Report No. 7214-PAK Pakistan Shelter Sector Review June 23, 1989 Country Department I Europe, Middle East and North Africa FOR OFFICIAL USE ONLY i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I~~~~~~~~~~~~~~~~~~~~~~ I~~. . '-) Document of the Wtorld Bamk This document has'a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank autiiorization. CURRENCY EQUIVALENTS Currency Unit = Rupee (Rs) Rs 1.00 = US$0.05556 US$ 1.00 = Rs 17.50 (December 1987) FISCAL YEAR July 1 to June 30 MEASURES AND EQUIVALENTS mm = millimeter (0.039 inches) m = meter (3.28 feet) km = kilometer (0.62 miles) km2 = 1 square km (0.386 square miles) ha = hectare (2.47 acres = 10,000 square meters) mgd = million gallons per day) 1 mgd = 4,546 m3/day) Marla = 225 square feet (20.91 square meters). In some areas I marla = 272 square feet (25.3 square meters) Kanal = 20 marla PRINCIPAL ABBREVIATIONS AND ACRONYMS ADB - Asian Development Bank ADBP - Agricultural Development Bank of Pakistan ADP - Annual Development Program ARV - Annual Rental Value DA - Development Authority DKA - Department of Katchi oadis FATA - Federally Administe ed Tribal Areas FMHW - Federal Ministry of Housing and Works GDP - Gross Domestic Product GOP - Government of Pakistan GOPUNJAB - Government of Punjab HBFC - House Building Finance Corporation HDA - Hyderabad Development Authority HPEPD - Housing, Physical and Environmental Planning Department HPPD - Housing and Physical Planning Department KDA - Karachi Development Authority KMC - Karachi Municipal Corporation KSDP - Karachi Special Development Project LDA - Lahore Development Authority LGRD - Local Government and Rural Development Department MCB - Muslim Commercial Bank MCO - Mobile Credit Officer MNA - Member of the National Assembly MHW - Ministry of Housing and Works MPA - Member of the Provincial Assembly NRA - National Housing Authority NWFP - Northwest Frontier Province PLS - Profit and Loss Sharing SBP - State Bank of Pakistan FOR OMCALM USE ONLY PAKISTAN SHELTER SECTOR REVIEW Table of Contents Page No. EXECUTIVE SUMMARY ................... i I. INTRODUCTION . A. Background .1 B. GOP 1987/88 Shelter for the Shelterless Program. 2 C. Program Proposed by Government. 4 D. Issues Raised by the Above Programs. 4 H. HOUSING MARKET TRENDS. 9 IIL. COPING WrrH KATCHI-ABADI GROWTH .15 A. The Problem and the ronstraints. 15 B. Early Attempts at Katchi-Abadi Upgrading 18 C. The Expansion of Katchi Abadi Upgrading under the Five Point Program .19 n. URBAN LAND AND SHELTER DEVELOPMENT. 22 A. Background and Constraints .22 B. More affordable Land Use, Infrastructure Standards and Plot Pricing .28 C. The Plot Allocation Process .29 D. The Land Acquisition Process .30 E. The system of Planning and Providing Trunk Infrastructure .31 F. The Need to Encourage Private Sector Development .31 G. A National Land and Shelter Development Program .32 This report is the result of a World Bank mission undertaken to Pakistan in December 1987, comprising Messrs. Wright, Ingram, Read (World Bank) Afshar, Gilmore, Kardar, Khalid, Lee and Pasha (Consultants), in collaboration with USAID (Gardner-Consultant). The report was reviewed with Government in March 1989. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. V. RURAKL HOUSING .......................................... 33 A. Introduction . ......... ..... ................. 33 B. Possible Constraints in Rural Housing Markets ................................. . 33 C. Present Rural Programs ..................... 37 D. Developing a More Sustainable Rural Housing Strategy . . 37 E. Progtam Design ........... 39 VI. HOUSING FINA.NCE . .. ............................ 41 A. The House Building Finance Corporation (HBFC) . . ................... ............... 41 B. Other Housing Finance . . 46 C. Strengths and Limitations of the Existing Housing Finance System . . 47 D. The Proposed Strategy for Housing Finance ............................ 49 VII. A MORE BALANCED SHELTER DEVELOPMENT PROGRAM ..................................... .. 57 ANNEXES 1. Statistical Annex . . . 62 2. Results of a Survey of Katchi-Abadis in Karachi ...71 3. Analysis of Land and Infrastructure Standards and Plot Pricing ....... ...................... 79 4. Land Acquisition ............................... . 92 5. Institutional Structure and Program Finance ..... 97 MAP IBRD - Map of Pakistan showing major centers - i - PAKISTAN SHELTER SECTOR REVIEW EXECUTIVE SUMMARY i. A new initiative has been launched in Pakistan to improve the standard of living for low income groups throughout the country by increasing allocations to the social sectors in FY87 to 45% of the Annual Development Program (ADP). The program includes new efforts for rural and urban land development and a new initiative to finance housing for low income groups. This initiative, to be coordinated by the newly created National Housing Authority (NHA), includes a first phase program to finance 100,000 houses in 1987/88 to be followed by a proposed second phase effort to finance 300,000 housing units in 1988/89 and beyond. This paper has been prepared in response to this latter proposal. ii. The program's objective is to improve the standard of living of low income groups in Pakistan. There is a recognition that housing standards have not improved along with economic growth and of the need to take decisive action to remedy the situation. In its present form, however, the proposed program may not fully achieve its objectives. By focusing solely on credit for housing construction, the program would fail to address the main constraints to housing development which are identified in this study and have been observed during the implementation of World Bank supported urban development projects in Pakistan. These include a lack of secure land tenure by the urban and rural poor, an inadequate supply of affordable developed land for new housing and a plot allocation system which is biased against the the poor. To make large amounts of housing credit available without at the same time addressing the main constraints to housing development would be very costly and less effective than a more balanced approach. iii. The new emphasis on social sectors comes at a time when the budget deficit increased from 5% of GDP in the early 1980s to almost 9% in FY87. The maintenance of long-term economic growth will require budget deficit reduction or slowing down the growth of expenditure and increased resource mobilization. Therefore, although the new priority for shelter is needed, new investments in social infrastructure would have to be carefully planned, adopting least cost solutions accompanied by improved pricing and fee structures to recover costs. In this context, the cost of the proposed program should be carefully considered. Subsidized interest rates and a high degree of politicization in program implementation are likely to limit opportunities for cost recovery. - ii - Housing Market Trends iv. While the urban population grew at 4.7f per year from 1961 to 1981 the housing stock grew at only 3.7f per year. Several trends can be identified from this data including that the supply of new land appears to be the most important constraint to housing development. V. Households have increased in size from 5.5 persons in 1961 to 6.5 persons in 1981, up an average of 1.03 persons in urban areas and 0.87 persons in rural areas. Much of the increase has been in the number of children per household. The number of persons per room remained unchanged for all Pakistan, down slightly in cities and up slightly in rural areas. This appears to have been because the house average size increased significantly in urban areas (from 1.8 rooms to 2.6) while average rural house size increased slightly (1.6 to 1.8). Thus, while household size has increased, overall crowding has remained about the same because units have gotten larger. vi. All of the increase in household size is not accounted for by an increase in the average number of dependant children. The number of potential household heads (married, widowed, divorced over 15) per unit increased from 2.54 to 2.71 overall (2.61 to 2.68 urban; 2.52 to 2.72 rural). This is a fairly significant indication of a tightening in the housing market in both rural and urban areas. It is surprising that this has occurred during a 20 year period in which real per capita income grew by 4.5% per year. Units have grown larger during the period, but there seems to be a constraint to an increase in their number. The constraint appears not to be in building materials supply, since houses have expanded and the price of construction material has increased less than general inflation. It appears, rather, to be in the supply of land for new housing. Coping with Katchi-Abadi Growth vii. The most evident consequence of the lack of affordable land for new housing has been the growth of katchi-abadis, unplanned, unserviced urban squatter settlements. In 1985, 5.5 million people, 18% of the urban population, lived in katchi-abadis. In many cities the katchi-abadi population has grown faster than general population growth. Local governments have lacked the resources to upgrade katchi-abadis, and the fragmentation of responsibility among sector institutions has impeded the implementation of upgrading programs. The lack of legal tenure impedes households' willingness to improve housing and their ability to use houses as collateral to borrow for improvements. viii. A World Bank assisted project in Karachi is providing basic infrastructure and services in 18 katchi-abadis with 600,000 residents and enabling them to purchase 99 year land leases. Although planning is advanced, implementation has been slow, and only 18,000 of the 175,000 eligible households have purchased leases. An important reason for the rather limited number of lease purchases is probaly that many residents feel they already have sufficient de facto tenure. - iii - ix. A National Katchi-Abadi Cell under the Ministry of Housing and Works (MHW) has been eatablished to monitor an expanded national program which will be implemented by Provincial Katchi-Abadi Directorates. The Government has given residents proprietary rights, and 50% of the cost of improving services is to be recovered. However, funds for this purpose are limited. In Punjab, for example, upgrading of all katchi-abadis may cost as much as Rs 1.5 billion over the next three years, but only Rs 60 million is in the 1987/88 ADP. Urban Land and Shelter Development x. Large amounts of land will have to be urbanized during the next few years, but the system for converting rural to urban land is not working well. The largest government urban shelter program has been area development schemes (serviced plots) in large and medium cities (Rs 2.1 billion budgeted in 1987/88). About 76,000 plots were developed in 1984/85 compared to the roughly 187,000 new urban households that year, 103,000 of which were in large and medium cities. However, few urban land development projects are affordable to low income groups. The few that are involve land priced far below market levels (e.g. land in Karachi is priced at 10% of market value). Vacancy is a serious problem. Of 120,000 plots developed since 1970 in Sind, only 3,000 were occupied by 1984. A limited program of 3 marla (75m2) plots for the urban poor is being implemented. As designed, however, the plots are only affordable with large subsidies. xi. A major effort is required to adopt affordable standards for land, infrastructure and housing development. For example, land for road space could be reduced from levels above 40% of project area to 15%-20%, and residential area increased. With careful design, the minimum plot size of 70 square meters could be reduced and still be acceptable to low income groups.l/ Under present regulations, very little land can be sold at market prices for high income and commercial/industrial purposes to reduce the price to low income beneficiaries. Only with a more flexible approach to development standards and differential pricing can land be developed affordably at the required scale. xii. Plots are allotted by "ballot" or random selection. Since land is underpriced, many speculators enter the balloting, and many plots are held off the market. Market pricing of plots would reduce the incentive to speculate, but more effective plot occupation covenants would also be necessary. xiii. The public sector cannot possibly meet all the need for urban land development, but there are many factors constraining private land I/ The Government of Punjab is of the view that for the prevailing social and climatic conditions, 70 square meter plots may be too small to be feasible. The Government of North West Frontier Province is of the view that these smaller plots are feasible and acceptable. - iv - developers. The impediments to private development should be more fully identified and a program to remove them should be prepared so that the private sector can make a larger contribution. Rural using xiv. The lack of secure tenure may also be an imoortant constraint to home improvement and construction in rural areas. At least as 26X of rural households in Punjab may live on the property of estate owners, and they may be hesitant to invest in shelter. There may also not be as much on-plot space for expansion as might be expected because of the large space requirements of rural households. High land prices in irrigated areas may make land for new houses unaffordable for low income groups. Because of the social value of land, many landowners may be unwilling to sell. xv. The Government is implementing a program to develop 2.2 million unserviced seven marla (175 square meters) plots. In Punjab, 550,000 plots have already been allotted. It is, however, using government land with little cost recovery. Also, some of the land appears to be in inappropriate locations for low income groups (e.g. far from villages; low-lying commons areas, etc.). The strategy is not replicable; the supply of government land is almost exhausted. xvi. More information is needed about rural land markets and the normal village expansion process to design a more sustainable rural residential land and shelter program. It would be important to find means of providing tenure to the many landless rural households to give them an incentive to improve their houses. The provision or upgracing of services for existing villages may be the next priority, but a well designed shelter loan program could also facilitate home construction and improvement. Housing Finance xvii. Organized housing finance accounts for less than 20% of housing investment and is dominated by the House Building Finance Corporation (HBFC). Most housing is financed directly from private savings and informal borrowing, about which little is known. xviii. HBFC receives a line of credit from the State Bank of Pakistan (SBP) (Rs 1,5 billion in 1987/88) which it lends to beneficiaries using a Profit and Loss Sharing (PLS) formula und_r the Islamic Banking system. HBFC has earned an average return of 6% to 7% on loans and has paid an average 3% to 3.5% to SBP. In 1986/87 about 15,000 loans were sanctioned for Rs 1.2 billion. HBFC finances housing units of up to 225 square feet lending Rs 90 per square foot (to a total of Rs 200,000), although the unit cost of housing can be two to three times this amount. xix. There are several problems with the present system. The PLS formula used by HBFC has reduced the returns to HBFC from before the introduction of Islamic Banking in 1979 (when HBFC earned 9%-10? from borrowers). Thus, while Islamization has generally introduced more market pricing of capital, HBFC's returns have been reduced. HBFC is therefore very dependant on concessional SBP financing; funds are exhausted well before the end of each financial year. xx. HBFC does benefit some low income households, but the requirements that beneficiaries make high downpayments, that they hold secure title and that housing comply with building regulations effectively exclude most low income households. HBFC administers some low income schemes, a flood relief program, a home improvement program in old Lahore, ._nd a loan program under recent Government initiatives. These have not been very successful (90% deliquency in flood program; low disbursements in Lahore). xxi. HBFC's profits have declined from Rs 11.8 million in 1983 to Rs 755,000 in 1986 due to lower margins under PLS. There is serious delinquency (though data is scarce) which causes cash flow problems. When HBFC cannot repay, SBP deducts arrears from drawdowns, further reducing the size of each year's program (net disbursements down from Rs 1,400 million in 1986/87 to Rs 1,050 million in 1987/88. xxii. The present system is costly and its size is limited to available concessional funding. It is inequitable, subsidizing a privileged few. Subsidies may invite inefficient land use. There is a need, therefore, for more market pricing of loans so that additional sources of funding can be tapped. The development of market instruments may take time; it would be facilitated by an improved macro-economic environment. xxiii. Short to Medium Term Improvements. The design of special housing finance programs targeted to low income groups should be improved. HBFC's financial condition should not be undermined by forced participation in risky loan programs. Government should cover part of such risk. HBFC should take a more active role in designing loan programs to be efficient (e.g. more flexible loan amounts and collateral, incremental construction) and improving community education. Collection systems (follow-up, monitoring) also require further attention. xxiv. Medium to Long-Term Improvements. The PLS formula used by HBFC could be revised to recover more of the large profit accruing to homeowners. Regulatory and/or legislative changes should be considered to improve resource mobilization (e.g. to allow borrowing from insurance companies and private savings mobilization). With support, HBFC could evolve towards a more market oriented institution, but in the long term a public-private sector mix may be preferable. Private institutions could begin with limited access to credit markets and, as they perfect lending instruments, they could mobilize private savings. HBFC would focus more on lower income groups. A detailed plan should be prepared of the legal and regulatory framework to support private housing finance. vi - The Proposed Program xxv. In parallel with the provision of credit for housing, a modified National Shelter Development Program would address the main development constraints in the sector, the inadequate supply of developed land, the lack of access by the poor to secure land tenure, and the present unaffordable standards for land, infrastructure and housing development. The newly created National Housing Authority would coordinate and finance such a program and provide technical assistance to implementing agencies at the provincial and local levels. It could be structured along the lines of present programs, consisting of the following: xxvi. National Community Improvement Program. This would be a modified version of the present katchi-abadi improvement program expanded from present levels to cover a significant percentage of the 5.5 million people living in katchi-abadis within a five year period. It could also be expanded to many medium and small towns and to areas which legally occupied but which have low service levels. Such a program would include normalization of tenure, affordable engineering standards, recovery of land cost, recovery of service cost through utility rates and property taxes, and strengthening of provincial monitoring capacity. The program would also include home improvement loans which would be channeled through a financial intermediary such as HBFC or a commercial bank. xxvii. Issues which would be addressed in designing such a program include: (a) how could procedures for normalizing tenure be improved? (b) what are the constraints to including katchi-abadis in the municipal tax system as they are improved? (c) what iAplementation arrangements are appropriate for cities of various sizes? and (d) what are the most feasible opticns for improving and regularizing tenure in katchi-abadis on private land? xxviii. National Land and Shelter Development Program. This would be an extension of present area development schemes with more efficient design and a credit component. Unit costs would be higher than past land development programs because land would be purchased at market prices to make the program replicable, but more appropriate standards and innovative pricing techniques would keep the program affordable to the urban poor. The program would include home materials and construction loans provided through a financial intermediary for low income plot recipients as well as for households owning serviced plots outside project areas. xxix. Issues which would be resolved in designing such a program include: (a) what institutional arrangements would be required, including channeling of technical assistance and financing from the national to the local level? (b) what changes would be required in the regulations for plot allocation? (c) what changes in land acquisition procedures and in the planning and provision of trunk infrastructure would facilitate this type of program? The program would not meet the needs of all new urban households. It would be important, therefore, to assess how the private sector could make a greater contribution over time to urban land development. - vii - xxx. Rural Land and Shelter Development. This would make credit available to rural households to finance house construction and/or priority shelter related needs such as land acquisition, on plot sanitation and building materials for new construction or upgrading, with flexible loan amounts and repayment periods. There is a need for large-scale beneficiary education, which would reqrl.rethe involvement of local government officials, but beneficiary selection and loan disbursement and collection should be through a financial intermediary such as a commercial bank or the Agricultural Development Bank of Pakistan (ADBP) in order to minimize political interference. The program should be initiated on a pilot basis and expanded gradually as institutional capacity and experience indicate. xxxi. Several issues would be addressed including: (a) what priority do rural households assign to house construction as opposed to other needs such as socia'l services, water supply and sanitation? (b)would a rural public services upgrading program be more appropriate in combination with a rural housing program? (c) where households do not own the land they occupy, is it possible to normalize their tenure? (d) what is the normal process for village extension and how can it be facilitated to allow households to acquire new land for residential purposes? (e) is the supply of building materials a constraint to house construction and, if so, what can be done? (f) what construction techniques are used in rural housing and how can credit facilitate them? xxxii. Housing Finance. "he most immediate priority would be the improvement of HBFC's oper.' ons, especially if it is to implement large housing credit programs for low income communities. This would inciude improved design of special credit programs, more flexibility in collateral requirements and an increased effort to improve collections. xxxiii. Over the longer term, it would be important to establish an environment which would foster the development of private as well as public housing finance. This would include the development of market priced lending instruments, the establishment of a regulatory environment for expanding resource mobilization and the reduction of credit risk. xxxiv. A possible role for international agencies in such a program is suggested in paragraphs 7.11 to 7.15. PAKISTAN SHELTER SECTOR REVIEW r. INTRODUCTION A. Background 1.01 Pakistan has faced an enormous challenge in housing its growing population, especially since independence. In 1951 only 6 million people or 17% of the total population was in urban areas. By 1981 this had risen to 24 million or 282. While housing needs in rural areas have been intense, it is in the cities where population and densities are growing most rapidly that the greatest pressure is usually felt. Given other development priorities, however, it has not always been possible to devote as much resources to housing investment as necessary. Gross Domestic Capital formation in the Housing Sector has seldom surpassed lOX of total investment (see Annex 1, Table 1). In many countries the percentage of total investment in housing ranges from 15% to 252. 1/ 1.02 Recognizing the need to improve the living conditions of the nation's poorest households, the Government announced in late 1985 a major development program (designated the Five Point Program) for which a total of about Rs 23 billion would be allocated in FY 1987-88. This includes an allocation of about Rs 2 billion to extend land ownership and improved housing to low income groups. For urban dwellers, the program includes a scheme to upgrade existing katchi abadis, illegal squatter settlements and a program to provide three marla (75m2) plots to the urban poor. For the rural poor there is a Seven Marla Program to provide about 2.2 million residential plots of 175 m2 each. The Government estimates that about 33% of rural and 20w of urban dwellers are below the poverty level. Since the announcement of the Five Point Program, there has been a significant increase of financial allocations to the social sectors, especially to education. In FY87, 45% of the Annual Development Program (ADP) was for the items of the Five Point Program, a 17% increase over the previous year's expenditures. 1.03 This Shelter Sector Review was prepared in response to the announcement of important new initiatives by the Government of Pakistan to provide housing loans to the urban and rural poor. There is a general agreement about the fundamental objectives of this program. The increased attention to living conditions, as well as literacy, education, helth and other basic services for lower income groups, is an important recognition of the need for more equitable service provision. This paper assesses the new program proposals within the context of existing land and housing development programs. Suggestions are made for improving program design to meet important equity objectives more effectively within likely economic constraints. 1/ Estimates of housing deficits can be difficult to compare because they involve deffering assumptions, but they can indicate the order of magnitude of the problem. Zaki (in "Housing Conditions in Pakistan: 1960-80", the Pakistan Economic Review, Volume XX, 1981) estimates a housing shortage of about 3 million units in 1980. - 2 - B. GOP 1987/88 Shelter for the Shelterless Program 1.04 The Government announced a program to finance a total of 150,000 houses for the shelterless during 1987/88 (see Table 1.1). This first phase of the "Shelter for the Shelterless" program has several components. It will construct 75,000 nucleus houses throughout the country to be financed with Zakat (religious charity) funds and given free of charge to the very poor (Mustahqueen). The house construction, about two per Zakat Committee (there are about 37,000 committees), will be supervised by Tehsil/Taluka Zakat Housing Committees consisting of the Assistant Commissioner, the Chairman of the Tehsil Zakat Committee and an engineer from a Government technical department to be entrusted with the work. In urban areas, local governments or development authorities will be involved in implementation. 1.05 Three subprograms will be financed from existing MNA/MPA (elected officials) development budgets. First, 20,000 houses costing Rs 25,000 each will be built in rural areas and allotted to shelterless plot owners, some presumably beneficiaries of Seven Marla schemes (see para 5.09). District Construction Committees are being established in rural areas to execute the schemes and District Housing Committees, consisting of elected representatives and the Deputy Commissioner and chaired by a member of Parliament, will screen applicants. Houses are to be allotted with Rs 5000 downpayments and the remainder recovered in 40 biannual installments with no interest. 1.06. The second subprogram consists of 20,000 rural housing improvement loans of Rs 10,000 each. Beneficiaries will also be selected by the District Housing Committees. The loans will be released in four tranches after verification of construction by a local committee consisting of the local Member of the Union Council, the Chairman of the Zakat Committee and the head of the loecl primary school. These loans will also be interest free. 1.07 The third subprogram consists of Rs 10,000 interest-free loans which will be provided to an additional 15,000 beneficiaries for new house construction in urban areas with the remaining cost (developed land and construction) to be paid by beneficiaries. Beneficiaries are presumed to include recipients of three marla plots (see para 4.12). District Housing Committees will be responsible for selecting beneficiaries and recovering loans. Municipal councils and development authorities will contract for house construction, estimated at Rs 25,000 per unit, collecting the balance of Rs 15,000 from the beneficiaries in installments as construction proceeds. - 3 - Table 1.1: EXISTING AND PROPOSED GOVERNMENT SHELTER PROGRAMS 1987/88 Housing Loan Program Program Funding Source Beneficiaries Units Nucleus Houses constructed Zakat very poor by government supervised by Charity Funds (Nustaqueen) chosen 75,000 Tehsil/Taluka Zakat Housing by 37,000 Zakat Committees Committees Houses costing 25.000 Rs. GOP Shelterless rural plot 20,000 (5,000 down & balance recovered owners selected by District interest free in 20 years; Housing Conmnittees possibly on 7-marla plots Rural housing improvement GOP Rural poor selected by 20.000 loans up to Rs. 10.000 District Housing Committees Rs. 10,000 interest free loans GOP Urban Poor, including 3 marla 15.000 for urban housing units plot recipients, selected by costing Rs. 25,000 with remainder District Housing Committees to be paid by beneficiaries HBFC housing construction HBFC HBFC 20,000 loans of Rs. 10.000 under (from normal normal terms. allocations) TOTAL .. 150.000 Proposed Loan Program (1988/89 and Next Two Years) Proaram Funding Source Beneficiaries Units_ Loans of Rs. 15,000 to house- GOP Rural/urban households 300,000 holds for construction and distributed proportionately upgrading; 6X profit; in lower 1/3 of income Total cost of Rs. 5 billion, distribution. covering 1989-1991. - 4 - 1.08 A separate subprogram calls for the provision of 20,000 loans of Rs 10,000 each for urban house construction by HBFC with recoveries based on their normal PLS formula. The overall shelter program, except for the Zakat and HBFC subprograms will be coordinated and monitored by the newly created National Housing Authority (NHA). Housing authorities have also been proposed to monitor the program at the provincial level. The total cost of the initial program to the Zakat Committees will be Rs 2 billion; to Government and HBFC it would be about Rs 1 billion. C. Program Proposed by Government 1.09 A second phase program has been outlined towards the ultimate objective of constructing one million homes. The program would provide financing for 300,000 houses during the three years beyond 1988. It has been proposed for possible international funding. The estimated cost of the program to Government is about Rs 5 billion (1987 prices). The program would make loans of Rs 15,000 each to beneficiaries in the bottom 1/3 of the income distribution. The loans, distributed in proportion with the population, would be for new house construction and upgrading in rural and urban areas. The program would be monitored by the NHA. Implementation would be by District Housing Committees, chaired by a District Housing Officer, and by Tehsil Housing Committees chaired by the local Member of Parliament. Tehsil Housing Committees would receive and evaluate loan applications and select beneficiaries by ballot. Loan disbursements would be released in three stages upon verification of works by a local housing committee, also chaired by an elected representative. Disbursements and collections would be through a commercial bank. Loans would be recovered with 6% nominal profit in 45 biannual installments. D. Issues Raised by the Above Programs 1.10 As mentioned above (para 1.03), there is no disagreement about the fundamental objective of these programs. There are, however, three types of concerns about whether these proposals would be the most effective means of reaching the objective: 1. The Constraints to Housing SupplY 1.11 The proposed program emphasizes credit for the construction of housing superstructures by low income households who already own land. As presently formulated, the proposal does not address the main constraints to housing development identified in this paper and observed during the implementation of World Bank supported urban development projects in Pakistan in recent years. First, there is an assumption that large numbers of katchi-abadi dwellers would qualify for housing credit. In fact, the process of aormalizing tenure in katchi-abadis is progressing slowly due to institutional rigidities, which may limit the demand for housing credit (see Chapter III). Second, building materials purchase and house construction appear to be less serious constraints to housing development than the supply of serviced residential land. Underlying the constrained supply of serviced land are land use and infrastructure standards which are not affordable to most households. There is also an implicit assumption in the new proposals that there are large numbers of urban households who already own plots and would like to build houses on them. In fact, the plot allocation system does not appear to allocate plots to those in need (see Chapter IV). Third, there also appears to be an inadequate supply of residential land in rural areas (see Chapter V). 1.12 A cost effective shelter program would have to deal with the real constraints in the housing sector. The proposed National Housing Policy recently outlined by the Environment and Urban Affairs Division of the Ministry of Housing and Works suggests many policy measures which would address these constraints. These include improved urban land acquisition and allotment policies, more intensive land use and more efficiently planned low income urban and rural community improvement programs. 2. Implementation 1.13 The first phase program to construct housing in small groups throughout the country will be difficult and costly to manage. The change in the proposed second phase program to provide loans for housing rather than direct construction by government agencies is therefore a step in the right direction, Since it is usually more economical for low income individuals to arrange for self-help or small contractor construction themselves. Additional flexibility should be considered to recognize that households' priorities may differ. For example, many low income katchi-abadi or rural village dwellers may wish to limit their borrowing to high priorities such as sanitary facilities, water connections or roofing material loans. 1.14 The proposed program of 100,000 units costing Rs 1.7 billion per year should be compaied with HBFC's current yearly capacity of 22,500 loans to middle and upper income households with clear tenure totaling about 1.5 billion. The Agricultural Development Bank of Pakistan (ADBP) makes about 100,000 loans per year for which it has a staff of over 6000, including some 2000 mobile collection officers. These comparisons indicate that the task of administering the proposed programs may have been underestimated. 1.15 Both programs have a high level of political involvement, especially in beneficiary selection and loan disbursement. Under the proposed program, a District Housing Officer would chair the District Housing Committees instead of an elected representative, which may reduce political influence somewhat, but the risk of political interference and consequent poorly targeted implementation and inadequate collections would remain very high. 3. The Cost of the Program 1.16 The cost of the program should be considered in relation to other housing and urban development expenditure. The total cost of the first phase program, Rs 3 billion, is significant, roughly equal to the total amount being spent in FY 87/88 on all other public housing development programs (see Table 1.2). Much of the Rs 2 billion cost would be borne by the Zakat, but this is likely to deplete available Zakat funds and not be replicable. - 6 - 1.17 The second phase proposal to finance 300,000 houses over the next three years would cost about Rs 1.7 billion per year. This is the amount of loans only, however. Table 1.3 illustrates what the cost of the program would be per year if the target of one million houses were achieved by FY91. Assuming additional costs for basic services and administration, such a program would cost Rs 6.7 billion per year, about two times this year's housing development budget. Most of this expense would be additional to the programs already being financed. It is clear from the magnitude of these figures that, whatever the ultimate size of the shelter program, it would be essential to minimize unit costs and to maximize cost recovery. 1.18 There is no obvious reason to offer households subsidized financing, especially given the inability of the government to maintain such a program over time and the high cost of funds to government (see para 6.22). Since so little financing is available, most households would be better off if market rate financing were available to them. The proposed movement from no return loans to 6% is an improvement but still far short of the cost of funds to government. The ADBP currently achieves a 12% return in loans to farming households plus a 32 service charge (waived for prompt repayment). 1.19 In spite of the need, this is a difficult time to increase public expenditure for new programs. Per capita income has been growing at about 3% to 4Z per year, but the fiscal deficit has also been increasing - from 5% of GDP in the early 1980's to almost 9% in FY87. The long-term climate for growth and the ability of Government to support development are threatened by the large and growing 1eficits. A substantial budget reduction effort is required together with a large resource mobilization effort in order to bring the deficit down to levels consistent with sustainable long-term economic growth. A shift in Government spending towards more social infrastructure expenditure (both ADP and recurrent) is important, but in the present economic context it will have to be accompanied by better investment programming and improved pricing and fees to cover costs. The proposed investments in shelter and related infrastructure should therefore be well planned, least cost solutions, and, to the extent possible, both capital and recurrent costs should be recovered from beneficiaries. 1.20 Some type of housing credit program would probably enable low income households owing plots and those acquiring tenure in katchi-abadis to build better quality houses more quickly. The availability of credit to all groups would increase the rate and improve the quality of construction and enable even some low income groups to undertake priority home improvements. (Housing finance is discussed in detail in Chapter VI.) However, to put large amounts of resources into housing credit without addressing the main constraints in the sector would be more costly than necessary. The Government would in fact be paying the high cost of inefficient policies elsewhere in the housing sector. It is important, therefore, to take action to relieve the main constraints to housing development in parallel with any program of housing credit. It would also be important to consider carefully the implementation and program cost issues raised above. 1.21 The following chapters outline in more detail the market trends and the ongoing programs in the housing sector and the lessons learned for future development. Chapter VII outlines a possible more balanced program of addressing constraints in the sector in parallel with the provision of credit. Table 1.2: Develooment Budget 1987/88 Housing Sector (Millions of Rupees) Scheme Baluchistan NWPF Punjab Sind Federal Total Total(X) 7 Marla Scheme 45.0 42.0 0.0 157.6 0.0 244.6 7X Katchi-Abadi Program 38.7 1.0 60.0 343.9 0.0 443.6 13X Area Development 4S.1 10.0 604.6 1448.8 0.0 2108.5 63X Civil Serves. Housing 25.2 60.5 241.7 35.0 213.0 575.4 17X Total Public Housing 154.0 113.5 906.3 1985.3 213.0 3372.1 100X Urban Development 193.4 296.4 2765.4 3920.6 981.1 8156.9 Total Development 1507.9 2642.3 9516.1 6361.2 30815.5 50843.0 Public Urban Housing as % All Urban Dev. 56% 24% 33% 47% 22% 38% Total Public Housing as % Total Oev. 10% 4% 10% 31% 1% n Table 1.3: Indicative Cost of Shelter Program Assumina 1 Million Houses Financed by FY91 (Millions of Rupees) Baluchistan NWPF Punjab Sind Federal Total Annual no. of units -Urban 2347 5773 45253 28582 0 81955 -Rural 12673 32579 118726 37399 0 201378 -Total 15021 38353 163979 65981 0 283333 Annual Cost (Million Rupees) -Urban 55.5 136.4 1069.1 675.2 0.0 1936.2 -Rural 299.4 769.7 2804.9 883.6 0.0 4757.6 -Total 354.9 906.1 3874.0 1558.8 0.0 6693.8 Proposed cost as % of Current Housing 231% 798% 427% 79% 0% 199% Current Total 24% 34% 41% 25% 0% 13% Proposed urban cost as % of current urban housing 51% 191% 118% 37% 0% 62% Proposed rural cost as % of current rural housing 665% 1833% N/A 561% N/A 1945% INDICATIVE COST OF SHELTER PROGRAM FY89 TO FY91 850.000 Units @ Rs 15.000 plus Rs 7,500 basic services plus 5% administration/supervision = Rs 20,081.25 million = Rs 6.693.75 million per year - 9 - II. HOUSING MARKET TRENDS Background 2.01 Pakistan's population has been growing rapidly. During the 1970's population grew at a rate of about 2.9X. As in most developing countries, urban population is growing faster than general population growth, rising at a rate of about 4.4% per year with the rural population increasing at 2.3% per year during the same period (see Annex 1, Table 1.03). High urban population growth in Pakistan is largely the result of high natural population growth. Fertility in urban areas has been slower to decline than in the cities of other countries with comparable incomes, and, as elsewhere, high natural growth has been compounded by rural-urban migration. 2.02 Table 2.1 shows the recent growth in the number of households in Pakistan. During the early 1980's Pakistan's population was growing by about 360,000 households per year, roughly an additional 140,000 new urban households and 220,000 new rural households each year. There is, thus, considerable pressure to accommodate new households as well as to improve the living conditions of existing households. For the urban areas, this means accommodating a population roughly equal to that of Faisalabad each year into cities which are already crowded and poorly serviced. New rural households can be housed in a much larger and less densely populated network of rural settlements, but even here the large absolute number of households to be housed constitutes a considerable challenge. 2.03 There is a widely held perception in Pakistan that the county is not coping with the housing challenge. One statistic which is widely quoted is that while the urban population grew between 1961 and 1981 at an annual rate of 4.7X, the urban housing stock grew during the period at an annual rate of only 3.72. These statistics may be cause for concern, but it is important to understand some of the trends behind them. 2.04 To identify accurately underlying trends in housing markets it is necessary to examine housing market behavior over fairly long periods because activities such as housing construction can demonstrate great variability on an annual basis. Fortunately. Pakistan has a high quality census which can be used to identify trends over the period 1962 to 1981; this data has been supplemented with information gathered from a number of additional sources. The Increase in Household Size 2.05 Although the population data in Table 2.1 suggest the possibility of a housing shortage, the growth rate of households has been lower than the growth rate of the population; household sizes have obviously increased over this period. From 1961 to 1981 average household size in Pakistan increased from 5.5 to 6.5 persons. Much of this increase appears to reflect demographic factors rather than housing market tightness. Although the 1961-81 increase in household size has been somewhat greater in rural areas (up 1.03 persons) than in urban areas (up 0.87 persons), average household sizes are still greater in urban than in rural areas, a pattern which reverses the one usually observed. - 10 - Table 2.1: Pogulation and Household Growth Area Year Annual rate 1961 1981 of growth Population ALL 39.442.439 82,055.097 3.73% URBAN 9.614.004 23.841,471 4.65% RURAL 29,828.435 58.213.626 3.49% Households ALL 7.159.634 12.587.648 2.86% URBAN 1,647.412 3.554.174 3.92% RURAL 5.512,222 9.033.475 2.50% Source: 1961 and 1981 Census of Population Table 2.2: Household Size and Unit Size Year ALL URBAN RURAL Persons per Household 1961 5.51 5.84 5.41 1981 6.52 6.71 6.44 Persons per Room 1961 3.3 3.1 3.3 1981 - Unadjusted 3.5 3.2 3.6 1981 - Adjusted 3.3 2.6 3.6 Rooms per unit/household 1961 1.7 1.8 1.6 1981 - Unadjusted 1.9 2.2 1.8 1981 - Adjusted 2.0 2.6 1.8 Source: 1961 and 1982 Census of Housing, Census of Population. - 11 - Table 2.3: Size Distribution of Units Area No. of Rooms Average Rooms 1 2 3+ Per Household 1961 ALL 59.9 23.9 16.2 1.7 URBAN 56.1 24.7 19.2 1.8 RURAL 6i.1 23.7 15.2 1.6 1981 ALL 51.5 29.6 18.9 1.9 URBAN 42.6 31.5 25.9 2.2 RURAL S5.1 28.8 16.1 1.8 Source: Derived from 1961 and 1981 Census of Housing. Table 2.4: Population and Household Growth Area Person per Children (1S & Married; Widowed, Household Never Married 15+ Divorced; all 1S+ 1961 ALL 5.51 2.97 2.54 URBAN 5.84 3.23 2.61 RURAL 5.41 2.89 2.52 1981 ALL 6.52 3.81 2.71 URBAN 6.71 4.02 2.68 RURAL 6.44 3.73 2.72 1981 MINUS 1961 ALL 1.01 0.84 0.17 URBAN 0.87 0.79 0.07 RURAL 1.03 0.84 0.20 Source: Derived from 1961 and 1A81 Census of Population - 12 - The increase in household size mainly results from the greater number of children accompanying the high birth rate, but part of the increase in household size seems to be due to housing market conditions, a topic dealt with below (para 2.11). 2.06 Two common indices of housing market conditions, persons per room and the average number of rooms per unit, have not shown deterioration. The 1981 census used nearly the same definitions for housing and households as the 1961 census, but kitchens were not counted as rooms in 1981 whereas they were counted as rooms in urban areas in 1961. Accordingly, the number of rooms was adjusted in 1981 to make it comparable with the 1961 definition. The index, shown in Table 2.2, indicates that the number of persons per (adjusted) room was essentially unchanged for all Pakistan from 1961 to 1981. In urban areas, however, persons per room fell while in rural areas it increased. This suggests that housing conditions improved somewhat in urban areas and deteriorated somewhat in rural areas. 2.07 Table 2.2 also shows as a measure of unit size the average number of rooms per dwelling unit. Unit size has increased from 1961 to 1981 with urban units growing much more than rural units. Since averages are sensitive to extreme values, it is possible that average unit size has gone up because a few households have been constructing very large units. To check for this, Table 2.3 shows the distribution of units by the number of rooms. It is clear that the proportion of one room units has decreased overall, most substantially in urban areas. Therefore it appears that the increase in average unit size is real and not just the result of relatively few large houses. 2.08 The housing market outcomes observed so far indicate that housing units have gotten somewhat larger, that units are overall as crowded in 1981 as they were in 1961, and that household size has increased. There are also unexpected differences in housing consumption between urban and rural areas, with urban areas having fewer persons per room but larger households than in rural areas. Persons per room is only an imperfect index of housing condition, however. Rural households undoubtedly have larger plots than urban households and could build more room if they wished. The larger average household size in urban areas in 1981 is due entirely to the fact that urban households have more members over 15 years of age who have never married. This may be because urban dwellers marry later than rural dwellers, or to a more frequent incidence of live-in servants and paying visitors in urban as opposed to rural households. From 1961 to 1981 the number of children per household increased more rapidly for rural than urban households, and by 1981 rural households had slightly more children than urban households. Changes in the relative composition of rural and urban households are likely to continue, and rural households may soon be larger than urban households. 2.09 Although housing market conditions are not as bleak one might draw from paragraph 2.03, it is surprising that housing consumption in Pakistan has barely increased over a 20-year period. During this same 20-year period, per capita real income has increased by nearly 90 percent (Pakistan's per capita real income grew by 4.52 per annum in the sixties and 1.9b per annum in the seventies). Housing is a normal good with an - 13 - income elasticity of demand which normally falls between 0.5 and 1.0; with Pakistan's income growth, housing consumption would normally increase quite significantly over 20 years. The fact that housing consumption has not increased with income merits further analysis. Higher Number of Potential Household Heads 2.10 In order to explore further the pattern of housing demand, Table 2.4 presents a summary of household composition in Pakistan in 1961 and 1981. The table classifies household members into two groups. The first is comprised of children less than 15 years of age plus those 15 and over who have never married. This group represents those dependents (children and young adults) normally expected to be living at home with their parents. The second group is comprised of persons 15 years and older who are, or have ever been, married. This group would be composed of parents and other adults who might be expected to set up households on their own if they had an opportunity to do so. The number of such potential household heads per household would vary with housing market conditions, whereas the number of dependents per household would be much less sensitive to housing conditions. This number of potential household heads per household will be used as as index of housing market tightness. 2.11 Table 2.4 shows several interesting patterns when household members are categorized into the two groups. First, in 1961 the number of potential household heads per household was greater in urban than in rural areas, which implies that urban housing markets were tighter than rural housing markets at that time. Second, by 1981 this pattern had changed and the number of potential heads was essentially the same in urban and rural areas, implying that housing markets in the two areas had become almost equally tight, During this time both markets tightened, however. Finally, the decomposition of the increases in household size (1981 minus 1)61 in Table 2.4) shows that 84 percent of the increases in household size is accounted for by dependents and only 16 percent is accounted for by the increase in potential heads. 2.12 What accounts for the increase in the number of potential heads per household in a country where per capita income has nearly doubled? One obvious explanation might stem from the increased life expectancy in Pakistan which would increase the number of elderly who would be likely to move in with a child. Although the number of elderly has certainly increased in Pakistan, the rest of the population has increased equally fast. As a result, the proportion of Pakistan's population that is over 60 has remained essentially constant at 6.93 percent in 1961 and 6.99 percent in 1981. The increase in potential heads per household is not related, therefore, to an increase in the elderly or to any similar demographic explanation. Constraints in Land Markets Indicated 2.13 It seems more likely that the housing market in Pakistan is encountering a constraint to expansion. The constraint seems to be operating on the number of dwelling units and not on their size, since dwelling units have become larger over time (see para. 2.07). Moreover, when other data about the functioning of the housing market in Pakistan are examined, there is no evidence of constraints operating on the housing - 14 - construction industry. The prices of construction material (see Annex 1, Table 1.05) have increased in pace with general inflation; the construction industry is competitive and many households do their own construction. 2.14 Observation of housing trends also points to a constraint operating on the land market -'. The most glaring symptom of the inability of land markets to supply an adequate number of serviced plots has been the growth of katchi abadis, unplanned unserved urban squatter settlements. The roughly 200,000 person per year growth of Karachi's squatter population represents a growth rate approximately twice that of Karachi's overall population. Over two mill1en people, 37% of Karachi's population, live in katchi abadis. The katchi-abadi population of other cities is estimated to be about 251 of total population on average. The program for improving katchi-abadis is assessed in more detail in the following chapter. - The possibility of other constraints to housing production such as rent control (see para 4.17). This and other possible constraints should be further examined. - 15 - III. COPING WITH KATCHI-ABADI GROWTH A. The Problem and the Constraints 3.01 About 20% of Pakistan's urban population in 1985 lived in katchi-abadis. Katchi-abadis are illegal squatter settlements; most residents do not have legal tenure. Those enumerated in the survey shown in Table 3.1 on March 23, 1985 had at least 40 houses. About 5.5 million persons lived in the 2302 katchi-abadis surveyed, and 872 of the katchi-abadis were on state owned land. By far the largest number of katchi-abadi dwellers are in Sind, 3.4 million people, most of whom live in Karachi. 3.02 Katchi-abadis are a market response to the inability of the present land development system to provide sufficient affordable developed land. They exist for a number of reasons, foremost of which is the lack of developed land priced to be affordable to low income households and allocated equitably. There has been a pattern to katchi-abadi development which has included protection by organizers against eviction, the development of basic street layouts and even the provision of minimal services by organizers. An informal land market has even developed, although tenure is not formally registered. 3.03 Annex 2 shows information from a recent survey of katchi-abadis in Karachi. Since many of the katchi-abadis in Karachi are older, they have the characteristics of more established communities. Many of the residents have been in Karachi for some time (61% since birth or since 1955). Many of the houses are fairl. large (on plots of 60 square yards or more), and many are made of permanent materials. The survey snows, however, that it is in access to municipal services that the katchi-abadis are set apart from more establis.k%d communities (only 452 of houses have piped water; 57% throw garbage in street). Many cities have made at least partial attempts to upgrade katchi-abadis by improving some basic services (water supply, sanitation, drainage, etc.) and by transferring legal tenure to residents of the land they own. These programs typically encounter several types of problems. 3.04 The first is the lack of adequate local resources to finance capital improvements as well as recurrent costs and maintenance. Local tax revenue, including the property tax, has scarcely kept up with inflation and population growth. Cost recovery in the water supply sector has been poor ani financial planning needs improvement. Only a sniall percentage of katchi-abadi households pay for water services, generally at the low end of the tariff scale which is inadequate to recover costs. Cost recovery in slum upgrading projects has been negligible. The difficulty of recovering capital costs in slum upgrading is compounded by - 16 - Table 3.3: SummarX of Katchi-Abadis in Baluchistan (December 1987) No. of Area in Housing Population Katchi Abadi acres Units Private 11 2.247 17.380 133,286 State/Municipal 1 3,000 2.000 lS.000 Municipal 8 1,446 1.866 14.233 Private/State 14 1,209 :.083 8.273 Railway/State 3 800 7.000 90.000 State 28 2,420 7.879 51.203 TOTAL 66 11.132 38.198 327.99S Table 3.4: Katchi-Abadis in Puniab (December 1985) No. of Katchi Household Population Abadis Units of Katchi Abadis Katchi-Abadis 205 40.808 2,94,945 of Federal Government Land. Katchi-Abadis 504 1.44.252 8.71,300 on Provincial Government Land. Katchi-Abadis 49 13.410 76.981 on Municipal Land. Katchi-Abadis 64 15,611 1.01,489 on Combined Land (Partly Private & Partly State Land). Katchi-Abadis 80 16,197 1.20.857 Purely on Private Land. TOTAL 902 2,30,278 14,6S.572 - 17 - Table 3.3: Sumary of Katchi-Abadis in Baluchistan No. of Area in Housing Population Katchi Abadi acres Units Private 11 2,247 17.380 133.286 State/Municipal 1 3,000 2.000 15.000 Municipal 8 1.446 1.866 14,233 Private/State 14 1.209 2.083 8.273 Railway/State 3 800 7.000 90.000 State 28 2,420 7.879 51.203 TOTAL 66 11,132 38.198 327.995 Table 3.4: Katchi-Abadis in Punjab No. of Katchi Household Population Abadis Units of Katchi Abadis Katchi-Abadis 20S 40,808 2.94,945 of Federal Government Land. Katchi-Abadis 504 1.44,252 8,71,300 on Provincial Government Land. Katchi-Abadis 49 13.410 76.981 on Municipal Land. Katchi-Abadis 64 15,611 1,01.489 on Combined Land (Partly Private & Partly State Land). Katchi-Abadis 80 16,197 1.20,857 Purely on Private Land. TOTAL 902 2.30,278 14,65572 - 18 - the difficulty of mobilizing resources to pay for services and to maintain any improved areas. By definition, as illegal occupiers, katchi-abadis dwellers do not pay property tax and hence make no contribution to the costs of general urban services. Only a small percentage of katchi-abadi households pay for water services, generally at the low end of the tariff scale which is inadequate to recover costs. As a result, investments in urban infrastructure are very dependent on the federal budget, and operational budget needs are not balanced by receipts, resulting in poor services and financial strains on local administration budgets. 3.05 Property tax is an important source of municipal revenue, and hence an indirect means of paying for services, but there are inconsistencies in the existing property tax system which exempt many properties from tax. Low-valued properties are exempt (e.g. in the Punjab, owner-occupied properties with an annual rental value (ARV) of less than Rs 1,200 and rented property with ARV less than Rs 720). Properties are tax-free for the first three years they are on the property tax register. Even most katchi-abadi houses would not be exempt from property tax if ARV were assessed on the basis of market rents, but assessed values are not consistent with market values. These and other inconsistencies result, for instance, in only 30% of properties in the newer parts of Lahore being liable for property tax, and only 50% in the Lahore Walled City. As katchi-abadis are improved and regularized, they too should begin to meet some share of the cost of urban services by participating in the property taxation and user charge system. 3.06 A further problem is institutional. Efficient service delivery is often undermined by the fragmentation of responsibility for urban service delivery. Implementation and management are fragmented between development authorities, municipal governments and water authorities. B. Early Attempts at Katchi-Abadi Upgrading 3.07 A large katchi-abadi upgrading and tenure regularization program is being undertaken with World Bank assistance in Karachi in 18 katchi-abadis on 3400 acres of land through the Karachi Special Development Project (KSDP). The project will benefit about 600,000 people. In addition to improvements in basic urban infrastructure and services, the project is intended to provide security of tenure in the form of 99 year leases. As part of KSDP, the Asia Development Bank also is supporting the upgrading of the two largest katchi-abadis on the fringe of Karachi, in Orangi and Baldia. 3.08 To implement the program, the Karachi Municipal Corporation (KMC) has created a new Department of Katchi-Abadis (DKA) responsible for the design and award of contracts, the supervision of construction contracts (except those involving utilities), cost recovery, community development and tenure regularization. Previously these functions had been under separate agencies and coordination had been difficult. - 19 - 3.09 The project initiated the concept of charging for development costs of plots in katchi-abadis. Costs were to be recovered through the sale of 99 year leases. Lease rates were set to permit full cost recovery for on-site infrastructure, related off-site infrastructure and community facilities. Land was priced at Rs 5 per square yard. Although this was considerably less than the market rate, it was intended to introduce the principle of charging for publicly owned land. 3.10 Little of the physical work has been undertaken as yet. Surveys, plans and engineering designs have been undertaken for almost all householders. The lease sale, now raised to Rs 15 per square yard, would have to be further raised to Rs 35 per square yard to achieve cost recovery. Although some 175,000 households would be eligible to take leases upon payment of the lease rate, only 18,000 have actually done so. It would appear than many katchi-abadi dwellers do not feel the need to purchase tenure. Residents have long been protected by katchi-abadi organizers, and recent political statements may also be interpreted as giving residents de facto tenure. C. The Expansion of Katchi-Abadi Upgrading under Government Programs 3.11 Many of these initial developments with katchi-abadi upgrading have now reached national attention. A National Katchi-Abadi Directorate has been established in the Ministry of Housing and Works to monitor and coordinate katchi-abadi upgrading activities throughout the country. Most of the spending authority is located at the provincial and local levels. 3.12 In July 1986, the Government declared that all katchi-abadi residents occupying land as of March 1985 would be given proprietary rights. Occupants would have to pay development charges, but people making payment within three months would get a 502 discount. Katchi-abadi cells have been established in each provincial government to fund and administer the program. Most schemes will be designed by development authorities and local councils. They will include water supply, sewerage, storm drainage, electricity and road surfacing. 3.13 Table 3.4 shows a breakdown of the katchi-abadis in Punjab. The Government of Punjab Katchi-Abadi Cell has estimated that the cost of upgrading the 902 katchi-abadis will be about Rs 700 million. This figure, representing only Rs 478 per beneficiary, is low by the standards of previous upgrading projects in Pakistan (average of about Rs 1000 per beneficiary; Table 3.5), especially when one takes into account that many katchi-abadis in medium towns are unlikely to have the partial infrastructure systems which have limited capital costs in some of the earlier up-grading efforts. In addition, the government intends to provide schools, dispensaries and community centers. Full upgrading is likely to require some Rs 1.5 billion in Punjab alone. 3.14 In 1986/87 the Government of Pakistan provided a grant of Rs 260.8 million for katchi-abadi upgrading in Punjab. Funding for this year is reduced to Rs 60 million. Table 3.6 shows the amounts which the Provinces have budgeted for katchi-abadi upgrading in 1987/88, which are - 20 - Table 3.5: Per Capita Casts of Katchi Abadi Upgrading (All Costs adjusted to 1987 Prices) Rs Per Capita Karachi Katchi Abadi Upgrading (KMC Actual) 850 Upgrading in Lahore (estimate) 500 - 660 Upgrading on Periphery of Lahore (estimate) 750 - 1,000 Estimate for Katchi upgrading in Prime Minister's program 662 Estimates in Punjab Urban Project Lahore 733 Silbot 1,920 Multan 1,066 Table 3.6: Amounts Budgeted for Katchi Abadi Upgrading In 1987/88 (Rs Millions) Baluchistan NWFP Punjab Sind Total 38.7 1.0 60.0 343.9 443.6 - 21 - substantial reductions from 1986/87 levels. It appears that, due to funding constraints, Federal Government funds will no longer be provided as additional to provincial ADP budgets. Much of the relatively large expenditure budgeted for in Sind represents the commitment to support katchi-abadi upgrading under the KSDP. With cities growing at rates of over 4% per year, and the katchi-abadi population increasing at rates of twice this in many cities, the present level of funding for upgrading may only be enough of keep up with the growth of katchi-abadis. 3.15 Slum upgrading in Pakistan and elsewhere has proven a very effective way of meeting the priority needs of low income groups. Unlike many land and shelter development programs, the beneficiaries are readily identifiable as poor. After employment, their priorities are usually public services. The survey in Annex 2 tends to support this conclusion. Slum upgrading programs usually have a high economic rate of return; that is, they provide services which beneficiaries value highly in relation to cost. For all these reasons, a larger program addressing the backlog in improvement and regularization in cities of all sizes would be justifiable. It could also include many areas of cities where residents have legal tenure but lack adequate services. 3.16 A larger community improvement program would be based on the following principles: (a) arrangements would be made for fully transferable and legal tenure (including collaterability) of titled properties before improvement works commence; (b) upgrading would take place after obtaining agreement of the local community and would meet affordable engineering standards; (c) the cost of land would be recovered at or near market prices; (d) the cost of services would be recovered through water and sewerage rates and the property tax system; (e) improvements would keep pace with householders' preparedness to pay for land leases and services; (f) the program would include the strengthening of provincial level directorates of katchi-abadis; and (g) a line of credit would be made available for house upgrading loans secured by the plot lease certificate. - 22 - IV. URBAN LAND AND SHELTER DEVELOPMENT A. Background and Constraints 4.01 As indicated in Chapter II, the inadequate supply of serviced land appears to be the most important constraint to an increased supply of housing. Adequate mechanisms do not exist to convert land from rural to urban residential use. There is also evidence of similar constraints to the supply of residential land in rural villages (see Chapter V). The size of the land development task in urban areas should not be underestimated. Even conservative estimates indicate that by the year 2000 new urbanized areas equal to the present size of Lahore will have to be created every eighteen months to two years in order to accommodate the new urban population. 4.02 In addition to the limited program of housing credit through the HBFC (see Chapter VI), the Government's main initiative in the urban shelter sector has been the development of serviced plots (area development) for housing. Table 4.1 shows the amounts budgeted for area development projects in the four provinces in 1987/88. Fortunately, during the Sixth Plan period, the construction of housing units by government agencies has been largely curtailed which has allowed government agencies to concentrate on what most consider to be the greatest priority: the provision of developed land. Most formal sector urban land development in Pakistan is undertaken by the urban development authorities which have been established in the metropolitan cities as well as in many secondary cities. In the largest cities of Lahore and Karachi, there is also some private and cooperative sector land development. In the last year, some provinces have also established special Three Marla (75 square meters) area development schemes to develop plots for the urban poor. 4.03 The available statistics show several interesting facts about land development. As shown in Table 4.2, the actual numbers achieved fall short of the targets, but it is not unusual for development targets to be overly ambitious. More important problems are that the number of plots actually produced falls short of accommodating urban growth, the r 'ts are not designed to be affordable to the majority of the urban pupulation and a very large percentage of the plots produced are not occupied within a reasonable period. Table 4.1: Development Budget for Area Development 1987/88 (Rs Millions) Baluchistan NWFP Punjab Sind Total 45.1 10.0 604.6 1448.8 2108.5 - 23 - Table 4.2: Allotment or Sale of Serviced Plots Compared with Targets and Performance of Key Agencies (000's) Punjab LDA Sind KDA NWFP PDA TOTAL URBAN Target Target Target TARGET ACHIEVED /1 /2 /1 /3 /1 /4 /5 /5 5th Plan 1978-79 2.6 18.2 0 70 47 1979-80 2.6 0 0 60 51 1980-81 0 10.0 3.5 57 46 1981-82 0 0 0 59 n/a 1982-83 0 0 0 n/a n/a 6th Plan 1983-84 33.8 0 64.9 10.0 10.3 0.4 85 51 1984-85 33.8 0.1 64.9 0 10.3 0.3 100 76 1985-86 33.8 1.4 64.9 0 10.3 8.3 n/a n/a 1986-87 33.8 0.2 64.9 0 10.3 0.2 174 n/a 1987-88 n/a n/a n/a n/a n/a n/a 150/6 7th Plan URBAN & RURAL 1988-89 100 1989-90 100 1990-91 100 1991-92 350 1992-93 350 Notes: 1) "Pakistan Urban Sector Profile" ADB November 1985 Appendix C. 2) Informal notes from LDA, CMP office 1987, however "Housing Shortage During the 6th and 7th Plan Period" - LDA undated CMP memo states 43,000 units provided 1981-88, i.e., 6,142 average p.a. 3) Extract from incomplete copy of document believed to have been prepared by KDA. 4) Compiled from Hayatabad Project files. 5) "National Housing Policy;" National Committee appointed by the Environment and Urban Affairs Division, Ministry of Housing and Works. December 1986, pp. 9, 10 and 24. 6) "Provision of 300,000 Housing Units under the Prime Minister's Programme" (PC-1) National Housing Authority November 1987. - 24 - 4.04 About 76,000 new plots were created in 1984-85, the last year for which data is available. During the same year, roughly 187,000 additional households had to be accommodated in cities. It would probably be more useful to compare plot production with the growth of metropolitan and secondary cities, since most of the development authorities are in these areas. These cities added a new population of about 103,000 households during 1985, a much more favorable comparison with formal sector plot production. 4.05 A more important problem is affordability. Of the large numbers of plots that have been produced by development authorities, a high percentage has not been affordable by low income groups. In Peshawar, for example, no plots smaller than 5 marlas (125 square meters) were available in the Hayatabad Township development for the first five years; the smallest plots have been priced at only about 25% less per marla than the largest plots in the best locations. Even where plots are fairly small, land use and infrastructure standards make plots unaffordable to the poor (see paras 4.09-4.13). 4.06 Table 4.3 shows another serious problem: of the roughly 120,000 low cost plots created since 1970 in Sind, only about 3000 had been occupied by 1984. Plots frequently lie unoccupied for years; this phenomenon is observed throughout Pakistan. 4.07 To address Pakistan's housing problems, there should be a priority effort to increase the supply of affordable developed residential land. Such an effort must address a number of issues: The first concerns the need to use land use and infrastructure design standards and pricing patterns which are affordable to all income groups, including the lower ones. The second concerns the plot allocation process which does not effectively require plots to be occupied. The third concerns the land acquisition process which in many cities is an important constraint to the public sector's land development programs. The fourth concerns the often inadequate systems for planning and providing trunk infrastructure at the city-wide level which constrains both public and private sector development. Finely, there is the system for private land development. It is becoming increasingly obvious that the public sector cannot by itself provide all the developed land required. The private sector must have a greater role. Although little is known about the private sector operates, it appears to be heavily constrained. 4.08 Government does not have the funds to subsidize land and shelter development on a large scale; it is also difficult for government to maintain large subsidized loan programs (see Chapter VI). Thus, if shelter programs are to meet the scale of the need, the capital cost of projects must be largely recovered from beneficiaries, at market rates of interest. This means that the design of new developments must be based on a thorough analysis of how much households are prepared to pay for housing. For example, it has been estimated that in Peshawar low income households with incomes of Rs 900 per month are in about the 20th percentile of the income distribution. They can only afford to pay about Rs 135 per month towards a house (see Annex 3). - 25 - Table 4.3: KDA/HDA Schemes for Low-Income Groups Since 1970 (Plots up to 120 sq. yards) Scheme Year Total no. of Actually Remarks Announcement residential occupied plots as of 10/1984 KARACHI Metroville 1 1974 4.133 approx. 700 Plots sold in 1974 to lower-middle income groups. Hetroville 2 1980 4.379 approx. 2200 Plots used for resettlement of Bihari refugees. Metroville 3 1979 3.200 2 Plots sold in ;979. I4etroville 4 1979/80 3,867 0 Area was encroached upon in early 1980's. Shah Latif Town 1979 43.891 1 15,000 plots sold in 1979. 10.000 in 1981 and 10.000 in 1983 (by ballot). Deh Surjani 1980 47,000 0 Out of 47,000 plots, 20.000 allotted to private developers in 1983 for low-cost schemes with completed housing units. HYDERABAD Qasimabad 1980 1.242 21 Gulshan-e-Shahbaz 1981 11.280 0 60/80 sq. yards plots were allotted on a highly subsidized basis. SLKKUR Gulshan-e-Iqbal 1975 924 0 TOTAL 119,916 2.924 Source: Hyderabad Development Authority - 26 - Table 4.4: Rate of Delivery of Plots in Hayatabad Township. Peshaw U Together wit Oficial and Market Price Levels Year No. Sold/Allotted Prices per Har.a (Rsi VACANT UNDER 1/ 2k lk 1Om Sm TOTAL 2k 1k 1v. 'Zm CONSTR. 1981 397 497 1154 1443 3491 1850 1700 '550 'ie 68% 12% 1982 17 - -1 30 46 1983 -1 - - 8 7 1984 86 172 67 72 397 2875 2530 2050 2000 95% 4% 1985 62 116 72 65 315 1986 772 1781 2529 3284 8366 3500 2900 2600 1987 7 29 24 li 178 4200 3500 3200 TOTAL 1340 2595 3845 5020 12800 Current market prices per marla (Average per Year) 191 371 549 717 1828 13000 10000 10000 k = kanal m = marla Note: Official pricing policy stipulates: "1 kanal plots shall be sold at 20% profit 10 marla plots shall be sold at cost. 3-S marla plots shall be sold with 10% subsidy." I/ As of 1986 - 27 - Table 4.5: Comarison of Alternative Land Use and Infrastructure Standards and Pricina Policies (1) (2) (3) Surjani Town Surjani town Swathi-South non-profit with market Sites & Services Scgnario Karachi Peshawar Mtinimum plot size (m2) 7OM2 70m2 S0m2 Plot price/ household of minimum (Rs) 7.000 9,240 10.000 Percent of households who canr.ut afford 3X 7X 12X No. of minimum sized plots per ha 13.7 13.7 3.1* Plots per ha 29.9 29.9 77.9 Plot area (X) 45 45 70 Facilities (M) 11 11 16 Roads. etc. (M) 44 44 14 Land Cost (per marla) 571 S.000 S.000 Land Development Cost per square meter (gross) 53.7 53.7 89.6 Land Development Cost per square meter (net) 118.33 118.33 128 Price charged for conmmercial land (per marla) 5.300 33,000 n/a Price of largest residential plots (per marla) 5.000 16,216 20,000 *19.3 if 75m2 plots are included. - 28 - B. More Affordable Land Use, Infrastructure Standards and Plot Pricing 4.09 Urban land development planning in Pakistan typically follows a hierarchical superblock pattern. An example of such a plan is shown at a scale of 1:5000 in Annex 3 for the Surjani Town Project being developed by the Karachi Development Authority (KDA). It comprises 4176 residential and 512 commercial plots in a large superblock of about 1 x 1.5 kilometers. Residential plots are sized 70m2, 104m2, 200m2, and 336m2 with eight blocks of flats on sites of 6600m2. Table 4.5, column I summarizes the land use, infrastructure standards and plot pricing in that project. The minimum plot size of 70 m2 is affordable to very low income groups and the project is profit#?ble to KDA in spite of a very low percentage of saleable area (45X), because of tne very low price charged for land (Rs 571/marla). 4.10 Land was acquired very cheaply by KDA some time ago. If a market price had to be charged for land, this project, as designed and priced, would not be affordable to KDA. The regulations of most development authorities limit the amount of markup possible on higher income plots and therefore the amount of price reduction possible on smaller plots (see note on Table 4.3). If the price of the larger residential plots and commercial plots could be raised to market levels (column 2, Table 4.5), the KDA project could be kept affordable to low income groups. The price of commercial land would also have to be raised from Rs 5,300 to Rs 33,000 per marla which may not be marketable. The KDA received much of its land a long time ago. Most development authorities in Pakistan do not have a large reserve of land; they must acquire land at a market price. If land has to be purchased at a market price, standards such as using 44X of project area for roads would not be affordable. 4.11 Table 4.5, Column 3 outlines how projects could be designed to be affordable even when a market price has to be paid for land using the example of a project designed for Peshawar. The amount of road space has been reduced by using a more efficient design consisting of individual modules, or cul-de-sac clusters, along local distributor roads (see Annex 3). This allows 702 of the project area to be used for plots. The designers also took care to design value into commercial and high income residential plots, putting them in good locations on larger roads. Smaller low income plots are placed in less desirable locations in the interior of the project area. The combined effect of intentionally designing valuable areas and using differential pricirg enables the smaller plots to be affordable to very low income groups. In the Peshawar example, it also helped to make materirls loans affordable by low income beneficiaries for house construction. Effective differential pricing does not imply taking from one group and giving to another. It is an accurate pricing system for relatively high and low valued plots. Accurate pricing also helps to eliminate windfall gains to beneficiaries who receive below-market priced plots, and thus it eliminates some of the incentive for speculation (see para 4.15). - 29 - 4.12 As yet only a few Three Marla Schemes have been initiated in urban areas. In Peshawar, three marla plots are being provided by subdividing some of the open space in the Hayatabad project. In Punjab a Three Marla Scheme is being developed at Bhopatian outside of Lahore for low income provincial government workers. This latter scheme raises problems because of its distance from the city. It is unlikely that the urban poor would be able to commute such a distance to work. Indeed, the scheme is undersubscribed even though the reserved plots are to be sold for only nominal amounts, Rs 1739 per marla. The Government of Punjab can afford this type of scheme in financial terms because it already owned the land but, since very little land remains in government ownership, the program is not replicable. In economic terms program design should not depend on land ownership. Land should be used in accordance with its opportunity cost, which is best reflected in the market price. The Bhopatian scheme uses the same uniform planning and pricing approach described above. It would itot be affordable if a markct price had been paid for land without modifying the planning and pricing. 4.13 A particularly innovative scheme in Hyderabad has reduced cost to low-income households by iaitroducing incremental service development. The initial, development included only street layouts and community services. As the community develops, residents are contributing towards the provision of individual services, street surfacing, etc. C. The Plot Allocation Process 4.14 In most schemes plots are allocated by "ballot" or random selection of beneficiaries, and very little differential pricing is allowed under existing regulations. For example, in the Hayatabad Township, Peshawar, larger plots can be sold for only 20Z profit; 10 marla plots must be sold at cost, and 3-5 marla plots can be sold with a l0X subsidy. In fact, since the cost of land is calculated on the basis of the price paid at the time of acquisition, there is likely to be a large difference between market price and sales price, especially if the land was acquired some years back. Table 4.4 shows, that in the case of Hayatabad, even the one kanal plots with 20Z mark-up are selling for only 322 of market value. 4.15 With such a large potential gain, there is little wonder that large numbers of applicants enter the "balloting" process. Althougn applicants may be screened to verify income, low income participants who are allotted plots have a large incentive to sell the plots to land speculators and realize an immediate capital gain. Many of the middle and higher income applicants probably have little intention of building houses on the plots, and, although there are sometimes covenants requiring owners to occupy plots within a given time period, these are rarely enforced. Similar to the Sind figures, the data in Table 4.4 for Hayatabad shows that 68Z of plots alloted before 1983 were vacant in 1986. 4.16 There are several reasons for holding vacant plots for such long periods. In an inflationary economy serviced land -nay be preferable to - 30 - many alternative investments. The land may be a form of security for many families who intend to build at some later date or to make it available to their children. There is no property tax levied on vacant land; capital gains on the sales of vacant land are not taxable. In addition, because financial markets are underdeveloped, there are limited investment alternatives. These factors are critical in making land investment a haven for illegal or "black" money, muc" of which is said to be invested in vacant land. 4.17 Other factors constrain landowners from building housing for rent. In Punjab, for example, rent control was established by the Rent Control Ordinance of 1959. The ordinance does place limits on the amounts of rent a landlord can charge. More seriously, however, it constrains the landlord's ability to repossess or evict a tenant. Even where there is just cause such as non-payment of rent or damage to the house it can take 15 months or more to evict a tenant and the legal costs are high. This appears to be a serious constraint to the house rental market and housing supply. 4.13 Several actions could help to reduce the large resource loss implied by vacant serviced plots. The more accurate pricing of plots would reduce the incentive of windfall capital gains to potential land speculators. Morn effective covenants could be enforced requiring plot occupation and house constrtction within a reasonable time. The very successful low income area development project in Hyderabad is demonstrating that this is possible (see para 4.13). After only one year, there are already 2000 families on site and building a new community. A tax on vacant urban land should also be seriously considered. This is justifiable because of the infrastructure investment which must be maintained. It would also encourage more efficient urban development; landowners not wishing to build would be encouraged to sell to those who do. Rent control legislation and particularly the eviction process should be reviewed to minimize the restrictions which are a disincentive to investment in rental housing. D. The Land Acquisition Process 4.19 Land acquisition in Pakistan on the periphery of urban areas is expensive, particu.larly in Punjab and other areas where land has a high agricultural potential. Prices of Rs 5,000 per marla (Rs 200 per m2 or about US$ 45,000 per acre) are not uncommon. The prices are high, beyond the value of the land from agricultural use. Under the existing Land Acquisition Act, public agencies must reimburse landowners on the basis of market value. In order to make urban land more affordable, there have been attempts to limit the amounts that could be paid to landowners. In addition, in the Punjab until 1985, there was an interesting experiment where the original landowners were given plots of developed land as part of their compensation. The law was not considered successful and Punjab reverted to the Land Acquisition Act of 1894 which uses market value as the basis for compensation. 4,20 Within the existing framework there are probably ways to improve the definition of "market value" and to improve the acquisition process - 31 - itself so that it takes less time (see Annex 4). If market value is to remain the basis of most land acquisition, it is all the more imperative for public agencies to begin using the more efficient planning and pricing techniques discussed above (paras 4.09-4.11). E. The System of Planning and Providing Trunk Infrastructure 4.21 Even where land is available for development and if development standards were more affordable, there are often long delays in the provision of the required trunk infrastructure which frustrate public and private sector development and leave large areas of cities unserviced. The present system of master plan preparation takes a long time and is difficult to implement. The system tends to be of little relevance to the urgent task of providing infrastructure. 4.22 The plans emphasize the long term use of land fu: various purposes and the provision of trunk infrastructure (main roads, trunk sewers, water mains, drainage mains, etc.) based on population projections. The establishment of land use and infrastructure standards is a physical planning exercise. Standards are set in accordance with established criteria, but cost or budget constraints are seldom taken into account. This is due usually to the lack of interaction between physical planners and engineers, on the one hand, and economists and financial planners on the other hand. 4.23 A preferable approach would be to design Outline Structure Plans, based on the best data available, outlining a framework for urban expansion over a 20 year period. This would include a main road and trunk infrastructure network together with broadly defined residential areas. Detailed expansion areas would be designed ~-n relation to the Outline Structure Plan showing five year growth increments. This would leave flexibility for dealing with the growth areas as the city expands. Infrastructure would be designed based on improved interaction between physical planners and financial planners. Physical plans would thus be related to actual development budgets. 4.24 This type of planning process has been adopted in Lahore and Peshawar, to guide future expansion and at the same time expedite project design and execution. The approach is soon to be used in six secondary cities of NWFP. Although the Lahore Structure Plan has not been amplified or improved since it was prepared eight years ago, it has proved to be a valuable tool for coordinating the infrastructure development programs of the Lahore Development Authority, the water authority and the private sector. F. The Need to Encourage Private Sector Development 4.25 Even with some of the improvements outlined above, it would be difficult for the public sector to supply all of the developed land needed. Private developers are active mainly in Karachi and Lahore, and they deal mainly in upper and middle income land development and house construction. Private developers face a number of hurdles in getting the - 32 - necessary permits. They must demonstrate clear title to land, which may be difficult, they must get approvals from development authorities that proposals conform to standards (this can take up to 18 months, and existing standards may preclude development for all but the rich), they must advertize to invite public objections, and they must get no objection certificates from a number of government agencies. Constrained by the expensive standards and faced with numerous bureaucratic procedures, it is not surprising that only illegal private developers operate for the low income market. 4.26 It would be important to review the many constraints, high standards and complex procedures faced by private developers. Unnecessary constraints should be eliminated and standards made more affordable so that private developers can make a larger contribution to the supply of developed land. Likewise, it is important to understand better the quasi-legal and illegal development process. It is serving an important share of the market. With better understanding and some modification and assistance, it could do so more effectively. This may be especially true in many small and medium towns. G. A National Land and Shelter Development Program 4.27 The newly established National Housing Authority (NHA) is in a strong position to lead a national program to develop more efficient area development schemes with affordable standards and pricing and improved allotment procedures. Schemes would be planned and implemented at the local level. A parallel home-loan program, through a financial intermediary such as HBFC (see Chapter VI) could advance loans for home construction to plot recipients. The NRA would be a source of funding and technical assistance, and it would monitor program implementation. - 33 - V. RURAL HOUSING A. Introduction 5.01 As noted in Chapter II, rural housing conditions appear to have gotten worse during the past few decades. The number of persons per household increased from 1961 to 1981 from 5.4 to 6.4 persons. A more important indication of housing market tightness is that the number of married, widowed and divorced persons per household increased even faster in rural areas than in urban areas (see Table 2.4). 5.02 There is much talk of rural "shelterless" families, but few families are completely shelterless. Overcrowding, the poor quality of housing, and the lack of secure tenure in many rural areas appear to be the more immediate problems. In 1980, 59% of rural households lived in shelters with walls of unbaked brick and mud finish, compared to 18Z in cities (see Table 5.1). A further 25Z of rural households lived in houses with walls of baked brick or stone with mud finish. Many houses in both groups are likely to need improvement. Eighty-eight percent of rural households had no access to potable water. Lack of access to adequate sanitation is also widespread. 5.03 The fact that the rural housing market appears to have tightened during a period when per capita incomes in rural areas have increased indicates that some of the same types of constraints are probably operating in rural as in urban housing markets, particularly land market constraints. The increase in overcrowding in rural areas and the poor condition of much of the rural housing stock suggests that all or some combination of the following is happening: (a) home improvement and extension is inhibited by the lack of secure tenure; (b) existing plots have no more space for home extension; (c) it is difficult and expensive to acquire more land for new houses or plot extension; and (d) the supply of affordable building materials may be limited. There is evidence to suggest that all of these factors are playing a part. B. Possible Constraints in Rural Housing Markets Land Tenure 5.04 In 1980, 17Z or 1.5 millicn rural households did not own the land upon which their houses were built. Other estimates have placed this number at 26% or 2.5 million households (see Table 5.2). Many Punjab villages, for example, are situated in "Abadi Deh," areas which have long been designated as villages but which remain legally the property of estate owners. There is some controversy about whether the landlords or the tenants own the houses in these areas. Households who lack secure tenure may be hesitant to improve or expand their houses. Recent increases in agricultural mechanization and a corresponding decrease in the demand for labor may compound the insecurity of tenure felt by rural tenants. - 34 - Table 5.1: Rural Housing Conditions (Z of households) Urban Rural Outer Walls Baked Brick/Stone with Mud Finish 36 25 Unbaked Brick with Mud Finish 18 59 Roof Guder/Beam, Wood, Baked Brick, etc. 62 85 RCC 26 1 Tenure Conditions Owner-occupied 68 83 Rented 22 2 Rent Free on Others' Land 10 15 Access to water and Sanitation Access to Potable Water 77 22 Waterborne Waste Disposal 48 4 - 35 - Table 5.2: Landless Rural Households* (1980) 2 (thousands) Total Rural Households 100 9,623 No Land Owned 26.2 2,521 Own Less than 0.4 ha 32.0 3,079 (i.e. with little effective agricultural land) Total Landless 58.2 5,600 Source: Khan, Mahmood Hasan (1987) "Rural Poverty in Bangladesh, India, and Pakistan: Profiles and Policies" Pakistan Institute of Development Economics, Islamabad. Table 8, p. 33. * 6.44 persons per household Table 5.3: The Availability of Land foi Seven Marla Schemes in Baluchistan No. of Districts Hectares State Land Available 12 1,458 Private Land Identified 6 613 No Land Available 3 Schemes Underway in 5 TOTAL 2,071 Total State Land Available 1,458 Plot Area (say 602 of total) 874 No. of 7 Marla Plots possible (*) 49,000 Total Number of Plots to be Distributed by 1990 (**) 110,000 Shortfall 61,000 (65% of Target) (*) 1 Marla = 25.3m2; 7 Marla = 177.1m2 (**) Target for Pakistan: 2.2 million Target for Baluchistan: 52 of total 110,000 Plots. Source: Calculated from information provided by Government of Baluchistan, December 1987. - 36 - On-Plot Space 5.05 Since rural plots are generally bigger than urban plots, one might naturally assume that rural houses could be more easily expanded. However, rural households may have more need for space, especially for the storage of grain, animals, etc. This fact and the observation of many rural villages suggest that the available space for house expansion may be more limited in rural areas that might be expected. Land for New Housing 5.06 The availability of land for new hoi g is not usually thought to be a major constraint for rural shelter d,.-elopment. However, the limited available information on rural land prices, ownership, and tenure suggest that rural land may not be as easily available for shelter as one might assume. Irrigated land, especially in Punjab but also in the more densely populated parts of other provinces, has a high value and a high price. Prices are also said to have been escalated by remittances from migrant workers who have a high propensity to buy land and because of its investment and social value. Rural land ownership is highly concentrated; the power and prestige traditionally associated with large landholdings make landowners reluctant to sell. Wealthy landowners have little economic need to sell. In Baluchistan, complex traditional land and water rights and obligations between leaders and the community make individual land sales difficult. Construction Materials and Technologies 5.07 Many urban building materials such as cement and steel have increased in price and are difficult for the rural poor to afford. At the same time, many traditional building materials, such as wood for roof construction and agricultural waste-fired "bhatti" bricks, which would otherwise be used in rural areas, also appear to be expensive and in short supply. Part of the problem may have been a lack of emphasis given to local construction materials manufacturing since they are not used in government construction. The present housing construction and loan program (see paras 1.04-1.08) is being undertaken by government engineering departments which are not likely to accept the use of traditional materials which would give an impetus to their production. 5.08 The above points tell only part of the story. More information is needed about rural land markets and the availability of building materials. For these and possibly other reasons, however, land appears to be unavailable and/or unaffordable to many of the rural poor. This constraint is also reflected in the difficulty the government is having in obtaining rural land for Seven Marla Schemes (see para 5.09). - 37 - C. Present Rural Programs 5.09 As part of the 1985 Five Point Program, the Government launched a program to create 2.2 million 7 marla (175 square meters) plots for allotment free of charge to landless rural families between 1986 and 1990. In Punjab, 550,000 plots have already been allotted. A similar Five Marla Scheme was undertaken by a previous Government in 1975 under which about 600,000 five marla (125 square meter) plots were allotted to rural households in Punjab before it was discontinued. 5.10 The new program is making land available to a large number of households, but it has several drawbacks. For budget reasons, the land already allotted was already in Government ownership. In many cases, village common (Charagah and Baquia) land is being distributed. It is not clear whether this land is appropriate for residential purposes or whether it is required for other purposes. Some of the sites being subdivided appear to be far away to be logical extensions of existing villages; some may be in flood prone areas. The intended beneficiaries may have difficulty settling there. 5.11 The designs of these schemes have typically used a simple grid pattern. More efficient land-use, such as narrower roads and cul-de-sac patterns, might have contributed to lower per unit development costs. Very little infrastructure is provided, raising, in particular, questions about possible flooding, sanitation and water supply. 5.12 The largest problem facing the program is its lack of long-term replicability. With the supply of government owned land almost exhausted, there is increased pressure to use government land in less appropriate locations. Land acquisition will be necessary but costly. It is estimated that land for the 500,000 plots required to complete the program in Punjab would cost Rs 2.5 billion which would be difficult to finance, especially since there is little cost recovery element in the program. Punjab's. 1987/88 budget for seven marla schemes is only Rs 166 million (see Table 5.3 for example of Baluchistan). D. Developing a More Sustainable Rural Housing Strategy Land Tenure 5.13 More information is needed about rural land ter.ure and how the lack of tenure may limit households' willingness to improve their houses or their ability to use houses as collateral for loans. Possible ways of normalizing tenure should be explored as part of any rural shelter effort. - 38 - Land Development 5.14 Given the lack of access of the poor to rural land, the objectives of the Seven Marla Program are basically sound. However, the design of a more sustainable program would require a better knowledge of the market for non-urban residential land. For example, how do villages normally expand? How is land subdivided? How does it exchange hands and at what cost? What measures would be necessary to complement local efforts to develop rural residential land? Such information may enable the design of a program which facilitates the normal expansion of villages, rather than replaces it, and which recovers much of the cost. The availability of infrastructure and services should also be considered but these would best be resolved in the context of broader rural services programs. Services such as water supply, sanitation and storm drainage may be higher priorities than housing structures. House Construction 5.15 The Government's latest housing credit programs (see paras 1.04-1.08), which include rural areas, have evolved from direct construction to programs of loans to plot holders for house construction, but even more flexibility is required. Instead of requiring construction to standard specifications, it would be better to develop low cost specifications for critical components of a housing unit such as on-plot sanitation, rain-resistant plaster, inexpensive but durable roofs, smokeless and efficient stoves, etc. Such specifications could be made to correspond to each step of traditional incremental construction or improvement of housing units, and information could be made available through local government personnel. Building Materials 5.16 Rural materials industries have been given little emphasis in the recent past. Assistance to the development of low cost materials would help rural households to cope with rising construction costs. Limestone, plentiful in parts of Baluchistan, could be a cheap substitute to cement in mortar and plaster. Fast growing trees in village woodlots may meet both roof construction and fuel needs. Improvements to agricultural waste-fired "bhatti" brick kilns, widespread in PuLijab, may provide a more economical alternative to oil-and-coal fired bricks, with little loss in quality. The Government's overall rural program implies a large commitment to construction, not only in housing but in schools, health centers and other government buildings. It would be important to review the specifications for these buildings to ensure that feasible opportunities to provide an impetus to rural materials production are not overlooked. Institutional Development and Rural Housing Finance 5.17 The institutional arrangements for the recently announced and proposed Government housing credit schemes which would be spread throughout rural areas have a high degree of political involvement, especially in beneficiary selection and loan disbursement (see paras 1.04-1.08). Although some degree of involvement by local government - 39 - personnel is desirable, especially for disseminating information, a more objective loan organization and disbursement system would be important to ensure proper utilization and recovery of loans. 5.18 The Agricultural Development Bank of Pakistan (ADBP) has been mentioned as a possible conduit for rural housing loans. It has more than 10 years of experience in rural credit programs. By the end of FY87 it had 36 regional offices and 235 branches throughout Pakistan. It also operates a team of 1259 Mobile Credit Officers (MCOs) who visit customers in their villages, providing personalized banking services. ADBP attributes both high repayment rates and the successful utilization of loans to the system of MCOs, which reached over 39,000 of the 47,600 villages in Pakistan during FY87. The Bank's recovery ratio since inception to the end of FY87 was 922. 5.19 In FY87, over 65% of the 98,000 loans made were less than Rs 25,000 in value. Although virtually all loans were secured (only 0.42 were personal surety loans). The ADBP is now implementing a pilot program of small loans to the landless poor, rural women and small farmers for income generating schemes. These are offered with some relaxation of normal banking procedures. Loans under the program, which is part of the Five Point Program, can be on the basis of personal surety (maximum amount Rs 25,000 for a maximum period of 18 months) instead of, or in addition to, being secured on real property. The ADBP should be considered as an intermediary for rural housing loan programs in order to achieve greater objectivity and distance from. the political process. 5.20 Effective rural credit schemes would have to be more flexible than urban housing loan programs. Loan amounts, and therefore repayments, would probably have to be low and focus on materials that beneficiaries cannot easily obtain or construct. Thus, they may be targeted to priorities such as on-plot sanitation, corrugated sheet roofing, etc. Small amounts repaid over fairly short periods on somewhat flexible terms are what most rural householders are used to. A program of small, well targeted loans may be the most efficient way to assist the present pattern of self-help incremental construction. E. Program Design 5.21 Any rural housing credit program should be approached cautiously. It would be best to begin on a pilot basis in areas where rural incomes are higher and there is a larger chance of success. The present program which seeks to cover all of Pakistan in the first phase is likely to encounter implementation and/or loan recovery problems. 5.22 Loans should be flexible, limited and well targeted. They should allow for flexible repayment in accordance with households' ability to repay. A large amount of public education would be required, and this could well require participation of local government officials. However, beneficiary selection and cost recovery should be the responsibility of a financial intermediary such as the ADBP. - 40 - 5.23 To complete the design of any rural housing credit program, several questions should be answered through further research: with the supply of government land almost exhausted, how can land best be made available for rural residential purposes? How do rural land markets function, and how can they be made to function better for the benefit of the poor? How can landless rural households gain tenure to the plots they occupy? Are rural public services more important than housing improvements and, if so, would a rural services upgrading and tenure normalization program (a type of rural katchi-abadi improvement), possibly with some well targeted housing loan component, be more appropriate? What would be the most important intermediary for housing credit in rural areas and under what terms? - 41 - VI. HOUSING FINANCE 6.01 In 1986/87, credit from the House Building Finance Corporation (HBFC) financed only about 14% of gross domestic capital formation in Housing (see Table 6.1), and in recent years the combined credit allocation for housing by all the nationalized commercial banks has been only about 10% of HBFC'S total allocation. Thus, formal sector housing credit, in addition to being very limited in relation to total housing investment, under 202, is dominated by HBFC which is the only specialized housing finance institution in Pakistan. Most housing is financed directly from private savings and from iniormal borrowing, about which very little is known. 6.02 Two issues characterize HBFC's recent program. First, operating in an environment of high nominal and real interest rates, HBFC has maintained a low interest rate structure, lending partly at negative real rates of return, which has made housing affordable to a range of beneficiaries. This has been possible only because most of HBFC's borrowing from the State Bank of Pakistan (SBP) is at highly subsidized rates. This dependence on subsidized borrowing has severely constrained the size of HBFC'S program. Second, there appear to be problems with targeting. Although HBFC has not compiled data on the income levels of its borrowers, high standards and eligibility requirements appear to limit the number lower income beneficiaries. A. The House Building Finance Corporation (HBFC) 6.03 HBFC was established in 1952 as a specialized housing finance institution. Its head office is in Karachi; there are 58 district offices under the supervision of 7 zonal offices. HBFC's authorized and paid-up capital is Rs. 200 million, all of which is owned by the Government of Pakistan and the (SBP). The Corporation is managed by a Board of Directors with eight members consisting of a Chairman who is also the Managing Director, two representatives of the Federal Government (Ministry of Finance and Ministry of Housing and Works), one from the SBP and one from each of the four provincial governments. 6.04 Source of Funds. HBFC receives most of its funds as extensions of credit from the SBP. It raises no funds in the market either from borrowings or deposits. Debentures guaranteed by the Federal Government and sold by HBFC to financial institutions used to provide a small percentage of HBFC's funds, but this was discontinued after 1979. 6.05 The allocation of credit from the SBP is subject to a limit fixed by the National Credit Consultative Committee which was Rs 1,450 million for 1987-88. Loans from SBP are repaid in equal installments over 10 years. Prior to the introduction of Profit and Loss Sharing (PLS) under Islamic Banking in 1979, nominal interest earned by HBFC from borrowers averaged 92-102 and HBFC paid interest on loans from the State Bank of Pakistan of about 8.52. Since 1979, the PLS formula used by HBFC yielded - 42 - Table 6.1: HBFC Credit as a Percentage of Housing Investment Total HBFC credit as Credit by Investment 2 of Housing Year HBFC in Housing Investment (Rs. Millions) (Rs. Millions) 1975-76 413.3 1341.2 30.8 1976-77 340.2 1709.0 20.0 1977-78 416.5 2034.6 20.5 1978-79 664.4 2272.9 29.2 1979-80 882.2 3003.0 29.4 1980-81 1058.1 3850.3 27.5 1981-82 1444.8 4501.0 32.1 1982-83 1265.6 5898.5 21.5 1983-84 1695.4 6395.0 26.5 1984-85 1496.4 6940.0 21.5 1985-86 1624.1 7705.0 21.1 1986-87 1168.6 8130.3 14.4 Table 6.2: NBFC Financial Position Statent BALANCE SHEET AS OF 30 JUHE 1986 IE 19sti Rupees 198S Rupees T 1906 RUDees 198S Rupees CAPITAL AND LIABILITIES PROPERTY AND ASSETS Authori2ed CASH AND USAK BALANCES 10 99.002.641 73.833.931 40 share of Rs. 5.000.000/each 20ci 0.0 LoO 200.eo00000 BANK ACCOUNTS 11 35.563.976 (1.256.976) Issued, subscribed and paid-up 3 200,000,000 200,000.000 INVESTMENTS - unquoted 12 11,100.000 11.100,000 RESERVES AND SURPLUS 4 37,024,192 41,770,918 INVESTMENT UNDER RENT SNARING SCHEME 7.178.121.017 6.089.983.059 LOAHS S 9.008.999,041 8,067,423.054 LOANS - secured 14 2.084,537.233 2.212,479,961 LOAN FROM INTERNATIONAL DEVELOPMENT ASSOCIATION 10.1 100$000.000 100.000.000 LOAN - Lahore Development Authority 10.1 94.000.000 94.000.000 OTHER LIABILITIES AND PROVISIONS OTHER ASSETS Accrued interest 6 160.768.060 184.819,633 Work-in-progress/ 1S Creditors and provisions 7 12.462.252 10,875.838 housing projects 75,702.101 86.170.036 Taxation - net 20.001.995 14,499.818 Advances, deposits & prepayments 16 63.972.399 48.039.934 Accrued interest 17 6.587,049 5.245.973 Share of profit payable to State Bank of Pakistan 8 442.247.600 382.068.119 Stock of stationery and forms 306.176 290.0s5 63S.479.907 592.263.403 146.S67.725 139.746.002 CONTINGENCIES AND COMM ITMENTS 9 - - FIXED CAPITAL EXPENDITURES fixed Assets - book value 18 19,796.S18 20,407,769 10.061.503.140 9.001.417.i75 Capital Work-in-progress 19 127.S99.I O4 1t01926.So3 147.396.422 122.334,272 NET ASSETS IN BANGLADESH 20 265.214.126 259.227.126 P.R.R.F.C. (LIABILITIES) 21 19,.627.601 19.627.601 10,061,503.140 9.001.447.375 P.R.R.F.C. (ASSETS) 21 19.627.601 19.62761.6 - 44 - an average return of only 61 to 7% from borrowers, although the system has recently been revised and the average return is now expected to be closer to 81. Under the PLS system, the amount of profit paid by HBFC to SBP is based on the profit HBFC earns, and recent profits paid to SBP have been equivalent to a return to SBP of 32 to 3.51. 6.06 HBFC's legislation gives it fairly wide borrowing powers. With the approval of the Federal Government, it may borrow or issue bonds or debentures but it currently obtains funds only from SBP. HBFC's Board of Directors has recently requested management to explore alternative ways of raising funds, but this is unlikely to be possible as long as returns paid by HBFC are less than returns from alternative investments. 6.07 Lending. HBFC finances housing units of up to 2250 square feet. It lends up to Rs. 90 per square foot for houses and RS. 80 per square foot for apartments in multi-storied buildings with a maximum loan of Rs. 200,000, recently increased from Rs. 150,000. Since housing can cost two to three times these amounts, the loan to value ratio decreases for larger, more expensively finished housing. Applicants' eligibility for borrowing is determined by calculating their repayment capacity which can go up to 50X of income. An applicant must have clear title to a plot of land. 6.08 The average loan size in 1986-87 was about Rs. 77,000. Housing standards are fairly high in Pakistan, and HBFC only lends for housing which meets the established standards. The average loan of Rs. 77,000 typically represents 40-45% of the cost of the average unit. Thus, the cash contribution would have to be of the order of 's. 95-100,000 which effectively excludes many households from the program. Housing costing as little as Rs. 30,000 to Rs 45,000 is being financed in Karachi with loan amounts of Rs. 22,000 to Rs. 30,000. 6.09 In 1986/87, only 15,148 loans were sanctioned for Rs. 1,169 million, down from an average of around 20,000 loans in previous years. Loan approvals for 1987-88 are expected to be back up to the 20,000 level. There have also been small separate lending programs, such as a special flood relief program. HBFC has agreed to lend for home improvement in the old city of Lahore under a World Bank assisted project (a total of Rs 20 million), but this work has not yet been initiated. HBFC is participating in the FY88 low income housirg program under the Prime Ministers Five Point Program. 6.10 After 1979 under the PLS system, borrowers or "partners", as they are called, were charged a percentage of the "profit". This was defined as the estimated rental income of that part of the house which corresponds to the HBFC loan (i.e. the percentage of total house cost financed by HBFC times the estimated rental income). Monthly installments of principal repayment were added to the profit charge and both were collected over a 15 year period. HBFC has very recently revised its method of calculating repayments so that all monthly payments will be equal over the life of the loan (similar to a level payment mortgage). To calculate repayment schedules, it has grouped borrowers into four income groups with an - 45 - additional four categories related to unit size, four to loan amount and four to areas of the country. The PLS lending formula also provides for the recapture of part of the capital gain upon resale of the house, but, since there are few resales before loans are paid off, this is in fact not done. Returns to HBFC are also reduced because no "profits" are charged for periods of up to 18 months on construction advances because there is no rental value to the borrower. 6.11 Before the introduction of PLS, interest was charged on a sliding scale from 8% to 12% (average of 9%-102), depending on the size of the loan. After the introduction of PLS, the average nominal return dropped to 6%-7%. However, under HBFC's revised method of calculation the nominal return now ranges from 6.752 to 12.5% depending on the category of the loan. Thus, although there are recent improvements, while Islamization of the financial system has tended to introduce a more market-oriented pricing of capital, it has led to lower nominal rates of return by HBFC, even negative real rates in some years. 6.12 Financial Results. HBFC's profit and loss statements show a deteriorating financial situation (see Table 6.2). Before tax profits have declined from Rs 11.8 million in 1983 to Rs. 755 thousand in 1986. As HBFC's portfolio becomes increasingly dominated by PLS loans (over 802 of its outstanding loan balance is now in PLS loans), net profits could continue to decline. Instead of interest, HBFC is charged by SBP on the basis of its own estimated profit. The profit sharing formula used to calculate the amount payable to SBP allows a smaller mark-up than under the previous interest bearing loan system. HBFC's management is concerned that as the pre-1979 loan portfolio is phased out it will cease to earn the level of surpluses it did previously. 6.13 The above financial trends are compounded by other factors. HBFC has a large number of outstanding loans one month or more delinquent. About 99X (Rs 730 million) of loans made under a flood relief program are in default. At present, there is no available breakdown of delinquent loans by age. HBFC does have foreclosure powers under its own Act and under the Land Revenue Act, and it is instituting a more vigorous program of pursuing delinquent borrowers. For example, recent enforcement measures in Islamabad were followed by increased payments. Despite a reluctance to press as hard as possible under the law, management is confident that about 90% of delinquent debts will be recovered eventually. Under the PLS system, a percentage of the accrued profits must be paid by HBFC to SBP. Since bad debts are not yet being written off, profits must be shared with SBP whether or not they are collected. This has contributed to cash flow problems. 6.14 Cash flow problems are also caused by uneven maturities in HBFC's assets and liabilities. Loans to HBFC by SBP carry 10 year maturities; HBFC lends to beneficiaries for 15 years, so HBFC recovers principal more slowly from beneficiaries than it must repay SBP. To cover the difference, SBP deducts the amount due and unpaid from each year's credit allocation. In 1987-88, for example, there is an authorized line of credit from SBP to HBFC of Rs. 1,450 million, but Rs. 400 million will be deducted from this amount by the SBP to cover unpaid principal. - 46 - 6.15 Given HBFC's total dependence on concessional government credit and its cash flow and potential profitability problems which limit the net amount available for housing finance, it is not surprising that it is able to meet only a small percentage of demand. In recent years HBFC has run short of funds and had to suspend new loan sanctions several months before the beginning of the next financial year. B. Other Housing Finance 6.16 Commercial Banks. Commercial banks include the five nationalized commercial banks, 17 foreign commercial banks and four specialized banks which together form the group of scheduled banks. The nationalized commercial banks dominate the banking system with over 90% of deposits, 85% of advances and 82% of commercial banks' after tax profit. Funds allocated by the National Credit Consultative Committee in 1987/88 to commercial banks for housing credit to the general public amounted to Rs. 190 million compared to Rs. 1,833 million for HBFC. Sums advanced are in excess of Rs. 150,000 but less than Rs. 300,000, presumably to reach a higher level than HBFC. Loans are also given on somewhat more expensive terms. The commercial banks recently dropped the PLS formula for housing because it was too complicated and now lend on a "mark-up" basis. 6.17 Deposits With Builders. An interesting symptom of the lack of formal financing mechanisms for housing is the growth in the use of advance deposits by builders. Under some schemes, builders accept deposits in monthly installments for up to five years before households are eligible to occupy housing. The fact that households are willing to leave large amounts on deposit without interest for long periods in an economy with a high opportunity cost for capital is symptomatic of both the high demand for housing and the lack of other financing mechanisms. These schemes depend heavily on the credibility of the builder and connote some risk since a type of fiduciary responsibility is created with very little regulation. Employer Loans and Allowances 6.18 Many employers offer some form of assistance to employees to obtain housing. The Federal and provincial governments make house building advances to employees who have at least ten years of service. The maximum loan is equivalent to 36 months' pay with repayment over twenty years. Loans are interest free to all non-gazetted civil servants; they carry market rates for higher level officers. 6.19 Private companies frequently make housing loans available to their employees. Data are not available showing the aggregate of such loans in the country. However, a 1983 survey indicated that loans of up to Rs 150,000 were available, although amounts of Rs 50-60,000 were more typical, and they are usually repayable in 10 to 15 years. Loans are typically interest free or at highly subsidized rates for lower level employees and at more market level. - 47 - C. Strengths and Limitations of the Existing Housing Finance System 6.20 Strengths. HBFC dominates the existing housing finance system. In listing the limitations of the existing system, two factors must be borne in mind: (a) HiFC has been a participant in several government-sponsored and socially oriented programs such as lending to flood relief victims where cost recovery has been problematical; and (b) foreclosure on delinquent loans is difficult. Some of HBFC's problems might not have occurred if it operated as private institution. However, despite the constraint of operating as government institution, it has developed some strengths: (a) Although not as extensive as the commercial banks, HBFC does have network of 58 branches throughout the country, particularly in urban centers. It has recently concluded arrangements with the Muslim Commercial Bank to handle disbursements and to receive payments from borrowers, thus extending its coverage; (b) HBFC has competent staff trained in all phases of residential lending; (c) HBFC has well developed lending procedures which, although they could be improved, operate reasonably effectively; (d) HBFC has expanded its computer capability and is developing improved systems for loan servicing and management information. Limitations a. The System Constrains Overall Housing Investment 6.21 The Government controls the allocation of credit and it has used this tool to ration credit away from housing investment. Given other priorities and the difficult fiscal situation in Pakistan (see para 1.19), it would be difficult to argue that Government should allocate more financing for housing. However, the present rapid urbanization, increasing per capita income and poor housing conditions imply a large demand for housing and potential high rates of return in housing investment. While it is impossible to say what portion of informal financing comes from household savings, intra-family loans or high cost borrowings on the informal market, the use of privately placed loans, accumulated savings in low-yielding financial assets or asset liquidations (e.g. jewelry) is at high opportunity cost. Given the limitations in govenmment funds for housing, any expansion in formal housing finance would have to be funded from private savings, but the existence of a large heavily subsidized housing program financed at negative real rates operating in a broad range of market segments constitutes almost insurmountable competition for market oriented housing finance. - 48 - b. The Present System is Costly 6.22 Formal sector housing finance is at great cost to the Government. Net disbursement by the SBP to HBFC was Rs. 1,400 million during FY 86-87 and is budgeted for Rs. 1,050 in FY 87-88. Since these are advances of credit and not direct government expenditure, the cost is obscured. It does not figure in the government budget deficit, but it is equivalent to about 3.52 of the budget deficit. Loans to non-bank financial institutions such as HBFC constitute a significant share of total loans by SBP to financial institutions (36% as of December 1985), and loans for housing finance (mainly to HBFC) constituted 732 of SBP credit to non-bank financial institutions. The return to SBP from its advances to HBFC has been 32 to 3.5% at a time when the marginal cost of domestic borrowing by the Government is 152 to 16%. 6.23 This very costly pattern of housing finance continues at a time when Pakistan's economy is under increasing strain. The Sixth Five Year Plan (FY 1984-88) established yearly targets for growth in the money supply of 13X-14% under the assumption that 6.5% would be accommodated by real growth in GDP and there would be an average annual inflation rate of about 6.52. Growth in money supply has roughly met the target in spite of rapid increases in credit extended to government (14% p.a.), public enterprises (13% p.a.) and the private sector which includes non-bank financial institutions (21%), because of the contractionary effect of a substantial decline of net foreign assets. Because Pakistan cannot continue accommodating credit expansion by running down foreign assets, reductions in credit expansion are likely to be necessary. Pressure for reduced credit expansion is likely to fall upon non-bank financial institutions like HBFC with the expectation that they should tap the private savings market. c. The Present System is Inequitable 6.24 A privileged few receive very low cost housing credit while the vast majority of housing is financed in the much higher cost informal sector. Although some lower income households benefit, many are effectively excluded because the housing financed by HBFC must meet high official standards, because borrowers must have clear tenure and because the borrowers must make large initial investments. d. The System Encourages Inefficient Use of Real Resources 6.25 Highly subsidized housing finance, in combination with other subsidies, particularly subsidies in the price of developed land, may be affecting the structure of cities adversely. Experience elsewhere has shown that when subsidized inputs are available for land and shelter development there is a tendency to use more lavish planning and building standards than would be used under market conditions (i.e. larger plots, more open space, more finished housing units). This is undoubtedly the case in some of the newly developing areas of large cities of Pakistan where fairly expensive development standards are being applied (see paras 4.09-4.13). - 49 - e. An Opportunity is Missed to Contribute to Financial Stability 6.26 The Islamization of the financial system has had the important effect of freeing financial rates of return on most types of lending from previous below-market levels, with the notable exception of housing. However, one of the potential drawbacks of the system is that, since returns on investments are determined on an ex post rather than an ex ante basis, an additional element of risk is introduced into financial transactions. Lenders' portfolios are taking on a quasi-equity character which may imply higher returns in periods of economic expansion but which also imply considerable down-side risk in periods of contraction. This greater risk may make resource mobilization more difficult from risk-averse savers. Savers seeking security could turn away from financial instruments, leading to disintermediation. There may be a risk of losses to financial institutions in a down cycle of the economy. One way to reduce this risk would be to enable financial institutions to maintain diversified portfolios with some potentially less risky assets, such as housing loans. If mortgage loans were priced more accurately, many financial institutions such as insurance companies and pension funds, would wish to hold them to balance the risk of their portfolios. D. The Proposed Strategy for Housing Finance Objectives for Housing Finance 6.27 A Greater Market Orientation for Housing Finance. Government has indicated a desire to increase housing finance, but a sustainable expansion of housing finance is only possible if it is based on market incentives. To achieve this, housing finance must be more closely integrated with financial markets, and yields on investments in mortgage credit, adjusted for risk, must evolve towards being competitive with alternative investments. The development of market-oriented housing finance in Pakistan would be greatly facilitated if a more appropriate economic policy environment could be developed at the same time. This would include especially measures to reduce the government deficit and to ease the pressure of government borrowing on credit markets. 6.28 Responsiveness to Low Income Groups. Housing finance should be provided equitably to the broadest segment of low income groups possible. Any available subsidies should be targeted to the lowest income groups. The need to reach low income groups is frequently cited as a reason not to introduce market oriented housing finance. However, if combined with more cost effective standards and more flexible lending criteria, market oriented finance would have the potential for financing a much larger housing program for a broader range of income groups than the present more narrowly focused program. Short to Medium Term Improvements 6.29 The most immediate priority would be the improvement of HBFC's operations, especially if it is to be the implementing agency for some of the Government's priority housing credit programs. HBFC is receiving - 50 - conflicting messages, first that it must implement risky programs and second that it must be more market oriented to receive funding. This conflict should be reconciled. 6.30 Improving the Design of Targeted Housing Finance Programs. HBFC has participated in special housing finance programs, including a flood relief program, the program to upgrade housing in central Lahore and the Five Point Program. Management has approached these programs reluctantly because of their political orientation and the likelihood of default. As a government agency, HBFC is likely to be asked to sponsor future housing finance programs for low income groups. However, Government should not undermine HBFC's financial position or its ability to move towards more market oriented sources financing. HBFC should not be required to finance loans for beneficiaries from whom collection is not feasible. Government should agree to cover at least part of any losses from such programs. For its part, HBFC should make a greater effort to design loan programs to be more appropriate for low income groups. This would include greater flexibility on loan amounts and collateral (see below) and improved outreach and community education. 6.31 Modification of Collateral Requirements. HBFC requires that housing only be financed which meets local building codes and for which legal tenure can be established and liens placed. This effectively excludes housing loans in katchi-abadis where tenure has not been normalized and home improvement loans in central cities where ownership records have not been maintained. The large effective downpayment requirements also exclude many lower income households. HBFC's reluctance to lend in circumstances of questionable legality and tenure is understandable. Nevertheless, other types of guarantees could be considered as collateral such as group co-guaranties or employer guarantees. HBFC is involved in a World Bank assisted program to improve houses in the old city of Lahore where tenure is difficult to document which provides a good opportunity for devising alternative collateral arrangements. The possibility of lending for incremental construction should also be explored by HBFC, even though houses may not conform fully to local building codes. 6.32 Improving Collections. Even without changing the pricing of HBFC housing loans, the cost of the program could be reduced by improving collections. HBFC has recently set up a new system utilizing the Muslim Commercial Bank (MCB) to receive loan payments. When installments are not made in a timely manner, MCB notifies HBFC which starts a follow-up procedure at the branch level. For both its new loans and old loans HBFC has now instituted a more stringent follow-up. Another step to improve collections is the maintenance of better data on delinquency. With the expansion of data processing through its new computer installation, HBFC now has the capacity for better monitoring. It is important for HBFC to continue to demonstrate its ability to improve of collections, especially as the system evolves towards being more market oriented. - 51 - Medium to Long-Term Improvements 6.33 Over the longer term, it would be important to establish an environment which would foster the development of private as well as public housing finance. 6.34 Development of Market Priced Lending Instruments. Even under present conditions, there is scope for movement towards the market pricing of loans. Housing is a very profitable investment in Pakistan, especially given the pressure on urban housing markets and the constrained supply. The formula now used to calculate HBFC's share of the profit could be adjusted to increase HBFC's share to a market rate of return, without harming beneficiaries. This may lead to a type of adjustable repayment to reflect that "profits" increase in nominal terms over time. 6.35 The difficulty of assigning a profit sharing concept to construction financing (see para 6.09) should be reviewed, in order to design a formula for recovering a fair part of the benefit to a borrower during construction. 6.36 Establishment of Regulatory Framework for Expanding Resource Mobilization. Much of the appeal of market-oriented housing finance is that private savings could be tapped directly for housing. This is likely to be difficult in the short term until adequate lending instruments and institutional capacity are developed. Changes may be required to enable housing finance institutions to accept new types of loans and deposits. 6.37 Mortgage loans which are appropriately priced with a minimum of risk are likely to be attractive investments for financial institutions such as insurance companies and pension funds which have a natural preference for long-term assets. The mobilization of resources from these types of institutions would be a logical first step in resource mobilization for market-oriented housing finance. Over time, however,the resource requirements of market oriented lenders would probably be greater than would be available from existing financial intermediaries. Mortgage lenders would then have to gain direct access to household and corporate savings. The ability of housing finance institutions to attract household deposits will depend on the range of services they can provide as well as the returns they can offer. Further analysis is required of appropriate sources for housing finance and the changes in legislation or regulations required to enable broader resource mobilization for housing finance. 6.38 Reduction of Credit Risk. In order to attract market-oriented credit into housing finance, returns must be comparable to other investments, adjusted for risk. The high levels of delinquency and default presently associated with HBFC's present program would tend to preclude investment from other than government sources. Means of reducing credit risk, therefore, require careful consideration. HBFC has powers to recover loans under both its own Act and under the Revenue Act. Nevertheless, there is a need to assess possible changes in legislation to provide improved mortgage instruments and foreclosure procedures. New legislation would probably be required for private mortgage lending - 52 - institutions. In some instances changes in the administrative procedures of courts may help facilitate foreclosure and reduce political interference. 6.39 Institutional Development. The market pricing of housing credit, as well as the other elements listed above, would be essential to any housing finance strategy in Pakistan, whether it involves a continuing reliance on public housing finance institutions such as HBFC, whether the private sector is given the primary role in housing finance or whether some mix of the two approaches is used. In principle, HBFC could evolve towards a more market oriented financial institution. Although this could be difficult, HBFC should, nevertheless, strive to operate as close to market as possible. This would not preclude an expanded role by the private sector in housing finance. 6.40 In the long run a system of mixed institutional responsibility could develop with the private sector gradually taking an increasing role in mobilizing resources and financing housing. Private sector institutions could begin mortgage lending with limited access to credit markets. As they perfect lending instruments and build public confidence, they could mobilize resources directly from private savings. A detailed plan should be prepared of the legal and regulatory framework required to support private housing finance, including a timetable for implementation. Under such a scenario, the public sector would play a reduced role in housing finance with its orientation focussed more on the lowest income groups but with the least amount of subsidy possible. - 53 - Table 6.3: HBFC - Profit and Loss Account for the Year Ended 30 June 1986 1986 Rupees 1985 Rupees INCOME Interest 160,817,555 178,376,655 Share in rental income 313,692,155 245,693,332 Other income 53,024,265 42,706,312 TOTAL 527,533,975 466,776,299 EXPENDITURE Interest and bank charges 144,750,943 170,712,286 Establishment expenses 47,188,205 43,475,016 Rent, rates, taxes and insurance 4,188,205 4,802.699 Postage, telephone and telegram 1,213,212 1,240,254 Printing and stationery 1,366,721 1,767,154 Legal and Professional charges 780,652 686,925 Audit fee 30,000 30,000 Depreciation and repairs to fixed assets 3,211,951 3,006,200 Provision for doubtful receivable 13,000,000 3,000,000 Other expenses 2,997,185 3,821,140 218,727,783 232,541,674 308,806,192 234,234,625 SHARE OF PROFIT PAYABLE TO STATE BANK OF PAKISTAN 308,050,736 234,196,864 NET PROFIT BEFORE TAXATION 755,456 37,761 PROVISION FOR TAXATION Current 5,502,182 1,518,880 Prior years - 13,204,690 5,502,182 14,723,570 LOSS AFTER TAXATION ( 4,746,726) (14,685,809) UNAPPROPRIATED PROFIT BROUGHT FORWARD 14,797,815 29,483,624 UNAPPROPRIATED PROFIT CARRIED FORWARD 10,051,089 14,797,819 - 54 - Table 6.4: Size Distribution of Loans by HFBC No. of Cases Amount (Rs in Million) Size of Loans 1985-86 1986-87 1985-86 1986-87 Up to Rs 20,000 310.0 142.0 5.6 2.4 (1.6) (0.9) (0.3) (0.1) 20,001 - 40,000 1,843.0 2,255.0 110.4 77.1 (9.2) (14.9) (6.8) (6.6) 40,001 - 60,000 6,536.0 3,722.0 338.4 203.6 (32.7) (24.6) (20.8) (17.4) 60,001 - 100,000 8,256.0 6,029.0 685.6 504.5 (41.3) (39.8) (42.2) (43.2) 100,001 - 150,000 3,024.0 3,000.0 484.0 381.0 (15.1) (19.8) (29.8) (32.6) TOTAL 19,969.0 15,148.0 1,624.1 1,168.6 Figures in brackets are percentages. - 55 - Table 6.5: ComParison of Rates of Return on Advances by HBFC and Other Financial Institutions 1985-86 1986-87 1. Scheduled Banks (Average) Overall 10.91 10.71 Private Sector 11.21 10.85 2. ADBP Short-Term 12.00 12.00 Medium and Long-Term 12.00 12.00 3. Cooperative Societies Short-Term 12.00 12.00 Medium and Long-Term 12.00 12.50 4. PICIC Foreign Currency Loans 14.00 14.00 Local Currency Loans 7.50* 3.00* 5. NDFC Foreign Currency Loans 14.00 14.00 Local Currency Loans 14.00* 14.00* 6. HBFC (Average) 6 - 7.00 * On mark-up basis. ** Rates under PLS as prescribed by SBP. Table 6.6: House Buildina Finanec Cornoration Head Office Karachi Disbursement Year Wise and 2one Wise Years Total | Karachi Hyderabad Quetta Multan Lahore Islamabad j Peshawar 1981-82 1344.70 506.00 120.43 13.19 142.56 228.15 233.73 100.64 1982-83 1150.30 384.73 122.67 15.40 98.38 119.92 136.40 272.80 1983-84 1385.00 393.90 213.30 27.00 141.21 234.19 245.19 130.21 1984-85 1300.71 329.32 176.70 17.90 146.21 259.18 230.54 140.86 1985-86 1314.00 382.4w 184.04 10.SO 130.41 228.38 244.74 133.52 1986-87 2020.93 582.03 200.84 15.61 241.05 353.61 199.66 228.13 TOTAL 8S1S.64 2578.39 1017.98 99.60 899.82 1423.43 1490.26 1006.16 Years Southern Reaian Horthern Reaion Total 1981-82 639.62 705.08 1344.70 1982-83 522.80 1150.30 1150.30 1983-84 634.20 7S0.80 13dS.OO 1984-8S S23.92 766.79 1300.71 1985-86 576.95 737.05 1314.80 1986-87 798.48 1222.45 2020.93 TOTAL 3695.97 4819.67 8515.64 - 57 - VII. A MORE BALANCED SHELTER DEVELOPMENT PROGRAM 7.01 As discussed in paragraphs 1.19-1.21, a balanced shelter development program would address the major constraints to housing development in conjunction with the provision of housing credit. The objective of such a program wotdd be to improve the shelter conditions of low income households. At the same time it would address the priorities of making land development more affordable through reduced standards and improved pricing and allocation techniques and normalizing tenure in katchi-abadis and rural areas. Given the present fiscal situation, it would also be important to minimize the cost to government and maximize cost recovery. Such a program could be structured alon,- the lines of existing shelter-related programs, with improved developmer, impact and targeting an4d with each program complemented by an appropriately designed credit component. 7.02 Table 7.1 illustrates the possible components of a National Low-Cost Shelter Development Program. The amounts shown are quite large (US$ 336 million per year compared to 1987/88 development budget for housing of about US$ 187 million). The ultimate program size would have to be carefully considered and would depend on budget constraints and implementation capacity. The program could consist of the following components: National Community Improvement Program 7.03 This would be a modified version of the present katchi-abadi improvement program expanded from present levels to cover a significant percentage of the 5.5 millioru people estimated to be living in katchi-abadis within a five year period. It could also be expanded to many medium and small towns where people live in similar conditions. Such a program would focus on households' priorities for clear tenure and services and other principles outlined in paragraph 3.16: affordable engineering standards, recovery of land cost, recovery of service cost through utility rates and property taxes, and strengthening of provincial monitoring capacity. Households receiving normalized tenure would be eligible for home improvement loans channeled through a financial intermediary such as HBFC or commercial banks. 7.04 Issues which would be addressed in designit.g such a program include: (a) how could procedures for normalizing tenure be improved and how could beneficiaries'incentive for buying tenure be increased? (b) what are the constraints to including katchi-abadis in the municipal tax system as they are improved? (c) what implementation arrangements are appropriate for cities of various sizes? and (d) what are the most feasible options for improving and regularizing tenure in katchi-abadis on private land? National Land and Shelter Development Program 7.05 A National Land and Shelter Development Program would be an extension of the present area development schemes. As outlined in Table 7.4, it would begin at present levels and grow gradually over a five - 58 - year period. It would demonstrate affordable land, infrastructure, and housing development standards. The program would include home materials and construction loans for low income plot recipients as well as for households owning serviced plots outside project areas. These would be provided through a financial intermediary such as HBFC or a commercial bank. Unit costs would be higher than past programs because land would be purchased at market prices to make the program replicable. More innovative design and pricing techniques (see paras 4.09-4.13) would, however, keep the program affordable to the urban poor. The unit costs shown in Table 7.1 (based on Peshaw.r) may be less for smaller cities. In some cases Government may already own land, which would reduce the required capital outlay, but should nonetheless be sold at market prices. 7.06 Issues which would be resolved in designing such a program include: (a) what institutional arrangements would be required, including channeling of technical assistance and financing from the national to the local level? what changes would be required in the regulations for plot allocation (see para 4.14)? over the longer term, what changes in land acquisition procedures and the planning and provision of trunk infrastructure would be feasible to facilitate this type of program? Without much greater expense, a government program could not possibly meet the needs of all new urban households. It would be important to enable the private sector to make a greater contribution to urban land development by eliminating unnecessary constraints (see para 4.25). Rural Land and Shelter Development 7.07 A Rural Land and Shelter Development Program would make limited amounts of credit available to rural households to finance priority shelter related needs such as land acquisition, on plot sanitation, roofing materials, etc, with flexible loan amounts and repayment periods. It would focus on providing land tenure and basic shelter needs. There is a need for large-scale beneficiary education, which would require the involvement of local government officials, but beneficiary selection and loan disbursement and collection should be through a financial intermediary such as the Agricultural Development Bank of Pakistan (ADBP) in order to minimize political interference. An extensive program would have to be based on improved information and experience. The program should, therefore, be initiated on a pilot basis and expanded gradually as institutional capacity allows and experience indicates. 7.08 There are several issues which would be addressed in designing and implementing a rural land and shelter program. These include: (a) what priority do rural households assign to house construction as opposed to other needs such as social services, water supply and sanitation? would a rural public services program be appropriate in combination with a rural housing program? (b) where households do not own the land they occupy, is it possible to normalize their tenure? (c) what is the normal process for village extension and how can it be facilitated to allow households to acquire new land for residential purposes? (d) is the supply of building materials a constraint to house construction and, if so, what can be done? (e) what construction techniques and processes are used in rural housing and how can credit facilitate them? Table 7.1: ILLIIR V COST QF ATIA lOW-MST S'ILTED 1900)og9 1900)90 ouet192 1'e1m92 1i721J Tatala munt esneficiaries' AwAt Seneftciaries' _.mt _eeiciariesf Amot ftntficiaries' Awit SaWiciaries' _t gemficiaries (As millions) thousnds) (ts millions) (thousands) (s millions) (thousw'ds) lRs mtllions) (thousands) 0tR mrillions l (th4Aat) (Us millios) thtousads) linttana Cmmitw Kathi Abadi Uperadinh' so0 27n 1,000 sse 1.5se e25 1.500 s21 I'&ge U25 S.000 3.o (41) (82) (121) (123) (123) (402) Katchi Abadi NM rovmnt Loans' 13 91 270 1I2 40U 272 408 272 40S 272 1,0 I.on (14) (27) (41 (41) (41 (41) Natt91 Urba Lmi ui Shlter -Area Dwelo itql 22.000 240 2.2S0 275 2.500 Soo 2,750 t358 I.'00 uS 12,600 1.515 436) (41) (48) (50) (55) 1227) bterials, Loans' 182 18i 1n 207 204 228 228 2S0 246 271 1.028 1.142 (27) (31) (54) (58) (41) (121) Other Urban Noe lprovennt and Construction Loans' 375 161 S28 23S 710 335 750 S38 70 3SS 3180 1.401 (25) (35) (SO) (50) (S0) (210) LP Rural Slter 0 Now Zroveent/ Construction Loans' 200 13 400 270 600 400 750 S0o 7s0 500 2,700 1.005 (20) (40) (60) (78)(7) (270) 3.372 SIS 4,630 1.330 1.9S9 160e 6.383 1.99S 86.51 2.,2U 26.995 *.0o4 (122) (90) (270) (298) (303) (I ) Footnotes. I Figures in parentheses are thousands of households. 2 197/8 prices. Unit Cost assures Rs 1,500 per capita. plus 105 design nd supervision. plus 105 contingences, say Rs 1.20 per capits. ' Asses 338 of households (average urban Ni site = 6.7 persons) ulli receive lans averaging Rs 10,000. ' Average plot cost including land, infrastructure, 10O des1gn and supervision: Rs S5,000 ' Assung 7S of beneficieries (lower-middle and lower Inctm households) get loans averaging Us 6,000 Asuoing an average loan of Rs 15.000 Assuming As 10,000 Per household - 60 - Housing Finance 7.09 In housing finance, the most immediate priority would be the improvement of HBFC's operations, especially if it is to be the implementing agency for some of the Government's housing credit programs. This would include movement towards market rates of return on all housing loan programs, improved design of special credit programs, more flexibility in collateral requirements and an increased effort to improve collections (see paras 6.30 - 6.32). 7.10 Because Government lacks the resources to continue providing virtually all formal sector housing finance, over the longer term it would be important to establish an environment which would foster the development of private housing finance. This would include the development of market priced lending instruments, the establishment of a regulatory framework for expanding resource mobilization and the reduction of credit risk (see paras 6.34-6.40). Possible Role for International Assistance 7.11 The eventual size and scope of a national program could be quite large. There would be important institutional development requirements at the national and provincial levels. Because of this, assuming there is agreement on the overall strategy in the sector, initial international assistance would best be focused on priority needs and geographic areas within each category of the program. 7.12 Community Improvement Program. An initial phase could be considered for international assistance. To be manageable and to provide an example for subsequent larger-scale implementation, it should focus on a limited number of cities where there is a high probability of success (i.e. efficient local organizations, tenure of government land can easily be transferred to beneficiaries). The initial program should include representative cities of all sizes to gain experience necessary for later program expansion. Technical assistance could be provided to help a national agency such as the NHA to establish program criteria and to assist provincial and local agencies to prepare subprojects, including improvements in municipal financial flows and cost recovery. 7.13 Land and Shelter Development Program. International assistance could support an initial program focused on improved land and shelter development schemes in a limited number of cities. To gain initial project experience rapidly, cities would be selected where land is already in public ownership, local institutions are efficient and there is a good probability of success. During implementation of the first phase, land acquisition and institutional strengthening could be undertaken in other cities which would be included in later phases. Technical assistance would be required by the NHA to establish program criteria, including regulatory changes necessary to establish more efficient land development systems, t. 3ist with the analysis of long term development issues, such as land acquisition and participation of the private sector, and to assist local agencies with the preparation of the schemes to be financed. - 61 - 7.14 Rural Shelter. International assistance could finance the initial phase of a rural home loan program which would serve limited areas to gain knowledge and experience before expanding to a nationwide level. In the preparation of the project, the land tenure and project design issues listed in paragraph 5.23 would be addressed as well as the selection of an appropriate intermediary. Technical assistance could be provided through the NHA to assist with analysis and project design. Financial assistance would be channeled through the financial intermediary with community liaison assistance form LGRD. 7.15 Housing Finance. In the short term, international financial assistance could support housing loan programs in urban areas through a financial intermediary, possibly HBFC, to complement the program components outlined in paragraphs 7.12 and 7.14. Technical assistance could be provided to the financial intermediary to help design and implement more effective low income loan programs addressing the issues in paragraphs 6.30-6.32 and especially the need to improve collections. Technical assistance and financing could also be provided over the longer term to help with the design of a more market oriented housing finance system with a greater contribution from the private sector (see paras 6.34 - 6.42). - 62 - Annex 1 PAKISTAN SHELTER SECTOR REVIEW Statistical Annex Table 1 Gross Domestic Capital Formation in the Housing Sector Table 2 Summary of Census Data Table 3 Pakistan - Historical and Current Population Table 4 Projected Population to Year 2000 - High and Low Estimates Table 5 Wholesale Price Indices of Major Construction Materials Table 6 Prices charged by Karachi Development Authority on Plots of Different Size in Various Schemes Table 7 Typical Land Price Increases in Lahore Table 8 Price Indices of Construction Inputs 1975-83 _ 63 Annex I Table I PAKISTAN SHELTER SECTOR REVIEW Gross Domestic Capital Formation in the Housing Sector 1975-76 TO 1986-87 (Rs in Million) Year Investment 7. of GDCF Price Index Real 7. Growth in in of Building Investment Housing Economy Maintenance in Housing (1975176 = 100) 1975-76 1341.2 100.0 1341.2 10.9 1976-77 1709.0 114.9 1487.4 8.4 1977-78 2034.6 6.7 126.2 1612.2 8.4 1978-79 2272.9 6.9 151.7 1498.3 -7.0 1979-80 3003.0 7.3 197.5 1520.5 1.5 1980-81 3850.3 9.0 204.7 1880.9 23.7 1981-82 4501.0 9.2 201.4 2234.8 18.8 1982-83 5898.5 10.4 200.5 2941.9 31.6 1983-84 6395.0 10.1 216.5 2953.8 0.4 1984-85 6946.0 9.7 229.0 3033.2 2.7 1985-86 7705.0 9.5 240.1 3209.1 5.8 1986-87 8130.3 8.6 250.1 3250.8 1.3 Source: Pakistan Economic Survey, 1985-86 -64- Annex_l Table 2 SHELTER SECTOR REVIE. Sumary of Census Data (Excludes F.A.T.A) Item 1961 19*1 ALL UR8AH RURAL ALL URBAN RURAL Popn. 39.442.439 9.614.004 29.828.435 82,055.097 23.841,471 S8.213.626 Popn.tlS 16,737.962 3.956,564 12.781.398 36.519,289 10.250.182 26.279.107 Papn.60+ 2.733,933 531.669 2.202.264 S.733.662 1.323.131 4.410.531 Popn.1s+ Har'd. 15.887.378 3.766.765 12.120.613 31.306.760 8.760,792 22.545.968 Popn. 15+ W1D 2.292.074 519.593 1.772.481 2.797.077 769.768 2.027.309 Popn. 1S+ N.M. 4.525.025 1.371.082 3.153.943 11.431,971 4,060.729 7.371.242 Single Households 7.159,634 1.647.412 5.512,272 12.587.648 3.554.173 9.033.475 Housing Units 7.815527 1.698.900 6.116.627 12,587.648 3.554.173 9.033,47S Avg. Nh. Size S.S1 5.84 5.41 6.52 6.71 6.44 Pers./H.U. 5.05 5.66 4.88 6.52 6.71 6.44 Pers./Room 3.3 3.1 3.3 3.5 3.2 3.6 Rooms/Unit 1.9 2.2 1.8 Rooms/Hh. 1.7 1.8 1.6 1.9 2.2 1.8 Persons/Family 4.7 4.6 4.7 Notes: WbD = widowed and divorced N.H = never married Source: 1961 Census of Housing; 1961 Census of Population 1981 Census of Housing; 1981 Census of Population January 1988 - 65 - Annex I Table 3 pAKISTAN SHELTER SECTOR REVIEW Historical and Current PoRulation (000) J22j 1972 12ffi1 1/ 1971-81 1985* Intercensal (mid-yerl Growth Rate (X) METROPOLITAN CITIES Karachi 3.515 5,208 4.5 6.297 Lahore 2.170 2.953 3.E 3.404 Subtotal 5.68S 8.161 (ave) 4.0 9.701 SECONDARY CITIES Faisalabad 823 1,104 3.3 1,278 Rawalpindi 61S 795 2.9 905 Hyderabad 629 752 2.0 827 Hultan 539 732 3.5 852 Gryranivala 324 601 7.1 803 Peshawar 273 566 8.4 795 SialkaL 204 306 4.4 366 Sargodha 200 291 4.2 349 Quetta 158 286 6.8 370 Islamabad 77 204 11.4 320 Subtotal 3.842 5.633. (ave) 4.3 6,872 OTHER CITIES 6.755 10,047 4.5 12,135 TIOAL URBAN 9.614 16,218 23,841 4.4 28.708 TOTAL RURAL 29,828 49.091 60.472 2.3 67.472 TOTAL POPULATION 39.442 65.309 84.251 2.9 96,180 PERCENT URBAN 24% 25% 28% 30% * Assuming continuation of 1972-1981 growth rates. 1/ 1981 population totals differ from Table 1.02 which excludes Federally Administered Tribal Areas (F.A.T.A.) - 66 - Annex 1 Table 4 PAKISTAN SHELTER SECTOR REVIEW Proiected Population to Year 2000, High and Low Estimates (000) 2000 High Low 1985 Estimate Estimate METROPOLITAN CITIES 9,700 17,242 16,128 Karachi (6,296) (11,190) (10,468) Lahore (3,404) (6,052) ( 5,660) SECONDARY CITIES 6,870 13,470 12,600 OTHER CITIES 12,135 23,169 21,672 URBAN TOTAL 28,705 53,881 50,400 TOTAL RURAL 67,472 95,788 89,600 TOTAL POPULATION 96,180 149,669 140,000 PERCENT URBAN 30 36 36 URBAN GROWTH RATE (% P.A.) 4.3 3.8 RURAL POPULATION GROWTH RATE (% P.A.) 2.4 1.9 TOTAL POPULATION GROWTH RATE (2 P.A) 3.0 2.5 Note: Urban growth rate calculated as follows: Metropolitan cities: national average growth rate plus 1.0% p.a. Secondary cities: national average growth rate plus 1.59% p.a. Other cities: national average growth rate plus 1.41% p.a. Wholesale Price indices of ?a-ior Construction Materials Iron Bars/ Wire/ Paints/ Building PRICE INDEX Year Ceent Bricks/Blocks Sheets Cables Tmber Varnishes Material Overall WPI 1976-77 123.9 121.0 116.2 100.0 97.9 101.0 114.9 111.9 1977-78 150.7 134.6 121.4 100.0 109.7 113.4 126.3 120.8 1978-79 211.9 151.2 137.8 104.4 137.9 135.9 151.7 128.6 1979-80 317.2 187.9 162.2 137.6 184.7 167.4 197.5 144.7 1980-81 318.0 204.8 170.8 151.9 199.3 171.9 204.7 163.7 1981-82 362.9 212.3 144.0 138.2 184.2 169.9 201.4 176.2 1982-83 374.8 215.3 138.8 138.2 183.6 168.9 200.5 182.3 1983-84 412.7 241.9 151.8 122.4 186.5 173.9 216.5 201.4 1984-85 418.5 249.0 172.4 120.9 206.8 170.0 229.0 208.9 1985-86 240.1 1986-87 250.1 I * Source: Pakistan Statistical Year Book Pakistan Economic SUrvey. 1985-86 aIL - 68 - Annex1 Table 6 PAKISTAN SHELTER SECTOR REVIEW Prices Charged bY the Karachi Development Authority an Plots of Oifferent Size in Various Schemes (Price per Sq. yard) Scheme/Plot Size 1972-7S 1976-87 SHAH LATIF 60, 80 SO 75 120 SO 100 GULSHAN-E-IQBAL 80 9 80 120. 240 12 80 400 14 80 600. 1000. 1500 20 80 GHULISTAN-E-JAUHAR 60, 80 60 100 240, 400 125 200 600 250 250 METROVILLE - I 80 27.50 120 30 - 240 32 - 400 35 - METROVILLE - II 60. 80 30 so 120 30 100 METROVILLE - III 60. 80, 120, 240 - 80 NORTH KARACHI 80 and above 8 50 KORANGI, MALIR, DRIGH. AND QASBA 80. 120 8 35 240 and above 8 50 SJRJANI 60 100 120 150 240. 480 200 CLIFTON 80 18 100 120 and above 40 100 - 69 - Annexg Table 7 PAKISTAN SHELTER SECTOR REVIEW Tvoical Land Price Increases in Lahore Price oer Kanal (Rs.) Areas 1980-81 1982-83 1986 1987 Residential Area of Upper Income groujU: Gulberg - Main Boulevard 500.000- 800,000- 1.500.000 2.500,000 550.000 1.000,000 - H.M. Alam Road 300.000- 400.000- 800,000 1,500,000 (another main rd) 350.000 450,000 - Side Lanes 250.000 300,000- 550.000- 850.000- 350.000 600.000 1,000,000 Residential Area of Middle To Upper Income Groups: Model Town 150.000- 200,000- 450,000 600.000- 175.000 225,000 700,000 Residential Area of Middle Income Grou: Township Z50,000 300.000 400,000 500.000 Allama Iqbal Town NOT DEVELOPED THEN 200,000 400,000 500.000 Samanabad 200.000- 325,000- 420,000 500.000 2S50.000 360.000 Commercial Area (in Middle Income Areas): Ferozepur Road (Price per Marla) 35.000- 75,000- 150,000- 90.000 100.000 200,000 Wahdat Colony (Price per Marla) 20.000 30.000 40.000 Commercial Area (Upper Income Areas): In Gulberg there is no commercial land available but average sized shops (on 1 to 1 1/2 Marla) are being sold for Rs. 1,600.000 compared to Rs. 800,000 in 1986. Notes: i) Prices have been obtained for areas only where land is available and transactions are being conducted. ii) 1 Kanal = approximately 500 sq. yds. iii) 20 Marlas = 1 Kanal Annex l Table 8 PAKISTAN SNELTER SECTMR REVIEW PRICE INDICES OF CONSTRUCTION INPUTS 1975-83 a/ Input Unit 1975-76 76-77 77-78 78-79 79-80 80-81 81-82 82-83 b/ Construction c/ Price 322 373 393 403 507 568 616 645 Index 100 116 122 125 157 176 191 200 Labour Manday 42 53 59 62 71 83 94 94 100 126 ^ 147 169 196 223 223 Skilled Manday 28 36 41 44 50 57 63 66 100 128 147 156 179 204 225 235 Unskilled Manday 14 17 17 18 21 25 26 28 100 123 124 130 149 180 185 199 Timber Log Cubic 1781 2477 2030 2078 2252 2449 3126 3531 (Shioshan) metre 100 139 114 117 126 140 176 198 Cement 50kg 24 30 36 38 56 58 64 66 (Portland, Grey) 100 125 149 159 234 241 264 275 Mild Steel Bars M ton 3644 4426 4045 4134 5371 6049 4526 4422 (1/20 dia) 100 121 111 113 147 166 124 121 Bricks 1000 245 266 285 299 439 469 482 465 (Coal Fired) 100 108 116 122 179 191 196 189 Motor Spirits litre 2.56 2.70 2.80 3.13 3.96 4.83 5.08 5.72 (Truck Transport) 100 105 109 122 155 189 198 223 a/ All Lahore prices from National Engineering services Pakistan Ltd, NESPAK Price Index, (Feb. 1983:24-25) except for motor spirits, oil and coal which are Pakistan prices from GOP (1983-84) Pakistan Economic Survey co b/ NESPAK July-Dec. 1982 figures for labour, timber, cement, steel, bricks c/ 1969-70 = 100 - 71 - Annex 2 PAKISTAN SHELTER SECTOR REVIEW Results ot a Survey of Katchi-Abadis in Karachi This survey was carried out by the Applied Economics Research Center, University of Karachi on behalf of the Karachi Development Authority, Master Plan Department, in the second half of 1987. The results show some of the most recent information available about the occupants of Katchi-abadis in Karachi. Many of the occupants have been in Karachi for some time, either since birth (292) or since 1955 (32X). The median household income of the sample was Rs 1700 per month. The average household income of the sample was Rs 2470 per month. A very high percentage (841) of households consider themselves owners of the dwellings they occupy, although this appears to include homeowners on land for which they may not have fully normalized tenure. By far the largest source of financing for house construction is savings (see Table 5). Roughly 202 of housing cost was financed from loans, but there is no information whether these loans are from formal institutions such as HBFC or from informal sources. Tables 6 and 7 show several important aspects of existing housing. They tend to be on fairly large plots, with 67X on plots of 60 square meters or more. A high percentage of houses have walls, roofs and floors of permanent materials. This is probably indicative of the age of the katchi abadis being surveyed. The limited availability of water supply, solid waste removal and other municipal services (Table 9) is indicative of the priority the residents of katchi-abadis are likely to have for improved municipal services. - 72 - Annex 2 Tables 1 and 2 PAKISTAN SHELTER SECTOR REVIEW Table 1: Household Characteristics Average Number per Household Age (yrs) Male Female Total 0-5 0.66 0.64 1.30 6-10 0.64 0.56 1.20 11-15 0.4E 0.45 0.91 16-20 0.52 0.43 0.95 21-30 0.66 0.53 1.19 31-4G 0.39 0.55 0.74 41-50 0.30 0.22 0.52 51-60 0.18 0.11 0.29 61 and Above 0.12 0.07 0.19 TOTAL 3.93 3.36 7.29 Table 2: Period of Migration to Karachi By Birth 29 Up to 1955 32 1955-1965 11 1966-1975 15 1976 Onwards 13 100 - 73 - Annex 2 Tables 3 and 4 PAKISTAN SHELTER SECTOR REVIEW Table 3: Distzibution of Household Income in Katchi-Abadis Monthly Household Income Income Group 2 Up to 500 Rs 2.4 501- 1000 19.1 1001- 1500 21.9 1501- 2000 18.7 2001- 2500 9.8 2501- 3000 9.7 3001- 3500 3.9 3501- 4000 4.3 4001- 4500 2.1 4501- 5000 2.4 5001 and Above 5.5 TOTAL 100.00 Average Household Income: 2470 Rs Table 4: Tenurial Status of Properties in Katchi-Abadis X of Housing Units Owner-Occupied 84 of which: property constructed 52 property purchased 18 property inherited 14 Rented 16 of which: pugree*/rented 1 monthly rented 15 *This corresponds to the payment of a relatively large lump sum payment of Key money, followed by relatively small monthly rent payments. - 74 - Annex 2 Table 5 PAKISTAN SHELTER SECTOR REVIEW Sources of Housing Finance for Katchi-Abadi Properties % of Households** Average financing X Using per Property by of Source Source (Rs) Finance A. If Property Constructed Savings 81 16,781 62 Loans 40 5,800 21 Remittances 3 1,655 5 Others* 10 2,781 11 27,017 100 B. If Property Purchased Savings 76 13,167 56 Loans 34 5,226 22 Remittances 2 1,413 6 Others* 17 3,702 16 23,508 100 *Other sources include sale of jewelry and other assets grants from relatives, etc. **The column total for the four sources may exceed 100% because of the use of multiple sources of finance for a property. _ 75 . Annex 2 Table 6 PAKISTAN SHELTER SECTOR REVIEW Basic Characteristics of Properties in Katchi-Abadis a. Number of Rooms: No of Rooms X of HousinR Units 1 30 2 41 3 17 4 8 5 and Above 4 TOTAL 100 b. Plot Sizes: Plot Size (Sq. yds) X of Housing Units up to 60 sq. yds 33 61-80 28 81-120 25 121-240 12 241 sq. yds and Above 2 TOTAL 100 c. Age of Properties: Age (yrs) S of Housing Units up to 5 yrs 24 6-10 yrs 26 11-15 yrs 15 16-20 yrs 14 21-25 yrs 5 26 yrs and above 16 TOTAL 100 - 76 - Annex 2 Table 7 PAKISTAN SHELTER SECTOR REVIEW Housing Characteristics a. Walls: % of Housing Units Materials Cement Plastered 51 Unplastered Cement Block 39 Stone/Mud/Wooden Blocks 2 Thatch/Chatai/Tin 2 TOTAL 100 b. Roof: 2 of Housing Units Materials RCC 19 Iron Girders with Tile 47 Asbestos/Tin Sheets Wooden Beams with Tile Asbestos/Tin Sheets 29 Mud/Thatch/Chatai 5 TOTAL 100 c. Floor: X of Housing Units Type Katcha 22 Pucca 78 TOTAL 100 - 77 - Annex 2 Table 8 PAKISTAN SHELTER SECTOR REVIEW Extent of Repairs and Improvements of Properties (over last five years) a. % of Households Making Repairs or Improvements: 64 b. Distribution by Cumulative Expenditure: Expenditure (Rs) Up to 500 Rs 21 501- 1000 Rs 18 1001- 3000 Rs 26 3001- 5000 Rs 17 5001-10000 Rs 11 10001 and above 9 TOTAL 100 Average Expenditure per Household: 4,853 Rs - 78 - Annex 2 Table 9 PAKISTAN SHELTER SECTOR REVIEW Access to Municipal Services a. Piped Water X of Housing Units Piped Access Through: KMC 45 Own Tubewell 2 47 b. For KMC Water Suoply. Hours of Supply Daily: Hours X of Housing Units up to 2 hours 45 3- 4 hours 31 5 - 6 hours 10 7- 8 hours 5 9 -12 hours s 13 - 16 hours 1 17 - 24 hours 5 TOTAL 100 c. If Adeauate PiRe Water not Available. Source of Water: Source X Shisti (Water Carrier 7 Water Tanker 9 Community Tap 70 Others (Multiple Source) 14 TOTAL 100 d. Nature of Garbaae Disposal: Thrown in Street 57 Thrown in Garbage Dump 24 Others (Janitors, Dumping in Open Plot. etc.) 19 TOTAL 100 Extent of Collection by KMC: 25 from Garbage Dump (X) e. Access to Electricity % of Housing Units with Electricity: 72 f. Access to Gas % of Housing Units with Piped Gas: 33 - 79 - Annex 3 Page 1 of 3 PAKISTAN SHELTER SECTOR REVIEW Analysis of Land Use and Infrastructure Standards and Plot Pricing This Annex uses a computerized methodology to illustrate how projects can be designed to be more affordable using improved planning techniques. These techniques help to identify more economical land use standards and infrastructure standards, together with improved plot pricing patterns. Planning Methodology for Johar Town, Lahore 1. Attachment 1 uses Johar Town, Lahore, to illustrate the difference between two approaches which have been used for planning in Pakistan: the traditional superblock approach which tends to use land inefficiently, and the more flexible "affordable modules" approach. The upper plan, which carefully followed the main roads in the Lahore Structure plan, was prepared by planners as a series of superblocks at a scale of 1:2000. It measured an unwieldy 2 meters square and showed every plot, well over 9000. Unfortunately the plan was found to be defective in drainage, with inadequate measures for flood control. It had an undesirable plot distribution and an unnecessarily scattered range of community facilities with no economies of scale. It was considered impractical to correct these defects in a short period of time. 2. A revised plan was prepared based on nine module types, each of two hectares, which were designed at 1:5000, simply by indicating roads and module types. With the help of a hand-held computer, a plan was developed rapidly that combined a viable overall infrastructure and flood protection system with public open space and the precise mix of plots required. These were related to four viable local centers. Analysis of Suriani Town Project, Karachi 3. Attachment 2 shows a typical hierarchical superblock plan from the Surjani Town Project outside Karachi being developed by KDA. It is reproduced at a scale of 1:5000 for direct comparison with later examples from Peshawar at the same scale. It comprised 4176 residential and 512 commercial plots in an overall superblock of about I x 1.5 kilometers bounded by two 90m and two 60m wide roads. The overall superblock is subdivided into 6 superblocks and a strip of mixed development, comprising flats and commercial units, with 25 meter roads, converging on an open space at the center. - 80 - Annex 3 Page 2 of 3 4. The six superblocks each contain a single plot size selected from one of four standard plots. They are based on the same conceptual model, at a reduced scale with 12 meter access roads converging on smaller central areas schools, facilities or local open space. In some instances, an even smaller third microcosm of the concept is repeated to give further open space or local facilities, at the junction of roads of 6 or 12 meters width, irrespective of length. 5. Residential plots are all in blocks of 50-180m length and are sized 336m2, 200m2, 104m2 and 70ml2 with eight blocks of flats on sites of 6600m2. On the basis of measuring the project area from the centerline of the surrounding toads, an excessive 44% is used for roads. 6. In order to facilitate realistic comparisons between projects and undertake sensitivity tests within them, a computer model had been developed which demonstrates the implications of alternative standards and pricing strategies. It also gives a rough indication of affordability in relation to income levels 1' Attachment 3 from the computer model shows that the project appears to be highly affordable. This is because of the very low land price of Rs. 571 per marla. The price used was based on the cost of land acquisition many years before. 7. Attachment 4 shows what the scenario would have been if the project had been implemented on land purchased at a price of Rs. 5000 per marla - typical at the edge of most cities. In order to break even in this case, with affordable housing plot prices, commercial prices would have had to increase to Rs. 33,000 per marla with the average price of serviced land doubling to Rs. 15,639. 8. It is at this point that the inappropriate standards should come under close scrutiny, such as the 44% of land reserved for roads. If a market price must be paid for acquiring land, which is usually the case in Pakistan, the whole range of land use and infrastructure design standards must be reexamined and these should be balanced with an appropriate pricing policy. One would also have to assess carefully whether the proposed 20% downpayments are feasible. If not, additional cost reduction measures may be required. Swati South Project, Peshawar 9. Another case can be used to illustrate the use of affordable modules to expedite the planning process. Typical modules of one hectar each were prepared for the planning of the Swati South Project in Peshawar are shown on Attachmeat 5. Overall plans for the entire project are shown on Attachments 6 and 7. -" The most recent version of the model has been programmed for IBM compatible Personal Computers and is explained in a manual: A Model for the Preparation of Physical Development Alternatives for Urban Settlement Projects, prepared by Agnes Bertaud for the Economic Development Institute, World Bank, 1985. - 81 - Annex 3 Page 3 of 3 10. This approach has the merit of providing a high degree of flexibility at the conceptual stage accompanied by fairly accurate costing for affordability analysis. The laborious and repeatedly abortive task of making detailed adjustments to match local site constraints can be deferred to a later design stage provided that major constraints are dealt with in the design concept. 11. The relative autonomy of individual modules, or cul-de-sac clusters, along local distributor roads, makes possible a much lower standard of internal road and other infrastructure. This is because of the absence of through traffic or long lengths of infrastructure in culs-de-sac, which are unavoidable in block and superblock systems. Module autonomy is also the essential prerequisite, on both practical and marketing grounds, for incremental construction of infrastructure and houses by lower income groups. 12. The financial benefits of these lower standards are a lower capital cost and especially a lower cost per plot. This saving can be further enhanced for lower income groups because of the wide differentials in the value of serviced land that would be reflected on the open market. 13. Attachment 8, for the same project in Peshawar as Attachments 5-7, shows an affordability analysis without differential pricing. Attachment 9 includes a system of differential land pricing of commercial plots (not shown in the printout). The higher price charged for commercial lowers the price charged to lowest income groups from Rs 5000 to Rs 3300 per marla. Both printouts show realistic market prices of Rs. 20,000 per marla for housing plots over seven marlas in size and affordable prices for the smaller plots. These range from only Rs. 3300 to about Rs. 6000 per marla for almost 90% of the plots in the lower priced areas of the scheme and Rs. 5000 to Rs. 8000 for over three-quarters of the plots which are priced at or near cost. Half the plots are four marlas or below, costing Rs. 6600 - Rs. 22,500 and reaching from about the 10th percentile of households in existing low income areas to the 38th percentile of the total urban population, depending on choice of house. 14. Both of the above examples assume a core house would be provided. If further economies were necessary, the greatest potential economies would be from reducing the costs of houses through self-help incremental building. Attachment 10 assumes that materials loans only would be given for plots below seven marlas. A Rs. 3000 building loan and a two-maria plot would thus be affordable down to the 12th percentile, which includes about 60% of households in the informal sector. Over three-quarters of the plots in this project fall within this level of affordability. _ _ r I!if~~~~~~~~I~~~~. III ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~0~~~9....9.......... "'U I -7-- t4 rtx - 83 -A " tt 2 \ 2 ,~ *I -0'W' 0*1' ,.'R 9. O, ,' 0 9 -I a v. .- 0 - q ) g 9Xt ^_ . 8 A -- T - _______ II" **d S LIO I Ir I i'] res1 ~~~t Oa; 1 t1 1: s' -9"' e_ _ _________1__ _, 1~~~~~~~~~~~~~~~~11 -84 - Annex 3 Attachment 3 SURJANI TOWN SECT-5 KARACHI LOW LAND COST AND PLOT PRICES FOR BREAK- EUEN) NON-PROFIT SCENARIO KARACHI DEUELOPnENT AUTHORIlt SCHEltE NO. *I Ga1.: Jan. £90 OPTIONS - 1 2 3 4 S 6 7 8 TOTAL nEAN -… HOUSING PLOTS -----> FLATS SHOPS SHOPS IND-COr SIZE: MRRLAS 6 3 4 a 13 I60 IS 7 7 SO.fETRES 152 ?0 104 1i8 336 4008 373 175 175 NO. OF PLOTS 2160 931 484 672 e 102 281 133 4697 x TOT. NUHBER 46w 20# 94 14-4 ow 4w 4w 3w 1880 w TOTAL AREA 214 14w 1l1 32# 4w 184w 5. 3w 188 LAND USE HECTARES PLOT AREAS 45# 15.12 9.68 6.e0 22.58 3.28 6.93 3.52 2.43 71.36 FACILITIES IS4 16.93 RosDS ETC. 44# 69.01 PRGJECT AREA leo8 (ATTRIBUTAsLE TO NEU USERS) 157.36 PRICING SCHEflE RUPEES PER rsARLA 1¶IL.RUPS LAND ONLY 1449 022 1233 1644 1644 2e54 1742 1742 1712 41.35 INFRASTRCT 2858 168 2517 3356 3356 4196 3558 3558 3436 94.44 PRICE -liar 4487 258- 3750 So8 s0oe 6258 8380 5 330 5288. RUPEES PER UNIT. PLOT PRICE 267?2 7080 15688 39680 6720818o8e83 7S500 37100 36456 HOUSE CON a . o O a ° 8 ° 8 88.88 MATRL.LOAN o o 8 o 8 a o 8 88.eo TOTAL 26782 7088 18600 38600 67280 1.9en 78 MM8 3780e 36486 125.78 AFFORDABILITY ANALYSIS READY CASH 1488 3120 11888 26808 TOTAL LOAN 8680 12480 27728 40328 * REPAYnENTS/M SS 123 2V3 397 x INCOME ROD 15.' 15S 30w 30# !NCCIE em 368 028 918 1324 PERCENTAGES OF HOUSEHOLDS WHO CANNOT AFFORD TO REPAY LOANS SIND URBAN HHLDS 3" 8x Sw 25w HOUSINB FINANCE ASSUn1PTICNS - LOANS FOR HOUSES AND nATERIALS DEPOSIT s VAL 15" 284 30# 4eP FEE x PAYr4ENT 3w 34 3# 3x INTEREST RATE ex Su ex ex PERIOD 15Y ISY ISY I5Y - LOANS FOR SERVICED LAND DEPOSIT x uAL 208 2e8 308 48w FEE w PAYrENT 3x 34 3x 3w INTEREST RATE 89 Ow ox 6O4 PERIOD 1SY ISY IsY IsY COSTS cImL.RUPEES) ON-COSTS APPLIED IN PRICE BUILD-UP W IlL.RUPS.) ROADS t SP 36% 21.69 UATER SUP. 16 10.82 BASE CONTINCENCIES DvStSUP INTERIJ TOTAL SEWERAGE 23S4 14.19 COST PHYSIC. PRICE iRDrIN rAINT FINAKCE DRPINAGE 22x 13.53 154 Ox 5w ex 15X ELECTRICS. 2" 1.38 SOLID UST. Su 0.00 MAT.LOANS 8.80 8.eo BUILDINGS o.eo 8.00 8.88 8.88 8.ee 8.88 TOT. INFRA 100w 60.81 9 6.01 8.12 8.88 3.50 0.88 11.01 e4.44 LAND R 571nla 35.86 ----> 38.s65.39 41.35 LESS SUBSIDY . o.o LAND-tINFRA 86.77 (Co._eS 1984) TARGET FOR COST RECOUERY 125.73 1904 INCOIE DATA RUPEES PER MONTH PERCENTILES 10S4 20- 30" 4e0 58" 68w 70% 080 88.4 SIND URBAN HiLDS 840 1130 1438 1680 1850 2250 2638 3330 4588 -85- Annex 3 Attacbment 4 SURJANI TOWN SECT-5 KARACHI IMPUTED LAND AND HOUSE UALUES INTRODUCED W ITH CROSS-SUBSIDY FROM COMMERCIA.L PLOTS KARACHI DEUELCPnENT AUTHORITY SCHEMSE NO. 41 Dato: Jla. 1988 OPTIONS 1 2 3 4 5 6 7 8 TOTAL rEAN < ------HOUSING PLOTS ------ FLATS SHOPS SHOPS IND/COM SizE rAPRLAS 6 3 4 8 13 160 15 7 7 SO.nETRES 152 70 104 Is8 336 4Ce8 375 175 175 NO. OF PLOTS 2160 831 404 672 8 182 201 133 4637 %4 TOT. NUMBER 46' 20 934 14" Z- 4" 4" 3% Io'e X TOTAL AREA 21" 14" Is 32"- 4x 1 ex 5- 14 loo LAND USE HECTARES PLOT AREAS 45% 15.12 9.68 s.88 22.59 3.20 6.83 2.52 2.'3 71.36 FACILITIES 11" 16.93 ROADS ETC. 44% 69.01 PROJECT AREA 100' (ATTRIBUTABLE TO NEU USERS) 157.36 PRICING SCHEME RUPEES PER nARLA MIL.RUPS LAND ONLY 12680 2676 3649 4827 16216 25765 26757 26757 26757 361.93 INFRASTRCT 2353 624 851 1126 3784 6012 6243 6243 6243 84.44 PRICE /flat 15639 3300 480s 5953 20808 31777 338e8 33800 3308Z RUPEES PER UNIT PLOT PRICE 95034 3248 18720 47148 2668005s84320 435ees 231eze 231018 HOUSE +CON 45667 a o 16s58e 228ee0 a 0 0 e 214.50 r.ATRL.LOPN 2563 3880 6080 a a a 0 0 0 12.e7 TCTpL 143270 12240 24720 212148 48ss88 5.e8M 4350ee 23180e 2318e8 672.94 AFFORDABILITY VANALYSIS READY CASH 1849 3744 63644 195528 TCTAL LOAN 10332 20876 148583 293260 REPAYRENTS/M 102 2e7 1463 2983 x INCOME ROD 15" 15% 30C 30-' INCOrnE nM 683 1379 4877 9631 PERCENTAGES OF HOUSEHCLDS U.NO CANNOT AFFORD TO REPAY LOANS SIND URBAN HHLOS 7x 27A 9Z% 339 HOUSING FINANCE ASSUMPTIONS - LOANS FOR HOUSES AND MIATERIALS DEPOSIT s4 UAL 04 e8 3e-4 40e FEE K PAYr.ENT 3x 3X 3n 3A INTEREST RATE 8x 8e 8x 8c PERIOD 15Y ISV 15Y ISY - LOANS FOR SERUICED LAND DEPOSIT f UAL 20- 20" 30% 42- FEE '4 PAYMENT 3x 314 34 3-4 INTEREST RATE 8x 8e 8- 8Y PERIOD sY 1SY IV5Y I5Y COSTS (MIL.RUPEES) ON-COSTS APPLIED IN PRICE BUILD-UP CrIL.RUPS.) ROADS * SP 36% 21.69 UPTER SUP. 16" 10.02 BASE CONTINCENCIES DEStSUP INTERIN TOTAL SEWERAGE 23- 14.13 COST PHYSIC. PRICE ADMIN rAINT FINANCE DRAINAGE 22" 13.53 15" ex 5x o0 15X ELECTRICS. 2" 1.38 SOLID UST. e- 0.00 MAT.LOANS 12.07 12.07 BUILDINS. 177.64 0.e0 8.89 0.00 27.38 214.50 TOT.INFRA 100" 60.81 -----> 60.81 9.12 0.00 3.50 0.Q0 11.01 84.44 LAND R 5000ona314.72 --------> 314.72 47.21 361.93 PROFIT 0.28 LANDtINFRA 375.53 TARGET FOR COST RECOUERY 672.34 139e4 INCOnE DATA RUPEES PER nONTH PERCENTILES le8 20" 304 40% 50 6e8 70% 88" sex SIND URBAN HHLDS 940 1138 1430 1680 135a 2258 2630 3330 4500 -86 - AM 3 Attacb tt 5 THREE TYPICAL DE OWPNE?NT NODUILS PESHAWAR * * '>-4. _ . .*. s ! it I I I i I I § I ! I | t § ,!, i ! [ l; I . T T f *f fI *1 .. -4T L Fe - * qt "/' si l; F$am ON i If~~~~~~~~~~~~~~~~~~~~~~tl or~ 0 Nift W sO(E *ESS S SEINE DOW CAU )SIE OM "Lo .sr *iPI O;I,It R * .I A .. I-w; H-i I~~~~ . 1. ,,: ,. .; s_. - s-' . -. - 87 - 3 Att8oCEmt 6 H USUIN DEMANT PAA *Scale 1:5000 Q4~~~~~~~~~~~~~~~~~~3" rt ........... ' I0.1% ms go a.~~~~~3 O~~ ~ -aa . C- I ....4woX'I-'. '5-^ /! '1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~6 Am=e 3 - 88 - t 7 A' : 4. t -I -~ ~ ~ ~ ~~~~~~~~~~UTISUl o .; 5EcOI¢ $tEWI - *' -b I'- / o t,2O;l.800& ,A*NOLI o -. 0Mum MANHOLE $1, oatiS .4i t -89- Annex 3 Attachment 8 SWATHI-SOUTH SITES & SERUICES UNSUBSIDCISED PESH4AAR DEUELLOt¶ENT PLANNING MROGRAMME FOR GOVERN.EhT OF PAKISTAN, UNDP 6 UNC^4S BY PDA DIRECTORATE OF URBAN PLPNNIN3 MItH GlLnoRC HANKEY KIRKE Dte Ian. 188 OPTIONS rEAN 1 2 3 4 5 S 2 TOTAL SIZEt nfLARAS 8 2 3 4 5 7 10 15 28 so.nETRES 128 58 75 198 125 175 28e 37S se5 NO. OF PLOTS 128 672 288 649 236 219 14 42 2247 a TOT. NWI8ER 6Y 3e# 13" 298 11" 18' 1t 2Y 188n x ToTAt. AREA 2" 17" 18O 28" 14" 18" 2" 7s le8 * - * LAND USE HECTARES PLOT AREAS 78# 0.64 5-04 2.Bs 8.18 q.13 5.'8 8.53 2.19 28.83 FACILITIES 16" 6.50 ROADS ETC. 14" 5.88 PROIECT AREA 188" (ATTRIBUTABLE TO NEU USERS) 41.27 PRICING SCHEME RUPEES PER MRLA hIL.qUPS LAND ONLY 8214 3588 3598 5757 5757 8976 14393 14393 14383 84.s2 INFRAsTRCT 3288 14e2 1482 2243 2243 3497 5687 5627 5607 36.98 PRICE enar 11414 5eo 588 SN O 12473 2808z 22808 2=388 RuPEES PER UNIT PLOT PRICE 58781 1o29 15888 329o 49888 87311 288888 38e8wo 48*eec FOUSE *CON 42267 6888 6e8o 2888 25088 258 1658see8 2288 558288 94.98 ATRL.LOAN a a 8 o o a o 9 8.88 TOTAL 188868 16898 21888 528o0 65888 11221s 365ce8 52eee2 5828Mo 226.88 AFFORDABILITY ANALYSIS READY CASH 26847 2288 4200 15608 1IS5M 33683 18e8O 2088 382222 TCTAL LOAN 74821 12088 16880 36498 45500 70616 25559i 38128o 57e88Z REPAVfENTSiII 732 126 166 359 448 ?7S 2517 3874 5616 x INCOnE ROD 18" 15" 15 15' 28' 22" 32* 38x 3-' INCOrE 0n 3883 941 18e3 2381 2241 8723 r 18246 19718 PERCENTAGES OF HOUSEHOLDS UNO CANNOT AFFORD TO RtEPAY I I TAKHAL PAYAN PSH 10" 47x 83" 813 SC a33 94" 99S ISLArABAD PSNLR. 12x 38" 71Y 658 S' 83S. 84' 938 NLFP URBAN HhLDS OM 13Y 57'Y 53Y c 82X 938 9Sx' HOUSING FINANCE ASSUrIPTIOaS - LOANS FOR HOUSES AND 4ATERIALS DEPOSIT a UAL 27" 20" 28" 304 30 380-4 30' 4o8 40' FEE Y PAYrTENT 2 3" 3Y 3'. 3 3gs 3# 3Y 3s INTEREST RATE 8x ON Ox u'. 8 e0 8it 8" ex PERlOD IsY I5Y 15Y ISV 15Y I5Y I5Y 15Y 15Y - LOANS FOR SERVICED LAND DEPOSIT v VAL 27X 28" 28" 30S. 38x 38" 328' 'e2 48e' FEE i PAYZIENT 3S 3" 3Y 3'. 3Y 3M 3Y 3 32' 3x INTEREST RATE 8' ex 8" 8Y' SY 8Y 8 8S 8" PERIOD ISY 15Y s5Y ISY ISY ISY ISY ISY IsV COSTS (IL. RUPEES$ ON-COSTS APPLIED IN PRICE BUILD-UP (HIL.RUPS.) ROADS ) SP 38" 18.28 UATER SUP. 31w 8.20 BASE CONTINCENCIES DES*SUP INTERIn TOTAL SEWERAGE 8x 2.42 COST PHYSIC. PRICE ADIlIN MAINT FINANCE SANITATION 3" 2.34 151' ex 51 go 15# ELECTRICS. 12# 3.17 SOLID UST. 1 0.22 IAT.LOANS 8.88 8.88 BUILDINGS 78.65 9.e8 3.83 8.88 12.3- 84.88 TOT.INFRtA 1O9 26.63 --------- 26.63 3.83 9.e8 1.53 8.89 4.82 36.98 LAND R SeddMn& 82.54 --------> 02.54 12.38 84.92 PROFIT 9.88 LAND-jINFRA 198.17 (CCcsw4 c. 1887) TARGET FOR COST RECCOERY 226.88 1386 INCOGE DATA RUPEES PER hONTH PERCENTILES l' 2e# 38'. 4ex 58" 608 70'. 92" so8 TAKHAL PAYAN PSH 699 8e8 898 1888 1508 1988 2889 212 309ea ISLAnMAAD PSHNUR. 888 8e8s 1S8M 1288 1580 1888 2388 30c9 388e LJFP URBAN HHLDS 18se8 183e8 15 188e 21ee 25e3 3ee8 3208 5400 - 90- Annex 3 Attachment 9 SWATHI-SOUTH SITES & SERVICES CROSS-SUBSIDISED BY COMMERCIML PLOTS PKSHA;APR UCLOPnENT PLANNING PROGRAMME FOR GOUERNMENT OF PAKISTAN# UNOP L UNCHS BY PDA DIRECTORATE OF URBAN PLANNINS WITH GILMCRE HASKEY KIRKE Date: Ian. 1988 OPTIONS ttEAN 1 2 3 4 S 6 7 0 TOTAL SIZE: MARLAS 5 2 3 4 5 7 to 15 20 SO.METRES 123 50 75 103 125 175 250 375 5se NO. OF PLOTS 228 672 288 648 236 219 14 42 2247 X TOT. NUMBER 6.4 324 13, 2S-Y 11% 1" e I s 2" 1 S0 x TOTAL AREA 24 17 je"4 284 1 4% IS4 2% 7s soee LAND USE HECTARES PLOT AREAS 70n 0.64 5.04 2.98 8.10 4.13 5.'8 0.53 2.18 28.05 FACILITIES 16" 6.50 ROADS ETC. 14< 5.88 PROJECT AREA loom (ATTRIBUTABLE TC NEU USERS) 41.27 PRICING SCHEME RUPEES PER n.ARLA ?lIL.RUPS LAND ONLY 6345 2375 2375 3238 3238 4284 14333 14383 14383 73.33 INFRASTRCT 2472 S25 825 1262 1262 1668 S687 5607 5607 28.57 PRICE hMar 8818 3303 33e2 4522 4580 5553 22808 22000 22828. RUPEES PER UNIT PLOT PRICE 45348 66e0 SSe 188z8 22500 41671 220282 3228o2 4028e HOUSE tCON 42267 6e£a 6220 2ee88 25202 25WO 165202 22020o 55220a 84.98 MATRL.LOAN 8 0 0 0 e 0 0 0 e8 .00 TOTAL 87615 12620 15300 3a020 47508 66671 365220 5200eo ss8202 186.87 AFFORDAS!LITY ANALYSIS REA-DY CASH 23384 2520 3188 114Z0 14250 22221 109520 200e20 380222 TOTAL LOAN 64232 18282 12722 26622 33250 46672 25552* 312022 578080 REPAY1ENTS4M 462 99 125 262 328 462 2517 3274 5616 x N0r.E ROD 19Y 15t4 15" 15 2e2 22%x 384 3a4 324 INCC.E 41 2457 662 835 1747 1638 2299 8331 10246 18719 PERCENTAGES OF HCUSCHOLDS LHO CANNOT AFFORD TO REPAY LOANS TAKHAL PAYAN PSH 12"c 17% 58" 54X 82" 93Y 94-4 99x ISLA!¶ABAD FSH'R. 8Y lix 56X 53X 69X 53" 94X 90% NS'FP URBAN HHLDS 6"C .9x 30% 34-4 54" 92Y 83"4 98 HOUSING FINANCE ASSURPTIONS - LOANS FOR HOUSES AND nATERIALS DEPOSIT " VAL 27x 22" 22" 32" 38" 32-4 30. 40% 40" FEE 9 PAYM¶ENT 3%4 3% 3% 3% 3% 3" 3% 3% 3" INTEREST RPTE 8% 84 8e 8" 8e 8u 84 8" 8"x i PERIOD 15Y 1SY 15Y ISY 15Y ISY ISY ISY 15Y - LOANS FOR SERVICED LAND DEPOSIT x UAL 27% 202 208 38-4 3e2 32-A 3ex 42" 42" FEE S PAYMENT 3.- 3% 3% 3% 3- 34 3% 34 3% INTEREST RATE 8eX % 8" 8% 8" 0" 8% 8Y 84 PERiOD 15Y 15Y 15Y 15Y 15Y 15Y 15Y 15Y 15Y COSTS (MlIL.RUPEES) CN-COSTS APPLIED IN PRICE BUILD-UP (MIL.RUPS.) ROADS t SP 39% 18.28 WATER sUP. 31x 8.20 BASE CONTINCENCIES DEStSUP INTERIM TOTAL SEuERAGe 8" 2.42 COST PHYSIC. PRICE ADMIN r1AINT FINANCE SANITATION s" 2.34 15% ox * 5s Ox 1"3 ELECTRICS. 12" 3.17 SOLID LIST. I4 8.22 MAT.LOANS o.ea o.8a BUILDINGS 78.65 8.ee 3.93 0.00 12.39 84.98 TOT. INFRA 122" 26.63 ------> 26.63 3.99 0-e. 1.53 8.e8 4.82 36.98 LAND R 50001ra 82.54 --------) 8i2.54 12.38 94.92 LESS SUeSIDY -38.02 LAND-INFRA 103.17 (CCst.A D4c.1987) TARGET FOR COST RECOUERY 196.87 1386 IN'OltE DATA RUPEES PER MONTH PERCENTILES I814 20% 30" 40% 50" 68" 70% 00" 52" TAKHPL PAYAN PSH 6e0 9e0 soa lee1 1s5a 1800 28O 2leo 3000 ISLAMABfD PSHi:JP. 8se 1see 1eee 1208 1See 182e 2300 3282 390 t-FP URBAN HHLDS 1820 13e8 1522 18e0 2188 2520 3080 3300 54e0 - 91 - Annex 3 - Attachment 10 SWATHI-SOUTH SITES & SERUICES SELF-HELP INCREMENTAL HOUSE CONSTRUCTION AND CROSS-SUBSIDISED BY COIMMERCIAL PLOTS' PESHAWAR DCVELOPrENT FLANNING PROGRA.¶r- FOR GOVERN .ENT CF PAKISTAN# UNOP S UNCHS BY PDA DIRECTORATE OF URBAN PLANNIN3G U.TF GILrORE HANXEY KIRXKE Da'.: an. 1388 OPTIONS MEA¶N 1 2 3 4 5 6 7 8 TOTAL SIZE: MARLAS 5 2 3 4 5 7 10 15 2D SQ.M.TRES 123 5s 75 1e0 125 175 25e 375 sce No. OF FLOTS 128 672 26S 648 286 219 14 42 2247 u TOT. NUrMBER 6-' 3% I 33A¶ 23% 1 1% 10% 1 V 2% 182% * TOTAL AREA 2%4 17%4 le 28% 14's 19* 2% 7' lee% LAND USE HECTARES PLOT AREPS 74% 0.64 5.04 2.s8 8.10 4.13 5.48 8.53 2.10 28.8S FACILITIES 1 6% 6.50 ROADS ETC. 14% 5.88 PRO:ECT AREA loeA (ATTRIBUTABLE To N-U USERS) 41.27 PRICING SCt4ErE RUPEES PER IIARLA MIIL.RUPS LAND ONLY 6345 2375 2375 3238 3238 4294 14393 14393 14393 73.33 INFRASTRCT 2472 825 325 1262 1262 1663 5627 g627 5687 28.57 PRICE /Mar 9818 3qqe 33e0 4520 4502 5953 280oo 20280 20808 RUPEES PER UNIT PLCT PRICE 45348 6620 830o 1se80 22590 41671 208280 388228 40eQ80 HOUSE COCN 32359 8 0 a 8 252CO 165820 228288 550288 69.22 IIATRL.LOAN 3567 3828 3000 6202 6080 8 8 o 8.02 TOTAL 73273 86e8 12820 24808 28522 66671 365800 522802 958seo 178.13 AlFFORDASILITY ANALYSIS READY CASH 145e2 1328 1388 5420 6750 2eee2 128500 2e8$22 380020 TOTAL LOAS 64771 8280 18320 1862e 21750 4667a 255520 312e28 5728e0 REPAYMENTSeM 411 82 1e8 183 214 460 2517 3274 5616 * INCOcE ROD 18% 154 15% 15, 2e2 22% 32e . 30% 3e% INCCOE /rI 2184 544 717 1222 1271 2299 8381 18246 18719 PERCENTAGES OF HOUSEHOLDS UHO CANSOT AFFCRD TO REPAY LOANS TAKHAL PAYAN PSH 3% 13% 44% 41x 92% 93% 94% 398 ISLAMABAD PSHUR. 6, 8% 4c£4 33% 69% 93% 94% 9Sa NUFP URBAN HHLDS 5% 72 17%4 12% 54% 92% 93% 39m HOUSING FINANCE ASSUrPTIONS - LOANS FCR HOUSES AND MATERIALS DEPOSiT %UAL 7# 04 e% x 8-% e4 304 328 40% 4a# FEE X PAYMENT 3% 3X 3x 3% 3% 3% 3X 3% 3% INTEREST RATE 8% 8% 8%m 8x 8e e%A FX 8x 8e PERIOD XsY 15Y 1SY 1sY 15Y ISY 15Y 15Y ISY - LOANS FOR SERUICED LAND DEPOSIT X VAL 27% 2e% 2e% 3OX 302 32% 32% 48% 4ex FEE x PAY.ENT 3S 3%A 3% 3% 3* 3% 3% 3X 3X INTEREST RAT- 8% 8%4 8%A 9 8ex ex 8Y 8% 8x PERIOD .5Y 15Y 15Y 15Y 15r ISY 15Y 15Y ISY COSTS CIlL.RUPEES) ON-COSTS ARPLIED IN PRICE BUILD-UP (lIL.RUPS.) ROADS t SP 33% 10.28 WATER SUP. 31% 8.20 BASE CONTINCENCIES DES-eSUP INTERIn TOTAL SEJERAGE 39 2.d2 COST PHYSIC. PRICE FDIN rMAINT FINANCE SANITATION 3% 2.34 15X OX 5% e% 15% ELECTRICS. 12% 3.17 SOLID liST. 1% 2.22 MAT.LOANS S. 2 8.82 BUILDINGS 56.'S 8.e0 2.82 a.e2 e.98 6e.22 TOT.INFRA 1ee% 26.63 ------…> 26.63 3.93 8.e9 1.53 e.e2 4.82 36.98 LAND R 5=C/t:a 82.54 --------> 62.54 - 12.38 94.92 LESS suesIDY -32.80 LAND4INFRA e1.17 (Cost.ed Dec.1987) TARGET FCR COST PECC1JERY 178.13 19e6 INCOrE DATA RUPEES PER rONTH FERCENTILES l% 20% 3e' 40. - 50'4 6Z24 70e 80-4 9e% TAKHAL PAYAN PSH 60 3ee, se0 1000 15eO 18ee 2e80 21ee 3002 iSLArPEAD PSHL.R. eea 109e 1PC8 1200 1520 18e2 2300 3e2e 3ee0 I:mFP PROAN HHLDS 1eee 13e0 158e 189e 2120 25W8 3e2e 33ee 5400 - 92 - Annex 4 Page 1 of 4 PAKISTAN SHELTER SECTOR REVIEW Land Acquisition 1. The Land Acquisition Act of 1894 has been and remains. at least for rural land, the principal law in all the provinces, although for urban land, local laws and policies are used which have interpreted its provisions differently. This legislation is as follows: (a) West Pakistan Amendment Ordinance 1969; (b) Punjab Acquisition of Land (Housing) Act 1973 (repealed 1985) - Section 9; (c) NWFP Urban Planning Ordinance 1978 - Section 50; and (d) Baluchistan Acquisition of Land (Housing and Development Schemes in Rural Areas) Act 1974 - Section 7 and Baluchistan Ordinance No. 7 of 1976 - Section 6. 2. Land is to be acquired in return for fair adequate compensation. No single formula for determining fair and adequate compensation has been evolved, although "market value" is specified in Section 23 of the 1894 Act and Section 4(i) stipulates that it should be that prevailing on the date of publication of notification. 3. With the exception of Sind, provincial legislation has awarded compensation on the basis of two years average sales prices preceding the date of notification. (Islamabad has a more complicated formula for determining market value in order to take into account the effect on land values of the Master Plan of 1958, which has had a far greater impact than master plans for other cities). 4. In addition to so-called market value, both Sind and Punjab, increase compensation by 152 for land acquired compulsorily for public purposes or 252 for private (company) purposes, under the West Pakistan Amendment Ordinance 1969. 5. Major objections have arisen where ceilings have been put on valuations. The maximum amounts of compensation payable in Punjab and Baluchistan have amounted to Rs 20,000 and Rs 50,000 per acre respectively and overall totals of Rs 100,000 and Rs 500,00 respectivel- If compensation entitlement ir Baluchistan exceeds Rs 500,00 -e balance is payable in installments bearing 6X interest. - 93 - Annex 4 Page 2 of 4 6. It is understood that land owners around Peshawar prefer compulsory acquisition to negotiation for urban land purchase because of the higher than market prices obtainable under the Provincial Urban Planning Ordinance of 1978. This may be due to the opportunities for falsification of records. One major disadvantage with the now in force 1894 Act is that the acquiring authority is obliged to deposit the full value of the land early in the acquisition process which is a major deterrent to private sector land development 7. The actual procedure for acquisition varies slightly. In NWFP, for example, where land ownership records were last updated over 60 years ago, the procedure under the 1978 Ordinance is as follows: (a) The land acquisition collector of PUDB is notifiur of land identified for acquisition. (b) The Collectorate identifies khasra numbers and issues Section 50 notices which notify owners of the Government's intention to buy and forbids sale or development of the land. This also allows the Government to survey and stake out the land. (c) The Collectorate reviews revenue records for the areas where land is to be acquired to obtain cos' for recent tran3fers of the land in the same area. (d) When the Government wishes to take possession of the land, Section 52 notices are issued stating this and defining areas to be acquired. At this time the collector invites claims for compensation from owners. (e) The price is fixed based upon: - market value determined on basis of two years average sale; - certain damages sustained by owners during acquisition; and - any reasonable expenses incurred as a result of acquisition. 8. The 1973 Punjab Acquisition of Land (Housing) Act with its related Exemption Policy was an interesting and constructive experiment until its repeal in 1985, mainly caused by opposition in the courts over low compensation and long delays. Its provisions are still being applied for such LDA projects as M A Johar Town - still at the planning stage and other projects now further advanced. - 94 - Annex 4 Page 3 of 4 9. In essence, the Act provided for compensation to be in the form of one or more serviced plots amounting to 302 of the expropriated landholding, with servicing costs offset against the compensation payable. The "exempted plots" are then disposed of by the owners. These have been cases where owner attempt to negotiate at 40% of the holdings returned to them. 10. This enforced partnership between public authority and private sector appeared to have the merit of assembling many small landholdings without the heavy up-front costs of expropriation. It was hoped that land for housing would be acquired in a timely manner, and each of the many individual landowners, with inadequate capital for undertaking development themselves, would be compensated by saleable assets. The Authority, meanwhile was expected to be able to generate a profit from the sale for housing, commercial and institutional plots from the balance of the land. This was, of course, after allowing for non-marketable uses, such as roads and public open space and, sometimes, cross-subsidized low-income plots. 11. In contrast to the theory, actual experience with the Exemption Scheme has turned out rather different. Small landowners found themselves deprived of their livelihood for at least two years, without compensation, while large landowners were quick to exploit the loopholes in the scheme to maximize their entitlement to the five standard sizes of plots (see Table 1-Annex 4). 12. It is understood that this highly skewed breakdown resulted from many owners identifying the 105 square meter plot entitlement as potentially the most profitable, subdividing their land accordingly into areas of about 280 square meters and selling it to proxies. The large number of 420 square meter plots arose automatically from larger landholdings of up to 2.1 hectares which could attract a maximum of fifteen 420m2 plots. 13. It was taken for granted by all landowners that prevailing development standards would be maintained whereby 1055 sq. m. plots would front onto nine meter roads and be attractive to middle income purchasers, whilst 420 sq. m. plots would front onto 45 meter roads with divided highways, service roads and green strips etc. 14. These development standards, together with the prevailing standards for schools and other community or commercial facilities, more than absorbed the 70X of the land for the public authority's own use. Thus the object of the scheme was frustrated. 15. Since the repeal of the 1973 Punjab Acquisition Act in 1985, no further attempts have been made by the Lahore Development Authority at land expropriation for housing and no consensus appears to exist in the Punjab's on acquisition policy. - 95 - Annex 4 Page 4 of 4 16. One idea under consideration as a basis for negotiated land acquisition by the Lahore Development Authority would be to use a modified version of the previous Exemption Policy coupled with more efficient planning. Given the innovative nature of the public/private land development schemes, it would be worth while redesigning the schemes with more efficient land use and pricing. Some recommendations that might be considered to improve the process include: (1) A time limit of not more than six months should be fixed for the completion of acquisition proceedings. A special tribunal headed by a person not below the rank of District and Session Judge should be set up at provincial levels to decide the appeals against the acquisition proceedings within a period of not more than six months. The decision by the special tribunal should be final. (2) Due to rapid increases in the cost of land and difficulties in assessment of average cost, development agencies, while preparing plans for guided development, should acquire strategic parcels of land in advance of the actual development. Annex 4 Table 1 PAKISTAN SHELTER SECTOR REVIEW The scheme provided for plots to be allocated is as follows: Exemgted Plot Allocation Table Original Plots area entitlement Plots to be (Maximum Landholding under the Exemption Policy Allocated area gain oossible) below 10 marlas cash compensation only none 10 - 13.3 " 3-4 marlas (63-84m2) 72m2 (14X) 13.3 - 20 0 4-6 " (84-125m2) 105m2 (25X) 20 - 26.7 U 6-8 * (12S-167m2) 160m2 (28X) 26.7 - 30 a 8-9 (167-l88m2) 72+105m2 (6X) 30 - 36.7 0 9-11 * (188-230m2) 105+105m2 (12%) 36.7 - 43.3 " 11-13 * (230-272m2) 250m2 (fl) 43.3 53.3 a 13-16 " (272-334m2) 72+250m2 (18X) 53.3 - 60 u 16-18 * (334-376M2) 105+250m2 (6X) 60 - 66.7 0 18-20 " (376-418m2) 160+250m2 M9X) 66.7 - 76.7 0 20-23 (418-481m2) 420m2 (0) 76.7 - 80 w 23-24 * (481-502m2) 72+420M2 (2M) 80 - 1000 24-300 ( S02-6271m2) IS x 420m2 maximum In the case of Johar Town (Phase II) in Lahore, for instance, the number of exempted plots were established as follows: 72 566 105M2 7016 160.2 481 250m2 442 420M2 1206 Total 9711 - 97 - Annex 5 Page 1 of 5 PAKISTAN SHELTER SECTOR REVIEW Institutional Structure and Program Finance 1. The shelter sector programs of the Government of Pakistan involve three tiers of Government: the Federal Government determines shelter sector policies and provides development finance from general revenues; the Provincial Government organizes programs, implements schemes where local agencies lack the capacity and, currently, provides program resources in the form of state-owned land and ADP allocations to cover much of the development costs; and local agencies (local government departments, development authorities, water and sanitation/sewerage authorities), especially in urban centers, undertake works including infrastructure provision, and operate and maintain services which they finance from tariffs and local taxes. Structure of Local Government 2. Pakistan has a federal system of Government under the 1973 constitution, with considerable responsibilities devolving to the provinces. The four provinces of Baluchistan, Northwest Frontier Province (NWFP), Punjab and Sind together with other federating units (Azad Kashmir and Federally-administered Tribal and Northern Areas) are divided administratively into 18 divisions, and in turn 84 districts. The districts are further divided into 276 tehsils or taluqas. The current numbers of local bodies in Pakistan is shown below, broadly divided between urban and rural. Distribution of Local Bodies in Pakistan Balu- Total All Punjab Sind NWFP chistan Pakistan Urban Councils Municipal Corporations 7 3 1 1 12 Municipal Committees 63 24 11 10 108 Town Committees 131 94 21 17 263 Rural Councils District Concils 27 13 13 16 69 Union Concils 2240 562 434 176 3412 TOTAL 246 £2EL 48 Z38AA - 98 - Annex 5 Page 2 of 5 For urban areas, Town Committees are established on the folloving basis: (a) Punjab--5,000 to 20,000 population; (b) Sind--up to 25,0(v population; (c) Baluchistan-5,000 to 10,000 population (no criteria in NWFP). Municipal Corporations and Committees are formed as follows: (a) Punjab-population of 20,000 to 500,000, above the latter becoming a corporation; (b) Sind--up to 500,000, above which a corporation; (c) Baluchistan-between 10,000 and 100,000, above the latter becoming a corporation; (d) NWFP have no criteria. For rural areas, each district is administratively under a Deputy Commissioner or political agent, appointed by the civil administration, whose functions are concerned with revenue, administration and maintenance of law and order. At district level, developmental responsibilities are with the elected district, union and urban councils. Provincial sectoral departments have representation at district level, and report to district coordination committees. 3. Social sector programs (such as shelter and related services) are primarily a provincial responsibility; implementation however is through or by the elected councils and assemblies, with the federal government formulating policy and coordinating, or through provincial level departments. 4. Local Government and Rural Development (LGRD) Departments are responsible at provincial level for the district and local government administration system; the Provincial Planning and Development Boards evaluate public investment programs and control the allocation and management of the provincial Annual Development Program (ADP) funding allocations. Provincial Housing and Physical Planning Departments (H&PP) have responsibility as lead technical departments for urban and rural public services, either as agents for local government bodies or acting as provincial level housing development departments. 5. In rural areas, the functions and powers delegated to the Union and District councils are of two broad categories: (a) implementation of development activities (including project preparation and land acquisition) and (b) development program formulation planning, and monitoring. For this, they have delegated powers to: (a) levy taxes; (b) approve budgets; (c) to approve development projects within limits. In urban areas, town, municipal, and corporation committees have broad delegated responsibilities, including provision and maintenance of public works, services, welfare and primary education. The urban committees have power to levy taxes, rates and fees on buildings, municipal services (including water supply and drainage sewerage), vehicles, and on the export and import of goods from local areas. Shelter Sector Respcnsibilities 6. The Federal Ministry of Housing and Works (FMHW), including the Environment and Urban Affairs Division and the Housing and Physical Planning Division, is the Federal Ministry with responsibility for the shelter sector. In August 1987, the National Housing Authority (NHA), an autonomous - 99 - Annex 5 Page 3 of 5 federal organization, was formed with the objective of implementing the "Shelter for the Shelterless" component of the Prime Minister's 5-Point Program. The first phase of this program, unlike the other shelter programs under the FMHW which focus on the provision of services land, is concerned in the first instance with the development of housing superstructures. The FMHW and the NHA, together with the House Building Finance Corporation (see Section VI) constitute the focus at national level of shelter programs in Federal Government. The NBA has designated existing provincial H&PP departments to act on its behalf at provincial level (see below) while the HBFC has a widespread provincial network of branches. It is not intended to set up independent NBA units at provincial level, but rather to use the existing HP&P organizations. Provincial level katchi-abadi directorates have also recently been set up at provincial level, generally within the existing LGRD departments--in some instances, municipal corporations (e.g. the Karachi Municipal Corporation) have also created katchi abadi department (overall coordination and monitoring is with the federal katchi-abadi cell). 7. Implementation of shelter programs is managed by provincial Governments. Each Province makes its own organizational arrangements, but typically the provincial departments involved in shelter sector programs include the Department of Housing and Physical Planning (in Punjab this includes Environment), the Department of Local Government and the Board or Revenue (which manages state land holdings and title registration). 8. In Punjab, the Housing, Physical and Environmental Planning Department (HPEPD) develops residential land through its Housing and Physical Planning Department (HPPD) and is the provincial department which also oversees the activities of the Development Authorities and Water and Sanitation Agencies, which undertake this work in the major urban centers (see below). The Secretary (HPEPD) has also been designated as the senior official representing the National Housing Authority at provincial level; HPEPD is tne provincial counterpart of the NHA. The Local Government and Rural Development Department (LGRD) is responsible for local government administration. It also has a Directorate of Katchi Abadis which coordinates and adm:nisters the Katchi Abadi improvement program. (Program execution is also undertaken by Development Authorities and municipal councils). The Board of Land Revenue administers all transfers of state land and manages title registration of privately-owned land and therefore has been closely involved, through the District organization of Deputy Commissioners, in rural shelter programs. 9. Development Authorities (DAs), which are semi-autonomous agencies of Provincial Government, and in limited cases municipalities execute shelter and area development programs in major urban centers. They have not to date been involved in large-scale construction of housing units except for limited numbers of flat (apartment) units in Lahore and Karachi; DAs have been formed in Karachi, Hyderabad (Sind); Multan, Lahore and Faisalabad (Punjab); Quetta (Baluchistan); and Peshawar (NWFP), with responsibility - 100 - Annex 5 Page 4 of 5 within the respective metropolitan municipal area for development planning and implementatioa, including area development ('tousing and commercial), katchi abadi and slum improvement (upgrading), building control, road development and traffic engineering and water supply, sewerage and drainage. They are set up as autonomous authorities within existing departments (HPEPD). 10. It was on this basis that under the 5PP, the Boards of Revenue and the Civil Administration were allocated responsibility for non-urban constituents (the 7 marla schemes) and the HPPs and DA responsibility for the 3 marla urban schemes. 11. The Zahat Committees have no shelter experience or responsibility, and have become involved principally as a socially-oriented short-term measure to bring more resources to the sector. 12. The above highlights the many involved parties, and the need to ensure project and policy responsibilities would be clearly agreed at the start of a strengthened program. Public Finance 13. Public finance in Pakistan is arranged so that lower tiers of Government rely on subventions from the Federal Government, particularly for development expenditure. Provincial governments are especially reliant on subventions from the Federal Government: in recent years Provinces have required deficit finance for current account expenditure as well. Annual Development Programs (ADPs) are virtually wholly financed by federal loans, with the balance of financing coming from provincial contributions, foreign loans and federal grants. (In the 1987/88 budgets, shelter sector grants for Katchi Abadi and 7 Marla programs are significant amounts of the total Federal Development Grant allocation to the provinces.) This takes no account, however, of the contribution being made to shelter programs by the provinces through the provision of state land to shelter shemes. This land is transferred from existing holdings, and it does not appear as a financial transaction in the cash-based province accounts. 14. Urban development authorities finance a substantial proportion of their urban development programs from sales of developed land. (In Punjab in 1986/87, the proportion of total development finance derived in this way was 43%.) In addition, development authorities execute provincial programs, some of which achieve partial cost recovery from beneficiaries as a supplement to federal grants. In certain provinces, development authorities are also responsible for water and sanitation in large cities.. Other provinces, use autonomous agencies responsible to the Department of Local Government. Water and sanitation agencies are partially financed from water tariffs, although system development is usually funded through ADP loans. In recent years the failure to revise tariffs has placed many of these agencies in a poor financial position. - 101 - Annex 5 Page 5 of 5 15. Municipal Corporations finance the operation and maintenance of a number of urban services (road maintenance, drainage, street lighting, primary school buildings) from local taxation and fees. Capital account schemes undertaken by the corporations are financed mainly from local sources, and Federal/Provincial grants and loans. Financial Management 16'. The financial management practices of the provincial and the semi-autonomous bodies differ. Provincial departments form part of the regular budgetary process of Provincial Government. Their expenditure is financed by budget allocations, and any income they receive is paid into consolidated province revenues. The financial information produced by provincial governments show stewardship of resources rather than information for management. 17. The semi-autonomous agencies, on the other hand, are more self-financing and their income is retained either to fund operational costs or to invest in development. For land development, this presents the opportunity to reinvest and to revolve funds. Due to their more commercial orientation, the semi-autonomous agencies have considerably better developed financial reporting systems than Provincial Government. 18. Local Government financial management is centrally regulated. This determines the style of their cash-based accounting and reporting. The annual budget is the principal instrument of financial planning and control. The financial systems are effective in controlling expenditure but do not form an adequate base for policy making and reporting. IBRD 16248R1 64' 72- U. S. S. R. j j_4CH I NA PAKISTAN * NATIONAL CAPITAL o CITIES AND TOWNS - NATIONAL ROADS .IPRIMARY AND SECONDARY ROADS A ) RAILWAYS ......... ...... 4 AIRPORTS - \ Lire Corwro/ INTERNATIONAL BOUNDARIES RIVERS walp,ndi JAM CM KASIR 32' 32( j A F GHAN ISTAN J 1 htt ' -_._*_._ _._/ g~~~~~~~~~~~~ N DIA -28 28- ISLAMIC L ) r3uy REPUBLIC OF L. IRAN _ kku 1/0 200 300 400 r ( nWv d- of 7 A. BW24 fto a~~~~~~~~~~~~~~w W VWI WOOl W FIWW -24- dW ; .:- - 0. 24- e*lww,mwl a, 06ccptaoco C SW0dbkf 6fl ~~ 7. °If72c 64' 6M 72- MAR04 1987