68138 Knowledge PAPeRS Financing the Urban Expansion in Tanzania Design: miki@ultradesigns.com Cover photos: Barjor E. Mehta Knowledge PAPeRS Financing the Urban Expansion in Tanzania Zara Sarzin and Uri Raich January 2012, No. 15 Urban Development Series Produced by the World Bank’s Urban Development and Local Government Unit of the Sustainable Development Network, the Urban Development Series discusses the challenge of urbanization and what it will mean for developing countries in the decades ahead. The Series aims to explore and delve more substantively into the core issues framed by the World Bank’s 2009 Urban Strategy Systems of Cities: Harnessing Urbanization for Growth and Poverty Alleviation. Across the �ve domains of the Urban Strategy, the Series provides a focal point for publications that seek to foster a better understanding of (i) the core elements of the city system, (ii) pro-poor policies, (iii) city economies, (iv) urban land and housing markets, (v) sustainable urban environment, and other urban issues germane to the urban development agenda for sustainable cities and communities. © 2012 THE WORLD BANK The statements, �ndings, and conclusions expressed in this paper are those of the authors only, and do not reflect the view of the Board of Executive Directors of the World Bank, or of the governments they represent. Contents Acknowledgements v Acronyms vi Introduction 1 Estimating the Cost of the Urban Expansion 7 Current Spending on Infrastructure in Urban Areas 13 Own Source Revenues 13 Intergovernmental Transfers 14 Other Sectoral Spending 16 Conclusions and Recommendations 19 Annex: Derivation of Unit Costs 23 Land Preparation 23 Roads and Drainage 23 Electricity Supply 25 Water Supply 25 Sanitation 26 Solid Waste Disposal 26 Summary 27 References 35 List of Tables Table 1. Typology of Urban Settlement Patterns in Tanzania 4 Table 2. Definition of Service Levels 8 Table 3. Modeled Plot Sizes 8 Table 4. Approximate Number of New Urban Households per Annum 11 Table 5. Summary of Per Plot Costs per Household and per Person 12 Table 6. Summary of Costs 12 Table 7. Source of Revenues of Urban and Rural LGAs, 2006/07 13 iii iv Financing the Urban Expansion in Tanzania Table 8. Urban LGA Own-Source Revenue Sources, 2006/07 14 Table 9. Intergovernmental Fiscal Transfers in Tanzania, FY2005/06 15 Table 10. Urban LGAs Intergovernmental Transfers, 2006/07 16 Table 11. Total spending in urban areas (US$) 17 Table 12. Infrastructure Financing Mechanisms 21 Table 13. Unit Costs for Land Preparation 23 Table 14. Unit Costs for Road Construction: All Gravel Engineered (“Minimum Service�) 24 Table 15. Unit Costs for Road Construction: Access Roads Gravel Engineered (“Average Service�) 24 Table 16. Unit Costs for Road Construction: All Paved Bitumen (“Full Service�) 24 Table 17. Unit Costs for Electricity (Single Phase Line) 25 Table 18. Unit Costs for Electricity (Three Phase Line) 25 Table 19. Unit Costs for Public Water Kiosks 25 Table 20. Unit Costs for Water Reticulation (Excluding Trunk Infrastructure) 26 Table 21. Unit Costs for Sanitation Options 26 Table 22. Unit Costs for Solid Waste 27 Table 23. Summary of Unit Costs by Service Level 27 Table 24. Summary of Unit Costs by Plot Size in Tshs 28 Table 25. Total Cost in Tshs Million per 150,000 New Urban Households 29 List of Figures Figure 1. Urbanization Rates in Mainland Tanzania 2 Figure 2. Demand and Supply of Urban Plots in Dar es Salaam 3 Figure 3. Urban Access to Infrastructure 3 Figure 4. Cost per Hectare by Service Category and Level 9 Figure 5. Cost per Plot by Service Category for Average Service Level 10 Figure 6. Cost per Plot for Various Residential Densities and Service Levels 10 Figure 7. Aggregate Cost for 150,000 Urban Households in Tshs Millions 11 Acknowledgments This paper draws substantially on a background paper prepared for the World Bank by Profes- sors Nnkya and Kombe of Ardhi University, Tanzania. The authors wish to thank Judy Baker (WBIUR), Matthew Glasser (OPCRS), Barjor Mehta, Solomon Alemu (AFTUW), Zhiyu Jerry Chen (ENCVF), Pankaj Gupta (FEUFS), Elijah Luhanga (AFTEG), Yonas Mchomvu, Dieter Schelling (AFTTR), Caroline van den Berg (MNSWA), Vivien Foster (AFTSN) and Jamie Boex (Urban Institute) for their valuable feedback and guidance and for sharing with us key material that made this paper possible. v vi Financing the Urban Expansion in Tanzania Acronyms CBG Capacity Building Grant CIUP Community Infrastructure Upgrading Project GPG General Purpose Grant LGA Local Government Authority LGDG Local Government Development Grant System LGTP Local Government Transport Programme MID Ministry of Infrastructure Development MOWI Ministry of Water and Irrigation PMO-RALG Prime Minister’s Office -Regional Administration and Local Government TANESCO Tanzania Electric Supply Company Tshs Tanzanian Shillings Introduction The urban transition is well under way in Tanzania—and urbanization is projected to continue at a rapid pace. According to census data, the urban population in mainland Tanza- nia increased from 5.7 percent (685,092 people) in 1967 to 22.6 percent (7.6 million people) in 2002.1 The level of urbanization is even higher if one considers “urban� from a population density perspective, with more than 33 percent of the mainland population (approximately 11 million people) living in high density, “urban� areas in 2002 (World Bank 2009, 14-15). Urbanization in Tanzania is projected to continue at a rapid rate. According to United Nations’ population projections, the percentage of people living in urban areas is likely to grow from 24 percent in 2005 to 38 percent in 2030. The urban population is expected to grow at more than twice the rate of the population as a whole so that by 2030, it is estimated that more than 25 million Tanzanians will be living in urban areas (see figure 1).2 The demand for urban land significantly exceeds the formal supply—and the gap is wid- ening. There are major challenges associated with acquiring planned plots, evidenced by the large gap between the number of applications and the number of plots that have been allocated (Kironde 2006, 462). With about half a million new urban residents each year, and assuming an average household size of 3.53, Tanzania would need to produce more than 143,000 plots per annum to satisfy demand.4 However, Tanzania may be producing as few as 6,000 formal plots per year, meeting only a small fraction of the demand. Failure of the plot production system in Tanzania is a major factor in the proliferation of informal and illegal settlements (World Bank 2009, 97). In large cities, such as Dar es Salaam, some 70 to 80 percent of resi- 1 Rates of urban population growth in Tanzania are not geographically even, reflecting different regional drivers of urbanization. For example intercensal growth rates between 1988 and 2002 are highest in Arusha (6.9%), Bu- koba (5.2%), Dar es Salaam (4.7%), Tarime (4.5%), Songea (4.4%), Dodoma (4.2%), Mbeya (4.1%), Kigoma (4.1%) and Morogoro (4.1%). It is not surprising that Dar es Salaam (Tanzania’s primate and port city and the country’s commercial hub), Dodoma (the capital city), Songea (a transport hub), and Morogoro (a transport hub at the intersection of major east-west and north-south corridors) feature on this list. The remaining urban localities are border cities and towns that benefit significantly from trade with neighboring countries. Moreover, due to their size, Tanzania’s largest cities account for the lion’s share of new urban residents. For example, Dar es Salaam, which accounted for 31 percent of the urban population in 2002, also accounts for more than one third of projected urban population growth to 2025. 2 Projections drawn from the UN’s World Population Prospects imply that the urban population will grow by an additional 17.8 million people from 2002 to 2030. While the estimated rate of urbanization may vary depend- ing on the projection methodologies or definition of “urban�, the trend is incontrovertible. 3 In 2002, the average household size was 4.1 in urban areas (4.8 in rural areas) compared to 4.5 (5.4 in rural areas) in 1998. This decline is consistent with global trends, which show that as countries develop, household sizes fall. For example household sizes in Europe, the United States and Canada declined from around 5 members in the mid 19th century to 2-3 members in the mid 1990s (Bongaarts 2001, 4-5). Additionally, urban households are typically smaller than rural households; as households move from an agricultural family structure, where children are an asset because they constitute part of the household’s labor force, to an urban family structure where children multiply costs, family sizes decline. As Tanzania develops, it is likely that the average urban household size will decline. Since this paper is focused on urban planning in the medium term, it uses an average household size of 3.5. 4 If the locations of newly supplied plots are not well aligned with the spatial distribution of urban growth, then the targeted number of new plots would need to be even higher, say 150,000 to 200,000 plots per annum. 1 2 Financing the Urban Expansion in Tanzania Figure 1. Urbanization Rates in Mainland Tanzania Source: World Urbanization Prospects (Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat 2007) dents live in informal areas because of an inadequate supply of planned, surveyed and serviced land parcels (see figure 2). Moreover, investments in urban infrastructure have not kept pace with urban popula- tion growth, resulting in poor or declining access to urban services in several sectors. As illustrated in figure 3, during the intercensal period from 1998 to 2002 urban access to piped water supply fell from 80 to 70 percent and urban access to flushed toilets declined from 13 to 12 percent, with the majority of the urban population (83 percent) still relying on traditional pit latrines for sanitation. These declines are related to the rapid rate of urbanization which oc- curred during the intercensal period and raised the urbanization level from 18 to 22.6 percent. Despite investing in new urban connections, service providers have failed to keep up with urban population growth. In the energy sector, urban access to electricity increased slightly from 27 percent in 1998 to 34 percent in 2002, but it is still extremely low relative to the 50 percent access rate for Sub Saharan Africa (World Bank 2009, 25-26). Rapid urban population growth poses significant challenges for the spatial planning, fi- nancing and provision of urban infrastructure and services, whether these are provided publicly or privately. Continued population growth will further multiply demands for infra- structure and services in Tanzania’s urban areas, where local authorities are already struggling to fulfill their mandates for urban service delivery. In residential areas, urban population growth is likely to lead to increased densification in both planned and unplanned areas and the expan- sion (sprawl) of mostly unplanned settlements on the peripheries of cities and towns. A brief typology of urban settlement patterns in Tanzania is presented in table 1: Tokyo, Japan Financing the Urban Expansion in Tanzania 3 Figure 2. Demand and Supply of Urban Plots in Dar es Salaam Source: Lugoe, 2008 Figure 3. Urban Access to Infrastructure Source: The Urban Transition in Mainland Tanzania (World Bank 2009, 25-26) 4 Financing the Urban Expansion in Tanzania Table 1: Typology of Urban Settlement Patterns in Tanzania Factors to consider in Process of costing infrastructure Urban Settlement Pattern Characteristics Densification requirements Established before the Planned and surveyed. Increase in occupancy Increasing demand for mid 1980s; typically rate. infrastructure commensurate Serviced, but centrally located. with increased densi�cation, infrastructure Horizontal or vertical e.g., installation of sewerage inadequate or poorly expansion of property systems, upgrading of water maintained. to accommodate supply systems and expansion increasing number of of parking facilities. people per dwelling. Established from the Planned and surveyed. Little densi�cation. Servicing of plots including mid 1980s; typically in provision of roads and Planned Services not provided peripheral areas. drainage on existing residential areas before plots allocated. reserves, upgrading of water Developers have and electricity distribution taken responsibility for systems, provision of solid infrastructure provision waste collection services (roads, water and and provision of community electricity) on an ad facilities. hoc basis and at high cost. Many dwellings remain unoccupied. Consolidated, dense Unplanned and not Increase in occupancy Increasing demand for settlements. surveyed. rate. infrastructure commensurate with increased densi�cation. Little infrastructure Horizontal or vertical provision. expansion of property Installation of water and to accommodate sewerage distribution systems, increasing number of upgrading electrical supply people per dwelling. systems, provision of parking facilities. Unconsolidated, less LGAs provide Subdivision and sale Provision and upgrading of Unplanned dense settlements. basic services in of plots to developers infrastructure and community residential areas collaboration with (unregulated). facilities. communities (most commonly schools). Residents have taken responsibility for infrastructure provision (roads, water and electricity) on an ad hoc basis and at signi�cant cost. Peri-urban villages No basic infrastructure Subdivision of rural Provision of all basic (green �eld). and services. plots. infrastructure. Source: Background paper prepared by Nnkya and Kombe. Financing the Urban Expansion in Tanzania 5 Spatially planned investments are required in a number of sectors, both within and out- side of the service delivery mandates of urban local government authorities. A basic infra- structure investment package includes roads, non-motorized transport (dedicated walkways and bicycle lanes), storm water drainage, water and sanitation, electricity and liquid and solid waste collection. Under Tanzanian legislation, local government authorities (LGAs) are re- sponsible for urban planning, district and feeder roads, drainage, solid waste collection and liquid waste collection from on-site sanitation systems. Electricity falls within the mandate of the Tanzania Electric Supply Company (TANESCO), the electricity utility, while water and sewerage fall within the mandate of the regional water and sewerage authorities. This paper seeks to develop estimates of the net cost of the urban expansion in Tanzania. The paper focuses on developing estimates of the cost of planning and servicing land for new residential urban settlement. It does not attempt to develop detailed estimates of the cost of addressing infrastructure backlogs which would include the retrofitting of basic urban infra- structure to unplanned areas. On the revenue side, estimates of current spending in urban areas takes into account urban LGA budgets as well as sectoral spending in urban areas (e.g., in the water sector). A number of assumptions were made to estimate both costs and revenues. These assumptions have been set conservatively, therefore, estimates of the net cost of the urban expansion are considered to be on the lower end of the possible range. These estimates can inform discussions on future investment needs of urban LGAs. They may also serve as the basis for analyzing options for financing each type of infrastructure; for determining what might be financed publicly versus privately, what might be financed with current funds versus credit, what might be financed with local versus national funds; and to determine the spatial arrangement of infrastructure to maximize efficiency, equity and sustain- ability. 5 1 Estimating the Cost of the Urban Expansion This paper focuses on estimating the cost of providing vice levels. The basic package of infrastructure includes: (1) a basic package of urban infrastructure to service new technical services such as valuation and compensation, de- residential neighborhoods. While urban population tailed land use planning and cadastral surveying; (2) roads growth will multiply demands for infrastructure spanning and storm water drainage; (3) water supply and sanitation; commercial, industrial and residential areas of the city, (4) solid waste collection; and (5) electricity. The package this paper focuses on residential areas only. Additionally, does not include other public services, such as health, edu- while urban population growth can potentially lead to the cation or recreational facilities, nor does it include the cost densification of existing (planned and unplanned) urban of administration or management of urban services. The areas as well as the growth of new urban settlements, this combination of these services is aggregated to arrive at the paper focuses on infrastructure provision in new urban total cost of planning and servicing new urban residential settlements, with cost estimates based on the assump- development. Within each category of urban infrastruc- tion that these areas will be planned. This is a significant ture, the costs are estimated separately for three different simplifying assumption since in Tanzania the majority of levels of service. Service levels have been constructed with urban development is unregulated. reference to the regulations governing urban development in Tanzania and extant conditions. These are: (1) minimum The costs of proactively dealing with new urban develop- or basic level of service; (2) average level of service; and (3) ment are assumed to be lower than the cost of upgrading. full service level. The definition of services provided at each The cost of retrofitting infrastructure to existing unplanned level is summarized in table 2. settlements and to planned, un-serviced settlements is a complex exercise that requires identifying available infra- The model considers different residential densities for structure in these areas and requirements to upgrade or new urban development with a range of plot sizes from provide infrastructure in order to fill service gaps. A back-of- 100 to 1,200 square meters. Tanzanian urban planning the-envelope estimate of upgrading unplanned settlements standards provide for the following densities and plot siz- is included toward the end of this section. It appears that in es: (1) high density (400m2 to 800m2); (2) medium den- most cases the cost of providing infrastructure to new urban sity (800m2 to1,200m2); and (3) low density (1,200m2 to areas is less than that for upgrading informal settlements, since one avoids having to deal with existing structures, 2,000m2). In this exercise, the lower end of each range was and the adverse topography, location and layout that often used, i.e., 400m2; 800m2 and 1200m2.5 In addition, to characterize unplanned settlements. For example, a 2007 test whether it would be possible and more affordable to paper on the cost of upgrading in Brazil estimates that the 5 Large plot sizes of 1,200m2 have been included for illustrative pur- costs of upgrading infrastructure in unplanned settlements poses only, however they are unsustainable unless provided on a free are two to three times the costs of servicing a formal-sector market basis with households providing for their own needs. Hous- subdivision (Akibo 2007). The paper therefore engages in a ing policy would ideally focus on plot sizes of around 100m2 to 400m2, with plot sizes of 100m2 to 200m2 envisaged as social hous- simplifying assumption that the costs of proactively dealing ing. More sustainable urban typologies should also be explored, in with new urban development are low relative to the cost of particular multi-family dwellings: low rise (e.g., three story) dwell- dealing with urban expansion after the fact. ings might accommodate 20-40 families per hectare; and medium rise (e.g., five story) dwellings might accommodate 40-80 families per hectare. While multi-family dwellings are more complex to man- The cost estimates cover a basic package of urban infra- age and are not common in Tanzania, nevertheless they would enable structure and services at minimum, average and full ser- more sustainable urban densities that can be serviced more efficiently. 7 8 Financing the Urban Expansion in Tanzania Table 2: Definition of Service Levels Type of service Basic Average Full Includes valuation, compensation, Includes valuation, compensation, Includes valuation, compensation, land use planning and cadastral land use planning and cadastral land use planning and cadastral Land preparation survey costs. These costs apply survey costs. These costs apply survey costs. These costs apply uniformly regardless of the level of uniformly regardless of the level of uniformly regardless of the level of service. service. service. Assumes that all roads are gravel Assumes that district/local Assumes all roads are constructed engineered with lined storm water distributors are paved or to paved/bitumen standard with Roads and drains. bituminized but access roads are lined storm water drains. drainage gravel engineered. All with lined storm water drains. Public kiosk or stand pipe or deep House connection (excluding trunk House connection (excluding trunk Water supply well/bore hole. infrastructure). infrastructure). House connection. The cost of House connection. The cost of House connection. The cost of supplying a typical neighborhood supplying a typical neighborhood supplying a typical neighborhood with energy includes the use of 11 with energy includes the use of 11 with energy includes the use of 11 Electricity supply KV lines along the road lengths, KV lines along the road lengths, KV lines along the road lengths, a substation and single phase a substation and single phase a substation and single phase service line for each plot. service line for each plot. service line for each plot. Liquid waste Ventilated Improved Pit latrine Septic tank and soakaway pit. Localized or central sewer system. disposal (VIP). Communal masonry container Communal masonry container Door-to-door collection by a local Solid waste located within 250 meters walking located within 250 meters walking authority or an agency contracted disposal distance. distance. to collect solid waste. urbanize with smaller plot sizes, the paper also considers Table 3: Modeled Plot Sizes plot sizes of 100m2 and 200m2. Population Number density per km2 In arriving at unit cost estimates for infrastructure, it of plots per (assuming is assumed that in a typical planned neighborhood, at Average plot size net residential household size (m2) hectare of 3.5) least 65 percent of land is devoted to residential plots 1200 5 1,750 (net residential area) while the remaining 35 percent is 800 8 2,800 used for infrastructure, such as roads and drainage, and 400 16 5,600 community facilities, such as primary schools, markets 200 32 11,200 and open spaces.6 While this is consistent with the urban 100 64 22,400 development planning regulations in Tanzania, it con- trasts with a typical dense informal settlement, where the The total area of the neighborhood layout used for this net residential area is well above 80 percent due to the fact exercise is 137.5 hectares with a total road length of that there are rarely proper roads and public facilities. If 20.71 kilometers. Land for basic facilities (primary school, it is assumed that at least 65 percent of land is used for nursery school, playground, active and passive open spaces) residential development, a net residential hectare would is intended to serve residents accommodated in detached accommodate 5, 8 and 16 plots at low, medium and high houses on individual plots, each plot accessible by a road densities respectively. The smaller plot sizes of 100m2 and with a minimum width of 10 or 15 meters. These access 200m2 translate to 32 and 64 plots per net residential roads feed into a network of 20-meter wide local distributor hectare. The modeled plot sizes are given in table 3. roads which in turn feed into a 30-meter wide network of district distributor roads. The total length of all the roads in 6 This is based on Tanzanian planning regulations. However, a lower percentage (e.g., 50 percent) would provide greater scope for future the neighborhood is 20.71 kilometers. These assumptions investments in roads, public spaces, and socio-economic infrastructure. are held constant across all residential densities. Estimating the Cost of the Urban Expansion 9 Unit costs rise steeply depending on the level of service 6. Unit costs are significantly lower for higher density provided. Unit costs were estimated for each category of residential developments and rise steeply with increasing service, for three levels of service (minimum, average and plot size. full). For the water sector, the unit costs exclude any al- lowance for bulk infrastructure costs, and for the energy Aggregate costs for the urban expansion in Tanzania sector, they exclude generation costs. As a consequence, were estimated assuming there are approximately for all service levels, road and drainage infrastructure 150,000 new urban households every year. According comprise the major portion of the cost per hectare (51 to the UN World Population Prospects, the percentage to 84 percent depending on the service level).7 As can be of people living in urban areas is likely to grow from 24 seen in figure 4, the estimated costs per residential hectare percent in 2005 to 38 percent in 2030. Tanzania’s urban vary significantly by service level. Detailed unit cost esti- population is expected to grow at more than twice the mates are included in Annex. rate of the population as a whole, so that by 2030, more than 25 million Tanzanians are likely to be living in urban Per plot costs are higher for larger plot sizes. Unit areas.9 Based on these projections, the approximate num- costs were also calculated by plot size, and these are pre- ber of new urban residents per annum was calculated. For sented by service category for an average level of service example, from 2010 to 2015, it is estimated that there will in figure 5 and by various residential densities in figure be 534,800 new urban residents each year. Assuming an Figure 4: Cost per Hectare by Service Category and Level8 Source: The Urban Transition in Mainland Tanzania (World Bank 2009, 25-26) 7 The World Bank’s Africa Infrastructure Country Diagnostic (AICD) also indicates that roads account for the heaviest weight in the infrastructure bundle (41percent), followed by power (28 percent), water and sanitation (29 percent), and ICT (1 percent). The AICD analysis uses different assumptions to those used in this paper—it includes both energy generation costs and bulk water costs. 8 The high cost of an “average� liquid waste disposal service is based on the high unit cost of providing septic tanks and soakaway pits for each urban plot. 9 Because the UN calculations do not use a consistent or standardized definition of “urban area� but rely on each country’s definition, they must be carefully interpreted. More conservative calculations for Tanzania (Bocquier 2005) estimate that by 2030, the urban population will increase only to 29 percent, which would be equivalent to 16.6 million people living in urban areas, or 6.6 million more than at the present time. 10 Financing the Urban Expansion in Tanzania Figure 5: Cost per Plot by Service Category for Average Service Level Source: The Urban Transition in Mainland Tanzania (World Bank 2009, 25-26) Figure 6: Cost per Plot for Various Residential Densities and Service Levels Source: The Urban Transition in Mainland Tanzania (World Bank 2009, 25-26) Estimating the Cost of the Urban Expansion 11 average household size of 3.5 people, the number of new (see table 4). Not all households would require new land urban households during this period is roughly 152,800 for development, so this represents an upper limit. Table 4: Approximate Number of New Urban Households per Annum Year Urban Period Urban annual New urban Approximate Approximate population growth rate (%) Residents new urban new urban (thousands) (thousands) residents per households per annum annum 2005 9,313 2010 11,487 2005-10 4.2 2,174 434,800 124,229 2015 14,161 2010-15 4.2 2,674 534,800 152,800 A principal driver of cost is the density of urban devel- costs would range from Tshs 189 billion to 443 billion opment. New urban development would accommodate a (see table 25 in Annex). For the purposes of determining range of different residential densities or plot sizes. How- the net cost of the urban expansion, the paper takes the ever, for simplicity and to facilitate comparisons across median assumption that new urban residents are accom- plot sizes, figure 7 presents aggregate costs for 150,000 modated in 400m2 plots with an average service level at a new urban households, assuming all development occurs cost of Tshs 959 billion, or US$820 million. These figures at a uniform density. For example, if all 150,000 house- illustrate the benefits of high-density urban development holds were accommodated in low-density settlements for cities with constrained fiscal resources. (plot sizes of 1,200m2), the aggregate costs would range from 1,412 billion Tanzanian shillings (Tshs) to Tshs The cost of retrofitting urban infrastructure to existing 5,231 billion (depending on the service level). Similarly, unplanned settlements is estimated at US$250 million if all 150,000 households were accommodated in high- per annum, assuming costs are spread over the next 30 density settlements (plot sizes of 100m2) the aggregate years. The UN estimates that the per capita cost of up- Figure 7: Aggregate Cost for 150,000 Urban Households in Tshs Millions Source: The Urban Transition in Mainland Tanzania (World Bank 2009, 25-26) 12 Financing the Urban Expansion in Tanzania grading is US$670 per person.10 This assumption seems Tanzanians living in unplanned settlements in 2001. Ap- reasonable for Tanzania, given that the estimated cost of plying the rough estimate of US$670 per person for up- formal urban development for high density development grading, suggests that upgrading all existing unplanned (100m2 to 200m2) ranges from US$307 to US$491 (for communities would cost over US$7 billion. Spreading basic service levels), and it can be expected that the cost upgrading over the next 30 years would mean an average of regularizing and upgrading unplanned settlements is annual price tag of US$250 million. higher. According to UN data, there were 11,030,600 Table 5: Summary of Per Plot Costs per Household and per Person Cost per Cost per Cost per Cost per Cost per 100m2 plot 200m2 plot 400m2 plot 800m2 plot 1200m2 plot Service Service level (US$) (US$) (US$) (US$) (US$) Total per plot / household Minimum 1,076 1,718 2,892 5,240 8,057 Average 1,506 2,902 5,470 10,605 16,768 Full 2,527 4,945 9,555 18,777 29,842 Total per person Minimum 307 491 826 1,497 2,302 (3.5 per household) Average 430 829 1,563 3,030 4,791 Full 722 1,413 2,730 5,365 8,526 The total cost of urban development is estimated to be be no new middle- or upper-income demand. Second, it US$1,070 million per annum. This includes the cost of does not include the cost of upgrading existing planned accommodating new urban residents in new residential settlements which account for a small percentage of ur- areas (at medium density and average level of service) ban residential areas and where some infrastructure and together with the per annum cost of addressing service services are already in place. Third, it does not include the gaps in unplanned settlements (assuming this cost is rehabilitation or repair of old urban infrastructure that is spread over 30 years). This figure is on the low side for reaching the end of its useful life. And it does not include four main reasons. First, it assumes that all new residents any estimates of current or future operations and mainte- will be accommodated on small plots and that there will nance costs.11 Table 6: Summary of Costs US$ million per annum Average plot size 100m2 400m2 1200m2 New urban expansion (cost of 150,000 new residential plots 226 820 12 2,515 at an average service level) Upgrading unplanned settlements13 250 250 250 Upgrading planned settlements14 Not estimated Operations and maintenance costs Not estimated Total 476 1,070 2,765 10 Estimates of unit costs of upgrading vary significantly across regions. Unit costs are also very dependent the geographical scope, population density and targeted levels of service among other variables. 11 Recent studies suggest that the cost of operating and maintaining infrastructure assets amount to about half of capital costs (World Bank 2009, 79). 12 This is the annual cost of servicing 150,000 plots, using the median assumption that residents are accommodated in 400m2 plots with an aver- age service level. This results in a cost of Tshs 959 billion, or US$820 million. 13 This is the rough estimate of the annual cost of upgrading existing unplanned settlements assuming upgrading is spread over 30 years. Existing unplanned settlements are typically high density (smaller plot sizes) and upgrading typically establishes only a minimum (basic) level of services. Therefore the cost of upgrading is not directly comparable to the estimated cost of new urban expansion which is based on the assumption that residents are accommodated in 400m2 plots with average service levels. 14 Since only four percent of urban settlements in Tanzania are planned, upgrading costs are likely to be relatively minor. 2 Current Spending on Infrastructure in Urban Areas The cost of financing the urban expansion in Tanzania percent. While the majority of Tanzania’s population lives needs to be considered in the context of current expen- in rural areas, even in per capita terms, transfers to rural diture on infrastructure in urban areas. The previous LGAs are 21 percent higher than those to urban LGAs section presented conservative estimates of the investment (World Bank 2009, 76). A pro-rural bias in the transfer in infrastructure required to keep pace with new urban system is not unusual or problematic in itself because ur- residential growth. To a small degree, urban infrastructure ban LGAs potentially have sound tax bases that can be is already being financed through urban LGA budgets and tapped to pay for urban infrastructure. However, Tanza- direct spending by utilities, line ministries and other agen- nian LGAs have experienced the progressive elimination cies in urban areas. of several own-source revenue instruments (including the elimination of business licenses as a significant source of The current framework for local revenues and intergov- revenue in urban LGAs) and have not been very efficient ernmental transfers cannot accommodate the signifi- at tax collection. cant financing needs of Tanzanian cities. Urban LGAs do not have sufficient revenues to invest significantly in Own Source Revenues new infrastructure. In the 2006/07 fiscal year, total urban LGA revenue (including both own-source revenue and LGA revenue comes from both own-source revenues intergovernmental transfers) was only $189 million. Total and intergovernmental transfers. In recent years, there expenditure was $153 million, of which only 28 million has been an increase in overall revenues and expenditures (18 percent) was spent on capital investment. In Tanzania, at the LGA level and an increasing reliance on central gov- most national taxes are collected from urban areas while ernment transfers relative to own-source revenues. Total most intergovernmental transfers go to rural areas. A high LGA revenue (including own-source revenue and inter- proportion of the domestic budget comes from revenues governmental transfers) increased dramatically from Tshs that are generated in urban areas. For example, in 2005/06 255 billion in 2001/02 to Tshs 920 billion in 2006/07. the domestic budget was nearly Tshs 2 trillion, of which Intergovernmental transfers increased from 79 percent of 83 percent came from Dar es Salaam alone. In contrast, total LGA revenues in 2001/02 to 93 percent in 2006/07 the lion’s share of fiscal transfers is directed to rural LGAs. while own-source revenue decreased from 21 percent of For example, total intergovernmental fiscal transfers in total LGA revenues in 2001/02 to 7 percent in 2006/07. 2006/07 were Tshs 859,468 million, of which urban The shares of total LGA revenue are roughly 20 percent LGAs received 18 percent and rural LGAs received 82 urban and 80 percent rural (see table 7 below). Table 7: Source of Revenues of Urban and Rural LGAs, 2006/07 Total Urban LGAs Urban Share Rural LGAs Rural Share Total Revenue (Tshs millions) 920,879 189,209 20.5% 731,671 79.5% Own-Source Revenue (Tshs millions) 61,411 34,870 56.8% 26,541 43.2% Intergovernmental Transfers (Tshs millions) 859,468 154,339 18.0% 705,130 82.0% Population in 2002 (people) 33,461,849 7,554,838 22.6% 25,907,011 77.4% Source: Authors’ calculations based on LOGIN data and 2002 census data. 13 14 Financing the Urban Expansion in Tanzania Urban LGAs derive 18 percent of their total revenues tives for the creation and growth of private enterprises. from own-source revenues. Prior to 2003, LGAs in Due to the perceived inefficiency and equity impacts of the Tanzania were empowered to define their own local tax local tax system, a series of “reforms� were implemented in structure. The Local Government Finances Act of 1982— 2003 and 2004. These significantly reduced the revenue which defines the financial framework for LGAs in Tanza- raising authority of LGAs and minimized the importance nia—allowed an ‘open list’ or permissive approach to local of local revenues in the overall intergovernmental fiscal taxation. This gave LGAs wide powers to impose taxes, framework. Currently, LGAs derive their own-source rev- levies and fees and set rates within their local jurisdictions. enues from a handful of instruments. Urban LGAs derive By 2003, the number and type of local revenue instru- 18 percent of their total revenues from own source rev- ments varied significantly from one LGA to another and enues. More than 80 percent of those revenues come from there were also large variations in the rates imposed by just three sources: the service levy (38.5 percent), fees and LGAs on similar tax bases. Local taxes were criticized for charges (28.4 percent) and property rates (15.0 percent) being inefficient and regressive and for creating disincen- as can be seen in table 8 below. Table 8: Urban LGA Own-Source Revenue Sources, 2006/07 Total Property Produce Service Hotel Fees and Other own rates Land rent cess levy levy Licenses charges revenues revenues Revenue (Tshs millions) 5,216.5 734.6 201.1 13,414.5 643.2 520.0 9,889.4 4,251.4 34,870.3 Share 15.0% 2.1% 0.6% 38.5% 1.8% 1.5% 28.4% 12.2% 100.0% Source: Authors’ calculations based on LOGIN data. Intergovernmental Transfers zania has implemented a unified system for the provision of capital funding to LGAs called the Local Government In Tanzania 18 percent of total intergovernmental Development Grant System (LGDG). LGDG provides transfers is allocated to urban LGAs, and this accounts both capital development grants (CDG) to assist LGAs to for more than 80 percent of their total revenue. Taken as construct new infrastructure and to rehabilitate existing whole, almost 95 percent of all urban and rural LGA rev- one and capacity building grants (CBG) to improve the enues come from intergovernmental transfers. The central capacity and performance to meet the CDG minimum government provides four different types of intergovern- conditions. In addition, to these discretionary transfers, mental transfers to LGAs: (1) recurrent block grants, con- LGDG provides several sector specific capital transfers. sisting of a series of formula-based transfers in a number of These are summarized in table 9. grant-aided sectors, including the agriculture, education, health, roads and water sectors; (2) unconditional General Recurrent transfers account for the majority of inter- Purpose Grant (GPG), providing funds for council admin- governmental transfers. For all LGAs, recurrent transfers istration and to equalize councils with low revenue bases; account for 84 percent and capital transfers for 16 percent. (3) other recurrent transfers (subventions) from ministries, In urban LGAs the proportion is very similar: 82 percent departments and agencies for recurrent expenditures in a of recurrent transfers and 18 percent of development number of sectors including agriculture, health, roads and transfers. These percentages account for Tshs 126,429.6 HIV/AIDS15; and (4) capital development grants and funds. million in recurrent transfers and Tshs 27,909.1 million With regard to capital transfers, the Government of Tan- in capital transfers (see table 10). In per capita terms this 15 Agriculture: Agriculture Sector Development Program (ASDP), Health Sector Basket Fund (HSBF), Roads: Roads (Fuel Levy) Fund, HIV/ AIDS: TACAIDS and Global Fund Subventions. Current Spending on Infrastructure in Urban Areas 15 means that in 2005/06 each urban resident received 2,854 alone. The remaining 30 percent was distributed amongst Tshs or US$2 in capital transfers. With regard to the sec- water (2.8 percent), roads (4.2 percent), agriculture (2.2 toral allocation of transfers, the most significant grants percent), a general purpose/administrative grant (GPG) are for the education and health sectors. In 2005/06, 70 and the LGDG. percent of the total transfers went to those two sectors Table 9: Intergovernmental Fiscal Transfers in Tanzania, FY2005/06 Budget item Tshs millions Proportion of Total Recurrent Block Grants Education Grant 295,335 Health Grant 67,890 Agriculture Grant 7,516 Roads Grant 6,666 Water Grant 11,214 GPG (incl. local admin) 64,211 Total Grants 452,831 71.81 Sector Basket Funds and Subventions Education Subventions 27,297 Health Subventions 19,159 Roads Subventions 11,576 HIV/AIDS Subventions 5,934 TACAIDS Subventions 2,106 Other Subventions 11,107 Total Subventions 77,180 12.24 Development Grants and Funds LGDG 34,493 Education 18,585 Health 6,044 Roads 8,479 Water 6,512 Agriculture 6,422 Local Admin. 3,029 TASAF 2,390 Other Capital Funds 14,648 Total Capital Funds 100,602 15.95 Total (Recurrent plus Development) Transfers Recurrent Transfers 530,011 84.04 Capital Transfers 100,602 15.95 Total Transfers 630,613 Source: Authors’ calculations based on data from PMO-RALG, Consolidated LGA Finance Statistics, FY 2005/06 (LOGIN) 16 Financing the Urban Expansion in Tanzania Table 10: Urban LGAs Intergovernmental Transfers, 2006/07 Recurrent Transfers Development Grants Total Transfers Transfers Share Transfers Share Transfers 126,429.6 81.92% 27,909.1 18.08% 154,338.6 Source: Authors’ calculations based on LOGIN data. Other Sectoral Spending of the LGAs budgets. The development budget, which is applied to the capital requirements for the road network, In addition to LGA spending, line ministries and other does not include own-source revenues, capital grants (i.e., agencies also spend directly in urban areas, i.e., off LGDG) and funding for roads infrastructure from other LGA budgets. Such spending is notable in the cases of wa- development programs (e.g., agricultural development ter, roads and electricity and therefore needs to be taken programs). However, most of the development budget into account to assess the total resources being spent in is spent on rural roads, and only US$0.94 million goes urban areas. Sectors do not always record data for specific for the rehabilitation of urban roads (United Republic of locations, and their definitions of “urban� are not always Tanzania 2008, 40-41). consistent across sectors. Therefore, the following analy- sis uses a number of simplifying assumptions to arrive at In the water sector, capital investments in urban wa- reasonable estimates that represent orders of magnitude ter supply and sanitation amounted to US$41 million rather than precise figures. in FY2007/08. In addition, about US$29 million went toward recurrent expenditures, bringing overall spending In the roads sector, there is limited direct spending by in the sector to US$70 million.17 The majority of expen- the Ministry of Infrastructure Development (MID) diture in the water sector comes from three sources: (1) in urban roads. The Local Government Transport Pro- Ministry of Water and Irrigation (MOWI); (2) LGAs; gramme (LGTP) establishes a budget of US$64.7 million and (3) regions. In terms of recurrent expenditures at the in FY 2007/08 for the maintenance and development of central level, these have been declining mainly because a LGA roads including US$51.7 million for maintenance, larger proportion of the expenditure is going through the US$9.5 million for development, and US$3.4 million regions and other ministries. The development budget has for capacity building (United Republic of Tanzania 2008, increased rapidly since 2000. Its composition has changed 41). Some items appear in the annual budgets of the significantly, partially because the decentralization policy LGAs and some in the annual budgets of PMO-RALG has increased the share of local and regional governments and MID. If the total maintenance budget is allocated at the expense of MOWI. The LGAs share in water sec- to urban areas according to the length of the urban road tor funding has increased from 17 percent in 2005 to 25 network, then maintenance spending in FY 2007/08 in percent in 2008. Most of the funding from LGAs and the urban areas was approximately US$5.3 million.16 How- regions is dedicated to rural water. Conversely, the largest ever, the road maintenance budget is financed through proportion of the MOWI’s budget is allocated to urban the Road Fund and is therefore already counted as part water supply. 16 The total road network length in Tanzania is 85,000Km. Out of the total, 28,891km cover trunk and regional roads. The other 56,625km covers “local government roads�. From that, 29,536km are district roads, 21,191km are feeder roads and only 5,898km are urban roads. 17 The amount recorded as recurrent expenditure comes from block grants mainly going to the financing of recurrent expenses. However, these grants are not earmarked for these expenses so there is a possibility that they might partly finance capital invest- ments. There is no data on the composition of these transfers and it is believed that that proportion going to capital investments is very small which suggests that the sector may actually be spending slightly more on capital investment than is estimated. Current Spending on Infrastructure in Urban Areas 17 The central government and TANESCO spend US$42 allocated share of the spending of key sectors directly in million a year in capital investments in energy in urban urban areas. Capital spending represents one third of the areas. In Tanzania, energy is the sector that spends the most total spending in urban areas, which is close to US$350 on infrastructure (close to 3 percent of GDP) at about 41 million a year. The other two thirds go to recurrent ex- percent of total infrastructure spending. From that, Tanza- penses. Note, however, that while a high proportion of nia’s central government spends 36 percent on capital in- sectoral spending in urban areas goes to finance develop- vestments, 9 percent on wages and 55 percent on other cur- ment expenses, development spending accounts for only rent expenditures. The spending structure of TANESCO 20 percent of LGA budgets. The capital spending in ur- is similar: 30 percent on capital, 4 percent on wages and ban areas is equivalent to roughly US$10 per capita (as- 55 percent on other current expenditures.18 Taken together, suming 11 million urban residents).19 According to the the central government and TANESCO’s overall spend- previous section the total investment required to keep up ing accounts for close to 3 percent of the country’s GDP with urbanization is estimated at US$1,070 million per (Briceno-Garmendia and Foster 2007). If one prorates that annum. Subtracting US$111 million (what is currently amount across the total urban population which has ac- being spent yearly) there is a net gap of US$959 million cess to the electricity network, then approximately US$119 per annum. This estimate does not take into account bulk million is spent on energy in urban areas. infrastructure costs including the cost of electricity gen- eration (which will be much higher than electricity dis- In aggregate, current spending on infrastructure in ur- tribution), and so the actual investment gap will be much ban areas amounts to roughly US$111 million a year. higher. The final section of the paper considers some pos- table 11 combines the total urban LGA budget with an sible strategies to begin to close this gap. Table 11: Total spending in urban areas (US$) Source Development Recurrent Total LGAs* 27,717,844 125,563,054 153,280,898 Water** 41,315,955 29,076,141 70,392,097 Roads*** 940,000 5,387,100 6,327,100 Energy**** 41,514,738 77,363,615 118,878,353 Total 111,488,537 237,389,910 348,878,448 * Data from 2006/07. Authors’ calculations based on LOGIN data. ** Data from 2007/08 (World Bank 2008). *** Data from 2007/08 (United Republic of Tanzania 2008) and (United Republic of Tanzania 2008). **** Data from 2000-2005 annual average (Briceno-Garmendia and Foster 2007). 18 TANESCO presents a large quasi-fiscal deficit (this is calculated by adding up fiscal transfers and tax exemptions to any operating losses) amounting to 1.8 percent of GDP. The bulk of the deficit is related to under-pricing. 19 This is consistent with data collected from a sample of African cities, which suggests that the average expenditure per person per year hardly exceeds $10. South Africa is an exception, with Cape Town at $1,163 and Durban at $1,152. (World Bank 2009) 3 Conclusions and Recommendations Urban LGAs should be spending more on infrastruc- to strengthen the framework for resource mobilization and ture. Demographic projections suggest that urbanization infrastructure financing in urban areas. in Tanzania will continue at a rapid pace. While the exact predictions vary, the trend is incontrovertible. This is a It is much cheaper in the long run to proactively plan positive development for Tanzania since agglomeration for new urban development than to retrofit infrastruc- economies and concentration are critical for productivity ture to rapidly growing unplanned settlements. While growth and national prosperity and can benefit popula- estimates of the cost of dealing with new urban growth tions in both urban and rural areas. However, to realize this may be daunting, the “do-nothing scenario� is likely to be potential, Tanzanian cities need to make significant invest- much more costly. Putting aside the obvious imperative ments to expand service coverage and to avoid problems to deal early with the problems of slum formation, over- of slum formation, urban poverty and congestion, which crowding and urban poverty, and focusing on the direct constrain urban productivity gains and act as bottlenecks costs of financing urban infrastructure, studies suggest to growth. Presently, Tanzanian cities may be trapped in that the cost of upgrading existing informal settlements a vicious circle since their lack of infrastructure prevents is much higher than the cost of providing infrastructure them from converting the expanding urban populations to new urban areas, possibly in the order of two to three into a broader tax base that could generate resources that times the cost of servicing a formal-sector subdivision in turn could be used to finance urban infrastructure. The (Akibo 2007). low and stagnant coverage of household services in Tan- zania’s urban areas (see figure 3) exerts a major social and To expand service coverage at a faster rate, a more prag- economic toll. The annual cost of investment in urban matic approach would be to target minimum (basic) infrastructure and services to accommodate Tanzania’s or average service levels rather than full service levels. growing urban population is in the order of US$1,070 The World Bank’s Arica Infrastructure Country Diag- per annum, and the previous sections estimates that the nostic notes that few countries are expanding coverage at financing gap is in the order of US$959 million per an- rates high enough to outstrip urbanization. Most African num. While the scale of these investment requirements countries have pursued a strategy of providing heavily is daunting, the corresponding return on investment in subsidized first-best options, such as piped water and elec- terms of urban productivity gains and resource generation tricity, which has tended to bankrupt and debilitate sector are potentially enormous. institutions without bringing about any significant ac- celeration of coverage. In addition, public subsidies have Several key strategies could be pursued to close the in- largely bypassed the neediest groups (World Bank 2009, frastructure financing gap in urban areas. This section 34). The analysis in this paper corroborates these conclu- proposes a number of potential strategies to begin to close sions, demonstrating that the costs of servicing a new ur- this gap, focusing on recommendations supported by the ban area are sensitive to the level of services provided, in analysis in this paper. The central recommendations of the particular the standard of road construction. For example, paper are to manage the cost of the urban expansion by: the cost of road construction in a neighborhood ranges (1) proactively dealing with new urban growth; (2) target- from Tshs 4.1 billion, if all roads are gravel engineered, ing lower average service levels; and (3) adopting smaller to Tshs 20.7 billion, if the roads are paved to bitumen minimum plot sizes. The paper also calls for greater efforts standard. 19 20 Financing the Urban Expansion in Tanzania A key recommendation of this paper is to adopt small- quired to accommodate urban growth. Even on very con- er minimum plot sizes for urban development which servative assumptions, LGA capital spending in urban would significantly reduce the cost of future urban areas is tiny in relation to the financing needs, in the order growth. The costing exercise illustrates the financing im- of US$2.50 per capita (US$27.7 million for approxi- plications of varying residential densities or plot sizes, and mately 11 million urban residents). In order to meet the demonstrates that the cost of expanding infrastructure is significant requirements for infrastructure investment, a highly sensitive to population density. figure 6 shows that number of financing alternatives could be explored. In the smaller the plot size (i.e., the denser the settlement), terms of own source revenues, serious efforts need to be the lower the per plot cost of providing infrastructure made to improve the efficiency of existing local revenue and services. This is because the network or bulk costs are instruments while at the same time exploring other new spread across a larger number of plots. For smaller plot or expanded revenue sources including user fees. Once sizes of 100m2 and 200m2, the costs of providing infra- new or expanded revenue sources are in place, with an structure and services per plot (particularly at minimum appropriate regulatory framework, responsible local bor- and average service levels) are much more affordable. For rowing in Tanzania could be possible in selected LGAs. example, for a 200m2 plot, the cost of providing a mini- In addition, since central government transfers play and mum package of services (i.e., communal water points, will continue to play a central role in local government ventilated pit latrine, solid waste transfer points, electric- financing, attention should be given to ensuring an equi- ity, gravel engineered roads and lined storm water drains) table and efficient transfer system that does not discourage would not exceed Tshs 2 million.20 The 400m2 plots cur- local efforts to generate revenue. The literature on local rently considered as high density could be categorized as infrastructure financing is broad and familiar, but an over- medium density while above 800m2 could be low den- view of the main instruments in place or with potential sity.21 A minimum plot size of 100m2 to 200m2 should be to finance the urban expansion in Tanzania is provided considered. in table 12 below. Finally, since urban space is not solely under the control of local authorities but rather a spatial On the revenue side, there is a spectrum of possible territory in which numerous actors interact, it would also strategies to mobilize financing for infrastructure de- be important to coordinate the financial contributions of velopment. Tanzanian LGAs currently lack the resource other actors, such as utilities and line ministries that also base to adequately finance the infrastructure that is re- invest in urban areas. 20 Private developers may be prepared to meet this cost, especially if the amount could be paid by installments over time. However, considering affordability as a dynamic reality, infrastructure ought to be provided at a standard and in such a way that allows staged upgrading as affordability increases, at the same time ensuring less operational and maintenance cost. It is for this reason that earth roads and unlined storm water drains are not considered as a realistic scenario and standard of urban development. Providing a pit latrine where it may pollute underground water is a more expensive option than providing a sewerage system. 21 Infrastructure and service package for medium density and low density could include: bitumen/gravel standard roads and lined storm water drains, water supply at a standard of house connection, electricity connection, house to house collection of solid waste, and a sewerage system. Currently such areas use septic tank and soak away system, the cost of which is at least ten times the cost of a sewer system. Besides, some of the septic tanks and soak away pits do not work properly because of high water table among the reasons. Conclusions and Recommendations 21 Table 12: Infrastructure Financing Mechanisms Financing Mechanism Comments Central It is likely that for the foreseeable future transfers from the central government, which currently account for government more than 80 percent of LGA budgets in Tanzania, will remain the most signi�cant source of funding for local transfers infrastructure. It is critical that transfers are adequate to cover the provision of services for which LGAs are responsible. A targeted urban grant could be provided, either through the LGDG system or outside of it. Such a grant should address all urbanizing areas of Tanzania, not just those that are already within the boundaries of an urban LGA. Alternatively, the existing transfer formulae could be revisited to take into consideration urban investment needs and urban revenue potential. Local taxes There is scope to signi�cantly improve the ef�ciency of existing own-source revenue instruments, in particular the property tax and city service levy, which are not living up to their potential. Although the share of prop- erty taxes increased from 6 percent in 2002 to 10 percent in 2005/06, it still accounts for a relatively small proportion of own-source revenue. In most urban LGAs, many properties are not on the valuation rolls, and efforts could be made to update valuation rolls focusing on high value properties. Urbanizing areas outside of urban LGAs could also be brought into the property tax net of LGAs, either by encouraging neighboring “rural� LGAs to collect property taxes or by expanding the boundaries of urban LGAs to reflect actual urban settlement patterns. In addition, Tanzania’s property rates are much lower than in many countries, and there is scope to increase rates signi�cantly. To protect the poor, properties under a certain value could be completely exempted from property tax, or a tax free threshold could be applied to all properties. In addition to the increase in tax rates, local revenue bases should be expanded. An enormous and expand- ing number of transactions take place informally without directly contributing to the tax economy. Thus, there is signi�cant potential to enlarge the tax base. Special district and/or special assessment �nancing could be explored so that those who bene�t most from infrastructure and services help pay for those investments. This mechanism is especially popular in many countries in business districts and in upper- or middle-income areas, freeing up other public resources to �nance investments that bene�t the poor. Development impact fees and capital contribution fees could be instituted to require commercial, industrial and high-end residential development to pay its way. Further efforts could also be made to explore other new or expanded local revenue sources, such as an ap- propriate business tax. Municipal Once new or expanded revenue sources are in place, and LGAs and other service providers are able to gen- debt erate revenues beyond those needed for current operations, responsible local borrowing would be possible, within an appropriate regulatory framework. In a number of decentralized countries, borrowing for local infrastructure is seen as ef�cient and equitable. There are a number of technical justi�cations for its use. First, because borrowing is in many cases the eco- nomically appropriate way to �nance capital outlays. Technically, in terms of both allocative ef�ciency and intergenerational equity, it often makes sense to �nance long-term investment projects by borrowing rather than relying on current public savings or transfers. Second, �nancing by means of capital markets frees LGAs from the uncertainties of central governments grants and loans. Third, in many countries, borrowing seems the only �nancing mechanism available because it is clear that neither sub-national governments’ savings on current account, nor central government grants, nor private sector provision, nor foreign aid will suf�ce to meet infrastructure investments needs. Last, the operation of credit markets provides (at least in theory) a much needed disciplining mechanism at the local level: sub-national government �scal performance will be remunerated or penalized by the market. In practice, however, great care must be taken before LGAs can freely borrow as a loosely regulated bor- rowing framework can result in �scal indiscipline and eventually in macroeconomic imbalance. The main principle that must be observed is to prohibit borrowing to �ll current budgetary de�cits. Borrowing to cover current account de�cits has just the opposite intergenerational effects of paying for capital expenditures from current revenues. It shifts to the future the costs of services enjoyed by today’s taxpayers. continues 22 Financing the Urban Expansion in Tanzania Table 12: Infrastructure Financing Mechanisms (continuation) Financing Mechanism Comments Municipal At the moment, however, the use of credit does not seem to be appropriate for Tanzania’s LGAs as credit debt (cont.) markets require mechanisms of control that are yet inexistent. There is a general lack of reliable data; systems for disclosing general information and audits of the public accounts are weak; �nancial markets are fragile; and there is a lack of instrument to assess �nancial risks. Direct central Electric, water, and telecommunication service investments in most urban areas are already �nanced by provision separate entities, and this approach could be expanded to other cost recovery (or fee �nanced) services such as solid waste and transport. These entities could be publicly or privately owned. Public Private Exploit potential for public-private partnerships and other modalities for private sector �nancing of infra- Partnerships structure. The use of public-private partnerships is one additional mechanism to �nance local infrastructure. Generally, in such partnerships, the private sector assumes a big proportion of the original investment and risk in exchange for future gains. The public sector makes some concessions generally in the form of land or tax bene�ts to enter into the partnership. The potential of such partnerships seems great, but in practice there is still a lot to be learned. Such partnerships require a substantial change in the nature of the LGAs from being service providers to become contract managers of sometimes highly complex contracts. Land-Based The underlying principle of this �nancing mechanism is that the bene�ts of infrastructure projects are capital- Financing ized into land values. Thus, as long as the spatial distribution of project bene�ts can be internalized within a well de�ned bene�t zone, it is economically ef�cient to �nance infrastructure projects by tapping the incre- ments in land values resulting from them. Some advantages of this mechanism include its capacity to generate revenue up front, reducing dependence on debt, and also developing a system of impact fees differentiated by the levels of infrastructure costs to be incurred—rather than using general revenues. In Tanzania, these instruments have not yet been put to practice, but in a speci�c situation and with the proper regulation, they can become viable tax mechanisms. There are numerous instruments for land �nancing, including: betterment levies, developer land sales, sale of development rights, etc. (For a full discussion on land-based �nancing, see Peterson 2009). Land-based �nancing is a potentially good mechanism in cases of rapid urban growth, like that seen in Tanzania, as land prices tend to rise rapidly, creating opportunities to generate signi�cant revenue. In these cases of rapid urban growth, urbanization usually expands into the surrounding fringe converting rural into urban land. As we have seen, converting rural land into urban serviced land is quite expensive, and land- based �nancing is a good mechanism to �nance the investments. In addition to �nancing the provision of new urban serviced land, these instruments can be used for the �nancing of basic urban services, such as water supply, drainage and street pavement. One way in which this has proven to be effective is by earmarking the proceeds from public land sales for the municipal capital budget and even to speci�c budget items such as water, drainage, etc. Despite the revenue generating potential of these instruments, they should be used with care, both because land transactions are prone to occur without full transparency and accountability (off budget) and because at the end of the day, land sales cannot continue inde�nitely, so smart, time-based decisions about present versus future investments should be made. Source: Africa Infrastructure Country Diagnostic (2009, 83), Peterson (2009) and authors’ analysis. Annex: Derivation of Unit Costs The derivation of unit costs is drawn from a background and contractors involved in the implementation of vari- paper prepared for the World Bank by Professors Nnkya ous infrastructure projects. The exercise also draws on the and Kombe of Ardhi University, Tanzania. ‘20,000 plots project’ in Dar es Salaam. Unit costs for infrastructure were obtained from vari- Land Preparation ous sources. These included Ministry of Lands and Hu- man Settlement Development, Ministry of Infrastructure Table 13 sets out the unit costs used for valuation, com- Development, National Construction Council, TANES- pensation, land use planning and cadastral survey. Once CO, Dar es Salaam Water and Sewerage Corporation, land has been identified as suitable for a planning scheme, Dar es Salaam Water Supply Company, Dar es Salaam a system of land compensation is implemented. This is in City Council, the Community Infrastructure Upgrad- order that the existing interests on the land can be cleared. ing Project (CIUP) in Dar es Salaam, and consultants Table 13: Unit Costs for Land Preparation Cost per neighborhood Cost per hectare Cost per 400m2 plot (Tshs) (Tshs) (Tshs) Valuation 550,000,000 4,000,000 250,000 Compensation 618,750,000 4,500,000 281,250 Land use planning 2,750,000 20,000 1,250 Cadastral survey 37,510,000 272,800 17,050 Total 1,209,010,000 8,792,800 549,550 Roads and Drainage ity lines. The estimates take into consideration the cost of clearing rights of way, removing top soil, replacing it with The cost of road construction is influenced by a number appropriate materials and raising the road profile above of factors including: (i) space standards; (ii) construction surrounding ground levels. Considering future mainte- standards; (iii) topography; (iv) presence of existing util- nance requirements, the basic minimum standard of road ity lines (e.g., water and sewerage pipes, electricity and construction in terms of materials is taken to be gravel telephone lines) which might need to be relocated; and (v) engineered with lined storm water drains. availability of construction materials. The unit costs used in this paper assume that roads and drains are provided Assuming an average construction standard of seven for the first time and there are no costs of relocating util- meters wide carriage way22, 1.5 meter shoulders on each 22 Prevailing standards for carriageways and rights of way for various categories of roads are shown in the table below (Kironde 2006, 465). Type of road Right of way (metres) Carriageway (metres) Primary distributors 30 – 40 7 – 10 District distributors 20 – 30 7 – 10 Local distributors 15 – 20 5 – 7.5 Access roads (shopping areas) 20 10 Access roads (industrial areas) 18 – 20 7 – 10 Access roads (residential areas) 10 – 20 5–7 Local distributors (residential areas) 10 – 20 6 Pedestrian access 10 2 23 24 Financing the Urban Expansion in Tanzania side and lined storm water drains, the estimated cost per it is established that the total length of all roads, provid- kilometer of road is between Tshs 0.8 – 1.0 billion for ing access to every plot, is 20.71 km, of which 7.58 km paved/bitumen road and between Tshs 0.15 – 0.2 billion serve as district and local distributor roads (with a right of for gravel engineered road. In view of the prevailing infla- way of 30 and 20 meters respectively) while the remaining tion rate of nearly 10 percent, the costing exercise uses the 13.13 serve as access roads (right of way of 15 or 10 me- upper limits of these estimates, i.e., Tshs 1.0 billion per ters wide). Assuming the basic minimum level of service kilometer for paved bitumen roads and Tshs 0.2 billion that all roads are gravel engineered, the cost for provid- for gravel engineered roads. With reference to a real life ing roads per hectare is Tshs 30 million. This implies Tshs case of a neighborhood (137.5 hectares) layout designed 1,882,727 per plot in high-density development. for a project to deliver residential plots in Dar es Salaam, Table 14: Unit Costs for Road Construction: All Gravel Engineered (“Minimum Service�) Kilometers per Cost per Cost per hectare Cost per 400m2 plot Roads neighborhood neighborhood (Tshs) (Tshs) (Tshs) Distributor 7.58 1,516,000,000 11,025,455 689,091 Access roads 13.13 2,626,000,000 19,098,182 1,193,636 Total roads 20.71 4,142,000,000 30,123,636 1,882,727 If all district/local distributors are paved or bituminized 3,445,455 per plot, and gravel engineered is Tshs 19 mil- while access roads are gravel engineered (the typical sce- lion per hectare or Tshs 1,193,636 per plot in high-density nario for urban areas in Tanzania), the cost per hectare development. This translates to a cost of Tshs 74 million for the bituminized roads is Tshs 55 million or Tshs per hectare or Tshs 4,639,091 per plot. Table 15: Unit Costs for Road Construction: Access Roads Gravel Engineered (“Average Service�) Kilometers per Cost per Cost per hectare Cost per 400m2 plot Roads neighborhood neighborhood (Tshs) (Tshs) (Tshs) Distributor 7.58 7,580,000,000 55,127,273 3,445,455 Access roads 13.13 2,626,000,000 19,098,182 1,193,636 Total roads 20.71 10,206,000,000 74,225,455 4,639,091 Assuming an ideal situation whereby all roads are paved/ is Tshs 0.15 billion implying a cost of Tshs 9,413,636 per bitumen standard, the cost for providing roads per hectare plot in high density (400m2) development. Table 16: Unit Costs for Road Construction: All Paved Bitumen (“Full Service�) Kilometers per Cost per Cost per hectare Cost per 400m2 plot Roads neighborhood neighborhood (Tshs) (Tshs) (Tshs) Distributor 7.58 7,580,000,000 55,127,273 3,445,455 Access roads 13.13 13,130,000,000 95,490,909 5,968,182 Total roads 20.71 20,710,000,000 150,618,182 9,413,636 The estimated cost of providing roads and storm water In arriving at unit cost estimates, a challenge has been the drains do not include costs of providing bridges and shift- general lack of up-to-date and standard unit rate for road ing of utility lines (water pipes, electricity and telephone construction and the fact that cost of road construction lines). According to the Tanzania Roads Agency (TAN- depends on many factors to include the source of con- ROADS) the cost of relocating existing infrastructure is struction materials, standard of construction, topography about 30 percent of the total cost of roads construction. of the area where a road is being constructed and whether there is infrastructure and utility lines to be relocated. Annex: Derivation of Unite Costs 25 Electricity Supply ing a substation with a cost of Tshs 18 million. The cost per hectare in this case is estimated to be Tshs 4.6 mil- It is assumed that in a typical neighborhood (137.5 ha), lion which translates to Tshs 290,591 per 400m2 plot. If electricity is supplied using 11 KV lines which cost Tshs a single phase service line which costs Tshs 385,619 per 30 million per kilometer. Assuming electric supply lines plot is added, the total cost per 400m2 plot is estimated along all the roads length of 20.71 kilometers, the total to be Tshs 676,210, while cost per hectare would be Tshs cost per neighborhood would be Tshs 639 million includ- 10.8 million. Table 17: Unit Costs for Electricity (Single Phase Line) Electricity Kilometers of line Cost per Cost per hectare Cost per 400m2 plot per neighborhood neighborhood (Tshs) (Tshs) (Tshs) KV lines 20.7 621,300,000 4,518,545 282,409 Substation 18,000,000 130,909 8,182 Single phase service line 848,361,800 6,169,904 385,619 Total 1,487,661,800 10,819,359 676,210 However, if a three-phase electricity connection is desired, three-phase connection fee per plot is Tshs 772,893. the cost will be accordingly higher, considering that a Table 18: Unit Costs for Electricity (Three Phase Line) Kilometers of line Cost per Cost per hectare Cost per 400m2 plot Electricity per neighborhood neighborhood (Tshs) (Tshs) (Tshs) KV lines 20.7 43,842,656 318,856 19,928 Substation 10,064,651 73,197 4,575 Three phase service line 2,040,436,200 14,839,536 927,471 Total 2,094,343,507 15,231,589 951,974 Water Supply kiosk for every 4 hectares, so as to limit walking distance to a water kiosk to a maximum of 100m, then 34 water For the purposes of this paper, two service levels are con- kiosks are needed. This implies a cost of Tshs 5,156,250 sidered: (1) communal water kiosks (less than 100 meters per neighborhood. To supply water to the public water from the dwelling); and (2) house connection (metered points, 10cm water pipes need to be installed along the water supply piped to the dwelling). local distributor roads (3km long) for a cost of Tshs 30 million per neighborhood. The total cost is therefore Tshs If communal water kiosks are used, the cost per kiosk, ex- 255,682 per hectare or 15,980 per plot in the case of a cluding distribution pipes, is Tshs 150,000. Assuming one high-density development. Table 19: Unit Costs for Public Water Kiosks Number of Cost per hectare kiosks per Cost per (one kiosk every Cost per 400m2 plot23 Water neighborhood neighborhood four hectares) (Tshs) Communal water kiosks 34.4 5,156,250 37,500 2,344 Water supply pipes 30,000,000 218,182 13,636 Total 35,156,250 255,682 15,980 23 16 plots per hectare. 26 Financing the Urban Expansion in Tanzania Cost for piped water connection per plot varies depend- between 15-30 meters depending on the location of a ing on the location of the house being connected. Assum- house in a particular plot. Assuming the upper end of this ing that there are water distribution pipes along all access range—30 meters from the connection point—the cost roads, the maximum distance for connecting water ranges per plot is Tshs 240,000 or Tshs 3.8 million per hectare. Table 20: Unit Costs for Water Reticulation (Excluding Trunk Infrastructure) Cost per neighborhood Cost per hectare Cost per 400m2 plot Water (Tshs) (Tshs) (Tshs) Reticulation costs24 528,000,000 3,840,000 240,000 Sanitation are in a high-density development, this translates to Tshs For the purposes of this paper, three sanitation options are 48 million per hectare. This amount is more than what is considered, depending on which, the cost can vary signifi- needed to provide a sewerage system per hectare. Depen- cantly: (1) ventilated pit latrine; (2) septic tank; and (3) dence on septic tank instead of a sewerage system is thus waterborne sewerage. The choice depends on population not a result of lack of affordability but lack of a strategy to density, soil and water table conditions and other factors. mobilize resources from developers and coordinate their servicing initiatives. A well-built ventilated pit latrine (the lowest standard of infrastructure for excreta disposal) is estimated to cost The estimated cost of providing sewerage pipes is Tshs 30 Tshs 250,000 per unit (i.e. per plot). Assuming 16 plots million per kilometer. Assuming sewer pipes are laid along per hectare in high-density development (400m2 plot), all 20.71 kilometers of road in a neighborhood, the total the cost per hectare is Tshs 4 million. cost per neighborhood would be Tshs 621,300,000 which translates to 4,518,545 per hectare and Tshs 282,409 per Where water borne sanitation is used, a septic tank and plot. Like all the other estimated costs, this estimate ex- soak away pit are commonly used. The current cost of cludes trunk sewerage infrastructure (including treatment building a septic tank and soak away pit is estimated to plants) which according to Dar es Salaam Water and Sewer- be Tshs 3 million per unit. Assuming 16 plots per hect- age Company could be about 30 percent of the unit rate. Table 21: Unit Costs for Sanitation Options Cost per Kilometers per neighborhood Cost per hectare Cost per 400m2 neighborhood (Tshs) (Tshs) plot (Tshs) Pit latrine25 550,000,000 4,000,000 250,000 Septic tank and soak away pit 26 6,600,000,000 48,000,000 3,000,000 Sewerage system 20.71 621,300,000 4,518,545 282,409 Solid Waste Disposal per unit, implying Tshs 80,000 per hectare or Tshs 5,000 The lowest level of solid waste disposal is a communal per plot. The same standard applies for an average level masonry container located within 250 meters walking of service while a full level of service would be house-to- distance. This translates to one container per 10 hectares. house collection by a local authority or an agency con- The cost of providing this is estimated to be Tshs 800,000 tracted to collect solid waste. 24 Assumes an average plot size of 400m2 25 Assumes an average plot size of 400m2 26 Assumes an average plot size of 400m2 Annex: Derivation of Unite Costs 27 Table 22: Unit Costs for Solid Waste Cost per Number of units neighborhood Cost per hectare Cost per 400m2 plot per neighborhood (Tshs) (Tshs) (Tshs) Communal masonry container 13.8 11,000,000 80,000 5,000 Summary Table 23: Summary of Unit Costs by Service Level % Total Level of Cost per Cost per service Type of service Description Hectare (Tshs) Hectare Land preparation Valuation, compensation, planning and survey 8,792,800 16% Roads and drainage Gravel engineered with lined storm drains 30,123,636 55% Water supply Public kiosk or stand pipe or deep well/bore hole 255,682 0% Basic Electricity supply House connection 10,819,359 20% Liquid waste disposal Ventilated Improved Pit latrine (VIP) 4,000,000 7% Solid waste disposal Communal masonry container 1,000,000 2% Total 54,991,477 100% Land preparation Valuation, compensation, planning and survey 8,792,800 6% Roads and drainage Access and local distributor roads: gravel engi- 74,225,455 51% neered with lined storm water drains / Local/district distributors: paved/bitumen standard with lined storm water drains. Average Water supply House connection (excluding trunk infrastructure) 3,840,000 3% Electricity supply House connection 10,819,359 7% Liquid waste disposal Septic tank and soak away pit 48,000,000 33% Solid waste disposal Communal masonry container 80,000 0% Total 145,757,613 100% Land preparation Valuation, compensation, planning and survey 8,792,800 5% Roads and drainage Paved/bitumen with lined storm water drains. 150,618,182 84% Water supply House connection (excluding trunk infrastructure) 3,840,000 2% Full Electricity supply House connection 10,819,359 6% Liquid waste disposal Localized or central sewer system 4,518,545 3% Solid waste disposal Door to door collection Not estimated Total 178,588,886 100% 28 Financing the Urban Expansion in Tanzania Table 24: Summary of Unit Costs by Plot Size in Tshs Cost per 100m2 Cost per 200m2 Cost per 400m2 Cost per 800m2 Cost per plot plot plot plot 1200m2 plot Service Service level (Tshs) (Tshs) (Tshs) (Tshs) (Tshs) Minimum 137,388 274,775 549,550 1,099,100 1,758,560 Land Average 137,388 274,775 549,550 1,099,100 1,758,560 preparation Full 137,388 274,775 549,550 1,099,100 1,758,560 Minimum 470,682 941,364 1,882,727 3,765,455 6,024,727 Roads Average 1,159,773 2,319,545 4,639,091 9,278,182 14,845,091 Full 2,353,409 4,706,818 9,413,636 18,827,273 30,123,636 Minimum 3,995 7,990 15,980 31,960 51,136 Water Average 60,000 120,000 240,000 480,000 768,000 Full 60,000 120,000 240,000 480,000 768,000 Minimum 393,609 530,914 676,210 966,801 1,315,510 Electricity Average 393,609 530,914 676,210 966,801 1,315,510 Full 393,609 530,914 676,210 966,801 1,315,510 Minimum 250,000 250,000 250,000 250,000 250,000 Liquid waste Average 3,995 141,205 282,409 564,818 903,709 Full 3,995 141,205 282,409 564,818 903,709 Minimum 1,250 2,500 5,000 10,000 16,000 Solid waste Average 5,000 5,000 5,000 5,000 5,000 Full 5,000 5,000 5,000 5,000 5,000 Minimum 1,256,923 2,007,543 3,379,467 6,123,316 9,415,934 Total Average 1,759,764 3,391,439 6,392,260 12,393,901 19,595,870 Full 2,953,401 5,778,712 11,166,805 21,942,992 34,874,415 Annex: Derivation of Unite Costs 29 Table 25: Total Cost in Tshs Million per 150,000 New Urban Households 100m2 plot 200m2 plot 400m2 plot 800m2 plot 1200m2 plot Service Service level (Tshs millions) (Tshs millions) (Tshs millions) (Tshs millions) (Tshs millions) Minimum 20,608 41,216 82,433 164,865 263,784 Land acquisition Average 20,608 41,216 82,433 164,865 263,784 and preparation Full 20,608 41,216 82,433 164,865 263,784 Minimum 70,602 141,205 282,409 564,818 903,709 Roads Average 173,966 347,932 695,864 1,391,727 2,226,764 Full 353,011 706,023 1,412,045 2,824,091 4,518,545 Minimum 599 1,199 2,397 4,794 7,670 Water Average 9,000 18,000 36,000 72,000 115,200 Full 9,000 18,000 36,000 72,000 115,200 Minimum 59,041 79,637 101,431 145,020 197,326 Electricity Average 59,041 79,637 101,431 145,020 197,326 Full 59,041 79,637 101,431 145,020 197,326 Minimum 37,500 37,500 37,500 37,500 37,500 Liquid waste Average 599 21,181 42,361 84,723 135,556 Full 599 21,181 42,361 84,723 135,556 Minimum 188 375 750 1,500 2,400 Solid waste Average 750 750 750 750 750 Full 750 750 750 750 750 Minimum 188,539 301,131 506,920 918,497 1,412,390 Total Average 263,965 508,716 958,839 1,859,085 2,939,380 Full 443,010 866,807 1,675,021 3,291,449 5,231,162 29 you did previously and I’ll have another to add, or I can send you the full list - can you do a search and replace in the text so that when a table or figure is refer- enced, it is in lower case; e.g., page 1, para 1, “see figure 1� becomes “see figure 1� 30 Financing the Urban Expansion in Tanzania -- my mistake that I didn’t catch it before sending to you. That’s all I have. We may get more comments form the author as well. References Africon. (2007). Cost Estimate for Urbanisation in Dar es Peterson, George E. (2009). Unlocking Land Values to Salaam. Pretoria, South Africa. Finance Urban Infrastructure. Washington, DC, USA: World Bank (PPIAF). Akibo, A. (2007). Basic Cost of Slum Upgrading in Bra- zil. Global Urban Development, vol. 3, no. 1. Online: Population Division of the Department of Economic and http://www.globalurban.org/. Social Affairs of the United Nations Secretariat. (2007). The 2006 Revision and World Urbanization Prospects: Bocquier, Philippe. (2005). World Urbanization Pros- The 2007 Revision. World Population Prospects. http:// pects: An Alternative to the UN Model of Projection esa.un.org/unup (accessed August 27, 2008). Compatible with the Mobility Transition Theory. Demo- graphic Research, vol. 12, no. 9, 197-236. The World Bank. (2008). Rapid Budget Analysis for An- nual Review Synoptic Note. Technical Note. Washington, Bongaarts, John. (2001). Household Size and Composi- DC, USA: The World Bank. tion in the Developing World. Population Studies, vol. 55, no. 3, 263-279. _______ (2009). Africa Infrastructure Country Diagnos- tic Part 1. Washington, DC, USA: The World Bank. Briceno-Garmendia, Cecilia, and Vivien Foster. (2007). More Fiscal Resources for Infrastructure? Evidence from _______ (2009). The Urban Transition in Mainland Tan- East Africa. Washington, DC, USA: The World Bank. zania. Washington, DC, USA: The World Bank. Kironde, J.M. Lusugga. (2006). The regulatory frame- United Republic of Tanzania. (2008). 10 Year Transport work, unplanned development and urban poverty: Find- Sector Investment Programme (TSIP) Phase I 2007/08 ings from Dar es Salaam, Tanzania. Land Use Policy, vol. – 2011/12. Dar es Salaam, Tanzania: Ministry of Infra- 23, no. 4, 460–472. structure Development. Lugoe, Furaha Ngeregere. (2008). Assessment of Main _______ (2008). Local Government Transport Pro- Urban Land Use Issues in Tanzania. Washington, DC, gramme (LGTP) Phase 1 (2007-2012). Prime Minister’s USA: The World Bank. Office Regional Administration and Local Government. 31 Financing the Urban Expansion in Tanzania 33 Previous knowledge papers in this series Lessons and Experiences from Mainstreaming HIV/AIDS into Urban/ Water (AFTU1 & AFTU2) Projects Nina Schuler, Alicia Casalis, Sylvie Debomy, Christianna Johnnides, and Kate Kuper, September 2005, No. 1 Occupational and Environmental Health Issues of Solid Waste Management: Special Emphasis on Middle and Lower-Income Countries Sandra Cointreau, July 2006, No. 2 A Review of Urban Development Issues in Poverty Reduction Strategies Judy L. Baker and Iwona Reichardt, June 2007, No. 3 Urban Poverty in Ethiopia: A Multi-Faceted and Spatial Perspective Elisa Muzzini, January 2008, No. 4 Urban Poverty: A Global View Judy L. Baker, January 2008, No. 5 Preparing Surveys for Urban Upgrading Interventions: Prototype Survey Instrument and User Guide Ana Goicoechea, April 2008, No. 6 Exploring Urban Growth Management: Insights from Three Cities Mila Freire, Douglas Webster, and Christopher Rose, June 2008, No. 7 Private Sector Initiatives in Slum Upgrading Judy L. Baker and Kim McClain, May 2009, No. 8 The Urban Rehabilitation of the Medinas: The World Bank Experience in the Middle East and North Africa Anthony G. Bigio and Guido Licciardi, May 2010, No. 9 Cities and Climate Change: An Urgent Agenda Daniel Hoornweg, December 2010, No. 10 Memo to the Mayor: Improving Access to Urban Land for All Residents – Fulfilling the Promise Barbara Lipman, with Robin Rajack, June 2011, No. 11 Conserving the Past as a Foundation for the Future: China-World Bank Partnership on Cultural Heritage Conservation Katrinka Ebbe, Guido Licciardi, and Axel Baeumler, September 2011, No. 12 Guidebook on Capital Investment Planning for Local Governments Olga Kaganova, October 2011, No. 13 Knowledge PAPeRS This publication has been financed in part by the Cities Alliance. For more information about the Urban Development Series, contact: Urban Development and Local Government Unit Sustainable Development Network The World Bank 1818 H Street, NW Washington, DC 20433 USA Email: Urbaninfo@worldbank.org Website: www.worldbank.org/urban/cipguidebook January 2012, No. 15