Report No. 35824-BD Bangladesh Dhaka: Improving Living Conditions for the Urban Poor April 16, 2007 Sustainable Development Unit South Asia Region Document of the World Bank PrefaceandAcknowledgements This report was prepared by a team led by Judy L. Baker (World Bank, FEU). Team members who participated in the field work and contributed background papers commissioned for this report include: Nazrul Islam, (University o f Dhaka and Centre for Urban Studies), "Dhaka Land and Housing", Somik Lall, (World Bank, DECRG), "A Poverty Profile for Dhaka"; David L e Blanc and Robert Buckley (World Bank, TUDUR), "Dhaka Urban Poverty: Land and Housing Issues"; Bernice Van Bronkhurst (World Bank, LAC Region), "Crime and Violence", Sabina Faiz Rashid (BRAC University) and Yasmeen Hossain (London School o f Economics), "Constraints to Delivering Services to the Urban Poor Living in Slums in Dhaka, Bangladesh", Claire Salmon, (University of Savoie, France), "The Situation o f the Poor in the Working Population o f Dhaka.", This study also drew on independent work camed out by Rashid and Mannan, 2004, "The Heterogeneity o f the Urban Poor: Political-Economy and Social Conditions in Urban Slums" and Aparajeyo, 2002 "Slum Observatory Survey." Surveys that were commissioned for this study were carried out by Aparajeyo (Crime and Violence), and the Local Government Engineering Department (LGED), (Spatial Poverty Mappingof the Dhaka Metropolitan Area). Data analysis was done by Asha Sundaram and Jonny Anderson. The Coalition for the Urban Poor (CUP) assisted greatly inorganizing meetings with its membersand with local communities. The work was carried out underthe guidance o f Sonia Hammam, Sector Manager, (SASEI) and Christine Wallich, Country Director for Bangladesh. The team benefited from comments from the two peer reviewers, Christine Kessides and Hassan Zaman, as well as from Patricia Annez, EnriquePantoja, Jaehyang So, Catherine Tovey, and Nilufar Ahmed. The report also benefited greatly from the inputs received from participants at the Consultation Workshop which discussed the findings o f the report extensively on Fibruary 14 and 16, 2006. 3 Table of Contents Executive Summary........................................................................................................... iv Chapter 1: A Poverty Profile for Dhaka ............................................................................ 1 I.Introduction................................................................................................................ 1 I1.Poverty and Inequality inDhaka................................................................................ 2 I1. Charactensticsofthe poor.......................................................................................... . . 6 I11.PerceptionsofPoverty.............................................................................................. 9 Chapter 2: Employment andPoverty............................................................................... 13 I.Introduction.............................................................................................................. 13 13 . I11. Employment and the Poor....................................................................................... I1. General Characteristicsofthe Labor Force inDhaka.............................................. 14 A. .General characteristics....................................................................................... 15 B. Employment Sector and Occupation ................................................................... 16 C. Gender. employment andpoverty........................................................................ 19 D. Participation of children in the laborforce ........................................................ 20 F. Un-and underemployment .................................................................................. 21 E. Workinghours and income.................................................................................. 21 IV.Prospectsfor EmploymentGeneration andPolicy Recommendations .................. 22 A. Trends andprospects on the supply side of Dhaka's labor market .................... 27 V. Improvingincome earning opportunities for the poor............................................ B. Trends andprospects on the demand side of Dhaka's labor market ..................22 31 Chapter 3: Shelter for the Poor ........................................................................................ 33 I.Background............................................................................................................... 33 A. A Fast Growing. Poor Megacity ......................................................................... 33 B. High Land Prices Have Excluded the Poorfrom Ownership of Land and Housing ..................................................................................................................... 35 I1. Main policy constraints............................................................................................ 37 A. Unresponsiveness toprice signalsfor a largeportion of the land in the city.....37 B. An enabling housing policy without enablers ..................................................... 39 C. Failures of coordination and services at the local level ..................................... 44 D. Thelimited role of NGOs in housingfor urbanpoor ......................................... 45 I11. Improving Shelter Conditions for the UrbanPoor.................................................. 46 Chapter 4: The Challenges of Service Delivery for Dhaka's Poor.................................. 50 I.TheCurrentSituation.............................................................. ................................ 50 A. Basic Services...................................................................................................... 51 B. Service Providers................................................................................................. 52 I1. Constraintsto Service Delivery................................................................................ 55 A. Lack ofgovernment policy.................................................................................... 55 B. Evictions of slum residents .................................................................................. 55 C. ThePervasive Role of Mastaans......................................................................... 56 D. Other constraints................................................................................................. 56 I11. ImprovingService Delivery for the Urban Poor..................................................... 58 Chapter 5: Crime and Violence and the UrbanPoor ....................................................... 61 I.BackgroundandIntroduction.................................................................................... 61 I1.Manifestations of crime andviolence affecting the urbanpoor inDhaka............... 62 A. Main Survey Findings.......................................................................................... 63 i I11.Relationship between crime andviolence and social. institutional. and legal issues ....................................................................................................................................... 70 IV.Addressing Crime andViolence............................................................................. 72 References......................................................................................................................... SpeciJicrecommendationsfor crime and violenceprevention in Dhaka .................73 80 List o fBoxes Box 2.1: The "missing" women inthe poor income groups............................................ 18 Box 2.2: Post MFA and entry o f China inthe WTO. their impact on the poor inDhaka? A summaryofthe literature.............................................................................................. 29 Box 3.1: Slum ImprovemenWpgrading Programs o f the National Government ............43 Box 3.2: NGO programs targetedto poor renters............................................................. 46 Box 4.1: MakingThings Work inHealth Care................................................................ 54 Box 4.2: Case Study from CUS: Slum Eviction inAgargaon ....................................... 56 Box 4.3: MakingThings Work inWater and Sanitation................................................. 57 Box 4.4: Case Study: Lessons Learned from DSK's Water DeliveryProject in Conjunction with DWASA............................................................................................... 60 Box 5.1: Community Police Stations inMumbai's slum................................................. 78 Box 5.2: Crime and Violence Prevention Components inBank-financed Integrated Slum Upgrading Operations: ...................................................................................................... 79 List of Tables Table 1.1: Slums covered inthe Aparajeyo Survey............................................................ 2 Table 1.2: Characteristics of the poor in Dhaka................................................................ 6 Table 1.3: Distanceto Public Services for Slum Dwellers (inmeters) ............................. 8 Table 1.4: Percentage of total (monthly) householdexpenditure inTakas for Dhaka SMA ............................................................................................................................................. 9 Table 2.1:Employment Characteristics of Householdsby Income Group ...................... 16 Table 2.2: Sectoral distribution o f the male workers inDhaka SMA by income group (in percentage) ......................................................................................................................... 17 Table 2.3: Main Jobs of Male Workers inDhaka............................................................. 17 Table 2.4: Sectoral distribution of female workers inDhaka SMA by income group (in percentage) ........................................................................................................................ 18 Table 2.5: MainJobs of Poor Female Workers ................................................................ 19 Table 2.6: Hours and wages of the poor, by occupation and gender ................................ 21 Table 2.7: Unemployment and Underemployment inDhaka SMA byIncome Group ....22 Table 2.8: Comparison o f Average Monthly Wages betweenDhaka and RuralAreas in 2000 (inTaka) ................................................................................................................... 24 Table 3.1: UrbanPopulationand Poor Population inDCC and DMDP........................... 34 Table 3.2: Housing Tenure ofthe UrbanPoor, 1995........................................................ 36 Table 3.3: Relative importance o fHousing Delivery Sub-sectors inDhaka ................... 39 Table 4.1:Administrative structure of service deliveryinDhaka.................................... 53 Table 5.1: Percentage o fRespondents .affectedby crime and violence........................... 64 11 Table 5.2: Nature o f Crime and Violence Occur inthe Slums (Multi-response) .............64 Table 5.3: The Severity o f Crime and Violence (Multi-response) ................................... 65 Table 5.4: Respondent's recommendation regarding combating existing Crime and Violence (Multi.response) ................................................................................................ 69 Table 5.5: Linking urbanspatial and non-spatial interventions: The KfWI City o f Cape Town KhayelitshaViolence Preventionthrough UrbanUpgradingProject .................... 77 List.o f Figures Figure 1.1: SlumFormation inthe DCC Area. 2005......................................................... 4 Figure 1.2: Slum Ownership. Dhaka. 2005........................................................................ 5 Figure 1.3: Imageof flood slums inflood zone. Dhaka .................................................... 5 Figure 1.4: Access to Public Toilets .................................................................................. 7 Figure 3.1: Map o f LandUse inDhaka City. 1984 .......................................................... 40 Figure3.2: Map o fLandUse inDhaka City. 2004 .......................................................... 41 List o f Annexes Annex 1: Poverty Profile Background Tables and Figures ............................................. 92 Annex 2: Supplemental Data on Labor............................................................................. 95 Annex 3: Supplemental Information on Land and Housing........................................... 118 Annex 4: Supplemental Data on Crime and Violence ................................................... 126 ... 111 EXECUTIVE SUMMARY Introduction Dhaka i s the fastest growing mega-city inthe world, with an estimated 300,000 to 400,000 new migrants, mostly poor, arriving to the city annually.' Its population is currently around 12 million and i s projected to grow to 20 million in2020, making it the world's third largest city. Most migrantscome from rural areas insearch o f opportunities which can provide new livelihood options for millions, translating to improvements in living standards. Their contribution to Dhaka's economic growth is significant, as they provide muchneeded labor to manufacturing, services, and other sectors. This migration, however, also adds tremendous strain on an already crowded city with limited inhabitable land due to the city's topography, limited infrastructure, and a low level o f public services. The draw o f Dhaka i s no surprise - it is a dynamic city and has attracted substantial industrial investments, particularly in the Readymade Garment (RMG) industry, which has created demand for workers and services. The city is, however, increasingly characterized by large slums, poor housing, excessively high land prices, traffic congestion, water shortages, poor sanitation and drainage, irregular electric supply, unplannedconstruction, increasingair pollution and poor urban governance which results ingrowing problems o flaw and order. The country does not have a-comprehensive policy on urbanization and urban poverty. There are between 16 and 40 different bodies involved in one way or another in urban matters inDhaka with little coordination and planning. As a result, there are major gaps in services and infrastructure ranging from weak electrical supply to inadequate land and housing options, and major traffic congestion. The poor are particularly affected as they do not have the resources to find alternatives for meeting their basic needs. While a new committee for UrbanDevelopment was set up andhas achieved some progress inthe coordination process, it focuses on solving day-to-day problems and not on medium to longterm strategy. This study reflects a comprehensive look at poverty in Dhaka with an aim to provide the basis for an urban poverty reduction strategy for the Government o f Bangladesh, local authorities, donors, and NGOs. While the needs in Dhaka are enormous, this study focuses on analyzing those critical for the poor - understanding the characteristics and dynamics o f poverty, issues o f employment, land and housing, basic services, and crime and violence. This analysis will provide a platform for developing recommendations for policy reform as Dhaka endeavors to meet the growing challenges and urgent needs o f the urban poor. Each chapter concludes with a set o f recommended priorities for poverty reduction. ' BangladeshBureau o f Statistics. iv Povertyin Dhaka As one travels through Dhaka, the pervasive poverty is evident, as is the inequality between rich and poor. Poverty affects roughly one in three residents. The average person in the wealthiest quintile consumes more than 5 times the consumption expenditure o f the average person inthe poorest quintile. The poor mainly live in slums scattered throughout the city, with close to 80 percent o f slums located on privately owned land creating considerable institutional challenges in terms o f basic service provision.2 Housing structures tend to be o f poor quality, and access to basic infrastructure services is low. For the poorest quintile, only 9 percent o f households have a sewage line, and 27 percent obtain water through piped supply (compared with 83 percent o f the wealthiest). Spatial mapping shows that only 43 o f the 1925 identified slums are within 100meters o f a public toilet. An estimated 7,600 households live in slums that are within 50 meters of the river and are in frequent risk of being flooded. As would be expected, poor households tend to have more people, particularly children, than non-poor households. The poor are substantially less educated and have lower school attendance rates than the non-poor. Social services can be quite far for some slum dwellers exacerbating problems o f access. Only seven percent o f slums have a public health clinic and 26 percent have a government school. Perceptions ofpoverty show a hierarchy among the poor linkedto income, as well as an individual's occupation and power, position and networks. Residents in focus groups refer to the many hardshipso f living in slums and enormous stresses o f everyday life. Employmentand the Poor Dhaka's poor work in a range o f sectors providing muchneededlabor to the city. Much o f this employment i s in the informal sector. Poor male workers are mainly employedas productionworkers (including rickshaw pullers and other transport workers) and trade workers (street vendors, retail trade, etc.). Half o f the poor female workers are employed as domestic workers or garment workers. Wages are generally low, though domestic work appears to be the lowest pay. The unemployment rate for the poor i s almost double than that of the non-poor. Underemploymentis an evenbroader problem, affecting 20 percent ofhouseholds. Child labor i s quite prevalent, with approximately 20 percent o f all children between ages 5-14 working. Most o f the child workers are between 10 and 14 years old, with similar proportions for boys and girls. In the poorest households with child workers, earnings from the children are significant, representingabout one third o f total householdincome. 'LGEDSurvey of Slums, 2005. V Improving income-earning opportunities for the poor will be dependent on ensuring a vibrant labor market inDhaka. Several factors are likely to influence the labor market over the coming years. On the supply side is the issue o f continuing migration which will both put additional pressure on the labor market, and also bring new capital, new skills andnew entrepreneurs. A second issue is the role o f education which can raise the productivity o f the poor and their income-earning potential. While education levels o f the poor in Dhaka are far better than in the rest o f the country, the average level o f education o f the poor i s low compared with other developing countries. Improving education levels in Dhaka and Bangladesh as a whole will be critical to keeping the country competitive inglobal markets. On the demand side are the consequences o f the removal o f the Multi-Fiber Agreement (MFA) quotas on the readymade garment (RMG) sector at the end o f 2004. While there are a number o f complex factors, the impact o f which are not fully known, there i s still a risk that some jobs might be lost in the future. An estimated 500,000 workers are employed in the RMG sector in Dhaka. Most o f these workers are women, who provide critical support to their families. Substantial numbers also work in related services industries. While the first half o f 2005 has not seen a decline in RMG exports, the future is unclear. Ifthere is a decline in demand as i s projected, Dhaka's labor force will be directly affected. A second issue on the demand side is the'development o f self employment through expanding opportunities for micro-credit in urban areas. ' Bangladesh i s known for its successes in micro-credit through the Grameen Bank and others, though these programs have generally not beenoperating inurbanareas. While a few have begunoperating inDhaka, there i s substantial scope for expanding micro-credit opportunities, particularly with the participation o fNGOs. Shelter for the Poor Secure shelter is a major challenge for Dhaka's urban poor. As migrants continue to amve, they often end up in illegal settlements on precarious lands with major environmental concerns. The slums are located throughout the city with few services offered at high prices through middlemen, also called musclemen or mastaans, using illegal methods. Slum evictions take place periodically, with no resettlement plans. The constant threat o f eviction adds to the stresses o f every day life for the urban poor. Land prices in the Dhaka City Corporation (DCC) area are very high making it impossible for the poor to afford to purchase land in the open market. Part o f the reason for high land prices stems from the fact that the City suffers from a shortage o f land due to its topography. This shortage has driven prices up to levels that are considered high even by developed country standards. The real scarcity o f developable land is exacerbated by an artificial scarcity, stemming primarily from the current use o f substantial amounts o f public land in downtown Dhaka. Large plots are taken up by university land and residences, a deserted airport, and an army cantonment. In addition, close to 10 percent o f the city's housing stock i s publicly owned and held in low-rise buildingon very expensive land. vi The Government has developed a National Housing Policy which reflects an enabling approach to land and housing markets. This Policy, however, has not been implemented effectively. Among the inconsistencies inimplementation are the estimated 70 percent o f urbandevelopment inthe City that is informal, large scale evictions o fslum dwellers with no relocation plans, and buildingon environmentally sensitive public lands such as flood or retention ponds. Part o f the lack o f enforcement o f the policy i s related to the fact that the institutional capacity for city management i s weak. The roles o f DCC, the PlanningAuthority RAJUK, and line ministries are not clear and coordination is very limited. Services for the Poor Access to basic services such as water, sanitation, electricity, health and education by the poor is limited. Because most slums are not recognized as legal lands, the Government, NGOs and Donors generally do not provide services in these areas. As a result, a parallel structure has emerged with mastaans providing services for a high fee. They are self appointed leaders who set up committees, maintain links and have patronage from local and national political leaders, government official and local law- enforcing agencies. As with housing, the administrative structure related to the public delivery o f services i s complex. Public services are delivered by a mix o f central and local agencies, with limitedresources, weak administrative capacity and little coordination. DCC, which i s responsible for a wide range o f services, cannot perform their functions adequately due to severe resource constraints and limited authority. DCC i s dependent on central government for financial grants and for staff appointments, and has limited role in city planning or physical development. Interviews with donors, government agencies and NGOs revealed three major constraints to providing service delivery to urban slum residents. These include: i)a lack o f government policy giving slum residents rights and access; ii)the eviction o f slum residents; and iii)the role o f mastaans inthe absence o f formal government structures. The lack o f government policy on urbandevelopment and urbanpoverty results in a conflicting power structure between the DCC and the national level, and a lack o f authority to work in slum areas. The urgentneed for a clear policy and mandated agency to address the challenges o f urban development, urban poverty and service delivery i s evidenced by current conditions. The constant threat o f evictions prevents agencies from investingin infrastructure for health, education, water and sanitation services. All agencies mentioned this as a barrier for effective service delivery. Mastaans are firmly rooted in the slums and are often the only service provider. Residents report paying high amounts for the services, much higher than the rates offered by the utilities, and regularly face extortion for fear o f physical harm or eviction ifpayoffs are not made. vii A series o f other constraints were also mentioned by implementing service providers in NGOs, Government and Donor organizations. Among these are resources, lack o f appropriate infrastructure -- particularly schools and health clinics, limits in technical capacity in the various agencies, difficulties in working with the various agencies, and concerns about discerning between effective and ineffective NGOs. Crime and Violence inDhaka's Slums Crime and violence has become a major problem in Dhaka, particularly affecting the poor. It has huge c,osts to the City in terms o f the judicial system, health care, foregone earnings, costs on private security, loss o f competitiveness, loss o f jobs and productivity. The non-monetary costs on the population are also considerable interms o f levels o f fear and mistrust, anxiety, and lack o f social cohesion. A survey o f crime and violence in four slums commissioned for this study provides new data on the prevalence and characteristics o f crime and violence. An astounding 93 percent o f respondents inthe survey report that they had been affected by crime and violence over the last 12 months with 33 different types o f crime identified by the respondents. Among the most commonly reported crime and violence are toll collection, mastaan-induced violence, drug and alcohol business, land grabbing, gambling, violence against women and children, illegal arms business, arson in slums, murder and kidnapping, and domestic violence. These results vary somewhat across slums. Most of the crimes are reported to occur within the slum, and about 35 percent within the household compound. Other highrisk areas are the roadside, markets, and the route to and from work. Violence against women, particularly domestic violence, i s very high and likely underreported in the survey. This is consistent with other studies on violence against women inBangladesh. There appears to be a deep lack o f trust by the urban poor in the justice system and police. The residents in the slums do not perceive much security and do not believe the justice system adequately serves their needs. In fact, some perceive that the system discriminates against slum dwellers. A majority o f victims (60 percent) state that the incidence o f crime and violence is not reported to anyone. When reports are made, it i s to community leaders and family members. Only 3 percent say that they reported the incident to the police and even fewer report to Ward Commissioners. The survey indicated that the police took action against a reported perpetrator inonly 1percent o f all cases. A Strategy Forward ... The problems o f poverty in Dhaka are enormous and thus any poverty reduction strategy for the city will require major policy shifts on many fronts. Overall improvements in the city's performance will contribute to economic growth which will have positive impacts for the poor. The city desperately needs strong management, better ... Vlll coordination across agencies, effective and transparent local government, improvements in infrastructure, better access to basic services, enabling reforms in land and housing markets, credit markets for the poor, and a reduction o f the disabling problem o f crime andviolence. Prioritizing the long list o f pressingneeds and identifyingviable mechanisms for implementation presents a daunting challenge. This will require coordinated actions from many stakeholders, but more importantly, strong political commitment to addressing the growing problems o f the urbanpoor in Dhaka within the context o f the overall problems of poverty in the country. The priorities for poverty reduction in Dhaka outlined below emerged from the analysis inthis report and through a consultativeprocess heldinDhaka in February 2006 including representatives from Government, NGOs, researchers, the private sector and donors. Actions are needed to ensure, in general, the right environment to promote growth and poverty reduction in Dhaka, and in particular to improve access to employment, services, and a reduction in crime and violence for the poor. These actions, outlined below and detailed inthe following chapters, shouldbe framed by an urban development strategy for Dhaka, and aligned with the country's Poverty Reduction Strategy Paper (PRSP). Simultaneously, overall reforms o f the judiciaVpolicing systems, reductions in crime and violence, and transparent ways of attaining services are necessary for Dhaka to continue to provide an enablingurban environment to its citizens. Policy Priorities Developing and implementing a comprehens;ve strategy for urban poverty reduction. Such a framework should address the needs o f the poor within the context o f overall urban development and the PRSP. Essential elements include identifying institutional roles and responsibilities for national and local governments as well as the NGO and private sector, creating an appropriate political and legal environment, and ensuring adequate resources for implementation. This strategy should be prepared through a consultative process led by the Office o f the Prime Minister, in coordination with key ministries and agencies such as the Ministryo f Local Government and Rural Development (MLGRD), as well as with the Planning Department (MOF) team overseeing PRSP implementation.As part o f implementation, MLGRD should establish an urban unit within LGD to review and develop policies and ensure implementation o f urban issues, including urbanpoverty reduction. One o f the key objectives of the Urban Poverty Strategy should be to improve income earning opportunities for the urbanpoor. Reducing poverty inDhaka will be linked to generating enough remunerative employment opportunities for new entrants and for the unemployed. Beyond the overall policy o f improving the infrastructure'inthe city to permit manufacturing growth and attract investment, 1x other priorities include: i)identifying new areas for employment generation; ii) improving the skills o f workers; iii)increasing access to credit for the poor to enable an expansion o f self employment activities; and iv) facilitating increases in female labor force participation through access to child care. Implementing institutional mapping, reform and capacity building of key agencies / institutions affecting urban growth and poverty reduction. Initially, GOB should focus on RAJUK and DCC, as part of a comprehensive effort by MLGRD to strengthen local government and agencies influencing Dhaka's development. In general, with clear roles and responsibilities, agencies should be given authority, but made accountable for implementation. Extensive training in urban management and in technical areas will be necessary for agencies at the local level to do theirjobs effectively. Implementing the National Housing Policy to ensure shelter for thepoor. The National Housing Policy provides an enabling framework for addressing land and housing markets in Dhaka, and enforcing basic property rights. The implementation o f this policy, as with an urban poverty reduction strategy, will require an adequate institutional framework with clear leadership, as well as roles andresponsibilities. The Government should establish an inter-agency committee including key actors such as LGED, National Housing Authority, Ministry o f LanMinistry o f Housing and Public Works, and RAJUK. Inthe medium term, the committee should be led by a clearly mandated agency with authority to implementthe necessary reforms. To improve the shelter conditions for the urbanpoor, it will be critical to keep up with Dhaka's growth and address the massive land and housing problems of the urban poor, which will require reforms on many levels. In addition to the importance o f implementing an Urban Development Strategy for Dhaka and the National Housing Policy (including enforcement o f basic property rights), other priorities include: i)developingmechanisms for better accountability for land use and better coordination o f services inthe city; ii)strengtheningpublic institutions to implement land and housingpolicies and foster coordination between agencies; iii)initiatingpilotshelterprojectsinpoorareasinpartnershipwithNGOs;andiv) addressing specific legal and regulatory issues that apply to zoning, land subdivisionandbuilding. Improving service delivery and access to infrastructure for the urban poor. To address the obstacles to delivering services to Dhaka's slums, recommendations from the review and consultations point to two overarching priorities: first, developing and implementinga policy o f urbanpoverty reduction which provides the mandate for working in slum areas, and second, strengthening the role o f local municipalities to deliver services. Other priorities include: i) improving accountability and oversight; ii)enforcing law and order to curtail illegal activities linked to extortion by the mastaans; iii)improving coordination between service providers; iv) exploring and implementing alternative delivery X mechanisms for some services (e.g. distance learning, mobile health units, etc.); and v) buildingcapacity for service providers through training. Infrastructure improvements will equally be needed to address the major deficiencies inproviding access to basic services to the poor living in slum areas. Such improvements will require an enablingpolicy and institutional framework as mentioned above, as well as substantial capital investment. To ensure a comprehensive and coordinated approach, DCC should work in close coordination with key agencies and ministries such as Dhaka Water and Sewerage Authority (DWASA), Dhaka Electricity Supply Authority (DESA), Titas, Ministry o f Communications, Ministry of Education; Ministry o f Health and FamilyWelfare, and Ministry o fSocial Welfare. Addressing crime and violence in slum areas. The general importance o f enforcing law andorder inDhaka i s highlightedabove. Importantly, however, the problems o f crime and violence, as well as the illegal activities o f the mastaans, which are pervasive particularly in slum areas. In addition to systemic judicial/policing reform, specific recommendations for crime and violence prevention stemming from international experience point to the need to: i)adopt a community based approach; ii)recover or create public space in slums and integrated slum upgrading; iii)initiate programs for the prevention o f violence against women; iv) improve trust between the police and community; v) focus attention to the needs o f children and youth; and vi) focus attention to issues o f drug and alcohol abuse. Most o f these efforts require concerted effort by the Ministry o f Women and Children Affairs, but community involvement i s essential. These are critical priority areas not just for the urban poor but indeed for urban development overall. Development partners stand ready to support GOB in implementation o f this complex agenda to improve the conditions o f the urban poor. First i s support at the policy level which would facilitate the formulation and implementation o f major reforms. Second i s continued support at the program and project level to include investments in infrastructure, service delivery systems, and human capital targeted to the poor o f Dhaka. Because o f the many complexities inworking inDhaka, a series o f new pilots could be initiated in slum areas which would be carefully monitored and evaluated so that the most appropriate approaches could be scaled up over time. . x i CHAPTER 1: A POVERTY PROFILEFORDHAKA I.Introduction Dhaka's population is estimated at around 12 million, or about one third o f Bangladesh's urban population. Its primacy i s both a source o f concern and ~pportunity.~ Concern, because the population growth creates significant urbanmanagement challenges interms o f provision of public services, local public goods and amenities. Opportunity, because the city provides new livelihood options for millions o fmigrants, translating into improvements intheir living standards. Dhaka is growing rapidly --between 1995 and 2000, the city grew at an average rate o f 4.24 percent per year. Much o f its growth stems from migration, with 46 percent o f its 1991 population born outside the metro area. Rural to urban migration i s attributed to extreme rural poverty and landlessness, and large urban-rural wage differentials (Ullah, 2004). Already strained to provide services and quality o f life to existing urban residents, Dhaka i s likely to face tremendous challenges in expanding existing infrastructure and avoiding deterioration o f living standards due to congestion, pollution, and lack of basic services. These inadequate services and worsening environmental conditions disproportionately affect the poor, many of whom live inslums. This chapter presents an overview o f poverty in the Dhaka Metropolitan Area based on data from several sources: i)the 2000 round o f the Household Income and Expenditure Survey (HIES), conducted by the Bangladesh Bureau o f Statistics; ii)the 2002 Slum Observatory Survey carried out by Aparajeyo-Bangladesh (NGO); iii)2005 spatial mapping data developed for this study and carried out by the Local Government EngineeringDepartment (LGED); and iv) qualitative data on the perceptions o f poverty carried out by Rashid and Mannan in2004. The HIES survey is a national survey which covered 7440 households inthe 2000 round in the five divisions o f Dhaka, Chittagong, Barishal, Rajshahi and Khulna. The sample i s stratified into Statistical Metropolitan areas, Rural areas and Urban Municipalities. Seven hundred households in the survey are from the Dhaka Statistical Metropolitan Area (Dhaka SMA). The slum survey carried out by Aparajeyo-Bangladesh is part o f their `Slum Observatory' project which runs an annual household survey infour different slums since 2002 (Table 1.1). The slums differ in age and type of settlement. The sample includes 1000 households, approximately 250 from each slum. The respondents were randomly selected between both users and non-users of Aparajeyo-Bangladesh services and ' measures were taken to ensure the inclusion o f different occupational groups in the sample. The survey attempts to re-interview the same households in subsequent years. Dhaka's primacy i s considerably higher than other countries inthe region. Inthe region, Mumbai accounts for 4.1 percent o f the Indian urbanpopulation, and Karachi accounts from 21percent o f Pakistan's urban population. 1 Slum Name Thana Ward Number Area Bizly Mohalla Mohammadpur 42 Mohammadpur Bou Bazar Hazaribagh 48 Dhanmondi Shahid Nagar Lalbagh 60 Azimpur Sujat Nagar/Nazml MollahBastee Pallobi 06 Pallobi The Poverty Mapping was carried out by LGED. The work developed spatially detailed composite GIS based digital maps o f i)facility locations (schools, health clinics, roads, public standpipes and public toilets); ii)slums in the DMA; and iii)a digital attribute dataset o f the features. This drew on two data sources. The first was a physical survey carried out from September 2004-February 2005 to collect the location o f each slum, boundary o f slum, and attributes including year o f establishment, land ownership, number o f households, water source, sanitation, electricity, educational facilities in slum areas, and health facilities in slums. Locations were captured with GPS devices. The second data source was satellite image processing which involved geo- referencing the data, preparation o f layered digital maps, and linking attribute data with the GIS database. This data set will also serve as a planning tool for DMA. In-depthqualitative research was carried out by Rashid and,Mannanintwo slums in Dhaka, Beguntila and Moderbasti, and one in Chittgagong (New Shaheed Lane), supported by DFID.4 The field work was carried out in 2003 including wealth ranking, 15 focus group discussions with men, women and children. Separate informal discussions were held with leaders, key informants and individuals in the community. An additional 10case studieswere carriedout with individuals from the categories ofless poor, more poor, and very poor. While the multiplicity o f data sources adds richness to the analysis, the sources are not directly compatible and thus it i s not possible to use all the data sources for each subtopic o f analysis. Also noteworthy, is the fact that this study did not explicitly covered pavement dwellers inthe analysis as it would have required a specially designed survey. Further analysis is recommended on this subgroup among t'hepoor. 11. Povertyand Inequalityin Dhaka Poverty estimates for Dhaka range depending on the methodology used. While measuring urban poverty presents a number o f challenges given its multidimensionality, for the purposes o f this study, we use a standard consumption expenditure measure as it i s broadly accepted as an objective approach which can be used for a range o f socio- economic variable^.^ Rashid and Mannan, 2004. See Coudel, Hentschel and Wodon, inWorld Bank, 2002, and Baker and Schuler, World Bank, 2004. 2 The poverty lines for Dhaka are those used in the 2000 HIES preliminary report issued by the Bangladesh Bureau o f Statistics (BBS and the World Bank Poverty Assessment, 2002). They are calculated on the basis o f the Cost o f Basic Needs approach, representing basic needs food consumption for the lower poverty line, andboth food and non-food consumption inthe upper line.6For the Dhaka Statistical Metropolitan Area, the lower poverty line (extreme) i s (Taka) Tk. 649 and the upper poverty line i s Tk. 893. Based on this methodology, approximately 28 percent of Dhaka's population or 3.36 million people were classified as poor, and 12 percent as extremely poor in 2000. This estimate o f the poverty head count index based on the HIES is conservative relative to estimates previously reported in the literature which places between 37 and 48 percent o f Dhaka's population under the poverty line.7 An explanation for this i s attributed to the difference in data sets and samples (Salmon, 2004). Estimates based on the Labor Force Survey (LFS) use income data rather than consumption (income data tends to under- represent household welfare), and cover an additional two thanas in the Dhaka SMA which are particularly poor. Regardless o f which poverty estimate is used, the characteristics o fthe poor inDhaka do not change.8 The incidence o f poverty is substantially lower than the national headcount o f 50 percent, rural areas (53 percent), as well as other main cities o f Chittagong (46 percent) and Khulna (50 percent).' The number o f poor people is, however, higher inDhaka than inother cities given its size. Poverty incidence in Bangladesh decreased by 9 percent during the nineties, and an even higher amount, 14 percent, in the Dhaka Division. This encouraging trend, however, i s dampened given that the number o f poor people has stayed the same. The progress in reducing poverty was equal across urban and rural areas, though rural areas did better inreducingthe depthand severity o fpoverty. Inequality is higher in Dhaka than the country as a whole, and other cities. The Gini coefficient inDhaka is 0.37, compared to 0.31 for the country, 0.29 for Chittagong and 0.35 for Khulna." This inequality is dramatic when comparing average household consumption which i s 5 times higher for the richest quintile as compared to the poorest (Table 1.2)." The food requirementi s based on 2122kcal per day per person. 7See for example, Khan and Siddique (2000). Islam, (2004) proposes Tk. 4500 (US2.50) per day as the poverty line, resulting ina headcount of 40 percent. * See Salmon, 2004 for a profile of the poor basedon the 2000 LFS. See World Bank, 2002, BangladeshPoverty Assessment loKernel Densityhas been calculatedusing analytic population weights. 'IQuintiles havebeencalculatedfor SMA Dhaka based on per capita householdexpenditureusing householdsamplingweights. Household sampling weights have also beenusedinall survey estimationso f means. 3 Figure 1.1: Slum F~~~~~~~inthe DGGArea. 2005 Figure 1.2: SlumOwnership, Dhaka, 2005 KC zone Not DCC zone Land Owner * No Data . Rwate Government i N Many slums are located in low lying areas near the river and are prone to flooding. Approximately 7,600 households in 44 slums live within 50 meters of the river and are at risk o fbeing flooded (see Figure 1.3). Figure 1.3: Image of Flood Slums inFloodZone, Dhaka 5 11. Characteristics of the poor As would be expected, poor households tend to have more people, particularly children, than non-poor households. In the four slums o f the slum observatory study, more than halfo f the household members are under the age of 19 (56 percent). The poor are substantially less educated, and have lower school attendance rates than the non-poor. Only 40 percent o f household heads among the poor have achieved more than 5 years of schooling, and only 6 percent have hadmore than 10years (HIES). I I=poorest I 2 3 4 I 5=richest Mean HHsize (No. of 5.50 4.99 4.84 4.86- 4.53 individuals) (0.13) (0.18) (0.17) (0.16) (0.15) Average number of rooms per 1.24 1.52 1.76 2.35 3.40 household (0.06) (0.08) (0.09) (0.12) (0.24) M e a nNo. of children (< 15 yrs) 2.62 2.11 1.83 1.87 1.74 (0.10) (0.09) (0.08) (0.09) (0.08) Mean No. of adults (15 to 64 yrs) 2.89 3.10 3.20 3.30 3.23 (0.12) (0.14) (0.13) (0.12) (0.12) M e a nNo. of old (>64 yrs) 1.13 1.06 1.32 1.11 1.08 Cement 0.07 0.20 0.30 0.53 0.83 HempMay 0.04 0.02 0.01 0.00 0.00 Size of Housing (Sq. Ft.) 1 201.1 294.4 370.9 545.9 967.3 (29.71) I(29.70) 1(30.2) I (43.0) I (80.9) I Sewerage line 0.09 I 0.21 I 0.24 I 0.40 0.47 Water Piped Supply 0.27 0.48 0.59 0.77 0.83 Tube wells 0.73 0.5 1 0.40 0.23 0.17 Electricity 0.88 0.91 0.96 100.0 100.0 Telephone 0.0 0.0 0.03 0.09 0.35 The housing stock for the poor i s o f low quality with a majority living in houses built with temporary kutcha materials made of hemp or hay walls. Access to basic 6 services i s a major problem for the poor as outlined below and discussed in depth in Chapter 4. Infrastructure services. Basic infrastructure services are limited inpoor areas of Dhaka with the exception o f electricity. Seventy percent of households under the poverty line do not have access to piped water and use tube wells as their main water source. Ninetypercent do not have access to a sewage line (Table 1.2). The spatial data on slums is consistent with the above, showing low access to services, though somewhat variable depending on the service, and zone (Table 1.3). For example, only 43 o f the 1925 identified slums are within 100 meters o f a public toilet (Figure 1.4). In fact, inZones 2, 5, 8, 9, and 10, there are no public toilets within 100 meters of slum settlements. The largest slum, Korali Basti inBanani, with more than 12,000 households, does not have a single public toilet or health clinic. Problems o f poor sanitation are made worse by the highpopulation density inslums, which have considerable implications for transmission o f communicable diseases and other negative externalities. The Slum Observatory data suggests that service delivery in slums located on government owned land i s worse than inslumsonprivatelyownedland.l4 Figure 1.4: Access to Public Toilets 101 -5M)o l4For example, 87 percent of slumhouseholdson government owned land did not have a bathroom while the correspondingfigure was only 49 percent for householdsinprivately owned slum land. 7 Table 1.3: Distanceto Public Services for Slum Dwellers (inmeters) Data Source: LGED Slum survey, 2005; Analysis by Lall. Social Services. Accessing schools and health clinics is also limitedfor the poor. School enrollment for children in the poorest quintiles i s 45 percent compared with 76 percent for ages 5-18. As with infrastructure services, the location o f health and education facilities varies across zones. Only 7.3 percent o f slums have a public health clinic, and 26 percent have a government school (LGED). Inthe absenceofpublicproviders,NGOpresence is strong- 20percent ofslums have a NGO operated clinic and 27 percent have a NGO operated school. In the Slum Observatory slums, 76 percent of those interviewed reported that the schools existing in their slum were run by NGOs. However, there was not much awareness among slum dwellers regarding other services being provided by NGOs - such as health, micro-credit or skill-training facilities. Within slums, the use of health services varies across income groups. Only 19 percent o f households with monthly incomes less than Tk. 2000 reported receiving treatment from modernclinics inthe slum (Slum Observatory Survey). However, in case o f the two higher income groups (Tk. 2000-5000/month and over Tk. 5000), about 30 percent o f households report that they had used local health clinics. Interms o f maternal health, a lower proportion o f households in the lowest income category reported that expectant mothers had utilized the services o f doctors during pregnancy. Further, while 84 percent o f the households in the 5000+ income category knew about family planning measures, only 68 percent o f households in the lowest income group had such information. Householdassets and spending. The poor spend the majority o f their budget on food (62 percent). Other major expenditures are housing (14 percent), and other non- food items (14 percent) (Table 1.4). Spendingpatterns are quite different as compared to the non-poor who spend a much lower proportion o f their householdbudget on food (32 8 percent), but a higher proportion on housing (24 percent), other non-food items (17 percent), transport (10 percent), and health and education (11percent) (Table 1S). Both household assets and savings appear to be minimal, though this does vary from household to household (Slum Observatory Survey). Table 1.4: Percentage o f Total (monthly) HouseholdExpenditure inTakas for Dhaka SMA StandardErrorsinparentheses;Data Source: HIES Survey, 2000 111. Perceptionsof Poverty Perceptions o f poverty from the perspective o f slum dwellers can be gleaned from a recent field study camed out by Rashid and Mannan in two slums in Dhaka City (Beguntila and Moderbasti) and one in Chittagong (New Shaheed Lane)." Beguntilla i s a relatively new slum, established in 1999 and Moderbasti i s older, established approximately 20 years ago. New Shaheed Lane i s about 15 years old. The summary below provides some information on all three, but largely focuses on perceptions from the two Dhaka slums. Poverty i s perceived by slum dwellers as a shortfall in income, as well as a number o f socio-cultural, political and economic factors as described below. The main indicator raised during focus groups was income, but was linked to an individual's power, position and networks, and occupation. During the household listing and wealth ranking, households were divided into 3 categories as described below. Being a leader or having a salaried job placed individuals inthe highest category, while those with `lower' jobs were ranked inthe less well off categories. 1. `Less' Poor characteristics: 0 Have a secure income (e.g., landlords, secure labor- DCC/railway workers; government workers, permanent jobs, drivers o fprivate companies, family members abroad sending money) *5Mannan,F andRashid, SF (2004), The Heterogeneityofthe UrbanPoor: Political-Economyand Social ConditionsinUrbanSlums.University of Southampton, UK.Fundedby DFID (project number: KAR ProjectR8028) 9 e Eat three meals a day e Less dependency on others e Householdwith a highnumber o fworking members e Belong to the socially respectable class -includes locally powerful individuals -mastaans, rich landlords e More education compared to others e Send children to school e Good intra householdrelationship e Fewer young children within the household e Strong social capital - strong networks inside and outside the slum 0 Political linkages -strong political links e Are leaders o f the slum/ respected by all e Basic facilities inthe household (water and sanitation, electricity, access to health services etc.) e Some assets (better housing, T.V. radio, beds, furniture, valuable utensils, good cloths, somejewelry, poultry, etc.) 2. `More' Poor include those who do not have secure income andwholly depend on their physical labor. Ifthey cannot work they cannot eat. Characteristics include: e Rickshawhan pullers(not owners o frickshaws) e Day laborers e Self-employed - vendors. e Insecure income (e.g., petty business, garment workers) e Cannot have three meals a day e Are more dependent on others e Have a household with a less number o fworking members e Belong to the socially less important class e Have no or little education compared to others e Cannot send children to school e Tensions inintra household relationship e More minor children within the household e N o social capital -poor networks inside and outside the slum e No political linkages e Less empowered section e Not having basic facilities properly (water and sanitation, electricity, access to health services etc.) e Fewer assets (poorer housing, no T.V., radio, beds, furniture, valuable utensils, good clothes, lessjewelry etc.) 3. The `Poorest' include those who do not have regular income and are extremely vulnerable. Characteristics include: 0 Beggars, Widows, Elderly and the Disabled Female-headed households with small children without any male support N o secure income (e.g. erratic employment, daily laborers, begging) 10 e Have one meal a day (iflucky) e Are more dependent on others e Have a householdwith number o funemployed members e Belongto the socially less prestigious class e Have no education compared to others e Cannot send children to school e Tension within intra household relationship/absence o f male member and/or adult children e More minor children within the household e No social capital e No political linkages e Less empowered section e Less access to basic facilities (water and sanitation, electricity, access to health services etc.) e No assets or little assets (no land, housing) Causes of Poverty. Perceptions o f the causes o f poverty varied between men and women. Both men and women attribute the main cause o f poverty to low income. Men also stressed external factors, networking, lack o f jobs, and low social prestige as the main causes o f poverty. For women, the absence o f male support was a main cause. Desertion was seen a major crisis for women not only because o f the loss o f income, but also the loss o f social, cultural and physical protection. It i s also generally thought to be higher than the data reflect. In the relatively anonymous environment o f urban slums, mamage break-ups are difficult to stop as families and community relationships are not as strong as inrural areas. Land and Housing. Security o f tenure is mentioned as a high priority by the poor, particularly those living in the newer slum area (Beguntilla). They refer to the constant fear o f eviction, extortion by mastaans, and the political connections requiredto acquire land. Those who are government workers, railway workers, and slum leaders appear to benefit from certain rights and entitlements. The majority o f slum dwellers, however, do not have such connections and thus perceive having few if any rights to secure tenure. Services. Residents in focus groups refer to the many hardships o f living in slums. These stresses are enormous. For women, there are additional pressures due to cultural norms which affect where and how they can bath, use latrines, and find drinking water. Accessing water can be very time consuming, physically demanding, and expensive. Water i s typically purchased from private tube well owners and from those with access to legal and illegal connections. Rates can be as high as 15 times the official unit rate. It is not uncommon for it to take 2 hours to collect water (Feroz, Ahmed, 2004). With regard to latrines, there are long lines for the facilities which are dirty, badly maintained and lack privacy. The long waits lead to heated exchanges among residents, the lack o f hygiene contributes to illness, and women face security risks if using the 11 latrines at night. Electricity supply and collection o f fees are reported to be controlled by local mastaans and can be very expensive for residents. Employment. The employment conditions varied somewhat across slums. In Beguntila, located on the margins o f the city, most slum residents work in the informal sector as rickshaw pullers or wage laborers. The slum i s considered to be isolated from the main markets in Pollabi and Mirpur which heavily impacts on income earning opportunities. The slum is an estimated 20-30 minutewalk from the nearest big market. .In contrast, Moderbasti is, more centrally located near to the railway station, main highway and waterway. A significant number o f residents are employed in government jobs and more skilled labor which is attributed to the fact that the slum is much older and more established, and more centrally located. Residents spoke o f choosing to live inthis slum because o f its prime location with leads to betterjob opportunities. Regardless o f location, finding employment, particularly for informal jobs, is closely linked to social networks. Common occupations for menthat were mentioned include day laborers, vendors, vegetable gardeners, rickshaw pullers, bus conductors, drivers, garment factory workers, shop-keepedassistants, and begging (for the poorest/disabled). Rickshaw pulling was the most popular because it requires no skills and i s easily available. Income is estimated at Tk. 40-200 per day, depending on the number o f hours worked and ifthe worker owns the rickshaw. The better paidjobs are perceived to be drivers (who can earn up to Tk. 5000-6000 per month), garment workers (particularly for men working as operators earning up to Tk. 3500 per month). For women, the perceived options were far more limited and included garment factory workers ( Tk. 700-950 per month), domestic help (Tk.300-500 per month), vegetable vendors (Tk: 50 daily), and sweepers (Tk. 500-800 per month). Beggingis reported to be a coping mechanism for abandoned women with children who can earn Tk. 250 per week. The following chapters discuss the key issues for Dhaka's poor inmore detail. 12 CHAPTER EMPLOYMENT POVERTY 2: AND I.Introduction Dhaka is a city o f opportunity for many. Migrants flock to the city in search o f jobs and better lives. The appeal o f the capital city is strong with prospects for employment ina range o f sectors, particularly services and industry. The labor market i s dynamic, with entrepreneurs amving every day hoping to carve out a decent living for themselves andtheir families. Yet for the estimated 28 percent o f the city's poor, earning a living sufficient to meet the basic needs o f themselves and their families is a continual struggle. Low wages, underemployment, unemployment and low skills levels are all challenges faced by the poor inthe labor market. Understanding and addressing these problems is key to poverty reduction. This chapter presents analysis o f the characteristics o f the labor market and links with the poor, with potential areas for policy change. The analysis is mainly based on the 2000 Labor Force Survey.16 For the purposes o f this study we focus on the Dhaka Statistical Metropolitan Area (DMA) which includes Dhaka (also called the Dhaka City Corporation (DCC) area), and the periphery area.17 While this data i s insightful and provides a good basis for analysis, there are some limitations: it is not current and the situation inDhaka i s rapidly changing; there are no data on travel to work, no panel data, and there are shortcomings with the classification o f domestic workers living with their employers (Box 2.1). 11. GeneralCharacteristicsof the Labor Force in Dhaka The labor force in Dhaka grew twice as fast as the country as a whole in the late nineties largely due to continuing migration and increasing female labor force participation. Between 1996 and 2000, Dhaka's labor force grew by 15 percent as compared with 7 percent for the country as a whole. There were an estimated 3.5 million people in the Dhaka S M A labor force in 2000, representing 59 percent o f the total working age population in the city (10 years and over) - 84 percent o f the male and 33 percent o fthe female working age population. 18 16More detailed analysis i s included inthe background paper prepared for this study (Salmon, 2004). The list o fthanas (municipalities) covered is included in Annex, Table A2. The sample o f DMA i s composed o f 1500households, and 6225 individuals. The sub-sample o f 861 children between ages 5-10 received a specific questionnaire on their possible employment and school attendance. I8Thisrefers to the extended definition o fthe labor force which includes any person aged 10years and over who was either employed (worked at least one hour ina week) for pay or profit who was with /without pay or profit during the reference period as economically active. It includes own household economic activities. By contrast, the usual definition excludes these own household activities. The Standard unemployment rate is absolutely unemployedtotal labor force. The "Proposed" unemployment rate i s total unemployed persons (absolutely unemployed + unpaid family workers)/total labor force. 13 The main sector o f employment is services, which employs more than two-thirds o f the city's population, substantially higher than the country as a whole (25 percent), and the second main city Chittagong, SMA (50 percent). Industry represents 20 percent o f all employment inDhaka with about half o f that inthe garment sector. This accounts for 28 percent o f female employment, similar to Chittagong but higher than other cities. Interestingly, the remaining 11percent work in agriculture, which i s explained by the fact that a significant part o f the landinDhaka SMA i s non-urban. This land i s found on the peripheryof the City where a large share o f the low income population lives. The share of formal employment in Dhaka SMA i s much higher than the rest o f the country at 51 percent including 14 percent in the public formal sector and 37 percent private (37 percent). For the country as a whole, formal sector employment only reaches 20 percent. As expected, education levels are higher in Dhaka SMA than for the rest o f the country with more thanhalf o f the employedpopulation aged 15 years and over attaining a level o f education higher than Class VI. This compares with only 25 percent for the country. Female labor force participation i s considered relatively high compared to other South Asian countries, though lower than most African countries and some East Asian countries. In Dhaka SMA women represent 33 percent o f the labor force as compared with 18 percent inNew Delhiand 14percent inCalcutta (Census o f India, 2001). unemployment in Dhaka i s between 7.4 percent and 10.4 percent depending on the rate used (standard or proposed). Assuming the proposed rate, this is slightly lower than the country as a whole (11percent), but higher than the cities o f Khulna (7 percent) and Rajshahi (8.6 percent). Underemployment, while an issue in the labor market in Dhaka at 16 percent is substantially lower than the rest of the country (35 percent). This i s explained by the fact that underemploymenti s more common for agricultural workers. Underemploymentrate for Chittagong is 23 percent of the labor force and 27 percent in Khulna. 111. Employment and the Poor For the purposes o f the poverty analysis in this chapter, the population was classified into three groups: Poor. The individuals who belong to the three lower deciles with per capita expenditures below the official lower poverty line. Near Poor. The individuals who belong to the deciles 4 and 5 and live on the margins o f poverty. Their expenditures per capita are higher than the lower poverty line, Underemploymentrate i s people working less than 35hper weeldemployedlabor force. Unemployment and underemploymentrate is total unemployedpersons + underemployedpersons/totallabor force. 14 but it is likelythat many ofthese households fall inand out o fextreme poverty depending on seasonal earnings and other conditions inthe labor market. Non-poor. Individuals who belong to the deciles 6 to 10 with expenditures per capita over the poverty line enabling them to meet their basic needs. A. General characteristics A number o f characteristics of the poor in the labor market emerge from the analysis o f the LFS survey. A household is more likely to be poor if the head o f the household is unemployed or underemployed, works in the informal sector, and i s a day laborer, or works intransport or agriculture. Women and children inpoor households are more likely to enter the labor force, and work in low quality jobs, than in non-poor households. As was seeninChapter 1, the demographics o fthe householdare closely linkedto the probabilities of being poor. While a large, number o f dependents increases the probability o f being poor as would be expected, the structure o f the household i s also highly important. The presence o f a highproportion o f male adolescents aged 9 to 14 in the household decrease the probability to be poor though there i s no significant effect o f the share o f girls aged 9 to 14 on the probability to be poor. While it is common for young girls inpoor households to work, a larger share are not paid and a significant share that are working, are not counted as a member o f their own household, but as member o f their employer's householdwhere they work as live-inmaids. This creates a possible bias inthe data (Box 2.1). In contrast, the higher the share of adults, male or female, the lower the probability o f being poor. Among the other variables, education stands out - the more educated the head o f the household, the lower the probability o f beingpoor. Estimations also show that the age o f the head also has explanatory power once the effects o f the structure o f the household are controlled for. The older the head, the lower the probability of the household being categorized as poor. As expected, the fact of owning no assets - such as a rickshaw, a sewing machine - increases the probability o f being poor. 15 Table 2.1:Employment Characteristics o fHouseholds by Income Group IIVariable boor* NearDoor Non-Poor IIII Mean Mean Mean II INumberof observations 454 296 750 IHousehold Size 5.13 4.68 4.40 IRatio o f earnershousehold size 0.30 IIIII 0.34 1IIII 0.36 III % male workers unemployed or under-employedl 16 15 11 % of households where the wife is working 29 30 23 % of income earned by female workers 16 12 9 YOofhouseholds where a childis working 28 18 16 % of income earnedby children workers 6 2 1 Y ofhousehold heads employees 35 39 45 O h o f household heads day laborers 13 5 1 % of household heads self-employed 44 50 40 % of household heads inthe formal sector 43 49 63 1%1%1%1o f household heads intranmort 1 2 0 1 1 7 1 7 o f household heads intrade 18 I 28 I 29 o f household heads inagriculture II l l I 8 I 5 III % o fhousehold heads inmanufactunng I 1 1 1 1 0 1 9 1 % of household heads inconstruction 5 4 3 % of household heads infinance 1 2 7 Annex 2. * Poor represents Deciles 1-3;Near Poor Deciles 4-5; Non-Poor, Deciles 6-10 B. Employriiertt Sector arid Occiipatioir More than two thirds o f male workers from poor households are found in two categories o f jobs - production workers (including transport laborers such as rickshaw pullers) and trade workers (street vendors, retail trade). Poor male workers are also over- represented as agricultural workers. Agriculture and transport activities are almost exclusively employment for the poor, which i s not the case for trade. 16 Table 2.2: Sectoral Distribution o fMale Workers inDhaka SMA by Income Group (in Notes: extendeddefinition ofthe labor force. Table 2.3: MainJobs o fMale Workers inDhaka I Poor male workers All Male workers Messengerloffice boy 4 3 Working proprietoriretail trade 10 15 Other sales workers 25 17 Latrine cleaner 4 2 Farmer & other agricultural activities 12 8 Rickshaw puller 10 5 lothertransportworkers 111 16 /Production worker (except transport) 123 118 II Miscellaneous 2 26 Total 100% 100% Employment options for poor female workers are even more limited. Given gender norms and lower education levels, women are excluded from a large range o f sectors and occupations such as transport, services, and trade (Annex 2, Table A2.18). Poor female workers are over-represented in the agricultural sector and the household services, once Live-in Domestics o f the better-off households were excluded (Box 2.1). The surprisingly strong importance of the agricultural sector (19 percent o f the female labor force but 25 percent o f the employment of the poor female workers) is both explained by the geographical contour o f the sample and the fact that the extended definition o fthe labor force i s used.17 17 I Box 2.1: The "missing ''women in thepoor income groups The analysis o f female workers in Dhaka i s constrained by a shortcoming in the data set. Live-in domestic workers are counted as a member o f the household o f their employers who tend to be in the upper deciles. They are therefore, not counted as members o f the households where their families are, which are largely poor. This distorts data on the characteristics o f female employment, the correIation o f poverty and domestic work, and earnings among poor families. A number of studies show that most female workers hand over a large proportion -or even all-of their pay to the head o f the family (Elson, 1999; Amin & ai., 1997; Ward & al., 2004, Kabeer, 1991). According to Amin & al. (1997), female workers are strongly expected to contribute toward household expenses regardless o f whether they live at home or independently. Yet this data does not appear to be captured in the income o f poor households. Data on the destination of incomes o f the gains o f the live-in-domestics i s not available in the LFS data set. Approximately 16 percent o f child workers work as non-relatives in the better-off households and thus are misclassified. A majority of poor female workers are employed intwo types ofjobs, as domestic workers (16 percent) and garment workers (32 percent) (Table 2.5). As discussed above, it is noteworthy that the share of live-in domestics amongst, the poor income groups is probably highly underestimated as most o f them are counted as members o f the better-off households. If one assumes that all live-in maids would come from the three lower deciles, the share of maids among the poorest female workers would reach about 35 percent instead of 16 percent. Garment workers tend to come from the middle deciles (4 and 5), with the financial contribution to household income helping to escape extreme poverty. Table 2.4: Sectoral distribution of female workers inDhaka SMA by income group (in I Formal public sector 2.7 14.8 20.5 14.1 17.1 Formal private sector 22.7 33.7 48.7 32.3 41.0 Informal sector 53.7 51.4 30.7 53.2 41.8 ~ N o n institutional sector 0.3 0 0.1 0.3 0.1 Total 100.00 100.00 100.00 100.00 100.00 18 Table 2.5: MainJobs o fPoor Female Workers Notes:Employedpopulation, 10years andover *Share of maids statisticallyunder-estimatedamongthe low income groups. See text. C. Gender, employment andpoverty The entry to the labor force o f women i s associated with level o f income, education, employment characteristics o f men in the household, and the presence o f children inthe household. Approximately 30 percent o fwomen inpoor households work as compared with 23 percent inthe wealthiest households. Given the lower participation and lower income levels than men, the financial contribution o f earnings to the household are also particularly low. LnDhaka SMA, only 8 percent o f total income i s brought home by women. Interestingly, the probit models showed that the employment characteristics o f the head (male) or o f the other men o f the household play a part in the probability for a woman to enter the labor force with women more likely to work when there are males in the household who are un-or underemployed. Currently, married women are more likely ,to work than divorced women or widows. Mamed women are less likely to work as domestic workers - live-in or live-out. Garment factories tend to employ young single women. Studies show that for a large proportion o f young women the decision to enter into employment, was made by the parents (Amin & al. (1997, Zaman, 2001, Kibria, 2001). Garment work i s also seen by women and their households as an alternative to early mamage as reported in Amin and al. (1997). With regard to education, ,low levels o f education o f women among the poor are associated with a higher participation in the labor force. Women with an intermediate level o f education (class 6 to 10) are more likely to work in the garment sector than to be non-workers. The lowest education levels are highly associated with the probability o f beingemployed as a domestic worker. The likelihood o f entering the labor force drops when young children (under 5) are present in the household. This would indicate potential for increased labor force participation with adequate day care options for mothers With young children. Infact, the 19 probit models show that the presence o f other women in the household explains the probability o f female employment, particularly when considering the extended definition o f the labor force. D. Participation of children in the laborforce Approximately 20 percent o f children between the ages 5-14 are found to be w~rking.'~ Most o f the children workers are between 10 and 14 years old. Surprisingly, this proportion i s similar for girls and boys (Annex 2, Table 2.21). Most boys are found in the trade sector and most girls in the household services. Manufacturing employed about 25 percent o f all child workers (equal for boys and girls). At first glance, the proportion o f child workers seems to be almost the same across per-capita income deciles. This is linked to a measurement problem explained in Box 2.1. Child workers are found to work around 34 hours per week (33 hours for the girls and 36 for the boys). However, there i s a great heterogeneity within this variable.. First, the younger workers - aged under 10 years old - work less than the older children, i.e. around 17 hours per week. Secondly, the working time i s substantially higher for the children living with their employer as they were found to work, on average, 58 hours per week. This mainly affects girls that are live-in domestics. Around one third of all child workers are paid with most o f them 10 years old and above. The youngest children are very rarely paid. In the poorest households with child workers (the lowest two deciles) earnings from the child workers represent a signi5cant contribution o f one-third o f the total income o f the household.20 On average, for all the deciles, the earnings of the children represent around 20 percent o f household income within the households where there are paid children. The average monthly income o f a paid child worker was Tk. 848 inDhaka in2000. Probit models corroborate this relationship between the level o f poverty of the household and the child labor (Annex 2, Table 2.22). Among the household characteristics playing a role inthe decision to enter the labor force for the children, three variables stand out: the share o f adult males inthe household, the share o f adult males un- or underemployed and the presence o f very young children in the household. Children, particularly boys, are more likely to enter the labor force when the share o f males over 15 years old i s low or when the share o f male over 15 years either underemployed or unemployedis high.Inboth cases, it seems that childrenenter the labor force because all other economic resources available inthe householdhave already been used. l9LFS2000. These estimates are considerably different from those inthe HIES2000 data. Inthe HIES, only 5 percent o f all children aged 5-14 were reported to be employed (urban and rural), and another 10 percent were looking for employment. This large discrepancy inparticipation rates may be due to differences inthe definition o f labor force participation inthe two data sets. The LFS uses a more extended *'On definition o f the labor force than the HIES. average, the earnings o f the children represent only 2 percent of the total income o f the whole households o f Dhaka SMA. 20 E. Workinghours and income There is much variation in the number o f hours worked and earnings among the poor (Table 2.6). For example, on average, a poor rickshaw wallah works 8 hours more thanthe "urban farmer" (about 55 hours / week versus 47 hours / week). The differential is far more striking for women. Female live-in domestic work 22 hours more per week thangarment workers. Moreover, the remuneration o f average live-in domestics is by far the lowest. Monthly wages among the extreme poor rank from about Tk. 700 - for the live-in domestic-to about Tk. 2,800 - for the transport workers. By comparison, the average level o f monthly wages inDhaka is Tk. 4,159 for males and Tk. 1,600 for the females for all income groups. Regardless o f the job, the remuneration o f poor female workers i s lower than that o fpoor male workers. Poor male workers Poor female workers Rickshaw Farmer Vendor Live-in Maid Garment Puller & domestics.* worker other transport workers Numberofobservations 111 70 136 98 38 79 Working time 54.9 h. 46.9 h 50.7 h 60.4h 43.5 h 38 h Percentageofpop. working 27% 19% 27% 45% 11% 11% more than 60 hoursper week Monthly wage Tk. 2,837 Tk.2,120 ** Tk. 1,688 Tk. 703 Tk. 731 ** Tk. 1,125 ** Self net income Tk. 2,126 Tk. 3,571 ** Tk. 2,730 F. Un-and underemployment Unemployment and underemployment are major problems for the poor.21 Unemployment i s almost double that o f the non-poor regardless o f which definition is used ranging from 10-14 percent for the poor. Underemployment affects 20 percent o f households (Table 2.7). Standardunemploymentis defined as the number ofunemployedas a proportionof the total labor force. The "proposed" unemploymentrate i s the total unemployed(unemployedplus unpaid family workers) as a proportion of the total labor force. Underemployment i s those working less than 35 hours per week, as a proportion o f the employed labor force. 21 The The precarious The Total poor non-poor Unemployment rate . - 10.0% 8.1% 5.4% 7.4% Unemployment rate proposed 14.3% 10.5% 7.9% 10.4% Underemployment 21.2% 16.3% 12.7% 16.0% This is the case for both men and women. Unemployment affects about 9.5 percent o f the poorest male workers versus around 5 percent o f the wealthiest workers. For females, about 25 percent o f the poor are unemployed compared to 12 percent for the non-poor.22 Underemployment affects almost half o f poor women versus one-third o f those who live inthe upper deciles. As the live-in domestics work longer hours than the average, the fact o f omitting them from the upper deciles tends to narrow the gap between the different income groups. If they were statistically reintroduced in their own households, the differences o f rates o f unemployment and underemployment across deciles might disappear inthe female group. IV. Prospects for EmploymentGenerationand PolicyRecommendations The above analysis demonstrates a strong link between poverty and employment. It is widely accepted that highrates o f productive and remunerative employment growth and highrates o f economic growth contribute to poverty reduction. Ina labor-surplus economy likeBangladesh, accessing productive employment is one o f the main routes to escaping poverty. In the case o f Dhaka, the challenge i s to create enough remunerative employment opportunities for all the new entrants and the unemployed in order to decrease poverty. Among the priority areas are: i)identifying new areas for employment generation; ii)improving the skills o f workers; and iii) increasing access to credit to enable an expansion o f self employment activities. Different factors are likely to influence these priorities for Dhaka's labor market. On the supply side i s the issue o f migration, and the role o f education. On the demand side, are the consequences o f the Multi-Fiber Agreement (MFA) removal on the RMG sector, and opportunities for expansion o f micro-credit in urban areas. These are discussed indetail below. A. Trends andprospects on the supply side of Dhaka's labor market i.MigrationtoDhaka The average annual growth rate o f the urban population in Bangladesh was around 3.5 percent during the last decade (BBS, 2001). During the same period, the growth o f the rural population was 1.5 percent per year. One could estimate that at the 22Based on the proposed definition o f unemployment. 22 current rate o f urbanization, about one-third o f the whole population will move to urban areas by 2010. As Dhaka is one o fthe main destinations for migrants, the city is projected to grow by 7 million people to 19.5 million inthe next 10 years.23 One o f the main reasons for this dramatic growth i s the constant influx o f rural migrants. Previous estimates show that the rural-urban migration rate contributed between three-fifths to two-thirds o f the urban growth rate in Banglade~h.~~ The rural to urban migration rate has not slowed down since the middle o f the eighties reaching 4.9 percent duringthe eighties and 5.9 percent inthe nineties.25 A number o f studies in Dhaka have found that when asking people why they migrated, the main reason was to find a job. This i s consistent with the literature from other countries.26 The principal "push" factor reported by migrants in Dhaka was insufficient job prospects in villages. At the same time, the perception o f a high probability o f getting ajob and earning a higher income in urban areas are the main pull factors. Other push factors include rural poverty, land erosion in their village, and perception o f better education, health services, and social amenities in the City.27 Many migrants interviewed in these studies report that they had nowhere to go after their land was destroyed by floods. In such conditions, Dhaka or other large cities are seen as possible places to try to earn a living. "When my father left my family for good, we had no other option but to come to Dhaka and find ways to survive. In our village we do not have any work. I made the journey with my brother and my two sisters, because everyone said that Dhaka is like a paradise where everyone canfind employment in the garmentfactories (...) ". A garment worker in Dhaka, cited by Oxfam (2004). The labor market in Dhaka versus rural areas. The comparison o f wages and employment characteristics between the rural areas and Dhaka suggests a significant wage gap between urban and rural wages providing incentives to migrate (Table 2.8). Average wages are significantly higher inDhaka and other cities than wages prevailing in the villages (LFS 2000).28This is true for all education levels, all sectors and both sexes. On average, accounting for differences in cost o f living between the capital and the rest o f the country, wages are approximately 40 percent higher in Dhaka and net self employment incomes are approximately 21percent higher. 2 3 C O Hand ACHR, 2000. ~ 24UN,1993 25According to the samplevital registration system, BBS. These figures do not includemigrationto foreign countries. 26According to recent surveys carried out inBangladeshon this topic, (see Aparajeyo-Bangladesh, 2003; Hossain, 2001; Bhuyan & al., 2001; COHRE and ACHR, 2000). 27 seeAparajeyo-Bangladesh,2003; Hossain, 2001; Bhuyan & al., 2001; COHRE andACHR, 2000). 28Cost of living differences are calculatedusing the Cost of Basic Needs (CBN) poverty lines, defined by geographic areas. Two correctionsare made for comparisonon wages and net selfemployment incomes. See Annex Tables A3. 23 The wage differential is the highest for the most educated workers, and those who work inthe construction, trade and finance sectors. The differential is more pronounced for female wage-earners than their male counterparts. By geographic area, differentials are greatest between Dhaka and the rural Rajshahi and Pabna areas, and substantial between Dhaka and other urban areas. For example, the differential between Dhaka and urban Chittagong for both wages and net self employment income is about 50 percent. Paradoxically, wage differentials are not very large for the manufacturing sector. Once differences incost o f living are taken into account, there i s only an 11percent differential for work inmanufacturinginDhaka as compared to rural areas. Dhaka (a) Rural areas (b) Rural'areas(b) Ratio (ah) Ratio (ah) Correction 1 Correction 2 Correction Correction [Average monthlv wage bv sector Agriculture 1768 1389 1208 *1.27 *1.46 Manufacturing 2329 2426 2100 0.96 1.11 Construction 6273 *24 10 *2057 *2.60 *3.05 Trade 3000 *1950 *1676 *1.54 *1.79 Transport 3918 *3476 *2978 I *1.13 I *1.32 Finance 7270 4951 4266 I 1.47 I 1.70 Community services 4507 4442 3781 1.01 1.19 Householdservices 1370 1088 940 1.26 1.46 Male ' 4159 2891 2841 1.44 1.46 Female 1600 1181 1015 1.35 1.58 The gap between agricultural wages and industrial wages has beenwidening since the mid-eighties (BBS). The wage gap reached about 40 percent at the beginning of the 24 nineties versus almost 100percent at the end o f the nineties. This gap does not, however, necessarily reflect actual wages for new migrants coming to Dhaka. A large proportion o f new migrants find employment in the informal sector, where wages are lower than in the formal industrial sector. Moreover, the cost o f living is higher in the city than in the rural areas, which decreasespurchasingpower. The perceived probability o f getting ajob inthe city greatly affects the decision to migrate. Comparative data on unemployment and underemployment in Dhaka versus rural areas is mixed. Underemployment i s far lower inDhaka than inthe rural areas (22 percent lower), and unemployment depends on which definition is used. Using the proposedunemployment rate (which includes the unpaid family workers), unemployment is lower in Dhaka (10.4 percent) than in rural areas (11.3 percent). The official unemployment rate shows the opposite - higher unemployment in Dhaka than in rural villages (7.4 percent versus 3.3 percent) - except for the highly educated. Another issue is the reliability o f the information on which prospective migrants are basing their decision. One study found that about 80 percent o f migrants mentioned prior migrants as the principal source o f information on the job prospects inthe city. The remaining migrants had no explicit reason for assigning the probability o f getting ajob in the citg9 Inpractice, migrants' perception o f the probability o f getting ajob inthe cities was generally reported to be very optimistic compared to the actual experiences. About two-thirds o f migrants reported that they had serious difficulties in finding a job after migration and that the procgss was particularly long. There are possible explanations for the misperception o f migrants' expectations with regard to findingjobs or wage levels. First, they view certainjobs as available and "well-paid", particularly in the garment industry and the public sector.30 Second, studies have shown that there is a feeling among the rural population o f an urban bias where urban areas benefit from priorities in terms o f public expenditures, access to good jobs, amenities, etc. 31 Consequences of rural-urban 'migration for the poor in Dhaka. The question o f whether the rural-urban migration i s beneficial or detrimental i s a controversial issue in the literature. It can provide labor needed for urban industrial growth, contribute to urban services, and provide opportunities for many. Remittances sent to rural areas can also provide substantial assistance to families. On the other hand, it can drain rural areas o f skilled individuals and the influx o f migrants to cities, particularly at a rapid pace, can strain the urban infrastructure, environment and labor markets. 29 Bhuyan & a1 (2001 '30 The development o f the RMG sector has played a very important part inthe economic development o f Dhaka but also in its demographic growth. The strong growth o f Dhaka that has been observed this last decade is for a large part due to the dramatic growth o f the RMG sector. 31 See Bhuyan & al, 2001 it is noteworthy that, before migration, around half the respondents report that they thought o f a probability o f between 0.5 and 0.75 o f getting ajob inthe city. Given that the differentials o f wages are particularly wide, the perceived employment opportunities remain highenough to pushpeople to migrate (Bhuyan & al., 2001). 25 Many migrants coming to Dhaka end up in slums where living conditions are particularly grim. New migrants cite a number o f major drawbacks once they arrive to urban areas: overpopulation, polluted environment, lack o fjobs, and deteriorating law and order (Bhuyan & al, 2001). However, in spite o f these negative factors, migrants do not express a desire to go back to their villages inmost surveys. In fact, evidence shows that, on average, after migration, monthly household expenditures o f migrant households increased by 40 percent. (Bhuyan & al, 2001). This substantial increase inearnings seems to compensate for most o f the drawbacks o f life inDhaka. For city managers, the dramatic growth of Dhaka has generated a number o f problems including providing adequate water-supply, sanitation and basic services, the management o f garbage, the increasing risk o f criminality and violence, and the deterioration o f environmental conditions. To address the influx o f migrants, the GOB has undertaken several policies such as forced evictions (discussed in Chapter 3) and the . Ghore Fera (Back to home program). This program encourages people to returnto their village by offering them loans for income generating activities in their village, and expenses for transportation and resettlement, though has not been considered very successful. As long as job prospects in cities look promising to the rural poor, many o f those who can will continue to migrate. ii. Theroleofeducation Investments in human capital for the poor can play a major role in boosting economic development. The increase o f human capital can raise the productivity o f the poor and their income earning potential. While.education levels in Dhaka are far better than inthe rest o f the country, the average level o f education remains low compared with other developing countries. Literacy rates for workers aged 15 and older are only 52 percent among poor male workers and 33 percent for female workers. Moreover, vocational training i s very limited, regardless o f income group. Less than 10 percent o f male workers and 6 percent o f the female workers have received technical or a vocational training (Annex 2, Table A2.8). Education costs can be prohibitive for the poorest. While most reports show that households - even poor households - demonstrate a strong willingness to send their children to school by investing huge sums in their education, the poorest o f them are limited by financial constraints. Even though the government has the country-wide Compulsory Primary Education Program, and the Food-For-Education (FFE) program mainly in rural areas, education remains very expensive for a poor family - both in monetary terms and interms o f opportunity cost. Vocational-technical training also remains very low relative to labor market needs. The quality o f instruction i s also cited to be low (Knowles, 2001). There are few linkages between the output o f the system and the demand for trained manpower. Significant improvement in this type o f education appears to be greatly needed, 26 particular176 through the development o f a private system o f training and vocational education. B. Trendsaridprospects or1 the demand side of Dhaka's labor market The extent to which the growth o f a sector contributes to employment generation can be analyzed through the employment elasticity o f sectors. The estimate o f output or value-added elasticity measures the responsiveness o f employment with respect to a change in output or value-added. High employment elasticity in a sector usually means potential for job creation. For the economy o f Bangladesh as a whole, estimates o f the employment elasticity with respect to sectoral growth show strong disparities among sectors. The service sector appears to be the most employment friendly with an elasticity over 1 (Table 5.4). The manufacturing sector i s somewhat more controversial and i s discussed below. Some o fthe discrepancies are due to data issues (Salmon, 2004). i.Employmenttrendsinthemanufacturingsector The manufacturing sector represents about 17 percent o f the total labor force in Dhaka SMA, with the RMG sector accounting for 12percent. Its development interms of output or value-added i s generally considered to be highly important to generating additional job opportunities in order to absorb the additional urban labor force.37 A disaggregated analysis o f the employment elasticity relative to value-added and output for 23 categories shows that the highest elasticities (higher than 0.75) included relatively low skilled industries such as textile industries,. bricks, and ship breaking (Rahman & Islam, 2003). Among them i s the Readymade Garment (RMG) industry - with an employment elasticity o f 0.85 to 0.96 with respect to val~e-added.~~ ii.HowdoestheRMGsectoraffectpoverty? Since the beginning o f the eighties, the RMG sector has undergone considerable growth interms of production, exports and employment. At the beginning o f the 2000's, the sector provided 76 percent o f Bangladesh's foreign earnings.33 RMGs employed about 200,000 workers in 1985 which grew to approximately 2 million in2003. Much of this job growth has been in Dhaka, which now employs about one-fourth o f all garment workers. Most garment workers are women (approximately 90 percent), who support an estimated 10 million dependents (Ward, 2004). Moreover, related service activities are estimated to employ another 2 to 5 million people dependingon the source (Ward, 2004; Ahmed and Sattar, 2004). 32Rahman& Islam(2003) for period 1980-1998 33Onabout 3,600 factories that operate inBangladesh, some 800-900 engage indirect exports, while the remainder works on subcontracts (Ward, 2004). 27 At a microeconomic level, there are a number of studies showing improvements inliving conditions for females who have entered into the garment sector (Kabeer, 1991; Amin & al, 1997). Working ina garment factory is seen by thousands o frural women as a way to escape extreme poverty (Kibria, 2001). Women inpoor urbanhouseholds report that garment employment i s perceived as a good opportunity compared with other types o f unskilled employment available to women with low levels o f education in urban areas (such as domestic work). The financial contribution o f Dhaka's garment workers accounts for one-third o f their family income (LFS, 2000). Wages are, on average, significantly higher than inthe other sectors offered to women. Several studies have also found that the development o f the garment sector has played a significant role in contributing to the empowerment o f women, particularly among the poorest o f them (Kabeer, 1991; Amin & al, 1997). Despite the positive benefits o f the RMG sector, recent growth and prospects for future development are less encouraging. Performance over the past few years has demonstrated that Bangladesh's garment sector was highly dependent and, thus, vulnerable to external shocks such as international events and changes in trade agreements. This industry, seasonal by nature, has become increasingly irregular interms o f employment since 2001. In August 2001, a few months after the U.S Trade Development Act-2000 became effective (which gave duty-free access and trade preference to African and Caribbean countries), more than half o f the Bangladeshi factories reported to have no orders. After September 11, 2001, orders declined rapidly so that by December 2001, it was estimated that nearly 1,300 factories closed and 400,000 women lost their jobs (Ward, 2004). A number of studies have focused on the effects o f closures o f garment factories finding it was extremely difficult for these women to find anotherjob giventhe lack o f otherjob opportunities inDhaka (Shefali, 2002). Additionally, external competition has increased since the entry of China in the WTO and the end of the MFA (Multi-Fiber Agreement) in December 2004. The MFA had provided Bangladesh with a guaranteed market in North America under the quota regime.At the end of2004, all quota restrictions were abolished. It is difficult to quantify the impact o f the many factors that the phase out of the MFA will have on employment in the RMG sector (Box 2.2). In fact, the reduction o f RMG exports is anticipated to affect all the households through the depreciation of the real exchange required to offset the decline in export earnings and through the overall reduction inlabor demand (Amdt, & al, 2002). A general equilibrium model o f the Global Trade Analysis Project estimates that the decrease of garment production could reach more than 10 percent.34 This then would translate to a decrease in direct employment by about 5 percent, and indirect employment o f another 5 percent. Other simulations predict a 25 percent decline in RMG export (excluding knitwear) leading to a 6 percent decrease in wage payments to unskilled 34See https:llwww.gtap.agecon.purdue.edu/resourcesldefault.asp 28 female labor in non-agriculture (Amdt & al, 2002). Regardless o f which estimate, the losses will hit the poor working inthis sector the hardest. A large part o f the job creation in the urban labor market is linked to the RMG sector. Given the termination o f the MFA, Bangladesh will have to invest inmakingthis sector sufficiently competitive to keep its market shares. This necessitates large investments to improve product quality, increase efficiency, modernize technology, and ensure that products are competitively priced (WTO, 2000). However, to meet these investment needs, it is unclear ifBangladeshi entrepreneurs will have the managerial and risk-taking capacity to handle such a level o f investment (Sobhan, 2002). This will require involvement o f financial institutions. With regard to the labor force, these investments are also likely to imply an improvement of average skills, which would necessitate the strengthening o f the education system andtraining. Box 2.2: Post MFA and entry of China in the WTO,their impact on thepoor in Dhaka? A summary of the literature The post-MFA situation is still evolving though it i s clear that these changes will likely have an adverse impact on Bangladesh in general, and particularly on Dhaka. A summary o f the literature points to the following strengths and weaknesses o f the Bangladeshi garment sector.35 Favorable characteristics: -Bangladesh has a c(l.wantic entrepreneurship, n cheap and skilled labor force. Bangladesh has a coiisiderable comparative advantage in the price o f its manpower. Its labor cost in spinning and weaving i s only 3 percent o f that o f the United States. Moreover, given different trade arrangements (but also a certain heterogeneity in the quality) the price of a shirt exported from Bangladesh to the EU i s 2.9 euros verstis 5.9 euros for export from China.36 (Annex 2, Table A2.31) -Bangladesh has improved the qisality of its products since the mid-eighties f Dowlah, 1999). -Bangladesh has denzonstruted a significant iinprovenzent in its "Revealed Comparative Advantage" (RCA) in the nineties in all garment products, except for non- 35 Also see End of MFA Quotas, Kej, Issues cinri Sttutegic Optionsf'br Bungkudesh Readymade Garnient Itidusti?, 2006, World Bank Report No. 34964. 36 The literature has taken different approaches in dealing with the questionof market shares. The most ffequent approachhas beenthe application of General Equilibrium Models, mainly through Global Trade Analysis Project (GTAP) either to simulatethe impact of the China`s entry into the WTO on developing countries (Gilbert and Wahl, 2000, )Morrison.2001 for a summary of CGE-basedmodels and references) or to simulatethe impact ofMFA quota removal (see Walkenhorst, 2003 for a survey of quantitative studies). 29 knit men's outwear (Shafaeddin, 2004)''. Much o f this i s attributed to a lack o f diversification and special agreements with the EU and the US. -Bangladesh has reduced its original dependence on importsfor its` intermediate inputs. Although local net export earnings retained within the country from the RMG sector were only 23 percent o f gross exports at the beginning of the 1990's, and reached 37 percent in2000. Less favorable characteristics: -The perforiirance of the Bangladeshi RMG sector has stronglv depended on favorable trade agreements. The Bangladeshi exports to the EUhave benefited from both GSP arrangements and the Lamy EBA agreement which permitted duty free access.38 Bangladesh has also benefited from large quota allotments with the U.S. -Current rules of origiir requirements penalize clothingproducers of Bangladesh. For a T-shirt to originate in Bangladesh under the EU's rules, i t must either have undergone two stages o f transformation there (from yam to fabrics, and fabrics to clothing), or have used fabrics from other South Asian countries and added more value in Bangladesh than in any other contributing country (Oxfani, 2004). In reality, the usual value added at the assembly stage performed in Bangladesh only reaches 25-35 percent, which i s low. Moreover, due to the weakness o f indigenous cotton crop and underdeveloped capacity in spinning and weaving, Bangladesh uses fabric inputs from other developing countries. -As e.xp'ort quotas are less restrictive for Bangladesh than for China atid Iridiu, Bangladesh willface conipurativelv greater competition from China and India under a quota p e e regime (Islam, 2001 and Lips & ul., 2003).j9 Bangladesh has free access to one o f its most iiliportant export markets (the EU), thus further trade liberalization will worsen the Bangladeshi positions with respect to its competitors. In importing countries, a phase-out o f export quotas will likely result in a drop in the prices o f apparel from India and China more than that from Bangladesh. -Bangladesh 's competitors also have low hourly wages. China, India, Pakistan, Sri Lanka and Indonesia also have low labor costs in the clothing industry (Table A2.3 1). -Bangladesh and Chinu have a veiy siniilai- export structure (Shafaeddin, 2004). This creates high competition, particularly inouter garments. (Annex 2, Table A2.3 1) -Bungladesh szgfers j?om the weakness of its backward linkuges. Bangladesh still imports the majority o f yam and textile necessary to the garment industry. This implies very long lead times (120-150 days, in comparison with 12 days in India). Moreover, after the end o f the MFA, prices o f yam and textiles may increase for Bangladesh if exporting countries redirect these products to their own garment industries. -Bangladesh gament sector suflei-s fioin internal problems, such as inadequate infrastructure, and unreliable energy supplies, which contribute to high costs. 37RCA is defined as the share o f a specific product intotal exports of a comtry relative to the share o f the same product inworld trade. A ratio exceedingunity indicatesthat the country has a comparative advantage inthat product. 38The Everything But Arms (EBA) Initiative eliminated quotas and tariffs on all imports into the EL from the 49 least developedcountries, with the exceptionof arms and munitions. EBA became effective in March2001. 39Quotas are also less restrictive for Pakistanand Nepal. 30 iii.DevelopmentofSelfEmploymentthroughMicro-credit Previous sections have shown that in Dhaka about 33 percent o f the labor force is self-employed. Most o f them are men. (40 percent are men versus 16 percent women). This would indicate substantial demand for microcredit. There i s also supporting literature on the positive impacts o f micro-finance on poverty reduction. (Khandker, 2003) It is estimated that some 19.3 million individuals received help from various micro-finance programs in 2004.40 Much o f the micro lending i s done through four NGOs, Granieen Bank, BRAC, ASA and Proshika, which accounted for approximately 86 percent o f micro-finance lending. Most micro-credit recipients are women (about 90 percent o f borrowers), and live in rural areas (about 90 percent). It i s only recently that the major micro-finance providers have begun to target the urban poor (e.g., BRAC, Proshika, ASA), and the coverage i s still low. Grameen Bank continues its focus in rural areas. Bangladesh has a relatively positive experience with microfinance in rural areas and i s known globally for its successes. This has not, however, reached the same coverage in Dhaka and other urban areas. In the context of rapid urbanization, there appears to be substantial scope for reaching the urban poor including youth who may have difficulties entering the labor market. V. Improvingincome earningopportunitiesfor the poor While rural migrants continue to come to Dhaka in search o f employment, finding remunerative employment i s a major challenge for the poor. Jobs tend to be low paying, and do not provide much security. To cope, additional household members, particularly women and children, enter the labor market to earn what they can. The prevalence o f child labor i s found to be particularly highamongst the poorest households. Recent trends have shown that the growth o f Dhaka's labor force is far from tapering o f f given the high rate o f rural-urban migration. Simultaneously, the development o f one o f the main fornial employment providers o f Dhaka - the garment sector - i s jeopardized by the MFA removal. In this context, the balance o f Dhaka's labor market strongly depends on policy measures taken on the national scale. Among these are: Iinplenientirzg policies to encourage growth though diversifiiiig riianufacturing beyond the RMG sector. Some of the areas that hold potential for strengthening Dhaka's labor market include food processing, assembly industries such as electronic goods, toys, construction, etc. and in the services sector, the development o f data processing and telecommunicationboth for domestic and export markets. 40 TheEconomics and Governance of NGOs in Bangladesh, 2005, World Bank. 31 Investing in iizfrastructure. Firms in Dhaka and Chittagong interviewed in a recent investment climate survey rank electricity supply as a major battlement to doing business and to growth.4' Poor water supply and traffic congestion are also major constraints to growth in Dhaka. In addition, access to land was also mentioned as a severe expansion bottleneck for 40 percent o f firms interviewed in Dhaka (see Chapter 3). There are signs of some firnis moving outside o f Dhaka which has the potential for addressing the land constraint, though such a transition on a larger scale will be dependent on adequate infrastructure and the availability o f skilled labor. Improving skills of workers. There i s substantial international evidence that investments in education and training can raise productivity and increase the income earning potential of the poor. Education levels for the urban poor remain low relative to other countries, and thus providing affordable options could have hugebenefits. Increasing access to credit. On a smaller scale, the expansion o f the access to credit by poor urban households should also be an important policy focus. Many NGOs and Micro Finance Institutions in Bangladesh have extensive experience in rural areas which could be implemented in Dhaka and other urban areas. This could play a significant role in financing employment creation by new micro-enterprises and in helping the urbanpoor to develop new income generating opportunities. Expanding access tu child care progranis. There is also potential scope for facilitating increases in female labor force participation through access to child care. Evidence from other countries has shown that access to child care through NGO or Government facilities can allow wonien the flexibility to enter the labor market, ultimately increasing household income. 4 1See Lall, 2005. The Survey was carried out by the Bangladesh Enterprise Institute and contains 1001 firms inDhaka and Chittagong. 32 CHAPTER3: SHELTERFORTHE POOR The city's poor live in rudimentary housing, and only one-in-twenty live in permanent housing. Rents are high, living space i s crowded, and the provision o f amenities, while difficult to measure with any precision, i s almost certainly low given the temporary nature o f most housing. Conditions that are already bad appear to be getting worse as rents continue to rise and migration to the city continues. Dhaka's rapid growth, large size, topography, environmental conditions, and problems o f governance exacerbate already complex land and housing issues for the poor. The number o f slum dwellers i s projected to grow to 8 million over the next decade. With this growth, policy makers will need to address land and housing as a top priority. This chapter covers the main land and housing policy issues in Dhaka, with a focus on the effects these policies have on the poor. Section Icovers the main policy issues o f land and housing prices, the national housing policy, and institutional roles. Section I1 presents main policy constraints and Section I11 outlines recommendations aimed at improving the functioning of these markets. A major constraint in the analysis i s the scarcity o f recent data on land ownership, rents, prices, and the supply o f various basic amenities. This lack o f information i s due largely to the informal nature o f urban developments in Dhaka. It i s also consistent with the almost complete absence o f government in the operation o f housing and land markets in the city, even in enforcing basic property rights. Given the paucity o f recent data the paper has relied on empirical reference points that are somewhat dated, complemented with assessments and observations by Bangladeshi experts and market participants. Updating the information base would be very useful indeepeningthe analysis o f these issues.42 ' I.Background A. A Fast Growing, Poor Megacity Dhaka i s known as one of the fastest growing cities of the world. Its growth has beenparticularly rapid since 1972, after its transformation from a provincial capital to the national capital o f the newly independent country, Bangladesh. The average annual growth rate o f the city's population was 6 percent during the 1974-2001 period. Rural to urban migration remained the most dominant factor o f the population growth, but natural increase was also high. Dhaka (or Greater Dhaka) emerged definitively as a mega-city in the new millennium.It is currently the 1lthmost populous city inthe world with over 12 million people and i s projected to move up to the 6th position with 18.4 million people in 2010 42For example, financing surveys of living conditions ofthe urbanpoor, such as the one undertakenin 1995 for the Study ofUrbanPoverty inBangladeshby the Centrefor Urban Studies(Islam, 1997) would be a very highreturn investment.Collecting data on public housingprograms runby the various public agencies and consolidating them would also be worthwhile. 33 (UN,WorldUrbanizationProspects, 1999), with only Tokyo, Mumbai, Lagos, Sao Paolo and Mexico City being larger. The projected population size for 2010 and 2016 respectively, has been estimated at 9 million and 10 million for the Dhaka City Corporation (DCC) area, and 14.88 million and 18.00 million for the Dhaka Metropolitan Development Plan(DMDP) area (Table 3.1). UrbanPopulation UrbanPoor Source: Islam(2004) Can Dhaka sustain such a growth rate in thefuture? Dhaka i s the focus o f most o f the country's politica!, administration, economic and social attention. Dhaka's primacy i s still on the rise, both in terms o f population and in functional terms. It i s the most centrally located and most accessible city o f the country. With a large national population (140 million in 2004) still growing rapidly and still predominantly rural, the population pressure on the primate city remains very highand appears likely to continue. The city also concentrates private investments in key sectors. For example, 80 percent o fthe garment industry, which accounts for 80 percent o f all exports and employs about 2 million people inthe country, is located in Dhaka. Experts also mention the fact that public spending and investment (inparticular, inhealth and education facilities), and public policydecisions have traditionally beenskewed infavor of Dhaka over other large or medium-sized cities since the independence o f Bangladesh. The concentration o f , investment and decision-making power in Dhaka, in turns, appears to fuel the on-going high rates of migration to the City. But, little empirical information or analysis is available on why this trend does not follow a self-correcting pattern in which the high prices and congestion in Dhaka make other cities more attractive investment and hence job locations. The continued growth o f Dhaka can eventually lead to negative implications for economic growth and poverty reduction.43 Empirical evidence shows that very large 43 Many local experts believe that the city's poverty problems cannot be addressed unless its growth rate i s reduced and would favor programs attempting to subsidize recent migrants to leave the city which have beenperiodically implemented. While the city's growth rate is unusually high, the international experience 34 population concentrations may have adverse effects on growth (Henderson, 2000), 44 and are generally driven by a set o f broader policies that lead to corruption and less equitable growth (Glaeser, 1999).45 B. High Land Prices HaveExcluded thePoorfrom Ownership of Land and Housing Housing Prices. Residential land values in prime locations o f Dhaka range between US $ 30 and $ 60 per square foot, similar to prices reported in other regional cities such as Hyderabad, Kuala Lumpur or New Delhi.46Nevertheless, land prices are high compared to those found in developed countries. For example, areas in the US where land prices exceed $60 per square foot are rare (Housing finance in Bangladesh, World Bank, 2004). These prices make it impossible for the poor to purchase land in the open market within the DCC area. The cheapest ready-to-build plot within DCC is priced at Tk. 500,000 per katha (i.e., 720 square foot lot) or US $ 12 per square foot. Normally RAJUK does not grant buildingpermission on lots smaller than 1050 square foot. This would cost Tk. 729,000 (US $ 12,600), which is equivalent to nearly 20 years o f income for an average poor household (Tk. 3000 per month). The cost o f housing would be additional. In any case, such small lots are hardly available inthe open market, and only rarely supplied by the government in a subsidized market. Consequently, it i s estimated that 97 percent o f the urbanpoor inthe city do not own any land.47 New Housing. With regard to the types of new housing produced, formal sector private developers generally serve only the upper and middle income groups. Even their "lower cost" housing products sell for approximately 1 million Takas (US $ 17,000). With a 50 percent down-payment and under current credit conditions, the unit can be repaid with monthly payments o f Tk. 5,000. Under an already high payment-to-income ceiling of one-third, the required minimum income o f the household would be of Tk. 15,000, which means that only those inthe top 30 percent o fthe distributiono f income in Dhaka could afford to purchase new housing. Thus, formal, industrially developed units are unaffordablenot only to the poor, but to the vast majority o fthe population. i s that subsidy programs designed to induce out-migrations from cities are rarely effective. More fundamental incentives are at work. 44 The literatureon the relation betweeneconomic growth andurbanconcentration suggest that national resources canbe squandered due to excessive concentration ofpopulation andinvestments inone city (Henderson 2000). Excessiveprimacy often leads to a decreaseo f competitionamong cities to attract investments, increasing commuting, congestion, andlivingandproductioncoststo excessive levels. It also leads to increases incosts o f land, housing andurban services, and decreases inproductivity o fbusinesses and quality o furban services. Thus, economic growth is impacted negatively. 45 See Lall, 2005 for a discussion o f this and simulations for the Bangladesh context. 46 2003 by Seraj and Akin. 47 Ina study for the Government o fBangladeshandADB in 1995,Islamet al. found that only 3.2 percent of the urban poor inDhaka owned the plot o f land on which their dwelling unit was located. 35 Existing Housing. It is estimated that less than 20 percent o f the poor o f Dhaka are owners. The last statistical survey, conducted in 1995 (Islam et. al, 1997) found that 16 percent o f the poor in Dhaka were owners; 56 percent were tenants; 8 percent were rent free dwellers, and nearly 20 percent were squatters or illegal occupants (Table 3.2). The low ownership rate inDhaka stands incontrast with comparable Indian cities such as Delhi or Bombay,48 and with cities o f other regions such as the Middle East and North Africa or Latin America where ownership of the dwelling is by far the dominant status even for the poor. * Table 3.2: HousingTenure o fthe UrbanPoor, 1995 Dhaka All Urban Hardcore Moderately All Hardcore Moderately All poor (%) poor (%) poor poor (%) poor (%) poor Owner 16.4 16.2 16.3 28.8 25.5 27.7 Tenant in 42.4 49.6 45.4 45.6 48.4 46.6 private house Government II 5.6 1 5.2 5.0 I 7.6 I 5.9 Source: Islam, et. al, 1997 Housing Quality. ,Access to housing for the poor in DCC is mainly limited to private slums and squatter ~ettlements.~~ Only 5 percent o f the poor of Dhaka inthe 1995 study lived in permanent or pucca houses, 22 percent lived in semi pucca houses, 41 percent in temporary units and another 32 percent in rudimentary kutcha units or jhupris (Islam et. al, 1997). Slums: Estimates and Definitions. Slums (Bastees), are substandard housing built on privately owned land. Squatter settlements (or Bastuhara colonies) are 'substandard housing built illegally on publicly owned land, including railway tracks. A third category comprises the sites o f the homeless / pavement dwellers (Bhashaman Janogosthi), estimated at around 120,000. The recent mapping o f slum settlements (Chapter l), that slums are located all around the City. The share of private slums shows i s thought to have been increasing over time because o f the rapid increase of the city. limits, and because evictions from government land have limited the growth o f squatter settlements.+ 48 Although no directly comparable statistics are available, the ownership rates among households living in one -room dwellings, which i s a proxy for the poorest groups o f the population, were 47 percent and 59 percent inDelhi and Bombay respectively, and higher for larger dwellings (http:l/www.indiastat.comi). 49 A large number o fthe poor (such as household workers, drivers, security guards etc.) live as part o fthe middle class or upper class households and as such benefit from almost all utilities and urban services. 36 Incomparison to other cities inthe region (India), the proportion o f slumdwellers (in the broader sense) in Dhaka is similar to Kolkata, less than in Mumbai, and greater thaninthree other major Indiancities, Delhi, Chennai, and Bangalore(Figure 3.1). Figure3.1:Estimatedproportion ofpopulation livinginslums inIndian cities andDhaka Percentage of city population living in slums Bangalore Delhi Chennai Kolkata Dhaka Greater Mumbai I 0 10 20 30 40 . 50 60 Source: For Indiancities, Indiastat. For Dhaka :Islam(2004). HousingRent Levels. Inprivate slums and `mess' units the poor pay regular rent to a landlord or home owner. Inthe case o f squatter settlements, occupants generally have to pay "tolls" to mastaans (musclemen) and agents o f employees o f the land owning authorities. Rents in the private slums are reported to be high. Islam (1985) found that slums dwellers in Dhaka City were paying higher rent per square meter than non-slum households, even though the latter usually benefited from a much better physical environment and level of services. Typical rents today would be around Tk. 500 for a single person, Tk. 1000 or Tk. 1,500 for a family. Many poor households pay up to Tk. 1500 (US $ 26) for a small (less than 10 square meters) two-room unit in slums with water, electricity, latrine and gas. This spending accounts for approximately 14percent o f the poor's household budget (HIES, 2001). This figure i s considerably higher than that observed in large Indian cities where the average urban household spends 7 percent o f their income on housing, even in high cost locations such as the state o f Maharasthra where Mumbai is located. As a consequence of tight budget constraints and relatively high rents, the poor inDhaka usually live in very small accommodations (2 or 3 square meters per person similar to the situation inMumbai). 11.Main policyconstraints A. Unresponsivenesstoprice signalsfor a largeportion of the land in the city The very high level o f land prices is the major constraint o f urban policies in Dhaka. Land prices appear to be comparable to those in suburban New York, whereas the average income o f Dhaka residents i s a hundred times lower. For example, the Bangladesh Institute o f Planners estimates that land in Dhanmondi, the most expensive 37 area in Dhaka, i s Tk. 3,50O/square foot or roughly $60. Similarly, land in Gulshan i s priced at Tk. 3,194hquare foot and Baridhara at Tk. 3,00O/square foot. Even more remarkable i s that land in the least expensive residential areas in Dhaka, is valued at Tk. 1,600, around $27 per square foot. These areas include Uttara, Pallabi, and Shyamali, which are on the periphery o f the city. Incomparison, landpricesinmostcitiesoftheUnitedStatesaremuchlower.50In a study o f land and housing costs in US cities, Glaeser and Gyorko (2003) report the following prices (in $/square foot) for a sample o f cities: Boston $13.10; Chicago $14.5; Los Angeles $30.4; Miami $11.1; New York $32.3; San Diego $26; San Francisco $64. These prices reflect a mix o f land shortages and development constraints (landuse zoning regulation), but only in a few cases i s developed land priced higher than in pai-ts of Dhaka. Landprices in Khulna have been estimated to be similar to those inDhaka, and are about 15 percent lower in Chittagong (World Bank, 2004). In any case across the metropolitan areas, land values are extraordinarily high. In fact, the least expensive middleclass housing inthe country is estimated at about Tk. 700,000, which is more than 10 times the median income (World Bank, 2004). The main reason for such high land prices is topographical, resulting in land supply constraints. Because o f the city's location there i s a real scarcity of developable land, which is reflected in the very high density o f Dhaka (20,000 persons per square kilometer, and up to 100,000 in some areas o fthe city). Dhaka is situated in a flood plain and surrounded by rivers which periodically flood the lowlands aroundthe city. The local topographical configuration led to a city development following a north-south pattern, rather than as a circle, as shown in Map 2. Most o f the high lands immune to flood are already built. Newly developed areas are built on the flood plain by progressivL filling. This real scarcity i s compounded by an artificial scarcity, stemming primarily from the lack ofutilization o fpublic landwithin the city. Large amounts ofpublic landin central Dhaka remain undeveloped. Government-owned land (called mas) i s now scarce (at most 1500 ha, but only 430 would be free of occupation today). Other locations are occupied by Government-related activities with very little or no use for construction (this i s the case of the Tejgaon airport, the military cantonment, and land tracks left for future construction o f public buildings). This may characterize as much as 20 percent o f the landin the inner city area as shown inMap 2. Other areas, also centrally located, such as the public housingstock are developed at very low densities giventhe landprices. Finally, there i s a gap between technically usable land and formally usable land. The latter i s defined by MJUK master plans. The current approach i s obsolete, resulting inareas potentially developable for residentialuse beingnot zoned as such. A consequence o f this scarcity o f formal serviced land has been the development of the city almost entirely by the informal sector over the last 20 years. As shown in Maps 1 and 2, most o f the new urbanization in the last 20 years has occurred at the 50 Also consider that per capita GDP inthe UnitedStates on a PPP basis i s 20 times that of Bangladesh. 38 margin o f the city, with little or no attention paid to planningregulations or any planning enforcement.51 This dominance o f the informal delivery systems over the formal system is also highlighted by a breakdown o f the housing stock into sub-categories. The informal sector is estimated to have produced 85 percent o f the 1.0 million housing units in the D C C D M A area (Table 3.3). While self-construction accounts for halfo f the total stock, slums andsquatter settlements represent the other 35 percent o f it.By contrast, the public sector's contribution to the housing system inDhaka has been around 100,000, while that o f the formal private sector (real estate companies or developers) has been estimated to account for around 48,000. The contribution of the cooperative sector i s very small, probably less than 2000 units. Thus, the total formal sector contribution to housing in Dhaka would be only about 150,000 units, or about 15 percent o f all units in the central city or D C C D M A area. Table 3.3: Relative importance o f Housing Delivery Sub-sectors inDhaka Sub-sector Number ofUnits inDCC/DMA Formal Sector Public housing 100,000 Private Housing 48,000 Cooperative 2,000 Total Formal Sector 150,000 Informal Sector Private Housing 500,000 Slum 200,000 Squatter Settlements 150,000 Total Informal Sector 850,000 Total 1,000,000 Source: Estimatesfrom Islam, 2004 B. An enabling housingpolicy without enablers i. The national Housing Policy In 1993, the Government of Bangladesh approved a National Housing Policy which was updated in 1999 and once again in 2004. The Policy describes the role o f the government in the housing sector as that o f a "facilitator" or "enabler" rather than as a "provider." This means its role i s to increase access to land, infrastructure services and credit; to ensure availability o f buildingmaterials at a reasonable price especially for the low and middle income groups; and to create and promote housing finance institutions. Actual construction o f housing i s to be left to private sector developers, private individuals, and NGOs (GOB, 2004). 51 Other problem relate to land titling, the registration of real estateproperty, and the coexistenceof different forms of land ownership. As these issues havebeen developedelsewhere (Syed Mynuddin Hussain,2004), this paperdoes not elaborate on them. 39 .-. Figure.3.1: Map_ _ofLandUseinDhakaCity, 1984 ..- . Source : Salma Shafi (1992). 40 Figure3.2: Mapof LandUse inDhaka City, 2004 ___ Source : Salma Shafi (2004). 41 The policy calls for special consideration for improvement o f low income settlements and slums and squatters. It has clear recommendations in support o f ensuring proper rehabilitation in cases o f slum eviction. The policy also makes statements on strategies for improving land supply, infrastructure, building materials on technology, housing finance institutions, legal and regulatory frameworks and institutional arrangements. Finally, it also recommends the formation o f a National Housing Council and a National HousingAuthority (NHA) inorder to implementthe policy objectives and programs o f the government. The latter has been formed but not the Council. The stated policy approach appears to be sound and i s consistent with the recommendations made in the World Bank's Housing Policy Paper: Enabling Markets to Work (1993). Nevertheless, important aspects o f it are not implemented. ii. Continuing Slum Evictions without Resettlement Eviction o f squatters from public land has been a continuing practice o f the government policy. Annex 3 has a list o f all evictions that have taken place in Dhaka since 1975, estimated at around 135. While the number o f slum dwellers evicted in each case i s not known, the list clearly shows that evictions have not slowed down since the adoption o f the National Housing Policy. The 2004 large scale eviction in Agargaon affected an estimated 40,000 slum dwellers. Evictions from squatted public land often occur without resettlement plans. As a result, evicted households, some o f whom have lived for more than 20 years in the same settlements, have little choice but to join the cohorts o f new migrants to the city on the urbanfringe. This movement inturn fuels the process o f landfilling and land grabbing. The systematic filling of the water retention ponds within the city area makes it more prone to flooding. Squatted areas rarely upgrade, because the threat o f eviction impedes investment inhouses.52 Thus, inthe longrun, the absence o f a relocation policy results in an implicit status quo development strategy for the city as a whole, which presents high environmental risks. iii. Fragmented and mis-targeted public interventions in housing The roles of the mainpublic actors intervening inthe housing sector are not clear. Those involved include the planning agency RAJUK, the National Housing Authority (NHA), the Urban Development Directorate, and the Public Works Department of the Ministry o f Housing.53 The NHA has jurisdiction over all Bangladesh, issued the National Housing Policy, and i s supposed to be responsible for its implementation. It does not, however, have specific means for enforcing the policy. The NHA also manages 52W h i l e the squatter settlements are obvious examples o f the effect o f uncertainty of property right on investment in housing, the same phenomenon i s also mentioned for public-private partnerships in housing, where the Government wants to recover the land after a certain number o f years and entertains uncertainty about the exact nature o f rights on the land, deterring investment by the private sector. 53Other players include the Ministry o f Land, which owns public land all over the country, and the Prime Minister's office, which directly manage some habitat projects all over Bangladesh. 42 housing projects. By contrast, RAJuK's jurisdiction covers only the Dhaka Metropolitan Area, and RAJUK's mission has never been to provide housing for the poor (Annex 3, Box A3.2). InDhaka, the division between the interventions ofall these actors appearsto be based largely on a geographical basis, determined by the holdings o f land acquired by them at different points in time. Precise data on the extent and composition o f public housing programs are lacking though public housing rograms serve only a very small fraction (less than 10percent) o fnew housing demand. p4 Few if any o f the public housing programs are targeted to the poor. RAJUK has been known to provide serviced lots o f a minimal size (100 square meters -1.5 khatas), which given land prices in the city makes them totally unaffordable for poor hou~eholds.~~The Land Ministry i s currently working on a project "for the poor and landless," consisting o f 15,000 small flats inhigh-rise buildings. Considering the costs of finished flat units, there is little chance that the officially-targeted households will be able to pay for them. Thus, units built through public programs are likely to be bought by (lower) middle income households, unless very generous subsidies - amounting to several times the annual income of beneficiaries - are delivered to a small share o f the population that can be served. Such programs operate more like lottery schemes than housingassistance. In conclusion, little has been achieved in terms of the implementation o f the National Housing Policy. Property rights have not been assured. Public interventions have been fragmented, have not targeted the poor, are based on unworkable design standards, and are not coordinated. The current policy environment is not one in which additional public resources will necessarily benefit the poor. Improving the housing conditions o fthe urbanpoor will require substantial policy reforms. 1Box 3.I:Slum Improvernent/Upgrading Programs of the National Government The Government o f Bangladesh beganimplementingupgrading program inurban slums inthe mid eighties with UNICEF. This programhas been active in some 25 cities, including Dhaka. Originally the program had a strong physical component like improvement o f drains and sanitation. It had no housing component. Later the World Bank, the ADB and major NGOs like CARE supported some slum improvement programs. None in Dhaka (DMDP area) however, has any housing component. All programs together have made very little impact on improvement o f slums in Dhaka, due to the massive scale o f the problem. j4According to experts, the only occurrenceof landbeing given to the poor on a large scale was in 1975, when 250,000 squatters were evicted frompublic land. 75,000 were relocatedinTongi, Mirpur,and Dembra.Those areas are now hrtherdeveloped and part of the modemcity. 55For example, plots servicedby RAJUKinthe new town of Purbachal,presentlybeing developedbeyond the Baluriver inthe east ofDhaka,will be of 3,000 square feet on average. 43 Dhaka City Corporation itself established its own Slum Improvement Department in the early 9Os, but fund allocations to this department has always been very small, although over 1/3'd o f DCC's population lives in slums. The NHA i s the new form o f the former Housing and Settlements Directorate. The latter agency managed some housing programs after the partition, mainly for displaced people coming from India. The NHA, established in 2001, has been implementing some small projects for the low income families and i s now trying to help a private company to implement a resettlement project inmulti storied buildings inMirpur, Dhaka. C. Failures of coordination and services at the local level At the local level, the lack o f coordination between all actors intervening in the housing sector is mentioned by Bangladeshi experts and international reviews as the source o f multiple inefficien~ies.~~Concrete manifestations o f the lack o f coordination are the absence o f planning enforcement, lack o f services and infrastructure, and inadequate timing in land and housing project developments which causes disruption o f services (Sarker, 2004). Among the causes of this lack o f coordination i s the limited role that the Dhaka City Corporation (DCC) plays in urban management. DCC does not manage planning, does not decide on the use o f public land located within the city, and has no power over the public utility companies within its jurisdiction. While a precise census would be necessary, figures quoted range from 16 to 40 different institutions intervening at one stage or another in urban matters inDhaka. Purely land related issues such as landrecords and land transfers are dealt with by the Ministry o f Land and Deputy Commissioner o f Settlements under the Ministry o f Works. Housing related issues are managed by the National Housing Authority and RAJUK. The Dhaka City Corporation (or other Municipal Authorities within DMDP area) is responsible for housing or property tax collection. The Department o f Environment has the responsibility o f controlling environmental regulation as these affect land and housing development. The utility service authorities like Dhaka Water and Sewerage Authority (DWASA), Dhaka Electricity Supply Authority (DESA) and Titas Gas Company are each responsible for providing their respective services to the households requiring these. The Dhaka City Corporation and the other Municipalities provide conservancy services. There are several other agencies involved in housing development. The performance o f the Dhaka Development Agency, RAJUK, officially in charge o f planning, planning enforcement, delivery o f building permits, and land 44 servicing, is also a major cause o f inefficiency. A review o f the role o f MJUKwould be a top priority for policy reform. Some progress seems to have been achieved recently in the coordination process, however. Two years ago, a new Committee for UrbanDevelopment was created, to solve the main problems o f coordination among urban actors in Dhaka.s7The Committee is chaired by the Principal Secretary to the Prime Minister. The role o f the Committee, however, has been in solving day-to-day problems, not in setting a medium or long-term vision for the development o f Dhaka. A major consequence o f the lack o f governance and accountability concerns access to utilities. In response to the unavailability o f services such as electricity by formal means in the illegally built zones, parallel systems for accessing utilities develop." As discussed in Chapter 4, maastans often arrange delivery o f services to the illegal area with the utility companies. Utilities are then sold back to resident renters at high prices. The resources generated are shared between the middlemen, bribed utility company employees, and government officials. As a result, not only can the utility companies not cover their costs, but the poor end up paying higher prices for services thanthe non-poor. D. Thelimited role of NGOs in Jtousitigfor urbanpoor NGOs and microfinance institutions have been impressively present in rural Bangladesh for at least two decades, and have achieved international recognition as effective and transparent pro-poor institutions. Some of them, such as Grameen Bank, have devoted significant resources to housing programs. Grameen has made more than 600,000 housing loans in rural areas in Bangladesh. .However, NGOs have not been intensively involved in housing programs in Dhaka City. Some NGOs are confined by their charters to work in rural areas (G~-ameen).'~ Others work in urban areas, but have prioritized income generating activities as a main focus.6o The very high land prices in Dhaka are cited as a main reason for this limited involvement. They preclude replication o f rural interventions in housing o f the Grameen type, in which no land cost i s covered and housing loans are given to households having already a piece o f land, or to homeowners for home improvements.61 57For example, it is this group which enforced local participation into the drafting o f Detailed Areas Maps prepared by RAJUK. 58Unitsbuiltwithout title or buildingpermit cannot intheory have access to utilities. 59The present charter o f Grameen forbids its direct involvement inurban areas. It helps other NGOs (such as Shoptika or PKSF) to operate inDhaka. Grameen is presently seeking to have its charter revised to be able to operate inurban areas. 60For example, BRAC activities inurban areas include income generating activities micro finance; health care (tuberculosis control program inpartnership with the Government); Micro enterprise lending programs; a pilot program for rickshaw pullers sponsored by the World Bank; and a pilot program for retrenched garment workers. 6'The order o f magnitude o f the loan would be Tk. 10,000 (US $ 170), paid back in 10 years. This corresponds to monthly payments o f around Tk.240 (US $4), that is, 20 times less than the loan for a 45 NGOs also find working in Dhaka high risk. The threat o f eviction, legal complications related to land ownership, the ability o f middlemen to extract rent from the poor, and the lack o f credibility o f the Government's commitment, render investments in those areas intrinsically too risky. While NGO involvement in housing programs has been limited, some interesting projects do exist that could potentially be scaled up. A promising approach to providing shelter solutions to the poor is the type o f projects run by Nan Uddyog Kendra (NUK), which has arranged rental accommodation in dormitories for female garment workers. NUK offers a comprehensive and cost-effective hostel model. Recently, BRAC has also engaged in similar types o f projects, with the support o f the government. Those projects offer affordable solutions for the poor, particularly single women.62 Box 3.2: NGOprograms targeted topoor renters , Currently NUK i s operating three Hostel and Development Centre for women garment workers inDhaka city with 600 seat-capacity, and two Student Hostels with 150 seat capacity in Dhaka and Mymensingh. All the hostels are operating on a non-profit basis inhired premises. Recently, BRAC also engaged in this activity and "Grihayan Tahabeel" (Housing Fund) o f the Government o f Bangladesh, administered through Bangladesh Bank, approved Tk 100 million to BRAC for building hostels for 2,500 women workers o f garment factories inUttara, Dhaka with 1percent interest, repayable in20 years. BRAC has another program centered on housing, targeted at single women renters. BRAC staff act as middlemen between landlords and single women who otherwise would not be able to find a place to live. They lend Tk. 5000 to cover the upfront costs o f rental (3-4 months o f rent as a security deposit). The woman will then be able to mobilize her own social network and sublet rooms or space to other women, which will provide a source o f income to repay the loan. 111. ImprovingShelter Conditionsfor the Urban Poor Dhaka's rapid growth i s likely to continue and thus understandingthe factors and policies that contribute to this growth will help to shape the future. Main challenges facing Dhaka in the next years with regard to land and housing for the poor are: coping with a sustained arrival o f poor migrants, adopting more proactive planning policies to formal unit describedabove. This gives a sense of the gap betweenthe reach of formal, developer-built housingsolutions, and low-income housing. According to a number of Bangladeshiobservers, single woman inBangladeshi s likely to suffer discrimination on anumber of fronts. Inparticular, they have a hardtime finding an accommodation, are subjectto all kind of abuse and as such constitute a privileged target of the NGOs. 46 address the environmental risks, finding a way to enforce property rights so that land grabbing does not take place, ensuring that any eviction from public or private property is done in a way that recognizes the rights o f those affected; and improving the accountability o fthe public sector as a whole inthe delivery o f basic services. N o program o f reforms seems likely to gain momentum until the Government's commitment to honor basic property rights with respect to land and housing i s credible. This credibility is difficult to achieve in a market where land and housing prices are unaffordable for most o f the population. In the current context, the uncertainty o f property rights, the perceived corruption and inefficiencies o f public institutions in the sector, and the weakness o f the local government constitute significant impediments to the hnctioning o f the housing market. They also impede further involvement o f NGOs and the private sector in the delivery o f urban services for the poor. Inparticular, unless actions are taken to mend the institutional setting described inthe above sections, projects of city extension to the east (Dhaka Eastern Bypass) can only result in the same deficiencies as those which are witnessed as an on-going process inthe city. Recognizing that a number o f basic policy changes are necessary, main reforms are needed in the policy and institutional framework. Some o f the proposed changes will take a considerable amount o f time to be implemented, but would result in a more effective and enabling policy environment, ultimately contributing to improvements in shelter conditions for the urbanpoor. Among these are: Implementing theNational Housing Policy. This is a priority for any sustainable improvements for Dhaka's poor. A key element o f this policy i s in the enforcement o f basic property rights. Destruction o f squatter settlements should be limited to necessary actions o f redevelopment or infrastructure building and should be accompanied by relocation plans for evicted squatters. Strengthening public institutions to implement the National Housing Policy. Much stronger and more credible public institutions are needed for this enabling strategy to work. The multiplicity o f public agencies intervening in housing programs is inefficient and could be reduced with each agency having clear roles and responsibilities, operating under the umbrella o f the National Housing Policy. If the National Housing Authority continues to be the overseer o f the National Housing Strategy, it would be more effective if they focus on a policy and regulatory role rather than implementing housing projects. Developing mechanisms for better accountability for land use and better coordination of services in the city. A first step would be undertaking an inventory o f public land belonging to different ministries and agencies within the city. This inventory could be the basis for a strategic plan for public land use in the future. Vacant public land in excess o f critical city needs could be released on the market. Solutions for promoting housing upgrading and investment on squatted tracks o f public land could include giving proper titles to long-term residents with programs targeted to low income households. 47 Promoting coordination between DCC, RAJUK- the planning authority, line ministries and utility agencies in urban projects in Dhaka. While ad hoc committees can be a viable short or medium-term solution to coordination problems, inthe long run, the Dhaka City Corporation should be given more means to achieve pro-poor goals. This long-term goal should be explicitly linked to any progress inthe decentralization process inthe country. Strengthening theperformance of RAJUK. A full reform o f RAJUK is needed. This could be based on the separation o f the multiple functions it now exerts (planning, planning enforcement, land servicing, and housing projects), which generate conflicts o f interest and are not implemented efficiently at present. Planning functions should ultimately depend on the local governments concerned, and could be assumed by a body gathering Dhaka City and the municipalities within the current RAJUK jurisdiction (5 municipalities). Intra-city planning in Dhaka should be reinforced and DCC given increased planning authority within its boundaries. Planning enforcement should be given to another body to avoid conflicts o f interest. Within DCC it should be done by the technical services o f the city. By the same token it shouldbe given adequate means and manpower. The private sector should be allowed to compete on level ground for land servicing though there i s room for a public land services, and the right o f eminent domain for public utility projects. The rights of owners, however, have to be better protected (revise the current procedure for power o f eminent domain and make it fair and transparent). Construction o f housing units for medium or high income households should be left to the private sector. Low-income housing projects could be done by a specific entity, but RAJUK i s not the best candidate since its mandate was never focused on the poor. Initiating pilot projects in poor areas in partnership with NGOs. NGOs have not been active in the housing sector due to perceived risks. However, NGO programs such as N U K ' s projects o f dormitories for female garment workers, BRAC's program for poor renters have worked well. It would be useful to foster the piloting o f additional programs aimed at improving land and housing conditions for the poor with careful monitoring and evaluation so they can be scaled up over time. Other potential partners include Dhaka City Corporation and utility companies. DCC officials seem eager to do something for the poor; however, the DCC owns very little land in the city. A pilot project could focus on the upgrading o f the semi-slums in the 4th class city workers quarters around the city. Dhaka Water Supply and Sewerage Authority (DWASA) has 4,000 acres o f land o f its own surrounding water plants, in which WASA employees could be housed. However, most o f the land has been squatted 48 over time. More broadly, there are 500,000 public employees in Dhaka city, most o f which live inslums, which could be the target o fpilot projects runbyDCC. Improving legal and regulatory framework From existing material and discussion with experts, there i s clear scope for efficiency gains in the land and housing markets, due to constraints caused by the existing laws and regulations applying to zoning, land subdivision, and building. In a first stage, it is worth examining possible "quick gains" which could arise from the removal o f well-identified constraints which increase the cost o f formality and reduce the fluidity o f the housing market. Those include: Suppressing o f minimumlot size for land subdivisions; 0 Allowing for the possibility o f bringing utilities to housing units built without title / buildingpermit; Lowering registration fees for land; Lowering transfer fees for landandhousing; 0 Reviewingo f the existing regulations, with the aim o f removing the main obstacles to the smooth fimctioning o f the market. 49 CHAPTER4: THECHALLENGESOF SERVICEDELIVERYFORDHAKA'S POOR The provision o f services inDhaka is an enormous challenge particularly in light o f its rapid growth. There are severe deficiencies in both access and quality, particularly for the poor. Delivery systems are complex, inefficient and non-transparent with many households payingbribes to ensure services.63 The poor are particularly vulnerable as many service providers do not work in slum areas. Access to health, education, power, water supply, sanitation and waste disposal i s limited for the urbanpoor. For those services that do exist, the quality is low, and costs can be prohibitive. With Dhaka's rapid growth, there is an urgent need to find sustainable solutions to meet the basic needs o f the population. This chapter focuses on the current challenges o f service delivery for the poor in Dhaka, and the range o f constraints faced by service providers. Information is drawn from in-depth structured interviews with approximately 20 representatives from Government, NGOs and Donors working in the field o f service delivery to the urban poor.64 A full review o f service delivery for each sector i s beyond the scope o fthis study, though i s recommended for hture research. I.TheCurrentSituation Bangladesh's Constitution states that the Government o f Bangladesh i s responsible for creating an environment for its citizens that allows them to improve their quality o f life through access to health, education and other basic services. The country does not, however, have an explicit policy on urbanization and urban poverty. Without a policy framework, there i s no mandate or priority givento dealing with the problem. Furthermore, two common perspectives among policy makers affect decision making with regard to the urban poor. First, it i s assumed that providing access to basic services such as health care, education, sanitation and water supply to squatter settlements will result in increased migration. Second, i s the perception that the urban population is better off than the rural population. Consequently, urban poor and slums have received limited attention from the donors, NGOs and Government agencies until recently. The services that do exist are delivered by a mix o f Government, NGOs, donors, and private individuals (mastaans) who often use illegal methods. These services are limited, and generally perceived to be o f low quality. An extensive study on urban service delivery carried out in 2002 looked at user's satisfaction in four cities including Dhaka.65Overall, less than 20 percent o f the households surveyedare satisfied with eight 63World Bank,2002, Improving Governancefor Service Delivery. 64See Backgroundpaper for list of interviews. 65World Bank, Proshika, and Survey andResearchSystem, 2002, Bangladesh Urban Service Delivery, A Score Card. 50 out of eleven surveyed services. Among poor households in Dhaka, less than 5 percent were satisfied with 7 basic services, and only 16 and 17 percent were satisfied with education andhealthcare, respectively. A. Basic Services Education. The right to education inBangladesh is written into the Constitution, a right for all children to receive basic primary and lower secondary education (World Bank, et al, 2002). Providers include governmental, NGO, religious, community and private schools. The largest proportion o f children inpoor households who got to school inurbanareas attend government schools. Private schools are the next destination and a greater proportion o fpoor children go to private schools than NGO schools, reflecting the fact that not many NGO schools operate inurbanareas.66 Despite the multiple options, the urban poor are less likely to send their children to school. Ina sample of slum areas inDhaka, 58 percent o f 6-7 year olds do not attend school (Aparajeyo, 2002). The reasons given range from inadequate school facilities, distance to school, and lack o f funds for the tuition, lack o f parents' awareness, inappropriate school timing, and a floating/migratory lifestyle. Drop out rates for primaryschools are also estimated to be substantially higher inurban areas thanrural due to slum evictions inurban areas and a needto work. From the score card survey, many households experience a number o f problems, both getting their children admitted to school and the experience their children have while inschool. InDhaka, 59 percent reportedproblems with limited number o f seats, 16 percent with failed admission tests, and 10 percent with school authorities demanding `donations.' The reported payment o f fees and for those inthe lowest income group was 865 Taka in 2002, and 2327 for the second quintile which can be prohibitive for some households. 67 Health Care. About half o f all health services are providedby the public sector and halfby the private sector. InDhaka less than 54 percent o fhouseholds report having access to health facilities, for the poor this would be substantially lower. Furthermore, the clinics and hospitals that are accessible do not have the capacity to adequately care for the needs of patients. Only 12 percent of all urban poor report getting medical services from the government service centers, mostly because they are unable to take the time out o f their daily chores or time off work to visit a clinic (HIES). Hospitals and clinics are typically not located close to slum areas, and the waiting line for patients i s extremely long. Focus groups carried out as part o f the score card survey show that slum dwellers feel that they are deprived of proper health facilities, and that they do not get good care from doctors and nurses or the necessary medicines. They report long delays saying that it could sometimes take a month for a patient to get to see a doctor at the hospital. 66World Bank, Proshika, and Survey andResearchSystem, 2002, Bangladesh Urban Service Delivery, A Score Card. 67Ibid. 51 Cost can also be prohibitive for the poor. For example, while hospitals are supposed to be free, slum dwellers report that they are required to make payments. Residents report that no service is available without `speed money.' The magnitude o f these `fees' can be relatively large. Households report having to pay to get admitted, to see a doctor, and for medicines and other essential medical supplies.68 Similar to the situation with education facilities, NGOs and donors remain reluctant to set up clinics inside the slums incase there i s an eviction and resulting loss o f infrastructure. Water and Sanitation. Water and sanitation are delivered through pipe connections to homes, public taps, and tube wells. In poor communities, the poor generally rely on private sources, and to a more limited extent, WASA. Hardware, including facilities such as latrines and water points, are a primary need for'these communities. Studies in urban slums report women and girls often waiting for hours to access the limited tube wells available for a large number o f families. The government's guidelines mandate that water should be provided to all urban &reas and all people covered by each municipality. Inpractice, this is not the case. Slum dwellers reported paying as much as Tk. 2 per bucket o f water, several times the price paid by those who had legal connections (the official price charged to consumers inDhaka was Tk. 4.33 per 1,000 liter^).^' B. Sersice Providers Government institutions. The administrative structure related to service delivery is complex. Services are deliveredby a mix o f central and local agencies, with limited resources, weak administrative capacity and little coordination. Dhaka City Corporation (DCC), which i s responsible for a wide range o f services listed inTable 4.1, cannot perform their functions adequately due to severe resource constraints and limited authority. DCC i s dependent on central government for financial grants and for staff appointments. It also has no role incity planning or physical development. There are 10 zones in the city, with 90 directly elected ward commissioners who are members o f the DCC having both policy making and input monitoring roles. Ward commissioners, however, do not have adequate resources, training, or staff to carry out their roles effectively. Budget resources are insufficient to meet demand. Local resources come from property taxes, some cost recovery, and government transfers. There are inefficiencies with each and the resources are well below what is needed to cover the cost of service provision. For DCC, 63 percent o f resources come from own revenue, and 37 percent from Government grants (2003-2004 Budget documents). For the poor, services are even more limitedas Government agencies do not work inslum settlements. 68hid. 69Ibid. 52 Table 4.1:Administrative structure o f service deliverv inDhaka Agency Services Sources of Financing Central government agencies Education, health, legal, Taxes, donor funds judicial, police, land regstration, Dhaka City Corporation Sanitation, solid waste Property taxes, conservancy, disposal, road buildingand lightingand water tariffs, fees, maintenance, street lighting, fines, rental income, traffic signaling, parks, government grants, donor playgrounds, graveyards, funds slaughter houses, market places, street addressing, provision o f nominal stipends to primary education, slum improvement, mosquito prevention Pourashava Sanitation, solid waste Propertytaxes, conservancy, disposal, road buildingand lightingand water tariffs, fees, maintenance, street lighting, fines, rental income, traffic signaling, parks, government grants, donor playground, poverty funds alleviation, slum improvement, planning, etc. WASA (Water and Sewage Drinlungwater, sewerage Water tariffs, loans, Authority) government grants, donor funds R4JuK Planning and development o f Sale o f lands, government physical infrastructure, grants including housing Specialized authorities: Civil works, housing, physical Government budget, donor PWD, NHA, DPHE, LGED develoDment funds DESA (Dhaka Electric Supply E1ectricity Authoritv) I Titas Gas Supply Sources: Chowdry, 2004, Worl Bank 2002 NGOs. The role of NGOs in delivering services in Bangladesh is significant. The biggest NGOs have traditionally worked in rural areas, with a relatively small number working in Dhaka. Thirty NGOs were estimated to work in Dhaka in 2003 though this i s thought to be growing (INTRAC). The Coalition for the Urban Poor, an umbrella organization o fUrbanNGOs, has 44 members. As NGOs have become increasingly prevalent, there is also concern o f a diminishingaccountability between elected central and local Government representatives and clients. Weaknesses inthe regulatory framework and in financial accountability raise questions as to the fiduciary risks o f further expansion (World Bank, 2005, The Economics and Governance o fNGOs inBangladesh). 53 Mastaans. Services are also provided by individuals (mastaans) who provide services to slum residents by circumventing the system, in return for a fee. They facilitate "illegal" connections to essential services like electricity and water, and in the process are able to benefit by diverting resources away from the state. All slums are controlled by mastaans, who have different levels o f authority and hierarchy (See Wood, 1998; Islam, 1996; Paul-Majumder et al, 1996; Rashid, 2004). There are the local level mastaans who control sections o f a slum, and more aggressive mastaans who have authority over the lower level leaders o f the slum. Both are generally considered to be involved in illegal activity such as extortion and violence. They will organize into gangs/groups and demand money from the urban poor who live under their `protection' intheir "slum area." Many mastaans are reported to have close links with politicians, municipal authorities, and the police in the neighborhood. Political parties rely on links with local mastaans for electoral support and re-election of particular candidates in an area. A mastaan's power base i s hrther consolidated ifthe political partyhe belongs to forms the government. Thus, patronage relationships characterize slum politics, which extend from the slum all the way into the local authority andpolitical parties." Box 4.I: Making Things Work in Health Care BRAC has examined ways o f overcoming problems inhealth care delivery to urban poor. Despite all the constraints, they have managed to implement a program that has proven successful. BRAC's essential health care program employs community health volunteers called Shastho Shebika (BRAC 2004). The health volunteers go door to door making house-calls inpoor communities. BRAC Village Organization members handpick the volunteers. Each one i s assigned to 300 households on average visiting 15 households per day. The volunteers are trained to treat and recognize ten o fthe most dangerous and common diseasesinBangladesh. While the volunteers are not salaried by BRAC, they are able to make an income through the sale o f essential health commodities, such as drugs, contraceptives, birth delivery kits, iodized salt, hygienic soap, sanitary napkins and vegetable seeds. To provide for a working collaboration inall aspects o fhealth care, they provide assistance on government health initiatives, such as government immunization centers and distributingVitamin-A capsules. Ifa patient needs further medical attention, the Shastho Shebikas refer them to BRAC's health centers or public sector secondary- level health facilities. Through this set-up, BRAC i s able to reach a large percent o f the poor population despite constraints such as working women who have no time to seek medical attention, and those affected by slum evictions, and lack o fpermanent infrastructure. ' O A recentreportina newspaper found that 60percentofMembersofParliamentinthe gqvernmenthave l i n k s with smugglingor criminalelements inthe country(NewAge, 2004). 54 11. Constraints to Service Delivery Three major constraints in providing service delivery to urban slum residents were identified in interviews with donors, government agencies and NGOs. These include: i)lack o f government policy giving them rights and access ii)the eviction o f slum residents iii)and the role o f mastaans in the absence o f formal government structures. A. Lack of governmentpolicy The lack o f an appropriate countrywide policy dealing specifically with urban poor was cited by NGOs, donor agencies, and even some government divisions as a major constraint to working in urban slums. There i s no central government policy that mandates individual divisional policies or a national strategy on this subject. Some specific government agencies are able to include individualmandates in their provisions; however, few chose to do so, as there i s no pressure or incentive to do so. The State's ambivalence towards urban policy i s manifest in the conflicting dual metropolitan power structure. Though the city Corporation is autonomous and its Mayor and Ward Commissioners are elected by direct votes by the city-dwellers, its power i s controlled by the Ministry o f Local Government, Rural Development and Cooperatives (MLGRD). The municipal governments do not have sufficient control and leadership over municipal affairs (Rashid, 2004). As a result, the repeated attempt by the Dhaka City Corporation (DCC) to create a city government under which all utilities and services within city would be the jurisdiction o f DCC failed during the past Awami League government tenure. The lack o f a local coordinating body such as DCC leads to a lack o f inter-agency coordination and poor governance between services delivered by WASA, DESA, Titas Gas, BRTC and other private transport agencies. B. Evictions of slum residents As discussed in Chapter 3, land tenure is a growing problem inDhaka, where 70 percent o f the poor have access to a very small percentage of land, while the remaining 30 percent o f the city's population holds 80 percent of the land. Eviction i s always a threat for service delivery, with the poor considered illegal and therefore not able to access formal services. The insecurity o f land tenure leads to major barriers to effective service delivery not only for slum dwellers, but for implementing agencies as well. With the constant threat of evictions, Government agencies, NGOs, and donors are reluctant to invest capital in erecting permanent structures ifthey stand to lose them ifthe slum gets evicted. UNICEF faced this problem repeatedly with their education centers where some are reportedto have beenbulldozed. 55 Box 4. 2: CaseStudyfrom CUS: Slum Eviction in Agargaon One o f the largest slums, based inthe area o fAgargaon, was settled on government landtwenty years ago. After 20 years o f undisturbedsettlement bythe urban poor, a contract was issued to construct new government institutions on that land. This meantthe immediateeviction o fthousand o f slumdwellers. A number o fNGOs had projects implemented inthe slum, such as Plan International, which was runningwater and sanitationprograms, healthcare, and non-formal education. They lost a lot incapital investment when the slum was evicted. Other projects, such as successful micro-credit programs, were able to continue as the implementingNGOs were able to track down community members and resume repayment o f loans when the families were resettled. The uncertainty of the security o ftenure presents a difficult challenge for those agencies wanting to work inthe slum areas. C. ThePervasive Role of Mastaans As mentioned, the gap in service provision has been filled by mastaans who usually control the acquisition o f and provision o f amenities, such as latrines, tube-wells, water and electricity connections as well as interventions by NGOs in the slum (Rashid, 2004). Most slum residents build networks and links with mastaans who act as brokers, assisting them with access to basic services for a high fee. Most often, these utilities are acquired through illegal connections. Because mastaans are often the only service provider in slum areas, the situation for residents i s dire as they report regular extortion, and fear of physical harm or eviction if payoffs are not made (Rashid, 2004). Additionally, mastaans hamper the effective service delivery o f outside NGOs to slum areas for fear they will be a competitor to the services they provide. This creates a major obstacle to reaching the poor. D. Other constraints Beyond the constraints mentioned above, there are additional constraints that were raised inthe interviewsby NGOs, Donors and Government agencies. Among these are: Lack of appropriate infrastructure, particularly schools and health clinics, to meet the needs of the population. The existing infrastructure simply does not meet the needs of the population making effective service delivery impossible. 0 Limited involvement of municipalities and lack of technical and financial capacity. With both o fthese factors, it i s difficult to work with local governments inimplementingservice delivery. 0 Mixed experiences working with NGOs. While some NGOs have made great strides in working in slum areas, others do not have adequate technical capacity which limits the work that they do. There have also been concerns about the 56 governance and transparency for some NGOs based on a few unfavorable experiences inrecent years. Challenges in identzjjing the urban poor. Slum settlements can be heterogeneous indensely populatedurbanareasmakingit difficult to identifypriorityareas. Funding. NGOs rely heavily on donors for funding. The NGOs express fi-ustration over continual changing processes, more limited funds available to NGOs, and donor coordination problems. Challenges in working with Government agencies. This includes difficulties in working through layers o f government bureaucracy, as well as a lack of accountability and transparency in government processes. Governmentpolicy priority towards ruralpoor. Frequent strikes. Political rivalries have led to major disruptions in every day life. Frequent hartals or national strikes shut the entire country down and make it difficult to proceed with program implementation. Box 4.3: Making Things Work in Water and Sanitation Despite the fact that there is no central government strategy on providing for the urban poor, the Local Government Division (LGD) took the lead on designing and implementing a Pro-Poor Strategy. The strategy holds great promise and i s a first step in providing services to poor communities: LGD's Pro-Poor Strategy The Unit for Policy Implementation o f the Local Government Division designed a Pro-Poor Strategy for the Water and Sanitation Section that i s to be implemented in2005. Inthe interest o f decentralizingthe government, the local government institutions Union Parishads and the City Corporations will be given authority over this project. They will identify extreme poor and poor households based on a set o f eligibility criteria. Based on their selection, the government will provide a subsidy for the installation o f water and sanitation services. Individual households will be expected to contribute 50 percent o f what non-poor households would contribute to a government project o f this magnitude, with a cap at Tk. 500. This payment canbe made ininstallmentswith a cap at paying Tk. 25 per month for the hardcore poor households. Sanitation systems that are covered under this project are defined as one hygienic latrine per household. If this i s not possible, a maximum of two households can share one latrine, and in the event o f community latrines, each should provide for a maximum o f 10 people. Drinkingwater facilities should provide 20 litres per capita per day. The water source should be within 50 metres o f households and meet the national water quality standards. The communities will be given authority to choose the technology o f implementation o f these services. While this Pro-Poor Strategy i s a step in the right direction, there i s no specific provision for the urban poor within the program. Rural and urban poor have different needs and their infrastructure set-up varies. There i s very little land space in urban areas, while rural areas can spread. out over many kilometers. The daily threats and fears o f the two groups vary as well, and income levels and sources vary. However, the Local Government Division i s very hopefid of this strategy, and believes that soon other government divisions will implement similar strategies. 57 111.ImprovingService Delivery for the UrbanPoor Inorder to achieve sustainable poverty reduction inDhaka and other urban areas, Bangladesh will need to confront the issues o f service delivery. There are vast improvements needed in every sector. This will require a major shift in current government policies and practices. Yet the Government cannot do it alone. The most effective programs for service delivery appear to be those built on partnershipsbetween NGOs, Local Government, Donors and the private sector. Ensuringthe enforcement of the rule o f law for those illegally providing services is also essential. Finally, clarifying and thereby empowering the roles and responsibilities o f local institutions will enable themto tackle the challenges of service delivery as is done inother countries. As pointed out in a previous World Bank Report (2002), service delivery will also require increasing partnerships with the private sector, changing the structure o f incentives for sefvice providers and consumers, increasing the involvement o f users and other stakeholders in planning, providing and monitoring service delivery by institutionalizing user surveys, and institutionalizing a system o f accountability that makes service providers answerable to service consumers. The interviews with Government, Donors and NGOs led to a number o f recommendations which are summarized below. Focusing attention on addressing urban poverty. The lack o f an explicit policy for urbanpoverty reduction has been raised numerous times as a major constraint. While this requires a major political commitment, developing a strategy through a consultative process including stakeholders i s important to focusing attention to the growing problems o f poverty in Dhaka and other Cities. An effective policy could provide a strategy for ' addressing poor areas, including the mandate and specific guidance for prioritizing the delivery o f services to the urban poor and dealing with the issue o f legal land tenure. Clearly identifying roles and responsibilities within Government, as well as for partners outside the Government (e.g. NGOs, private sector) will helpto ensure implementation. Strengthening the role of local municipalities. Currently many basic services are runby Central ministries who are not always able to respondto needs at the local level. There i s much scope for strengthening the role o f DCC in urban management, and playing a greater role in improving interagency coordination with the various development authorities responsible for infrastructure development and service provision. Improving accountability and oversight. There i s a need to hold agencies accountable for delivering quality services, with a mechanism for allowing grievances to be handled. Approaches used in other countries include a performance based monitoring system with credible indicators and feedback system and the introduction o f incentives in public sector organizations. , 58 Enforcing law and order. While Mastaans do provide services to the poor who have no other options, the practices o f extortion are illegal, violent and can detract from city development. Such practices can only be considered a detriment to foreign investment inDhaka. There i s a substantial role for legitimate service delivery by private providers, with the public sector ensuring a regulatory framework that ensures transparency, and adherence to the rule o f law. Coordinating between serviceproviders. Given the range o f services and service providers who face the same challenges, it would be beneficial to 'foster coordination across institutions and to identify potential synergies. All stakeholders mentioned the need for improvements in coordination. This includes improved coordination between government agencies, Donors, and NGOs. This could be achieved through activitiessuch as stakeholder workshops and capacity building. Developing alternative payment schemes. Cost recovery schemes are viable as evidenced by the fact that the poor already do pay substantial amounts for services. The development of alternative schemes could provide more flexibility for the poorest ensuring that they will be able to afford services intimes o f need. For example, a scheme designed as a savings plan that would enable pre-payment towards services or purchase o f vouchers when resources are available would help to mitigate shocks such as health emergencies or seasonal shifts in employment. Utilizing alternative delivery mechanisms. For some services, alternative delivery mechanisms may provide a way to reduce cost and improve access. This could include distance learning programs, mobile health units, and shared water points where groups o f households could join in purchasing a water connection. Households would then be responsible for organizing payment, and the operations and maintenance o f the equipment. Building capacity through training. Implementation capacity could be strengthenedat many levels including central and local government institutions, NGOs, and community members. Training in implementation issues, administration, and technical capacity would ultimately benefit the quality ofthe work that i s carried out. 59 Box 4.4: CaseStudy: Lessons Learnedfrom DSK's Water Delivery Project in Conjunction with DWASA There are several important lessons learned from DSK's negotiations with DWASA to provide water to urbanslum communities. Among them: i)There is awillingness andability among theurbanpoorto pay for basic service delivery (Ahmed, 2003). The myth that providing service delivery to poor communities requires heavy subsidies is disproved. Additionally, DWASA benefited from the increased revenue, which acted as a motivator to continue the working relationship. Moreover, DWASA has now initiated their own "Urban Water and Sanitation Initiative for Dhaka's Urban Poor" that i s to be implemented with finding from Plan, UNICEF, Water Aid, and DFIDin collaboration with local NGOs. ii)Mediation and negotiations for the rights ofthe urbanpoor by anNGOcan be successful. It was only through DSK's continued negotiations with D W A S A that water points were supplied to slum communities. iii)Communityownershipiscriticalforthesustainablesuccessofanyproject. Thefirst attempt for a water point in a slum was taken over by the local mastaan. The second attempt put management o f the water point in the hands o f the community themselves, who did not allow a take-over by the mastaan. It i s almost impossible for an outside NGO, such as DSK to extort any control or pressure o f slum leaders; the control has to come from within the slum community. iv) If a community is already accustomed to making installment payments or regular payments o f any kind, such as through microcredit loans, they are more likely to make timely payments for services such as water. Community ownership should be taken to the extreme goal o f leaving the entire process up to community members, from negotiations with DWASA, to implementation o f the water point and keeping local power seats out o fthe process. v) As with any successfil project, the cooperation between the field staff and upper level management i s essential. The commitment o f senior level managers is as important as the commitment o f field-based staff. 60 CHAPTER 5: CRIME AND VIOLENCE AND THE URBAN POOR I.BackgroundandIntroduction Crime andviolence have beenidentified as a key issue for Dhaka, andparticularly affects the urban poor. Accurate data on levels o f violence in Dhaka are scarce - but available information indicates that while levels o f lethal violence (homicide rates) are much below some o f the very violent urban centers in Africa and Latin America, overall levels o f violence are serious and perceived to be 0: the rise. The sheer scale and diversity o f crime and violence in the poor slums o f Dhaka means that it has become `routinized' or `normalized' into the hnctional reality o f life, and affects every aspect o f daily life for the urban poor. Reflecting general public opinion, daily newspaper reports attest to the extent o f the problem. The situation in Dhaka was described the following way in a 2000 publication: "...Dhaka has emerged as a city o f crime, insecurity and political violence.. .social unrest, violence, theft, robbery, looting, murder, hijacking, arson, throwing o f acid on innocent females, raping o f minor girls, possession and use o f illegal arms, illegal rent/toll collection, frequent traffic congestion, etc. phenomenally increased over the years and have now become a way o f life in Dhaka city." (Siddiqui et az, 2000). The situation in Dhaka i s thought to be worse than in the rest o f the country. There are few studies or statistics on crime and violence for Dhaka or other cities in Bangladesh. However, a recent IFPRI study conducted a small household survey in 14 slums inthe city of Dinajpur. The area i s thought to have a lower crime rate than Dhaka, and yet some form o f crime or violence affected one o f every six households in these slums in the previous 12 months (Garrett and Ahmed, 2004). In a 1995 survey of the urban poor 24 percent o f households described the security situation in their area as poor and 8 percent as very poor. The corresponding numbers for Dhaka were 32 and 14 percent respectively (Nazrul Islam et. aZ., 1997). Crime and violence are no longer considered just a `social' or `law and order' problem, but also as an obstacle to. development which erodes the poor's human, physical, and social capital, and with high associated economic costs. Whilst estimating the costs of crime and violence inDhaka is beyond the scope o f this study, these include costs to the judicial system, health care costs, foregone earnings, costs on private security, loss o f competitiveness, and loss o fjobs, production. (See Annex 4, Box A4.1). Crime and violence affect all levels o f society: the rich and - even more - the poor, women and men, and young and old. Urbancrime and violence also generate a climate o f fear. The fear of crime and violence are `serious threats to the stability and social climate o f cities, to sustainable and economic development, the quality o f life and humanrights'. (UN-Habitat Safer Cities, 2002). Ways to address crime and violence and strategies to deal with this public "epidemic" i s an area'that i s increasingly becoming a priority - internationally as well as inBangladesh - for government and citizens alike. The traditional response to crime and violence has been one of control and repression. Studies (e.g. Greenwood et al, 1998) 61 have shown, however, deterrents such as the severity o f punishment and the strength o f the police force alone have a limited impact. It is often argued that crime and violence is the output o f a complex set o f factors including rapid urbanization, persistent poverty and inequality, social exclusion, political violence, the more organized nature o f crime, and the growth o f an array o f illegal activities. As such, experiences in the US, Europe, and increasingly from developing counties themselves, have shown that crime and violence can be substantially reduced through well-plannedprevention strategies that go beyond traditionalpolice responses -especially at the local level. This chapter presents some o f the main manifestations o f crime and violence in Dhaka - particularly as they affect the urbanpoor - by presenting the main findings o f a qualitative and victimization study carried out in four Dhaka slums for this study. It will also briefly examine some o f the social, institutional, and legal aspects to the crime and violence situation inDhaka. Inthe final section, policy recommendations are presented. 11. Manifestations of crime and violence affecting the urban poor inDhaka Whilst there i s some information on the types o f crime and violence that routinely face the urban poor in Dhaka from anecdotal evidence, newspaper reports, and a few academic and qualitative studies from around Bangladesh - the various dimensions o f crime and violence such as the range, prevalence rates, costs, and associated risk factors are not fully understood. Routine police statistics - that are often usedto give some sense o f prevalence rates in particular - are especially unreliable in the case o f Dhaka. There appears to be a very high rate o f under-reporting o f crime and violence, particularly by the urbanpoor due to the costs involved o f lodging a complaint, legally, bribery and also due to the threat o f retaliationby police (UNDP, 2002; Siddiqi, 2000). Study Methodology: In an attempt to fill in some o f the gaps in existing studies and police data, a victimization study was carried out in the four slums included in the `Slum Observatory' project (explained in Chapter 1) in the 2004/05 round. The study asked key questions about the types o f crime and violence in slums in Dhaka; how prevalent crime and violence are; the perceptions and the degree o f fear o f crime and violence; the costs o f crime and violence and their impact on households; and lastly how the poor respond to crime and violence and information on their views about the police, victim support services, etc. The four slums located in Pallabi, Mohammadpur, Lalbagh and Rayer Bazar are quite distinct. They differ inage and type o f settlement and have quite different levels o f violence - Shahid Nagar reported to be the most violent slum, followed by Mohammadpur, after which the other two come with some distance. Aparajeyo- Bangladesh has been providing services to children and mothers for some years in these areas and the NGO has an in-depth knowledge o f and an established relationship o f trust with these communities. As part o f the study design, a series o f focus group discussions were held in the different slums with community leaders and residents to explore perceptions o f crime and 62 violence and potential responses to the survey questions. In addition, a series o f life histories were collected. Study Limitations: Whilst the study provides a wealth o f interesting data and information, a number o f limitations were encountered. A key challenge o f this type o f study is to ensure full confidentiality in order to gain frank responses. In most cases the interviewers report that full cooperation was received from respondents when discussing the general problems o f crime and violence in their slums. Many respondents seemed to enjoy having outsiders visit them and ask their opinions. However, due to the crowded conditions in the slums it was impossible to guarantee confidentiality. In most cases the head o f the households (including many female-headed households) were interviewed, and bystanders or family members often listened in and joined the discussion. This resulted in an unwillingness to report personal victimization information. So while high levels o f crime and violence were widely acknowledged and discussed, few respondents admitted to having been a victim themselves and the interviewers reported a real fear amongresidents inrespondingto personal victimization questions, e.g. `have you been a victim o f robbery over the last 12 months'. This does not appear to have been a problem in the Dinajpur survey and may reflect the higher levels of insecurity in the slums of Dhaka where few o f the respondents were willing to take the considerable and real personal risk o f being seen as an "informer" in spite o f repeated assurances o f anonymity andconfidentiality bythe interviewers. On the other hand- and to our surprise - there was a very highrate o f reported domestic violence. This i s contrary to international experience which tends to find that in order to get information on domestic violence special surveys and methodologies are needed - separately from general victimization studies. Again, this may reflect the fact that domestic violence and violence against women are extremely widespread and considered a part o f normal life in these communities - carrying no legal or social sanctions and to be freely discussed. A. Main Survey Findings Extent and variety of crime and violence: The results from the focus group discussions and surveys indicate a wide range of different types o f crime and violence that occur in the slums and confirmed the complex, multi-faceted, and all-present nature o f the phenomenon. The vast majority of respondents, some 93 percent, said that they have been affected by crime and violence over the last 12 months with 33 different types o f crime identifiedby the respondents (Table 5.1). Among the most commonly reported crime and violence are toll collection, mastaan-induced violence, drug and alcohol business, land grabbing and violence, gambling, violence against women and children, illegal arms' business, arson in slum, murder and kidnapping, and sexual, physical, and psychological torture and abuse of wife. During FGDs and recording o f life histories in particular, the widespread problem o f alcohol and drug abuse - including serious heroin addiction - was often linked to crime and violence (trafficking, crime to support habits) and domestic violence. 63 Slum ShahidNagar Raver Bazar Mohammadpur Pallabi Total Affected 250 249 213 216 928 100% 99.6% 85% 86% 93% Not Affected 0 1 37 34 72 0% 0.4% 15% 14% 7yo In addition, many other categories were identified such as `using children in political meetings and picketing', `loafers'dens', and acid throwing. Whilst not giving an indication o f how serious the presence o f each of these types o f crime and violence i s in the various slums, it clearly indicates the extent to which every aspect o f the daily lives o f the slumdwellers is affected by acts o fcrime and violence (Table 5.2). -r z lccur inthe Slums (Multi-response) Slum Shahid Rayer Moham Pallabi Nagor Bazar mad I pur I Total (perce:of respoi mts) affectedbv crime and violence 100% 100% 85% 86% 93% Clash amongmastaadterrorist groups 86 58 88 80 78 Politicalviolence 83 53 78 76 73 Toll collection 80 50 66 48 61 Carry and distribution of drugs, alcohol and substance 54 58 56 53 55 Committing crime under the influence of intoxicants 53 58 49 57 54 Sexualabuse 44 73 48 49 53 Kidnapping 50 52 36 50 47 Conflict betweenhusbandandwife 45 81 82 74 71 Farce inthe name ofjustice 49 27 58 56 47 Land grabbingunder political shelter 67 37 63 64 58 ~ Establishpolitical control 25 41 68 44 45 Theftipilferage 76 52 72 67 67 Mugging 45 68 69 55 59 Acid throwing 37 24 18 36 29 Conflict between house owner and tenant 56 27 56 43 45 Gambling 38 61 36 38 43 Torturing womedgirls for dowry 56 68 54 48 56 ~ Disturbing the community under the influence of cannabis, 46 57 54 48 51 alcohol & heroine ' Forcefully conducting sex trade with adolescent girls 19 27 20 31 24 Kidnapping girls 15 40 41 24 30 Selling illegal and intoxicating substances 67 67 60 47 60 Usingchildren inpolitical meetingsandpicketing 62 75 77 74 72 Harassment andregular toll collection by police 50 58 48 39 49 Violence and abuse of girls working as domestic aides 39 74 67 54 58 Loafers den 57 72 77 65 68 Known mastaadgangsden 51 58 42 38 47 Kidnap people from other places and holding them inthe 12 20 21 23 19 slumfor ransomimugging Exploding handbombs 33 22 67 32 37 64 Type o f Crime and Violence Reported by Respondents ,,,,,,,,7 Shahid Rayer Moham Pallabi o fyFpents)56 Illegal fire a m business and hiding arms inthe slum 53 Arson inslums to gain control 33 53 55 35 44 Collecting regular toll from shopkeepershawkers by police 36 50 51 30 42 and their aides Harassment by police inthe name of House Search 38 72 67 45 55 Collecting toll inthe name o f arbitration by local political 41 60 52 52 51 leaders Establishillegal casino to arrange circudfairs 38 71 42 41 48 Operate illegal financial transactions through local clubs 40 53 25 38 39 Table 5.3: The Seventy of Crime and Violence (Multi-response). Shahi Nagor Rayer Bazar Mohammadpur I Pall )i N Severity Severity Severity Severity scale N Scale N Scale N scale Murder 134 Hand bomb explosion 94 2 163 2 223 202 1 Use o f arms 184 2 161 2 190 174 2 Toll collection 207 2 166 3 193 187 2 Political conflict 166 3 217 4 203 182 3 Sexual abuse 73 4 216 4 222 225 3 Drugand substances trade 156 2 224 2 182 190 2 Harboring mastaans to 40 3 199 3 190 174 2 control slum 4 173 4 160 204 2 Women trafficking and 55 sex trade 2 146 2 143 2 2 1 9 9 These results also confirm the high level o f control o f mastaans over the slums and community life. This was also found in the Dinajpur survey. There seems to have been a city-wide rise in mastaani preying on city dwellers and collecting money from businesses, bus terminals, construction worksites, and slums over the last decade. According to a report prepared in 1998, there were 30 different groups, employing about 65 30,000 people, active in Dhaka City (GSS, 1998 as cited in Siddiqui et al, 2000). They are mostly area based, bus terminal based, and college/university based (in particular Dhaka University and Jagannath University College). It i s estimated that alcohol and drugs such as hashish, heroin, phensidyl, pethidine, etc. are sold at about 5000 points all over Dhaka City (Siddiquiet al, 2000). Political violence generally takes the form o f clashes between the police and opposition political groups or between supporters o f the government and opposition political parties. This i s particularly noted during hartals (general strikes), processions, demonstrations and political meetings. Other political crimes are secret killings, looting, destruction o f property, arson, and noting (Siddiqui, et. al, 2000). The survey findings also indicate the strong links between political parties and criminals living in the'slum areas - regularly paying (or compelling) slum dwellers, including children, to participate inpolitical rallies, protests, and open-air meetings. Perceptionsof safety. Overall, about 31percent o f respondents feel that they are not safe within their community though this varies greatly between slums. In Shahid Nagar, 64 percent o f respondents reported to feel not at all safe in their community, followed by 33 percent in Mohammedpur, and 19 percent in Pallabi. The respondents from Rayer Bazar feel quite safe intheir slum, only 6 percent reporting to not feel safe at all. (Annex 4, Table A4.1). Interestingly, much o f this i s attributed to the fact that the slum in Rayer Bazar i s reported to have a very strongly active Imam, who does not `tolerate' immoral or criminal behavior inhis area. These results reflect differentials in reported crime data as well as international patterns (JI'srban violence, which usually find that crime and violence are concentrated in a limited number o f `hotspot' neighborhoods and can also differ quite dramatically within different areas of the same slums. ''Whilethese are often some of the poorest slums, many very poor slums do not have serious problems o f crime and violence. This has real implications for the targeting o f crime and violence preventionand reductionpolicies. During focus group discussions with community leaders from around the city of Dhaka the following areas were identified as the most unsafe in declining order: Vasanti, Agargoon, Lalbag, Mohammedpur, Bapura, Ghattala, Kalsi, Kalayanpur, Taltala, Mesul. Where, when, and by whom?The survey results show that crime and violence occurs in a number of different places, most o f it within the slum and about 35 percent reported to happeninside the household compound (which i s usually sharedby a number o f different families). Other high risk areas are the roadside, the market side and the way to and from work. Female garment workers are particularly vulnerable on their route to work and will often walk in groups to provide some protection or even stay the night at the garment factory. Crime and violence appear to occur throughout the day; however, evening and night are `peak times'. When asked who is involved (perpetrator) in crime "Policedatawerecollectedfortheyear2004fromallfourThanaPoliceStationsrepresentedinthe survey. Unfortunately, the data were not available specifically for each o f the slums within these areas, but do clearly demonstrate differentials between areas. 66 and violence the most commonly mentioned are the mastaans, from both inside and outside the community, husbands (mainly involved in domestic violence), law enforcing agencies, and neighbors and `others'. `Others' include thieves, muggers, robbers, extortionists, teasers, snatchers, murderers, arsonists, etc. The respondents indicate that many poor slum dwellers are involved with activities such as the arms trade, drug trade, trafficking, andblack marketing as a means to maintain their livelihoods. Crime and violence against women. Violence against women was repeatedly raised during FGDs and the qualitative fieldwork as one o f the most serious manifestations o f violence affecting poor women in Dhaka. This includes physical and psychological violence inflicted by the husband or the husband's family (the latter often dowry-related), burning o f wives, acid attacks, suicide as a result o f physical and psychological torture, as well as violence against women inthe workplace, on the way t?- and from work, etc. Of our total sample (about 50 percent female respondents), 30 percent said to have been a victim o f domestic violence inthe past 12 months - in fact it was the highest reported incidence o f crime or violence reported in the survey. This i s likely still significantly under-reported as this was not a specialized domestic violence survey. The causes most frequently reported include marital problems, confiscation o f wife's money, polygamy, dowry related issues, and not taking responsibility o f children. See Annex 4, Table A4.2 for a full list; These findings reflect what i s known about violence against women in Bangladesh. According to the UNFPA (2000) Bangladesh has the second highest incidence o f violence against women in the world. A survey carried out in 1997 by the Ministry o f Women and Children's Affairs with the assistance o f the Government of Denmark found that not only i s violence against women widespread, it is also widely accepted: 50 percent of the women interviewed thought it the husband's right to beat his wife; 85 percent felt it was right to hit the wife if she was disobedient; 80 percent o f the women respondents felt that if a manrapes a women the woman should marrythe rapist. As mentioned above, part o f the relatively highresponse rate on violence against women in our survey may be explained by this high level of social acceptance of the p h e n ~ m e n o n . ~ ~ Women, particularly poor women, face social and economic exclusion in Bangladeshwhich is evident through the high levels of violence against women. While the constitution guarantees principles o f non-discrimination on the basis o f gender, as regards `family matters' women are subject to "Shariah" or religious law. Bangladeshis a signatory to CEDAW (Convention on the Elimination of all forms o f Discrimination Against Women) and despite the introduction o f a number o f laws such as the `Ordinance o f Acid-throwing Act", the `Ordinance o f Women and Child Repression Act', and the revised `Ordinance o f Family Act' - all aiming to assist women - the main problems are the lack o f enforcement o f these laws and lack of access for women to the legal system. 72 Factstaken froma World Bankpresentation: ViolenceAgainst Bangladeshi Women: afav facts and figures. 67 Costs and impact of crime and violence in the slums. The survey results show that there are significant financial costs associated with the crime and violence in the slums. Costs related to medical treatment, loss o f productivity due to injuries and direct financial costs due to the collection o f `tolls' - many garment workers report routinely having to pay 10 percent o f their wages in `protection' for example. Other costs include the illegal land renting fees, electricity and water bill fees, `tolls' collected by both local gangs and the police in the name o f providing `security', losses through thefts and muggings, etc. Unfortunately the study was unable to quantifythe total monetary costs to the poor o fthe crime and violence they routinely face. Thiswould certainly be an area of interest for further research. The non-monetary costs are also considerable. It is clear that inmany o f the more violent slums there i s lack o f social cohesion and considerable levels o f fear and mistrust. This negatively affects social capital and prevents community members from organizing. Many respondents report being afraid o f moving about their community after dark - which particularly impacts women. Annex 4, Table A4.3 lists some o f the negative effects o f the high levels o f violence in the slums as reported by the respondents. If anything, these responses indicate that whilst crime, violence, and fear are `routinized' and part o f daily life, they do nevertheless have an enormous impact on the poor's well- being and their impacts on children are o f particular concern. Some o f the impacts mentioned include: children sufferinghigh levels o f insecurity and mental anxiety; family disruptions due to domestic violence; tremendous pressure on the parentso f growing girls due to the risks to them; working women getting womed and scared o f sexual abuse; mental anguish due to unemployment status; and constant fear and insecurity due to use o f arson, arms, and explosives. Reported and non-reported crime and violence and access to justice. A majority (60 percent) o f crime and violence is not reported to anyone. Of the reported numbers, only 3 percent is reported to the police. When crime was reported, it was mainly reported to community leaders (16 percent) and family members (18 percent). The very low reporting rate to law enforcement agencies or Ward Commissioners, less than2 percent reflects the extraordinary low levels o ftrust inthese agencies. (See Annex 4, Table A4.5). The survey data also indicate that in 88 percent o f reported cases no action was taken and perpetrators went fully free. The police took action against a reported perpetrator in only 1 percent o f all cases. The lack o f reporting to the police or official action undertaken reflects the low prevalence of a formal preventive, investigative, or judicial authority in these slums. In 7 percent o f the reported cases the community took some kind o f action - reflecting the respondents' preference for and higher confidence in local shalish (See section 3 below). The role of politicalleaders and institutions and Law Enforcement Agencies in dealing with crime and violence. The links between the incidence of crime and violence and the role o f political leadershstitutions are perceived to be strong. (Annex 4, Table A4.4). Respondents indicate that the leaders o f different political parties 68 patronize the criminal activities and violence with the intention to control the slums through `terrorists' and mastaans. By controlling the slum dwellers, the politicians command a strong voting base, collect rent from the dwellers and income from other illegal activities, ensure the participation o f slum dwellers in various political meetings, gathering, rallies and demonstrations, and use them in political conflicts and confrontation. Duringthe FGDs it became clear that there i s a widespread feeling that the political commitment to reduce violence i s only rhetoric. At the same time, law enforcement agencies are hardly seen to play any more positive role incombating crime and violence. Police are seen to maintain a low profile in responding to crime and violence because o f their deep relationships with the offenders. Inmost cases the victims do not lodge complaints against the criminals infear of further retaliation by them because they often enjoy the protection o f the law enforcement agencies. It is also widely reportedthat inaddition to legitimate fees, the law enforcement agencies often collect bribes from victims just to record a case. When slum residents were asked what can be done to improve the crime and violence situation in their communities, the results range from better employment and income generation opportunities, to improved social welfare services and education and training, to improvedjustice, law and order, and governance (Table 5.4). Table 5.4: Respondent's recommendationregardingcombating existing Crime and Violence (Multi-response). Slum - - Shahic Vagor Rayei hzar Moham .adpur - -Pa rbi k N % N % N % - - N % N % Job opportunity 227 90 197 79 23 1 93 210 84 865 87 Training 118 47 81 33 178 72 186 74 563 56 Undertake different development 138 55 132 53 193 78 159 64 622 62 activities Better governance 146 58 137 55 189 76 181 72 653 65 Compel the police to behave lawfully 132 52 144 58 185 74 156 62 617 62 Stop political crimes 141 56 189 76 220 88 154 62 704 71 Punish criminals through speedy 142 56 139 56 191 77 156 62 628 63 trials Develop strong public welfare 120 48 163 66 181 73 174 70 638 64 leadership Adopt all possible measures to 58 23 96 39 117 47 149 60 420 42 prevent women trafficking Take strong actions against sex trade 94 37 107 43 127 51 140 56 468 47 Strong control over black marketing 104 41 144 58 136 55 69 28 453 45 and drug trade Control toll collection 148 59 180 73 171 69 - - 125 50 624 63 All 252 25 248 25 `249 25 - - 250 25 999 100 Source: Slum Observatory f wey, 214 69 111.Relationshipbetween crime andviolence and social, institutional,and legal issues Various frameworks to understand crime and violence and to develop corollary short-, medium-, and long-term policy responses have been developed. One prominent approach comes from the field o f public health and is based on the identification and addressing o f risk factors.73 The risk factors inthis model are usually classified in three groups: individual factors, domestichousehold factors, and societal factors. The available information on Dhaka shows that this model i s consistent with the characteristics o f crime and violence - they are a complex mix o f interpersonal, social, economic, and structural factors. Crime and violence i s carried out by a variety o f different individuals and groups and ranges from widespread domestic violence, to disorganized `opportunistic' street crime, to the extremely well organized - and who operate for motives that range from hunger, to greed, andto politics. Justice System. It is clear from the survey findings that there is an almost complete lack o ftrust by the urbanpoor inDhaka inthejustice system. This i s confirmed by the comprehensive 2002 UNDP report on Human Security in Bangladesh which finds that many laws and practices o f the criminal justice system in Bangladesh are `anti-poor' and have a far more harmful effect on the poor and disadvantaged than on other sectors o f society, despite constitutional guarantees to the contrary. Consequently, many citizens, especially the poor, do not feel adequately protected. The report finds that: many laws are open to interpretation and to the use o f discretionary powers, and that the current bail system favors the wealthy and influential and penalizes those without money or influence. Thus the poor are over-represented in prisons whereas many wealthy or privileged individuals guilty o f crimes manage to escape the process o f law; that the legal grounds for arrest without warrant are too broad and allow the police to exercise a high degree o f discretion. Most preventive detention cases never reach the courts, whilst o f those that do, almost 90 percent are found to be illegal; that victims o f rape and other sexual violence continue to be further victimized by a legal systemthat discriminates against victims and protects the guilty; that laws relating to dispossession o f immovable property, public nuisance, and vagrants are not implemented in-line with their original intent, and often discriminate against and harm the poor; and that in addition to loopholes in laws and procedures and lack o f awareness on judicial issues among the population, there i s a severe implementation gap. (UNDP, 2002). Specific recommendations are made in order to improve the laws and legal procedures and to ensure that they do not discriminate against the poor. Access to justice for the poor and legal aid. The enormous costs of going to court, the delays in court proceedings, and the lack o f legal aid facilities has made the 73 Whilst looking at risk factors i s very important, various authorshave highlighted the needto also look at the strengths and assets or `protective factors' ofindividuals and communities, notjust their problems.See International Centre for the Prevention of Crime, 2000; Moser, 2000) 70 judicial system virtually inaccessible for the vast majority o f the poor. The main providers o f legal aid to the poor are NGOs such as BLAST (Bangladesh Legal Aid and Services Trust) and BELA (Bangladesh Environmental Lawyers Association). Despite constitutional provisions stating a citizen's right to legal defense, there i s currently no legal obligation for the State to provide legal aid if the pleader or accused i s unable to afford the fees(UNDP, 2002). Shalish (informal courts). It is estimated that nationally about two thirds o f disputes never enter the formal court process and are either settled at the local level, through informal settlement by local leaders or a community court, or remain unsettled. The results from our survey indicate a much lower use o f this informal settlement mechanism (8 percent), however, it was still the most used form o f seeking justice inthe Dhaka slums. Since settlement o f cases through formal courts i s both time consuming and expensive, Shalish councils provide a welcome alternative that could be strengthened and supported. However, it i s also recognized that the institutional affiliations and political backgrounds o f court officials and community leaders means that their relations to other institutions, particularly the police, local political leaders, and wealthy and influential individuals open these shalish courts to outside influence and may affect their functioning (UNDP, 2002). Dhaka Metropolitan Police. Weak law enforcement, impunity, and human rights violations are central to the spread o f violence. It i s clear from our survey that the poor have extremely low expectations o f the police. As a result, they generally prefer the intervention o f community leaders to resolve conflict among community members, family members,or preferto not report incidents at all. They will only go to the police as a last resort. Concern is widely expressed about the complicity o f the police with the mastaans; torture o f suspects in police custody; demanding and taking o f bribes for `protection' or to lodge a complaint; poor level o f police responsiveness to the poor; the vulnerability of witnesses, with witnesses complaining o f having to pay transport expenses, being subject to police pressure to distort facts, and some being threatened by the other side in the case. There is a lack of any type of special witness protection program. On the other hand, it is also generally agreed that the size o fthe police force is too small, that police lack educational qualifications, professional training, and sufficient pay and allowances. Specific recommendation for reform include better training, rationalization o f the police staffing structure, monitoring and performance indicators, and reviewing salary structures and incentives. It is found by our and other studies that the poor do want a police presence in certain areas such as markets. The trend towards community policing i s welcome and should be the focus o f the Dhaka Metropolitan Police to facilitate community participation in crime and violence prevention and to reduce the vulnerability o f the poor (UNDP, 2002, Siddiqui, 2000). See Box 5.1 for an interestingand successful example incommunity policing from Mumbai. Duringthe last year, the Government formed the Rapid Action Batallion (RAB) which i s a separate police unit that i s mandated to provide immediate response to law and 71 order incidents. The RAB has been given some extra-judicial powers for combating crime and violence and its officers are better trained and better equipped in terms of weapons and transportation than the police. They are also generally considered to be `clean' and relatively free o f corruption and political influence. The RAB functions, in large part, on `tip-offs' and over the last year has killed a number o f professional criminals or mastaans that were involved in a variety o f criminal activities during `cross- fire'. As a result it appears that the RAE3's actions have contributed to some improvement in the overall crime and violence environment as some criminals appear to have gone underground or have become more circumspect about their activities. Consequently, the RAE3 i s popularly well regarded. There are however serious concerns related to the extra- judicial nature o f many o f the actions taken by the RAB - denying suspects access to the due justice process - and overly relying on `informants' leaves it open to be easily manipulated. IV. AddressingCrime andViolence This chapter has provided a briefreview o fthe main findings o f a survey on crime and violence and the urban poor in Dhaka and has shown that urban violence is now widely recognized as a serious development problem. This is, however, a relatively new area o f research and intervention, with serious data constraints as well as very limited systematization o f information either analytically or operationally. The traditional response to increasing levels o f crime and violence has been one o f control or repression. This approach focuses on addressing the problem after the crime or violent act has been committed. It is usually related to `toughening' up the legal and justice system, increasing policing resources and capacities, and introducing harsher penalties inan effort to deter and repress crime and violence. Under this approach, crime and violence are seen as the responsibility o f the police and the courts. Most countries battling high levels o f crime and violence, however, find that these measures alone are ' not sufficient to have a significant impact on crime and violence. This i s often accompanied by a loss o f confidence inthe criminaljustice system, whilst public concern about crime and violence remains high (ICPC, 2000). A second - and complementary - policy response is one o f crime and violence prevention. The basic premise is to stop the crime or violent act from occurring in the first place by understanding and addressing the causes o f crime and violence, the risk factors associated with them, as well as constructing safer communities by building on their strengths (e.g. level o f community organization), rather than exclusively focus on a community's problems. Furthermore, prevention i s a much more cost effective option than repression. Thus while there i s no `silver bullet' to the problem o f urban crime and violence and little systematic evaluation to date, international experiences with different policy approaches suggest that an integrated framework combining their various strengths can lead to a successful urban crime and violence preventionheduction strategy. This 72 framework for local action should be based on a participatory and comprehensive diagnostic,andinclude some elements of: 0 judiciaYpolicing reform74 ensuring that order, fairness, and access to due - process i s maintained inthe day to day activities o f the community and reducing the public fear o fcrime; 0 social prevention- targeted multi-agency (multi-sectoral) and community driven programs that address the causes andrisk factors o f crime and violence; 0 situationalprevention-measures that reduce opportunities for particular crime andviolence problems through urbanspatial interventions and slumupgrading. 0 Lastly, it requires a shared local vision, strong leadership, political commitment,and an action plan for the short, medium,and long-term. From regional and international experience it i s evident that one o f the most effective entry-points for crime and violence prevention is the local level. It is the level o f government closest to the people andwhere projects can be designedto target the specific needs o f the local community. This i s also where the day-to-day delivery o f services happens. These services improve people's quality o f life and build better living environments. Many o f these services are also the basic elements o f crime and violence prevention. However, effective local government action requires all the municipal services to work together, rather than in isolation. It requires support from the different sectors inthe community such asjustice, health, education, media, police, social services, the private sector, NGOs, and other CBOs such as religious organizations. And importantly, it requires support from higher levels o f government and links between the national level, state, region or province. See Table 5.5 for a summary o f an interesting example o f a local multi-sectoral crime and violence prevention initiative from Capetown, South Africa. Specific recommendationsfor crime and violenceprevention in Dhaka Adopting a community-based approach. Working with residents' associations, women's savings and credit groups, CBO's, communities, and NGOs i s essential to identify the communities' priorities, problems, and strengths and in order to build the virtually nonexistent trust between those agencies responsible for the maintenance o f law and order and the prevention o f crime and violence (the police) and the communities they are meant to serve. Interesting methodologies in participatory community mapping for crime prevention are being developed and used in South Africa and Latin America and can be adapted for use in Bangladesh. Initial results o f these processes indicate that the process allows people to understand that crime does not occur randomly but that it happens in certain and predictable places. The process also has the ability to empower communities and to act together with the police in order to prevent and reduce violence crime (See e.g. Lieberman and Coulson, 2004, World Bank, 2003). 74 W s t the World B a n k can not be drectly involved in policing issues such as police reform, indnectlp it can recogruze the importance of the role of police and judicial system, work on judicial reform, and encourage crime and vlolence prevention partnerships that include the police. 73 Investing in more research and better information systems: Much more information and understanding i s needed about the types o f crime and violence taking place in Dhaka; associated risk factors; costs and impacts; prevalence rates; and trends. A first step would be to create a "Violence Observatory", in partnershipwith Government, the Police, forensic/legal medical institute, public health authorities (injury prevention/surveillance systems), and a research institute. The hnctiono fthis kindo f center i s to systematically collect data about crime and violence from different sources and from around the city as well as to carry out periodic victimization surveys. This information is then analyzed and used to inform short-, medium-, and long- term policy decision about `hotspots', crime trends, priorities, targetingo f social prevention activities, etc. Improving inter-institutional coordination: crime and violence in Dhaka arises from a complex mix o f causes that require a coherent, integral, multi-sectoral approach and the participation o f various public/state, and civil society agencies. Crime and violence prevention is not just the role o f the police or the judicial sector, and a local champion who will assume leadership and coordination inthis area i s crucial. There i s also much scope for an increased role for NGOs. Creating public space and implementing integrated slum upgrading approaches. There has long been an understanding o f the link between `crime and grime' or rather o f the link between the incidence o f crime and violence and poor infrastructure, poor living conditions, lack o f adequate communications technology, poor public lighting, inequitable distribution o f facilities and amenities, lack o f maintenance, etc. Therefore, one important violence prevention strategy is the regularization and upgrading o f slums integrating `design for safety' principles as well as social prevention activities. Many public spaces in Dhaka and in its slums have lost their traditional function o f promoting socialization, recreation, and citizenship as they have been invaded by the informal economy, mastaans, delinquents, drug dealing and consumption activities, etc. There i s virtually no public or recreational space available in the slum areas. Physical upgrading o f public and slum. spaces will promote community appropriation o f these spaces and prevent crime and violence from occurring inthese places. This would include: i)a selection o f `hotspot' areas for intervention; ii)diagnosis o f relevant problems and o f neighborhood perceptions o f public space and insecurity in these areas; iii)assessment o f already existing and related programs and new initiatives; iv) creation o f a local partnership and strategy that includes the community, residents' associations, NGOs, community leaders, and state agencies. See box 5.2 for examples o f how the World Bank i s piloting the integration o f a Crime and Violence prevention component in more traditional integrated slum upgrading programs in Latin America and the Caribbean. 74 0 Improving trust between ,the police and the community and improve the judicial system and deal with impunity. Community policing should be the focus o f the Dhaka Metropolitan Police to facilitate community participation in crime and violence prevention inthe slums and to reduce the vulnerability o f the poor. Inaddition, action is needed to address the anti-poor biases in laws and their implementation; speed up judicial processes - possibly through working with shalish informal community courts and community policing programs; and to improve access to legal aid by working with legal aid providing NGOs. Training for police on implementingnew approaches such as community policing i s also needed. 0 Implementing programs for the prevention of violence against women. Domestic and violence against women needs to be addressed through stepped up programs on introducing legal awareness, literacy and advocacy campaigns, legal aid, andbetter training o f the police andjudicial system so that women victims o f violence can access the justice system without fear o f further victimization. Introducifig women complaint officers in police stations could reduce the intimidation to report acts o f violence. In addition, programs such as the `One- Stop Crisis Centre ' are providing valuable experiences in providing integrated services to women and children victims o f domestic violence. Further support and engagement of the women's movement is essential as well as working with and channeling resources for the prevention o f domestic violence with community-based women groups such as lending and credit co-operatives. Addressing domestic violence will require an attitude change in society and will take time. Investing in a wide spread public education targeted at specific groups could beginto foster that change. 0 Focusing attention on the needs of children and youth. In order to prevent crime and violence in the medium and long term, much attention and programs are needed that target the needs o f children and youth, boys and girls. International experience has shown different violence prevention programs to be effective for different groups, in different types o f communities, and in different countries. Still, based on reports, evaluations, and research studies, certain types o f efforts seem warranted. These can take the form o f school-based, community- based, or family-support programs such as (Guerra, 2005): Individual levelprograms 3 Universal, high-quality childcare andpreschool education for young children k Socialskills andproblemsolving interventions for primaryand secondary school children k Life skills training and civic education for adolescents Building relationships: programsfor families, peers, and mentors k Homevisitationprograms for parents ofinfantsand young children 75 Parent school partnerships that encourage parental involvement in children's education and learning Programs and policies to prevent child abuse and maltreatment Parenttraining programs, particularly those that help families utilize resources Gangprevention programsthat combineprevention, intervention, and suppression Mentoring programs that provide role models for at-risk youth School and communityprograms Improving the quality o f education, including universalizingprimary school Emphasizingcooperative learning and student/family engagement Developing community policing programs that are sensitive to local conditions Buildinginfrastructurewithin communities to provide opportunities for youth and families to engagement inpositive activities (recreation, learning, employment) Buildingphysical and social capital incommunities through collaborative efforts Investing in drug and alcohol abuseprevention/treatment. Giventhe known role of alcohol and drugs as risk factors for violence, both from international evidence as well as coming strongly from our fieldwork in Dhaka - decreasing availability and ease o f access to both should contribute to reductions in violence - particularly domestic violence. Although large scale studies are scant, thert :: evidence that reducing alcohol availability i s related to a drop inviolence rates. In addition, alcohol and drugs prevention and treatment programs specifically targeting the poor are needed. 76 Table 5.5: Linking urbanspatial and non-spatial interventions: The KfW/ City of Cape Town KhayelitshaViolence Prevention through Urban Upgrading Project Spatial Types of violence and manifestations I Spatial and non-spatial violence prevention o r reduction manifestation interventions Shebeens Shebeen Violence Shebeens (bars) relocated to sites where social & police control is Assault more efficient Murder Alternative opportunities for socializing where alcohol is . Rape .------------------------------------------------------- controlled Drug & alcohol-related violence Business code o f conduct by Shebeen owners' association Murder; domestic & child abuse Domestic spaces Domestic Violence More houses o f refuge & counseling facilities Assault Police stations equipped with trauma facilities & female officers Rape ._________----______-------------------------- Facilities for conflict resolution Child abuse 0 Police receive training in handling domestic violence cases Emotional abuse Awareness-raising campaign on domestic rights Open public space Gender-Related & Economic Violence 0 Improved street lighting & visibility Openfields Rape Widely spread & functional telephone system Narrow lanes Empty stalls Robbery Rape-relief centers & self-defense training Assault Safe walkways Murder Vegetable stalls locked at night ._________----_____--------------------- 24-hour internal public transportation system More visible police patrolling& neighborhood watches Banks & A T M s Economic Violence ....................................................... Increased access to banking & safe deposit places RobberyOn long routes to ` banks & ATMs More visible Dolice Datrolling & neighborhood watches I I Income generation ~ Informal housing Economic Violence Income generation Burglary inunsecured homes More & better visible police patrolling & neighborhood watches Sanitary facilities Gender-Related & General Violence Sewers installed & outside toilets phased out Rape on narrow paths to & from 0 Refuse containers provided with smaller lids so bodies cannot be outside sanitary facilities dumped Refuse containers used to dump ._________---------------------------------------- Communal sanitary facilities supervised corpses Schools School Violence Schools should be declared gun-free zones Theft o f property Schools should be protected against theft & keep out guns by Vandalism and gangsterism installing better fencing, metal detectors & guard dogs Physical violence Guarded schools could then double as safe off-street playgrounds Possession o f drugs & weapons after hours ~ Roads & Road & Transport Violence Stations declared gun-free zones (metal detectors & lockers) transport Taxi-violence-related deaths 8~injuries Jobs & services must be brought closer to residents to reduce Robbery ._________------ excessive transport needs ....................................... Assault Trains need to be accompanied by police Sexual harassment 8~assault by drivers Conflict management / development programs for taxi-violence Source: Feasibility Study' (2002). 77 Box 5.I:Community Police Stations in Mumbai s slum. The police inMumbai, India, have started a partnership with community organizations to provide police services inthe `slums', areas that traditionally had received few police services but where they were muchneeded. By September 2004,65 slums inMumbai had `slum policepanchayats', each made up o f ten representatives from the s l u m (seven women, three men) and a local police officer. Each community representative i s a police sayayak (helper), and wears a photo-badge authorized by the Police Commissioner, but they are appointed by residents' associations, not by the police. The community also makes available a room in each slum, where the police are based and which also serves as an office for the policepanchyar. These panchyats are responsible for policing in their area. They establish a permanent partnership between the police and the s l u m residents. Residents get to know their local police constables, and this also ensures more police accountability to the local population. Inturn, the police know that they have partners working in each slum. The decision to have a majority of women on eachpolice panchyar is inrecognition o f the fact that women are disproportionately the victims o f crime, and often face problems o f domestic violence. In addition, there are strong savings and credit groups formed by women slum and pavement dwellers and these support the policepanchyars. Eachpolicepanchyat is opened with a public celebration. The community volunteers help patrol the settlement to maintain law and order. They also seek to resolve disputes before they escalate into violence or other crimes. Slum inhabitants can bring disputes to the police panchyat, which meets every day, and under the auspices o f the local police many complaints and conflicts are resolved. The police panchyats have proved to be able to resolve many issues such as domestic quarrels and disputes between neighbors over plot or house boundaries. They also help to prevent disputes from escalating into violence or problems o f public order. For slum inhabitants, they are also much quicker and easier to use than going to the police to lodge a formal complaint. They also feel more confident about making complaints, as they know some of the community volunteers and the police. This also frees up police time to allow them to concentrate on serious crimes. The volunteers are clear that they do not have police powers and that dispute resolutions are indertaken by the group as a committee, with details o f all cases discussed being carefully recorded. In some slums, the police volunteers have also brought pressure on local people who are illegally brewing and selling alcohol to close down, as a way o f reducing drunkenness and the violence to which it often contributes. The local savings and credit groups have also supported the people who previously made illegal alcohol to develop new livelihoods or have rewarded them with new houses. Without the community volunteers, the police would find it almost impossible to control this. One o f the key characteristics o f the police panchyats i s that they can be implemented o n a very large scale without additional resources from the government. They first started because the Police Commissioner promoted the idea, but they are sustained in each locality because they meet the needs o f the police: they get free helpers, a safe location within each slum from which to work, and a system that resolves many small disputes without their involvement. They are also rooted in local representative organizations in slums, where the stressors that contribute to violence and crime are obvious - the overcrowded poor quality homes and the lack o f infrastructure (e.g. water and sanitation) and civic amenities (schools, open space). Source: Environment and Urbanization, V o l 16, No 2, October, 2004. 78 Box 5.2: Crime and ViolencePrevention Componentsin Bank-financed Integrated Slum Upgrading Operations: The World Bank is operationalizing local crime and violence prevention in some o f its urban operations through the development o f specific project components inthe LAC region. These components do not intend to solve the country's or city's overall crime and violence problems, rather they seek to take advantage o f the infrastructure and social investments taking place and mainstream prevention at the local level into the overall projects. The components focus specifically on the reduction o f the very high levels o f homicide, youth violence, and associated risk factors in the participating neighborhoods. The components adopt a municipal/ urban renewal approach with a preventive, multi-sectoral, and local strategy through activities that are complementary to, coordinated with, but go beyond traditional police responses. Particularly important are the synergies between the infra-structure, upgrading, and the `situational prevention' - and the community-based `social prevention' activities. The overall objective is a comprehensive intervention at the neighborhood level that is also closely coordinated with other relevant municipal, government, and non-governmental programs addressing crime and violence and their associated risk factors in these neighborhoods. The components generally have six subcomponents: -Diagnostics: Crime and violence mapping o f the micro areas using police statistics and where possible using GIS systems; victimization section in the baseline surveys, willingness-to-pay for increased safety, and; community based and situational diagnostics. -Situational prevention: measures that reduce opportunities for particular crime and violence problems through spatial interventions such as Crime Prevention Through Environmental Design (CPTED) methodology and urban renewal. This method is mainstreamed in the infrastructure works o f the projects through the training o f the archltects, engineers and other technical staff. This methodology i s very new in Latin America but has been successfully piloted in Chile and in the Bank-financed PROMETROPOLE project inPemambuco, Brazil. -Capacity-building, training and technical assistance in multi-sectoral crime and violenceprevention to the participating government agencies, municipalities, CBOs.. -Complementary investments and activities: A fund for complementary investments and activities: the neighborhood residents work with the partnei agencies and the technical staff o f the project to develop a plan for C&V prevention and use these funds to implement the prioritized subprojects and community programs that are not already covered by one o f the other project programs or partner programs. The menu for these may include: social infrastructure investments - such as recreation centers, community facilities, and public lighting not financed through the infrastructure works components, and; Social prevention activities - such as life skills, job skills, and parenting skills training, conflict resolution training, homework clubs, sports and arts-based recreational activities, victim support, and domestic violence prevention. -Community Organizers: The role o f these technical experts in community organization and crime and violence prevention is - at the neighborhood level - to: carry out community-based diagnostics; formulate participatory community safety plans and strategies; liaise and coordinate with other relevant agencies and associations, in particular with Community Safety Councils and the Police; coordinate closely with the works to ensure integration o f CPTED principles; identify and work with youth at -risk in the community; organize and mobilize the community around the concept o f safety though community campaigns (e.g. community clean-up\painting days, community safety festival); initiate additional projects such as summer camp for at-risk youth, etc. -Monitoring and Evaluation component: Evaluations o f the components have been designed and will be carried out Whilst we have as yet no data from these Bank-financed projects, a few similar community- based integrated interventions have yielded dramatic results. The `Fica Vivo' program in the Morro das Pedras favela inBelo Horizonte. Brazil resulted ina 45 Dercent reduction inhomicide rates inits f r s t vear. 79 REFERENCES Absar, S. S. 2001. "Problems Surroundings Wages: the Ready Made Garment Sector inBangladesh." Labour and Management in Development Journal 2:2-17. AsianDevelopment Bank (ADB).2001a. Key Indicators 2001: Growth and Change in Asia and the Pacijic. Asian Development Bank:Manila. .2001b. "Women inBangladesh". Asian Development Bank: Manila. Afsar, R. 2000. Rural Urban Migration in Bangladesh. Causes, Consequences and Challenges. Dhaka: University Press Limited. .2003. "Internal Migration and the Development Nexus, the Case of Bangladesh." Paper presented at the Regional Conference on Migration, Development and Pro-poor Policy Choices, Dhaka, June 22-24. Ahmed, A., F.Naher, and J.Garrett. 2003. Crime and Violence in Dinajpur. Evidence from a SHAHAR Survey. Final report submitted to CARE-Bangladesh. Ahmed, I. (unknown) "Violence, Mastanocracy and Chronic Poverty." (mimeo.) Universityo fDhaka Department o fInternationalRelations: Dhaka. Ahmed, S. and Z. Sattar. 2004. "Trade Liberalization, Growthand Poverty Reduction: The Case o f Bangladesh." World Bank Report IDP-190. Washington, D.C. Amin, S. et al. 1997. "Transition to AdulthoodofFemaleWorkers: Some Evidence from Bangladesh." PopulationCouncil Policy Research Division Working Paper 102. New York. Amin, S. et al. 2003. "Does MicrocreditReachthe Poor andVulnerable?Evidence fromNorthern Bangladesh."Journal oJDevelopment Economics 70(1):59-82. Amin, S. et al. 2004. "Poverty and Other Determinants ofChildLabor in Bangladesh." Southern Economic Journal 70(4): 876-892. Arndt, C. et al. 2002. "Opportunities and Challenges inAgriculture and Garments: A General EquilibriumAnalysis o f the Bangladesh Economy." International FoodPolicy Research Institute (IFPRI) Discussion Paper 107, Bangladesh and the World Trade Organization (WT0)Project. Washington, D.C. Barkat, A. and S Akhter. 1999. "Human Deprivation inthe Urban Slums and Squatter Settlements inBangladesh:A Rapid Survey."(mimeo.) Dhaka. 80 Bangladesh Bureau o f Statistics (BBS) Labor Force Survey. 1991. Report on Labor Force Survey: 1990-91. Bangladesh Bureauo f Statistics: Dhaka. .1996. Report on Labor Force Survey: 1995-96. Bangladesh Bureau o f Statistics: Dhaka. .2000(a).Report on Labor Force Survey: 1999-2000. Key Data. Bangladesh Bureau o f Statistics: Dhaka. .2000(b). Labor Force Survey 2000. Data. Dhaka: BBS. .2001. Population Census 2001, Preliminary Report. Bangladesh Bureau o f Statistics: Dhaka. .2001. Household Income and Expenditure Data, 2001Bangladesh Bureau o f Statistics: Dhaka Berger, Louis Group, Inc. "Strategic Transport Plan for Dhaka," working paper #3: Population projections to 2024 in STP area. Bangladesh Consultants Ltd, June 2004. Bhuyan, A. R. et al. 2001. "Rural UrbanMigration andPoverty: The Case for Reverse Migration inBangladesh." Centre on Integrated RuralDevelopment for Asia and the Pacific (CIRDAP) Focus Study 10.Dhaka. Buvinic, M.,A, Morrison and M.Shifter. 1999. "Violence inthe Americas: A Framework for Action." InA. Morrison, eds. Too Close to Home: Domestic Violence in theAmericas. Washington, D.C.: Inter-American DevelopmentBank (IDB) and John Hopkins UniversityPress. Buvinic, M.and A. Morisson. 1999. ViolencePrevention - Technical Notes 1to 6. Inter- American Development Bank (IDB) Sustainable DevelopmentDepartment:Washington, D.C. Caldwell, B. et al. 2002. "Why D o the Children o f the Poor Die inDhaka, Bangladesh?"Population Research and Policy Review 21(3):159-178. Cameron, J. 1996. "The Challenge of Combining Quantitative Methods inLabour Force and Livelihoods Analysis. A Case-Study o f Bangladesh."Journal of International Development 8(5): 625-653. Centre for Urban Studies. 2001. "Report o fthe Task Force on Urban Governance." National Policy Forum:Dhaka. Chaudhury, N.and J.S. Hammer. 2003. "Ghost Doctors: Absenteeism in BangladeshiHealth Facilities." World Bank Policy ResearchWorking Paper 3065. 81 Washington, D.C. Center for Housing Rights and Evictions (COHRE) and ACHR Asian Coalition of Housing Rights (ACHR). 2000. "Forced Evictions inBangladesh", Mission report of the Centreon Housing Rights andEvictions, Switzerland andthe Asian Coalitiono fHousing Rights, Bangkok. Deaton, A.1997. TheAnalysis of Household Surveys - A Microeconometric Approach to Development Policy. Baltimore: John Hopkins UniversityPress. Deichmann, U.et al. 2003. "Information-Based Instruments for Improved Urban Management." World Bank Policy Research Working Paper 3002. Washington, D.C. Delap, E. 2001. "Economic and Cultural Forces inthe Child Labour Debate. Evidence from UrbanBangladesh." Journal of Development Studies 37(4): 1-22. Dowlah, C.A.F. 1999. "The Futureof the ReadymadeClothingIndustryof Bangladesh inthe Post-UruguayRoundWorld." WorldEconomy 22(7):933-953. Dubois, J.L. 1997. "Monitoring UrbanPoverty at the BangladeshBureauo f Statistics: Design and Implementation Issues for an Adaptive 1-2-3 Survey." Development Institutions et Analyses de Long Terme (DIAL) Document de Travail DT/97/13. Paris. Elson, D. 1999. "Labor Markets as GenderedInstitutions: Equality, Efficiency and EmpowermentIssues." WorldDevelopment 27(3):621-627. Evans, T. G.,et al. 1999. "Demystifying Nonparticipation inMicrocredit: A Population-BasedAnalysis." WorldDevelopment 27(2):419-430. Fouquin,M.et al. 2004. "Mondialisation et regionalisation: le cas des industriesdu textile et de 1'habillement.Tentre d'Etudes Prospectives et d'Informations Internationales (CEPII) Document de Travail 2002-08. Paris. Garrett, J. and A. Akhter. 2004. "Incorporating Crime inHousehold Surveys: A Research Note." Environment and Urbanization 16(2):139-151. Gilbert, J., and T. Wahl. 2000. "Applied General EquilibriumAssessments ofTtade Liberalization inChina." Paperpresentedat the Workshop on China's Accession to the WTO. CPB NetherlandsBureaufor Economic Policykalysis and Institute of Qualitative and Technical Economies ofthe Chinese Academy of Social Science, Beijing, October 19-20. Greenwood, P. et al. 1996.Diverting Childrenfrom a Life of Crime. Measuring Costs and Benefits. Santa Monica: RandCorporation. 82 Guerra, N.2005. Youth Crime Prevention. World Bank Community Based Crime and Violence Prevention inUrbanLatin America and the Caribbean Program. Washington. D.C. Harpham,T. and M.Tanner eds.1995. UrbanHealth in Developing Countries. London: Earthscan. Harris, J. R. and M.P. Todaro. 1970. "Migration, Unemployment and Development: A Two-Sector Analysis." American Economic Review 60( 1):126-42. Hefferson, Robert and Salma Shafi, "Strategic Transport Plan for Dhaka", working paper #2: Existing land use and future scenarios, Robert Hefferson and Salma Shafi, July 2004. Hoek-Smith, Maria. 1998. "Housing Finance inBangladesh, Improving access to HousingFinance by Middle and Lower Income Groups." Report prepared for the Governmentof Bangladesh, Ministry o f Local Government, Rural Development and Co- Operatives and UNDP/UNCHS (Habitat). Hossain, M.Z. 2001. "Rural-Urban Migration inBangladesh: A Micro-Level Study." Paper presented at the International Union for the Scientific Study o f Population (USSP) General Conference, Salvador, August 20-24. Huq-Hussain, S. 1996."Female Migrants inan Urban Setting-the Dimensions of SpatialPhysical Integration - the Case o fDhaka." Habitat International 20( 1):93- 107. Hussain, Syed Minuddin. 2004."Bangladesh LandLaw Study."(mimeo.) International Labor Organization (ILO) Employment Sector. 2002. Women and Men in the Informal Economy: A Statistical Picture. International Labor Organization: Geneva. Islam, M.H.,Chowdhury, Q.A. 2003. "Slums inDhaka City: a Situation Analysis o f Four Selected Slums of Dhaka City."(mimeo.)Aparajeyo Bangladesh: Dhaka. Islam, N. 1985. "The Poor's Access to Residential Space in an Unfairly Structured City, Dhaka." Oriental Geographer 29-30 (1&2):37-46. . 1998. Human Settlements and Urban Development in Bangladesh. Dhaka: Universityo fDhaka. . 1999. Urbanization in Bangladesh and the Growth of Dhaka: Land Use,Poverty and Governance. Maulana Abul Kalam h a d Institute o f Asian Studies: Calcutta. . 2004. "The Poor's Access to Land and Housing inDhaka." Paper written for the World Bank. Dhaka Urban Poverty Study. 83 Islam, N.and A. UlahAhsan, eds. 1996. Urban Bangladesh - Geographical Studies. University of Dhaka: Department of Geography, Urban Studies Programme.Dakha. Islam, N.ed. 2000. Urban Land Tenureand housingfinance in Bangladesh. UN-Habitat, Center for UrbanStudies. Dhaka. Islam, N.and S.A. Shafi. 2004. "Solid Waste Management and the Urban Poor in Dhaka." Forum on UrbanInfrastructure and Public Service Delivery for the UrbanPoor Regional Focus: Asia, New Delhi, June 25-26. Islam, N.,N.Huda, F. Narayan, and P. Rana, eds. 1997. Addressing the Urban Poverty Agenda in Bangladesh, Critical Issues and the 1995Survey Findings. Dhaka: University Press. Islam, R. 2003. "Labour Market Policies, Economic Growth and Poverty Reduction: Lessons and Non-Lessons from the Comparative Experience o f East, South-East and South Asia." International Labor Organization Recovery and ReconstructionDepartment, Issues in Employment and Poverty Discussion Paper 8. Geneva. Kabeer, N. 1991. Tultural Dopes or Rational Fools? Women and Labour Supply inthe BangladeshGarment Industry."European Journal of Development Research 3(1):133- 160. Khan,A.R. 2001. "Employment Policies for Poverty Reduction." International Labor Organization Recovery and ReconstructionDepartment, Issues inEmployment and Poverty Discussion Paper 1.Geneva. Khan, H.R.and Q.I. 2000. "Urban Water ManagementProblems inDeveloping Siddique Countries with Particular Reference to Bangladesh." Water Resources Development 16(1): 21-33. Khandker, S. 2003. "Microfinance andPoverty: EvidenceUsingPanelData from Bangladesh." World Bank Policy Research Working Paper 2945. Washington, D.C. Kibria, N.2001. "Becoming a Garment Worker: The Mobilization o f Women into the GarmentFactories ofBangladesh." InR. Sobhan, ed. Globalisation and Gender: ChangingPatterns of Women's Employment in Bangladesh. Dhaka: University Press Limited. h a u l , F M.and M.A. Ramirez. 2002. TheImpact of ChildAbuse and Family Violence on the Probability of Future Family Violence, Educational Attainment, and Labour Outcomes: TheMexican Case. Mexico, D.F. Knowles, J.C. 2001. "Bangladesh: Public Expenditure Review, Education Sector." (mimeo.) Backgroundpaper for the World Bank. 84 KPMG. 2002. "Dhaka EasternBypass Concept Paper."(mimeo.) Report for the KPMG Infrastructure Investment Facilitation Center (IIFC). Kruger, T., K.Landmanand S. Liebemann.2001. Designing Safer Places: A Manualfor CrimePrevention Through Planning and Design. CSIR: Pretoria Liebemann, S. andJ. Coulson. 2004 "Participatory Mappingfor Crime Preventionin South Africa - Local Solutions to Local Problems." Environment and Urbanization 16(2):125-134. Lips, M.et al. 2003. "Textile andWearing Apparel Sector Liberalization - Consequences for the BangladeshEconomy." Paperpresented at the 6th Conferenceon Global Economic, The Hague, June 12-14. Majumder, P. et al. 1996. Thesquatters ofDhaka City: Dynamism in the Life of Agargaon Squatters. Dhaka: UniversityPress Limited. Mannan, F and Rashid, SF (2004), "The Heterogeneity of the Urban Poor: Political- Economy and Social Conditions inUrbanSlums," Univeristyof Southampton, UK. Fundedby DFID (Project Number:KARProject, R8028). Mahmud,S. 1997. "Women's Work inUrbanBangladesh: Is There an Economic Rationale?" Development and Change 28(3):235-260. Matin, I. D. Hulme.2003. "Progrkns for the Poorest: Learning from the IGVGD and Program inBangladesh." World Development 31(3):647-665. Mehra R. and S. Gammage. 1999. "Trends, Countertrends, and Gaps inWomen's Employment." World Development 27(3):533-550. Morrison, W. 2001. Y3ummaries o f Studies of the Impact o f WTO on China's Trade". Organization of Economic Cooperation andDevelopment (OECD) Center for Cooperation with Non-Members. Paris. Moser, C. and C. McIlwaine. 2000. Urban Poor Perceptions of Violence and Exclusion in Colombia, World Bank. Washington, D.C. Moser, C. et al. 2005. "Violence, Fear, and Insecurity Among the UrbanPoor inLatin America." InM.Fay ed., The Urban Poor in Latin America. World Bank. Washington, D.C. Nabi, R. et al. 1999. "Consultation with the Poor. Participation Poverty Assessment in Bangladesh." Bangladesh: NGO Working Group on the World Bank. Osmani, S.R. 2003. "Exploring the Employment Nexus: Topics inEmploymentand Poverty." Report preparedfor the Task Force on the Joint ILO-UNDP Programme on 85 Employment and Poverty. Oxfam. 2004. "How Rich-Country Protectionism inTextiles and ClothingTrade Prevents Poverty Alleviation." Oxfam Briefing Paper'60. Oxford. Paul-Majumder, P. and A. Begum. 2000. "The Gender Imbalancesinthe Export Oriented GarmentIndustryinBangladesh." World BankPolicyResearchReport on Gender and Development, Working Paper Series 12. Washington, D.C. Rahman, M.et al. 1999. "Child Labor inBangladesh: A Critical AppraisalofHarkin's Billandthe MOU-Type SchoolingProgram." Journal of Economic Issues 33(4):985- 1003. Rahman, M.andD.Bhattacharya.2000. "Bangladesh Experiencewith Trade and InvestmentLiberalisation. A Perspectiveon Poverty Alleviating Implications." In Liberalisation and Poverty: Is Therea Virtuous Circle. India: Consumer Unity and Trust Society (CUTS). Rahman, M. M. 2002. "Problems of the NGOs in Housing Urban Poor in Bangladesh.'' Habitat International 26(3):433-45 1. RahmanR. Iand K.M.Nabiul Islam. 2003. "Employment Poverty Linkages: Bangladesh." International Labor Organization, Issues inEmployment and Poverty Discussion Paper10, Geneva. Rashid and Mannan, 2004, "The Heterogeneity o f the Urban Poor: Political-Economy and Social Conditions in Urban Slums" and Aparajeyo, 2002 "Slum Observatory Survey." Ravallion, M.and Q.Wodon. 2000 "Does Child Labour Displace Schooling? Evidence on Behavioural Responses to an Enrollment Subsidy." TheEconomic Journal 11O(502): 158-175. Rau,M.et al. 2003. Espacios UrbanosSeguros: Recomendaciones de disenoy gestion comunitariapara la obtencidn de espacios urbanos seguros. Santiago de Chile: Ministerio de Vivienda y Urbanismo. Remenyi, J. 2000. "Poverty Reduction andUrbanRenewalThrough UrbanAgriculture and Microfinance: A Case Studyo fDhaka, Bangladesh." (mimeo.) GeelongVictoria: Deakin University, School o f Australian and International Studies. Roy, A.N., Jockin, A., Javed, A. 2004. "Community Police Stations inMumbai's Slums." Environment and Urbanization 16(2):135-138. Salmon, C. 2002. "Performance ofthe Bangladeshi Labor*Marketduring the Nineties." 86 Background paper for the Asian Development Bank and the World Bank. Salway, S. et al. 1998. "Livelihoods and Health Among the UrbanPoor of Dhaka. Descriptive Findings from the UrbanLivelihoods Study. ULS Study Report. Institute for Development Policy Analysis and Advocacy, Dhaka. Salway, S. et al. 2003. "A Profile of Women's Work Participation Among the Urban Poor o fDhaka." WorldDevelopment 31(5):881-901. Sarker, N.C., Huq, N.A. 2004. "Dhaka City: RoadDigging andPublic Sufferings." Oriental Geographer 48(2) Sen, B.2003. "Drivers of Escape and Descent: Changing HouseholdFortunesinRural Bangladesh." WorldDevelopment 31(3):5 13-534. Seraj, T.M. and S. Afrin, 2003. "Solving Housing Problem Through Private Sector Real Estate Development inDhaka." InN.Islam, ed. Water and Sanitationfor Cities, World Habitat Day 2003 Souvenir, Dhaka: Centre for UrbanStudies (CUS), Bangladesh Institute of Planners (BIP). SethuramanS.V. 1998. "Gender, Informalityand Poverty: A GlobalReview." International Labor Organization Working Paper. Geneva. Shafaeddin, S.M.2004. "Is China's Accession to WTO Threatening Exports of Developing Countries ?" China Economic Review 15(2):109-144. Shafi, S. A.1999. "Study on Financing o f Urban Low-Income Housing." Report for the AsianDevelopment Bank. Sheltech ConsultantsLtd. Shaw, M.2000. TheRole ofLocaI Government in Community Safety. International Centre for the Prevention o f Crime: Montreal. Shefali, M.K.2002. "Impact o f International Trade Regime on Female Garment Workers inBangladesh." Paperpresented at the InternationalWorkshop on "Globalization, Trade Liberalization and Economic GrowthinAsia: Should Labour ad Environmental Standards be Part ofthe Equation?The Case o f Bangladesh." University o fNew England, Australia: October 3-4. Siddiqui, K. ed. 1994.Local Government in Bangladesh. 2nd edition, Dhaka:University Press Limited. .1997.Land Management in South Asia. A Comparative Study. Dhaka:University Press Limited. Siddiqui, K.et al. 2000. Overcoming the GovernanceCrisis in Dhaka City. Dhaka: UniversityPress Limited. Singha D.2001. "Social Intermediation for the UrbanPoor inBangladesh" Paper 87 presented at Department for International Development (DFID)Regional Livelihoods Workshop: Reachingthe Poor inAsia, May 8-10. Shefali, M.K.andN.U.Kendra. (undated) "Women and shelter development: perspectives andreality inBangladesh context." Paper presented at the UnitedNations Foundation and the Quest Ltd Seminar, Dhaka, December 6-7. Sobhan, R.2002. "Bangladesh inthe Next Millennium:Between Promise and Fulfillment."Centre for Policy Dialogue: Dhaka. Ullah,A.K.M.A. 2004. "Bright City LightsandSlums ofDhakaCity: Determinantsof Rural-Urban Migration inBangladesh." Migration Letter I( 1):26-41. UnitedNations (UN).2001. World UrbanizationProspects: The2001Revision. United Nations. New York. UnitedNationsDepartment o fEconomic and Social Affairs (UNDESA).(undated). Commodity TradeStatistics Database. United Nations Department o f Economic and Social Affairs. New York. UnitedNationsDevelopment Programme(UNDP). 2002.Human Security in Bangladesh. In Search of Justice and Dignity. UnitedNations Development Programme Bangladesh. Dhaka. .2004. Human Development Report. United Nations Development Program. New York. UnitedNationsFamily PlanningAssociation (UNFPLA). 2000. State of the World's Population. United Nations PopulationFund.New York. UnitedNations Habitat (UN-Habitat). 2002. Safer CitiesApproach. UnitedNations Habitat. New York. Walkenhorst, P. 2003. "Liberalising Trade inTextiles and Clothing: A Survey o f Quantitative Studies." Working Paper of the Trade Committee (2003)/2 Organization for Economic Co-Operation and Development: Paris. Ward, K.,et al. 2004. "The Effects o fGlobalEconomics Restructuring on Urban Women's Work and Income-Generating Strategies inDhaka, Bangladesh." Critical Sociology 30(1):63-102. Werner International. 2003. 2002Labor Cost Comparison Report. World Health Organization (WHO). 2002. WorldReport on Violence and Health. World Health Organization. Geneva. 88 Wodon, Q.,T. 1997. "Food Energy Intake and Cost of Basic Needs: Measuring Poverty inBangladesh."Journal of Development Studies 34(2):65-101. .1999. "Micro-determinants of Consumption, Poverty Growth, and Inequality in Bangladesh." World BankPolicy ResearchWorking Paper 2076. Washington, D.C. World Bank. 1996. "Bangladesh: Labor Market Policies for Higher Employment." World Bank SouthAsia Country Department 1, Country OperationsDivision, Report 13799- BD.Washington, D.C. . 1998(a). "Bangladesh, FromCounting the Poor to Makingthe POOFCount." World Bank Report 17534-BD. Washington, D.C. . 1998(b). Bangladesh 2020.A Long-run Perspective Study. Bangladesh Development Series, World Bank and BangladeshCentre for Advance Studies, Dhaka: UniversityPress Limited. .1999(b). Bangladesh. Key Challengesfor theNextMillennium. WorldBank: Washington, D.C. .2002. "Poverty inBangladesh: Building on Progress." World Bank Report 24299-BD. Washington, D.C. .2003.(a) "Improving the Investment Climate in Bangladesh World Bank and " BangladeshEnterprise Institute. .2003(b). A Resource Guidefor Municipalities: Community-Based Crime and ,Violence Prevention in Urban Latin America. World Bank, Washington D.C. .2004(a) Housing Finance Reforms in Bangladesh. SouthAsia Finance and Private Sector, World Bank. Washington, D.C. 2004(b). World Development Report 2005, World Bank: Washington, D.C. Zaman, H.2001. "Paid Work and Socio-Political Consciousness o f GarmentWorkers in Bangladesh." Journal of Contemporary Asia 3l(2): 145-160. Zaman, H.2004. "The Scaling-up o f Microfinance inBangladesh:Determinants, Impact, andLessons." World Bank Policy Research Working Paper 3398. Washington, D.C. 89 Annexes 90 Annex 1: PovertyProfileBackgroundTables and Figures Figure Al.1: CumulativeDistributionof Per Capita Expenditure,Dhaka 0 I I I , 1 500 1000 1500 2000 2500 Monthly PCE Cumulative distribution of PCE-lower poverty line upper poverty line FigureAl.2: Povertyand ExtremePovertyHeadcount,UrbanBangladesh 60 1 $ 50 c , I 0 Dhaka 1 mchittagong, i - 50 6 1 Q .E 40 .-=> D c 30 0 Extreme (food) Poverty Poverty 91 FigureA1.3: Access to Water for Dhaka's Poor I % households below the poverty line 80 70 60 50 40 30 20 10 0 I Piped water Tube wells FigureA1.4: Housingqualityand povertyincidence Substandard housing Upperpoverty line 0 5000 10000 15000 Monthly PCE other housing poor quality housing upper povertyline 92 FigureA1.5: Kernal Densityof Expendituresin UrbanAreas -1 I ,. 6 8 10 12 lngcexp Chittagong .................Khulna % households below poverty (upper) line BricWcement Hemp/Hay Material of Wall 25.95 39.46 Material of Roof 7.03 4.32 Table A1.2: Inequalityof distributionof consumption expenditureby quintile Ouintiles I % o f consumotion exoenditure accruing to this auintile Dhaka Pune Bangalore (India) (India) Poorest 20% 8.35 7.8 8.59 2 11.55 9.79 13.25 3 15.80 16.57 12.08 4 22.25 22.44 27.91 Richest 20% 42.61 43.4 38.17 93 Annex 2: SupplementalData on Labor TableA2.1: Populationaged 15 years and over and working populationin the Dhakadistrict. in thousands. 1996-2000 Dhaka District Usual % Extended Yochange definition change definition 1996 2000 1996 2000 Population, aged 15 years and over n.a. 5612 n.a. 5612 Population, aged 15 years and over, male n.a. 2848 n.a. 2848 Population, aged 15 years and over, female n.a. 2764 n.a. 2764 Active population, 15 years and over 2632 3108 18% 2891 3318 15% Active population, 15 years and over, male 2079 2365 14% 2089 2404 15% Active population, 15 years and over ,female 553 742 34% 802 913 14% Sexual distribution Male 79% 76% 72% 72% Female 21% 24% 28% 28% Labor force participation* Total 55% - 59% Male 83% - 84% Female 27% - 33% Bangladesh Population, aged 15 years and over n.a. 74203 n.a. 74203 Population, aged 15 years and over, male n.a. 38325 n.a. 38325 Population, aged 15 years and over, female n.a. 35877 n.a. 35877 Active pop. 15 years and over 36054 40734 13% 50337 53519 6% Active pop. 15 years and over, male 30666 32175 5% 31268 33458 7% Active pop. 15 years and over ,female 5388 8558 59% 19069 20061 5% Sexual distribution Male 85% 79% 62% 62% Female 15% 21% 38% 38% Labor force participation Total 55% - 72% Male 84% - 87% Female 24% - 56% Source: different LFS issues. Notes: *: Numbero f economically active persons to the population o f 10years and over. 94 Table A2.2: Main characteristics of Dhaka's labor force compared with the overall country. 2000 Bangladesh Dhaka Bangladesh Dhaka Descriptionof the labor force (thousands) Sectoraldistributionof the employedpop. Total population, 10 years and over 91643 6442 Agriculture 62.3 11.8 Total population, + 10 years, Male 47640 3249 Manufacturing 7.4 17.1 Total population, + 10 years, Female 44003 3193 Trade 12 23.2 Transport 4.6 9.3 Active pop., + 10 years 60297 3509 Community services 7.4 21.7 Active pop., + 10 years, Male, 37487 2460 Household services 2.6 9.9 Active pop., + 10 years, Female 22809 1049 Other sectors 3.6 7.0 Employedpop., + 10 years 58071 2770 Agriculture 62.3 11.8 Employed pop., + 10 years, Male 36134 2291 Industry 10.3 20.6 Employedpop., + 10years, Female 21937 959 Services 27.4 67.6 Labor force participation rate (total) 65.8% 54.5% Formal public sector 4 14.2 Labor force participation rate (male) 78.7% 75.7% Formal private sector 13.9 36.9 Labor force participation rate (female) 51.8% 32.9% Informal sector 81.7 48.7 Non-profit instit. sect. 0.3 0.1 Distributionof the activepopulationby gender Status of employedpopulation Male 62.0% 70.1% Self-employer 32.4 33.5 Female 38.0% 29.9% Employer 0.2 0.3 Unemploymentand underemployment Employee 13,3 48,8 Unemployment rate (standard) 3.7% 7.4% Unpaid family worker 36.7 12.9 Proposed unemployment rate 11.0% 10.4% day laborer 17.6 4.6 Underemployment rate 35.3% 16.0% Educationlevel of the employedpop.(+ 15) Un-& underemploymentrate 39.2% 22.2% No education 49.4 30.9 Class I-V 24.8 17.6 Class VI-X 16.1 21.9 SSCiHSC 6.5 15.4 Degree & above 3.2 14.2 Source: 2000 LFS Notes:extended d e f ~ t i o on fthe labor force inDhaka SMA. Population aged 10years and over. Table A2.3a: Sectoral distribution of employment by gender Sectoral Sectoral Sectoral % o f women in distribution o f distribution o f distribution o f the sector male employment female employment employment (both sexes) Agriculture 8.9% 19.0% 11.9% 47.3% Manufacturing 13.6% 25.5% 17.1 % 44% Construction 3.6% 0.6% 2.7 Yo 6.8% Trade 32%% 2.4% 23.2 % 3% Transport 12.7% 1.O% 9.2 Yo 3.2% 95 Finance 4.6% 1.O% 3.5 % 8.4% community services 20.3% 25.0% 21.7 % 34% Householdservices 3.8% 24.3% 9.9 % 72.7% Source: 2000 LFS Notes:extendeddefinitiono fthe labor force inDhaka SMA. Population aged 10years and over. Table A2.3b: Some key figures on female employmentin developingcountries Women's economic activity rates % o f Yo o f women in women in the manuf. the services Sector South Asia 43% Pakistan: 36%, India: 42%, Sri Lanka: 43%, 41% 15% Nepal: 56%, Bangladesh: 66% East Asia and South East 57% Malaysia: 48%, Philippines: 49%, 44% 41% Asia Indonesia: 56%,China: 72%,Thailand: 73%, LatinAmerica and 42% Brazil: 43%, Venezuela: 43%, 36% 48% Caribbean Sub Saharan Afsica 62% Cote d'Ivoire: 44%, Niger: 69%, 28% 36% Mozambique: 82%, Date and source 2002 ** 2002 ** 1990's * 1990's * Source: * Mehra and Gammage (1999). **UNDP (2004) Table A2.4: Labor force in the main cities of Bangladeshin 2000 Bangladesh Dhaka Chittagong Khulna Rajshahi Usual definition Labor force participation rate (total) 49.2% 51.6% 50.2% 47.4% 51.6% Labor force participation rate (male) 73.5% 74.3% 73.3% 70.7% 73.0% Labor force participation rate (female) 22.8% 28.3% 25.5% 23.0% 27.6% Distribution o f the active population Male 78.0% 72.9% 75.4% 76.2% 74.7% Female 22.0% 27.1% 24.6% 23.8% 25.3% Extended definition Labor force participation rate (total) 65.8% 54.5% 56.1% 52.3% 57.7% Labor force participation rate (male) 78.7% 75.8% 75.1% 71.5% 74.2% Labor force participation rate (female) 51.8% 32.9% 35.7% 32.3% 39.3% Distribution o f the active population Male 62.0% 70.2% 69.2% 69.8% 67.8% Female 38.0% 29.9% 30.8% 30.2% 32.2% Source: Data computed from 2000 LFS. Notes:StatisticalMetropolitanAreas. Population aged 10years and over. 96 Table A2.5: Sectoralcomposition of the employed labor force inthe 4 main cities DMA Cluttagong Khulna Rashahi Agriculture 11.8 21.8 10.0 27.2 Manufacturing 17.1 24.0 20.1 10.3 Garment 11.9 11.4 7.4 6.2 Trade 23.2 18.8 20.4 21.0 Transport 9.3 13.2 11.6 11.5 Community services 21.7 7.9 15.3 18.5 Householdservices 9.9 7.8 9.7 3.6 Other sectors 7.0 6.4 12.9 8.1 Total 100 100 100 100 Agriculture 12% 22% 10% 27% Industry 21% 29% 30% 17% Services 67% 49% 60% 56% Total 100 100 100 100 Note: extendeddefinition of the employed labor force aged 10 years and over. Source: 2000 LFS Table A2.6: Unemploymentand underemploymentin the 4 main cities of Bangladesh Bangladesh Dhaka Chittagong Khulna Rajshahi Unemployment rate (standard) 3.7% 7.4% 5.0% 3.5% 7.3% "Proposed" unemploymentrate 11.O% 10.40h 11.2O/O 7.0% 8.6% Underemploymentrate 35.3% 16.0% 21.7% 23.3% 27.3% Unemployment and underemploymentrate 39.2% 25.2% 25.6% 26.0% 32.6% Source and notes: Salmon, based upon 2000 LFS. Definitions are the following: Standard unemployment rate = absolutely unemployedhotal labor force; "Proposed" unemployment rate = total unemployedpersons (absolutely unemployed 7 unpaid family workers)/total labor force; Cnderemployment rate = people worhng less than 35h per weekemployed labor force; Unemployment and underemploymentrate = total unemployedpersons 7 underemployedpersons/totallabor force. 97 TableA2.7: Distributionof income and expenditureper capita, hiesversus Ifs (in taka) 2000 LFS incomeper 2000 LFS expenditures 2000 HIES expenditures capita per capita per capita percentile Centile centile centile D1 500 440 647 -Zl: 649- D2 625 583 --Zl: 649- 801 D3 750 666 963 D4 870 1141 -ZU: 893- D5 1000 933 -Zu:893- 1360 D6 1250 1003 1583 D7 1600 1333 1832 D8 2166 1800 2425 D9 3333 2500 3243 Source: author's calculationson the 2000 LFS. Notes:Z1: low poverty line, Zu: upperpoverty line. Table A2.8: Generalcharacteristicsof the householdby income group. Variable T h e poor The precarious The non-poor Mean Std.Dev Mean StdDev Mean StdDev Numberofobservations 454 296 750 Size of the household 5.13 1.73 4.68 I,87 4,40 1,95 Size of the house (acre) 20.80 81.80 37.30 160,60 58,45 172,40 Size of the land (acre) 25.24 186.22 31.30 114,08 58,03 332,57 % ofhouseholdsowning their house 0.28 0.45 0.27 0,45 0,34 0,47 % ofhouseholdswith livestock 0.18 0.38 0.16 0,37 0,lO 0,29 % of householdsreportingno asset 0.81 0.39 0.78 0,41 0,59 0,49 YOofhouseholdheadedby a woman 0.08 0.27 0.05 0,23 0,09 0,29 Age of the mother 32.08 8.36 31.67 8,90 33,52 10,44 Age of the father 40.49 10.55 40.25 I1,17 42,20 11,82 YOof workers inthe household 0.64 0.23 0.65 0,25 0,59 0,26 Number of earners per household 1.56 0.83 1.58 0,80 1,58 0,95 Numberofmalesinthe working age 1.38 0.89 1.44 0,86 1,59 1,OO Ratio number of eamersisize of the household 0.30 0.17 0.34 0,19 0,36 0,22 % of heads of the householdwithout any education 0.50 0.50 0.38 0,49 0,12 0,33 % of heads ofthe householdwith educationlevel 1 i. 0.22 0.41 0.22 0,41 0,11 0,31 % of heads o fthe householdwith educationlevel 2 0.19 0.39 0.22 0,41 0,21 0,40 % ofheads of the householdwith educationlevel 3 0.09 0.29 0.18 0,39 0,57 0,SO % of wives without any education 0.58 0.49 0.46 0,50 0,14 0,35 % of wives with education level 1(class 1-5) 0.16 0.37 0.19 0,39 0,14 0,34 % of wives with education level 2 (class 6-10) 0.12 0.32 0.19 0,39 0,24 0,43 % of wives with educationlevel 3 (HSCiSSC& above) 0.15 0.35 0.17 0,38 0,49 0,50 Source: 2000 LFS 98 Table A2.9: Probit models. estimationsof the probabilityto be poor. marginaleffects 1 2 3 4 Marg.eff Std err. Marg.eff Sid err. Marg.eff Std err. Marg.eff Std err. Size of the household 0.032*** 0.008 0.039*** 0.007 0.036*** 0.009 0.056*** 0.010 Age ofthe head -0.002 0.001 -0.002* 0.001 -0.003 * 0.002 -0.004** 0.002 Yo of boys 5-9 0.016 0.128 0.029 0.129 -0.087 0.159 -0.076 0.176 %ofboys 10-14 0.122 0.I23 0.149 0.167 0.338*** 0.350*** 0.410*** 0.525*** % ofgirls 5-9 -0.024 0.131 0.026 0.129 0.022 0.171 0.143 0.179 %ofgirls 10-14 -0.066 0.121 0.042 0.119 -0.297** 0.151 -0.146 0.163 YOo f women 15-59 0.681*** 0.119 0.117 0.141 0.153 0.554*** 0.898*** 0.800*** %ofwomen over 60 -0.360 0.222 -0.240 0.215 0.270 -0.640** 0.276 0.719*** % of men over 15 0.112 0.109 0.131 0.141 0.807*** 0.627*** 0.994* ** 0.870*** %ofmen un- or underemployed 0.543*** 0.133 , 0.469*** 0.131 0.688*** 0.163 0.655*** 0.175 Head in the formal sector 0.035 -0.06 1* 0.034 0.041 -0.075* 0.044 0.104*** 0.136*** Headday laborer 0.409*** 0.070 0.305*** 0.079 0.403*** 0.049 0.288*** 0.075 Heademployee 0.020 0.038 0.054 0.038 0.005 0.046 0.061 0.049 Head in sectors up 0.044 -0.068 0.069 0.065 -0.046 0.086 0.172*** 0.260*** Head inthe construction sector -0.128** 0.045 -0.088 0.052 -0.181** 0.076 -0.073 0.088 Head inthe trade sector 0.092*** 0.030 -0.065** 0.031 -0.079** 0.039 -0.024 0.043 Head in the agriculture 0.008 0.048 -0.017 0.044 0.017 0.061 -0.058 0.063 Head in the community services -0.014 0.050 0.052 0.058 -0.075 0.058 0.049 0.065 Head in the public sector 0.043 -0.119** 0.046 -0.155** 0.059 -0.075 0.066 0.165*** Head in public community services 0.204** 0.092 0.180** 0.097 0.163* 0.080 0.128 0.091 Head in householdservices 0.039 0.084 0.085 0.084 -0.069 0.099 0.014 0.096 Head in the manufacturing sector 0.006 0.118 -0.029 0.106 -0.013 0.151 -0.065 0.156 Head inthe formal manufac.sector 0.014 0.128 0.088 0.139 0.027 0.160 0.141 0.161 Household without any asset 0.154*** 0.026 0.111*** 0.026 0.276*** 0.031 0.235*** 0.035 Headwith education level 1 (class 1-5) 0.029 0.046 0.082*** 0.136*** Headwith education level 2 (class 6-10) 0.025 0.039 0.160*** 0.320*** Headwith educationlevel 3 (HSC:SSC & 0.025 0.03 1 above) 0.345*** 0.592*** Number of observations 1500 1500 1500 1500 Percent correction predictions Poor (+ precariousfor 3 & 4) 72.7 75.0 65.2 65.1 Non-poor (+ precarious for 1 & 2) 50.0 53.1 55.7 54.5 Notes: * significant at 10% level. ** significant at 5% level. *** significant at 1% level. Dependent variables: Estimations 1 & 2: Probability of being poor (deciles 1 to 3), Estimations 3 & 4: probability of being poor or precarious (deciles I to 5). 99 Table A2.10: Prevalence of poverty by sector, both sexes (in percentage) % o fpoor ?4o fprecarious % o fpoor & precarious Agriculture 42.7 22.1 64.8 Mining 33.3 11.1 44.4 Manufacturing 33.1 24.6 57.7 Garment 35.5 23.6 59.1 Energy 9.1 18.2 27.3 Construction 37.8 24.3 62.2 Trade 25.9 21.5 47.4 Transport 40.0 26.4 66.4 Finance 8.4 9.5 17.9 Community services 27.6 14.6 42.2 Household services 27.0 12.0 39.0 Householdservices (without live-in domestics)* 45.5 19.8 66.3 Public sector 21.1 16.9 38.1 Private sector** 24.7 18.7 43.4 Informal sector** 37.6 21.3 58.9 Notes:Employedpopulationaged 10years and over, extended definition o fthe labor force. *see text ** this distribution does not take into account the problemo the live-in maids. f 100 Table A2.11: Prevalenceof poverty by thana # o f YOofpoor % o f Sectors over-represented inthe thana obs. precarious Cantonment 178 36% 8% Construction, transport, household services Demra 535 46% 17% Transport, trade Dhanmondi 226 5% 8% Community and household services Thana 43, dt Dhaka* 109 67% 14% Agriculture Gulshan 294 22% 18% Manufacturing Keraniganj 515 53% 18% Agriculture and manufacturing Kotwali 237 38% 14% Community services Lalbagh 491 31% 18% Trade and manufacturing Mirpur 629 16% 15% Community and household services Mohammadpur 172 42% 28% Transport ,manufacturing, construction Motijheel 211 25% 21% Community services Ramna 208 4% 1% Household services and finance Sabujbagh 425 10% 16% Household services and finance Savar 182 23% 33% Agriculture and construction Sutrapur 458 40% 20% Manufacturing and trade Tejgaon 271 17% 37% Transport andmanufacturing Uttara 131 58% 16% Agriculture Bandar* 117 32% 26% Manufacturing Narajanganj 415 34% 18% Manufacturing Gazipur Sadar 317 51% 35% Agriculture Thana 50, DtNarajanganj* 104 39% 42% Manufacturing Notes: author's calculations based upon the 2000 LFS. The number of observations includes all persons aged 5 to 99. *: small sub-sample: number of workers less than 50. Table A2.12: Workingtime in Dhaka sma by income group: average and distribution Working time The poor The precarious The non-poor Total Average 46.3 hours 48.2 hours 49.6 hours 48.3 hours <15 hours 5.5% 3.1% 3.7% 4.1% 15-19 hours 3.2% 2.2% 1.1% 1.9% 20-29 hours 9.7% 7.6% 6.3% 7.6% 30-39 hours 8.1% 7.6% 4.9% 6.4% 40-49 hours 27.2% 29.9% 32.3% 30.3% 50-59 hours 18.2% 23.4% 20.5% 20.3% 60-69 hours 16.8% 15.7% 17.9% 17.1% >70 hours 11.O% 10.1% 13.0% 11.8% Total 100% 100% 100% 100% Notes:this distribution does not take into account the problemofthe live-inmaids. Source: 2000 LFS 101 Table A2.13: Share of living-in domestics (LID) in Dhakasma by incomegroup The The precarious The non-poor poor % o fnon-relative members amongst female workers l 44.32** * 1 1 1 1 N o education 20.501*** 19.568*** 1.375 Education, level 1 22.31*** 14.811*** 13.722** * 2.458 *** Member o f the family 1 Non-relative 107.02* ** Unmamed 1 1 1 1 Mamed 0.193*** 0.043O* ** 0.205*** 0.65 1 Divorce or widow 0.505 0.607 0.638 0.985 N o children under 5 I I 1 1 Presence o f children under 5 0.580** 1.822 0.492*** 0.472*** YOo f male inage o f working 0.258 * 0.I96 0.275 1.146 % of male un-or underemployed 0.61I 0.535 0.673 0.697 Head inthe informal sector I 1 1 1 Head inthe formal sector 0.296*** 5.940** * 0.156*** 0.825 Head day laborer I I 1 1 Head employee 1.785 0.147* 3.111** 1.891** Head self employer 1.242 0.I26* * *** 1.312 1.369 Head, education l e v e b l I I 1 1 Head, education level 1 1.259 0.024* ** 1.361 1.595** Householdreporting asset I I 1 1 N o asset 0.919 0.272 *** 1.106 1.044 D6-D7, without female contribution I I 1 1 Dl-D3, without female contribution 1.397 0.074 * * * 1.919* 2.584*** D4-D5; without female contribution 0.491* 0.338** 0.732 1.152 Options included inthe logit multinomial 0: non-worker 0: non-worker 0: non-worker 0: non-worker Initalics, the reference group 1: garment 1: garment 1:garment 1:garment Underlined, the estimation presented 2: maid 2: live-inmaid 2: live-in maid 2: live-in maid Number ofobservations: 2658 3: other 3: live-out 3: live-out 3: live-out maid - maid maid 4 :other 4 : other 4 :other Notes: * significant at 10%level. ** significant at 5% level. *** significant at 1% level. Women aged 10 years and over. In italics, risk ratios that could not be strictly interpreted given the inclusion o f L-1.Mamong the observations. 105 TableA2.21: Labor force participationrate of childrenand sector distributionof their employment Boys Girls Both Labor force participation amongst the 5-14 21% 19% 20% years old Sector of employment o f child workers Agriculture 13% 20% 16% Manufacturing 25% 27% 26% Trade 34% 0% 18% Transport 9% 0% 5% Community services 8% 9% 8YO Household services 9% 43% 25% Other 2% 2% 2% Total 100% 100% 100% Source: computed from 2000 LFS. 106 Table A2.22: Determinantsof childwork. probitmodel. marginaleffects. 12 13 14 15 16 All obs. All obs. Boys Girls Girls. without without non- non- relative relative Female -.0521 ,0418 -.0865** ,0422 Age .I175*** ,0236 .0934*** ,0215 .1202*** ,0352 .0832** ,0336 .0591** ,0258 -.0027** ,0011 -.0019* .0010 -.0028 .0018 -.0013 ,0016 -.0008 .0013 No education .3898*** ,0648 .3353*** ,0657 .5015*** ,0923 .2937*** ,0853 .2633*** ,0858 Education, level 1 .1218*** ,0198 .1066*** ,0177 .1684*** ,0320 .0903*** ,0251 .0674*** . ,0194 Son -.0423 ,0318 -.0543* ,0285 -.0810* ,0512 Daughter -.0239 ,0312 .0011 ,0308 -.0003 ,0320 -.0004 ,0246 Non relative member .4854*** ,1247 ,1674 ,1437 .6788*** ,1402 Presenceof children -.268*** ,0395 -.2653*** ,0392 -.2424*** ,0530 -.3095*** ,0596 -.2557*** ,0560 aged under 5 * % ofmale in age of -.2427*** ,0593 -.2053*** ,0536 -.2491*** ,0961 -.1979** ,0772 -.1471** ,0611 working % of male un- or .4157*** ,0759 .3488*** ,0675 .7456*** ,1261 -.0637 ,1156 -.0710 ,0923 underemployed Head in the informal ,0238 ,0172 .0302** ,0157 ,0241 ,0267 .0296 ,0232 .0280 ,0185 sector Head self employed ,0166 ,0181 .0115 ,0163 ,0069 ,0271 ,0062 ,0237 ,0083 .0191 Headday laborer .0947*** ,0432 .0826*** ,0391 .1350** ,0682 ,0407 ,0516 ,0322 ,0415 Mother, educationlevel 2 .0392* ,0245 .0340* ,0218 ,0172 ,0329. .0658** ,0377 .0485* ,0297 Mother, educationlevel 3 .0371* ,0244 .0297 ,0222 ,0353 ,0407 .0327 ,0289 .0199 ,0234 Father, education level 2 .0613*** ,0272 .0509** ,0242 ,0177 ,0353 .0941*** ,0403 .0729*** ,0327 Father, educationlevel 3 -.0061 ,0190 -.0036 ,0171 ,0024 ,0312 -.0138 ,0242 -.0072 ,0191 Household reporting no -.0282* ,0175 -.0278* ,0160 -.0339 ,0271 -.0382 ,0249 -.0310 ,0200 asset Decile 1- 3 without child .0511*** ,0191 .0480*** ,0170 .0507* ,0295 .0675*** ,0273 .0564*** ,0215 contribution Decile 4- 5 without child -.0133 ,0207 -.0062 ,0185 -.0199 ,0294 .0029 ,0327 ,0144 ,0258 contribution Number of observations 1839 1776 917 922 873 Percent of correct prediction Entry 48.7 43.9 42.0 40.4 37.6 No entry 75.2 75.5 77.3 83.5 82.1 Notes: * significant at 10%level. ** significant at 5% level. *** significant at 1%level. Children aged 5-14 years. 107 Table A2.23: Comparison of wages betweenDhaka and the rural areas in 2000 (in taka) Dhaka (1) Rural areas (2) (111(2) Average monthlywage 3312 1835 1.80 Average monthlywage by level of education No education 1323 902 1.47 Class Ito V 1971 1407 1.40 Class VI to X 2783 1866 1.49 SSC-HSC 3964 3496 1.13 Degree and above 7696 4533 1.70 Average monthly wage by sector Agriculture 1768 1059 1.67 Manufacturing 2329 1706 1.37 Construction 6273 1771 3.54 Trade 3000 1360 2.21 Transport 3918 2519 1.56 Finance 7270 3604 2.02 Community services 4507 3227 1.40 Household services 1370 783 1.75 Average monthlywage by gender Male 4159 2132 1.95 Female 1600 887 1.80 Source: 2000 LFS. Extended definition o f the labor force. Employed population aged 10 years and over. Table 5.2: Comparison of un- and underemp:qlment rates betweenDhaka and rural areas Dhaka (1) rural areas (2) (1)-(2) Unemploymentrate 7.4% 3.3% 4.1 Proposedunemploymentrate 10.4% 11.3% -1.1 Underemployment 16.0% 38.6% -22.6 Unemploymentrate by level of education no education 3.2% 1.6% 1.6 Class Ito V 5.9% 3.4% 2.5 Class VI to X 8.3% 5.0% 3.3 SSC-HSC 13.5% 10.5% 3 Degree and above 6.0% 8.3% -2.3 Proposedunemploymentrate by level of education no education 5.5% 9.1% -3.6 Class Ito V 10.0% 12.9% -2.9 Class VI to X 13.6% 14.4% -0.8 SSC-HSC 14.9% 13.5% 1.4 Degree and above 6.7% 9.9% -3.2 Source: 2000 LFS. Extended definition o f the labor force. Employed population aged 10 years and over. 108 Table A2.24: Education levels of workers by income group (in percentage) The poor The precarious The non-poor Total Male workers Literacy rate 52.3 58.75 88.68 72.16 No education 43.3 35.8 9.1 24.4 Class I-V 26.2 22.8 10.6 17.6 Class VI-X sscmsc 22.0 21.7 24.6 23.3 6.1 13.8 22'2 17.4 Degree & above 2.1 5.6 30.2 17.2 Technical & 8.4 8.4 10.8 9.6 Vocational training Female workers Literacy rate 32.6 46.9 60.7 (73.6) 49.3 No education 66.8 50.7 35.5 (21.4) 48.1 Class I-V 14.3 23.O 17.1(20.8) 17.4 Class VI-X sscmsc 14.3 18.4 20.2 (24.8) 18.0 3.9 6.9 14.0 (18.4) 10.0 Degree & above 0.5 0.7 11.9 (14.6) 6.2 Technical & 9.9 1.5 6.1 (7.1) 6.3 Vocational training Notes:Employed workers aged 15 years and over. Figures inparenthesis indicate the correspondingshare for the female workers excludingthe LID. Source: 2000 LFS. Table A2.25: Estimates of the employment elasticity with respect to sectoral growth. annual average Annual average Employment Sources growth of value- growth o f elasticity added employ. 1991-2000 1991-2000 1991-2000 Agriculturea 3.3% 0.9% 0.27 Salmon (2002) In;fustrya 7.2% -0.8%