68829 INDIA: Municipal Financing Requirements – Water, Sewerage and Solid Waste INDIA: Municipal Financing Requirements – Water, Sewerage and Solid Waste June 2010 SOUTH ASIA URBAN & WATER UNIT BACKGROUND STUDY FOR HIGH POWERED EXPERT COMMITTEE 1 INTRODUCTION 3 A. MUNICIPAL FINANCING NEEDS: WSS AND MSW SECTOR 5 B. URBAN WATER SUPPLY SERVICES 21 C. URBAN SEWERAGE SERVICES 40 D. MUNICIPAL SOLID WASTE 52 E. INVESTMENT NEEDS -A PART OF A COMPLEX PICTURE 65 ANNEX I: METHODOLOGICAL APPROACH 68 ANNEX II: SAMPLE DESCRIPTION 85 ANNEX III: INDIA URBAN POPULATION FORECASTS 86 ANNEX IV: MAIN RESULTS 93 ANNEX V: PROFESSIONALIZATION FOR SUSTAINABLE WSS DEVELOPMENT 99 ANNEX VI: CROSS-COUNTRY COMPARISONS 101 2 Acknowledgements This study was conducted as part of the Non-lending Technical Assistance (NLTA) to the High Powered Expert Committee (HPEC) on Urban Development, as an input to the HPEC estimation of expenditure (investment and O&M) requirements for urban Water Supply and Sanitation (WSS) and Municipal Solid Waste (MSW). The study has been prepared by a team comprising Elisa Muzzini (TTL), Gabriela Aparicio and Karan Rajput with contributions from Bill Kingdom, Shruti Garga, Deena Magnall, Vasudha Sarda, Martina Tonizzo under the supervision of Junaid Ahmad. The cost models have been developed in close collaboration with Dr. Isher Alwhawalia (HPEC Chair), Ramesh Ramanathan and Prof. Om Mathur (HPEC members), Ranesh Nair and Shubhagato Dasgupta (HPEC consultants), Prof. Usha Ragupathi and Chetan Vadya (National Institute of Urban Affairs) and technical support from a panel of experts comprising Maruthi Mohan, Bill Kingdom, Oscar Alvarado, N. V.V. Raghava, Guillermo Yepes for WSS and Sampath Kumar and Da Zhu for MSW. A number of HPEC meetings were held to review and discuss the results of the study. Comments and contributions at various stages of the study have been provided by peer reviewers Steve Karam, Christine Kessides and Paul Noumba, and Augustin Pierre Maria, Luis Andres, Dan Biller, Sudeshna Banerjee, Songsu Choi, Fook Chuan, Celine Ferre, Matt Glasser, Santiago Guerrero, Kristen Hommann, Dan Hoornweg, Pete Kolsky, Fernanda Nunez, Eduardo Perez and Giovanna Prennushi. JNNURM Project Appraisal Notes and Detailed Project Reports have been made available to the team by HPEC for the purpose of the study. The team is grateful to Caroline van den Berg and Sasha Danilenko for providing access to the IbNet database, and the Tamil Nadu Urban Development Fund (TNUDF), the Karnataka Urban Water Supply and Drainage Board (KUIDFC) and the Andhra Pradesh Municipal Development Project, Municipal Strengthening Unit for sharing their project data with the team. The team would also like to thank Oscar Alvarado, Shruti Garga, Raghu Kesavan and N. V.V. Raghava for coordinating the data collection. Support to the team has been provided by Michelle Lisa Chen and Sunita Singh. 3 Introduction 1. The paper presents the main results of the cost models developed as an input the High Powered Expert Committee (HPEC) on Urban Development in the estimation of investment and O&M requirements for urban water and sanitation (WSS) and municipal solid waste (MSW) for the period 2007-31. Each cost model builds on the methodological approach developed by HPEC in consultation with sector experts. The service standards adopted for the cost models, and the main assumptions, are the results of a number of consultations undertaken by HPEC with support from sector experts. The cost models are designed as tools that allow linking the various building blocks of the cost estimation to one other, and testing the impact of the main model assumptions on the overall investment requirements. A cross- country comparison is also conducted to benchmark the key service standards adopted in the models against international experience. 2. The paper is organized in five main sections and six Annexes: Section A presents the overall methodological approach adopted for the cost estimation and the consolidated results of the three costs models; Section B, C, and D discusses the methodology and main results of the water supply, sewerage and solid waste cost models respectively; Section E outlines the main policy implications emerging from the cost estimation; Annex I discusses in detail each step of the methodological approach and the main assumptions of the three cost models; Annex II describes the sample data and the data sources; Annex III explains the methodology for the urban population forecasts; Annex IV present in more detail the results of the cost estimation; Annex V describes the methodology for the estimation of the costs of low professionalization in the WSS sector; finally, Annex VI illustrates the main results of the cross- country comparison of sector indicators. 4 A. Municipal Financing Needs: WSS and MSW Sector A.I India Urban Population and Infrastructure Trends 3. India’s urban population is estimated to double in size from 2001 to 2031. The urban population of India, estimated at 286 million in 2001, is expected to reach 627 million by 2031, equivalent to 40 percent of the Indian population.1 Megacities (with population above 5 million) will also double in size over the same period, from 61 million to 133 million. Indian cities with population between 1 and 5 million will register the highest absolute increase in urban population, from 46 to 126 million, equivalent to an increase from 15 to 20 percent in their share of India urban population (see Figure 1 and Figure 2). The average annual population growth rate for urban India is expected to stabilize at 2.5 percent per annum, in line with the population growth rate recorded over the period 1995-2000, although below the record growth of 3-4 percent registered in the previous decades. Figure 1: Urban Population by Class Size (million), Figure 2: Urban Population by Class Size (million), 2001-2031 2001 and 2031 700 700 600 600 133 500 500 I.A : > 5m I.A : > 5m 400 I.B: 1-5m 126 400 I.B: 1-5m I.C: 1m-100k 300 I.C: 1m-100k II: 50-100K 300 61 195 II: 50-100K 200 III: 20-50k 46 III: 20-50k IV+: <20k 200 IV+: <20k 100 99 54 100 68 28 0 35 27 52 0 2001 2031 Source: Estimates based on UN World Urbanization Prospects 2007 and Census of India. 4. The challenge of a fast growing urban population is compounded by the high backlog in urban service delivery. Infrastructure deficits in urban areas are large and growing. Universal water access for urban population in India has yet to be realized. To date, virtually no city in India has 24/7 piped water supply. The water supply in Indian cities is characterized by limited hours of access per day and, in some cases, alternate day access. In the case of sanitation, the national average for sewerage network coverage is only 33 percent (based on 2001 census data), with some States receiving virtually no service. Although 300 urban centers have sewerage systems, most of these systems only partially cover their 1 Estimates based on UN population forecasts. The methodology for the estimation is described in Annex III. 5 populations. More than one third of the urban population relies on septic tanks as a form of sanitation. Additionally, treatment facilities exist in only 70 cities and the services are rudimentary at best. The present level of solid waste management is similarly dismal. There is no public system of waste collection from the source in Indian cities. As a result, street sweeping of waste has become the primary de facto method of waste collection. Furthermore, barring a few exceptions, there are no sanitary landfills in India, posing serious public health and environmental concerns. Uncovered solid waste is instead dumped haphazardly within or outside cities. 2 5. A comprehensive and updated assessment of urban infrastructure investment requirements for India is not available. There has been no systematic bottom-up assessment of service delivery and expenditure norms for urban India since the seminal Zakaria Committee’s Report of 1963.3 The most recent Government of India (GoI) assessment of infrastructure investment needs for urban India, the Rakesh Mohan Committee’s Report (also known as the 1996 Infrastructure Report) departed from the Zakaria Committee’s methodology by taking a macro or top-down approach to the estimation of urban investment needs.4 The rest of this section provides a short summary of the methodologies adopted by the 1963 and 1996 seminal GoI studies. 6. The Zakaria Committee’s Report (1963). The Zakaria Committee adopted a demand- driven approach to estimate physical norms (service standards) and financial norms (per capita investment requirements) for urban India. The standards were derived based on actual data collected from a sample of cities on level of urban basic services, demand for services, cost for provision, services maintenance and municipal finances. The assessment included water supply, sewerage, storm water drainage, urban roads and footpaths.5 For example, per capita water consumption was estimated to range between 45 and 270 liters per day depending on city size.6 The Zakaria Committee’s financial norms adjusted for inflation are still widely applied today as a benchmark for assessing infrastructure requirements in Indian cities. 7. The Rakesh Mohan Committee’s Report (1996). The Rakesh Mohan Committee adopted a top down approach to the estimation of infrastructure investments for urban India over the period 1996-2006. The exercise involved full-scale macroeconomic projections based on assumptions about expected economic growth for the Indian economy. The underlying 2 3i Network. 2006. ‚India Infrastructure Report‛. New Delhi. 3 Committee of Ministers constituted by the Central Council of Local Self Government (1963), ‚Augmentation of Financial Resources of Urban Local Bodies‛ (also known as the Zakaria Committee’s Report). 4 Expert Group on the Commercialization of Infrastructure Projects (1996). ‚The India Infrastructure Report – Policy Imperatives for Growth and Welfare‛ (also known as the Rakesh Mohan Report or the India Infrastructure Report, 1996). 5 Solid waste was not included in the cost estimation. 6 M.P. Mathur, Rajesh Chandra, Satpal Singh and Basudha Chattopadhyaya (2007). ‚Norms and Standards of Municipal Basic Services in India‛. National Institute of Urban Affairs Working Paper 07- 01. 6 principle of the estimation is that infrastructure investments will only take place in a policy environment that is investor friendly and transparent. The Rakesh Mohan Committee estimated the cost of urban infrastructure at Rs 28,000 crore or USD 5.6 billion (1996 prices), over the period 1996 -2006 across the three key services of water, sanitation and roads. Of this, the investment share for urban water and sanitation was estimated at Rs 15,523 crore, or USD 3 billion, which is significantly below what is considered, 15 years later, as an appropriate estimate of investment requirements in the WSS sector. With increasing volume of data made available from the Jawaharlal Nehru National Urban Renewal Mission (JNNURM), there is now sufficient project data available to re-visit the previous estimations of urban investment needs. 8. JNNURM, established in 2006, is the only centrally sponsored scheme for urban infrastructure improvement in India. The Jawaharlal Nehru National Urban Renewal Mission (JNNURM) is a centrally sponsored scheme launched in 2006 with the objective of improving urban infrastructure through a combination of investments and urban reforms. The duration of the scheme is seven years (2006-2012). JNNURM comprises four separate windows of assistance (or ‚Sub-missions‛), which share the same principles of engagement but differ in their investment focus and target cities. There are two windows of assistance with a focus on urban infrastructure investment: the Urban Infrastructure and Governance (UIG) scheme and the Urban Infrastructure Development Scheme for Small and Medium Towns (UIDSSMT). The UIG window targets 65 mission cities, selected based on their strategic importance for urban development. The UIDSSMT window is open to all other urban centers in India, excluding the 65 UIG cities. All four windows of assistance are project-based schemes, as funds are disbursed against sanctioned urban projects.7 9. The high level of funds committed under the JNNURM program signals a significant un-met demand for infrastructure investment in urban areas. As of June 2009, the total central budget allocation to the two JNNURM infrastructure windows amounts to about USD 8.6 billion which are expected to be disbursed over a period of seven years (2006-2012). In the majority of the States, the central allocation is closed to be fully committed. When States’ and ULBs’ contributions are included, total UIG and UIDSSMT commitments amount to about USD 14 billion. UIG is the window of assistance with the largest central allocation (USD 6.3 billion) and commitments (USD 4.7 billion), and also the best performing window in terms of disbursement (USD 1.7 billion). Projects sanctioned under the UIG window are sector-specific investments; water supply and sewerage account for about 63 percent of committed funds in the 65 mission cities, followed by urban transport projects (18 percent). The average size of a UIG project is USD 21 million. Projects sanctioned under the UIDSSMT are small-scale integrated urban projects, with an average project size of USD 4 million.8 The project data approved under the JNNURM UIG window have been utilized for the cost estimation based on information provided in the JNNURM Project Appraisal Notes. 7 Eligible cities are provided with a menu of investment options and requested to prepare City Development Plans to prioritize investments within the eligible sectors. 8 Analysis based on MoUD data. 7 A.2 Methodology: The Building Blocks of the Models 10. This section discusses the overall methodological approach adopted for the estimation of investment and O&M requirements for urban WSS and MSW over the period 2007-31. For each sector, a cost model has been developed as a tool to support HPEC in the estimation. The objectives of the cost models are the following: (i) to provide a framework for exploring the linkages between the various building blocks of the cost estimation and the total expenditure (investment and O&M) requirements; (ii) to compare the financial costs associated with alternative service standards, and (iii) to facilitate the updating and refining of the estimates as the data sample expands. The rest of the Section discusses the building blocks of the cost models, and how they relate to one other. Investment requirements 11. The estimation of investment requirements is based on the following four building blocks: (a) service standards, (b) cost drivers; (c) unit/per capita costs; and (d) three investment components (backlog or ‚unmet demand‛, demand growth and asset re-placement). The methodological approach adopted for the investment estimation, and the relationship between the four building blocks and the total investment needs, is summarized in Figure 3 below. Figure 3: Cost Models – Building Blocks Demand-side Supply -side • Supply option • Per capita consumption Assumptions • Coverage/ Level of service - 2006 level • Efficiency - 2007-31 growth • Economic life of assets Cost drivers • City size Service standards • Population density Unit/Per Capita Costs Investment components • 2006 backlog (un-met demand) • Project data • Demand growth (2007-31) Estimation • Cost simulation • Asset re-placement Investment requirements 2007-31 8 Service standards 12. The first step in the investment cost estimation is the setting of service standards or targets, based on both demand and supply considerations. The service standards adopted for the cost estimates have been determined by HPEC through a consultative process. A cross- country comparison is conducted as part of the study to benchmark these service standards vis-à-vis levels of service in comparable countries. 13. The main demand standard incorporated into the cost model is per capita consumption (level and growth). For all services, per capita consumption is highly dependent on income growth and pricing policies. In the absence of reliable data on the elasticity of demand vis-à-vis price and income, a cross-country comparison is conducted to estimate the expected level of WSS demand and solid waste generation in lower-middle income countries. While a full Willingness-to-Pay (WTP) analysis is beyond the scope of this study, an affordability test is carried out to estimate the share of household budgets that could be allocated toward to the costs of service provision. 14. The supply-side standards that are incorporated into the models are the following (a) supply option (e.g. network versus off-network water supply systems), (b) coverage rate, (c) level of service (e.g. private water connection versus stand-posts) and (d) economic life of the assets. Efficiency considerations are reflected in the cost estimation to the extent possible. For example, the investment cost estimation for the water sector assumes an efficient level of water leakage (equivalent to 20 percent of water production). In addition, the assumption related to the economic life of the assets is based on efficiency considerations – e.g., an economic life of 30 years for water and sewerage assets implies an efficient operation and management of the WSS systems. 15. Service standards affect the cost estimation through several channels. First, service standards determine the sample of projects available for the cost estimation. For example, 24/7 water supply continuity is one of the main targets for the water sector. Given that very few projects in India are designed to deliver 24/7 water supply, a cost simulation is conducted to complement the limited sample of 24/7 pilot projects. Second, the demand standards are used as reference to estimate the project beneficiaries (defined as the number of people that would be served by the incremental capacity delivery by a project). For example, the per capita water consumption standard of 135 lpcd (plus an allowance for efficient leakage) is used as the reference to estimate the number of beneficiaries of JNNURM water production projects.9 16. The service standards also determine the size of the investment components (backlog, demand growth and asset replacement). First, investment requirements for asset replacement are estimated based on the economic life of the assets. Second, investment backlogs are related to the supply standards. For example, assuming 24/7 water supply continuity as one of the sector standards increases the backlog investment requirements, given that intermittent water 9 This is done by dividing the incremental project capacity by the demand standard. 9 supply is currently the norm for the vast majority of the India urban population. Finally, investment requirements for demand growth depend on the level of per capita consumption and its growth over time. Cost drivers 17. The spatial pattern of urbanization is one of the key determinants of the unit costs of service provision. For example, everything being equal, it is more expensive on a per capita basis to provide piped water supply to low-density and small urban settlements than megacities. Modeling the relationship between cost drivers and unit investment costs would require the estimation of cost functions, which is beyond the scope of this work. An attempt has however been made to model the impact of selected cost drivers (namely city population and density) on investment requirements by estimating unit costs by city size class. The classes adopted for the estimation are in line with the census classification, with two main differences. Fist, the first census class (including cities with more than 100,000 inhabitants) is split into three sub-categories (> 5m, 1-5m, 100k-1m). Second, the last three census categories (IV-VI) have been aggregated into one class, which include small towns with less than 20,000 inhabitants. 18. Unfortunately, the number of project data available for small and medium towns at this stage is limited, and not representative, given that the cities receiving JUNNURM UIG funds are concentrated in the first Census class (> 100,000). Hence, only investment requirements for all urban India are presented as part of the study. In spite of the limited sample size, clear correlations between unit costs and population size and density are identified in most of the models, and are discussed as part of the study. 19. Other important cost drivers related to city topography could not be systematically captured in the cost models, but relevant findings are documented to the extent possible in the study. For example, the unit costs of composting are significantly higher in coastal cities, because weather conditions require that the compost plant be covered. The presence of economies of scale is also tested based on project data for all sectors. Per capita and unit costs 20. The per capita/ unit costs are estimated based on a sample of JNNURM UIG projects complemented by a data collection in selected States where the Bank has ongoing projects (Karnataka, Tamil Nadu and Andhra Pradesh). As a first step, JNNURM Project Appraisal Notes have been reviewed by sector experts to build a sample of representative projects. A combination of engineering and statistical criteria has been used to screen for outliers. As a second step, the Notes have been analyzed to compute project costs by sub-sectors. For example, solid waste project costs are estimated separately for collection and transportation, processing and disposal. As a third step, unit and per capita costs have been estimated by city size class based on the project’s design-year incremental capacity and number of beneficiaries. 21. Given the very limited number of water supply projects designed to deliver 24/7, a cost simulation has been conducted to estimate the cost of 24/7 water supply up-gradation and 24/7 10 distribution extension. The methodological approach for the cost simulation is described in detail in Annex I. Investment components 22. Investment requirements are estimated for the period 2007-31. The previous GoI Rakesh Mohan’s Committee Report estimated investment requirements up to the year 2006; hence 2006 is chosen as the base year of the model. The following three main investment components over the period 2007-2031 are considered for the cost estimation:  2006 backlog. The backlog is equivalent to the un-met demand for the 2006 base year. The main data sources for the estimation of backlogs are the City Development Plans complemented by the 2001 Population Census data.  Demand growth over the period 2007-31. Demand growth is estimated based on urban population forecasts and per capita consumption level and growth. Industrial demand growth in cities with more than 500,000 inhabitants is also included in the estimation of water investment requirements. Urban population forecasts over the period 2007-2031 are based on UN estimates.  Asset re-placement. Replacement requirements are estimated based on the economic life of the assets. As for the unit costs, the investment components are estimated by city size class. Total investment requirements 23. Total investment requirements are estimated by multiplying unit or per capita costs by investment components for each sub-sector. Investment requirements are expressed as a band rather than as a point estimate, to account for the variance in the unit/per capita cost estimates. The band is based on a 90 percent confidence interval for the unit/per capita costs. The mechanics of the estimation of backlog investment requirements for solid waste collection and transportation (C&T) is presented in Figure 4 below as an example. The Figure shows the relationship between unit costs /per capita costs and backlog (un-met demand). 11 Figure 4: Total investment Requirements (C&T Backlog) Operation & Maintenance 24. Operations & Maintenance (O&M) costs are estimated on an annual basis over the period 2007-2031 for the served population. Annual O&M costs are computed by multiplying unit O&M costs by total volume, which is estimated based on the coverage rate and per capita demand in a given year. Hence, annual O&M costs are expected to escalate as connection rates increase from the current level to full coverage. For the purpose of the estimation, coverage rates are simulated based on the assumption that full coverage would be achieved by 2031. Unit O&M costs are estimated based on a sample of projects and expert advice. It has to be noted that O&M costs depend to a significant extent on local conditions, and the variation in O&M costs across localities couldn’t be captured as part of the cost modeling. Strengths and limitations of the cost estimation 25. The main strengths of the methodological approach adopted for the estimation of investment requirements are (a) the establishment of clear linkages between service standards and per capita/unit costs; and (b) the modeling of important cost drivers related to the spatial patterns of urbanization, namely city population and population density. 26. The main drawback of the exercise is the data limitation. Although the recent launch of JNNURM has significantly increased the amount of project data available for WSS and solid waste, information on small and medium towns (with population below 100,000) is still very limited. Hence, the sample of project data that has been collected is too small to allow the estimation of investment requirements by city size class with a reasonable level of accuracy. 12 27. Second, only information on approved JNNURM project costs is available at this stage, given that the JNNURM program has only recently started, and the project completion rate is still relatively low. Anecdotal information suggests that cost escalation represent a significant share of total costs, up to 20-30 percent. As more and more JNNURM projects reach closure, it would be important to compare approved costs with actual costs, and determine the main drivers of cost escalation. This will allow ascertaining the extent to which cost escalation is due to causes outside the control of implementing agencies, rather than inefficiencies in procurement and implementation. 28. Finally, land costs are not included in the estimation of investment requirements, although they are likely to represent a significant share of investment costs in the MSW sector. Information on land costs is not available from JNNURM projects, given that JNNURM does not cover the cost of land acquisition. Such costs are therefore not reported in the JNNURM Project Appraisal Notes. 13 A.3 Municipal Investment and O&M Requirements- WSS and MSW 29. Overall investment requirements for urban WSS and MSW for the period 2007-31 range from Rs 4,637 to 6,785 Bn (2009 prices), equivalent to USD 103-151 Bn. The point estimate for the total investment requirements is Rs 5,711 Bn, equivalent to USD 127 Bn. The investment requirements for the residential water and sewerage sectors are comparable: the investment requirements for the residential water sector ranges from Rs 2,035 to 3,139 Bn (USD 45-70 Bn), while the investment requirements for the sewerage sector ranges from Rs 1,913 to 2,544 Bn (USD 43-57 Bn).10 The MSW investment requirements are in the scale of Rs 368 to 607 Bn (USD 8-13 Bn). Finally, the investment requirements for the industrial water sector ranges from Rs 321 to 496 Bn (USD 7-11 Bn). The sector with the highest cost variation relative to the size of the sector’s investment requirements is MSW, followed by water; while sewerage has the least cost variation. Investment requirements by sector are presented in Figure 5. 30. The confidence intervals for the total investment requirements (Rs 4,637 - 6,785 Bn) reflect the variability in per capita investment costs (PCIC). The variation of per capita costs observed across projects is large; thus, the actual PCIC of any particular future project may be higher or lower than the average PCIC estimated from the sample. This PCIC variation is used to construct a 90 percent confidence interval for the total investment requirements. Thus, 90 percent of the times, total investment requirements will be contained whithin the confidence interval, depending on how large or how small the actual PCIC turn out to be. 31. The PCIC for the water sector is comparable to the PCIC for the sewerage sector as the interval estimates overlap. The water PCIC ranges from Rs 3,363 to 5,188 (including both water production and distribution). The PCIC for sewerage (network and treatment) ranges from Rs 3,101 to 4,124. The PCIC for solid waste varies from Rs 323 to 515, including collection & transportation, processing and disposal. The water sector has the largest per capita annual O&M cost (Rs 501), followed by MSW (Rs 190), the sewerage sector the smallest (Rs 102). See Figure 6. 32. Demand growth is the main driver of investment requirements. The largest component of total investment requirements is residential demand growth, which accounts for 44 percent of the total investment requirements. Industrial demand growth accounts for an additional 7 percent of total capital requirements. The backlog, or unmet demand, accounts for 35 percent of investment requirements, while assets replacement accounts for the remaining 14 percent. While demand growth is the main investment component for both water and sewerage, assets replacement is the main driver of investment requirements for solid waste. Figure 7 shows the share of the total investment requirements accounted for by each of the three investment components. 10 See Box 2 for methodological issues that need to be taken into account when comparing water and sewerage investment requirements. 14 33. The investment trend is estimated on a five year basis coinciding with the Planning Commission’s five year plans. For the purpose of estimating the profile of the investments, the assumption is made that full service coverage would be achieved by the end of the 15th Plan (2027-32). The investment requirements are found to increase over time. The first phase of the investment trend, coinciding with the 11th Plan of the Planning Commission, amounts to Rs 961 Bn. The last phase, coinciding with the 15th Plan, amounts to Rs 1,404 Bn. The average investment per Plan is Rs 1,142 Bn (in 2009 prices). The investment trend is presented in Figure 8. 34. On average, for the period 2007-31, investment requirements account for about 1 percent of GDP per plan. The five year trend of investment requirements is estimated as a share of GDP, assuming a 7 percent real GDP growth rate over 2007-31. Investment requirements as a share of GDP decrease over time. The first investment, coinciding with the 11th Plan represents 1.6 percent of GDP. The last investment, coinciding with the 15th Plan accounts for 0.6 percent of GDP. On an annual basis, investment requirements on average account for 0.21 percent of GDP. The results are overall comparable with investment requirements estimated based on alternative, but complementary top-down models (See Box 1). Results are presented in Figure 9. 35. Annual O&M costs are expected to more than double in real terms over the period 2007-31 as a result of the increase in coverage. O&M costs are calculated annually for the period 2007-31. The data shows an increasing trend in annual O&M costs, as indicated in Figure 10. O&M costs are expected to increase from Rs 178 Bn in 2007 to Rs 609 Bn in 2031 (in 2009 prices). The water sector accounts for the largest share of total O&M costs (70 percent of the total)11, followed by MSW (20 percent of total). The sewerage sector accounts for the smallest share of total O&M costs (11 percent of the total). 36. For the median urban person, annual O&M costs account for about 6.2 percent of per capita expenditure.12 Overall, the affordability analysis suggests that there is scope for recovering O&M costs from the connected population. The water sector accounts for the largest budget share (3.9 percent of per capita expenditure), followed by sewerage (1.5 percent), and solid waste (0.8 percent). Hence, user fees in line with O&M costs for the WSS sector would represent 5.4 percent of per capita expenditure for the median urban person. The estimated WSS expenditure budget share is broadly in line with internationally recognized affordability norms for the WSS sector. 11 The residential water sector accounts for 55 percent of O&M costs, the industrial water sector for 15 percent. 12 The daily per capita expenditure for the median urban person is estimated at Rs 26 for the year 2004/05 based on National Sample Survey (NSS) data. The per capita expenditure is extrapolated to the year 2009 taking into account annual growth rate in per capita expenditure and inflation. The annual per capita expenditure for the year 2009 for the median households is estimated at Rs 12,726. 15 Box 1: Investment Requirements - Benchmarking of Results Fay and Yepes (2003) estimate infrastructure investment requirements for the period 2000-10 based on a model that relates demand for infrastructure with economic growth. Chatterton and Puerto (2006) refine Fay and Yepes’ estimation for the South Asia region. Results from both of these prior studies are not directly comparable to those obtained from the cost models for various reasons. First they use a top-down rather than a bottom-up methodology. Secondly, they are not country specific, they cover both urban and rural areas, and they do not include rehabilitation / upgrading costs. In spite of these differences, comparing the results of the models may provide useful insights on the order of magnitude of the investment requirements based on alternative, but complementary methodologies. Estimates obtained by Fay and Yepes for 2005-2010 and estimates obtained by Chatterton and Puerto for 2006-2010 are compared with estimates from the model for a comparable time-frame, such as the 11th Plan (2007-2012). According to the cost models, expected annual investment needs for water and sanitation (WSS) for the 11th Plan are equivalent to 0.29 percent of GDP. Estimates from Fay and Yepes (2003) for South Asia amount to 0.4 percent of GDP for the same two sectors. Estimates from Chatteton and Puerto (2006) are equivalent to 0.49 percent of GDP for India (See Figure 11). As expected, investment estimates based on the (bottom-up) urban cost models are lower than estimates based on the top-down models, which cover both urban and rural areas. It is however worth noting that the difference between the two models can’t be considered a proxy for rural investment requirements give the important methodological differences between the two models listed above. In addition, the investment requirements for the 11th Plan period, as estimated based on the bottom-up models, depend on the profiling of the investments. For the purpose of the estimation, it is assumed that full coverage will be achieved by the year 2031. In addition, the bottom-up models also estimate water production investment requirements for industrial customers, which are not included in the top down models by Fay and Yepes (2003) and Chatterton and Puerto (2006). Source: Fray and Yepes (2003); and Chatterton and Puerto (2006). 16 Box 2: Comparison of Water and Sewerage Investment Requirements In general, capital expenditures for the sewerage sector are more expensive than capital expenditures for the water sector. As a result, most countries have achieved a higher level of piped water coverage compared to sewerage network coverage. For example, while high income countries have 96 percent piped water coverage on average (See Table 25); they only have 70 percent sewerage network coverage (See Table 29). Thus, it may seem surprising that the models’ point estimate for water investment requirements is higher than the point estimate for sewerage investment requirements, although the interval estimates overlap . However, there are a number of explanations for this result. First, water and sewerage PCICs are estimated based on slightly different methodologies. The PCIC for sewerage is estimated from project data, for both network and treatment cost components. The PCIC for water is estimated partially based on project data (for the water production component) and partially based on a cost simulation (for the water distribution 24/7 component), given that very few projects in India are designed to supply water 24/7. On average, estimates based on project data are more vulnerable to cost escalation than estimates based on cost simulations, as the project costs data are based on cost estimates at project design, rather than actual costs (which were not available at the time the study was conducted). A sample of sewerage projects in Karnataka suggests that costs at contract award are 20 percent higher than costs at project design (net of inflation). As a result, the under-estimation may affect sewerage more than water. Second, the PCIC for water distribution (24/7 standards) include house connections and metering costs (equivalent to Rs 2500 per household, or Rs 500 per capita) that are likely to be borne by consumers. When, beneficiary contributions are excluded, the per capita cost for water decline by Rs 500. Third, due to high PCIC variability, it is more appropriate to compare interval estimates, rather than point estimates. When interval estimates are considered, there is no statistical difference between water and sewerage investment requirements, as the two intervals overlap. 17 Figure 5: Investment Requirements 2007-31, Rs Bn (2009 prices) 3,500 3,139 3,000 2,544 2,587 Rs Bn (2009 Prices) 2,500 2,229 2,000 2,035 1,913 1,500 1,000 607 496 500 487 408 368 321 0 Water (Residential) Sewerage Solid Waste Water (Industrial) Note: Confidence Intervals shown above reflect variability in per capita investment costs (PCIC). Figure 6: PCIC (Rs/capita) and PC annual O&M, by Sector13 Figure 7: Investment Requirements 2007-31, By Component, Rs Bn (2009 prices) Demand Growth (Industrial Water) 7% Replacement 14% Backlog 35% Demand Growth (Residential) 44% 13Per Capita Investment Costs in the solid waste sector is expected to increase over time as a result of growth in per capita waste generation. The average over 2007-31 is presented in the Figure. 18 Figure 8: Investment Requirements, by Five-year Plan, Rs Bn (2009 prices) 1,500 1,404 1,276 1,250 152 1,142 1,094 108 961 978 97 1,000 104 534 62 61 491 750 433 446 374 397 500 676 718 599 250 525 520 556 0 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007 - April 2012 - April 2017 - April 2022 - April 2027 - Average per plan March 2012 March 2017 March 2022 March 2027 March 2032 Solid waste Sewerage Water Figure 9: Investment Requirements, by Five-Year Plan, Share of GDP 2.00% Solid waste Sewerage Water 1.57% 1.50% 0.10% 1.21% 0.61% 0.08% 0.96% 1.03% 1.00% 0.80% 0.08% 0.49% 0.09% 0.07% 0.62% 0.41% 0.38% 0.31% 0.07% 0.50% 0.24% 0.86% 0.64% 0.55% 0.21% 0.49% 0.42% 0.32% 0.00% 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007 - April 2012 - April 2017 - April 2022 - April 2027 - Average Average March 2012 March 2017 March 2022 March 2027 March 2032 per plan per year Figure 10: Annual O&M Costs, Rs Bn (2009 prices) 609 584 559 536 600 513 491 470 450 430 500 RS Bn (2009 prices) 410 392 374 356 338 400 321 305 289 274 259 245 231 300 217 204 190 178 200 100 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 Solid waste Sewerage Water (Residential) Water (Industrial) 19 Figure 11: WSS Annual Investment Requirements (Share of GDP): Benchmarking of Results 0.80% Sewerage Water 0.68% Expected Annual Investment Needs 0.60% 0.49% 0.40% 0.42% 0.40% 0.29% 0.19% 0.31% 0.20% 0.12% 0.21% 0.26% 0.18% 0.17% 0.00% India, Urban South Asia, South Asia, India, Urban and Rural Urban and Rural Urban and Rural % of GDP % of GDP % of GDP % of GDP (2007-11) (2005-10) (2006-10) (2006-10) Cost Models Fay and Yepes (2003) Chatterton and Puerto (2006) 20 B. Urban Water Supply Services B.I Urban Water Service Standards DEMAND SUPPLY Per capita residential water 135 lpcd Supply option Piped water supply consumption (2007-31) (distribution) Annual growth in industrial 7% Level of service 24/7, private connections water demand (2007-31)14 (distribution) Target coverage 100% Efficient leakage 20% of water production Assets’ economic life 30 years 37. The service standards that are incorporated into the water supply cost model are defined along the following four dimensions: (a) supply option; (b) level of service/ continuity of supply; (c) efficient leakage and (d) per capita consumption norm. The cost model is based on the target of full piped water supply coverage through private connections, on a 24/7 basis. The per capita consumption norm is assumed at 135 liters per capita per day (lpcd) as an average for the period 2007-31, in line with MoUD benchmarking standards and with consumption data for lower-middle income countries.15 Including an allowance for efficient leakage (20 percent of water production), the per capita production norm is set at 168 lpcd. The level of efficient leakage is in line with MoUD benchmarking standards, and consistent with the leakage level in well run water utilities in developed countries.16 The same targets are assumed across all urban India, although the model allows to differentiate standards across city size classes, and to conduct a sensitivity analysis to estimate the impact of changing standards on total investment requirements. 38. The service standards incorporated into the model should be in line with consumers’ expectations. A cross-country comparison is thus under-taken to benchmark India WSS performance with respect to comparable countries. Although a full Willingness-To-Pay (WTP) 14 For cities with more than 500,000 inhabitants. 1515 See Ministry of Urban Development (2008), Handbook of Service Level Benchmarking. Delhi. See also Figure 15 and Table 27 in Annex VI. 16 It is however worth noting that the current level of leakage in Indian water utilities is well above the efficient level. Leakage levels are estimated at about 40 percent of water production, based on anecdotal evidence (the true level of leakage is unknown given the widespread lack of metering). Hence, reducing leakage levels from the current level to an efficient level would entail significant reforms in the management and governance structure of Indian water utilities. 21 study is not carried out as part of the study, a number of tests are conducted to asses the extent to which the model assumptions are in line with consumers demand. Cross-country comparison 39. Private Connections. About 49 percent of urban households are estimated to have private (within premises) access to water supply in urban India based on WHO/Unicef data collected as part of the Millennium Development Goal’s Joint Monitoring Program (JMP). The estimate is in line with the average access rate for urban India based on 2001 Population Census data (48 percent). Based on census data, the access rate varies from 57 percent in megacities, with population above 5 million, to 31 percent in towns with less than 20,000 inhabitants (see Figure 12). The urban India access rate is below the average for lower-middle income countries, estimated at 70 percent based on a sample of 43 countries (see Figure 13 ). For example, lower- middle income countries such as like the Philippines and China perform significantly better than India, with private connections covering 69 and 87 percent respectively of total urban households. WHO/Unicef estimates are broadly consistent with estimates based on IbNET’s utility data, which report access rates of the order of 70 percent for lower-middle income countries based on data from 773 utilities, including both private and shared connections (see Table 25 in Annex VI). 17 40. Continuity of supply. India is one of the countries with the lowest standards for water continuity of supply, with an average of only 4 hours of water supply per day in urban areas. The average number of hours of water supply is 16 hours in lower-middle income countries, and 23 hours in upper-middle income countries (see Figure 14). 41. Per capita water consumption. Data on per capita water consumption in urban India are difficult to obtain because of the widespread lack of metering and the uncertainty surrounding the level of water losses. The per capita consumption standard set by MoUD as part of the WSS benchmarking exercise (135 lpcd) is adopted for the model. The MoUD standard is within the appropriate range identified by the cross-country comparison for comparable countries. Based on utility data, per capita water consumption in urban India is estimated at 132 lpcd based on a sample of 8 utilities18. The average includes both private and shared connections and public taps. Efficient pricing is controlled for by excluding from the sample utilities that don’t recover their operating costs. Based on IbNet’s utility data, the average level of per capita consumption for lower- and upper-middle income countries is estimated at 103 and 162 lpcd respectively based on a sample of 707 utilities. (See Figure 15). Assuming a level of efficient leakage of 20 percent, a per capita production norm of 168 lpcd is adopted for the model. 42. It is difficult to estimate how per capita consumption will evolve and respond to income growth and efficient pricing, given that most utilities currently do not charge the full 17 See the International Benchmarking Network for Water and Sanitation Utilities (IbNET), http://www.ib-net.org/ 18 Data for India obtained from the 2007 Benchmarking and Data Book of Water Utilities in India 22 economic cost of service provision to their customers. While income growth may increase demand for water, the introduction of efficient pricing may deter further increases in consumption. Based on a survey of 1,100 households conducted in Dehra Dun in the northern Indian State of Uttar Pradesh in 1995, the price elasticity for individually connected users was estimated at --0.31 (a 10 percent increase in the price of water would lead to a 3 percent decrease in water consumption for households with individual connections). Household income was found to have a larger impact on water use: water use increases 4.1 percent as household income increases 10 percent.19 Given that the net impact of price and income effects are expected to counterbalance each other to some degree and vary across localities, no change is assumed in the per capita consumption norm of 135 lpcd over the period 2007-31. Willingness-to-pay 43. Ability to pay can be considered an upper bound estimate of WTP for improved service delivery. An affordability analysis is conducted to estimate the ability-to-pay of the median urban person for private water connections. The results indicate that there is significant scope for recovering O&M costs of private water connections from residential customers, although subsidies to cover the capital expenditure of water connections would be required. Based on a per capita consumption norm of 135 lpcd, water charges in line with O&M costs would represent 3.9 percent of per capita expenditure for the median urban person. The budget share is below the affordability norm, which is generally estimated at 3-5 percent of household budgets. A recent World Bank study on the ‚Review of Effectiveness of Rural Water Supply Schemes in India‛ finds that even among rural households there is WTP to cover the O&M costs for private connections. Rural households’ WTP for house connections is estimated to range from Rs 40-58 per household/month, or 1 – 2 percent of monthly household income.20 The results of the analysis suggest that the standard of house connection is broadly in line with consumers’ demand. 44. Coping costs can be considered a lower bound estimate of WTP for continuous water supply. The coping costs of intermittent water supply can therefore be estimated as a lower bound for the WTP of 24/7 up-gradation. The empirical evidence suggests that the coping costs of intermittent water supply are significant across all income groups. The main coping costs are investment in water-related equipment for households with private connections and the value of time spent fetching water for households that rely on public taps. See Box 3. 19 See Choe, Varley and Bijlani (1995). ‚Coping with Intermittent Water Supply: Problems and Prospects‛. Activity Report N. 26. 20 World Bank (2008), ‚Review of Effectiveness of Rural Water Supply Schemes in India‛ Sustainable Development Unit. South Asia Region. June. 23 Box 3: Household Coping Costs of Intermittent Water Supply Most of the urban population in the developing world receives intermittent water service. While it is evident that unreliable water supply worsens the quality of service provided to customers, the full extent of the problem is not easily recognized. Indeed, all consumers incur costs to cope with intermittent water supply. For households with access to piped water supply, the main coping costs include investments in water- related equipment such as water storage, and the monthly maintenance costs of operating such equipment. The coping strategies, for households who can afford it, include increasing the households’ water holding capacity, enhancing water pressure and purifying water via water tanks, electric pumps and water filters. Such coping strategies require large lumpy investments. Indeed, according to a survey in Dehra Dun, the average cost for all coping equipment is Rs 4,905 and Rs 3,688 (in 2009 prices) for individuals with private and shared connections, respectively. Moreover, the average payment for coping mechanisms, including capital investment and regular maintenance and operation, is as high as Rs 71 and Rs 52/household/month (Rs 3.9 – Rs 4/m3) for private and shared connections, respectively. For households that rely on public taps, the main coping cost is the value of the time spent queuing in line to fetch water. Households that rely on public taps generally belong to a low-income class and cannot afford the capital investment or access to credit to finance major equipment purchase. Their capital equipment expenditures are limited to small containers and the like. However, the opportunity cost of the hours spent fetching implies loss of wage income. In Dehra Dun, this cost equals Rs 7 per hour 21 (in 2009 prices). Moreover, since the average household spends 3 hours per day collecting water, the lost income potential equals almost Rs 356/household/month (approximately 10 percent of their monthly income). Thus the total coping costs (including storage and lost time) amount to Rs 367/household/month (Rs 82/m3) for the regular season. Coping costs are large compared to the tariffs paid. For example, connected customers in Dehra Dun are billed on average Rs 4/m3 (in 2009 prices); however, once the private investment in water storage is taken into account, the actual costs are over Rs 8/m3. Similarly, although public tap users pay no cash to the water utility, the real costs arising from the opportunity cost of the time spent queuing are over Rs 94/m 3 in the dry season. As a result, the poor pay higher real costs for water than those who are connected, due to the limited options that the poor have in adopting alternative coping strategies. Such large coping costs (which can be considered a lower bound estimate of WTP) suggest that even the poorest households are willing and able to pay for improved service delivery. Customers are willing to pay for continuous water supply. The present situation of low cost recovery is usually attributed to high prices; however, low cost recovery can also be interpreted as an indication of customers’ dissatisfaction with the performance of the water utility. Indeed, it has been estimated that in Dehra Dun, households would be willing to pay an additional Rs 5/m3 for continuous supply above the Rs 4 that they are billed on average (for a total willingness to pay of Rs 9/m3). Source: Choe, Varley and Bijlani (1995), Yepes, Ringskog and Sarkar (2001) and Zérah (2000). 21 Based on 80 percent of hourly wage rates among public-tap users 24 Figure 12: Urban Piped Water Supply Coverage in India, by City Size Class 70% 57% 55% 60% 48% 48% 50% 41% 37% 40% 31% 30% 20% 10% 0% Class I.A Class I.B Class I.C Class II Class III Class IV+ Weighted Average (7) (27) (360) (404) (1,164) (2,415) (4,377) Source: 2001 Population Census. Notes: Number of urban centers/towns in parenthesis. Figure 13: Urban Piped Water Supply –Selected Countries (2006) – Private Connections 96 90 96 88 87 100 84 70 69 80 59 Coverage as a % of population 60 49 48 43 34 40 20 20 7 0 Mexico India Nigeria Brazil China Indonesia Vietnam Bangladesh South Africa Philippines Pakistan Low income High income Upper middle Lower middle income income (43) (36) . (47) . (42) . Upper middle income Lower middle income Low income Source: WHO/Unicef. Notes: For income groups sample size (number of countries) in parenthesis. Figure 14: Urban Water Continuity of Supply – Selected Countries (2004-08) 24 20 16 12 hrs/day 8 INDIA 4 0 0 5,000 10,000 15,000 20,000 25,000 $ GNI per capita 2008 (PPP) Source: IBnet. Notes: Based on a sample of utilities for each country. 25 Figure 15: Urban Per Capita Water Consumption – Private/Shared Connections and Public Taps 250 179 162 202 171 197 200 132 122 114 103 127 110 150 97 97 72 76 77 100 50 0 Lpcd (Pakistan) India South Africa Mexico Indonesia Shanghai (Philippines) (India) Upper middle Lower middle China Low income (Indonesia) Vietnam High income Bangladesh Delhi Philippines (China) Karachi Jakarta Manila income (142) (494) . income (213) . (201) . Upper middle inc. Lower middle income Low income City Level Source: IBnet; Water in Asian Cities; 2007 Benchmarking and Data Book of Water Utilities in India Notes: Based on a sample of utilities. Number of utilities available for each income groups in parenthesis. 26 B.II Urban Water Supply Cost Model Methodology 45. Urban water investment requirements are calculated for both (A) residential customers and (B) industrial customers. O&M costs are estimated separately on an annual basis for both customer groups. The rest of the Section discusses the methodological approach adopted for the cost estimation. A summary of the methodology is presented in Table 14. A detailed step- by-step description of the approach and data sources is provided in Annex I and II. 46. Residential customers – The methodology for estimating investment requirements for residential customers consist of the following four building blocks: (1) service standards, which are described in detail in the previous section; (2) investment components, (3) cost drivers and (4) Per capita investment costs (PCIC).  There are three main investment components. The first investment component corresponds to the backlog or un-met demand, defined as the percentage of the current population that is un-served, in relation to the service standard. The second component corresponds to demand growth, defined as the population that will require service over the period 2007-31. Finally, the last component corresponds to assets replacement, which is the cost of replacing outdated assets.  Per capita investment costs (PCIC) are calculated based on project data and cost simulations. Three separate PCIC are estimated for (a) production; (b) distribution extension (based on 24/7 standards) and (c) 24/7 up-gradation.  The two main cost drivers associated with the spatial pattern of urbanization are city size and population density. The correlation between cost drivers and PCIC is explored by estimating PCICs by city size class. 47. Investment requirements are calculated as the product of each one of the investment components times the PCIC of the corresponding sub-sector (production or distribution) or the PCIC for 24/7 up-gradation. Overall, the methodological approach is depicted in the diagram presented in Figure 16 below. The diagram presents a breakdown of investment requirements based on cost component (production, distribution and 24/7 up-gradation) and investment component (backlog, demand growth and asset re-placement). The investment requirements include: total investment requirements for backlog [1+ 2+ 5]; total investment requirements for demand growth [3+4]; and total investment requirements for assets replacement (water production) [6]. Investment requirements are estimated as a band based on a 90 percent confidence interval for PCICs.22 48. Industrial customers – Investment requirements for industrial customers are calculated at the city-level (for cities with more than 500,000 inhabitants). First, industrial water demand is 22There is a one-to-one relationship between PCICs and total investment requirements – i.e. a 10 percent increase in PCICs leads to a 10% increase in total investment requirements. 27 forecasted based on the percentage of water production capacity dedicated to industrial customers and demand growth. Second, city-level investment requirements are estimated based on unit water production costs. Finally, total investment requirements are computed as the product of city-level investment needs and the total numbers of cities in each city class. 49. O&M costs are calculated on an annual basis for the served residential population and industrial customers. Figure 16: Urban Residential Water Supply Cost Model - Methodology 28 Figure 17: Urban Water Supply Cost Model Methodology: Building Blocks Residential Water Supply (1) Service Standards 100% piped water supply (private connections); per capita water consumption at 135 lpcd over 2007-31. (2) Investment Components 2006 Backlog Demand Growth (2.1) (2.2) (2.3) Assets Replacement (Un-met demand) 2007-31 Production [1] Production [3] Production [6] Definition: percent of base-year un- Definition: Incremental water Definition: Assets are assumed to met residential water demand based demand over the period 2007- have a 30 year economic life. on a per capita water production 2031. norm of 168 lpcd. Source: Forecasted yearly by Source: City Development Plans applying UN population growth (CDPs) (Sample size 67 obs.) rates to the 2001 Census population. 24/7 Up-gradation [5] Definition: percent of base-year urban population without access to water supply on a 24/7 basis. Source: Equivalent to the entire Indian urban population connected to piped water supply (virtually no city in India has 24/7). Distribution Extension [2] Distribution Extension [4] There are no assets replacement Percent of the base year urban Same as production (see [3]) costs for distribution, because population without access to piped replacement costs of existing water supply (within premises). assets are accounted for in 24/7 Source: Census Data (Sample size: up-gradation. all UA, cities and towns). (3) Per Capita Investment Costs (PCIC) (3.1) Production (3.2) 24/7 Up-gradation (3.3) Distribution Extension Definition: Costs for source Definition: Costs of rehabilitating Definition: Costs for distribution, augmentation, treatment, and the existing distribution network to storage, connection and metering. transmission. PCIC are calculated as achieve 24/7. PCIC are calculated PCIC are calculated based on a cost project costs over beneficiaries. based on a sample of pilot 24/7 simulation, which includes the cost Project beneficiaries are defined as project data and a cost simulation. of extending the network to the number of people that could be The cost simulation assumes that connect additional people (50% of served at the level of the per capita 50 percent of the existing network the per capita cost of the production norm (168 lpcd) given is to be replaced, and includes the connected population, as some of the incremental capacity generated costs of connection, metering and the network is already in place), by project. storage. and the cost of connection, Source: 27 JNNURM Project Source: JNNURM Project Appraisal metering and storage. Appraisal Notes; data collection Notes; data collection from Source: CDPs (Sample size 31 obs.) from Karnataka; and CDPs (Sample Karnataka, and CDPs (Sample size size 41 obs.) 33 obs.) 29 Industrial Water Supply (1) Service Standards Industrial production requirements are estimated to increase in line with the average annual growth rate for industry value added over the past 10 years (7 percent per annum). (2) Investment Components (2.2) Demand Growth 2007-31 It is assumed that 20% of the city-level production capacity is dedicated to industrial customers (based on expert estimate and CDPs). Demand growth is estimated based on annual growth rate of 7 percent. Sources: JNNURM, CDPs, UN forecasts. (3) Unit Costs (3.1) Production Unit production costs are estimated based on JNNURM project data. O&M (Residential and Industrial) Annual O&M costs are calculated based on the volume of water produced for the served residential population and industrial customers. Source: O&M unit cost estimates have been made available by sector experts based on recent project data. 30 B.III Urban Water Investment and O&M Requirements 50. Investment requirements for urban water supply range from Rs 2,356 to 3,634 Bn (2009 prices), or USD 52-81 Bn, as shown in Figure 19. This estimate accounts for both residential and industrial investment requirements. The point estimate is Rs 2,995 Bn, or USD 67 Bn. Industrial water needs account for 14 percent of total water investment requirements. Cost variation is roughly the same in both sub-sectors (the coefficient of variation for distribution is .8 compared to .7 for production). Confidence intervals reflect the variability in per capita investment costs (PCIC) across projects. The methodology for the computation of interval estimates for investment requirements is discussed in Annex I. Investment requirements by sub-sector and cost component are reported in Figure 18. 51. The total PCIC for residential urban water supply (including production and distribution) ranges from Rs 3,363 to 5,188 (2009 prices). The PCIC for distribution extension ranges from Rs 2,164 to 3,491 (based on a 90 percent confidence interval), with a point estimate equal to Rs 2,828. The PCIC for water production ranges from Rs 1,199 to 1,697, with a point estimate equal to Rs 1,448. The backlog for water production is estimated to be 35 percent of the base-year urban population; the backlog for water distribution extension is higher, at 52 percent. PCIC and backlog percentages are reported in Figure 20. 52. The PCIC for 24/7 up-gradation is estimated to range from Rs 1,873 to 3,153, with a point estimate of Rs 2,513 per capita. The backlog for 24/7 up-gradation is estimated at 48 percent, which is equivalent to the piped water supply coverage rate for the year 2006 (See Figure 20). The costs of 24/7 up-gradation are estimated for the current population only (backlog), given that all future investment in the distribution network are assumed to be made based on 24/7 standards.. Investment requirements for 24/7 up-gradation are estimated based on a cost simulation and a sample of six pilot projects. The per capita costs for the pilot projects are reported in Table 24 in Annex IV. 53. Demand growth is the main driver of water supply investment requirements. The largest share of capital expenditure is associated with demand growth, which accounts for 56 percent of the total investment requirements. Demand growth from residential customers account for 42 percent of total investment requirements, and demand growth for industrial customers for an additional 14 percent. The second largest cost component is backlog or unmet demand (for residential customers), which accounts for 34 percent of total requirements. More specifically, the un-met demand for distribution extension accounts for 17 percent of total investment requirement; the un-met demand for water production 5 percent, and the backlog for 24/7 up-gradation 12 percent. Replacement costs account for 10 percent of the total capital expenditure. Figure 21 shows the share of the total water investment requirements accounted for by each of the cost components. 54. The profile of the investments is estimated based on the assumption that full coverage would be achieved by the end of the 15th plan. Investment requirements are estimated at five year intervals, which correspond to the Planning Commission’s five-year 31 Plans. Investment requirements are estimated to increase over time. For the first phase of the investment trend, coinciding with the 11th plan of the Planning Commission, investment requirements are estimated to be Rs 525 Bn (in 2009 prices). The last investment, in the amount of Rs 718 Bn, coincides with the 15th plan. The investment trend for the water sector is presented in Figure 22. 55. The investment requirements as a share of GDP decrease over time. The five year trend of investment requirements for the water sector is calculated as a share of GDP, based on the assumption that full coverage will be achieved by the end of the 15th plan. Results are presented in Figure 23. On average, for the period 2007-31, the investment requirements account for 0.55 percent of GDP per plan (based on a 7 percent real GDP growth rates). The first investment, coinciding with the 11th plan represents 0.9 percent of GDP. Instead, the last investment, coinciding with the 15th plan represents 0.32 percent of GDP. 56. Total annual O&M costs are estimated to triple over the period 2007-31 as a result of the expected increase in coverage. O&M costs are calculated annually. The cost estimation indicates an increasing trend in annual O&M costs, as shown in Figure 24. Annual O&M costs for urban water supply are estimated to increase from Rs 135 Bn in 2007 to Rs 420 Bn in 2031, or from USD 3 to 9 Bn (in 2009 prices), as a result of the expected increase in coverage. The per capita annual O&M expenditure is estimated at 501 Rs/capita. 57. Distribution extension (based on 24/7 standards) accounts on average for 52 percent of residential urban water supply investment requirements. Production accounts for the second largest share (34 percent) followed by 24/7 up-gradation for the connected population (14 percent). The relatively higher cost share associated with distribution extension is explained by the fact that distribution costs account for about 70 percent of total capital expenditure on a per capita basis. In addition, the backlog for distribution extension (estimated at 52 percent) is significantly higher than the un-met demand for water production (35 percent). 32 Figure 18: Urban Water Supply Investment Requirements Residential Water Supply 2006 Backlog Demand Growth Assets replacement (Un-met demand) 2007-31 Total Production Production [1] Production [3] Production [6] (Residential) Rs Bn: 160 Rs Bn: 428 Rs Bn: 288 Rs Bn: 875 $Us Million: 3,554 $Us Million: 9,511 $Us Million: 6,390 $Us Million: 19,455 Total 24/7 Up-gradation 24/7 Up-gradation [5] (Residential) Rs Bn: 369 Rs Bn: 369 $Us Million: 8,210 $Us Million: 8,210 Distribution Distribution Total Distribution Extension [2] Extension [4] Extension (Residential) Rs Bn: 522 Rs Bn: 819 Rs Bn: 1,342 $Us Million: 11,608 $Us Million: 18,209 $Us Million: 29,817 Total Assets Total Backlog Total Demand Growth Replacement TOTAL RESIDENTIAL (Residential) (Residential) (Residential) Rs Bn: 1,052 Rs Bn: 1,247 Rs Bn: 288 Rs Bn: 1,342 $Us Million: 23,371 $Us Million: 27,721 $Us Million: 6,390 $Us Million: 29,817 Industrial Water Supply Demand Growth TOTAL INDUSTRIAL (Production) Rs Bn: 408 Rs Bn: 408 $Us Million: 9,075 $Us Million: 9,075 Total Water Supply (Industrial + Residential) Total Backlog Total Demand Growth Total Assets Replacement TOTAL RESIDENTIAL+ (Residential + Industrial) (Residential + Industrial) (Residential + Industrial) INDUSTRIAL Rs Bn: 1,052 Rs Bn: 1,656 Rs Bn: 288 Rs Bn: 2,995 $Us Million: 23,371 $Us Million: 36,796 $Us Million: 6,390 $Us Million: 66,557 33 Figure 19: Urban Water Investment Requirements, 2007-31, Rs Bn (2009 prices) 4,000 3,634 2,995 Rs Bn, 2009 prices 3,000 2,356 2,000 1,000 0 LOWER BOUND AVERAGE UPPER BOUND Note: Confidence Intervals shown above reflect variability in per capita investment costs (PCIC). Specifically, the actual total investment requirement will be contained within the interval 90 percent of the times, given that actual PCIC costs may differ from the avereage estimate from the sample of projects. Figure 20: Per Capita Investment Costs (Rs/capita) and Backlog (%), by Sub-sector 4,000 3,491 3,153 3,000 2,827 PCIC (RS/person) 2,513 2,000 1,697 2,164 1,873 1,448 1,000 1,199 0 Production 24/7 Up- Distribution gradation Extension Figure 21:Urban Water Investment Requirements, By Component (Rs Bn) Backlog (24/7 Demand Growth upgradation) (Industrial) 12% 14% Backlog (production) Asset Replacement 5% 10% Backlog (dist extension) 17% Demand Growth (Residential) 42% 34 Figure 22: Urban Water Investment Requirements, by Five-Year Plan (Rs Bn) 1,000 Industrial - Demand growth 900 Residential -Asset Replacement (production) 800 Residential -Demand Growth Rs Billion (2009 Prices) 700 Residential - Backlog (production and dist extension) Residential - Backlog (24/7 Upgradation) 127 600 127 61 49 55 500 32 61 39 400 110 35 290 326 300 172 214 246 200 100 136 136 136 136 136 0 74 74 74 74 74 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007- March April 2012- March April 2017- March April 2022- March April 2027- March 2012 2017 2022 2027 2032 Figure 23: Urban Water Investment Requirements, Share of GDP (2007-31) 1.0% Industrial Asset Replacement Demand Growth Backlog 0.9% 0.1% 0.8% 0.2% 0.6% 0.1% 0.55% 0.5% 0.5% 0.3% 0.1% 0.42% 0.1% 0.3% 0.1% 0.32% 0.2% 0.1% 0.2% 0.3% 0.2% 0.3% 0.1% 0.11% 0.3% 0.2% 0.2% 0.0% 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007 - April 2012 - April 2017 - April 2022 - April 2027 - Average Average March 2012 March 2017 March 2022 March 2027 March 2032 per plan per year Figure 24: Annual O&M Costs (Rs Bn) and Production Coverage Trends (%) 100% 500 80% 400 Coverage (%) 60% 300 Rs Bn 200 40% 100 20% - 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 Annual O&M Coverage (production) % 35 B.IV Capital Works Unit Cost Analysis Cost drivers 58. City population size and population density are both important cost drivers of urban water supply investments. The cost model makes an attempt to capture the correlation between cost drivers and unit costs, by estimating investment requirements by city size class. While the data sample is not large enough to allow the estimation of investment requirements with sufficient accuracy for each city size class, significant trends and correlation emerge from the analysis. 59. The analysis indicates a slightly negative correlation between city population (or city size) and the unit costs of water production for the entire data sample. The main explanatory variable is distance to water sources, which tends to be significantly longer in megacities, compared to other urban localities. In the data sample, the correlation is driven by two megacities projects, which have significantly higher than average production costs. The first project in Mumbai involves the construction of a dam to store water and transmit water by a tunnel. The second project in Chennai involves the construction of a transmission main at a significant distance from the city (pumping is 14.6 km and gravity main is 34.5 Km). These two JNNURM water production projects are excluded from the sample as their unit costs are significantly above the average, and they are not considered representative for urban India. Although both projects are deleted from the sample due to engineering reasons, one of them also fits the statistical definition of an outlier (its values is more than three standard deviations away from the mean). If the two outliers are removed, no correlation is found between city population and the unit cost of water production. The data also show limited economies of scale in water production if these two outliers are excluded from the sample. This is in line with conventional thinking that economies of scale in water production are generally not as large as in water distribution. (See Figure 25 and Figure 26). 60. The prevailing water source for JNNURM projects is surface water. Only one project (for Rajkot city) relies on groundwater as the main source of water. The prevalence of surface water is consistent with the most recent trends in urban India, as surface water is overtaking groundwater as the major source of water supply across urban areas, because of declining groundwater levels. 61. Population density is the main cost driver for water distribution costs. Small cities have lower population density and therefore higher per capita distribution costs, compared to large cities. The per capita length of the distribution network calculated based on CDP data is significantly higher in small and medium towns compared to large cities. As a results, per capita investment costs increase significantly from Class I.A to Class IV+ cities, although the magnitude of the increase cannot be estimated with sufficient accuracy given the limited number of observations. (See Figure 27). 36 62. The main cost driver that explains variation in O&M costs is the water head, as higher heads imply higher power charges, which are estimate to account for about 40 percent of total O&M costs based on a sample of project data. Maintenance costs are estimated to account for about 10 percent of total O&M costs. Also, large cities are found to have higher unit O&M costs than the average: O&M costs are estimated to vary from 12 Rs/m3 in Class I.A to 3 Rs/m3 in Class IV+ (see Table 4 in Annex I). The trend is mainly explained by the fact that large cities tend to rely on more distance sources of water supply, compared to other urban localities. Sensitivity analysis on service standards 63. A sensitivity analysis is conducted to assess the impact of changing service standards on project costs. The main standards in the model that have a direct impact on costs are (a) the per capita production norm and (b) the supply option (piped water supply with private connections). 64. Per capita production norm. Assuming a lower per capita production norm for small and medium towns is shown to have only a marginal impact on total investment costs – for example, lowering the per capita production norm from 168 to 88 lpcd for towns with less than 50,000 inhabitants (Class III and IV+) would only marginally decrease total investment requirements, from Rs 2,995 Bn to 2,939 Bn (see Figure 28). This is explained by the fact that a lower per capita production standard would affect the cost for water production (which account for about 30 percent of total investment costs), but not for water distribution. A lower per capita consumption norm for small and medium towns wouldn’t affect the optimal distribution system capacity for the following two reasons. First, the distribution network is characterized by economies of scale, mainly related to the cost of excavating and re-surfacing the roads. Given the large economies of scale, the distribution system is generally over- designed. For example, if per capita water consumption in a small town is 70 lpcd, the optimal design solution may be to build a distribution network for 100 lpcd to achieve economies of scale. Second, conservative assumptions have been made to estimate the mix of pipes required for the extension of the distribution network in small and medium towns, based on information reported by the towns themselves as part of the data collection and expert advice. The baseline model assumes an average mix of pipes of 250-350mm diameter, with a minimum size of 100mm diameter. Based on a cost simulation, it is estimated that in a town of 50,000 inhabitants, a distribution network with a mix of pipes of 250-350 mm diameter would deliver a daily per capita consumption in the range of 70-135 lpcd. Assuming a lower standard of per capita consumption would therefore not affect the mix of pipes and the cost estimates for water distribution. 65. Supply option. The second main standard incorporated in the model is the supply option (piped water supply with private connections). A full sensitivity analysis has not been conducted to estimate the impact on costs of adopting standposts as a service standard rather than private connection in small towns (for a given level of water consumption), because of limited data available on the cost of standposts. The savings associated with standpipes are related to distribution costs, and more specifically to the costs of house connection and tertiary 37 distribution. Considering that small towns (with less than 50,000 inhabitants) account for only 19 percent of urban India population based on 2009 estimates, the impact of a lower distribution standard in small towns is likely to be modest. Figure 25: Unit Production Costs (Rs/m3) and City Size (2009 prices) 50,000 Unit cost (Rs/m3) 40,000 30,000 20,000 10,000 0 - 2 4 6 8 10 12 14 16 18 City size (m) Notes: Two projects excluded from the sample for engineering reasons (project not representative) and statistical reasons (outlier with value more than 3 standard deviations above the mean). Figure 26: Unit Water Production Costs (Rs/m3): Economies of Scale (2009 prices) 35,000 30,000 25,000 20,000 Unit cost (Rs/m3) 15,000 10,000 5,000 0 - 100 200 300 400 500 Incremantal capacity (MLD) Notes: Two projects excluded from the sample for engineering reasons (project not representative) and statistical reasons (outlier with value more than 3 standard deviations above the mean). 38 Figure 27: Per Capita Distribution Costs (Rs/capita) and City Size (2009 Prices) 12,000 Per Capita Costs (Rs/capia) 10,000 8,000 6,000 4,000 2,000 0 - 2 4 6 8 10 12 14 City size (m) Figure 28: Sensitivity Analysis: Water Per capita Production Norm 4,000 3,635 3,570 3,000 2,995 2,939 Rs Bn, 2009 Prices 2,000 2,356 2,308 1,000 0 Production norm 168 lpcd (City Class III and IV+) Production norm 88 lpcd (City Class III and IV+) Assumption 1 Assumption 2 39 C. Urban Sewerage Services C.I Urban Sewerage Service Standards DEMAND SUPPLY Per capita sewerage 108 lpcd Supply option Sewerage network generation (2007-31) (collection) Target coverage 100% Assets’ economic life 30 years 66. Urban sewerage investment requirements are estimated based on the target of full network and treatment coverage. Per capita sewerage generation is assumed at 108 lpcd, which is equivalent to 80 percent of per capita water demand. The norm is considered appropriate for middle-income countries with an average per capita water demand of 135 lpcd. Hence, no growth is assumed in sewerage generation on a per capita basis over the period 2007-31. The economic life of the assets for both network and sewerage is assumed to be 30 years. This implicitly assumes an efficient operation and maintenance of the assets. Should the assets not be properly maintained, the economic life of the assets would shorten significantly. 67. A cross-country comparison is conducted to cross-check the targets vis-à-vis service levels in comparable countries. The percentage of urban India population with access to sewerage network is 33 percent based on 2001 population census data. Access varies significantly across city size classes, from 53 percent in the megacities to 14 percent in small towns with less than 20,000 inhabitants (see Figure 29). On average, sewerate network coverage for urban India (33 percent) is significantly below the average for lower-middle income countries, estimated at 56 percent for a sample of 15 countries, based on UNSTATS data. Sewerage treatment coverage in urban India is estimated at 28 percent, slightly below the 36 percent average for lower-middle income countries, based on a sample of 13 countries. (See Figure 30 and Figure 31). On average, the cross-country comparison indicates that coverage rates for urban sewerage are significantly below access rates for urban piped water supply across all countries’ income groups. The average sewerage coverage rate for high income countries is 77 percent for network and 70 percent for treatment based on a sample of 38 and 37 countries respectively. These results are in contrast with the almost universal coverage in piped water supply in high income countries (see Figure 13). 68. The model assumes an increase in sewerage coverage in urban areas from 33 and 28 percent for network and treatment respectively to 100 percent over the period 2007-31. For the 40 simulation of the investment profile, the assumption is made that full coverage would be achieved by the end of the 15th Plan. 41 Figure 29: Urban Sewerage Network Coverage in India, by City Size Class 60% 53% 46% 40% 33% 28% 20% 20% 15% 14% 0% Class I.A Class I.B Class I.C Class II Class III Class IV+ Weighted Average (7) (27) (360) (404) (1,164) (2,415) (4,377) Source: 2001 Census. Notes: Sample size (number of urban centers) in parenthesis. Figure 30: Urban Sewerage Network Coverage - Selected Countries (2000-07) 98 100 87 83 77 80 68 55 60 56 60 48 46 40 33 15 19 20 0 High Upper Mexico South Brazil Lower Jordan Morocco Armenia China India Paraguay Low Income Middle (2005) Africa (2006) Middle (2004) (2007) (2006) (2004) (2001) (2007) Income Income (2007) Income (38) (23) . (15) . (5) Upper Middle Income Lower Middle Income . Source: UNSTATS; 2001 Census for India. Notes: Sample size (number of countries) in parenthesis. Figure 31: Urban Sewerage Treatment Coverage- Selected Countries (2000-07) 100 80 80 70 57 60 52 43 35 36 34 33 40 26 28 26 20 5 0 High Upper South Mexico Brazil Lower Morocco Jordan India Armenia China Iraq Low Income Middle Africa (2005) (2006) Middle (2007) (2004) (2006) (2004) (2005) Income Income (2007) Income (37) (20) . (13) . (4) Upper Middle Income Lower Middle Income . Source: UNSTATS; CDPs for India. Notes: Sample size (number of countries) in parenthesis. 42 C.II Sewerage Cost Model Methodology 69. This Section discusses the overall methodological approach for the cost estimation. A summary of the approach is presented in Figure 33. A detailed step-by-step description of the methodology is provided in Annex I. 70. The methodology for estimating investment requirements for urban sewerage consists of four building blocks: (1) service standards, which have been described in detail in a previous section; (2) investment components, (3) cost drivers (city population and density) and (4) Per capita investment costs (PCIC). Urban sewerage investment requirements are calculated for residential customers only.  There are three main investment components. The first investment component corresponds to the backlog or un-met demand, defined as the percentage of the base year population that is un-served, given the chosen service standards. The second component corresponds to demand growth, defined as the population that will require service over the period 2007-31. Finally, the last component corresponds to assets replacement, which is the cost of replacing outdated assets.  Per capita investment costs (PCIC) are calculated based on project data. Two separate PCIC are estimated for (a) network and (b) treatment.  The correlation between city size and population density and PCICs is explored by estimating PCICs by city size class. 71. Investment requirements are calculated as the product of each one of the investment components times the PCIC of the corresponding sub-sector (network and treatment). The overall methodological approach is depicted in the diagram presented in Figure 32 below. The diagram presents a breakdown of investment requirements based on sub-sector (network and treatment) and investment component (backlog, demand growth and assets replacement). The investment requirements include: total investment requirements to close the backlog, or un- met demand [1+ 2]; total investment requirements for demand growth [3+4]; and assets replacement costs [5]. Investment requirements are estimated as a band based on a 90 percent confidence interval for PCICs.23 72. O&M costs are calculated on an annual basis for the served residential population. 23There is a one-to-one relationship between PCICs and total investment requirements – i.e. a 10 percent increase in PCICs leads to a 10 percent increase in total investment requirements. 43 Figure 32: Sewerage Cost Model Methodology 44 Figure 33: Sewerage Cost Model Methodology – Building Blocks Investments (1) Service Standards Supply: 100% target coverage; Demand: per capita sewerage generation of 108 lpcd (80% of per capita water consumption norm) (2) Investment Components 2006 Backlog Demand Growth (2.1) (2.2) (2.3) Assets Replacement (un-met demand) 2007-2031 Network [1] Network [3] Network [5] Definition: % of base year Definition: Demand from Definition: Assets are assumed urban population without incremental urban population to have a 30 year economic life access to wastewater network. over the period 2007-2031. prior to replacement. Source: Census Data (Sample Source: Forecasted yearly by size: all UA, cities and towns) applying UN population growth rates to the 2001 Census population. Treatment [2] Treatment [4] Treatment [5] Definition: % of base year Same as for Network (above). Same as for Network (above). urban population without access to wastewater treatment. Source: City Development Plans (CDPs) (Sample size 80 obs.) (3) Per Capita Investment Costs (PCICs) (3.1) Network (3.3) Treatment Definition: Estimated using the formula Definition: Same as for network (left) PCIC = total sub-sector project cost / project beneficiaries. Project beneficiaries are defined as the number of people that can be served by the project given the level of the per capita norm (108 lpcd) and the incremental capacity generated by project. Source: 47 JNNURM Project Appraisal Notes Source: Same as for network (left) (Sample size 79 complemented by data collection in Karnataka, projects). Tamil Nadu, and Andhra Pradesh (Sample size 113 projects). A O&M 1 Annual O&M costs are calculated based on wastewater generation for the served residential population. Source: O&M unit cost estimates have been made available by sector experts based on recent project data. V 45 C.III Urban Sewerage Investment and O&M Requirements 73. Total investment requirements for urban sewerage range from Rs 1,913 to 2,544 Bn (2009 prices), or USD 43-57 Bn. The point estimate for the investment requirements is Rs 2,229 Bn, or USD 50 Bn, as shown in Figure 35. This value accounts only for residential investment requirements. The sub-sector with the highest cost variation is network (the coefficient of variation is 1 compared to .7 for treatment). The confidence intervals reflect the variability in per capita investment costs (PCIC) across projects. The methodology for the computation of interval estimates for investment requirements is discussed in Annex I. Investment requirements by sub-sector and cost components are reported in Figure 34. 74. Network investment accounts for 66 percent of total investment requirements. The total PCIC for urban sewerage is estimated to range from Rs 3,101 to 4,124. The PCIC for network accounts for 66 percent of total costs on a per capita basis, ranging from Rs 2,019 to 2,727 with a point estimate equal Rs 2,373 per capita. The PCIC for treatment ranges from Rs 1,082 to 1,398, with a point estimate equal Rs 1,240 per capita (in 2009 prices).The backlog for network is estimated at 67 percent, the backlog for treatment is only slightly higher at 72 percent. PCIC and backlog percentages are reported in Figure 36. 75. Demand growth accounts for 47 percent of total investment requirements. Figure 37 shows the share of the total sewerage investment requirements accounted for by each of the cost components. The largest share of the cost is associated with demand growth (47 percent of total). The second largest cost component is the backlog or un-met demand, which accounts for 39 percent of the sewerage costs. Replacement costs account for 14 percent of the sewerage investment costs. 76. Investment costs are estimated at five year intervals, in line with the Planning Commission’s five-year Plans. The investment trend for the sewerage sector is presented in Figure 38. Investment requirements are estimated to increase over time. The first phase of the investment trend coincides with the 11th Plan of the Planning Commission, and the investment is estimated to be in the amount of Rs 374 Bn, or USD 8 Bn. The last investment, in the amount of Rs 534 Bn (USD 12 Bn), coincides with the 15th Plan. 77. On average, for the period 2007-31, investment requirements account for 0.4 percent of GDP per plan (based on a 7 percent real GDP growth rate). The five year trend of investment requirements for the sewerage sector is calculated as a share of GDP. Results are presented in Figure 39. The investment requirements as a share of GDP decrease over time. The first investment, coinciding with the 11th Plan represents 0.61 of GDP. The last investment, coinciding with the 15th Plan represents 0.24 percent of GDP. 78. Annual O&M costs for the sewerage sector are estimated to increase exponentially over the period 2007-31 as a result of increase in coverage. O&M costs are calculated annually based on per capita sewerage generation and coverage in any given year. As shown in Figure 46 40, annual O&M costs are forecasted to increase from Rs Bn 8 (USD 0.18 Bn) in 2007 to Rs Bn 64 (USD 1.4) in 2031. Per capita annual O&M costs are estimated at Rs 102. Figure 34: Urban Sewerage Investment Requirements Investments (Residential) 2006 Backlog Demand Growth Assets Replacement (Un-met demand) 2007-31 Network [1] Network [3] Network [5] Total Network Rs Bn: 559 Rs Bn: 700 Rs Bn: 214 Rs Bn: 1,473 $Us Million: 12,423 $Us Million: 15,549 $Us Million: 4,765 $Us Million: 32,737 Share of Total 66% Treatment [2] Treatment [4] Treatment [5] Total Treatment Rs Bn: 314 Rs Bn: 356 Rs Bn: 85 Rs Bn: 756 $Us Million: 6,987 $Us Million: 7,917 $Us Million: 1,895 $Us Million: 16,790 Share of Total 34% Total Assets Total Backlog Total Demand Growth Replacement TOTAL Rs Bn: 873 Rs Bn: 1,056 Rs Bn: 300 Rs Bn: 2,229 $Us Million: 19,401 $Us Million: 23,466 $Us Million: 6,660 $Us Million: 49,527 Share of Total 39% Share of Total 47% Share of Total 13% 47 Figure 35: Urban Sewerage Investment Requirements, 2007-31 (2009 prices), (Rs Bn) 3,000 2,544 2,500 2,229 2,000 1,913 Rs Bn, 2009 Prices 1,500 1,000 500 0 LOWER BOUND AVERAGE UPPER BOUND Note: Confidence Intervals shown above reflect variability in per capita investment costs (PCIC). Figure 36: Per Capita Investment Costs (Rs/capita) and Backlog (%), by Sub-sector 75% 72% Backlog (%) 67% 50% 25% 0% Network Treatment Figure 37: Urban Sewerage Investment Requirements, By Investment Driver, (Rs Bn) Assets Replacement 14% Backlog 39% Demand Growth 47% 48 Figure 38: Urban Sewerage Investment Requirements, 5 year trends, Rs Bn (2009 Prices) 600 Assets Replacement Demand growth Backlog 500 87 74 400 49 Rs Bn (2009 Prices) 40 49 300 242 272 182 209 151 200 100 175 175 175 175 175 0 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007- March April 2012- March April 2017- March April 2022- March April 2027- March 2012 2017 2022 2027 2032 Figure 39: Urban Sewerage Investment Requirements, Share of GDP 0.75% Assets Replacement Demand Growth Backlog 0.61% 0.08% 0.49% 0.50% 0.05% 0.41% 0.25% 0.38% 0.04% 0.31% 0.22% 0.05% 0.24% 0.25% 0.18% 0.19% 0.15% 0.29% 0.12% 0.08% 0.22% 0.17% 0.15% 0.11% 0.08% 0.00% 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007 - April 2012 - April 2017 - April 2022 - April 2027 - Average Average March 2012 March 2017 March 2022 March 2027 March 2032 per plan per year Figure 40: Annual Urban Sewerage O&M Costs (Rs Bn) and Coverage Trends (%) 100% 70 64 Annual O&M (Rs Bn) 62 59 57 60 Network coverage (%) 80% 54 52 Annual O&M (Rs Bn) 49 47 50 45 42 40 60% Coverage 38 40 36 34 32 29 27 30 40% 25 23 21 20 18 20 16 14 12 20% 10 8 - 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 49 C.IV Capital Works Unit Cost Analysis Cost drivers 79. The results indicate that some economies of scale are present in both network and treatment sub-sectors. The unit costs are negatively correlated with the incremental project capacity, with a negative correlation of 35 percent (See Figure 41). The cost model makes an attempt to capture the correlation between unit costs and cost drivers, such as city population size and population density, by estimating investment requirement by city size class. While the data gathered is not large enough to estimate investment requirements with sufficient accuracy for each city size class, significant correlation emerge from the analysis. Larger and more densely populated cities tend to have lower costs on a per capita basis for sewerage network than small, low-density towns. The per capita network of sewerage collection almost doubles from Class I.A to Class IV+. 80. It is worth noting that the sewerage cost model is based on project cost estimates rather than actual costs. This may have led to an under-estimation of the per capita costs, as cost escalation at contract award and implementation is reported to be an an issue in the sector (based on a sample of sewerage projects in Karnataka, costs at contract award are estimated to be about 20 percent higher than costs at project design stage). The scope for under-estimation is more limited in the water sector, compared to the sewerage sector –water distribution costs which account for the bulk of water capital expenditure are estimated based on a cost simulation, rather than on project data. Sensitivity analysis 81. A sensitivity analysis is conducted to assess the impact of assuming different standards on total investment requirements. As in the case of the water sector, a sensitivity analysis is carried out to assess the impact of lowering the per capita sewerage generation norm for small towns with less than 50,000 inhabitants. A reduction of the norm from 108 lpcd to 70 lpcd would lead to a negligible decline in the point estimate for the overall investment requirements from Rs 2,229 Bn to Rs 2,030 Bn (See Figure 42). 82. Sewerage network is the supply standard adopted for the cost estimation. Septic tanks are the lower supply standard considered for the sensitivity analysis. The capital expenditure of a septic tank is in the range of Rs 12,000 – 15,000. An additional per capita cost of Rs 1,000 is required for the proper disposal of effluents. The overall per capita cost of septic tanks is therefore Rs 3,400 -4,000 (based on an average household size of five). Septic tanks also require cleaning every alternate year at a cost of about Rs 1,000, equivalent to an annual per capita O&M cost of about Rs 500. 83. Based on the results of the cost model, septic tanks (with a per capita cost of Rs 3,400- 4,000) are an economical option only for Class II, III and IV+ cities (i.e. cities and towns with 50 population below 100,000). The cost savings are relatively limited on a per capita basis.24 In addition Class II-III-IV+ account for only 28 percent of the urban population (as of 2009). Assuming septic tanks as a service standard in Class II-III-IV+ cities is therefore not expected to have a significant impact on the overall urban investment requirements for the sewerage sector. Figure 41: Treatment Unit Costs and Incremental Capacity (MLD) 25,000 20,000 Unit cost (Rs/m3) 15,000 10,000 5,000 - - 100 200 300 400 500 600 700 800 Incremental capacity (MLD) Figure 42: Sensitivity Analysis: Changes in Sewerage Per capita Production Norm 3,000 2,544 2,330 Rs Bn, 2009 Prices 2,229 2,000 2,030 1,913 1,730 1,000 0 Production Norm 108 lpcd (City Class III and IV+) Production Norm 70 lpcd (City Class III and IV+) Assumption 1 Assumption 2 24The limited cost savings may be partially due to the possible under-estimation of sewerage network costs mentioned in Para. 77 51 D. Municipal Solid Waste D.I Municipal Solid Waste Service Standards DEMAND SUPPLY Per capita waste 02-06 Kg/cap/day Target C&T coverage 100% of solid waste generation (2006) generation Per capita annual 1.3 percent Target compost share 65%-50% of MSW growth rate generation Target landfill share 30-45% of MSW generation Assets’ economic life 10 yrs (C&T), 5 yrs (landfill at half of original unit costs) 84. The cost models is built on the following demand and supply standards: 85. Per capita MSW generation is estimated to range between 0.2 to 0.6 Kg/capita/day across Indian cities and depending on city size (See Figure 64 in Annex IV).25 On average per capita waste generation is estimated to increase at 1.3 percent per annum on a per capita basis over the period 2007-31. The assumption is in line with references provided in the empirical literature.26 Coverage for collection and transportation (C&T) is estimated to increase from the current level of 60 percent to 100 percent. 86. A cross-country comparison is conducted to benchmark waste generation patterns across a number of comparable countries. The per capita waste generation for Indian cities that is reference in the literature is significantly below the average for comparable countries (see Figure 43 and Figure 62 in Annex VI ). This implies that solid waste generation is expected to increase significantly as India’s process of economic and demographic transformation unfolds. The benchmarking also indicates that the urban coverage rate for C&T (estimated at 60 percent based on CDPs) is below the expected value for a sample of comparable countries (see Figure 44 and Figure 63 in Annex VI). 87. The prevailing solution for waste processing in Indian cities is composting, while sanitary disposal is virtually non-existent, with the exception of a limited number of cities. The international comparison suggests that the percentage of processed and landfill waste varies to 25 Estimated are based on 1995 NEERI data, as reported in 3i Network (2006), ‚India Infrastructure Report‛. 1995 data are updated to the base year of the model (2006) assuming a 1.3 percent annual growth rate in solid waste generation. 26 See for example Shekdar, A.V., 1999. ‚Municipal solid waste management – the Indian perspective‛. Journal of Indian Association for Environmental Management 26 (2), 100–108). 52 a significant extent across countries. The type of processing also depends on local conditions. Incineration is for example the most widespread solution in countries such as Japan where land is extremely scarce (See Figure 45). On average, the most widespread solution in both low and middle income countries is sanitary disposal in landfill sites. 88. The target compost share for the purpose of the estimation is assumed to range from 65 to 50 percent of solid waste generation. The assumption is made that the sector will gradually move from the prevailing target compost share of about 65 percent to 50 percent within a period of 10 years. Fifty percent is assumed to be the optimal target share for composting, as it is broadly in line with the share of compostable waste based on available empirical evidence. The average compostable share ranges significantly across countries, from 20 to 80 percent, based on UNSTAT data for a sample of countries (see Figure 46). In India, it is estimated that 35 to 55 of municipal waste material is organic waste, which can be converted into useful compost.27 The current widespread Indian practice of mixed composting (i.e. the composting of un-segregated waste) has led to a much higher target share of composting, which is estimated at 65 percent. Mixed composting is however not considered sustainable in the long-term. The assumption is therefore made in the model that the target compost share would decline over time to levels that are in line with the share of compostable waste. 89. The target landfill share is assumed to range from 30 to 45 percent of total solid waste generation. The assumption is made that the sector will move from the prevailing target landfill share of 30 percent to 45 percent. While very few cities safely dispose of MSW, the construction of landfill sites is in the pipeline in a number of cities. It is therefore expected that the optimal landfill share will increase over time, in spite of the land scarcity. Alternative target rates can be specified as model assumptions, to assess the impact of different solid waste policy options on investment requirements. For example, the model allows testing the cost savings associated with increase in recycling, which would reduce the target compost and landfill share. 90. The following assumptions are embedded in the cost model with respect to the economic life of the assets: C&T assets are expected to have an economic life of ten years, while the replacement costs for landfill are estimated at five year intervals, in line with information provided by JNNURM project data, which cover landfill capital expenditure for a planning horizon of five years.28 Replacement costs for composting are not included, as they are expected to be covered by compost sales. 27 See for example World Bank (2008). ‚Improving Municipal Solid Waste Management in India‛. Washington DC. 28 The replacement costs for landfill are estimated at 50 percent of the original unit costs, based on the assumption that certain fixed costs (e.g. road to the landfill sites) would only be incurred once. 53 Figure 43: Municipal Waste Generation vs. Figure 44: Population Served by Waste GNI per capita (1995) Collection vs. GNI per capita (2000-07) 1.2 100 Waste Generation (kg/cap/day) 1 80 Population Served (%) 0.8 0.6 60 India India 0.4 40 0.2 0 20 0 1,000 2,000 3,000 4,000 0 25,000 50,000 75,000 GNP per capita ($) GNI per capita ($) Source: World Bank (1999). Source: UNSTATS; CDPs for India Figure 45: Treatment and Disposal of Municipal Solid Waste (2000-07) Unknown (%) Composted (%) Recycled (%) Incinerated (%) Landfilled (%) 100 80 60 40 20 0 Algeria (2003) Chile (2006) Canada (2004) Japan (2003) Armenia (2007) Yemen (2007) Uganda (2006) United States (2005) Lithuania (2007) Colombia (2005) Morocco (2000) Korea, Rep. (2004) Syrian A. Rep. (2003) China (2003) Madagascar (2007) Mexico (2006) Australia (2003) Tunisia (2004) Poland (2007) High income Upper middle income Lower middle income Low income Source: UNSTATS. 54 Figure 46: Composition of Urban Solid Waste 1995 (Asian Countries) Others Metal Glass Plastic Paper Compostables 100 80 60 40 20 0 Lao PDR China Malaysia India Singapore Hong Kong Indonesia Philippines Nepal Myanmar Sri Lanka Japan Thailand Bangladesh High Income Middle Income Low Income Source: World Bank (1999). Notes: Classification based on 1995 GNP per capita. 55 D.II Municipal Solid Waste Methodology 91. This Section presents the overall methodological approach for the estimation of investment and O&M requirements for MSW. A summary of the approach is presented in Figure 48. A detailed step-by-step description of the methodology is provided in Annex I. 92. The methodology for estimating investment requirements consists of four building blocks: (1) service standards, which have been described in detail in the previous section; (2) investment components, (3) cost drivers (city population size and density) and (4) unit costs.  There are three main investment components. The first investment component corresponds to the backlog or un-met demand, defined as the percentage of the base year waste that is not collected & transported, properly processed and disposed. The second component corresponds to demand growth, defined as the incremental volume of waste that will need to be collected & transported, processed and disposed over the period 2007-31. Finally, the last component corresponds to assets replacement, which is the cost of replacing outdated assets.  Unit costs are estimated based on project data. Three separate unit costs are estimated for (a) C&T; (b) processing (composting) and (c) disposal.  The correlation between unit costs and city population size and population density is assessed by estimating unit costs by city size class. 93. Investment requirements are calculated as the product of each one of the investment components times the unit costs of the corresponding sub-sector (C&T, processing and disposal). Overall, the methodological approach is depicted in the diagram presented in Figure 47 below. The diagram presents a breakdown of investment requirements based on cost component (C&T, processing and disposal) and investment component (backlog, demand growth and asset re-placement). The investment requirements include: total investment requirements for the backlog [1+ 2+ 3]; total investment requirements for demand growth [4+5+6]; and assets replacement costs [7]. Investment requirements are estimated as a band based on a 90 percent confidence interval for the unit costs.29 94. O&M costs are calculated on an annual basis based on the total solid waste collected, transported and disposed. 29There is a one-to-one relationship between unit costs and total investment requirements – i.e. a 10 percent increase in unit costs leads to a 10% increase in total investment requirements. 56 Figure 47: MSW Cost Model Methodology 57 Figure 48: MSW Cost Model Methodology – Building Blocks Investments (1) Service Standards Supply: 100% target C&T, 65-50% target compost share; 30-45% target landfill share, Demand: 0.2-0.6 Kg/capita/day solid waste generation, 1.3% annual growth rate (2) Investment Components 2006 Backlog Demand Growth (2.1) (2.2) (2.3) Assets Replacement (un-met demand) 2007-2031 Collect & Transport [1] Collect & Transport [4] Collect & Transport [7] Definition: percent of base-year Definition: Incremental municipal Definition: Assets are assumed to municipal waste generation that is waste generation to be collected have a 10 year economic life prior not collected and transported. and transported over the period to replacement. Source: City Development Plans 2007-2031. (CDPs) (Sample size 51 obs.) Source: Estimated based on UN population forecasts, annual per capita growth in MSW generation. Processing [2] Processing [5] Definition: percent of base-year Definition: Incremental municipal There are no assets replacement urban waste generation that is not waste generation to be processed costs for processing because these properly treated. based on optimal compost share costs are generally funded by Source: City Development Plans (2007-2031). compost sale revenues. (CDPs) (Sample size 49 obs.) Source: Same as C&T. Disposal [3] Disposal [6] Disposal [7] Definition: percent of base-year Definition: Incremental municipal Definition: Assets are assumed to urban waste generation that is not waste generation to be disposed have a 5 year life prior to properly disposed. based on optimal landfill share replacement. Assets replacement Source: Equivalent to the entire (2007-2031). costs are estimated at 50% of Indian urban population (virtually Source: Same as C&T. original unit costs. no landfills in India). (3) Unit Costs (3.1) Collect & Transport (3.2) Processing (3.3) Disposal Definition: Estimated dividing the Definition: Estimated dividing the Definition: Estimated dividing the total project costs for C&T by the total project processing cost over total landfill project costs over the volume of waste generation in the the volume of waste processing estimated volume of waste project design year. capacity of the compost plant. disposed in the project design year. Source: 22 JNNURM Project Source: Same as for C&T (left) Source: Same as for C&T (left) Appraisal Notes (PANs); data (Sample size 27 obs.) (Sample size 33 obs.) collection from Karnataka; (Sample size 40 obs.) O&M Z Annual O&M costs are calculated based on total solid waste collected &transported and disposed (O&M costs for treatment are not included as they are generally covered by compost revenues). Source: O&M unit cost estimates have been made available by sector experts based on recent project data. V 58 D.III Municipal Solid Waste Sector Investment and O&M Requirements 95. MSW investment requirements are in the range of Rs 368-607 Bn. The point estimate for the MSW sector is Rs 487 Bn (in 2009 prices), or USD 11 Bn. The 90 percent confidence interval for the investment requirements ranges from Rs 368 Bn to Rs 607 Bn (USD 8.2 to 13.5 Bn), as shown in Figure 50. The processing sub-sector has the largest cost variation, followed by disposal; while C&T has the smallest cost variation. The methodology for the computation of interval estimates for investment requirements is discussed in Annex I; confidence intervals reflect the variability in per capita investment costs (PCIC) across projects. Investment requirements by sub-sector and cost components are reported in Figure 49Figure 34. 96. For the period 2007-31, the average PCIC for MSW is estimated to range from Rs 323 to Rs 518. As shown in Figure 51, the PCIC for the solid waste sector is expected to increase over time as a result of growth in per capita waste generation. The PCIC for C&T exhibits an increasing trend, with values ranging from Rs 117 per capita in 2007 to Rs 159 per capita in 2031, and an average value of Rs 137 per capita. The PCIC for disposal is also increasing over time. The average PCIC for disposal over 2007-31 is Rs 152 per capita; the value for 2007 is Rs 96 per capita, and the value for 2031 is Rs 187 per capita. The PCIC for treatment is somewhat flatter overtime, with an average value of Rs 131 per capita. This is due to the fact that the target share of per capita solid waste generation that is expected to be sent to compost plant is assumed to decline over time from 65 to 50 percent. 97. Asset replacement is the main investment component. Figure 52 shows the share of the total sewerage investment requirements accounted for by each of the cost components. The largest share of investment requirements is associated with assets replacement, which accounts for 42 percent of the total investment requirements. The second largest cost component is demand growth, which accounts for 39 percent of total capital expenditure requirements. Backlog accounts for the remaining 19 percent of MSW capital expenditure. 98. The disposal sub-sector accounts for the largest share of investment requirements. The disposal sub-sector accounts for 42 percent of total capital expenditure, followed by C&T (37 percent) and processing (21 percent). Disposal has the highest backlog (100 percent), followed by processing (93 percent) and C&T (41 percent). See Figure 51. 99. The investment trend is estimated based on the assumption that full coverage would be achieved by the end of the 15th Plan. Investment requirements are expected to increase over time. The first phase of the investment trend coincides with the 11th Plan of the Planning Commission, and the investment is estimated to be in the amount of Rs 62 Bn (in 2009 prices). The last investment, in the amount of Rs 152 Bn, coincides with the 15th Plan. The investment trend for the solid waste sector is presented in Figure 53. 100. Investment requirements as a share of GDP decrease over time. The five year trend of investment requirements for the solid waste sector is computed as a share of GDP, based on the assumption of 7 percent real GDP growth. On average, for the period 2007-31, the 59 investment requirements account for 0.08 percent of GDP per plan. The first investment, coinciding with the 11th Plan represents 0.10 percent of GDP. The last investment, coinciding with the 15th Plan represents 0.07 percent of GDP. Results are presented in Figure 54. 101. Annual O&M costs are estimated to increase over time as a result of the increase in coverage. There is an increasing trend in annual O&M costs, as shown in Figure 55. Annual O&M costs for the MSW sector are estimated to increase from Rs 31 Bn (USD 0.7 Bn) in 2007 to Rs 124 Bn (USD 2.8 Bn) in 2031 (in 2009 prices). The per capita annual O&M is estimated at Rs 190. Figure 49: MSW Investment Requirements Investments (Residential) 2006 Backlog Demand Growth Assets Replacement (Un-met demand) 2007-31 Total Collect & Collect & Transport [1] Collect & Transport [4] Collect & Transport [7] Transport Rs Bn: 14 Rs Bn: 58 Rs Bn: 109 Rs Bn: 182 $Us Bn: 315 $Us Bn: 1,297 $Us Bn: 2,422 $Us Bn: 4,034 Share of Total 37% Processing [2] Processing [5] Total Processing Rs Bn: 51 Rs Bn: 54 Rs Bn: 106 $Us Bn: 1,139 $Us Bn: 1,206 $Us Bn: 2,345 Share of Total 22% Disposal [3] Disposal [6] Disposal [7] Total Disposal Rs Bn: 26 Rs Bn: 76 Rs Bn: 98 Rs Bn: 200 $Us Bn: 508 $Us Bn: 1,698 $Us Bn: 2,169 $Us Bn: 4,453 Share of Total 41% Total Assets Total Backlog Total Demand Growth Replacement TOTAL Rs Bn: 92 Rs Bn: 189 Rs Bn: 207 Rs Bn: 487 $Us Bn: 2,040 $Us Bn: 4,201 $Us Bn: 4,592 $Us Million: 10,832 Share of Total 19% Share of Total 39% Share of Total 42% 60 Figure 50: MSW Investment Requirements, 2007-31 (2009 prices), (Rs Bn) 800 Rs Bn, 2009 prices 607 600 487 400 368 200 0 Lower Bound Average Upper Bound Note: Confidence Intervals shown above reflect variability in per capita investment costs (PCIC). Figure 51: Per Capita Investment Costs (Rs/capita) and Backlog (%), by Sub-sector 125% 100% 100% 93% Backlog (%) 75% 50% 41% 25% 0% C&T Processing Disposal m Figure 52: MSW Investment Requirements (2007-31), by Component, (Rs Bn) Backlog 19% Assets Replacement 42% Demand Growth 39% 61 Figure 53: MSW Investment Requirements, 5 year trends (Rs Bn) 175 Assets Replacement 152 150 Demand Growth Rs Billion (2009 Prices) 125 Backlog Investments 104 108 82 100 48 46 75 62 61 50 18 13 44 52 26 30 38 25 18 18 18 18 18 0 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007- March April 2012- March April 2017- March April 2022- March April 2027- March 2012 2017 2022 2027 2032 Figure 54: MSW Investment Requirements, Share of GDP Backlog Demand Growth Assets Replacement 0.10% 0.10% 0.09% 0.03% 0.08% 0.08% 0.07% 0.07% 0.02% 0.04% 0.03% 0.05% 0.04% 0.03% 0.04% 0.04% 0.03% 0.03% 0.03% 0.02% 0.03% 0.02% 0.02% 0.02% 0.02% 0.00% 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan April 2007 - April 2012 - April 2017 - April 2022 - April 2027 - Average Average March 2012 March 2017 March 2022 March 2027 March 2032 per plan per year Figure 55: Annual MSW O&M Costs (Rs Bn) and Coverage Trends (%) Annual O&M Coverage C&T Coverage Disposal 100% 140 124 119 114 109 120 80% 104 99 95 Coverage (%) 100 90 86 60% 82 78 74 80 Rs Bn 70 66 63 59 55 60 40% 52 49 45 42 39 36 34 40 31 20% 20 - 0% 2011 2007 2008 2009 2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 62 D.IV Capital Works Unit Cost Analysis Cost drivers 102. The most common technology for processing solid waste in India is composting. The results indicate that there are no significant economies of scale in processing. The main cost driver for composting is geography. When roof cover in coastal areas is to be done costs are significantly higher, as it is the case in Trivandrum, Chennai and Imphal. (See Figure 56). City size and density are not shown to be significantly correlated with unit costs. This may be related to the fact that land costs are not included in the investment cost estimation (JNNURM does not cover the cost of land acquisition), as land costs tend to be significantly higher in large cities compared to small and medium towns. Sensitivity analysis 103. A sensitivity analysis is conducted to assess the impact of changing service standards on investment requirements. Assuming a different combination of target compost and landfill share does not have a major impact on overall project costs. For example, assuming a target composting and landfill share equal to the transition share (65 percent for composting and 30 percent for landfill) reduces total investment costs by roughly 5 percent, from Rs 487 to Rs 469 Bn. Growth in per capita waste generation has a more significant impact on overall capital expenditure. For example, increasing the per capita growth rate from 1.3 to 1.6 percent is estimated to increase overall investment requirements by 13 percent, from Rs 487 to Rs 558 (see Figure 57 and Figure 58). 63 Figure 56: Unit Processing Costs (Rs/TPD): Economies of Scale (2009 prices) 3,500,000 Non Coastal Cities 3,000,000 Coastal Cities 2,500,000 Unit Cost (Rs/TPD) 2,000,000 1,500,000 1,000,000 500,000 - - 100 200 300 400 500 600 Incremental Capacity (TPD) Note: One project has been excluded from the graph because its plant capacity is more than three standard deviations away from the average. Figure 57: Sensitivity Analysis: Changes in Share of Waste Treated and Disposed 607 584 600 Rs Bn, 2009 Prices 487 469 400 368 355 200 0 Target: Composting 50% and Landfill 45% Target: Composting 65% and Landfill 30% Transition: Composting 65% and Landfill Transition: Composting 65% and Landfill 30% 30% Assumption 1 Assumption 2 Figure 58: Sensitivity Analysis: Changes in Growth Rate of Solid waste Generation 700 695 607 600 558 Rs Bn, 2009 Prices 500 487 400 422 300 368 200 100 0 Per Capita Solid Waste Growth Rate of Per Capita Solid Waste Growth Rate of 1.3% 1.6% Assumption 1 Assumption 2 64 E. Investment Needs -A Part of a Complex Picture 104. The results of the cost estimation give an indication of the magnitude of the physical investment requirements in the urban WSS and MSW sectors, and the growing O&M requirements as sector coverage increases. Total investment requirements as a share of GDP range from 1.6 to 0.6 percent over the period 2007-31, with an average of 0.9 percent. While the shares are purely indicative, as they depend on the profile of the investments, they are significantly above the current share of capital expenditure in municipal services, and therefore raise important questions with respect to the sectors’ absorptive capacity. Is the current policy framework appropriate to support the much needed scale-up of municipal investments? What sector interventions are needed to minimize the costs and maximize the impact of physical investments? How to ensure that these investments translate into effective and efficient service delivery? 105. Physical investments are a necessary component for the delivery of municipal services in an efficient, effective and sustainable manner. However, investments alone are not sufficient. These investments will not deliver their potential unless they are complemented by actions to strengthen autonomy & accountability of service providers, enhance incentives, and improve professionalization. The rest of this section exemplifies how incentives and professionalization, combined with adequate investments, can work together in improving the delivery of urban WSS services. 106. Current institutional arrangements for urban WSS service provision in India do not reflect good international practices where well run public water companies share the characteristics of reasonable financial and managerial autonomy, clear accountability to stakeholders and a strong customer orientation. Moving to such an institutional framework will take time, and States and cities need support and guidance on how to make such changes. 107. First, people respond to incentives. The current institutional arrangements in the WSS sector provide low incentives with regards to service delivery to customers and (operational and capital) efficiency. It is worthwhile considering the impact appropriate incentives might have on making maximum use of investment funds. In the United Kingdom, where a high incentive regulatory structure is in place, unit capital expenditure to deliver the required levels of service reduced by almost 25 percent over the 10 year period from 1994-2003. The impact of such levels of savings in India would be significant. See Box 1 for more details on the England and Wales’s regulatory system. 108. Second, high quality technical, managerial and commercial skills are needed to make the most of existing assets, and to ensure that new investments provide good service for their full economic life. Inadequate skills will mean sub-optimal performance of assets, which result in poor quality of service to customers. This in turn will make it more difficult to engage with customers in the steps needed to turn around ailing service providers – including for example, improving commercial operations of the providers, and reducing losses. 65 109. Inadequate skills, coupled with insufficient revenues, also mean that assets are not properly operated and maintained leading to shortened asset lives. It has been estimated that the economic loss associated with shortened asset lives in the urban WSS sector amounts to Rs 438 Billion (USD 10 Billion) over the next forty years (2009 prices). See Box 2 for details. While this is a coarse assessment, it is clear that investment in human capacity can provide significant pay back - not just in terms of deferring future capital expenditures, but also in improving quality of service to customers. In other words the less obvious need to invest in professionalization of the sector must be addressed alongside the more obvious need to invest in assets. 66 Box 4: Economic Regulation in the England and Wales WSS Sector An institutional environment that provides utilities with incentives for efficiency improvements can gradually lead to a reduction in the unit cost of service provision, and ultimately reduce the capital expenditure requirements for the WSS industry as a whole. The England and Wales’ WSS sector is a case in point. Following the 1989 privatization of the WSS sector, England and Wales has created an institutional and regulatory environment that incentivizes efficiency improvements. In the England and Wales WSS sector, the hallmark of economic regulation is the use of price caps combined with yardstick or comparative competition. The price caps set the maximum prices that utilities are allowed to charge over a five year period. Water prices are set at the efficient level that allows water companies to finance their functions, including the earning of an adequate return on investments. An autonomous economic regulator, the Office of Water or Ofwat, was established to minimize political interference in the price-setting process. Since the water sector was privatized, Ofwat has undertaken four reviews of water charges – the 1994, 1998, 2004, and 2010 Periodic Reviews. Price-cap regulation is one of high-powered incentives: companies have an incentive to reduce their costs beyond the efficiencies expected by the regulator, as they can retain the financial benefits for distribution to their stakeholders. The efficiency incentives generated by price cap regulation are based on a system of yardstick competition that rewards companies that achieve efficiency gains above the industry average, and push the industry’s revealed efficiency frontier forward. The system of yardstick competition established by Ofwat covers both inputs - operating and capital expenditure - as well as outputs – as measured by a series of quality of service indicators. For operating and capital expenditure, Ofwat’s efficiency assessment is based on a combination of econometric techniques and unit costs analysis. A comparative capital unit cost approach, known as the ‚cost base‛ was developed by Ofwat to compare capital expenditure across the industry and identify those companies that appeared to be more efficient at procuring capital assets than others. Companies with higher unit capital costs were considered to have more scope for savings in their expenditure projections that companies with lower capital unit costs. The analysis conducted by Ofwat as part of the 1998 and 2004 Periodic Review indicates that substantial efficiency gains have been achieved by the industry as a whole since economic regulation was introduced. Over the period 1994-98, unit capital expenditure decreased by 10 percent for water services and by 15 percent for sewerage services across the industry as a whole.30 Over the period 1998-2003, unit capital expenditure declined by an additional 15 percent for water services and by 10 percent for sewerage services. 31 Improved procurement processes and the use of innovative techniques were reported by the companies as the main drivers of capital expenditure reductions. 32 Improvements in efficiency are reflected in future price limits resulting in lower bills for consumers. Reductions in capital works unit costs also create space for a larger maintenance program and other improvements to be carried out at no extra costs for consumers. 30 Ofwat (1998), ‚Capital Works Unit Costs in the Water Industry: An Analysis of the June 1998 Water Company Cost Base Submissions‛. 31 Ofwat (2004), ‚Capital Works Unit Costs in the Water Industry: Feedback on Our Analysis of the March 2003 Water Company Cost Base Submissions‛. 32 It is important to note that efficiency savings explain a large part, but not the totality, of the industry reduction. Changes in companies’ methodologies contribute in some part to the variances in the capital unit costs. 67 Annex I: Methodological Approach Urban Water Supply STEP-BY-STEP METHODOLOGY Urban water investment requirements are estimated for both residential and industrial customers for the period 2007-31. O&M requirements are computed on an annual basis over the same period. Investment requirements are estimated separately for water production (source augmentation and transmission) and distribution (network, storage and metering). The methodological approach adopted for the estimation consists of the following main steps: Step I: Define sector targets to be achieved by 2031 Defining service standards based on demand and supply considerations is the first step for estimating investment requirements for urban water supply. The cost model is based on the assumption of homogeneous service standards across urban India. The supply-side service standards embedded in the cost model are the following: (a) 100 percent target piped water supply coverage (for both water production and distribution), (b) 24/7 water supply continuity, and (c) a 20 percent leakage. The demand-side service standards are the following: (d) a per capita residential consumption norm of 135 lpcd, and (e) a 7 percent annual increase in industrial water demand for cities with population above 500,000. Step II: Classify Sample Cities based on Population All cities for which the data is available are classified into six categories on the basis of their population as reported in the census. The first census class, including all cities with population above 100,000, is further divided in three sub-categories: Class I.A (megacities, > 5 million), Class I.B (1-5 million) and Class I.C (100k-1 million). 33 The classification is based on Urban Agglomeration (UA) population whenever applicable. The rationale for the classification of sample cities into size classes is to capture differences in capital expenditure norms across the spectrum of urban cities. See Table 1. Table 1: City Size Classes Class I.A 7 Mega-cities, > 5 million Class I.B 1 – 5 million Class I.C 100,000 – 1 million Class II 50,000 – 100,000 Class III 20,000 – 50,000 Class IV+ < 20,000 33 The seven mega-cities are Delhi, Mumbai, Hyderabad, Kolkata, Chennai, Bangalore and Ahmadabad. All megacities currently have population above 5 million. Note that Ahmadabad is classified in Class I.A but was not yet a megacity in 2001 being its population below 5 million. 68 Step III: Estimate Urban Households Water Backlogs The backlog percentages for water distribution are calculated based on 2001 Population Census data. Water distribution backlogs are defined as the percentage of the current urban population with no private water connection. It is assumed that the 2001 backlog percentages, as estimated based on census data, apply to 2006.34 Backlog percentages for water production are estimated based on City Development Plans (CDPs)’ data, most of which were prepared around 2006 when the JNNURM program was launched. Water production backlogs are computed as the difference between city-level production capacity (net of industrial water use), as reported in the CDPs, and residential production requirements. To ensure comparability across cities, residential production backlogs are estimated based on the per capita production norm of 168 lpcd across all city size classes. The backlog for 24/7 up- gradation is assumed equal to the entire urban population connected to water supply, given that virtually no Indian city currently benefit from 24/7 water supply continuity. The backlog percentage for each of the six classes of cities is calculated as the average backlog for all cities in that category weighted by city population. Similarly, the backlog percentage for all urban India is estimated as the average backlog for all city size categories weighted by the total urban population in each category. Step IV: Project Urban Population by City Size Class The 2001 census population is taken as the base year population for the cost model. Population forecasts are based on estimates provided by the United Nations Population Division of the Department of Economic and Social Affairs (DESA) in its regular statistical publication, the World Urbanization Prospects (2007 revision). For each city class, population is forecasted by applying the UN population growth rates to 2001 census population figures over the period 2001-2031. There is no complete alignment between the six Census of India city classes, as reported in this study, and the UN population classes. More specifically, the UN projection model provides estimates for only 5 broad city classes – the lowest class including all cities with population below 500,000. As a result, the same growth rate is applied to all Indian cities with population below 500,000. Step V: Calculate 2006 Backlog Population and Incremental Urban Population (2007-2031) The 2006 population for each class of cities is multiplied by the corresponding backlog percentage to calculate the total backlog population for production, 24/7 up-gradation and distribution extension. The total additional population over the period 2007-2031 for each class of cities is calculated by deducting the 2006 population from the projected population of the year 2031. 34 Census data is used to estimate distribution backlogs instead of CDPs for the following two reasons: first, CDPs are only available for a small sample of cities (most of which are in Class I.A-I.C), second a few cities, including megacities, do not include slum area in the estimation of the coverage rate. In spite of the significant difference in the coverage rate estimated based on the two data sources, using the CDPs rather than the census figures for the backlog estimation has a negligible impact on total investment requirements. 69 Step VI: Estimate Unit and Per Capita Investment Costs for Residential Water Production The unit and Per Capita Investment Costs (PCICs) for production are calculated based on JNNURM project data. The PCICs are computed by dividing total JNNURM production costs by the target beneficiaries. Beneficiaries are computed by dividing the project design-year incremental capacity by the 168 lpcd production norm (this methodology ensures a consistent approach in the definition of project beneficiaries across all JNNURM projects, given that a number of JNNURM Project Appraisal Notes do not provide information on design-year project population). The unit costs for water production are estimated by dividing total project cost by the project incremental capacity. Step VII: Estimate Water Production Investment Requirements for Industrial Customers The investment requirements to meet the production requirements of industrial customers are estimated for all Indian cities with population above 500,000. The assumption is made that about 20 percent of the production capacity of Indian cities is dedicated to industrial customers (based on expert estimates and average percentages reported by CDPs). Industrial production requirements are estimated to increase in line with the annual growth rate for industry value added (at about 7 percent per annum). City-level production investment requirements for industrial customers are then estimated for each city in the sample based on JNNURM average unit production costs for city size classes. Total investment requirements are computed by multiplying city-level investment needs by the total numbers of cities in each city class, estimated based on UN forecasts. Investment requirements for industrial water production are only calculated for cities with population above 500,000, because UN estimates on numbers of cities are only available for Class I.A and Class I.B cities and a sub-set of Class I.C cities (1m - 500k). Step VIII: Estimate PCIC for 24/7 Up-gradation Estimating the per capita costs of upgrading the existing distribution network to achieve 24/7 water supply continuity is methodologically complex, given that virtually no Indian city has 24/7 water supply with the exception of a few pilot projects (e.g. Karnataka towns and Nagpur in Maharashtra). It is also complex because the solution to delivering 24/7 is a mix of rehabilitation of old assets (e.g. to fix leaks in old pipes and service connections), new assets (e.g. improved network layout, creation of district meter areas for leakage management), and (importantly) significantly improved distribution system management. Assumptions have therefore been made on the type of investment required to achieve 24/7 water supply across the various categories of cities and towns. In the absence of project data, a cost simulation exercise has been conducted to estimate the PCIC for 24/7 up-gradation. The cost simulation exercise has been complemented by cost data based on a small number of 24/7 up-gradation pilot projects.  Estimating the cost of partially replacing the distribution system (cost simulation). Many of the water distribution networks in Indian Cities are old and poorly constructed. The share of the distribution network that needs to be replaced to achieve 70 24/7 varies from city to city. In the absence of accurate city-level information on the conditions of the water distribution assets, the estimate is based on the assumption that on average 50 percent of the distribution network needs to be replaced. The same replacement share is applied to water storage and household connection/metering costs. The replacement share is assumed to range from 40 percent for Class I.A cities to 60 percent to Class III and IV+. The variation across city size classes takes into account the fact small cities may have to replace a higher share of the network because of the lower maintenance standards, compared to the largest cities. For the purpose of the cost simulation, the per capita optimal storage capacity is assumed to be equivalent to one third of per capita water consumption. The cost of partially replacing the distribution network is estimated on the basis of information provided by the City Development Plans on the length of the distribution network per connected person, and the diameter of pipes for the primary, secondary and tertiary network. Variation in the length of the network per connected person across cities captures differences in population density across city size classes. The cost of pipes of different diameters has been provided by experts based on recent project data. Metering and storage costs are estimated based on cost norms gathered through the data survey and expert estimates.  Estimating the costs of achieving 24/7 water supply (project data). Very few JNNURM projects are designed to achieve 24/7 water supply continuity. Nevertheless, a number of 24/7 water pilot projects have been recently carried out across a number of cities in India. The actual and proposed costs for a small sample of 24/7 projects have been collected and included in the model (see Table 24). All 24/7 up-gradation pilot projects envisage a partial replacement of the distribution network. The share of the distribution network that is replaced to deliver 24/7 varies across projects, from 90 percent in Hubli to 40 percent in Nagpur. Step IX: Estimate PCIC for Extending the Distribution Network (based on 24/7 Standards) Delivering 24/7 water supply to the un-connected urban population requires an expansion of the existing distribution system (based on 24/7 standards), household connection and metering installation and the building of an optimal level of water storage capacity. The following methodological approach has been to calculate distribution extension PCICs for the sample cities. First, in a given city the per capita distribution network requirements are estimated based on the length of the distribution network for the served population (including primary, secondary and tertiary network). It is assumed that the average per capita distribution pipes’ extension that is needed to connect the un-served population is equal to 50 percent of the per capital length of the distribution pipe for the served population (given that part of the network is expected to be already in place). The per capita cost of extending the distribution network is estimated based on cost norms for a mix of water pipes. Second, the per capita cost of household connections, metering and water storage are computed based on the methodology described above as they are necessary conditions for the provision of 24/7 water supply. Step X: Estimate Total Investment Needs for the Period 2007 – 31 Water investment requirements are estimated as the total of the following cost components: 71  Investments for 2006 backlog (un-met demand)  Investments for demand growth (2007-2031)  Capital re-placement costs for residential water production  Production investments for industrial customers Investments for backlog: The backlog population for production, 24/7 up-gradation and distribution extension are multiplied by the PCIC for the respective cost components to estimate total backlog investments. Investments for demand growth: To calculate the investment requirements for the additional urban population, the PCIC for production and distribution extension (24/7 standards) are multiplied by the incremental urban population, as any additional urban population will need investment in both production and distribution. To simulate the profiling of the investment, the incremental investments are computed on a five year basis. Assets replacement costs (production): Assets replacement costs for residential water production are estimated based on the assumptions that assets fully depreciate in 30 years. For example, 1981 assets would reach the end of their economic life and would need to be replaced in 2011, while in 2021 (2031), replacement will be required for the 1991 (2001) connected population. Production investments for industrial customers: Production investment costs for industrial customers for cities with population above 500,000 have been estimated and included in the total investment requirements. To account for the fact the most JNNURM approved costs are expressed in 2006 prices, the total investment needs are then converted in 2009 prices. Step XI: Estimate Annual O&M Needs for the period 2007 – 31 Annual O&M requirements are estimated based on total volume of water production in each year for residential and industrial customers. The total annual O&M requirements are estimated to increase over time as the connection rate increases. O&M unit cost estimates have been made available by sector experts based on recent project data. Distance to water sources is the main factor affecting O&M unit costs – the further away are water resources, the higher the O&M costs. Variation in unit costs across cities reflect the fact that large cities are generally further away to water sources than small and medium towns. Step XII: Conduct Sensitivity Analysis Investment requirements are calculated as a band rather than a single point estimate. The band reflects the reliability of the investment estimates, which in turn depends on the accuracy of the PCICs, backlog percentage and population forecasts. Among these three factors, the PCIC is the most critical variable affecting investment estimates – a 10 percent increase in PCIC leads to an equivalent increase in investment requirements – while variation in the other two variables 72 lead to a less than proportional change in investment requirements. The standard deviation is calculated to measure the variation of the PCICs for each of the sub-sectors.35 Based on the standard deviation, a 10 percent confidence interval for the total PCIC is estimated. Given that a change in the PCIC leads to a proportional change in total investment requirements, the low and high estimate boundaries of the confidence interval for the PCICs can be converted into a range of values for the investment requirements. Step XIII: Simulate the profile of the investments In order to simulate the profile of the investments, investment requirements are estimated in line with the 5-year Planning Periods. For example, the 2006 base year corresponds to the period April 2006-March 2007, and the first five year interval (2007-2011) correspond to the 11th Plan (April 2007 - March 2012). For the purpose of the simulation, it is assumed that full coverage would be achieved by the end of the 15th Plan (2027-2032). 35The standard deviation is a good proxy of the accuracy of a variable. A low standard deviation indicates that the data points tend to be very close to the average value of the distribution, whereas high standard deviation indicates that the data are spread out over a large range of values. The standard deviation also depends on the sample size: everything else being equal, the larger the sample size, the more accurate the estimates, and the lower the standard deviation. 73 Urban Sewerage STEP-BY-STEP METHODOLOGY Sewerage investment requirements for the urban population are estimated for the period 2007- 31. O&M requirements are computed on an annual basis over the same period. Investment requirements are estimated separately for network and treatment. The methodological approach adopted for the estimation consists of the following main steps: Step I: Set Service Targets for Urban Sewerage Defining service standards based on demand and supply considerations is the first step for estimating investment requirements for urban sewerage. The cost model is based on the assumption of homogeneous service standards across urban India. The standards embedded in the cost model are (a) 100 percent target coverage rate for both sewerage collection and treatment (supply-side) and (b) a per capita residential sewerage generation of 108 lpcd, i.e. 80 percent of per capita water consumption, over the period 2007-2031 (demand-side). The objective of the cost model is to estimate the investment requirements to meet the specified sector targets. Step II: Classify Sample Cities based on Population All cities for which the data is available are classified into six categories on the basis of their population as reported in the census. The first census class, including all cities with population above 100,000, is further divided in three sub-categories: Class I.A (megacities, > 5 million), Class I.B (1-5 million) and Class I.C (100k-1 million). 36 The classification is based on Urban Agglomeration (UA) population whenever applicable. The rationale for the classification of sample cities into size classes is to capture differences in expenditure norms across the spectrum of urban cities. See Table 2. Table 2: City Size Classes Class I.A 7 Mega-cities (> 5 million) Class I.B 1 – 5 million Class I.C 100,000 – 1 million Class II 50,000 – 100,000 Class III 20,000 – 50,000 Class IV+ < 20,000 36 The seven mega-cities are Delhi, Mumbai, Hyderabad, Kolkata, Chennai, Bangalore and Ahmadabad. All megacities currently have population above 5 million. Note that Ahmadabad is classified in Class I.A but was not yet a megacity in 2001 being its population below 5 million. 74 Step III: Estimate Backlog Percentages for Urban Sewerage Backlog percentages for sewerage network are estimated based on 2001 Population Census data. It is assumed that the 2001 backlog percentage applies to the base year 2006.37 Backlog percentages for treatment are estimates for a sample of cities based on City Development Plans, most of which were prepared around 2006 when the JNNURM program was launched, and information provided through data collection. The estimation of treatment backlogs is based on a demand-driven per capita sewerage generation norm of 108 lpcd. The treatment backlog at the city level is computed based on the difference between current treatment capacity, as provided in the CDPs, and treatment requirements to meet the per capita norms. The average backlog percentage for each of the six classes of cities is calculated as the average backlog for all cities in that category weighted by city population. Similarly, the backlog percentage for all urban India is estimated as the average backlog for all city size categories weighted by the total urban population in each category. Step IV: Forecast Urban Population over the Period 2001-2031 The 2001 census population is taken as the base year population for the cost model. Population forecasts are based on estimates provided by the United Nations Population Division of the Department of Economic and Social Affairs (DESA) in its regular statistical publication, the World Urbanization Prospects (2007 revision). For each city class, population is forecasted by applying the UN population growth rates to 2001 census population figures over the period 2001-2031. There is no complete alignment between the six Census of India city classes, as reported in this study, and the UN population classes. More specifically, the UN projection model provides estimates for only 5 broad city classes – the lowest class including all cities with population below 500,000. As a result, the same growth rate is applied to all Indian cities with population below 500,000. Step V: Calculate 2006 Backlog and additional population (2007-31) The 2006 population for each class of cities is multiplied by the corresponding backlog percentage to calculate the total backlog population for both network and treatment for each class of cities for the year 2006. The total additional population over the period 2007-2031 for each class of cities is calculated by deducting the 2006 population from the projected population of the year 2031. Step VI: Estimate Unit and Per Capita Investment Costs (PCIC) Unit and Per Capita Investment Costs (PCICs) are calculated separately for each of the two sub-sectors: (a) network and (b) treatment. For each sub-sector, project costs are divided by the beneficiary project population to calculate the PCIC. For cities with more than one project, the city-level PCICs are estimating by averaging out the PCICs for all projects in that city. Project 37 Census data is used to estimate network backlogs instead of CDPs for the following two reasons: first, CDPs are only available for a small sample of cities (most of which are in Class I.A-I.C), second a few cities do not include slum area in the estimation of the coverage rate. However, using the CDPs rather than the census figures for the backlog estimation has a negligible impact on total investment requirements. 75 beneficiaries are calculated by dividing the project design-year incremental capacity by the per capita sewerage generation norm (108 lpcd). This is to ensure a consistent approach in the definition of project beneficiaries across all JNNURM projects, given that a number of JNNURM Project Appraisal Notes do not provide information on design-year project population. When no specific reference is made in the Project Appraisal Notes, pumping costs when present are allocated to the treatment sub-sector. Step VII: Calculate Total Investment Requirements for the period 2007-2031 Investment requirements include the following cost components:  Investments for 2006 backlog (un-met demand)  Investments for demand growth (2007-2031)  Assets replacement costs Investments for backlog: The backlog populations for network and treatment are multiplied by the PCIC for the respective cost components. The two cost components are then added to compute the total investment costs for the backlog. The baseline scenario is based on the assumption that full coverage is achieved by 2031. Investments for demand growth: To calculate the financing required for providing sewerage services to the additional urban population, the total PCIC costs for network and treatment is multiplied by the incremental population over the period 2006-2031. The underlying assumption is that any additional urban population will need to have access to all sewerage services (i.e. network and treatment). Assets replacement costs: The replacement costs for sewerage network and treatment are estimated based on an economic life of the assets of 30 years. For example, it is assumed that the assets replacement for the population covered in 1981 will be made in 2011, assets replacement for the population covered in 1991 (2001) will be made in 2021 (2031). All project costs and investment requirements are converted in 2009 prices. Step VIII: Calculate Annual O&M Costs for the period 2007-2031 Annual O&M requirements costs are calculated based on annual sewerage generation, which is estimated based on projected coverage rate and a per capita sewerage generation norm of 108 lpcd. The total annual O&M requirements are estimated to increase over time as the connection rate increases. The unit O&M costs are estimated based on sector estimates and Bank’s project data. Step IX: Conducting Sensitivity Analysis Investment requirements are calculated as a band rather than a single point estimate. The band reflects the reliability of the investment estimates, which in turn depend on the accuracy of the PCICs, backlog percentage and population forecasts. Among these three factors, the PCIC is 76 the most critical variable affecting investment estimates – a 10 percent increase in PCIC leads to an equivalent increase in investment requirements – while variation in the other two variables leads to a less than proportional change in investment requirements. The standard deviation is calculated to measure the variation of the PCICs for each of the two sub-sectors.38 Based on the standard deviation, a 10 percent confidence interval for the total PCIC is estimated. Given that a change in the PCIC leads to a proportional change in total investment requirements, the low and high estimate boundaries of the confidence interval for the PCICs can be converted into a range of values for the investment requirements. Step X: Simulate the profile of the investments In order to simulate the profile of the investments, investment requirements are estimated in line with the 5-year Planning Periods. For example, the 2006 base year corresponds to the period April 2006-March 2007, and the first five year interval (2007-2011) correspond to the 11th Plan (April 2007 - March 2012). For the purpose of the simulation, it is assumed that full coverage would be achieved by the end of the 15th Plan (2027-2032). 38The standard deviation is a good proxy of the accuracy of a variable. A low standard deviation indicates that the data points tend to be very close to the average value of the distribution, whereas high standard deviation indicates that the data are spread out over a large range of values. The standard deviation also depends on the sample size: everything else being equal, the larger the sample size, the more accurate the estimates, and the lower the standard deviation. 77 Municipal Solid Waste STEP-BY-STEP METHODOLOGY Municipal solid waste investment requirements for the urban population are estimated for the period 2007-31. O&M requirements are computed on an annual basis over the same period. Investment requirements are estimated separately for Collection & Transportation (C&T), processing and disposal. The methodological approach adopted for the estimation consists of the following main steps: Step I: Set Service Targets for Urban Solid Waste Management Defining the service standards based on demand and supply considerations is the first step for estimating investment requirements for municipal solid waste. The demand-side standards embedded in the cost model are (a) per capita waste generation ranging from 0.2 to 0.6 kg/capita/day, (b) 1.3 percent annual per capita solid waste generation.39 The supply-side standards are the following: (c) a target collection and transportation coverage of 100 percent, (d) a target compost share of 65-50 percent; (e) a target landfill share of 30- 45 percent. For composting, the assumption is made the MSW sector will gradually move from the prevailing target compost share of 65 percent to 50 percent within a period of 10 years. For landfill, it is assumed that the sector will move from the prevailing 30 percent target to 45 percent over a period of 10 years. From 2017 onward, the target compost and landfill share are therefore expected to be 50 and 45 percent respectively. Step II: Classify Sample Cities based on Population All cities for which the data is available are classified into six categories on the basis of their population as reported in the census. The first census class, including all cities with population above 100,000, is further divided in three sub-categories: Class I.A (megacities40, > 5 million), Class I.B (1-5 million) and Class I.C (100k-1 million). The classification is based on Urban Agglomeration (UA) population whenever applicable. The rationale for the classification of sample cities into size classes is to capture differences in expenditure norms across the spectrum of urban cities. See Table 3. 39 See Shekdar, A.V., 1999. ‚Municipal solid waste management – the Indian perspective‛. Journal of Indian Association for Environmental Management 26 (2), 100–108. 40 The seven mega-cities are Delhi, Mumbai, Hyderabad, Kolkata, Chennai, Bangalore and Ahmadabad. All megacities currently have population above 5 million. Note that Ahmadabad is classified in Class I.A but was not yet a megacity in 2001 being its population below 5 million. 78 Table 3: City Size Classes Class I.A 7 Mega-cities, > 5 million Class I.B 1 – 5 million Class I.C 100,000 – 1 million Class II 50,000 – 100,000 Class III 20,000 – 50,000 Class IV+ < 20,000 Step III: Estimate Backlogs for Urban Solid Waste Backlog percentages for solid waste C&T, processing and disposal are estimated based on City Development Plans. The backlog percentage for each of the six classes of cities is calculated as the average backlog for all cities in that category weighted by city population. Similarly, the backlog percentage for all urban India is estimated as the average backlog for all city size categories weighted by the total urban population in each category. Step IV: Forecast Urban Population over the period 2001-31 The 2001 census population is taken as the base year population for the cost model. Population forecasts are based on estimates provided by the United Nations Population Division of the Department of Economic and Social Affairs (DESA) in its regular statistical publication, the World Urbanization Prospects (2007 revision). For each city class, population is forecasted by applying the UN population growth rates to 2001 census population figures over the period 2001-2031. There is no complete alignment between the six Census of India city classes, as reported in this study, and the UN population classes. More specifically, the UN projection model provides estimates for only 5 broad city classes – the lowest class including all cities with population below 500,000. As a result, the same growth rate is applied to all Indian cities with population below 500,000. Step V: Calculate 2006 Backlog (un-met demand) and demand growth over 2007-31 The 2006 waste generation estimates for each class of cities is multiplied by the target C&T, compost and landfill shares to estimate the total amount of waste that would need to be composted and sent to landfill sites in the base year. The waste volume for C&T, processing and disposal is multiplied by the corresponding backlog percentages to calculate the total backlog waste (un-met demand) for the year 2006. The additional waste generation for the period 2007-31 is estimated based on population forecasts and a 1.3 percent annual increase in per capita waste generation. The target C&T, processing and disposal shares are applied to total waste generation to forecast the total volume of waste that would need to be collected & transported, composted and sent to a landfill site in a given year. Step VI: Estimate Unit Investment Costs Unit Investment Costs are calculated separately for each of the three sub-sectors: (a) C&T; (b) processing and (c) disposal. For each sub-sector, project costs are divided by the corresponding project incremental capacity (expressed in volume of waste) to calculate unit costs. For collection and transportation, the volume of waste is assumed to be equal to the 79 design-year solid waste generation for the project area. The assumption is in line with the specification in the Project Appraisal Notes that the project would include the replacement of the existing equipment, which has reached the end of its economic life. For composting, the incremental waste volume is estimated based on the design-year size of the compost plant, which is reported by most Project Appraisal Notes.41 For landfill, the incremental waste volume is available for only a few projects. For the available projects, on average the landfill share is equivalent to 32 percent of the design-year solid waste generation. The share is applied to estimate volume of waste that is disposed for projects for which data is not available. For cities with more than one project, the city-level unit costs are calculated by averaging out the unit costs for all projects within that city. Per Capita Investment Costs (PCIC) are computed based on unit costs and per capita waste generation. Step VII: Estimate Total Investment Requirements for the period 2007-2031 Investment requirements include the following cost components:  Investments for 2006 backlog (un-met demand)  Investments for demand growth (2007-2031)  Assets replacement costs Investments for backlog: The un-met demand for C&T, processing and disposal is multiplied by the unit costs for the respective cost components. The three cost components are then added to compute the total backlog investment costs. Investments for demand growth: To calculate the investment requirements for demand growth, the unit costs for collection and transportation, processing and disposal are multiplied by the incremental waste volume that would need to be collected & transported, processed and disposed over the period 2007-31. The incremental waste volume is estimated based on population forecasts and increase in per capita waste generation. Assets replacement costs: Assets replacement costs for collection and transportation are estimated based on an asset life of 10 years. This implies that, for example, the stock of vehicles purchased in 2001 would need to be replaced in 2011. Assets replacement costs for disposal are estimated based on a five year interval at a unit cost equal to 50 percent of the initial cost. This is consistent with the fact that landfill costs provided in the JNNURM Project Appraisal Notes are generally for a 5 year period. However, the incremental costs of landfill are going to be lower than the original cost (50 percent of the initial costs) because some initial investments (e.g. roads) will not be needed in the future. Assets replacement costs for processing are not included, as they are generally funded by compost sale revenues. 41 On average, the size of the compost plant is about 48 percent of the design-year waste generation capacity for the project area. For the three projects that do not provide information on the size of the compost plant, the incremental processing capacity of the project is estimated based on the 48 percent share. 80 While land costs cannot be estimated, land requirements for landfill sites are computed for the year 2031 based on a norm of 150 acres per 150 Tpd. All project costs and investment requirements are converted in 2009 prices. Step VIII: Estimate Annual O&M Costs for the period 2007-2031 Annual O&M costs for both C&T and disposal are estimated based on total waste volume. O&M costs for processing are not considered because they are covered by compost sale revenues. The average unit O&M cost is estimated at Rs 1,200 per ton of waste generation (Rs 1,000 per ton for C&T and Rs 200 for disposal) based on a sample of project data and expert estimates. Step VIII: Conducting Sensitivity Analysis Investment requirements are calculated as a band rather than a single point estimate. The band reflects the reliability of the investment estimates, which in turn depends on the accuracy of the PCICs, backlog percentage and population forecasts. Among these three factors, the PCIC is the most critical variable affecting investment estimates – a 10 percent increase in PCIC leads to an equivalent increase in investment requirements – while variation in the other two variables leads to a less than proportional change in investment requirements. The standard deviation is calculated to measure the variation of the PCICs for each of the three sub-sectors.42 Based on the standard deviation, a 10 percent confidence interval for the total PCIC is estimated. Given that a change in the PCIC leads to a proportional change in total investment requirements, the low and high estimate boundaries of the confidence interval for the PCICs can be converted into a range of values for the investment requirements. Step XIII: Simulate the profile of the investments In order to simulate the profile of the investments, investment requirements are estimated in line with the 5-year Planning Periods. For example, the 2006 base year corresponds to the period April 2006-March 2007, and the first five year interval (2007-2011) correspond to the 11th Plan (April 2007 - March 2012). For the purpose of the simulation, it is assumed that full coverage would be achieved by the end of the 15th Plan (2027-2032). 42 The standard deviation is a good proxy of the accuracy of a variable. A low standard deviation indicates that the data points tend to be very close to the average value of the distribution, whereas high standard deviation indicates that the data are spread out over a large range of values. The standard deviation also depends on the sample size: everything else being equal, the larger the sample size, the more accurate the estimates, and the lower the standard deviation. 81 MODEL ASSUMPTIONS General  Cities are classified into six categories on the basis of their population as per the census data. The first Census Category (Class I), which include all cities with more than 100,000 population, is split into the following three sub-classes: Class I.A: 7 Megacities; Class I.B: 1-5 million and Class I.C: 100,000-1 million.  For the classification of cities, where applicable, the population of ‘urban agglomeration’ is used instead of the population within the area of municipal corporation/ municipality. This is because the City Development Plans (CDPs) have been prepared for the urban agglomeration (and not just the area under the municipal corporation/ municipality). For cities where the population figures for urban agglomeration are either not available or not applicable, population figures for the municipal corporation/ municipality are used. This is because the backlog percentages for these cities were specifically for the area under municipality and not the urban agglomeration.  The base year for the estimation of investment needs is 2006. The model uses year 2031 as the forecast year.  The backlog percentages that are estimated based on the City Development Plans are generally for the year 2006. The backlog percentages that are estimated based on 2001 population census data are assumed to apply to the base year 2006 (for example, for the estimation of the water distribution backlog, it is assumed that the share of urban population with private water connection in 2001 is the same as the share in 2006). The sewerage backlog data collected directly from the cities in Karnataka, Tamil Nadu and Andhra Pradesh may not necessarily be as of 2006; when this is the case, it is assumed that the same backlog percentage applies to the year 2006.  When the CDPs express the backlog for coverage in terms of area, it is assumed that the population is evenly distributed within the area; hence the same percentage is used to determine the population backlog.  Investment needs are converted in 2009 prices, given that most JNNURM approved projects costs are in 2006 prices. The rate of inflation used for conversion of 2005-06 prices into 2008-2009 prices is 30 percent. This is derived using the whole sale price index (non- food items) sourced from Reserve Bank of India. For the sake of simplicity, it is assumed that all project costs are in 2006 prices when the date of project approval is not available.  The exchange rate that is used in the model was: USD 1 = INR 45. 82 Water  The average per capita production norm is assumed to be 168 liters/per capita/day across all city size classes. The average per capita consumption standard is 135 lpcd.  For cities with population above 500,000, industrial water production is estimated to account for about 20 percent of total water production, based on City Development Plans information for a sample of cities.  For cities with population above 500,000, industrial water production is estimated to increase at 7 percent per annum, in line the average industry value added for the last ten years.  The assumption is made that on average 50 percent of the distribution network would need to be replaced to deliver 24/7 water supply in Indian cities. The replacement share ranges from 40 percent for Class I.A cities to 60 percent to Class III and IV+.  . It is assumed that the average per capita distribution pipes’ extension that is needed to connect the un-served population is equal to 50 percent of the per capital length of the distribution pipe for the served population  Unit costs for O&M are estimated based on a number of projects provided by sector experts. The unit costs for each city size class are reported in Table 4. Table 4: Unit Water O&M Costs (Rs/m3) Class Unit Cost (Rs/m3) Class I.A 12 Class I.B 9 Class I.C 7 Class II 7 Class III 5 Class IV+ 3  Unit costs of pipes are based on project data provided by sector experts. The costs are reported in the Table 5 below. Table 5: Unit Cost of Pipes (Rs/m) Diameter Cost 100 1,200 150-200 1.857 250-300 2,514 350-400 3,965 450-550 5,524 600 8,650 650 9,739 700 and above 10,827 83  Based on project data provided by sector experts, unit costs of storage are estimated at 5 Rs/l for storage capacity above 500 MLD, 8 Rs/l for storage capacity of 150-500 MLD, and 12 Rs/l for storage capacity below 150 MLD. Per capita storage requirements are estimated at one third of per capita water consumption.  For the estimation of replacement costs, the economic life of water production assets is assumed to be 30 years. Sewerage  For the estimation of replacement costs, the economic life of network and treatment assets is assumed to be 30 years.  The unit O&M costs are assumed to be equal to 1.6 Rs/m3 for treatment and 1 Rs/m3 for network based on a sample of projects provided by sector experts. Solid Waste  Per capita waste generation is estimated based on 1995 data as reported in the 3i Network 2006 Infrastructure Report, assuming an annual growth rate of 1.3 percent. Table 6: Per Capita Waste Generation (2006) Class Gr/capita/day Class I.A 577 Class I.B 404 Class I.C 288 Class II 242 Class III 242 Class IV+ 242 Source: Estimates based on 3i Network. 2006. ‚India Infrastructure Report‛. New Delhi.  The economic life of the assets assumed for the estimation of assets replacement costs is summarized below Table 7: Per Capita Replacement Costs C&T Every 10 years Processing No re investment Disposal Every 5 years at 50% of original unit cost  The target compost share is assumed to be gradually declining from the prevailing 65 percent to about 50 percent over a 10 year period. In parallel, the target landfill share is expected to gradually increase from 30 to 45 percent. 84 Annex II: Sample Description Table 8: Project Sample Size, by Sector I.A I.B I.C II III IV+ TOTAL Megacities 1-5m 100k-1m 50-100k 20-50k <20k Water Production 9 16 13 2 1 0 41 24/7 Up-gradation 6 8 8 1 10 0 33 Distribution extension 5 7 8 1 10 0 31 Sewerage Network 11 21 33 13 30 5 113 Treatment 15 22 23 10 4 5 79 Solid Waste C&T 3 15 15 2 5 0 40 Processing 3 15 8 0 1 0 27 Disposal 3 15 10 1 4 0 33 Table 9: Backlog Sample Size (CDP’s and Census), by Sector I.A I.B I.C II III IV+ TOTAL Megacities 1-5m 100k-1m 50-100k 20-50k <20k Water Production (CDPs) 6 23 21 2 10 5 67 Distribution extension 7 27 360 404 1,164 2,415 4,377 (Census) Sewerage Network (Census) 7 27 360 404 1164 2415 4,377 Treatment (CDPs) 6 18 29 11 12 4 80 Solid Waste C&T (CDPs) 7 18 16 0 7 3 51 Processing (CDPs) 7 19 14 0 6 3 49 Disposal (CDPs) 6 23 21 2 10 5 67 85 Data Sources Data Sources for Projects SEWERAGE Network Treatment JNNURM Project Appraisal Notes 41 46 Tamil Nadu Urban Development Fund (TNUDF) 24 23 Karnataka Urban Water Supply and Drainage Board 43 5 Andhra Pradesh Municipal Development Project, 5 5 Municipal Strengthening Unit WATER Production JNNURM Project Appraisal Notes 35 Karnataka Urban Water Supply and Drainage Board 3 Projects provided by sector expert: Ponnur, Kadiri, and 3 Eluru. WATER Distribution Cost Simulation (based on 22 CDPs, Karnataka Urban Infrastructure Development and Finance Corporation 31 ‚KUIDFC‛ data for 9 Karnataka cities, and expert advice) WATER Up-gradation Cost Simulation (based on 26 CDPs and expert advice) 26 JNNURM 24/7 water pilot projects (Detailed Project 3 Reports: Kolkata, Allahabad, and Nagpur) Karnataka 24/7 water pilot projects: Hubli, Gulbarga, 3 Belgaum. SOLID WASTE C&T Processing Disposal JNNURM Project Appraisal Notes 22 22 22 Multiple Sources (including: Karnataka Urban Development and Coastal Environmental Management 18 5 11 Project (KUDCEMP); Infrastructure Development Consultancy Karnataka). 86 Data Sources for Backlog Source Sector 46 CDPs for JNNURM mission cities Sewerage Treatment; Water Production; 9 CDPs for small and medium towns in Karnataka Solid Waste C&T, Processing and Disposal Census data Water Distribution; and Sewerage Network Data sources for O&M Source Sector Estimates provided by sector experts based on data from Water Hyderabad city and from multi village schemes 15 projects in Karnataka (Vrishabhavathi; Koramangala Sewerage and Challaghatta; and Hebbal) 87 Annex III: India Urban Population Forecasts 110. Indian urban population is expected to double in size from 2001 to 2031. Based on UN estimates, the population of Indian cities is expected to reach 627 million by 2031, equivalent to 40 percent of the Indian population.43 Over the same period, the population of Indian megacities (with population above 5 million) is estimated to double, from 61 million in 2001 to 133 million in 2031. The second largest category of Indian cities (with population between 1 and 5 million) is expected to record the highest absolute increase in urban population, from 46 to 126 million over the 30-year period. As a result, the share of Indian urban population residing in cities with 1-5 million population is expected to increase from 15 to 20 percent over the period 2001-2031 (see Figure 1 and 2).44 111. The annual population growth rate for urban India is expected to stabilize at about 2.5 percent per annum over the period 2001-31. The forecasted growth rate is in line with the population growth recorded over the period 1995-2000, although below the record growth of 3- 4 percent registered in the previous decades. Cities with population between 1 and 5 million are expected to grow at a significant higher growth rate than the national average, of about 3.4 percent per annum. The growth rate of cities below 1 million, currently below national average, is forecasted to steadily increase to reach 2.6 percent by 2020. Megacities are expected to grow in line with the national average, although their growth rate will experience a decline from the current level of 4.0 percent to 1.9 percent in 2031. Unfortunately, the UN data available does not allow distinguishing the sources of population growth – i.e. re-classification (i.e. cities switching to a higher size class), natural population growth and migration. See Figure 59 and Table 10. 43 The 2001 urban population of India is estimated at 196 million, based on UN estimates. 44 Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat, World Population Prospects: The 2006 Revision and World Urbanization Prospects: The 2007 Revision, http://esa.un.org/unup. 88 Figure 59: India Urban Population Growth Rates, 2001-2031 5% 4.5% 4% 4.0% 3.0% 3% 2.7% 2.6% 2.3% 2.6% 2% 1.9% 1.4% 1% 1.0% 0% 00-05 05-10 10-15 15-20 20-25 25-31 All classes I.A : > 5m I.B: 1-5m I.C: 1m-100k II-IV+: < 100k Source: World Urbanization Prospects 2007 and Census of India. Table 10: Average Annual Growth Estimates, 2001-2031 Class/ Year 2001-05 2005-10 2010-15 2015-20 2020-25 2025-31 2001-31 Class I.A >5m 4.0 3.3 1.9 2.9 1.9 1.9 2.6 Class I.B 1-5 m 4.5 3.5 3.8 2.7 3.0 3.0 3.4 Class I.C 1 m-100,000 1.4 1.7 2.3 2.4 2.8 2.7 2.3 Class II 50-100,000 Class III 20-50,000 1.0 1.8 2.4 2.4 2.6 2.6 2.2 Class IV+ < 20,000 All Classes 2.3 2.4 2.5 2.6 2.6 2.6 2.5 Source: World Urbanization Prospects 2007 and authors’ calculations. Notes: See Table 3 for assumptions. Methodology 112. Population forecasts are based on estimates provided by the United Nations Population Division of the Department of Economic and Social Affairs (DESA) in its regular 89 statistical publication, the World Urbanization Prospects (2007 revision).45 The World Urbanization Prospects is a database reporting past, current and future urban population for each country in the world and their major agglomerations. The database is revised and updated every two years. The latest revision was published in 2007. Being the most comprehensive database on urbanization currently available, the UN data is largely used and referred to for urban population trends and projections. The UN relies on data produced by national statistical offices, and adopts national definition of urban areas. Historical urban population trends are based on and fully consistent with Census of India statistics, and adopt the concept of urban agglomeration. The UN urban population projections are based on the basic assumption that urbanization slows down with growing urbanization. A projection model is built based on the intrapolation and extrapolation of urban-rural growth differentials. 113. Population forecasts for urban India are based on 2001 population census figures and UN growth rate estimates by city class. For each city class, population is forecasted by applying the UN population growth rates to 2001 census population figures over the period 2001-2031. Unfortunately, there is no complete alignment between the Census of India city classes, as reported in this study, and the UN population classes (see Table 2 below). More specifically, the UN projection model provides estimates for only 5 broad city classes – the lowest class including all cities with population below 500,000. The Census of India classification, as adopted in this study, is more fine-grained, with the lowest class including all towns with population below 20,000. As a result, the same growth rate is applied to all Indian cities with population below 500,000 given that forecasts for individual classes are not available. 45 World Urbanization Prospects. The 2007 Revision. United Nations Department of Economic and Social Affairs (DESA) Population Division - Population Estimates and Projections Section. http://esa.un.org/unup/ 90 Table 11: Census of India versus UN Classes Census Classes UN classes Class UN.1 10 -5m Class I.A >5m Class UN.2 >5m Class I.B 1-5 m Class UN.3 1-5 m Class UN.4 1 m - 500,000 Class I.C 1 m - 100,000 Class II 50 - 100,000 Class UN.5 < 500,000 Class III 20 - 50,000 Class IV+ < 20,000 114. The forecasts are based on the following assumptions:  2001 census population for the two largest city classes (Class I.A and Class I.B) are taken from the World Urbanization Prospects database, given that there is a perfect match between these two census classes and the first three UN classes.46 The population figures reported in the UN database are sourced from the Census of India 2001. For the other city classes that do not match the UN classes, 2001 population figures are sourced directly from the Census of India website.47  An exponential growth rate is assumed to forecast urban population, in line with the methodology applied by the UN.  For the period 2025-30, UN provides projections only for entire urban India, with no breakdown by city class. Given that the urban India population growth rate for the period 2025-30 is estimated to be the same as the growth rate for the period 2020-25, growth rates for individual classes are also assumed to be equal to the growth rates of the 2025-30 period.  UN estimates are only available up to the year 2030. Population figures for 2031 are projected assuming the same annual growth applied to the period 2025-2030.  Urban population forecasts for the period 2001-2031 by city class are reported in Table 12 below. 46 http://esa.un.org/unup/ 47 http://www.censusindia.gov.in/Census_And_You/area_and_population.aspx 91 Table 12: Urban Population, 2001-2031 (million) Class 2001 2005 2006 2009 2010 2015 2020 2025 2030 2031 I.A 61 72 75 82 85 93 108 119 130 133 I.B 46 55 57 63 66 79 91 106 123 126 I.C 99 105 107 112 114 128 144 166 190 195 II 28 29 29 31 32 36 40 46 52 54 III 35 37 37 39 40 45 51 58 66 68 IV+ 27 28 28 30 30 34 39 44 50 52 TOTAL 296 326 333 358 367 416 473 538 611 627 Notes: (1) Separate growth estimates for Class II – IV+ are not available. (2) UN estimates are only available up to the year 2030. Population figures for 2031 are projected assuming the same annual growth applied to the period 2025-2030. (3) For the period 2025-30, UN provides projections only for urban India, with no breakdown by city class. Given that the national urban population growth rate for 2025-30 is estimated to be the same as the growth rate for 2020-25, growth rates for individual city classes for 2025-30 are also assumed to be equal to the growth rates for 2020-25. 92 Annex IV: Main Results Investment Requirements Table 13: Total Urban Investment Requirements, by Sector 2006-2031 (2009 prices), Rs Bn Rs Bn WATER SEWERAGE SOLID WASTE Residential Backlog 1,052 873 92 Additional requirements 1,248 1,056 189 Assets replacement 288 300 207 Total residential 2,587 2,229 487 Industrial Additional requirements 408 - - Total industrial 408 - - TOTAL 2,995 2,229 487 Table 14: Urban Water Investment Requirements, by Sub-sector 2006-2031 (2009 prices), Rs Bn 24/7 Distribution Production Total Rs Bn Up-gradation extension Residential Backlog 160 369 522 1,052 Additional requirements 428 - 820 1,248 Re-placement 288 - - 288 Total residential 875 369 1,342 2,587 Industrial Additional requirements 408 - - 408 Total industrial 408 - - 408 TOTAL 1,284 369 1,342 2,995 93 Table 15: Urban Sewerage Investment Requirements, by Component 2006-2031 (2009 prices), Rs Bn Network Treatment Total Rs Bn Residential Backlog 559 314 873 Additional requirements 700 356 1,056 Re-placement 214 85 300 TOTAL 1,473 756 2,229 Table 16: Municipal Solid Waste Investment Requirements, by Component 2006-2031 (2009 prices), Rs Bn Collection & Processing Disposal Total Rs Bn Transportation Residential Backlog 14 51 26 92 Additional requirements 58 54 76 189 Re-placement 109 - 98 207 TOTAL 182 106 200 487 94 Backlog Table 17: Backlog Percentage, by Sector I.A I.B I.C II III IV+ Urban Megacities 1-5m 100k-1m 50-100k 20-50k <20k Population Percent of total population Water Production 46% 29% 18% 29% 56% 62% 35% 24/7 Up-gradation 57% 55% 48% 41% 37% 31% 48% Dist. extension 43% 45% 52% 59% 63% 69% 52% Percent of total population Sewerage Network 54% 54% 72% 80% 85% 86% 67% Treatment 47% 54% 77% 88% 96% 100% 72% Percent of total waste Solid Waste C&T 13% 48% 41% 41% 65% 75% 41% Processing 88% 94% 93% 93% 100% 100% 93% Disposal 100% 100% 100% 100% 100% 100% 100% 95 Per Capita / Unit Costs Table 18: Per Capita and Unit Costs, Water Investment O&M Per Capita Costs Unit Costs Per capita annaul (Rs/capita) (Rs/m3) Costs (Rs/capita) Production 1,448 7,558 - 24/7 Up-gradation 2,513 - - Distribution Extension 2,828 - - TOTAL 4,276 - 501 Table 19: Per Capita and Unit Costs, Sewerage Investment O&M Per Capita Costs Unit Costs Per capita annual (Rs/capita) (Rs/m3) Costs (Rs/capita) Network 2,373 6,526 - Treatment 1,240 3,368 - TOTAL 3,613 9,894 102 Table 20: Per Capita and Unit Costs, Solid Waste Investment O&M Per Capita Costs Unit Costs Per capita annual (Rs/capita) (Rs/TPD) Costs (Rs/capita) C&T 137 319,458 175 Processing 131 578,559 - Disposal 152 835,826 15 TOTAL 421 1,733,844 190 Note: per capita costs are the average for the period 2007-2031. 96 Investment Profile Table 21: Profile of Total Investment Requirements, by Sector April 2007- April 2012- April 2017- April 2022- April 2027- March 2012 March 2017 March 2022 March 2027 March 2032 TOTAL 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan Rs Bn 2007-2011 2012-2016 2017-2021 2022-2026 2027-2031 2007-2031 Water 525 520 556 676 718 2,995 Sewerage 374 397 433 491 534 2,229 Solid waste 62 61 104 108 152 487 Total 961 978 1,094 1,276 1,404 5,711 Table 22: Investment Requirements (Share of GDP), by Sector April 2007- April 2012- April 2017- April 2022- April 2027- March 2012 March 2017 March 2022 March 2027 March 2032 TOTAL 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan Rs Bn 2007-2011 2012-2016 2017-2021 2022-2026 2027-2031 2007-2031 Water 0.9% 0.6% 0.5% 0.4% 0.3% 2.7% Sewerage 0.6% 0.5% 0.4% 0.3% 0.2% 2.0% Solid waste 0.1% 0.1% 0.1% 0.1% 0.1% 0.4% Total 1.6% 1.2% 1.0% 0.8% 0.6% 5.2% Table 23: Profile of Household Coverage, by Sector (As of the end of the period) April 2007- April 2012- April 2017- April 2022- April 2027- March 2012 March 2017 March 2022 March 2027 March 2032 11th Plan 12th Plan 13th Plan 14th Plan 15th Plan 2007-2011 2012-2016 2017-2021 2022-2026 2027-2031 Water Production 75% 84% 90% 96% 100% Sewerage Network 53% 69% 82% 92% 100% Treatment 49% 66% 80% 91% 100% Solid waste C&T 74% 85% 92% 97% 100% Disposal 46% 72% 85% 94% 100% 97 24/7 Pilot Projects Table 24: 24/7 Water Pilot Projects Costs PCIC City Class Beneficiaries Investment Source (Rs M) (Rs/capita) (Partial) Distr. Kolkata I.A 252 218,000 1,157 JNNURM replacement (Partial) Distr. Allahabad I.B 1,623 615,360 2,638 JNNURM Replacement I.B 3,470 2,350,000 1,477 Dist replacement JNNURM Nagpur (40%) Distr. Replacement Hubli I.C 123 65,765 1,865 KUIDFC (90%) (Partial) Distr. Gulbarga I.C 40 17,500 2,291 KUIDFC Replacement (Partial) Distr. Belgaum I.C 98 41,650 2,353 KUIDFC Replacement 98 Annex V: Professionalization for Sustainable WSS Development 115. Inadequate human capacity to manage the urban WSS sector carries a very high opportunity cost. Low professionalization48 in the WSS sector leads to inadequate maintenance of the assets, which as a result fail to deliver their full economic life. Low professionalization also translates into sub-optimal WSS system operation, with direct impact on service to customers. The opportunity to introduce cost-recovery pricing schemes is limited when quality of service does not meet consumers’ expectations, and willingness-to-pay is low. This leads to a vicious circle which drastically reduces the prospects for the sustainable development of the WSS sector: the poor financial performance of the water utilities becomes a major constraint to sector professionalization and capacity building, which is in turn a necessary condition for improving the financial prospects of the sector. The monetary value of inadequate capacity is particularly high in countries with very large investment requirements, like India, as low professionalization negatively impacts the rate of return of new investments, by escalating asset replacement costs. 116. The economic benefits of scaling up professionalization in the urban WSS sector are estimated at about Rs 438 billion (USD 10 billion) over the next forty years, in 2009 prices (see Box 5).49 The estimation is based on the assumption that the low professionalization of the WSS sector results into a 10-year reduction in the optimal economic life of the assets. This implies that the all new assets will have a reduced economic life, unless steps are taken to improve sector capacity. By improving the professionalization of the WSS sector, investing in capacity building would extend the life of the assets to their full economic value and defer assets replacement costs. A real discount rate of 5 percent is used to annualize the monetary gains of scaling up capacity building. The economic benefits are lower, but still substantial when a higher real discount rate is applied - the estimated benefits are about Rs 212 (USD 5 billion) based on a 10 percent discount rate. The results indicate that the economic gains of improving WSS sector capacity are expected to be significantly above the costs, and sizeable net economic benefits can therefore be reaped by investing in capacity building. 117. What steps are required to improve the professionalization in the WSS sector? The promotion of continuing professional development and management training for utility staff, the introduction of professional accreditation, the design of university and vocational course, and the establishment of active professional associations are among the important steps that need to be taken to enhance the professionalization of WSS utilities in India. In parallel, there is a need to improve the capacity of political and administrative decision makers who, by their actions, affect the ability of utility professionals to do their jobs. 48 Deemed to include all WSS professionals and owners/oversight agencies involved in the process of delivering, or facilitating the delivery of WSS services. 49 Note the estimated benefits may slightly change as the PCIC are finalized. 99 Box 5: Methodology for estimating the economic benefits of professionalization in the WSS sector The economic benefits of investing in the professionalization of the WSS sector are estimated based on the cost models for urban water and sewerage. The two scenarios of optimal and low professionalization are compared. Under the scenario of optimal professionalization (Scenario I), the assets have a full economic life of 30 years. Under the scenario of low professionalization (Scenario II), the life of the assets is shortened to 20 years. Assets replacement costs are estimated under the two scenarios over the period 2006-2050 for the water sector (production and distribution) and the sewerage sector (network and treatment). A 5 percent real discount rate is used to annualize the assets replacement costs. The difference in the Net Present Value between the two scenarios is equivalent to the economic benefits of investing in capacity building to reach optimal professionalization in the WSS sector. A sensitivity analysis is carried out based on alternative discount rates. The estimation is sensitive to the profile of the investments. The assumption is made that investments will ramp up in the WSS sector so that full coverage is achieved by the end of the 15th Plan. The water cost model is built under the assumption that a significant part of the existing distribution network has already reached the end of its economic life, and need to be replaced to deliver 24/7 water supply continuity. The assumption is made that the required replacement of the existing distribution network will be completed by 2021. Hence, for the estimation of the economic benefits of professionalization, assets replacement costs for water distribution are calculated starting from the year 2021. Under the scenario of low professionalization, the newly built distribution assets would need to be replaced in 2041 (assuming a 20 year economic life of the assets). Under the scenario of optimal professionalization, the assets would need to be replaced in 2051 (assuming a 30 year economic life of the assets). 100 Annex VI: Cross-country Comparison of Service Performance Table 25: Urban Piped Water Coverage (2006 and 2004-08) URBAN URBAN (SELECTED UTILITIES) [3] Private Connections [1] Private and Shared Connections [2] 2006 2004-08 Obs.[4] High income[5] 96 96 335 Upper middle income 90 85 1,790 Mexico 96 84 38 Brazil 88 80 603 South Africa 84 86 45 Lower middle income 70 72 773 China 87 91 167 Philippines 69 46 46 India 49 - - Pakistan 48 - - Indonesia 34 - - Nigeria 7 34 3 Low income 43 54 426 Vietnam 59 60 225 Bangladesh 20 48 38 Notes: [1] Coverage is defined as piped water coverage into dwelling, yard or plot. [2] Coverage is defined as the share of population with private or shared connections as a percentage of total population under the responsibility of the water utility. [3] Country level coverage rates are calculated as the average of selected utility data for the period 2004-2008. The coverage rates are for urban areas, as the utilities for which data is available are primarily urban. [4] The number of observations is equal to the number of utilities times the number of years for which data is available. [5] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Sources: [1] WHO/UNICEF (http://www.childinfo.org/water_data.php) [2] IBNET (http://www.ib-net.org/) 101 Table 26: Urban Continuity of Water Supply (2004-08) Duration of Water Supply (Hrs/day)[1] COUNTRY LEVEL URBAN (SELECTED UTILITIES) [2] 2004-08 Observations[3] High income[5] 24 781 Upper middle income 23 3,380 Brazil 24 1,993 South Africa 24 45 Mexico 21 38 Lower middle income 16 834 China 24 167 Philippines 21 46 Indonesia 20 7 Pakistan 10 12 Nigeria 9 12 India 4 10 Low income 19 576 Vietnam 22 272 Bangladesh 11 38 CITY LEVEL[4] 1995-96 Shanghai (China) 24 Jakarta (Indonesia) 18 Manila (Philippines) 17 Lagos (Nigeria) 6 Delhi (India) 4 Karachi (Pakistan) 3 Notes: [1] The duration of water supply is defined as the average hours of service per day. [2] Country level duration of water supply is calculated as the average of utility data for the period 2004-2008. The indicator is for urban areas, as the utilities for which data is available are primarily urban. [3] The number of observations is equal to the number of utilities times the number of years for which utility data is available. [4] The indicator is calculated for a city's main utility in 1995-1996. [5] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Source: [1] IBNET (http://www.ib net.org/) [4] Second Water Utilities Data Book – Asian and Pacific Region (http://www.adb.org/Documents/Books/Second_Water_Utilities/default.asp) and Africa Infrastructure Country Diagnostic (AICD) WSS Survey (http://www.infrastructureafrica.org/) 102 Table 27: Urban Per Capita Water Consumption (2001 and 2004-06) – Private/Shared Connections and Public Taps Per capita Water Consumption (lpcd) [1] COUNTRY LEVEL URBAN (SELECTED UTILITIES) [2] 2004-2006 Observations[3] High income[5] 179 142 Upper middle income 162 494 South Africa 202 10 Mexico 171 18 Brazil - - Lower middle income 103 213 India 132 8 Philippines 122 40 Indonesia 114 6 China 72 88 Low income 97 201 Vietnam 97 157 Bangladesh 76 17 Per capita water consumption (lpcd) [1] CITY LEVEL [4] 2001 Shanghai (China) 251 Karachi (Pakistan) 197 Manila (Philippines) 127 Delhi (India) 110 Jakarta (Indonesia) 77 Notes: [1] Per capita water consumption is defined as the volume of total annual residential water consumption (sold through private and shared connections and public taps) as a fraction of the population served. [2] Country level per capita water consumption is calculated as the average of utility data for 2004-2006. The per capita water consumption is for urban areas, as the utilities for which data is available are primarily urban. Controlled for water pricing by including in the sample only utilities with operating cost coverage ratio (revenues/ operating costs) > 1 [3] The number of observations is equal to the number of utilities times the number of years for which utility data is available. [4] City level per capita water consumption is the value for a city's main utility in 2001. Water pricing is not controlled for at the city level. [5] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Sources: [2] IBNET (http://www.ib-net.org/) and 2007 Benchmarking and Data Book of Water Utilities in India (http://www.adb.org/documents/reports/Benchmarking-DataBook/default.asp [4] Water in Asian Cities - Utilities Performance and Civil Society Views (http://www.adb.org/Documents/Books/Water_for_All_Series/Water_Asian_Cities/default.asp) 103 Table 28: Urban Per Capita Water Consumption (2004-06) - Private and Shared Connections Per Capita Water Consumption (lpcd) [1] COUNTRY-LEVEL URBAN (SELECTED UTILITIES) [2] 2004-2006 Observations[3] High income[4] 181 139 Upper middle income 126 292 Colombia 105 63 Poland 102 89 Lower middle income 108 192 Philippines 122 40 China 70 88 Low income 96 150 Vietnam 97 114 Bangladesh 84 17 Notes: [1] Per capita water consumption is defined as the volume of total annual residential water consumption (sold through private and shared connections) as a fraction of the population served. Controlled for water pricing by including in the sample only utilities with operating cost coverage ratio (revenues/operating costs) >1. [2] Country level per capita water consumption is calculated as the average of utility data for the period 2004-2006. The per capita water consumption is for urban areas, as the utilities for which data is available are primarily urban. [3] The number of observations is equal to the number of utilities times the number of years for which data is available. [4] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Sources: [1] IBNET (http://www.ib-net.org/) 104 Table 29: Urban Sewerage Network Coverage – Selected Countries (2000-07) Population connected to wastewater collecting system (%) 2000-07 Countries High Income 77 38 Upper Middle Income 55 23 Mexico (2005) 68 South Africa (2007) 60 Brazil (2006) 48 Lower Middle Income 56 15 Jordan (2004) 98 Morocco (2007) 87 Armenia (2006) 83 China (2004) 46 Iraq (2005) 26 Paraguay (2007) 15 Low Income 19 5 Notes: [1] Urban sewerage network coverage is defined as the percentage of the resident population connected to the wastewater collecting systems (sewerage). [2] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Source: UNSTATS (http://unstats.un.org/unsd/environment/Time%20series.htm#Waste) Table 30: Urban Sewerage Treatment Coverage – Selected Countries (2000-07) Population connected to urban wastewater treatment (%) [1] 2000-07 Countries High Income 70 37 Upper Middle Income 43 20 South Africa (2007) 57 Mexico (2005) 35 Brazil (2006) 26 Lower Middle Income 36 13 Morocco (2007) 80 Jordan (2004) 52 Armenia (2006) 34 China (2004) 33 Iraq (2005) 26 Low Income 5 4 Notes: [1] Urban sewerage treatment coverage is defined as the percentage of the resident population whose wastewater is treated at wastewater treatment plants. [2] Economies are divided according to 2008 GNI per capita, calculated using the World Bank Atlas method. The groups are: low income, $975 or less; lower middle income, $976 - $3,855; upper middle income, $3,856 - $11,905; and high income, $11,906 or more. Source: UNSTATS (http://unstats.un.org/unsd/environment/Time%20series.htm#Waste) 105 Figure 60: Urban per Capita Water Consumption: Selected Countries (2004-06) Private and Shared Connections 200 181 160 126 122 105 102 108 96 97 120 84 70 80 Lpcd 40 0 High Upper Colombia Poland Lower Philippines China Low Vietnam Bangladesh income middle middle income income income (139) (292) (63) (89) (192) (40) (88) (150) (114) (17) Upper middle income Lower middle income Low income Source: IBnet. Note: Number of utilities in parenthesis. Figure 61: Urban Water Unit Operational Costs, Selected Countries (2004-08) 1.50 0.94 0.94 1.10 0.42 0.52 0.70 0.30 0.29 0.21 0.21 0.17 0.16 0.18 0.14 0.06 0.30 -0.10 US$/m3 income Low income Mexico Nigeria Bangladesh Vietnam China South Africa Indonesia India (equals Rs 13) Upper middle Lower middle Philippines High income income (247) (1,010) (25) (35) (712) (5) (165) (45) (2) (7) (382) (239) (38) High Upper middle income Lower middle income Low income Source: IBnet. Notes: 2008 Prices. Number of utilities in parenthesis. Figure 62: Urban Solid Waste Generation Rate 1995 (Based on a sample of cities) 6 Waste Generation Rate GNP per Capita 40,000 5 30,000 4 $ GNP per Capita Kg/cap/day 3 20,000 2 10,000 1 0 0 Source: World Bank (1999). 106 Figure 63: Municipal Waste Collected vs. GNI per capita (2000-07) 4 3 kg/capita served/day 2 1 India 0 0 25,000 50,000 75,000 GNI per capita ($) Source: UNSTATS; CDPs for India. Figure 64: Urban Waste Generation in India 1995 0.7 14 0.6 12 0.5 10 Waste Generation Rate Urban Population (Millions) 0.4 8 (Kg/cap/day) 0.3 6 0.2 4 0.1 2 0 0 Indore Pune Kochi Nagpur Kanpur Madras Ludhiana Bangalore Vadodara Jaipur Delhi Bhopal Visakhapatnam Lucknow Patna Calcutta Hyderabad Madurai Varanasi Surat Coimbatore Bombay Ahmedabad Source: World Bank (1999). 107 References Africa Infrastructure Country Diagnostic (AICD) WSS Utility Database (http:// www.infrastructureafrica.org/) Asian Development Bank (2004). ‚Water in Asian Cities - Utilities Performance and Civil Society Views‛ Water for All Series 10. (http://www.adb.org/documents/ books/water_for_all_series/Water_Asian_Cities/default.asp) Asian Development Bank (1997). ‚Second Water Utilities Data Book – Asian and Pacific Region‛ (http://www.adb.org/ Documents/Books/Second_Water_Utilities/default.asp) Census of India (2001). Chatterton, Isabel and Olga S. Puerto (2006). ‚Estimation of Infrastructure Investment Needs in the South Asia Region‛ World Bank. Choe, Varley and Bijlani (1995). ‚Coping with Intermittent Water Supply: Problems and Prospects‛. Activity Report N. 26. Committee of Ministers constituted by the Central Council of Local Self Government (1963). ‚Augmentation of Financial Resources of Urban Local Bodies‛. Expert Group on the Commercialization of Infrastructure Projects (1996). ‚The India Infrastructure Report – Policy Imperatives for Growth and Welfare‛. Fay, Marianne and Guillermo Yepes (2003). ‚Investing in Infrastructure: What is Needed from 2000 to 2010?‛, World Bank Policy Research Working Paper No. 3102 Hoornweg, Daniel and Laura Thomas (1999). ‚What a Waste: Solid Waste Management in Asia‛ Working Paper Series No. 1. Urban Development Sector Unit, East Asia and Pacific Region. The World Bank, Washington, D.C. Pp. 43. IbNET Database (http://www.ib-net.org/) 3i Network (2006). ‚India Infrastructure Report‛. New Delhi. Mathur, Mukesh, Rajesh Chandra, Satpal Singh and Basudha Chattopadhyaya (2007), ‚Norms and Standards of Municipal Basic Services in India‛. National Institute of Urban Affairs Working Paper 07-01.Ministry of Urban Development (2008). ‚Handbook of Service Level Benchmarking‛. Delhi. Ofwat (1998). ‚Capital Works Unit Costs in the Water Industry: An Analysis of the June 1998 Water Company Cost Base Submissions‛. 108 Ofwat (2004). ‚Capital Works Unit Costs in the Water Industry: Feedback on Our Analysis of the March 2003 Water Company Cost Base Submissions‛. Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat (2006). ‚World Population Prospects: The 2006 Revision and World Urbanization Prospects‛ (http://esa.un.org/unup). Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat (2007). ‚World Population Prospects: The 2007 Revision and World Urbanization Prospects‛ (http://esa.un.org/unup). Shekdar, A.V. (1999). ‚Municipal solid waste management – the Indian perspective‛. Journal of Indian Association for Environmental Management 26 (2), pp. 100–108. UNSTATS Database (http://unstats.un.org/unsd/environment/Time%20series.htm#Waste) WHO/UNICEF Database (http://www.childinfo.org/water_data.php) World Bank (2008). ‚Review of Effectiveness of Rural Water Supply Schemes in India‛ Sustainable Development Unit. South Asia Region. June. Report 44789. WSP (2007). ‚2007 Benchmarking and Data Book of Water Utilities in India‛ (http://www.adb.org/ documents/ reports/Benchmarking-DataBook/default.asp) Yepes, Guillermo, Klas Ringskog and Shyamal Sarkar (2001). ‚The High Costs of Intermittent Water Supplies‛. Journal of the Indian Water Works Association. July. Zérah, Marie-Hélène (2000). ‚Household Strategies for Coping with Unreliable Water Supplies: the Case of Delhi‛. Habitat International, Volume 24, Issue 3, pp. 295 -307. Zhu, Da, P. U. Asnani, and Chris Zurbrugg (2008). ‚Improving Municipal Solid Waste Management in India – A source book for Policy and Practitioners‛. The World Bank, Washington DC. 109