary production level have encouraged farmers in those countries to invest in better production, handling, and transport technologies.138 Figure 1.50. Markups for boro rice and brinjal across divisions, 2019 a. Boro rice b. Brinjal retail prices 1.25 2 Price of Brinjal ‫ ە‬Superior Quality 1.2 Markup of Boro‫ە‬Medium 1.15 1.5 1.1 1.05 1 1 2019m1 2019m7 2020m1 2020m7 2019m1 2019m7 2020m1 2020m7 Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet Mymensingh Source: Authors’ calculation using DAM (2019). Note: Markups are defined as the ratio between the retail price and wholesale price of the item in the same market. 138 World Bank (2020a). 93 BANGLADESH RURAL INCOME DIAGNOSTICS Stronger credit and insurance markets can expand rural growth potential Recent years have witnessed a continued expansion in access to credit, with NGOs having a large presence in this market. Between 2015 and 2017, the share of rural households with access to a bank account increased from 30 to 50 per- cent.139 140 About 35 percent of rural households reported obtaining a loan in the past year (Figure 1.51a). Loans are mainly obtained from NGOs, with ASA, Grameen Bank, and BRAC being the most commonly used. The share of loans from informal sources, including moneylenders, is small (Figure 1.52). Figure 1.51. Use of credit in rural areas a. Percentage of rural households obtaining b. Percentage of households that obtained a loan in the past 12 months, by division agricultural credit, by districtt 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% itt al g Kh a Ra na Ra hi t h r an lhe pu on ak es Ch aris a ul jsh ng ad Dh ag l B Sy B gl 0.58 – 0.67 Bay of Bengal 0.49 – 0.58 ra 0.40 – 0.49 Ru 0.31 – 0.40 All rural Only rural poor 0.22 – 0.31 Source: Authors’ calculations using HIES 2016 (panel a) and BBS (2018) (panel b). The poor in rural areas report using credit at rates close to the rural average (about 33 percent for all households in rural areas, compared to 35 percent for the rural poor). The sources of credit used by the poor diȂer little from those accessed by better-oȂ rural households, though ASA and Grameen Bank are used less oȅen by the richest quintile. Findex (2015, 2017). 139 Despite the expansion of credit, Bangladesh has a relatively low share of domestic credit to private 140 sector as percentage of GDP. In Bangladesh this share is 45 percent, while in Vietnam, India, and South Asia as a whole the corresponding ‫ޔ‬gures are 137 percent, 50 percent, and 47 percent, respectively (WDI, 2020). 94 ASSESSING CONTEXT AND HETEROGENEITY Figure 1.52. Sources of loans for rural households by consumption quintile 12% 10% 12% 13% 13% 16% 7% 6% 6% 6% 14% 6% 12% 13% 4% 14% 13% 16% 17% 15% 15% 15% 14% 14% 22% 21% 23% 22% 21% 17% 26% 26% 24% 21% 23% 18% 1 2 3 4 5 All ASA Grameen Bank Other NGO BRAC Other micro finance establishment Krishi Bank/Rajshahi Krishi Bank Relatives Money lender Other Source: Authors’ calculations using HIES 2016. Solid progress in strengthening credit markets has not reached those in the middle of the demand spectrum. There is no evidence that simply liȅing credit constraints will allow faster income growth. However, perceived lack of access to ‫ޔ‬nance is identi‫ޔ‬ed as the third most important obstacle to doing business in Bangladesh. Lack of credit is systematically reported by farmers, key actors along value-chains, and household-based enterprises as a constraint for growth. According to HIES 2016 data, household-based enterprises reported lack of credit as their main constraint to growth. About 70 percent of households with enter- prises mentioned inadequate capital or credit as their most important growth barrier. This is observed across all divisions. About 83 percent of households with enterprises reported ‫ޔ‬nancing operations with their own resources (for exam- ple assets, inheritance, or savings). Nearly half of small and medium enterprises (SMEs) report being fully or partially hindered in their access to credit. Mid-size enterprises in particular are underserved by the formal ‫ޔ‬nancial sys- tem. Banks are mostly concentrated in urban areas and target larger corporations, while micro‫ޔ‬nance targets micro-businesses. Meanwhile, the middle segment remains underserved, with 48 percent of SMEs being fully or partially constrained 95 BANGLADESH RURAL INCOME DIAGNOSTICS in access to credit. 141 The gap between the needs of SMEs and the funds available is estimated at about US$ 3 billion, and the interest rates that these businesses pay are about 4 percentage points higher than those available to larger enterpris- es.142 A recent study points out that limited access to credit curbs the growth pros- pects of medium-sized enterprises in sectors with export potential (e.g., footwear, plastics, light engineering).143 Financial inclusion for farmers is focused on short-term or seasonal loan prod- ucts. This may constrain investments in productive inputs or switching to high- er-value products. Some of the constraints preventing greater ‫ޔ‬nancial inclusion for farmers include: (i) eligibility criteria requiring land ownership or guarantees from the landlord to use land as collateral; (ii) perception of high risks driving high interest rates; and (iii) a limited oȂer of ‫ޔ‬nancial products and services tailored to farmers.144 Entrepreneurs in key value chains important for agricultural diversi‫ޔ‬cation also cite access to credit as a major business obstacle. About 37 percent of traders along four key value chains (brinjal, Pangash ‫ޔ‬sh, milk, and chicken) reported limited access to ‫ޔ‬nancing and the cost of ‫ޔ‬nancing as major or severe problems for their business. A further 24 percent identi‫ޔ‬ed them as minor or moderate problems. Access to ‫ޔ‬nancing ranked as the third top problem for traders in these value chains, aȅer road conditions and roadblocks.145 Finally, women in Bangladesh have less access to and control over ‫ޔ‬nance than do men. Analysis of Findex 2017 data reveals that only 35.8 percent of women had a bank account, compared to 64.6 percent of men. According to BIHS data, less than 4 percent of rural women and 11 percent of men used savings for pro- ductive purposes, though men were more likely than women to buy land. In addi- tion, women have limited control over household ‫ޔ‬nancial assets and a limited role in deciding the use of these assets. Use of mobile banking by women is very limited.146 141 World Bank (2020b). 142 World Bank (2020b). 143 Sumi and Reaz (2020). 144 World Bank (2020a). 145 Shilpi et al. (2015). 146 SolotaroȂ et al. (2019). 96 ASSESSING CONTEXT AND HETEROGENEITY Formal insurance markets are underdeveloped, and negative income shocks experienced by rural households are largely managed with own resources.147 148 As previously discussed, due to lack of options, the poor rely more on reducing food consumption when facing a shock. Response to large-scale weather events and natural disasters by government and development partners has been focused on dealing with ex-post losses. HIES data shows that government help is reported as a method for coping with shocks by 33 percent of households. Limited agricul- tural insurance constrains the ability to manage these risks ex-ante. Compared to India, weather insurance options are very limited. Most schemes, particularly around weather-index insurance, have operated on a small scale.149 Important factors inhibiting development include the capacity of the regulator, sectoral frag- mentation, the weak quality and capacity of many companies, and an erosion of public trust in the sector.150 Limited insurance markets may be discouraging farmers from investing in high- er-return but higher-risk production. Commercial insurance in Bangladesh could help small- and medium-scale farmers stabilize and increase their crop income by up to 41 percent, if insurance unlocks credit and adoption of high-yielding variet- ies.151 If farmers currently growing aman local or boro HYV switched to higher yield- ing varieties (aman HYV or boro hybrid, respectively), the increase in expected yield would largely compensate for the increase in input costs. A randomized control trial152 assessed the eȂectiveness of an index insurance product designed to help smallholder farmers in Bangladesh manage crop production risk during the mon- soon season. The study ‫ޔ‬nds that purchasing insurance yields both ex ante risk management eȂects as well as ex post income eȂects on agricultural production practices. The risk management eȂects lead to an expansion of cultivated area with concomitant increases in agricultural input expenditures during the monsoon sea- son. The income eȂects lead to more intensive rice production during the subse- quent dry season, with more intensive use of both irrigation and fertilizers, resulting in higher yields and higher total rice production. 147 The insurance sector in Bangladesh is heterogenous and fragmented, consisting of 62 companies that are a mix of public corporations and private entities. The sector is several times smaller than in many of Bangladesh’s aspirational peers, such as Malaysia and Thailand. Within Bangladesh’s insur- ance sector, around three-quarters of the market consists of life insurance products. 148 As previously noted, Bangladesh’s ‫ޔ‬nancial sector is characterized by low penetration of ‫ޔ‬nancial products and services beyond current accounts. Savings products represent just 18 percent of avail- able ‫ޔ‬nancial products (PWC 2019). 149 Aheeyar et al. (2019). 150 World Bank (2020a). 151 World Bank (2015b). 152 Hill et al. (2019). 97 SECTION 2. Analyzing opportunities and constraints for rural income growth As the previous section showed, rural households pursue multiple livelihood strategies to raise their incomes. A strategy to support these households thus requires a balanced approach that enhances opportunities to accelerate growth in agriculture, but also opportunities for non-farm and migration income. This strategy needs to pay particular attention to location in a context of growing spatial income disparities, and it needs to tap the potential of women to earn higher incomes. This section describes the opportunities to boost income growth in rural areas and identi‫ޔ‬es constraints to taking advantage of those opportunities. The discus- sion builds on the descriptive analysis presented in section 1 and on an extensive literature review, cited where relevant. It is important to emphasize that the role of macroeconomic factors, general business environment constraints, and gover- nance challenges, which are also relevant to promote income growth, are outside the scope of this analysis. In addition, the diagnostic does not provide detailed analysis of constraints to fertility reduction, health, nutrition, and education, nor does it identify policy priorities to address these. Raising incomes in agriculture Agriculture in Bangladesh needs to grow faster and become more poverty reducing. Despite the country’s economic transformation, agriculture still accounts for a large share of rural incomes. The recent slowdown in agricul- tural growth is an important reason why poverty has stagnated or increased in the Western areas of the country, so the performance of agriculture is also central for reducing spatial welfare disparities. Given Bangladesh’s limited scope for arable land expansion and the current high intensity of land use, 99 BANGLADESH RURAL INCOME DIAGNOSTICS agricultural growth will need to come from a combination of higher agricultural productivity, greater diversi‫ޔ‬cation toward high-value crops, and developing a modern agri-food supply chain.153 While potential for diversi‫ޔ‬cation through arable land expansion is limited, diversi‫ޔ‬cation could be achieved through higher production intensity and sub- stitution of crops. World Bank (2020a) indicates that intensi‫ޔ‬cation could poten- tially be increased in almost all areas of the country. This requires closing invest- ment gaps as well as addressing the current policy environment. Diversi‫ޔ‬cation through crop substitution can be achieved through closing the current yield gap. Evidence suggests that diversi‫ޔ‬cation potential in agriculture is considerable across all farming systems.154 Considering that climate change is a major risk for the sector, the realization of agricultural growth potential hinges upon the intro- duction of climate-smart practices. This will include use of inputs and technolo- gies that are more resource eȂicient, introduction of crops that are more resilient, and the modernization of agribusiness storage and logistics. Currently, Bangladesh obtains low yields in many agricultural products and remains uncompetitive in these outputs. Bangladesh yields are below regional and world averages, with the exception of rice paddy, mangoes, and guavas. Bangladesh also underperforms in most livestock items (Table 2.1). Bridging these yield gaps will be essential to increase the production and cost competi- tiveness of agricultural products. A weak business environment contributes to these gaps (Box 2). Closing yield gaps in rice remains important. The overall growth of the agricul- tural sector is linked to growth in rice production, which has been lagging. There is room to close remaining yield gaps in aman and aus rice to support freeing up suitable boro rice areas for a diverse range of high-value and more sustainable crops, while accelerating the development of the livestock, poultry, and ‫ޔ‬sher- ies subsectors. The gaps between agronomic potential yield obtained in research ‫ޔ‬elds and yields in current farm ‫ޔ‬elds are considerable, ranging from 25 percent for boro to 75 percent for aus and aman paddy.155 153 World Bank (2020a). 154 Masylkanova (2020). 155 World Bank (2020a). 100 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH Table 2.1. Comparison of national crop and livestock yields, annual average, 2014-2018 Crops (MT/ha) Areca Chilies, Mangoes, Onions, Rice, Bananas Potatoes Wheat nuts dry guavas dry paddy Bangladesh 0.8 17 1 23 10 20 5 3 World 1 21 2 9 19 21 5 3 Afghanistan -- -- -- -- 12 15 3 2 Bhutan 1 -- 5 5 2 10 4 2 India 2 36 2 9 16 22 4 3 Nepal 3 16 5 8 -- 14 3 3 Pakistan -- 4 2 10 13 22 4 3 Sri Lanka 3 1 -- 10 16 16 4 -- Livestock (kg/an) Eggs, Milk, Milk Cattle Chicken Duck Goat meat hen (cow) (goat) meat meat meat Bangladesh 2 205 38 71 0.7 1 7 World 10 2449 87 219 2 1 12 Afghanistan 2 494 44 180 0.8 -- 13 Bhutan 6 1387 -- 103 1 -- 9 India 13 1588 167 103 1 1 10 Nepal 5 612 67 85 0.8 0.8 12 Pakistan 5 1230 106 196 1 1 17 Sri Lanka 12 979 116 136 1 1 20 Source: FAOSTAT. The expanding domestic market for high-value food provides an opportunity to increase incomes by diversifying output into higher value-added crops, live- stock, and ‫ޔ‬shing products. With rising urbanization and increasing incomes, dietary patterns in Bangladesh are changing from cereals to nutrient-dense, high-value agricultural commodities, and tŚĞƌĞŝƐĂƌĂƉŝĚůLJĞdžƉĂŶĚŝŶŐůŽĐĂůŵĂƌ- ŬĞƚ ĨŽƌ ĨƌƵŝƚƐ͕ ǀĞŐĞƚĂďůĞƐ͕ ƉƌŽƚĞŝŶƐ͕ ĂŶĚ ƉƌŽĐĞƐƐĞĚ ĨŽŽĚƐ͘156 Since 2000, average daily rice intake per person has declined by 20 percent - from 459 grams in 2000 to 367 grams in 2016 - while consumption of vegetables, fruits, ‫ޔ‬sh, meat, onion, and eggs increased considerably over the same period (Figure 2.1). If the current 156 Dizon et al. (2019). 101 BANGLADESH RURAL INCOME DIAGNOSTICS downward trend in rice consumption continues, total consumption of rice will decline by a further 9 percent by 2030. A projection of food demand shows that total demand for eggs, fruits, meat, and ‫ޔ‬sh will grow 50 percent or more by 2030, compared to current levels.157 Figure 2.1. Change in food intake between 2000 and 2016 Egg Edible oil Onion Meat Fish Fruit Vegetables Wheat Pulses Milk and milk products Rice -25% 0% 25% 50% 75% 100% 125% 150% 175% Source: Authors’ calculations using HIES 2000 and 2016. Projections show room for increasing outputs of non-rice products without compromising the supply of rice. As noted, the scope for improving rice yields is substantial, given current yield gaps, especially for aman and aus rice. With a continued focus on agricultural research to push the production frontier and build agriculture’s resilience to climate and biophysical stress, there is signi‫ޔ‬cant potential to release land for much-needed crop diversi‫ޔ‬cation without compro- mising the production of rice and food security targets.158 A comparison of gross margins between rice and selected non-rice cereal, fruit, and vegetable crops indicates that non-rice production can provide higher pro‫ޔ‬ts than rice. Sergeant and GraȂham (2020) estimated and compared gross margins for crops including rice, maize, soybean, gourd, potatoes, tomatoes, and mango (Figure 2.2). They found that many crops oȂer higher average margins than rice. This explains why many farmers are gradually diversifying toward non-paddy crops. 157 Ahmed et al. (2021). 158 'ĂƵƚĂŵ ĂŶĚ &ĂƌƵƋĞĞ ;2016Ϳ͖ tŽƌůĚ ĂŶŬ ;2020ĂͿ͘ 102 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH Figure 2.2. Projected gross profit margins for rice and selected other crops in Bangladesh (US$/ha), 2020 ǘǕǕǕ ǗǚǕǕ ǗǕǕǕ ǖǚǕǕ ǖǕǕǕ ǚǕǕ Ǖ ȒǚǕǕ Boro *+ Aus Maize Soybean Gourd ,*1, Potato Mango Rice Livestock Vegetables feed crops Total Revenue Total Cost Gross Margin Source: "/$"+1+! /ƛ%*țǗǕǗǕȜǽ The livestock subsector is small, but with signi‫ޔ‬cant potential. With the excep- tion of poultry, livestock farming is largely smallholder, with low productivity and relatively slow growth compared to other subsectors. Growth in the livestock subsector has high potential to create rural jobs and livelihood opportunities for women, youth, and the vulnerable. About 80 percent of rural households are engaged in either livestock or poultry production, though mostly for own con- sumption. As shown in section 1, women make up a large share of the agricul- tural labor force, and they are mainly involved in livestock and poultry systems. In addition, a growing share of rural households without cultivable agricultural land are engaged in livestock.159 Appendix 3 summarizes some of the main opportunities and constraints in the livestock sector. The ‫ޔ‬sheries sector has emerged as a major growth driver, consistently contrib- uting 6 percent of gross domestic product for the last decade. Fisheries have also become the second-largest foreign currency earner aȅer the garment industry. The 2016 HIES, administered by the BBS, showed that the share of landless rural agricultural house- 159 holds involved in the livestock subsector increased to 66 percent in 2010, from 37 percent in 2000. 103 BANGLADESH RURAL INCOME DIAGNOSTICS According to the 2019 agriculture census, 2.8 percent of households depended on ‫ޔ‬shing at some point during the year. The sector comprises three subsectors: aquaculture (56 percent of total production), inland capture ‫ޔ‬sheries (28 percent), and marine and coastal capture ‫ޔ‬sheries (16 percent). Aquaculture has experi- enced rapid growth during the past 30 years, amounting to a 25-fold increase in farmed ‫ޔ‬sh production.160 Bangladesh was the fourth-largest aquaculture pro- ducer globally, aȅer China, India, and Vietnam.161 Nearly all aquaculture produc- tion (94 percent) in Bangladesh is consumed domestically.162 World Fish has estimated that another 1 million hectares of inland waters in Bangladesh could potentially be developed for productive ‫ޔ‬shing. This would encompass ‫ޕ‬ood lands (70 percent of the additional area), rivers and estuaries (10–15 percent), and inland closed waters (10–15 percent). Bringing this esti- mated incremental area under well-managed ‫ޔ‬sh production could raise gross revenues from ‫ޔ‬shing by US$2 billion; generate an additional 1.5 million full-time equivalent jobs, mostly among the rural poor; and create signi‫ޔ‬cant upstream and downstream indirect jobs and enterprises. Appendix 3 summarizes some of the main opportunities and constraints in the ‫ޔ‬sheries sector. Raising non-farm and migration incomes Improving quality job creation across the non-farm sector is central to reap the bene‫ޔ‬ts of the demographic dividend and reduce vulnerability. Increasing job creation in manufacturing remains central, as this has been a major driver of pov- erty reduction in the past. Boosting agricultural productivity is also important for non-farm job creation through its multiplier eȂects in manufacturing (food pro- cessing) and services (food service economy and domestic trade/marketing). This is an important pathway for structural transformation, as higher labor productiv- ity also allows freeing up from agriculture to shiȅ into these activities.163 Supporting the growth of secondary cities and other urban areas can help boost rural incomes and reduce poverty. Varying degrees of urbanization across districts strongly shape non-farm employment opportunities and pov- erty reduction, including in rural areas. To date, Bangladesh’s urbanization 160 Hernandez et al. (2017). 161 FAO (2016). 162 Hernandez et al. (2017). 163 Shilpi and Emran (2016) ‫ޔ‬nd that Bangladesh’s agricultural productivity growth has supported strong non-farm job growth and transformation toward wage jobs, many of them in social services located in nearby urban areas. 104 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH and economic development have been unbalanced and largely concentrated around Dhaka and Chittagong. Supporting economic activity in and around the country’s smaller cities oȂers an important lever to raise non-farm incomes among rural households. Section 1 highlighted that industry and service jobs are largely informal, of low sophistication, and vulnerable. Bangladesh has seen manufacturing growth in peri-urban areas of Dhaka, but job growth in second- ary cities and other urban areas has thus far been limited. Other urban areas are not yet attracting industrial investment, and access to quality jobs is limited outside the capital.164 Gaps in access to electricity and education have substantially narrowed between Dhaka and other urban areas, creating an opportunity for higher pro- ductivity and job creation in secondary areas. Urban areas outside Dhaka and Chittagong have traditionally had substantially lower levels of education and rel- atively poor access to health care and electricity, discouraging industrial invest- ment and job growth in smaller cities. While gaps remain, particularly in terms of water and sanitation, progress has been important in reducing these infrastruc- ture and human-capital gaps. For instance, in 2000, only 65 percent of households in urban areas outside Dhaka and Chittagong divisions had access to electricity, compared to 91 percent in urban Dhaka division. In 2016, more than 9 in 10 house- holds in other urban areas enjoyed access to electricity (Table 2.2), However, as previously highlighted, accessibility gaps remain, creating high transaction costs for economic activity outside the country’s main corridor. Table 2.2. Access to basic services and level of education, 2016 Access Access Household Access to Secondary to to piped has mobile Literacy and electricity water sanitation phone rate above Urban Dhaka 98% 57% 49% 97% 70% 28% Urban Chittagong 92% 20% 45% 95% 64% 24% Other urban 90% 11% 27% 94% 70% 29% Source: Authors’ calculations using HIES 2016. Note: Access to electricity, piped water, sanitation, and mobile phone are estimated at household lev- els. An individual is considered literate if she/he can write a letter. Literacy rate and Secondary and above rates are based on individuals older than 18. Dhaka and Chittagong refer to the divisions, not only the metropolitan areas. 164 Farole and Cho (2017). 105 BANGLADESH RURAL INCOME DIAGNOSTICS Exploiting local comparative advantages can support income growth in urban areas outside the Dhaka and Chittagong corridor. Some earlier policy eȂorts tried to bring jobs to people by changing relocation incentives for ‫ޔ‬rms in Bangladesh. However, these eȂorts have not succeeded in counterbalancing the powerful agglomeration forces that anchor many ‫ޔ‬rms in Eastern areas and in turn encour- age people to seek jobs there.165 The current agglomeration economies indicate that local ‫ޔ‬rm growth and not the reallocation of ‫ޔ‬rms will drive growth in sec- ondary cities and other urban centers. The agro-processing sector in Bangladesh provides an opportunity for income growth and development of secondary cities and other urban areas. Secondary and other urban centers are well placed to host value-added processing of agricul- ture and can be critical links to support raising agricultural productivity and rural incomes. The current spatial distribution of agro-processing is more even across the country than is the case for other non-farm activities, highlighting its potential to create non-farm income opportunities in a more decentralized way. Muzzini and Aparicio (2013) highlight that agro-processing activities have signi‫ޔ‬cant potential, as they provide nearby production areas with easily accessible markets (Figure 2.3).166 The sector has the capacity to create supply chains by linking diȂerent actors across service sectors such as trade, marketing, and transportation, among others. Agro-processing can also expand the market for the already well-settled agricultural sector.167 EȂorts to strengthen value addition in agriculture can be thought of as complementing broader eȂorts to expand quality job growth in urban areas. Agro-processing also has the potential to attract international invest- ment and increase export diversi‫ޔ‬cation. Foreign direct investment (FDI) in domestic markets can help raise productivity by strengthening domes- tic competition and integrating domestic suppliers into their value chains. Similarly, outward-bound FDI by Bangladeshi investors allows ‫ޔ‬rms to grow larger, become more competitive through exposure to foreign competition, and acquire new knowledge and technology. The COVID-19 crisis will lower FDI ‫ޕ‬ows in the short term, but foreign investment will arguably be even more important for Bangladesh’s ability to leverage exports in its recovery from the crisis.168 While ready-made garments dominate exports, agricultural and related agro-processed goods represent the country’s third export cluster.169 165 Muzzini and Aparicio (2013). 166 Muzzini and Aparicio (2013); Farole and Cho (2017). 167 Rahian et al. (2017). 168 World Bank (2020a) 169 Atlas of Economic Complexity. https://atlas.cid.harvard.edu 106 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH Fish and crustaceans currently account for 50 percent of Bangladesh’s agro-pro- cessing exports, but export products also include tea, vegetables, tobacco, ‫ޕ‬owers, fruits, spices, and dry food. FDI in food products has increased in recent years, under- scoring the potential of agro-processing to attract international capital (Figure 2.4). Figure 2.3. Economic base of secondary cities DiȂerence between local sectoral growth and national growth, 2001-2009 10 9 Location quotient (2009) Legal and 8 accounting 7 6 Chemicals Hotels and Transport and 5 restaurants equipment Publishing 4 Wholesale 3 and retail Financial 2 services Jute 1 Agro-processing 0 -200 -100 0 100 200 300 Important sectors that Small declining Important growing Potential demand attention sectors sectors emerging clusters Source: Muzzini and Aparicio (2013). Figure 2.4. Evolution of foreign direct investment in agro-processing, 2010-19 900 25% 800 Percentage of total net FDI 700 20% 600 MIllion of US$ 15% 500 400 10% 300 200 5% 100 0 0% 11 12 14 15 16 17 18 19 3 01 0 0 0 0 0 0 0 0 -2 -2 -2 -2 -2 -2 -2 -2 2 2- 10 11 13 14 15 16 17 18 1 20 20 20 20 20 20 20 20 20 US million Share Source: Authors’ calculations using Bangladesh Bank (2019). 107 BANGLADESH RURAL INCOME DIAGNOSTICS The previous section highlighted the importance of migration for income growth and poverty reduction in Bangladesh. The COVID-19 crisis may recon‫ޔ‬gure some opportunities and obstacles in this domain. However, the evidence indicates that expanding migration opportunities across areas of the country can be a way to spread the bene‫ޔ‬ts of this process. More balanced urbanization and job growth in areas outside Dhaka and Chittagong are important elements to consider, as a main motivation for migration is the pursuit of greater economic opportunities. Raising women’s incomes Increased female labor force participation can support higher incomes. Using panel data, Gautam and Faruqee (2016) show that female labor force participa- tion is associated with about 14 percent higher household income. Heath and Mobarak (2015) report that greater opportunities for women in the manufacturing sector are associated with delayed marriage and childbirth and expanded female participation in education. The study shows that: (a) young girls become more likely to be enrolled in school as garment job opportunities that reward literacy and numeracy arrive; and (b) older girls become more likely to work outside the home in villages located in proximity to garment industry facilities. Women can ease pressures in the agricultural labor market. Crop diversi‫ޔ‬ca- tion and the use of agricultural labor are strongly linked.170 Due to high migra- tion from rural areas to both domestic cities and abroad, the rural labor market in Bangladesh is tightening, as evidenced by the agricultural wage growing faster than the general wage rate.171 This may in part be driving women’s increasing participation in agriculture. The changing gender composition of the agricultural workforce implies the need to tackle constraints that still hinder some women from taking full advantage of these opportunities. Constraints to increasing rural incomes Current evidence indicates a series of constraints to take advantage of the opportunities outlined above. We classify these constraints in ‫ޔ‬ve areas: (i) pol- icy focus and regulations; (ii) investment gaps; (iii) market-related constraints; (iv) knowledge and information constraints; and (v) gender norms (Figure 2.5). 170 Hoque and Ahmed (2020). 171 Bangladesh Economic Review (2019). 108 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH This grouping aims to organize and clarify the discussion. It is important to note, though, that some constraints may span multiple categories, so this classi‫ޔ‬ca- tion is mostly for presentational purposes. Figure 2.5. Constraints on rural income growth Policy focus and Productivity and faster regulations diversification of agriculture Gender Investment Non-farm income norms gaps PRIORITIZED IMPACT ON CONSTRAINTS Migration income Market Knowledge failures and capacity Women's income constraints Source: Authors. Policy focus and regulations Moving forward, redirecting agricultural policy and support will be key to bal- ance the incentive structure across crops and promote diversi‫ޔ‬cation. The pre- vious section noted that rice and wheat production carries much lower price risks than growing other high-value agricultural commodities, as the price volatilities for cereals are much lower than those of non-cereal agricultural commodities.172 Since the 1980s, one reason behind lower price risks for cereals is the sustained agri- cultural policy focus on ensuring food security.173 Today, agricultural policy con- tinues to maintain incentives for rice production through input subsidies, output price support, and price stabilization. Such policy support for growing cereals has largely contributed to the country’s achieving self-suȂiciency in rice production. However, this approach also creates a bias against the production of non-cereal commodities. Farmers continue to grow cereals due to the low risks, despite lower gross margins compared to other options. Addressing bias in agricultural support against some high-value, non-rice crops can level the playing ‫ޔ‬eld and allow farm- ers to seek higher return opportunities without compromising food security goals. Ahmed et al. (2021). 172 For details on agricultural policy reforms and their impacts on agricultural growth, see Gautam and 173 Faruqee (2016). 109 BANGLADESH RURAL INCOME DIAGNOSTICS Current fertilizer subsidy policy promotes the overuse of fertilizer but does not translate into additional output. The overuse of fertilizer raises environmental, health, and even productivity risks, with land degradation. The cost of agricultural subsidies is substantial, with the largest share allocated to fertilizer subsidies. In FY2018-19, around USD 1 billion were allocated to subsidized fertilizers and other agricultural inputs, about 40 percent of the total budget allocation for the agricultural sector.174 This suggests an opportunity to redeploy support to other activities, such as R&D, that could boost productivity across crops more sustainably. More generally, support to farmers can also be provided in alternative ways that are less distortion- ary and that do not encourage the overuse or ineȂicient use of inputs involving large negative externalities. This support can take the form of providing incentives to use better, more resource-eȂicient and environmentally friendly inputs and practices. A more robust seed sector is also central to support productivity and diversi‫ޔ‬- cation. Recent research outlines steps to streamline seed regulations and develop a more robust seed sector. Studies have laid out steps to reduce remaining barri- ers in the seed sector.175 First, public-sector research must focus more strongly on crops other than rice. Second, authorities need to level the playing ‫ޔ‬eld for the private sector to help spur the development of better varieties of open-pollinated crops, including through public-private partnerships. Third, policy makers can rationalize the cumbersome, lengthy process for registering new varieties, giving attention to harmonizing varietal testing with India and other countries. Current regulations are largely favorable for the private sector’s participation in the seed industry, but further opportunities exist to remove cumbersome restrictions. The private seed sector is prohibited from breeding and producing seeds for noti‫ޔ‬ed crops.176 Another issue is the lack of a seed quality control mech- anism. The seed market is supplied with low-quality seeds at lower prices, which increases farmers’ mistrust toward certi‫ޔ‬ed seeds.177 Measures that could improve farmers’ access to and use of quality seeds include: removing the regulation that prevents private-sector actors from breeding and producing seeds for noti‫ޔ‬ed crops; increasing public-private coordination on quality seed supply for noti‫ޔ‬ed crops; and increasing market transparency regarding the quality and eȂective- ness of seeds. The Enabling the Business of Agriculture (EBA) index for Bangladesh indicates signi‫ޔ‬cant room for improvements in terms of seed regulations (Box 2). 174 Bangladesh Economic Review (2019). 175 Naher and Spielman (2014) and World Bank (2019b). 176 Gautam and Faruqee (2016); USAID (2014). 177 BRAC (2019); LTS (2019). 110 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH Box 2: Enabling the Business of Agriculture in Bangladesh Overall Supplying Registering Securing Using Sustaining Protecting Trading Accessing score seed Fertilizer Water machinery livestock plant health food ‫ޔ‬nance Bangladesh 44.5 18.5 47.1 20.0 49.9 56.7 60.0 43.7 60.0 India 62.2 73.2 57.8 20.0 79.0 33.3 60.0 74.6 80.0 Sri Lanka 50.2 74.0 64.1 0.0 91.1 91.1 60.0 25.4 40.0 Nepal 49.0 67.2 11.1 40.0 90.6 18.3 60.0 74.4 30.0 Pakistan 48.9 67.8 58.7 30.0 37.4 33.3 20.0 73.7 70.0 Vietnam 61.4 44.6 94.0 60.0 69.5 75.0 50.0 58.3 40.0 Afghanistan 31.5 18.5 73.4 20.0 0.0 0.0 40.0 70.2 30.0 The Enabling the Business of Agriculture (EBA) index measures how regulation aȂects the livelihood of domestic farmers.178 EBA indicators show that Bangladesh is lagging behind its peers in South Asia in many areas. Two important indicators concern the supply of seed and the registering of fertilizers. These indicators measure the quality of existing regulations, as well as the time and cost required to register new seed varieties and fertilizers. Results are signi‫ޔ‬cantly poorer in Bangladesh than in other South Asian countries. The EBA report shows that Bangladesh’s seed sector is highly regulated, limiting private participation. In the case of fertilizers, Bangladesh has ensured almost universal access to fertilizers for its farmers. However, bureaucratic procedures create obstacles to improving fertilizer quality and introducing new varieties of fertilizers. Bangladesh’s trading food metric is also relatively low but improving due to recent reforms. This indicator measures the time and cost required to obtain permits to trade agricultural goods, as well as the quality of regulation in the food system. Recently, Bangladesh has begun to publish the oȂicial fee schedule for phytosanitary certi‫ޔ‬cates online and in its legislation. This simple innovation facilitates access to crucial informa- tion for farmers, in turn facilitating trade in agricultural products. Trade policies contribute to the unbalanced use of mechanization. Constraints behind the unbalanced used of mechanization have been identi‫ޔ‬ed on the policy, supply, and demand fronts.179 Factors related to supply include prices and the availability of spare parts. Demand-side factors include farm- ers’ knowledge of how to use machinery and their ability to access ‫ޔ‬nance. 178 World Bank (2019b). 179 Alam and Khan (2017). 111 BANGLADESH RURAL INCOME DIAGNOSTICS An important constraint identi‫ޔ‬ed in the literature is an unresolved tension between the government’s dual policy objectives of increasing farm mechani- zation and supporting the local manufacturing sector. To achieve the ‫ޔ‬rst objec- tive, policies facilitate mechanization through elimination of trade barriers and provision of subsidies, loans, and tax exemptions for machinery. To advance the second objective, the policy imposes trade barriers on imports of spare parts (e.g., 40-50 percent import taxes).180 Transport policies lack coordination, and road safety enforcement is weak, increasing congestion and transport costs. Several ministries and agencies responsible for the transport sector have overlapping functions, and agencies’ service delivery and regulatory mandates may con‫ޕ‬ict. DiȂerent agencies have their own infrastructure development plans, with little coordination between them. In addition, weak enforcement of regulations regarding the licensing of drivers and vehicles defeats the purpose of the regulations. Unquali‫ޔ‬ed drivers driving heavy vehicles provide facilitation payments to obtain licenses. Modi‫ޔ‬ed or un‫ޔ‬t trucks are issued ‫ޔ‬tness certi‫ޔ‬cates. Transport operators oȅen overload their trucks by more than 50 percent. These practices have resulted in unskilled drivers, substandard trucks on the road, and overloading which aȂects the quality of road infrastructure and increases congestions costs.181 Weak local-level functions and capacity also constrain service delivery and local economic development. Dhaka’s dominance as the country’s main urban and economic center re‫ޕ‬ects not only its role as the nation’s capital, but also policy choices that tend to bolster centralization. The fact that political, administrative, and regulatory decision-making powers are concentrated in Dhaka has facilitated the disproportionate development of the city over other urban areas.182 In order to meet high-end market standards and develop a strong food pro- cessing sector, Bangladesh needs to improve its food safety regulatory and oversight systems. Food safety and standards are a serious issue for agricul- ture in Bangladesh, due to unbalanced and excessive use of fertilizer and pes- ticides, along with a lack of good agricultural management practices. Food safety challenges have hurt Bangladesh’s food sector in the recent past, both in domestic and export markets. Despite previous eȂorts to address these issues, compliance monitoring and law enforcement remain weak and irregular. 180 World Bank (2020a). 181 Herrera Dappe et al. (2019). 182 Bird et al. (2018). 112 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH The government, for instance, does not have accredited laboratories for the full range of necessary tests and relevant products.183 It is also important to support the private sector to adopt appropriate food safety standards and practices.184 Investment gaps Diversi‫ޔ‬cation toward non-paddy crops requires closing irrigation gaps in some geographic areas and also enabling farmers to improve ‫ޕ‬ood, drainage, and overall water control. The existing irrigation infrastructure can be reori- ented to support other crop needs. The extensive access to water for irrigation on the edge of the ‫ޔ‬elds is especially useful during the dry season and for other crops such as fruits and vegetable that are sensitive to water levels. World Bank (2020a) show that farms with irrigation access are more diversi‫ޔ‬ed. Bangladesh is subject to signi‫ޔ‬cant ‫ޕ‬ooding during parts of the year, which provides space for crop and aquaculture production. However, there are also considerable risks from deep ‫ޕ‬ooding, erosion, and drainage problems. In addition, in the post-monsoon period, surface water availability is signi‫ޔ‬cantly constrained. Combined with erratic pre-monsoon rainfall, this can cause serious soil mois- ture de‫ޔ‬cits, which must be addressed through irrigation. Between June and September, when about 90 percent of rainfall takes place, farmers need drain- age and ‫ޕ‬ood protection structures. In non-monsoon seasons, farmers need dependable irrigation. Existing small-scale pump irrigation can technically support required water control. However, declining groundwater availability is aȂecting pumping costs, as well as the availability of irrigation water, par- ticularly in the northwest and in Dhaka.185 The irrigation system, which began with large public irrigation projects, has gradually moved to private enterprises. Therefore, tackling declining groundwater availability and closing investment gaps will require both public and private eȂorts. Bangladesh has succeeded in putting basic infrastructure in place, but gaps remain very large. The 2019 Global Competitiveness Report ranks Bangladesh 114th out of 141 countries on infrastructure quality, below the regional aver- age. The lack of adequate transport and logistics infrastructure and reliable power supply leads to high production costs that constrain overall income growth. The country’s total infrastructure needs through 2040 are estimated at US$608 billion. These needs include investment in infrastructure expansion, 183 World Bank (2020a). 184 Ahmed et al. (2021). 185 World Bank (2020a). 113 BANGLADESH RURAL INCOME DIAGNOSTICS modernization, and maintenance. Raising private ‫ޔ‬nance will be essential. In addition, given high climate-change risks, adaptation measures need to be priced into future infrastructure assets.186 Closing electricity coverage gaps and improving the reliability of the supply can strengthen productivity, facilitate diversi‫ޔ‬cation, and support non-farm employment. Access to electricity is still far from universal in Bangladesh, and electricity supply is unreliable, with repeated outages. This is particularly con- straining for perishable goods and food processing, as it damages equipment and requires supplementing the grid with generators. It also creates challenges to maintain cold chain operations along the value chain. Compared to other countries, Bangladesh lags behind in electricity access and supply quality. The country remains below the South Asia average in terms of the share of the rural population with access to electricity (88 percent).187 Despite improvements in generation capacity, less than 80 percent of existing pow- er-generation capacity is operational, and power cuts are frequent. The 2019 Competitiveness Index of the World Economic Forum ranked Bangladesh 108th in terms of electricity access and 68th in terms of electricity supply quality, among 141 countries.188 The central issue in the energy sector is lack of cost recovery and ineȂiciencies. These elements undermine the sector’s sustainability and discourage greater private sector participation. >ŝŬĞ ŵĂŶLJ ŽƚŚĞƌ ĐŽƵŶƚƌŝĞƐ͕ ĂŶŐůĂĚĞƐŚ ƐƵďƐŝĚŝnjĞƐ ƚŚĞ ĐŽŶƐƵŵƉƟŽŶ ŽĨ ĞůĞĐƚƌŝĐ ƉŽǁĞƌ ĂŶĚ ŶĂƚƵƌĂů ŐĂƐ͘ Despite gradual increases since 2009, retail tariȂs (which average 6 Takas per kilowatt hour ҙkWhҚ) do not cover cost (8 Takas/kWh). dŚĞ current electricity rate for farmers is 25 percent lower than the cost of the electricity supply itself and is lower than for individual users and businesses (US$0.03 per kWh for farmers against US$0.04 for individual con- sumers and US$0.08 for businesses). dŚĞ ĐŽŶƐƵŵĞƌ ƉƌŝĐĞ ŽĨ ĞůĞĐƚƌŝĐŝƚLJ ŝƐ ƐĞƚ Ăƚ Ă ůĞǀĞů ǁĞůů ďĞůŽǁ ŝƚƐ ůŽŶŐ-ƌƵŶ ŵĂƌŐŝŶĂů ĐŽƐƚ ŽĨ ƉƌŽĚƵĐƟŽŶ͘ dŚŝƐ ŚĂƐ ůĞĚ ƚŽ Ă ƐŝƚƵĂ- ƟŽŶ ŝŶ ǁŚŝĐŚ ƚŚĞ ŐŽǀĞƌŶŵĞŶƚ ŶĞĞĚƐ ƚŽ ƉƌŽǀŝĚĞ Ă ĚŝƌĞĐƚ ;ďƵĚŐĞƚͿ ƐƵďƐŝĚLJ ƚŽ ƚŚĞ ĂŶŐůĂĚĞƐŚ WŽǁĞƌ ĞǀĞůŽƉŵĞŶƚ ŽĂƌĚ ;WͿ͕ ƚŚĞ ƐƚĂƚĞ-ŽǁŶĞĚ ƵƟůŝƚLJ ƌĞƐƉŽŶƐŝďůĞ ĨŽƌ ƐƵƉƉůLJŝŶŐ ĞůĞĐƚƌŝĐŝƚLJ ŝŶ ƚŚĞ ĐŽƵŶƚƌLJ͘ dŚĞ ŐĂƉ ďĞƚǁĞĞŶ ƚŚĞ ƐƵƉƉůLJ ĐŽƐƚ ĂŶĚ ƚŚĞ ƚĂƌŝī ĞŶƚĂŝůƐ Ă ƐƵďƐŝĚLJ ƚŚĂƚ ŝƐ ĚŝƌĞĐƚůLJ ƚƌĂŶƐĨĞƌƌĞĚ ĨƌŽŵ ƚŚĞ ŐŽǀĞƌŶŵĞŶƚ ƚŽ ƚŚĞ W͘ 186 World Bank (2020b). 187 WDI 2018. 188 World Bank (2020). 114 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH dŚŝƐ ƚƌĂŶƐĨĞƌ ǁĂƐ h^Ψ800 ŵŝůůŝŽŶ ŝŶ 2012͕ ĞƋƵŝǀĂůĞŶƚ ƚŽ 0͘6 ƉĞƌĐĞŶƚ ŽĨ ƚŚĞ ĐŽƵŶ- ƚƌLJ͛Ɛ 'W͘ /Ŷ 2015͕ ŝƚ ŚĂĚ ŝŶĐƌĞĂƐĞĚ ƚŽ ĂůŵŽƐƚ h^Ψ1͕200 ŵŝůůŝŽŶ͘189 tŚŝůĞ ƚŚĞ ƐƵďƐŝ- ĚŝĞƐ ŚĞůƉ ƉŽŽƌ ŚŽƵƐĞŚŽůĚƐ͕ ĂůŵŽƐƚ ĞǀĞƌLJ ŚŽƵƐĞŚŽůĚ ƌĞĐĞŝǀĞƐ ĞůĞĐƚƌŝĐŝƚLJ ƐƵďƐŝĚŝĞƐ͕ ĐƌĞĂƟŶŐ ĂŶ ŝŵƉŽƌƚĂŶƚ ůĞĂŬĂŐĞ ƚŽ ŶŽŶ-ƉŽŽƌ ŚŽƵƐĞŚŽůĚƐ͘190 In addition, the obso- lete state-owned power plants suȂer from large ineȂiciencies, requiring almost twice the amount of gas for the generation of one unit of electricity compared to combined cycle generating plants The price paid for gas by power producers in Bangladesh has long been among the lowest in the world. At current prices, domestic reserves are declining, given lack of incentives for international private developers to invest in exploration.191 The quality of transport infrastructure needs to improve. Despite gains in connectivity, the overall quality of the road network remains poor. Congestion and ‫ޕ‬oods are a challenge for maintaining a good-quality network. Inland waterway transport terminals are inadequately equipped, and airport infra- structure and railways need upgrading and modernization due to years of insuȂicient maintenance. It is also important to improve infrastructure main- tenance practices.192 There is limited ‫ޔ‬scal space to fund Bangladesh’s transport investment needs. Annual transport infrastructure investment needs are estimated to be as high as 5 percent of GDP.193 In contrast, total public investment in infrastructure has been less than 2 percent of GDP over the last decade. High construction costs compared to other countries are a challenge for Bangladesh (Figure 2.6). The country’s elevated construction costs re‫ޕ‬ect expensive imported inputs, geo- logical factors, and governance issues in infrastructure construction.194 The lack of regular maintenance also increases costs, as the full reconstruction of poorly maintained roads is, on average, at least three times more expensive than ade- quate routine maintenance.195 189 Sadeque and Bankuti (2017). 190 HIES (2016). 191 World Bank (2020b) 192 Herrera Dappe et al. (2019). 193 Andres, Biller, and Herrera Dappe (2013); World Bank (2015). 194 Herrera Dappe et al. (2019). 195 World Bank (2005). 115 BANGLADESH RURAL INCOME DIAGNOSTICS Figure 2.6. Cost of building roads in selected countries 8000 7000 (USD per kilometer) 6000 5000 Cost 4000 3000 2000 1000 0 sh a tio n an sia ey a es in di ra sia in de rk st ne n Ch In ip s Tu ki la Ru do ill Pa ng Ph In de Ba Fe Source: Herrera Dappe et al. (2020), based on Asian Development Bank (2019). Note: Road construction costs for a four-lane urban arterial road, including traȂic-controlled intersec- tions. Market-related constraints Improving land rental markets is important to better deal with land fragmen- tation. Many farmers actively rely on land rental agreements, but land rental markets are subject to limited rental tenures and do not seem to be working eȂi- ciently. Active rental markets are important, and many landless are able to engage in agriculture by relying on rental agreements. Evidence shows that larger farm sizes are correlated with more diversi‫ޔ‬cation. However, typical land rental tenure in Bangladesh is limited to three years and largely informal, which hinders diver- si‫ޔ‬cation into certain crops that require sizeable upfront investments and a longer time to obtain returns (e.g., tree crops). Existing rental conditions also discourage the intermittent use of cover crops to replenish the soil, as farmers are not sure they will bene‫ޔ‬t. Increasing access to ‫ޔ‬nance can support agricultural diversi‫ޔ‬cation and non-farm incomes. The ‫ޔ‬nancial system in Bangladesh faces challenges to adequately support ‫ޔ‬nancial intermediation. Distortions and governance issues in the dominant banking sector, and the underdeveloped equity and bond markets, hinder ‫ޔ‬nancial intermediation, particularly for long-term ‫ޔ‬nance. Banks’ ability to provide long-term ‫ޔ‬nance is constrained by asset-li- ability mismatches. The banking sector largely relies on short-term deposits 116 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH (roughly 75 percent are for less than one year), which limits their ability to lend long term.196 Financial inclusion currently remains focused on short-term or sea- sonal loan products which oȅen have constraining eligibility requirements and high interest rates that do not re‫ޕ‬ect the actual risk of investments. Access to ‫ޔ‬nance could help tackle cash constraints that may hinder migration. For the poor, lack of liquidity and insuȂicient knowledge about the migration pro- cess seem to be important barriers to grasping migration opportunities. Two ran- domized experiments in Rangpur division that incentivized poor households to send a seasonal migrant to an urban area showed that a small cash incentive led to a large increase in the number of seasonal migrants, that migration increased house- hold expenditure by a third on average, and that households that sent a migrant continued to be more likely to migrate in future years. Liȅing the liquidity constraint allowed households to undertake a riskier choice (migration), and the knowledge gained from the process lowered costs to migrate again (despite the absence of another cash incentive). The interventions also showed that larger-scale emigration increased wages and work hours in the village of origin, indirectly bene‫ޔ‬ting other residents who stayed behind. The results suggest that it is more cost eȂective to oȂer migration loans to the same number of poor households concentrated in fewer villages than thinly spread across many villages, as having a neighbor who is migrat- ing encourages households to send a migrant, as well.197 Developing insurance markets can encourage diversi‫ޔ‬cation. Given the inherent risks involved in agriculture, the insurance market needs to develop to provide mechanisms for better risk management. Advances in this area can encourage higher-return but higher-risk activities. Increasing competitiveness in logistics service markets can contribute to reduc- ing transport costs. Trucking and inland water shipping operations are provided by thousands of operators, many of them in small ‫ޕ‬eets or single vehicles. Drivers’ unions and associations set prices and manage access to cargo. These entities limit the interaction between service providers and shippers and limit full market competition. In this context, there are limited incentives to provide high-quality services or innovate, given that providers are not rewarded for the quality of their services. Because of the low quality of services, many manufacturing ‫ޔ‬rms elect to address their logistics needs in house, which increases transaction costs.198 196 World Bank (2020b). 197 Bryan, Chowdhury and Mobarak (2014, 2016). 198 Herrera Dappe et al. (2019). 117 BANGLADESH RURAL INCOME DIAGNOSTICS The most widely used vehicle for transporting agricultural produce between major towns is a 7-ton truck which has an average cost of US$0.095/ton/km. To put these costs into perspective, comparable transport costs in India and Pakistan are US$0.025 and US$0.022/ton/km, respectively.199 Knowledge and capacity constraints Low education limits income growth. In the medium term, it is essential to continue improving rural education and skills. Low literacy makes it harder for farmers to evaluate new opportunities and learn unfamiliar techniques. Rice is a crop that farmers understand well, which likely in‫ޕ‬uences the continued prefer- ence for rice growing. Reducing knowledge gaps related to non-rice production is central to promote diversi‫ޔ‬cation. Moreover, improving productivity in non-farm activities requires tackling shortfalls in skilled labor. The demand for skilled labor in agro-processing has been accelerating with the sector’s expansion. In many cases, food processing industries cannot attain their desired productivity due to an acute shortage of skilled workers.200 In addition, low literacy and skills limit people’s access to migration opportunities. Despite substantial yield gaps in almost all non-paddy crops, agricultural R&D remains disproportionately focused on rice. The current R&D portfolio misses opportunities to improve agricultural productivity and diversi‫ޔ‬cation through knowledge creation. Bangladesh’s agricultural research expenditure is about 0.35 percent of GDP, higher than India, Pakistan, Cambodia, and Vietnam.201 However, the current R&D portfolio can be more balanced to support crops other than rice. For instance, the share of agricultural researchers engaged in horticulture research, for instance, is one of the lowest in the region. In addi- tion, greater eȂorts could also be made to support private-sector research. Lack of eȂective seed quality control measures along with existing regulatory barri- ers, particularly approval and certi‫ޔ‬cation processes for new seeds, hinder the acceleration of private-sector participation in R&D for non-rice and non-crop agriculture.202 If such barriers were removed, expanded private-sector partic- ipation could help develop varieties and agricultural technologies tailored to conditions in Bangladesh.203 199 Herrera Dappe et al. (2019). 200 Rahian et al. (2017). 201 FAO (2020). 202 World Bank (2019b). 203 World Bank (2020a). 118 ANALYZING OPPORTUNITIES AND CONSTRAINTS FOR RURAL INCOME GROWTH Inadequate extension systems weaken agricultural productivity and diversi‫ޔ‬ca- tion. Improved extension could provide more timely and relevant information to farmers. Particularly in a context of low literacy, large numbers of farmers spread across small plots pose a clear challenge for extension work. The Department of Agricultural Extension’s resources and current extension techniques and mate- rials need improvement to meet these challenges. Extension eȂorts need to reach around 15 million farms but make very limited use of aggregation models. Outreach diȂiculties are compounded by the current lack of articulation between extension eȂorts provided by the public and private sectors. Neither formal aggregation mechanisms in agriculture (e.g., cooperatives) nor informal mechanisms are widely used in Bangladesh, in contrast to many other countries. Limited aggregation mechanisms create challenges for the delivery of extension services, access to ‫ޔ‬nance, developing joint infrastructure, and aggre- gating products for marketing. These gaps lead to additional logistics costs and make it harder for local eȂorts to gain momentum. Monitoring of food safety and quality in Bangladesh is also diȂicult in part due to the lack of formal aggregation models such as producer groups, cooperatives, contract farming, and others. Accompanying recent improvements in education and basic services in second- ary cities, strengthening local government capacity is a priority for sustained economic development. Local governments in Bangladesh are a central piece of the overall governance structure and have statutory responsibility for service delivery at the village level. However, local authorities operate in a constrained policy and institutional environment. The Government of Bangladesh has made signi‫ޔ‬cant eȂorts to improve local governance systems by devolving powers to lower levels through greater autonomy and increased resources. However, chal- lenges in governance, resource management, and institutional capacity remain a major constraint.204 Gender norms Female labor force participation is constrained by both demand and supply fac- tors. The main constraints include: diȂerential access to productive inputs and assets compared to men; women’s role in home-based and caretaking activities; market failures and institutions; and social norms constraining women’s mobility.205 204 World Bank (2017). 205 Rahman and Islam (2013). 119 BANGLADESH RURAL INCOME DIAGNOSTICS Women’s role in home-based and caretaking activities reduces their participa- tion in education and paid work. The prevalence of early marriage is still high in Bangladesh compared to other South Asian countries, which increases the likeli- hood that women will leave school or exit the labor force due to increasing house- hold responsibilities. About 1 in 3 women (30 percent) not attending secondary school cite family chores and marriage as the reasons for not attending.206 These factors also reduce women’s rate of engagement in paid employment. Restrictions on women’s ability to inherit property and limited access to credit constrain women’s participation in the labor market.207 The sharia law applying to the Muslim majority prescribes that sons inherit twice as much as daughters, while for Hindu and Buddhist minorities daughters have no inheritance rights. This might aȂect women’s capacity to access credit because of the resulting lack of collateral.208 Evidence suggests that increasing women’s access to microcredit has contributed to raising their income-generating capacity and empowerment.209 The development of micro‫ޔ‬nance has helped increase female labor participation by expanding opportu- nities for women in rural areas to create businesses in livestock, poultry, and small textiles. The presence of NGOs employing women as health workers, community leaders, and teachers has also contributed to greater female labor participation.210 However, challenges remain in women’s control of ‫ޔ‬nances. Mobility constraints limit women’s ability to engage in non-farm activities and the commercialization of farm products. About 7 in 10 rural women work in their residence or in areas near their homes, partly re‫ޕ‬ecting mobility constraints.211 Women are also less likely than men to sell produce in markets.212 206 Genoni et al. (2019). 207 Cameron, Dowling, and Worswick (2001); Gonzales et al. (2015); Kabeer (1990); Amin (1997); World Bank (2012); Klugman et al. (2014); Srivastava and Srivastava (2010); NeȂ, Sen, and Kling (2012); SolotaroȂ et al. (2019). 208 Raihan and Bidisha (2018). 209 Rahman and Islam (2013). 210 Khandker, Samad, and Khan (1998); Raihan and Bidisha (2018). 211 World Bank (2012); Amin (1997); Kabeer (2001); Asadullah and Wahhaj (2016). 212 Anderson and Eswaran (2009). 120 SECTION 3. Setting priorities for action: Which constraints should be tackled first? Since there are many constraints on rural income growth, it is important to set priorities for which should be addressed ‫ޔ‬rst. The previous section outlined a series of constraints that, if addressed, could help Bangladesh take advantage of opportunities to boost rural incomes. This section of the RID prioritizes these con- straints based on two main criteria: (i) strong evidence that current performance in the area is falling short; and (ii) strong evidence that tackling the constraint can substantially improve rural income growth. The evidence used in the assessment comes from two sources: the descriptive analysis undertaken for this study and an extensive literature review. The prioritization of constraints considers the strong linkages between agricul- ture and non-agricultural domains. As the previous sections showed, a potential way to increase non-farm incomes, given strong agglomeration economies and the spatial distribution of ‫ޔ‬rms, is to expand activities along agricultural value chains, such as agro-processing. Increasing productivity and diversi‫ޔ‬cation in agriculture can enhance value addition. Therefore, the prioritization presented in this section primarily emphasizes agricultural constraints related to those issues. One underlying element behind the identi‫ޔ‬cation of constraints is their ability to build resilience in a context of high risks and climate change. Additionally, the analysis focuses on tackling elements that allow for a levelled playing ‫ޔ‬eld across agricultural products and economic sectors. Additional factors considered in prioritization include the time likely to be required to address a constraint and evidence that the constraint constitutes a bottleneck: that is, that other obstacles to rural income growth can be more easily removed if this barrier is tackled ‫ޔ‬rst. In some cases, constraints aȂect a compar- atively narrow range of rural income opportunities but are prioritized because of the large share of households that could bene‫ޔ‬t from their removal. 121 BANGLADESH RURAL INCOME DIAGNOSTICS Table 3. 1 summarizes priority constraints. Other constraints that have not been assigned priority for action are summarized in Table 3.2. The constraints in Table 3.2 are very important but have not been prioritized due to the time they may take to resolve or the lack of evidence that tackling them rapidly would increase rural incomes in the time horizon considered by this diagnostic. As discussed, key ele- ments related to human capital development, macro and business environment, and governance are outside the scope of this prioritization exercise. Policy focus and regulations Among the constraints identi‫ޔ‬ed in the previous section, policies around rice production are especially pivotal. A strong policy focus on supporting rice pro- duction has been important to raise production of this staple food and ful‫ޔ‬ll Bangladesh’s goal of reducing food insecurity risks. However, as income growth and urbanization continue, demand is expected to diversify to other, more nutri- ent-rich and higher-value foods. To meet this growing demand, policies need to be fully aligned to sustain growth in productivity as well as a more diversi‫ޔ‬ed agriculture. Evidence indicates that, given changing demand patterns and higher incomes, this can be done without compromising the food-security goal. Fertilizer subsidies are a speci‫ޔ‬c area where budget resources could be repur- posed for higher productivity. Currently, these subsidies are being overused without generating additional returns. Important bene‫ޔ‬ts can also be obtained by rebalancing R&D toward non-rice crops. A stronger seed market can support additional gains in productivity. A rebalanced policy environment could incentiv- ize farmers to diversify their production to maximize expected returns, without necessarily disincentivizing rice production. Given the substantial bias of current agricultural policies, this shiȅ has signi‫ޔ‬cant potential to raise productivity and spur diversi‫ޔ‬cation. Investments in transport policy merit priority. Tackling coordination issues for better infrastructure maintenance, improved logistic services, and eȂective enforcement of existing regulations can enable higher returns on ongoing trans- port investments by reducing congestion costs. These costs are an important component of the transaction costs in key value chains important for diversi‫ޔ‬- cation (e.g., horticulture, livestock, ‫ޔ‬shing) and constrain the growth of urban areas outside Dhaka and Chittagong. Thus, a reduction of these transaction costs can support the development of high-potential value chains and boost non-farm income growth across the country. 122 SETTING PRIORITIES FOR ACTION Investment gaps Water management, transport, and electricity remain central for rural income growth. Given ‫ޕ‬ood threats and the need for reliable water supply in oȂ-monsoon seasons, better irrigation and drainage remain central for productivity and diver- si‫ޔ‬cation in agriculture. Investments in transportation and electricity are also key, although the Government of Bangladesh has taken action to narrow these gaps. Transport and electricity investments should focus on closing remaining coverage gaps while supporting infrastructure maintenance and the quality of electricity. Market-related constraints Within market constraints, evidence indicates that improving the functioning of land markets is central. This is a fundamental constraint, given the country’s average land size and limited room for expansion. Land markets need to become more eȂicient to help create economies of scale for investment and more formal commercialization of production and value addition. Even though credit markets are underdeveloped, in particular for small businesses and farmers aiming to scale up productive activities, they appear more developed than in comparator countries, and there is no clear evidence that tackling this constraint alone can boost income growth. Therefore, the issue of credit markets is not prioritized. Given substantial climate-related risks in Bangladesh, insurance markets need to develop to provide mechanisms for better risk management that can encourage higher-return, higher-risk activities. However, given the extent of climate-related risks, we prioritize areas that can help to build resilience. Knowledge and capacity constraints The adult population in Bangladesh’s rural areas still has a low human-capi- tal base. This constrains people’s opportunities to adopt more productive agricultural activities and access sources of non-farm income. Fully resolving human-capital constraints, including closing skill gaps, will take time. Areas that can be tackled in the short term include strengthening extension services to achieve better use of inputs and promote diversi‫ޔ‬cation. Gender norms Despite improvements, women’s potential for income growth remains largely untapped. Norms around asset ownership, mobility, and early responsibilities in care-giving activities constrain women’s ability to engage in more productive, 123 BANGLADESH RURAL INCOME DIAGNOSTICS paid farm and non-farm activities. Some of these elements will move slowly with time. However, without tackling the speci‫ޔ‬c constraints women face, addressing the other constraints previously discussed may not result in income growth for Bangladeshi women. Therefore, we prioritize these constraints, as a large share of the population is expected to bene‫ޔ‬t from them. Table 3.1. Constraints recommended for priority action Income growth Strength of evidence: Constraint drivers aȂected Size and impact of constraint Policy focus and regulations 1. Policy focus on support- Productivity • Strong evidence of substantial bias. ing rice Diversi‫ޔ‬cation • Less evidence concerning impact on Input subsidies, output growth. Given the size of the bias, price support, and price impact is expected to be large. stabilization. • Very important for areas still heavily Research and development invested in rice but that have potential for diversi‫ޔ‬cation: particularly the have been centered around Northwest and areas southeast of paddy rice. Chittagong. 2. Seed market regulations • Evidence that regulations constrain Constrain supply of quality Productivity access to and use of quality seeds. seeds. • Evidence that past eȂorts on this front Diversi‫ޔ‬cation were bene‫ޔ‬cial for productivity. 3. Transport policy Productivity • Strong evidence that these issues are Current policies lack coor- Diversi‫ޔ‬cation present and constrain connectivity. dination and are largely • Less evidence that liȅing the coordi- Non-farm income focused on infrastructure nation constraints alone will change development, neglecting income growth. Given the size of the quality. Enforcement of problem, impact is expected to be large. road safety is weak. Current policies constrain connectivity by increasing congestion and transaction costs. Opportunity to rebalance focus from development of infrastructure to mainte- nance and quality. 124 SETTING PRIORITIES FOR ACTION Income growth Strength of evidence: Constraint drivers aȂected Size and impact of constraint Investment gaps 4. Electricity and transport Productivity • Strong evidence that these gaps are Diversi‫ޔ‬cation present and constrain income growth. • Evidence that past expansions have Non-farm income been linked to higher incomes and lower poverty. 5. Irrigation and drainage Productivity • Strong evidence that these gaps are Irrigation gaps in Rangpur. Diversi‫ޔ‬cation present and constrain productivity and diversi‫ޔ‬cation. Shortage of reliable water. • Evidence that past eȂorts on this Lack of good drainage and front have enhanced productivity and ‫ޕ‬ood protection. diversi‫ޔ‬cation. Market-related constraints 6. IneȂicient land markets Productivity • Strong evidence that farm fragmen- Severe land fragmentation. Diversi‫ޔ‬cation tation dampens productivity and investment. Short rental terms. Gender Female income gaps. • Given evidence that larger farms are Non-farm income more eȂicient, impacts are expected to be important. • Important across all areas, particu- larly in the Western divisions and for women. Knowledge and capacity constraints 7. Weak extension services Productivity • Evidence that this is a gap. Inadequate extension sys- Diversi‫ޔ‬cation • Evidence that addressing it would tems in a context of small improve productivity. farms and low literacy. Gender ч. Gender norms Productivity • Strong evidence of a substantial gap. Strong norms around asset Diversi‫ޔ‬cation • Strong evidence of lower participation ownership, mobility, and and lower productivity due to gen- Female income caregiving responsibilities. der-biased selection into activities. Women are largely involved • Given the size of the gap, income in livestock but not in its eȂects are expected to be large. commercialization. 125 BANGLADESH RURAL INCOME DIAGNOSTICS Table 3.2. Constraints not prioritized but important to tackle Income growth Strength of evidence: Constraint drivers aȂected Size and impact of constraint 1. Low literacy and skill Productivity • Strong evidence that a large gap levels persists. Diversi‫ޔ‬cation • Evidence that higher education Low levels of education expands farm and non-farm Non-farm income limit rural people’s ability to income, migration, and paid female adopt technologies, pursue Migration employment – all linked to higher nonfarm activities, Female income incomes and less poverty. or migrate. • This is very important but a slow-moving constraint. Demand- side elements also need to be addressed to be able to absorb a higher-skilled population. 2. Weak food safety Diversi‫ޔ‬cation • Strong evidence that there is a gap. standards • Scant evidence to justify prioritiza- Non-farm income tion, as traditional markets seem Monitoring of food safety to work well, and current demand and quality pose challenges. is largely driven by the domestic Compliance monitoring and market. law enforcement remain • In more diversi‫ޔ‬ed areas, this weak and irregular. also involves livestock and ‫ޔ‬sh producers. 3. Uncompetitive logistic Productivity • Strong evidence that there is a services large gap. Diversi‫ޔ‬cation • Strong evidence that this increases Lack of competition results transaction costs. Non-farm income in low-quality services. • Limited evidence on the impact on rural incomes if tackled in isolation. 4. Concentration of Migration • Strong evidence that migration is migration networks concentrated spatially. • Evidence that migration supports The spatial pattern of higher incomes and less poverty. migration creates network • This is a slow-moving constraint. eȂects. Networks need to be fostered in current low-migration areas to lower migration costs. 126 SETTING PRIORITIES FOR ACTION Income growth Strength of evidence: Constraint drivers aȂected Size and impact of constraint 5. Import and export Productivity • Evidence that existing policies may policies related to create contradictory incentives Diversi‫ޔ‬cation mechanization around mechanization. • Less evidence of impact regarding Constrain balanced income growth. mechanization. • Unclear from the evidence what types of machinery are most relevant and what threshold scale would make investment in expen- sive livestock machinery viable. 6. Credit and insurance Productivity • Strong evidence that credit mar- markets kets are underdeveloped for micro Diversi‫ޔ‬cation enterprises and farmers. Collateral requirements, • Strong evidence that insurance Migration high interest rates, and poor markets are underdeveloped. diversity in ‫ޔ‬nancial prod- Female income • Strong evidence that liȅing liquid- ucts and services. Access to Non-farm income ity and insurance constraints has ‫ޔ‬nance challenges SMEs. positive impacts on investment. Underdeveloped insurance • Impacts across all areas and for markets. Liquidity con- women. straints for migrants. • However, no clear evidence that tackling these elements in isolation will boost rural incomes. 7. Weak local governance Non-farm income • Strong evidence of a substantial bias. Weak local governance • Less evidence of growth impacts. around service delivery con- Recent trends in urban poverty strains locally driven growth. performance, growth in agro-busi- nesses, and ‫ޔ‬rm reallocation suggest a large eȂect. • This is a slow-moving constraint. 127 SECTION 4. Identifying policy solutions This section outlines a series of policy lines that can help in liȅing the priority constraints on rural income growth identi‫ޔ‬ed in the report. Two main criteria were used to select policy areas: (i) The measures are supported by a strong evi- dence base; (ii) The policies can achieve strong results within ‫ޔ‬ve years. In the cur- rent state of evidence, some policy directions are clearer and more speci‫ޔ‬c than others.213 On some fronts, such as tackling gender norms or strengthening insur- ance markets, continued experimentation can be important to tailor approaches to the country context and for scalability. The focus of this analysis is on boosting rural household incomes in the short to medium term. Longer-term eȂorts related to building human capital, improv- ing the business environment, or addressing national governance constraints are clearly important but lie outside the scope of this study. Evidence of existing successful approaches in Bangladesh was collected from evaluation documents, recent high-quality assessments, and through stakeholder consultations. Results were cross-referenced with extensive literature reviews on the eȂectiveness of existing policies and potential alternatives. The analysis highlights areas where evidence suggests that the policies now being used to improve rural incomes in Bangladesh can be enhanced to achieve greater impact. Liƚ constraints related to polic4 focus and regulations Review and reform input subsidies, especially on fertilizers. Leverage extension services to make fertilizer use more eȂicient. These actions could help remove two important constraints on higher farm yields, especially for boro paddy: (i) overuse and imbalanced use of fertilizer, resulting in declining soil fertility; and (ii) inadequate farm knowledge and practices. 213 The COVID-19 crisis has limited the country-level consultations that could be undertaken to inform these policy recommendations. To address that challenge, this report has relied on remote consulta- tions and information from Government of Bangladesh reports (including discussions for the prepara- tion of the 2020 Bangladesh Development Forum to inform the country’s next ‫ޔ‬ve-year development plan), together with recent World Bank assessments that were based on more extensive consultations. 129 BANGLADESH RURAL INCOME DIAGNOSTICS Rebalance the agricultural R&D budget toward non-rice products. The domi- nant focus of the current R&D budget is on paddy rice, despite substantial yield gaps aȂecting many non-rice outputs. Support greater coordination between government and the private sector on research priorities. Through research institutes and universities, the government has a large role in shaping research priorities in agriculture. Greater coordination between private- and public-sector agencies would help inform the choice of these priorities. This is already being done to some extent via organizations such as the Krishi Gobeshona Foundation, which provides a platform for coordination and information sharing on climate change.214 Revisit regulations and actions to expand private participation in the seed sector. Seed sector liberalization during the late 1990s contributed to important achievements in diversi‫ޔ‬cation toward higher-value crops. However, there are still constraints on private-sector investment in this area. Removing them will require revisiting regulations that prevent private-sector actors from breeding and producing seeds for noti‫ޔ‬ed crops;215 increasing public-private coordination on quality seed supply for noti‫ޔ‬ed crops; and increasing market transparency regarding the quality and eȂectiveness of seeds in the market. Use policy and regulation to reduce high transport and logistics costs.216 Some policy areas that can be tackled in the next ‫ޔ‬ve years include: • Assess roles to improve coordination across ministries and agencies with a voice in transport. At the national level, there are multiple ministries and agencies responsible for the transport sector with overlapping functions. DiȂerent agencies have their own modal infrastructure development mas- terplans, with limited coordination between them. • Review outdated policies and regulations to move beyond infrastructure development, strengthening focus on the quality of services. Current sectoral policies do not focus on upgrading services. Modern national-level policies in other countries seek to improve infrastructure, integrate modes of transport, and improve the quality of services. 214 World Bank (2020a). 215 Noti‫ޔ‬ed crops are rice, wheat, potato, jute, and sugarcane, which are considered important in terms of food and national security. Seed segments for these crops are highly regulated. 216 These recommendations build on the extensive analysis conducted by Herrera Dappe et al. (2019). 130 IDENTIFYING POLICY SOLUTIONS • Fully implement the Competition Act, which oȂers legal instruments to prevent collusion in the logistic services market and price ‫ޔ‬xing by industry players. • Step up enforcement of regulations to prevent road accidents and over- loading of trucks. Regulations regarding the licensing of drivers and vehi- cles are meant to reduce accidents, but weak enforcement defeats the pur- pose of the regulations. Weak enforcement of regulations has resulted in a large majority of drivers being unskilled, substandard trucks on the road, and overloading of cargo. Liƚ constraints related to investment gaps Invest in high-eȂiciency irrigation coverage and ‫ޕ‬ood protection, particularly in hilly and highland divisions. Invest in good drainage, particularly in low-lying areas where production of non- paddy crops is diȂicult even with good drainage. Continue eȂorts to promote the shiȅ away from boro paddy, as this crop is rap- idly depleting groundwater storage. Invest in closing electricity gaps and improve the reliability of the supply. This includes leveraging private investment in generation to further increase power supply from cheap, clean energy sources and strengthen transmission and distri- bution capacity.217 Invest in transport infrastructure, with an important focus on maintenance. Liƚ constraints related to markets Revisit land markets and tenures. Review the agricultural land rental market to identify speci‫ޔ‬c market failures that impede longer-term rentals of agricultural land. Liƚ constraints related to knowledge and capacities Revise budget and capacities to deliver extension services. The current DAE sys- tem lacks adequate resources to deliver services to the high number of Bangladeshi farmers who could bene‫ޔ‬t. There is also a need to train DAE extension staȂ on new 217 World Bank (2020b) 131 BANGLADESH RURAL INCOME DIAGNOSTICS agricultural knowledge and techniques for delivery training. More support is needed for private sector-led training programs (e.g., PRAN). Closer integration between DAE and the private sector could get greater impact from limited resources.218 Liƚ constraints related to gender norms219 Despite improvements, women’s potential for income growth remains largely untapped. Norms around asset ownership, mobility, and early caregiving respon- sibilities constrain women’s ability to engage in more productive paid farm and non-farm activities. While there is evidence of progress, particularly for younger cohorts, liȅing these constraints will require a comprehensive approach that tack- les many barriers at once. In the short run, policy levers include: Invest in programs to make women’s participation in livestock more productive. The livestock subsector is important, given the large share of women involved. Expanding the use of mechanized operations could bring substantial bene‫ޔ‬ts. Engage men and communities to promote gender equality. There is evidence from evaluations in Bangladesh and other countries that interventions engaging young men and communities to promote gender equality can: support women’s empowerment in ‫ޔ‬nancial decision making,220 increase women’s registration of land,221 change views about women’s asset ownership,222 and promote better reg- ulations on women’s land rights.223 These eȂorts can also be tailored to empower women around value chains, particularly in livestock, where gender norms con- strain their more productive participation. Institute legal reform for women’s inheritance of land, and address social norms that discourage women’s land ownership and registration. Recent strategies with evidence of success include: • Reduce land registration fees for women and streamline the registration process. This strategy has been used successfully in Nepal and India and could provide lessons for Bangladesh. 218 World Bank (2020a). 219 These recommendations build on the extensive analysis and consultations by SolotaroȂ (2019). 220 Doyle et al. (2018). 221 Lastarria-Cornhiel et al. (2014). 222 Jahan et al. (2016). 223 Mueller et al. (2017). 132 IDENTIFYING POLICY SOLUTIONS • Promote joint titling of land and property. Some evidence suggests that land formalization programs requiring joint registration between both spouses can support female land ownership (e.g., Philippines, Vietnam, India).224 A quasi-experimental study of Rwanda’s pilot program, in which married females were registered as co-owners of land by default, showed signi‫ޔ‬cant impacts on women’s land rights among married-couple households.225 Target interventions to increase women’s control of ‫ޔ‬nancial resources. Evidence from multiple countries suggests that actions need to tackle not just market failures related to banking institutions or products, but also constraints related to gender norms.226 Programs also need to consider the potential risks of altering intra-household dynamics. Thus, experimentation and learning are a necessary ‫ޔ‬rst step to scale up interventions. Experimentation can be rapidly pur- sued in two areas, based on evidence of prior success: • Programs to provide savings and other ‫ޔ‬nancial products solely in wom- en’s names. There is substantial evidence that such programs can support women’s empowerment, promote acceptance of women’s control over ‫ޔ‬nancial resources, loosen restrictions on women’s mobility, and encour- age their engagement in other economic activities.227 • Models to incentivize women’s use of mobile technologies. Despite the expansion of mobile banking in Bangladesh, women’s use of these tools is very limited. Recent studies suggest that improving women’s use of mobile phones for banking can increase their ‫ޔ‬nancial inclusion.228 The ability to manage ‫ޔ‬nances through mobile ‫ޔ‬nancial services rather than physically traveling to a bank branch can help overcome constraints related to mobil- ity and time.229 However, evidence on eȂective interventions to promote uptake is limited. 224 SolotaroȂ (2019). 225 Ali et al. (2015). 226 Banerjee et al. (2015); Karlan et al. (2016). 227 Ambler and De Brauw (2017); Cheema et al. (2016); Field et al. (2016). 228 Gammage et al. (2017). 229 SolotaroȂ (2019). 133 Closing remarks This Rural Income Diagnostic (RID) set out to answer the question: “What are the main opportunities and constraints to faster, sustained income growth for poor and vulnerable households in rural areas of Bangladesh?” The objective was not to examine how to increase rural GDP, but to show how people living in rural areas can enjoy higher incomes in the near future. The diagnostic incorporated results from a series of recent technical assessments, complementing these studies with descriptive analysis and benchmarking. The analysis shows that Bangladesh has been on the right path for poverty reduc- tion, but that the country now faces new challenges that will require renewed eȂorts. Bangladesh’s gains against poverty have been impressive but also poten- tially fragile, as many families continue to move in and out of poverty and have limited means to cope with income shocks, which are common in a country highly exposed to extreme weather events and climate change. Rural incomes have been rapidly transforming, and many households now depend on the good performance of both agriculture and non-agricultural sec- tors. However, the transformation is not complete: rice still largely dominates agricultural production, and growth in the sector has recently been slow and less poverty reducing. OȂ-farm income opportunities and the bene‫ޔ‬ts of urbanization have been more limited in some areas, particularly the Western divisions and areas outside the Dhaka-Chittagong corridor. In addition, the earning potential of half of Bangladesh’s working-age adults, women, remains largely untapped. Progress on closing spatial disparities and achieving an inclusive economic transformation requires a balanced rural development strategy that accelerates growth in agriculture but also increases opportunities for non-farm and migration income. This strategy needs to pay special attention to location in a context of growing spatial income disparities, and it needs to bolster rural women’s capacity to earn higher incomes. 135 BANGLADESH RURAL INCOME DIAGNOSTICS Among the central results emerging from the RID is that agriculture needs to grow faster and become more poverty reducing. Despite Bangladesh’s remarkable economic transformation, a large share of households in rural areas still depend on agriculture for the bulk of their income. The recent slowdown in agricultural growth is an important reason why poverty has stagnated or increased in Western areas of the country, so the performance of agriculture is also central for reducing spatial welfare disparities. Boosting growth in agriculture has additional signif- icance, because the sector generates a multiplier eȂect for growth in rural non- farm incomes. This is particularly important to reduce spatial inequalities, given the current distribution of non-farm activities across the country. Considering Bangladesh’s limited scope for arable land expansion and the current high inten- sity of land use, agricultural growth will need to arise from a combination of higher agricultural productivity, greater diversi‫ޔ‬cation toward high-value crops, and modernizing the agri-food supply chain in the medium term. EȂort to increase productivity in agriculture have to be accompanied by a more balanced urbanization process driven by local comparative advantages that can expand oȂ-farm income opportunities. The bene‫ޔ‬ts of manufacturing-driven growth have been largely concentrated around the Dhaka-Chittagong corridor. In addition to broad eȂorts to increase high-quality job growth across urban and rural areas, supporting the development of food processing and higher-value agri- culture can support higher income opportunities in the rural non-farm sector that are less connected to the main cities. The RID has identi‫ޔ‬ed a series of key constraints aȂecting rural income growth whose removal can accelerate Bangladesh’s development. Needs already high- lighted in the literature include rebalancing agricultural policies to incentivize faster agricultural diversi‫ޔ‬cation. Also crucial are the government’s ongoing eȂorts to close infrastructure gaps in irrigation, transport, and electricity. However, the analysis presented here shows that these actions alone, while vital, may not suȂice. Policy also has to tackle constraints that have been less widely recognized to date. First, while investments in education continue to gradually improve Bangladesh’s human-capital base, innovative short-term solutions are needed to close knowledge gaps in a context of high illiteracy. Land markets also need to work better to overcome land fragmentation. Finally, gender norms remain a fundamental constraint holding women back from increasing their earn- ings and contributing all they could to Bangladesh’s progress. 136 CLOSING REMARKS As the country’s new ‫ޔ‬ve-year development plan shows, many government ini- tiatives are moving assertively in the right direction. The RID maps directions for action that can inform and support government eȂorts in the short term, based on the best available evidence. At the same time, the diagnostic makes clear that Bangladesh’s leaders will need to go beyond known solutions. In an uncertain global context, the nation’s policy makers will be called on to set forth their prac- tice of bold experimentation. Bangladesh’s next generation of policy solutions in rural development will re‫ޕ‬ect a more spatially tailored approach able to reduce poverty, protect the vulnerable, and create conditions for inclusive income and welfare gains throughout the country. 137 APPENDIX 1 Summary statistics by quintile 1 2 3 4 5 All BIHS 201ч Average monthly per capita income 2,900 3,675 4,599 5,683 7,777 4,879 Median monthly per capita income 2,433 3,164 3,806 4,456 6,089 3,701 Average monthly per capita consumption 2,283 3,202 4,062 5,211 8,622 4,630 Median monthly per capita consumption 2,159 2,949 3,755 4,873 7,375 3,716 HIES 2016 Average monthly per capita consumption 1,730 2,432 3,084 4,014 7,585 3,756 Median monthly per capita consumption 1,794 2,434 3,071 3,970 6,149 3,079 Source: BIHS 2018 and HIES 2016. Bangladeshi Takas. APPENDIX 2 Summary statistics by division Barisal Chittagong Dhaka Khulna Rajshahi Rangpur Sylhet All BIHS Average monthly 3,949 4,604 4,732 5,381 6,093 4,649 3,776 4,880 per capita income Median monthly 3,200 3,333 3,666 4,250 4,435 3,666 3,000 3,701 per capita income Average monthly per capita 4,651 5,727 4,748 4,329 4,189 3,699 4,646 4,630 consumption Median monthly per capita 3,570 5,414 3,877 3,386 3,404 2,962 3,233 3,716 consumption HIES 2016 Average monthly per capita 3,734 4,043 3,875 3,730 3,553 2,982 4,352 3,756 consumption Median monthly per capita 3,148 3,475 3,140 3,042 2,913 2,428 3,580 3,079 consumption Source: BIHS 2018 and HIES 2016. Bangladeshi Takas. 139 BANGLADESH RURAL INCOME DIAGNOSTICS APPENDIX 3 Opportunities and constraints in the livestock and fisheries sectors Given recent attention to Bangladesh’s livestock and ‫ޔ‬sheries subsectors, includ- ing substantial public investment to address key challenges, the RID has not conducted any additional analysis in this area. This appendix summarizes recent evidence on opportunities and constraints for these subsectors. Livestock subsector The livestock sector has experienced relatively low growth over the last decade. According to government data from the Department of Livestock Services (DLS), Bangladesh’s livestock population grew at an annual rate of about 2.24 percent from 2010-11 to 2019-2020, with the major growth coming from the poultry sector. In the same period, the annual growth rates for cattle, goats, and other ruminant populations were 0.54 percent, 0.91 percent, and 0.79 percent, respectively. In fact, the poultry subsector has shown real dynamism in recent years, with an expansion of commercial poultry production. Between 2001 and 2011, broiler meat production and egg production grew 8.8 percent and 7.7 percent annually.230 The poultry industry is moving towards self-suȂi- ciency in meeting local demand for meat and eggs. In contrast, annual growth in milk productivity was negative, at 1.5 percent, between 2000 and 2016.231 As the livestock subsector grew slowly in recent decades, imports of dairy products have been increasing. In FY15-16 alone, the country spent around US$249 mil- lion to import dairy products. Livestock supply chains in Bangladesh are largely informal and depend on smallholders, typically with very low productivity levels. Seventy per- cent of dairy farmers are small holders, who together produce 70-80 per- cent of the country’s total milk supply. The red meat value chain is the least developed, with only two formal meat processing enterprises operating in the country, whereas the poultry chain is the most commercially developed. The milk supply chain is also dominated by informal traditional markets, and the presence of formal commercial processors in the value chain is limited. 230 Hamid et al. (2017). 231 Masylkanova (2020). 140 APPENDIX The few commercially developed value chains for red meat, poultry, eggs, and milk are centered around the big cities (Dhaka and Chittagong). Despite its low contribution to GDP, the livestock sector employs a large share of Bangladesh’s labor force and oȂers signi‫ޔ‬cant opportunities for job creation. As incomes rise and diets shiȅ to animal protein, demand for livestock products is growing rapidly. The sector can play a critical role in improving rural incomes and nutritional status in the country. Unlocking the growth potential of the livestock sector will require addressing the following challenges: i) poor animal husbandry practices and low penetration of high-yielding breeds; ii) the safety and eȂiciency of value chains are limited by poor post-harvest and market infrastructure; iii) a high incidence of diseases results in high rates of animal losses; iv) inadequate food safety and quality control result in substantial public-health risks; v) weak livestock extension and advisory services, along with lack of coordination among public and private actors; and v) the intensity of greenhouse gas emission in live- stock production is at unsustainable levels, contributing to an outsized environ- mental footprint for the sector. Fisheries subsector The ‫ޔ‬sheries subsector showcases a private sector-led diversi‫ޔ‬cation success story, with Bangladesh having become the world’s third-largest producer of inland capture and ‫ޔ‬ȅh in inland culture ‫ޔ‬sheries.232 The country has massive marine, coastal, and inland water resources for ‫ޔ‬sh production. Within the sub- sector, inland capture ‫ޔ‬shery is the largest component, accounting for 56 per- cent of total production and 3.91 million ha of water spread areas (WSA). Inland aquaculture is the second largest component, at 28 percent of total production and WSA of 0.79 million ha. Marine ‫ޔ‬shery accounts for 15 percent of national production. Commercial shrimp production has been scaled up considerably and generates the country’s third-highest export earnings, aȅer the textile sec- tor and the leather sector. Species of shrimp and prawn with higher commercial values were adopted. The extent of marine ‫ޔ‬shery is likely to increase, as the country has negotiated a new maritime boundary with India, which provides an additional 11.9 million ha of maritime water with great potential for the expan- sion of marine ‫ޔ‬shery. 232 FAO (2018). 141 BANGLADESH RURAL INCOME DIAGNOSTICS Inland aquaculture has undergone a substantial transformation in the last three decades. The farmed-‫ޔ‬sh market experienced a 25-fold increase during this period, driven by domestic demand.233 Between 2004 and 2014, ‫ޔ‬shpond area and the number of ‫ޔ‬sh farmers increased by 30.4 percent and 63 percent, respectively, while ‫ޔ‬sh production rose by 117.4 percent.234 This impressive success in aquacul- ture was mainly driven by the private sector. Numbers of wholesale markets, feed dealers, and ‫ޔ‬sh traders have more than doubled, and capital intensity in the sector has increased substantially over the period. Expansion of the rural road network, rising aȂordability, and access to new technology have been the crucial factors enabling sustained high growth of inland aquaculture. The domestic market has been the main trigger behind aquaculture transformation in Bangladesh, as over 90 percent of farmed ‫ޔ‬sh (excluding shrimp) are sold on the domestic market. Despite remarkable performance, sustainability of inland capture and cul- ture ‫ޔ‬sheries is at risk due to environmental and ecosystem degradation that has resulted from destructive ‫ޔ‬shing practices, over‫ޔ‬shing of species with high commercial values, and poor resource management.235 Climate change poses additional risk to the sector’s sustainability, especially for coastal aquaculture. The Bangladesh Delta Plan 2100 (BDP2100) emphasized the development of sus- tainable, socially acceptable, economically viable, and environmentally friendly technology to support the continued development of aquaculture. The BDP2100 highlights two constraints for aquaculture in the country. Coastal aquaculture (shrimp, ‫ޔ‬n‫ޔ‬sh, crab, and others) is not ‫ޕ‬ourishing in line with its potential, due to poor investment. In addition, extension services provided to aquaculture farmers remain inadequate. 233 Rashid and Zhang (2019). 234 Rashid and Zhang (2019). 235 UNCTAD (2017). 142 References Aheeyar, Mohamed, Sanjiv de Silva, and Sonali Senaratna Sellamuttu. 2019. “Making index-based ‫ޕ‬ood insurance socially inclusive in Bangladesh: challenges and options.” Colombo, Sri Lanka: International Water Management Institute (IWMI). CGIAR Research Program on Water, Land and Ecosystems (WLE). 6p. (WLE Brie‫ޔ‬ng Series No. 25). Ahmed, Akhter U. 2016. “Description of the Bangladesh Integrated Household Survey.” USAID and IFRPI, Washington, DC. Ahmed, F., D. Roy, M. Yanez-Pagans, and N. Yoshida. 2017. “Design of a Multi-Stage Strati‫ޔ‬ed Sample for Poverty and Welfare Monitoring with Multiple Objectives: A Bangladesh Case Study.” World Bank Research Working Paper, World Bank, Washington, DC. Ahmed, F., Y. Arias-Granada, M.E. Genoni, M. Yanez-Pagans, N. Yoshida, D. Roy, and A. Latif. 2019. “Description of the OȂicial Methodology Used for Poverty Estimation in Bangladesh for 2016/17.” Background paper for the Bangladesh Poverty Assessment 2010-2016/17, World Bank, Washington, DC. Ahmed, M., and M. Gautam. 2020. “What Drives Rural Incomes in Rural Bangladesh? The Evolution of Farm, Non-Farm and Migration Incomes.” Background paper for “Bangladesh Rural Income Diagnostic,” World Bank, Washington, DC. Ahmed, Mansur; Saint-Geours, Jean; Gitau, Ciliaka. 2021. "Promoting Agrifood Sector Transformation in Bangladesh : Policy and Investment Priorities." International Development in Focus;. Washington, DC: World Bank. Alam, Monjurul, and Imanun Nabi Khan. 2017. “Agricultural Mechanization: Status, Challenges and Opportunities in Bangladesh.” Chapter 2 in T.R. Gurung, W. Kabir, and S.M. Bokhtiar (eds.). " %+&71&,+#,/201&+)"$/& 2)12/) +1"+0&Ɯ 1&,+&+ "$&,+ǽ SAARC Agriculture Centre, Dhaka, Bangladesh, pp. 41-70. 143 BANGLADESH RURAL INCOME DIAGNOSTICS Alam, Shamsul, Jaap de Heer, and Giasuddin Choudhury, eds. 2018. “Bangladesh Delta Plan 2100: Baseline Studies on Agriculture, Food Security and Nutrition,” Volume 4. General Economics Division, Bangladesh Planning Commission, Ministry of Planning, Government of the People’s Republic of Bangladesh. Ali, Daniel Ayalew, Klaus W. Deininger, Markus P. Goldstein, and Eliana La Ferrara. 2015. “Empowering Women through Land Tenure Regularization: Evidence from the Impact Evaluation of the National Program in Rwanda.” Development Research Group Case Study. Washington, DC: World Bank. Ambler, Kate, and Alan De Brauw. 2017. “The Impacts of Cash Transfers on Women’s Empowerment: Learning from Pakistan’s BISP Program.” Social Protection and Labor Discussion Paper 1702, Washington, DC: World Bank. Amin, S. 1997. “The poverty–purdah trap in rural Bangladesh: implications for women’s roles in the family.” "3"),-*"+1+!%+$" 28(2): 213-233. Andres, L. A., D. Biller, and M. Herrera Dappe. 2013. "!2 &+$,3"/166),0&+$,21%0&ȉ0 Infrastructure Gap. Washington, DC: World Bank. Asadullah, M. Niaz, and Zaki Wahhaj. 2016. “Missing from the Market: Purdah Norm and Women’s Paid Work Participation in Bangladesh.” IZA Discussion Paper No. 10463. https://www.iza.org/publications/dp/10463/missing-from-the-market-purdah-norm-and- womens-paid-work-participation-in-bangladesh Asfar, R. 1999. “Rural-Urban Dichotomy and Convergence: Emerging Realities in Bangladesh.” Bangladesh Institute of Development Studies, Dhaka. Banerjee, Abhijit, Esther Du‫ޕ‬o, Nathanael Goldberg, Dean Karlan, Robert Osei, William Parientĝ, Jeremy Shapiro, Bram Thuysbaert, and Christopher Udry. 2015. “A Multifaceted Program Causes Lasting Progress for the Very Poor: Evidence from Six Countries.” Science 348 (6236): 1260799. Bangladesh Bank. 2019. “Foreign Direct Investment (FDI) in Bangladesh.” Survey Report January-June, 2019. Statistics Department. Bangladesh Bureau of Statistics (BBS). 2018. “Report on Agriculture and Rural Statistics 2018. Agriculture and Rural Statistics Survey (ARSS) Project 2017.” Dhaka: BBS. 144 REFERENCES Bangladesh Bureau of Statistics (BBS). 2019. “Report on Agriculture and Rural Statistics. Agriculture and Rural Statistics Survey (ARSS).” Dhaka: BBS. Barrett, Christopher, Luc Christiaensen, Megan Sheahan, and Abebe Shimeles. 2017. “On the Structural Transformation of Rural Africa.” Journal of African Economies 26(i1): i11–i35. Bayes, Abdul. 2007. “Impact Assessment of Jamuna Multipurpose Bridge Project (JMBP) on Poverty.” Dhaka: Japan Bank for International Cooperation. Bird, Julia, Yue Li, Hossain Zilur Rahman, Martin Rama, and Anthony Venables. 2018. ,4/! /"1"/%(ǽ"4/+"3"),-*"+1/!&$*014/!. Washington, DC: World Bank. Blankespoor, Brian, M. Shahe Emran, Forhad Shilpi, and Lu Xu. 2018. “Bridge to Bigpush or Backwash? Market Integration, Reallocation, and Productivity EȂects of Jamuna Bridge in Bangladesh.” World Bank Policy Research Working Paper No. 8508. World Bank, Washington, DC. Blankespoor, Brian and Nobou Yoshida. 2010. “Market Accessibility in Bangladesh,” Exploratory analysis. World Bank, Washington, DC. BPC (Bangladesh Planning Commission). 2018. Bangladesh Delta Plan 2100—Bangladesh in the 21st Century: The Best Giȅ for the Future Generations by the Present Generation. Dhaka: BPC. BRAC. 2020. “Livelihoods, Coping, and Support During COVID-19 Crisis” PPRC-BIGD Rapid Response Research. Bryan, G., S. Chowdhury, and A. M. Mobarak. 2014. “Under-Investment in a Pro‫ޔ‬table Technology: The Case of Seasonal Migration in Bangladesh,” Econometrica 82(5): 1671-1748. Bryan, G., S. Chowdhury, and A. M. Mobarak. 2016. “No Lean Season. Encouraging Seasonal Migration to Address Income Insecurity.” Policy Brief. https://www.poverty-action.org/sites/default/files/publications/No-Lean-Season-Migration- Policy-Brief.pdf 145 BANGLADESH RURAL INCOME DIAGNOSTICS Cameron, Lisa A., J. Malcolm Dowling, and Christopher Worswick. 2001. “Education and Labor Market Participation of Women in Asia: Evidence from Five Countries.” Economic "3"),-*"+1+!2)12/)%+$"49 (3): 459-477. CazzuȂi, Chiara, Andy McKay, and Emilie Perge. 2017. “Commercialization in agriculture, 2006-14.” In Finn Tarp, ed., /,41%Ǿ 1/2 12/) /+0#,/*1&,+ +! 2/) %+$" &+ Vietnam. Oxford: Oxford University Press. Cheema, Iȅikhar, Simon Hunt, Sarah Javeed, Tanya Lone, and Sean O’Leary. 2016. “Benazir Income Support Programme: Final Impact Evaluation Report.” Oxford: Oxford Policy Management Limited. https://www.opml.co.uk/‫ޔ‬les/Publications/7328-evaluating- pakistans-flagship-social-protection-programme-bisp/bisp-final-impact-evaluation-report. pdf?noredirect=1. Christiaensen, Luc, and Will Martin. 2018. “Agriculture, structural transformation and pov- erty reduction: Eight new insights.” World Development 109. Dercon, Stefan, and Douglas Gollin. 2014. “Agriculture in African Development: Theories and Strategies.” ++2)"3&"4,#"0,2/ " ,+,*& 0 6(0). Dizon, Felipe Fadullon, Md Mansur Ahmed, Deepika Nayar Chaudhery, Nazmul Hoque, Varun Joshi, Manal Musta‫ޔ‬z, Firdousi Naher, Julie Ting Ting Perng, Aneeka Rahman, Jillian Waid, and Zetianyu Wang. 2019. “Food for Improved Nutrition in Bangladesh.” Washington, DC: World Bank. Dorosh, Paul, and Shahidur Rashid. 2012. “Bangladesh Rice Trade and Price Stabilization: Implications of the 2007/08 Experience for Public Stocks.” IFPRI Discussion Paper 24. Doyle, Kate, et al. 2018. “Gender-Transformative Bandebereho Couples’ Intervention to Promote Male Engagement in Reproductive and Maternal Health and Violence Prevention in Rwanda: Findings from a Randomized Controlled Trial.”,+" 13 (4). Egger, E.-M., J. Litch‫ޔ‬eld, and T. Siddiqui. 2015. “Migration and Social Networks: Evidence from Bangladesh.” SSRN Scholarly Paper ID 3300794. Social Science Research Network. The Economist. 2013. “Women and property rights: Who owns Bangladesh?” August 20, 2013. Message posted to http://www.economist.com/blogs/banyan/2013/08/ women-andproperty-rights 146 REFERENCES FAO (Food and Agriculture Organization). 2018. “The State of World Fisheries and Aquaculture 2018 - Meeting the Sustainable Development Goals.” Rome: FAO. Farole, Thomas, and Yoonyoung Cho. 2017. “Bangladesh Jobs Diagnostic.” World Bank, Washington, DC. Field, Erica, Rohini Pande, Natalia Rigol, Simone Schaner, and Charity Troyer Moore. 2016. “On Her Account: Can Strengthening Women’s Financial Control Boost Female Labor Supply?” Working Paper, Harvard University, Cambridge, MA. Gammage, Sarah, Aslihan Kes, Liliane Winograd, Naziha Sultana, Sara Hiller, and Shelby Bourgault. 2017. “Gender and Digital Financial Inclusion: What Do We Know and What Do We Need to Know?” Washington, DC: International Center for Research on Women. Gautam, Madhur, and Rashid Faruqee. 2016. 6+*& 0 ,# 2/) /,41% &+ +$)!"0%ǿ 201&+&+$,3"/16"!2 1&,+ǽDirections in Development. Washington, DC: World Bank. doi:10.1596/978-1-4648-0876-0. Genoni, M.E., S. Dev Bhatta, U. Sharma, B. E. Khaltarkhuu, and L. Maratou-Kolias. 2019. “Equity in education outcomes and spending for Bangladesh.” Background paper for the Bangladesh Poverty Assessment 2010-2016/17, World Bank, Washington, DC. Genoni, Maria Eugenia, Afsana IȂat Khan, Nandini Krishnan, Nethra Palaniswamy, and Wameq Raza. 2020. “Losing Livelihoods: The Labor Market Impacts of COVID-19 in Bangladesh.” World Bank, Washington, DC. Gregg, C., and E. von Uexkull. 2011. “Skills for Trade and Economic Diversi‫ޔ‬cation (STED) in Bangladesh: The Case of Pharmaceuticals and Agro-Food.” Geneva: International Labour Organization. The Growth Lab at Harvard University. The Atlas of Economic Complexity. http://www.atlas. cid.harvard.edu. Goldin, Claudia. 1995. “The U-Shaped Female Labor Force Function in Economic Development and Economic History” in T. Paul Shultz, ed. +3"01*"+1&+,*"+ȉ0 2*+ -&1)+! ,+,*& "3"),-*"+1. Chicago: University of Chicago Press, pp. 61-90. 147 BANGLADESH RURAL INCOME DIAGNOSTICS Gonzales, Christian, Sonali Jain-Chandra, Kalpana Kochhar, and Monique Newiak. 2015, “Fair Play: More Equal Laws Boost Female Labor Force Participation.” StaȂ Discussion Note, SDN/15/02. Washington, DC: International Monetary Fund. GSMA. 2018. “Country overview: Bangladesh mobile industry driving growth and enabling digital inclusion.” Haggblade, Steven, Peter Hazell, and Thomas Reardon. 2007. /+0#,/*&+$ 1%" 2/) ,+#/* ,+,*6ǿ--,/12+&1&"0+!%/"10&+1%""3"),-&+$,/)!. Baltimore: Johns Hopkins University Press. Hamid, M., M. A. Rahman, S. Ahmed, and K. M. Hossain. 2017. “Status of the Poultry Industry in Bangladesh and the Role of the Private Sector for its Development.” Asian Journal of ,2)1/6 &"+ " 11 (1): 1-13. Hanson, G.H. 2010. “International Migration and the Developing World.” In +!,,( ,# Development Economics, Vol. 5, edited by Dani Rodrik and Mark Rosenzweig, 4363-4414: Oxford: North-Holland. Heath, Rachel, and Seema Jayachandra. 2016. “The Causes and Consequences of Increased Female Education and Labor Force Participation in Developing Countries,” NBER Working Paper Series. Working Paper 22766. http://www.nber.org/papers/w22766 Heath, Rachel, and Ahmed Mush‫ޔ‬q Mobarak. 2015. “Manufacturing Growth and the Lives of Bangladeshi Women.” Journal of Development Economics 115: 1–15. Hernandez, R., B. Belton, T. Reardon, C. Hu, X. Zhang, and A.U. Ahmed. 2017. “The Ҁquiet revolution’ in the aquaculture value chain in Bangladesh.” Aquaculture 493: 456–468. doi: 10.1016/j.aquaculture. 2017.06.006. Herrera Dappe, Matias, Charles Kunaka, Mathilde Lebrand, and Nora Weisskopf. 2020. ,3&+$ ,/4/!ǿ,++" 1&3&16+!,$&01& 01,201&++$)!"0%ȉ02 "00ǽWashington, DC: World Bank. doi:10.1596/978-1-4648-1507-2. Hill, R., and J. Endara. 2019a. “Assessing the role of spatial and structural changes on pov- erty reduction in Bangladesh from 2000 to 2016.” Background paper for the Bangladesh Poverty Assessment 2010-2016/17, World Bank, Washington, DC. 148 REFERENCES Hill, R., and J. Endara. 2019b. “Understanding poverty trends in Bangladesh: insights from decomposition analysis.” Background paper for the Bangladesh Poverty Assessment 2010-2016/17, World Bank, Washington, DC. Hill, R., and Z. Rahman 2019. “Poverty in Urban Bangladesh.” Background paper for the Bangladesh Poverty Assessment 2010-2016/17, World Bank, Washington, DC. Hill, Ruth. 2018. “Rural Income Diagnostics: A framework” Washington, DC: World Bank. Hill, Ruth, and Maria Eugenia Genoni. 2019. +$)!"0%,3"/1600"00*"+1ǿ  &+$)!+! "4 /,+1&"/0&+,3"/16"!2 1&,+. Washington, DC: World Bank Group. Hill, Ruth Vargas, Neha Kumar, Nicholas Magnan, Simrin Makhija, Francesca de Nicola, David J. Spielman, and Patrick S. Ward. 2019. “Ex ante and ex post eȂects of hybrid index insurance in Bangladesh.” Journal of Development Economics 136: 1-17. Hoque, M., and M. Ahmed. 2020. “Agricultural Diversi‫ޔ‬cation in Bangladesh: Trends, Spatial Patterns, and Drivers.” Background paper for “Agriculture Strategies Development in Bangladesh,” Agriculture and Food Practice, World Bank, Washington, DC. Hossain, M., and A. Bayes. 2009. 2/) ,+,*6+!&3")&%,,!0ǿ +0&$%10#/,*+$)!"0%. Dhaka: A. H. Publishing House. Hossain, M., and A. Bayes. 2018. 2/)/+0#,/*1&,+ǿ +0&$%10#/,*+$)!"0%. Dhaka: University Press Ltd. Hossain, M., B. Sen, and Y. Sawada. 2013. “Jobs, Growth and Development: Making of the ҀOther’ Bangladesh.” Background Paper for the World Development Report. World Bank, Washington, DC. Hossen, M A. 2019. “Mechanization in Bangladesh: Way of Modernization in Agriculture.” International Journal of Engineering Trends and Technology 67 (9) (Sep 2019): 69. Jahan, Ferdous, Asif M. Shahan, Shameem Reza Khan, Akter Uz-zaman, Shakila Sharmin, and Mishkat Jahan. 2016. “Gender Quality Action and Learning Program: Documenting the ҀTransformations Achieved’ and the ҀProcess of Achieving Them.’” Unpublished working paper. 149 BANGLADESH RURAL INCOME DIAGNOSTICS Kabeer, N. 1990. “Poverty, purdah and women’s survival strategies in rural Bangladesh.” In %" #,,! .2"01&,+ǿ -/,Ɯ10 3"/020 -",-)"Ȅ Edited by H. Bernstein, M. Mackintosh, C. Martin, and B. Crow. London: Routledge, 134-148. Kabeer, N. 2001. “Con‫ޕ‬icts over credit: reevaluating the empowerment potential of loans to women in rural Bangladesh.” World Development 29(1): 63-84. Karlan, Dean, Jake Kendall, Rebecca Mann, Rohini Pande, Tavneet Suri, and Jonathan Zinman. 2016. “Research and Impacts of Digital Financial Services.” NBER Working Paper Series w22633, National Bureau of Economic Research, Cambridge, MA. Khaki M., E. Forootah, M. Kuhn, J. Awange, F. Papa, and C.K. Shum. 2018. “A study of Bangladesh’s sub-surface water storages using satellite products and data assimilation scheme.” Science of The Total Environment 625 (1 June 2018): 963-977. Khandker, S. R., H. A. Samad, and Z. H. Khan. 1998. “Income and employment eȂects of micro-credit programmes: Village-level evidence from Bangladesh.” The Journal of Development Studies 35(2): 96-124. Klugman, J., L. Hanmer, S. Twigg, T. Hasan, J. McCleary-Sills, and J. Santamaria. 2014. Voice +!$"+ 6ǿ*-,4"/&+$4,*"++!$&/)0#,/0%/"!-/,0-"/&16. Washington, DC: World Bank. Lanjouw, Peter and Gershon Feder. 2001. “Rural Non-farm Activities and Rural Development. From Experience towards Strategy.” World Bank. Rural Strategy Background Paper 4. Lastarria-Cornhiel, Susana, Julia A. Behrman, Ruth Meinzen-Dick, and Agnes R. Quisumbing. 2014. “Gender Equity and Land: Toward Secure and EȂective Access for Rural Women.” In "+!"/ &+ $/& 2)12/"ǿ ),0&+$ 1%" +,4)"!$" -, edited by A. R. Quisumbing, R. Meinzen-Dick, T. L. Raney, A. Croppenstedt, J. A. Behrman, and A. Peterman. Springer Science & Business. Mahmud, M. and Y. Sawada. 2014. “Infrastructure and Well-being: Employment EȂects of Jamuna Bridge in Bangladesh.” Working paper, University of Tokyo. Masylkanova, K. 2020. “Improving Resilience and Quality of On-Farm Incomes and Productivity through Agricultural Diversi‫ޔ‬cation.” Background paper for “Agriculture Strategies Development in Bangladesh.” Agriculture and Food Practice. Washington, DC: World Bank. 150 REFERENCES Ministry of Finance. 2019. +$)!"0%  ,+,*&  "3&"4 ǗǕǖǞ. Dhaka: Bangladesh Government Press. Minot, N., M. Epprecht, T. T. T. Anh, and L. Q. Trung. 2006. “Income Diversi‫ޔ‬cation and Poverty in Northern Uplands of Vietnam.” Research Report 145, International Food Policy Research Institute (IFPRI), Washington, DC. Mueller, Valerie, Lucy Billings, Tewodaj Mogues, Amber Peterman, and Ayala Wineman. 2017. “Filling the Legal Void? Impacts of a Community-Based Legal Aid Program on Women’s Land-Related Knowledge, Attitudes, and Practices.” 5#,/! "3"),-*"+1 Studies. Muzzini, Elisa, and Gabriela Aparicio. 2013. “Bangladesh: The Path to Middle-Income Status from an Urban Perspective.” Washington, DC: World Bank. doi:10.1596 /978-0-8213-9859-3. NeȂ, D., K. Sen, and V. Kling. 2012. “The Puzzling Decline in Rural Women’s Labor Force Participation in India: A Reexamination.” GIGA Working Paper No. 196. GIGA German Institute of Global and Area Studies. PWC. 2019. “Potential for growth: Transforming Bangladesh’s insurance sector.” https:// www.pwc.com/bd/en/assets/pdfs/research-insights/2019/potential-for-growth.pdf Quisumbing, A. R., N. Kumar, and J.A. Behrman. 2018. “Do shocks aȂect men’s and wom- en’s assets diȂerently? Evidence from Bangladesh and Uganda.” "3"),-*"+1 ,)& 6 "3&"4 36(1): 3-34. Raihan, Selim, Alberto Lemma, Bazlul Haque Khondker, and Farazi Binti Ferdous. 2017. “Bangladesh Sectoral Growth Diagnostic.” The Asia Foundation. Rahman, Rushidan I., and Rizwanul Islam. 2013. “Female labor force participation in Bangladesh: trends, drivers, and barriers.” ILO Asia-Paci‫ޔ‬c Working Paper Series. Raihan, Selim, and Sayema Haque Bidisha. 2018. “Female empowerment stagnation in Bangladesh.” Research paper. Economic Dialogue on Inclusive Growth in Bangladesh. EDIG Research Paper Five. Rahman, H.Z. 2014. “Urbanization in Bangladesh: Challenges and Priorities.” Paper pre- sented at Bangladesh Economists’ Forum, June 21–22, 2014, Dhaka. 151 BANGLADESH RURAL INCOME DIAGNOSTICS Rashid, S. and X. Zhang. 2019. “The Making of a Blue Revolution in Bangladesh: Enablers, Impacts, and the Path Ahead for Aquaculture.” Washington, DC: International Food Policy Research Institute. Ravallion, M., and M. Huppi. 1991. “Measuring changes in poverty: a methodological case study of Indonesia during an adjustment period.” %",/)!+( ,+,*& "3&"4 5(1): 57-82. Roberts, Mark, Brian Blankespoor, Chanan Deuskar, and Benjamin Stewart. 2017. “Urbanization and Development: Is Latin America and the Caribbean DiȂerent from the Rest of the World?” Policy Research WP 8019. World Bank, Washington, DC. Sayeed, K. A., and M.M. Yunus. 2018. “Rice prices and growth, and poverty reduction in Bangladesh.” FAO: Rome, Italy. Sen, B., M. Ahmed, M. Yunus, and Z. Ali. 2014. “Regional Inequality in Bangladesh in the 2000s: Re-Visiting the East-West Divide Debate.” Bangladesh Institute of Development Studies. Shilpi, F., M. N. Hoque, J. Yao, and H. Uddin. 2015. “Connecting Farmers to Consumers: Trading in High-Value Products in Bangladesh.” Background paper for 6+*& 0,#2/) Growth in Bangladesh, Agriculture Global Practice, World Bank, Washington, DC. Shilpi, Forhad, and M. Shahe Emran. 2016. “Agricultural Productivity and Non-Farm Employment: Evidence from Bangladesh,” World Bank Policy Research Working Paper No. 7685. World Bank, Washington, DC. SolotaroȂ, Jennifer L., Aphichoke Kotikula, Tara Lonnberg, Snigdha Ali, Rohini P. Pande, and Ferdous Jahan. 2019. ,& "01,%,& "0ǿ+$)!"0%ȉ0 ,2/+"6&+,*"+ȉ0 ,+,*&  Empowerment. International Development in Focus. Washington, DC: World Bank. doi:12.1596/978-1-4648-1374-0 Srivastava, N., and R. Srivastava. 2010. “Women, work, and employment outcomes in rural India.”  ,+,*& +!,)&1& )""()6: 49-63. Sumi, Hosna Ferdous, and M. Masrur Reaz. 2020. “Building Competitive Sectors for Export Diversi‫ޔ‬cation: Opportunities and Policy Priorities for Bangladesh.” Washington, DC: World Bank Group. 152 Timmer, Peter, and Selvin Akkus. 2008. “The Structural Transformation as a Pathway Out of Poverty: Analytics, Empirics, and Politics.” Center for Global Development Working Paper 150. Tillan, Pablo, M.E. Genoni, A. Velasquez, and J. Endara. 2020. “Understanding the impli- cations of migration for rural income growth in Bangladesh.” Background paper for Bangladesh Rural Income Diagnostic. World Bank, Washington, DC. TU Delȅ. 2018. "0&)&"+ ",#1%"/+0-,/1"14,/(&++$)!"0%ǿ /-%Ȓ%",/"1& )+! Simulation Based Approach. Report commissioned for a World Bank study, World Bank, Washington, DC. Uchida, Hirotsuga, and Andrew Nelson. 2010. “Agglomeration Index: Towards a New Measure of Urban Concentration.” UN WIDER Working Paper No. 2010/29. United Nations. 2018. ,/)! /+&71&,+ /,0-" 10ǿ %" ǗǕǖǝ "3&0&,+. Department of Economic and Social AȂairs, Population Division. UNCTAD. 2017. “Fishery Exports and The Economic Development of Least Developed Countries: Bangladesh, Cambodia, The Comoros.” Geneva: UNCTAD. USAID (United States Agency for International Development). 2014. “Bangladesh: Inclusive Growth Diagnostic.” Washington, DC: USAID. Verick, S. 2014. “Female labor force participation in developing countries.” IZA World of ,/. Wallace, T. 2013. Property Rights and Women’s Economic Participation in Bangladesh. August 28, 2013. Message posted to http://www.cipe.org/blog/2013/08/28/proper- ty-rightsand-womens-economic-participation-in-bangladesh/comment-page-1/#.Uv1MmvldVzc World Bank. 2012. Gender Equality and Development. World Development Report. Washington, DC: World Bank. World Bank. 2015a. “Bangladesh – More and better jobs to accelerate shared growth and end extreme poverty. A systematic country diagnostic.” Washington, DC: World Bank. World Bank. 2015b. “Bangladesh: Agricultural Insurance Solutions Appraisal Technical Report.” World Bank, Washington, DC. BANGLADESH RURAL INCOME DIAGNOSTICS World Bank. 2017. Local Governance Support Project – 3. Project Appraisal Document. World Bank, Washington, DC. World Bank. 2019a. “Bangladesh: Seizing the Opportunity,” World Bank Group Policy Notes. World Bank, Washington, DC. World Bank. 2019b. “Enabling the Business of Agriculture 2019.” Washington, DC: World Bank. doi: 978-1-4648-1387-0. World Bank. 2019c. “Bangladesh: Climate-Smart Agriculture Investment Plan.” World Bank, Washington, DC. World Bank. 2019d. “Productive Jobs for Bangladesh’s youth: Constraints in wage-employ- ment and the way forward.” World Bank, Washington, DC. World Bank. 2020a. “Agricultural Diversi‫ޔ‬cation and Modernization Strategy: Promoting Agri-Food Sector Transformation in Bangladesh.” World Bank, Washington, DC. World Bank. 2020b. Bangladesh: Country Private Sector Diagnostic.” World Bank, Washington, DC. Zaedi, Md Sumon, Katsuhiko Demura, Yasutaka Yamamoto, Daisuke Sawauchi, Kiyotaka Masuda, and Tomoaki Nakatani. 2009. “Dairy Industry in Bangladesh and Structures of Milk Vita.” %""3&"4,#$/& 2)12/) ,+,*& 0 64(May):169–80. 154