DISCUSSION PAPER Report No.: UDD 19 HOUSING AND FINANCIAL INSTITUTIONS IN DEVELOPING COUNTRIES by Bertrand Renaud December, 1982 Urban Development Department Operational Policy Staff World Bank The views )resented here are those of the author, and they should not be interpreted as reflecting those of the World Bank. ABSTRACT This paper presents an overview of issues arising in the development of housing finance institutions in developing countries. The housing problem is defined as the necessity to reconcile the three partially conflicting objectives of affordability for households, viability for financial institutions and resources mobilization for the expansion of the sector and of the national economy. The impact of the dynamics of the housing market and of the capital market on the growth prospects of housing finance institutions are examined. It is shown that only a minor proportion of annual housing investment (much less than 20% in the best cases) comes from financial institutions. The causes and consequences of this limited reach of housing finance institutions are discussed and some of the major issues encountered in the development of housing finance systems are identified. Much emphasis is placed on the need for governments to realize that they have two major but distinct responsibilities in the sector: the development of financially viable institutions and the provision of minimum basic urban services to the entire residential sector. Table of Contents HOUSING AND FINANCIAL INSTITUTIONS IN DEVELOPING COUNTRIES Page No. I. INTRODUCTION........................1 A. Objectives of this Paper.g..... e.g...... *.......1 B. What is the Housing Finance Problem?...................... C. Current Context of Housing Finance Work...................4 D. Organization of the Paper. .......6 II. THE URBAN ECONOMY, THE RESIDENTIAL SECTOR AND HOUSING FINANCE............................. . .........8 III. DYNAMICS OF LDC HOUSING MARKETS AND CONTRIBUTION OF FINANCIAL INSTITUTIONS...oooooo*eooeoooooe*oooeooo oe*@. .....13 A. Introduction.oo............oooo oooe-e.....o*013 B. Three-Tier Structure of LDC Housing Markets..............13 C. Illustrations: Tunisia and Other Country Experiences....14 1. The Tunisia Housing Market...........................14 2. Limited Reach of Financial Institutions in other Countries ... ...1 D. Progressive Housing Investment in the Informal Sector....23 1. Policy Questions Raised by the Informal Sector.......23 2. Progressive Housing in the Informal Sector...........23 3. Non-Financial constraints on the supply of housing...27 E. Implications of the LDC Market Structure for Housing 1. Multiple benefits of institutional housing finance...28 2. Policy Implications..oeog * .. ... gggg.0.0gQ... 0.30 3. Financing Low-Income Housing..... ........31 IV. HOUSING FINANCE INSTITUTIONS AND CAPITAL MARKETS.............36 A. Introduction. .. .....36 B. Level of Economic Development and Supply of Financial Resources.... eeoeoeeoeoo .............o oooo............37 C. Resources Mobilization and Constraints on Housing Finance... o o , oo...o. . , . o .. , , o ;o . o e o 41 1. Household income and savings....... **&o*....... 45 a. Income Level, Income distribution and Savings....45 b. Ownership Opportunities and Saving Response......47 c. Employment Conditions and Income Stability.......51 2. Resources Mobilization and Housing Policy............53 3. Benefits of Positive Real Interest Rates.............56 D. Some Policy Implications....... .*******oo**o*ooo......61 V. DEVELOPING A HOUSING FINANCE SYSTEM...... ..... 63 A. B. Financing Housing Compared to Other Sectors..... .....64 C. Structure and Functions of Housing Finance Systems.......65 1. Some Determinants of Institutional Structure.........65 2. Four Country Profiles.....0... 000000000. ........70 3. Functions of Housing Finance Systems.................74 D. A Consolidated View of Housing Finance Policy Issues....75 E. Sources of Funds: Issues in Resources Mobilization....80 1. Major Issues for Voluntary Schemes .. ........80 a. Real Interest Rates.... ooo...o...v.o.o*e 80 b. Inflation and Indexation. .......................81 c. Contractual Savings Schemes....,,*oove*aee*ez..83 d. Institutional and Corporate Deposits............85 e. Branching and Quality of Services.ooooe........85 2. Issues in Mandatory Schemes... ...oo*ooooo*. *.....86 a. Schemes Imposed on Individuals....oo o........87 b. Mandatory Schemes for Businesses................91 c. Regulations imposed on Financial Institutions....92 3. Government Transfers to Housing Finance..............93 a. Borrowing from the Central Bank..................94 b. Budget Allocations to Housing....................95 c. Tax Exemptions and Tax Expenditures.,............95 F. Uses of Funds** ......96 1. Conventional Lending... os***oo*so.oo.. o. .o....96 2. Capital Investments in Subsidiaries and other G. Critical Factors for Housing Finance Systems.............99 1. Policy Environment and Government Regulations........99 2. Income Level and Development Opportunities..........103 3. Financial Deepening and Specific Low-Income Housing 4. Encouraging Appropriate Institutions.,ooeosoo ......107 VI. CONCLUSION:.. . o... . ..... . o ... .. . ... o.. .109 A. Appropriate Housing Finance Strategies....eee......109 B. Areas for FurtherAnalysis..... .........112 SUMMARY Analyses of housing finance in developing countries are frequently limited and fragmentary. As a first step to remedy this tendency toward partial treatment, this paper presents an overview of the various issues that are frequently arising during the development of housing finance systems. The housing finance problem is defined here as the need to reconcile the three partially conflicting objectives of affordability for households, viability for financial institutions and resource mobilization for the expansion of the sector and of the national economy. The dynamics of the residential sector in LDC's leads typically to a three-tier market structure: (1) a high to middle-income "legal" private sector, (2) a heavily subsidized, mostly middle-income public sector and (3) a large, rapidly growing low-income "informal" private sector which is far from disorganized and forms residential zones that are developed at high density without infrastructure or community facilities. While the annual output of the housing sector is quite significant, housing finance institutions play a limited role and finance much less than 20 percent of annual housing investment in the best cases. Their limited reach is confined mostly to the two top tiers of the market, especially the second tier. While there a-a important non-financial constraints on the supply of housing, housing finance institutions should play a central role in structuring the very large annual volume of housing investment that takes place without any institutional support. They can improve the mobilization of resources in financial form because homeownership is the highest priority in asset formation for most households. On the investment side they can help in promoting sectoral objectives by financing the more desirable types of projects and by providing resources on a scale large enough for the production of infrastructure and services that individual households in the informal sector cannot provide for themselves. The question is how one can move from one-shot financing operations by groups of relatives and friends typical of the informal sector to the institutional financing of housing. If there are large, unmet financial needs on the housing market, the growth of housing finance institutions is conditioned by the level of development of the capital markets which are themselves closely associated with the level of development of a country. Compared to other sectors, the financing of housing presents specific problems: it involves households instead of businesses and this leads to substantial risks and high transaction costs compared to the size of loans; it requires a significant intermediation efforts because housing finance institutions are expected to lend long even though they are borrowing short; also, because of this extensive term intermediation housing finance is extremely sensitive to inflationary environments. The single most significant flaw of the housing finance institutions currently confined to serving the second tier of the housing market is their limited ability to mobilize resources effectively. In particular, public sector financial intermediaries are too frequently operating as simple development funds allocating public resources mobilized elsewhere. Even if it is possible to present a generalized diagnosis of the housing finance problem in developing countries it is far from obvious that there is a single model for the development of a housing finance system. Plans for the rationalizatioi of the housing finance sector are dependent on national priorities regarding overall financial development including macro-economic policies, priorities in the allocation and pricing of financial resources, and national preferences regarding institutional development. What should be clearly kept in focus is that LDC governments have two distinct roles to play regarding the financing of housing. The first role is to encourage financial innovations for the provision of housing finance to households on a financially viable basis. This should increase the reach of institutions into the upper income part of the informal sector, especially for self-employed households. The second objective consist in developing specific programs for the lower income households who could still contemplate some form of home-ownership with a major focus on urban infrastructure and services. This second role goes beyond housing -finance proper and overlaps with the broader issue of financing urban development. It requires a more thorough understanding of the role of subsidies and cross-subsidies in LDC housing than is presently the case. This overview of housing finance issues in developing countries closes with a listing of areas where further work is needed regrouped under five headings: housing finance system development, resources mobilization, flexibility in housing finance, financing low- income housing, and improving institutions and the production of I. INTRODUCTION - A. Objectives of this Paper This paper outlines a framework for the review of housing finance issues in developing countries. Experience shows that discussions of problems of housing finance and the examination of possible ways to finance housing are too often fragmentary. As long as the scope of a financial problem is confined to a specific publicly supported project and to the question of how best to finance a limited number of housing units, a partial approach taking existing financial conditions as given can be quite adequate. But as soon as concerns become sectoral and imply a series of projects, the type of lending actions and t'e types of institutions that should be encouraged become important issues requiring more comprehensive evaluations. To remedy this tendency toward partial treatment, a first step is to give a comprehensive overview of the various issues that are frequently arising during the development of housing finance systems. Thus this paper has two main objectives: first, it delineates the scope of housing finance work and clarifies its nature; second, it identifies the main policy and operational housing finance issues in developing countries. In the necessary trade-off between comprehensiveness and depth of analyses the emphasis in this paper is on identifying the main issues and how they- relate to one another and on presenting a more systematic way to think about housing finance in developing countries. -2- B. What is the Housing Finance Problem? There is more than one way to look at housing finance and according to the viewpoint chosen significantly different answers are given to the question of what constitute "the" housing finance problem. From the viewpoint of a household the problem is the possibility of obtaining a loan at affordable terms. For ministry of housing officials the problem is the lack of resources to carry out public housing programs. For international agencies with a mandate to make loans which will reach the lowest 50 percent of the household income distribution, the problem is to develop sustainable financial programs for low income housing. From the viewpoint of ministries of finance and Central Banks the problem is to prevent financial instability and to maintain confidence in the financial system. National pla.nning agencies are interested in the contribution that housing finance can make to the mobilization of resources and their effective use. In the case of a capital market analyst the housing finance problem is that of term transform-tion in an inflationary environment. This paper examines how these various definitions relate to-one another and looks at housing finance institutions as financial intermediaries. The focus on financial institutions has immediate implications. First, as their name would suggest, financial intermediaries constitute the interface between on the one hand the housing market and on the other, capital markets. This is not as trite a reminder as may seem at first sight considering that one of the -3- recurrent problems of policy discussions and consequently of programmatic activities is that condicions on one or the other of the two markets are being ignored. It is a common occurrence to read a housing finance sector review which details the financial institutions operating in a country but provides little or no indication of the types of households being served nor of the exact role of institutional financing in the total housing supply in the country. Conversely, many LDC housing sector reviews deal with the structure and behavior of the housing market and tend to treat housing finance issues in a limited and superficial way. Most housing sector reviews are essentially concerned with the imbalance between the growth in the number of households and the number and types of units produced annually. Their discussion of financial problems focus on affordability and on the difficulties that households face to achieve home ownership. They tend to ignore institutional issues in the development of financial services and the constraints that the level of development of national capital markets imposes on the type of institutional financing which can be developed. What they consider as housing finance is limited to the development of financial techniques that could make housing project affordable to beneficiaries, basically in five ways: by lowering interest rates, lengthening maturities of 7mortgages, using graduated payments, increasing the equity base through - various subsidies or cross-subsidies, reducing the cost of houses by lowering standards or a combination of all these. Interest in the -4- housing finance system as whole is limited to the question of how more resources can be made available for low-income housing. 1- Another important point concerning housing finance is that, from a policy perspective, LDC governments have two distinct responsibilities. The first one is the regulatory responsibility of encouraging the development of viable housing institutions and maintaining the credibility of the banking system. The second one is to address effectively the needs of low income households and in particular to find ways to finance their housing needs. An overview of housing finance in developing countries cannot be limited to a discussion of the financing of low-income housing, it must also consider the problems of developing. effective financial intermediaries. The broad definition of- the housing finance problem which underlies this paper is how to reconcile three partially conflicting objectives: affordability for the households, viability for the financial institutions and resource mobilization for the expansion of the sector and the national economy. C. The-Current Context of Housing Finance Development Developing and developed countries alike are presently going through extremely difficult economic conditions. Following the two consecutive oil shocks of 1974 and 1979,. many countries have been suffering from a deleterious mix of inflation (often very high), rising 1/ One of the best treatments of this limited approach is found in the book by N.O. Jorgenson, Housing Finance For Low-Income Groups, General Printers, Ltd., Nairobi, Kenya, 1977. 240p. -5- energy costs, sluggish growth of export markets with the volume of world trade actually shrinking for the first time in many years, deteriorating terms of trade, high current accounts deficits and debt repayments that sometimes are reaching half of export earnings. Anticipating an extended period of scarce capital resources, effective investment strategies and a practical framework for domestic development are of critical importance. Good planning and institutional reform to utilize domestic resources more efficiently are important to the housing sector which must often compete with other sectors receiving higher priority from the government. However, the possibility of inducing net real savings mobilization through the encouragement of home ownership as an additional reason for encouraging the growth of housing finance systems remains an unsettled issue. In.spite of a more difficult economic environment, the urban population of developing countries will continue to grow at very high rates over the next twenty years. Because so much of the most dynamic economic activities takes place in cities, an efficient urban environment is necessary to improve the competitive international portion of a country as well as its domestic economy. As a part of a more efficient and more effective financing of urban development the development of a strong and adaptable system of housing finance capable of meeting the expanding needs of the urban economy is important for several concurrent reasons: -- residential investment is directly tied to urban infrastructure investment and the internal efficiency and productivity of cities; -6- -- in most developing countries a very high proportion of residential investment is provided through private individual initiative without the support of effective and responsive financial institutions. The existing institutions are typically serving a minority of households at a high cost with limited results. The informal financing of housing generally leads to major gaps between housing construction proper and the needed complementary urban infrastructure to the great detriment of urban development; - meeting the housing needs of a large low-income labor force will require sustained efforts and stronger institutions; - the fragmented nature of housing finance within the informal sector and the narrow reach and. effectiveness of existing institutions reduces the mobilization of domestic savings in the form of financial assets from most of the household sector of the economy; - to improve the potential of their housing finance system, LDC's need a clearer understanding of the various alternatives open to them both in terms of viable institutions and of effective technical tnnovations. Their development. plans for a housing finance system cannot ignore either the constraints affecting the entire financial sector or those related to the operatiom of the housing market. D. Organization of the Paper To clarify further the meaning of housing finance used in this paper and to show in what manner housing finance can be distinguished from other forms of urban development finance, Part II discusses briefly the relation between the urban economy, the residential sector and -7- housing finance. Then Part III discusses the dynamics of the residential sector in LDC's, it outlines the typical patterns of housing investment and shows the scope of institutional housing finance. It asks the questions of why institutional financing represents such a small proportion of total housing investment and of whether housing finance constitutes the main constraint on expanding the supply of housing. If there are other constraints, how do they interact with financing? Having considered the first of the two markets in which housing finance intermediaries have to operate, Part IV turns to the capital markets. It focuses on three questions: What is the relationship between the level of economic development and the growth of financial markets? What are the typical obstacles to mobilizing household savings in financial forms? What are the constraints that the capital markets imposes on the growth of housing investment and on the financing of low income housing? Given the focus on institutions, Part V is the longest and the most important. It examines the factors that distinguish housing finance from the financing of other sectors of the economy. It presents. a consolidated view of the wide diversity of policy issues that are typically encountered in LDC housing finance, and examines the main ones in mobilizing funds and in using them. The conclusion reviews the questions encountered in the paper and makes a provisional attempt at identifying the "entry points" for -8- the improvement of housing finance systems in developing countries, given the two distinct policy responsibilities that governments have to meet. -9- II. THE URBAN ECONOMY, THE RESIDENTIAL SECTOR AND HOUSING FINANCE The urban economy is an essential part of the national economy: the leading economic sectors, innovations, and flows of goods and services circulate through the network of cities and the efficiency of its cities is vital to the competitive position of a country's economy. Even in low income countries like India where only 24% of the population is urban and about 76% of the labor force is in agriculture over 60 percent of GDP originates outside agriculture. The economy of a given city can be conveniently viewed as divided into two major components: its economic base and its residential sector. The economic base consists of all the business firms (large and small) that are generating income for the residents of the cities. Some of them are oriented towards exports, others are producing intermediate goods for these export firms or are serving the needs of local residents. In particular, construction, utilities, and service firms have a dual role: they provide the infrastructure for the firms and the housing and neighborhood services of the community. It is through the provision of infrastructure and labor- services that the interactions between the economic base and the residential sector of a c.ty have a major impact on the competitive position of the city. Inadequate infrastructure, e.g. inefficient urban transport systems and interregional terminals, poorly designed road networks, the lack of water and sewage facilities, unreliable utilities, all limit the growth potential of firms on which they impose serious - 10 - costs. And the greatest majority of businesses, even the majority of manufacturing firms, numerically as well as economically, cannot insulate themselves from these problems through location in industrial estates. Because there are occasional misunderstandings in the discussion of urban financial problems, it must be clear that housing finance in this paper is limited to the financing of the residential sector and does not cover the broad variety of problems arising in the financing of urban development. The priority given here to the residential sector covers the financing of land, infrastructure and housing and reflects the need to improve the design and implementation of urban operations to meet the rapid expansion of LDC's shelter needs for the majority of low-income households. The question then is how it is possible to improve the way the household sector can finance its own urban requirements By contrast the totality of,urban development needs includes,. in addition to housing and the residential sector, employment and industry, offices, shopping and commercial facilities, transportation, health and social services, and other local services including leisure and recreation. These non-residential urban investment requirements typically involve different financing.arrangements and cost-recovery methods. To put it differently, if primary urban infrastructure relates to the needs of the entire city (for instance water reservoirs and water mains), secondary infrastructure relates to district level investment and tertiary infrastructure to neighborhood facilities and dwelling - 11 - connections, the residential sector is or could be directly accountable for tertiary investment and could share the cost of secondary and primary urban infrastructure with commercial and industrial users of urban facilities. The multiplicity and complexity of financialarrangements between local government and central government, the public and the private sector, the resources involved and the much longer planning horizon generally required have lead to a great variety of financial arrangements for non-residential investments which deserve separate treatment. A comprehensive accounting of total urban investment as defined above has been found unfeasible because most of it is more easily recorded under industrial sector activities or lumped together with other accounts in government investment programs. On the other hand, national accounts do report gross investment in residential construction exclusive of land which is a non-depreciable asset, and of part or all of off-site services investment which is accounted for under "other construction" or "social overhead" in national accounts. There is a non-linear relationship between the share of housing in GNP and the level of income of a country, the range being about 1.5 -percent of GNP at the lowest levels of development to a maximum of about 7 percent for some middle and upper income countries which are urbanizing very rapidly. Given the significant infrastructure and land components missing from these estimates the financial resources required by the residential sector are seen to be quite important. - 12 - Thus, housing finance is directly related to the residential sector and indirectly to the provision of the complementary neighborhood and city-wide infrastructure which individual households cannot provide for themselves. But the role of the housing finance system does not stop at the provision of shelter, it is also very important to the mobilization of domestic savings. In most developing countries, the household sector provides between 40 and 60 percent of domestic savings, the rest being provided either by businesses* or the public sector. -. Shelter being one of the first priorities of households and a leading reason for savings, the housing finance sector can and should play a central role in the mobilization of resources by households. Some of these savings can then be used for the immediate production of new urban infrastructure while the balance can be used to expand the directly productive base of the economy, providing employment to urban and rural populations. Housing finance systems in LDC's are seriously deficient in coverage and in quality and do not fulfill their three major functions toward the residential sector which are:. to improve financial. discipline within the sector and use resources moreeffectively, to stimulate efficient technology and appropriate planning-by making- ' resources available to innovative suppliers and. to mobilize household. savings in financial form. 1/ See for instance V.V. Bhatt and Jacob Meerman, "Resources Mobilization in Developing Countries: Financial Institutions and Policies", World Development, 1978, Vol.6, No.1, pp.45-64. - 13 - III. DYNAMICS OF LDC HOUSING MARKETS AND CONTRIBUTION OF FINANCIAL INSTITUTIONS A. Introduction Even though the focus of this paper is on financial institutions and not on the problems as seen from the viewpoint of households, housing finance problems in developing countries cannot be evaluated without a prior examination of the dynamics of the housing market and of the various ways households meet their housing needs. In this section three questions are addressed: what is the typical structure of LDC housing markets? How much do we know about progressive housing investment, especially in the informal sector? What are the implications of the dynamics of LDC housing markets, especially for the financing of low-income housing? B. Three-Tier Structure of LDC Housing Markets The dynamics of the residential sector in LDC's leads to a three- tier market. I. First, at the top of the income scale we have hou-seholds able to afford housing of high quality in fully serviced neighborhoods of low density which make up a substantial-proportion of the residential land of cities. The financial needs of these.groups are met either through specializing housing finance institutions or through 1/ Obviously, further market segmentation of these three categories exists and should be used for detailed planning. See, for instance, the concept of a stock-user matrix in Paul Strassman, The Transformation of Urban Housing, The Johns Hopkins University Press, Baltimore and London, 1982. - 14 - other means as their resources are substantial. Then there is a relatively narrow strata of middle-inc.ome, middle-class households who are the main users of specialized housing financial institutions, particularly public ones. This group is the major beneficiary of available public subsidies and is predominantly composed of civil servants or wage and salary earners working for large private companies and public sector corporations. The third and numerically largest group consist of low-income households whose housing is provided by the private sector often in a clandestine and illegal fashion. The net result is that in LDC's only a very small part of annual housing investment receives financing from formal institutions. C. Illustrations: Tunisia. and Other'Countries -Keeping in mind that there are significant quantitative variations from country to country according to their levels of income and the structure of their capital markets and financial policies, the typical LDC housing market structure can be better understood by looking at a specific country such as Tunisia. 1. The Tunisia Housing Market Tunisia is a small country of about 6.5 million people with a per capital income of $1120 in 1979 which grew at a high rate of 4.8%-a year- over the period of 1960-1979, with a fairly high literacy level of 62%. It is already more urban than rural (52% urban in 1980), with a declining rural population, still a rare occurrence in developing 1/ Much of the evidence presented here is based on the 1982 Tunisia Housing Sector Review done by World Bank staff. - 15 - countries. These are conditions which could a priori be considered favorable to the development of a well structured system of housing finance. The evolution of the housing stock over 1975-1980 shows a net loss of 24,500 units in rural areas and a net gain of over 129,000 units in urban areas. But these net figures hide a complex combination of new construction, upgrading, subdivision and demolition of units. In fact 316,000 units were constructed: 199,000 in urban areas and 117,000 in rural areas over the period 1975-80 as follows: Construction of Units Urban Areas Rural Area Total new units 199,000 117,000 316,000 demolition 60,000 141,500 201,500 net gain 139,000 -24,500 114,500 total stock 605,000 515,000 1,120,000 Modification of Units upgraded/enlarged 21,000 21,000 42,000 subdivided units 72,000 22,000 94,000 The LLumber of modified units does not affect the total number of new units, but it certainly implies a significant housing investment concentrated mostly in urban areas. In urban areas, in spite of the construction of 200,000 units and a 27% net expansion of the number of units, occupancy rates have - 16 - increased markedly, yielding an average of 6.7 persons per unit or 2.4 per habitable room. The number of families per dwelling rose from 1.09 in 1975 to 1.16 in 1980. Over 40% of new housing consisted of one-room dwellings built by the private sector resulting in an increase in the number of urban families liv'ing in one-room units from 17% in 1975 to 26% in 1980. But the good quality of construction led to a significant reduction in substandard units. A key aspect of the housing supply situation is that about 52% of urban housing built between 1975 and 1980 by-passed municipal development regulations and outpaced the ability of the public sector to guide and control urbanization. This is all the more striking in that, compared to less dynamic and less urbanized countries, Tunisia has developed an extensive system of institutions to structure the housing sector: housing development agencies (SNIT), land development agencies (AFH), redevelopment agencies (ARRU), financial agencies (CNEL) and various financial programs (such as.FOPROLOS, FNAH). Nevertheless, present Tunisia still exhibits the typical three-tier housing market structure of developing countries: (1) A high to middle income "legal" private sector; (2) a heavily subsidized, mostly middle-income-public sector and (3) a large, rapidly growing low-income "informal" private sector which could be further broken down into a "clandestine" -or - "unauthorized" part of fairly good quality and "slums." Most housing developed by the "legal" private sector is of relatively low density (15-30 dwelling units per hectare) with a resulting high infrastructure development cost. This sector is largely - 17 - self-financed and is therefore highly sensitive to the evolution of incomes and the propensity to save of the population. Assuming a 1980 cost of dwelling to be of the order of TD 4,600 (one Tunisian dinar is worth U.S. $2.00) and financing to a 15-year loan at 10% interest resulting in a monthly shelter cost of TD 50, it appears that only about 20% of the urban population would be able to acquire housing through this sector. The effectiveness of the public sector's participation in the hdusing sector in Tunisia in 1975-1980 is the outcome of its triple role of developer, lender and subsidizer. As a developer it controlled 41% of housing investment and produced 26% of new4 housing. As a lender it provided financing equivalent to 8.8% of total housing investment, in addition it provided direct subs'ii,es equal to an additional 5.3%. The Ministry of Housing-is authorized to license private developers to construct both private and public housing. Prior to 1975 there were only a dozen licensed developers, at present there are over 70. Their role is marginal, however, as they built only 3,700 dwellings whose value was less than 4% of housing investment. In most cases privately developed subsidized projects are undertaken on AEH land on behalf of either CNEL savers or the employees of such large public or semi-private agencies as SOMEDE, STEG, the Societe Tunisienne de:Banque, Tunis Air, etc. The ability of current public programs to address the housing needs of lower income households is difficult to assess in the absence of precise household income data and information on the propensity to - 18 - save at various income levels. However, it is possible to construct indicators of need that can be compared with the various levels of subsidization proposed by public programs. Population targeted by programs vary considerably; yet all programs, except those intended for rural or slum relocation housing anticipate an equity of over TD 1,000. Considering that median household incomes in urban areas are the order of TD 1,450 per annum, 75% of state aided housing constructed between 1975 and 1980 was not reaching low-income households. The "informal" sector is far from disorganized, although it is illegal insofar as it does not adhere to existing development standards and bureaucratic procedures and is often built on land either owned by the government or whose title is unknown. In contrast to the shanty towns of the early 1970's, it develops large residential areas of good quality which are well maintained. These de facto residential zones are developed at high density and without infrastructure or community facilities. The quality of construction is improving continually, in part because of tht work of small builders who are willing to take on jobs of any scale. This type of housing is generally incremental: initially a single-room is built and the plot fenced in; at an estimated-cost of TD 850 this minimum development is still on the order of 1.3:times the- average annual income of these households. As resources'become available or employment of the household head more stable, building materials are stockpiled in the courtyard. Finally, a permanent - 19 - compound wall is built and the dwelling enlarged over a period of three to Eive years. To summarize this description of the typical three-tier structure of housing markets in LD's we can look at the Tunisian situation over the period of 1975-1980 from three convergent viewpoints: (a) the relative contribution of each tier to total housing output; (b) the situation along the household income distribution of beneficiaries of each tier; (c) the scope of public and private financial institutions in the financing of total housing output. This structure is presented in summary form in Table 1. In terms of volume of investment it controls or regulates, the public sector is dominant in Tunisia which is not the case in less urbanized countries, but because a significant volume of urban infrastructure is involved, the number of housing units affected is much less.. What is most striking is the undeveloped state of financial institutions mobilizing resources for housing: only 7.7% of the "legal" private sector investment represents institutional financing. This ratio rises only to 34.4% for the public sector investment and 16.7% of total housing. If public sector subsidies are excluded the share of financial institutions drops to 11.4%. The detailed breakdown is presented in FIGURE 1. 2. Limited Reach of Financial Institutions in Other Countries The limited role of institutional housing finance in the provision of housing found in Tunisia is matched by the experience of - 20 - Table 1 THREE-TIER STRUCTURE OF A HOUSING SECTOR: TUNISIA 1975-1980 (1) (2) (3) (4) Institutional Financing Share of (including sub- Housing Number sidies) Investment of units Households (excluding down- (millions of TD) (1,000) Served payments) I. "Legal" Private urban 268 urban 58.3 100th to 80th TD 21 million Sector rural 4.3 rural 6.l percentile (7.7%) total 272.3 total 64.4 (top 20%) (34.2%) (20.5%) II. Public Sector urban 265 urban 38.3 90th to TD 113 million Controlled Con- rural 63.2 rural 42.1 30th (34.4%) struction and (41.2%) (25.6%) percentile (including Financing subsidies) III. "Informal" Sector urban 144.6 urban 96.4 50th per- zero (urban & rural) rural 50.9 rural 72.8 centile & total 195.5 total 169.2 below (24.6%) (53.9%) IV. TOTAL - TD 796 314.0 TD 134 million .million (16.8% of (100%.) column one). - 21 - Figure 1 TUNISIA HOUSING FINANCE SYSTEM 1975-1980 (Millions of Tunisian Dinars) Private Household Financing (496) (downpayment) (62.3%) Private Sector Investment Exclusively (517) (65%) Institutional Financing (21) (loans) (2.7%) Total Value of Housing Investment (795) Private 100% Household (166) downpayments (20.8%) Public Sector Controlled (70) Investment loans (279) (8.8%) (35%) Public Sector (113) Financing (14.2%) direct (43) subsidies (5.4%) 21 + 70 = 11.4% Share of loans in total financing: 796 i.e., Role of Institutional financing in the provision of housing. (If subsidies are included, the percentage rises to 16.8%.) Source: Tunisia Housing Sector Review 1982, World Bank (1 T.D. equals US$ 2.00). - 22 - other countries. In Thailand in 1981, financial institutions provided less than 33% of total investments: 76% of the institutional funds went to upper income groups, 15% to middle income groups and only 6% to low- income groups. In the Philippines in 1977 only 25.5% of housing investment was financed by institutions with a distribution of beneficiaries as skewed toward high-income groups as in Thailand. In Portugal over the three-year period 1976-1978, the average financing of housing investment was only 20%. There is only limited evidence about fluctuations in the level of financing over time in LDCs. During the three years for which the ratio is available for Portugal it ranged between 6.5% and 35.4%. Because the United States are considered one of the housing markets most easily financed, the evolution of the level of institutional financing is interesting to trace. It has been as follows: Percentage of Share of New Housing Year Institutional Financing in Total Financing 1970 65.3 24.6 1972 84.6 24.4 1975 65.5 19.7 1977 75.6 20.7 1981 44.4 16.4 The significant drop in the level of financing in 1981 'is a-reflection of the drastic decline in the degree of isolation of housing finance institutions from the rest of capital markets under conditions of high interest rates. It will be noted also that in the United States a high proportion of funds went to refinancing existing units. - 23 - D. Progressive Housing Investment in The Informal Sector 1. Policy Questions Raised by the Informal sector The high proportion of housing that is built in the third tier of the market without relying on institutional finance in developing countries raises several major questions regarding the development of housing finance policies: How much is understood of this progressive housing investment process whose outcomes are everywhere to be seen? Is this kind of housing supply apprppriate or does it have serious shortcomings? Is housing finance the only constraint on expanding the supply of housing? The main one? What are the significant policy implications of progressive housing investment, especially regarding the financing of low income housing? 2. Progressive Housing Investment in the Informal Sector Without a prior examination of progressive housing investment in the informal sector, it is not possible to discuss effectively the formulation of housing finance policies toward low income households. Over the last 10 years progress has been made in understanding the dynamics of the informal housing markets and in identifying the .conditions which are favorable to investments both in the production of new housing and the upgrading of existing ones. 1 As could be seen in 1/ For an analysis of the production of new housing see Emmanuel Jimenez, "The Economics of Self-Help Housing: Theory and Some Evidence from a Developing Country", Journal of Urban Economics, Vol.11, 1982, pp.205-208. On upgrading see Raymond J. Struyk and Robert Lynn, "Determinants of Housing Investment in Slum Areas: Tondo and Other Locations in Metro Manila", The Urban Institute, Washington, D.C., August 1982 (mimeo). -24- the case of Tunisia, progressive investment can be defined as a method of housing construction or upgrading characterized by staged development. It is a mode of investment in which the infrastructure and part of the house are built by a contractor in a fairly short period of time. Then the rest of the shelter is completed incrementally by the household which hires a contractor and/or laborers according to its current income position and construction supply conditions. 1/ The flexibility and diversity of housing arrangements that can be encountered in developing countries is remarkable. For instance, a variety of tenure arrangements can be encountered in most rapidly growing cities: renting a room, renting a complete unit, owning the. structure built on a site over which the household has no legal control (squatting); owning the structure but holding a lease on the-site; owning both the structure and the land and subletting parts of it; owning and occupying the unit and the land. A wider dispersion of income levels than in advanced economies has spawned this greater diversity of housing arrangements which can be interpreted according to four different dimensions of housing demand: the choice of tenure, the housing services provided, the demand for land, and the demand for . structure. When a householdl must make a housing decision it has to: identify the combination of the four elements of demand that is most- 1/ See Douglas H. Keare and Scott Paris, Evaluation of Shelter Programs for the Urban Poor, Principal Findings, World Bank Staff Working Paper, No.547, 1982. - 25 - satisfactory. It considers its employment conditions and available income, the size of the family, the savings and assets that it has accumulated, the intra-family transfers on which it can rely as well as what it can borrow in the informal or the formal financial market. In principle, the tenure decision is clearly the most critical one for housing finance: it is the one which differentiates housing consumption from investment in the assets that generates the housing services. However, in LDC's the lesser degree of tenure security may prevent the use of the property as a collateral for conventional or, rather, institutional financing. It can be expected that informal financial arrangements will be tailored to the various shades of tenure that are found in the large low-income housing markets of LDC's, but very little is known yet about the various defects and market inefficiencies associated with these arrangements, particularly with regards to the critical impact that the order in which various decisions are made has on the investment. One such important decision in the investment behavior of homeowners, at the low end of the income scale, is the taking of tenants who rent one room and share the unit with the family of the homeowner. Experience shows that these homeowners who rent rooms Lavest differently from those who do not. Renters. appear to play a crucial role in progression of housing investment as they allow owner-investors to control a greater amount of housing equity sooner than would otherwise be possible, until the time when they reach the required level of sustained income. The renting of rooms is quite significant to low - 26 - income housing policy, for instance, in the Tondo area of Manila, over 40% of owners share their units with renters.-L. It is clear that prohibition against room rental in low income housing programs is not well advised, but not enough is known about this observed investment strategy to formulate better low income housing policies. In spite of the flexibility of staged investment in the informal sector there are problems of lumpiness to overcome in achieving home ownership. Minimum consumption levels exists even in slum room rental and, as is the case in rich economies, the very poor appear to devote a much greater proportion of their meagre resources to shelter than is the case for even moderately better off groups. Minimum resources are required to move from one type of tenure to another. A household may be squatting because it has enough resources for the structure itself, but the legal control of the land is beyond its reach and therefore accepts uncertain tenure over the land because it reduces the price of this asset. 1 An economically successful household maychange status over time from room renter to squatter; then from owner of the structure and renter of the land to owner of both. While climbing these various steps the household may also increase the level 1/ Research and Analysis.Division, Executive Staff. Housing Consolidation Study, Tondo Foreshore, Dagat-Dagatan Development Project, Report Series 80-2, National Housing Authority,.Manila, Philippines 1980 and David L. Lindauer: The Tondo Project: Whom have we served" DEDRB, The World Bank, 1980 .(draft mimeo). 2/ See Harold Dunkerley, et al. Urban Land Policy Issues and Opportunities, World Bank Staff Working Paper No. 283, May 1978; (Two Volumes). - 27 - of housing services by upgrading the quality of the unit and expanding its floor area, but no obvious path in the progression has yet been documented. 3. Non-Financial Constraints on the Supply of Housing In addition to low income levels and the difficulty of saving underscored by the pervasive pattern of progressive investment, there are other constraints of a regulatory nature which can seriously restrict the supply of housing. Whenever they are encountered, these regulatory constraints affect all income categories of the market with the possible exception of the highest income groups which can afford high quality housing. Inadequate land registration, excessive land development standards, rent control, and excessive construction standards frequently increase the difficulty of providing financing to middle-income as well as low income families. For instance, rent control in Bombay had had a very severe impact on the quality and total stock of housing available. The implementation of the 1976 Land Regulation and Ceiling Act of India has had an effect opposite to that intended by the -legislators and has reduced the supply of land for housing in every city. Because a better housing finance system is expected to - increase the supply of housing and to improve its quality, it is useful to consider financial services as an input to the production of housing in a manner comparable to labor, land or construction materials and standards. It is generally the case that a series of improvements in - 28 - the provision of these complementary inputs is required in order to facilitate the supply of finance and to increase housing output. Inappropriate regulations, instead of improving investment in the informal sector, make it more difficult and accentuate fragmentation in the progressive investment process. E.. Implications of the LDC Market Structure for Housing Finance, The review of the Tunisia housing market has shown that only a small proportion of new housing investment benefits from institutional financing and the discussion of progressive housing development has clarified why this happens. The obvious questions then are whether financial institutions are needed at all and what are appropriate directions for the development of a housing finance system. 1. Multiple Benefits of Institutional Housing Fance Housing finance institutions are needed in develo.ping countries for a multiplicity of reasons. Their growth has a positive impact on the quality and quantity of urban development, on resource mobilization, on the efficiency of markets, on the growth of a. very important sector of the economy and eventually on homeownership. From an urban development viewpoint there is.much scope for- improving the sector. At present, middle-class and low-income . households are forced by poor regulations and lack of-financing to develop their housing in piecemeal inefficient ways. They can produce. structures of fairly good quality, but they have major problems with the provision of clear and secure tenure, water supply and sanitation and other urban infrastructures that each household individually cannot - 29 - provide for itself. If more than 50% of total housing supply bypassed local government institutions in Tunisia it was partly because of inappropriate public regulations and also because the undeveloped housing finance system could not mobilize financial resources on 'the scale needed for the provision of at least on-site infrastructure. The shortage of serviced land would be lessened if residential development by private as well as public developers could be supported by large scale, flexible institutional financing over all phases of developmeht: land development, construction, interim financing and mortgage financing. There is a circular causation between the provision of secure, serviced land and increased savings by households as will be discussed in Part IV. At present most household savings are not mobilized in financial form. Homeownership is the highest priority for asset information for most households in any country and housing finance programs should be developed to trigger and mobilize household savings with the help of efficient financial instruments. Present methods of financing involve high risks and large transaction costs compared to the sums involved; they are neither very good for the user of funds nor for the provider of funds who should be offered more flexibility, more security and higher yielding saving instruments. In the progressive.- development method of housing there is no requirement that a young - household move through each investment step in a predetermined way over the life cycle of the family. In fact the object of public policy could be described as making several of these steps unnecessary by reducing - 30 - existing constraints in the supply of land, public services and the provision of housing finance. Finally, by increasing investment opportunities housing finance stimulates also the growth a sector that generates between 15 and 30% of total national investment every year and is a leading generator of low to medium skills jobs requiring only a limited amount of foreign exchange compared with other investment sectors. 2. Policy Implications The dynamics of the housing market as it is observed in developing countries has strong implications for housing finance policies toward the informal sector. The heterogeneity of the informal sector requires two distinct but complementary policy directions: at the upper end of the informal market the objective is to increase the reach of the legal sector into the low-to-middle income clandestine market. This requires a redirection of financial institutions concurrently with the reduction of irrealistically high standards, a review of the impacts of land regulations, improved land administration to clarify ownership conditions, and a review of occupancy and tenure protection laws which virtually preclude households from being evicted for non-payment of- debts such as mortgages. Increasing the reach of the legal sector could be compared to a portfolio allocation problem. As can be seen in Table 1, Tunisia could increase the number of units produced without even expanding the financial resources required if smaller, cheaper units could be traded- - 31 - off for larger more expensive ones. At present, the legal private sector control 34.2% of housing investment and produces only 20.5% of the units; even worse, public sector controlled units absorb 41.2% of resources and represent only 25.6% of all the units produced; together the legal private sector and the public sector absorb 75.4% of resources to produce 46.1% of the units. If expansion of financial resources together with resource reallocation at the margin could be achieved, the impact on the number of units produced would be quite significant. At the lower end of the income scale the situation is quantitatively and qualitatively different. The problem is not to develop commercially oriented viable financial institutions. The priority lies in mobilizing resources and developing specific programs with a major focus on urban public infrastructure - on t: e "public goods" - which households cannot provide for themselves It is an open question*whether both upper-end and lower-end programs can be executed by the same institution: sources of funds, populations to be served and types of output are all different. 3. Financing Low-Income Housing The problems of financing low income housing continue to be vexing ones for national governments. The World Bank policy views regarding housing and urban residential investment can be summarized in five points: (a) to improve substantially the low-income housing. situation; (b) to provide affordable standards of construction and infrastructure; (c) to insure the financial viability of the institution involved in the sector and the replicability of the operation; (d) to - 32 - improve cost recovery in order to eliminate, or at least reduce significantly the need for subsidies from public funds; and (e) to maximize the contribution of urban investments to the local and the national economy. Part of the difficulties in formulating programs arises from the lumping together by public policies of a wide spectrum of significantly different households. As just noted, if there is scope for significant improvements in the "clandestine" or "unauthorized" sector of the markets which is made-up of low-middle class households, the problems raised by the residents of slums are quite different. In general the latter have neither the equity base nor the capacity to borrow to achieve some form of fully legal ownership. Only a minority among them could contemplate financing on terms comparable to the better-off groups in the "unauthorized" settlements. The problem is not to finance housing but basic infrastructure, sanitation and neighborhood services which are the first priority over the housing shell itself. The minimum amount of services that they deserve is in the nature of a a "merit good" and typically involves some form of direct subsidy or cross-subsidy. The progressive investment pattern and the wide variety of- methods used in the informal sector to mobilize resources lead to the - central question of LDC housing finance: to what extent and under what conditions can housing finance institutions grow out of the extensive interpersonal networks which represent up to eighty percent of housing investment resources? In other words, how do yoq correct the mismatch - 33 - between the financial services that institutions can provide on one hand and the equity base and borrowing capacity of LDC households? There are familiar and important limitations to the income qualifications required from potential borrowers: an adequate level of income, regular stable employment, a verifiable income and satisfactory collateral, i.e., collateral in the form of conventional marketable assets. There are also restrictions on loan terms which make access to financing difficult such as: minimum sizes of loans that are too large, high downpayments in absolute terms and small loan-to-house value ratios, very long maturities out of all proportion with small low-income time horizons, requirements for fixed and regular amortization payments, inconvenient office locations, high housing finance cost burdens, complex loan terms and conditions which are difficult to understand and/or to comply with.' Addressing this question more fully requires a discussion of resource mobilization and capital markets in developing countries. However, at the end of the review of the housing market one can already anticipate two types of barriers to institutionalization. The first one is-technical and relates to economies of scale and efficiency in the pro-vision of financial services. The second could be labelled socio- political and relates to the reluctance of many participants -of the informal financial markets to rely on public sector organizations or 1/ For a review of this mismatch problem see United Nations, Non- Conventional Financing of Housing For Low-Income Households, Department of International Economic and Social Affairs, ST/ESA/83, New York, 1978. - 34 - government regulated banks. Even for business activities most potential borrowers would rather join trade organizations and enter into financial arrangements run by a private group for several reasons but most particularly to avoid contact with tax administrations. This is particularly true at lower levels of development. Public policy is expected tb encourage the growth of rotating credit associations, credit unions and cooperative movements out of present interpersonal networks based on common trade, family, religion and provirce of origin. Such an objective is very difficult to pursue because in most countries public authorities restrict the definition of housing policy to what the government and public agencies can do in housing and to the allocation of public resources. For international organizations, it would be-even more difficult to set-up "a non- conventional housing credit intermediary" where "efforts should first be devoted to organizing formal and/or informal groups." I In India, for instance., official support to the housing cooperative movement has led to its increasing identification with public sector activities and dependence on public resources. Managers of state level cooperatives identify themselves easily with officialdom and cooperative:membership is essentially middle- and upper-middle. class and has for maim-objective to gain access to closely regulated resources such as serviced land or preferintial financing. 1/ U.N., "Non-Conventional Financing of Housing", Recommendations number 6, page 109, op. cit. - 35 - The question of how one can move from one-shot financing operation by groups of relatives and friends to the institutional financing of low-income housing must now be examined from the viewpoint of capital markets. -36 - IV. HOUSING FINANCE INSTITUTIONS AND CAPITAL MARKETS A. Introduction The examination of the housing market side of housing finance intermediaries shows already that policies towa-rd the financing of housing must address two distinct problems: one is to deepen the reach of financial institutions into the upper part of the informal sector, the other is to locate the resources necessary to provide the lowest income groups with the basic services (the public goods) which they need and which public policy has determined they should have. Institutional financing is not the only constraint on expanding the supply of housing; there are also regulatory and income constraints operating on the housing market, it .is now necessary to consider the extent to which capital market constraints also constribute to making institutional financing such a small proportion of total housing investment in most LDC's. The anticipated benefits of .financial intermediation are already clear, but what are the characteristics of housing finance that differentiate it from the other sectors? What are the functions that a housing finance system is expected to play? How is its growth-linked to the overall level of financial development in a country? To'the development of long-term credit markets? In other words,what is the - dynamics of capital markets and what is its impact on -hou-ing finance?-- - 37 - B. Level of Economic Development and Supply of Financial Resources At any given time, the pool of financial resources available in domestic capital markets is finite and from the viewpoint of finance ministries or national planners to give more resources to housing, or industry, or agriculture, means to give less to some other sector. It is beyond the scope of this paper to-discuss in detail the factors determining the total volume of financial savings at various levels of development. But many housing policy reviews would have benefitted from taking closer account of the level of development of the capital markets, their actual size, and the degree of priority given to housing in comparison with other sectors. Too often projections of future housing output based on needs are inconsistent with the financial resources available, even after taking into account the role of the informal sector. To characterize the level of financial development of various economies, theorists often refer to their "financial depth" and to the extent of "financial repression". Financial depth reflects the degree of development of the financial sector and it is expected to increase with the level of income. It shows how far a country has progressed fromzsufficiency in the rural areas, to barter, to trade based on money, to fiat money and beyond to the use of a whole range oftfinancial instruments. But two countries with similar levels of per. capita GNP . donot show the same degree of financial depth and financial repression refers to the inability of financial services to grow as rapidly as could be expected in view of the growth of the economy. Financial - 38 - repression occurs when lenders and borrowers by-pass official institutions due to over-regulation .and is essentially caused by attempts to maintain nominal interest rates below inflation rates as well as other regulations attempting to force interest rates below the cost of capital. Thus one expects the degree of financial depth of an economy to be a first indicator of the extent to which the housing finance system is developed and of the degree to which it would be possible to make improvement, without requiring also major changes in other parts of the system. In the case of housing finance institutions, financial repression leads to a larger role for the informal sector than economies of scale and transaction costs in the production of financial services might justify. The International Finance Corporation (IFC) has compiled ratios for selected countries in 1977 to to facilitate comparisons of financial depth and illustrate its link with the income level. I/ Table 2 presents these ratios grouped by region and ranked according to per capita GNP. The indicators presented are: 14-1 money, "broad money" (i.e., M7--financial aggregate compound of currency, all deposits and any financial instrument with maturities of one year or less);-- "commercial bank assets," "other deposit bank assets," which include housing finance institutions, savings and otner deposit-intermediaries such as postal savings systems, "Total Deposit Bank Assets," which is 1/ Capital Markets Department, IFC, Country Financial Systems Profile for Selected Countries: Methodology, Tables and Diagrams. Washington, D.C., July 1980. - 39 - the sum of the first two, and finally "Total Financial Aggregates" defined as the sum of total deposit bank assets and outstandin securities. As could be expected, the monetization of the economy rises rapidly with the level of development. However, while Mi and M7 rise most rapidly at low to modest levels of development, ml begins to level off relatively soon; at high income levels it even declines as the technology permits other means of payment. These patterns are also applicable to "broad money" M7 although they are less distinct and the range of variation is significantly wider. The deepending of the financial system with economic growth is also quite obvious, but the relationship is not linear with GNP because the ratio of total financial aggregate to GNP is also influenced by the degree of financial repression, inflation and hyper inflation. For instance, in the East Asia region, financial depth increases with per capital GDP very markedly (see Table 2, column 6). On the other hand, the degree of financial depth in Latin America is less than could have been expected because of inflation as seen for Chile and Brazil. Similarly, the degree of government control, credit allocation and manipulation of interest rates appears to lead to much smaller financial systems as appears to be the case for Turkey. This situation can be contrasted with a more positive approach to the development of financial services in Thailand and tfte Philippines which has resulted in greater financial depth at much lower levels of income. Another factor at play is the role of foreign banks as in the - 40 - Table 2 INCOME LEVELS AND DEVELOPMENT OF FINANCIAL SYSTEMS (Financial Depth) (1) (2) (3) (4) (5) (6) (7) Other Total Broad Commercil. Deposit Deposit Total Region and "M-1" Money Bank Bank Bank Financial Per Capita Country Money "M-7" Assets Assets Assets Aggregates GNP 1978 per cent of GNP, 1977-78 -,US dollars East Africa Kenya 21.8 32 28 8 36 57 330 West Africa Senegal 22.3 28 50 n/a 50 52 340 Nigeria 18.7 34 32 n/a 32 39 560 Ivory Coast 24.2 33 29 6 35 '37 840 EAP Indonesia 10.0 17 19 n/a 19 19 360 Thailand 11.3 42 42 6 48 63 490 Philippines 9.0 41 44 4 48 74 560 Korea 10.8 42 64 38 102 123: 1,160 South Asia India 13.7 43 38 2 40 58 180 EMENA Morocco 35.1 39 27 5 32 36 670 Jordan - 66 55 8 63 113 1,050 Turkey 20.2 23 22 (-) 22 22 1,200 Portugal 41.4 82 58 19 77 n/a 1,990 Spain 28.0 75 52 26 78 122 3,470 LAC Bolivia 10.2 - 14 16 1 17 17 510 Colombia 12.9 23 17 6 23 29 850 Ecuador 15.4 23 31 (-) 31 41 880 Dominican Republic 12.4 29 32 3 35 n/a 910 Mexico 10.3 25 11 15 26 32 1,290 Chile 4.7 16 21 8 29 40 1,410 Brazil 12.2 36 22 20 42 57 1,570 Argentina 6.6 29 28 3 31 35 1,910 Industrial Countries United Kingdom 15.4 56 121 3 124 199 5,030 Japan 31.1 116 74 56 130 204 7,280 France 25.8 87 56 33 89 112 8,260 Netherlands 23.3 89 62 34 96 135 8,410 Germany 16.7 73 36 64 100 145 9,580 U.S. 16.7 66 67 32 99 221 9,590 -41 - case of the African countries represented here. Unfortunately, the degree of development of long-term credit markets which is especially important to housing finance cannot be established from the data of Table 2. Still it is worth noting that there appears to be a non-linear positive relationship between level of development and the proportion of assets in GNP in the same manner that the share of housing output 1/ increases with the level of income as mentioned earlier. - In spite of the direct relationship between level of economic development and the growth of financial institutions, the variability in the volume of finance savings. within a group of countries of comparable per capita income is great; it is also significant within a country over time. The growth of housing finance institutions is dependent on the rise of the national volume of financial resources as much as it contributes to it. It is necessary to consider now the problem of resources mobilization which is one of the three major elements of what has been defined earlier as the housing finance problem (see p.4). C. Resources Mobilization and Constraints on of Housing Finance There are essentially six ways to mobilize resources in an --economy: (a) public sector savings through taxation and, a positive balance in the government current accounts, (b) savings by public -sector enterprises (especially in the energy sector), (c) inflatiqnary_ financing, (d) private savings by the corporate sector, (e) private 1/ See Also Leland S. Burns and Leo Grebler, The Housing of Nations: Analysis and Policy in a Comparative Framework, London, McMillan, 1977 and B. Renaud,"Resource Allocation to Housing Investment: Comments and Further Results", Economic Development and Cultural Changes, Vo. 28, No.2, January 1980. -42 - savings by the non-corporate sector including personal savings, and (f) foreign borrowings. The relationship between aggregate domestic savings and financial savings is not rigid: a country can have a good savings rate of the order of 20% to 25% of GNP with financial savings significantly below that ratio. The gap is generally due to the dominant role of savings retained by businesses and households in the form of non-financial assets. The level of aggregate domestic savings and the methods chosen to mobilize resources constitute macro-economic constraints of significance to the potential growth of housing finance institutions. If the national savings rate is very low the economic environment is not going to be favorable to the growth prospects of housing finance institutions. In addition, financial authorities will not pay them much attention because the problem is likely to be the lack of control over public sector expenditures and poor management of public enterprises and that. will demand all their attention,. In addition, the housing finance system tends to be small and underdeveloped comparea to the banks and the other financial intermediaries. On the other hand, to the extent that housing institutions rely on private non-corporate.savings, they depend on the healthiest source of savings in the economye- For instance, a recent study of private savings mobilization-_- based on 1973 data for 26 African countries shows that the ratio of-. gross domestic savings to GNP varied enormously that year from -2.2 percent in Upper Volta to 33.7% in Algeria. Variations in public - 43 - savings where even wider and private savings which represented 83% of total domestic savings for the entire sample played a stabilizing role for the level of aggregate savings. In Mali, where public savings were negative, private savings where positive and amounted to 5.7 percent of GNP. In Nigeria, aggregate savings were 29.8 percent and public savings were 18.6 percent, implying a less dominant investment role for private savings that year. 1/ When governments are resorting to inflationary finance the prospects for a healthy housing finance system are dim because inflation is essentially a tax on money balances, which discourages financial savings. The housing finance sector is often more severely affected by such policies than other sectors because governments typically establiski nominal interest rates for housing below inflation rates in the name of social objectives and for political reasons: ttle consequence is that interest rates are low but there is little mortgage money available to lead. Another important macro-economic constraint wita which the housing sector has to contend is the perception of housing as a consumption item and the low priority that the sector receives. in national plans. This low degree of priority accorded to housing is 1/ Note again that high aggregate levels of savings do not imply that all or even the greatest proportion of savings mobilization took place through the financial sector. In Africa, in particular, the informal sector is very important. For more details, see Sani Geadah, Private Savings Mobilization in African Countries, International Monetary Fund, August 1981, DM/81/57. - 44 - essentially a political decision which is dictated oy national preferences, the level of urbanization of the country and its rate of increase. The opportunity cost of capital and the rate of return to other types of construction activities are such that the share of GNP going to residential construction is low (1% to 1.5%) at low levels of economic development and rises with the income level and the associated level of urbanization (up to a maximum of 6.5 to 7.0%). There are common features to resources mobilization in. developing countries. Public sector savings through taxation and a positive balance in the government current accounts have typically not exceeded 3% of GDP in the past except in countries such as Korea and Brazil during their period of very high growth. Savings through public sector enterprises have proven very elusive since many utilities and public sector firms are prone to running deficits. Corporate savings have not played a dominant role because of the smallness of the sector in many LDC's. Inflationary financing, while always tempting to governments has too many drawbacks for the growth of the financial system and the external competitive position of the economy. Finally, reliance on foreign borrowings is now running up against.sizeable- foreign debt burdens and expanding, foreign loan amortization requirements in most LDC's during this decade.. For all these reasons the non-corporate sector is of crucial significance: as noted earlier (p.9) in most low-income and middle- income countries it represents between 40% and 60% of domestic savings. It is the surplus sector where savings exceed investments - 45 - permitting the growth of the public and the corporate sectors which are deficit sectors. Raising the household level of saving and improving allocation through efficient resource mechanisms is very important. The housing finance system has its role to play in utilizing savings in coordinated fashion with other financial institutions serving small and medium-size family-controlled businesses. It is necessary to develop convenient financial instruments, to offer facilities for borrowing at reasonable terms and to provide financial assets as an alternative to retaining earnings and savings in inefficient low return forms. 1. Household Income and Savings a. Income Level, Household Income Distribution and Savings The possibility of mobilizing personal savings through a formal housing finance system is dependent on the per capita income level of a country. In addition, the income distribution has a strong influence on the organization and behavior of the housing finance system. This obvious constraint is clearly seen in the pattern of savings by household income deciles in three Asian countries with significantly different per capita income levels Sri-Lanka, Malaysia and Taiwan. We have the following patterns in household surveys: . - : Expenditure First Decile Average Savings .- - . Per Capita Share of Lowest with Net Ratio cof . Income 1979 40% Households Savings Households Sri-Lanka (1970) $230 21.2 8th -0.3% Malaysia (1974) $1,450 13.8 5th 15.5 Taiwan (1973) $1,900 20.6 1st 15.1 (Source: Pravin Visaria, Poverty and Living Standards in Asia, October, 1980, The World Bank) In Sri-Lanka where per capita income is low, it was very difficult in 1970 to save at all levels of income. and only the 8th, 9th and lUth top decile show positive savings. In Malaysia in spite of a much higher level of income, only the sik top deciles are consistent savers as a group, partly because the income distribution is more skewed than in Sri-Lanka and Taiwan, two economies noted for their high degree of economic equality. In fact, in Taiwan all household deciles were saving in this 1973 survey. Even though Malaysia and Taiwan show comparable average saving ratios over all households, the housing finance system that could be. developed in Taiwan on the basis of the existing structure of financial savings could be more extensive and more diversified than in Malaysia. In fact, a housing finance system mobilizing household savings directly can have a broader social base than is indicated by the net savings per decile shown by surveys because within each income decile a significant proportion of households are able to save. In Sri-Lanka 11.3% of households in the lowest, decile saved, in Malaysia 15.5% and in Taiwan 80.1%. Evidence for Mexico and other Latin American countries shows similar conditions. In the specific case of low-income households, the number that could:be served by general financial institutions will generall-yno.t be large in spite of the significant proportion of households that is savings in cash form for essentially three reasons: (a) their incomes are very irregular, (b) the preference for direct interhousehold transfers is very high and the proportion of dissavers remains higher - 47 - than that of savers, and (c) the transaction cost of relying on formal financial institutions may be very high because available methods of savings collection and loan origination are not adapted to the special needs of very low-income groups. 1/ b. Ownership Opportunities and Saving Response During the last ten years of low income housing experiments it has been confirmed that home ownership can be a powerful motivator for saving even by very poor households. Surveys of beneficiaries of low- income housing projects who obtained their housing under hire-purchase agreements show that they were able to mobilize a significant amount of resources for down payments through interhousehold transfers and that consumption patterns were modified significantly to maintain monthly payments. For instance in India, in the case of households with income not exceeding Rs. 350 or US $38 per month--the social group designated as EWS: "Economically Weaker Section"-the structure of the down payment was found to be: 2/ 1/ See D. Kaufmann and D. Lindauer, "Basic Needs, Inte_rhousehold - Transfers and the Extended Family", Urban and Kegional Economies Report No. 80-15, DEDRB, World Bank, for similar evidence in the Philippines and El Salvador. And also D. Kaufmann, Social Interaction As a Strategy of Economic Survival Among the Urban Poor, Harvard Ph.D. Dissertation, 1982. 2/ Mulkh Raj, Allotment of a House on dire-Purchase and its Impact on Savings and Consuption Expenditures, HUDCO, New Delhi, June 1982. -48- - Own Savings 43Y - Loans from Friends and Relatives 40% - Loans from Banks 6% - Withdrawal from Provident Fund 4% - Gifts 3% - fortgage or Sale of Property 4% It was strongly felt by the analyst that the 40% share of down payments due to interhousehold transfers could not have been mobilized for other purposes than housing. It was also found that the purchase of a house leads to an increase in savings and a noticeable fall in the percentage of expenditires on entertainment, clothing, transport, etc. Host importantly, the zero savings rate prior to purchase became positive even after excluding monthly payments as a form of savings-investment. 'Monotoring of projects in El Salvador suggests also that household secondary workers had more incentive to work to cover increased housing payments. The figures for the EWS group in India were as follows: Percent of Percent of Income 6 Months Income 6 1onths Expenditures Before Purchase After .Purchase (1) Food Items 63.8 - 56.8 (2) Education, Entertainment, Clothing, Books, Transport 30.4 19.0 (3) Rent or Monthly Payment 5.8 17.9 Savings zero 6.3 - 49 - These data show that under very adverse conditions the desire to own housing with the upward social movement that it implies can be a powerful saving motivator. However when considering the savings propensity of very low-income groups at any given time, one should differentiate between the ability of a minority of poor households to mobilize savings for housing and the fact that, as a group, households in low-income deciles are not able to save much even in the form of non- financial assets. 1/ While the objective of owning a house can trigger positive savings on the part of poor households which have the actual opportunity of acquiring a unit, as seen in the Indian survey, a significant saving capacity already exist as one moves up the income scale. The problem there is that inflation, inadequate financial services, inaccessible facilities or low levels of literacy combine with traditions to induce savings in real, safe and generally very illiquid assets. A survey o potential depositors in Morocco which covered mostly salaried workers showed that 63% had accummulated savings. 2/ Among those who accepted to give estimates of these savings the structure of savings showed that only 17% took the form of financial savings deposits. The reported breakdown was: 1/. A comparable unwarranted extrapolation has sometimes been used to -justify slum clearance: it was said that because a minority of slum residents (10% to 15%?) in a certain area can afford standard housing elsewhere in the city but choose not to move, all other data area residents could also afford to move and, therefore, the slum could be cleared. 2/ Source: Credit Immobilizer et Hotelier, Casablanca, Morocao. Unpublished survey, June 1980, Dh 1.00 = US$ 0.187 as of December, 1981; 59% of the survey respondents had a monthly income under Da 2,000. - 50 - Real Estate 21% Jewelry 18% Precious Objects: 9% Monetary Savings: 52% of which - cash kept at home 26% - cash in the bank 57% - National Savings Bank (C.E.N.) 5% - Portal Savings 5% - Others 7%' It is an open question whether such a portfolio distribution of savings is consistent with the relative risk of various types of assets. The large proportion of monetary savings in the form of cash balance shows why it is tempting for so many governments to rely on inflationary financing of their programs.. As could be expected, the portfolio composition changed with the volume of assets: Cash Balances Class Level or Financial Real Other of Assets Savings (.) Jewelry (%) Estate (%) Valuables (%) 0 - 10,000 58.8 24.8 16.5 10 - 20,000 64.1 27.5 - 8.3 20 - 30,000 53.9 24.0 17.6 4.4 30 - 40,000 59.5 7.4 14.5 18.5 40 - 50,000 42.8 8.3 48.9 50 - 100,000 36.7 4.5 58.8 - 100,000 + 19.8 4.6 72.5 3.1 By comparison, the poverty line in Morocco in 1980 was estimated to include 28 percent of the population and monthly household incomes were estimated by the World Bank as follows: . Second Lowest Decile (D2): DR 800/month Third Lowest Decile (D3): DH 930/month Fifth Decile (D5): DR 2,000/month - 51 - The positive relationship between the, level of savings and the proportion of real estate in the asset portfolio of Moroccan housetiolds, except for the top 5% which have a very different behavior, is a pattern which prevails almost everywhere. The reliance on cash balances, jewelry and other valuables at lower levels of savings suggests the difficulty that a financial system has in offering safe and dttractive types of financial assets and the unrealized potential for expansion of a housing finance system. c. Employment Conditions and Household Income Stability The type of jobs held by household members and the others sources of household income have a crucial impact on the diffusion of housing financial services among households and on the kind of clientele that housing finance institutions usually serve. A major dichotomy exists between wage and salary earners who are the prime target of housing finance institutions and self-employed workers who experience considerable difficulties in gaining access to housing finance. For instance, in many African countries the small size of the formal sector and its very high content of public sector employment has had a significant impact on the types of housing finance systems that have developed. Some estimates of public sector and formal employment in these countries are: - 52 - Percentage of "Formal" Share of Public Sector employment in the Employment in the County Year Active Population "Formal" Sector Ghana 1972 10.1% - 73.9% Tanzania 1974 6.3% 66.4 Zambia 1976 14.2% 71.5% Ivory Coast 1970 10.2% - Kenya 1977 12.5% 41.7% Malawe 1976 9.6% 39.2% Uganda 1970 5.9% 42.2% (Source: World Bank, Accelerated Development in Sub-Saharan Africa, 1981, p. 41) Because of the continued effect of a low level of income on financial savings and of the small proportion of formal sector employment, most African housing finance systems remain at an embryonic stage. In addition, because of the very high percentage of public sector employees in the formal sector, whatever housing finance programs exist have been tailored to their needs. the consequence is that the social housing programs involving elements of income redistribution have been turned de facto into fringe benefits and transfers to public sector employees. In countries where the majority of the housenolds draw their income from the informal sector in an irregular way from month to month, financial institutions need to, adapt their operations to meet the need of their clientele. Housing banks which are already short of funds are not interested in self-employed households as potential borrowers because they find it too difficult to determine income sources and borrowing capacity. Since they have little hope of receiving a mortgage in later years, self-employed households lose interest in making deposits in these housing finance institutions. The paradox is that - 53 - because of the irregularity of their earnings self-employed workers do tend to have high marginal savings rates which are not capitalized upon by financial institutions. In addition to the source and regularity of income there are other factors affecting household savings such as wealth, interest rates, the price level and its stability, demographic characteristics and the family life-cycle, the socio-economic characteristics of the population and interhousehold transfers as well as the financial structure of the country. In developing new financial instruments for a housing finance system these factors should be reviewed and the evidence on local conditions carefully evaluated. Even though there are still many grey areas on the determinants of savings behavior many housing finance reviews tend to stress qualitative institutional and legal aspects at the expense of more complete quantitative evaluations. 1/ 2. Resources Mobilization and Housing Policies If there is one characteristic of the housing finance work done by the World Bank during its first generation of projects that stands out, it is the almost exclusive concentration on accessibility to financing and on the problems of beneficiaries. 2/ Discussions concerned with savings mobilizatio. tave dealt with the- determi ants of 1/ For a review of savings behavior, see Katrine W. Saito, The Determinants of Savings Behavior: A Survey of the Evidence, Studies in Domestic Finance No.35. Public and Private Finance Division, D.E.D., The World Bank, 1977. 2/ See World Bank, Learning by Doing: World Bank Lending for Urban Development, 1972-1982, Washington, D.C., 1983. 54- household savings and the proportion of income that could go into housing. Little attention has been paid to savings mobilization by financial intermediaries and the problem of developing viable systems of savings for housing finance. This limited attention paid to savings mobilization resulted from the priority given to designing affordable projects for the majority of the urban population and to the almost exclusive involvement with public sector construction agencies rather than with financial institutions. Whenever financial institutions were actively involved, it was mostly in a mortgage loan servicing capacity. In addition to the high priority given to designing appropriate urban projects for low-income groups a more fundamental factor may have been at work in many countries: it is that savings mobilization is generally inconsistent with policies of low-interest lending. As was pointed out recently by Vogel in the case of agricultural finance: "Savings mobilized by financial institutions at relatively high cost cannot realistically be on-lent at low rates of interest. In addition, officials of financial institutions are likely to find it more pleasant to bargain with their own government or international donors on conditions for obtaining subsidized resources than to. face the task of mobilizing savings from the tural: - 55 - population."1/ He goes on to say that "the desire to maintain subsidized low interest rate lending policies and not the arguments against savings mobilization, is the main reason for the neglect of savings mobilization." In all urban projects the level of interest rates has been a difficult issue to resolve because of the great reluctance to raise them oR the part of the country agencies involved and also because the broader system-wide financial implications of significantly positive interest rates cannot be treated easily within the context of a relatively limited, single urban project. The question has been the same everywhere: "should the international lender recommend positive, real lending rates to the institutions through which it lends for a specific project regardless of the rates being charged by other financial institutions in the country?" In the urban sector the necessity for international leaders to move from demonstration projects to fully developed sector lending combines with the scarcity of international capital and high interest rate level to make it necessary to look at savings mobilization and housing finance system in a more systematic way. Internally, the- -scarcity of foreign capital and high debt levels are leading,member countries to take a harder look. at. the mobilization of domest-ic - resources: given the 1/ Robert C. Vogel Savings Mobilization: The Forgotten Half of Rural Finance, World Bank/AID Colloquium on Rural Finance, Washington, D. C., September 1981. - 56 - low foreign exchange component of urban investment, LDC governments expect urban investments to draw essentially on domestic resources. 3. Benefits of Positive Real Interest Rates Nowhere has the confusion of public policy objectives between the expansion of a viable housing finance system and the distinct and specific needs of public programs for the lowest income groups been as harmful as in the setting of inadequate interest rates for the entire housing sector. Positive real interest rates to,savings depositors constitute a crucial requirement for the development of financial services., Poorly designed low interest rate lending policies for the housing sector prevent resources mobilization and have four undesirable consequences: they have a negative impact on income distribution, they distort resource. allocation, they affect the viability of financial institutions and they blunt project incentives. A policy of low interest loans typically leads to a concentration of a small number of relatively large loans to better-off households. Low interest lead to an excess demand for credit which financial institutions will be obliged to ration as best as tney can. Confronted with an imbalance between their assets (loans).and,. liabilities (deposits) the most rational strategy is to-minimize the cost, and the risk of providing financ -J services by avoiding bQrrowers without good collateral and who are costly to serve. If regulations impose on housing finance institutions to target some of their loans to low-income borrowers, financial institutions will also attempt to ration loans away from them by raising transaction costs (limited location of - 57 - branches, reduced hours, extensive paperwork, etc.). The discussion of housing finance found in the World Bank's staff appraisal reports consistently comment that existing institutions are confining their resources to a very narrow group of middle-income households. Low interest rates (i.e., negative real interest rates) prevent housing finance institutions from fulfilling their role as financial intermediaries. If they are forced to lend at unrealistic low rates, housing finance institutions will have little incentive for savings mobilization and will rely on central bank rediscounts, government loans and/or budgetary allocation if they are in the public sector and loans from international donors. It is much faster and much cheaper to them than opening and maintaining branches close to depositors. Essentially these housing finance institutions stop functioning as true financial intermediaries collecting relatively small deposits from a broad base of depositors while maKing loans that are on average larger to a more limited number of borrowers; they are also unable to mobilize long-term credit from the capital markets on a competitive basis. The lack of services to small dep.-itora.deprives low-income households from*earning positive real interest-on-the small - amount of savings that they typically keep for emergency needs. .The cash that they keep is then exposed -to the inflation tax and the other inflation hedges that they have such as jewelry and other valuables involve such high transaction costs for conversion into cash that savings are discouraged. As the Indian surveys of project beneficiaries - 58 - show, when safe savings vehicles are made available to them even low- income households can make remarkable saving efforts. The second impoztant reason for strengthening financial intermediaries, and among them housing finance institutions is that, when the right services are available, resources are moved away from non-productive investments such as the large proportion of inflation hedges noted in the case of Moroccan savers. However, improving resources mobilization by providing adequate returns to the depositors of housing finance institutions is difficult to implement in heavily regulated banking environments. Aggressive savings mobilization through housing finance institutions is generally discouraged by national economic planners as simply leading to a transfer of resources to the housing sector at the expense of higher priority sectors. To this it can be replied that if depositors are moving their deposits it must be because they are better off and that it will be the institutions earning the higher returns on their assets which will be able to compete most effectively for deposits and household savings. The scope for more effective savings mobilization through the housing.finance system has sparked lively -debates in advanced countries on the merits of specialized housing ----finance systems in order to insulate the housing and fekridential-7 1/ construction sector from the impact of inflation and business cycles.- 1/ See for instance the reports of the OECD Financial Market Committee on Housing Finance, Present Problems, OECD, Paris 1974 and Flexiblilty in housing Yinance, UECD, Paris 1975. - 59 - In housing finance, savings mobilization has been closely tied to government social objectives and suosidies. Arbitrarily low interest rates are a threat to the viability of housing finance institutions. This threat explains the great reluctance that market-oriented institutions have either to developing or to taking over low-income programs sponsored by the government in many countries. In particular, even if they have attracted massive volume of funds by providing positive real returns to depositors as was the case in Korea with the "Workers Assets Formation Savings Program," they may not be permitted to lend these funds at rates that would provide positive spreads. The Korea housing Bank survived financially because about 50% of the resources so collected were then mandatorily redeposited in the National Investment Funds supervised by the Bank of Korea, carrying high interest and used for heavy industry investment programs. In addition, the government directly subsidized the interest paid on deposits and cheap resources were also mobilized through the compulsory sale of bonds with considerably negative real interest to the purchasers of expensive houses, cars, etc. This convoluted form of intermediation for non- housing objectives has recently been replaced by a single structure.with two separate windows under the same KHB roof: a market-oriented-YKHB- ..,structure and a National Housing Fund or low-income housing financed- through public subsidies. One of the reasons for this reorganization is that housing and urban investment has recently been assigned a much higher degree of priority to stimulate the Korean economy and generate employment through the construction sector. - 60 - The third benefit of household savings mobilization through housing finance institutions is that financial intermediaries who serve a large number of depositors and borrowers have a much more accurate knowledge of their current and potential markets, are more efficient in their loan origination and servicing procedures and are less likely to incur high rates of delinquency and default than institutions which are acting as pass-through agencies for public funds. This is why there is always a great deal of interest in the growth of housing financial intermediaries mobilizing resources directly from the public such as building societies, savings and loan institutions and mutual savin4s banks. They are more apt to know their local market thoroughly and to be effective at cost recovery than agencies depending, on public funds or international donors for their viability and the expansion of their activities. Also, potential delinquent borrowers are likely to be more responsible if the funds they have received originate from their own community. But the development of local level intermediaries must oe closely tailored to the rest of the banking and housing finance environment. There seems to be little alternative to these full local, housing finance intermediaries if the needs of self-employed households are to be met. Households with. income that are eitherirregilak or difficult to verify are generally excluded by institutions which rely on indirect savings mobilization. Because they have no way to establish their credit through their past deposit record, these households are simply ignored in lending operations. Contractual savings schemes have - 61 - been one of the leading method to remedy this problem; they could be complemented effectively by mortage insurance programs. Fourthly, and finally, the lack of stress on savings mobilization in individual projects can make long-term housing programs indefinitely dependent on series of central government transfers or on international loans to pursue their objectives. The level of activity of housing finance institutions is no longer a function of the potential of domestic market conditions but of the availability of external resources. Can it then be said that institutional development objectives are being achieved? D.. Some Policy Implications Several important implications flow from this review of capital markets and their impact on the growth of housing finance institutions. First, housing finance policies are subordinated to the overall capital market development policies followed by a country. The degree of sophistication in reforms of the financial system that can be proposed is closely related to the level of income in the country. In particular, interest rates policies at the macro-economic level and - their differentiation across sectors will determine to a great extent the viability of housing institutions which should develop into genuine financial intermediaries as opposed to remaining simple pass-tthrough-: institutions for public resources. Second, the desirability of mobilizing resources through the housing finance system reinforces the earlier finding that workable finance development strategies consist of two complementary but distinct - 62 - parts. On one hand, there should be programs for the development of viable institutions that will reach the better-off households of the informal sector which are already producing unserviced housing of relatively good quality. On the other, there should be separated programs for the worse-off part of the informal sector who cannot have access to ordinary housing finance without public assistance. Third, serving the financial savings and investment needs of the low-to-middle class households which are self-employed is an obvious priority need. Since the desire for a house is a strong motivator of savings, the release of both financial and non-financia.1 regulatory constraints on the supply of housing should be carried out jointly. Only then could the full benefit of financial intermediation to the housing sector be achieved. Fourth, policy makers should review more systematically the objectives that they are trying to achieve when they subsidize housing for the lowest income groups. In particular, the merits of low nominal interest rates in an inflationary environment should be more systematically reviewed: What can they achieve in the short-run and the in long-rua? - 63 - V. DEVELOPING A HOUSING FINANCE SYSTEM A. Introduction Having first looked at the types of financial needs that the dynamics of LDC housing market generates in Part III and then having looked at the savings base available for financial support in Part IV the present section goes into a more detailed presentation of the issues encountered in the development of housing finance intermediaries and of housing finance system. It raises questions such as: What differentiates housing finance from the financing of other sectors? Even if there is one diagnosis of the LDC housing finance problem is there only one model of housing finance development? What are the functions that a housing finance system should perform? How should financial intermediaries mobilize their resources? Must they be deposit non-bank institutions? What are the ways to protect a housing finance system against inflation? Should housing finance be provided by specialized institutions or be part of general purpose financial intermediaries? Should housing finance institutions be run by the -public sector? -Should market lending operations and government low- income housing programs be run through two windows of the same -institution or two separate institutions? Some of these. ue_stions can. be answered fairly completely while others are still actively debated. In this overview, the emphasis is on the presentation of a structured and consolidated view of these housing finance issues. -64- B. Financing Housing Compared to Other Sectors In addition to exhibiting the symptoms of lack of depth and financial repression which characterizes its broader financial environment, the housing finance sector is confronted with specific problems tied to the nature of its operations: (a) It must deal with households on the leading as well as on the deposit side and this leads to substantial risks and high transaction costs compared to the size of the loans involved. In addition, the legal value and the liquidity of the collateral varies greatly and servicing costs as well as payment collections are high. (b) The structure of lending presents major problems of intermediation: more than most others, housing finance institutions are expected to lend long even though they are. borrowing short. (c) Because of this extensive term intermediation they are extremely sensitive to inflationary environments and need to have sufficient flexibility to operate successfully under. a wide variety of economic and monetary conditions. (d) Because of the long-term nature of their operations public confidence in housing finance institutions is closely_related to their level of capitalization and their effectiveness .at loan recovery; public confidence is a prime consideration of banking .auperviso.rs;_-_ (e) Another element that differentiates the housing finance sector from the financing of agriculture, industry or the export sector is that it finances a very long term fixed investment producing a flow of services over several decades, which, given the low degree of - 65 - mobility in many LDC housing markets, tends to be consumed by owner- residents instead of being sold on the open market. This direct consumption of services by home-owners goes a long way in explaining the high proportion of self-financing in the sector. C. Alternative Structures and Functions of Housing Finance Systems 1. Some Determinants of Institutional Structure A wide variety of institutions can be involved in housing finance. In developing countries most of these are specialized institutions because commercial banks often are either not interested in or prohibited from getting involved in the sector. On the private sector side, one encounters thrift institutions (savings and loans, building societies, mutual savings banks, credit unions, housing cooperatives, etc.), mortgage bankers, commercial banks, credits fonciers and other types of development banks as well as non-bank financial intermediaries (private pension funds and life insurance companies). On the public sector side, one encounters public sector housing banks, targeted government funds, social security systems and provident funds. The number of institutions encountered in a country reflects local conditions and the economics of providing financial services such as the minimum cost of entry into the business and economies of scale in providing services through branches. This number is influenced by the total resources of the country, its population, its degree of urban concentration and geographical scale. It will also reflect the structure of financial policies in the country and the regulatory - 66 - environment. Some countries encourage consolidation of institutions in contrast to others which favor decentralization. In a number of developing countries where the government wants to encourage housing there is only one institution organized by the government as either who.lly owned by the central government or dominated by it through a majority of equity shares and control of the majority of board members. Such an approach can be caused by a policy of government control of the economy and the financial system or by impatience with the time required to develop a fully private housing finance system, often both. If one were to perform a flow-of-funds analysis of various housing systems to trace the origin of financial resources and their use, two fundamental systems would be seen. One system based on the- mobilization of deposits directly from the general public which could be called the British-US model in recognition of its historical origins; the other, based on the indirect mobilization of financial resources through the sale of debentures of various maturities to the capital markets, could be called the Continental-European model. Each one of these two basic models has its advantages and shortcomings. In the building society or savings and loans model, the institutions solicits deposits of funds from the general public to lend to households that want to built. Borrowers do not need to be depositors and savers do not necessarily plan to take a mortgage in the future. Each institution is faced with its own term-transformation problem and tries to induce long term deposits by offering a variety -of savings instruments. - 67 - A building-society or savings-and-loans system constitutes a specialized system of housing finance based on independent local institutions for which national regulatory institutions and specialized tax advantages have developed over time. It originates from a banking tradition of decentralization and local control reenforced by the regulatory framework. This model of housing finance presents many attractive features for developing countries: the typical S&L is relatively simple to manage and is a community based institution which, like credit unions, is a natural extension of the informal "rotating credit associations" prevalent in many societies. However, because it performs term intermediation within a single market, borrowing short and lending long for a single activity, it is very sensitive to inflation. The U.S. housing finance system which, until now has been the most complete and most developed S&L system and has been used as a model in many countries, is undergoing very rapid changes. It is becoming less specialized and its evolution since 1974 shows that the reason there were so many small S&L institutions was not so much their successful adaptation to changing needs or their innovative capacity as to the fact that they have been sheltered from competition by State and Federal regulations. Breaking away with traditional resources mobilization methods, mortgage bankers have been much more enterprising than the S&L's and have tapped directly into the capital markets (life insurance companies and retirement funds). They have expanded very rapidly providing almost one third of all U.S. home mortages in 1981. -68- The opposite archetype for housing finance systems is the Continental European model. It is based on specialized banking institutions which mobilize resources for home mortgages in the national capital markets or even abroad. Their major advantage is that they have a much easier time matching maturities between their assets and liabilities. They are also in better position to take advantage of economies of scale for the production of financial services to lower their administrative costs. On the other hand, they may have to pay a higher rate for their loanable funds than what S&L would pay to household depositors. Most importantly, it is clear, in this alternative approach, that the capital market investors will dictate the terms of the mortgage instruments that,will be offered to borrowers. Their growth in LDC's will also depend on the willingness of regulatory authorities to let them have easy access to the fairly limited capital markets. In practice a variety of institutions can be present within the same country and they tend to play a complementary role. For instance, in Germany, the Bausparkassen is a system of contractual savings which complements the mortgages provided by commercial banks. In Latin America, Savings and Loans patterned afterthe U.S. experience coexists with specialized housing banks which do not mobilize their resources directly from the general public. To illustrate the diversity of housing finance systems encountered around the world, Table 4 shows the results of a 1981 survey Table 3 INSTITUTIONAL DIVERSITY IN HOUSING: SPECIALIZED INSTITUTIONS IN SELECTED COUNTRIES 1981 1/ Specialized Number Total Assets per Assets Assets per Assets as Total Institutions of Assets Institution per Capita Percent of Region and Country Population (Number) Offices (US$ M (millions) Office (US$) GDP East Africa Kenya 15.9 3 10 160 53.3 16.0 11 2.41 Zimbabwe 7.4 3 48 856 285.3 17.8 116 18.45 Malwi 6.0 1 6 15 15.0 2.5 3 1.05 West Africa Ivory Coast 8.6 1 - 400(est) 400.0 - 47 4.03 Nigeria - 1 10 - - - - - East Asia/Pacific Indonesia - 1 3 - - - - - Philippines 47.9 87 182 273 3.1 1.5 6 0.79 Korea 38.5 1 116 1506 1,506.0 13.0 39 2.57 Sonuth Asia IndLa 673.2 1 8 568 568.0 71.0 0.84 0.36 Morocco 20.1 1 9 495 495.0 55.0 24.6 3.11 Turkey 45.4 - 1 134 134.0 - 3 0.20 1ortugal 9.8 3 297 8499 2833.0 9.5 867 36.73 Tunisia 6.4 1 16 240 240.0 15.0 38 2.88 Latin America Bolivia 5.6 12 36 77 6.4 2.1 14 2.42 Colombia 26.7 10 58 2,400 240.0 41.4 90 7.60 Ecuador 8.4 11 43 289 26.3 6.7 35 2.83 Domtnican Republic 5.4 15 63 400(est) 26.7 6.3 74 6.45 Peru 17.6 19 187 1,356 71.4 7.3 77 8.23 Brazil 118.7 92 8,382 25,965 282.2 3.1 219 10.7 Argentina 27.7 26 80 952 36.6 11.9 35 1.43 Indust.ralized Countries United Kingdom 55.6 251 6,454 117,958 470.0 18.3 2,110 26.64 Japan - - - - - - - - France 53.5 3 83 2000(est) 666.7 24.1 38 0.32 Germany 60.9 31 n.a 63,361 2043.9 n.a 1,040 7.65 Canada - 60 - - - - - - United States 227.3 4,347 22,135 663,844 152.7 30.0 2.921 25.71 Source: International Union of Building Societies and Savings Associations, 1982 Fact Book, Chicago, Illinois 1982. TF-iThe information presented in this table is indicative and not exhaustive. It is based on reports made to the International Union by the national associations. It covers individual building societies, savings association and specialized home financing inotitutions. The number of institutions listed and the coverage of assets does not represent the totality of resources available to the housing finance sector. - 70 - of specialized institutions. I/ The data presented is only indicative of conditions in each country rather than exhaustive. It covers individual building societies, savings institutions and specialized home banking institutions and ignores mortgage lending by other institutions. For instance, the share of savings assets controlled by specialized institutions varied between 3% for Argentina and 46% for the UK in a sample of 25 reporting countries. The share of the housing mortgage market they served varied between 5% in New Zealand and 96% in South Africa. 2/ What Table 4 illustrates is the degree of development of various systems as reflected by the ratio of Assets to GDP; concentration measured by the number of institutions; the diffusion of services measured by the number of offices and branches and assets per capita; the scale of operations measured by assets per institutions and assets per office. 2. Four Country Profiles A quick glance at India, Morocco, Brazil and the Philippines can illustrate the variety of systems that can be encountered between the two poles of either fully specialized institutions based on intermediation between the capital markets and the borrowing public or of institutions dealing directly with the public both on the savings collection side and the lending side. 1/ See, International Union of Building Societies and Savings Associates, 1982 Fact Book, Chicago, Illinois 1982. 2/ IUBSSA, 1982 Fact Book, Table 11, p.19. - 71 - India is an example of a yet undeveloped housing finance system, operating in a centralized environment of close government control and relying on the capital markets rather than directly on the public for its resources. At present the greatest proportion of housing financial needs (over 95%) are met through informal arrangements. National institutions are just in the process of expanding their activities throughout the states. The capital markets of India operate under a system of credit allocation and tight regulations by the Reserve Bank of India. The housing finance system consists of two institutions at the national level. One, HUDCU (Housing and Urban Development Corporation) is a public agency serving the lower income groups drawing long-term resources essentially from two nationalized insurance companies as well as benefitting from some budgetary inputs. The second, HDFC (Housing Development Finance Corporation) is a very recent housing corporation specializing in finance. At the moment it is serving a population of income higher than the groups served by HUDGO. It is mobilizing its resources from the regulated capital markets and is confronted with a decision whether to move into the direct collection of deposits from households. Morocco, which has been influenced by the.French institutional experience, provides an illustration of the continental model at an intermediate stage of development. There is only one financial institution the CIH (Credit Immobilier et Hotelier) involved in housing finance at present, both for the middle class and low income group. - 72 - The CIH does not rely directly on household savings collected through an extensive network of branches, rather it relies on medium- and long-term securities subscribed to by the CDG (Caisse de Depot et de Gestion) and insurance companies. The CIH floated 23% of all long-term issues in 1978 and 28.4% in 1981. The actual intermediation role is played by the CDG which collects the deposits from the postal savings system, the social security system, and the savings banks. This CDG centralization of deposits implies close control of the government over the allocation of long-term credit. To a great extent the CDG has substituted itself for competitive capital markets for long-term credits. At present, the diffu-sion of financial services to the housing market of MHorocco is limited and covers less than 15% of annual housing investment. To the'extend that the CIH has reoriented itself toward meeting low income needs, the mode of expansion of financial services to the rest of the population and alternative forms of resources mobilization remain open questions. Brazil has a housing finance system which includes elements of both the Continental model and the UK/US model. A housing finance system (SFH - Sistema Financerio da Habitacao) was initiated with the National Housing Bank as the lead agency (BNH - Banco Nacional da Habitacao) financing public local executing agencies (COHAB - Companhias da Habitacao Popular) as well as supervising the activities of a private sector consisting in Federal and State Savings Banks, Real Estate Credit Companies and Savings and Loans Associates (SBPE - Sistema Brasilevio de Emprestino e Poupanca) which is one of the four largest in the world and - 73 - the largest in Latin America. On one hand, BNH solves its long-term credit and intermediation problems by receiving an earmarKed share of social security funds (FGIS - Fundo de Garantia do Tempo de Servico). On the other hand, the SBPE system mobilizes resources directly from households. At present the Brazilian SFE constitutes a dynamic and powerfully structured system in a banking environment where government intervention can be quite significant since government financial institutions provided 51% of all lending in 1978 and in addition 1/ contributed 44% of the funds lent by other institutions. -- The Philippines provide an illustration of a system based on a decentralized approach to banking and reliance on S&L's for housing finance. But this approach strongly influenced by the US experience has led to a shallow and fragmented pattern of extremely small institutions which experience serious difficulty in mobilizing resources and in meeting their long term credit needs. It is a system where the differentiation of financial functions has run ahead of needs, which will sooner or later undergo some consolidation. Government impatience with the inability of S&L's to mobilize enough resources from the public and difficulties in term-intermediation have led to the creation in 1978 of the Home Development Mutual Fund, first a voluntary provident system made mandatory in 1980 becoming effective in 1983. This new housing finance system is still in the rationalization phase. 2/ 1/ The World Bank, Brazil Financial Sector Review, 1979. 2/ The World Bank, The Philippines Housing Finance, A World 6ank Country Study, Washington, D.C. 1992. - 74 - These four brief illustrations could be pursued farther to show that the level of incomes, the degree of urbanization, income distribution, the national financial philosophy and institutional tradition have been very important factors in the growth of these HF systems. In addition, further analyses would show that vigorous government in support of financial innovations is crucial whether there is a preference for a decentralized or a centralized approach to finance. In the first case, government support might lead to a set of sectoral regulations and incentives, in the second it is likely to lead to new specialized public financial institutions. 3. Functions of Rousing Finance System The comparative analysis of housing finance systems in developing countries has barely begun. Even for a single country, the evaluation of housing finance systems requires criteria which, unfortunately, are not well developed. The first step lies in the determination of the range of functions which these systems are expected to perform. It should then be possible to develop methods and measures to evaluate how well these functions are fulfilled by existing institutions. Typically a housing finance system is expected to: (a) mobilize household savings into the mortgage and home improvements loan markets; (b) provide maximum incentives for increasing the volume of financial savings into the economy; (c) allocate the supply of loanable funds among households; - 75 - (d) provide policy controls over the allocation of resources between the urban sector and other sectors of the economy; (e) direct the demand for housing and community facilities toward unused or ineffectively used resources; (f) stimulate efficient methods in planning and construction of residential areas which might require large lump-sum investments; (g) improve the financial and commercial evaluation of projects; (h) extend financial services to all segments of tile population in particular self-employed and lower income households; (i) reallocate funds -from relative surplus to relative deficit areas; and (j) faciL.tate the flow of domestic and international resources into priority areas. D. A Consolidated View of Housing Finance Policy Issues To understand the nature and current operations of a housing finance system the first step is to perform a flow of funds analysis, which would have to be specific to each country. Given the wide diversity of of institutions and their extensive interactions arising in the development of housing finance systems, when it comes to policy issues, a natural way to classify them is by reference to a balance- sheet structure. Whether the problems arise at the level of an institution, a program or a specific instrument, financial results can always be consolidated with respect to the sources of hunds (tne - 76 - liability side) the uses of funds (the asset side) and the profitability oE the system and its viability (the net worth element). The consolidation of housing finance issues is presented in the following three tables (4, 5, 6). In Table 4, concerning the mobilization of resources, the issues are listed according to the source of funds and whether they imply voluntary schemes , mandatory schemes or government transfers. The typical instruments used, the policy objectives, the problems encountered and the typical actions generally taken are listed. On the lending side (Table 5), the various assets of housing finance institutions are listed together with the type of activities financed and, as in the case of the sources of funds, the policy objectives, the problems and the types of action that can be considered. Finally, the elements of the capital base and profitability of the housing finance system are outlined in Table 6 since they reflect the lending capacity and the viability of the system. Ultimately, the confidence that the public has in housing finance institutions rests.on their net worth. .The viability of a housing finance system is dependent on a balanced management of its assets and liabilities. As noted earlier, insufficient attention is being paid t savings mobilization or to the impact of lending regulations on the liability side. Resources mobilization is considered first in Section E, and the issues arising in the use of these funds in section F. - 77 - Table 4 A CONSOLIDATED VIEW OF HOUSING FINANCE PROBLEMS AND POLICIES SOURCES OF FRWS (The Liability Side) POLICY O.1EICTIVFS PROPLElS TYPES OF A-TIO! TINCARY SQlCE IS From 1ndividuial - demand deposits - encourage use of ba-ikinG - inconvenient to house- - improve bra-ching structure - time deposits facilitie- holds - review regulations - encourage sarings - not competitive with Baiks - installment - improve torns of liability - must be tallored to local - cha.ie a=tuarial structure deposits,can- - raise level of not savings unatable income. Tunns nrot - consider interest subsidies tractual savin3s coi:potitLve wLth infornal Fur certain groups market - housing lotterieq - tnpa wider, low Income - encourages speculative - a national decision. More and related scheimes market based on national attitude in low Income groups attractive deposLt terms preferred preferences From InstLtutiuns - debentures - improve terienS of Liabilities- debentures tenns not - a financial policy decision which (housinl bonds - increase flow of resource -oLpetLtLve with other type can be reviewed etc.) into houiing of debentures or approved volume too small !?ULSORy SCHFIES )nLaLvidal - retiremfent pay - rechannal satings through - in.:omplete coverage of scheme- review aztuarial structure deduccion fonal finaicLl syste: - ineqitable ue of resource - re.*ew Lending terms - increase housin fina-ice - use at funds for non-housing - review rates of return to resources purposes saers - improve term structure of liabilities InstLtutions - reulation of - irprove flow4,of resources - cost of resources channeled - revie4 L=plicit sib3tdy struccure retirtment funds into housing to housing higher than in l.ght or na:Ional a-:d uran - insurance companies- Improve ten! stcucture of ruturns provided. tmplict Cbjectives etc. HU liabilities cross-subsidies not reviewed - diversify portfolios adequacely VE.LIE,T TIAmSFERS B orro-in, frum - improve postilon of Mt? - inflationary method - determine conditions under which Centrul Bank system - teras significan:ly below the system can move closer to - provide cheap resources mearKt rate stru:ture competitive markec conoitions - freqjently a curpensarton fur excess reAulactons elsewhere - main benefiaries are higher income groups ') Subsidles (t) C epiCtt subsidLes: - to addreas particularly - never large eno-:-h to meet review in context of financial budget allocations severe low Income identified necd, benefiting policies objective aad hou3ing problems only a minority obj Lc tive (ii) implicit i - increase resource level - A toc approazh which ray be . ca< ex-pC1n1s to 4W inconsistent with overall institutions financiat policiei - indirect tncrea;e of capital cost of o-zher sectors * tax exemption to - increase access to housing - a redIstribucive a,proach should be rejected if not yet in individual borruwers by lowering financial coti facurnl e ive h incotg eroups place. Should be contained 1o wmeans - open ended cormrmnt or deduction ceLings otherwise - misallocation of resources tAx exemptions to other - to encoura e support to - freqently upi-ended tax - should be reviewed, eliminated in financial instituttons housing investments or expenditures r-ti.nitudie favor of market related incentives or corporations as compensations for - resource mLsallozation, compulsory investments hi?h =acroeconomic cost - 78 - Table 5 A CONSOLIDATED VIEW OF HOUSING FINANCE PROBLEMS AND POLICIES B: UsS OF r-:!S (the Ass-ets Sie) POLICFi 1dTCl,E P.3 :E.DI; TYPE.S 9? AC-LION housin: '.-ii: New UJnite co incr:4c supply of now - .loaa criceria LiLted to - redaze loart anaunc units high unit stad.ardi, high - sca-ilacdize tenas inc4)u households - make 11F systua responsi*e - small rutn.ber of Lnrge loa-s %r blundin, of funds - loa% tenns decided )y - altertaco aortgage invesc- sources of funds not house- meats hold cype - alternate mortga3e invest- - inflation aents ExistInZ units increase efficient allocation - often not available - determine what level of HF of the housing stock development would permit such loans for home improve- facilitate discontinuous - high loai origination cost - use bulk loans to inter- ment investment In housing and risk mediate agency haadling screeninZ and spreadin- risk Coi-nercial Loans: housing stabilize construction phase - may not be required from HP - based on local conditions developers of housin;, increase developer syscan if funds atatlable capitalization for commercial banks coserciAl improve neighborhood quality - not considered a-i appro- - use to influence type and developers and employ-ent local priate function of l location of cot:mer.:ial system developmun:/raLse rate of return of portfolio for cross-subsidication. material improve efficiency of building- maut be consistent with production industry housing policy objectives Infrastructure Loans: facilitate ovevall urban * terls of loan (low rate. - accordin; to local conditions: development, increase urban long maturity) not consisren: goverr:nent g,rarcecs, capitcal efficiency with iaSIltcy structure of part-cip.2:ion In specialized UW systen. institutons, indLrect lendi.% % Chrough ;overivren: dopertures SECLTITIES Government Securities: * liquidity regulations * high propor'tton of savin,s * baic on national priorities require purchase of approved subsidicd 'V HF sysie. closely tLed to clarificaLaa securities vithlrawn to meet other diqs:t-n: controversy. development objectives (industrial investment, national debt, etc.) Other Securities: managerial objectives of a a function of the * dictated oy financial positive interactions with orientacion of Hi system institutional structure and other parts of finance (ccntralized vs. de- policies system centralized structure). PRMMSES AND large number of branches - maiaaerial probtm of cost - piggy-backing on other extztir EQUIPN-T needed for access to loans effcctiveness and capital networks and sav-igs mobilization in-obilization (but * reliance on mobile units inflation Ledge) other CAF!ZAL .t AFFILIATED - land corporations encourae the development of * haphazard process not * esta'lish a central agency fo: CO:CZRNS - savings banks efficient specialized Lnsti- reflecting housing plolicy control: - housing co- tutions policy objectives supervision, a jditing, provis, operative- asso- coordinacion of policies of charters, seed capital, ciations sharing of rare ma.-agerial technIcal assistance, etc. - etc. resources - 79 - Table 6 A CONSOLIDATED VIEW OF HOUSING FINANCE PROBLEMS AND POLICIES C: CAPITAL BASE & PROFITABILITY (The Net Worth Element) POLICY OBJECTIVES PROBLEMS TYPE OF ACTION Authorized Capital - determines the statutory - may be a limit to - review the need for capital range of activities of each lending activity increase, determine feasibility institution Paid In Capital - provides resources for the - actual paid-in capital - determine feasibility of capital institution small percentage of increase (budget authorization) authorized amount Legal Reserves protection against risk - raised high to meet - determines whether lower and stabilization of macroeconomic objectives level still consistent with financial system . broad national objectives Voluntary Reserves based on mna.-agement objectives - level inadequate in review finaacial stability under planning new schemes sharply different situation Earnings/Profits indicator of effectiveness - none require subsidies on merger if of (1)assett/liability inadeq-ate. A major element of maragement (2) internal public confidece in RF institution: management of institutions and formal banking systeu E. Sources of Funds: Issues in Resources MIobilization There are only three methods of collecting funds for housing finance: to rely on voluntary schemes, to resort to mandatory schemes or to depend on government transfers. In developing countries it is frequent for an institution to rely on all three methods at the same time, but each source of funds tend to generate specific issues. 1. Major Issues for Voluntary Schemes There are at least five important policy issues in the voluntary mobilization of resources: providing positive real rates of interest; the indexation of resources to protect individuals and institutions against the impact of several inflation as well as fluctuations in interest rates; the stabilization of individual deposits through contractual schemes; the attraction of institutional and corporate long-term deposits; and branching and the quality of financial services provided. a. Real Interest Rates The primary requirement for effective savings mobilization from households, of any income level, is to provide them with a real positive return on their deposits. The raising of a low nominal interest rates will not trigger any significant response from depositors as long as these new levels still yield negative real rates. The Korean experience with the Workers Assets Foundation Schemes offering rates competetive with the curb market rates (after adjustment for risks) shows that such a positive real interest rate strategy is effective. 81- However, this strategy implies that the overall interest rate structure be adjusted, otherwise a subsidy element from tfte government may be required.1/ In addition, if the overall interest structure is not reviewed, high deposit rates for housing may simply trigger a displacement of savings. b. Inflation and Indexation ("Monetary Correction") Long maturities are required for urban investment and inflation is the curse of housing finance. Because so many LDC governments find it convenient to resort to inflation to mobilize resources and because the open economies of LDC's also import inflation from large dominant advanced economies, LDC housing finance systems must offer savings vehicles adapted to inflation, / Four countries stand out for their extensive use of "indexation" or "monetary correction:" Brazil, Chile, Colombia, and Israel. Other Latin American countries like Peru have -been very interested in the technique of indexation to protect individual deposicors and the capital base of lending institutions. Originally, it was expected that indexation would contribute to the control of inflation by increasing contractual and voluntary 11 -In the case of Korea the government was willing to provide short- term interest subsidy time because another objective was to mop up excessive consumer liquidities partly due to foreign remittances from the Middle East. 2/ In this period of worldwide inflation and high interest rates it is hard to believe that the original concept of Credit Foncier was developed in the 1350's on the basis of private 60-year mortgages at 2% interest. - 82 - savings, accelerating tax payments and reducing reliance on inflationary finance. However, indexation has turned out to be a means to "live with inflation" in these countries. Of the three Latin American countries, Brazil is the country which has made use of indexation most extensively since the financial plan of 1964. However,-large discretionary chianges have prevented 1)0 percent links to the price level particularly regarding exchange rates, tax brackets and financial instruments not all of which are indexed. The areas where indexation is most full or automatic involve long-term transactions, long-term loans and Treasury bonds, social security deposits and housing financei/ Monetary correction has succeeded in helping the development of housing finance systems and the rate of urban investment has improved significantly .above previous levels. However, this procedure is far from being universally approved. First, it is a cure of the symptoms and not of the causes of inflation. Second, monetary correction induces severe distortions between indexed and non-indexed areas of the economy. Third, as the Brazilian case shows what started as a three- 1/ For more complete descriptions and discussions of the Brazilian experience see among others: Jack D. Guenther, The Role of Indexing in Brazil's Economic Policies, IMF, IM/74/111, November 1974, Albert Fishlow; Indexing Brazilian Style: Inflation Without Tears? Brookings Panel on Economic Activities, April, 1974; World Bank, Brazilian Financial System Review (Report No. 2790-a-BR, November 1980); Roger J. Sandilands, Monetary Correction and Housing Finance in Columbia, Brazil and Chile, Gower Publishing Co. Farnborou6h, UK, 1980. - 83 - year transitional plan runs the risk of becoming a permanent financial feature of the economy. A more complete review of indexation applied to housing finance would be desirable. Even after indexation, interest rates have not been free to move completely with market forces, the rates have been set by the government. Modifications have been introduced from time to time reducing the real interest rate, sometimes making it negative, ceilings have been imposed, taxes have been imposed on the earnings of depositors and exempted on the interest paid by borrowers. Loving averages have been used when inflation accelerated. Tinkering with the index base has occasionally caused severe drains of savings trom the housing finance system. c. Contractual Savings Schemes In order to mobilize resources at a low cost which would permit future lending at low interest, many countries have been interested in developing contractual savings schemes along the German model of the Bausparkassen or the French Savings-for-Housing Schemes (Epargne-Logement) originally inspired by the German experience. Such programs essentially generate a specialized financing system where the guarantee of a future housing loan at low cost encourages personal savings at deposit rates that may not be high. The idea was originally derived from cooperative and mutualist credit associations. The applicability of the concept of contractual-savings to a new environment must be carefully evaluated. In the German case two crucial factors for success are often overlooked: first, a very stable - 84 - price level during the post World War II period until recently favored long-term contracts; second, the availability of complementary loans arranged by lenders guaranteed that housing would effectively be bought at the end of the savings period. One could mention a third and self- evident factor: at any period of time there were more depositors than borrowers. The remarkable success of the French version of contractual savings scheme after 1966 also stimulated interest in many countries. In tho French system savers are not obliged to use their savings for housing. Three conditions were necessary for the success of the program for which 4 out of 10 French households signed up: savings rates that were competitive with other forms of deposits, a high proportion of savers who were not using their right to a housing loan, a small but growing and open-ended interest subsidy to the scheme. The scheme developed successfully in a mild inflationary environment and complementary loans were available because housing was excluded from credit controls in periods of tight monetary policies. By contrast, the first attempts at contractual savings in Morocco and Tunisia did not develop well because all these conditions were not met and, at the end of their savings phase, contractors could not buy housing with the small loan offered to them compared to the value of a housing unit. A systematic evaluation of the potential of contractural savings scheme seems desirable in many countries as this approach seems to be one of the most promising ones to finance housing for the self- employed. As was seen earlier, such households are typically excluded - 85 - from standard borrowing in most countries because the verification of their income and the determination of their credit worthiness is eitner impossible or administratively too costly. Through a contractual savings program self-employed households, who make up a substantial share of the labor force in every country, have the opportunity to establish their credit worthiness. This group of workers could also expand substantially the resource base of the housing finance system. d. Institutional and Corporate Deposits Housing finance institutions can raise resources by floating debentures if rates offered are competitive with other debentures and compatible with interest charged on loans. One area of significant potential lies with industrial corporations trying to develop staff housing to stabilize their labor force. These corporations could not afford to finance such projects internally but often can develop mutually beneficial plans.with housing finance institutions. This approach has been a source of rapid growth for HDFC in India. e. Branching and Quality of Financial Services Financial intermediation is a service the quality of which has a major impact on the mobilization of securities. Branching is one of the most important decisions for a growing institution. As noted earlier, under serious financial repression housing finance institutions--especially public ones--find it much easier to rely on central bank rediscounts at low nominal rates, on budget transfer, or international donors that reach out for their intended clientele. Premature branching can compound the managerial problems of a young - 86 - institution as proved to be the case with the National Mortgage Bank of Nigeria which was asked to open brancies in each of the 22 states at once, thus diluting its scarce managerial resources. Low cost alternative strategies to reach household depositors exist such as mobile branches in low density areas, agents working on commission such as was done for the "pigmy deposits" of the Syndi,cate BanK of India and is presently being considered by HDFC also in India; reliance on the more extensive 'networks of commercial banks is also feasible and leads to a non-specialized housing finance system. In France, the leading institut.on for the "savings-for-housing" program is the Credit Agricole which has the largest network of branches (including in large cities!). The economics of branching idst be carefully reviewed. The two constraints for branching are the (a) se,' vices provided be competitive with those of informal money lenders; and (b) that the branch be self-supporting within a reasonable period of time after opening. 2. Issues in Mandatory Schemes Mandatory savings schemes have developed all over the world, particularly in Latin American countries where domestic savings were considered insufficient to achieve the growth and employment objectives pursued by governments and, in addition, private institutions encountered serious term-intermediation problems accentuated by inflation. Households save contractually in a variety of forms: life insurance pensions (both public and private), repayment of mortgages, repayment of principal on loans for durable goods, employee termination - 87 - compensation and related forms of payroll savings. Given the need for long-term financing in urban residential investment these various forms of contractual savings look very attractive to housing policymakers. Whenever possible they have attempted to earmark some of these long-term funds for housing or to change them from contractual to mandatory schemes, often both. Mandatory savings schemes are generally not well-documented and have not been studied systematically in most countries. Issues of particular interest are (a) the degree of substitutions between discretionary and mandatory savings; (b) the impact of .inflation; (c) the interest yields of these schemes compared to other forms of savings; (d) their redistributive impact through preferential access given to limited groups of beneficiaries; (e) the relationship between periodic changes in income, consumption, prices and discretinary saving in contrast to mandatory savings; (f) the degree of effectiveness in the use of resources; and (g) loan recovery. a. Schemes Imposed on Individuals Essentially all mandatory schemes applying to individuals are based on the employment relationship, they consist of mandatory financial contributions prorated to wages and salaries which are either applied unilaterally on the employee or on the employer, or both. In LDC's where wage and salary employment is not widespread, the benefit and costs of such schemes are bound to be unevenly spread and the schemes become easy mechanisms for income transfers as well as opportunities for internal cross-subsidies. - 88 - Among the best known systems are Brazil's, the Philippines' and Mexico's. In the case of Brazil the rapid growth of the Housing Finance System (Sistema Financeiro de Habitacao) lead by BNH (Banco Nacional de 1abitacao) is based ,on a 1966 law mandating employers to contribute part of employee salaries into a special account supervised by the Central Bank: the FTGS (Funda de Garantia de Tempo de Servicio) which is a part of the social security system. In 1967 it was decided that BNH would manage the Employee Garantee Fund (FGTS). At present BNH total assets are well over $10 billion equivalent and the resources that it controls though its operations are more than double that amount. This large resource base has allowed BNH to function as a comprehensive urban development bank for the financing of integrated urban development, local government planning, sanitation, construction, electricity and highways. In addition, BNH works closely with a second tier of financial institutions in particular the savings and loan system. SBPE (Sistema Brasileiro de Emprestimo e Poupanca). Because of its influential role on other systems, the Brazilian system should be more closely evaluated for its long-term prospect when the Employee Garantee Fund (FGTS) will have to fulfill its original mandate as a social security system and beneficiaries begin to retire. The uses of resources and the loan recovery experience in various types of investment should be also carefully studied since many low-income groups may perceive the FGTS as a tax or a government redistribution program and may be reluctant or unable to maintain regular repayments. -89 - In Mexico, a similar system was set up in 1972 with INFONAVIT (the Instituto del Fondo Nacional de la Vivienda para los Trabajadores). Its resources come from a 5 percent payroll tax on employers. In addition, the government as an employer contributes 5 percent of the wages for its 700,000 employees. The payment by companies is deductible for corporate income tax purposes. The financial viability of the system needs to be closely examined since loans to workers were granted without any downpayment at an interest rate as low as 4 percent with a 10 to 20 year term plus life insurance on the mortgage paid by INFUNAVIT. The exorbitant benefits implied by the scheme are such that the government had to rely on a lottery to allocate the benefits. In the Philippines, the housing finnce system has recently been reorganized and the original 1978 voluntary program popularly known as "Pag IBIG" of the House Development Mutual Fund (HDMF) was made mandatory after July 1981. Under the scheme all employed persons in the formal sector of the economy see a share of their income deducted at the source. This money is complemented by an equal matching amount from the employer and is deposited into the HDMF fund. The funds deposited earn a minimum of 7.5% compounded annually and the total amount is to be retrieved 20 years later at maturity. Originally, employees contributed 1%; in 1982 they are contributing 2%; they are expected to contribute 3% in 1983 and thereafter, Participants are given two mutually exclusive options: to receive the lump sum payment after 10 years or to be given a 25-year home mortgage from an accredited bank at 9% for a maximum of - 90 - $100,000 per member with up to three primary members being entitled to pool their rights. In addition to the new HDMF scheme the Philippines housing finance system can also draw on the resources of the private social security system (SSS) or those of the separate Government Services Insurance System (GSIS). In 1980, the family sources of funds of the Philippines housing finance system were: - Social Security System (private sector) 20% (declining) - GSIS (public sector employees) 42% - Commercial Banks 32% - Savings Banks and others 6% Clearly mandatory savings form the core of the formal Philippines housing finance system. The 1982 Philippines Housing Finance report, after reviewing the current st ucture of the Philippin:s system, expressed concern about the actuarial stability of the present system, the viability of the current programs, the degree of protection workers included in the program could actually expect and the fairness of the system to the extent that a limited number of high-income beneficiaries could deplete quickly all the resources available. 1/ 1/ The World Bank, The Philippines Housing Finance, A World Bank Country Study, 1982. - 91- In general, compulsory schemes based on salaries are very attractive to government because they minimize the cost of mooilizing resources and give the central administration total control over a large pool of resources which they can.allocate according to politically accepted priorities. One way to evaluate these schemes is to compare the quality of the financial services they provide with other type of locally available housing finance services, since beneficiaries must receive a fair return for their loss of control over their own resources. b. Mandatory Schemes for Businesses Distinct from the contributions that employers have to maKe by matching employee mandatory contributions, there are additional programs requiring employers to invest the equivalent of 1% of their wage bills for housing. Such schemes have been used in various countries such as France, Italy or Japan. Indian public corporations are investing significantly in worker housing and in former African British colonies the right to migration to towns as well as to housing was tied to employment. While experiences in industrial worker housing are extremely diversified, the mobilization of business resources may be of some value provided that a great deal of flexibility be left to individual firms about the form of their investment. It can be made directly (through own construction, acquisition of property, loans and subsidies to employees) or indirectly (through payment to non-profit organizations, professional associations, building societies, family allowance funds or low-income housing programs). - 92 - It is not clear who would benefit from such benefits in many developing countries. One must note also that employer-provided housing is frequently overlooked as a significant component of public housing policy. For instance, in India, public corporations have very large investment programs which might well include 1 to 3 percent of the resources going to housing. The aggregate impact and nature of such publicly finance housing has yet to be quantified and evaluated. c. Regulations Imposed on Financial Institutions One of the most attractive ways for housing finance institutions to raise long-term resources is to have the government make it commercial banks, life insurance companies and retirement systems to earmark a fixed--preferably large--proportion of their resources for housing. Then it is no longer necessary to develop extensive networks to collect savings directly from households, the resource base becomes relatively predictable and, because the regulated financial institutions have no choice for the use of their funds, their rates might be significantly lower than the opportunity cost of capital and prevailing market rates. In India for instance, the greatest part of the formal resources flowing into housing finance come from the nationalized Life Insurance Corporation which is required to invest every year a fixed proportion of its net accretions into housing. In such a context, the resource mobilization strategy for the chief financial officer of a housing finance institution then becomes a bureaucratic game of getting the central bank to increase the quota allocated to housing and to have - 93 - ones debentures rated as acceptable for liquidity and reserve requirements of commercial banks and other financial institutions. The benefit of such mandatory schemes are that they improve the flow of resources into housing, they improve the term structure of housing finance liabilities by their long maturities and allow a diversification of sources. The drawbacks are that the cost of resources channeled to housing are higher than the returns actually provided to the forced buyers of debentures and that the implicit cross- subsidies which are part of such arrangements are seldom reviewed adequately. In some countries, like Korea, the purchase of five-year housing debentures at low nominal interest and often substantial negative real rates of interest has been forced on the households who are deemed to be privileged beneficiaries of urban growth. These mandatory housing bonds purchases are applied to land transactions, automobile purchases, purchases of high-value houses and a significant number of other items. These compulsory bond proceeds are then allocated to a National Housing Funds to finance public sector low- income housing programs. This implicit form of taxation does not show up in budgets but involves administrative costs. In essence, these bonds are government-mandated, private-sector transfers. 3. Government Transfers to Housing Government transfers and subsidies to housing via the housing finance system do not represent the totality of subsidies which could be made available to households but they constitute a si,nificant part of -94 - it. A comprehensive review of the housing subsidy issue should be a high priority in urban sector work as there are significant gaps between the rethoric of the countries and the realities of the programs with which the World Bank is involved.. Housing finance subsidies are one of the three channels available for subsidies, the other two being demand side subsidies (in the form of income or rental subsidies) and supply side subsidies (through land costs, construction costs and favnrable taxation of construction companies). Subsidies to the cost of capital can involve Interest rates, taxation, the treatment of capital gains and depreciation. a. Borrowing from the Central Bank As a compensation for government intervention in their lending programs many housing finance institutions--especially in countries with central credit allocation systems--enjoy privileged access to cheap funds through a line of credit at the Central Bank or the Treasury, for instance in Tunisia and Korea. These lines of credit tends to be a standard feature of public housing banks. In addition to being a source of inflation they are not conducive to tight cost control of operations. In fact the level of profitability of the institutions through which public sector programs are being financed often drops to a low level of priority as their managers are evaluated on the basis of a wide variety of other "public sector" criteria. - A Central Bank's line of credit provides cheap resources and improves the position of the housing finance system. The terms offered are significantly below prevailing market rates and could be seen as a - 95 - compensation for excess regulations elsewhere in their housing finance operations. The main beneficiaries are higher income groups which are also the main users of formal housing finance. b. Explicity Subsidies - Budgjv Allocations to Housing Budget allocations to housing are rather infrequent in most LDC countries. They are meant to address particularly severe emergency low-income needs and typically fall short of the resources that appear to be required in part because the demand for free funds will always exceed their supply. A more appropriate use of budgetary allocations is for the financing of the paid-in capital of public sector housing finance institutions which will expand their borrowing capacity to finance the sector. c. Implicit Subsidies: Tax Exemptions and Tax Expenditures The most extensive form of subsidy to housing finance systems are tax exemptions. These open-ended "tax expenditures" will grow with the institutions and after a few years become firmly embedded in the structure of housing finance systems. They may take the form of tax exemptions to the housing finance institutions themselves, of exemptions to individual savers or of exemptions to other financial institutions or corporations purchasing housing debentures. Their objectives are to increase the level of resources in the formal housing finance system, to increase access to housing by lowering capital costs and to encourage support to housing investment or to act as compensation for mandated low-income housing investments. Such tax subsidies are often triggered by short-term difficulties and the level - 96 - of resources involved may appear to be small at first, but as the housing sector grows these implicit entitlements become politically very sensitive and are almost impossible to remove. Ad hoc tax incentives may be inconsistent with the overall objectives of financial policies, they may indirectly increase capital costs to other sectors, they imply a redistributive approach favoring high-income groups whoch are the main and largest users of the housing finance system. In financial sector work and in more specific reviews of housing finance the estimation of the subsidies involved through budget allocations, tax exemptions and direct interest rates subsidies is of primary importance.l/ In the 1982 Philippine Housing Finance Review it was estimated that tax expenditures on the new housing finance system during the first year of operations were already almost equal to budget allocations to public housing programs. F. Uses of Funds: 1. Conventional Lending As noted previously, the insufficient development of conventional lending for housing has two major consequences: poor financial discipline over housing investments in the form of technical and commercial evaluation of projects submitted for financing reduces 1/ In a recent review of the U.S. housing finance system it was estimated that on a sector-wide basis Federal tax expenditures for housing in FY81 anounted to a staggering figure of US 32.U45 billion in a single year. By comparison, during the same year the volume of financing going to new housing investment was estimated at US $36 billion, see Chapter 13, Housing and Community Development, The Reagan Experiment, The Urban Institute, Washington, D.C., 1982. - 97 - the ability of financial institutions to encourage technical innovations or to steer housing investment in desirable directions. Regarding housing loans, in addition to the bias tov3ard relatively few large mortgages for high cost and high income units, there is not enough attention paid to making loans for home improvement. In a period of rapid urban growth a substantial amount of investment takes the form of upgrading of existing units for rental accompanied by land use intensification. Little effort is made to evaluate the kind of profitable financial services which could be developed in the form of short-term loans for such operations. The other major flaw of conventional lending programs is the systematic exclusion of a large percentage of self-employed households with good earning power. A third area of conventional lending that should be systematically reviewed is the financing of housing developers through whom more efficient types of housing can be provided. At the moment many housing fihance institutions focus too exclusively on individual borrowers,.they could also provide support to housing cooperatives and building societies. The financing of commercial development to improve the quality of neighborhoods is another area that offers great potential for profitable lending and opportunities for project level cross-subsidies to improve housing for lower-income groups as is systematically done in Korean public housing projects for instance, and many Bank integrated urban development projects. The use of appropciate financing to induce private sector firms to supply more low-income housing is an area which is just - 98 - beginning to be explored by the World Bank. It offers the promise of a more adequate division of responsibi-littes between on one hand the public sector which would define objectives, identify priority areas for investment and facilitate financing and on the other private sector firms which are inherently more flexible and more adaptive than public sector construction agencies in implementing projects. Given the need for large amount of low-cost if not necessarily very low-income housing the potential of the private sector to develop large amounts of low-cost 1/ housing to meet rapidly growing needs should be developed.- Finally, because of the lack of well-articulated housing policies, housing finance institutions have seldom been asked to support the development of more standardized residential construction products2/ nor to improve the efficiency of residential construction firms. There is little evidence that commercial banks are frequently solicited either. 2. Capital Investment in Subsidiaries and other Institutions In some countries with fairly developed housing finance systems some of the largest housing finance institutions have shown a 1/ The idea that housing agencies, whether national or state, should move toward the role of promoters with the private sector becoming the developer operating inside a set of performance specifications is being explored through various projects in Morocco, Pakistan, the Philippines, Manila, Egypt, Korea and Kenya. 2/ Confusion should not be made here between the standardization of building inputs and industrial housing. Industrial housing around the world has proven itself just as ineffective in providing low- cost/low-income housing as new towns have been unsuccessful in absorbing large proportions of rural-urban migcation at lcw cost; frequently for the same reasons, - 99 - tendency to invest capital in other types of public or private corporations operating in the urban sector such as land corporations, savings banks, even providing the capital for housing cooperative associations, etc. The objectives are to encourage the development of efficient specialized institutions to improve the coordination of policies and to share scarce managerial resources. It is not possible to be dogmatic regarding such capital investments, however, when the institutions providing the original capital have been entirely public institutions, very frequently the new affiliated concerns have not been as effective as originally anticipated. The role of a central agency for policy control responsible for supervision of the various housing finance institutions, auditing, provision of charters, seed capital, technical assistance, etc. is generally a useful question to review before determining how housing sector policy and housing finance policies can be brought more closely together. At present, coordination among housing institutions, when it is effective, takes place mostly among public sector institutions which may then be quite reluctant to allow new private institutions to enter the sector. G. Critical Factors for the Growth of Housing Finance Systems 1. Policy Environment and Regulatory Role of Governments The policy environment is the foremost factor in the growth of a housing finance system in three ways: by being favorable to financial development, by being supportive of housing objectives and by providing consistent and coherent actions in support of the shelter sector. - 1u - Arbitrary controls over the interest rate structure, as well as bad monetary and fiscal policies can have a very severe impact on financial resource mobilization. When the financial sector as a whole doe not fulfill its resources mobilization role well, what can be the prospects for housing finance institutions? There is no set of government regulations more extensive than that of financial institutions. 1/ The Central Bank exercises close control over banks and non-bank financial intermediaries. In addition there are special regulations of housing finance with several objectives and often "special circuits" to isolate the housing sector through "preferential" measures. This regulatory environment is replete with conflicting objectives: on the one hand housing is expected to be encouraged but resources should not be drawn away from other priority investments; low-income needs should be the government's priority concern but badly needed middle-class savings should be encouraged; depositors should be encouraged but interest rates should not be too high lest they push up the entire rate structure; low-income households should be given affordable mortgages but they should pay the opportunity cost of capital. When giving a low priority to housing finance governments overlook two important aspects of housing: one is the high proportion of tertiary and secondary urban infrastructure investments in housing 1/ See Arvind Virmani, The Nature of Credit Harkets in Developing Countries, A framework for Policy Analysis, World Bank Staff Working Paper, No. 724, June 1982. - 101 - needed at low-income levels and the difficulty of financing it privately because of their many "public goods" dimensions. Financial . intermediation to increase the scale of projects and facilitate the provision of this infrastructure is very important. The other point is the frequent confusion between long-lasting housing units and the annual flow of services which they produce. If the rates of return on housing were small or zero there would be no landlord left in the cities which is not the case, and the implicit rental value of owner-occupied dwellings would disappear too. Traditionally the central banks of LDC's have taken a very strict regulatory attitude rather than stressed their developmental role and they have tended to be quite wary of financial innovations. One of the difficulties in changing a financial structure is that because money is fungible, a comprehendive framework to appraise the joint effects of specific policies on the financial system at large -- or on housing finance in particular - need to be developed. There is abundant evidence that the actions of one policy can be unintentionally blocked by another policy. Even in the absence of a systematic framework, country specific evaluations of trade-offs between policies and interactions between institutions within a housing finance system are nonexistent. Even if the financing needs of urban residential investment are large and rapidly growing, the evolution of specific financial intermediaries must take place in a way that will not promote further fragmentation of the capital markets. Too often policy recommendations - 102 - have been limited to specific institutions and reforms of the financial structure have not been based on a comprehensive data analysis of savings and of the flow-of-funds structure. Isolated operations typically lead to further fragmentation. In particular a housing finance system with a large number of institutions may be quite incomplete when a well-balanced system serving all its functions may require a small number of institutions; in many LDC's a proliferation of institutions may simply waste scare managerial resources. Proposals for creating new institutions with specialized functions are too frequently the outcome of limited evaluations in the absence of an adequate prior evaluation of the broader macro-economic context and the capital market structure of the country. The creation of a new institution must be based on more than the assumptions that increased competition cannot hurt and that some financial innovations might follow. The three principles of financial regulation are clear. They are: competition, efficiency and the development of long-term finance. In contrast, the trends in actual regulatory environments are not so evident. Dominating the scene at present is the debate on the merits of universal banking vs. specialized financial institutions. The new emphasis in developing and developed countries is to move in the direction of multipurpose financial institutions which are expected to provide long-term financing, to increase competition and to lower the cost of intermediation. In the new environment, financial policies would consist in broad sector regulations defining the framework within which any multipurpose institution would have to operate. Such an - 103 - approach contrast with prevailing situation in LDC's where financial regulations tend to be highly specific to each specialized institution. In the case of housing finance the verdict on universal banking is far from clear. In fact, the limited IFC experience with housing finance suggests that in a multipurpose institution the housing function will lose out because of the serious housing finance long-term intermediation problem compared to other commercial forms of bank lending. Further evaluation of the merits of multi-purpose banking is clearly necessary. 2. Income Level and Development Opportunities Quite clearly, in countries at low levels of income, especially in sub-Saharan Africa, the level of demand for financial services is not very high and the effective demand for housing not very strong. This explains the frequent emphasis on government housing institutions which function de facto as pass-through agencies for government or even international funds. The extent to which government housing banks could function as apex lending institutions encouraging the growth of credit unions or other form of financial association better suited to the need of the self-employed majority has not been well explored. The quality of management as well the lack of clarity of housing policy have also delayed such a form of public support to financial innovation. In middle-income countries opportunities for the development of financial services are much greater. Higher household income levels together with a greater density of demand for financial services and a - 104 - larger volume of transactions create economies of scale adequate for the growth of financial intermediaries, even if there remain sharp geographical differences in the growth of these services. As the volume of institutional financial services increases there are further opportunities for the development of differentiated market functions. Among them the development of a secondary mortgage market to improve the flow of long-term resources into housing and to give more flexibility to financial operations may be an important one. 3. Financial Deepening and Specificity of Low-Income Housing Needs In the housing finance sector there has been a major and continuing confusion between "public policy" toward housing and the housing finance sector on one hand and "public sector activities" regarding housing on the other. This confusion parallels the typical housing policy evolution that one can observe over time in LDC's. At first, there is no explicitly articulated housing policy, then "housing policy" means what the public sector does through public agencies for housing and, finally, the government realizes that public sector investment will remain a very small part of total housing investment and that the incentives and regulations of the quantitatively dominant . private sector activities are essential to the overall success of housing-policy. It is essential to distinguish clearly between the need. to increase the market penetration of housing finance institutions into the higher segments of the informal market (financial deepening) from the financing of public services in very low income residential areas - 105 - which, because of poor or no cost recovery, rely on government subsidies or cross-subsidies. The deepening of financial services requires coordinate policies to release non-financial constraints on the supply of housing (especially confused land titles and rent controls) as well as better financial services. It may include the development of mortgage insurance, new deposit methods and mortgage loans to the self-employed, and the offering of a broader ranges of services for all phases of residential investment (land development, construction loans, interim financing, mortgages). Also in periods of rapid urban growth a substantial amount of investment takes the form of upgrading of existing units for rental accompanied by land use intensification. Often, little effort is made to evaluate the kind of profitable financial services which could be developed in the form of short-term loans for such operations. Under stable prices, there are several convergent reasons not to expect the demand for mortgage credit by low-income households to be very large. First, the income elasticity of demand for housing services has been found consistently smaller than one. Second, given the low- income elasticity of demand, the demand for mortgage credit can be expected to be even smaller since most dwellings cannot have 100 percent financing. Third, irregular incomes tend to reduce further the desired consumption of housing. For these reasons, the residents of "slums and squatters settlements" raise a different kind of problem from those of - 106 - "clandestine settlements". Only a minority among them could realistically contemplate some form of legal ownership at any period of time on terms comparable to the better-off group in "unauthorized" or "clandestine" settlements. The problem then is not to finance housing but rather basic infrastructure, sanitation and neighborhood services which are the poor's first priority over the housing shell itself. The minimum amount of services that they desire is in the nature of a "merit good" and involves some form of direct subsidy or cross-subsidies, typically channelled through public sector construction agencies. There has been some question as to whether these construction agencies should increase the scope of their responsibilities. Should they handle all low-income activities under one roof? For instances, in Thailand, the National -Housing Authority which was originally a construction agency has been drawn into the loan origination and servicing business to disburse the funds it received from international donors. Is this a viable long-term orientation? If greater stress is placed on resource mobilization rather 0 than on the production of housing only, it seems desiraole to separate the financial function from the construction function since the types of services provided are sharply different. As noted earlier, the loan origination function will be greatly strengthened if potential borrowers are already known as regular savings depositors. The separation of the financing phase from the construction phase will make it easier for the f low-income middle-class to have choices in its housing rather than being confronted by a de facto public sector monopoly construction agency of - 107 - the kind that the SNIT-CNEL interlocking structure in Tunisia seems to be producing. 4. Encouraging Appropriate Institutions At least three separate levels of issues must be differentiated in designing appropriate institutions. First, there is the need for appropriate financial intermediaries as seen in the context of financial deepening and the differentiation of financial functions performed by type of activities and borrowers as financial market growth permits. Second, and most important for the public sector, is the choice of public instruments for low income housing policies. The unsatifactory performance of both financial and construction functions by public housing agencies already noted, needs to be reviewed. The third group of issues is related to the functions that should be performed by public financial institutions and in particular the role of apex institutions. A full examination of the role of apex housing finance institutions, their potential strength and current limitations is a high priority because country after country have found it necessary to create such organizations. What should be their role in housing sector development? How can they advance sectoral objectives? What kind of impact should they have through the use of their funds? On the types of projects financed? On cheir geographical distribution? How can-they promote the deepening of the financial structure? By establishing links with other types of institutions? By institutionalization functions performed by the informal curb market? Could they enter into - 108 - cooperative agreements with regional banks to improve the geographical spread of housing finance services? Could they influence positively long-term debentires? Should they take equity participation in other institutions? Are they financially viable? Essentially, publicy sector apex housing finance institutions should be evaluated along five major lines:- their impact on sector development; their contribution to deepening the financial structure; resource mobilization; their management, organization and procedures; their financial position in the short-term and the longer-term. - 109 - VI. CONCLUSION A. Appropriate Housing Finance Strategies This overview of housing finance in LDC's has started with a definition of the housing finance problem as the need to reconcile the three different objectives of affordability for households, viability for financial institutions and resource mobilization for national economic planners. The evidence presented has shown that there should be a clear policy differentiation between the responsibility of government regarding the financial stability and deepening of housing finance institutions and those concerning the specific needs of the lower income groups. The role of government should be divided into two distinct activities. The first is to encourage financial innovation to provide housing finance tq households on a financially viable basis; this implies organizing thte market and generating financial services for fully serviced residential investment. The second objective consist in developing specific programs for the low income households with a major focus on the urban infrastructure that individual households cannot provide for themselves but which is central to efficient long-term urban development. This second role goes much beyond housing finance proper and overlaps with the broader issue of financing urban development. The need to have differentiated objectives can best be met if a central public agency for policy control can become responsible for the supervision of the various institutions, their crediting, the - 110 - provision of charters, seed capital, technical assistance, etc. It should help determine how housing sector policies and housing finance policies can be brought more closely together. At present whatever coordination takes place is generally limited to public sector institutions and ignores much of the activities of private institutions.. While housing finance institutions will expand over time, the primary objective in raising the level of financial savings may-not be so much to increase national savings but to spread the benefits of financial intermediation and to reduce the degree of fragmentation of the economy where old and new technologies with sharply diverging rates of return might be found, to facilitate larger scale investment and to limit the constraints imposed by self-financing on the adoption of more efficient, higher return production techniques. In addition, in the case of housing, there is evidence that increases in savings rates do also occur with housing investment as home ownership is a powerful incentive for households to reduce consumption and increase savings. An important point of entry for expanding housing finance is the largely unser-vTiced group of self-employed households who have serious difficulties in gaining access to housing finance institutions. Too many housing finance programs have been geared exclusively to a minority of salary and wage earners and systematic efforts should be made the majority of non-salaried housefolds. The choice of strategy will depend on the level of economic development of each country. - 111 - The central issue of low-income residential investment is finance -- low cost finance. There is considerable evidence that most subsidized credit in a financial system goes to wealthier households and that in urban development higher-income groups have been very adept at using cheap resources and low-cost projects. It is also true that many projects have relied on low-cost (subsidized) counterpart public funds and have achieved much less than full cost recovery; a familiar ploy being to claim that direct recovery was not practicable but that indirect recovery through taxation and user charges was feasible when in fact tax collections and local governments have consistently proven to be less than perfectly efficient. Plans for the development of a housing finance system or the rationalization of the existing institutional structure are dependent on national priorities regarding the overall development of the financial sector including macroeconomic policies, priorities in the allocation and pricing of financial resources and national preferences (or more exactly past experiences) regarding alternative forms of institutional development. For national financial authorities, who have to play a crucial developmental role, housing is only one of the major sectors to be financed, they are also quite concerned with industry and agriculture. Since the actions of one policy can be unintentionally blocked by another policy it is important that proposals for reforms and development of the housing finance systems be congruent with the overall financial development strategies of the country. Ideally, reviews of the housing finance sector should be part or follow comprehensive - 112 - financial sector reviews. In that case, the place of the housing sector in the national economy and the development process as well as the levers that it can provide to help achieve national development objectives can be fully understood. Housing finance has for too long been treated in isolation from other development finance needs. B. Areas for Further Analysis It is not surprising that an overview of housing finance in LDC raises more questions than it provides answers. To help.meet the identified policy needs, further analysis would, be desirable on the following topics: 1. Housing Finance System Development a. What are the stragegic choices for the development of a housing finance system? What are the relative merits of direct vs. indirect savings mobilization for term intermediation in housing finance. b. What are the specific problems of public housing finance institutions at low levels of development in meeting their role in sector development, apex lending, appraisal capacity, etc.? c. Are there many alternative directions for financial deepening in middle-income countries? What are the circumstances under which specialized secondary mortgage market institutions might be needed? - 113 - d. What are the relative merits of specialized housing finance institutions as opposed to financing by multifunctional banking institutions? 2. Resources Mobilization a. What are the best ways of expanding services to the self- employed? In particular, what are the strength and limitations of contractual savings schemes? b. What are the strengths and shortcomings of mandatory savings schemes? Issues of particular interest are: the degree of substitution between discretionary and mandatory schemes; the impact of inflation; the comparative yield of these schemes compared to other savings; their redistributive impact; their effectiveness in the use of resources. 3. Flexibility in dousing Finance a. Term intermediation in an era of volatile interest rates and inflation is a widespread financial problem. A comprehensive review of indexation experiences in housing finance would be very desirable. b. Considerable new knowledge has been accumulating since 1975 on Alternative Mortgage investments (AMI's) as a means of improving access to housing finance. A more comprehensive evaluation would be desirable regarding: their impact on the viability of lending institutions; their acceptability on the mortgage markets and the - 114 - extent to which investor preferences determine the choice of AMI; their regulatory and legal implications. 4. Financing Low-Income Housing a. Clearly the most urgent issue concerns the role of subsidies in low-income housing. A review of available experience should help answer the question of whether subsidies can (instead of should) be avoided. A fuller examination of the actual experience regarding subsidies, in their direct form as very low-cost finance or in their indirect form as incomplete cost recovery, would be very helpful to determine where the priorities are in this current period of slow economic growth and increased capital scarcity. Should public resources be limited to certain activities such as infrastructure provisions? It would appear that as soon as one goes beyond general principles, the level of economic development, the level of urbanization, the degree of development of urban institutions, the political priorities of the government are determinant. It is quite noticeable that past a certain level of development, public low-cost financing of urban residential investment for low-income groups - in effect subsidies - has been considered necessary in every country. Then, what are the second-best approaches? - 115 - b. Further studies of progressive housing investment are needed. In particular, they should help determine the role of non-financial constraints, their impact on the demand for financial services and on the type of financial services to be provided. 5. Improvi.ng Institutions and the Production of Financial Services There are many unanswered questions regarding the problem encountered in the production of financial services in LDC's such as economies of scale in branciing management, internal organization, procedures, financial planning, implementation of resource mobilization programs and the development of project lending.