70228 An Evidence-Informed Response to Slum Settlements A Learning Note1 June 2011 SUMMARY This note provides city-level policymakers with basic information about the use of data to inform programs and policies that deal with slums. The goal is to enable officials to work with technical experts and staff to put data collection efforts in place. The note (1) describes different indicators of deprivation of slum dwellers and their settlements (2) discusses data sources and methods (3) suggests ways to categorize these settlements based on these data, and (4) examines policy responses and the data required for implementation. Case studies and examples are included. INTRODUCTION – WHY DATA MATTER With close to one billion slum dwellers today and projections that over 90 percent of world population growth until 2030 will occur in the developing world, it is now clear that shelter strategies of cities that are experiencing rapid population growth need to advance on two fronts. First, for those already resident in the city, slum upgrading needs to be significantly scaled up and become much more programmatic rather than a series of isolated projects. And second, if they are to stem th e rate of new slum formation, cities need to become much more proactive providing access to new affordable housing opportunities for households joining their ranks each month through migration or natural increase. Designing scaled-up policy-based approaches requires much better information than currently exists on slums and on the broader land markets in which they are found. Very often national and municipal governments do not have a firm grasp of: where all of the slums are; how many slum households/ dwelling units exist; at what rate the slum population is growing; what is the spatial pattern of slum growth; who owns the underlying land on which the slums exists; what forms of land/housing tenure slum dwellers have; how many of the slum settlements are located on environmentally sensitive land or land that is unserviceable; what is the market value of the land occupied by slums; what level of infrastructure services slum dwellers have access to; socio-economic profile information of slum dwellers etc. Yet, this kind of evidence is critical for the development of public policies to address slums and to promote better functioning of housing and land markets. Why bother to collect data? The use of reliable and timely evidence enables policymakers to identify the most deprived areas of the city. A rapid diagnosis is essential to focus attention in selected urban areas in order to target scarce resources in a more effective way. Data provide a way to categorize slums in order to identify slum characteristics and design the right policies for the right set of circumstances. This promotes a more efficient allocation of resources and investments. Finally, data provides officials and policymakers with a way to engage with all the interested public, private, city-level and community-level parties and move the politically sensitive issues of slums to neutral ground. This learning note prepares policymakers to engage with their technical experts and staff in designing and implementing an evidence-informed response to the challenges of existing and future slum 1 Based on an earlier paper authored by Dr. Carlos de la Espriella. This version incorporates comments from Robin Rajack, André Herzog and Uwe Deichmann all of the World Bank and Remy Sietchiping of UN-HABITAT, includes additional sources, and was revised by Barbara Lipman, consultant to the World Bank. Acknowledgements also are due to Alfredo Stein, Claudia Varela and Anders Herngren for useful discussions and constructive comments, and to Juan Ricardo Mejía and Luis Alberto Quintero for background material about Medellin and Bogota, respectively. 1 settlements. To that end, the note discusses a set of indicators and data analysis techniques needed in the identification and characterization of slums as well as relevant lessons and examples. In addition, the noted discusses the data requirements of various policy responses. 1.1 INDICATORS OF DEPRIVATION OF PEOPLE The deprivation of people who reside in slums takes many forms such as lack of adequate housing (discussed in more detail in the next section) food, education, health and basic services. No one set of measures of well-being are appropriate for all circumstances. However, this section discusses some of the basic, most common measures that can provide insights to policymakers. The most commonly used poverty measures are those based on the per capita income or consumption of a household.2 Such monetary indicators aim to assess whether households can afford to buy a very basic basket of goods at a given point in time. There are many ways to define the value of this ‘basic basket’ but they all have in common that the basket contains a minimum of goods essential for the household: food (often distinguished by their nutritional contributions), housing, water, clothing , transport etc. This is used to establish a line under which households who earn less than the income needed to afford this basket of goods are considered poor. The number of households falling below this line is an indication of the poverty rate. Another, related, way to consider poverty within a slum neighborhood is to look at the poverty gap. This is a measure of how far below the poverty line households fall. As some will be close to the line and others far below, this gives an indication of the distribution and severity of poverty among households within the community. The poverty gap measure also provides a sense of the resources that will be required to improve the conditions of residents. Other commonly used income and poverty measures take account of the distribution of income or income inequality. Often done on a city-wide basis, this measure indicates the average income of different parts of the city and organizes the income by quintile or decile. Areas where most residents fall into the lowest income ranges may be cited as priorities for limited government resources. As an example of poverty measures, this table from Karachi shows the poverty rate in various types of neighborhoods. Of note, even though the rural fringe neighborhoods have the highest poverty rate, only 7 percent of the poor (column 2) live in this type of neighborhood. Thus, it is important to know both the incidence of poverty (the rate) as well as how many of the total poor live in a certain location. Table 1: Poverty indicators for neighborhoods in Karachi, Pakistan Poverty Poor Households (hhs) % Population Rate (%) as % of All Poor hhs Accounted for in City Clusters Old City 61 18 12 Old Settlements 56 10 7 Korangi 49 12 10 Site 59 27 18 Service Areas 54 11 8 Rural Fringe 67 7 4 Affluent 15 16 41 TOTAL 100 100 _____________________________________________________________________________ Source: Altaf et al.3 2 Hentschel, J. and R. Seshagiri, (2000),The City Poverty Assessment: A Primer, World Bank, pp. 2-3. 3 Altaf, M.A. [et al] (1993), “Poverty in Karachi: Incidence, Location, Characteristics, and Upward Mobility�. Pakistan Development Review. 32: 159-70. 2 In addition to income, there are other indicators of the wellbeing of households such as health and education indicators. They directly measure the well-being of the population – to what degree people can lead healthy and long lives, and what level of education they have reached. Several such outcome indicators concentrate on children – the group in society for whom it is of great importance to be well-fed and healthy because their whole life will depend on their first few years. A widely used indicator is the percentage of all children below the age of five that are malnourished, i.e. children not having grown sufficiently for their age (chronic malnutrition), or children that do not weigh enough given the height they have obtained (acute malnutrition). 4 Other outcome measures are the rate of children that die as infants or in their childhood, the maternal mortality rate, or the life expectancy of the city population. Education indicators could include school attendance among children, the percentage of adults that can read and write, or jobs obtained by school graduates. Deprivation indicators can be quite diverse and include such measures as crime and violence rates (robbery, homicide, domestic violence etc.), child labor, or discrimination in the city or specific neighborhoods. As Hentschel and Seshagiri note, if a group such as a certain ethnic background or gender is discriminated against in the workplace, resulting in lower wages, or if they are discriminated against in public life by the courts or the police, this is a deprivation in itself. 5 Often these types of indicators are difficult to measure, but the effects on the lives of slum residents are nevertheless profound. That said, some progress has been made measuring residential segregation using an “index of dissimilarity� which estimates the difference in proportion of population groups in a given territory compared to that in the city as a whole. The results of these estimations are instrumental in the design of policy responses that seek to integrate population groups. For example, the index was used to estimate housing differences in the medium-sized city of Liberia, in Costa Rica, in order to identify deprived urban areas and take specific actions to raise housing conditions to city standards. 6 However, as Moreno points out in a UN-HABITAT study, most income and well-being indicators fall short in their ability to incorporate concepts such as freedom, human rights, participation and other issues related to governance that are essential to human development. “The usual poverty measures are extremely limited in that they fail to include issues of social exclusion including the lack of opportunity to participate in society or lack of social linkages and connections. It can also mean disadvantage in access to resources, human misery, dependency, social instability and economic [disenfranchisement].�7 1.2 INDICATORS OF DEPRIVATION OF PLACE Perhaps the most conspicuous aspects of slum settlements are the poor quality of housing and lack of services for residents. Here, a good starting point for ascertaining typical housing conditions in slum settlements are the five indicators that, accordingly to UN-HABITAT, define an inadequate housing unit 8. These basic characteristics can be adapted to local circumstances. First, households that live in slums usually occupy dwellings that do not provide adequate protection to weather and climate conditions often are exposed to high morbidity and mortality risks. This condition is measured with the poor structural quality indicator. 4 The World Health Organization maintains a ‘Global Database on Child Growth and Malnutrition’ (www.who.org) which contains comparative information, data sources, and analytical information. 5 Hentschel, J. and R. Seshagiri, (2000), p 6. 6 de la Espriella, C (2009). A Technique for Small Geographical Scale Poverty Analyses: Its Application in the Case of Liberia, Costa Rica. Urban Studies, 46 (11), 2399-2421. 7 Moreno, Eduardo Lopez, Living with Shelter Deprivations: Slums Dwellers in the World UN-HABITAT p. 5. Undated. 8 A full description of each indicator can be found in UN-HABITAT (2004) Urban Indicators Guidelines. See also: UN Millennium Project (2005) A Home in the City, Task Force on Improving the Lives of Slum Dwellers. 3 Second, although many non-slum areas worldwide are also highly densely populated, density is still a key characteristic of slums. Reduced space per person is often associated with certain categories of health risks and therefore considered a key criterion to define an inadequate housing unit. This condition is measured with the overcrowding indicator. Third, a supply of clean water is necessary for life and health; yet, almost 2 billion people lack access to adequate water or can only obtain it at high prices. Slum households are rarely connected to the network and can only rely on water from vendors at vastly higher cost than the tap price. Lack of access to safe water implies greater burden on people — mostly women — to collect water from available sources. It also means increased risk of water-related diseases and diminished quality of life. This condition is accounted for with the inadequate access to safe water indicator. Fourth, lack of sanitation is a major public health problem that causes diseases, sickness and death. Highly infectious, excreta-related diseases such as cholera still affect whole communities in developing countries. Diarrhea, which is spread easily in an environment of poor hygiene and inadequate sanitation, kills about 2.2 million people each year, most of them children under five. This condition is measured with the indicator of inadequate access to improved sanitation. Lastly, slum dwellers are often exposed to unlawful eviction. UN-HABITAT defines secure tenure as “the right of all individuals and groups to effective protection by the State against unlawful evictions�. 9 The insecure residential status indicator can be used to measure this condition. It should be noted, however, that measuring lack of tenure security in a meaningful way can be complex. Recent work in Luanda Angola, for example, suggests that residents may feel secure despite the lack of official, legal documents. As the majority of property transactions in the city mainly occur through “informal� markets most residents considered that sales documents and property transactions papers provided them with some measure of security even though these documents are unlikely to be recognized in a court of law. In fact, a recent survey indicated that property disputes in slum settlements were quite rare. However, it is not clear what protection residents would have against large-scale evictions 10. Assuming data are available to create these measures at the local level, how might city officials interpret them? As Moreno points out, at present, one out of three urban residents in the developing world are slum dwellers. New empirical information shows that not all of them suffer the same degree or magnitude of deprivation. This is due to the fact that not all slums are homogenous; some in fact, provide better living conditions than others. 11 The degree of deprivation depends on how many of the five shelter deprivations are associated with a particular slum household. Based on the number of deprivations, UN-HABITAT has grouped the slum dwellers into three categories: i) Moderately deprived, are those households living with one shelter deprivation. This could be, for instance, lack of access to improved water or sanitation; it could also be that the household lives in non-durable housing structures or with insufficient living space. ii) Severely deprived, are households living with two shelter deprivations that can be a combination of different possibilities; for example, lack of sanitation and insufficient living space, or lack of durable housing structures and lack of improved water; etc. iii) Extremely severely, are households living with three or more shelter deprivations. Here also different combinations of slum conditions exist. For example, a household can lack improved water and sanitation and live in a house that is not deemed to be durable as n ational standards. 9 Despite being an imperfect definition, it includes both legal and perception aspects of tenure security; there have been various efforts to calculate this condition, for example, the Expert Group Meeting on Urban Indicators 2002 (pg.8) proposed an alternative measure of tenure security that uses five separate indicators. 10 Development Workshop (DW) (2011) The Case of ANGOLA:Strengthening Citizenship through Upgrading Informal Settlements Cross-Country Initiative, World Bank report. 11 Moreno, p. 14. 4 Another set of commonly used indicators are “access indicators� which are used to capture other aspects of living conditions. These include access to schools, healthcare and hospitals, and public and private transport – all of which are related to location of slums. Such data also provides insights on the reach of government nutritional aid and social assistance programs as well as the actual distribution of beneficiaries of city programs. 12 Moreover, policymakers can design responses that bring urban services closer to slum settlements, such as providing connections to utilities, transportation of children to existing schools, the relocation of schools and police stations to slum areas, or the provision of bus routes to service slums. Information collected on conditions and services can be depicted in a graphic as shown in Figure 1 which captures data from Yerevan, Armenia. If residents in all the neighborhoods had access to all seven services, the outline of the figure would be bordered in blue. Instead, the figure shows that Yerevan neighborhoods have excellent police protection and very good sanitation and electricity coverage (although there is room for improvement). Public transport exists but its coverage is somewhat limited. In contrast, the city fares extremely poorly when it comes to the provision of water, establishment and maintenance of road networks, and solid waste disposal. Figure 1: Indicators of housing conditions and services in neighborhoods in Yerevan, Armenia Water 1 Police 0.8 0.6 Sanitation protection 0.4 0.2 0 Solid waste Road network Electricity Public transport 13 Source: Habitat for Humanity International (2009) 2. TAKING THE MEASURE OF DEPRIVATION – SOURCES OF DATA In order to be useful in addressing slum challenges, data must be collected that captures the situation on the ground, and that informs policymakers of the best approach in response. Ideal data will never exist. But some types of data are best to describe certain situations – there are times an aerial view of the landscape may be informative; other times, household-level survey data may be required. Table 2 presents a list of indicators, variables, key data and data sources for measuring deprivation and to help categorize slums so that appropriate policy responses can be designed. For any given piece of information, there are a number of ways to collect data directly or to consider proxies that can give a roughly reliable sense of magnitude. Existing sources of data should be consulted first, and then, if inadequate or out-of-date, new data collected to fill in the gaps. 12 Hentschel, J. and R. Seshagiri, p. 5. 13 Habitat for Humanity International, Global Housing Indicators projects, based on 2009 survey data in Yerevan, publication forthcoming. 5 Table 2: Indicators, variables and data sources used to measure deprivation Indicators Variables Data sources Data for income and poverty indicators Censuses, household surveys, household-level Income and poverty - Poverty rate (incidence), poverty gap, income distribution, and interviews; mapping of poverty data indicators income inequality Data for health and education indicators Censuses of height and weight, nutritional - Infant mortality rate, maternal mortality rate, life expectancy, Health of children analyses of food baskets, mapping of clinics malnutrition Education Specialized household surveys of parents, - literacy rate, years of schooling students and schools; Data for inadequate housing indicators - Dwellings built in a hazardous location (e.g. exposed to earthquakes, Poor structural landslides or floods; located on garbage-mountains, around high- Censuses, household surveys and primary data quality industrial pollution areas; or other high-risk zones such as railroads, obtained via direct surveys in targeted areas, (at least one of the airports, energy transmission lines), enumeration surveys, remote sensing, land- following two - Dwellings built with non-durable materials in relation to quality of use mapping including demarcation of flood variables) materials used for wall, floor and roof; and the compliance with local zones and other hazards 14 building codes, standards and bye-laws 15 Overcrowding - Dwellings that accommodates more than three persons per room Censuses, household surveys and primary data - Water for drinking from pipes, public tap, borehole or pump, Censuses, household surveys, primary data, Inadequate access to protected well, protected spring or rainwater, is inaccessible, its and data from public, parastatal or private safe water 16 quantity is insufficient, or its price is unaffordable companies in charge of water supply Censuses, household surveys, primary data, Inadequate access to - Household lack access to facilities that hygienically separate human and data from public, parastatal or private improved sanitation excreta from human, animal and insect contact at accessible price companies in charge of sanitation Insecure - Households without documentary evidence of the household’s legal Property register departments and household residential status claim for tenure, surveys of alternative documents; information (at least two of the - The eviction experience of the household, 17 from neighbors, property brokers following variables) - Reported perceptions of insecurity of the head of household Data for other quality of life indicators Crime and violence Police and court records, mapping of transit - Reported and un-reported criminal activity; arrests; police services Transit access routes and employment centers; utility - Access to jobs and education via public transit City services companies; neighborhood-level surveys and - Hours of electricity; type/frequency of garbage disposal (electricity, garbage) interviews Data for social exclusion indicators - Lack of access to social and city services, credit Review of city per capita expenditures across Inequitable - Perceptions of residents neighborhoods; focus groups and surveys of resources; resident - Perceptions of residents by non-residents employers, lending institutions, residents, and perceptions non-residents Source: Based on de la Espriella (2010) and Hentschel and Seshagiri (2000) The advantages of using secondary, or existing, sources of data are that the cost of data collection is minimized. National level censuses or household surveys (Living Standard Measurement Surveys or Employment Surveys) where available may provide a starting point, although local information for informal settlements usually is lacking. In addition to traditional government censuses, other local sources may include utility companies, universities, and NGOs working with community groups as well as multilateral donor organizations working in-country. Another benefit of using existing 14 The observance of these variables requires skills to determine hazardous areas, expertise on durability of local construction materials and techniques, and specific knowledge about the legal condition and the land use plan. 15 A room is defined as a space in a dwelling enclosed by walls to a height of at least two meters, of an area large enough to hold a bed for an adult, and that is at least 4 m2. Therefore, the concept of a room can be applied to spaces such as bedrooms, dining rooms, living rooms, studies, habitable attics, servants’ room, kitchen, etc. 16 This indicator requires adapting definitions of ‘inaccessibility’, ‘sufficient quantity’ and ‘affordable’ to the local context. 17 This variable requires a classification of tenure documents, as clearly all documents are not of equal weighting (e.g. a deed of lease in the name of a deceased parent may be considerably weightier than a property tax receipt). Tenure documents may also vary in weightiness from country to country. In that respect, this would require someone to calibrate the most commonly held documents in a country by asking a few local exper ts what each major category of document allows in that country (e.g. is it good as a protection against eviction; can it be used as a collateral for credit; can it be used to trans fer the property to someone else, etc.). 6 data is that reaching out to organizations such as these is a way to develop and solidify working partnerships (See Box A below for an example). Box A: Gauteng City-Region Observatory (GCRO), in South Africa The Gauteng City-Region Observatory (GCRO) was established in 2008 as a partnership between the University of Johannesburg (UJ), the University of the Witwatersrand, Johannesburg (Wits) and the Gauteng Provincial Government (GPG). Its goal is to provide the foundations for better planning, management and co-operative government, based on improved data, information, analysis and reflective evaluation. Its main activities include: (1) to provide direct assistance to government through short to medium-term work; (2) to develop structures, processes and interventions to connect government to academic expertise, or to help academic specialists reach decision makers if their work has policy implications, (3) to collect and store strategically useful data, surveys, development of new data sources, GIS mapping and analysis work, and assemble information into innovative data products; (4) to provide medium to longer-term applied research; (5) to carry out academic research; and (6) to promote partnerships and networks. Despite its short history, the GCRO has already published a number of evidence-based maps and reports, and has organized policy-related seminars. Additional tools such as interactive maps are being developed. This is an inspiring example that promotes the use of evidence in planning and brings together local governments, academia and interested stakeholders around issues of urban planning. Source: www.gcro.ac.za When secondary data for the construction of variables and indicators mentioned above is not available, primary data collection may be needed. Household surveys are a very important data resource. However, the sample of households must be carefully chosen so that the results of the survey nevertheless describe living conditions in the neighborhoods accurately. Sampling should be based on mapping of actual settlements, including newly-formed informal ones. The sample size (the number of households interviewed) will vary with a number of factors, among them (a) the indicator that is to be measured. For example, it is much more difficult – and therefore the sample size has to be larger -- to estimate average income of households than it is to estimate the percentage of households with water connection; (b) the level at which the city policy maker needs the information. Intuitively, a city -wide average electricity connection rate will require fewer households to be interviewed than averages in each of 20 sub-districts in the city; and (c) the population of the city or neighborhood. As a rule of thumb, cities that have carried out household surveys have had a minimum sample size of 1,500 to 2,000 households. 18 Smaller districts could be covered by 100 or so households. For neighborhood level data gathering, the questionnaire shown in 3 may be used as a guide. Table 3: Slum identification questionnaire Dwelling & Resident characteristics Date: Household members & ages Name of interviewee: Household members school and work journey and attendance Location in hazardous location (observation): Compliance with building codes (observation): Durable materials in walls (observation): Durable materials in floors (observation): Durable materials in roof (observation): No. of people living in the dwelling: No. of rooms: Main source of drinking water: Quality of drinking water: Intermittence of drinking water: Monthly price of drinking water: Type of sanitation facility (observation): Type of ownership: Own Documents on occupancy, Experiences of eviction: Perception of tenure insecurity: building and/or land; pay rent to purchase, rental or ownership: whom? Source: de la Espriella (2010) 18 Hentschel and Seshagiri, p 8 7 Qualitative methods can take a number of forms: in town-hall meetings or discussions with certain groups or their representatives; communities can rank what they consider the causes of poverty, individual interviews can investigate the problems w omen or children in households; discussions with people at their jobs can reveal information about employment, commuting, access to credit, and other issues related to livelihood. With respect to city services, canvasing residents can help determine priorities for services in the face of limited resources, for example, provision of toilets, or bus service along certain routes. And, as Hentschel and Seshagiri suggest, such assessments can get to a certain type of information which other sources normally cannot capture, for example the incidence and effect of domestic violence. New technologies provide another, often cost-effective way, to capture conditions in established slums and identify areas where new slums are forming or likely to form. Geographical Information Systems (GIS) allow the mapping of data collected from secondary sources or new household surveys, for example, mapping poverty rates across the city to identify high-poverty areas. Another type of application is mapping changes in real estate values to see where the city is expanding on the periphery or to identify areas where informal settlements are likely to come under the pressure of commercial developments (see Figure 2 below). Rapid advances in the availability of spatial images, particularly through Google Earth, allow officials to take a “bird’s eye view� of the situation on the ground in slums, including identifying new areas of high density informal settlements. Remote sensing technology enables the identification of areas suffering from deforestation or in danger or flooding. All of these methods – secondary data collection, primary data collection, qualitative methods, and new technologies have advantages and disadvantages. Directly speaking with residents, whether through a census, survey, or smaller focus group, has the advantage of putting officials and their staff representatives in touch with residents. In this way, policymakers can get a first-hand look at the conditions and circumstances of slum dwellers – what and how often they eat, how they get to work, etc. This can potentially lead to an ongoing dialogue in which policies are developed with the preferences of residents foremost in mind. There are advantages, too, in combining both quantitative and qualitative information through these direct methods. The disadvantages are that surveys are very labor intensive and the cost of carrying them out can be prohibitive. This may result in the surveying of some poor neighborhoods and the omission of others with quite different needs. Also, good sampling techniques must be used in the case of surveys and strong recruitment policies in the case of interviewing smaller groups of residents or community leaders. Otherwise, information on how typical or widespread conditions are may be misleading. Data technologies also come with relative advantages and disadvantages. Mapping and remote sensing provide a perspective that may not be possible with on-the-ground surveys. They sometimes provide a politically neutral way to deal with issues such as demarcating boundaries and pointing out hazardous conditions to residents. On the other hand, aerial views have limitations. Visibility is limited and images often need to be ground-truthed based on actual surveys or discussions with residents. Conditions inside buildings, such as overcrowding, cannot be captured. Employment centers can be identified but the time and effort needed by individual workers to commute to them will need to be verified with surveys or focus groups or by researchers taking residents’ journey to work themselves. Realistically, a combination of the data collection methods described above will need to be employed. The important point is that collection of data need not be dismissed out of hand as being too expensive. Many options for data collection are available. Alternative techniques exist that are less costly and time-consuming. For example, much can be learned from Google Earth maps which are virtually free with computer access. Information may already exist but need to be gathered from local NGOs, central government or international agencies. Finally, as discussed above, the cost can be 8 reduced by using suitable proxies, the application of sampling techniques, the involvement of communities, or the association with strategic local partners. Figure 2: Mapping the value of real estate transactions in Med ellin, Colombia (2004-2007) Value or real estate transactions in 2004 Value or real estate transactions in 2007 3. CATEGORIZING SLUMS AND THEIR NEEDS How can the data collected lead to a deeper understanding of the needs of slums and su ggest the right policy response? It is important to develop a typology, i.e., categorize, slums by the most prevalent issues they are facing. This allows a broader-based response to address the underlying causes in some cases and more the development of more specific, targeted interventions in others. For example, it may be the case that lack of secure tenure or fear of eviction arises across-the-board of residents in most slums; however, lack public transit and access to jobs and schools may a more typical concern for residents in newer settlements in the periphery of the city. Analyzing these common and specific needs will help city officials decide what policies need to be put in place (to provide tenure security, for example) and where and how to allow scarce resources (to create new minibus routes, for example). The typology used to categorize slum neighborhoods can vary from simple to complex, and will be determined by the resources and expertise available to do the analysis. A basic useful classification can be done based on the knowledge of city staff and in co llaboration with residents. More in-depth analyses can be carried out using detailed household surveys, GIS in combination with satellite photos, 9 engineering analyses and other data. This process is best explained with a detailed example. 19 A recent study in Luanda, Angola created a categorization of nine types of slum settlements (“musseques�). Two of these are shown in Figure 3 below: Figure 3: Aerial maps, satellite photos, and on-the-ground views of Luanda slum settlements Old Musseques Peripheral Musseques Source: Development Workshop (DW) Angola, for World Bank, 2011 Settlements of the Old Musseques type were estimated to account for about 40 percent of the population in Luanda; the Peripheral Musseques are where roughly 21 percent of the city population live. Household surveys were conducted in sample old and peripheral neighborhoods (bairros). Although both are inhabited by the poor, the characteristics and needs of each type of community are quite different. The Old Musseques neighborhoods are more than 35 years old and are constructed along informal settlement patterns. They are adjacent to industrial districts and commercial centers. Water access is mainly through public standpipes and private tanks. The areas lack integrated infrastructure but are partially electrified and many residents use dried pit latrines. Most people live in ad hoc building structures and precarious houses. Improvements are done mainly by residents, but slowly. In comparison, the Peripheral settlements tend to be less than 10 years old. As the category name implies, they are located on the periphery of the city. They have grown in unplanned, irregular street patterns. Water access through cistern trucks and private tanks linked informally to pipe connections . There is no formal electricity (private generators or meters are used) and no formal sewage system (high use of septic tanks). Houses are of poor quality and are of precarious construction. No sign ificant public 19 Development Workshop (DW), op. cit. 10 investment to improve infrastructure has occurred but considerable investment in construction has been made by residents and private investors. Surveys among residents revealed that those in one of the Old Musseques perceived that they were living in a stable neighborhood because a high proportion of the residents have lived within the bairro for 21-30 years. The great majority (85%) have a high sense of security about their tenure, with the most common housing documents being a “purchase and sale declaration�. However, this sense of security is being threatened by the fact that the bairro is situated in a region that has been zoned for industrial use, and any expansion of the industrial sector in the future may mean that the residents may have to relocate. Residents are connected to the city in that public services such as electricity and water have been put in place by the government and NGOs. Also, the community values the fact that the location of the bairro is close to the city, as well as the fishing harbor. In contrast, interviews with residents in one of the Peripheral Musseques revealed that the bairro has grown rapidly during the last five years, but is characterized by a high level of informality . Most basic services are provided by the informal markets and most of the residents earn their living by working in the informal market. Residents have poor access employment opportunities as well as basic services. There is little sense of safety among the inhabitants, and the level of violence and crime is high, especially at night. Consequently, land values are low. The bairro does not have any urban-grid planning that would facilitate infrastructural upgrading.Some 87 percent of the residents reported that they have a high sense of security of tenure though, in reality, land tenure security remains weak. Clearly, these data aid city officials a great deal in formulating the appropriate policy responses. Conditions in the older, established musseques require additional upgrading of infrastructure such as toilets and more electrical connections. As there is some threat from gentrification, some form of tenure recognizing the long-term occupancy of residents would afford them greater protection from eviction. In the peripheral neighborhoods, basic infrastructure is lacking. Depending upon whether they are on government land or private land, an innovative approach to servicing the land might be called for, including land pooling or guided development (both described below). These residents, too, would benefit from a system of incremental recognition of their tenure. In addition, greater police protection is needed for the peripheral area as well as the extension of bus or minivan services. In both neighborhoods, some relocation may be required – in the older area because of pressures of gentrification and in the peripheral area because of the need to consolidate land for infrastructure. Likely all residents – in these neighborhoods and throughout the many slums areas in the city – also could benefit from broader efforts to deliver health and social services and improve the prospects for children’s education. 4. DATA-DRIVEN IMPLEMENTATION OF APPROPRIATE POLICIES As just noted, data serve as a basis for designing a tailored set of policy responses to slum challenges, and also helps target scarce resources in a more effective way. For example, relocation may be required when informal settlements are shown to be located in dangerous or environmentally sensitive areas; upgrading of infrastructure may be appropriate in well-established settlements whose residents depend on nearby jobs or customers for their home-based businesses; land sharing between private owners and residents may be another solution for settlements on contested land in prime, central locations. Still other measures, such as public transit or attraction of employers, may be required for newer settlements on the urban periphery located far from jobs and economic opportunity. Some of the major policy levers to deal with slums are discussed here followed by the data requirements for implementation. 11 In-situ slum upgrading specific activities include, among others, land and shelter regularization, connection to utilities, and provision of urban services. The advantages for slum dwellers are that households can continue to live in their neighborhoods with their current social networks, they receive new or better quality services such as water and electricity and public transit, and their upgraded homes and neighborhoods provide a better platform for the delivery of social services, he alth clinics and police protection. They are likely to feel less disenfranchised from city government. Meanwhile, the city benefits by incorporating slum districts into normal city functions because this will generate more revenues through tax collection and payment for utilities as well as better health and education outcomes and a more mobile labor force for employers. For example, the Peñaloza administration (1998-2000) in Bogotá, Colombia, focused on connections to utilities, construction of schools, community and health centers, and public space in and near slum areas, accompanied by the construction of a mass transport system (Transmilenio) and a comprehensive bicycle route network (see Figure 4). Bogotá’s achievements were recognized in the tenth edition of ‘Biennale de Venezia’ and Bogotá received the Golden Lion Award for cities. 20 Figure 4: Upgrading of urban services in Bogotá, Colombia Mass transport system Library and community center (Tunal Sewage (Ciudad Bolivar district) (Transmilenio) district) Source: Alcaldía Mayor de Bogotá (2009) Avance formulación y coordinación interinstitucional de la subdirección de Barrios; and archives of Carlos de la Espriella. Even if full-scale upgrading cannot be accomplished, some measures can be taken with limited resources. For example, as an alternative to conventional land administration street addressing, now implemented in over 50 African cities, maps existing settlement patterns regardless of the land tenure or ownership complexity and translates that spatial information into a GIS that city administrators have been able to use for a wide variety of purposes ranging from postal delivery to epidemic monitoring. The practical result is that a relatively simple intervention can make a significant different for the delivery of urban services, by making it easier to locate residents, operate e mergency services, and enforce collection for user-pay services, such as water and electricity. Land sharing, mentioned earlier, is a mechanism to accommodate commercial development on lands occupied by slum dwellers, without evicting existing land occupants who have the right to remain on site. This is achieved by agreeing to divide (share) a plot, so that a developer is given the right to build on 20 Biennale de Venezia 2006 recognised that in the last decades the city of Bogotá has been addressing the problems of social inclusion, education, housing and public space, and described it as a beacon of hope for other cities, whether rich or poor. 12 one portion of the site and land occupants are re-housed on another portion of the same site, with a promise of secure tenure on their new plots or in their new, usually denser, housing. It includes several preconditions: a booming property market to push landowners to make concessions; well -established communities to strengthen the power of slum dwellers; an outside organization (e.g. public agency, strong NGO) to play the role of intermediary; sufficiently large land to accommodate residential and commercial uses safely, and in compliance with local regulations; and financial feasibility for residents, landowners and developers. Two examples of land sharing were analyzed by Rabé (2010). He compared the cities of a Phnom Penh, in Cambodia and Bangkok in Thailand, and found that the success of this mechanism depends a great deal on community organization and third party intermediation. 21 Land Pooling/Readjustment (LP/R) and Guided Development (GLD) are both options for financing and providing infrastructure in underserved areas. Urban LP/R is a technique for managing and financing the subdivision of selected urban-fringe areas for their urban development. In each LP/R project, a group of separate land parcels are consolidated for their unified design, servicing, and subdivision into a layout of roads, utility service lines, open spaces and building plots. The sale of some of the plots is then used for project cost recovery and the now smaller – but higher value – lots are distributed back to the landowners in exchange for their rural land. The approach is widely used in Japan, South Korea and Taiwan, and is being transferred to the developing countries of Southeast and South Asia. LP/R projects are mainly undertaken by local governments. The attraction of LP/R for landowners is that they can share in the land value gains from efficient development. For local government s, it ensures efficient urbanization of land at no cost because the project site does not have to be purchased and the cost of the infrastructure and subdivision can be financed with a short/medium term loan and then quickly recovered by the sale of some of the new building plots. GLD is a land management technique for guiding the conversion of privately owned land in the urban periphery from rural to urban uses through the provision of infrastructure. GLD is undertaken in partnership with landowners. Landowners participate by donating land and by paying betterment levies on the now improved and more valuable land. The GLD proceeds with the preparation of an infrastructure development plan that follows the existing plot boundaries. To finance the scheme, a l oan is initially taken out to build the infrastructure, which is paid from betterment levies provided by landowners either on annual installments or in lump sum if they sell the land. GLD has been implemented widely in Bangkok and Cairo, on a limited basis in Guinea, and has been proposed but not fully implemented in Indonesia and Ecuador. GLD is best used in locations where urban areas are expanding into areas of privately owned land and where government is not able to control the process of land transformation from rural to urban use. By installing public infrastructure in areas where government considers growth can best be channeled, GLD can encourage landowners to realize the increased value of their land resulting from its subdivision and servicing by e ither selling it to a developer or subdividing and developing it themselves. Relocation of residents includes options such as free resettlement, partial financial or in kind compensation, and eviction without compensation. Often resisted by residents, this option is sometimes unavoidable, especially in cases where data indicate that settlements are located in hazardous locations. Dangers may come from natural disasters such as severe flooding, mudslides, or falling rocks or from man-made conditions such as polluted soil and water or waste dumps. Successful relocation efforts ensure that urbanized land and/or adequate housing are available, and that urban services reach new areas, including transport connectivity to livelihood. 21 For more details, see for instance Rabé, P. (2010) Land Sharing in Phnom Penh and Bangkok: Lessons from four decades of innovative slum redevelopment projects in two southeast Asian “boom towns� Paper presented at the policy workshop "Examining The Places We Live: Slums & Urban Poverty in the Developing World�, held at the Woodrow Wilson Center, Washington, DC, USA (April 30th) 13 Lastly, it should be noted that ill-considered policies and regulations often are the cause of lack of affordable land and housing, and a chief culprit in the formation of slums. When overly ambitious plot sizes and road widths are required on new housing developments, units end up too expensive for low- income residents. Strict limits on the height, type, and use of buildings are sometimes put in place a vain effort to discourage both density and further in-migration to the city. Similarly, restrictions limit conversions of land from rural to urban uses. These policies rarely work as intended. Instead, population growth increases the demand for both land and housing and these unaffordable costs force a significant proportion of the urban population into “unauthorized� means of accessing land, including occupying land in dangerous locations. A step that can be taken at little or no cost is to appoint a task force to undertake an audit of standards, regulations, approval procedures, and staffing practices for urban land management. The purpose of the audit will be to review the existing regulations, standards and administrative processes with an eye toward eliminating those regulations and processes that are unnecessary, outdated, unsuitable, inappropriate, obstructionist and that throw up barriers to access to land for those who need it. As part of the audit, the task force should (1) identify the regulations and st andards that protect the public interest, especially relating to health and safety and the environment, (2) suggest appropriate revisions to planning and building standards, regulations and administrative procedures relating to the processes of registering land, or applications for development, (3) identify and eliminate unnecessary and unneeded rules and regulations that restrict access to more affordable land and adequate housing, and (4) provide incentives for professional agency staff to ensure effectiv e administration. The following table (Table 4) summarizes some of the key data requirements for implementing the approaches to slum upgrading and slum prevention outlined above. Table 4: Relationship between data and slum policy options Policy options Data complexity Main data requirements In-situ slum - Land and Shelter High (practices in formal and - Detailed land and housing characterization upgrading regularization informal property markets) - Resident perceptions of tenure security - Utility connections High (data coordination with - Excellent project monitoring and coordination utility companies) - Technical skills to identify hazardous areas, and - Provision of urban services Medium knowledge of durability of construction materials and techniques Land sharing High (data coordination and - Careful land ownership study Land pooling support to proposed project by - Clear understanding of the proposal by all project actors Guided all actors involved) - Knowledge about legal tenure conditions and land use Development - Careful analysis of historical data to plan ahead for likely growth scenarios Relocation - Free resettlement Low - Detailed data on the proposed project and on location, - Partial compensation Low services and amenities in and around the proposed project - No compensation Low - Clear understanding of the risks of the present location and the proposed project by households Regulatory - Revision of urban limits Low - Review of relevant codes and regulations Audit - Revision of urban norms Low - Careful analysis of historical data to plan ahead for likely - Revision of housing norms Low growth scenarios CONCLUSION This learning note is intended to introduce to city policymakers the basic concepts of using data to inform decisions about policies and programs dealing with slums. This note does not advocate data collection for data collection’s sake. Nor does it suggest that data need to b e exhaustively (or expensively) collected in order to proceed. Rather, the point is that much is to be gained by making efforts to collect information by talking directly to residents (through informal meetings and focus groups or more formal surveys) and from the many low-cost sources of data that are now available. An important part of the process is not to just to find out about residents and the conditions in which they 14 live, but involve them in discussions about what the information means – no matter whether the source is a survey, a map or a satellite photo. In this way, key decisions about dealing with slums are based on evidence and the process of implementing those decisions is based on partnership and mutual understanding. SOME HELPFUL RESOURCES The following web links and studies provide international experiences useful for the development of an evidence-informed response to slum challenges. Websites  The Slum Managerial Information System of São Paulo, Brazil, HABISP (www.habisp.inf.br)  The Gauteng city-region's observatory, South Africa (see Box B) (www.gcro.ac.za)  The Slum Upgrading Information System of Medellín metropolitan area, Colombia, SIM (servicios.metropol.gov.co/sim/LoginN.aspx)  The ESPON HyperAtlas of the European Spatial Planning Observation Network (www.espon.eu/mmp/online/website/content/tools/912/index_EN.html)  The l’Observatoire des Territoires : Indicateurs in France (www.territoires.gouv.fr/indicateurs/portail_fr/indicateurs_fr/p4_libre.php)  The Policy Research Initiative in Canada (www.policyresearch.gc.ca/) STUDIES  The Informal Settlement Growth Pattern Model for Dar es Salaam, Tanzania 22  The Peripherisation Model tested in São Paulo, Belo Horizonte, Buffalo and Boston 23  The Logic for Modeling Spontaneous Urban Change for Suzhou region, China 24  The Tool for Strategic Territorial Planning for Liberia, Costa Rica, TSTP) (www.planning4equality.com). 25  Development Workshop (DW) (2011) The Case of ANGOLA:Strengthening Citizenship through Upgrading Informal Settlements, Cross-Country Initiative, World Bank report.  Rabé, P. (2010) Land Sharing in Phnom Penh and Bangkok: Lessons from four decades of innovative slum redevelopment projects in two southeast Asian “boom towns� available at www.worldbank.org/urban  Moreno, Eduardo Lopez, Living with Shelter Deprivations: Slums Dwellers in the World UN -HABITAT Undated.  Hentschel, J. and R. Seshagiri, (2000),The City Poverty Assessment: A Primer, World Bank 22 Blair, H. (2000). Participation and Accountability at the Periphery: Democratic Local Governance in Six Countries. World Development, 28 (1), 21-39; and Montgomery, M.; R. Stren; B. Cohen & H. Reed, eds. (2003). Cities Transformed. Demographic Changes and Its Implications in the Developing World. National Academies Press. Washington, DC, USA 23 Barros, J. (2004). Urban Growth in Latin American Cities. Exploring urban dynamics through agent-based simulation. PhD thesis University College of London, UK. 24 Xie, Y.; M. Batty & K. Zhao (2005) Simulating emergent urban form: Desakota in China. Working paper. CASA Working Papers (95). Centre for Advanced Spatial Analysis (UCL), London, UK. 25 de la Espriella, C. (2009). Pro-poor Planning: A Tool for Strategic Territorial Planning and a Conceptual Framework Drawn from Studies in Colombia and Costa \d Rica. HDM, Lund, Sweden. 15