C O X ’ S B A Z A R Inclusive Growth Diagnostic THE WORLD BANK Contents The World Bank 1818 H Street N.W. Washington, D.C. 20433 USA Phone: (202) 458-1876 Email: feedback@worldbank.org Website: www.worldbank.org Acknowledgements 15 This report is a product of staff of the World Bank. The findings, interpretations, and con- List of abbreviations 17 clusions expressed herein do not necessarily reflect the views of the Board of Executive Executive Summary 21 Directors of the World Bank or the governments they represent. 1. Introduction 33 The World Bank does not guarantee the accuracy of the data included in this work. The 2. Fundamentals: People, land, and infrastructure 45 boundaries, colors, denomination, and other information shown on any map in this work Demographics and density 45 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. Human capital and living conditions 49 Education 49 Health 58 Food security and living conditions 64 Geography 68 Connective infrastructure and accessibility 73 3. Economic outcomes: Jobs, livelihoods, and incomes 81 Structure of the Cox’s Bazar economy: Economic activity and firm composition 81 Agriculture and fisheries 83 Services and industry 89 Work and livelihoods in Cox’s Bazar 98 Employment patterns and dynamics 98 Income sources and livelihoods 106 Firm performance and earnings of the self-employed 110 Migration and remittances 114 The COVID-19-related economic slowdown in Cox’s Bazar 116 Work and livelihoods among the recently displaced Rohingya population 118 4. Accelerating inclusive growth: Constraints and opportunities 123 Targeting constraints to inclusive growth 123 Constraints to human capital accumulation 124 Constraints to productive inclusion in the labor market 129 Constraints to private sector activity and entrepreneurship 131 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Contents Identifying opportunities 137 Figure 2-4: Share of school types, by upazila, Cox’s Bazar, 2011 53 Localized comparative advantage 137 Figure 2-5: Share of students by type of institution and upazila, Cox’s Bazar, 2011 53 Economic connectivity and infrastructure enhancements 143 Governance and service delivery 155 Figure 2-6: School attendance rates before and after the 2017 Rohingya influx, Humanitarian assistance and local economic activity 158 host children and Rohingyas 57 5. Areas for policy action 163 Figure 2-7: Female-male school attendance ratio gaps, before and after 2017 Rohingya influx, host community and Rohingya 58 Policy context, challenges, and opportunities 163 Figure 2-8: Net school attendance rate after influx, hosts in high- and Policy recommendations 167 low-exposure areas 58 Early investments in productive potential 168 Strengthening productive capacity 169 Figure 2-9: Travel times to health care facilities, by population share 61 Expanding economic opportunities 173 Figure 2-10: Travel time to health care facilities, Teknaf and Ukhia versus Bridging evidence gaps 176 other upazilas 61 References 179 Figure 2-11: Women’s age at first marriage: Bangladesh, Chittagong division, Annex 1: Additional tables and figures 191 and Cox’s Bazar, 2019 62 Figure B3-1: Uses of land, districts in Chittagong division and nationally 70 Annex 2: Methodology note - Cox’s Bazar accessibility analysis 213 Figure B3-2: Risk of cyclones and storms and average annual rainfall in Cox’s Bazar 71 Annex 3. Nighttime light data and economic activity 225 Figure B3-3: Deforestation in Kutupalong camps, May 2017 versus May 2020 72 Annex 4. Pattern of employment in Cox ’s Bazar: Economic Census 2013 and Population Census 2011 228 Figure 2-12: Average travel time to markets of any size by level of education and upazilas 76 Figure 2-13: Average travel time to growth centers by level of education 76 Figure 3-1: Economic performance among Bangladeshi districts (1): share of national non-agricultural firms in relation to population 82 Figure 3-2: Economic performance among Bangladeshi districts (2): share of national fish production in relation to population 82 Figure 3-3: Economic performance among Bangladeshi districts (3): List of figures share of total net cropped area in relation to population 83 Figure B1-1: Poverty rate, districts in Chittagong division versus Bangladesh, Figure 3-4: Economic performance among Bangladeshi districts (4): 2000 - 2016 40 share of total crop production in relation to net cropped area 83 Figure B1-2: Poverty rate, upazilas in Cox’s Bazar, 2010 small area estimates 40 Figure 3-5: Share of crops in total agricultural production, different agroecological zones, Chittagong division, 2018 86 Figure B1-3: Monthly per capita expenditure (Tk), upazilas in Cox’s Bazar, 2010 small area estimates 41 Figure 3-6: Share of crops in total agricultural production, Cox’s Bazar and comparator areas, 2018 86 Figure 2-1: Demographic pyramids, Bangladesh versus Cox’s Bazar, 2016 49 Figure 3-7a: Sectoral composition of non-agricultural firms and employment, Figure 2-2: Educational attainment, adults (18+), 2019 (low-exposure areas, Bangladesh, 2013 89 high-exposure areas, and Rohingya camps) 51 Figure 3-7b: Sectoral composition of non-agricultural firms, Cox’s Bazar, 2013 89 Figure 2-3: Female-male gaps in educational attainment, adults (18+), 2019 (low-exposure areas, high-exposure areas, and Rohingya camps) 51 Figure 3-8: Sectoral composition and employment, Bangladesh, 2013 90 Figure 3-9: Sectoral composition employment, Cox’s Bazar 2013 90 4 5 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Contents Figure 3-10: Share of non-agricultural firms by main activity -Cox’s Bazar district, Figure 3-31: Average composition of monthly income, by quintile, low-exposure Chittagong division, and Bangladesh, 2013 91 and high-exposure areas, 2019 109 Figure 3-11: Cox’s Bazar firms, as a share of division and national firms, Figure 3-32: Median net monthly revenue of firms: Cox’s Bazar, Chittagong, diverse sectors, 2013 91 and Bangladesh, 2016 110 Figure 3-12: Where are firms located in Cox’s Bazar? Share of district firms Figure 3-33: Monthly labor income in Cox’s Bazar: Wage employment versus by upazilas, 2013 92 business profits (average), 2019 110 Figure 3-13: Distribution of firms by sector within Cox’s Bazar upazilas, 2013 92 Figure 3-34: Mean monthly earnings by area and sector, Cox’s Bazar, 2019 112 Figure 3-14: Average size of non-micro firms (by number of workers), Figure 3-35: Main sectors of work, by area and employment type, Cox’s Bazar, 2019 112 Cox’s Bazar and comparator areas, 2013 94 Figure 3-36: Share of wage employment by sector and self-reported firm size, Figure 3 15: Average size of non-micro firms (by number of employees), Cox’s Bazar upazilas, high-exposure versus low-exposure areas, 2019 113 2013 94 Figure 3-37: Most wage workers are employed in very small enterprises: Self-reported Figure 3-16: Average age of firms, by firm size, Cox’s Bazar, 2013 94 firm size among wage workers, high-exposure and low-exposure areas, 2019 113 Figure 3-17: Ownership type, by firm size, Cox’s Bazar, 2013 97 Figure 3-38: Share of wage employment by sector, gender, and self-reported firm size, high-exposure areas, 2019 113 Figure 3-18: Ownership type, non-micro firms (>=10 workers), Cox’s Bazar, Chittagong, and Bangladesh, 2013 97 Figure 3-39: Share of wage employment by sector, gender, and self-reported firm size, low-exposure areas, 2019 113 Figure 3-19: The normality of informality in Bangladesh 97 Figure 3-40: Monthly international remittance flows to Bangladesh from Figure 3-20: Share of workers with written contracts, Cox’s Bazar, 2019 97 wage workers abroad, 2019-2020 (millions of US dollars) 115 Figure 3-21: Share of host-community men and women working in different sectors, Figure 3-41: Characteristics of households that receive or do not receive remittances, Cox’s Bazar, 2019 100 high- and low-exposure areas, Cox’s Bazar, 2019 115 Figure 3-22: Share of individuals working, by activity and gender, Cox’s Bazar Figure 3-42: Share of households receiving remittances, by income quintile, (Bangladeshi households), 2019 100 Cox’s Bazar, 2019 115 Figure 3-23: Sectors of employment, high-exposure versus low-exposure males, Figure 3-43: Labor market indicators, Rohingya population in camps, 2019 119 Cox’s Bazar, 2019 101 Figure 3-44: Share of employment by sector, Rohingya men and women in camps, 2019 119 Figure 3-24: Sectors of employment, high-exposure versus low-exposure females, Cox’s Bazar, 2019 101 Figure 4-1: Key constraints faced by non-agricultural enterprises: Bangladesh, Chittagong, and Cox’s Bazar, 2016 132 Figure 3-25: Probability of employment by sector and level of education, high-exposure versus low-exposure areas, 2019 102 Figure 4-2: Uses and sources of finance for business: Bangladesh, Chittagong, and Cox’s Bazar 133 Figure 3-26: Share of individuals with secondary jobs, Cox’s Bazar, 2019 105 Figure 4-3: Use of machines among manufacturing firms, Cox’s Bazar, Chittagong, Figure 3-27: Types of contracts used for secondary jobs, Cox’s Bazar, 2019 105 and Bangladesh, 2013 134 Figure 3-28: Share of individuals with secondary jobs, by sector, Cox’s Bazar, 2019 105 Figure 4-4: Education level of firm owners, Bangladesh, Chittagong, and Cox’s Bazar, 2013 135 Figure 3-29: How types of contracts differ for primary and secondary jobs, Figure 4-5: What kind of work for the highly skilled? Main job, by workers’ education levels, high-exposure and low-exposure areas, Cox’s Bazar, 2019 105 Cox’s Bazar, 2019 136 Figure 3-30: Average composition of monthly income (last 30 days) from Figure 4-6: Firms’ main markets and customer types: Bangladesh, Chittagong, agricultural and non-agricultural sources, by per capita income quintile, and Cox’s Bazar 136 low-exposure and high-exposure areas, 2019 108 Figure 4-7: Distribution of exporting firms by districts, Bangladesh, 2013 143 6 7 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Contents Figure 4-8: Distribution of exporting firms by upazilas in Cox’s Bazar, 2013 143 List of maps Figure B6-1: Expenditure allocation (Tks) in proportion to school-age population Map 1-1: Bangladesh and Cox’s Bazar district: Major roads and seaports 35 (0-14 years), by district 156 Map 1-2: Locations of recently displaced Rohingya camps in Teknaf Figure B6-2: Per capita allocated health expenditure (Tks), by district 156 and Ukhia upazilas 36 Figure B6-3: Per capita allocated local governance expenditure, by district 157 Map 1-3: Population density by zila, estimated 2018 (does not include Figure B6-4: Allocated agricultural expenditure per acre of cropped area, by district 157 Rohingya population) 37 Figure 4-9: Share of total funding by clusters in Cox’s Bazar humanitarian response, Map 1-4: Poverty headcount by zila, 2016 (upper poverty line) 37 2017-2019 159 Map 1-5: Nighttime lights as a marker of economic activity, Bangladesh Figure 4-10: Evolution of funding in Cox’s Bazar humanitarian response, 2010-2019 159 and Cox’s Bazar, 2019 39 Figure 4-11: Quadratic fit of monthly nightlight intensity around growth centers over time 161 Map 2-1: Population of Cox’s Bazar district, by upazila, pre-2017 Rohingya influx 47 Figure 5-1: Framework for policy recommendations 168 Map 2-2: Population of Cox’s Bazar district, by upazila, post-2017 Rohingya influx 47 Figure A1-1: Bangladesh agroecological zones 194 Map 2-3: Population density by upazila, before 2017 influx, including pre-2017 displaced Rohingya 48 Figure A1-2: District diversification in coastal plains and northern hills agroecological zone 195 Map 2-4: Population density by upazila, after 2017 influx, including newly displaced Rohingya 48 Figure A1-3: District diversification in river and estuarine flood plains agroecological zone 195 Map 2-5: Travel times to primary schools 55 Figure A1-4: Share of rural population by upazila, Cox’s Bazar 195 Map 2-6: Travel times to secondary schools 56 Figure A2-1: A gravity model of union-level accessibility to high quality jobs in Cox’s Bazar 213 Map 2-7: Travel times to tertiary education 56 Figure A2-2: Nodes and edges in a network 214 Map 2-8: Estimated travel times to health center facilities, Cox’s Bazar district 60 Figure A2-3: Simplifying a junction 215 Map 2-9: Cox’s Bazar road transport network 69 Figure A2-4: Simplifying a small network 215 Map 2-10: Cox’s Bazar population, camps, and road transport network 69 Figure A2-5: Nodes, edges, origins, and destinations in a network 216 Map 2-11: Estimated travel times to Chittagong city 74 Figure A2-6: Populated places / origins (in blue) around Cox’s Bazar town (HRSL 2018) 216 Map 2-12: Accessibility to growth centers in Cox’s Bazar 74 Figure A2-7: Health center destinations in Cox’s Bazar 217 Map 2-13: Accessibility to proposed Matarbari port and energy complex 78 Figure A2-8: Accessibility statistics were aggregated at various levels, Map 4-1: Estimated travel times to Matarbari port, with upgrade of key roads 145 and occasionally further subdivided by demographic indicators within them 218 Map 4-2: Estimated travel times to Matarbari port, with upgrade of key roads and ferry 146 Figure A2-9: Proposed transport investments in Cox’s Bazar 219 Map 4-3: Estimated travel times to Matarbari port, with upgrade of key roads, ferry, Equation A2-1: The negative exponential model 220 and AH41 (N1) Highway 146 Figure A4-1: Employment shares in Cox’s Bazar, Chittagong, and Bangladesh, 2011 230 Map 4-4: Estimated travel times to Cox’s Bazar Sadar, with upgrade of key roads 147 Figure A4-2: Employment shares by upazila, 2011 230 Map 4-5: Estimated travel times to Cox’s Bazar Sadar, with upgrade of key roads and ferry 147 Figure A4-3: Education level and sector of employment, Cox’s Bazar 235 Map 4-6: Estimated travel times to Cox’s Bazar Sadar, with upgrade of key roads, ferry, and AH41 (N1 Highway) 147 Figure A4-4: Main sectors of employment by zilas in Chittagong Division 235 Map 4-7: Unions with significant presence of large firms, Cox’s Bazar 148 8 9 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Contents Map 4-8: Large firm accessibility indices (markets weighted by large firms), Table 3-6: How people sort to different kinds of employers, low-exposure pre-transport investments 149 and high-exposure areas, Cox’s Bazar, 2019 103 Map 4-9: Large firm accessibility indices (markets weighted by large firms), post-transport Table 3-7: Average number of months and weekly hours allocated to primary investments 149 and secondary jobs, waged and non-waged workers, Cox’s Bazar, 2019 104 Map 4-10: Travel times to growth centers, (current) pre-investment 150 Table 3-8: Income sources and average incomes, low-exposure versus high-exposure areas within Cox’s Bazar, 2019 106 Map 4-11: Travel times to growth centers, post-investment 150 Table 3-9: Weekly and hourly wages in primary and secondary jobs, Cox’s Bazar, Map 4-12: Market accessibility index (growth centers), pre-investment (unweighted) 151 2019 (averages, in Takas) 111 Map 4-13: Market accessibility index (growth centers), post-investment (unweighted) 151 Table 3-10: Education and sector of employment, recently displaced Rohingya, 2019 119 Map 4-14: All firms accessibility indices (markets weighted by firms), pre-transport Table 3-11: Income sources and average incomes, recently displaced Rohingya investment 152 in Cox’s Bazar, 2019 120 Map 4-15: All firms accessibility indices (markets weighted by firms), post-transport Table 4-1: Reasons for never attending school, high-exposure and low-exposure areas, investment 152 Cox’s Bazar, 2019 124 Map 4-16: High-quality jobs accessibility indices (markets weighted by high-quality Table 4-2: Reasons for dropping out of school, by age and gender, high-exposure job numbers), pre-transport investment 153 versus low-exposure areas, 2019 125 Map 4-17: High-quality jobs accessibility indices (markets weighted by high-quality Table B4-1: Median expenditure on education by quintile, Cox’s Bazar, Chittagong division, job numbers), post-transport investment 153 and Bangladesh, 2016 128 Map 4-18: Travel times to tertiary education, pre-transport investment 154 Table B4-2: Components of educational expenditure, by education level, Map 4-19: Travel times to tertiary education, post-transport investment 154 Cox’s Bazar, 2016 128 Map 4-20: Location of growth centers in Cox’s Bazar 161 Table 4-6: Main activities and size of export firms, Cox’s Bazar, Chittagong, and Bangladesh, 2013 142 Table B7-1: Cox’s Bazar airport traffic 160 List of tables Table 4-4: Growth centers in Cox’s Bazar district, by distance from Rohingya camps 161 Table 2-1: Average number of students per school for different school types, Table A1-1: Road speeds by type 191 Cox’s Bazar upazilas 54 Table A1-2: Incidence of crime in the neighborhood, as reported by CBPS 2019 respondents 191 Table 2-2: Average student-teacher ratios, different school types, Cox’s Bazar upazilas 54 Table A1-3: Exposure to trauma events among CBPS 2019 respondents 192 Table 2-3: Malnutrition indicators for Bangladesh, Chittagong division, Table A1-4: Trauma symptoms reported by CBPS 2019 respondents 193 and Cox’s Bazar district, 2007-2019 59 Table A1-5: Firm density by upazila 196 Table 2-4: Market accessibility index - Ranking of Cox’s Bazar upazilas, 2010 75 Table A1-6: Non-agricultural establishment size distribution, Bangladesh vs Chittagong Table 3-1: Size of land holdings, Cox’s Bazar and comparator areas, 2008 87 vs Cox’s Bazar 197 Table 3-2: Intensiveness of fish production in ponds, Cox’s Bazar and comparators, 2017 88 Table A1-7: Size-wise distribution of firms, by sector - Bangladesh, Chittagong, Table 3-3: Firm size (official versus IGD classification of enterprises) 95 and Cox’s Bazar 198 Table 3-4: Labor force participation, Cox’s Bazar, Chittagong division, and Bangladesh 98 Table A1-8: Sector-wise distribution of firms, by firm size - Bangladesh, Chittagong, and Cox’s Bazar 198 Table 3-5: Sectoral composition of employment: Bangladesh, Chittagong, and low- and high-exposure areas of Cox’s Bazar 99 Table A1-9: Upazila-wise distribution of firms by firm-size groups 199 Table A1-10: Number of firm by upazila, firm-size groups, and sector 200 10 11 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Contents Table A1-11: Distribution of firms by upazila and by firm-size group within each sector 202 Table A4-9: Number of workers and share of industry and services employment, Pekua 241 Table A1-12: Share of firm-size groups among total firms, by sector and upazila 203 Table A4-10: Number of workers and share of industry and services employment, Ramu 242 Table A1-13: Breakdown of “Other industry” and” Other services” categories Table A4-11: Number of workers and share of industry and services employment, Teknaf 243 for non-micro enterprises (more than 10 employees), by upazila 205 Table A1-14: Constraints on access to education among persons who never attended school, bottom 40 and upper 60, by gender, high- and low-exposure areas 207 Table A1-15: Constraints on access to education among persons who dropped out List of boxes of school: bottom 40 and upper 60, by age group and gender, high-exposure areas 207 Box 1: Poverty in Cox’s Bazar 39 Table A1-16: Constraints on access to education among persons who dropped out Box 2: Data sources for this diagnostic 41 of school: bottom 40 and upper 60, by age group and gender, low-exposure areas 208 Box 3: Land availability and environmental risks in Cox’s Bazar 45 Table A1-17: Share of individuals who dropped out of school, by type of school, Box 4: Household education expenditure in Cox’s Bazar 127 quintile, and gender, high- and low-exposure areas 209 Box 5: Modeling accessibility 144 Table A1-18: Distribution of firms by sector and upazila, Cox’s Bazar 209 Box 6: Analyzing district-level public expenditure in Bangladesh 156 Table A1-19: Share of firm by market for goods 210 Box 7: Airport activity in Cox’s Bazar since the Rohingya influx 159 Table A1-20: Distribution of exporting firms by size, Cox’s Bazar, Chittagong, and Bangladesh 210 Table A1- 21: Share of vulnerable and secure jobs among all workers in each geographic unit 211 Table A1-22: Distribution of total vulnerable and secure jobs in Cox’s Bazar, across upazilas 211 Table A2-1: Network segment speeds 217 Table A2-2: Investment scenarios 219 Table A2-3: Data sources 221 Table A2-4: Categorizing service vs. industrial workers 223 Table A3-1: Nightlight intensity regression results 227 Table A4-1: Number of non-agricultural workers and shares by firm size and sector 231 Table A4-2: Employment distribution and shares by sectors and upazilas 232 Table A4-3: Shares of female, male, and total employment in industry and services, Cox’s Bazar (main activities) 234 Table A4-4: Employment shares by sector, Bangladesh, Chittagong, and Cox’s Bazar 236 Table A4-5: Number of workers and share of industry and services employment, Chakaria 237 Table A4-6: Number of workers and share of industry and services employment, Cox’s Bazar Sadar 238 Table A4-7: Number of workers and share of industry and services employment, Kutubdia 239 Table A4-8: Number of workers and share of industry and services employment, Maheshkhali 240 12 13 Acknowledgements This report was prepared by a team led by Nandini Krishnan (Senior Economist, Poverty and Equity Global Practice - South Asia Region, World Bank) and including: Pablo Antonio Tillan (Consultant, Poverty and Equity Global Practice - South Asia Region, World Bank), Joaquin Endara (Consultant, Poverty and Equity Global Practice - South Asia Region, World Bank), Arti Grover (Senior Economist, Finance, Competitiveness and Innovation, World Bank), and Robert Banick (Consultant, Poverty and Equity Global Practice - South Asia Region, World Bank). Overall guidance for the report was provided by Mercy Tembon (Country Director, Bangladesh and Bhutan, World Bank), Zoubida Allaoua (Regional Director, Equitable Growth, Finance and Institutions - South Asia Region, World Bank), Andrew Dabalen (Practice Manager, Poverty and Equity Global Practice - South Asia Region, World Bank), Dandan Chen (Operations Manager, World Bank), Yutaka Yoshino (Program Leader, Equitable Growth, Finance and Institutions, World Bank). The team gratefully acknowl- edges feedback and guidance provided throughout the drafting of the report by Ikechi Okorie (Senior Operations Officer and Rohingya Response Coordinator, Bangladesh, World Bank) and Suleiman Namara (Lead Social Protection Specialist, Social Protection and Jobs - Africa Region, World Bank). Alexander Irwin (Consultant, Health, Nutrition and Population Global Practice, World Bank) also contributed at several stages to the revision and editing of the report. The report core team is grateful for technical inputs and feedback received during formal and informal reviews, provided by: Maria Eugenia Genoni (Senior Economist, Poverty and Equity Global Practice - Middle East and North Africa Region, World Bank), Afsana 15 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Khan (Consultant, Poverty and Equity Global Practice - South Asia Region, World Bank), Arshia Haque (Consultant, Poverty and Equity Global Practice - South Asia Region, World List of abbreviations Bank), Carlos Rodriguez Castelan (Lead Economist, Poverty and Equity Global Practice - Africa Region, World Bank), Johannes Hoogeveen (Practice Manager, Poverty and Equity Global Practice - Middle East and North Africa Region, World Bank), and Kevin Carey (Lead Economist, Office of the Regional Director, Equitable Growth, Finance and Institutions - Middle East and North Africa Region, World Bank). The report core team is grateful for the technical feedback provided by: Peter D’Souza (Senior Economist, Foreign, Commonwealth and Development Office, Government of the United Kingdom) and Nick Harvey (Senior Humanitarian Adviser, Foreign, Commonwealth and Development Office, Government of the United Kingdom). This report was funded with aid from the Government of the United Kingdom. The team also thanks the World Bank – UNHCR Joint Data Center on Forced Displacement, which financed the Cox’s Bazar COVID- 19 phone monitoring surveys analyzed in this report. Map data copyrighted OpenStreetMap contributors and available from https://www.open- ADB Asian Development Bank streetmap.org. ARI Acute respiratory infections ARRS Agricultural and Rural Statistics survey BAMBEIS Bureau of Educational Information and Statistics BBS Bangladesh Bureau of Statistics BDHS Bangladesh Demographic and Health Survey CBPS Cox’s Bazar Household Panel Survey CIESIN Center for International Earth Science Information Network DGHS Directorate General of Health Services DHS Demographic and Health Surveys EMRCR Emergency Multi-Sector Rohingya Crisis Response ESCAP UN Economic and Social Commission for Asia and the Pacific EU European Union FLFP Female labor force participation GAGE Gender and Adolescence: Global Evidence GDP Gross Domestic Product GIEP Guideline on Informal Education Program GoB Government of Bangladesh GOST Geospatial Operations Support Team 16 17 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC List of abbreviations GW Gigawatt SMEs Small and medium enterprises HIES Household Income and Expenditure Survey SNSP Safety Net System for Poorest HRSL High Resolution Settlement Layer SPS Sanitary and Phytosanitary ICT Information and communication technology TBT Technical Barriers to Trade IHS Inverse hyperbolic sine TFP Total factor productivity IOM International Organization for Migration Tk Bangladeshi Taka ISCG Inter Sector Coordination Group TVET Technical and vocational education and training ISIC International Standard Industrial Classification UN United Nations IT Information technology UNDP United Nations Development Program JOSM Java OpenStreetMap Editor UNESCO United Nations Educational, Scientific and Cultural Organization km Kilometer UNHCR United Nations High Commissioner on Refugees LCFA Learning Competency Framework and Approach UNICEF United Nations Children’s Fund LFP Labor force participation USAID United States Agency for International Development LFS Labor Force Surveys VAT Value added tax LGED Local Government Engineering Department WASH Water, sanitation, and hygiene LGIs Local government institutions WB World Bank MICS Multiple Indicator Cluster Survey WBG World Bank Group MSMEs Micro, small, and medium enterprises WDI World Development Indicators NGOs Non-governmental organizations WDR World Development Report NOAA United States National Oceanic and Atmospheric Administration WFP World Food Programme NTL Nighttime lights WTO World Trade Organization NTMs Non-tariff measures OCHA Office for the Coordination of Humanitarian Affairs (United Nations) OD Origin-destination OSM OpenStreetMap RHD Roads and Highways Department RMG Readymade garment SAE Small Area Estimation SARI ITC Severe Acute Respiratory Infection Isolation and Treatment Centres SDGs Sustainable Development Goals SEZ Special economic zone 18 19 Executive Summary Background Over the past two decades, Bangladesh has achieved an economic transformation enabling formidable reductions in extreme poverty and remarkable human development progress. Between 2000 and 2015, Bangladesh lifted more than 25 million people out of pov- erty. However, the structural transformation of the country’s economy remains incomplete, and economic growth has not benefited all regions and population groups equally. The district of Cox’s Bazar, in southeastern Bangladesh, is an instructive context to under- stand how long-standing and newer growth opportunities and constraints manifest at the local level, remote from Bangladesh’s major growth poles of Dhaka and Chittagong. Potentially exacerbating Cox’s Bazar’s pre-existing development challenges, the district is hosting a large influx of displaced Myanmar nationals (Rohingya). More than 884,000 peo- ple have crossed into Bangladesh from Myanmar, the vast majority since August 2017, more than doubling the population living in the Cox’s Bazar upazilas of Teknaf and Ukhia, which had higher poverty rates than the rest of the district prior to the arrival of Rohingya. The local economy of Cox’s Bazar district cannot spontaneously generate the growth and jobs needed to accompany such a rise in population density. The district’s potential for inclusive growth continues to be constrained by its lack of integration to the national economy and the latter’s growth drivers. Beyond physical connectivity, the district is poorly connected with growth sectors in economic terms, with the current economic structure comprising largely of low-productivity services and agriculture. Several factors limit the inclusivity of the current growth model, based on export-oriented, labor-intensive man- ufacturing. Key constraints affecting Cox’s Bazar include: poor human capital and skills; barriers to women’s economic participation; and a business environment that favors older, established, larger firms to the detriment of new, small firms which tend to be dynamic and innovative. Consequently, local growth opportunities which leverage the district’s natural endowments, such as tourism and aquaculture, remain largely unrealized. 21 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC E x e c u t i v e S u m m a r y As the response to the Rohingya crisis moves to the medium term, a fresh assessment effort, any new growth sectors in the national economy risk doing the same. Currently, of local development challenges and options is needed. The post-2017 humanitarian and Bangladesh’s major firms, mostly based in Dhaka and Chittagong are unlikely to move host government response in Cox’s Bazar district was successful at meeting the basic needs operations from those centers to Cox’s Bazar. Moreover, fledgling local comparative advan- of the Rohingya population. Humanitarian assistance has been estimated to account for 84 tages in Cox’s Bazar, for example in tourism, need policy shifts to facilitate foreign direct percent of the total per capita consumption of displaced Rohingya in 2019. This response investment and an upgraded regulatory framework to promote ecological sustainability. has evolved as conditions and population needs change, and it will continue to do so. The Similarly, for any expansion in the district’s fisheries sector, for example in shrimp exports, humanitarian effort has the potential for generating new economic opportunities for the policy changes are needed to help local industry meet export standards. host population but can only be effectively leveraged once the district’s structural develop- ment challenges are addressed. Support for recently displaced Rohingya and host commu- More generally, the majority of Cox’s Bazar’s small, informal firms are disadvantaged by nities forms part of a broader development agenda for Cox’s Bazar district. the challenging business environment at the national level. The business environment in Bangladesh favors established, connected firms and sectors, and disadvantages new This diagnostic seeks to understand the implications of new and pre-existing drivers and entrants, including young, small establishments and investors trying to expand or start constraints to inclusive growth in Cox’s Bazar in a context of important data and evi- their business. Moreover, access to finance is a pervasive constraint for firms in Cox’s Bazar. dence gaps. The diagnostic: (i) analyzes Cox’s Bazar’s economy before the recent Rohingya Around 60 percent of firms in Cox’s Bazar report credit to be the major impediment to busi- influx; (ii) identifies changes in key factor markets and how they are related to the influx; ness, compared with 40 percent of firms in Chittagong and Bangladesh. More than 80 per- (iii) analyzes key constraints to current and future growth and poverty reduction; and (iv) cent of firms in Bangladesh report that they use their own sources of finance; the same is identifies data, evidence gaps and areas for intervention. The value addition of this diag- true for about 90 percent of firms in Cox’s Bazar. Continued access and quality issues have nostic comes through new analysis of existing and recently collected datasets, combined limited businesses’ ability to leverage digital technologies, with less than 1 percent of busi- with geospatial analysis on travel times and accessibility, to provide insights at district and nesses in Cox’s Bazar reporting the use of information technology in their daily operations sub-district level. in the last Economic Census. As a diagnostic based on currently available data, this report prepares the way for a Cox’s Bazar’s economy cannot readily harness new economic opportunities because of future second phase of work. Currently, key evidence gaps remain that prevent the identi- its low human capital and skills base. With a large share of illiterate adults and a weak fication of specific economic sectors for investment and impede quantifying negative and education system, Cox’s Bazar remains poor in human capital. The lack of locally available positive spillovers of increased humanitarian assistance and the Rohingya influx on the skilled labor may be one reason why the local economy primarily relies on low-productiv- local economy. A second phase of work will aim to fill existing data gaps and foster dialogue ity agriculture and services and has not been able to effectively leverage promising geo- with stakeholders (local government, private sector, development partners, and humani- graphic and economic endowments for tourism, hospitality, or aquaculture. In turn, given tarian agencies) to build consensus on areas for intervention. the structure of the local economy, it is not surprising that there are limited returns to edu- cation until tertiary level. Financial pressures and social norms are the major constraints that keep Bangladeshi children in Cox’s Bazar from attending school and force them to drop out of school early. Key findings and evidence gaps Economic inclusion through productive and remunerative labor market participation Cox’s Bazar remains disconnected from existing forces of growth and income generation for both men and women is constrained by low educational attainment, limited access in Bangladesh. Travel times from the growth poles of Dhaka and Chittagong are too long. to well-paying jobs, and physical distance from the country’s growth poles. These con- Poor transport infrastructure makes it costly for firms to be based in Cox’s Bazar and for straints are further compounded for women through differential access to productive local workers to reach jobs outside the district. Unions around Chakaria have some con- inputs and assets compared to men; women’s role in home-based and caretaking activ- nectivity with Chittagong, but Teknaf and Ukhia, bearing the brunt of increased population ities; market failures and institutions; and social norms constraining women’s mobility. density, will remain largely disconnected even after planned infrastructure upgrades. Women’s potential to generate incomes and engage in productive, paid work outside the home and the farm is further constrained by prevailing norms around asset ownership, The local private sector is largely disconnected from the national growth model, which home- and care-related responsibilities, and mobility. has relied on export-oriented, labor-intensive manufacturing. The readymade garment industry boom at the national level has largely left Cox’s Bazar behind. Absent concerted 22 23 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC E x e c u t i v e S u m m a r y Finally, the demographic profile of the population in Cox’s Bazar underscores the need If fostered appropriately, two secondary factors could transform the growth trajectory of for basic investments in early childhood interventions, maternal and child health, and Cox’s Bazar. The first is the Matarbari energy complex and deep seaport. Proactive mea- expanding access to electricity. The population of Cox’s Bazar district was already younger sures are needed to connect the Matarbari complex to Cox’s Bazar district physically and in than the national average, and this has been reinforced by the arrival of an even younger employment terms. This requires identifying the skill profiles needed in the complex and Rohingya population. Among both populations, the relatively early age at marriage and investing in corresponding local skills development, including for value chains such as trans- motherhood can pose risks to maternal and child health. The district’s children lag behind port and storage. Links are possible to the fledgling growth cluster in Chakaria and some in key markers of early childhood development, with a higher propensity to be malnour- northern unions. Large, export-oriented firms remain unlikely to move a significant share of ished than the national average. Limited access to improved sources of drinking water and their operations to Matarbari unless the district has the necessary infrastructure to connect to sanitation affects both hosts and displaced Rohingya and has important implications for international markets and Dhaka and Chittagong. Finally, expanding power transmission and child nutrition. Even among the host community, access to electricity averages only 11 distribution capacity can directly benefit the host community in the district. hours a day. Expanding access to electricity, including through the use of renewables, and clean cookstoves can yield disproportionate benefits for children and women. Investing in An additional secondary factor is the economic potential linked to the Rohingya camps the early years of the district’s young population will be critical to address inequality and and the inflow of humanitarian and development assistance. Evidence points to increas- increase the productive potential of the population. ing economic activity near the camps, proxied by the growth in nighttime lights. Indeed, Cox’s Bazar is one of few Bangladeshi districts outside of Dhaka and Chittagong displaying signs of growth on this proxy indicator in recent years. The aid economy appears to be gen- erating new types of work for the host community, not necessarily restricted to the imme- Growth drivers diate environs of camps. The presence of humanitarian workers and organizations in the district is likely to spur greater demand for housing, office space, transportation services, In light of these constraints, and based on the existing evidence base, this report iden- restaurants, and hospitality services, and for local facilitation such as translation services. tifies four sets of key growth drivers in the district. These may be classified into major Potential exists to increase local procurement for the humanitarian effort, if the district’s growth drivers, which aim to leverage pre-existing growth opportunities and ease struc- economy can reliably cover basic needs for displaced Rohingya. This will require supporting tural constraints to inclusive growth, and secondary growth drivers, which take advantage local farmers and fishers to create well-managed producer and marketing organizations. of emerging opportunities. Table ES1: Inclusive growth drivers for Cox’s Bazar, potential payoffs, Concerted efforts to promote local comparative advantage offer a first major growth and constraints driver. These efforts, for example in tourism and hospitality and aquaculture, can lever- age the district’s natural endowments, while ensuring and promoting ecological sustain- ability. Activating these growth opportunities will require a conducive business environ- Potential drivers of inclusive growth Growth potential Constraints to inclusive growth ment that promotes ecologically responsible investments, provides a level playing field, Tourism • High-value international orienta- • Lack of infrastructure, branding and establishes linkages with the local economy. Investments in connecting and facilitat- tion, ecologically sustainable • Lack of skills ing infrastructure will help develop these value chains and linkages. These will also need to • Job creation be accompanied by investments in specific labor market skills for the host community, so Aquaculture • Export oriented • Small scale; need certification, • Job creation; links to assistance quality standards that new job opportunities in these sectors are accessible to the local population. economy • Lack of skills/technology adop- tion, infrastructure The second major growth driver identified in the report is improving connectivity within Connective • Improved travel times for people • Reliance on road transport, lim- Cox’s Bazar, and from the district to the rest of the country. For the district to leverage its infrastructure and goods, access to jobs and ited volume capacity, congestion services natural endowments, its transport network and infrastructure will need to be developed Humanitarian • Job creation, ancillary services • Lack of coordination with govern- through the use of multiple modes, while increasing capacity to handle high traffic volumes assistance • Increased demand for local ment investments and reducing travel times for people and goods. Within the district, the lack of connective production • Need aggregation and capacity building for scale infrastructure currently limits access to existing clusters of economic activity and growth, Matarbari port and • Backward and forward linkages • Inherently capital intensive new growth potential, and equitable access to services including tertiary education. energy complex can create jobs and growth • Needs skills, building linkages with local firms, connective infrastructure 24 25 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC E x e c u t i v e S u m m a r y Development interventions by multilateral agencies such as the World Bank are Figure ES1: Key policy recommendations Figure ES1. Key policy recommendations designed to support both host communities and the displaced. By facilitating partner- ships between the humanitarian community and government, development agencies can Investing early Strengthening Expanding support investments in service delivery and monitoring in the district, while strengthening in productive productive economic Bridging evidence gaps national systems. Local government institutions need greater capacity in last-mile service potential capacity opportunities delivery and advocacy for local people’s needs in development priorities. Access to clean Human capital Private sector led Disaggregated, water, improved and skills job creation timely, reliable data Taken together, these findings call for a comprehensive, evidence-based, multi-sector sanitation, and statistics approach to improve inclusive growth and welfare in Cox’s Bazar. This includes raising electricity living standards by investing in portable assets such as health and education; remov- Maternal and child Resilient Market integration New analytical work ing distortions in the local investment climate; and creating a level playing field for the health livelihoods and connectivity district’s private sector, with access to adequate services and infrastructure. Improving physical and economic connectivity to growth opportunities, while investing in local Green Resilient Inclusive people’s capacities and skills, will open a wider set of economic opportunities for all in Cox’s Bazar. Policy recommendations Early investments in productive potential The report’s policy recommendations aim to foster inclusive economic growth in local Access to clean water, improved sanitation, and electricity communities by increasing the productive capacity of the population and its range of eco- • Inclusive – Expand access to private sources of clean water, and reduce reliance on nomic opportunities, while investing in children early to ensure a firm foothold for their shared sources, particularly in host communities close to Rohingya camps. future potential. These recommendations focus on ways to expand the economic pie, as • Inclusive – Broaden access to improved sanitation facilities across the district. well as the ability of different groups to benefit from that growth. Policy recommendations • Inclusive – Increase water, sanitation, and hygiene (WASH) investments in camps to focus on areas with a comparatively solid evidence base, while encouraging investments in reduce reliance on shared facilities. more and better data. • Inclusive – Promote investments in electricity distribution and transmission capacity to increase the number of hours of grid electricity across the district, particularly in host The recommendations follow the Green, Resilient, and Inclusive Development (GRID) communities close to camps. framework. Given the district’s natural endowment and its exposure to climate risk, all • Green – Invest in solar and wind-based energy generation to expand access to electric- development interventions must, at a minimum, do no ecological harm and, where feasi- ity. Improve coordination between international organizations and local government to ble, invest in pathways to ecologically responsible and sustainable livelihoods. At the same expand programs and subsidies to increase the use of solar panels. time, investments are needed to build resilience in the local economy and livelihoods, • Resilient – Modify the scheme of national electricity prices to achieve a cost recovery enabling populations to bear risk and uncertainty without eroding productive assets and rate, which is essential to the sustainability of the system. capital. Finally, leveling the playing field in terms of access to services, jobs, and growth • Resilient – Strengthen local government mandates, allowing community preferences to opportunities is essential to tap into the productive potential of all residents of the district, be reflected in budget allocations and expenditures, particularly outside Municipal and and build resilience among the Bangladeshi and Rohingya communities. City Corporations. • Resilient – Strengthen links and communication between local government entities and humanitarian agencies to better align resource use with local needs and strengthen institutional capacity to respond to development needs. 26 27 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC E x e c u t i v e S u m m a r y Maternal and child health Resilient livelihoods • Resilient – Expand nutritional programs among hosts, including early detection • Green - Review and reform input subsidy policies with a special focus on fertilizers. of child malnourishment and programs for good nutrition practices among young Complement with extension services for more efficient fertilizer use, and environ- mothers, awareness and adherence to vaccinations, and pre- and post-natal care. mentally friendly alternatives for improving soil quality. This can increase resilience among vulnerable host households in the context of • Inclusive - Expand mechanization for seed establishment, crop protection, irriga- COVID-19 in the short term, and of undernourishment in the medium term. tion, and harvesting, particularly among small farmers. • Inclusive – Increase coordination between humanitarian actors and local govern- • Inclusive - Expand the collateral registry’s mandate to include movables and immov- ment to expand nutritional programs already present in camps to host communi- ables as collateral. This will broaden access to credit. ties. This can guarantee access to basic nutrients for all children. • Resilient - Expand infrastructure projects to protect populations from environmen- • Inclusive – Expand social assistance support to female-headed households, particu- tal disasters. This will increase the resilience of local communities. larly those headed by young mothers, so that they do not have to trade off caring for • Resilient - Develop the insurance sector to expand access to insurance instruments young children and earning a living. among farmers and households. • Inclusive – Expand programs to close immunity gaps among children living in camps, • Green - Improve environmental and forest regulations to manage climate risk. and protect against future infectious outbreaks through scale-up and strengthening • Green - Given climate-change and environmental risks, sector-specific measures will of routine immunization services. be needed to help farmers adapt their cropping systems and fisheries activities. Investing in productive capacity Expanding economic opportunities Human capital and skills Private sector-led job creation • Inclusive – Provide pro-poor scholarships and conditional cash transfers to women, • Resilient/Green – A package of coordinated interventions are needed to radically the economically disadvantaged, and students at higher risk of dropping out. change the orientation and earnings potential of tourism and aquaculture. To realize • Resilient – Promote business and vocational skills programs to foster self-employ- the potential of the tourism sector, concerted effort is needed, including in attracting ment in service-related activities. foreign investment, infrastructure and information and communication technology • Inclusive – Pilot and expand implementation of the Myanmar curriculum for Rohingya (ICT) services, marketing, and environmentally sustainable tourism infrastructure children in camps, while easing mobility and safety concerns to increase enrollments. and planning. Fishing and aquaculture development could be fostered, if comple- • Inclusive – Provide education certification for primary and secondary school com- mentary investments are made to facilitate storage, transport, marketing, and qual- pletion for Rohingya children. ity and standards assurance and certification. • Resilient – Expand programs that: provide psychosocial support to Rohingya youth • Inclusive – Expand initiatives to use locally sourced and procured food for food assis- and adolescents; expand awareness of and access to sexual and reproductive health tance in Rohingya camps. services; and support survivors of sexual and gender-based violence and trauma. • Green – Create conditions for hosts to take advantage of rising demand for local • Inclusive – Improve school learning environments and teacher-student ratios to products due to the Rohingya. A larger local market reduces transaction and mar- boost education quality and reduce dropouts. keting costs for perishable products and, by limiting reliance on imports, has the • Inclusive – Promote support from development partners for government efforts to potential to reduce the carbon footprint. This could spur diversification in produc- strengthen human capital and skills. This includes encouraging private-sector entities tion and encourage local farmers to invest in productive improvements. involved in infrastructure and tourism to develop employment-oriented skills and • Green – Expand low-skill job opportunities in farming, construction, and environ- vocational training programs, better preparing youth for employment in these sectors. mental restoration close to camps. • Resilient – Invest in market-relevant skills for migrants. This can boost migration’s • Resilient – Engage the private sector in humanitarian assistance by sharing techno- potential as a welfare driver, reducing pressure on local labor markets. logical capabilities and expertise, adapting business models to sell goods and ser- • Resilient – Implement a system for ongoing real-time health surveillance. vices to the Rohingya. 28 29 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC E x e c u t i v e S u m m a r y Market integration and connectivity Government and partners can act now to leverage opportunities for inclusive growth • Inclusive – Increase tracking and monitoring capacity for shipments. Introduce in Cox’s Bazar, while filling evidence gaps. The results of this diagnostic can inform that a regulatory system to ensure that large and small transport operators can meet work. Government can play a critical role in coordinating private, public, humanitarian, appropriate quality and safety standards. and development actors to leverage local growth potential and help capitalize on the dis- • Resilient – Promote investment in storage, transport, marketing, quality and stan- trict’s natural advantages. Meanwhile, important data and evidence gaps remain and will dards assurance, certification, and other sector-specific needs for tourism and hos- need to be filled to better understand: (i) how the local economy is evolving in response pitality, aquaculture, and high-value crops. to the Rohingya influx; (ii) sector-specific challenges to growth for the local private sector; • Resilient – Upgrade infrastructure and ICT services for the international business and (iii) the potential for humanitarian and development interventions to work at scale to clientele. improve the livelihoods of hosts and the displaced. • Inclusive – Ease connectivity challenges so that Cox’s Bazar’s northern unions can exploit their comparative advantage as a hub for non-agricultural activities. • Inclusive – Connectivity investments focused on upgrading existing networks can lower the cost of accessing jobs, inputs, and markets, while better connecting south- ern Cox’s Bazar to more economically vibrant northern unions. • Inclusive – Adapt the policy and regulatory framework on infrastructure develop- ment to include multiple transport modes, service quality, and road safety. • Inclusive – Improve quality and expand access to digital infrastructure in the district through fiber-optic infrastructure, 4G capacity expansion, and telecom towers. Bridging evidence gaps • Strengthen statistical capacity to produce and share subnational expenditure data. This will help policy makers and stakeholders better understand how expenditure relates to health, education, and other outcomes and can enable efficiency gains in public spending. • Foster investments in data and evidence on constraints to firm entry, growth, and dynamism that are specific to the district. Similarly, invest in data and evidence on the potential for improved linkages with local businesses to deliver new invest- ments in tourism and aquaculture, as well as on how government, humanitarian, and development investments may affect job creation in Cox’s Bazar. • Promote research on how the Rohingya influx has affected service delivery. Apply research results to inform an appropriate policy response. • Invest in research to generate evidence on new employment opportunities for women and school dropouts in the camp-related economy, as well as on how more inclusive vocational programs might be linked to productive opportunities for work. • Promote research to understand how the large influx of humanitarian assistance has affected local host communities, both in terms of potentially increasing competition for low-skill jobs, and providing new work and income-earning opportunities for hosts, including better-educated youth. 30 31 CHAPTER 1. Introduction Over the past two decades, Bangladesh has achieved an economic transformation enabling formidable reductions in extreme poverty and remarkable human develop- ment progress. The expansion of labor-intensive manufacturing and exports, primarily garments, has driven the country’s economic gains, which have supported sustained per capita income growth. Between 2000 and 2015, Bangladesh lifted more than 25 million people out of poverty. Robust jobs growth has been accompanied by increasing labor force participation among women, which rose by 10 percentage points between 2003 and 2016, from 26 to 36 percent (Farole and Cho 2017). However, the structural transformation of Bangladesh’s economy remains incomplete, and since 2015, unsolved pre-existing constraints and emerging risks (including the COVID-19 pandemic) have threatened to slow the country’s progress. A large share of the workforce is un- or under-employed, and average educational attainment is low. Agriculture still accounts for 40 percent of employment, despite its declining contribution to economic growth. Urban poverty reduction has stagnated, while a suboptimal business environment and regulations limit private sector growth (Zafar et al. 2020). Connectivity challenges make it harder to integrate markets within Bangladesh and link the country to regional and global markets. Growth in the agricultural and service sectors has slowed, while deteriorat- ing competitiveness in the readymade garment (RMG) industry, the absence of diversified exports, and the impacts of COVID-19 have constrained job creation in manufacturing. To date, asset accumulation and income diversification from the rural sector have been the main drivers of household welfare gains. However, the marginal return of accumulation is reaching a limit, and constraints to taking advantage of these returns are slowing growth and poverty reduction. 33 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 1 – I NTRO D U C T I ON These structural issues are not uniform across the country and could exacerbate regional Map 1-1: Bangladesh and Cox’s Bazar district: major roads and seaports3 Map 1-1. Bangladesh and Cox’s Bazar district: major roads and disparities. Subnational disparities affect not just poverty reduction and local economic seaports2 growth, but also resilience and the capacity of subnational entities to manage exogenous shocks such as climate-change-related disasters or demographic shifts. Indeed, equity and spatial development considerations are becoming increasingly important in Bangladesh, given the need to foster new sources of inclusive growth, leverage local endowments, and mitigate local challenges in light of the reemerging East-West welfare divide (Hill and Genoni 2019). The district of Cox’s Bazar is an instructive context to understand how long-standing and newer growth constraints and opportunities in Bangladesh manifest at the local level— and how local action can accelerate national agendas.1 Opening growth paths in Cox’s Bazar can boost livelihoods for all district residents while advancing country-wide eco- nomic inclusion and development goals. Leveraging these opportunities require under- standing Cox’s Bazar’s distinctive geography and development trajectory. Historically, the district’s location at the southeastern tip of Bangladesh and its lack of connectivity to Dhaka the major growth poles of Dhaka and Chittagong (Map 1-1) have constrained its growth options. Combined with its relatively low endowment in assets and human capital, the dis- trict’s distance from urban drivers of growth and job creation has largely excluded Cox’s Bazar from the garment-industry boom. Mongla Cox’s Bazar’s location, bounded by the Bay of Bengal to the south and the west and shar- Chittagong ing a border with Myanmar to the east, also determines potential comparative advan- tages for tourism and international trade, which have remained largely untapped. This Road classifications Martarbari (proposed) locational advantage allows for the planned construction of a new energy complex and Primary deep-sea port at Matarbari, in the district’s Maheshkhali upazila (subdistrict). This is one Secondary Cox’s Bazar Tertiary of a number of fast-track mega-projects which are being prioritized by the Government of Bangladesh.2 These projects are critical to national growth, improving the competitive- Major sea port ness of Bangladesh’s exports, and expanding access to international trade. Source: Outline shape file, Bangladesh Bureau of Statistics; transport data © OpenStreetMap contributors. Cox’s Bazar is hosting a large influx of recently displaced Myanmar nationals (the Rohingya), whose presence signals new inclusive growth challenges and opportunities. Cox’s Bazar hosts more than 884,000 displaced Myanmar nationals, of whom 725,000 Rohingya have crossed into Bangladesh from Myanmar since August 2017. The influx has more than doubled 1 Administratively, Bangladesh is divided into eight divisions and 64 districts, with each district further the population living in the Cox’s Bazar upazilas of Teknaf and Ukhia (see Map 1-2 for location divided into upazilas or sub-districts. Each upazila is comprised of a number of unions, consisting of of main camps), which had higher poverty rates than the rest of the district prior to the influx several villages, with the exception of urban and metropolitan areas, which are designated as pau- (Map 1-3 and Map 1-4 show zila (district) level pre-Rohingya population density and poverty rasavas (municipalities) or city corporations. Chittagong and Cumilla are the only two city corpora- estimates; see also Box 1). tions in Chittagong division (of 12 in the country). Within Cox’s Bazar district, there are four metropoli- tan areas – in Cox’s Bazar Sadar, Chakaria, Maheshkhali, and Teknaf. 2 These projects include, for instance, the Padma Multi-Purpose Bridge Project; Dhaka Metro Rail Project and the Dhaka Elevated Expressway; and the construction of single line dual gauge railway 3 All transport data used in the maps in this report are from Open Street Map,© OpenStreetMap track from Dohazari to Cox’s Bazar via Ramu and Ramu to Gundum, near Myanmar. contributors. 34 35 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 1 – I NTRO D U C T I ON Map1-2: Map 1-2. Locations Locations of of recently recently displaced displaced Rohingya Rohingya camps camps in Teknaf in Teknaf and Ukhia Map 1-3: Population density by zila, Map 1-4: Poverty headcount by zila, and Ukhia upazilas upazilas as of May 2018 estimated 2018 (does not include 2016 (upper poverty line) Map population) density by 1-3. Population Rohingya Map 1-4. Poverty headcount by zila, estimated 2018 (does not zila, 2016 (upper poverty line) include Rohingya population) 0 75 150km 0 75 150km Camp 4, 9, 17 / Kutupalong Balukhali 623,350 Camp 14 / Hakimpara 30,680 Balukhali MS Myanmar Dhaka Dhaka Camp 15 / Jamtoli 45,425 Camp 21 / Chakmarkul 12,800 Camp 16 / Bagghona / Potibonia 22,400 Chittagong Chittagong Camp 22 / Unchiprang 21,600 Cox’s Bazar Cox’s Bazar Camp 23 / Shamiapur District population density Poverty head count ratio 13,100 Population per km2 Percentage of population below the poverty line 85 500 1200 2000 8200 2 15 30 45 72 Teknaf Source: World Bank staff estimates based on Source: World Bank 2020b, based on HIES 2016. Population Census 2011. Bay of Bengal Highly localized within these two upazilas, the Rohingya influx and the arrival of sig- nificant aid resources have brought major changes to a district that, before the influx, Camp 25 / Alikhali 9,550 reported poor development outcomes relative to national averages. Prior to the Rohingya Camp 24 / Leda 35,700 influx, Cox’s Bazar already had a comparatively large share of its population working in the Camp 26 / Nayapara primary sector (43 percent versus 38 percent nationally), lower literacy rates (55 percent Dhaka 70,850 versus 60 percent nationally), and less access to electricity (52 percent versus 76 percent Chittagong Camp 27 / Jadimura nationally) (Household Income and Expenditure Survey, HIES, 2016). Additionally, the 14,230 sudden increase in population density, unaccompanied by growth, places unprecedented pressures on the area’s natural resources in a context of high vulnerability to natural calam- 0 1.5 3km ities like cyclones and floods.4 Highways Roads International boundary Collective zila Collective zila with host community Myanmar Upazila Union 4 Both Cox’s Bazar and Chittagong have been identified as South Asian cities which are at risk of all four Source: Inter Sector Coordination Group, Cox’s Bazar, 2018. major hazards: flooding, earthquakes, landslides, and cyclones (Ellis and Roberts 2016). 36 37 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 1 – I NTRO D U C T I ON The Rohingya influx has been accompanied by a large-scale humanitarian response, 1-5:Nighttime Map1-5: Map Nighttime lights lights as as a a marker marker of economic of economic activity, activity, Bangladesh Bangladesh and in a context of weak local governance that has made it harder for host communities to and Cox’s Cox’s Bazar, 2019 2019 Bazar, make their voices heard (Fox and Menon 2008). Between 2017 and 2020, an average of US$634 million annually was allocated to the Rohingya crisis response.5 The humanitar- Bangladesh Cox’s Bazar ian effort has been largely successful in delivering basic needs and food security to the displaced population (World Bank 2020c). However, local governments have limited funds and capacities to collaborate actively in the response. In general, district and local govern- ments in Bangladesh have a limited role in charting policy and shaping public investments. As a result, there are often few avenues for local communities to influence policy agendas that may affect their welfare. In particular, elected local government representatives rarely participate in the identification, appraisal, approval, implementation, and monitoring of investment projects funded through the Annual Development Plans (World Bank, forth- coming). In this context, durably aligning the interests of displaced Rohingya people and local host communities in Cox’s Bazar has proved challenging. Camps Deciles Deciles 1 1 Growing economic activity around Rohingya camps may already be advancing economic 2 2 convergence within Cox’s Bazar and may be contributing to a positive national trend. 3 3 4 4 Bangladesh’s two major urban growth poles, Dhaka and Chittagong, remain outliers in 5 5 6 6 their concentration of economic activity. However, there is some recent evidence which is 7 7 suggestive of increasing economic activity in some districts, including Cox’s Bazar. Using 8 8 9 9 nighttime lights (NTL) as a proxy of economic activity, the concentration of economic activ- 10 10 ity in the Dhaka-Chittagong corridor is readily evident, as are marked regional disparities Source: World Bank staff calculations using United States National Oceanic and Atmospheric Administration (NOAA) in economic density (Map 1-5). However, over the period 2014 to 2019, the Gini coefficient nighttime light intensity data. of the average NTL intensity between districts by year fell from 0.44 to 0.37, and from 0.28 Note: Deciles created using nighttime light intensity from March 2019, excluding values below 0.1. to 0.20 if the top 10 percent of districts in NTL density (primarily in the Dhaka-Chittagong corridor) are excluded. In 2014, Cox’s Bazar ranked 46th out of 64 districts in NTL intensity (proxying economic activity), whereas by 2019, its rank had improved to 33rd. Among all the districts in the country, Cox’s Bazar was the only one that moved up in the ranking by more than 10 places in this period. Most of the NTL intensity change in Cox’s Bazar took Box 1: Poverty in Cox’s Bazar place after 2017, in areas near the Rohingya camps and along the road connecting Cox’s Bazar Sadar to the upazilas of Teknaf and Ukhia. This is suggestive of increasing economic activity (and electrification), potentially related to the influx of Rohingya and associated Cox’s Bazar as a whole has relatively low poverty rates compared to the humanitarian assistance and aid flows. national average, but some of its sub-districts are much poorer. According to Bangladesh’s latest poverty assessment, the national poverty headcount rate (upper poverty line) was 24.5 percent in 2016, while the poverty rate in Cox’s Bazar district stood at 16.5 percent (World Bank 2019a), among the lowest in Chittagong division. In fact, between 2000 and 2016, the district’s poverty rate dropped 24 percentage points (Figure B1-1). However, significant disparities exist within Cox’s Bazar. According to 2010 small area poverty estimates, prior to the Rohingya influx, the two upazilas of Teknaf and Ukhia had high poverty rates (38.2 and 37.8 percent respectively). These rates placed Teknaf and Ukhia, 5 United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA) Financial Tracking Service. “Bangladesh: 2020 Joint Response Plan for Rohingya Humanitarian Crisis (January- December).” https://fts.unocha.org/appeals/906/summary 38 39 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 1 – I NTRO D U C T I ON along with Maheshkhali, on par with some of the poorest districts in the coun- B1-3: Monthly Figure B1-3. Figure Monthly per capita expenditure per capita expenditure (Tk), upazilas in (Tk), upazilas in Cox’s try (Figure B1-2). As Figure B1-3 shows, average per capita expenditure in Cox’s Cox’s Bazar, 2010 small area estimates Bazar, 2010 small area estimates Bazar Sadar is almost 20 percent higher than in Teknaf and Ukhia, and these dif- ferences are statistically significant. A simple average of per capita expenditure Cox's Bazar Sadar 2577 between Teknaf and Ukhia is still 10 percent lower than the average for Cox’s Chakaria 2427 Bazar Sadar, Chakaria, Pekua, and Ramu. Pekua 2295 Ramu 2288 Kutubdia 2229 Figure B1-1. Figure Poverty rate, B1-1: Poverty rate, districts districts in in Chittagong Chittagong division division versus versus Ukhia 2191 Bangladesh, 2000 Bangladesh, - 2016 2000 - 2016 Teknaf 2160 Maheshkhali 2104 0 500 1000 1500 2000 2500 3000 74,80 63,50 57,80 Source: Ahmed et al. (2010). 53,15 52,20 48,86 45,70 44,80 42,00 40,60 39,30 34,40 32,89 31,90 32,00 29,38 28,76 24,52 23,67 16,42 13,52 12,95 10,38 8,10 This diagnostic seeks to understand the implications of new and pre-existing drivers and constraints to inclusive growth in Cox’s Bazar. The diagnostic will: (i) analyze Cox’s Bazar’s economy before the recent Rohingya influx; (ii) identify changes in key factor markets and ria g r ni a li sh ur ur i ri n at za how they are related to the influx; (iii) analyze key constraints to current and future growth ha on ill ba ha Fe dp ip de am ba m Ba ak ag ar ch hm an la Co an ng nd No x's itt ra and poverty reduction; (iv) pinpoint opportunities to accelerate inclusive growth, address- ng Ch ks m Ra ag Ch Ba Co Ba La ah Kh ing the needs of host communities and displaced Rohingya; and (v) identify areas for in- Br 2000 2016 tervention along with data and evidence gaps. The value addition of this diagnostic comes Source: Staff calculations using HIES 2000, HIES 2016-17, and Census 2001. through new analysis of existing and recently collected datasets, combined with geospatial Note: Figures present the national upper poverty rate by division and district. The poverty rate for 2000 was calculated using Small Area Estimation (SAE) analysis on travel times and accessibility, to provide insights at district and sub-district lev- el (Box 2). As a diagnostic based on currently available data, this report prepares the way for a future second phase of work. That phase will focus on filling existing data gaps and involve dialogue with a range of stakeholders (local government, private sector, develop- Figure B1-2. Figure B1-2: Poverty Poverty rate, rate, upazilas upazilas in Cox’s Bazar, in Cox’s Bazar, 2010 small area ment partners, and humanitarian agencies) to build consensus on areas for intervention. 2010 small estimates area estimates Maheshkhali 40.2 Teknaf 38.2 Ukhia 37.8 Box 2: Data sources for this diagnostic Ramu 34.3 Kutubdia 31.1 Pekua 30.9 The Cox’s Bazar Inclusive Growth Diagnostic draws on several existing and Chakaria 28.5 newly available datasets that allow for analysis at the national, regional, dis- Cox's Bazar Sadar 26.2 trict, and sub-district level. These include census data and sample surveys, as Source: Ahmed et al. (2010). well as administrative and geospatial data. The report’s principal data sources are the following: 40 41 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 1 – I NTRO D U C T I ON The Cox’s Bazar Household Panel Survey (CBPS) was designed to assess the sources are the Annual Yearbook of Agricultural Statistics (BBS 2017), District implications of the 2017 Rohingya influx into the Bangladeshi district of Cox’s Statistics Cox’s Bazar (2011), Annual Primary School Census (2018-2019), and Bazar on the living standards and welfare of the host population. The CBPS the Agricultural and Rural Statistics survey (ARRS 2018), as well as data from the is the result of a partnership between the Yale Macmillan Center Program on Bangladesh Bureau of Educational Information and Statistics (BAMBEIS 2018), Refugees, Forced Displacement, and Humanitarian Responses (Yale Macmillan the Demographic and Health Survey (DHS 2014), and the Multiple Indicator PRFDHR); the Gender & Adolescence: Global Evidence (GAGE) program; the Cluster Survey (MICS 2019). Poverty and Equity Global Practice of the World Bank; and the State and Peacebuilding Fund (SPF) administered by the World Bank. The sampling strat- Spatial analysis has been conducted using data on administrative boundar- egy for this survey aimed to produce reliable statistics for the Rohingya living ies (Local Government Engineering Department [LGED] 2018), schools (all in camps and the host population living in Cox’s Bazar district. To distinguish types) (LGED 2018), OpenStreetMap (OSM late April 2020), and High-Resolution between host communities that are more or less affected by the arrival of Population Density Maps (Facebook and CIESIN 2018). The latter is a gridded the Rohingya, the survey’s sampling strategy uses a threshold of three hours’ population model distributing 2018 projections of census data per union to pixels walking time from a campsite to define two strata for hosts, in addition to the deemed inhabited by a country-specific convolutional neural network built by stratum comprised of Rohingya in camps: (i) host communities with potentially Facebook. Essentially, the model allocates population only to where residential high exposure to the displaced Rohingya, and (ii) host communities with poten- buildings are detected, leaving uninhabited areas blank, improving accuracy. tially low exposure. The CBPS baseline implemented in 2019 and subsequent phone surveys are representative at the strata level. The Household Income and Expenditure Survey (HIES) is the main official source of information about household consumption, poverty, and income in Bangladesh. HIES 2016/17 data was collected from April 2016 through March This diagnostic follows in spirit the framework put forth in World Development Report 2017 and, unlike previous rounds of the survey, is representative at the district 2009: Reshaping Economic Geography (WDR 2009). The WDR argued that the economic level and division level - by rural and urban areas. integration of lagging or underdeveloped areas can be fostered through interventions that promote density and agglomeration economies, reducing effective distance between labor The Economic Census 2013 is the most comprehensive enumeration of the and the areas where the returns to labor are the highest, and reducing barriers to integra- full set of economic units belonging to Bangladesh. The third economic census tion of markets—or division—within and across countries. Applying this framework to Cox’s was conducted between March and May 2013 across the country and aimed to Bazar district, this report examines the constraints and opportunities facing the district in measure the structural changes in Bangladesh’s economy during the preceding light of an exogenous increase in population density, in the context of lagging socio-eco- decade (BBS 2015a). The Economic Census includes all non-agricultural estab- nomic development and limited economic linkages between the Rohingya and the host lishments in the country. community. Cox’s Bazar, particularly the southern sub-districts that are most affected by the population surge, remain distant from national growth poles in terms of transport con- Bangladesh’s Population Census 2011 is the most complete data set containing nectivity and of effective links to the economic sectors that have driven growth, exports, information about the size of the country’s population, as well as socio-economic and employment in Bangladesh. However, the district enjoys two new opportunities to and socio-demographic characteristics. The population census provides data at turn this state of play into an advantage. These are the planned investments in and around national, division, zila, and thana-upazila levels and distinguishes urban and rural the Matarbari port and energy complex, and the significant inflow of humanitarian assis- populations. The population census was conducted in March 2011 (BBS 2015b). tance into Teknaf and Ukhia. Importantly, these potential advantages for inclusive growth can only become operative on certain conditions: if targeted policies are in place to help This inclusive growth diagnostic also draws on other surveys and adminis- the people of Cox’s Bazar grasp the emerging opportunities; if institutions respond effec- trative data from the Bangladesh Bureau of Statistics (BBS). Among these tively to local preferences; and if there are favorable conditions for market forces to work. 42 43 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC This report is organized as follows: Chapter 2 examines the state of fundamental endow- CHAPTER 2. ments in the district, including demography, geography, and human capital. It also reviews the district’s existing stock of connecting infrastructure. How these endowments shape Cox’s Bazar’s economic structure and the welfare and livelihoods of its residents is the focus Fundamentals: of Chapter 3. Chapter 4 highlights constraints to inclusive growth and identifies opportuni- ties for investment and policy action to advance inclusive development. Chapter 5 presents People, land, and recommendations for policy and programming and suggests areas for further study.6 infrastructure This chapter examines fundamental endowments that shape growth potential and com- petitiveness in Cox’s Bazar. It highlights endowment gaps that have constrained the dis- trict’s success in advancing inclusive growth, but also distinctive assets that hold promise for the future. The chapter discusses four areas: demographics; human capital and living conditions; geography; and connective infrastructure. Demographics and density Cox’s Bazar district accounts for just 1.7 percent of Bangladesh’s total population, but the district’s population has been growing relatively fast. Cox’s Bazar district’s estimated population growth rate of 2.33 percent over the 2016-2021 period is more than one-and- 6 Some sections of this report could not be completed as planned, due to COVID-19-related travel and a-half times the national average of 1.39 percent, and the highest among zilas (districts) other restrictions. Affected components primarily include a planned assessment of the current deliv- in Chittagong division (BBS 2015). Assuming that the district’s population growth rate ery model of humanitarian assistance, along with documentation of pilot initiatives inclusive of the has remained steady in recent years, the estimated population of Cox’s Bazar (excluding host community. The diagnostic’s narrative has been updated to reflect developments in Cox’s Bazar recently displaced Rohingya) in 2019 was 2.5 million (BBS 2019). since the outbreak of the pandemic and the imposition of restrictions on some forms of economic activity. New data collection is planned in the second phase of work to understand how local economic activity has evolved since the 2017 Rohingya influx. 44 45 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE The arrival of displaced Myanmar nationals (Rohingya) in 2017 has dramatically altered Map 2-1: Population of Cox’s Bazar Map 2-2: Population of Cox’s Bazar the demographic profile of Cox’s Bazar. Cox’s Bazar’s location, particularly its long land Map 2-1. district, byPopulation of Cox’s upazila, pre-2017 Rohingya Map 2-2. district, Population by of Cox’s upazila, post-2017 Bazar influx district, by upazila, Bazar district, Rohingya influx by upazila, border with Myanmar, has made it a natural refuge for displaced Myanmar nationals. 9 pre-2017 Rohingya influx9 post-2017 Rohingya influx Within a four-month period beginning in August 2017, the most recent major influx of displaced Rohingya from Myanmar increased the population living in Cox’s Bazar by 31.7 0 20 40km 0 20 40km percent. These displaced people joined over 150,000 Rohingya who had already arrived in Cox’s Bazar since the late 1970s (UNHCR 2018).7 While the Rohingya community represents Dhaka C H I T TAG O N G Dhaka C H I T TAG O N G less than 1 percent of Bangladesh’s total population and 3 percent of the population of Chittagong Pekua Chittagong Pekua Chittagong division, it comprises 40 percent of the population of Cox’s Bazar district (rela- Kutubdia Kutubdia tive to 2011 census population estimates, which do not include Rohingya). While a 1 per- cent increase in population may have limited effects, a 40 percent increase in population Chakaria Chakaria Matarbari BANDARBAN Matarbari BANDARBAN can be expected to place substantial pressures on local infrastructure, markets for food (approximate) (approximate) basic necessities, and labor markets. Maheshkhali Maheshkhali Within Cox’s Bazar, the two upazilas most affected by arriving Rohingya have been Cox’s Bazar Cox’s Bazar Teknaf and Ukhia, which border Myanmar. Before the recent influx, around 45,000 and Cox’s Bazar Cox’s Bazar Sadar Ramu Sadar Ramu 123,000 Rohingya were already living in Teknaf and Ukhia, respectively. In the second half COX’S COX’S of 2017, the influx of displaced Rohingya increased the population living in Teknaf and BAZAR BAZAR Ukhia by 38 percent and 150 percent, respectively (The absolute numbers of newly arriving Ukhia Ukhia Rohingya reached some 125,000 persons in Teknaf and 600,000 in Ukhia.) This implies that the share of the Rohingya community in the total population of these upazilas rose sharply. In Ukhia, displaced Rohingya now account for as many as 4 out of 5 inhabitants. In Teknaf, 4 Teknaf Teknaf out of 10 persons are now displaced Rohingya (Map 2-1 and Map 2-2). In light of these shifts, Ukhia is now the most densely populated upazila in Cox’s Bazar district, overtaking Cox’s Bazar Sadar, and followed by Teknaf (Map 2-3 and Map 2-4).8 St. Martin St. Martin Dwip Dwip The population of Cox’s Bazar district is younger than the national average, and this has been reinforced by the arrival of an even younger Rohingya population. About 30 per- cent of Bangladesh’s population is made up of children in the 0–14 age group; for Cox’s Population Density (2018) Road classifications Rohingya camps thousands per square kilometer (including Rohingya) Primary Secondary Bazar district, this figure is some 10 percentage points higher, at about 40 percent (Figure Tertiary Ferry 2-1). The recently displaced Rohingya population is overwhelmingly young: 50 percent of 138 150 250 350 520 955 the arriving Rohingya are under 15 years of age. Combined, these demographic character- istics of host and Rohingya communities have important implications for education and Source: LGED, OpenStreetMap, HRSL, UNPD, ACAPS. Notes: Scalar adjustment of 2011 census based on UN Population Division growth estimates health-sector needs going forward (UNDP 2018). 9 Population for 2017 has been estimated by applying the estimated population growth rate to Cox’s Bazar district and using the share of each upazila in total district population from Population Census 2011. Since Population Census 2011 counted only Bangladeshi citizens, the pre-influx population 7 UNHCR estimates that there were 168,000 Rohingya in Cox’s Bazar prior to the 2017 influx. See https:// numbers presented here have been adjusted to incorporate the displaced population known to have data2.unhcr.org/en/situations. settled in Bangladesh before August 24, 2017. Data on growth rates is available in Population projec- 8 Population at upazila level has been estimated using the population growth rate for Cox’s Bazar tion of Bangladesh: dynamics and trends 2011-2061 (BBS 2015). Data on the displaced population included in Population projection of Bangladesh: dynamics and trends 2011-2061 (BBS 2015). present before August 2017 is available at https://data2.unhcr.org/en/situations 46 47 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Figure 2-1. Demographic pyramids, Bangladesh versus Cox’s Bazar, 2016 2-3.Population Map2-3: Map Population density density by Map 2-4.Population Map 2-4: Population density density by Figure 2-1: Demographic pyramids, Bangladesh versus Cox’s Bazar, 2016 by upazila, before 2017 influx, upazila, before 2017 influx, including by upazila, after 2017 influx, upazila, after 2017 influx, including including pre-2017 displaced including newly displaced pre-2017 displaced Rohingya newly displaced Rohingya 65 + Bangladesh Rohingya Rohingya 65 + Cox's Bazar 54 to 65 Bangladesh 54 to 65 Cox's Bazar 45 to 54 Bangladesh Dhaka 0 20 40km Dhaka 0 20 40km 45 to 54 Cox's Bazar Chittagong Chittagong 35 to 44 Bangladesh 35 to 44 Cox's Bazar C H I T TAG O N G C H I T TAG O N G 25 to 34 Bangladesh 25 to 34 Cox's Bazar Pekua Pekua 15 to 24 Bangladesh 15 to 24 Cox's Bazar Kutubdia Kutubdia 0 to 14 Bangladesh 0 to 14 Cox's Bazar Chakaria Chakaria -50% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50% Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) Female Male Maheshkhali Maheshkhali Source: World Bank staff calculations, HIES 2016. Cox’s Bazar Cox’s Bazar Cox’s Bazar Sadar Ramu Cox’s Bazar Sadar Ramu Human capital and living conditions COX’S COX’S BAZAR BAZAR Human capital is increasingly recognized as a critical input to inclusive growth (World Bank 2020a).10 In contrast to geography, for example, human capital is an endowment that Ukhia Ukhia governments can substantially improve through well-understood policy choices and pro- grams. The World Bank’s Human Capital Index (HCI) measures “the productivity as a future worker of a child born today, compared with what it could be if he or she had full health and Teknaf Teknaf complete, high-quality education” (World Bank 2020a). Bangladesh’s HCI performance lags somewhat behind the average for lower-middle-income countries and the average across the South Asia region. This is mainly due to the comparatively poor quality of education in Bangladesh and the continued prevalence of stunting. This section presents data on how St. Martin St. Martin Cox’s Bazar is performing in the two key domains of human capital formation, education Dwip Dwip and health, and in the provision of basic goods and services that contribute to the devel- opment of human capital. Population Density (2018) Road classifications Rohingya camps thousands per square kilometer (including Rohingya) Primary Secondary Education Tertiary Ferry 744 1000 1500 2500 3750 Cox’s Bazar has historically been among Bangladesh’s poorest-performing districts in terms of education. According to the 2011 Population Census, only five Bangladeshi dis- Source: LGED, OpenStreetMap, HRSL, UNPD, ACAPS. tricts had a lower adult literacy rate than Cox’s Bazar, and only Bandarban district was Notes: Scalar adjustment of 2011 census based on UN Population Division growth estimates ranked lower within Chittagong division. Educational attainment among adults remains 10 https://www.worldbank.org/en/publication/human-capital 48 49 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE low, with half of all Cox’s Bazar adults never having attended school, and only a third of The results suggest that Bangladeshi households living in Teknaf and Ukhia, closer to the adults in poor households able to read or write (USAID 2018). According to HIES 2016, Rohingya camps, were already accumulating human capital at lower rates than households only 53 percent of individuals older than 18 in the district are literate, below the national living further away in the district. In this context, the influx of a Rohingya population with average of 59 percent and among the lowest rates in the country. Forty-seven percent of even lower average rates of human-capital endowment risks intensifying competition for the adult population has never attended school, compared to the national average of 42 low-skill jobs in high-exposure areas. Compared to hosts, adult Rohingya have far lower percent and the Chittagong division average of 39 percent. Only 34 percent of the district’s educational attainment – with 90 percent of adults never having attended school or com- adult population has achieved primary education or beyond, compared to 39 percent on pleting less than primary education. In addition, gender gaps become large at secondary average in Bangladesh overall and 43 percent in Chittagong division (HIES 2016). and post-secondary levels, for the few Rohingya who did go to school. Cox’s Bazar lags not only on quantitative measures of educational attainment, but also Figure 2-2: Educational attainment, adults (18+), 2019 (low-exposure areas, on measures suggestive of educational quality, such as student-teacher ratios. According Figure 2-2. Educational high-exposure attainment, areas, and Rohingya adults camps) 13 (18+), 2019 Figure 2-2. Educational attainment, adults (18+), 2019 to preliminary data from the National Primary School Census 2019 and BANBEIS 2018, the (low-exposure areas, high-exposure areas, and Rohingya camps) (low-exposure areas, high-exposure areas, and Rohingya camps) 90% student-teacher ratio for Cox’s Bazar is among the highest in the division and in the coun- 90% try, at 42 and 68 students per teacher in primary school and secondary school, respectively. 62% 62% The National Student Assessment 2017 (BBS 2018b), reflecting standardized test results 50% 50% in mathematics and Bangla for children in grades 3 and 5 nationwide, places Cox’s Bazar 28% 28% among the three lowest-performing districts in the country.11 18% 12% 12% 18% 10% 12% 12% 8% 10% 7% 2% 1% 8% 7% 2% 1% Within Cox’s Bazar, the upazilas of Teknaf and Ukhia were already lagging in educational attainment prior to the influx of displaced Rohingya. A comparison of adult educational Never attended/ Primary Incomplete Secondary Never less attended/ than primary Primary complete Incomplete secondary Secondary and above attainment across areas which were more exposed to the Rohingya influx and those that less than primary complete secondary and above were further away within Cox’s Bazar (CBPS 2019) suggests that the high-exposure areas High-exposure Low-exposure Camps High-exposure Low-exposure Camps (mainly Teknaf and Ukhia) have a higher share of adults who have never attended school Female-male Figure 2-3: Female-male gaps gaps in educational in educational attainment, attainment, (18+), (18+), adultsadults (Figure 2-2).12 High-exposure areas also have lower educational attainment among adults Female-male gaps in educational attainment, adults (18+), 2019 (low-exposure 2019 (low-exposure areas, areas, high-exposure high-exposure areas, areas, and Rohingya and Rohingya camps) camps)camps) for post-primary school levels. Within the district, adult men are more likely to have some 2019 (low-exposure areas, high-exposure areas, and Rohingya 15% education than adult women, or in other words, 52 percent of men never attended school 15% relative to 55 percent of women. The overall pattern of higher adult male educational attainment at the district level appears to be driven by low-exposure areas (Figure 2-3), 7% 6% 7% 6% which have larger shares of male adults with education beyond primary (50 percent in 1% low-exposure versus 41 percent in high-exposure areas). On the other hand, the large share 1% of females with secondary incomplete in low-exposure areas contributed to narrowing the -1% -1% -1% -1% gender education gap at the district level. -3% -1% -1% -1% -1% -4% -3% -4% -7% -7% Low human capital among hosts living near camps raises the risk of competition with -11% -11% Rohingya for low-skill jobs. The findings just summarized on adult educational attain- Never attended/ Primary Incomplete Secondary ment within Cox’s Bazar reflect conditions before the recent influx of displaced Rohingya. Never less attended/ than primary Primary complete Incomplete secondary Secondary and above less than primary complete secondary and above High-exposure Low-exposure Camps 11 Specifically, the results flag Cox’s Bazar as the second-worst performer in Bangla Language and High-exposure Low-exposure Camps among the 10 worst in mathematics at grade 3. For grade 5, Cox’s Bazar had the third-lowest mean Source: World Bank staff calculations, CBPS 2019. among all Bangladeshi districts for both Bangla language and mathematics. Source: World Bank staff calculations, CBPS 2019. 12 To distinguish between host communities that are more or less affected by the arrival of Rohingya, the survey’s sampling strategy uses a threshold of three hours’ walking time from a campsite to define two strata for hosts: (i) host communities with potentially high exposure to the displaced Rohingya, 13 Religious education is included with the “Never Attended/Less than Primary” category. 1.8 percent and (ii) host communities with potentially low exposure. of the host population and 4.6 percent of the camp population report this type of education. 50 51 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE While district level public education spending per student generally tends to increase Sadar and other upazilas had far fewer, accounting for no more than 5 percent of all edu- with poverty rates across Bangladesh, Cox’s Bazar is an exception to this pattern, with cational institutions there (Figure 2-4). Despite the proliferation of NGO-run schools in low per-student expenditure relative to districts with similar poverty rates. Analysis of some parts of the district, such schools account for a disproportionately small share of public expenditure data on education finds a positive correlation between poverty and students (Figure 2-5), as NGO schools are generally small, enrolling 45 students each on per-student spending (Genoni et al. 2019). However, government expenditure per student average. In Teknaf and Ukhia, NGO schools were even smaller, reporting an average of in primary and secondary levels in Cox’s Bazar is significantly lower than in other districts 39 and 35 students each (Table 2-1).16 In general, student-teacher ratios in Cox’s Bazar with similar poverty rates (Bhatta et al. 2019). This relationship might be partially explained district were already high in 2011, averaging 70 students per teacher in primary school by the relatively higher number of students and schools in the area. Indeed, HIES data sug- (Table 2-2). These ratios were particularly high in some upazilas, including Teknaf (132), gest that Cox’s Bazar is among the top 10 districts in terms of students enrolled in primary Chakaria (81), and Maheshkhali (87). education in the country, but it still has fewer primary students than other districts in the division such as Brahmanbaria, Noakhali, Comilla, and Chittagong. Figure 2-4: Share of school types, by Figure 2-5: Share of students by type Educational spending appears to yield lower outcomes in Cox’s Bazar, relative to other upazila, Cox’s Bazar, 2011 of institution and upazila, Cox’s districts that spend similarly per student. Genoni et al. (2019) use data from BOOST Bazar, 2011 2014 and HIES 2016 to assess the relationship between public expenditure and educa- tional outcomes.14 While there is a positive correlation between spending and net atten- dance rates overall, this is only statistically significant at the secondary level. Cox’s Bazar has a relatively lower attendance rate when compared with other districts with similar expenditure per student at both primary and secondary levels. The district has the low- est percentage of children of primary school age enrolling in first grade (USAID 2018). Cox’s Bazar is among the districts with the lowest survival rate (about 65 percent), high- est dropout rate (about 35 percent), and lowest efficiency ratio, relative to districts with similar expenditures per student at the primary level (Genoni et al. 2019).15 Poor educational outcomes may be related to the type and size of educational insti- tutions in Cox’s Bazar. Data show a high prevalence of small, NGO-run educational insti- tutions, as well as high student-teacher ratios in formal schools in Ukhia and Teknaf in 2011, well before the Rohingya influx. This suggests that there was already a heavier reli- a ia r ia li a u af a az aria Ku ar es a li a u af hi da ha ha ku hi i ku m m ar bd bd kn kn d Uk ance on non-formal sources of education outside the public sector in these areas, which Uk Ra Ra Sa Sa Pe Pe hk hk ak ak Te Te tu tu es Ch Ch r ar Ku za oh oh may signal pre-existing stresses on the education sector (which are beyond the scope of Ba B M M x's 's this report to explain). Teknaf and Ukhia had 327 and 140 NGO-run schools respectively x Co Co in 2011, which accounted for at least half of all educational institutions in the two upa- zilas. In contrast, Kutubdia had 34 (roughly 20 percent of all schools), while Cox’s Bazar Primary Secondary College Madrasah NGO Other Source: BBS (2013). 14 See https://www.worldbank.org/en/programs/boost-portal. 15 The survival rate is the percentage of a cohort of pupils (or students) enrolled in the first grade of a given level or cycle of education in a given schools year expected to reach successive grades, regard- less of repetition. This rate is calculated following the UNESCO reconstruction cohort model. The coef- ficient of efficiency is an indicator of the internal efficiency of an educational system. It summarizes the consequences of repetition and dropout on the efficiency of the educational process in producing graduates. It is defined as the ideal (optimal) number of pupil years required (i.e., in the absence of 16 There is substantial within-district variation, with Cox’s Bazar Sadar NGO schools enrolling an aver- repetition and dropout) to produce a number of graduates from a given school cohort, expressed as a age of 762 students, suggesting they may be substantively different and not comparable. While it is percentage of the actual number of pupil years spent to produce the same number of graduates. The beyond the scope of this report, understanding the proliferation and role of NGO schools in the district coefficient of efficiency therefore ranges from a low of 0 to a high of 1. is an area for further research. 52 53 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Table 2-1: Average number of students per school for different school types, Map2-5: Map 2-5. Travel Travel times times to primary to primary Physical accessibility to schools does not Cox’s Bazar upazilas schools schools1717 appear to be a major determinant of poor 0 20 40km educational outcomes in Ukhia and Teknaf. Teknaf’s and Ukhia’s reduced levels of edu- Primary Secondary College Madrasah NGO Other Average, upazila Dhaka C H I T TAG O N G cational attainment come in spite of good Chakaria 385 559 248 177 100 338 Chittagong Pekua average levels of accessibility to primary and secondary schools, shown in Map 2-5 and Cox’s Bazar Kutubdia 307 987 1,012 497 762 783 477 Map 2-6. However, climate-related disrup- Sadar Chakaria tions and natural disasters in the area create Kutubdia 349 387 424 221 61 257 Matarbari BANDARBAN (approximate) accessibility problems for both students and Moheshkhali 473 631 375 422 30 80 447 teachers. Physical access to schools is often Maheshkhali constrained due to the low quality of roads Pekua 359 583 425 303 30 134 347 and traffic congestion (USAID 2018). This Cox’s Bazar Ramu 378 617 756 181 182 344 Cox’s Bazar Ramu suggests that the qualitative differences in Sadar schooling described above and/or the lack of COX’S Teknaf 583 476 274 246 39 175 BAZAR economic opportunity described in Chapter 3 are the main negative influences on human Ukhia 375 583 682 242 35 60 192 Ukhia capital formation in these upazilas. Total 388 668 574 269 45 524 297 Teknaf Geographic and infrastructural dispar- ities within Cox’s Bazar affect access to Table 2-2: Average student-teacher ratios, different school types, education in some unions. More than 80 Cox’s Bazar upazilas Primary school percent of the population of Cox’s Bazar St. Martin Minutes travel to primary school Dwip lives within 15 minutes of a primary Primary Secondary College Madrasah NGO Other Average, upazila With current transport infrastructure school, while more than 60 percent lives 0 5 10 20 in similar proximity to secondary and ter- Chakaria 81 47 21 35 17 60 tiary educational institutions. However, Note: Estimations based on an internal model of travel Cox's Bazar access to education is constrained in Cox’s 57 67 38 47 254 30 56 times (See Annex 2) and population distribution models Sadar from the High-Resolution Settlement Layer from Bazar’s northern and western unions by Facebook and the Center for International Earth Kutubdia 57 46 25 21 15 38 Science Information Network 2016. inadequate school facilities and underde- veloped roads. Within-union variation is Moheshkhali 87 71 18 37 10 5 56 high and primarily linked to proximity to the main north-south road, along which house- Pekua 69 40 20 31 30 22 51 holds and schools cluster. Far fewer schools serve the significant populations set among the surrounding paddy fields; large populations with access to few nearby primary Ramu 70 61 24 21 46 49 schools and secondary schools can be found in remote areas of Chiringa, Kuntakhali, Teknaf 132 55 11 40 39 65 and Saharbil unions in Chakaria; Kalarmaechhara and Hoanak unions in Maheshkhali; Ali Akbar Deil in southern Kutubdia; Harbang and Barno Bilchari in Chakria; and St. Martin Ukhia 70 40 30 35 35 6 51 Dwip in Teknaf (Map 2-5 and Map 2-6). Poor road infrastructure across all outlying areas Total 75 56 27 34 34 26 55 constrains access to the district’s centrally located universities, particularly for remote Kutubdia, Maheshkhali, and Teknaf (Map 2-7). Source: BBS (2013). 17 Travel times have been estimated using speed by type of road specified in Table A1-1, Annex 1. 54 55 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Map2-6: Map Travel 2-6. times Travel to secondary times to Map2-7: Map 2-7.Travel Traveltimes to tertiary times to tertiary Educational opportunities for displaced Rohingya children have improved since 2017 but schools secondary schools education education remain insufficient. Older children and youth are especially disadvantaged. At the onset 0 20 40km 0 20 40km of the 2017 influx, recently displaced Rohingya children living in camps had access only to non-formal education in learning centers operated by NGOs. By 2019, Pascaud and Panlilio C H I T TAG O N G C H I T TAG O N G (2019) reported significant improvements in education programming in Cox’s Bazar. For Dhaka Dhaka Chittagong Pekua Chittagong Pekua instance, attendance at learning centers had increased substantially. This was partly due to the construction of a larger number of local centers, alleviating previous concerns related Kutubdia Kutubdia to mobility and safety. Similarly, learning centers had improved their staffing to include Chakaria Chakaria more trained and dedicated teachers. Despite such gains, substantial challenges persist. Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) In 2020, the GoB authorized the use of the Myanmar curriculum for Rohingya children in camps, but its rollout has been severely hampered by COVID-19. Thus, the de facto curric- Maheshkhali Maheshkhali ulum in learning centers consists of only English, Burmese, math, and life skills classes.19 Cox’s Bazar Cox’s Bazar Without adequately structured curricula or grade progression, this model is not meeting Cox’s Bazar Cox’s Bazar the needs of young adolescents and youth, who are almost entirely left out of the system. Sadar Ramu Sadar Ramu COX’S COX’S BAZAR BAZAR School attendance rates among Rohingya children remain well below those among Ukhia Ukhia hosts at all levels of education. Due to discriminatory practices in Myanmar, nearly half the Rohingya children who arrived in Bangladesh had had no previous opportunity to engage in formal schooling (Guglielmi et al. 2020). Displaced individuals who arrived in Bangladesh Teknaf Teknaf before 2017 and lived in host communities accessed education in local private and govern- ment schools. But, in the absence of valid Bangladeshi documents (a prerequisite), they were unable to secure certification for their education, excluding them from future opportunities. Secondary school University St. Martin Dwip St. Martin Dwip Figure 2-6: School attendance rates Female-male attendance gaps have nar- Minutes travel to secondary school Minutes travel to tertiary school With current transport infrastructure With current transport infrastructure before and after the 2017 Rohingya rowed among children from the host influx, host children and Rohingyas community in primary and secondary 5 10 15 20 30 60 25 60 90 120 150 210 schools, compared to before the influx. 94% 90% Note: Estimations based on an internal model of travel Note: Estimations based on an internal model of travel Prior to the 2017 influx, in both primary and 82% times (See Annex 2) and population distribution models times (See Annex 2) and population distribution models secondary education, host girls were more from the High-Resolution Settlement Layer from from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Facebook and the Center for International Earth likely to attend school than host boys. The Science Information Network 2016. Science Information Network 2016. female-male attendance gaps in primary 59% and secondary schooling reached 9 and 12 50% Secondary net attendance rates among host children have increased slightly since the percent, respectively, although boys com- Rohingya influx, while tertiary attendance has fallen. These findings are based on com- pleted both educational levels at margin- parisons between HIES 2016 and the CBPS 2019 (Figure 2-6, standard errors are also shown ally higher rates than girls. Since the influx, 16,66% 18% in the figures below).18 relatively more boys in the Cox’s Bazar host community are attending primary and sec- 8% 2% ondary school (Figure 2-7). This is reflected 18 Two earlier reports based on focus group discussions and key informant interviews preceded the Primary Secondary Tertiary in the increase in overall attendance rates 2019 CBPS. They found that school attendance had decreased and dropout rates had increased among hosts, and that this was related to increased work opportunities related to the Rohingya influx (UNDP Host before Host after Rohingyas 2018 and USAID 2018). While the findings of the present report are qualitatively similar, this report 19 https://www.unicef.org /bangladesh/en/ relies on data from the 2019 CBPS baseline which is representative of Rohingya and host communities, Source: Authors’ calculations using HIES stories/expanding-education-rohingya-refu- and is more recent. 2016/17. gee-children-bangladesh 56 57 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE at both levels, particularly secondary. In contrast, in tertiary school, where boys were 19 Access to healthcare facilities varies across the district, with facilities clustered near percent more likely to attend than girls in 2016, this gap has fallen to 3 percent, stemming some Rohingya camps providing services for both displaced populations and hosts. from a decline in male tertiary school attendance since the influx. This is reflected, in turn, Cox’s Bazar currently has 43 healthcare facilities. At least 95 percent of the district pop- in lower overall net attendance rates in tertiary school, as measured by the CBPS baseline in ulation lives within 45 minutes of a healthcare facility (Figure 2-9), and the mean gap in 2019. In addition, the observed increase in secondary net attendance rates among hosts is travel times between low-skilled agricultural and high-skilled service workers (a proxy for driven by low-exposure areas, meaning upazilas other than Teknaf and Ukhia (Figure 2-8). better-off host households) is only seven minutes. As Map 2-8 and Figure 2-10 show, mean These patterns need further investigation. Within the Rohingya community, gender gaps in travel times to health services are lower for people in the southern parts of Cox’s Bazar school attendance are pronounced, particularly at the primary and secondary levels. than in the north. In Ukhia, a cluster of healthcare facilities around the camps serves hosts and displaced Rohingya. This cluster helps to provide better access for host communities Figure Female-male school 2-7: Female-male Figure 2-7. school Figure 2-8: Net Figure 2-8. Net school school attendance attendance in Teknaf and Ukhia. However, host communities may avoid such health services because attendance ratio gaps, attendance ratio gaps, before and before and rate after influx, hosts in high- rate after influx, hosts in high- and and of anti-Rohingya stigma (IOM and ACAP 2020), so effective healthcare access may be lower. 2017 Rohingya after 2017 after influx, host Rohingya influx, host low-exposure areas low-exposure areas community community and Rohingya and Rohingya Table 2-3: Malnutrition indicators for Bangladesh, Chittagong division, and Cox’s Bazar district, 2007-2019 93% 12% 94% 9% 9% 6%   Area 2007* 2011* 2012/13** 2014* 2017/18* 2019** 62% 51% National 43% 41% 42% 36% 31% 28% -3% -3% Stunting Chittagong 46% 41% 43% 38% 33% 27% -7% Cox’s Bazar 50% 35% 8% -19% 7% National 41% 36% 32% 33% 22% 23% -25% Underweight Chittagong 42% 37% 32% 36% 21% 23% Primary Secondary Tertiary Primary Secondary Tertiary Cox’s Bazar 41% 29% Host before Host after Rohingyas High exposure Low exposure National 17% 16% 10% 14% 8% 10% Source: World Bank staff calculations, CBPS 2019 and HIES 2016. Wasting Chittagong 18% 16% 9% 16% 8% 10% Cox’s Bazar 10% 10% Health * Estimates obtained from BDHS 2007, 2011, 2014, and 2018/18 rounds final reports. ** Staff estimates using MICS rounds 2012/13 and 2019. At the national level, Bangladesh has made impressive gains on critical health indica- Notes: Although BDHS 2017/2018 and MICS 2019 rounds were undertaken after the Rohingya influx in August 2017, these surveys do not include recently displaced Rohingya in their sample frames. tors that are especially important for long-term human capital accumulation, including All indicators are for children under five years old, following WHO (Child Growth Standards. Technical Report, Geneva: child nutrition. However, progress in Cox’s Bazar has been slower. Undernutrition is WHO, 2006. http://www.who.int/childgrowth/standards/Technical_report.pdf?ua=1). associated with nearly half of all child deaths worldwide, while many children who sur- Underweight is defined as children whose weight-for-age is more than two standard deviations below the median. vive early undernutrition suffer lifelong losses of cognitive capacity (Black et al. 2013). Data Stunting is defined as children whose height-for-age is more than two standard deviations below the median. Wasting is defined as children whose weight-for-height is more than two standard deviations below the median. show that Cox’s Bazar is lagging behind in reducing child stunting and the prevalence of underweight children, compared to Chittagong division and the national average (Table Congested living conditions, high population density, and limited sanitation within 2-3). Bangladesh and Chittagong almost halved the share of underweight children and the camps imply that the recently displaced Rohingya population remains highly vulnerable incidence of wasting and substantially reduced stunting between 2007 and 2019. Available to the spread of infectious diseases. A third of Rohingya households share a toilet with data for Cox’s Bazar suggest improvements in stunting and child underweight between more than 25 people, and two-thirds report sharing water facilities with more than 25 peo- 2012 and 2019, but prevalence remains higher than the national average, while progress ple (compared to 1 and 7 percent respectively among host households, CBPS 2019). on wasting has been stagnant. 58 59 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Map 2-8: Estimated travel times Infectious disease outbreaks have Figure Travel times 2-9: Travel Figure 2-9. to health times to care health care Figure Travel time 2-10: Travel Figure 2-10. to health time to care health care to health center facilities, emerged sporadically in and around some facilities, by population share facilities, by population share facilities, Teknaf and Ukhia versus facilities, Teknaf and Ukhia versus Cox’s Bazar district Rohingya camps but have so far been ade- upazilas other upazilas other 0 20 40km quately controlled. During 2017 and 2018, Bangladeshi health authorities detected 240 - 300 240 - 300 Dhaka C H I T TAG O N G outbreaks of several communicable dis- 180 - 240 180 - 240 Minutes to healthcare facilities Minutes to healthcare facilities Chittagong Pekua eases, including a diphtheria outbreak 120 - 180 120 - 180 Kutubdia among displaced Rohingya and the nearby 90 - 120 90 - 120 host community in Cox’s Bazar. These out- Chakaria 75 - 90 75 - 90 Matarbari BANDARBAN breaks induced substantial morbidity but (approximate) relatively few deaths. Prompt action from 60 - 75 60 - 75 Maheshkhali the Directorate General of Health Services 45 - 60 45 - 60 (DGHS) has so far limited the impacts of 15 - 30 15 - 30 Cox’s Bazar infectious outbreaks (Health Bulletin BBS 15 - 30 15 - 30 Cox’s Bazar Sadar Ramu 2018). However, evidence on the efficacy 0 - 15 0 - 15 of these campaigns is mixed; while the COX’S BAZAR number of measles and diphtheria cases 0% 10% 20% 30% 40% 0% 20% 40% 60% Ukhia detected in the camps had decreased District population share District population share through 2018, infections had not ceased Teknaf and Ukhia Other entirely. Studying the population across Teknaf Kutupalong Camp, Nayapara Camp, and Note: Estimations based on an internal model of travel Estimations based on an internal model of travel times times (See Annex 2) and population distribution models (See Annex 2) and population distribution models from makeshift settlements,20 Summers et al. from the High-Resolution Settlement Layer from the High-Resolution Settlement Layer from Facebook (2018) find high incidence rates of diph- Facebook and the Center for International Earth and the Center for International Earth Science Informa- Science Information Network 2016. tion Network 2016. theria despite vaccination efforts. With St. Martin Health facility (any type) Dwip the exception of unregistered Rohingya in Minutes travel to nearest health facility Kutupalong Camp, coverage with at least In the context of the COVID-19 pandemic, the Rohingya population and their hosts in With current transport infrastructure one dose of oral cholera vaccine was high. Cox’s Bazar remain at high risk. An outbreak in the camps would not only overwhelm exist- 0 15 30 45 60 84 An investigation into risk factors for acute ing health systems but also transmit rapidly due to the high population density, inadequate respiratory infections (ARI) among chil- water and soap supplies to maintain hygiene, limited ability to isolate infected individuals, Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models dren under 10 years in Rohingya camps and large household sizes (Truelove et al. 2020). A follow-up phone survey of a quarter from the High-Resolution Settlement Layer from found that about 21.6 percent of the 259 of CBPS households carried out between April 11 and April 17, 2020, showed that most Facebook and the Center for International Earth Science Information Network 2016. children studied showed acute ARI symp- respondents understand how COVID-19 is transmitted, yet three-fourths of households in toms (Oishi and Alam 2020). Immunity camps and half in host communities attend communal prayers, despite the risk of disease gaps persist despite several vaccination campaigns. This may reflect historically low vac- spread (Lopez-Pena et al. 2020). cination coverage rates among Rohingya, compounded by high birth rates that rapidly replenish the susceptible population. Humanitarian agencies and the Government of Bangladesh have joined forces to fight COVID-19 in Cox’s Bazar. Early in the pandemic, this included support for two Severe Acute Respiratory Infection Isolation and Treatment Centres (SARI ITC) and four quaran- 20 Three cross-sectional population-representative household surveys were conducted in 2018. These took place in Kutupalong (October 22–28), makeshift settlements (October 29–November 20), and tine centers. New ICU beds were added in Cox’s Bazar’s main district hospital, while WHO Nayapara (November 20–27). Sampling frames included all households in each area, regardless of coordinated with the GoB to expand testing capacity in the Field Laboratory at Cox’s Bazar whether they were registered with UNHCR. In Kutupalong and Nayapara, households were selected Medical College.21 Humanitarian agencies worked with the government to inform host using simple random sampling. Because of the large population residing in the makeshift settle- ments, households in these sites were selected using multistage cluster sampling; the Inter Sector Coordination Group provided block populations. 21 UNHCR Bangladesh. “COVID 19 Preparation/Response- 31 May 2020 “(#4). 60 61 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE and Rohingya communities about COVID-19 prevention through neighborhood-based While systematic data on the prevalence of child marriage among the Rohingya is approaches and mass-media outreach. At least 5,641 hand-washing stations were installed lacking, some studies have found that early marriage is likely to be more prevalent in public places within camps.22 than among the host community (Ainul et al. 2018). While the 2019 CBPS did not ask directly about age at first marriage, a significantly larger share (35 percent) of married Reproductive health, gender-based violence, mental health, and trauma Rohingya women in the sample are younger than 25, compared to married women in the host community sample, of whom 24 percent were under 25.23 A 2016 UNHCR report Early marriage and childbearing can severely impact the reproductive health and men- estimates that more than half of the Rohingya girls who have fled Myanmar since 2012 tal well-being of adolescent girls (Gordon, Jay, and Lee-Koo 2018). Across Bangladesh, were married before they were 18. Conservative religious and social norms about the use the median age at first marriage of women aged 20-49 was 16.3 years in 2017-18. In other of contraception and early marriage contribute to a lack of information about sexual and words, half of Bangladeshi women currently aged 20-49 were married before the age of 16. reproductive health, relatively high rates of fertility within camps, and maternal morbid- Fifty-nine percent of women aged 20-24 marry before the legal age of marriage, 18 years ity and mortality (Hasan-ul-Bari and Ahmed 2018). (Ministry of Health and Sports 2017). The median age at first marriage in Cox’s Bazar dis- trict and Chittagong division is only 17 (UNICEF 2019) (Figure 2-11). In Chittagong division, Available evidence points to a severe burden of mental health conditions among more than a quarter of girls aged 15–19 years are already married, and adolescent girls in displaced Rohingya, including children and youth. Qualitative studies, in-depth inter- Chittagong are more likely to marry a significantly older man than girls of their cohort in views, and focus group discussions have identified numerous factors adversely affecting the rest of the country (BBS and UNICEF 2015). the mental health of the displaced Rohingya. Contributing factors include the chronic stresses of poor living conditions, reliance on assistance, domestic violence, lack of physical safety, and the emotional scars of displacement. Corna et al. (2019) document Figure 2-11: Women’s Figure 2-11. Women’s age age at at first first marriage: Bangladesh, Chittagong marriage: Bangladesh, Chittagong division, division, and Cox’s Cox’s Bazar, mistrust, depression, PTSD symptoms, and sleeping problems among the markers of and Bazar, 2019 2019 distress. Many Rohingya women and girls survived sexual violence in Myanmar before fleeing to Bangladesh, resulting in persistent physical and mental trauma. Some months after arriving in Bangladesh, almost half of Rohingya children reported experiencing .1 indicators of distress and sleeplessness (IOM 2018). While concerted efforts to address mental health and trauma have been a part of the humanitarian effort, the 2020 Joint Response Plan for the Rohingya Crisis (UNHCR 2020) notes the urgent need for scaled-up psychosocial support for children under the age of 18. Density .05 More recent data collected as part of the CBPS effort confirm the prevalence of trauma and experiences of violence, particularly among the displaced Rohingya. As part of the CBPS, GAGE (Gender and Adolescence: Global Evidence) implemented mixed-methods data collection focusing on adolescents and their caregivers, with quantitative survey data complemented by qualitative research in three camps and two host community areas. 0 0 10 20 30 40 50 Fourteen percent of adolescents in the sample reported experiencing psychological dis- Age tress, with older adolescents twice as likely and adolescent boys more likely than girls to National Chittagong Cox’s Bazar experience distress (Guglielmi et al. 2020). Analysis of the trauma and mental health and the crime and conflict modules of the CBPS reveals that 1 out of 2 Rohingya reported hav- Source: World Bank staff calculations using Bangladesh MICS 2019. Note: Vertical dashed line at 15. ing been close to death, and 44 percent reported having experienced torture or combat sit- Notes: Although MICS 2019 was undertaken after the Rohingya influx in August 2017, these surveys do not include uations. While only 6 percent of Rohingya reported having personally experienced rape or recently displaced Rohingya in their sample frames. sexual abuse, the large majority had either witnessed it or heard about others’ experience. 23 Data from the 2015-16 Myanmar DHS indicate that the median age at first marriage for women aged 22 Inter Sector Coordination Group (ISCG). “COVID-19: Preparedness and response for the Rohingya 20-29 was 22.1 years, with lower ages at first marriage and first birth for women in Rakhine state. The refugee camps and host communities in Cox’s Bazar District Weekly Update #12.” ISCG, Cox’s Bazar, median age for women’s marriage in Rakhine State in 2016 was 20.7 years (Ministry of Health and 31 May 2020. Sports 2017; Ripoll 2017). 62 63 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE This difference between reporting about others’ experience versus one’s own is not uncom- limited variety of foods due to lack of resources. However, reports of deprivation decrease mon, given the sensitivity of this issue, as well as the social norms and stigma, particularly as the scale progressively moves towards indicators for moderate to severe hunger. Only 6 for women, associated with such a traumatic experience.24 percent of households report that members went at least an entire day and night without any food within the four-week recall period. Low-exposure households self-report higher Hosts have a lower incidence of traumatic events. However, hosts in areas closer to rates of deprivation compared to high-exposure households across indicators for moderate camps are more likely to have witnessed or heard about events related to imprisonment, to severe hunger but report lower rates of dissatisfaction in terms of their dietary diversity. combat, murder of strangers, or torture. This may refer to the experience of the Rohingya or their own. Among hosts, the most common symptoms of psychological stress and dis- The economic fallout of the COVID-19 pandemic has adversely affected access to food tress include feelings of powerlessness and lack of a future. Worryingly, 1 in 2 Bangladeshi among the host population. The first round of the CBPS high-frequency follow-up,27 imple- women and girls live in neighborhoods where harassment and physical and gender-based mented during the first half of 2020, showed that urban, low-exposure areas were more violence are issues.25 adversely affected than the more rural, high-exposure areas. Indeed, 50 percent of host households in low-exposure areas reported they were not able to purchase basic needs in the seven days prior to the survey, as opposed to 34 percent in areas closer to the camps. Food security and living conditions These impacts on consumption are correlated with larger labor-market shocks faced by low-exposure households (see also section 3 of this report). The higher incidence of diffi- Consumption patterns for hosts indicate broad access to basic foods, with no large dif- culty in basic food access among relatively more urbanized communities in Cox’s Bazar has ferences between low- and high-exposure areas. Overall, consumption patterns for hosts apparently been driven by (i) greater labor-market disruptions and losses in purchasing and displaced Rohingya indicate broad access to a range of basic food groups, and higher power; and (ii) limited scope for self-production of basic foods, coupled with high reliance than minimum required levels of caloric intake per capita per day.26 On average, host on market purchase of food. households living in high- and low-exposure areas consumed 10 of 12 basic food groups in the week preceding the interview, with the majority of the items being purchased. The The Government of Bangladesh is providing assistance to host communities to mitigate average caloric intake per person per day was 2,240 calories. Seventy-eight percent of the the impact of COVID-19. Fifteen percent of surveyed hosts reported receiving some form caloric intake for hosts came from market purchases, with the remainder coming from of assistance after March 1, 2020. Three-fourths of this assistance was newly received, that self-production or transfers/gifts. High- and low-exposure hosts report similar patterns of is, not part of previously running programs. In line with CBPS baseline findings, more than consumption, with the former reporting marginally better access to food groups and per 90 percent of this assistance came from the government. High-exposure hosts were mar- capita caloric intake. This pattern of more urbanized areas reporting lower food consump- ginally more likely to receive assistance from NGOs than low-exposure hosts. One source of tion relative to more rural regions is consistent with the consumption patterns reported NGO assistance is the World Food Programme (WFP), which started a district-wide support nationally in the HIES 2016 survey. program for hosts who are vulnerable due to COVID-19. This support includes in-kind food transfers and cash transfers. Overall, 75 percent of the assistance received was through dis- Measured food insecurity among the host population is relatively low, although many tribution of food and other basic needs, 22 percent through work or jobs programs where households report being dissatisfied with dietary diversity. Data from the 2019 CBPS can in-kind basic needs assistance was also provided, and 3 percent through cash transfers. be used to measure food security, classified into three stages based on severity of depri- vation: (i) Inadequacy and dissatisfaction: lack of dietary diversity; (ii) Moderate hunger: The welfare of the recently displaced Rohingya population remains primarily reliant on having to consume smaller or fewer meals than usual; (iii) Severe hunger: having no food humanitarian assistance, and the latter has been largely successful in ensuring access at home, going to sleep hungry, or going days without food. More than 2 out of 3 hosts to food and basic needs. Analysis of the food consumption module of the CBPS 2019 finds report either not being able to consume their preferred foods or having to consume a widespread and adequate access to food for Rohingya households living in camps, financed primarily through humanitarian assistance. Food consumption covers a wide range of food 24 See Table A1-2 and Table A1-3. types, but consists primarily of cereals, vegetables, fish, spices, and sweets, with low intake 25 See Table A1-4. of dairy products, meat, and eggs. On average, 85 percent of the Rohingya households in 26 For planning purposes, the World Health Organization (WHO) and the U.S. Committee on International camps consumed 8 or more food groups in a week, out of the 12 food groups considered. Nutrition recommend that an average of 2,100 kcal per person/per day be used as an initial reference The lowest range of food groups consumed was 5 or 6 groups, and this was reported by figure. Since implementation of revised Memoranda of Understanding (MoUs) among UNHCR, WFP, and UNICEF (UNHCR/WFP, July 2002; WFP/UNICEF, February 1998), the three agencies have adopted 2,100 kcal as their initial planning figure for calculating energy requirements and designing food rations. 27 The CBPS high-frequency follow-ups Round 1 was conducted between April and May 2020. 64 65 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE only 3 percent of households. Eighty-four percent of the calories consumed per capita by open market to fill nutritional gaps and meet other needs (WFP 2019a). However, in the displaced Rohingya were obtained from humanitarian aid. This was especially reflected in current COVID-constrained operating environment, while WFP has continued the deliv- the consumption of calories from oils and fats, legumes, and cereals (96 percent, 96 per- ery of food assistance, temporary adjustments have been necessary. Rohingya receiving cent, and 95 percent derived from aid, respectively). e-voucher assistance now receive a fixed basket of products to reduce crowding in food distribution centers.30 At the same time, WFP in coordination with the GoB has started a On average, households in camps consumed 2,352 calories per capita per day. About 60 one-off scheme for host communities. This scheme will benefit 105,000 households and percent of the displaced Rohingya households consumed more than the standard 2,100 consist of in-kind and cash transfers.31 calories per capita provided within a WFP food basket.28 The composition of the calories consumed is also similar to surrounding hosts, with 90 percent of calories coming from Access to shelter, sanitation, water, and electricity affects welfare and human capital the following sources: cereals (65 percent), oils and fats (12 percent), eggs (8 percent), and accumulation among Cox’s Bazar displaced Rohingya and hosts. Limited access to water legumes, nuts, and seeds (4 percent). Similar to host populations, the share of daily calo- and sanitation reflects low living standards for both hosts and displaced Rohingya. ries coming from fish, meat, and vegetables, along with other remaining food groups, is While no open defecation is reported in the CBPS, almost half of host households are using low, at 1-2 percent on average. Despite the low caloric share, households report consuming unimproved sanitation, with 40 percent having access to basic (improved and not shared) seven different types of vegetables on average in the past week, indicating access to a wide sanitation facilities, and 10 percent to limited sanitation facilities.32 According to World range of produce. Bank (2019c), only 2 percent of Rohingya have access to basic sanitation (improved and not shared), and 23 percent use unimproved sanitation facilities. There is a high reliance While food support provides the bulk of Rohingya families’ essential nutrition, house- on shared facilities in Rohingya camps. Only 4 percent of camp households have access to holds also report purchasing food. Almost all Rohingya households (99 percent) report private latrines, and around one-third of households are sharing these facilities with more purchasing at least some food items during the past week, although such purchases rep- than 25 households. resent only about 12 percent of total calories for the average household. The resources for these purchases could, in principle, come from stipends from cash for work programs, For both host and Rohingya populations, access to water through improved sources from bartering or selling items received as humanitarian aid, or from other resources.29 is widespread, but that access is shared to a large degree. Differences emerge between Nevertheless, there is a clear distinction in the kinds of foods that are largely obtained from host households living in high- and low-spillover areas. Households living closer to humanitarian assistance as opposed to those that are purchased: cereals, oils, legumes, Rohingya camps report almost 6 percentage points lower access to private sources of sweets, and eggs largely come from aid, whereas the remaining food types are largely drinking water than households living farther away. In addition, host households living reported as being purchased. close to camps are more likely than households farther away to share their water source with many other families (12 percent versus 4 percent sharing access with more than 25 Providing food assistance via electronic vouchers shows promise to further improve other households). Reliance on shared sources of drinking water is far more prevalent in nutrition among the Rohingya. While at the beginning of the influx most food aid was camps: 3 out of 4 displaced households share their drinking water source with more than based on an in-kind system, in the months prior to the COVID-19 pandemic, 95 percent 25 households. of food aid was transferred to e-vouchers. Filipski et al. (2020) study the effect of receiv- ing electronic vouchers versus food rations on the nutritional status of Rohingya chil- dren. The study finds that e-vouchers led to positive growth outcomes among children between the ages of 6 and 24 months, and that one of the main reasons behind bet- ter nutritional outcomes among e-voucher recipients was the flexibility that vouchers 30 WFP. “WFP in Cox’s Bazar - Perception Analysis - COVID-19 Response.” Brief, April 2020. https://www. wfp.org/publications/wfp-coxs-bazar-perception-analysis-covid-19-response allowed households in purchasing items beyond those provided by humanitarian organi- 31 WFP. “COVID-19 - Support to the Host Community - Cox’s Bazar.” Brief, November 2020. https://www. zations. While nearly two-thirds of the displaced Rohingya people had access to e-vouch- wfp.org/publications/wfp-bangladesh-covid-19-support-host-community-coxs-bazar ers at the time of the study, the rest were obliged to sell or barter their entitlements in the 32 Definitions for sanitation indicators follow standards set by the WHO-UNICEF Joint Monitoring Programme for Water Supply, Sanitation and Hygiene (JMP) 2017 update (UNICEF 2017). Improved sanitation refers only to the type of facility used – in the survey, “sanitary” and “pacca” latrines are 28 For 75 percent of households consuming above the 2,100 calories per capita threshold, the range of classified as improved facilities. Basic sanitation services are defined as use of improved sanitation calories per capita was between 2,100 and 3,000. facilities which are not shared with other households. Limited sanitation services are defined as use of 29 Forthcoming briefs will explore how food purchases are funded (e.g., by selling aid, through cash- improved sanitation facilities which are shared with other households. The survey does not collect the for-work, or from other sources). data required to classify facilities as safely managed. 66 67 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Access to electricity is widespread for hosts in Cox’s Bazar, although the availability of Kutubdia, and the proposed Martarbari port location are poorly connected to the Chittagong power varies substantially. Most Rohingya households do not have access to electricity. road. Martarbari and Maheshkhali are connected to Cox’s Bazar Sadar across the bay only On average, 87 percent of host households have access to electricity. While low-spillover by small, slow private launches unsuited for major commercial activities. South of the city, areas receive around 12 hours of electricity per day on average, high-spillover areas receive traffic is split along two small roads on either side of steep hills hugging the coastline. The only 6. Among Rohingya households, only 40 percent have access to electricity, relying coastal road, which passes through fewer villages, is primarily used for transporting aid to completely on solar-powered sources. the Rohingya camps (Map 2-9). However, as these roads were built for traffic prior to the influx, they are now both heavily congested and rapidly deteriorating under the weight of According to CBPS 2019, 8 out of 10 host community households were living in owned trucks ferrying goods to the camps. The geography effectively prevents the development dwellings, but the poor quality of construction materials reflects the area’s low living of additional routes, such that improvements will have to come from upgrades to existing standards and poses risks given high exposure to climate-related disasters. Host houses routes (Map 2-10). The Asian Development Bank is currently overseeing upgrades to the have, on average 2.5 rooms in both high- and low-exposure areas. However, differences coastal road to improve the delivery of aid to the camps. in housing quality between low- and high-spillover areas are pronounced. While 24 per- Map 2-9: Map 2-9.Cox’s Cox’sBazar Bazarroad transport road Map 2-10: Map 2-10.Cox’s Cox’sBazar population, Bazar road cent of host dwellings in high-exposure areas have walls of brick/cement, in low-exposure network network transport camps, and transport road transport network network areas this share rises to 37 percent. The share of households with brick/cement roofs in low-spillover areas (13 percent) is almost double that in high-spillover areas (7 percent) 0 20 40km 0 20 40km (World Bank 2019b). The Rohingya in Cox’s Bazar are also generally living in low-quality dwellings. Around 95 percent of Rohingya housing is built of bamboo, straw, polythene, or C H I T TAG O N G C H I T TAG O N G Dhaka Dhaka canvas materials. Rohingya dwellings are not only of lower quality than local houses but Chittagong Pekua Chittagong Pekua are also smaller. On average, Rohingya shelters have fewer than two rooms. In contrast Kutubdia Kutubdia to many displaced populations elsewhere, however, recently displaced Rohingya in Cox’s Chakaria Chakaria Bazar have largely remained in the camps where they were initially settled. Ninety-eight Matarbari BANDARBAN Matarbari BANDARBAN percent of the displaced Rohingya have not moved their residence outside the camps since (approximate) (approximate) they arrived (Genoni et al forthcoming). Maheshkhali Maheshkhali Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Geography Sadar Ramu Sadar Ramu COX’S COX’S BAZAR BAZAR Cox’s Bazar district is situated in the Chittagong Division of southeastern Bangladesh. It Ukhia Ukhia lies south of Chittagong district and west of Bandarban district. Cox’s Bazar is bounded by the Bay of Bengal on the south and west, Myanmar and the Naf river on the east. Comprising 2,491.9 square kilometers, Cox’s Bazar represents about 1.7 per cent of the total land area Teknaf Teknaf of Bangladesh, making it one of the country’s smallest districts. The land area of Cox’s Bazar district is part hilly and part coastal plain and islands, as the district straddles two Road classifications Other Road classifications Other agroecological zones, the Northern and Eastern Hills and the Chittagong Coastal Plain. The Primary Ferry Primary Ferry Secondary Secondary district is also characterized by one of the longest unbroken natural beaches in the world Tertiary St. Martin Dwip Tertiary St. Martin Dwip Minor and is a major domestic tourism destination. Population Density (2018) thousands per square kilometer (including Rohingya) Cox’s Bazar’s geography constrains transportation, accessibility, and development 138 150 250 350 520 possibilities. From the north, a single two-lane primary road connects the district to Rohingya camps Chittagong, the regional economic hub, and onwards to Dhaka. East-west connections are weak throughout the district, such that the eastern unions of Rama upazila, the populous Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information western upazila of Maheshkhali, western areas of Charkaria upazila, the island upazila of Network 2016. 68 69 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE The recent Rohingya influx has been concentrated in areas with relatively scant land avail- able for cultivation and other economic uses (Box 3-Figure B3-1). More than 60 per cent the path of monsoon rain33 and tropical cyclones.34 Every year the government of the land area in Cox’s Bazar district is either forest or unavailable for cultivation, com- of Bangladesh, together with international organizations, provides extra shel- pared to 40 per cent nationally (BBS 2017b). Compared with the district’s biggest upazila, ter and performs earthworks to help displaced and host populations prepare Chakaria, which also has the largest share of cultivable land, Teknaf and Ukhia, the two for monsoon season.35 The region is also susceptible to earthquake, wildfire, upazilas with the highest concentration of recently displaced Rohingya, have a relatively extreme heat, and tsunamis.36 smaller land area and a greater share of reserved forest (Teknaf 41 per cent, Ukhia 59 per cent) (BBS 2013). Thus, the recent influx has not only increased population density but also Figure B3-2. B3-2: Risk of cyclones and storms and average annual rainfall the need for fuelwood and shelter. This has compromised livelihoods through deforesta- in Cox’s Bazar tion and reduced access to land (Tallis et al. 2019). A. B. Box 3: Land availability and environmental risks in Cox’s Bazar Figure B3-1: Uses of land, districts in Chittagong division and Cyclone intensity Cyclone frequency (TS-H3) (number) nationally TS 2-3 Figure B3-1: Uses of land, districts in Chittagong division and nationally TD 3-4 H1 4-5 H2 5-6 Bandarban H3 6-8 Chittagong District C. D. Bangaldesh Cox's Bazar Chittagong Division 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Forest area Not available for cultivation Cultivable wasted area Current fallow area Single cropped Double cropped Storm surge Average annual rainfall Triple cropped Quadruple cropped (10 meter) (mm) Affected 2400 – 2800 Unaffected 2801 – 3000 Source: BBS (2018). 3001 – 3500 Source: Alam, Sammonds, and Ahmed (2019). The Rohingya displaced population has settled in an area with extremely high environmental risks. Bangladesh is among the seven countries in the world 33 Monsoon periods start in late May and gradually diminish between October and with the highest long-term climate risk indices (Germanwatch 2020). Flat and November. 34 For instance, previous to the largest influx, in May 2017, around 70 percent of shelters in low-lying floodplains make the country vulnerable to water-related natural camps were damaged by Cyclone Mora (ISCG 2017). risks, such as floods and storm surge, particularly in coastal areas. Southern 35 In 2018, around 150,000 Rohingya were estimated to live in areas that were at high risk districts with long coastlines on the Bay of Bengal, including Cox’s Bazar, are on of floods and landslides (UNHCR 2018). 36 See http://thinkhazard.org for a hazard profile of Cox’s Bazar 70 71 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Connective infrastructure and accessibility The incoming Rohingya population has exacerbated pre-existing environmental risks in Cox’s Bazar. According to the Ministry of Forests and UNDP (2018), the Especially where geography poses challenges, competitive modern economies invest most important and visible impacts of the influx are forest degradation, habitat in infrastructure to reinforce connectivity and accessibility. Connective infrastructure loss, fragmentation of territory for wildlife, soil erosion, ground water degrada- (roads, rail, waterways, and ferries, as well as digital infrastructure38) and accessibil- tion, and water scarcity, all of which increase climate vulnerability in the region ity (proxied in this report as travel times, which are determined by topography and the (Tallis et al. 2019) (Figure B3-3). Indeed, since the influx, 2,283 hectares of forest presence and quality of connective infrastructure) lower transport costs, increase mar- have been deforested, reducing forest coverage in the areas around Kutapalong ket access, decrease interregional price gaps, and enhance economic growth. They also camps by 18 percent (Hassan et al. 2018). Strategies and resources are needed to improve supply chain efficiency and increase population access to social services such as manage the increasing stress on the natural environment, as well as the conse- health and education (Donaldson 2018; World Bank 2019c). More generally, expanding the quences for the wellbeing of host and displaced populations. The Government coverage of and access to transport and digital infrastructure has the potential to expand of Bangladesh is working with international organizations and NGOs to promote access to markets, enhance capacity to manage risks, and boost productivity and income reforestation37 around the camps, helping to mitigate these risks. generation capacity. Figure B3-3. Deforestation in Kutupalong camps, May 2017 (top) Better connectivity and accessibility in Cox’s Bazar are key for regional economic growth Figure B3-3: Deforestation in Kutupalong camps, May 2017 (top) versus May 2020 (bottom) and can accelerate national growth. Rising wage demands from workers, tougher global versus May 2020 (bottom) price competition, and inefficient logistics have raised pressure on local producer costs A. in Bangladesh. This poses a growing challenge for the country’s economic model, which has relied on a competitive advantage in wages (Farole and Cho 2017; Herrera Dappe et al. 2020). Bangladesh needs to invest in its transport infrastructure, which has not kept pace with its growth – the country ranks 100th out of 141 countries in the World Economic Forum’s Global Competitiveness Index in transport, and even further behind in the rank- ing for road connectivity. Beyond the lack of adequate transport infrastructure, operation and maintenance of existing assets have been inadequate, even more so considering the repeated exposure to flooding and natural hazards. Developing a more inclusive growth model beyond the Dhaka and Chittagong corridor will require better transport and logistics systems to connect people to jobs and allow businesses to invest in areas of high return. B. Integration with the global economy will be facilitated by the development of a multi-modal and interconnected transport network that effectively and efficiently links Bangladesh’s seaports with more of the country’s interior (IFC 2020). In this context, improving connec- tivity infrastructure and access is crucial for future growth, and progress in Cox’s Bazar can have national impact. To date, high population density, poor road quality, and lack of alternatives to road transport have kept Cox’s Bazar relatively isolated from Bangladesh’s main economic centers. Map 2-11 shows estimated travel times from different areas of Cox’s Bazar to Bangladesh’s main economic gateway, Chittagong city. Connectivity varies within the Source: Google Earth (2020). 38 Households and firms in Cox’s Bazar are also disadvantaged by poor access to digital infrastructure. 37 See http://www.fao.org/bangladesh/news/detail-events/en/c/1200069/ As discussed in chapter 4, access and quality issues, compounded with the relatively high cost of inter- net connections, are a nationwide constraint, but are particularly salient to businesses in Cox’s Bazar, for which technology is the second most important constraint, after access to credit. 72 73 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE district and is affected by factors other than geographic distance. For example, the high- The market accessibility index in Table 2-4 reflects this inequality of connectivity within way connecting Dhaka with Bangladesh’s southeastern districts varies in quality along the district and underscores the relative isolation of Cox’s Bazar.40 This index measures its length.39 This increases travel times and reduces connectivity in areas of lower road each upazila’s cumulative access to every major market (defined as cities of 50,000+ pop- quality (Map 2-12). ulation) in Bangladesh, given current transportation infrastructure. However, the lack of connectivity primarily seems to affect access to national rather than local markets (Table 2-4). As Map 2-12 shows, with the exception of Maheshkhali, which seems to have major Map 2-11: Estimated travel times to Map 2-12: Accessibility to growth connectivity problems, access to growth centers or key multi-modal markets41 within each Chittagong Map city 2-11. Estimated travel times to centers in Accessibility Map 2-12. Cox’s Bazar to growth upazila is relatively good. Moreover, using the distribution of education as a proxy for Chittagong city centers in Cox’s Bazar inequality, we observe that not only growth centers, but all markets, are equally accessible 0 20 40km 0 20 40km for individuals of different levels of education (Figure 2-12 and Figure 2-13). C H I T TAG O N G C H I T TAG O N G Table 2-4: Market accessibility index - Ranking of Cox’s Bazar upazilas, 2010 Dhaka Dhaka Chittagong Pekua Chittagong Pekua Ranking in Cox’s Bazar Ranking in Chittagong Ranking in Bangladesh Kutubdia Kutubdia Chakaria Chakaria Kutubdia 7 92 481 Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) Maheshkhali 6 91 470 Maheshkhali Maheshkhali Teknaf 5 90 461 Ramu 4 86 441 Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Sadar Ramu Sadar Ramu Ukhia 3 85 430 COX’S COX’S Cox’s Bazar Sadar 2 75 393 BAZAR BAZAR Ukhia Ukhia Chakaria 1 73 383 Note: Ranking in Chittagong division ranges from 1 to 94, with 94 being the lowest rank. Ranking in Bangladesh ranges from 1 to 493, with 493 as lowest. Source: Blankespoor and Yoshida (2010). Teknaf Teknaf The influx of Rohingya has increased congestion and underlined the urgent need for better road infrastructure. During the first wave of displacement, Rohingya used roads, Growth centers bridges, and dams as shelter, causing damage to transport infrastructure (UNDP and UN St. Martin St. Martin Minutes travel to Chittagong Dwip Minutes travel to growth center Dwip Women 2017). Subsequently, the growing international relief workforce and the large- With current transport infrastructure With current transport infrastructure scale delivery of humanitarian supplies have further stressed the district’s congested road 90 120 180 240 447 0 10 20 30 45 90 transport infrastructure. The Roads and Highways Department estimates that road traffic in affected areas has more than doubled, damaging the main road between Cox’s Bazar Note: Estimations based on an internal model of travel Note: Estimations based on an internal model of travel and Teknaf in particular (UNDP 2018). times (See Annex 2) and population distribution models times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Facebook and the Center for International Earth 40 See Blankespoor and Yoshida (2010) for more details. This index was calculated with the negative Science Information Network 2016. Science Information Network 2016. exponential potential model for 202 cities with population from the 2001 census, with populations ranging from approximately 15,000 to 6,500,000 (Dhaka). 41 The more important markets in Bangladesh are characterized as having permanent multimodal structures including shops, banks, and storage facilities, among others, as well as managing a large volume of trade. Since the early 1990s, the Planning Commission of Bangladesh has identified these See Table A1-1 in the Annex for road and ferry speeds used for modeling, adapted from Blankenspoor 39 important markets as Growth Centers. These centers are intended to be the focal points of rural devel- and Yoshida (2010). opment (GoB 2005). 74 75 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Along with the Rohingya influx, climate risk affects connectivity in Cox’s Bazar. Transport The lack of multimodal transport nodes, the dominance of road transport, and the infrastructure is destroyed every year because of exposure to monsoon cyclones and floods. lack of capacity to handle high traffic volumes constrain both economic development Twenty-five percent of the total length of roads and 1.5 km of bridges and culverts need to and humanitarian action in Cox’s Bazar. On the one hand, the district faces rising trans- be rehabilitated to be accessible to traffic throughout the year (GFDRR 2018). To guarantee port costs and reduced competitiveness, with greater effects on relatively isolated areas continued delivery of vital humanitarian assistance to the Rohingya camps, every year42 (Herrera Dappe et al. 2020). On the other hand, the pressure on local transport infrastruc- international organizations and the Government of Bangladesh perform risk-reduction ture has weakened Cox’s Bazar’s logistic capacity, complicating effective humanitarian activities including re-paving of main roads, improvement of drainage systems, and earth- assistance (UNHCR 2019). work construction. Investment in transport infrastructure in Cox’s Bazar needs to increase. The factors just Figure 2-12: Average travel time Figure 2-13: Average travel time to discussed have boosted the demand for investments in infrastructure in Cox’s Bazar. Such Figure 2-12. Average travel time to Figure 2-13. Average travel time to to markets of any size by level of growth centers by level of education investment would also contribute to Bangladesh’s goal of increasing economic density in markets of any size by level of growth centers by level of education education and upazilas secondary cities, a priority for local development and national economic growth. education and upazilas Thanks to its natural comparative advantages, Cox’s Bazar can host infrastructure proj- Ukhia University ects of national and international importance. Bangladesh’s privileged geographic loca- tion between South and Southeast Asia creates a unique opportunity to benefit from the Teknaf Higher international movement of goods, services, and investment flows (Plummer, Morgan, and secondary Wignaraja 2016). Ambitious infrastructure projects have been planned in anticipation of Ramu the incoming demand from local international trade, as well as increasing demand related Secondary to the regional connectivity agenda (JICA 2016). Cox’s Bazar is positioned to play a key role Pekua in these advances. Maheshkhali Lower secondary The proposed construction of Bangladesh’s first deep seaport at Matarbari holds Kutubdia promise.43 Increased international trade and the concentration of 98 percent of cargo in Chittagong port has exceeded the port’s capacity.44 This directly impacts Bangladesh’s eco- Primary Cox’s Bazar nomic growth prospects. Accordingly, the country’s Seventh Five-Year Plan 2016-2020 has Sadar included the construction of port terminals financed by JICA. Along with the ship terminals, Chakaria No education the project includes the upgrading and construction of new local roads to improve connec- tivity to the port area (Map 2-13).45 0 5 10 15 0 5 10 15 20 25 Minutes travel to all markets (of any size) Minutes travel to growth centers No education Secondary Primary Higher secondary Lower secondary University 43 Originally, the plans for the energy hub at Maheshkhali Upazil called for development of up to 6 gigawatts (GW) of coal power plant capacity (in addition to some 3 GW of liquid natural gas-based generation), with the associated climate/environmental and financing challenges. The Government Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from of Bangladesh has recently announced that it will review the plan for coal power additions as part the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Network 2016. of a forthcoming Power/Energy Sector Master Plan, starting in early 2021. As such, some of the pre- viously planned coal-based generation in Matarbari may be postponed or cancelled. A project to invest in a Bay Terminal development at Chittagong port is also planned. This will help to reduce congestion not only in Chittagong port but also in Matarbari port in the future. 42 In 2020, disaster risk reduction efforts were suspended due to the COVID-19 lockdown. https://www. 44 https://www.joc.com/port-news/asian-ports/port-chittagong unhcr.org/news/briefing/2020/4/5e9ea77e4/covid-19-unhcr-warns-severe-implications-annual-mon- 45 The project also includes the development of special economic zones, logistic parks, and power soon-response-bangladesh.html plants. See JICA (2018). 76 77 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C HA P TER 2 – FUN D AMENTA L S : P EO P L E , L AN D , AN D I NFRASTRU C TURE Map 2-13. Accessibility to proposed Cox’s Bazar is a key link in other regional This plan will be financed by the Asian Development Bank, will employ workers from the Matarbari port and energy complex connectivity projects fostering interna- camps, and will include improvements to both roads and water supply while also increas- tional and domestic trade and growth. ing environmental resilience, strengthening disaster risk management, and improving Under international agreements in which energy infrastructure (RHD 2019).49 Dhaka 0 20 40km Chittagong Bangladesh participates, the country is part of five regional connectivity corridors.46 In Better rail connections may help relieve pressure on the district’s road system. Prospects C H I T TAG O N G 2009, Bangladesh and the UN Economic and of increasing cargo traffic, the development of Cox’s Bazar as a tourist destination, and the Pekua Social Commission for Asia and the Pacific influx of Rohingya have all contributed to rising stresses on roads. Considering this, the Kutubdia (ESCAP) signed an agreement on connect- Government of Bangladesh has started a project to connect Cox’s Bazar to the national ing the Asian Highway, which includes a key and sub-regional railway network (ADB 2016). The project, co-financed by the Asian Chakaria road going along the AH41 axis Chittagong- Development Bank, aims to extend the railway corridor from Chittagong to Cox’s Bazar. The Matarbari BANDARBAN (approximate) Cox’s Bazar-Teknaf (Plummer et al. 2016). project is part of the Asia Railway network and is expected to improve the district’s access The Asian Highway is especially relevant for to regional markets and trade. The plan also foresees extending rail lines to the Myanmar Maheshkhali Cox’s Bazar because it connects the district border and the Matarbari Port area.50 with other regional corridors terminating Cox’s Bazar in Chittagong and because it is the most Many of these capital-intensive and export-oriented investments are critically important Cox’s Bazar Sadar Ramu important arterial road for the Matarbari for Bangladesh’s economic growth, most directly through linkages with Chittagong and Development Project. Additional ongo- Dhaka. Indeed, the above-mentioned projects will likely lead to an increasing demand for COX’S BAZAR ing and proposed road projects also hold services in Maheshkhali and neighboring upazilas in Cox’s Bazar district (through higher promise for regional development. The demand for real estate, urban services, transport and communication services, and oth- Ukhia cross-border road network improvement ers). However, considering the capital intensity of the investments and the low skills and project includes the reconstruction of four human capital endowment of the district, the direct beneficiaries of these investments are bridges between Chittagong and Cox’s likely to be larger export-oriented business located in Chittagong and Dhaka. Targeted pol- Teknaf Bazar and is linked to the development of icy actions are needed to ensure that the people of Cox’s Bazar also benefit fully from these Matarbari port. Projects still at the proposal strategic investments. stage include the upgrading from two to four lanes of the National Highway from St. Martin Dwip Chittagong to Teknaf47 and the construction of an alternative route (the N1) connecting Minutes travel to Matarbari deep sea port With current transport infrastructure regional highways to improve connectivity48 (JICA 2018b). In addition to these projects, 0 60 120 150 180 214 and considering the increased congestion since the large-scale arrival of Rohingyas in Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models 2017, the Government of Bangladesh has from the High-Resolution Settlement Layer from initiated a project to upgrade the road seg- Facebook and the Center for International Earth Science Information Network 2016. ment connecting Teknaf with Cox’s Bazar. 46 Asian Highway, SAARC highway corridor, SASEC road corridor, BIMSTEC road corridor, BBIN MVA cor- ridor, Chittagong Port access from the North East, and India and BCIM economic corridor (JICA 2016). 47 “Due to the expected large-scale social environmental impact, financial source for the construction works of the project has not been confirmed yet” (JICA 2018b, page 3.17). 48 “This new road construction project has been proposed by the Cox’s Bazar Road Division of RHD. 49 For more information on the projects see JICA (2016), RHD (2019), and RHD (2018). This road will be a 20 km long Regional Highway connecting R170 to Z1132 as a secondary road of N1” 50 The project is part of the Railway Master Plan formulated by the GoB, ADB, JICA, and WB. It is (JICA 2018b, page 3.17). expected to be completed by 2025. See https://www.adb.org/projects/46452-002/main#project-pds 78 79 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC CHAPTER 3. Economic outcomes: Jobs, livelihoods, and incomes The previous chapter reviewed foundational endowments that affect prospects for inclusive growth in Cox’s Bazar. This chapter discusses Cox’s Bazar’s economy and how well it has been working to translate those endowments into inclusive development and welfare outcomes. The chapter starts by examining the structure of the district 's economy, focusing especially on the sectoral composition of economic activity and the characteris- tics of firms. The second part of the chapter explores the livelihoods that the local econ- omy enables for the people of Cox’s Bazar. Throughout the chapter, economic structures, trends, and outcomes in Cox’s Bazar are compared with those in Chittagong division and Bangladesh as a whole, providing a sense of what Cox’s Bazar could achieve. Structure of the Cox’s Bazar economy: Economic activity and firm composition The contribution of Cox’s Bazar district to the national economy must be measured using proxy indicators. These suggest that Cox’s Bazar’s economic contribution is not directly commensurate to its population, a common pattern among districts in Bangladesh. Official data on sub-national estimates of economic growth and output are not available in Bangladesh, requiring the use of imperfect proxy indicators. Such indicators provide indi- rect evidence that Cox’s Bazar, which accounts for 1.6 percent of Bangladesh’s population, may not be contributing commensurately to the country’s economic activity. In agricul- ture, for example, Cox’s Bazar district represented less than 1 percent of total Bangladeshi crop production in 2017 and around 7 percent of production in Chittagong division. Net cropped area in Cox’s Bazar district is 1 percent of total net cropped area in Bangladesh and 8 percent of net cropped area in the division.51 Within the agricultural sector, fish pro- The analysis in this section relies mainly on the 2011 Population Census, the 2013 Economic Census, 51 and the 2019 CBPS, as these are the only sources of statistical data that allow for inferences at the 80 81 4% Share of Shar 1% Cox's Bazar 3% 2% 1% Cox's Bazar 1% 0% 0% 0% 1% 2% 3% 4% 0% 1% 2% 3% 4% COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Share of working Chapte r 3population age – E C ONOM I C Share OUT C OMES : J OBS , of L I working age VE L I HOO D S population , AN D I N C OMES duction and livestock in Cox’s Bazar each account for 1 percent of national production and Figure 3-3. Figure 3-3: Economic Economic performance performance Figure 3-4. Figure 3-4: Figure Economic3 2:performance Economic 5 percent of division production. The Economic Census of 2013 found that industry and among Bangladeshi among Bangladeshi districts districts (3): (3): performance among districts among Bangladeshi Bangladeshi (4): service-sector enterprises in Cox’s Bazar represented 1 and 6 percent of such firms in the share of total net cropped area in share of total net cropped area in districts (2): share of total crop production fish share of national in country and division, respectively. relation to relation to population population production in relation relation to net cropped to population area Cox’s Bazar’s economic performance is generally on par with districts of similar popula- 4% Naogaon Mymensingh 6,0% tion, but with some unusual traits. Many Bangladeshi districts perform at similar levels on Dinajpur 4% Share of total crop production the indicators that proxy subnational economic contribution (Figure 3-1). The long coast- 5,0% Share of net cropped area 3% Rajshahi line of Cox’s Bazar should be a comparative advantage for sea-caught fish, and indeed, Bogra Comilla 4,0% Bogra the district contributes 20 percent of national production. But when considering total 3% Naogaon Rangpur national fish production, the district’s contribution is relatively low, with the exception 2% 3,0% Mymensingh of shrimp, where it represents a tenth of national production (Figure 3-2). Indeed, pond- Jhalokati Tangail 2% based fish production gives districts such as Comilla (on the Dhaka-Chittagong highway, 2,0% Cox's Bazar 1% in Chittagong division), Mymensingh district (in the division of the same name, home to 1,0% the fourth-largest city in Bangladesh), and Jessore (near Khulna) the edge in production. 1% Cox's Bazar Similarly, despite its reliance on agriculture as a source of employment, net cropped area 0% 0,0% in Cox’s Bazar is close to the average predicted by its population size (Figure 3-3), and crop 0% 1% 2% 3% 4% 0,0% 1,0% 2,0% 3,0% 4,0% production is below the average predicted by cropped area (Figure 3-4). That being said, Share of working age population Share of net cropped area Cox’s Bazar does display some distinct characteristics in terms of its economic structure, apparent in its above-average contribution in sectors including shrimp production, salt Source: World Bank staff estimates, based on Agricultural Yearbook 2017, Economic Census 2013, Population extraction, and some specific cash crops, further discussed below. Census 2011. Figure 3-1: Economic performance Figure 3-2: Economic performance Agriculture and fisheries Figure 3-1. Economic performance Figure 3-2. Figure 3 2: Economic among Bangladeshi districts (1): share among Bangladeshi districts (2): among Bangladeshi districts (1): share performance among Bangladeshi of national non-agricultural firms in share of national fish production in of national non-agricultural firms in districts (2): share of national fish Agriculture was a key sector in Cox’s Bazar before the Rohingya influx and remains an relation to population relation to population relation to population production in relation to population economic mainstay. On the 2011 Population Census, 50 percent of households in Cox’s Bazar reported agriculture as their main sector of employment, followed by 43 percent in 4% 12% Mymensingh services and 7 percent in industry. Beyond the structure of employment, a third of house- Mymensingh 11% 4% holds at that time relied on agricultural labor as their main source of income, with another Share of non-agricultural firms 10% 10 percent deriving income mainly from fishing activities.52 These shares are 11 and 4 per- Share of fish production 3% 9% Comilla 8% centage points higher than the division average, and 7 and 2 percentage points higher than 3% Jessore Comilla 7% the national mean.53 In addition, 76 percent of rural Cox’s Bazar households are involved in 2% 6% activities related to livestock and poultry (BBS 2018a). This highlights the relative impor- 5% 2% tance of the agricultural sector in terms of employment and incomes in Cox’s Bazar, prior 4% 1% Cox's Bazar 3% to the Rohingya influx. More recently, preliminary reports from the 2019 agricultural census 2% 1% Cox's Bazar 1% 0% 0% 0% 1% 2% 3% 4% 0% 1% 2% 3% 4% 52 The agricultural census defines “Agriculture labor households” as those whose major source of income during the preceding year was obtained by working as agriculture labor. Agriculture labor con- Share of working age population Share of working age population notes receiving wages either in cash, kind, or both for performing agricultural work on land operated by other holders (BBS 2010). 53 The differences between division and national average are lower when Dhaka and Chittagong are not Figure 3-3. Economic performance Figure 3-4. Figure 3 2: Economic considered. In this case, the differences between Cox’s Bazar and the division and national mean are 8 among Bangladeshi sub-district level. districts (3): performance among Bangladeshi and 2 percentage points, respectively. share of total net cropped area in districts (2): share of national fish relation 82 to population production in relation to population 83 4% Naogaon Mymensingh 6,0% COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES estimate that 41 percent of the district’s households are farming households, cultivating at Contrary to other districts in the northern hills and coastal plains, agricultural activi- least 0.05 acres of land.54 Agriculture persists as the mainstay of the Cox’s Bazar economy. ties in Cox’s Bazar are based on small production units.58 As discussed above (see Box 3), while dependent on agricultural activities, Cox’s Bazar district has relatively little cultivable Cox’s Bazar accounts for a significant share of Chittagong division’s production of cash land. Indeed, 90 percent of farms in the district measure less than 1 hectare (Table 3-1), crops such as tea, tobacco, betel nut, and betel leaf, suggesting potential for specializa- compared to roughly 80 percent of farms on average for this agricultural zone (BBS 2010). tion and diversification. Fruits, vegetables, and the crop group formed by tea, tobacco, Considering that Teknaf and Ukhia are home to 34 percent of the district’s forest land, and betel nut, and betel leaf represent 12, 7, and 9 percent of Cox’s Bazar’s total farming pro- that 35 percent of the district’s land is not available for cultivation (BBS 2018c), the south- duction, respectively. Among these crops, the district is one of the major contributors to ern upazilas appear to have a larger share of medium and large farms than other areas division production of tea, tobacco, betel nut, and betel leaf, accounting for a quarter of of the district. This could be an indicator of lower urbanization in those areas. The 2011 Chittagong’s production and cultivated area for these outputs. Population Census showed that, compared to an average of 76 percent rural popualtion in the other upazilas in Cox’s Bazar, 92 and 94 percent of the population lived in rural areas in While high and medium-elevation land in the district’s northern hills provides an oppor- Teknaf and Ukhia, respectively.59 tunity to increase agricultural diversification, farming in Cox’s Bazar remains dominated by rice cultivation. Comparing agricultural diversification among Chittagong districts in Land markets in Cox’s Bazar reflect the nationwide upward trend in the share of cultivated similar agroecological zones shows greater diversification in the northern hills and coastal land under tenancy.60 Following the national trend, the share of tenant households has plains than in low-lying lands such as estuarine and river flood plains (Figure 3-5).55 While increased from 12 to 20 percent in the district between 2008 and 2019. Various studies have the share of fruits and vegetables is only 25 percent in the division’s low-lying districts, these documented an increase in the share of tenancy in rural Bangladesh.61 However, despite the crops represent almost 40 percent of total crop production in the higher-elevation districts. rising trend, evidence shows that a better-functioning rental market in the country would 56 However, examining the crop structure of districts in the northern hills and coastal plains, allow households that are more efficient at cultivating to rent land, not only improving their Cox’s Bazar is still producing a lower share of high value-added crops than its neighbors.57 living standards, but also increasing aggregate productivity (Genoni et al. 2021). The farming sector in Cox’s Bazar is dominated by rice, which represents 69 percent of total district farm production (Figure 3-6). This compares to 49 percent nationally (BBS 2018c). Limited adoption of new technology constrains diversification and productivity in Even where agroecological characteristics create a relatively unfavorable environment for higher-value crops. Reforms since 1980 have included the distribution of small irriga- growing rice, lower yields and returns in other crops, along with the need to ensure food tion equipment and the elimination of import restrictions on agricultural equipment. supply, appear to create a bias towards rice production (Gautam and Faruqee 2016). As Such measures have facilitated the rapid adoption of mechanized irrigation across the noted in the Bangladesh Rural Income Diagnostic (Genoni et al. 2021), the high concentra- country. Irrigation has not only improved productivity but enabled farmers to introduce tion of rice as the crop of choice stems from the lower risks associated with its cultivation. multi-cropping systems and to plant during dry seasons (Gautam and Faruqee 2016). The policy environment for rice cultivation – including input subsidies, price management, However, while at national level 50 and 96 percent of the gross and net cropped areas and targeted research and development – is very favorable, reflecting longstanding politi- are covered by irrigation systems, in Cox’s Bazar, these shares are lower: 40 and 70 per- cal concern to promote food security. For alternative crops that are perishable (unlike rice), cent, respectively. At national level, 80 percent of irrigated land is covered by tube well commercialization poses additional challenges. systems, but in Cox’s Bazar, tube well systems cover only 21 percent of irrigated land, while 69 percent is irrigated with power pumps, and 10 percent still relies on traditional irrigation methods (BBS 2018c). 54 The agricultural census defines farm households as those cultivating at least 0.05 acres. Non-farm households are those who have no cultivated or operated land or who are cultivating less than 0.04 acres. 55 An agroecological zone is an area characterized by having homogeneous agricultural and ecological 58 Farm size groups are defined as follows. Small: 0.05 to 2.49 acres/0.02 to 1.007 hectares. Medium: 2.5 characteristics. Bangladesh has delineated 30 agroecological zones based on four elements: physiog- to 7.49 acres / 1.01 to 3.03 hectares. Large: more than 7.5 acres/more than 3.03 hectares. raphy, soil properties, land levels in relation to flooding, and agro-climate. See Figure A 1 in Annex 1. 59 See Figure A-4 in Annex 1. 56 Districts in Estuarine and River flood plains are Chandpur, Comilla, Feni, Brahmanbaria, Lakshmpur, 60 Tenancy is defined according to the criteria used for Bangladesh’s Agricultural Census. Tenant hold- and Noakhali. ers are defined as those having no owned land but operating land belonging to others on a sharecrop- 57 See Figures A1-2 and A1-3 in Annex 1 for a district-wise diversification pattern within each agroeco- ping basis or on other terms (BBS 2010). logical zone in Chittagong division. 61 See Hossain and Bayes (2009); Hossain and Bayes (2018); Genoni et al (forthcoming). 84 85 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-5. Share of crops in total Figure 3-6. Share of crops in total Figure 3-5: Share of crops in total Figure 3-6: Share of crops in total Table 3-1: Size of land holdings, Cox’s Bazar and comparator areas, 2008 agricultural production, different agricultural production, Cox’s Bazar agricultural production, different agricultural production, Cox’s Bazar agroecological zones, Chittagong and comparator areas, 2018 Non-farm Total Farm agroecological zones, Chittagong and comparator areas, 2018 Small* Medium* Large* division, 2018 households households division, 2018 Noakhali 35% 65% 88% 10% 1% 6% 3% 8% Brahmanbaria 45% 55% 89% 10% 1% 9% 7% 7% 22% 16% 19% 12% 17% Chandpur 41% 59% 96% 4% 0% 17% 9% Comilla 39% 61% 94% 5% 0% 7% 21% 11% 20% 2% 2% Feni 46% 54% 92% 8% 0% 23% 3% 7% Lankshmipur 35% 65% 93% 7% 1% 2% 1% Khagrachhari** 23% 77% 64% 31% 5% 65% 69% 61% Rangamati** 26% 74% 49% 42% 9% 54% 50% 49% Bandarban** 29% 71% 54% 38% 8% Chittagong** 72% 28% 91% 8% 1% Cox’s Bazar** 56% 44% 90% 9% 1% l p il in st st s ric er on r sh di oa ill za ta n H s ts l st iv ts la ls Chakaria*** 55% 45% 89% 10% 1% ric a isi n &c nH as er tric de Ba di r in iv n d la er co th is ai an gD x's ng & or r d rth pl ine Co Ba on Cox’s Bazar Sadar*** 71% 29% 92% 7% 1% he No ag od ar Ot flo stu ai N itt pl Ch Kutubdia*** 60% 40% 93% 7% 0% E Moheskhali*** 53% 47% 92% 7% 1% Rice Fruits Other crops (including jute) Tea,tobacco, betelnut & betel leaf Vegetables Pekua*** 57% 43% 91% 8% 1% Ramu** 45% 55% 85% 13% 1% Source: World Bank staff calculations (BBS 2018c). Teknaf*** 63% 37% 84% 14% 2% Ukhia*** 34% 66% 92% 7% 1% Farm households appear to have good access to markets in Cox’s Bazar, though lev- els remain below the national average. Sixty-four percent of farm households in Cox’s Northern hills and coastal 62% 38% 81% 17% 3% plains Bazar are selling their products in a haat bazar, compared with 85 and 77 percent at the national and division level, respectively. More importantly, 73 percent of the markets Chittagong division 48% 52% 89% 10% 1% where households sell their products are less than 4 km away, and 35 percent are less Bangladesh 47% 53% 84% 14% 2% than 2 km away (BBS 2018a). Indeed, agricultural markets in Bangladesh appear to be Note: *Shares over the total of farm households. **Districts included in northern hills and coastal plains. ***Cox’s functioning quite efficiently, with limited information asymmetries and low marketing Bazar upazilas. margins (Gautam and Faruqee 2016). Despite the physical proximity of markets to pro- Source: World Bank staff calculations, Agricultural Census 2008. ducers, marketing margins arise due to the quality of transport infrastructure and logis- tics costs. While evidence on market integration is scarce for perishable value chains, for non-perishable value chains such as rice, market integration appears to be better, in part The district’s fishing (particularly shrimp) sector presents an opportunity to enhance reflecting high levels of government intervention. exports and household income diversification. Fisheries and livestock sectors are import- ant because they act as stabilizers, create employment, improve food security, and contrib- ute to poverty reduction (Gautam and Faruqee 2016). Moreover, fisheries are Bangladesh’s 86 87 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES main agricultural export, making the sector important from the perspective of export Services and industry diversification.62 While the district’s overall fish production constitutes less than 1 percent of national fish production, it represents a fifth of national sea-caught fish, and Cox’s Bazar Data gaps limit our understanding of the non-agricultural economy in Cox’s Bazar, par- seems to have a comparative advantage in shrimp and prawn farming. Shrimp and prawn ticularly in terms of recent changes in the nature and composition of the firm economy. farming accounts for 76 percent of fisheries production in the district, 90 percent in the Much of our understanding of the non-agricultural economy at the sub-national level is division, and 10 percent of total national production (BBS 2018c). Pond production in Cox’s reliant on the 2013 Economic Census. While no longer current, these data identify some Bazar seems to be more developed, given the larger share of highly intensive ponds com- structural characteristics of Cox’s Bazar, which are distinct from the national economy, and pared to other districts in the division and relative to the division average. Despite these help understand how these differences may have shaped more recently measured changes apparent advantages, significant obstacles to the sector’s further development persist. in employment patterns in the district. These include the need to reform contracts and improve productivity, market access, trace- ability, and food safety systems (Toufique and Ahmed 2014). Moreover, fish production in The share of firms involved in key non-agricultural sectors in Cox’s Bazar was consis- ponds in Cox’s Bazar tends to be extensive rather than intensive, compared with other dis- tent with the national pattern in 2013. However, the composition of employment dif- tricts in the division (Table 3-2), lowering average productivity. fered between the district and the national level. Figure 3-7 shows these relationships. Nationally and in Cox’s Bazar district, the majority of firms were engaged in wholesale and retail trade, but a relatively smaller share of Cox’s Bazar’s firms were engaged in transpor- Table 3-2: Intensiveness of fish production in ponds, Cox’s Bazar and tation and storage (Figure 3-7a & Figure 3-7b). More substantial differences emerge in the comparators, 2017 composition of employment. Despite the large proportion of Bangladeshi firms dedicated Extensive Semi-intensive Intensive Highly Intensive to wholesale and retail trade, such firms accounted for only 34 percent of the country’s total employment. In contrast, trade accounted for 51 percent of employment in Cox’s Bazar, Bandarban 25% 75% 0% 0% suggesting that this sector was more labor intensive here than nationally (even excluding B.Baria 1% 58% 38% 3% Dhaka and Chittagong from the national average). The opposite pattern is evident in man- ufacturing. This sector accounted for 11 percent of firms but 29 percent of employment at Chandpur 1% 59% 41% 0% the national level. In Cox’s Bazar, the 14 percent of firms in manufacturing accounted for only 12 percent of jobs (Figure 3-7b & Figure 3-9). The composition of firms and employ- Chittagong 18% 64% 18% 0% ment in transport diverged in a similar way between district and national levels. Comilla 2% 42% 48% 9% Figure 3-7a. Figure Sectoral composition 3-7a: Sectoral composition of Figure Figure 3-7b. Sectoral composition 3-7b: Sectoral composition of Feni 3% 70% 26% 2% non-agricultural firms and non-agricultural firms, Cox’s Bazar, of non-agricultural firms and of non-agricultural firms, Cox’s Bazar, Khagrachari 10% 76% 14% 0% employment, employment, Bangladesh, Bangladesh, 2013 2013 2013 2013 Lakshmipur 1% 74% 25% 0% 7% 8% 6% Noakhali 3% 93% 4% 0% 11% 13% Rangamati 5% 92% 3% 0% 46% 47% 13% Cox’s Bazar 33% 48% 9% 10% 17% 7% Chittagong Div. 5% 61% 31% 3% 11% 14% Bangladesh 3% 46% 31% 20% Accommodation and food Transportation and storage Trade Manufacturing Note: Extensive <1.5 Metric tonne (MT)/Ha; Semi intensive 1.5-4 MT/Ha; Intensive 4-10mt/ha; Highly intensive +10 MT/Ha. Other service activities Others Source: World Bank staff calculations, Yearbook of Agricultural Activities 2017. Source: World Bank staff calculations, Economic Census 2013. Note: “Others” includes industries such as mining, construction, utilities, education, health, public administration, financial, professional services, information and communication, and real estate. “Other service activities” in Cox’s 62 See Atlas of Economic Complexity https://atlas.cid.harvard.edu/ Bazar include tailoring and hairdressing. 88 89 Figure 3-8. Sectoral composition and Figure 3-9. Sectoral composition employment, Bangladesh, 2013 employment, Cox’s Bazar 2013 Accommodation and food Transportation and storage Trade Manufacturing Other service activities Others Source: World Bank staff calculations, Economic Census 2013. Note: C OX’S“Others” BAZAR includes — INC L U S I V E such industries G R Oas WT mining, G N O S T I C utilities, education, health, public administration, H D I Aconstruction, C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES financial, professional services, information and communication, and real estate. “Other service activities” in Cox’s Bazar include tailoring and hairdressing. Figure 3-10. Share of non-agricultural Figure 3-8: Sectoral composition and Figure 3-9: Sectoral composition for another 30 percent (Figure 3-12). The Figure 3-10: Share of non-agricultural firms by main activity -Cox’s Bazar employment, Bangladesh, Figure 3-8. Sectoral 2013 and composition employment, Figure Cox’s Bazar 3-9. Sectoral 2013 composition composition of the private non-agricultural firms by main activity -Cox’s Bazar district, Chittagong division, and employment, Bangladesh, 2013 employment, Cox’s Bazar 2013 enterprise sector also varies by upazila, district, Chittagong division, and Bangladesh, 2013 with some having a larger industrial pres- Bangladesh, 2013 ence (Figure 3-13). These include Chakaria, 5% 8% which is home to three-quarters of the 15% 16% district’s RMG firms, and a quarter of firms 34% 8% engaged in salt extraction. In contrast, pri- 47% 46% 45% 51% 3% vate non-agricultural enterprises in Teknaf 9% 10% and Ukhia are dominated by the services 29% 12% sector, with trade representing more than half of service-sector firms in both upazi- 18% 19% 22% las. Along with Chakaria, Teknaf has the 7% 8% 10% Accommodation and food Transportation and storage Trade Manufacturing highest ratio of firms to population in the 7% 17% 8% Other service activities Others district.63 The main difference between 6% 9% 8% 11% Teknaf and the two largest upazila econ- 5% 3% 4% Source: World Bank staff calculations, Economic Census 2013. omies, Chakaria and Cox’s Bazar Sadar, is 0% 0% Note: “Others” includes industries such as mining, construction, utilities, education, health, public administration, in the composition of firms. In Teknaf, 90 r sio g sh za vi on n financial, professional services, information and communication, and real estate. “Other service activities” in Cox’s de Ba Di tag percent of all firms are in services, mainly la Bazar include tailoring and hairdressing. x's it ng Ch Co Ba wholesale and retail trade. In Chakaria and Cox’s Bazar Sadar, 66 and 84 percent of Extraction of salt Other services The RMG (ready-made garment) sector, the engine of Bangladesh’s recent growth, was firms are in services, respectively. Teknaf’s Transport Other industry an important source of economic activity in Cox’s Bazar, the district’s contribution to the most important industry is salt extraction, Trade Accomodation and Food RMG sector at the national level was modest. RMG garment and textile firms represented accounting for 4 percent of all firms in this Manufacture of a relatively larger share of firms in Cox’s Bazar than at national and division level (9, 3, and textiles and RMG upazila, followed by RMG at 3.5 percent. 4 percent respectively), but the district’s contribution to the national RMG industry was In contrast, RMG and the textile industry Source: World Bank staff calculations, Economic only 3 percent in 2013. On the other hand, while the share of firms in the salt extraction account for a quarter of firms in Chakaria. Census 2013. industry was negligible at national and division level, salt was the second most important Figure 3-11. Cox’s Bazar firms, as a share of division and national firms, diverse industry in Cox’s Bazar. Five percent of non-agricultural firms in Cox’s Bazar are engaged in sectors, Figure 2013 3-11: Cox’s Bazar firms, as a share of division and national firms, salt extraction, and these firms account for more than three-quarters of all salt extraction diverse sectors, 2013 firms in the country (Figure 3-10 and Figure 3-11). Trade is the main component of the dis- Other services trict’s service sector, followed by other services and hospitality (food and accommodation), consistent with findings at divisional and national levels. However, the composition of the Accomodation and food district’s industrial sector departed from divisional and national patterns. Within industry, Transport garment and textile firms represented 9 percent of all non-agricultural firms in Cox’s Bazar Trade and 64 percent of the district’s manufacturing firms, but only 3 percent of non-agricultural Other industry firms in Chittagong and nationally. The vast majority of garment and textile firms in Cox’s Manufacture of textiles and RMG Bazar are single-person enterprises. Extraction of salt 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% The distribution of non-agricultural firms by sector of activity was uneven across upa- Chittagong Division Bangladesh zilas in Cox’s Bazar and reveals a concentration of economic activity in northern areas of the district. Chakaria and Cox’s Bazar Sadar together account for almost half of all Source: World Bank staff calculations, Economic Census 2013. non-agricultural enterprises in the district, followed by Teknaf and Ramu, which account 63 See Table A1-5 in Annex 1. 90 91 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-12. Where are firms located in Cox’s Bazar? Share of district firms by Figure 3-12: upazilas, Share of district firms by upazilas in Cox's Bazar, 2013 2013 number of workers as non-micro enterprises (Table 3-3).64 Overall, Cox’s Bazar mimics national and division patterns, with more than half of non-agricultural firms having less than 5 employees. While in services the firm-size distribution is similar, manufacturing Chakaria firms in Cox’s Bazar are characterized by a much larger share of 1-2 person enterprises, 8% Maheshkhali relative to the national and division-level average.65 15% 27% Teknaf Cox's Bazar Sadar Non-micro firms are smaller in Cox’s Bazar than nationally. Larger firms tend to be older Pekua in the district and the country as a whole. The average size of non-micro firms in Cox’s 14% Ukhia Bazar is less than half the Bangladesh average of 46 (Figure 3-14). Firm size also varies 20% Utubdia across upazilas, with Ukhia having the smallest average, 16 employees, compared to 20 5% Ramu 8% 3% in Teknaf and 30 in Cox’s Bazar Sadar (Figure 3-15). At both national and district levels, non-micro firms have been operating for almost twice as long as micro firms, on average. In Cox’s Bazar, non-micro firms have been in operation for 18 years, on average, compared to half that for 1-4 person enterprises (Figure 3-16). Many authors have highlighted the impor- Source: World Bank staff calculations, Economic Census 2013. tance of new and young firms in job dynamics (Haltiwanger et al. 2013, 2017). Figure 3-13: Distribution of firms by sector within Cox’s Bazar upazilas, 2013 A dynamic firm environment is associated with more job creation, given that well-estab- Figure 3-13. Distribution of firms by sector within Cox’s Bazar upazilas, 2013 lished firms have a limited capacity to grow, and because the process of “creative destruc- Ukhia tion” allows the market to allocate production factors to more efficient enterprises (Farole Teknaf and Cho 2017). Firm size plays an important role for a firm’s survival or growth in the long run. Ramu For instance, using firm-level data from several developing countries, Goswami, Medvedev, Pekua and Olafsen (2019) show that high-growth firms are usually not small, but mid-sized firms. Maheshkhali Likewise, the literature suggests that exit rates are higher among smaller firms (Jovanovic Kutubdia 1982), and that financial constraints are especially relevant for young, small firms. More wor- Cox's Bazar Sadar rying, even the old non-micro firms are not large. This suggests that the “up or out” dynamics Chakaria observed for example in the United States (Haltiwanger, Jarmin, and Miranda 2010) are not playing out in Cox’s Bazar, indicating a lack of pro-competitive forces. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Extraction of salt Other services Accommodation and food Manufacture of Textiles and RMG Other industry Trade Transportation and storage 64 Bangladesh industry policy establishes the following criteria to classify firms by size: Manufacturing: Cottage: up to 9 workers or value of fixed assets excluding land and building less than Tk. half a mil- lion. Microenterprises: 10-24 workers or value of fixed assets excluding land and building between Source: World Bank staff calculations, Economic Census 2013. Tk 500.000 and 5.000.000. Small: 25-99 workers or value of fixed assets excluding laznd and building between Tk 5 million and 100 million. Medium: 100-250 workers or value of fixed assets excluding land and building in between Tk 100 million and 300 million. Large > 250 workers or of fixed assets Firm size, informality, sectoral concentration, and spatial distribution excluding land and building in excess of Tk. 300 million. Non-manufacturing sector: Small: 10-25 workers or value of fixed assets excluding land and building between half a million and 10 million. In terms of firm size, Cox’s Bazar reflects broad national and divisional patterns, with Medium: 50-100 workers or value of fixed assets excluding land and building between 10 million and 150 million. Large >100 workers or value of fixed assets excluding land and building in excess of Tk. very small, non-manufacturing enterprises accounting for the vast majority of all 150 million. In every case, if a firm falls in two categories, it will be classified according to the larger non-agricultural firms. Using official definitions of firm size by number of employees, 98 one (Bangladesh Industry Policy). For simplicity, in this diagnostic, we classify firms based solely on percent of all non-agricultural enterprises in Bangladesh fall under the smallest category, the number of workers. Given the preponderance of enterprises in the cottage category among both cottage enterprises, with less than 10 employees (in services, the proportion is 88 percent). manufacturing and non-manufacturing firms, we further disaggregate these, while pooling together The cottage enterprise category can be further disaggregated into 1-2 person enterprises, all firms with more than 10 employees into a non-micro category. 65 See Table A1-6 in Annex 1 for detailed composition of firms by size in Cox’s Bazar, Chittagong, and 3-4 person enterprises, and 5-9 person enterprises, treating all enterprises hiring a greater Bangladesh. 92 93 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-14. Average size of Figure 3-15. Average size of Figure 3-14: firms non-micro Average size (by of non-micro number of Figure 3-15: firms non-micro Average size (by of non-micro number of industry is another important cluster of larger firms, representing 6 percent of non-micro workers), Cox’s Bazar and Cox’s firms (by number of workers), employees), Cox’s Bazar upazilas, firms (by number of employees), Cox’s enterprises. In contrast, among smaller firms, employing between 3 and 10 workers, more comparator Bazar areas, 2013 and comparator areas, 2013 2013 Bazar upazilas, 2013 than 60 percent are engaged in trade.67 30 115 Table 3-3: Firm size (official versus IGD classification of enterprises) 25 24 21 20 Firm size 18 Firm size 16 16 (classification (official % firms in % firms in used in 58 classification) this report) 46 36 35 34 29 28 25 25 24   Bangladesh Chittagong Cox Bazar   Bangladesh Chittagong Cox Bazar 22 22 Manufacturing Brahmanbaria Cox's Bazar Noakhali Lakshmipur Chandpur Feni Comilla Bandarban Rangamati Khagrachhari Bangladesh Chittagong div. Chittagong sadar Ukhia Maheshkhali Chakaria Teknaf Kutubdia Pekua Cox's Bazar Ramu 1-2 workers 4.03% 6.64% 10.54% Cottage: <10 10.40% 13.63% 13.91% 3-4 workers 4.92% 5.63% 2.48% 5-9 workers 1.45% 1.36% 0.90% Micro Source: World Bank staff calculations, Economic Source: World Bank staff calculations, Economic enterprises: 0.21% 0.13% 0.06% Figure Census 3-16.Average age of firms, by 2013. Census 2013. 10-24 Note: Includes only firms with 10 or more workers. firm size, Cox’s Bazar, 2013 Small Figure 3-16: Average age of firms, In contrast to the national pattern, large enterprises: 0.30% 0.16% 0.07% Non-micro: 25-99 0.59% 0.36% 0.15% by firm size, Cox’s Bazar, 2013 >= 10 RMG and textile firms are not a distinguish- ing characteristic of Cox’s Bazar. Farole Medium: 0.04% 0.03% 0.02% 100-250 10 and Cho (2017) found that, at the national level, 4 out of 5 firms with more than 100 Large:>250 0.04% 0.04% 0.00% 5-9 employees are manufacturers, and around 50 percent of them are in the RMG sector. Services Workers 3-4 Considering that the average size of firms 1-2 workers 63.77% 53.74% 50.30% with more than 10 employees in Cox’s Bazar 2 is far below 100 (on average, 22 employees Cottage <10 87.62% 84.57% 84.68% 3-4 workers 17.22% 23.63% 25.85% work in large firms in the district), and that 1 5-9 workers 6.63% 7.20% 8.53% only 0.4 percent of the district’s RMG firms 0 5 10 15 20 have more than 10 employees, large RMG Small enterprises: 1.33% 1.38% 1.19% Years firms are very rare in Cox’s Bazar.66 10-49 Non-micro: 1.39% 1.43% 1.26% Medium: >= 10 Most of Cox’s Bazar’s non-micro firms are engaged in education, financial intermedia- 50-100 0.04% 0.04% 0.06% tion, and public administration, while micro firms are in “non-tradable” services. Most of the largest firms in Cox’s Bazar, employing 10 or more workers, operate in the educa- Large:>100 0.02% 0.02% 0.01% tion and other services sectors, which are mainly represented by firms related to financial Source of official classification shares: Economic Census Final Report, BBS (2015), Table S4. intermediation, government administration, and health activities. These two categories Source of report classification: Statistics produced by the team using micro data from census 2013. alone account for 60 percent of the district’s largest firms. The short-term accommodation 66 See Table A1-7 in Annex 1. 67 See Table A1-8 in Annex 1. 94 95 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Within Cox’s Bazar district, non-micro firms are mainly located in the upazilas of Cox’s registration. Moreover, the share is highest in Cox’s Bazar, where it is close to 100 percent. Bazar Sadar, Teknaf, and Chakaria. Most of these non-micro enterprises are in the service CBPS data show that, on average, only about 12 to 13 percent of the entire work force in sector, with 70 percent engaged in “education” and “other services” activities. Among the cat- Cox’s Bazar had a written contract in 2019 (Figure 3-20). egories included in “other services,” “financial services activities, except insurance and pen- Figure 3-17. Ownership type, by Figure 3-18. Ownership type, sion funding” accounts on average for 12 percent of non-micro firms in the district. The sec- firm size, Figure Cox’s 3-17: Bazar, Ownership 2013 type, by firm non-micro Figure 3-18: firms (>=10type, Ownership workers), non-mi- ond most important category under “other services” is “public administration and defense,” size, Cox’s Bazar, 2013 Cox’s Bazar, cro firms (>=10 Chittagong, andBazar, workers), Cox’s which represents 11 percent of firms across upazilas, on average. “Other industries” is the Bangladesh, Chittagong, and 2013 Bangladesh, 2013 third-largest group of firms characterized by the presence of non-micro enterprises. On aver- 32% 3% 3% 5% 28% age, 14 percent of firms hiring more than 10 employees are in this sector. Within this cate- 4% 20% 27% gory, “manufacture of furniture” and “manufacture of non-metallic mineral products (mainly 3% 25% 25% 25% 22% 22% 22% bricks)” represent on average 6 and 4 percent of non-micro enterprises, respectively.68 17% 19% 22% 17% 5% Relative lack of dynamism among Cox’s Bazar firms may also be related to character- 96% 95% 88% 9% 78% 6% istics such as family ownership and a lower share of private limited firms. Bloom and 9% 7% 6% 6% 6% 5% Reenen (2010) find that most family-owned firms are poorly managed, and hence show low 4% 32% levels of firm performance. In Bangladesh, 90 percent of microenterprises, as well as small and medium firms, are owned by individual families (Figure 3-17 and Figure 3-18). Only e td rs rs rs ke s m al ip d t r en ke or lu en yp Lt among large firms is there a higher share of government, private, and public limited com- Fa idu ke ke ke sh eL Go tity rs Pa ily w 10 p or m of r t ic or or or er at v rn 1w bl he di rtn 2w 4w 9w iv panies, as well as other types of entities. The picture in Cox’s Bazar differs from national Pu ve In Ot Pr 3- 5- and division patterns, however. While in Chittagong and nationwide, 1 out of 5 large firms Individual family Partnership Chittagong Bangladesh are single-family owned, this number rises in Cox’s Bazar to 1 out of 3 firms. Only 9 percent Private Ltd Public Ltd Cox’s Bazar of firms are private limited in Cox’s Bazar, roughly half the national rate. In addition, gov- Government Others type of entity ernment has a higher presence in large firms in Cox’s Bazar, compared to Chittagong and the country as a whole (Figure 3-18). As Farole and Cho (2017) point out, lack of dynamism Source: World Bank staff calculations, Economic Census 2013. Note: Following the Economic Census classification, “other type of entity” includes autonomous, foreign, joint in microenterprises is related to the role of enterprises as a livelihood strategy for house- venture, cooperative, NPI, and Non-Resident Bangladeshi enterprises, and others holds, given lack of other job opportunities. Figure 3-19. Figure The normality 3-19: The of normality of Figure 3-20: Figure 3-20. Share Share of workers with of workers with Informality is a prominent feature in Bangladesh and Cox’s Bazar, which is likely to informality Bangladesh 2013 in Bangladesh, informality in written contracts, written Cox’s Bazar, contracts, Cox’s 2019 Bazar, 2019 have deep economic implications. First, high informality implies widespread tax eva- 100% sion, hindering government’s ability to provide public goods. Second, informality may 14% 99% 98% 93% 89% distort firms’ decisions along important margins, such as the size of their work force. 84% 13% 14% Third, it allows less productive (informal) firms to compete with more productive (for- mal) firms, leading to misallocation of resources and potentially large total factor pro- 13% ductivity (TFP) losses (e.g., Hsieh and Klenow 2009).69 Figure 3-19 suggests that more than 90 percent of firms in Bangladesh have neither a tax identification number nor a VAT 12% 16% 11% 7% 2% 1% 0% 12% Yes No Yes No TIN Registered VAT Registered 11% 68 See Table A1-9, Table A1-10, Table A1-11, Table A1-12, and Table A1-13 in Annex 1 for a detailed distribution of non-agricultural firms by sector and upazila. Chittagong Bangladesh 11% 69 In contrast, informality can sometimes be beneficial to growth, as it provides de facto flexibility for Cox’s Bazar High exposure Low exposure firms that would otherwise be constrained by burdensome regulations. Therefore, understanding what proportion of firms belong to the informal sector is crucial for gauging the aggregate impacts of Source: World Bank staff calculations, Economic Source: World Bank staff calculations, CBPS 2019. policies on economic development. Census 2013. 96 97 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Work and livelihoods in Cox’s Bazar Recent data on the composition of employment in Cox’s Bazar confirms its greater reli- ance on services, particularly in low-exposure areas.71 Using data from the 2016-17 Labor Employment patterns and dynamics Force Surveys and the 2019 CBPS, Table 3-5 shows that both high- and low-exposure areas in Cox’s Bazar rely more heavily on education and health services for employment, com- The benefits of Bangladesh’s economic growth in terms of employment have been pared to the national and division average. Similarly, construction jobs also account for a unevenly distributed across the country. For example, Chittagong division’s female higher share of employment. High-exposure areas (primarily Teknaf and Ukhia) are charac- labor force participation increased from 29 percent in 2005 to 32 percent in 2016, but terized by a higher share of employment in agriculture, whereas services and industry are women’s engagement in the labor force in Barisal and Sylhet declined during the same more important in low-exposure areas (Figure 3-21). period (Labor Force Survey [LFS], various years). Within Chittagong division, evidence from the CBPS points to district-level disparities in key employment variables.70 Table The employment structure within Cox’s Bazar continues to be differentiated by gender 3-4 shows that labor force participation in Cox’s Bazar was below national and division and across space. In 2019, a quarter of working men were employed in agriculture, a fifth in averages, and female participation in the labor force was half that of Chittagong division industry, and more than half worked in service-related activities (Figure 3-21). For women, overall (World Bank 2019a). agriculture is the primary source of livelihoods, engaging half of all working women, with a third employed in services. When breaking down non-agricultural sectors, gender dif- ferences are more striking. Within industry, while a larger share of women are involved Table 3-4: Labor force participation, Cox’s Bazar, Chittagong division, in manufacturing and utilities, men are mostly performing construction activities (Figure and Bangladesh 3-22). Across service activities, women have a larger participation than men in “other ser- LFS 2016-2017 CBPS 2019 vices, education and health.” On the other hand, the male labor force is more diversified, with men’s share of employment being larger than women’s in trade and accommodation, Bangladesh Chittagong Cox’s Bazar transport, and other activities including non-classified waged and salaried workers. Male 83% 80% 71% Table 3-5: Sectoral composition of employment: Bangladesh, Chittagong, and low- and high-exposure areas of Cox’s Bazar Female 35% 32% 19% Bangladesh Chittagong Low-exposure High-exposure Agriculture 41% 39% 30% 41% Total 58% 54% 42% Trade and 16% 19% 19% 15% accommodation Industry 15% 13% 10% 7% Structural change accompanying economic growth in Bangladesh has not only Construction 6% 7% 11% 10% decreased the share of agriculture in GDP but also reshaped employment patterns, Transport 9% 9% 8% 11% with a progressive shift of jobs toward non-agricultural sectors. Using different data Services, education, 14% 13% 22% 17% sources, recent studies including the latest Poverty Assessment (Hill and Genoni 2019), health Bangladesh Jobs Diagnostic (Farole and Cho 2017), and Rural Income Diagnostic Total 100% 100% 100% 100% (Genoni et al. 2021) have confirmed that Bangladeshi individuals and households are Source: World Bank staff calculation, LFS 2016-2017, CBPS 2019. shifting away from agriculture. However, this process is progressive, and the sector still employs a substantial share of workers. 71 Lack of pre-influx data at district level, together with large standard errors regarding employment pat- terns in HIES 2016, prevent reliable pre-post influx comparisons. The high prevalence of informality and home-based work in Cox’s Bazar and across Bangladesh must also be considered. These factors suggest that labor force analysis based on the Economic Census could prompt misleading conclusions about trends. This report uses the Economic Census to demonstrate firm structural characteristics that are less 70 Note that Labor Force Survey data does not allow for district-level comparisons. HIES tends to likely to have changed significantly over the past decade. However, the report bases its descriptions of strongly underestimate female labor force participation relative to other sources, perhaps because the employment patterns on CBPS 2019 to the greatest extent possible. For a more detailed discussion of labor module is filled out by proxy. employment patterns using Economic Census 2013 and Population Census 2011, see Annex 4. 98 99 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-21. Share of host- community Figure 3-22. Share of individuals Figure 3-21: Share of men and women Figure 3-22: Share of individuals Figure 3-23: Sectors of employment, Figure 3-24: Sectors of employment, men and women working in different working, by activity and gender, Cox’s working in different sectors, working, by activity and gender, high-exposure Figure versus 3-23. Sectors oflow-exposure employment, high-exposure Figure versus 3-24. Sectors oflow-exposure employment, sectors, Cox’s Bazar, 2019 Bazar (Bangladeshi households), 2019 high-exposure versus low-exposure high-exposure versus low-exposure Cox’s Bazar, 2019 Cox’s Bazar, 2019 males, Cox’s Bazar, 2019 females, Cox’s Bazar, 2019 males, Cox’s Bazar, 2019 females, Cox’s Bazar, 2019 35% 70% 54% 60% 51% 30% 60% 50% 25% 50% 20% 40% 40% 15% 30% 30% 10% 20% 30% 5% 10% 26% 20% 0% 0% 21% re n s n th rs re n s n t th rs t 19% or or tie tie io io io io he he tu tu al al sp sp at ct at ct ili ili he he ul ul Ot Ot 10% tru tru od od an ut ut an ric ric n, n, Tr m m Tr g& g& ns ns Ag Ag tio tio m m Co Co rin rin ca ca co co du du tu tu ac ac 0% ac ac ,e ,e e& e& uf uf ice ice ad ad an an re n s n t th rs or rv rv tie io io he tu Tr Tr al M M sp Se Se at ct ili he ul Ot tru od an ut ric n, Tr m g& ns High exposure male High exposure female Ag tio m Co rin ca co Low exposure male Low exposure female du tu ac ac ,e e& Agriculture Industry Services uf ice ad an rv Tr M Source: World Bank staff calculations, CBPS 2019. Se Male Female Male Female Women and men also work for different types of employers, with NGOs providing an Source: World Bank staff calculations, CBPS 2019. important source of jobs for women in high-exposure areas. Women’s and men’s jobs are typically linked to different types of employers. The kinds of firms likely to hire women Employment differences between high- and low-exposure areas are more pronounced also change, depending on whether women live in low- or high-exposure areas. This is less for women than men. Looking at the composition of employment among males in high- the case for men. Men in all areas of Cox’s Bazar are mainly employed by small enterprises and low-exposure areas (using the 2019 CBPS), it is evident that individuals in Teknaf and and private institutions (58 and 20 percent, respectively, in high-exposure areas and 51 and Ukhia (high-exposure) have a larger share in low-skill activities such as agriculture and 22 percent in low-exposure areas). In contrast, women working in high-exposure areas are transport than is the case in other areas of Cox’s Bazar (Figure 3-23). On the other hand, mainly employed in NGOs, small enterprises, and other households (35, 36, and 17 percent a larger share of males in low-exposure areas are working in construction and other man- respectively), while women living in more urbanized, low-exposure areas are more likely to ufacturing, as well as in trade and accommodation and other service activities. Among work for government (20 percent of women in the labor force in low-exposure areas com- females, the differences between high- and low-exposure areas are more prominent. As pared to 8 percent in high-exposure areas) and less likely to work for NGOs (which employ previously mentioned, women’s employment is less diversified than men’s, and this low only 9 percent of the female labor force in low-exposure areas). level of diversification is even more pronounced in high-exposure areas (Figure 3-24). Indeed, the agriculture sector represents more than 60 percent of total female employment Most well-educated workers in low-exposure areas are employed by government in Teknaf and Ukhia. Other services—including personal services, NGO work, education, or private companies, while in high-exposure areas, NGOs employ two-thirds of all and health—emerge as the second-largest employer of women in high-exposure areas, fol- workers with at least secondary education. In both high- and low-exposure areas, the lowed by manufacturing and utilities. Women in low-exposure areas are less concentrated share of individuals hired by small enterprises decreases as worker education increases. in agriculture and are more likely to be employed in services than in agriculture. Indeed, while in high-exposure areas, 2 out of 3 individuals with primary education or 100 101 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES less are hired by small enterprises, this share decreases to less than half for those with Given the structure of the local economy, which relies on informal, small-scale, low-pro- incomplete secondary education and to 6 percent for high school graduates. NGOs were ductivity jobs in services and agriculture, it is not surprising that there are limited the main employer for those with more than secondary education in high-exposure returns to education until tertiary level.72 While individuals with secondary education areas (66 percent of this group). Among individuals with at least secondary education or less appear to earn more, on average, than individuals with no education, only tertiary in low-exposure areas, government and private companies were the main employers, education yields a statistically significant and positive impact in determining higher hourly accounting for one-third of such workers each. Among similarly well-educated workers wages. Furthermore, these returns to higher education are larger in areas closer to camps in areas close to Rohingya camps, two-thirds are employed by NGOs (Table 3-6). While than in areas farther from the Rohingya camps. Among the self-employed, earnings are pre- and post-influx comparisons are not possible, it is likely that NGOs working in or significantly higher for those engaged in services, and once this is taken into account, there close to the Rohingya camps have provided new work opportunities for better-educated are no distinguishably different returns to higher education (Table 3-7). host workers and women living in surrounding areas. Table 3-6: How workers sort into different kinds of employers, low-exposure Recent data confirm that agriculture is an important source of employment among the and high-exposure areas, Cox’s Bazar, 2019 less educated, while most people with more than secondary education are employed in services. Half of all employed workers with no education in high-exposure areas work in agri- Private office/ institution/ culture, compared to 38 percent in low-exposure areas (Figure 3-25). As education increases, company/ Small Government mill/*factory NGO Households enterprises reliance on agriculture decreases, with some shifting to industry and services. Among those with incomplete secondary education, more than half work in services, and for the better High Male 7% 20% 11% 4% 58% exposure educated, service jobs are the predominant source of employment (Figure 3-25). Female 8% 5% 35% 17% 36% Low Male 7% 22% 2% 17% 51% 3-25: Probability of employment by sector and level of education, exposure Figure 3-25. Female 20% 13% 9% 32% 26% high-exposure versus low-exposure areas, 2019 Never attended 4% 10% 8% 11% 66% 100% school Less than 90% 5% 20% 2% 9% 64% primary 30% High 80% 35% 38% Complete 43% 45% 43% exposure primary 7% 20% 5% 3% 65% 70% 57% 56% Incomplete 11% 8% 19% 25% 4% 43% secondary 60% 19% 87% 86% Secondary 21% 7% 66% 1% 6% 50% and above 27% 21% 21% 28% Never 40% attended 3% 11% 0% 34% 52% 20% 21% school 30% 59% Less than 46% 6% 16% 0% 20% 59% 20% primary 36% 36% 36% 27% Low Complete 10% 23% 23% 5% 6% exposure 9% 19% 0% 19% 53% primary 8% 8% 0% Incomplete 8% 28% 11% 9% 43% High Low High Low High Low High Low High Low secondary Never attended Less than Primary Secondary Secondary Secondary to school primary complete incomplete complete 33% 34% 5% 13% 15% and above Agriculture Industry Services Source: World Bank staff calculations, based on CBPS 2019. 72 Results based on regressions of log hourly wages (for wage workers) and log monthly earnings (for Source: World Bank staff calculations, CBPS 2019. the self-employed) on education, sector of work, gender, and age cohort. 102 103 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-26. Share of individuals with Figure 3-27. Types of contracts used Table 3-7: Average number of months and weekly hours allocated to primary Figure 3-26: Share of individuals with Figure 3-27: Types of contracts used secondary jobs, Cox’s Bazar, 2019 for secondary jobs, Cox’s Bazar, 2019 and secondary jobs, waged and non-waged workers, Cox’s Bazar, 2019 secondary jobs, Cox’s Bazar, 2019 for secondary jobs, Cox’s Bazar, 2019 26% 22% Non-waged Waged 58% 16% 67% Weekly 87% Month Month Weekly hours 93% 11% hours Primary Secondary Primary Primary Secondary Primary Secondary 42% 33% 7% 13% High-exposure High High Low Low High High Low Low 10 8 45 9 6 45 29 male exposure exposure exposure exposure exposure exposure exposure exposure male female male female male female male female High-exposure Wage employee 10 10 20 9 6 36 33 Self employed / own business female Source: World Bank staff calculations, based on CBPS 2019. Low-exposure 10 7 47 9 6 44 40 male Figure 3-28. Share Figure 3-28: Share of individuals of sectors for with Figure 3-29. Figure 3-29: How types Contract of contracts types for Low-exposure secondary jobs, by sector, Cox’s individuals with secondary jobs, differ for primary and secondary primary and secondary jobs, high- jobs, 10 9 17 9 8 37 17 female Bazar, 2019 high-exposure and low-exposure by gender and high vs low-exposure exposure and low-exposure areas, Cox's Bazar, 2019 2019 areas, Cox’s Bazar, Source: World Bank staff calculations, based on CBPS 2019. Note: Number of observations for secondary jobs are as follows: high-exposure 328; low-exposure 161. Among high-exposure workers: males waged 90; males non-waged 132; females waged 10; females non-waged 96. 3% 4% 7% 6% 9% 5% 11% 3% 4% 17% 15% 4% 3% 4% 17% 4% 8% 26% Reliance on secondary jobs for livelihoods varies within Cox’s Bazar and by worker edu- 13% 36% 42% cation. CBPS data show that, while 1 out 4 individuals in high-exposure areas engage in a 20% secondary activity to complement their main earning activity, only 15 percent of people in low-exposure areas do so (Figure 3-26). In both areas, most of these secondary activities involve running own businesses or self-employed activities (Figure 3-27). Most individuals 89% 85% 74% are engaged in agricultural work (73 and 63 percent for high- and low-exposure, respec- 66% 65% 57% 53% tively) or construction activities (11 and 16 percent for high- and low-exposure, respec- 51% tively). This could indicate that some of these jobs are seasonal in nature. When engaged in secondary activities, women are less diversified than men (89 and 85 percent are engaged in agriculture in high- and low-exposure areas, respectively) (Figure 3-28). On the other High High Low Low Primary employment Secondary employment Primary employment Secondary employment hand, while agriculture still dominates secondary jobs for males, the probability of working exposure exposure exposure exposure in either construction or service jobs is greater than for women. Individuals allocate a sim- male female male female ilar number of hours per week to work in secondary and primary activities, but secondary Agriculture Construction Services jobs are more likely than primary jobs to be based on a daily arrangement and, among High-exposure Low-exposure Other industry wage jobs, to occupy workers for fewer months compared to their primary jobs (Figure 3-27 Hour / Daily Weekly Irregular / Not fixed and Figure 3-29). Monthly / Quarterly / Fortnightly Source: World Bank staff calculation using CBPS 2019. 104 105 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Income sources and livelihoods Agricultural incomes represent a significant share of total incomes among poorer house- holds, with households in high-exposure areas more likely to be exclusively engaged in While the latest official poverty figures predate the 2017 Rohingya influx, available data agriculture.74 As Figure 3-30 shows, agricultural-sector income represents a higher share suggest that the pre-existing welfare gap persists between Teknaf and Ukhia and the other of total income among less well-off households, particularly in high-exposure areas.75 upazilas in Cox’s Bazar. Using data from the 2019 CBPS, per capita incomes in high-expo- Seventeen percent of households in high-exposure areas report earning only agricul- sure areas (primarily Teknaf and Ukhia) are 22 percent lower than in low-exposure areas ture-related income, compared to 12 percent in low-exposure areas. Other studies have (Table 3-8).73 High-exposure areas are also more reliant on incomes from cultivation, whereas also shown that the main livelihoods for vulnerable households in Teknaf were fishing and low-exposure areas report a larger share of households receiving remittance incomes. forestry,76 which is consistent with the larger share of agricultural incomes for the bottom 20 percent of the population in high-exposure areas. Differences in the composition of agri- cultural income between high- and low-exposure areas come largely from wages (Figure Table 3-8: Income sources and average incomes, low-exposure versus high- 3-31). High-exposure areas report a larger share of workers engaged in wage labor in the exposure areas within Cox’s Bazar, 2019 agricultural sector, and income from cultivation represents a larger share of income in high-exposure settings. Moreover, average wages are lower in agriculture, when compared Household income sources High-exposure Low-exposure with services and industry (CBPS 2019), confirming the pattern of lower welfare in high-ex- Wages 53% 53% posure areas. Finally, the shares of income from livestock, fisheries, and forestry are higher in the first and the last quintile (Figure 3-31). Income from cultivation 12% 7% While rice is the main crop in both high- and low-exposure areas, and rice-producing Income from livestock/fishing/forestry 5% 4% households report larger cultivated areas, poorer households are less likely to be culti- Income from non-agriculture business vating rice. Forty percent and 71 percent of households report cropping rice in high- and 14% 14% earnings low-exposure areas, respectively. The three main crops for hosts after rice are chili, potato, and betel. Half of rice producers in low-exposure areas crop only rice, compared with 20 Remittances 11% 15% percent of households in high-exposure areas. This is reflected in the average number of Asset earnings 1% 2% crops cultivated per household: 2 in low-exposure areas versus 1.6 in high-exposure areas. However, households in the bottom 40 percent of the income distribution are less likely to Pensions 0% 0% be cultivating rice in high- and low-exposure areas. Cash assistance from government 1% 1% Other 2% 2% Average per capita income 3,553 4,566 74 Income data was collected at the household level. To distribute total household wages into agri- culture and non-agriculture categories, the adult module data was used. If members of a household Average income 16,972 21,370 reported all wages to be earned in one sector, the total household wage was classified accordingly. If members reported different sectors for wages, the total household wage was distributed proportion- Source: World Bank staff calculation using CBPS 2019. Note: Average income in Takas. All indicators calculated using only households reporting income. Percentages indi- ally using the individual wages reported. If members reported sectors but not wages, and the house- cate share of households reporting income above zero. hold reported income from wages, wage income was distributed in equal proportions. (For example, if one adult reported working in agriculture, and another reported working in the non-agriculture sector, then 50 percent of the total wage income was assigned to each sector.) If respondents to the adult module did not provide any information on employment sectors, then wages were classified as not defined. Classification was carried out using International Standard Industrial Classification (ISIC) sec- 73 Ninety-four percent of households in the high-exposure strata of the CBPS live either in Teknaf or tors, complemented by the self-reported agriculture and non-agriculture sectors for the missing cases. Ukhia, with the remaining split between Naikhongchhori and Ramu. Within low-exposure areas, only 75 Agricultural income includes agricultural wages, income from cultivation, and income from live- 3 percent of the sample is from Teknaf and Ukhia, with the largest shares coming from Chakaria (38 stock. Non-agricultural income includes non-agricultural wages, income from non-agricultural busi- percent) and Cox’s Bazar Sadar (31 percent). As previously noted, Maheshkhali and Kutubdia are not ness, remittances, asset earnings, pensions, cash assistance, and others. included in the CBPS sample. 76 For example, see Tani and Rahman (2018). 106 107 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Figure 3-30: Average composition of monthly income (last 30 days) from agricul- 3-31. Average Figure 3-31: Figure Average composition of monthly composition of monthly income, by quintile, income, by per capitalow-exposure income tural and non-agricultural sources, by per secondary capita income quintile, low-exposure and high-exposure areas, 2019 quintile, low-exposure and high-exposure areas, 2019 Figure 3-30. Share of individuals with jobs, by sector, Bazar, 2019 areas, 2019 and high-exposure Cox’s 5 4 100% 4 13 8 13 11 11 6% 6% 7% 90% 14 17 14 6% 11% 13% 11% 8% 10% 8% 17% 18% 80% 5% 4% 19% 4% 16% 11% 19% 70% 14% 22% 11% 16% 3% 4% 6% 17% 24% 47 61 3% 60% 51 18% 75 79 13% 13% 5% 17% 11% 9% 66 56 70 6% 50% 67 73 14% 10% 5% 6% 8% 5% 14% 40% 11% 10% 13% 13% 30% 4% 30% 37% 32% 30% 5% 42% 20% 40 33% 34 31 27% 31% 28 26% 32% 10% 22 20 19 19 16 17 22% 15% 14% 13% 0% 10% 7% 6% 2% 5% Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 High-exposure Low-exposure Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 High-exposure Low-exposure Agricultural income Non-agricultural income Wage not defined Agricultural wages Non-agricultural wages Income from cultivation Income from livestock\fishing\forestry Remittances Asset earnings Source: World Bank staff calculation using CBPS 2019. Cash assistance from the government Others Note: Quintiles generated using per capita income. Agricultural income includes agricultural wages, income from cultivation, and income from livestock. Non-agricultural income includes non-agricultural wages, income from Wages not defined Pensions non-agricultural business, remittances, asset earnings, pensions, cash assistance, and others. “Wage not defined” Income from non-agricultural business earnings includes wages that could not be distributed between the other two categories due to lack of information. See footnote 59 for additional details on wage classification. Source: World Bank staff calculation using CBPS 2019. See footnote 74 for details on wage classification. In both high- and low-exposure areas, households in the bottom two per-capita income quintiles (bottom 40 percent of incomes) rely heavily on wages (Figure 3-31). Finally, cash and in-kind assistance is generally a small source of income. Such assistance Wages represent a higher share of total income and are more important for households is comprised primarily of health and education support and is largely provided by the in low-exposure areas, and at least 40 percent of households in the bottom two quintiles Government of Bangladesh. Households in high-exposure areas are more likely to report report wages as their only sources of income. In both areas, as incomes rise, households assistance, although assistance is not received regularly in any area. The top five assistance rely less on wages from agricultural and non-agricultural work, and more on earnings items reported by hosts are education (37 percent), health (33 percent), blankets/bedding/ from non-agricultural enterprises. Remittances are generally a more important income mosquito nets (26 percent), rice (22 percent), and cash (8 percent). Households in high-ex- source for households in the top income quintiles, particularly in low-exposure areas of posure areas report receiving 63 percent of their assistance from the GoB, 9 percent from the district. Incomes among the top quintile are also more likely to come from multiple WFP, and the remainder from various NGOs. Households in low-exposure areas report 87 sources, with at least 60 percent of households in this quintile in low- and high-exposure percent of assistance coming from the GoB. areas relying on two or more sources. 108 109 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Firm performance and earnings of the self-employed Analysis of incomes in secondary jobs suggests that entry into business and entrepre- neurial activities may be driven by the opportunity of better marginal earnings for wage Using revenue or sales as an indicator for firm performance, service and agricultural workers in high-exposure areas. Secondary employment is common in the rural areas of enterprises in Cox’s Bazar perform below national levels. Based on HIES 2016 household developing countries, particularly as the returns from a single job are low. Table 3-9 shows data, the median agricultural firm in Cox's Bazar earns77 around 6,000 Bangladeshi Taka that weekly earnings are larger for primary jobs and that wage workers earn more than per month, which is less than what a median agricultural firm earns in Chittagong (around the self-employed. It also confirms systematically higher earnings for men compared with 7,000) and Bangladesh as a whole (7,500) (Figure 3-32). In the case of services, firms in Cox’s women, and that women engaged in wage work earn more in high-exposure areas in both Bazar district earn 20 percent less than firms in Chittagong division can expect to earn. On primary and secondary jobs. The returns (hourly wages) to secondary wage work differ the other hand, a manufacturing firm in Cox’s Bazar is likely to achieve earnings similar to based on proximity to camps. While secondary jobs are paid more per hour than primary the national median but 12 percent less than the division median. jobs in high-exposure areas, the contrary is true in low-exposure areas. Within Cox’s Bazar, wage work generates higher monthly income relative to self-employ- Table 3-9: Weekly and hourly wages in primary and secondary jobs, ment in both low- and high-exposure areas, and for men and women. Using 2019 CBPS Cox’s Bazar, 2019 (averages, in Takas) data, monthly earnings can be compared between the self-employed and wage workers, across high- and low-exposure areas and by gender (Figure 3-33).78 Two main conclusions Waged Self employed emerge. First, self-employed women earn far less than self-employed men in high- and Weekly Hourly Weekly low-exposure areas. Second, wage work is a dominant choice for women and men in terms Primary Secondary Primary Secondary Primary Secondary of earned income. Incomes are higher in low-exposure areas relative to high-exposure job job job job job job areas, for wage work and self-employment, except for female wage workers, which may High exposure 2,531 1,596 79 135 1,531 907 reflect the prevalence of NGO jobs near Rohingya camps. Low exposure 3,360 2,150 144 73 1,953 1,047 Figure 3-32: Median net monthly Figure 3-33: Mean monthly labor High exposure 2,627 1,659 82 139 2,358 1,274 Figure revenue3-32. Median of firms: net Cox’s monthly Bazar, Figure 3-33. income Monthly in Cox's labor Bazar: wageincome in male revenue of firms: Chittagong, Cox’s Bazar, and Bangladesh, 2016 Cox’s Bazar: wage employment versus employment versus business profits High exposure 2,125 1,320 69 84 370 298 Chittagong, and Bangladesh, 2016 business profits (average), 2019 (average), 2019 female Low exposure 9,750 18,000 3,700 2,341 156 76 2,713 1,367 9,000 16,000 male 8,333 7,917 7,850 7,533 7,417 14,000 6,817 Low exposure 1,541 1,117 78 31 448 220 5,917 12,000 female 10,000 8,000 High exposure 6,000 2,079 1,379 72 120 1,021 653 agriculture 4,000 2,000 High exposure 0 2,186 1,231 81 217 1,142 933 Wage Self Wage Self industry Bangladesh Chittagong Cox's Bazar Employee Employed Employee Employed Agricultural Industry Services High exposure High Exposure Low Exposure 2,909 1,883 82 94 2,387 1,425 services Male Female Low exposure 2,032 1,191 114 72 1,311 905 Source: World Bank staff calculations from establish- Source: World Bank staff calculations, CBPS 2019. agriculture ments data in HIES 2016. Note: Results do not change qualitatively when median Note: Revenues trimmed at 1 percent of the distribu- values are used instead of means. Low exposure tion for each geographic area. 2,277 2,608 119 56 884 787 industry Low exposure 4,444 2,205 168 79 2,866 1,626 77 Net of expenses. services 78 The reported incomes are the averages of individual labor income reported in the adult module of Source: World Bank staff calculations, based on CBPS 2019. the CBPS questionnaire. 110 111 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Sector-level differences in earning opportunities for the self-employed point to the Overall, the probability of working in small firms is similar for men and women across potential of services. In both low- and high-exposure areas, service jobs yield the highest areas. Roughly 60 percent of women and men are likely to be employed in firms with less average monthly earnings, irrespective of type of employment (Figure 3-34). While wage than five employees in both low- and high-exposure areas (Figure 3-36 and Figure 3-37). work generally pays better than self-employment, among the self-employed, work in ser- Some interesting differences arise when examining sectoral shares. Within agriculture, vices pays substantially more than working in industry or agriculture. Among these sec- women in low-exposure areas are twice as likely as those in high-exposure areas to be tors, therefore, services seem to offer an opportunity to improve earning capacity through working in non-microenterprise firms. In the industry sector, males in high-exposure areas self-employment and entrepreneurship. Despite the possibility of higher earnings in ser- are more likely to be working in larger firms than men in low-exposure areas. At the same vices, self-employment in agriculture remains important, more for subsistence than as an time, women have a higher likelihood of working in larger firms than men, and these differ- entrepreneurial activity, given its lower earnings (Figure 3-35). ences are even higher for low-exposure women. Indeed, the share of wage-earning female workers employed in microenterprises within the industry sector in low-exposure areas is 3-34: Mean Figure 3-34. monthly earnings Mean monthly earnings 3-35: Main sectors of work, by Figure 3-35. half that in high-exposure areas (14 and 29 percent, respectively). by area and sector, Cox’s Bazar, 2019 area and employment type, Cox’s 2019 Bazar,Bazar, Cox’s 2019 2019 Figure 3-36. Figure Share of 3-36: Share ofwage wage Figure 3-37. Figure 3-37: Most wageworkers Most wage workersare are Figure 3-36. Share of wage Figure 3-37. Most wage workers are employment employment by by sector sector andself- and self- employed employed in in very very small small enterprises: enterprises: employment by sector and self- employed in very small enterprises: 20,000 90% reported reported firm firm size, size, high-exposure high-exposure self-reported firm self-reported firmsize amongwage sizeamong wage reported firm size, high-exposure self-reported firm size among wage Earnings (in Banlgadeshi Takas) 18,000 80% versus low-exposure versus low-exposureareas, areas,2019 2019 workers, workers, high-exposure high-exposure and and low- versus low-exposure areas, 2019 workers, high-exposure and Proportion of individuals 16,000 70% low-exposure exposure areas, areas, 20192019 low-exposure areas, 2019 14,000 79% 60% 59% 79% 58% 69% 12,000 59% 58% 69% 50% 57% 57% 54% 54% 42% 10,000 57% 41% 57% 54% 54% 46% 46% 42% 40% 43% 43% 41% 46% 46% 8,000 43% 43% 31% 30% 31% 21% 6,000 21% 4,000 20% 2,000 10% 0 0% 1-5 workers >5 workers 1-5 workers >5 workers High exposure Low exposure Wage Self Wage Self Wage Self Wage Self 1-5 workers >5 workers 1-5 workers >5 workers High exposure Low exposure Employee Employed Employee Employed Employee Employed Employee Employed High exposure Low exposure High exposure Low exposure High Exposure Low Exposure High Exposure Low Exposure Agricultural Industry Services 1 - 5 workers > 5 workers Agricultural Industry Services 1 - 5 workers > 5 workers Agricultural Industry Services Figure 3-38. Figure Share of 3-38: Share ofwage wage Figure 3-39. Figure 3-39: Share Shareof ofwage wage Figure 3-38. Share of wage Figure 3-39. Share of wage employment employment by by sector,gender, sector, by sector, gender,and and employment employment by by sector,gender, sector, gender,and by sector, and Source: World Bank staff calculations, CBPS 2019. employment gender, and employment gender, and self-reported self-reported firm firm size, size,high-exposure high-exposure self-reported self-reported firm firm size, size,low-exposure low-exposure self-reported firm size, high-exposure self-reported firm size, low-exposure areas, 2019 areas, 2019 areas, 2019 areas, 2019 areas, 2019 areas, 2019 Wage workers are more likely to be employed in very small enterprises if they are work- 9% 23% 9% 16% ing in the agricultural sector. Agricultural wage workers in high-exposure areas are 23% 32% 16% 37% 42% 52% 46% 45% 32% 37% 43% 52% 46% 45% 43% 42% more likely than those in low-exposure areas to be working in microenterprises. Wage 71% 71% 86% workers are more likely to be employed in enterprises hiring less than five employees, 91% 86% 77% 91% 84% particularly if they are engaged in agriculture (Figure 3-36). The pattern detected in the 77% 68% 84% 63% 58% 48% 54% 55% 68% 63% 57% 48% 54% 55% 57% 58% 2013 Economic Census and highlighted above—of employment in Cox’s Bazar being domi- 29% 29% 14% 14% nated by small firms—is confirmed in the 2019 CBPS. In the latter, firms with less than five Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female employees account for roughly 60 percent of employment in both low- and high-exposure Agriculture Industry Services Agriculture Industry Services Agriculture Industry Services Agriculture Industry Services areas (Figure 3-37). However, agricultural workers in high-exposure areas are more likely 1 - 5 workers > 5 workers 1 - 5 workers > 5 workers than those in low-exposure areas to be working in microenterprises. On the other hand, 1 - 5 workers > 5 workers 1 - 5 workers > 5 workers non-agricultural wage workers are more likely to be employed in non-micro enterprises in high-exposure compared to low-exposure areas. Source: World Bank staff calculations, CBPS 2019. Source: World Bank staff calculations, CBPS 2019. 112 113 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Migration and remittances than domestic in Cox’s Bazar, and remittances represent a higher share of total income for better-off households, since better-off households are more likely to report remittances, Migration is an important facet of livelihoods in Bangladesh. Outward international migra- especially from international migrants (Figure 3-42). tion from Cox’s Bazar broadly reflects national patterns. International migrants represent 3.8 percent of the population of Cox’s Bazar, similar to the national average of 3.4 percent Figure 3-40: Monthly international Figure 3-41: Characteristics of Figure 3-40. remittance Monthly flows international to Bangladesh from Figure 3-41. Characteristics households of not that receive or do (Population Census 2011). Cox’s Bazar accounts for 1.7 percent of total Bangladeshi inter- remittance flows to Bangladesh wage workers abroad, 2019-2020 from receive remittances, high- andnot households that receive or do low- national migrants. Ninety-eight percent of international migrants from Cox’s Bazar are male wage workers abroad, 2019-2020 receive remittances, high- and (compared to 96 percent of international migrants in Bangladesh as a whole), 87 percent (millions of US dollars) exposure areas, Cox’s Bazar, 2019 (millions of US dollars) low-exposure areas, Cox’s Bazar, 2019 are between the ages of 15 and 40 (versus 92 percent nationwide), and 94 percent migrate 2600 90% 1.2 for work. 14.6 percent of households in Cox’s Bazar report having an international migrant, 2400 80% higher than the national average of 12.2 percent and the Dhaka division average of 11.7 per- 1.0 70% Dependency ratio 2200 Million US dollar cent, but lower than the Chittagong division-level figure of 26.2 percent. While Bangladeshi Share of Heads 60% 0.8 2000 households sending international migrants have on average 1.22 migrants, in Cox’s Bazar, 50% 1800 0.6 this number is higher, at 1.35. Most international migrants from Cox’s Bazar have less than 40% 1600 secondary education. Fourteen percent have no education, 30 percent have primary educa- 30% 0.4 1400 tion, and 47 percent have less than secondary education. Compared to the national average 20% 1200 0.2 of 70 percent, 80 percent of international migration from Cox’s Bazar is to Gulf countries, with 10% 1000 0% - 50 percent going to Saudi Arabia and 25 percent to the United Arab Emirates. No remittances Remittances No remittances Remittances Jan-19 Apr-19 Jul-19 Oct-19 Jan-20 Apr-20 Jul-20 Oct-20 To date, Cox’s Bazar has received little domestic migration. Internal or domestic migra- tion is proxied by the share of the population in a district that reports being born elsewhere 2020 In million US dollar on the 2011 census. Only 1 percent of the Bangladeshi population in Cox’s Bazar was not High Low born in the district, which is lower than the national average. Within Bangladesh, 8 percent Head is female Literacy of Head of the population was born in a district other than the one in which they were counted Head work in last 7 days Dependancy ratio during the census. These estimates are much larger for populations living in Dhaka district Source: Staff calculation using Bangladesh Central Source: Staff calculations using CBPS 2019. A (50.6 percent) and Chittagong district (11.5 percent), urban centers which attract domestic Bank data. January 2019 to November 2020. household is considered to receive remittances if it migrants. Indeed, Cox’s Bazar is a net sender of domestic migrants based on this measure. declares strictly positive income from remittances. In 2011, an estimated 60,000 Bangladeshis born in Cox’s Bazar were living in other districts, whereas 22,000 Bangladeshis living in Cox’s Bazar were born elsewhere. Eighty-seven per- 3-42: Share Figure 3-42. Figure Share of households receiving of households receiving remittances, income quintilet, by income remittances, by quintile, cent of the population born in Cox’s Bazar that moved to another zila moved within the Cox’s Bazar, 2019 Cox’s Bazar, 2019 Chittagong division, and 10 percent moved to Dhaka division. 35% 30% International remittances represented 5.7 percent of Bangladesh’s GDP in 2018 (WDI). In 25% the context of COVID-19, the amount of remittances flowing into the country has fluctu- 20% ated heavily since January 2020. Figure 3-38 traces this pattern. Analysis of the most recent 15% household survey shows that 17 percent of households at the national level received some 10% type of remittance in 2016 and suggests that remittances are more likely to go to well-off 5% 0% households (Hill and Genoni 2019). Total Q1 Q2 Q3 Q4 Q5 Total Q1 Q2 Q3 Q4 Q5 Total Q1 Q2 Q3 Q4 Q5 Remittances Domestic International In 2019, remittances represented on average 11 and 15 percent of total household income High spillover Low spillover for households in high- and low-exposure areas within Cox’s Bazar, respectively. In addition, households receiving remittances are more likely to be female-headed and have Source: Staff calculations using CBPS 2019. A household is considered to receive remittances if it declares strictly fewer members (Figure 3-41). Finally, more households received international remittances positive income from remittances. 114 115 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES The COVID-19-related economic slowdown in Cox’s Bazar prior to the survey.  Specifically, 53 percent of high-exposure hosts and 67 percent of low-exposure hosts who reported being employed in the seven days prior to the survey The spread of the novel coronavirus poses substantial health and economic threats were temporarily absent from work, that is, not actively working. Unsurprisingly, an over- to Bangladesh. COVID-19 poses grave health and welfare risks in South Asia, a region whelming majority of temporarily inactive workers attributed the situation to COVID-19 characterized by weak health systems, high population density, reliance on insecure, work restrictions, with the highest prevalence of COVID-19 restrictions being reported in informal livelihoods, and limited safety net systems. These risks are further exacerbated high-exposure hosts (97 percent), compared to 87 percent in low-exposure hosts. On the in Bangladesh, one of the most densely populated countries in the world, particularly in other end of the spectrum, 85 percent of those who reported not working since January its large urban concentrations, such as Dhaka and Chittagong, and in areas of heightened were not employed during the baseline survey, either. That is, these are largely hosts who local density, such as Teknaf and Ukhia. The government of Bangladesh initiated coun- are not participants of the labor force. Taken together, these results indicate that tempo- try-wide lockdowns on March 26, 2020, and mobility restrictions were also imposed in 2021 rary job losses were attributable to the COVID-19-related slowdown in economic activity. following the second wave. Among those who reported being employed amid the lockdown, a higher proportion To monitor the evolving labor-market and economic impacts of pandemic-related lock- of women are seen to be actively working, meaning that the high rates of temporary downs, phone monitoring surveys have been rolled out in Dhaka, Chittagong, and Cox’s absence from work are driven by men. This could be explained by the nature of the jobs Bazar. The following paragraphs draw insights from the labor module of the first rapid that these two groups are typically engaged in, with many women performing more home- phone follow-up of the Cox’s Bazar Panel Survey. This follow-up survey was conducted in based work less likely to be affected by lockdowns, while men are likelier to be engaged in April and May 2020 (World Bank 2020c). It engaged a representative sub-sample of 3,005 service-sector activities such as transportation and trade. households out of the 5,020 surveyed at baseline. It was designed to track key factors and trends in current labor force participation, employment, unemployment, and income, in People temporarily absent from work were largely wage-based day laborers in agricul- comparison to baseline scenarios. The study aims to provide insights on the impact of the ture and services. Among this population, temporary work suspensions may easily lead ongoing crisis on the current labor market among host and Rohingya populations in Cox’s to permanent job losses. Sixty percent of non-wage workers, as opposed to 67 percent of Bazar. The labor module considers three employment periods, assuming that the country wage workers, reported being temporarily absent from work in the seven days prior to the started dealing with the impacts of the COVID-19 crisis from early March 2020. The labor survey. The higher rates of absence for wage workers were driven by low-exposure hosts, module allows for the identification of changes in work status and incomes from January among whom 70 percent of wage workers were unable to work, largely due to COVID-19- to March 2020, in early-mid March, and during the survey recall period (seven days prior to related restrictions. For high-exposure hosts, temporary absences among wage and non- survey dates falling in late April to mid-May 2020).79      wage workers were relatively more balanced. These trends could be explained by the sec- toral employment shares in high- and low-exposure regions, the former being more reliant As a consequence of the lockdowns and social distancing measures, unemployment on agriculture than the latter. Low-exposure areas are more reliant on non-agricultural jobs increased among Bangaldeshi households in Cox’s Bazar district. This was reflected in an in wholesale and retail trade, construction, and transportation industries, which were more increase in the share of individuals actively seeking work, and was accompanied by declin- severely affected by the crisis. Close to three-fourths (72 percent) of temporarily absent ing employment, particularly for women in low-exposure areas, which are more urbanized wage workers, across high- and low-exposure areas, reported being paid on a daily or and include Cox’s Bazar Sadar upazila. This increase in unemployment (and labor force par- weekly basis, implying broad engagement in informal sector jobs as day laborers who are ticipation) largely reflected previously non-participating women and secondary household likely not to be paid during their absences from work. Given that most jobs are informal, members entering the labor force, likely in search of additional sources of household income, many of these temporary absences may well translate into permanent job losses.  particularly in more urbanized parts of the district. The lockdown and its impacts have taken a differential toll on wage and non-wage work- Among those who reported themselves as still being employed, more than half had ers. Among non-wage workers and enterprise owners, a somewhat smaller share reported been temporarily absent from work, that is, employed but absent in the seven days being temporarily absent, but income losses since March 2020 were reported more widely in this group. Ninety-eight percent of these non-wage workers report running enterprises with five or fewer employees, which is a potential indicator of how the lockdown impacted 79 The findings from the follow-up are presented as a cross-sectional update on baseline adults. micro-enterprises. New data from recently completed phone surveys will provide addi- Panel comparisons on employment transitions for adults who have been part of both surveys also tional insights on how the economic shock associated with the pandemic has affected demonstrate consistent trends. For more recent results from COVID monitoring surveys, please refer to https://www.worldbank.org/en/country/bangladesh/brief/cox-s-bazar-panel-survey-briefs. household-based micro enterprises and wage workers in the country. 116 117 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Work and livelihoods among the recently displaced Rohingya population 3-43: Labor market indicators, Figure 3-43. Figure 3-44: Figure 3-44. Share of employment by Rohingya population in camps, 2019 sector, Rohingya sector, men and Rohingya men women in and women in Given current restrictions on work, the Rohingya people in camps have low labor force 2019 camps, 2019 camps, 2019 participation and are more reliant on informal work. Only 1 in 3 displaced individuals 78% were participating in the labor force (that is, were employed or actively seeking work in 38% 64% 64% 2019 (Figure 3-43). The majority of the workforce was comprised of men. While 64 percent 62% 33% of working-age men participated in the labor force, labor force participation among women 27% 27% was only 8 percent. Among the few Rohingya who are employed, the majority work in infor- 38% mal jobs, as non-agricultural wage labor and informal workers for independent employers. 36% 33% 16% Types of work differ by gender, with men employed in non-agricultural wage-labor jobs and 14% 22% 13% women in self-run, small-scale homebased activities (World Bank 2019a). Most recently 9% displaced Rohingya who work receive daily wages, at higher rates than Bangaldeshi work- 7% 6% 9% 4% 3% ers. Most employed Rohingya reported working for NGOs, 83 percent of employed men and 2% 0% 61 percent of employed women. All camp Male Female at e & re Se ons ties g n al n, t rs or n C ili in tio he he io tu residents od d Tr th io sp ut ur m Tra t ice ruc ul Ot ca & act an Among the few Rohingya who are employed, sectors of activity are differentiated by gen- ric du t uf Ag ,e an m der. While 88 percent of employment in camps is in non-agricultural activities, Rohingya M co rv ac women are 3 times more likely than men to be involved in farm activities (Figure 3-44). LFP Employment (as % of LFP) Within non-agricultural activities, men are mainly working in construction, as earth work- Unemployment (as % of LFP) Camp male Camp female ers (non-government), construction workers, masons, and other miscellaneous non-ag- Source: Staff calculations, 2019 CBPS. Note: LFP = labor force participation. ricultural day laborers. On the other hand, women are most likely to be employed in home-based manufacturing and education, for example, as tailors or teachers in camps. Both genders report working in health and social volunteering work, with women taking a Table 3-10: Education and sector of employment, recently displaced slightly higher share of these jobs than men. In addition, working Rohingya men are much Rohingya, 2019 more likely to be waged employees (79 percent) relative to women, who are more likely to be self-employed (60 percent). This disparity in wage and self-employment could be Never explained by the fact that Rohingya men and women are engaged in very different activi- attended Less than Complete Incomplete Complete ties (World Bank 2019a). school primary primary secondary secondary Agriculture 17% 10% 10% 1% 1% The types of activities in which the displaced Rohingya are employed, together with their low human-capital endowment and work restrictions, limit the potential for competi- Trade & 10% 11% 15% 18% 0% Accommodation tion with the host population in the labor market. Conditional on the overall low levels of education among the displaced Rohingya population, and their low rates of employment, a Manufacturing & 11% 14% 13% 8% 4% Utilities large share of less-educated, employed Rohingya work in construction, whereas more edu- cated Rohingya are primarily employed in health and education services in camps (Table Construction 36% 36% 42% 21% 2% 3-10). The low engagement of displaced individuals in agriculture, compared with hosts, Transport 3% 1% 0% 2% 0% the reliance on camp-based labor-intensive public works, and the dependence on NGO- related employment for the few educated Rohingya all suggest low labor-market friction Service, Education, 10% 12% 14% 41% 93% Health with the local host population. Others 13% 17% 6% 9% 0% Source: World Bank staff calculations, CBPS 2019. 118 119 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 3 – E C ONOM I C OUT C OMES : J OBS , L I VE L I HOO D S , AN D I N C OMES Given restrictions on work, the prevalence of secondary employment among Rohingya Reliance on humanitarian aid and limited opportunities for employment have shaped in camps is low, at 15 percent. Similar to hosts, Rohingya men are more engaged in second- the earned-income structure of displaced populations. While 53% of income sources come ary activities than women (16 versus 9 percent, respectively). Among these secondary jobs, from wages, 22 percent of their income stems from other sources, including assistance pro- most men work for wages across all sectors, while women primarily report being self-em- grams (Table 3-11). Remittances represent 10 percent of total household income for Rohingya ployed in agriculture. As with primary employment, better-educated Rohingya individuals households, on average. Within labor income, male wage workers enjoy higher weekly earn- report secondary work in health and education services. ings than their female compatriots. Similar to hosts, while weekly earnings are higher in pri- mary jobs, secondary employment pays more on an hourly basis, suggesting the presence of Despite the restrictions on movement and work, there is emerging evidence of some temporary or seasonal opportunities to earn additional income. dynamism within the camp economy, and of growing business opportunities around the camps. Filipski et al. (2020) found that the Rohingya have access to a range of active busi- Rohingya depend heavily on aid for their livelihoods. One hundred percent of Rohingya nesses within camps. Hosts and the Rohingya both operate businesses within the camps. households report receiving some kind of aid in the last year (99.92 percent in the last This suggests that settlement economies spring up not only through entrepreneurial drive, month). Under food aid, 100 percent of Rohingya report receiving rice, followed by oil but also when locals identify business opportunities and fill the demand. Regardless of (99 percent) and pulses (98 percent). Ninety-eight percent of Rohingya households also their location, the majority of the businesses were classified within the wholesale and retail report accessing health assistance. Various actors distribute aid, with WFP the most widely trade category, but the type of enterprises run by hosts and Rohingya tend to differ. While cited. Part of this aid is not directly used by beneficiaries; at least 21 percent of households trade and manufacturing businesses are mainly run by Rohingya, some activities inside report bartering some portion of the assistance received in the previous month. Most such camps, such as transportation businesses, are run predominantly by Bangladeshis. Despite items are exchanged for cash, and these transactions occur with other Rohingya and host opportunities for small businesses, Rohingya-run enterprises face greater constraints than communities alike. A rapid classification of other income sources80 reported in Table 3-11 their local counterparts: Rohingya-run businesses are smaller and less profitable, and reveals that, for Rohingya households reporting other sources of income, 87 percent are Rohingya workers are paid lower wages than host workers. Moreover, lending plays an related to or include some type of transaction with food aid. important role in sustaining these businesses. The transition to WFP’s SCOPE value-voucher modality, which allows for more dietary Table 3-11: Income sources and average incomes, recently displaced Rohingya diversity (20 items, of which 12 are fixed and 8 flexible) was underway during the CBPS in Cox’s Bazar, 2019 baseline survey period. By March 2020, shortly before government COVID-19 lockdowns were initiated, 72 percent of the Rohingya population had transitioned to value-voucher Household income sources Share of Rohingya HH modality. In addition, WFP, in collaboration with Relief International, had piloted a farm- er’s market in select camps. The aim was to provide greater access to a variety of foods, Wages 53% while allowing small host-community farmers to sell their produce directly in camps as an Income from cultivation 3% extension of the aid delivery system. Public-health measures related to COVID-19 disrupted Income from livestock/fishing/forestry 1% this experiment, however. From March 26 onwards, accessory operations such as farmers’ markets were halted, and camps shifted to an essential-operation-only modality, with all Income from non-ag business earnings 6% camp residents now reverting to receiving commodity vouchers: a fixed food basket with Remittances 10% consideration to broad food preferences and nutritional value. Asset earnings 0% The shift in modality for food support led to deteriorating perceptions about food Pensions 0% assistance during the pandemic. Ninety-six percent of camp households reported get- Cash assistance from government 4% ting food assistance from WFP in March 2020, but more than half of respondents reported receiving “less food” than usual. This perception of less food than before may have been Other 22% Average per capita income 910 80 In CBPS 2019, when a household reported other sources of income, a brief explanation of the source Average income 4,254 could be provided. Forty-eight percent of Rohingya households reporting income indicated receiving Source: World Bank staff calculation using CBPS 2019. income from other sources. A rapid classification of cited sources revealed that 87 percent involved Note: Average income in Takas. All indicators calculated using only households reporting income. HH = households. activities related to food-aid bartering or selling. 120 121 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC driven by the shift of delivery modality. A fixed basket of food may seem more restric- CHAPTER 4. tive, despite containing the same total monetary and caloric value of food entitlement received in the value-voucher modality, which provides more flexibility in basket compo- sition. Despite restrictions in the modality of food assistance delivery, there is evidence Accelerating inclusive of strengthened provision of WASH services to mitigate the potential spread of the coro- navirus. Hygiene and sanitation assistance mechanisms have clearly been enhanced in growth: Constraints and camps in response to the crisis, with 13 percent of households reporting receiving more services than usual. opportunities Post-baseline evidence from April-May 2020 suggests that employment among dis- placed Rohingya had declined significantly in that period, accompanied by a sharp rise in unemployment. Results from the rapid phone follow-up to the 2019 CBPS base- line suggest that employment dropped from 64 percent in 2019 to 23 percent in April- May 2020, while unemployment increased sharply, from 36 percent to 77 percent. Labor force participation increased in camps, mainly due to the rise in unemployment and fall in employment. The rising trend in unemployment was mainly driven by men, who are highly dependent on wage labor. This chapter uses the results from previous sections to analyze barriers to inclusive growth in Cox’s Bazar and locate leverage points for progress. Currently, several features of mar- However, these recently identified changes in the labor market cannot be entirely kets and institutions in Cox’s Bazar hold local people back from reaching their productive attributed to COVID-19. More than half of the male camp population of working age reported potential, make it harder for businesses to grow, slowing down the development of the dis- not having worked since January 2020, suggesting that the trend was driven by pre-COVID-19 trict economy overall. This chapter diagnoses key obstacles to equitable growth and shows factors. Among such factors is a September 2019 (post-baseline) government directive that opportunities exist to tackle them. It analyzes local comparative economic advantages banning cash-for-work programs in camps (World Bank 2020c). The restricted operational of Cox’s Bazar sub-districts and across economic sectors such as fishing and aquaculture, modality adopted by the camps in response to COVID-19 has also reduced the work gen- tourism, and manufacturing. It looks at how strategic improvements in connectivity, infra- erated inside the camps, which had previously been the main source of earned income for structure, governance, and service delivery may accelerate inclusive growth and provides Rohingyas. Since March 25, 2020, all non-critical camp operations have been suspended evidence that policy can leverage the ongoing humanitarian response in Cox’s Bazar to or reprogrammed, including the complementary food voucher scheme, farmers’ markets, unlock fresh economic opportunities for Rohingya people and host communities. self-reliance support, livelihood support, and shelter/non-food items activities – many of which had provided earning opportunities for the Rohingya population in camps. Targeting constraints to inclusive growth The chapter begins by highlighting select factors identified in the preceding analysis that constitute key barriers to inclusive development in Cox’s Bazar. The discussion foregrounds factors based on three considerations: (1) their substantial negative impact on fundamental conditions for inclusive growth is documented; (2) they are potentially actionable through well-understood, evidence-based policy options; (3) important steps to address them can be achieved in a short timeframe (1-3 years), laying foundations for lon- ger-term gains. The chapter discusses barriers under three headings: constraints to human capital accumulation, specifically education human capital; constraints to local people’s productive inclusion in labor markets, with special focus on women; and constraints to private-sector activity and entrepreneurship. 122 123 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Constraints to human capital accumulation81 Adolescents aged 16-18 are at the highest risk of dropping out of school in Cox’s Bazar. Two-thirds of all students who abandon their education drop out of school between the With a large share of illiterate adults and a weak education system, Cox’s Bazar remains ages of 16 and 18, with most other dropouts observed in the 12-15 age group. This pattern poor in human capital. This is an important reason the local economy continues to heav- holds across low- and high-exposure areas and for boys and girls.84 ily rely on low-productivity agriculture and services. Improving its human capital is cru- cial, as the district moves to leverage its promising geographic and economic endowments Girls’ reasons for dropping out of school vary across areas and by age. Social restric- for tourism, hospitality, aquaculture, and other fields that can power inclusive growth. It is tions and marriage are powerful drivers. While family and social restrictions play a vital to tackle economic constraints that make it difficult for low-income families to finance larger role in high-exposure areas relative to low-exposure areas (24 versus 15 percent education expenses. Doing so will expand educational opportunities—and future work respectively), marriage is more important in explaining female drop out in areas further options—for both females and males. away from camps (10 versus 24 percent in high- and low-exposure areas). These con- straints influence education decisions at different ages. While marriage can become a Financial pressures and social norms are the major constraints that keep Bangladeshi constraint for some girls as early as 15, this barrier is most relevant for 18-year-old girls children in Cox’s Bazar from attending school. Once children begin their education, (>70 percent of girls reporting this constraint are 18 years old). On the other hand, social these are also the main reasons they drop out. Gendered social norms strongly constrain norms start to discourage girls from pursuing their studies at early ages85 and remain girls’ educational opportunities. Only 5 percent of children aged 6 to 18 in Cox’s Bazar important thereafter (Table 4-2).86 have never attended school. However, about 1 in 5 school-aged children in high-exposure areas and 15 percent in low-exposure areas drop out.82 Among current school-aged chil- Table 4-2: Reasons for dropping out of school, by age and gender, high-exposure dren, the cost of education is the main constraint for boys and girls across areas. About 50 versus low-exposure areas, 2019 percent of children who drop out do so because their families are unable to bear the costs of schooling. Meanwhile, 1 out of 3 women who never went to school report that they were 6-18 years constrained by social restrictions, family pressures, or because of marriage (Table 4-1).83 High-exposure Low-exposure Male Female Male Female Table 4-1: Reasons for never attending school, high-exposure and low-exposure No money/too expensive 49% 48% 45% 43% areas, Cox’s Bazar, 2019 Family/social restrictions 2% 24% 2% 15% High-exposure Low-exposure Male Female Male Female For marriage 0% 10% 0% 24% Lack of money/food/need to work/help with 73% 48% 76% 48% Do not want to study more/completed 31% 8% 27% 9% family chores studies Disability/illness 2% 2% 0% 1% Must work/family chores 12% 3% 17% 6% Family/social restrictions/marriage 3% 33% 4% 34% No schools close to home 6% 7% 4% 6% Other 6% 7% 9% 3% Age (too old/too young) 8% 5% 8% 8% No need/no interest to study 5% 2% 5% 2% Other 2% 3% 3% 1% 84 The only exception is among girls living closer to camps, among whom 40 percent of dropouts occur between the ages of 12 and 15. Source: World Bank staff calculations using CBPS 2019. 85 The youngest girls in high- and low-exposure areas reporting this constraint were 11 and 12 years old, respectively. 81 Given data limitations, this sub-section focuses only on constraints to education. 86 These results are consistent with USAID (2018). In interviews with key informants and focus group 82 Similarly, 40 percent of adults (over 18) in high-exposure areas and 32 percent in low-exposure areas discussions with members of the community, USAID researchers found that livelihoods/earnings and never attended school, while 55 percent in high-exposure areas and 61 percent in low-exposure areas household budget were the main constraint for students who dropped out. The study also found that started school but dropped out. some students who abandoned school were attracted by comparatively high salaries in camp-re- 83 These gendered patterns are even more pronounced among older cohorts (aged 18+) who attended lated occupations. In the case of girls, the study mentions that early marriage might be related to the school but subsequently dropped out. spouse’s ability to achieve financial independence and solvency. 124 125 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Older than 18 among girls and in low-exposure areas.88 This may indicate that access to education is not the main problem, but that the additional costs associated with schooling—expenditures High-exposure Low-exposure for books, uniforms, and tutoring, for example—are too much for some households to Male Female Male Female bear. Eighty percent of household out-of-pocket educational expenditures are for expenses Family/social restrictions 3% 28% 4% 21% other than school fees (Box 4). No money/too expensive 46% 27% 40% 24% For marriage 1% 23% 2% 37% Do not want to study more/completed 29% 11% 29% 11% studies Box 4: Household education expenditure in Cox’s Bazar Must work/family chores 14% 6% 21% 4% Others 7% 6% 5% 3% The growth in incomes and consumption in the last two decades in Bangladesh Source: World Bank staff calculations using CBPS 2019. has been accompanied by an increase in the share of education expenditures Note: The number of observations were as follows. For high-exposure areas: Ages 6-18: males 246; females 272. Older in household budgets. In 2016, the share of education expenses in total con- than 18: males 898; females 793. For low-exposure areas: Ages 6-18: males 202; females:142. Older than 18: males 950; females 944. sumption had almost doubled with respect to 2000 (7.7 versus 4.3 percent, respectively) (Bhatta et al. 2019). By this measure, Cox’s Bazar performs above The current employment status of school dropouts is consistent with the main reasons the national average, with about 7 out of 10 households reporting expenditure reported for leaving education. While only 10 percent of girls who drop out of school are on education in 2016. Furthermore, Cox’s Bazar is among the 10 Bangladeshi employed, about 7 out of 10 boys who drop out are currently working. This suggests that districts with the highest share of households reporting education expendi- the returns to education net of the out-of-pocket costs for these children must be lower than ture (HIES 2016). However, despite having a relatively large share of house- the alternative wages and earnings, or that the need to help support their families in the holds reporting education expenditure, in Cox’s Bazar, these expenses’ share immediate-term overwhelmed the potential future benefit of continuing their education. in the average household’s total consumption is relatively lower than in other Chittagong division districts and the national average. In absolute terms, the median household in Cox’s Bazar spends 5 and 16 percent less than the median Economic constraints matter differentially in explaining school dropouts across the income household at national and division level, respectively. In 2016, while the median distribution and among age cohorts. More than half of children who dropped out between household in Bangladesh spent Tk. 802 per month on education (about Tk. 516 ages 6 and 18 in the bottom 40 percent of the income distribution in areas close to camps per student), in Cox’s Bazar this amount was Tk. 764 per month (about Tk. 384 reported that their families were unable to cover education costs (63 percent among girls per student). Furthermore, poor households still have substantially lower pri- and 57 percent among boys). This share is higher than among children from more affluent vate spending on education than richer households, though the gap in expen- households. In low-exposure areas, the situation is similar, but the intensity of this constraint diture per student is smaller in Cox’s Bazar. The lower spending of the poor also is slightly lower among girls. When comparing current school-aged children with cohorts of translates into a lower education share in their total budget (Bhatta et al. 2019). adults (older than 18), it appears that financial considerations are more important in explain- ing school dropouts among younger cohorts than they were for older cohorts.87 In the last 16 years, expenditure by level of education at national level has increased for all levels of education (Bhatta et al. 2019). Households in Cox’s While education costs clearly emerge as the main barrier for human capital accumu- Bazar are spending, by level of education, about the same as median national lation, most students who dropped out were enrolled in government-run educational and division households. Indeed, while expenditures at national level are institutions. This underscores the importance of costs other than school fees as barriers 300 Tk, 843 Tk, and 1338 Tk for primary, secondary, and tertiary education, to educational opportunity. Across the income distribution, boys were more likely to be in the district, the median household was spending 331 Tk, 836 Tk, and 1475 enrolled in government-run schools than girls, which may indicate a revealed preference Tk, respectively (Table B4-1). Similar to typical households at national level, among households that private schools are more appropriate for girls’ education. Private- in Cox’s Bazar, 20 percent of education expenditure goes to cover fees, but school attendance systematically increases in higher income quintiles, and it is greater See Table A1-14, Table A1-15 and Table A1-16 in Annex 1 for a detailed breakdown of constraints to 87 human capital accumulation across the income distribution. 88 See Table A1-17. 126 127 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Effective access to education for Rohingya children has improved but remains severely limited. The 2020 GoB decision to allow for education using the Myanmar curriculum this share increases as one moves upward in the education distribution. The may improve the quality of education inputs for Rohingya children, but COVID-19-related remaining 80 percent of the expenditure is to cover books, tutoring, transport, restrictions have delayed implementation of the plan. As previously discussed, a large- and other costs, such as those for uniforms, footwear, hostel, tiffin, internet/e- scale 2019 assessment of educational programming in Rohingya camps noted positive mail, schooling donation, and others (Table B4-2). trends, with more learning centers being built and improvements in staffing (Pascaud and Panlilio 2019). However, without an adequately structured curriculum, most adolescents Table B4-1: Median expenditure on education by quintile, Cox’s Bazar, Chittagong are still left out of the system. While the transition to the Myanmar curriculum will be a wel- division, and Bangladesh, 2016 come step, additional efforts will be needed to enroll and keep children in school, ensure minimum quality standards, and provide some form of educational certification. Median expenditure Median expenditure per student Ongoing policy dialogue and operations show positive momentum and offer opportu- Chittagong Cox’s Chittagong Cox’s Bangladesh Bangladesh nities to address shortfalls in education services for the Rohingya. However, substantial Division Bazar Division Bazar challenges remain. The main challenges for ongoing operations are clear. The NGO-led 1 315 323 326 202 190 180 delivery of education has resulted in a fragmented approach, offering services that remain suboptimal in scale and quality. The language of instruction is English, and Rohingya chil- 2 548 651 649 362 387 291 dren have not yet been able to access Bangla or Myanmar curricula, although rollout of the Myanmar curriculum is expected soon. Teacher capacities are limited and access to tech- 3 773 841 785 509 506 368 nology is restricted. So far, instruction has remained non-formal, with limited sequencing between years and no recognition of studies or accreditation of providers. 4 1,127 1,255 1,036 725 713 563 Constraints to productive inclusion in the labor market 5 1,933 1,755 1,099 1,310 1,143 907 Total 802 911 764 516 554 384 In Cox’s Bazar, productive and remunerative labor market participation for both men and women is constrained by low educational attainment, limited access to export-ori- Source: HIES 2016. ented, labor-intensive manufacturing jobs that have so far fueled growth in Bangladesh, and physical distance from the country’s growth poles. The reliance on subsistence agri- Table B4-2: Components of educational expenditure, by education level, Cox’s culture and low-value, informal service-sector jobs in Cox’s Bazar reflect the district’s low Bazar, 2016 productive, human-capital base and the lack of alternative employment opportunities. Fees Books Tutoring Transport Others Data from the 2019 CBPS shows that only 60 percent of host adults in Cox’s Bazar can read, one-third of the adult population has no schooling at all, and an additional 25 Total 21% 19% 15% 3% 43% percent of adults have only some primary education. Among adults ages 20-29, men in Cox’s Bazar show far lower levels of educational attainment than men in the same age Primary 14% 20% 7% 2% 57% group nationally. Taken together, these patterns suggest that almost 60 percent of the Secondary 22% 20% 21% 3% 35% adult population in host communities cannot access any type of skilled employment. Educational attainment levels are lower for host communities in areas of high exposure Tertiary 36% 20% 12% 6% 25% to Rohingya: 38 percent of adults in high-exposure areas never attended school, com- pared to 32 percent in low-exposure areas. Only 52 percent of adults in high-exposure Source: HIES 2016. areas can read, compared to 62 percent in low-exposure areas. The share of adults who received some secondary schooling is 10 percentage points higher in low-exposure areas than in high-exposure areas. 128 129 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES These constraints are further compounded for women and are evident in their low levels Constraints to private sector activity and entrepreneurship of labor force participation, particularly in market-oriented activities. Additional barriers affecting women include differential access to productive inputs and assets compared to The private sector in Bangladesh faces a challenging and deteriorating business envi- men; women’s role in home-based and caretaking activities; market failures and institu- ronment. The country’s ranking in Doing Business has fallen from 65th in 2016 to 168th out tions; and social norms constraining women’s mobility (Genoni et al. 2021). Within the host of 190 countries in 2020. This situation is reflected in other assessments of the regulatory community in Cox’s Bazar, low overall educational attainment is accompanied by signifi- environment, such as the World Economic Forum’s 2019 Global Competitiveness Index, cant gender gaps. Thirty-seven percent of adult women have no schooling, compared to 29 where Bangladesh ranks 109th of 141 countries on the “institutions” pillar. The business percent of men. Men are almost twice as likely as women to complete secondary school, environment favors established, connected firms and sectors, and disadvantages new while most women who do attend school drop out during secondary school. In particular, entrants, including young, small establishments and investors trying to expand or start women in high-exposure areas have poorer educational attainment on average than those their business. in low-exposure areas. A complex licensing environment makes it difficult to start or to close a business, as Women’s potential to generate incomes and engage in productive, paid work outside the documented in a recent IFC report (IFC 2020). Starting a business requires investors to home and the farm is further constrained by prevailing norms around asset ownership, navigate a complicated process involving more than 150 services from 34 line agen- home- and care-related responsibilities, and mobility (Anderson and Eswaran 2009). cies. An insufficient insolvency framework makes it difficult for firms to close down and Restrictions on women’s ability to inherit property inhibit their ability to start businesses creditors to collect on debts. Bangladesh performs very poorly in terms of the ability to and access credit for expansion, due to a lack of collateral. The expansion of microcre- enforce contracts, ranking second to last in this dimension on the Doing Business indi- dit finance to women has partially eased this constraint. This has increased female eco- cators. It takes four years on average to resolve a contract dispute in Bangladesh, and it nomic engagement in livestock, poultry, and small textiles manufacturing. Some women is estimated that associated costs make up two-thirds of the claim value. The difficulties also enjoy increased avenues of employment with NGOs as health workers and teachers of enforcing contracts may explain why large companies prefer to keep their operations (Raihan and Bidisha 2018). vertically integrated. According to CBPS 2019 data, 27.3 and 16.5 percent of women are participating in the Broadly speaking, in Cox’s Bazar, firms are constrained by four factors: (i) firm capabil- labor force in high- and low-exposure areas, respectively. Women living in low-exposure ities, (ii) access to finance, (iii) access to markets, and (iv) business environment and the areas are more likely than those in high-exposure areas not to participate in the labor mar- lack of a level playing field. Identification of these constraints is imperative to channel ket because of household responsibilities (84 versus 73 percent of women not in the labor interventions for the development of private-sector enterprises. force in low- and high-exposure areas, respectively), although this is the most frequently cited reason across both areas. Among women who are not in the labor force in high-ex- Using establishment data from HIES, Figure 4-1 plots the percentage of non-agricultural posure areas, a quarter report that this is because of social norms and family objections. firms that report having faced a constraint that affects their business performance, from among credit, technology and costs, raw materials, government regulations, and lack of Evidence suggests that easing constraints on Bangladeshi women’s decision making and customers. Insufficient finance is a key obstacle to firm growth (Malhotra et al. 2007), and expanding their control over assets and earnings in the livestock and aquaculture value it has been found that small firms face bigger challenges in obtaining finance compared chains can help close gender gaps in economic participation. For instance, women’s role to larger firms (Schiffer and di Mauro 2001; Beck et al. 2002).89 in livestock tends to be focused on home-based activities such as feeding and milking cows, raising small ruminants, and raising backyard poultry. Women’s role in marketing, Bangladeshi firms, and Cox’s Bazar firms in particular, feel most constrained by lack of their ability to access earnings and make decisions about these businesses are severely credit. Around 60 percent of firms in Cox’s Bazar report credit to be the major impediment constrained by limited voice, agency, and mobility; poor access to inputs and credit; and to business. The figure is around 40 percent for firms in Chittagong and Bangladesh. At lack of business skills (World Bank 2018b). Similarly, women’s participation in farming and fishery is concentrated in casual, unpaid work in the lower production segments of value chains, and their ability to participate in marketing and business management is similarly 89 Financing is important for firms because it helps in expansion of operations, innovation, and invest- constrained by norms and lack of control over assets and incomes (Shelly and D’Costa ing in production facilities and new staff (OECD 2006). However, many firms that are willing to expand find it difficult to obtain credit from financial institutions. This essentially constitutes the “financing 2001; World Bank 2018a). gap” faced by firms. This gap is more prevalent in developing countries than in advanced economies, where banks have developed various risk-management strategies for lending to firms (OECD 2006). 130 131 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES the national level, there is an estimated gap of US$2.8 billion between the financing needs they face. Data from the economic census suggests that only 1 percent of Cox’s Bazar of small and medium enterprises (SMEs) and the funds available to them. Moreover, the firms use information technology (IT) in their daily operations. Across Bangladesh, the scarce finance that is available to SMEs is offered at higher rates of interest than credit to share of firms using IT in daily operations was also low, about 4 percent. larger firms. These constraints are even more binding for female-owned and managed small enterprises. More generally, firms lack access to long-term finance, relying heavily on Figure 4-2: Uses and sources of finance for business: Bangladesh, Chittagong, lending from commercial banks, which is in turn constrained by banks’ own dependence Figure and 4-2. Cox’s : Uses and sources of finance for business: Bangladesh, Bazar on short-term deposits. As the recent Bangladesh Private Sector Diagnostic (IFC 2020) Chittagong, and Cox’s Bazar notes, other sources of long-term finance such as venture capital, private equity, and fin- 100% 100% tech remain significantly underdeveloped in Bangladesh. 90% 90% 4-1: :Key Figure 4-1. Figure constraintsfaced Keyconstraints faced by The challenges in accessing credit for 80% 80% by non-agricultural enterprises: non-agricultural enterprises: business are evident in Cox’s Bazar. Panel Bangladesh, Chittagong, and Cox’s Chittagong, and Cox’s A of Figure 4-2 shows the amount of capi- 70% 70% Bangladesh, Proportion of Firms Proportion of Firms 2016 Bazar, 2016 Bazar, tal used by businesses in Cox’s Bazar. More 60% 60% than 80 percent of firms have capital assets 0.6% 50% 50% of only between 600 and 6,000 USD when starting a business. One must also consider 40% 40% 0.5% the sources of finance that entrepreneurs 30% 30% Proportion of Firms 0.4% use (Panel B). More than 80 percent of firms in Bangladesh report that they use 20% 20% 0.3% their own sources of finance; the same is 10% 10% 0.2% true for about 90 percent of firms in Cox’s Bazar. The next available source of finance 0% 0% 0.1% is from entrepreneurs’ relatives, an infor- Up to 500,000 to More than 5 to 10 s t s es rs di ce er he tiv 500,000 5 million 10 million million re nd ur Ot la oc so le mal source. This suggests that there are BDT BDT BDT BDT Re icr re al 0% M rm n Ow significant credit market frictions affecting Fo it st & ia w re rs ed er a vt e co gy s ls g at r r firm performance. It is not clear whether go om Panel A: Use of capital assets Panel B: Sources of finance Cr lo m the no & st Cu O these frictions stem primarily from the sup- ch Te Bangladesh Chittagong Cox’s Bazar Bangladesh Chittagong Cox’s Bazar ply side (available liquidity in banks) or the demand side (poor quality of firm credit Source: World Bank staff calculations using HIES 2016. applications and insufficient collateral). Source: World Bank staff calculations using the 2013 Economic Census (use of capital assets, left panel) and the 2016 HIES (sources of finance, right panel). Access to and use of technology also appear to be major constraints for firms. While information and communication technology (ICT) penetration has been increasing rap- The use of capital and machines can be important for driving down firm costs, and existing idly in Bangladesh, the infrastructure for digital communications and services remains data suggests very limited use of mechanization and power in production processes: in underdeveloped. The country’s telecom industry has expanded to become the fifth-larg- Bangladesh generally and especially in Cox’s Bazar. Given that firms report costs to be a est mobile market in the Asia-Pacific region, and recent efforts by the Government of key constraint, and that capital is a critical part of firm production process, it is important Bangladesh, such as the Digital Bangladesh program, have helped expand the acces- to assess capital use among firms in Bangladesh. Plotting the share of manufacturing firms sibility and use of mobile and internet technologies. However, continued access and using fuel and/or power versus hand-operated machines in manufacturing processes, Figure quality issues have limited businesses’ ability to leverage digital technologies. This is 4-3 suggests that only 18 percent of such firms in Cox’s Bazar use power and/or fuel machines compounded by the relatively high cost of internet connections, the second highest in in their production. This figure is lower than the Bangladeshi national average, 30 percent, the South Asia region (IFC 2020). Figure 4-1 suggests that around 10 percent of all firms and much lower compared to firms in other developing countries. A large share of firms use operating in Cox’s Bazar report technology to be the second major business constraint only hand-operated machines as the means of producing goods or operating the firm. 132 133 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Figure 4-3: Use of machines among Identifying the appropriate instrument The performance of a firm depends on the education of the owner, and on how the edu- Figure 4-3. Use of machines among manufacturing firms, Cox’s Bazar, to improve firm productivity is inhibited cation system prepares students to be employers of labor.92 Technology adoption and the manufacturing firms, Cox’s Bazar, Chittagong, and Bangladesh, 2013 by lack of appropriate data. There is a financial performance of firms are positively correlated with firm owners’ education levels Chittagong, and Bangladesh, 2013 clear need to increase access to credit and (Barker and Mueller 2002; Farag and Mallin 2018; Kaur and Singh 2018). Better-educated 70% financing for firms in Cox’s Bazar, and to owners take measured risks, create more business ideas, and are well informed regard- promote the adoption of new technology ing their external environment. Analysis suggests that the most educated population in 60% and modernizing the means of production. Bangladesh sorts into wage employment, while entrepreneurship is pursued by those While there is evidence from other countries who cannot find a better wage employment opportunity. The absence of entrepreneurship on the varying efficacy of different finan- in the education curriculum and of private-sector-led programs for apprenticeship and Proportion of manufacturing firms 50% cial-support instruments, identifying the on-the-job training further limit entrepreneurial opportunities. binding constraints in the context of Cox’s 40% Bazar is a prerequisite to testing and rolling In Cox’s Bazar, entrepreneurship appears 4-4: Education level of firm Figure 4-4. out interventions.90 Similarly, the literature to be perceived as an inferior livelihood owners, Bangladesh, Chittagong, and 30% analyzing the effect of public support for choice for educated people. Owners of and Cox’s Bazar, Cox’s Bazar, 20132013 technology adoption on firm performance nearly 65 percent of district firms have is relatively scant, especially when focusing only attained primary or lower second- 70% 20% on developing countries.91 The instruments ary education, compared to 60 percent in 60% to be picked in the case of Cox’s Bazar will Bangladesh and Chittagong (Figure 4-4). 10% 50% Proportion of Firms depend on the base or initial level of tech- Only 20 percent of Cox’s Bazar firms are nology of firms, along with sector context. owned by people who have completed sec- 40% 0% Relatedly, the importance of ICT has to be ondary education, a slightly lower figure 30% aligned with the economic opportunities ne d le er than in Bangladesh and Chittagong. CBPS hi te ab w s ac ra po lic m ope that are present in the district and whether data enables comparison of the educa- 20% pp d an ta nd el No Ha the sectors that will drive growth are really tional background of wage earners versus Fu 10% Bangladesh Chittagong Cox’s Bazar ICT-facilitated sectors. Lack of detailed and entrepreneurs in Cox’s Bazar. Results show 0% recent data on firm performance, produc- that more educated individuals are likely to No Primary & Secondary education lower & above tivity, and value chains hinders the identifi- work as wage employees (Figure 4-5). This secondary Source: World Bank staff calculations using the 2013 Economic Census. cation of specific policy interventions. suggests that entrepreneurship in Cox’s Bangladesh Chittagong Cox’s Bazar Bazar, although it offers higher income, is not seen as an attractive opportunity for Source: World Bank staff calculations using the 2013 the more educated population. Economic Census. 90 New financial-support instruments such as early-stage equity investment; demand-driven financial supports (vouchers and public procurement programs); indirect instruments (fiscal incentives, loan guarantees); inducement instruments; and recognition awards have been used worldwide to help Problems in accessing markets may also be substantial and affect a firm’s performance, overcome firms’ financial constraints. A recent meta-analysis of 16 studies assessed the impact of growth, and survival in the long run. The discussion has so far focused on supply side SME financing programs in developing countries. It found positive effects on capital investments, firm factors such as access to finance, capital, and technology, but it is also well recognized performance, and employment, as well as insignificant effects on profitability and wages (Kersten et that demand shocks can positively influence firm growth (Woodruff 2018). Firms may face al. 2017). Other recent literature has shown that, while experimental evidence on grants documented several constraints when attempting to access product markets, due to the presence of high marginal return to capital within targeted firms, randomized experiments providing loans show a weaker impact on firm performance (Woodruff 2018). high transportation costs and other trade barriers, customers’ lack of information about 91 The promotion of technology adoption has become one of the main policies aimed at enhancing product characteristics (price and quality), and lack of trust in unfamiliar suppliers, among productivity in many countries in the world. There are several instruments to encourage technology other factors. More broadly, business relationships are vital for firm growth. However, these upgrading, including matching grants, thematic funding, guaranteed loans, targeted credit, public procurement programs, and fiscal incentives. While the impact of public support for technological upgrading is positive, its effects on firm performance are not always significant, due to the time hori- For example, see Magoutas et al. (2011) for evidence on the relationship between owner’s education 92 zon under analysis. and firm outcomes in Greek manufacturing. 134 135 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES relationships may not emerge efficiently due to networking frictions, such as lack of infor- Eighty percent of firms in Cox’s Bazar sell locally. The numbers are not significantly dif- mation or trust, implying a network-based barrier. One can gain insights by investigating ferent for either Chittagong and/or Bangladesh as a whole. However, a very small share the main markets for firms in Cox’s Bazar, in particular the extent to which firms sell only (2.5 percent) of firms are also engaged in selling to foreign enterprises. Interestingly, the locally (Figure 4-6, left panel). proportion of firms doing so in Cox’s Bazar is higher than in Chittagong and Bangladesh. HIES household establishments data shows the types of customers to which firms sell their Figure 4-5. What kind of work for the highly skilled? Main job, by workers’ Figure 4-5: What kind of work for the highly skilled? Main job, by workers’ products. This data suggests that an overwhelmingly large percentage of firms (around 75 education levels, Cox’s Bazar, 2019 education levels, Cox’s Bazar, 2019 percent) sell to households or individuals. Around 20 percent of firms sell to domestic enter- prises, whereas around 5 percent sell to foreign enterprises (Figure 4-6, right panel). 80% 70% Proportion of Individuals 60% 50% Identifying opportunities 40% 30% Options exist in Cox’s Bazar to tackle several of the major constraints to inclusive growth 20% that have just been discussed. This subsection presents directions for action based on 10% current evidence, while highlighting knowledge gaps for future research. The subsec- 0% tion first summarizes what is known about local comparative economic advantage in Cox’s No Primary & Secondary No Primary & Secondary Bazar sub-districts, focusing on economic sectors with strong local development potential, education lower & above education lower & above secondary secondary including fishing and aquaculture, tourism, and manufacturing. It then looks at opportu- High Exposure Low Exposure nities to unlock inclusive growth potential in Cox’s Bazar by developing connectivity and infrastructure. Next, it considers existing governance and service delivery capacities in the Wage employee Self employed district and ways to improve them. Finally, it discusses potential economic opportunities associated with the humanitarian response in Cox’s Bazar. Source: World Bank staff calculations, CBPS 2019. Localized comparative advantage Figure 4-6: Firms’ Figure 4-6. Firms’ main markets and main markets and customer types: Bangladesh, customer types: Chittagong, Bangladesh, Chittagong, and Cox’s Bazar Given data gaps, it is difficult to clearly identify productive activities with high poten- and Cox’s Bazar tial to accelerate inclusive growth in Cox’s Bazar. The lack of recent data on the value of Panel A: Markets Panel B: Customer type economic activities for Bangladesh, and especially in Cox’s Bazar district, complicates the 100% 100% identification of the sectors and/or products with the most growth potential. While there is 90% 90% some information on the quantity of industrial and service-sector establishments, volume 80% 80% of employment by activity, and other indicators, 2013 is the latest year of information. At Proportion of Firms Proportion of Firms 70% 70% a more granular level, firm productivity data is also missing, which makes it difficult to 60% 60% understand how well local establishments are performing, the challenges that they face, 50% 50% 40% 40% and possible links to national and global value chains. Overall, this makes it difficult to 30% 30% assess the productivity, comparative advantages, or growth trajectory pertaining to any 20% 20% activity. An enterprise survey is planned in Cox’s Bazar to help fill this critical data gap. 10% 10% 0% 0% While existing data sources do not allow for an analysis of the competitiveness of dif- Local Export Local & Export Household Domestic Foreign firms firms ferent sectors in Cox’s Bazar, there is some suggestive evidence of existing specializa- Bangladesh Chittagong Cox’s Bazar tion in certain types of non-agricultural economic activity in some upazilas. This may indicate localized comparative advantage. In all upazilas, the wholesale and retail trade Source: World Bank staff calculations using the 2013 Economic Census (markets, left panel) and the 2016 HIES sector accounted for the largest number of firms, ranging from 37 percent in Chakaria to (customer type, right panel). 136 137 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES 60 percent in Teknaf (2013 Economic Census). That being said, non-agricultural activity Teknaf, characterized by its disconnectedness from the rest of Cox’s Bazar, its long was relatively concentrated in the northern parts of the district, as highlighted in previous coastline, and its border with Myanmar, did not appear to have any clusters of special- chapters. Chakaria and Cox’s Bazar Sadar alone accounted for 47 percent of the district’s ization in 2013, perhaps because this spatial disconnectedness required the production non-agricultural firms, followed by Ramu and Teknaf, which accounted for another 29 per- of services and manufacturing locally. However, the economies of Teknaf and Ukhia may cent. In contrast, Ukhia and Maheshkhali were home to only 8 percent of non-agricultural well have been profoundly reshaped by the large influx of humanitarian assistance being firms each, while Pekua and Kutubdia had the smallest shares, 5 and 3 percent93. transported through and delivered within these upazilas. More recent data will be needed to understand this. Chakaria stands out as the home of the RMG and textile manufacturing sector in the district. This upazila alone accounted for more than a quarter of all non-agricultural enter- Fishing and aquaculture prises in Cox’s Bazar, and for three-quarters of all RMG and textile manufacturing firms in the district. 1 in 4 non-agricultural enterprises in Chakaria were engaged in RMG and textile Fishing and aquaculture development have the potential to create livelihoods and manufacturing. Other important non-trade sectors in Chakaria were salt extraction, trans- generate income in Cox’s Bazar. If organized well and in tandem with environmental port, and other services (including tailoring), each of which accounted for between one- standards, they can contribute to dietary diversity and nutrition, with a lower carbon fifth and one-fourth of all enterprises in the district. footprint than other animal proteins. With increasing sea levels and salinization of coastal land, a shift from land-based agriculture to aquaculture may be inevitable for coastal areas Cox’s Bazar Sadar, not surprisingly, was home to a diverse set of service activities. These in the country. Inland, farmed aquaculture has expanded substantially in Bangladesh, featured firms engaged in hospitality sectors - accommodation, food and tourism-related making it one of the world’s top five largest producers of inland capture and culture. This services (25 percent of such firms in the district); education (22 percent); and other services has been driven by expanding domestic demand. Over 90 percent of farmed fish (excluding including tailoring (27 percent). Firms in Cox’s Bazar Sadar also account for a substantial shrimp) are sold on the domestic market (Rashid and Xiang 2019). At the same time, com- share of the district’s manufacturing. In salt extraction and the “other industry” category, mercial shrimp production has also expanded, becoming the third-largest sector in terms including manufacture of wood and wood products except furniture, Cox’s Bazar Sadar is of export earnings. However, shrimp exports have been declining in recent years, due to home to roughly a quarter of all district firms. Within Cox’s Bazar Sadar, services includ- challenges in maintaining international food standards and traceability requirements (UN ing trade accounted for 84 percent of non-agricultural firms. This data predates the influx Conference on Trade and Development 2017), particularly due to the high concentration of of humanitarian assistance, organizations and workers to Cox’s Bazar, which will have subsistence farming with outdated practices, low productivity and product quality. increased the demand for housing, transport, and urban services. Cox’s Bazar has a comparative advantage at the national and international level in terms of Ramu, along with Chakaria and Ukhia, accounted for 80 percent of firms engaged in the shrimp cultivation and sea-caught fish. Natural characteristics such as saltwater endow the transport and storage sector in 2013. In addition to being home to a third of transport district with high potential for cultivating many types of fish. Currently, shrimp production firms, Ramu hosted a small cluster of firms engaged in manufacturing wood products and from Cox’s Bazar serves both local and international markets. Bangladesh exported $532.03 furniture. Although the latter accounted for only 7 percent of firms in the upazila, these million worth of fish and fishery products during FY 2016-17, of which almost 90 percent was represented 28 percent of all such firms in the district. contributed by shrimp (Department of Fisheries Bangladesh 2017). Besides export of crab and dry fish to Southeast Asia and the Middle East has potential for expansion. Maheshkhali accounted for 28 percent of all salt extraction firms in Cox’s Bazar, and 1 in 4 firms in the upazila were engaged in this activity in 2013. Maheshkhali was otherwise Shrimp aquaculture in coastal areas of Cox’s Bazar provides income options and pro- dominated by trading activities. While more recent data is not available, available data sug- motes food security. Fish represents an important source of protein for hosts as well as gests that this upazila is lagging behind others in the district, due both to limited connec- Rohingya. In this sense, the influx has increased the potential to develop the sector (FAO tivity and a limited set of work opportunities. Complementary investments in connective 2019). This could foster the absorption of many low-skilled workers, mainly in rural areas. infrastructure and urban services will be needed to take advantage of the large-scale capi- While it boosts job opportunities and incomes, developing the fish industry can also bring tal investments expected in the energy complex and deep seaport in Matarbari. foreign currency into the country. The establishment of high-tech firms in fish process- ing—particularly frozen and dry fish processing and shrimp cultivation and export—could further spur the region’s economic development (Lemma et al. 2018). 93 See Table A1-18 in Annex 1. 138 139 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Although 90 percent of total fish and fishery exports are dominated by shrimp, the shrimp first ten years (USAID 2019).94 The increased presence of humanitarian workers and devel- industry is hamstrung by several factors. These include consistently low prices, diseased opment actors in Cox’s Bazar, and the accompanying need for accommodation and office seed, value chain inefficiencies, traditional methods of farming, declines in shrimp process- space, has also increased earnings in the hotel industry and the rental value of real estate, ing plants, shortages of quality raw materials, and a lack of scientific culture (USAID 2019). which could further boost incentives to invest. At the same time, investments are being Three types of barriers to exporting shrimp, for instance to the European Union (EU), have made to upgrade the district airport to international standards. been documented: (i) government practices and regulations, including restrictive trade prac- tices, customs, and administrative entry procedures; (ii) technical barriers to trade, including Despite the obvious potential to develop tourism, structural challenges remain. For one, standards, testing and certification; and (iii) specific limitations, including import licensing, concerns about safety and security of tourists and occasional acts of hijacking and kidnap- exchange rate control, export restraints, measures to regulate domestic prices, and require- ping prevent a larger number of tourists from visiting and staying longer.95 Lack of branding ments concerning marking, labeling, and packaging (Naureen et al. 2006). and promotion also constrains tourism growth. The lack of promotion of the country as a tourist destination implies that Bangladesh continues to possess a somewhat negative Despite the potential for a large international market for dry fish, increasing non-tariff image abroad among potential tourists (USAID 2019). The business environment is also measures (NTMs) have emerged as a critical barrier to exports. These NTMs have emerged not conducive to increased private investment, and there are high direct and indirect barri- from the Uruguay Round of the Multilateral Trade Negotiations and Agreements on Technical ers to entry for new businesses due to imperfect competition. Finally, the recent Rohingya Barriers to Trade (TBT) and Sanitary and Phytosanitary (SPS) measures. The WTO, SPS, and influx and concerns about security may have a negative impact on the development of TBT agreements imposed a bound obligation to the exporting member countries to improve tourism in Cox’s Bazar as local and foreign investors may be more hesitant to make large food quality as per set international standard. However, the compliance cost related to SPS investments in the industry (Lemma et al. 2018). obligations is too high, and the Government of Bangladesh is reluctant or otherwise unable to meet the set criteria (Ahmed, Islam, and Shamsuddoha 2007). Manufacturing export clusters in Cox’s Bazar Tourism Manufacturing export firms in Cox’s Bazar are entirely concentrated in the RMG sector and in Chakaria. Manufacturing firms represent 14 percent of total firms in Cox’s Bazar, of which 3 Prior to COVID-19, tourism was the largest and fastest-growing industry in the world. percent96 reported exporting their products on the 2013 Economic Census.97 However, while And with the longest sandy sea beach in the world, Cox’s Bazar has the potential to this share is slightly larger than at the national and division levels, the district only represents become one of the world’s major tourist attractions. While recent years have witnessed 2 percent of Bangladesh’s total manufacturing export firms (Figure 4-7). Ninety-eight percent a huge expansion in hotels, motels, and restaurants in the district, infrastructure facilities of Cox’s Bazar firms that are selling their products in the international market are in the RMG and improved communications are yet to be developed to a commensurate level to foster sector. Furthermore, 84 percent of the exporting firms are located in Chakaria and 12 percent the district’s potential as a tourism hub (Lemma et al. 2018) in Ramu. Less than 1 percent of nonagricultural firms selling their products on the interna- tional market are located in Teknaf and Ukhia (Figure 4-8). The development of tourism in Cox’s Bazar would foster the growth of other sectors, par- ticularly transport, food, and accommodation, as well as retail trade and personal services. In contrast to national and Chittagong division patterns, most Cox’s Bazar export firms Tourism-driven growth creates additional opportunities for investment, development, have fewer than 10 employees. While 7 percent of exporting firms in the district have and infrastructure spending (Kyungmi, Muzaffer, and Sirgy 2013). Locally generated tax between 2 and 9 workers, 90 percent of exporting firms have 1 worker, and only 3 percent revenues can also increase through lodging and sales taxes, revenues from air travel and other transportation taxes, as well as taxes on business and fuel (Bhattacharjee, Polas, and 94 Siam International of Thailand will invest around $500 million out of the total budget of $3 billion Rahman 2018). Tourism and hospitality are also a labor-intensive source of growth and, for infrastructure development. Sabrang Tourism Park will be the first exclusive tourism park in the with the accompanying skills and training, could provide new, better quality jobs for local Cox’s Bazar district, encompassing an area of 1,027 acres. Sonadia Eco-Tourism Park in Maheshkhali is residents. Finally, it could provide a boost for the local handicrafts sector, which has a rich developing on 9,467 acres of land (Dhaka Tribune 2018; Daily Star 2018). heritage (Lemma et al. 2018). 95 Siddiqi Raquib, “Second SAARC Tourism Ministers meet ends with no breakthrough,” The New Nation, June 11, 2006; Amin Sakib-Din, “The role of tourism in Bangladesh economy,” The New Nation, December 6, 2006. Several ongoing initiatives seek to boost tourism in the district, including the Naf Tourism 96 See Table A1-19 in Annex 1. Park, Sabrang Tourism Park, and Sonadia Eco-Tourism Park, with a target to create 200,000 97 The census only asks this question to firms in the manufacturing sector, so this is an underestimate jobs, and full tax exemptions granted by the Bangladesh Economic Zones Authority for the of exporting firms. 140 141 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Figure 4-7. Firms’ main markets and Figure 4-8. Distribution of exporting have more than 10 employees. Most of these firms are in the RMG sector. The structure of customer types: Figure 4-7: Bangladesh, Distribution of exporting firms by Figure upazilas 4-8: in Cox’s of Bazar, Distribution 2013 exporting firms with access to international markets is completely different at national and division Chittagong, and Cox’s firms by districts, Bazar 2013 Bangladesh, firms by upazilas in Cox’s Bazar, 2013 level. 70 and 50 percent of exporting firms have more than 10 employees in Bangladesh 2% 3% and Chittagong, respectively (Table 4-6). Among them, 1 out of 4 exporting firms has more 3% than 100 employees.98 6% 35% 7% Among agricultural enterprises, the fisheries sector is a key export cluster in Cox’s Bazar. 84% As previously discussed, agricultural products are the third export cluster in Bangladesh. 8% In 2017, the sector represented 3 percent of total national exports, of which 35 percent was 13% comprised of frozen shrimp exports. While this is also an important export activity for Cox’s 24% Bazar, only 6 out of 162 Bangladeshi shrimp processing factories are located in the district.99 12% Table 4-6: Main activities and size of export firms in Cox’s Bazar, Chittagong, and Bangladesh, 2013 Dhaka Sylhet Cox's Bazar Sirajganj Gazipur Chittagong Chakaria Ramu Maheshkhali Pabna Others Narayanganj Pekua Teknaf Cox’s Bazar Sadar Activities Cox’s Bazar Chittagong Bangladesh Manufacture of leather and related products 1% 3% Source: WB staff calculations, Economic Census 2013. Manufacture of fabricated metal products 6% 3% Economic connectivity and infrastructure enhancements Manufacture of food products 7% 8% Manufacture of furniture 13% 7% The current state of transportation infrastructure prevents most of Cox’s Bazar district from profiting from the jobs and economic opportunities to be created by the proposed Other manufacturing 2% 18% 14% deep seaport in Matarbari. Average travel times to the port from upazilas in Cox’s Bazar is Manufacture of textiles and RMG 98% 55% 65% principally shaped by geography, with the southern upazilas remaining effectively discon- nected, even taking into account proposed improvements in road and ferry connections. Firm Size Cox’s Bazar Chittagong Bangladesh The latter improvements are described in detail further below. 1 worker 90% 19% 5% Current access to the proposed deep seaport at Martarbari is quite low, even for nearby 2 workers 4% 4% 2% areas in Chakaria, because of poor quality roads servicing the port site from the main highway and the surrounding communities. Further afield, the most direct commuting 3-4 workers 2% 10% 9% route across the Cox’s Bazar bay and up Maheshkhali’s central road is blocked by the lack of 5-9 workers 1% 17% 15% a dedicated ferry and the modest state of the road. Potential commuters would instead be forced to use the Chittagong road, adding 30-60 minutes in commuting time that effectively 10 plus workers 3% 50% 69% places the commute out of reach for most. Furthermore, there is a risk that, if multimodal Total number of export firms 358 2,245 16,988 logistic services are not enhanced, Cox’s Bazar would face negative exposures from the traffic between greater Dhaka and Chittagong,100 through increased congestion (the average speed Source: World Bank staff calculations using the 2013 Economic Census. on inter-city roads is only 30 kilometers per hour) and high levels of pollution (IFC 2020). 98 See Table A1-20 in Annex 1. 100 The majority of container freight movement takes place between greater Dhaka and Chittagong: https://www.unescap.org/sites/default/files/6-%20%20Sea%20Food%20Export%20from%20 99 70 percent of container traffic from Chittagong goes to Dhaka, with almost 95 percent going by road Bangladesh-Kabir.pdf. See also Mahmud (2018). (IFC 2020). 142 143 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Box 5: Modeling accessibility All the travel times, accessibility indicators, and economic gravity calculations dis- played in the maps and corresponding charts presented here are the result of World Bank staff calculations. These calculations leverage data from the following sources: It is possible to model the impacts of several transportation investment scenarios • Transport: OpenStreetMap on accessibility to services, economic integration, and equitable growth. To do so, • Population: Facebook and CIESIN 2016 (based on Bangladesh Bureau of analysts downloaded geospatial data for all roads and ferries in Cox’s Bazar (and Statistics Census 2011) onwards to Chittagong) from OpenStreetMap, assigned them speeds according to • Population, Economic Census: Bangladesh Bureau of Statistics 2013 their type,101 and used the resulting network to calculate travel times between all • Administrative borders, markets, services: Local Government Engineering origins (population points) and destinations (services, markets, and others). Department (LGED) 2019 The accessibility methodology, its theoretical underpinnings, and the data inputs To calculate the accessibility improvements from proposed investments, analysts are described in greater detail in the Annex. increased the speeds for selected road or ferry segments to their projected post-in- vestment levels and then re-calculated the origin-destination travel times. The pro- jected speed improvement is usually 2-3 times existing speeds, depending on the Map 4-1. Estimated travel times specific road segment. to Matarbari Map port,travel 4-1: Estimated upgrade with times to The proposed upgrades to key feeder of key roads Matarbari port, with upgrade of key roads shown in Map 4-1 would greatly The resulting origin-destination travel time figures are then used to prepare aver- roads improve access to the port for the whole 0 20 40km age travel times and accessibility indices for each destination under each scenario. district, although the establishment C H I T TAG O N G Average travel times record the mean travel time to the nearest service or place of Dhaka Chittagong Pekua of a rapid ferry across the Cox’s Bazar employment per administrative unit – unions in most cases. Accessibility indices Bay as shown in Map 4-2 is necessary to employ a more sophisticated measure of potential accessibility, which measures Kutubdia make commuting from northern Ukhia access to all the services, jobs, and population centers in the district, to better Chakaria viable. The maximum feasible commut- account for the cumulative nature of access. One category of accessibility indices, BANDARBAN Matarbari (approximate) ing distance for desirable, high-quality commonly known as gravity models, calculates destinations’ economic gravity by jobs in Cox’s Bazar district is unknown, weighting their potential accessibility measurements by their attractiveness and the Maheshkhali but previous research in Dhaka showed inverse of the travel-time distance between them, when calculating this cumulative residents endured mean commute times access. The analysis here follows Yoshida and Blankespoor (2010) in employing a Cox’s Bazar negative exponential model to calculate the gravity indices. More detail on the tech- Cox’s Bazar of approximately 70 minutes, with signif- Sadar Ramu nical methodology can be found in the Annex. icant shares tolerating times of up to 100 COX’S minutes (RSTP Household Survey 2014). BAZAR Three scenarios additional to the current transportation setup are considered and The ADB’s planned upgrades to the south- Ukhia visualized: ern road linking Cox’s Bazar Sadar, Ukhia, • The main roads servicing the Martarbari port and Maheshkhali upazila and Teknaf significantly complement are upgraded Teknaf these investments by extending the plau- • The above roads are upgraded, and a dedicated ferry line is set up con- sible commuting catchment of the deep necting Maheshkhali and Cox’s Bazar city across the bay Proposed road segments for upgrading seaport throughout Ukhia, as shown in • The above investments are made, and upgrades are made to the principal Minutes travel to Matarbari Map 4-3. Even with the full suite of invest- southern highway connecting Ukhia and Teknaf to Cox’s Bazar Sadar. deep sea port St. Martin Dwip With upgrades to port roads ments, commuting from eastern Ramu, southern Ukhia, and Teknaf will likely be 0 60 120 150 180 214 impossible, so residents of these areas are Note: Estimations based on an internal model of travel unlikely to be physically connected to any times (See Annex 2) and population distribution models 101 Typology adapted from the Bangladesh Roads and Highways Authority. See Table A2-1 in Annex 2. from the High-Resolution Settlement Layer from potential growth and employment oppor- Facebook and the Center for International Earth tunities from the investments in Matarbari. Science Information Network 2016. 144 145 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Map 4-2: Map 4-2.Estimated Estimated travel times travel to times Map 4-3: Map 4-3. Estimated Estimated travel times travel to times Map 4-4: Estimated travel times to A successful southern commuter strategy Matarbari port, with upgrade to Matarbari port, with upgradeof key Matarbari port, with upgrade to Matarbari port, with upgrade of key Cox’s Bazar Sadar, with upgrade of could have negative secondary effects on roads of keyand ferry roads and ferry key ferry, roads, of roads,and AH41 ferry, (N1) and Highway AH41 (N1) key roads 0 20 40km travel speeds. If a ferry successfully enables Highway C H I T TAG O N G commuting to Matarbari from southern Pekua 0 20 40km 0 20 40km Dhaka Chittagong Cox’s Bazar, increased congestion within Kutubdia Cox’s Bazar Sadar may result: the current Chakaria pier location requires northbound traf- Dhaka C H I T TAG O N G Dhaka C H I T TAG O N G Matarbari BANDARBAN Chittagong Pekua Chittagong Pekua (approximate) fic from Teknaf or Ukhia to route through Maheshkhali beachfront and downtown roads already Kutubdia Kutubdia congested from tourist and commercial traf- Cox’s Bazar Chakaria Chakaria Cox’s Bazar Ramu fic. At its worst, this could negate some or all Sadar Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) COX’S potential improvements to commute times BAZAR from the south. Further investment in traf- Maheshkhali Maheshkhali Ukhia fic-alleviating connecting roads, road widen- ing, pavement maintenance, traffic dividers, Cox’s Bazar Cox’s Bazar Teknaf Cox’s Bazar Cox’s Bazar Proposed road and signals within Cox’s Bazar Sadar may Sadar Ramu Sadar Ramu segments for upgrading be needed to mitigate these potential side COX’S COX’S Minutes travel to Cox’s Bazar With current transport infrastructure effects (Hussain and Mallick 2017). BAZAR BAZAR St. Martin Dwip Ukhia Ukhia 0 30 60 90 120 337 Map 4-5: Estimated travel times to Map 4-6: Estimated travel times to Teknaf Teknaf Cox’s Bazar Sadar, with upgrade of Cox’s Bazar Sadar, with upgrade of key roads and ferry key roads, ferry, and AH41 (N1 Highway) 0 20 40km 0 20 40km Dhaka C H I T TAG O N G C H I T TAG O N G Chittagong Dhaka St. Martin St. Martin Pekua Chittagong Pekua Dwip Dwip Kutubdia Kutubdia Minutes travel to Matarbari deep sea port Chakaria Chakaria Matarbari BANDARBAN Matarbari BANDARBAN Proposed road Proposed ferry Proposed AH41 (approximate) (approximate) segments for upgrading across CXB bay road for upgrading 0 60 120 150 180 214 Maheshkhali Maheshkhali Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Sadar Ramu Sadar Ramu Network 2016. COX’S COX’S BAZAR BAZAR Ukhia Ukhia By increasing accessibility from all sides, the proposed upgrades to key roads, ferries, and the AH41 highway collectively further increase the centrality of the Sadar to Cox’s Bazar. As Teknaf Teknaf seen in Map 4-4 and Map 4-5, the proposed northern road and ferry upgrades would integrate southern and central Maheshkhali with the Sadar’s markets and perhaps create commercial St. Martin St. Martin opportunities for farmers there. Map 46 shows that the AH41 highway upgrade would do the Minutes travel to Cox’s Bazar Dwip Dwip same for Ukhia. These also increase opportunities for tourists flying into Cox’s Bazar to more Proposed road Proposed ferry Proposed AH41 segments for upgrading across CXB bay road for upgrading easily access areas outside of the city. However, remote areas like Kutubdia, eastern Ramu, 0 30 60 90 120 337 and Teknaf will remain distant from the city even with these upgrades. Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Network 2016. 146 147 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Transport investments will support growth in some upazilas and northern unions in Pekua and Kutubdia have limited access to these clusters of growth. Map 4-7. Unions with significant With the proposed transport investments, the economic weight of unions surrounding Further analysis investigates the potential presence Map of large 4-7: Unions withfirms, significant Cox’s Bazar Sadar and parts of southern Maheshkhali increases significantly (Map 4-9), benefits of the planned investments for Cox’s Bazar presence of large firms, Cox’s Bazar whereas unions in Teknaf and Ukhia are affected only marginally. growth in the district and the spatial inclu- 0 20 40km Map 4-8. Large firm accessibility Map 4-9. Large firm accessibility sivity of growth. The approach simulates Map 4-8: Large firm accessibility Map 4-9: Large firm accessibility indices (markets weighted by indices (markets weighted by improvements in accessibility for large Dhaka indices (markets weighted by large indices (markets weighted by large Chittagong large firms), pre-transport large firms), post-transport firms, growth centers, and employment C H I T TAG O N G firms), pre-transport investments firms), post-transport investments Pekua investments investments clusters of good/secure jobs and documents 0 20 40km 0 20 40km any growth exposures by upazila. Large Kutubdia firms are defined as firms with more than Chakaria C H I T TAG O N G C H I T TAG O N G 10 workers. Growth centers are important Matarbari BANDARBAN Dhaka Chittagong Pekua Dhaka Chittagong Pekua (approximate) markets characterized by having perma- Kutubdia Kutubdia nent multimodal structures and managing Maheshkhali a large volume of trade, as identified by the Chakaria Chakaria Bangladesh Planning Commission. Secure Cox’s Bazar Matarbari (approximate) BANDARBAN Matarbari (approximate) BANDARBAN employment is defined as individuals Cox’s Bazar Ramu Sadar who are business owners or on a full-time Maheshkhali Maheshkhali COX’S contract. The workings of the accessibility BAZAR Cox’s Bazar Cox’s Bazar models, accessibility indexes, and gravity Cox’s Bazar Cox’s Bazar Ukhia Ramu Ramu models are described in Box 5 and further Sadar Sadar elaborated in the Annex. COX’S COX’S BAZAR BAZAR Teknaf Ukhia Ukhia Firms in northern and central Cox’s Bazar are best positioned to exploit transport investments. Large firms with the greatest Teknaf Teknaf Proposed road growth potential are only present in a hand- St. Martin segments for upgrading Dwip ful of unions in Maheshkhali, Chakaria, and Number of large firms Proposed ferry across CXB bay Cox’s Bazar Sadar (Map 4-7). In a spatial Proposed AH41 road for upgrading Low High statistical analysis,102 these firms dispro- St. Martin St. Martin Dwip Dwip Economic gravity Economic gravity portionately cluster in the urban unions Note: Estimations based on an internal model of travel Growth centers Growth centers around Cox’s Bazar Sadar, implying that times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from they effectively form a single economic Unions Unions Facebook and the Center for International Earth cluster in and around the Sadar upazila. Science Information Network 2016. Low High Low High The planned transport upgrades increase the economic weight of these urban unions Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from near Sadar, as well as the weight of Maheshkhali, where Matarbari is located. Map 4-8 the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Network 2016. shows the current accessibility of populated areas in Cox’s Bazar, averaged by union, to large firm clusters. Along with unions in Ukhia and Teknaf, southern parts of Maheshkhali Higher levels of investment in transportation will better integrate existing markets in the north-central core but fail to integrate peripheral markets in southern Cox’s Bazar, 102 In a Gedis-Ord Local GI* “hotspot analysis,” these unions showed the only statistically significant Kutubdia, and northern Chakaria. Upgrades to roads surrounding Martarbari would mod- (z-score > 1.96) spatial concentration of large firms. Adjoining unions were defined as belonging to the estly increase the accessibility of government-designated growth centers (major markets) same cluster “neighborhoods” for this analysis when defining spatial weights. 148 149 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES for residents in Maheshkhali and Chakaria (see Map 4-10 for pre-transport investment acces- Alone, the substantial proposed transportation investments cannot transform the cur- sibility versus Map 4-11 for post-transport investment accessibility). Additional upgrades rent geography of employment opportunities in Cox’s Bazar. Accessibility to non-agricul- to ferries would integrate southern Maheshkhali residents better with Cox’s Bazar Sadar’s tural employment remains clustered around northern and central Cox’s Bazar, even with commercial markets. By contrast, market accessibility indices barely improve in Ukhia and the improvements in access displayed in Map 4-13 and Map 4-15. These investments raise Teknaf, even taking into account the ADB’s proposed upgrades to the main southern AH41 the overall index of accessibility to jobs and boost the prospects of a few unions in particu- road. In line with improvements in access, notable increases can be seen in the economic lar, but don’t fundamentally resolve low levels of access in southern Cox’s Bazar, Kutubdia, weight of growth centers in Cox’s Bazar Sadar and southern Maheshkhali before (Map 4-12) or northern Chakaria. and after (Map 4-13) investments. Map 4-12: Market accessibility index Map 4-13: Market accessibility index Map 4-10: Travel times to growth Map 4-11: Travel times to growth (growth Map Market pre-investment centers), 4-12. accessibility index (growth centers), Market post-investment Map 4-13. accessibility index Map 4-10. Travel times to growth Map 4-11. Travel times to growth centers, (current) pre-investment centers, post-investment (unweighted) (growth centers), pre-investment (unweighted) (growth centers), post-investment centers, (current) pre-investment centers, post-investment (unweighted) (unweighted) 0 20 40km 0 20 40km 0 20 40km 0 20 40km Dhaka Dhaka Chittagong Chittagong Dhaka Dhaka Chittagong Chittagong C H I T TAG O N G C H I T TAG O N G C H I T TAG O N G C H I T TAG O N G Pekua Pekua Pekua Pekua Kutubdia Kutubdia Kutubdia Kutubdia Chakaria Chakaria Chakaria Chakaria Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) Maheshkhali Maheshkhali Maheshkhali Maheshkhali Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Cox’s Bazar Sadar Ramu Sadar Ramu Sadar Ramu Sadar Ramu COX’S COX’S COX’S COX’S BAZAR BAZAR BAZAR BAZAR Ukhia Ukhia Ukhia Ukhia Teknaf Teknaf Teknaf Proposed road Teknaf segments for upgrading Proposed ferry across CXB bay Proposed AH41 road Proposed road for upgrading segments for upgrading St. Martin Economic gravity St. Martin Economic gravity St. Martin Growth centers St. Martin Growth centers Dwip Dwip Growth centers Dwip Growth centers Dwip Minutes travel to growth center Minutes travel to growth center With current transport infrastructure With upgrades to port roads Unions Unions 0 10 20 30 45 90 0 10 20 30 45 90 Low High Low High Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Network 2016. Network 2016. 150 151 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Map 4-14. Map 4-14: All firms accessibility All firms accessibility Map 4-15: Map 4-15. All All firms firms accessibility accessibility of the district, followed by Pekua, Ramu, and Teknaf, which are home to 13, 15, and 10 indices (markets weighted indices (markets weighted byby firms), firms), indices (markets weighted by indices (markets weighted by firms), firms), percent of secure jobs, respectively. On the other hand, only 4 percent of all enterprise/ investment pre-transport investment pre-transport post-transport investment post-transport investment business owners and full-time workers are located in Ukhia. At the same time, Cox’s Bazar Sadar and Chakaria account for 47 percent of all vulnerable jobs in the district, with Teknaf 0 20 40km 0 20 40km in third place, hosting 14 percent of vulnerable workers in Cox’s Bazar.104 Map 4-16 shows Dhaka Dhaka Chittagong Chittagong that northern Cox’s Bazar currently enjoys much higher accessibility to relatively secure C H I T TAG O N G C H I T TAG O N G employment. Residents there have access to high-quality jobs in Cox’s Bazar Sadar, smaller Pekua Pekua firms throughout the north, and the university cluster in Ramu. After the full suite of pro- Kutubdia Kutubdia posed transport investment (Map 4-17), southern Maheshkhali and Cox’s Bazar are more Chakaria Chakaria tightly integrated with the job clusters in Ramu and the north, but accessibility continues Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) to lag in the south, apart from a modest increase in Teknaf. These maps indicate a need to Maheshkhali Maheshkhali create additional non-agricultural jobs in or closer to Ukhia and Teknaf. Cox’s Bazar Cox’s Bazar Map 4-16. Map High-quality jobs 4-16: High-quality jobs 4-17: High-quality Map 4-17. Map High-quality jobs jobs Cox’s Bazar Cox’s Bazar Sadar Ramu Sadar Ramu accessibility indices accessibility indices (markets (markets accessibility indices accessibility indices (markets (markets COX’S COX’S weighted by weighted by high-quality high-quality job job weighted by weighted by high-quality high-quality job job BAZAR BAZAR numbers), pre-transport investment numbers), pre-transport investment numbers), post-transport numbers), post-transport investment investment Ukhia Proposed road Ukhia segments for upgrading Proposed ferry across 0 20 40km 0 20 40km CXB bay Teknaf Proposed AH41 road Teknaf for upgrading Dhaka C H I T TAG O N G Dhaka C H I T TAG O N G Chittagong Chittagong Pekua Pekua Economic gravity Economic gravity Growth centers Growth centers Kutubdia Kutubdia St. Martin St. Martin Chakaria Chakaria Dwip Dwip Unions Unions Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) Low High Low High Maheshkhali Maheshkhali Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from Cox’s Bazar Cox’s Bazar the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Cox’s Bazar Cox’s Bazar Network 2016. Sadar Ramu Sadar Ramu COX’S COX’S While the share of insecure employment in the district is relatively low, the inability BAZAR BAZAR to access secure employment is more pronounced in certain upazilas.103 In particu- Ukhia Proposed road Ukhia segments for upgrading lar, northern parts of the district have greater access to better-quality jobs. Together Proposed ferry across CXB bay with proposed transportation investments, efforts are needed to create high-quality Teknaf Proposed AH41 road Teknaf for upgrading non-agricultural jobs in or nearer upazilas that now have limited access. Five percent of workers in Cox’s Bazar have an insecure job (are unpaid, part-time, or irregular workers), Economic gravity Economic gravity though this share is lower than at the national (8 percent) and division (6 percent) levels. Growth centers Growth centers St. Martin St. Martin Substantial differences exist across upazilas. Cox’s Bazar Sadar and Pekua have the larg- Dwip Dwip Unions Unions est shares of insecure workers in the district, at 8 and 11 percent, respectively. In Teknaf and Ukhia, only 4 and 3 percent of workers have insecure jobs. Assessing the distribution Low High Low High of the relatively insecure and secure jobs across upazilas, we find that 43 percent of all secure jobs (full time workers and business or enterprise owners) are located in the capital Note: Estimations based on an internal model of travel times (See Annex 2) and population distribution models from the High-Resolution Settlement Layer from Facebook and the Center for International Earth Science Information Network 2016. Secure employment is defined as being a business owner or on a full-time contract. It contrasts with 103 vulnerable, insecure employment for unpaid, part time, and irregular workers. 104 See Tables A1-21 and Table A1-22 in Annex 1. 152 153 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Infrastructure investments may make a modest contribution to long-term human cap- Governance and service delivery ital formation. As present access to primary and secondary schools is mostly very good, transportation investments will only exert a significant impact on access to tertiary educa- The Rohingya influx has been accompanied by a large-scale humanitarian response in tion (Map 4-18 versus Map 4-19). Upgrades to roads for Matarbari would modestly increase a context of weak local governance. Local governments in affected areas have limited accessibility to universities for residents of northern Maheshkhali and Chakaria, while a funds, functions, personnel, and capacity to manage the response.105 Between 2017 and ferry and the ADB-built southern road would similarly improve accessibility for southern 2020, funding of the Rohingya crisis response has averaged US$564 million,106 and has Maheshkhali, Ukhia, and Teknaf. However only for Maheshkhali and Ukhia would this likely been largely successful in delivering basic needs and food security to the displaced popu- reduce the mean travel time below a threshold of approximately 90 minutes. lation (World Bank 2020c, based on CBPS 2019). However, district and local governments in Bangladesh do not currently participate actively and contribute to decision-making and investments at their level. This implies that there are few avenues for citizens and the host Map 4-18: Travel times to tertiary Map 4-19: Travel times to tertiary Map 4-18. Travel times to tertiary Map 4-19. Travel times to tertiary community to shape the responsiveness of government programs and policies to their education, pre-transport investment education, post-transport investment education, pre-transport education, post-transport needs. Elected representatives of local government institutions rarely participate in the investment investment identification, appraisal, approval, implementation, and monitoring of investment proj- ects funded through the Annual Development Plans (World Bank, forthcoming). 0 20 40km 0 20 40km Dhaka Dhaka Chittagong Chittagong A World Bank report (World Bank 2020b) documents variable capacity at different levels of C H I T TAG O N G C H I T TAG O N G local government. Paurashavas, or municipal governments, generally have better capacity Pekua Pekua than local government institutions at the district level. Likewise, upazila parishads107 (pres- Kutubdia Kutubdia ent in both urban and rural settings) appear to be more capacitated than union parishads. In practice, neither upazila parishads nor union parishads have sufficient command over other Chakaria Chakaria government departments in their area. While lower-tier local governments are closer to citi- Matarbari BANDARBAN Matarbari BANDARBAN (approximate) (approximate) zens, capacity gaps limit their ability to advocate for local preferences and needs. Among the four paurashavas in Cox’s Bazar (Chakaria, Cox's Bazar Sadar, Maheshkhali, and Teknaf), the Maheshkhali Maheshkhali actual number of permanent staff is currently well below the “standard,” and there are seri- ous shortfalls in administration, engineering, and health and family planning departments. Cox’s Bazar Cox’s Bazar At the zila parishad level, as well, both staff and budgets are relatively small, in line with the Cox’s Bazar Cox’s Bazar Sadar Ramu Sadar Ramu limited portfolio of activities. At lower tiers of government, both staffing and mandates are COX’S COX’S larger. However, when comparing Cox’s Bazar to other districts, for paurashavas, the level of BAZAR BAZAR per capita expenditure in the district is twice as high as the national average, while for union Ukhia Ukhia parishads—the lowest tier of government, with the longest list of subjects in its mandate— expenditure is only 75 percent of the national average. Teknaf Teknaf An analysis of district-level public expenditure allocations suggests great unevenness across districts and relatively fixed allocations across districts under certain expenditure heads, irrespective of population size. In terms of education, health, agriculture, and local governance expenditures (to the extent that they can be spatially allocated), Cox’s Bazar St. Martin St. Martin does not appear to be attracting public resources on par with its population size (Box 6). Dwip Dwip University University 105 See Khan (2016). Minutes travel to university Minutes travel to tertiary school 106 OCHA Financial Tracking Service, Bangladesh: 2020 Joint Response Plan for Rohingya Humanitarian With current transport infrastructure With upgrades to port roads Crisis (January-December), https://fts.unocha.org/appeals/906/summary 20 30 60 120 150 210 107 An upazila parishad consists of a chairman, two vice-chairpersons (one of them a woman), chair- 5 10 15 25 60 90 men of all union parishads under the upazila concerned, mayors of all municipalities, if there are any, and women members of the reserved seat. 154 155 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES Box 6: Analyzing district-level public expenditure in Bangladesh Figure B6-3: Per capita allocated Figure B6-4: Allocated Figure local B6-3. Per capita governance allocated expenditure Allocated agricultural Figure B6-4.expenditure agricultural per acre local governance expenditure expenditure per acre of cropped The allocation of public resources at district level in Bangladesh appears highly (BDT), by district of cropped area (BDT), by district (Tks), by district area (Tks), by district uneven. Dhaka division receives the highest share of executed budget that can be 0 75 150km 0 75 150km spatially allocated, followed by Chittagong division. In fact, Dhaka receives the most executed budget, both allocated and unallocated, with a large component of unal- located budget spent on public services that are overwhelmingly concentrated in Dhaka (Figure B6-1 & Figure B6-2). Per capita allocated education expenditure is higher in Barisal and Khulna divisions and in the eastern part of Chittagong division than in other areas of the country. Cox’s Dhaka Dhaka Bazar district falls in the bottom of the distribution in terms of per capita allocations both at the national level and within the division. Per capita allocations in Rangmati and eastern districts in Chittagong division are high due to their small population, while per capita allocations are low in Cox’s Bazar, in part due to a relatively high Chittagong Chittagong concentration of population in this district compared to its eastern neighbors. Public resource allocations include a fixed component which may partially explain the pat- 6072.615 – 6743.528 Cox’s Bazar 7110.663 – 7800.019 Cox’s Bazar 5401.702 – 6072.615 6421.307 – 7110.663 terns observed (Figure B6-3 & Figure B6-4). 4730.789 – 5401.702 4059.876 – 4730.789 5731.951 – 6421.307 5042.595 – 5731.951 3388.962 – 4059.876 4353.239 – 5042.595 2718.049 – 3388.962 3663.883 – 4353.239 2047.136 – 2718.049 2974.528 – 3663.883 1376.223 – 2047.136 2285.172 – 2974.528 Figure B6-1: Expenditure allocation Figure B6-2: Per capita allocated 705.3098 – 1376.223 34.39666 – 705.3098 1595.816 – 2285.172 906.4598 – 1595.816 (BDT) Figure proportion inB6-1. to school-age Expenditure allocation health Figureexpenditure (BDT), B6-2. Per capita allocated (Tks) in proportion population to school-age (0-14 years), by district health by expenditure (Tks), district Within public spending categories, per capita allocation on health is relatively higher population (0-14 years), by district by district in Khulna, Barisal, and urban areas close to Dhaka, as well as in the eastern districts of Chittagong division. 0 75 150km 0 75 150km Allocated spending per capita under the local governance head is relatively similar but low among most districts, with the exception of Rangamati. Chittagong (in par- ticular Rangamati), Khulna, and Barisal receive high per capita budget allocations for local governance. However, most of this budget is for loans, capital transfers, and fixed assets, that is, spatially unallocated. A significant share of local governance Dhaka Dhaka expenditure (21 per cent), allocated and unallocated, is for construction of rural roads, with another 10 percent for other roads, highways, and bridges. Allocated agricultural expenditure is distributed using the net cropped area. The Chittagong Chittagong map shows a disproportional distribution of resources toward urbanized areas around Dhaka. In addition, except for Cox’s Bazar, the southeastern districts in Cox’s Bazar Cox’s Bazar Chittagong division are also relatively well positioned. Over the period 2011-18, 25 9573.037 – 10494.2 1670.368 – 1829.99 8651.871 – 9573.037 7730.705 – 8651.871 1510.747 – 1670.368 1351.125 – 1510.747 to 50 percent of allocated agricultural expenditure went to fertilizer subsidies. On 6809.539 – 7730.705 1191.504 – 1351.125 5888.373 – 6809.539 1031.882 – 1191.504 distributing these subsidies spatially (for 2016-17), besides east Chittagong, districts 4967.206 – 5888.373 872.2606 – 1031.882 4046.04 – 4967.206 712.639 – 872.2606 3124.874 – 4046.04 553.0175 – 712.639 in Khulna receive high spending per acre of cropped land. 2203.708 – 3124.874 393.3959 – 553.0175 1282.542 – 2203.708 233.774 – 393.3959 Source: World Bank staff calculations based on BOOST. 156 157 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES The humanitarian response in Cox’s Bazar is moving into the medium term, creating 4-9. Share Figure 4-9: Figure of total Share of funding by total funding by 4-10. Evolution Figure 4-10: Figure funding in of funding Evolution of in opportunities to build broad-based linkages with host community livelihoods and boost clusters in Cox’s Bazar humanitarian clusters in Cox’s Bazar humanitarian Cox’s Bazar humanitarian response, Cox’s Bazar humanitarian response, the local multiplier of aid. But the local government appears disconnected from the man- 2017-2019 response, 2017-2019 response, 2010-2019 2010-2019 agement of Rohingya affairs, which are entirely overseen by the national government. 100% 800,000 Local government institutions have a constitutional mandate to lead local development by 90% managing the work of public officials, maintaining public order, planning and delivering 80% 700,000 services, and levying taxes. As long as policies covering the Rohingya are determined at 70% 600,000 the national level, local governments and the host population may perceive themselves 60% Thousands of US$ as shut out from efforts to link humanitarian assistance with local development. The first 50% 500,000 40% phase of the UNDP-led development planning exercise could help bridge this gap between 400,000 30% local needs and centrally determined spending priorities and inform investment decisions 20% 300,000 by government and development partners. 10% 0% 200,000 Humanitarian assistance and local economic activity 2017 2018 2019 2020 100,000 Communication Education The most recent influx of displaced Rohingya from Myanmar necessitated an immediate, Shelter Wash 0 2004 2017 2018 2019 2020 Health Logistics large-scale humanitarian response, averaging over 600 million USD per year since 2017. Protection Others Food security and nutrition remain the single largest aid category, accounting for roughly Food security & nutrition 30 percent of all assistance. As noted above, this support has been largely successful in meeting basic food needs for the Rohingya population in the camps in Teknaf and Ukhia Note: 2020 Note: includes COVID-19 2020 includes funding. COVID-19funding. (Figure 4-9 and Figure 4-10). However, available funding falls short of the requirements Source: World Source: Bank staff World Bank calculation,OCHA. staffcalculation, OCHA.108 estimated by the United Nations (UN). The COVID-19 pandemic has restricted UN agencies’ ability to deliver a full range of assistance, narrowing the focus to life-saving humanitarian There is some indicative evidence that the influx of humanitarian and development aid. Simultaneously, the pandemic has generated new needs for support. assistance, together with the presence of workers and staff, has already shaped the local economy in meaningful ways. Air traffic between Cox’s Bazar and Dhaka has increased Such a large influx of humanitarian assistance can increase local economic activity and substantially (Box 7); there is increased demand for real estate and accommodation in generate a demand impetus through multiple channels. First, the humanitarian effort Sadar; and traffic flows have intensified on the Sadar-Teknaf main road that connects the has been supported by a large number of international and domestic staff working with airport and district headquarters to the camps. non-governmental organizations, humanitarian, and development organizations. They generate demand for local accommodation and hospitality services, travel to and from Cox’s Bazar to national and international headquarters, and transport materials from the district center to the Rohingya camps in Teknaf and Ukhia. Second, humanitarian assis- Box 7: Airport activity in Cox’s Bazar since the Rohingya influx tance of this magnitude requires procurement of food and service delivery supplies at The expansion of humanitarian assistance in Cox’s Bazar following the 2017 Rohingya scale. Much of this material continues to be procured internationally or through Dhaka influx has been accompanied by an increase in the number of international and and Chittagong. While these processes generate demand for transport services, new pilot domestic staff working with humanitarian and development organizations in the initiatives signal the potential to procure more assistance locally, expanding local incomes district; a rise in travel back and forth from Dhaka and other international airports; and consumption. Third, the assistance economy has the potential to create new jobs, and an increase in associated cargo operations (Table B7-1). This is evidenced in an especially for well-educated hosts, to provide facilitating services such as translation, but increase in airport activity in Cox’s Bazar. Between 2017 and 2018, passenger traffic more importantly to work in delivering services in camps. Finally, cash assistance delivered at the district’s airport increased 111 percent, and the number of passengers travel- to Rohingya in camps will probably not be spent entirely within camps, and is likely to be ing on the Dhaka-Cox’s Bazar route increased 40 percent (BBS 2019). spent in local shops, which will lead to higher levels of economic activity in and around the Rohingya camps. Data downloaded from https://data.humdata.org/dataset/e31467b1-0f37-40ea-b5be-558cf8c1b8aa 108 158 159 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 4 – A C C E L ERAT I N G I N C L US I VE G RO W TH : C ONSTRA I NTS AN D O P P ORTUN I T I ES 2-20.Location Map4-20: Map Location ofof growth growth centers in Table 4-4: Growth centers in Cox’s centers Cox’s in Cox’s Bazar Bazar Bazar district, by distance from Table B7-1: Cox’s Bazar airport traffic 0 20 40km Rohingya camps Percentage Change Dhaka C H I T TAG O N G Chittagong Number Freight/ Pekua Distance from camps of growth centers Passenger mail Air traffic Freight Air traffic (‘000) (tonnes) movements Passenger mails movements Kutubdia 5 km 5 2014 87 3541 3915 Chakaria Matarbari BANDARBAN 2015 107 2809 5452 23 -21 39 (approximate) 5 to 10 km 1 2016 154 2087 4852 44 -26 -11 Maheshkhali 2017 256 1676 5688 66 -20 17 10 to 15 km 2 2018 539 3834 7131 111 129 25 Cox’s Bazar Cox’s Bazar 15 to 25 km 3 Source: BBS (2018). Sadar Ramu COX’S 25 to 80 km 23 In response to increasing operations, and to support the development of Cox’s BAZAR Bazar, the GoB plans to upgrade the local airport to international status. In 2019, Ukhia Total in Cox’s Bazar 34 Bangladesh’s civil aviation authority began a runway expansion project to accom- modate international flights and fully loaded wide-body aircraft. This project com- Source: World Bank staff calculations based on LGED Teknaf and population location data. plements the construction of an international passenger terminal, which was com- Growth centers pleted in 2019. 109 Rohingya camps Road classifications Other Primary Ferry St. Martin Secondary Dwip Preliminary analysis using changes in nightlight intensity provides evidence of greater Tertiary Minor economic activity in markets near the Rohingya camps.110 There are 34 officially designated growth centers within Cox’s Bazar district, of which five are within five kilometers of one of Figure 4-11. Figure 4-11: Quadratic Quadratic fit fit of of monthly monthly However, major evidence gaps remain the Rohingya camps (Table 4-4 and Map 4-20). To explore the hypothesis that the expansion nightlight intensity around growth nightlight intensity around growth in assessing the local economic impact in assistance to respond to the Rohingya influx was accompanied by increased demand and centers over centers over time time of the humanitarian economy and its activity in local markets, researchers measured changes in monthly nightlight intensity in a potential to deliver widespread benefits 500-meter buffer around growth centers. The points are visualized in Figure 4-14. The dark 5 to host communities. These will need to IHS of total nightlight intensity blue line shows how markets within 5km of a Rohingya camp show more economic activity be filled. Specifically, a detailed under- 4 (proxied by NTL) after the arrival of the Rohingya (dashed vertical line) when compared with standing of changes in economic activity markets farther away. The results hold in regression specifications that are summarized in in the district since the influx is needed. 3 Annex 3. The main finding, which is robust to changes in specification, is that, while markets It should document the nature and scope further away from Rohingya camps also have a positive increase in NTL activity, markets clos- of new types of employment opportuni- 2 est to camps (within five kilometers) experience much larger increases. This provides some ties for hosts, focusing on Sadar, Ukhia, suggestive evidence of increased local demand and economic activity near Rohingya camps. Teknaf, and the main road connecting 1 jan/14 jan/16 jan/18 jan/20 Sadar to the Rohingya camps. The human- 109 http://caab.portal.gov.bd/site/page/748bfeaa-b00a-43f5-9523-39d7d4e169bc. | https://www.thefi- Linear fit and CI of NTL intensity within: itarian effort has already shifted towards nancialexpress.com.bd/trade/coxs-bazar-airport-runway-project-underway-1548219691 greater local integration, including direct 1-5 KM 5-10 KM 10-15 KM 15-25 KM 110 Nightlight intensity has been shown to have a strong correlation with economic activity and growth 25-80 KM investments and programming to promote (see for example Henderson et al. 2012; Donaldson et al. 2016). This metric has been used in the past to assess the impact of refugees on host community welfare (Alix-Garcia et al. 2018). Data on nighttime host community livelihoods and income Source: Staff calculation using NOAA nightlight intensity. lights (NTL) intensity over time from the U.S. National Oceanic and Atmospheric Administration (NOAA) Note: Vertical line is on August 2017. To reduce variance generation. Efforts have been made to link and include zeros, calculations use the inverse hyperbolic the assistance given to the Rohingya to are combined with several additional kinds of data, including growth center locations from the Local sine (IHS) of total nightlight intensity within the Government Engineering Department (LGED); host population location and counts from Facebook’s and 500-meter buffer around each growth center. Results are spending in local markets supporting local CIESIN’s High Resolution Settlement Layer (HRSL); and recently displaced Rohingya population counts. consistent using a logarithmic 160 161 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC producers (WFP’s farmers’ market pilot offers one example). Such initiatives hold prom- CHAPTER 5. ise to boost spending in local markets in the district and generate better incomes for the host community. More work is needed now to quantify the benefits. If in-kind food aid can be procured locally in larger proportions, Bangladeshi producers of cereals, produce, Areas for policy action eggs, poultry, and fish can stand to gain. To achieve this, they will need to deliver quality goods at scale. Policy context, challenges, and opportunities The district of Cox’s Bazar has faced an unprecedented, exogenously driven increase in population density due to the 2017 Rohingya influx. This increase in density is not a nat- ural, endogenously driven outcome signaling agglomeration economies and urbanization benefits. In Bangladesh, the latter forces continue to be concentrated in the megacities of Dhaka and Chittagong and their surrounding areas. Therefore, the local economy of Cox’s Bazar district cannot naturally generate (nor should it be expected to) the types of jobs, incomes, or growth that would otherwise accompany such increases in density. The dis- trict’s potential for inclusive growth continues to be constrained by its lack of integration to the national economy and the latter’s growth drivers. The district is also poorly connected with growth sectors in economic terms, with the current economic structure comprising largely of low-productivity services and agriculture. Poor human capital and skills, a busi- ness environment that favors older, established, larger firms to the detriment of new, small firms, and barriers to women’s economic participation all limit the inclusivity of the current growth model. Consequently, local growth opportunities which leverage the district’s nat- ural endowments, such as tourism and aquaculture, remain largely unrealized. Addressing these growth challenges may also pave the way for an additional growth impetus from rising local demand for food and basic necessities to support the Rohingya population, and the increase in accompanying humanitarian assistance. Cox’s Bazar’s pre-existing human capital endowments and local economic structure can- not readily manage the increased population density or effectively translate a potential demand impetus into a realized boost in economic activity. A large share of the district’s 162 163 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON adult population is illiterate. Poor educational quality and financial constraints combine a major constraint to firms in Cox’s Bazar. Beyond the connectivity challenges discussed with other factors to prevent school-age children from attending school and completing above, the lack of locally available skilled labor may limit the ability of the economy to their education. Overall, Cox’s Bazar’s human capital endowment remains poor. The local effectively leverage promising geographic and economic endowments for tourism, hos- economy is largely reliant on low-productivity agriculture and services, with a small con- pitality, or aquaculture. Productive participation in the labor market is limited by low centration of manufacturing and large firms in the northern unions. Finally, as noted, the educational attainment, limited access to well-paying jobs, and physical distance from district’s promising geographic and economic endowments – for tourism and hospitality or the country’s growth poles. These constraints are more binding for women and are com- aquaculture, for instance – have not yet been effectively engaged. pounded by women’s unequal access to productive assets, as well as by prevailing norms about women’s role and mobility outside the home. Human capital deficits begin early in The district remains effectively distant and disconnected from the existing forces of life. The demographic profile of the district, with a large and growing young population, growth and income generation in Bangladesh in at least two ways. First, high travel underscores the need for investments in early childhood and expansion of basic services. times isolate Cox’s Bazar from the growth poles of Dhaka and Chittagong. Existing trans- If delivered equitably, these investments can address existing deprivations, and set the port infrastructure and associated costs (including congestion) make it difficult for firms to stage for greater productive potential. be based in Cox’s Bazar and for local workers to reach jobs outside the district. Within the district, the northern unions around Chakaria have some connectivity with Chittagong, but In light of these constraints, and based on the existing evidence base, this report iden- Teknaf and Ukhia, facing the brunt of the increased population density, will remain largely tifies four sets of key growth drivers in the district. These may be classified into major disconnected even after planned infrastructure upgrades are in place. growth drivers, which aim to leverage pre-existing growth opportunities in tourism and aquaculture and ease structural constraints to inclusive growth, and secondary growth Second, the private sector in the district is largely disconnected from the country’s drivers, which take advantage of emerging opportunities. growth model, which has relied on export-oriented, labor-intensive manufacturing. The readymade garment industry boom at the national level has largely left Cox’s Bazar The first major potential growth driver centers around the comparative advantage and behind. It is likely that, absent concerted effort, any new growth sectors which emerge in natural endowment of the district, i.e., growth opportunities related to tourism, hospi- the national economy will do the same. Large conglomerates in Bangladesh, including tality and aquaculture. Concerted efforts are needed to leverage the natural endowments in the manufacturing sector, are comprised of a few, very large, old, and connected fam- of the district, while ensuring and promoting ecological sustainability. Activating these ily firms, many of which are based in Dhaka and Chittagong (Genoni et al. forthcoming). growth opportunities will require a conducive business environment to attract investment Given the existing benefits to agglomeration in these two urban centers, the preferential and foster ecologically sustainable development choices. Investments in connecting and treatment these firms can obtain through special economic zones (SEZs), and their ability facilitating infrastructure will help develop value chains and linkages with the local econ- to inhibit the entry of smaller, younger firms, it is unlikely that these firms would readily omy. The local labor market will not naturally be able to take advantage of these new work move to Cox’s Bazar. Moreover, fledgling local comparative advantages in Cox’s Bazar, say opportunities, unless investments are made in building sector-specific skills, in collabora- in tourism, will need a policy shift that brings in foreign direct investment to yield benefits tion with the private sector. at larger scale. A regulatory framework will also need to be in place to ensure that invest- ments in this sector are ecologically sustainable and environmentally friendly. Similarly, The second major growth driver identified in the report is improved connectivity and for any expansion in the fisheries sector—for example, in shrimp exports—an appropriate accessibility within Cox’s Bazar, and from the district to the rest of the country. Currently, policy framework will need to be in place and effectively implemented to allow local indus- connectivity within the district and beyond is constrained by congestion, reliance on road try to meet export standards and certifications. The political economy considerations and transport rather than multiple modes, and limited capacity for high traffic and cargo vol- competing priorities that have so far limited action in this policy space will need to change umes. Planned infrastructure investments generally continue to focus on expanding con- for the status quo to shift. nectivity between and to Dhaka and Chittagong, limiting the likely benefits for residents of the district. More generally, the majority of Cox’s Bazar’s small, informal firms are disadvantaged by the challenging business environment at the national level, and limited access to Two secondary factors have the potential to shift the growth trajectory of Cox’s Bazar, capital, digital technologies, and a skilled labor force. At the national level, new firms, if these opportunities are carefully fostered and linked to the local economy and pop- including young, small establishments and investors trying to expand or start their busi- ulation. These are the Matarbari energy complex and deep seaport, and international ness, are disadvantaged by the barriers to entry and growth and the lack of a level play- humanitarian and development assistance for the Rohingya response. First, the large- ing field. Access to finance, and reliance on own sources of financing for businesses, are scale, capital-intensive investments planned in and around Matarbari will not naturally 164 165 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON create a large number of jobs, and certainly not for the host community. The energy com- local people’s capacities and skills, will open a wider set of economic opportunities for all plex and port will need to be better connected to Cox’s Bazar district both physically and in Cox’s Bazar. Government can play a critical role in coordinating private, public, human- in terms of employment potential. This will require careful identification, in collabora- itarian, and development actors to leverage local growth potential and help capitalize on tion with the private sector, of the skills profiles needed, together with investments in the district’s natural advantages. local skills development for jobs in and around the port, including in related value chains such as transport and storage. There is potential to link to the fledgling growth cluster Finally, data and evidence gaps will need to be filled to guide future policy and inter- in Chakaria and some northern unions by identifying and fostering forward and back- ventions. Ongoing research is needed to understand: (i) how the local economy is already ward linkages. Large, export-oriented firms remain unlikely to move a significant share evolving in response to the Rohingya influx; (ii) sector-specific challenges to growth for the of their operations to Matarbari unless the district (including Sadar upazila) secures the local private sector; and (iii) the potential for humanitarian and development interventions necessary infrastructure to be well connected to international markets and to Dhaka and to work at scale to improve the livelihoods of hosts and the displaced. Chittagong. Beyond investments in airport and road infrastructure, there is a need to upgrade standards in the hospitality sector to support a business clientele, including improved ICT services. Policy recommendations The Rohingya influx has been accompanied by a significant inflow of humanitarian and development assistance to the district. This report has shown evidence of increasing The report’s policy recommendations aim to foster inclusive economic growth in local growth near the Rohingya camps, as proxied by the increase in nighttime lights. Indeed, communities through four sets of interventions: (i) foundational, early investments in Cox’s Bazar is among a limited number of districts in Bangladesh that display some signs the productive potential of the district’s youngest residents; (ii) strengthening the pro- of growth in recent years. The increasing share of women in high-exposure areas working ductive potential of the district’s workforce; (iii) expanding opportunities for work and in the NGO and related services sector could signal the emergence of new types of work for economic participation; and (iv) bridging key evidence gaps. Policy recommendations the host community, generated by the aid economy. These are not necessarily restricted under points (i), (ii), and (iii) focus on areas where the evidence base is relatively solid: to the areas around the Rohingya camps in Teknaf and Ukhia. It is reasonable to assume the need for increasing the productive capacity of the population, the range of economic that the increased presence of humanitarian workers and organizations in the district will opportunities available, and investing in children early to redress lifelong inequality of lead to greater demand for housing, office space, transportation services, restaurants, and opportunity. More generally, these recommendations focus on ways to expand the eco- hospitality services, and for local facilitation such as translation services. nomic pie and level the playing field so that different groups can access opportunities, achieving a more equitable distribution of the benefits of growth. Finally, there are import- The humanitarian effort has already shifted towards direct investments and pro- ant data and evidence gaps, particularly related to the implications of the Rohingya influx gramming to promote host-community livelihoods and income generation. Increased for service delivery and labor market competition and opportunities, which will need to be efforts are underway to link the assistance given to the Rohingya to spending on local bridged to better understand how to orient the policy response. products in local markets. Development interventions by multilateral agencies such as the World Bank are designed to support both host communities and the displaced. This chapter uses the Green, Resilient, and Inclusive Development (GRID) framework By facilitating partnerships between the humanitarian community and government, to organize key policy recommendations. All three elements of the GRID framework are development agencies can support investments in service delivery and monitoring in particularly salient for Cox’s Bazar district. The district simultaneously faces grave risks the district, while strengthening national systems. Local government institutions need from the consequences of climate change and, at the same time, relies on its natural greater capacity in last-mile service delivery and advocacy for local people’s needs in capital and endowment for growth impetus. Environmental sustainability and climate development priorities. change prevention, mitigation, and adaptation must be central to its development strat- egy. Building resilience within the population and the economic structure to bear risk Taken together, these findings point to the need for a comprehensive, evidence-based, and uncertainty will be critical to the sustainability of any growth strategy. Last but not multi-sector approach to improve inclusive growth and welfare in Cox’s Bazar. This least, inclusion – in terms of access to services, jobs, and productive opportunities – is includes raising living standards by investing in portable assets such as health and edu- essential to the effectiveness of any growth strategy, so that the benefits of the latter are cation; removing distortions in the local investment climate; and creating a level playing widely accessible. field for the district’s private sector, with access to adequate services and infrastructure. Improving physical and economic connectivity to growth opportunities, while investing in 166 167 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON Figure5-1: Figure ES1. Key Key policy policy recommendations recommendations • Green – Invest in solar and wind-based energy generation to expand access to elec- tricity. Improve coordination between international organizations and local govern- Investing early Strengthening Expanding ment to expand programs and subsidies to increase the use of solar panels. Bridging in productive productive economic evidence gaps • Resilient – Modify the scheme of national electricity prices to achieve a cost recovery potential capacity opportunities rate, which is essential to the sustainability of the system. • Resilient – Strengthen local government mandates, allowing community prefer- Access to clean Human capital Private sector-led Disaggregated, water, improved and skills job creation timely, reliable data ences to be reflected in budget allocations and expenditures, particularly outside sanitation, and statistics electricity Municipal and City Corporations. • Resilient – Strengthen links and communication between local government entities Maternal and child Resilient Market integration New analytical work and humanitarian agencies to better align resource use with local needs. Stronger health livelihoods and connectivity links between local government entities and humanitarian agencies could help bet- ter align resource flows with local needs. A robust social contract connecting the Green Resilient Inclusive state, non-state actors, business, and local government institutions is essential to improve service delivery in Cox’s Bazar. Improved service delivery would benefit tourism and local trade and ultimately boost the national economy. Maternal and child health Early investments in productive potential • Resilient – Expand nutritional programs among hosts, including early detection of child malnourishment and programs in good nutrition practices for young mothers. Bolster awareness and adherence to vaccinations, along with pre- and post-natal With 40 percent of the district’s population made up of children below the age of 15, and care. This will increase resilience among vulnerable host households in the context half the Rohingya population below the same age, the demographic composition of the of COVID-19 in the short term, and of undernourishment in the medium term. local population has important implications for early childhood interventions, maternal • Inclusive – Increase coordination between humanitarian actors and local govern- and childcare, and access to basic services such as clean drinking water, improved sanita- ment to expand nutritional programs already present in camps to host communities tion, and electricity. With such a large young population, relevant interventions will benefit to guarantee access to basic nutrients for children. many. Investing early in children has been shown to have major benefits over the lifetime. • Inclusive – Expand social assistance support to female-headed households, particu- Moreover, such investments are strategic, because they address inequality of opportunity larly those headed by young mothers, so that they do not have to trade off caring for at birth and can set the foundations for a more productive and inclusive future. young children and earning a living. • Inclusive – Expand programs to close immunity gaps among children living in Access to clean water, improved sanitation, and electricity camps and protect against future infectious outbreaks through scale-up and • Inclusive – Expand access to private sources of clean water and reduce reliance on strengthening of routine immunization services. Despite repeated vaccination shared sources, particularly in host communities close to Rohingya camps. campaigns, immunity gaps persist among children living in camps. This is partic- • Inclusive – Broaden access to improved sanitation facilities across the district. ularly true of diseases like diphtheria, which require serial vaccinations to achieve • Inclusive – Increase water, sanitation, and hygiene (WASH) investments in camps maximum protection (Feldstein et al. 2020). There is a need to close immunity to reduce reliance on shared facilities. Within camps, overcrowding and poor san- gaps and protect against future outbreaks by expanding and strengthening rou- itation and housing conditions, including a reliance on shared facilities, remain a tine immunization. concern from a public-health perspective, particularly in the context of the COVID-19 pandemic. Strengthening productive capacity • Inclusive – Promote investments in distribution and transmission capacity to increase the number of hours of grid electricity across the district, and particularly Low educational attainment and persistent gender gaps in outcomes remain a concern in host communities close to camps. for policy action. Tertiary education is still limited, especially for women and poor house- holds. Opportunities exist to tackle economic constraints that make it difficult for low-in- come families to finance education expenses. Doing so may improve education indicators 168 169 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON for both females and males. Out-of-pocket expenditures on uniforms, tuition, books, and private-sector engagement and input into their own programming and operations transport make up a substantial proportion of necessary education costs for households. focused on skills, training, and employment. This way, partners will better under- This suggests that, by itself, a no-fee policy for public schools may not suffice to eliminate stand the actual demand for skills and be able to design more appropriate programs. financial constraints on education. Closing the gender gap might require a longer process, Employment-oriented skills and vocational training, particularly for younger cohorts considering that social norms discouraging female education are deeply rooted. Along with of better-educated labor market entrants, may need to be specifically designed for access, the quality of education will also need to be improved to promote better learning and driven by private-sector entities involved in: outcomes and build skills in demand in the private sector. • Infrastructure mega-projects, including the planned Matarbari deep seaport and energy hub At the same time, given the district’s high exposure to climate and environmental risk, • Special economic zones in Maheshkhali together with local people’s limited ability to absorb risk, investing in households’ capac- • Tourism and hospitality services ity to mitigate and manage risk will be important for income generation and livelihoods • Resilient – From a medium-term perspective, investing in market-relevant skills for diversification. Agricultural incomes will remain central to livelihoods and welfare in the migrants can also boost the potential of migration as a driver for improved welfare, district. Currently, heavy reliance on rice cultivation runs counter to the district’s agroeco- reducing the pressure on local labor markets. logical advantages but is consistent with the more favorable policy environment for rice. • Inclusive – There is a clear need for psychosocial support to Rohingya youth and Complementary investments to increase agricultural productivity outside of rice could pro- adolescents, expanded access to and awareness of sexual and reproductive health mote a more diversified cropping pattern, including higher-value crops. As the latter are services, and continued support to survivors of sexual and gender-based violence likely to be perishable, mitigating risks associated with such diversification will be essential and trauma. for the strategy to bear fruit. • Inclusive – Services must be designed to fully cover current cohorts of children and deliver missed vaccine doses to older children. More generally, given risk factors and Human capital and skills living conditions in camps, a system for continuous real-time health surveillance is • Inclusive – Provide (pro-poor) scholarships to women and economically disadvan- needed in the medium term. taged students at secondary and higher level (Bhatta et al. 2019). • Inclusive – Implement the already-designed pilot initiative to provide secondary • Inclusive – Conditional cash transfers may be targeted to areas and social groups education for grades 6 to 9 in Myanmar language. This will expand access to edu- with lower educational attainment and higher risk of dropping out – including chil- cation for secondary school-aged adolescents. Overall, longer-term policy dialogue dren in high-exposure areas, girls, and teenagers. would need to address such issues as: (i) limited supply of trained teachers and • Inclusive – Strengthen measures to enhance the quality of the school learning envi- spaces for learning centers; (ii) consistency of funding; (iii) cultural restrictions that ronment and improve the teacher-student ratio. These steps can boost the quality discourage adolescent girls from attending school; and (iv) skills development and of education and reduce dropouts. opportunity to engage youth in the Rohingya camps. • Inclusive – Pilot and expand implementation of the Myanmar curriculum for • Inclusive – Adopt the Learning Passport model to formalize GoB commitment to Rohingya children in camps, while easing mobility and safety concerns to increase extend curriculum and develop inclusive, quality education and skills-development enrollments. Provide incentives to keep young adolescents and youth enrolled for adolescents and youth. The development partnership is based on a Rohingya in school. Gradually increase the number of grades of education with minimum Education Response Plan, endorsed by the major partners. quality standards and trained educators, continue nutrition support programs, • Inclusive – Expand the World Bank’s Reaching Out-of-School Children (ROSC II) and expand school feeding. Implement education certification for primary and project. So far, around 314,926 children and youth are studying at 3,236 Learning secondary school completion. Centers in 32 camps, based on an agreed Learning Competency Framework and • Inclusive – As the crisis transitions to the next phase, humanitarian and development Approach (LCFA) known as GIEP (Guideline on Informal Education Program). organizations, as well as the GoB, need to consider how education and vocational programs can be more inclusive of host community and Rohingya youth. The best Resilient livelihoods solutions may link youth-focused programs to productive opportunities for work. • Resilient – Financial and insurance instruments need to evolve, if they are to be • Inclusive – More generally, the education system may not currently be geared to meaningful sources of credit and insurance. invest in market-relevant skills for the local labor market, particularly tourism and • Reliance on property as collateral makes it harder for small farmers and micro hospitality-related services, as well as the international labor market. Development and small enterprises to access credit. This especially affects women and landless partners and multilateral agencies can support the GoB by incorporating 170 171 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON households. Expanding the collateral registry’s mandate to include movables and (through mobilization, outreach, and grievance-redress activities); contribute to cli- immovables as collateral will help expand access to credit. mate and environmental risk mitigation; improve camp living conditions through • Another key constraint is the lack of access to insurance instruments, particularly cleaner environments; and prevent anti-social behavior. weather-risk insurance, which is salient for Cox’s Bazar. • Resilient – Specific policy recommendations from the Bangladesh Rural Income • Green – From a medium-term perspective, better environmental and forest man- Diagnostic (Genoni et al. forthcoming) may be particularly relevant, including: agement is critical to managing risk in Cox’s Bazar. Fuelwood is the single greatest • Review and reform input subsidies policies with a special focus on fertilizers, source of household energy for hosts and initially also for the Rohingya, which led while complementing with extension services to promote more efficient use to increasing deforestation. Subsequently, the humanitarian effort has expanded of fertilizers. These actions could help lift two important constraints to reduc- access to alternative energy sources in camps, but hosts continue to rely on tradi- ing yield gaps at the farm field, especially for Boro paddy: (i) the overuse and tional sources such as fuelwood. imbalanced use of fertilizer, resulting in declining soil fertility; ii) inadequate farm • Green – The climate and topography of Cox’s Bazar mean local communities are knowledge and practices. exposed to multiple natural hazards and experience recurring extreme weather • Expand use of mechanization for seed establishment, crop protection, irriga- events. Vulnerable Bangladeshi communities in the district have long borne the tion (particularly high-efficiency irrigation technologies), and harvesting opera- brunt of cyclones, landslides, and flash floods. The Rohingya crisis has increased tions. Evidence on constraints affecting agriculture mechanization is outdated. the size of the population at risk and is creating new risks due to deforestation, Revisiting government subsidy and trade policies in setting incentives for invest- hill-cutting, and pressure on infrastructure. Ongoing efforts need to be supple- ment can be important. mented by medium-term responses to increase the resilience of local communities. • Improve irrigation systems. In Cox’s Bazar, tube well systems cover only 21 per- For instance, the Delta Management Plan points out the need to extend and improve cent of irrigated land, while 69 percent is irrigated with power pumps, and 10 per- cyclone shelters and strengthen anti-flood embankments. cent still relies on traditional irrigation methods (BBS 2018c). The lack of irriga- • Resilient – Climate change and the increasing number of extreme weather events tion may limit investments in alternative crops such as betel nut and leaf, which will require farmers to change or adapt their current cropping systems, as well as yield a higher return per unit of area cultivated than rice. The district appears to their fisheries activities. Complementary to policies to enhance crop diversification have an agroecological advantage in the crop, but it also needs frequent watering and income generation, actions are needed to improve resilience in the agricultural and fertilizer application, which may limit its adoption. Slopes are better suited sector. Examples include: for betel cultivation, so an expansion may not need to be at the cost of rice culti- • Developing field trials of climate-resilient cropping patterns and associated vation, but deforestation risks will have to be carefully managed. water management systems • Introducing technologies to adapt aquaculture activities, given increasing salin- Expanding economic opportunities ity in marine fisheries. • Green – Cox’s Bazar has an opportunity to invest in new renewable energy sources National-level constraints to firm growth and job creation are even more salient in Cox’s to boost the reliability and quality of electricity access, transition to a cleaner energy Bazar. These include lack of access to finance, limited technology adoption, and the need mix, and reduce costs. Options include using small-scale grids to expand access to to upgrade firm and entrepreneur capabilities. The lack of a level playing field and fair, remote and marginal areas; leveraging wind and solar energy sources; and limiting transparent regulatory regime has long posed a challenge to small- and medium-scale households’ reliance on firewood by investing in improved cookstoves.111 enterprises. Growth has hitherto been led by the export-oriented readymade garment • Resilient – Implement the Emergency Multi-Sector Rohingya Crisis Response industry, which has benefited from export processing zones and special industrial zones (EMRCR) and Safety Net System for Poorest (SNSP) projects, both of which have with guaranteed access to necessary infrastructure and a streamlined regulatory environ- components that address the economic and social resilience of the vulnerable ment. Cox’s Bazar is a case in point, where the absence of a vibrant RMG industry is accom- through their engagement in community services and workfare schemes. Under panied by a relatively stagnant private sector and few alternative sources of dynamism and these schemes, Rohingya households will participate in subprojects and activities job creation. The high degree of informality and the dominance of very small 1-2 person intended to: enhance community services for the vulnerable (including at-risk youth, enterprises also exemplify the constraints to growth and the limited perceived benefits of women, disabled persons, and the elderly); strengthen engagement mechanisms formalization. Most enterprises appear to be involved in secondary or subsistence activi- ties to complement household income, rather than entrepreneurial activities. Policy action will be needed to ensure that the private sector in Cox’s Bazar has a level playing field and For case studies in refugee camps, see the 2019 IFC report “Private Sector & Refugees: Pathways to 111 can take advantage of existing and emerging growth opportunities. Scale.” 172 173 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON There is significant potential to expand the local multiplier effect of humanitarian assis- infrastructure work is designed to help reduce climate vulnerability and disaster tance in Cox’s Bazar. Doing so can boost local incomes and livelihoods for hosts, while risks. Engaging working-age youth can contribute towards improved mental and improving sustainability and reducing the costs associated with delivering assistance. emotional wellbeing through participation in labor-intensive activities that also A new approach to the Rohingya response can provide hosts and the displaced with serve to enhance camp livability. enhanced services and economic opportunities, while strengthening links with national • Inclusive – Expanding the share of locally sourced and procured food assistance and local government institutions. delivered in Rohingya camps is more cost effective and allows for a diversification of the food basket for the Rohingya. Equally importantly, it creates a potentially large Finally, there is a clear need to invest in upgrading the transport network to foster effective source of demand for agricultural products in the local market. This is particularly connectivity within Cox’s Bazar and between the district and Bangladesh’s main growth true for seasonal and perishable produce and can provide an important boost for centers. These are even more important to ensure broad access to basic services, and to fos- local livelihoods. WFP has already made a number of changes in this regard, which ter backward and forward linkages of new growth opportunities within the local economy. should be pursued once COVID-19-related restrictions are lifted. Development assistance in Cox’s Bazar could also incentivize implementing agencies and service Firm growth and private sector-led job creation providers to expand the use of locally sourced products and hire qualified local • Resilient/Green – Promote foreign direct investment in tourism and hospitality. personnel. Concerted effort is needed, including in marketing and environmentally sustainable • Inclusive – The reliance of the Rohingya on assistance can be a source of significant tourism infrastructure and planning, if the potential of this sector is to be realized. and relatively consistent demand for locally produced food and non-food items. For • Resilient – Upgrade infrastructure and ICT services for the international business the host community, a larger local market reduces transaction costs and the costs clientele. of marketing for perishable products. This can enable diversification and encourage • Resilient – Fishing and aquaculture development could be fostered, if complemen- farmers to invest returns in productive improvements. tary investments are made to facilitate storage, transport, marketing, and quality • Resilient – The private sector can be engaged in humanitarian assistance. Firms may and standards assurance and certification. Relatedly, there is potential to develop share technological capabilities and expertise, adapt business models to sell goods the Vannamei shrimp variety, which is less prone to disease than the dominant tiger and services to the Rohingya, and integrate into value chains by working with both shrimp variety, but this would require international certification (IFC 2020). host community and Rohingya enterprises (IFC 2019). • Resilient – More generally, realizing the potential of aquaculture, particularly as an • Inclusive - Bangladesh’s pioneering role with micro-finance opens the door to export-oriented growth sector, will likely require a comprehensive approach. This explore pioneer micro-leasing of productive assets. This could potentially be com- may include bringing in foreign investment and expertise; establishing clear stan- bined with more attention to platform economies (such as Uber), which have the dards and implementing them uniformly; expanding access to technological know- potential to unite the supply by many small/micro businesses, for instance to create how; providing certification facilities; and vertical integration or cooperatives to access to export markets for firms that would otherwise be too small to access such expand the effective scale, market access, and investment ability of the numerous sophisticated markets. small producers in the sector. • Inclusive – Chakaria and the surrounding northern unions could emerge as a hub for Market integration and connectivity non-agricultural economic activity based on their current comparative advantage, conditional on resolving connectivity challenges. • Inclusive – Connectivity investments focused on upgrading existing networks will be • Inclusive – Services are a potential source for additional self-employment, particu- important. These can reduce the cost of accessing jobs, inputs, and markets for local larly among the better educated. Business and vocational skills programs may help people and firms. They can also better connect the southern parts of the district to foster some of these nascent activities. the more economically vibrant northern unions. • Inclusive – Build backward and forward linkages with the Matarbari investments and • Resilient – Well-documented national-level logistics constraints are equally import- international humanitarian and development efforts. ant in Cox’s Bazar. Priority issues include: • Inclusive – The Emergency Multi-Sector Rohingya Crisis Response (EMRCR) and • The policy and regulatory framework on infrastructure development should Safety Net System for Poorest (SNSP) projects have developed community work- expand to encompass integrating multiple modes of transport, improving the fare schemes. These programs aim at reducing the likelihood of at-risk youths’ quality of services, and improving road safety. participation in anti-social behavior by engaging them in workfare on basic infra- • Logistics services markets will benefit from a more level playing field and greater structure maintenance, as well as camp cleaning and maintenance activities. The competition among industry players. 174 175 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC C h a p t e r 5 – AREAS FOR P O L I C Y A C T I ON • Road cargo transport remains critical for domestic and international trade link- management and sanitation. More work is needed to generate evidence for an ages. Important agendas include increasing capacity to track and monitor ship- appropriate policy response. ments, ensuring that trucks meet quality and safety standards, and ensuring that • The weak mandates of local governments limit the degree to which community pref- trucks are appropriately loaded. These actions will help optimize investments in erences are reflected in budget allocations and expenditures, particularly outside transport infrastructure and reduce logistics costs at ports and major trade hubs. Municipal and City Corporations. The limited ability to disaggregate public expen- • Inclusive – Expanding access to and quality of digital infrastructure in the district diture data to different sub-national levels constrains analysts’ capacity to correlate through fiber-optic infrastructure, 4G capacity expansion, and telecom towers will expenditures with outcomes. Understanding the efficiency and efficacy of public be particularly important, if Cox’s Bazar is to leverage new growth opportunities. expenditures is critical to bridge gaps in achieving the Sustainable Development • Inclusive – There is a need for a comprehensive policy agenda to expand the cov- Goals (SDGs). This applies across Bangladesh’s districts and regions, and across erage, access, and use of digital technologies by households, firms, and farms. If upazilas and unions within districts. Efforts to address the SDG challenge will need successful, the agenda can increase these actors’ potential access to markets, to be based on intra- and inter-regional comparisons. enhance their capacity to manage risks, and ultimately boost their productivity and income-generation capacity. Bridging evidence gaps • Investments in data and evidence are needed to assess how the local labor market is evolving in response to the aid economy and seize emerging opportunities. A second phase of the present diagnostic work will attempt to fill some of the gaps and identify locally binding constraints to firm entry, growth, and dynamism. In addition to iden- tifying current and potential channels for the growth of businesses and jobs linked to the assistance economy, data and evidence are needed to quantify the job creation related to government, humanitarian, and development investments in the district. • While there is some suggestive evidence of increased economic activity around camps, more data and evidence are needed to understand how the influx of human- itarian assistance has affected local host communities: for example, by poten- tially increasing competition for low-skill jobs, while also providing new work and income-earning opportunities for hosts, including better-educated youth. • More work is needed to confirm whether Cox’s Bazar truly has a comparative advan- tage in salt extraction and what scope there is to promote this activity. • Given the generally low education and skill levels of the Rohingya, there may be opportunities to expand their engagement in labor-intensive activities in farming, construction, and environmental restoration in or near camps, without creating local competition. New data collection and analytical work can help understand the scope for these types of activities. • Analysis finds suggestive evidence of higher male dropouts in tertiary education in areas of high exposure to Rohingya camps. More work is needed to assess if this reflects more local job opportunities in the camp and humanitarian-aid economy for relatively well-educated hosts. While dropping out of school for these types of aid-economy jobs may make economic sense in a context of low returns to educa- tion, the sustainability of these jobs is an open question. • Pressure from domestic tourism and the influx of humanitarian assistance in the district pose pressures for the delivery of urban services, including solid waste 176 177 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References ADB (Asian Development Bank). 2016. “Proposed Multitranche Financing Facility and Technical Assistance Grant People’s Republic of Bangladesh: South Asia Subregional Economic Cooperation Chittagong-Cox’s Bazar Railway Project. Phase 1.” Mandaluyong City, Philippines: ADB. Ahmed, Dr Munir, Nazrul Islam, and Md Shamsuddoha. 2007. “Value Chain Analysis in the Dry Fish Production and Marketing of Post Harvest Fishery Products (PHFP) in the Coastal Belt of Bangladesh.” Ahmed, Faizuddin, et al. 2010. Poverty maps of Bangladesh 2010 (Vol. 2). Technical report. Washington, DC: World Bank Group. Ainul, Sigma, Iqbal Ehsan, Eashita Haque, Sajeda Amin, Ubaidur Rob, Andrea Melnikas, and Joseph Falcone. 2018. Marriage and Sexual and Reproductive Health of Rohingya Adolescents and Youth in Bangladesh: A Qualitative Study. Population Council. Alam, Akhtar, Peter Sammonds, and Bayes Ahmed. 2019. “Cyclone Risk Assessment of the Cox’s Bazar District and Rohingya Refugee Camps in Southeast Bangladesh.” Science of the Total Environment 704, Article 135360. 10.1016/j.scitotenv.2019.135360. Alix-Garcia, J., E. Artuc, and H. Onder. 2017. “The Economics of Hosting Refugees: A Host Community Perspective from Turkana.” Working paper. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/26269 Alix-Garcia, Jennifer, Sarah Walker, Anne Bartlett, Harun Onder, and Apurva Sanghi. 2018. “Do Refugee Camps Help or Hurt Hosts? The Case of Kakuma, Kenya.” Journal of Development Economics 130(C):66–83. Anderson, Siwan, and Mukesh Eswaran. 2009. “What Determines Female Autonomy? Evidence from Bangladesh.” Journal of Development Economics 90(2):179–91. 178 179 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References Barker, Vincent L., and George C. Mueller. 2002. “CEO Characteristics and Firm R&D Bhattacharjee, Amitab, Mohammad Rashed Hasan Polas, and Lutfor Rahman. 2018. Spending.” Management Science 48(6):782–801. “Challenges and Prospects of Tourism in Cox’s Bazar: An Empirical Study.” BBS (Bangladesh Bureau of Statistics). 2010. “Census of Agriculture 2008.” Dhaka: Black, Robert E., Cesar G. Victora, Susan P. Walker, Zulfiqar A. Bhutta, Parul Christian, Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Mercedes de Onis, Majid Ezzati, Sally Grantham-McGregor, Joanne Katz, Reynaldo Planning, Government of the People’s Republic of Bangladesh. Martorell, and Ricardo Uauy. 2013. “Maternal and Child Undernutrition and Overweight in Low-Income and Middle-Income Countries.” The Lancet 382(9890):427– -----. 2013. “District Statistics 2011. Cox’s Bazar.” Dhaka: Bangladesh Bureau of Statistics, 51. doi: 10.1016/S0140-6736(13)60937-X. Statistics and Informatics Division, Ministry of Planning, Government of the People’s Republic of Bangladesh. Blankespoor, Brian, M. Shahe Emran, Forhad Shilpi, and Lu Xu. 2018. “Transport Costs, Comparative Advantage, and Agricultural Development: Evidence from Jamuna -----. 2015. Population Projection of Bangladesh: Dynamics and Trends 2011-2061. Dhaka: Bridge in Bangladesh.” SSRN Electronic Journal 10.2139/ssrn.3176160. Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of the People’s Republic of Bangladesh. Blankespoor, Brian, and Nobuo Yoshida. 2010. “Accessibility Methodology Bangladesh.” Washington, DC: World Bank. -----. 2018a. “Report on Agriculture and Rural Statistics 2018. Agriculture and Rural Statistics Survey (ARSS) Project 2017.” Dhaka: Bangladesh Bureau of Statistics, Statistics and Bloom, Nicholas, and John Van Reenen. 2010. “Why Do Management Practices Differ across Informatics Division, Ministry of Planning, Government of the People’s Republic of Firms and Countries?” Journal of Economic Perspectives 24(1):203–24. Bangladesh. Corna, Francesca, Fahmida Tofail, Mita RaniRoy Chowdhury, and Cécile Bizouerne. -----. 2018b. “The National Student Assessment 2017. Grade 3 and 5.” Dhaka: Bangladesh 2019. “Supporting Maternal Mental Health of Rohingya Refugee Women during the Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Perinatal Period to Promote Child Health and Wellbeing: A Field Study in Cox’s Bazar.” Government of the People’s Republic of Bangladesh. Intervention 17(2):160. doi: 10.4103/INTV.INTV_28_19. -----. 2018c. “Yearbook of Agricultural Statistics 2017.” Dhaka: Bangladesh Bureau of Deichmann, U. 1997. “Accessibility Indicators in GIS.” New York: United Nations, Statistics Statistics, Statistics and Informatics Division, Ministry of Planning, Government of the Division, Department for Economic and Social Information and Policy Analysis. People’s Republic of Bangladesh. Delmelle, Elizabeth Cahill, and Irene Casas. 2012. “Evaluating the spatial equity of bus -----. 2019. “Statistical Yearbook Bangladesh 2018.” Dhaka: Bangladesh Bureau of Statistics, rapid transit-based accessibility patterns in a developing country: the case of Cali, Statistics and Informatics Division, Ministry of Planning, Government of the People’s Colombia.” Transport Policy 20(C): 36-46. Republic of Bangladesh. Department of Fisheries Bangladesh. 2017. “Yearbook of Fisheries Statistics Of-Bangladesh BBS and UNICEF (Bangladesh Bureau of Statistics and United Nations Children’s Fund). 2015-2016.” Dhaka, Bangladesh: Government of the People’s Republic of Bangladesh. 2015. “Multiple Indicator Cluster Survey 2012– 2013.” Dhaka, Bangladesh: BBS. Donaldson, Dave. 2018. “Railroads of the Raj: Estimating the Impact of Transportation Beck, T. A., A. Demirguc-Kunt, and V. Maksimovic. 2002. Financing Patterns around the Infrastructure.” American Economic Review 108(4–5):899–934. World: The Role of Institutions. Washington, D.C.: World Bank. Ellis, Peter, and Mark Roberts. 2016. “Leveraging Urbanization in South Asia: Managing Bhatta, Saurav Dev, Maria Eugenia Genoni, Uttam Sharma, Buyant Erdene Khaltarkhuu, Spatial Transformation for Prosperity and Livability.” Washington, DC: World Bank. Laura Maratou-Kolias, and T. M. Asaduzzaman. 2019. “Bangladesh Education Sector https://openknowledge.worldbank.org/handle/10986/22549 Public Expenditure Review.” Dhaka, Bangladesh: Government of the People’s Republic of Bangladesh. 180 181 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References FAO (Food and Agriculture Organization of the United Nations). 2019. “FAO Fisheries Sector GFDRR (Global Facility for Disaster Reduction and Recovery). 2018. “Rohingya Crisis 2017- Support Programme.” Rome: FAO. 2018. Draft Rapid Impact, Vulnerability and Need Assessment.” Washington, DC: GFDRR. Farag, Hisham, and Chris Mallin. 2018. “The Influence of CEO Demographic Characteristics GoB (Government of Bangladesh). 2005. “Rural Road Master Plan 2005.” Dhaka, Bangladesh: on Corporate Risk-Taking: Evidence from Chinese IPOs.” The European Journal of Government of the People’s Republic of Bangladesh. Finance 24(16):1528–51. Geurs, Karst, and Bert Wee. 2004. “Accessibility evaluation of land-use and transport strat- Farole, Thomas, and Yoonyoung Cho. 2017. “Job Diagnostic Bangladesh.” Washington, DC: egies: Review and research directions.” Journal of Transport Geography 12: 127-140. World Bank. 10.1016/j.jtrangeo.2003.10.005. Feldstein, L.R., S.D. Bennett, C.F. Estivariz, G.M. Cooley, L. Weil, et al. 2020. “Vaccination cov- Gordon, Eleanor, Hannah Jay, and Katrina Lee-Koo. 2018. “Adolescent Rohingya Girls in erage survey and seroprevalence among forcibly displaced Rohingya children, Cox’s Bangladesh: One Year On.” Policy brief. Clayton, Australia: Plan International/Monash Bazar, Bangladesh, 2018: A cross-sectional study.” PLOS Medicine 17(3): e1003071. University. http://docs.wixstatic.com/ugd/b4aef1_5c13068eb067467587b7a2420b- https://doi.org/10.1371/journal.pmed.1003071 c6ecad.pdf Filipski, Mateusz J., Gracie Rosenbach, Ernesto Tiburcio, Paul Dorosh, and John Goswami, Grover, D. Medvedev, and E. Olafsen. 2019. High-Growth Firms: Facts, Fiction, and Hoddinott. 2020. “Refugees Who Mean Business: Economic Activities in and around Policy Options for Emerging Economies. Washington, DC: World Bank. the Rohingya Settlements in Bangladesh.” Journal of Refugee Studies feaa059. doi: 10.1093/jrs/feaa059. Guglielmi, Silvia, Nicola Jones, Jennifer Muz, Sarah Baird, Khadija Mitu, and Muhammad Ala Uddin. 2020. “‘How Will My Life Be?’: Psychosocial Well-Being among Rohingya Fox, William F., and Balakrishna Menon. 2008. Decentralization in Bangladesh: Change Has and Bangladeshi Adolescents in Cox’s Bazar.” London: GAGE. Been Elusive. Paper 0829. International Center for Public Policy, Andrew Young School of Policy Studies, Georgia State University. Haltiwanger, J., R. S. Jarmin, and J. Miranda. 2013. “Who Create Jobs? Small versus Large versus Young.” Review of Economics and Statistics 95(2):347–61. Gautam, Madhur, and Rashid Faruqee. 2016. Dynamics of Rural Growth in Bangladesh: Sustaining Poverty Reduction. Washington, DC: World Bank. Haltiwanger, John, R. S. Jarmin, R. Kulick, and J. Miranda. 2017. “Contribution to Job, Output, and Productivity Growth.” Pp. 11–62 in Measuring Entrepreneurial Businesses: Genoni, Maria Eugenia. forthcoming. “Survey Design and Sampling: A Methodology Note Current Knowledge and Challenges, National Bureau of Economic Research Studies in for the 2019 Cox’s Bazar Panel Survey.” Washington, DC: World Bank. Income and Wealth. Chicago: University of Chicago Press. Genoni, Maria Eugenia, Saurav Dev Bhatta, Uttam Sharma, Buyant Erdene Khaltarkhuu, Hasan-ul-Bari, S. M., and Tarek Ahmed. 2018. “Ensuring Sexual and Reproductive Health and Laura Maratou-Kolias. 2019. “Equity in Education Outcomes and Spending In and Rights of Rohingya Women and Girls.” The Lancet 392(10163):2439–40. doi: Bangaldesh.” In Bangladesh Poverty Assessment. Facing old and new frontiers in pov- 10.1016/S0140-6736(18)32764-8. erty reduction. Volume II. Washington, DC: World Bank. Hassan, Mohammad Mehedy, Audrey Culver Smith, Katherine Walker, Munshi Khaledur Genoni, Maria Eugenia, Madhur Gautam, Mansur Ahmed, and Pablo Tillan. Forthcoming. Rahman, and Jane Southworth. 2018. “Rohingya Refugee Crisis and Forest Cover “Rural Income Diagnostic. Enabling Faster and More Equal Income Growth in Rural Change in Teknaf, Bangladesh.” Bangladesh.” Washington, DC: World Bank. Henderson, J. Vernon, Adam Storeygard, and David N. Weil. 2012. “Measuring Economic Germanwatch. 2020. “Global Climate Risk Index 2020.” Retrieved (https://germanwatch. Growth from Outer Space.” American Economic Review 102(2):994–1028. doi: 10.1257/ org/en/cri). aer.102.2.994. 182 183 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References Herrera Dappe, Matias, Charles Kunaka, Mathias Lebrand, and Nora Welsskopf. 2020. Kersten, Renate, Job Harms, Kellie Liket, and Karen Maas. 2017. “Small Firms, Large “Moving Forward Connectivity and Logistics to Sustain Bangladesh’s Success.” Impact? A Systematic Review of the SME Finance Literature.” World Development Washington, DC: World Bank. 97(C):330–48. Hill, Ruth, and Maria Eugenia Genoni. 2019. “Bangladesh Poverty Assessment. ‘Facing Old Khan, Niaz Ahmed. 2016. “Challenges and Trends in Decentralised Local Governance in and New Frontiers in Poverty Reduction.’” Washington, DC: World Bank. Bangladesh.” ISAS Working Papers 222:19. Hussain, Imran Bin, and Nobel Mallick. 2017. “Changes of Land Use with Urban Growth - A Kyungmi, Kim, Uysal Muzaffer, and Joseph Sirgy. 2013. “How Does Tourism in a Community Study on Cox’s Bazar, Bangladesh.” International Journal of Scientific & Engineering Impact the Quality of Life of Community Residents?” International Journal of Tourism Research 8(4):6. Management 36:527–40. IFC (International Finance Corporation). 2019. Private Sector and Refugees. Pathways to Lemma, Alberto Francesco, Maria Quattri, Jessica Hagen-Zanker, Selim Raihan, and Abu Scale. Washington, D.C: International Finance Corporation. Eusuf. 2018. “Bangladesh Economic Dialogue on Inclusive Growth.” EDIG Research Reports No. 4, “Strategies for Inclusive Growth in Cox’s Bazar” (4):68. -----. 2020. Bangladesh Country Private Sector Diagnostic. Creating Market in Bangladesh. Unleashing the Private Sector to Sustain Development Success. Washington, D.C: Lopez-Pena, Paula, C. Austin Davis, A. Mushfiq Mobarak, and Shabib Raihan. 2020. World Bank. “Prevalence of COVID-19 Symptoms, Risk Factors, and Health Behaviors in Host and Refugee Communities in Cox’s Bazar: A Representative Panel Study.” Policy Brief. New IOM (International Organization for Migration). 2018. “Rapid Mental Health and Psychosocial Haven, CT: Yale Y-RISE. http://yrise.yale.edu/prevalence-of-covid-19-symptoms-risk- Needs Assessment.” Geneva: IOM. factors-and-health-behaviors-in-host-and-refugee-communities-in-coxs-bazar-a- representative-panel-study/ IOM and ACAP. 2020. “Different and Unequal. How COVID-19 Affects Different Sex, Age, Ability and Populations Differently.” Geneva: IOM. Magoutas, A., G. Agiomirgiannakis, and T. Papadogonas. 2011. “Education and firm perfor- mance: Empirical evidence from Greece.” International Journal of Economic Research, ISCG (Inter Sector Coordination Group). 2017. “Situation Report: Cyclone Mora.” Cox’s 8 (2), 141-152. Bazar, Bangladesh: ISCG. Mahmud, Shakil. 2018. “Frozen Foods: Emerging Influential Export Earning Sector for JICA (Japan International Cooperation Agency). 2016. “Preparatory Survey on the Cross- Bangladesh.” Border Road Network Improvement Project. Final Report.” Dhaka, Bangladesh: JICA. Malhotra, M., Y. Chen, A. Criscuolo, Q. Fan, I. Hamel, and Y. Savchenko. 2007. Expanding -----. 2018a. “Preparatory Survey on the Matarbari Port Development in People’s Republic Access to Finance: Good Practices and Policies for Micro, Small and Medium Enterprises. of Bangladesh. Volume 1.” Dhaka, Bangladesh: JICA. Washington, D.C: World Bank. -----. 2018b. “Preparatory Survey on the Matarbari Port Development in People’s Republic Mathon, D., P. Apparicio, and U. Lachapelle. 2018. “Cross-border spatial accessibility of of Bangladesh. Volume 3.” Dhaka, Bangladesh: JICA. health care in the North-East Department of Haiti.” Int J Health Geogr 17, 36. https:// doi.org/10.1186/s12942-018-0156-6 Jovanovic, Boyan. 1982. “Selection and the Evolution of Industry.” Econometrica 50(3):649–70. Ministry of Environment and Forest, UNDP, and UN WOMEN. 2018. “Environmental Impact of Rohingya Influx.” Dhaka, Bangladesh: Ministry of Environment and Forest, Kaur, Rupinder, and Balwinder Singh. 2018. “CEO Characteristics and Firm Performance: A Government of Bangladesh. Study of Indian Firms.” Indian Journal of Corporate Governance 11(2):185–200. 184 185 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References Ministry of Health and Sports, Myanmar. 2017. Myanmar Demographic and Health Survey Shelly, A.B., and M. D’Costa. 2001. “Women in aquaculture: Initiatives of CARITAS 2015-2016. Naypyidaw, Republic of the Union of Myanmar: Ministry of Health and Sports. Bangladesh.” Global Symposium on Women in Fisheries, Sixth Asian Fisheries Forum, 29 November, Taiwan, China. Munoz-Raskin, Ramon, 2010. “Walking accessibility to bus rapid transit: Does it affect prop- erty values? The case of Bogotá, Colombia.” Transport Policy 17(2): 72-84. Summers, A., A. Humphreys, E. Leidman, et al. 2018. “Notes from the Field: Diarrhea and Acute Respiratory Infection, Oral Cholera Vaccination Coverage, and Care-Seeking Naureen, M., D. S. Kabir, M. S. Ali, and M. Kamal. 2006. “Status of the Shrimp Sector in Behaviors of Rohingya Refugees - Cox’s Bazar, Bangladesh, October-November 2017.” Bangladesh in the Context of HACCP and Trade Issues: A Review.” Morb Mortal Wkly Rep 2018;67(18):533-535. doi:10.15585/mmwr.mm6718a6 OECD (Organisation for Economic Cooperation and Development). 2006. “The SME Tallis, Heather, Cindy Huang, John Herbohn, Karen Holl, Sharif A. Mukul, and KAM Morshed. Financing Gap: Theory and Evidence.” Financial Market Trends 2006(2):89–97. 2019. “Steps Toward Forest Landscape Restoration in the Context of the Rohingya Influx.” Oishi, Shafiqua, and Nazmul Alam. 2020. “Risk factors associated with Acute Respiratory Infection (ARI) among children under 10-years in Rohingya refugee camp.” Recent Tobler, W. 1993. “Three Presentations on Geographical Analysis and Modeling: Non- Research in Science and Technology. 32-35. Isotropic Geographic Modeling; Speculations on the Geometry of Geography; and Global Spatial Analysis (93-1).” UC Santa Barbara: National Center for Geographic Páez, Antonio, Darren Scott, and Catherine Morency. 2012. “Measuring Accessibility: Information and Analysis. Retrieved from https://escholarship.org/uc/item/05r820mz Positive and Normative Implementations of Various Accessibility Indicators.” Journal of Transportation Geography: online first. doi: 10.1016/j.jtrangeo.2012.03.016 Toufique, K., and N. Ahmed. 2014. “Barriers to Developing the Fisheries Sector in Bangladesh.” Bangladesh Institute for Development Studies 1407. Pascaud, Yannick, and Rafael Panlilio. 2019. “Education Needs Assessment (ENA): Rohingya Crisis Response.” Washington, DC: World Bank. Truelove, S., O. Abrahim, C. Altare, S.A. Lauer, K.H. Grantz, A.S. Azman, et al. 2020. “The potential impact of COVID-19 in refugee camps in Bangladesh and beyond: A model- Plummer, Michael G., Peter J. Morgan, and Ganeshan Wignaraja, eds. 2016. Connecting Asia: ing study.” PLoS Med 17(6): e1003144. https://doi.org/10.1371/ journal.pmed.1003144 Infrastructure for Integrating South and Southeast Asia. Cheltenham: Edward Elgar. UNDP (United Nations Development Program). 2018. “Impacts of the Rohingya Refugee Raihan, Selim, and Haque Bidisha. 2018. “Female Employment Stagnation in Bangladesh.” Influx on Host Communities.” New York: UNDP. RHD (Roads and Highways Department). 2019. “Initial Environmental Examination UNDP and UN Women. 2017. “Rohingya Crisis in Bangladesh: Early Recovery Joint Emergency Assistance Project: Improvement of National Highway (N1) from Link Assessment.” New York: UNDP. Road (Cox’s Bazar) to Teknaf.” Dhaka, Bangladesh: RHD. UNHCR (United Nations High Commissioner for Refugees). 2018. “Operational Update 5-20 RHD, ADB (Roads and Highways Department and Asian Development Bank). 2018. 2018.” Geneva: UNHCR. “Resettlement Plan for Cox’s Bazar-Teknaf Road N1.” Dhaka, Bangladesh: RHD. -----. 2019. “Joint Response Plan for Rohingya Humanitarian Crisis. January December Ripoll, Santiago. 2017. “Social and Cultural Factors Shaping Health and Nutrition, Wellbeing 2019.” Geneva: UNHCR. and Protection of the Rohingya within a Humanitarian Context.” -----. 2020. “Joint Response Plan for Rohingya Humanitarian Crisis. January-December Schiffer, M., and Beatrice Weder di Mauro. 2001. “Firm Size and the Business Environment: 2020.” Geneva: UNHCR. Worldwide Survey Results.” Working Paper. World Bank - International Finance Corporation, Washington, DC. 186 187 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC References UNICEF (United Nations Children’s Fund). 2017. “Progress on Drinking Water, Sanitation -----. Forthcoming. “Political Economy Assessment and Analysis of Public Investment and Hygiene.” New York: UNICEF. Management Draft Report.” Washington, DC: World Bank. -----. 2019. “Multiple Indicator Cluster Surveys 2019.” New York: UNICEF. Yoshida, N., and U. Deichmann. 2009. “Measurement of Accessibility and Its Applications.” Journal of Infrastructure Development 2009;1(1):1-16. doi:10.1177/097493060900100102 USAID (United States Agency for International Development). 2018. “Rapid Education and Risk Analysis. Cox’s Bazar.” Washington, DC: USAID. Zafar, Prepared Ali, Iuliia Mironova, and Liane Lohde. 2020. “Bangladesh: Country Private Sector Diagnostic Concept Note.” Washington, DC: World Bank. -----. 2019. “Comprehensive Private Sector Assessment.” Washington, DC: USAID. Woodruff, Christopher. 2018. “Addressing Constraints to Small and Growing Businesses.” Working paper. International Growth Center (IGC), London. https://www.theigc.org/ wp-content/uploads/2018/11/IGC_ANDE-review-paper_final-revised.pdf. World Bank. 2009. World Development Report 2009: Reshaping Economic Geography. Washington, DC: World Bank. -----. 2018a. “Project Appraisal Document on a Proposed Credit to the People’s Republic of Bangladesh for a Sustainable Coastal and Marine Fisheries.” Washington, DC: World Bank. -----. 2018b. “Project Appraisal Document on a Proposed Credit to the People’s Republic of Bangladesh for a Livestock and Dairy Development Project.” Washington, DC: World Bank. -----. 2019a. “Cox’s Bazar Panel Survey. Brief III.” Washington, DC: World Bank. -----. 2019b. “Cox’s Bazar Panel Survey. Brief IV.” Washington, DC: World Bank. -----. 2019c. “Infrastructure Connectivity.” Japan G20 Development Working Group 13. -----. 2020a. “Bangladesh Human Capital Index.” Washington, DC: World Bank. -----. 2020b. “Cox’s Bazar Local Government Capacity and Service Delivery Assessment. Options of Using LGIs for Service Delivery, Local Economic Development and Disaster Management.” Washington, DC: World Bank. -----. 2020c. “Cox’s Bazar Panel Survey. Brief V.” Washington, DC: World Bank. -----. 2020d. “‘World Development Indicators Online Database.’” Washington, DC: World Bank. Retrieved (https://data.worldbank.org/data-catalog/world-development-indicators.). 188 189 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC ANNEX 1: Additional tables and figures Table A1-1: Road speeds by type Road classification (OpenStreetMap) Mean speed (k/h) Trunk; primary 55 Secondary 40 Tertiary 30 Residential; living street 20 Road; service; unclassified 15 Ferry 15 Track 10 Pier 5 All link roads, e.g. primary_link -5 Source: Based on Bangladesh road classifications, adapted from Blankenspoor and Yoshida (2010). Table A1-2: Incidence of crime in the neighborhood, as reported by CBPS 2019 respondents Gender Based Stratum Harassment Physical Violence Violence High exposure 57% 58% 47% Low exposure 57% 48% 38% Rohingya camp 36% 43% 40% Source: CBPS-2019. 190 191 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Table A1-3: Exposure to trauma events among CBPS 2019 respondents Table A1-4: Trauma symptoms reported by CBPS 2019 respondents Experienced Witnessed Heard about it A little Quite a bit Extremely High Low High Low High Low High Low High Low High Low exposure exposure Camp exposure exposure Camp exposure exposure Camp exposure exposure Camp exposure exposure Camp exposure exposure Camp Recurrent thoughts or memories Being close to 45% 40% 37% 19% 24% 45% 5% 6% 12% 33% 28% 53% 16% 15% 19% 21% 29% 17% of the most death hurtful or ter- rifying events Serious injury 29% 36% 33% 37% 28% 44% 25% 22% 16% Recurrent 35% 37% 49% 7% 9% 15% 1% 1% 2% nightmares Feeling Unnatural death detached or of family or 25% 17% 32% 9% 13% 11% 7% 21% 11% 29% 27% 39% 5% 9% 11% 1% 2% 2% withdrawn friend from people Unable to feel 30% 37% 37% 7% 9% 11% 1% 5% 2% Torture 16% 15% 44% 16% 14% 25% 37% 27% 19% emotions Feeling irrita- ble or having Murder of family 48% 44% 46% 15% 19% 14% 2% 4% 2% 12% 12% 35% 3% 6% 9% 8% 18% 10% outbursts of or friend anger Not wanting Imprisonment 12% 15% 14% 40% 28% 48% 43% 38% 32% to interact with others 15% 20% 20% 3% 6% 5% 1% 1% 1% outside the household Forced separation from 11% 9% 25% 6% 8% 11% 15% 23% 15% Feeling as family members if you don’t 28% 25% 35% 12% 13% 26% 3% 6% 7% have a future Having diffi- Combat 11% 16% 44% 30% 25% 40% 47% 41% 14% culty dealing situation 26% 32% 41% 4% 7% 12% 1% 1% 3% with new situations Forced isolation Troubled 4% 5% 30% 11% 10% 24% 39% 35% 27% by physical 39% 40% 42% 18% 20% 21% 5% 7% 5% from others problem(s) Feeling Lost or unable to 4% 5% 10% 8% 5% 22% 66% 55% 52% 31% 32% 37% 7% 9% 10% 1% 2% 2% kidnapped make daily plans Murder of Feeling that stranger or 3% 5% 10% 12% 8% 41% 67% 53% 41% people do not strangers understand 35% 31% 36% 12% 12% 15% 1% 2% 3% what hap- pened to you Rape or sexual Feeling others 2% 2% 6% 4% 7% 29% 74% 58% 61% abuse are hostile to 28% 30% 26% 9% 11% 6% 2% 2% 1% you Source: CBPS-2019. 192 193 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 A little Quite a bit Extremely A1-2: District Figure A1-2. Figure District diversification diversification in Figure A1-3. District Figure A1-3: District diversification diversification in in coastal plains and northern coastal plains and northern hills hills in river and estuarine river and estuarine floodflood plains plains High Low High Low High Low exposure exposure Camp exposure exposure Camp exposure exposure Camp zone agroecological zone agroecological agroecological zone agroecological zone Feeling that 4% 3% 3% you have no 100% 31% 29% 34% 12% 11% 13% 2% 3% 2% 6% one to rely 10% 8% 10% 7% 9% 7% 15% 6% 17% upon 90% 6% 10% 6% 18% 12% 24% Feeling no 80% 23% 24% 6% 12% 5% 41% 41% 41% 14% 13% 14% 2% 2% 2% 9% 18% trust in others 14% 34% 10% 70% 6% Feeling pow- 2% 15% 60% erless to help 35% 27% 32% 18% 17% 25% 7% 6% 8% 30% 3% 28% others 50% 45% 84% Spending time 40% 2% 1% 77% 69% 72% thinking why 30% 62% 64% these events 38% 32% 38% 14% 19% 37% 4% 8% 8% 53% 53% happened to 20% 1% 36% 38% you 10% 21% Source: CBPS-2019. 0% n g r i i ur a ni ria li ur ar at za on ba ha ill Fe dp ip am ch ba Ba m ag ar ak Figure A1-1: Bangladesh agroecological zones hm an Cu ra an nd ng x's itt No ag Ch ks m Ch Ba Ra Figure A1-1: Bangladesh agroecological zones Co Kh La ah Br 1 Old Himalayan Piedmont Plain Rice Vegetables Fruits Other crops Tea, tobacco, betelnut & betel leaf 2 Active Tista Floodplain 3 Tista Meander Floodplain 4 Karatoya - Bangali Floodplain Source: WB staff elaboration, Agricultural yearbook 2017. 5 Lower Atrai Basin 0 50 100km 6 Lower Purnabhaba Floodplain 7 Active Brahmaputra - Jamuna Floodplain Young Brahmaputra & Jamuna Floodplain 8 Figure A1-4: Share of rural population by upazila, Cox’s Bazar 9 Old Brahmaputra Floodplain 10 Active Ganges Floodplain Figure A1-4. Share of rural population by upazila, Cox’s Bazar 11 High Ganges River Floodplain 12 Low Ganges River Floodplain 91.1 86.8 13 Ganges Tidal Floodplain 84.4 83.9 83.5 80.4 80.4 14 Gopalganj - Khulna Beel 15 Arial Beel 16 Middle Meghna River Floodplain 17 Lower Meghna River Floodplain 18 Young Meghna Estuarine Floodplain 52.7 19 Old Meghna Estuarine Floodplain 20 Eastern Surma - Kushiara Floodplain 21 Sylhet Basin 22 Northern & Eastern Piedmont Floodplain 23 Chittagong Coastal Plain 24 St. Martin’s Coral Island 25 Level Barind Tract 26 High Barind Tract 27 North-Eastern Barind Tract 28 Madhupur Tract 29 Northern & Eastern Hills 30 Akhaura Terrace ria ia li a u a r a f da a ku hi m bd kn kh a Uk Ra Sa Pe ak tu Te sh International Boundary District Boundary Wide River Ch r Ku za e ah Ba Lake Dhaka city Mangrove forest M x's Co Source: BBS (2018c). Source: WB staff elaboration, Agricultural yearbook 2017. 194 195 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Table A1-5: Firm density by upazila Table A1-6: Non-agricultural establishment size distribution, Bangladesh vs Chittagong vs Cox’s Bazar Population N firms (000) Firm density Industry Bangladesh Chittagong Cox’s Bazar Kutubdia 2,992 125 24 1 worker 277,182 31% 89,875 45% 13,329 71% 2 workers 52,481 6% 10,184 5% 1,318 7% Maheshkhali 7,817 319 24 3-4 workers 398,934 45% 77,112 38% 2,786 15% Pekua 4,499 170 26 5-9 workers 114,784 13% 18,491 9% 1,091 6% 10-35 workers 28,033 3% 2,545 1% 137 1% Ukhia 7,835 207 38 more than 35 20,135 2% 2,459 1% 89 0% 891,549 100% 200,666 100% 18,750 100% Cox’s Bazar 23,143 821 42 Services COX’S BAZAR SADAR 19,529 458 43 Bangladesh Chittagong Cox’s Bazar 1 worker 3,170,779 47% 391,684 35% 27,718 36% Chittagong division without Chittagong 2 workers 1,727,217 25% 317,835 28% 15,809 21% 947,079 20,817 45 Sadar 3-4 workers 1,299,405 19% 304,561 27% 24,301 32% Chittagong division 1,327,629 28,451 47 5-9 workers 499,248 7% 93,805 8% 7,917 10% 10-35 workers 97,374 1% 17,761 2% 1,040 1% Ramu 13,193 270 49 more than 35 8,543 0% 1,317 0% 79 0% 6,802,566 100% 1,126,963 100% 76,864 100% Bangladesh 7,694,115 144,114 53 Total Bangladesh Chittagong Cox’s Bazar Bangladesh without Dhaka district and 6,650,981 124,385 53 Chittagong Sadar 1 worker 3,447,961 45% 481,559 36% 41,047 42.9% 2 workers 1,779,698 23% 328,019 25% 17,127 17.9% Chakaria 25,420 473 54 3-4 workers 1,698,339 22% 381,673 29% 27,087 28.3% 5-9 workers 614,032 8% 112,296 8% 9,008 9.4% Teknaf 14,329 263 54 10-35 workers 125,407 2% 20,306 2% 1,177 1.2% Source: WB staff elaboration, Economic Census 2013 and Population Census 2011. more than 35 28,678 0% 3,776 0% 168 0.2% 7,694,115 1 1,327,629 1 95,614 1 Source: Staff calculations based on Economic Census 2013. 196 197 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Table A1-7: Size-wise distribution of firms, by sector - Bangladesh, Chittagong, Manufacture and Cox’s Bazar Extraction of textiles Other Accommodation Other N of firms of Salt and RMG Industry Trade Transport and food Education services Manufacture Chittagong Extraction of textiles Other Accommodation Other of Salt and RMG Industry Trade Transport and food Education services Total 1 481,559 1% 7% 11% 40% 19% 1% 0% 22% Cox’s bazar 2 328,019 0% 1% 2% 44% 1% 33% 1% 17% 1 84% 87% 35% 31% 91% 3% 9% 48% 43% 3-4 381,673 0% 3% 17% 52% 3% 5% 3% 16% 2 4% 10% 6% 17% 2% 73% 10% 16% 18% 5-9 112,296 0% 1% 15% 52% 2% 3% 13% 15% 3-4 7% 3% 41% 40% 6% 18% 25% 26% 28% 10 5-9 4% 1% 15% 12% 1% 4% 37% 8% 9% 24,082 0% 3% 17% 7% 1% 3% 35% 33% plus 10 Total 1,327,629 0% 4% 11% 45% 8% 10% 3% 19% 1% 0% 3% 0% 0% 2% 18% 3% 1% plus Bangladesh Chittagong 1 3,447,961 0% 3% 4% 45% 30% 1% 0% 16% 1 84% 69% 35% 32% 83% 4% 5% 42% 36% 2 4% 5% 5% 24% 4% 79% 6% 23% 25% 2 1,779,698 0% 1% 2% 48% 7% 22% 1% 19% 3-4 7% 22% 45% 34% 11% 14% 31% 25% 29% 3-4 1,698,339 0% 4% 19% 47% 6% 5% 2% 16% 5-9 4% 2% 12% 10% 2% 2% 36% 7% 8% 5-9 614,032 0% 5% 14% 48% 3% 2% 12% 16% 10 10 1% 2% 3% 0% 0% 0% 22% 3% 2% 154,085 0% 13% 18% 6% 1% 1% 31% 29% plus plus Bangladesh Total 7,694,115 3% 8% 0% 46% 17% 7% 2% 17% 1 83% 46% 24% 44% 80% 6% 7% 43% 45% Source: Staff calculations based on Economic Census 2013. 2 4% 6% 6% 24% 10% 76% 6% 26% 23% 3-4 7% 29% 52% 23% 9% 16% 22% 20% 22% Table A1-9: Upazila-wise distribution of firms by firm-size groups 5-9 4% 12% 14% 8% 1% 2% 39% 8% 8% 10 N firms 2% 8% 4% 0% 0% 0% 26% 3% 2% plus Cox’s Source: Staff calculations based on Economic Census 2013. Bazar Chakaria Sadar Kutubdia Maheshkhali Pekua Ramu Teknaf Ukhia Total 1 14427 6993 558 1839 682 7215 5813 3520 41047 Table A1-8: Sector-wise distribution of firms, by firm size - Bangladesh, 2 2698 4092 375 1983 1726 3712 1411 1130 17127 Chittagong, and Cox’s Bazar 3-4 6516 5204 1251 2677 1298 1078 6441 2622 27087 Manufacture 5-9 1606 2665 753 1213 740 1063 487 481 9008 Extraction of textiles Other Accommodation Other N of firms of Salt and RMG Industry Trade Transport and food Education services 10 Plus 173 575 55 105 53 125 177 82 1345 Cox’s Bazar Shares 1 41,047 10% 17% 5% 34% 15% 1% 1% 18% Cox’s Bazar 2 17,127 1% 5% 2% 45% 1% 31% 2% 14% Chakaria Sadar Kutubdia Maheshkhali Pekua Ramu Teknaf Ukhia Total 3-4 27,087 1% 1% 8% 67% 1% 5% 2% 15% 1 35% 17% 1% 4% 2% 18% 14% 9% 100% 2 16% 24% 2% 12% 10% 22% 8% 7% 100% 5-9 9,008 2% 0.5% 9% 61% 1% 3% 10% 13% 3-4 24% 19% 5% 10% 5% 4% 24% 10% 100% 10 1,345 4% 0.4% 12% 9% 1% 9% 34% 30% plus 5-9 18% 30% 8% 13% 8% 12% 5% 5% 100% Total 95,614 9% 6% 5% 47% 7% 8% 3% 16% 10 Plus 13% 43% 4% 8% 4% 9% 13% 6% 100% Source: WB staff elaboration, Economic Census 2013. 198 199 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Table A1-10: Number of firms by upazila, firm-size groups, and sector Manufacture N of N of Extraction of textiles Other Accommodation Other Manufacture workers firms of salt and RMG industry Trade Transport and food Education services N of N of Extraction of textiles Other Accommodation Other workers firms of salt and RMG industry Trade Transport and food Education services Pekua Chakaria 1 682 192 52 14 168 1 14 6 235 1 14427 1200 5746 577 3437 1782 39 67 1579 2 1726 10 450 84 622 6 383 4 167 2 2698 8 186 23 775 46 1327 24 309 3-4 1298 1 7 62 869 4 116 24 215 3-4 6516 25 156 645 4160 197 251 100 982 5-9 740 37 7 114 428 4 9 45 96 5-9 1606 17 21 124 1024 11 42 174 193 10 Plus 53 0 0 15 5 1 2 20 10 10 Plus 173 0 2 19 4 1 5 96 46 Total 4499 240 516 289 2092 16 524 99 723 Total 25420 1250 6111 1388 9400 2037 1664 461 3109 Ramu Cox’s Bazar Sadar 1 7215 52 214 675 2769 2073 26 26 1380 1 6993 1057 685 257 2462 289 116 59 2068 2 3712 1 66 98 2226 68 827 33 393 2 4092 37 32 52 2212 23 992 78 666 3-4 1078 2 6 316 503 13 31 78 129 3-4 5204 71 37 600 3138 68 437 93 760 5-9 1063 1 2 90 687 5 12 133 133 5-9 2665 6 8 252 1713 27 119 184 356 10 Plus 125 1 0 21 14 3 1 36 49 10 Plus 575 41 1 65 56 6 98 132 176 Total 13193 57 288 1200 6199 2162 897 306 2084 Total 19529 1212 763 1226 9581 413 1762 546 4026 Teknaf Kutubdia 1 5813 533 406 101 2870 672 11 32 1188 1 558 305 0 8 64 51 0 8 122 2 1411 14 70 9 502 7 385 7 417 2 375 1 0 0 17 0 343 0 14 3-4 6441 28 16 186 4977 71 280 125 758 3-4 1251 32 4 75 826 16 17 34 247 5-9 487 2 4 87 169 4 42 139 40 5-9 753 59 0 27 508 2 17 52 88 10 Plus 177 0 1 20 41 0 19 52 44 10 Plus 55 0 0 8 1 0 0 25 21 Total 14329 577 497 403 8559 754 737 355 2447 Total 2992 397 4 118 1416 69 377 119 492 Ukhia Maheshkhali 1 3520 0 1 150 1792 1198 39 29 311 1 1839 966 34 104 358 46 1 8 322 2 1130 0 0 8 567 4 233 110 208 2 1983 144 2 23 852 6 746 6 204 3-4 2622 0 6 187 1786 10 94 149 390 3-4 2677 192 4 128 1789 5 76 34 449 5-9 481 0 1 70 197 2 11 120 80 5-9 1213 101 0 61 804 6 20 70 151 10 Plus 82 0 0 6 1 0 2 40 33 10 Plus 105 16 2 8 2 0 0 51 26 Total 7835 0 8 421 4343 1214 379 448 1022 Total 7817 1419 42 324 3805 63 843 169 1152 Source: WB staff elaboration, Economic Census 2013. 200 201 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Table A1-11: Distribution of firms by upazila and by firm-size group within Manufacture each sector N of Extraction of textiles Other Accommodation Other workers of Salt and RMG industry Trade Transport and food Education services Manufacture Teknaf N of Extraction of textiles Other Accommodation Other workers of Salt and RMG industry Trade Transport and food Education services 1 5813 9% 7% 2% 49% 12% 0% 1% 20% Chakaria 2 1411 1% 5% 1% 36% 0% 27% 0% 30% 1 14427 8% 40% 4% 24% 12% 0% 0% 11% 3-4 6441 0% 0% 3% 77% 1% 4% 2% 12% 2 2698 0% 7% 1% 29% 2% 49% 1% 11% 5-9 487 0% 1% 18% 35% 1% 9% 29% 8% 3-4 6516 0% 2% 10% 64% 3% 4% 2% 15% 10 Plus 177 0% 1% 11% 23% 0% 11% 29% 25% 5-9 1606 1% 1% 8% 64% 1% 3% 11% 12% Ukhia 10 Plus 173 0% 1% 11% 2% 1% 3% 55% 27% 1 3520 0% 0% 4% 51% 34% 1% 1% 9% Cox’s Bazar Sadar 2 1130 0% 0% 1% 50% 0% 21% 10% 18% 1 6993 15% 10% 4% 35% 4% 2% 1% 30% 3-4 2622 0% 0% 7% 68% 0% 4% 6% 15% 2 4092 1% 1% 1% 54% 1% 24% 2% 16% 5-9 481 0% 0% 15% 41% 0% 2% 25% 17% 3-4 5204 1% 1% 12% 60% 1% 8% 2% 15% 10 Plus 82 0% 0% 7% 1% 0% 2% 49% 40% 5-9 2665 0% 0% 9% 64% 1% 4% 7% 13% Source: WB staff elaboration, Economic Census 2013. 10 Plus 575 7% 0% 11% 10% 1% 17% 23% 31% Kutubdia Table A1-12: Share of firm-size groups among total firms, by sector and upazila 1 558 55% 0% 1% 11% 9% 0% 1% 22% Manufacture 2 375 0% 0% 0% 5% 0% 91% 0% 4% N of All Extraction of textiles Other Accommodation Other 3-4 1251 3% 0% 6% 66% 1% 1% 3% 20% workers sectors of Salt and RMG industry Trade Transport and food Education services 5-9 753 8% 0% 4% 67% 0% 2% 7% 12% Chakaria 10 Plus 55 0% 0% 15% 2% 0% 0% 45% 38% 1 57% 96% 94% 42% 37% 87% 2% 15% 51% Maheshkhali 2 11% 1% 3% 2% 8% 2% 80% 5% 10% 1 1839 53% 2% 6% 19% 3% 0% 0% 18% 3-4 26% 2% 3% 46% 44% 10% 15% 22% 32% 2 1983 7% 0% 1% 43% 0% 38% 0% 10% 5-9 6% 1% 0% 9% 11% 1% 3% 38% 6% 3-4 2677 7% 0% 5% 67% 0% 3% 1% 17% 10 Plus 1% 0% 0% 1% 0% 0% 0% 21% 1% 5-9 1213 8% 0% 5% 66% 0% 2% 6% 12% Cox’s Bazar Sadar 10 Plus 105 15% 2% 8% 2% 0% 0% 49% 25% Pekua 1 36% 87% 90% 21% 26% 70% 7% 11% 51% 1 682 28% 8% 2% 25% 0% 2% 1% 34% 2 21% 3% 4% 4% 23% 6% 56% 14% 17% 2 1726 1% 26% 5% 36% 0% 22% 0% 10% 3-4 27% 6% 5% 49% 33% 16% 25% 17% 19% 3-4 1298 0% 1% 5% 67% 0% 9% 2% 17% 5-9 14% 0% 1% 21% 18% 7% 7% 34% 9% 5-9 740 5% 1% 15% 58% 1% 1% 6% 13% 10 Plus 3% 3% 0% 5% 1% 1% 6% 24% 4% 10 Plus 53 0% 0% 28% 9% 2% 4% 38% 19% Kutubdia Ramu 1 19% 77% 0% 7% 5% 74% 0% 7% 25% 1 7215 1% 3% 9% 38% 29% 0% 0% 19% 2 13% 0% 0% 0% 1% 0% 91% 0% 3% 2 3712 0% 2% 3% 60% 2% 22% 1% 11% 3-4 1078 0% 1% 29% 47% 1% 3% 7% 12% 3-4 42% 8% 100% 64% 58% 23% 5% 29% 50% 5-9 1063 0% 0% 8% 65% 0% 1% 13% 13% 5-9 25% 15% 0% 23% 36% 3% 5% 44% 18% 10 Plus 125 1% 0% 17% 11% 2% 1% 29% 39% 10 Plus 2% 0% 0% 7% 0% 0% 0% 21% 4% 202 203 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Manufacture Table A1-13: Breakdown of “Other industry” and” Other services” categories for N of All Extraction of textiles Other Accommodation Other non-micro enterprises (more than 10 employees), by upazila workers sectors of Salt and RMG industry Trade Transport and food Education services Maheshkhali Chakaria Cox’s Bazar Sadar Kutubdia Maheshkhali Pekua Ramu Teknaf Ukhia Manufacture of 1 24% 68% 81% 32% 9% 73% 0% 5% 28% 1% 1% 2% 1% 2% 4% food products 2 25% 10% 5% 7% 22% 10% 88% 4% 18% Manufacture of 1% tobacco products 3-4 34% 14% 10% 40% 47% 8% 9% 20% 39% Manufacture 5-9 16% 7% 0% 19% 21% 10% 2% 41% 13% of wood and products of wood 1% 2% 1% 1% and cork, except 10 Plus 1% 1% 5% 2% 0% 0% 0% 30% 2% furniture; Pekua Manufacture of paper and paper 0% 1 15% 80% 10% 5% 8% 6% 3% 6% 33% products Printing and 2 38% 4% 87% 29% 30% 37% 73% 4% 23% reproduction of 1% recorded media 3-4 29% 0% 1% 21% 42% 25% 22% 24% 30% Manufacture of 5-9 16% 15% 1% 39% 20% 25% 2% 45% 13% rubber and plas- 1% 1% tics products 10 Plus 1% 0% 0% 5% 0% 6% 0% 20% 1% Manufacture of other non-metallic 1% 1% 4% 2% 8% 9% 3% Ramu mineral products 1 55% 91% 74% 56% 45% 96% 3% 8% 66% Manufacture of fabricated metal 2 28% 2% 23% 8% 36% 3% 92% 11% 19% products, except 1% 1% 7% 2% 1% machinery and 3-4 8% 4% 2% 26% 8% 1% 3% 25% 6% equipment Manufacture of 5-9 8% 2% 1% 7% 11% 0% 1% 43% 6% 8% 1% 4% 4% 17% 6% 3% 4% furniture 10 Plus 1% 2% 0% 2% 0% 0% 0% 12% 2% Other 1% manufacturing Teknaf Electricity, gas, steam and air con- 4% 1 41% 92% 82% 25% 34% 89% 1% 9% 49% ditioning supply 2 10% 2% 14% 2% 6% 1% 52% 2% 17% Water collection, treatment and 0% 3-4 45% 5% 3% 46% 58% 9% 38% 35% 31% supply Construction of 5-9 3% 0% 1% 22% 2% 1% 6% 39% 2% 1% buildings 10 Plus 1% 0% 0% 5% 0% 0% 3% 15% 2% Publishing 0% activities Ukhia Motion picture, 1 45% . 13% 36% 41% 99% 10% 6% 30% video and tele- vision program 0% production, sound 2 14% . 0% 2% 13% 0% 61% 25% 20% recording 3-4 33% . 75% 44% 41% 1% 25% 33% 38% Programming and broadcasting 0% 5-10 6% . 13% 17% 5% 0% 3% 27% 8% activities 11 Plus 1% . 0% 1% 0% 0% 1% 9% 3% Computer programming, 0% consultancy and Source: WB staff elaboration, Economic Census 2013. related activities 204 205 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Chakaria Cox’s Bazar Sadar Kutubdia Maheshkhali Pekua Ramu Teknaf Ukhia Table A1-14: Constraints on access to education among persons who never attended school, bottom 40 and upper 60, by gender, high- and low-exposure Financial service activities, except areas 14% 13% 16% 9% 4% 10% 14% 16% insurance and pension funding High exposure Insurance, reinsur- Bottom 40 Upper 60 Total ance and pension funding, except 2% 2% 4% 1% 0% 4% 1% 1% Female Male Female Male Female Male compulsory social security No money/too expensive 46% 56% 40% 59% 42% 58% Real estate Family/social restrictions 25% 1% 38% 4% 32% 3% 1% activities No schools close to home 7% 5% 6% 7% 7% 6% Law and Accounting 0% Age (too old/too young) 6% 10% 4% 7% 5% 8% Activities Must work/family chores 5% 12% 4% 11% 4% 11% Other profes- Lack of food 3% 6% 1% 3% 2% 4% sional, scientific 1% Others 8% 9% 7% 9% 7% 9% and technical activities Low exposure Veterinary 0% Bottom 40 Upper 60 Total activities Rental and leasing Female Male Female Male Female Male 0% 1% 4% 1% 1% 1% activities No money/too expensive 50% 65% 37% 55% 44% 60% Employment Family/social restrictions 27% 5% 38% 4% 33% 4% activities Age (too old/too young) 8% 9% 7% 8% 8% 8% Travel agency, tour operator, No schools close to home 6% 4% 5% 3% 6% 4% reservation ser- 0% vice and related Must work/family chores 3% 10% 4% 18% 3% 13% activities No need/no interest to study 2% 3% 2% 7% 2% 5% Office admin- Others 4% 5% 6% 6% 5% 5% istrative, office support and other 0% Source: WB staff elaboration, CBPS 2019. business support activities Public administra- Constraints on access to education among persons Table A1-15: tion and defense; compulsory social 6% 8% 13% 10% 9% 14% 7% 18% who dropped out of school: bottom 40 and upper 60, by age security group and gender, high-exposure areas Human health 2% 4% 5% 3% 2% 5% 1% 2% activities Dropped-out High exposure Residential care 6 to 18 years 0% activities Bottom 40 Upper 60 Total Sports activities and amusement Female Male Female Male Female Male 1% and recreation No money/too expensive 63% 57% 37% 46% 47% 49% activities Family/social restrictions 21% 3% 27% 1% 24% 2% Activities of membership 2% 1% 1% 3% Do not want to study more/com- 7% 24% 9% 34% 8% 31% organizations pleted studies Other personal Safety concerns 3% 0% 2% 1% 2% 1% 1% 1% 1% 2% 1% service activities For marriage 3% 0% 15% 0% 10% 0% Source: WB staff elaboration, Economic Census 2013. Must work/family chores 1% 8% 5% 14% 3% 12% Lack of food 1% 0% 1% 1% 1% 1% Others 1% 9% 5% 3% 3% 5% 206 207 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Dropped-out High exposure Table A1-17: Share of individuals who dropped out of school, by type of school, Older than 18 quintile, and gender, high- and low-exposure areas Bottom 40 Upper 60 Total High exposure Female Male Female Male Female Male Quintile 1 2 3 4 5 Total No money/too expensive 34% 56% 23% 43% 27% 46% Female Male Female Male Female Male Female Male Female Male Female Male Family/social restrictions 27% 1% 29% 4% 28% 3% Government 65% 70% 60% 73% 63% 64% 50% 65% 58% 60% 59% 65% For marriage 15% 0% 27% 1% 23% 1% Private total 17% 14% 23% 17% 27% 21% 31% 26% 28% 28% 26% 23% Do not want to study more/com- 11% 25% 11% 30% 11% 29% Private pleted studies 13% 11% 16% 14% 18% 14% 23% 16% 22% 21% 19% 16% (govt. grant) Must work/family chores 6% 10% 6% 15% 6% 14% Private (non- 4% 3% 6% 4% 8% 8% 9% 10% 6% 7% 7% 7% Lack of food 2% 4% 0% 2% 1% 3% govt. grant) Others 4% 4% 4% 5% 4% 4% NGO 2% 2% 3% 2% 1% 0% 2% 1% 2% 0% 2% 1% Source: WB staff elaboration, CBPS 2019. Madrasa 16% 14% 15% 8% 9% 15% 17% 8% 13% 12% 14% 11% Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Table A1-16: Constraints on access to education among persons who dropped Low exposure out of school: bottom 40 and upper 60, by age group and gender, low-exposure Quintile 1 2 3 4 5 Total Female Male Female Male Female Male Female Male Female Male Female Male areas Government 57% 65% 57% 61% 52% 59% 50% 53% 51% 60% 54% 59% Low exposure Private total 27% 25% 28% 27% 34% 26% 33% 32% 39% 31% 32% 29% 6 to 18 years Private 17% 17% 14% 14% 25% 18% 25% 25% 26% 16% 21% 18% (govt. grant) Bottom 40 Upper 60 Total Private (non- Female Male Female Male Female Male 10% 8% 14% 13% 9% 8% 8% 7% 13% 15% 11% 11% govt. grant) No money/too expensive 51% 57% 37% 36% 43% 45% NGO 3% 2% 3% 4% 0% 2% 1% 1% 0% 1% 1% 2% For marriage 18% 0% 29% 0% 24% 0% Madrasa 13% 8% 12% 8% 14% 13% 16% 14% 10% 8% 13% 10% Family/social restrictions 13% 0% 16% 4% 15% 2% Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Do not want to study more/ 9% 20% 9% 32% 9% 27% Source: WB staff elaboration, CBPS 2019. completed stu Must work/family chores 6% 15% 6% 20% 6% 18% Disability/illness 2% 2% 0% 3% 1% 3% Table A1-18: Distribution of firms by sector and upazila, Cox’s Bazar Lack of food 1% 6% 0% 2% 1% 4% Upazila’s share of total firms, by sector Others 0% 0% 3% 3% 2% 2% Manufacture Older than 18 N of Extraction of textile Other Accommodation Other firms of Salt and RMG industry Trade Transport and food Education services Total Bottom 40 Upper 60 Total Chakaria 27% 24% 74% 26% 21% 30% 23% 18% 21% 27% Female Male Female Male Female Male Cox’s Bazar No money/too expensive 35% 50% 18% 34% 24% 40% 20% 24% 9% 23% 21% 6% 25% 22% 27% 20% Sadar For marriage 30% 2% 41% 2% 37% 2% Kutubdia 3% 8% 0% 2% 3% 1% 5% 5% 3% 3% Family/social restrictions 19% 3% 23% 5% 22% 4% Maheshkhali 8% 28% 1% 6% 8% 1% 12% 7% 8% 8% Do not want to study more/ 8% 19% 13% 33% 11% 28% Pekua 5% 5% 6% 5% 5% 0% 7% 4% 5% 5% completed stu Must work/family chores 4% 20% 3% 22% 3% 21% Ramu 14% 1% 3% 22% 14% 32% 12% 12% 14% 14% Lack of food 1% 3% 1% 1% 1% 2% Teknaf 15% 11% 6% 8% 19% 11% 10% 14% 16% 15% Others 2% 3% 2% 2% 2% 2% Ukhia 8% 0% 0% 8% 10% 18% 5% 18% 7% 8% Source: WB staff elaboration, CBPS 2019. Total 95614 5152 8229 5369 45395 6728 7183 2503 15055 95,614.00 208 209 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 1 Sector’s share of total firms, by upazila Table A1-21: Share of vulnerable and secure jobs among all workers in each geographic unit Manufacture N of Extraction of textile Other Accommodation Other firms of Salt and RMG industry Trade Transport and food Education services Total Vulnerable jobs Secure jobs Chakaria 5% 24% 5% 37% 8% 7% 2% 12% 100% Chakaria 98% 2% 25,420 Cox’s bazar Cox’s Bazar Sadar 92% 8% 6% 4% 6% 49% 2% 9% 3% 21% 100% sadar 19,529 Kutubdia 98% 2% Kutubdia 13% 0% 4% 47% 2% 13% 4% 16% 100% 2,992 Maheshkhali 97% 3% Maheshkhali 18% 1% 4% 49% 1% 11% 2% 15% 100% 7,817 Pekua 89% 11% Pekua 5% 11% 6% 46% 0% 12% 2% 16% 100% 4,499 Ramu 93% 7% Ramu 13,193 0% 2% 9% 47% 16% 7% 2% 16% 100% Teknaf 96% 4% Teknaf 4% 3% 3% 60% 5% 5% 2% 17% 100% Ukhia 97% 3% 14,329 Cox’s Bazar 95% 5% Ukhia 0% 0% 5% 55% 15% 5% 6% 13% 100% 7,835 Chittagong 94% 6% Source: WB staff elaboration, Economic Census 2013. Bangladesh 92% 8% Source: WB staff elaboration, Economic Census 2013. Table A1-19: Share of firm by market for goods Totally Local Totally Export Local and Export Not Applicable Total Cox’s Bazar 82% 2.57% 0.10% 16% 13,441 Table A1-22: Distribution of total vulnerable and secure jobs in Cox’s Bazar, across upazilas Chittagong Division 90% 0.76% 0.41% 9% 192,299 Bangladesh 88% 1.15% 0.83% 10% 857,572 Vulnerable jobs Secure jobs Bangladesh (Not incl. Chakaria 22% 7% 88% 0.67% 0.66% 11% 736,270 Chittagong and Dhaka) Cox’s Bazar Sadar 25% 43% Chittagong Division (Not 90% 0.50% 0.17% 9% 133,679 incl. Chittagong zila) Kutubdia 5% 2% Source: WB staff elaboration, Economic Census 2013. Maheshkhali 10% 5% Pekua 6% 13% Table A1-20: Distribution of exporting firms by size, Cox’s Bazar, Chittagong, Ramu 11% 15% and Bangladesh Teknaf 14% 10% Cox’s Bazar Chittagong Bangladesh Ukhia 7% 4% Less than 10 workers 97% 50% 31% Source: WB staff elaboration, Economic Census 2013. Between 10 and 24 workers 1% 6% 8% Between 25 and 99 workers 2% 18% 37% Between 100 and 250 workers 0% 6% 8% More than 250 workers 0% 20% 15% Total N of export firms 358 2,245 16,988 Source: WB staff elaboration, Economic Census 2013. 210 211 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC ANNEX 2 Methodology note - Cox’s Bazar accessibility analysis July 2020 Figure A2-1: High-quality jobs accessibility indices (markets The Cox’s Bazar Growth Diagnostic makes Figure A2-1: A gravity model of union- weighted by high-quality job heavy use of accessibility maps and charts level accessibility to high quality jobs numbers), pre-transport investment to explain the relationship between access in Cox’s Bazar to services and employment and growth 0 20 40km potential in Cox’s Bazar (as in Figure A2-1, Dhaka C H I T TAG O N G Figure A2-7, and Figure A2-9). This note Chittagong Pekua explains how we prepared the accessibility Kutubdia analysis figures underpinning these visuals Chakaria and how the same process might be applied Martarbari BANDARBAN (approximate) elsewhere, especially in Bangladesh. The note pays special attention to any compli- Maheshkhali cations that might result from a change in Cox’s Bazar the scale of the analysis. Cox’s Bazar Sadar Ramu COX’S This note is meant as a resource for policy BAZAR makers and technical experts to under- Ukhia stand and apply the analysis outputs – not  as an exhaustive account of how to Teknaf replicate the analysis. Those interested in more technical detail should look at Economic gravity GOST’s tutorials at the GOSTNets Github Growth centers St. Martin Repository, this project’s code notebooks Dwip on GOST’s code repository, or the authors Unions cited herein.  Low High 212 213 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 2 Accessibility modeling process  The tradeoff is that holding all these nodes and edges in memory and calculating routes over them is significantly more taxing for computers when working with large networks. Calculating travel times using a network analysis Even with significant simplification, regional, national, or international country-scale analyses may have to be conducted on dedicated servers which can handle the associated All accessibility statistics were computed using origin-destination (OD) matrices generated in load. Exact calculation times depend on the complexity of the network vs. the power of a network analysis. Given a set of origins and destinations, OD matrices show the travel time the machine employed and are thus difficult to predict. Running the full multi-scenario from every origin to every destination over a transport network, using average travel times analysis routine in Cox’s Bazaar using a dedicated server with 64 GB RAM and 16 processors took 3-4A2-03 Figure Simplifying hours when a junction fully optimized. across different classes of transport links (main roads, small roads, unpaved tracks, ferries, etc.). The minimum time thus calculated represents the quickest possible access time to that type of destination and in the process reveals the nearest destination. In this case, origins We usually simplify the network by removing nodes and straightening edges below a set were populated places and destinations were cities, services, or places of employment.  density threshold. The exact threshold selected depends on the scale of the analysis area and the importance of accuracy in time readings. For Cox’s Bazar, we simplified segments The network in a network analysis is repre- under 50m, whereas for a national analysis we might simplify under 1km. Figure A2-3 and Figure A2-2: Nodes and edges in a Figure A2-03 Simplifying a junction Figure A2-2. Nodes and edges in a sented as a collection of nodes and edges, Figure A2-4 below show examples of simplifying an individual junction and a larger net- network network as in Figure A2-2, where u and v are distinct work, respectively. nodes and (u,v) is an edge. Both nodes and Figure A2-3: Simplifying a junction V edges can have properties representing their type, size, importance, length, associated speeds, ID number, and/or any other useful characteristic. Edge lengths are calculated in (u, v, data) meters and multiplied by the average meter U data = {ID: ...,length: ..., type: / hour speed associated with that type of ..., speed: ...} transport link to yield an average speed per (u, data) edge (see Table A2-1 for speed details). If data = {junction_type:..., desired, some nodes can be assigned “wait Figure A2-04 Simplifying a small network prperty_2: ...} times” to represent traffic signals, expected congestion, border crossing delays, etc. Figure A2-4: Simplifying a small network We prepare these networks from existing geospatial data for roads, ferries, paths, or other transportation services. Such geospatial data can come from many sources: official govern- ment sources, privately held datasets, open databases like OpenStreetMap, or even GPS Figure A2-04 Simplifying a small network traces from field workers. Network analyses always face a tradeoff between accuracy and performance. A more detailed network will represent edge lengths, network shapes, and the network’s “topol- ogy” – the connections between segments (edges) – more accurately. The contrasting complex and simple representations of a traffic roundabout in Figure A2-3 exemplify this. Additionally, because origins and destinations in the analysis are “snapped” to the near- est node, usually by calculating the straight-line walking distance to this node, a greater density of nodes provides greater spatial accuracy in terms of start/end destinations and greater walking time accuracy – which can be substantial if either is far from a node. Fewer nodes thus lead to longer snap distances and greater average walking times. 214 215 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 2 Representing origins and destinations We analyzed the accessibility of each origin to a variety of destinations: health centers, dif- ferent types of schools, different types of markets, and key places of employment like the Martarbari port. Each destination was represented as a point, as in Figure A2 7, where a red Origins and destinations can take many forms in a network analysis. Both are represented point represents each health center. as points that are “snapped” to the nearest node on the network. The length of the snap- ping distance is used to calculate a to-node walking time specific to each origin or destina- In a few cases, it was faster to walk directly to a destination from an origin than to travel tion. This walking time is added to the calculated network travel time from origin node to over the network – for instance, students walking across a field to a nearby school. Our destination node. analysis considers this possibility and assigns direct walking speeds to origins where direct walking times are lower than walk-to-network + on-network times, as with O3 - D1 and O5 - In a larger-scale analysis (e.g., nationally), origins might be villages, towns, cities, or even D2 in Figure A2-5. the geographic center of administrative units. Given the small spatial scale of the Cox’s Bazar analysis we used a more detailed set of origins drawn from the High Resolution Speeds Satellite Layer (HRSL), a gridded population model released by Facebook and CIESIN, which is described in greater detail under the Data Quality section. We achieved a high Road travel times were adapted from a sim- Figure A2-7: Figure A2-7. Health Healthcenter center degree of spatial precision in origin locations by representing each cell in HRSL’s 30m x 30m ilar analysis by Blankespoor and Yoshida destinations destinations inin Cox’s Cox’s Bazar Bazar grid as an origin – over 100,000 origins in total for the whole district – as shown in Figure (2010), with a 10 km/h downward adjust- A2-6. This high precision allowed us to aggregate accessibility information at almost any ment to account for traffic density and poor 0 20 40km level of detail with high confidence in the results – mouza level aggregation would have road conditions in Cox’s Bazar. Ferry times been possible, if spatial boundaries were available. were estimated conservatively to account C H I T TAG O N G Dhaka for probable delays. All speeds are summa- Chittagong Pekua Note that processing this number of origins was computationally taxing, and it would not be rized in Table A2-1 below.  Kutubdia feasible to employ the same HRSL-derived origins for a regional- or national-scale analysis. Chakaria Table A2-1: Network segment speeds Martarbari BANDARBAN (approximate) Figure Road class  Speed Figure A2-05: Nodes edges, A2-5: Nodes, and edges in a origins, A2-06: Figure A2-6: Figure Populated Populated places // places (from OpenStreetMap)  (km/h)  Maheshkhali network and destinations in a network origins (in blue) around Cox’s Bazar origins (in blue) around Cox’s Bazar town (HRSL 2018) (HRSL 2018) Trunk 55 town Cox’s Bazar Primary 55 Cox’s Bazar Ramu V Sadar O3 Secondary 40 O1 COX’S D1 V BAZAR Tertiary 30 O2 Ukhia Residential / Unclassified 20 (small paved roads) (D2, data) D2 Tracks (unpaved), Service 15 Teknaf O4 U roads data = {type: ..., size: ..., name: Ferry 15 O5 (O5, data) ..., walk_time: ...} data = {population: ..., Pier 4.5 St. Martin walk_time: ...} Walking 4.5 Health facility (any type) Dwip Walking Distance Link roads Minutes travel to nearest health facility Note: not the full extent of the analysis. With upgrades to key roads (e.g. primary_link, second- -5 ary_link) 0 15 30 45 60 84   216 217 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 2 The 4.5 km/h walking speed is a slight reduction from Tobler’s (1993) recommended 5.06 The scenarios and their corresponding speed upgrades are described north-to-south in km/h speed  over flat terrain; this accounts for non-linearity in routes to roads (around Table 2 and visualized (in reference to accessibility to Martarbari) in Figure A2-9.  paddy fields, etc.) and is consistent with the literature on walking speeds used by transport geographers (Munoz-Raskin 2010, Mathon et al. 2018, Delmelle and Casas 2012). Table A2-2: Investment scenarios We routed populated places within Cox’s Bazar to seven (types of) destinations:  Scenario  Upgrade  • Downtown Cox’s Bazar   Major upgrades to the main roads ser- Upgrade port roads to pri- • Downtown Chittagong  vicing the Martarbari port and the mary, Maheshkhali road to secondary adjoining Maheshkhali upazila • The proposed deep sea port in Matarbari  • The nearest health facility (of any type)  The above roads are upgraded, and a dedicated ferry line Is set up connecting Maheshkhali and Upgrade ferry to tertiary speed • The nearest educational facility (primary, secondary, and tertiary separately)  Cox’s Bazar city across the boy • Growth centers  The above investments are made, and upgrades • Markets of all sizes  are made to the principle southern highway Upgrade AH41 to secondary speed (the AH41 N1) connecting Ukhia and Teknaf to Cox’s Bazar Sadar. Facilities in neighboring Bandarban and Chittagong districts were included in the analy- sis  to ensure accuracy in border areas.  The final products were packaged into maps, as shown in Figure A2-1, Figure A2-7, and Figure A2-9, and charts, as in Figure A2- 8. Figure A2-9: Figure A2-9: Proposed Proposed transport transport Potential accessibility and gravity models investments investments inin Cox's Cox’s Bazar Bazar A2-08. Figure A2-8: Accessibility Accessibility statistics statistics The resulting access statistics were calcu- 0 20 40km We prepared a further analysis of potential aggregated at were aggregated variouslevels, at various levels,and lated per populated place (30 m2) and then accessibility  to growth centers using these and occasionally occasionally further further subdivided subdivided by aggregated up to the union, upazila, and OD matrices, in recognition of growth cen- C H I T TAG O N G by demographic demographic indicators indicators within within them district (zila) levels for analysis and visualiza- Dhaka Chittagong Pekua ters’ central role in propelling economic them tion. All aggregated results were population growth. Potential accessibility considers Kutubdia weighted, e.g., if half the HRSL origin points each origin’s accessibility to all destina- Ukhia Chakaria tions instead of just the nearest one. This in a union contained 2x the population, their average travel times would be weighted Martarbari (approximate) BANDARBAN is useful where access is cumulative and Teknaf double in the mean union / upazila value. multiple destinations are important, as Ramu Maheshkhali with markets, as compared to single-use Investment scenarios  destinations like schools. Transportation Pekua Cox’s Bazar Cox’s Bazar geographers have developed several math- Sadar Ramu Maheshkhali ematical models of potential accessibility, All of this analysis was first prepared for COX’S the literature and theoretical basis for the current transportation network setup, BAZAR Kutubdia which are well reviewed by Geurs and van then replicated for three additional trans- Ukhia Wee (2004) and Paez et al. (2012).  Cox’s Bazar Sadar portation investment scenarios, taking into account improved speeds over the Chakaria Gravity models are a variant of potential key roads being upgraded. This permitted Proposed road segments for upgrading Teknaf comparisons between different investment accessibility models that increase the 0 10 20 Proposed ferry options and opens the door for future anal- across CXB bay attractiveness of a destination according Minutes travel to growth centers Proposed AH41 yses comparing the cost-effectiveness of road for upgrading to a given attribute (population size, these investments in accessibility terms, Minutes travel to Martarbari St. Martin market importance, etc.) and decrease No education Secondary deep sea port Dwip Primary Higher secondary or the incorporation of travel-time savings With upgrades to port roads, CXB bay ferry and AH41. it inverse to the distance (measured in Lower secondary University into formal cost-benefit analysis. meters, travel time, etc.). There are several 0 60 120 150 180 214 mathematical variants of gravity models: 218 219 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 2 we followed the standard practice in economic development literature of employing the We employed growth centers as proxies for major economic centers. Growth centers are negative exponential model,1 which balances well the decay in attraction over a distance. markets designated by Bangladesh’s government for investment based on their high Major publications describing this approach in the economic development context come importance for growth potential. In some cases, this proxy relationship may be imperfect, from Deichmann (1997) and Deichmann and Yoshida (2009), with relevant recent applica- as when jobs are in fact based far from a growth center; locations of major employers tions by Blankespoor and Yoshida (2010) and Blankespoor et al. (2018). would improve the analysis in such cases. The negative exponential model is summarized in Equation A2-1.  An important note is that the coarser spatial resolution of unions presents unavoidable error when, for instance, a growth center lies at the border of two unions but is only Equation A2-1: The negative exponential model assigned firms from one.  Data sources  All geospatial population data was sourced from the High Resolution Satellite Layer (HRSL) released by Facebook and CIESIN for Bangladesh in 2018. The HRSL uses deep learning to Where  categorize populated vs. unpopulated places (principally by recognizing building roofs) in a Iine The negative exponential accessibility index for origin i  30 m2 grid for each country. Populations for the most detailed available administrative level from the latest census are then distributed to these populated grid cells and adjusted to the S j  Destination j  year in question (2018) using a country-specific scalar from the UN Population Division. dij  Travel time from origin i to destination j  The transport network data is all sourced from OpenStreetMap, a “Wikipedia of maps” plat- b  Distance decay function b  form commonly used by geographers. OSM’s quality in Cox’s Bazar is overall very high, as a  Relative attractiveness of the destination (e.g., number of high-quality jobs)  volunteers have created lots of data to assist with the humanitarian response. All health-fa- cilities data comes from OpenStreetMap and Bangladesh’s Local Government Engineering Department (LGED). All educational facilities and markets come from the LGED. Downtown We employ gravity models to prepare four different indices of accessibility to growth centers:   Chittagong, Cox’s Bazar, and the approximate port location were manually located by the analyst using OpenStreetMap as a reference layer.  1. Market accessibility index  2. Firms (all sizes) accessibility index  A full list of data layers is provided in Table A2-3. 3. Large firms accessibility index   4. High-quality jobs accessibility index.   Table A2-3: Data sources The former is an unweighted negative exponential model. The latter three weight the Data type Source Aggregation Level Notes attractiveness of the market by relevant economic census data for the union containing the Roads OpenStreetMap Lines growth center. We employed a distance decay which halved the attractiveness of a growth center over 60 minutes travel. Thus, a growth center with 5,000 high quality jobs 0 minutes LGED / OpenStreetMap away would be equally as attractive as a growth center with 10,000 such jobs 60 minutes Observed / UN Inter-Sector Health facilities Points gaps and away. Other distance decays were considered for different variables that ultimately were Coordination Group inconsistencies (merge) not included in the growth diagnostic – for instance, “bad-quality jobs” were given a dis- tance decay of only 30 minutes to reflect their lower attractiveness. Educational facilities, Growth Centers, LGED Points Markets 1 There are many other models for calculating potential accessibility, which are beyond the scope of this paper to review. The negative exponential model is often preferred because it degrades attractive- ness proportionally to travel times. 220 221 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 2 Table A2-4: Categorizing service vs. industrial workers Data type Source Aggregation Level Notes Mining and Quarrying Population (2018) High Resolution Satellite 30 m2 grid Derives population (origins and totals) Layer (CISEIN and figures from 2011 Facebook) census, adjusted Manufacturing just Garment via a scalar to 2018 Industry Manufacturing excluding Garment Employment category 2013 Economic Census Union (admin 4) Does not include breakdowns (1) (BBS) agricultural Electricity, Gas and Water Supply workers, has lower aggregate figures than census Construction Wholesale and Retail Trade, Repair of Motor Vehicles and Motor Cycles Employment category 2011 Census Upazila (admin 3) Has higher breakdowns (2) (BBS) aggregate figures than census Accommodation and Food Service Activities (Hotel and Restaurants) Transport, Storage and Communications Educational 2011 Census Upazila (admin 3) attainment (BBS) breakdowns Financial and Insurance Activities Services Real Estate Activities Public Administration and Defense, Compulsory Social Security Data quality Education All transport network segments are roads aside from three critical ferry routes linking Kutubdia, southern Maheshkhali, and St. Martin Dwip to the main network. Based on our Human Health and Social Work Activities review, in the vast majority of the study area the roads data seemed adequately dense and correctly classified. In remote rural areas, the network was at times underdeveloped or Community, Social and Personal Services classified incorrectly (e.g., as a rough track instead of a paved road). We manually cleaned   up roads in important areas, but a full cleaning was beyond the scope of this assignment; therefore, the network may give erroneous readings in some remote areas. These errors Caveats and limitations  seem to occur most often in unpopulated areas far away from the Rohingya camps area so reflect least on the host communities most affected by the Rohingya influx. The model outputs should be seen as travel under reasonably optimum conditions, i.e., nor- mal amounts of traffic. However, roads conditions in the study area are occasionally far from We believe the margin of error within the analysis is acceptable. Based on personal experi- optimal, with many delays, heavy traffic, and detours. We attempted to gather traffic data ence in the area, the times returned approximate realistic optimal travel conditions (e.g., 45 from Mapbox to calibrate travel speeds, but it was not suitably comprehensive in this area.  minutes from Ukhia to downtown Cox’s Bazaar). Official LGED roads data was actually less complete than OSM data and thus wasn’t employed. A major limitation of this analysis is that it provides no insights into qualitative access to health and education because of incomplete attributes in the source data. In reality, dif- The health and educational datasets are from shapefiles provided by the LGED office. ferent facilities provide different services to different populations, with different levels of quality – think maternal health facilities vs. hospitals, or well-resourced schools vs. poorly A breakdown of the definitions used within the Economic Census for service and industrial resourced schools. With better source data, a future analysis could assess these qualitative work is provided in Table A2-4 below. aspects of access. 222 223 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Tools employed ANNEX 3 All of the analysis was performed in Python, specifically leveraging the GOSTNets package developed by the World Bank’s Geospatial Operations Support Team (GOST). Additional Nighttime light data roads data was created in OpenStreetMap using its standard Java OpenStreetMap Editor (JOSM) tool. Data quality spot checks were performed in QGIS. and economic activity All maps were created using a combination of QGIS and Adobe Illustrator: a rough cut of each map was prepared in QGIS and then exported as a PDF for refinement in Adobe Illustrator. Chart visualization exclusively took place in Python, to allow bulk production of charts. Using nighttime lights (NTL) data, we provide evidence of higher economic activity in mar- kets run by the host population near the Rohingya camp, suggesting that the host com- munities near camps have seen more economic activity after the influx of the Rohingya. The hypothesis explored is that the sudden influx of people brings new resources to the region, especially by the expansion in humanitarian aid. The aid provides means to increase the demand and activity in local markets. In Cox’s Bazar, the international assis- tance that the Rohingya received includes in-kind and e-voucher transfers, plus different cash-for-work programs. The effects of refugee influxes on the welfare of host populations are complex. Alix-Garcia et al. (2017) suggest a framework separating effects via market mechanisms and outside market mechanisms. Within the market, there are effects via demand and supply in goods, services, and labor. The changes in demand and supply impact the prices and incomes of host populations, thus affecting their welfare. In this case, we focus on effects via a market mechanism. Principally, we argue that more economic activity after the influx might have increased the income of hosts. To estimate the relationship between Rohingya camp populations and local economic activity, we follow Alix-Garcia et al. (2018). Similar regressions were run to test the relation- ship between refugee populations and economic growth in markets near Rohingya camps. 224 225 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 3 Lightsit=α+βRohingyat*fd (km from campi) Table A3-1: Nightlight intensity regression results +γConectivityiy+ϕHost poputioniy +θi+μt+ϵit   (1) (2) (3) (4) (5) Variables IHS(Lights) IHS(Lights) IHS(Lights) IHS(Lights) IHS(Lights) where Lightsit is the inverse hyperbolic sine (IHS) of the sum of the VIIRS luminosity within 0 to 5 km * Dummy 0.255*** a 500 m buffer around growth center i in month t, Rohingyat is the monthly Rohingya pop- August 2017 (0.0409) ulation reconstructed using UNHCR and IOM publicly available data,2 Conectivity is the travel time from growth center to Cox’s Bazar Sadar (the district capital) interacted with a IHS of distance to -0.0226*** year γ fixed effect, Host popution is the IHS of the host population living in a 5 km radius Camp * Rohingya population (0.00475) of growth center i interacted with a year fixed effect γ, θi is a growth center fix effect, μt is a time fixed effect, and ϵit is an error term. Inverse of distance 0.205*** to Camp * Rohingya population (0.0256) As in Alix-Garcia et al. (2018), both parametric and semi-parametric measures of distance are used to identify the effect of displaced population on economic activity. The monthly 0.120*** 0 to 5 km * Rohingya Rohingya population is interacted with measures of distance of growth centers from population (0.0300) camps, denoted as fd(km from campi). The latter takes three different forms, two of which are parametric: first, the IHS of the distance from growth center i to the closest -0.0545 5 to 10 km * Rohingya camp; and second, the inverse of distance to the closest camp. The semi-parametric population specification is a series of dummy variables representing a distance range. In this case, (0.0440) the following ranges are used: 0 to 5 km, 5 to 10 km, 10 to 15 km, 15 to 25 km, and 25 -0.0985*** 15 to 25 km * to 80 km. Two different specifications of Rohingyat*fd (km from campi) are used for Rohingya population robustness. The process first follows a basic difference-in-difference approach using the (0.0329) major 2017 influx of Rohingya as an identification strategy, using August 2017 dummy 0.0276 0 to 5km dummyi. 0 to 5km dummyi is a dummy variable that takes value 1 if the 25 to 80 km * Rohingya population growth center is within 5 km of a camp, and August 2017 dummy is a dummy that (0.0263) takes value 1 after the major Rohingya influx in August 2017. Identification follows, as the IHS Total population 0.0889*** influx is uncorrelated with the location of growth centers. At a second stage, the process Rohingya in a buffer uses Rohingya 5kmit which represents the IHS of the Rohingya population within a 5km of 5000 m around GC (0.0130) radius of growth center i in month t. Observations 2,448 2,448 2,448 2,448 2,448 Table A3-1 in the annex shows the results from estimation of the different specifications. R-squared 0.884 0.883 0.885 0.886 0.885 Column 1 shows how economic activity at growth centers close to Rohingya camps grows Time trend YES YES YES YES YES after the influx. The coefficient of interest has a positive sign and is significant. Column 2 shows that, as distance from camps increases, economic activity decreases, and column 3 Accessibility control YES YES YES YES YES shows the same result using the inverse distance measure. More work is needed to explore Host size population the mechanisms operating. This analysis presents results suggestive of positive impacts of YES YES YES YES YES control the Rohingya population on hosts, to the extent that nighttime lights provide an adequate proxy for economic activity. Market FE YES YES YES YES YES Year Month FE YES YES YES YES YES Robust standard errors in parentheses 2 Where data gaps were identified in the monthly refugee population count at camp level, a linear *** p<0.01, ** p<0.05, * p<0.1 interpolation between months was used to complete the series. 226 227 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 ANNEX 4. Within Cox’s Bazar, upazilas showed wide variation in their main sectors of employment. While in Pekua and Maheshkhali, 7 out of 10 individuals relied on agriculture in 2011, in Pattern of employment Cox’s Bazar Sadar, the service sector was dominant (Figure A4-1). Among the southern upazilas, Teknaf was relatively more diversified: two-thirds of households in Ukhia are in Cox ’s Bazar: Economic classified as farm households, as opposed to a third in Teknaf. Teknaf was also more indus- trialized, with industry employing more than 15 percent of workers, as opposed to only 4 Census 2013 and Population percent in Ukhia. On the other hand, in Ukhia, 55 percent of individuals were employed in the service sectors, compared to 33 percent in Teknaf. Census 2011 While industry is the dominant employer of women outside agriculture in Bangladesh and Chittagong, in Cox’s Bazar, services are a more important source of non-agricul- tural jobs for women. 4 Among women working outside of agriculture in 2013, two-thirds were working in industry in Bangladesh as a whole. In Chittagong division, the proportion of women working in industry was even higher, at three-quarters. In contrast, in Cox’s Bazar, the main non-agricultural sector of employment for women was services, where more than half of working women were employed. Within services, tailoring, education, and retail trade were the main occupations for women in the district, representing 31, 23, and 18 percent of female service-sector employment. In industry, textile and RMG man- ufacturing were the main cluster for women’s employment, accounting for more than half of women working in industry in Cox’s Bazar. Manufacture of furniture and wood products represented roughly another 20 percent of women’s industrial employment in the district. Among men, non-agricultural employment is dominated by services, with a relatively small share of men relying on RMG and textile manufacturing in Cox’s Bazar. Male employ- ment was heavily dominated by service-sector activities at national, division, and district levels. Services accounted for 76, 77, and 86 percent of non-agricultural male employment Chittagong division as a whole is rapidly urbanizing, with an economy oriented to man- in Bangladesh, Chittagong, and Cox’s Bazar respectively. The main occupation among men ufacturing and export. In contrast, the latest reliable district-level estimates before the employed in services was retail and wholesale trade, representing more than 50 percent Rohingya influx show that Cox’s Bazar at that time still depended heavily on agriculture.3 of male service jobs in Cox’s Bazar. The pattern of employment in the industry sector also The non-agricultural economy of Cox’s Bazar is dominated by 1 and 2 person enterprises differed by gender. While salt extraction and furniture manufacturing employed 30 and 20 in the services sector. At the same time, Cox’s Bazar has a significantly smaller share of percent of men in industry, respectively, the RMG and textile branches employed only 13 employment in industry, when compared with national and division averages. This is percent of Cox’s Bazar’s male industrial workers. driven by a lower presence of RMG industries in Cox’s Bazar and, within industry, the high prevalence of 1-2 person enterprises engaged in RMG and salt extraction, relative to bet- ter-connected areas of Chittagong such as Chittagong zila or Feni. • 4 The Population Census underestimates female work (mainly in the agricultural sector) because it 3 Considering the scarcity of pre-influx data bases representative at zila and upazila level, and given is not designed to capture unpaid and housework employment. For this reason, employment com- that employment shares at district level for Cox’s Bazar using HIES 2016 had large standard errors, parisons by gender cannot be done using Population Census data without presenting misleading employment shares are calculated using the 2011 Population Census. However, recognizing that the information. For this reason, gender comparisons are based on non-agricultural employment using population census is not ideal for this purpose, we revert to LFS and HIES for estimating employment the Economic Census 2013. That being said, estimates at national and division level from different at the national and division level. While agriculture’s ranking holds across these sources, there are data sources suggest that agriculture and services are the main employers for females and males differences in point estimates. respectively. 228 229 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Figure A4-1. Figure A4-1: Employment Employment shares shares in A4-2. Employment Figure A4-2: Figure Employment shares shares by by Table A4-1: Number of non-agricultural workers and shares by firm size and Cox’s Bazar, in Cox’s Chittagong, Bazar, and Chittagong, and upazila, 2011 upazila, 2011 sector Bangladesh, 2011 Bangladesh, 2011 Industry Bangladesh Chittagong Cox’s Bazar 1 worker 277,182 4% 89,875 6% 13,329 31% Ukhia 40% 4% 55% 2 workers 104,962 1% 20,368 1% 2,636 6% 3-4 workers 1,548,556 21% 300,787 20% 10,688 25% 38% Teknaf 51% 16% 33% 46% 43% 5-9 workers 640,722 9% 103,421 7% 6,652 16% 10-35 workers 567,392 8% 50,681 3% 2,608 6% Ramu 56% 11% 32% more than 35 4,185,988 57% 948,546 63% 6,854 16% 7,324,802 100% 1,513,678 100% 42,767 100% 15% Pekua 72% 6% 22% 8% Services 14% Bangladesh Chittagong Cox’s Bazar Maheshkhali 72% 5% 23% 1 worker 3,170,779 19% 391684 13% 27718 13% 55% 2% 43% 2 workers 3,454,434 20% 635670 20% 31618 15% Kutubdia 49% 3-4 workers 4,327,199 26% 992506 32% 76643 35% 47% 40% Cox's Bazar 5-9 workers 3,581,696 21% 698644 22% 60763 28% Sadar 23% 7% 69% 10-35 workers 1,490,082 9% 272310 9% 15812 7% Chakaria 49% 10% 41% more than 35 870,966 5% 126003 4% 4757 2% 16,895,156 100% 3,116,817 100% 217,311 100% Banlgadesh Chittagong Cox's Bazar Total Agricultural Industry Services Bangladesh Chittagong Cox’s Bazar 1 worker 3,447,961 14% 481,559 10% 41,047 15.8% Source: Staff calculations, Population Census 2011. 2 workers 3,559,396 15% 656,038 14% 34,254 13.2% 3-4 workers 5,875,755 24% 1,293,293 28% 87,331 33.6% Despite having a similar firm-size structure, the pattern of non-agricultural employment 5-9 workers 4,222,418 17% 802,065 17% 67,415 25.9% by firm size is different in Cox’s Bazar compared with national and divisional levels. 10-35 workers 2,057,474 8% 322,991 7% 18,420 7.1% While in Cox’s Bazar only 38 percent of industry employment is in firms hiring more than more than 35 5,056,954 21% 1,074,549 23% 11,611 4.5% 5 individuals, in Bangladesh and Chittagong, 3 out of 4 workers are employed by non-mi- 24,219,958 1 4,630,495 1 260,078 1 cro enterprises in the industry sector. In the case of the service sector, on the other hand, Source: Staff calculations, 2013 Economic Census. Cox’s Bazar reflects the national and division pattern of employment being concentrated in small firms, with roughly 2 out of 3 workers engaged in service activities being employed by firms with less than 5 employees (Table A4-1). The differences in employment shares by Non-agricultural employment in Cox’s Bazar was mostly concentrated in Chakaria and firm size across industry and services suggest that larger firms in Cox’s Bazar have a lower Sadar upazilas. Half of individuals employed in non-agricultural activities are based in employment capacity than those at national and division level. Furthermore, considering Chakaria and Cox’s Bazar Sadar, which represent 22 and 26 percent of total non-agricul- that 8 percent of firms with more than 100 employees are in the RMG industry (Farole and tural employment in the district, respectively (Table A4-2). A second group comprised of Cho 2017), the differences in employment pattern by firm size between Cox’s Bazar and Teknaf, Ramu, and Maheshkhali represent a third of the district’s total non-agricultural the national average highlight the importance of the RMG industry as an employer among jobs (accounting for 14, 11, and 10 percent of such jobs, respectively). The distribution of larger firms for the country, but not to the same extent for the district. employees working in services and industries shows similar spatial patterns, but with a 230 231 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 peculiarity for Teknaf. While this upazila is home to 15 percent of service-sector workers, Employment patterns within the service sector were relatively homogeneous across it only hosts 8 percent of individuals employed in industry. In terms of non-agricultural upazilas, dominated by wholesale and retail trade, with education services following a employment composition, Teknaf and Ukhia are the two least-industrialized upazilas, with distant second. Wholesale and retail trade was the most important employer for males services representing 90 and 92 percent of their total non-agricultural employment. On the working in services, accounting for 62 percent of all men in the service sector, on average other hand, the district’s two most industrialized upazilas are Chakaria and Pekua, where (Table A4-3). Teknaf and Maheshkhali are the upazilas with the largest share of service-sec- industry represents 23 and 22 percent of total non-agricultural employment. tor workers involved in trading activities, roughly two-thirds. On the other hand, Sadar upazila has the lowest share, at 56 percent. The second-most-important activity within Table A4-2: Employment distribution and shares by sectors and upazilas the tertiary sector is education, mainly primary-education-related activities. Unlike trade, education represents an important share of female employment. Thirty-five percent of Share of total workers by Share of total non-agricultural female employment is in education services, compared to 7 percent of all Number of workers sector upazila workers non-agricultural employment for women and men. Ukhia has the largest share of non-ag- Industry Services Total Industry Services Total Industry Services ricultural employment in this sector, at 11 percent. Accommodation and food and tailoring services are also important employers across Chakaria 12,781 43,305 56,086 30% 20% 22% 23% 77% upazilas. The former, mainly represented by “tea stall” activities, absorbed 9 percent of the labor force engaged in services, on average. In Cox’s Bazar Sadar and Teknaf, short-term accommodation activities were also an important cluster of employment. These activi- CXB 10,811 55,807 66,618 25% 26% 26% 16% 84% ties have an opposite gender intensity. While tea stalls constitute a slightly higher share of employment for males than females, short-term accommodation represents a larger Kutubdia 1,790 10,974 12,764 4% 5% 5% 14% 86% share among women workers. Tailoring activities absorbed 5 percent of service workers across upazilas, representing only 4 percent of male service-sector jobs but a quarter of female employment in services. These activities are particularly important for women in Maheshkhali 4,155 21,438 25,593 10% 10% 10% 16% 84% some upazilas, representing roughly half of female service-sector employment in Teknaf and Ramu. In contrast, in Kutbudia and Ukhia, only 3 and 9 percent of women working in services were involved in tailoring activities. Pekua 3,342 11,945 15,287 8% 5% 6% 22% 78% The employment composition within industry varies across upazilas, perhaps reflecting difference in underlying economic structure. Manufacture of textiles and RMG is particu- Ramu 4,860 24,834 29,694 11% 11% 11% 16% 84% larly important in some upazilas. In Chakaria, Pekua, and Teknaf, this activity represented 54, 31, and 18 percent of total individuals engaged in industry. At the same time, in these upazilas, RMG was the most important female employment cluster, accounting for 82, 59, Teknaf 3,602 31,719 35,321 8% 15% 14% 10% 90% and 50 percent of female industrial workers. This sector is negligible in Kutubdia and Ukhia. Another important industrial sector in Cox’s Bazar is the “extraction of salt” industry, which employed 24 percent of industrial workers in the district. Salt extraction was relatively Ukhia 1,426 17,289 18,715 3% 8% 7% 8% 92% more important for Kutubdia and Maheshkhali, where it represented 52 and 68 percent of all industrial workers, followed by Sadar, Teknaf, and Pekua, where it accounted for a third of industrial jobs. “Manufacture of furniture” is another important employment sector in Cox’s Bazar 42,767 217,311 260,078 100% 100% 100% 16% 84% Cox’s Bazar, representing 18 percent of industrial workers. Within this industry, manufac- district ture of wooden products is the main activity. Across upazilas, its importance is relatively higher in Ukhia, where it represented 46 percent of industrial employment. Source: Staff calculations, 2013 Economic Census. 232 233 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Table A4-3: Shares of female, male, and total employment in industry and A4-3. Education level Figure A4-3: Figure level and and Cox’s Bazar had relatively low employment services, Cox’s Bazar (main activities) sector of employment, Cox’s Bazar sector quality, when compared with Chittagong and Bangladesh as a whole. According to 100% % of total Farole and Cho (2017), as the Bangladeshi % of female % of male 90% workers 80% economy has experienced structural 70% Extraction of salt 4% 30% 24% 60% transformation, average job quality has 50% improved. However, gains have been 40% uneven across space and gender. Analysis Manufacture of textile and RMG 57% 13% 24% 30% 20% of the 2013 Economic Census shows that Manufacture of wood and products of wood 8% 5% 6% 10% workers in Cox’s Bazar had an elevated 0% probability of working in agriculture or o d y e e ov ry ar et et holding lower-quality jobs such as day ss sch de ab a an ol e im d nd Manufacture of furniture 13% 20% 18% pl pl to tten om om pr an eco labor (Figure A4-3). This is due in part to the ra c nc rs y ve yi th ar ge district’s low average educational attain- Ne ar im Hi Manufacture of food (main rice milling) 4% 5% 5% nd Le Pr co Industry ment. In general, the lower the education, Se Agricultural Industry Services the lower the quality of jobs and the higher Manufacture of other non-metallic (main brick 4% 11% 9% the likelihood of workers’ being involved in block tiles) Source: Staff calculations, Population Census 2011. agriculture. Sixty-one percent of workers Manufacture of jewelry 3% 4% 4% with no education worked in the primary sector. At the other end of the educational distri- bution, 90 percent of people with higher education5 worked in industry and services. The manufacture of fabricated metal product continued importance of agriculture as a source of livelihoods in the district likely reflects 4% 6% 5% (mainly structural metal products) in part its lagging progress in human capital outcomes, described in Chapter 2. Other Industries 3% 6% 5% Trade 18% 64% 61% Additional figures and detailed tables supporting Annex 4 Transport 0% 4% 4% Figure A4-4. Main sectors of employment by zilas in Chittagong Division Figure A4-4: Main sectors of employment by zilas in Chittagong Division Accommodation and food 9% 9% 9% Education 23% 6% 7% Services Tailoring 31% 4% 6% Public administration 3% 2% 2% ni ria i ur r g n i r li a ar at za pu on ha ill a Fe ip rb ba hh am ba m nd ak ag m da Co an c ng sh x's a No itt ra Health 5% 1% 2% n Ch m Ra ag k Ch Ba Co La ah Kh Br Agricultural Industry Services Other services 11% 10% 10% Source: WB staff elaboration, Population Census 2011. Source: Staff calculations, 2013 Economic Census. 5 Individuals with higher education are those individuals who have completed higher secondary level or upper education level. 234 235 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Table A4-4: Employment shares by sector, Bangladesh, Chittagong, Table A4-5: Number of workers and share of industry and services and Cox’s Bazar employment, Chakaria Bangladesh Chittagong Cox’s Bazar % of Sectors % % Total % of % of % of % of % of % of % of % of % of Female Male Total Female Male workers total female males total female males total female males B Mining and quarrying 0.3% 0.4% 0.2% 0.3% 0.1% 0.3% 3.9% 1.6% 4.1% Manufacture of textile and RMG 4,174 2,730 6,904 82% 36% 54% C Manufacturing 29.5% 63.8% 22.7% 32.0% 74.8% 21.7% 12.2% 43.5% 9.1% Extraction of salt 10 1,418 1,428 0% 18% 11% D Electricity, gas, steam and air conditioning 0.2% 0.3% 0.2% 0.2% 0.1% 0.2% 0.2% 0.1% 0.2% supply Manufacture of rice/rice milling 85 279 364 2% 4% 3% E Water supply; sewerage, waste Manufacture of furniture 382 1,710 2,092 7% 22% 16% 0.1% 0.0% 0.1% 0.0% 0.0% 0.1% 0.1% 0.1% 0.1% management and remediation activities Industry Manufacture of jewelry, 70 320 390 1% 4% 3% F Construction 0.2% 0.2% 0.2% 0.1% 0.1% 0.1% 0.1% 0.0% 0.1% bijouterie and related articles Total Industry 30.2% 64.7% 23.4% 32.7% 75.1% 22.4% 16.4% 45.3% 13.6% Manufacture of wood and 208 348 556 4% 5% 4% G Wholesale and retail products of wood trade; repair of motor 34.2% 6.0% 39.9% 34.4% 3.3% 42.0% 50.9% 9.7% 55.0% vehicles and motorcycles Other industries 176 871 1,047 3% 11% 8% H Transportation and 7.6% 2.5% 8.6% 3.5% 1.1% 4.1% 3.2% 0.2% 3.5% storage Total industry 5,105 7,676 12,781 100% 100% 100% I Accommodation and 5.0% 1.0% 5.8% 6.9% 0.9% 8.4% 7.8% 4.8% 8.0% food service activities Trade 497 25,861 26,358 24% 63% 61% J Information and 0.4% 0.3% 0.4% 0.3% 0.2% 0.3% 0.4% 0.5% 0.3% communication Transport and communication 18 2,559 2,577 1% 6% 6% k Financial and insurance 1.9% 2.5% 1.8% 1.9% 1.9% 1.9% 1.6% 3.6% 1.4% activities Accommodation and food 113 3,770 3,883 5% 9% 9% L Real estate activities 0.2% 0.1% 0.2% 0.1% 0.1% 0.1% 0.2% 0.1% 0.2% M Professional, scientific Financial service activities, 0.7% 0.4% 0.7% 0.6% 0.1% 0.7% 0.4% 0.1% 0.5% except insurance and pension 125 476 601 6% 1% 1% and technical activities funding N Administrative and 0.6% 0.2% 0.7% 0.6% 0.1% 0.8% 0.7% 0.2% 0.8% support service activities Services Education 525 2,312 2,837 25% 6% 7% O Public administration and defense; compulsory 2.4% 1.6% 2.5% 2.4% 1.0% 2.7% 1.4% 1.8% 1.3% social security Tailoring services 536 1,680 2,216 26% 4% 5% P Education 6.0% 9.3% 5.4% 5.9% 7.0% 5.6% 5.8% 12.7% 5.1% Hairdressing and other beauty 21 1,258 1,279 1% 3% 3% Q Human health and treatment 1.7% 2.7% 1.5% 1.3% 1.6% 1.2% 1.4% 2.9% 1.2% social work activities R Arts, entertainment Other services 238 3,316 3,554 11% 8% 8% 0.1% 0.0% 0.2% 0.1% 0.0% 0.2% 0.1% 0.1% 0.1% and recreation S Other service activities 8.9% 8.6% 9.0% 9.1% 7.5% 9.5% 9.8% 18.1% 9.0% Total Services 2,073 41,232 43,305 100% 100% 100% Total Services 69.8% 35.3% 76.6% 67.3% 24.9% 77.6% 83.6% 54.7% 86.4% Source: WB staff elaboration, Economic Census 2013. Source: WB staff elaboration, Economic Census 2013. 236 237 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Table A4-6: Number of workers and share of industry and services employment, Table A4-7: Number of workers and share of industry and services employment, Kutubdia Cox’s Bazar Sadar % of % of Total % % Total Female Male Total % Female % Male workers Kutubdia Female Male Total Female Male workers Extraction of salt 271 3386 3657 13% 38% 34% Extraction of salt 1 934 935 0% 60% 52% Manufacture of Textiles 696 257 953 35% 3% 9% and RMG Manufacture of rice/rice milling 22 67 89 9% 4% 5% Manufacture of rice/rice 66 325 391 3% 4% 4% milling Manufacture of brick / block, tiles 41 109 150 17% 7% 8% Manufacture of non-metallic mineral 75 781 856 4% 9% 8% Manufacture of structural metal products n.e.c. 73 155 228 31% 10% 13% products Manufacture of Industry structural metal 101 555 656 5% 6% 6% Manufacture of furniture 62 149 211 26% 10% 12% products Industry Manufacture of wood Manufacture of jewellery and 123 299 422 6% 3% 4% 13 43 56 5% 3% 3% and products of wood related articles Manufacture of wooden 370 1330 1700 18% 15% 16% furniture and fixture Other industries 26 95 121 11% 6% 7% Manufacture of jewellery 75 385 460 4% 4% 4% and related articles Total Industry 238 1552 1790 100% 100% 100% Electricity, gas, steam and air conditioning 23 374 397 1% 4% 4% Trade 16 7148 7164 6% 67% 65% supply Other industries 216 1103 1319 11% 13% 12% Accommodation and food 7 885 892 3% 8% 8% Total Industry 2016 8795 10811 100% 100% 100% Financial service activities, except 34 119 153 13% 1% 1% Trade 857 30596 31453 17% 60% 56% insurance and pension funding Accommodation and 681 6419 7100 13% 13% 13% Education 148 647 795 56% 6% 7% food Financial service activities, except Services Public Administration 19 182 201 7% 2% 2% 322 1158 1480 6% 2% 3% insurance and pension funding Hairdressing and other beauty 3 433 436 1% 4% 4% Services Education 958 2886 3844 19% 6% 7% treatment Hospital activities 291 547 838 6% 1% 2% Tailoring services 8 344 352 3% 3% 3% Public Administration 216 1231 1447 4% 2% 3% Tailoring services 1406 1677 3083 28% 3% 6% Other services 30 951 981 11% 9% 9% Other Services 336 6226 6562 7% 12% 12% Total Services 265 10709 10974 100% 100% 100% Total Services 5067 50740 55807 100% 100% 100% Source: WB staff elaboration, Economic Census 2013. Source: WB staff elaboration, Economic Census 2013. 238 239 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Table A4-8: Number of workers and share of industry and services Table A4-9: Number of workers and share of industry and services employment, Maheshkhali employment, Pekua % of % of % % Total % % Total Female Male Total Female Male workers Female Male Total Female Male workers Extraction of salt 0 511 511 0% 20% 15% Extraction of salt 89 2744 2833 25% 72% 68% Manufacture of textile and RMG 496 526 1022 59% 21% 31% Manufacture of rice/rice milling 31 132 163 9% 3% 4% Manufacture of wood and 99 149 248 12% 6% 7% Manufacture of Textile and RMG 53 104 157 15% 3% 4% products of wood Manufacture of brick / block, tiles 101 247 348 12% 10% 10% Manufacture of wood and prod- Industry 46 200 246 13% 5% 6% ucts of wood Industry Manufacture of wooden furniture 109 755 864 13% 30% 26% and fixture Manufacture of furniture 61 332 393 17% 9% 9% Manufacture of jewellery and 9 103 112 1% 4% 3% Other industries 75 288 363 21% 8% 9% related articles Total Industry 355 3800 4155 100% 100% 100% Other industries 29 208 237 3% 8% 7% Total Industry 843 2499 3342 100% 100% 100% Trade 126 14215 14341 18% 69% 67% Trade 102 7682 7784 35% 66% 65% Accommodation and food 47 1829 1876 7% 9% 9% Accomodation and Food 28 1218 1246 10% 10% 10% Financial service activities, except 37 196 233 5% 1% 1% insurance and pension funding Financial service activities, except 11 46 57 4% 0% 0% insurance and pension funding Public administration and defense; compulsory social 55 197 252 8% 1% 1% Public administration and security defense; compulsory social 12 146 158 4% 1% 1% security Services Education 243 987 1230 35% 5% 6% Services Education 77 589 666 26% 5% 6% Hairdressing and other beauty 11 852 863 2% 4% 4% Hairdressing and other beauty treatment 3 442 445 1% 4% 4% treatment Tailoring services 133 866 999 19% 4% 5% Tailoring services 51 482 533 17% 4% 4% Other Services 41 1603 1644 6% 8% 8% Other services 10 1046 1056 3% 9% 9% Total Services 693 20745 21438 100% 100% 100% Total Services 294 11651 11945 100% 100% 100% Source: WB staff elaboration, Economic Census 2013. Source: WB staff elaboration, Economic Census 2013. 240 241 COX’S BAZAR — INCLUSIVE GROWTH DIAGNOSTIC Annex 4 Table A4-10: Number of workers and share of industry and services Table A4-11: Number of workers and share of industry and services employment, Ramu employment, Teknaf % of % of % % Total % % Total Female Male Total Female Male workers Female Male Total Female Male workers Manufacture of rice/rice milling 67 324 391 7% 8% 8% Extraction of salt 0 666 666 0% 23% 18% Manufacture of textiles and RMG 207 173 380 21% 4% 8% Manufacture of rice/rice milling 16 78 94 2% 3% 3% Manufacture of bamboo and cane Manufacture of Textile and RMG 361 285 646 50% 10% 18% 347 401 748 34% 10% 15% products Manufacture of bamboo and cane 34 141 175 5% 5% 5% Manufacture of brick / block, tiles 96 1444 1540 10% 37% 32% products Industry Manufacture of wooden furniture Manufacture of brick / block, tiles 13 569 582 2% 20% 16% 205 1001 1206 20% 26% 25% and fixture Manufacture of structural metal 35 169 204 5% 6% 6% Manufacture of jewellery and products 33 205 238 3% 5% 5% Industry related articles Manufacture of cutlery, hand 48 92 140 7% 3% 4% Other industries 53 304 357 5% 8% 7% tools and general hardware Manufacture of wooden furniture Total Industry 1008 3852 4860 100% 100% 100% 105 534 639 15% 19% 18% and fixture Trade 275 14084 14359 16% 61% 58% Manufacture of jewelry and 45 193 238 6% 7% 7% related articles Transport and communication 8 2318 2326 0% 10% 9% Other Industries 65 153 218 9% 5% 6% Accommodation and food 30 1846 1876 2% 8% 8% Total Industry 722 2880 3602 100% 100% 100% Financial service activities, except 91 235 326 5% 1% 1% insurance and pension funding Trade 323 20594 20917 17% 69% 66% Education 333 1319 1652 19% 6% 7% Accommodation and food 193 2188 2381 10% 7% 8% Services Health 103 257 360 6% 1% 1% Financial service activities, except 88 425 513 5% 1% 2% insurance and pension funding Hairdressing and other beauty 7 629 636 0% 3% 3% Services treatment Education 289 1693 1982 15% 6% 6% Tailoring services 784 722 1506 45% 3% 6% Tailoring 904 1620 2524 48% 5% 8% Other Services 121 1672 1793 7% 7% 7% Other Services 97 3305 3402 5% 11% 11% Total Services 1752 23082 24834 100% 100% 100% Total Services 1894 29825 31719 100% 100% 100% Source: WB staff elaboration, Economic Census 2013. Source: WB staff elaboration, Economic Census 2013. 242 243