68243 Housing Finance in Sri Lanka: Opportunities and Challenges November 2007 Sadiq Ahmed Sector Director Simon C. Bell Sector Manager Tatiana Nenova Senior Financial Sector Specialist with Kevin Villani Professor, Former Chief Economist, Freddie Mac and Lohita Karunasekera Consultant Finance and Private Sector, South Asia Region World Bank Table of Contents Executive Summary ....................................................................................................................... 1 1.0 The Housing Finance Sector in Sri Lanka: Descriptive Statistics and National Policies 1.1. Among Competing Development Priorities, Why Housing Finance? .................................................................. 6 1.2. Macroeconomic Policy Framework ..................................................................................................................... 6 1.3. Macroeconomic Risks to Mortgage Market Development ................................................................................... 7 1.4. The Policy Environment in the Housing Finance Sector ..................................................................................... 7 1.5. Where Does Housing Finance in Sri Lanka Stand and What Is The Road Ahead? ........................................ 9 1.6. Structure of the Report..................................................................................................................................... 11 2.0 The Housing Market, Home Needs, and the Current Demand for Mortgages 2.1. Housing Needs................................................................................................................................................ 12 2.2. Housing Stock Supply and Housing Turnover................................................................................................. 13 2.3. Construction Methods and Costs ..................................................................................................................... 13 2.4. Rental Supply and Vacancy Rates .................................................................................................................. 14 2.5. Precisely Measuring the Gap in Housing Investment ........................................................................................ 14 2.6. Demographics .................................................................................................................................................. 14 2.7. Effective Demand for Home Mortgages............................................................................................................ 14 2.8. Addressing the Housing Needs of Low-Income Groups.................................................................................... 16 3.0 Current and Potential Growth in Mortgage Supply 3.1. Main Providers of Home Mortgages ................................................................................................................ 18 3.2. Current Sources of Funds for Mortgage Lending .............................................................................................. 21 3.3. Growth Potential of Mortgage Lending Funds ................................................................................................. 22 3.4. Banking and Home Mortgage Sector Efficiency ............................................................................................... 24 3.5. Uses of Funds for Mortgage Loans .................................................................................................................. 25 3.6. Mortgage Products and Loan Terms ................................................................................................................ 26 3.7. Residential Development––the Multifamily Market ........................................................................................ 28 3.8. Lending by Microfinance and Informal Institutions .......................................................................................... 28 4.0 Constraints on the Development of the Home Mortgage Market Legal and Regulatory Framework 4.1. Land Registration and Titling ......................................................................................................................... 30 4.2. Foreclosure and Eviction ................................................................................................................................. 31 4.3. Taxation......................................................................................................................................................... 32 4.4. Property Valuation Standards ........................................................................................................................ 32 4.5. Land and Housing Price Movements ............................................................................................................... 32 4.6. The Credit Information Bureau ....................................................................................................................... 33 Risk Management 4.7. Operational Risk ............................................................................................................................................ 34 4.8. Investment Risk and Return............................................................................................................................ 34 4.9. Political Risk .................................................................................................................................................. 35 5.0 Policy Options: Elements of a Housing Finance Strategy 5.1. A Housing Finance Policy Paradox? .............................................................................................................. 36 5.2. A Growth Scenario ......................................................................................................................................... 36 5.3. An Inclusive and Sustainable Housing Finance System ................................................................................... 37 2 Annexes Annex A: Statistical Annexure Annex A1. Population, Income, and Price Levels Annex A2. Housing Stock in Sri Lanka by District: Type of Housing Unit, Type of Structure, and Tenure Annex A3. Building Units by District Annex A4. Total Assets of the Financial System Annex A5. Lending and Deposit rates Annex A6. Mortgage Products Offered in the Home Lending Market, Major Players Annex B: Strategies to Assist the Development of Low-Income Housing Annex C: Mortgage Operations and Operational Risk Management: Theory Basics Annex D: Secondary Mortgage Markets, Mortgage Capital Markets, and Securitization Annex E: Successful Housing Finance Development—Examples from Malaysia, Korea, and India Annex F: International Best Practice in Housing Policy Annex G: Longer-Term Focus Areas for Policy Makers 3 EXECUTIVE SUMMARY Sri Lanka has embarked on a gradual transition from a system of directed credit in a highly segmented market toward an integrated market-driven housing finance system. This transition has included an increased role of private universal banks in the immediate term and a functioning secondary mortgage market in the long term. To nurture home mortgage markets, this ambitious agenda would require a stable macroeconomy, low inflation, and careful fiscal policies. An active system of housing finance provides real economic benefits and positively affects savings, investment, and household wealth. It provides an investment option for long-term funds in the economy as an alternative to investment in treasury bonds. In turn, each dollar invested in the housing sector catalyzes economic activity in other sectors, exerting an indirect positive impact on employment levels, the retirement system, fiscal returns, and consumption. Housing finance enables households to accumulate assets that can provide the collateral for their investment needs, thus stimulating small business. Housing finance development boosts equitable economic growth and reduces poverty by improving living conditions, empowering the middle- and lower-income population, and strengthening communities. The mortgage lending market has been swelling by leaps and bounds in the past three years, at real annual rates of 10 to 30 percent. Housing prices have been increasing as well, fueling both construction and speculative or investment housing purchases. Prices may be driven up by perceived and projected shortages of urban land because of the extensive government holdings (85 percent of land is government owned). Speculative buying may be fueled by a real borrowing rate close to zero and the dearth of alternative investments. The 18 percent increase in real land prices is in line with alternative investments in equity markets. Thus, price behavior appears rational and a bubble in the housing market is not evident. National Housing Policies Housing policy focuses on improving government land use and maximizing the use of the existing housing stock by providing basic public services and upgrades. Sri Lanka national housing policies are defined by the Department of National Planning, and more recently, by the Ministry of Housing and Common Amenities, which has yet to become fully functional. Policies are implemented by the state institutions, such as the National Housing Development Authority and the Urban Development Authority. The share of state-owned housing institutions such as the State Mortgage and Investment Bank (SMIB), Housing Development Finance Corporation (HDFC), and National Savings Bank (NSB) has come down to about one-third of the mortgage market share, as the private sector has displaced the government as the primary housing finance provider. The two state-owned housing banks need further improvements in governance, management, and operational efficiency; should compete with each other and with private banks on lending to middle-income groups; could play a more active role in the provision of lower-income housing; and should improve the use of their share of the government budget to the wider benefit of the Sri Lankan society. The Housing Market, Home Needs, and Current Demand for Mortgages Housing needs are significantly larger than effective demand, because of poor access to housing finance (for middle-income groups) as well as high housing costs (for lower-middle-income groups). The housing demand of Sri Lanka’s population of 19.9 million, or roughly 5 million households nationwide, is currently counterbalanced by the existing 4,687,157 housing units. The national housing shortage is estimated at 350,000 units, or 7.5 percent of existing stock, and the annual increase in household housing needs is estimated at up to 100,000 units. The cost of construction has increased about threefold since 1990. To boost effective demand for homes to match actual national housing needs, housing finance availability is essential. The housing finance gap in Sri Lanka potentially includes up to half of the Sri Lanka population. This portion of the population is capable of servicing a mortgage loan but has no access to finance. Effective demand for home mortgages is limited to one in seven 1 households and is confined to the high end of the market (highest income decile). Low-cost housing, if financing were available, would be affordable to a considerably larger share of the population (the top 60 percent). Potential demand for housing finance by middle-income groups is much higher than their effective demand, and new lending approaches to expand access to finance are not undertaken on a sufficiently large scale either by the market or by state lenders. The demand for house rehabilitation and upkeep is only met in a limited manner. Currently, about one-third of the existing units are semipermanent and require upgrades, and about 0.7 percent of the households in the country live in shanties. Low-cost rental markets have not developed to address the demands of households who cannot afford homeownership. This is due in part to strong tenant protections against eviction and in part due to a culture of homeownership. This hurts the lowest-income households, especially in large cities. The government housing institutions and state-owned specialized banks are carrying out an outdated mandate to bridge single-handedly the housing and financing gaps. Following government liberalization and withdrawal of funding, their mission should be redefined, as they achieve further and increasing improvements in efficiency and governance, portfolio re-vamping, training, modernization / computerization, and updating of risk management skills, within the rules and risk management measures introduced by CBSL. Current and Potential Growth in Mortgage Supply Private commercial banks and NSB are satisfying the effective demand in the country. This demand is composed of the highest-income population, accounting for about two-thirds of total mortgage lending in the country. The two specialized state-owned housing banks, SMIB and HDFC, account for 8 percent and 7 percent of housing mortgage credit, respectively. Some of their housing lending is provided to middle-income groups (with explicit or implicit subsidies) and threatens to crowd out potential private sector supply to these groups. The Development Finance Credit Corporation (DFCC) provides funding for developers. Foreign-owned banks do not have a significant presence in the Sri Lankan housing finance market. In order for the mortgage market to rapidly expand beyond current effective demand and eat into some of the existing housing finance gap, adequate mortgage funding is needed. The required liquidity for fast growth cannot be provided by existing funding sources. Basic and robust secondary mortgage market solutions (such as covered bonds or a liquidity facility) would make this possible. In the longer run, once a sizeable primary mortgage market of a certain scale develops, securitization will become a viable option. Home mortgage lending accounts for about 3.5 percent of total assets of the financial system and for 5 percent of the banking system,1 which is low by comparison with other countries. Private sector credit in Sri Lanka has increased since 2004, prompting CBSL to tighten monetary policy to reduce high credit expansion. Historically, lending to the government and state-owned enterprises has been more than 10 times the amount lent to households for housing. The amount of funds available for housing finance does not permit a rapid expansion of primary mortgage markets and a stable basis for active secondary market development. Deposits provide one source of funding for mortgage lending that could comfortably fund a modest growth in home mortgages. Deposits satisfy the effective demand by high-income households but fail to go a long way toward closing the existing housing finance gap. For a more rapid expansion in mortgage lending, the long-term funding in the country is insufficient, and competes at a disadvantage with government deficit financing. The NSB, the largest deposit-holder in the economy, has about 80 percent of its assets invested in government securities. The two state provident funds, the Employee Provident Fund (EPF) and Employee Trust Fund (ETF), are 90 percent invested in treasuries, as are the private provident funds and life insurance companies. Foreign investment has not shown strong trends for growth. The banking sector is reluctant to expand mortgage lending to a wider middle-income group. This reluctance perhaps is due to a perceived lack of bankable opportunities and high entry costs into the new market segment or to a lack of sufficient credit information and adequate credit-scoring mechanisms to manage risk effectively. State-owned banks continue to have some problems with nonperforming loans (NPLs), particularly SMIB with 25 percent. Similarly, the two state specialized lenders leave substantial 2 room for efficiency improvement. That said, private licensed commercial banks (LCBs) generally exhibit an admirable level of operating efficiency for a country at this stage of development, although this efficiency remains below levels found in member countries of the Organisation for Economic Co- operation and Development (OECD). Private banks are experimenting with adjustable rate mortgage loans. Historically, state lenders have offered only fixed rate mortgages, which expose them to considerable market and interest rate risks. Maturities range from 15 to 25 years, whereas bank borrowing is short term, exposing the entire system to liquidity risk. At private LCBs and the NSB, the average loan is around SL Rs (Sri Lankan rupees) 1 million. The state housing banks do not go much further down the income scale—the average loan is SL Rs 0.25 to 0.6 million, falling short of reaching lower-income groups. Microfinance lenders go much lower—the average loan to small and medium enterprises (SMEs) is SL Rs 15,000 to 100,000. The mortgage market remains segmented by income groups, and middle-income households that would be viable mortgage borrowers under more advanced lending techniques do not have access to housing finance. Some of this demand is picked up on a small but fast-growing scale by microfinance institutions (amounting to 0.1 percent of financial institution assets) that currently fund housing repair and upgrades. These institutions have limited potential in the medium term to expand into the home loan market for the poor because of their currently small level of outreach. The proposed Micro Finance Institutions Act (expected to be enacted this year) will establish a sound regulatory regime for microfinance firms. The lending through microfinance entities likely will remain segmented from formal home mortgage finance for the foreseeable future. Yet, first attempts for bank downsizing exist in the country. Hatton National Bank, the largest private commercial bank, has opened 120 microfinance offices to provide small business loans in addition to its retail bank branches. Constraints on the Development of the Home Mortgage Market The housing finance sector needs to have a supportive regulatory framework that does not impede its growth but stimulates it. Sri Lanka is in the pilot stage of implementing a title registration system and a cadastre, in a few selected jurisdictions, to curtail boundary and ownership disputes (full implementation is planned for the next 15 years). Further advances are warranted, including property registration, collateral realization, modernization, data provision, better land use, and additional improvements in the residential development framework. Parate powers (that is, the ability of a lender to foreclose and sell a defaulted property without going to court) greatly improved the ease of foreclosure by banks on mortgage collateral (the collateral sale takes about four to five months). Yet the effectiveness of foreclosure is hampered by the weak eviction powers for most lenders. This explains the few cases of actual parate application that have been made so far in the country. For nonbank mortgage lenders without recourse to parate powers, the civil court process is long and complex (about three to five years). Stamp taxes are more distortive than other types of taxation, because they discourage sales and encourage underreporting of prices. As a result, recorded prices for private transactions may be but a third of the actual price. Market data is scarce, precluding fast-response policy decisions, as well as careful market analysis. The lack of property price data and fragmentary borrower records hampers loan origination and servicing. CBSL’s initiative for a property price index is well taken and requires support. Shortcomings with available computerization levels, technology, and staff training require action. These problems appear quite manageable, however, especially at private banks. The Credit Information Bureau (CRIB) provides detailed information on approximately 80 percent of loans made. The work of the bureau is constrained by technology limitations. A modernization project has begun to provide online access to CRIB credit information reports. The bureau is expanding the scope of the operation to include any credit provider, thereby including SME and microlending as well as a wider potential pool of users such as insurance and telecom providers, utilities, and the superannuation funds. CRIB is exploring strategies to provide value added services, such as credit scoring, fraud prevention, and consumer protection. A second potential constraint on mortgage market development is the ability of financial intermediaries to manage the risks of mortgage lending. Absent the technology and systems for efficient operations, banks would find it hard to profitably offer a rich range of relevant products. Some private lenders are well on this path, but most banks, especially SMIB and HDFC, lack credible risk management and operational efficiency procedures. State-owned specialized lenders cannot offer 3 adjustable rate loans, because they do not have a servicing system capable of making adjustments, and thus are subject to interest rate risk. This risk typically is reduced by funding mortgage lending with long- term debt instruments, such as privately placed general obligation bonds and mortgage-backed bonds. Experiments with private placements have occurred and realistically can develop into a secondary mortgage market in the long term. Liquidity is not expected to represent a systemic issue in Sri Lanka, although funding long-maturity mortgage loans with short-term borrowings can become a concern, especially if the Sri Lanka mortgage market undergoes a rapid expansion. Policy Options: Elements of a Housing Finance Strategy Luxury housing and home financing supply abounds. Some flats financed at subsidized rates are rented or immediately resold for profit. Simultaneously, many middle- and lower-income households have no access to home financing (and some have no homes) and face scarce market rentals, the efficiency of land use could be improved, and housing turnover is low. What would explain this seeming paradox? First, markets are segmented. Difficulties with recovering the money from collateral cause banks to rely on regular salary and other secure means of repayment for their lending, shying away from middle- income households and focusing instead on regular, mostly government, employees. Second, the playing field is not level for private mortgage lenders, and primary markets are not strong as a result. State-owned banks compete in sectors in which the private financial firms are active, instead of addressing failed markets that cannot be commercially viable. Subsidies are not precisely targeted at vulnerable groups and are instead chasing middle-income households, since government housing institutions have lost budget funding. These factors account for weak primary markets that cannot support securitization, and a housing finance gap that largely excludes middle- and lower-middle- income groups from access to affordable homes. There are two growth scenarios for the Sri Lankan housing finance market. First, the slower development relies on current effective demand by high-income groups, which is being met adequately by the private commercial banks and the NSB. These financial institutions can be expected to comfortably handle growing demand at 15 percent on the value of housing loans a year or more in the coming 10 years. The pace of growth in home mortgage credit should more than keep up with the pace of new construction. That scenario would leave a considerable gap in housing finance provision for middle- and lower-middle-income households. Second, a faster-growth scenario would require stronger institutional capacity, maintained macroeconomic performance, lower fiscal deficits, and further liberalization in the lending environment. A second significant defining factor for a rapid mortgage market expansion is the ability to evict and realize collateral. With collateral functioning only as a threat, and not as an asset of monetary value, mortgage lending will remain confined to the top income groups and will continue to co- rely on regular salary income and other assurances of palpable monetary value so lenders can ensure repayment. Market data, borrower records, and real estate pricing are critical factors for mortgage market development. Building an Inclusive and Sustainable Housing Finance System: Suggested Areas of Focus Macroeconomic stability is a required foundation for a well-developing housing finance system. In the case of Sri Lanka, inflation risk and government debt crowding-out are particular concerns that the National Development Strategy plans to tackle. A workable legal and regulatory system is central to providing an enabling policy environment for housing finance. Two legal prerequisites require emphasis from international experience. The first necessity to develop primary mortgage markets is a functional land registration and titling system. This work is ongoing in Sri Lanka. The second prerequisite to develop active and efficient mortgage markets, as suggested by international efforts, is data availability. In particular, property and house price indexes are required, as well as further improvements in the functioning of CRIB and its possible privatization. In international practice, private credit bureaus have provided credit information profitably and efficiently. 4 Policy makers should focus on a robust and enabling policy. The government should forge ahead in its current strategy of housing finance liberalization and continue to reduce government interventions in the market. Policies should be simple and parsimonious, not complex or costly, in keeping with existing financial institution capacity and the level of sophistication of the current financial infrastructure. The policy choice between generic commercial banks and specialized mortgage lenders is on the agenda. Practice suggests that integrating specialized banks into the general regulatory system could eliminate market segmentation. The state-owned mortgage banks (SMIB and HDFC) would benefit from a rethinking of their strategy and mission, as well as further improvements in their governance and operational efficiency. The rationale for having two separate institutions might merit rethinking as well, especially as they compete with each other and with the private sector. SMIB and HDFC were created at a stage when Sri Lanka’s mortgage markets were developing; however, the current growth shows that the private sector is capable of absorbing existing demand, except for lower-income groups. Specifically—  Strategic refocusing of SMIB and HDFC should be considered. The banks could undertake two sets of activities, which are not mutually exclusive: (1) fully competitive housing lending without any (explicit or implicit) subsidies or guarantees at a level playing field with the private sector; and (2) again on an equal footing as the private sector, be allowed to channel transparent, well- targeted subsidies exclusively to the lower-income groups (and unavailable to middle- or upper- income groups).  Depoliticized boards of directors would function better. They could be in a stronger position to make management accountable and could enforce the requisite modern internal controls.  Deeper efficiency improvements would require upgrades and human resources training, as well as modernization and computerization of existing operations.  Improved risk management and credit loss indicators would need to address the exposure to credit and market risks as well (perhaps via reductions in NPLs and a transition to adjustable mortgage rate instruments). Transparent, direct, and well-targeted subsidies to lower-income groups should replace existing interest rate subsidies, which are costly and distort the budget. Competition would be increased by making the subsidy for low-income housing available to the private lenders, at the same terms. Transparency would enable precise estimation of the subsidies’ future costs to the government. Existing subsidy programs were implemented at an early stage of housing market development and now require revision as financial liberalization and private mortgage growth have changed the country’s housing finance landscape. A detailed study of existing subsidy programs would identify potential budget savings and efficiency improvements via better targeting, transparency, and more efficient subsidy forms. More aggressive growth in primary mortgage markets is constrained by the availability of long- term funds. Such growth requires work on the term transformation issues. Middle-income household access to the mortgage market is further inhibited because of the inability of banks to lend without collateral, on the one hand, and the difficulties in realizing collateral value, on the other. Given these impediments, closing the housing finance gap for middle-income households suggests basic solutions, such as private placements of covered bonds by mortgage lenders or a liquidity facility. Securitization is an option in the long run, but it will not develop in a vacuum––it needs to base activities on a strong and active primary mortgage market of a significant scale, which currently does not exist in Sri Lanka. Whichever instruments are used, the policy success will be predicated on the ability to ensure a willing and able counterparty to the transaction that would invest long-term funds. Banks cannot finance from abroad. Whether it is a foreign investor or domestic institution (such as the EPF, ETF, private provident funds, and insurance companies) or the government (in the case of international success models of liquidity facilities), a counterparty is necessary for the market to develop. The considerable holdings of government debt by the financial system might be a potential concern when assessing the ability of the banking system to attract long-term mortgage funding. 5 Chapter 1 The Housing Finance Sector in Sri Lanka: Descriptive Statistics and National Policies With a per capita gross domestic product (GDP) of US$1,355 and a population of 19.9 million, Sri Lanka is on track to meet a number of the Millennium Development Goals. It has maintained a steady growth rate at around 5 percent for the past two decades in the context of civil conflict and a tsunami disaster. Much of this resilience has to do with the relatively well-developed human capital and the continuation of market-friendly reforms since the late 1970s. The government’s current National Development Strategy, provided in the Mahinda Chintana: Vision for a New Sri Lanka (2006), emphasizes an active buildup of infrastructure and domestic investment, including housing, while boosting economic growth to the 8.5 percent mark. This development scenario in the housing arena brings to the fore the importance of a well-functioning mortgage finance system, capable of servicing the housing finance needs of the country, and tapping into the savings and investment potential of its rapidly growing middle-income population. 1.1. Among Competing Development Priorities, Why Housing Finance? The most immediate benefits from a well-functioning market for home mortgages are increased availability of credit at a lower cost and improved consumer satisfaction with housing services across the income spectrum. Beyond this, functional mortgage markets are key to meeting urbanization and demographic challenges in terms of a growing housing demand and preventing slum proliferation. In Sri Lanka, population growth is projected at a low 1 percent in the coming decade, but in urban areas that rate will be three times as high. An active system of housing finance provides key economic benefits and positively affects savings, investment, and household wealth. In turn, each dollar invested in the housing sector catalyzes economic activity in other sectors, exerting a possible indirect positive impact on employment levels, the retirement system, fiscal returns, and consumption. The increased savings effect is particularly important for Sri Lanka, where the domestic savings rate tends to be low. Housing finance development boosts equitable economic growth and reduces poverty by helping households build assets, improving living conditions, empowering the middle- and lower-income population, and strengthening communities. Overall, developing mortgage credit, ensuring effective property rights, and removing obstacles to commercial lending for housing not only are the pillars of a sound housing strategy, but also match the broader initiatives of supporting private sector development, improving the investment climate, and increasing financial depth, which are the key elements of the government’s growth strategy. 1.2. Macroeconomic Policy Framework Figure 1 The stable economic growth and macroeconomic framework in the Market rates in Sri Lanka country has enabled the government to 25 make a bid for a significant jump in growth rates and a major dent in 20 poverty. Economic growth, averaging 5 LCB Mortage Rate percent for the past 25 years, was SMIB lending rate insufficient to reach many segments of 15 NHDA lending rate the population, however. The 2006 3-year T-bond yield National Development Strategy calls 10 Inflation for raising economic growth to 8.5 LCB Deposit Rate percent by 2010 and to more than 10 5 (AWDR) percent by 2016.2 The goal is to halve the percentage of the population living in poverty from 23 percent in 2002 to 0 2002 2003 2004 2005 2006 12 percent by 2015. This requires a significant increase in investment. In Source: Central Bank of Sri Lanka 2007. 6 addition to planned increases in public investment in infrastructure, the development strategy has to rely on a boom in private domestic and foreign investment. In turn, raising private investment requires the withdrawal of the state from certain economic sectors to allow for the expansion of the private sector and avoid crowding-out. The current fiscal budget deficit of 9.1 percent of GDP is planned to be transformed into a surplus, largely by cutting expenditures. Furthermore, public expenditures will be refocused toward infrastructure investments to support the planned business investment boom. This fits well with the government strategy of pulling out of the housing finance sector, to economize on budget as well as enable the private sector to fill in for housing investment, as government funds are instead channeled in full toward a complementary investment in infrastructure. In the past four years, the country’s inflation has increased from below 10 percent to more than 13 percent, as lending and mortgage rates have mirrored the inflation patters with a margin (except for the National Housing Development Authority [NHDA]). Real government bond rates are low, especially in 2005–06. Deposit rates have been close to zero in real terms (detailed rates are presented in annex table A5). 1.3. Macroeconomic Risks to Mortgage Market Development The ambitious agenda for mortgage market development leaves little room for error. The reforms require fiscal discipline and resource targeting, monetary stability, and an investment boom in productive capacity that is financed with domestic and foreign savings flows. Achieving a robust macroeconomic environment would protect home mortgage markets from a number of potential downside risks. Inflation undermines long-term contractual savings, the best and most important source of housing funds in many countries. Deposit intermediaries have safely funded fixed rate mortgages with short-term funding in many countries as long as the financial markets and inflation were stable. But the recent run-up in inflation in Sri Lanka caught the deposit financed mortgage lenders off guard. These lenders now realize that inflation may be somewhat unpredictable and interest rates may be volatile, and in the future, they will have to manage the resulting risk. The recent boom in the construction of condominiums and rapidly rising land and house prices may well be one source of this concern. The “Financial System Stability Review� (Central Bank of Sri Lanka [CBSL] 2006a) specifically addresses this issue and concludes it warrants monitoring. Fiscal and monetary policy discipline, while stringent, must be maintained. In this sense, an additional risk is imposed by the high level of government debt held by the financial system and banks in particular. Government budget financing should be careful not to crowd out alternative private sector uses of financing, including housing finance. High national savings rates are crucial to fuel these investment increases. An additional factor of importance not covered under the National Development Strategy is the need to generate long-term contractual savings or an alternative investment strategy for the two public provident funds, the Employee Provident Fund (EPF) and the Employee Trust Fund (ETF), which could be major sources of long-term funding for housing and other investment needs. The “Financial System Stability Review� cites the growing mismatch between mortgage maturity and funding maturity as an additional source of risk to the financial system. Housing production is a big user of scarce economic resources. Government investments must stay focused on infrastructure and avoid outlaying large resources in housing that could be successfully provided by the private sector. The “Road Map for Monetary and Financial Sector Policies in 2007 and Beyond� (CBSL 2007) cites this risk as "particularly challenging." 1.4. The Policy Environment in the Housing Finance Sector Sri Lanka has embarked on the gradual transition from a system of directed credit in a highly segmented market toward an integrated housing finance system. This transition has included an increased role of private commercial banks in the immediate term and a functioning secondary mortgage market in the long term. Past policies centered on government-owned institutions, nationalized land, and housing 7 provision programs diverted scarce fiscal resources away from other potentially more productive investments and likely contributed to the less-than-desired rate of economic growth.3 Housing subsidies have been used to help households acquire formal sector housing, curb the expansion of informal settlements, and upgrade existing informal housing. These efforts have not always been successful, however, and ultimately have not closed the existing gap between estimated housing "needs" and the existing stock. Presently, the uneven availability of subsidized mortgage finance in the public sector hampers the private production of middle- and lower-income housing. Housing policy in the National Development Strategy is divided into three components. The first component is to enhance and encourage private sector participation. The second is to put government land to higher and better uses, using the proceeds from the increased land value to pay for housing for displaced dwellers and low- to middle-income housing construction. The third component is to maximize the use of the existing housing stock by providing basic public services and upgrades. The central government is no longer directly involved in the construction of houses, although inevitably it is involved indirectly through the land disposition and development process of its self-funded agencies. New production programs for selected groups are frequently proposed. The major state-owned participants in the housing finance markets together account for about a third of the mortgage market share. This system has increasingly become unviable because of competing government priorities and fiscal pressures. The government has recognized the need to develop a self- sustaining housing finance market. Increasingly, commercial banks and specialized financial institutions are investing in housing. Implementing institutions for the government housing policy have somewhat lost the sharpness and clarity of their mission and direction, together with most of their government funding. For example, the NHDA’s board of directors is politically appointed and its management is tied to their dictates. The state institutions do not have adequate governance, management, and operational efficiency (though for some state banks it is improving, such as Bank of Ceylon and People’s Bank) to play a sufficiently active role in the provision of lower-income housing beyond its current contribution, or to better utilize the government budget to the wider benefit of the Sri Lankan society. Sri Lanka national housing policies are defined by the Department of National Planning, and more recently by the Ministry of Housing and Common Amenities, which is yet become fully functional. These executive organs facilitate and coordinate the operation of the government housing policy implementing institutions. The existing housing policy implementing agencies and state housing banks are insufficiently focused on lower-income housing, failing to address market failures in that market, as well as misdirecting (implicitly or explicitly) subsidized loans to middle-income markets and crowding out potential viable commercial alternatives. The National Housing Development Authority (NHDA) is the national government entity that implements government housing policy. It was established in 1979 to assist the development of homes for the rural poor and to provide the necessary infrastructure surrounding new housing developments, including roads and water. In addition, in its early days, NHDA built rental housing for the urban poor. Because rents were kept below maintenance costs, most developments deteriorated, forcing the sale of units. After government subsidies were curtailed in the mid-1990s, the NHDA was forced to redirect its operations to more viable middle-income population groups. NHDA manages substantial assets of mostly urban land owned by the government, using it for housing development projects. This development is typically done in a joint venture with a private sector developer, as NHDA contributes the land in return for a say in the distribution of the newly constructed housing and a share of the profits. Current production is mostly urban, multifamily, and middle income.4 The deep subsidy and imprecise targeting of need-based housing provision allows buyers to abuse the system. NHDA engages in urban renewal and joint ventures on private land. Their comparative advantage in each case is the ability to resettle existing tenants without the lengthy court process facing private developers. People who are resettled are offered one unit per household and are credited with the value of the forfeited land and property against the acquisition price. For such developments, average unit prices run from SL Rs 800,000 to 1,200,000, with a down-payment of SL Rs 7,500. Financing is secured mostly through state-owned banks, but private financing is also available at market rates. NHDA experimented in 2007 with direct subsidies to buyers financed from straightforward land sales to developers. 8 The Urban Development Authority (UDA), much like NHDA, has a primary responsibility for addressing low-income needs, but it has an exclusively urban focus. UDA does not currently receive budget transfers, except for special projects. It manages and develops government-owned land and infrastructure, frequently in cooperation with private developers (in which case it provides the land) or municipal governments (for which it develops municipal government land). UDA bears a significant focus on housing needs of displaced households. Similar to NHDA, most low-income programs have been curtailed. For example, past strategies of low-cost high-rise construction (14-story buildings) for low-income groups have been abandoned because of high maintenance costs, in favor of five-story walkup structures. Several organizations are active in low-cost housing development. The National Building Research Organization (NBRO) undertakes research on alternative construction materials and focuses its activity on housing provision in flood plains and landslide areas. The Centre for Housing Planning and Building provides advisory services in the planning, design, and building of houses and promotes cost effective housing operations. Real Estate Exchange Ltd. assists with urban shanties upgrading. The Plantation Human Development Trust upgrades houses in the estate sector, where about 6 percent of the population lives, and housing facilities are mainly provided in line rooms. 1.5. Where Does Housing Finance in Sri Lanka Stand and What Is the Road Ahead? In developed economies, housing as an asset makes up from 20 to 50 percent of the reproducible wealth (figure 2). It is a major motivation for household savings and significantly influences household consumption. In addition, it affects inflation, financial depth, labor mobility, and the balance of payments, as well as budgets through taxes and subsidies. In Sri Lanka, formal financial sector mortgage assets represent 3 percent of GDP, as compared with a world average of 14 percent, and 4 percent in India, 2.5 percent in Bangladesh, 0.6 percent in Pakistan, 22 percent in Malaysia, 16 percent in Thailand, and 12 percent in China (figure 2). Figure 2 Housing Finance Depth in Sri Lanka, Region and World Comparison 70% 60% 50% 40% 30% 20% 10% 0% EU - 15 Colombia Czech Rep. Chile Iran Peru China Mexico Jordan Croatia Poland Ghana Canada Tunisia Russia Malaysia Thailand Morocco Slovakia Sri Lanka Hong_Kong Hungary Brazil USA Latvia Bolivia Slovenia Housing debt to GDP (% ) Source: World Bank 2006. The potential to develop mortgage financing in Sri Lanka is high and housing demand (including effective demand) is substantial. Housing lending has seen remarkable growth rates in recent years (figure 3). Sri Lanka’s population of 19.9 million constitutes currently about 5 million households. According to Census data for 2001, the national housing shortage is estimated at 350,000 units. Projections estimate annual increases in household housing needs at up to 100,000 units per year. In addition to new constriction, the housing market demands significant housing upkeep. Of the existing total stock (4,687,157 units), about a third is categorized as semipermanent, improvised, or unclassified, and needing substantial improvement. Sri Lanka’s housing needs, particularly for low-income groups, were further accentuated by the 2005 tsunami disaster, which left at least 110,000 houses damaged, of which only 61,000 had been rebuilt by 9 end 2006. Of the estimated US$1 billion (4.5 Figure 3 percent of GDP) in tsunami-related damages that Sri Lanka suffered, housing losses are estimated at US$300–350 million (World Bank 2005a). National housing needs are therefore substantial. Effective demand, however, has been much smaller than housing needs, because it is determined by the ability and willingness of households to pay for a home. Roughly, one in seven households effectively demanded a mortgage by the end of 2006, that is, individuals amounting to 3 to 4.5 percent of the population had taken out mortg+ages.5 To boost effective demand to Source: CBSL 2006. match actual national housing needs, the availability of housing financing is essential. Major commercial banks granted 98,000 loans during 2006, amounting to SL Rs 34,227 million. The State Mortgage and Investment Bank (SMIB), the Housing Development Finance Corporation (HDFC), and National Savings Bank (NSB) granted 35,137 housing loans valuing SL Rs 12,187 million in the same year. Another strategy to close the disparity between effective demand and housing needs is to lower the cost of housing. Section 2.7 shows that a majority of the population could afford low-income housing at current prices, if financing were available and credit screening were appropriate.6 Housing finance does not only involve mortgage lending on new construction or purchased property. Substantial funds are also needed for housing rehabilitation, although this is rarely provided through mortgage financing. Microfinance lenders tend to finance housing improvements directly via unsecured loans. They finance housing upkeep indirectly, because many small enterprises operate out of the home, and enterprise loan funds eventually finance home upkeep and other uses. The current market is composed of the wealthier bankable households with regular salaries (blue rectangle at the top of figure 4). Figure 4’s vertical axis shows the population by income deciles, from 0 percent (bottom income decile) to 100 percent (top decile). Access to housing finance can be expanded in several ways. First, from the top down, population and mortgage market growth will naturally enlarge the market (shaded blue rectangle).7 The lower half of the market Figure 4 consists of households who cannot afford to Sri Lanka Housing Market at a Glance purchase a new home but Bank already own one (green Grow th Bank and shaded green 100% GAP Market Serviced by rectangles). Microfinance 90% mortgages companies currently are Bank responsible for financing 80% Serviceable Finance housing maintenance and 70% GAP (market growth) upgrades of a small 60% Potential Not serviced but MFI portion of homes and Market bankable (new small business (shaded 50% Finance GAP housing) green rectangle). The gap 40% Serviced by MFIs between the population 30% Housing serviced by commercial Cost MFI banks and that financed 20% GAP Not serviced but Market bankable by microlending is the 10% Rental (upgrades) housing finance gap (red 0% Market Not bankable rectangle) and represents GAP the focus of this report. This gap includes households Source: Authors’ calculations. 10 that can afford to buy new low-cost housing but that are not considered bankable by commercial institutions. This gap also includes families who own their home and are in need of upkeep funds but who have not yet been reached by the limited scope of the microlenders. Downscaling by banks, new credit-scoring techniques not relying on collateral, and better credit information would push some of the red rectangle households up to the bankable category. Microlending expansion could eat away at the gap from below, but its current limited reach suggests only a modest potential for expansion in the medium term. The lowest deciles of income include households that cannot afford new housing but that already own homes and need funds for house upkeep (green rectangle). Improvements in basic housing construction technology would lower home costs and allow these households to increasingly afford new homes, using them as collateral for financing. In the very long run, a possible widespread reach of microfinance institutions (such as exists in Bangladesh, for example) would expand access to finance into the low- income sector. The lowest yellow area represents households that do not own and cannot afford to buy a home, and households that require government or extended family aid or other short-term solutions. In the long run, international experience shows that the households without access to the financial system can secure housing via a well-developed low-cost rental market. 1.6. Structure of the Report The purpose of this study is to identify contributing factors and constraints to more aggressive growth and deepening of the housing finance markets in Sri Lanka. This undertaking is an attempt to provide basic qualifying measures of the housing finance market, as well as to focus attention on further data needs. The availability of this data would enable crisper diagnosis and better policy targeting in the sector. This report is by no means an exhaustive analysis of all aspects of the housing finance sector, nor is it intended to be. Rather, it highlights the most palpable strengths and barriers to progress, in view of the current needs and opportunities in the country. The work focuses on recent performance of the mortgage markets, and to keep the paper succinctly directed toward identifying promising financial sector avenues for growth, it dedicates relatively little attention to nonfinancial factors and reforms affecting these markets. These pages offer a discussion of low-income housing realities and some policy options, but this is not the primary focus of the study. The report is structured as follows. Chapter 2 provides an overview of the current housing markets, home needs, and demand for mortgages in Sri Lanka. Chapter 3 reviews existing mortgage supply, sources of funds, growth potential of the home mortgage market, and the main mortgage lenders and products on offer. Chapter 4 highlights the legal and regulatory structure, focusing, inasmuch as is relevant to mortgages, on registration and titling, collateral realization, taxation, valuation, property price movements, and credit information availability. This section also discusses risk management issues in the housing finance sector. Chapter 5 presents the elements of a housing finance strategy for Sri Lanka that serves to raise awareness of the mortgage market potential and lists various policy approaches and payoffs involved. 11 Chapter 2 The Housing Market, Home Needs, and Current Demand for Mortgages 2.1. Housing Needs The housing demand of Sri Lanka’s population of 19.9 million, or roughly 5 million households nationwide, is currently counterbalanced by the existing 4,687,157 housing units. Shortages have developed, especially in urban housing. About half of the Sri Lankan population will live in urban areas by 2010. Currently, about 0.7 percent of the households in the country live in shanties. About 6 percent of households are housed in estates, where housing facilities are mainly provided in line rooms. Single houses account for 80 percent of the population’s housing, another 5.4 percent live in flats and 3.3 percent live in attached houses. Official estimates put the gap in housing provision at 350,000 housing units.8 According to the Ministry of Housing and Plantation Infrastructure, the overall shortage of houses in Sri Lanka could reach 650,000 units by 2010. About one-third of the housing stock, more than 1 million units, is estimated to require improvement or rehabilitation as well as better public service. The National Development Strategy calls for greater utilization and upgrading of this stock as well as increases in infrastructure services, rather than replacement. Market failures preventing the closure of the housing gap can be attributed to a faulty regulatory framework (land and titling, eviction rules, etc); underdeveloped housing provision market; improperly targeted directed government programs; and housing finance problems, which are the ultimate focus of this report. In the sphere of housing finance, this gap is explained by the following main factors. High interest rates and housing prices preclude low- and middle-income households from affording a house. Low penetration of banks and microfinance institutions into clientele from these income groups is another major factor for the housing gap. This gap is predicated by weak credit information, the lack of technology, and a perceived riskiness for lending to nonregularly salaried employees, as well as problematic enforcement of eviction provisions. Conversely, drastic improvements in education and literacy have helped Sri Lankans to feel more comfortable with banks and financial instruments, including mortgages. At 41.6 percent of the population, the poverty headcount at US$2 per day is high in Sri Lanka (World Bank 2007a).9 Low savings rates through formal financial intermediaries have contributed little to the growth of mortgage financing (see table 1), and household expenditure on housing is around 11 percent (CBSL 2006b). No "savings for housing" schemes are available, but several banks run well- advertised contests for the best savers, giving away such prizes as cars and other items. Table 1. The Savings Rate in Sri Lanka Year Domestic savings rate National savings rate 2000 17.4 21.5 2001 15.8 20.3 2002 14.4 19.5 2003 15.9 21.6 2004 15.9 21.6 2005 17.3 23.4 2006 17.1 23.4 Source: Central Bank of Sri Lanka 2007. Note: 2006 figures are provisional. 12 2.2. Existing Housing Stock and Housing Turnover Annex table A2 provides more detailed data on the housing stock by province. Tenure statistics divulge an underdeveloped rental market––nationwide, 76 percent of the housing units are owner occupied and 5.4 percent are rented. An additional 1.3 percent of housing is encroached by squatters. In Colombo, 71.6 percent of the housing units are single houses, 9.2 percent are flats, and the corresponding figures for attached house/annex and row house/line room are 8.1 percent and 6.4 percent, respectively. Huts and other types of housing units constitute only 1.7 percent. The majority of housing units (72.3 percent) is owned by a member of the household; 17.5 percent of households live in rented or leased housing units and 4.8 percent households live rent-free. Another 1.5 percent of households live in encroached housing units. In addition to housing units, Colombo features collective living quarters (8,477), institutions (2,599), and nonhousing units (89,349), which bring the total number of building units to 608,103 (annex table A3). Housing market turnover rates are extremely low. This exacerbates the eviction problem and makes the disposition of real estate owned by lenders through foreclosure more difficult. The lack of housing turnover also makes the initial appraisal more problematic. Turnover is a more acute problem in rural areas. 2.3. Construction Methods and Costs Both the initial quality of construction and ongoing maintenance and repair are somewhat deficient in Sri Lanka. Most of the new housing stock being added is anticipated to be maintained and repaired by condominium associations. The stock formerly owned by the government and not fully converted remains the responsibility of the government to maintain and repair. Maintenance of government- constructed housing is inadequate. Much of the existing stock has been self-constructed. Few large scale developers construct single-story detached or semidetached dwellings. In fact, there are few professional contractors, because people typically act as their own general contractor. This makes the quality of every house both unique and problematic. Households are generally in no position to evaluate and manage the work of subcontractors the way a general contractor should, nor to evaluate material quality. People tend to stick to traditional construction methods, including concrete, clay, and brick construction, which are labor-intensive and time-consuming. Table 2. Cost Indexes of Housing Construction Year Cost Index 1990 100.0 2000 184.2 2001 204.8 2002 213.9 2003 229.8 2004 254.6 September 2005 303.6 Source: Institute of Construction, Training, and Development 2007. The cost of construction has increased about threefold since 1990, inhibiting growth of construction of houses (table 2). Building materials that registered substantial price increases since 1990 include sand (1070 percent), timber (568 percent), and bricks (678 percent). Labor cost has increased by nearly 250 percent during this period. Current supply problems include shortages of high-quality clay for bricks and sand for concrete production. Innovations in low-income housing construction include, for example, the use of cement instead of clay tiles on roofs. The tsunami relief effort introduced many new types of structures. But none of these cost-saving strategies have taken hold in domestic construction. More work on reducing costs of construction is needed. 13 2.4. Rental Supply and Vacancy Rates There is virtually no rental market for the people who need it most––that is, middle- and low-income renters. Rental units are especially lacking for households that cannot afford to purchase housing or that are not bankable or reached by microfinance lenders. An active rental market is important to facilitate speedy and undisputed evictions, because courts are less reluctant to support lender rights when eviction is less traumatic, requiring only a move into a readily available vacant rental housing unit. 2.5. Precisely Measuring the Gap in Housing Investment An alterative view on housing cautions against overestimating housing needs. National income and output accounts do not capture owner-financed equity investments in owner-occupied housing, nor the returns to this investment over time. In countries like Sri Lanka, where all income groups partially or fully construct their own homes, GDP accounts also miss a large part of the (equity) savings and investment in the housing stock. Many analysts looking at national income and product accounts may falsely believe that emerging market economies may be underinvesting in housing, when the opposite may be occurring, simply because of the failure to measure informal housing construction and owner contributions to formal construction. From the perspective of households in such economies, especially lower-income households, adding on to a dwelling may be the best or only way to save—particularly given the dearth of alternative savings vehicles and investment alternatives, and limited rental housing opportunities. Some policy makers (and analysts) may view the nascent development of the home mortgage market as a means to increase housing investment to levels necessary to close the gap between housing needs and the existing stock. The mortgage market can help close this gap by financing the construction of new replacement housing units. Hence, the government might perceive a need to provide the “missing� housing. In contrast, the market solution, espoused by Sri Lanka’s National Development Strategy, points to upgrades of existing housing, especially for lower-income groups. Taken from the perspective of housing finance, this distinction implies that “housing finance� cannot be equated to “mortgage financing,� especially in developing countries. Although housing finance is discussed, particularly with reference to microfinance lending products and lower-income housing solutions, this report focuses on mortgage financing. 2.6. Demographics Population growth in Sri Lanka is one of the lowest in the developing world, at 1.2 percent in 2005. The population is aging, however, with 69 percent of the population in the 15 to 54 age-group (compared with 55 percent in 1990) and 15 percent of the population in the more than 55 age-group (up from 9 percent in 1990). Urban housing shortages are further adversely affected by expected increases in urbanization trends. In the past decade, urbanization has not accelerated significantly, hovering just above 15 percent since 2000 (World Bank 2007a). In the coming decades, the urban population is expected to grow at around 3 to 4 percent annually. Consequently, the demand for housing in urban areas will account for most of the demand for home mortgages. Although housing density will need to increase because of the rising price of land, the traditional preference for ownership is likely to result mostly in multistory condominium growth. 2.7. Effective Demand for Home Mortgages Table 3 presents the average income per decile of the population (also see figure 5). Affordability of mortgage finance, quite independently of bank willingness to lend, refers to the ability of a household to comfortably and safely cover mortgage repayments. Using a set of typical mortgage terms, at 16 percent, a 14 15-year mortgage, a mortgage monthly payment of SL Rs 14,680 is needed to finance a house valued at SL Rs 1 million. In other words, using industry standard policies, this mortgage requires a monthly income of SL Rs 24,467 (or SL Rs 294,000 annually). Similarly, at 16 percent, a 15-year mortgage for a house valued at SL Rs 500,000 must be financed by an annual income of SL Rs 220,000. At the high end of the market, this means that more than 10 percent of the population finds luxury housing affordable. For comparison, in Bangladesh, mortgage financing is affordable for 5 percent of the population (World Bank 2005b). Table 3. Average Income by Population Decile Population decile Average income 1st (bottom) decile 13,043 2nd decile 29,644 3rd decile 48,616 4th decile 49,801 5th decile 68,773 6th decile 86,559 7th decile 103,160 8th decile 132,803 9th decile 182,605 10th (top) decile 470,740 Source: Department of Census and Statistics 2003. Using an industry standard of mortgage lending up to three times annual income, a low-cost house of SL Rs 175,000 could be afforded by an individual with an annual income of SL Rs 58,000, which leaves the bottom 40 percent of the population out of the affordability range for housing.10 Housing construction at these low costs has not been implemented on a large scale in the country, however. Figure 5 Cumulative Income by Decile 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 Source: World Bank 2007. 15 2.8. Addressing the Housing Needs of Low-Income Groups Until market segmentation is history and rental markets develop, thus making commercial solutions possible, the lowest-income group’s housing needs require artificial support. Donors and government programs in Sri Lanka have addressed the issue. One such example is the North East Housing Reconstruction Project (NEHRP), funded by the International Development Agency (IDA). The project designated $75 million to the reconstruction of 4,900 houses in eight districts (phase I) and identified 13,000 beneficiaries for further civil works in phase II. NEHRP also addresses and supports capacity building, construction issues, and support activities. As another example, the IDA-funded Puttalam Housing Project has designated $26 million for about 5,800 new houses and the completion of 1,200 half- completed and semipermanent houses. This project aims to (1) meet the housing needs of the conflict- affected population in Puttalam District, with the caveat that internally displaced persons in Puttalam would be integrated into their current location of choice; and (2) provide drinking water, sanitation, and internal roads for displaced and selected nondisplaced persons. Box 1 discusses the ongoing Tsunami Emergency Recovery Program in Sri Lanka. Box 1. Tsunami Emergency Recovery Program This relief operation makes US$75 million (of which US$30 million would be a grant) IDA funds available through an Emergency Credit/Grant to support recovery and reconstruction efforts in a wide variety of sectors, including health and education, over a two- to three-year period. A major component of the program is housing (US$40 million), which consists of reconstruction grant assistance, technical assistance, and capacity building. The housing construction operation is intended to finance the repair or reconstruction of the tsunami-damaged houses of the poorest households in the nine tsunami-affected districts and relevant tsunami-affected divisions. It further supports the government’s plan of creating a 100-meter setback in the south and west, and a 200-meter setback in the north and east, to reduce vulnerability to future disasters. Implementing such a setback requires extensive consultation with the communities affected and a complex process to acquire and allocate serviced land and ensure service delivery. The capacity building activities include training of construction workers training of homeowners for autoreconstruction; design and implementation of a communications initiative; establishment of a grievance redress mechanism at the local, divisional, and district levels and land management issues; and continuous social and environmental monitoring. Technical assistance is offered on quality control and auditing; housing damage assessment and social verification survey by the District Task Forces; facilitation of the building materials market and village rehabilitation committees; and support for partner organizations. The district units of the recently approved IDA-financed North East Housing Reconstruction Program (NEHRP) will implement the program in the north and east. District units will be set up in the south and west to implement the program there. Source: World Bank 2007. Existing government programs directed at housing provision for the very poor include those offered by the NHDA, UDA, and the NBRO (see section 1.4). Because these entities currently are not funded from the government budget, they have shifted their focus toward commercially viable income groups, at the relative neglect of the lowest-income population, which the market also fails to reach. Some lower-income housing programs remain operational in the portfolios of these state entities. NHDA offers loans of up to a maximum of SL Rs 50,000 (but more typically of SL Rs 25,000) to enable beneficiaries to upgrade or build their homes. The loans usually are accompanied by some technical assistance provided by a pool of advisers skilled in self-building techniques. The NHDA facilitates the development of the necessary infrastructure around new housing developments, including roads and water. Loans are required to be repaid over a 5- to 15-year period at an average interest rate of 11 percent (ADB 2005). Among existing low-income loans, low recovery has been a major problem. NHDA further provides some direct building material assistance grants, such as roofing sheets. It implements urban redevelopment with their own funds (46,021 housing units provided in 2006) and has planned 100,000 units for the next five years (5,000 for 2007). The Department of National Planning is planning to rehabilitate urban shanties with free water supply and sanitation services. The Plantation Human 16 Development Trust plans to improve 250,000 line rooms for plantation workers as well as make other improvements for the internally displaced population. Real Estate Exchange Ltd. plans to uplift 66,000 shanties in urban areas over the next 10 years. The Sustainable Township Program has been introduced in Colombo, under which slum dwellers have been given the opportunity to voluntarily exchange their dwelling and land space for re-housing in high-rise apartment blocks. Various subsidy programs are in existence, most of which are interest rate based (Chapter 5 reviews the optimal subsidy and cautions against those based on interest rates). The government has introduced a housing loan scheme called the Affordable Housing Finance loan scheme for low- and middle-income households, which offer 3 percent loans. The housing loan scheme operated for EPF members by state banks, provides loans at 4 percent per year. Government employee housing loans benefit from an advance in the amount of five times the employee’s salary up to SL Rs 20 million at 4 percent annual interest rate. Various housing loan schemes are operated by government-owned institutions for their employees as well. These subsidies are expensive––better targeting and flexible subsidy types would lower costs significantly and achieve the same impact. Moreover, subsidies were established when housing markets were underdeveloped, and the need for some of the subsidies might have faded away with financial liberalization and private mortgage growth. Governments frequently make a poor choice of housing subsidy instruments. Subsidies are often in an interest rate form and at fixed rates in an inflationary environment. The state subsidizes the debt rather than the housing itself, robbing households of incentives to repay their loans fast. Subsidies would apply to the life of the loan, without making adjustments for increases in household earnings, which increases the homeowner’s cost without adding any benefit. Competition would be increased by making the subsidy for low-income housing also available to the private lenders, and at the same terms. Finally, subsidies are rarely transparent, making future costs to government hard to predict. In response to destruction caused by civil war and the tsunami disaster, the government has undertaken housing projects to resettle refugees and internally displaced persons, numbering 443,000 from the tsunami disaster alone. For the housing sector in particular, the 2005–07 activity plan for public investment includes initiatives on Housing for Public Servants, Urban Development and Low-Income Construction Housing, Rural Community Development and Estate Infrastructure, and the NEHRP. These efforts fall short of meeting low-income housing needs. When the housing needs of the urban poor remain unmet, they are forced into informal settlements with shelter conditions that violate land development regulations and housing construction codes. Urbanization, economic growth, and in some cases financial globalization have, in many cities, made land acquisition costs for building affordable housing prohibitive for the landless urban poor, including new migrants from poorer rural areas. The 2004 UN report on the implementation of human settlement’s upgrades addresses the policy issue of housing the urban poor and provides lessons learned and policy options (see annex B for a summary of the report’s main findings). Some of these strategies are discussed in the recommendations to this report. 17 Chapter 3 Current and Potential Growth in Mortgage Supply 3.1. Main Providers of Home Mortgages Sri Lanka has 23 licensed commercial banks (LCBs), of which two are state owned, and 14 specialized banks, all state owned. Annex table A4 presents the asset structure of the Sri Lanka banking system. Private domestic LCBs accounted for 45 percent of assets and deposits in 2006 (the respective shares of state-owned banks were 40 percent and 42 percent). The four major private banks (Hatton Bank, Commercial Bank, Seylan Bank, and Sampath Bank) account for 81 percent of the assets, 88 percent of the deposits, and 83 percent of the loans granted (CBSL 2006b). Table 4. Market Share for Commercial Banks, by Ownership (percent of total LCB advances) 1998 2006 (percent) (percent) State owned banks 54.2 37 Private domestic banks 34.5 49 Foreign banks 11.3 14 Source: Central Bank of Sri Lanka 2007. Note: LCB = licensed commercial bank. The two state-owned commercial banks, Bank of Ceylon (BOC) and People’s Bank (PB), have been recapitalized and restructured in the late 1990s, achieving performance improvements.11 Further improvement efforts are needed in operational efficiency and information technology, among other areas. BOC reports comparatively better financial performance. The two banks are incorporated under their own statutes rather than the banking law, reflecting their mandates to meet the policy objectives of providing financial services throughout the country and to target government priority sectors rather than pursue pure commercial objectives. Both banks hold considerable Treasury-paper portfolios. Their market share has declined over the last few years, but the two banks still account for 37 percent of commercial banking assets (see table 4). Table 5. Distribution of LCB Banks and Bank Branches, 2006 Total Number of LCB Branches and Other Outlets 3,516 Branchesa 1,530 Domestic Bank Branches 1,491 Main Branches 1,166 Kachcheri Branches 22 Extension/Pay Offices/Service Counters 290 Overseas Branches 13 Foreign Bank Branches and Other Outlets 39 Branches (including extension offices and subbranches) 31 Other Outlets 8 Pawning Centers 186 Student Savings Units 1,800 Source: Central Bank of Sri Lanka 2007. Note: LCB = licensed commercial bank. a. Includes head offices; excludes pawning centers and student savings units. 18 LCBs continue to advance in terms of growth of assets, expansion in delivery channels, improvement in risk management, risk absorption capacity, and use of information technology. The private commercial banks are expanding their branch networks. The LCBs operate through a network of 1,530 branches and 1,986 other service outlets (see table 5). Supporting this network there are 1,127 automated teller machines (ATMs) and 8,753 electronic fund transfer facilities at the point-of-sale machines, which contribute toward greater accessibility of banking services. Sixteen banks provide full-fledged Internet banking facilities. Commercial bank holdings of government paper are not as high as other financial institutions, standing at about 45 percent of their assets. Collectively, the foreign-owned banks account for 5 percent of commercial bank assets, and 14 percent of advances. Of the 12 foreign banks in Sri Lanka, only HSBC and Standard Chartered, with 10 and 9 branches, respectively, compete in the retail market, focusing on its high end and in the Colombo area. Housing finance is not an active area of operation of foreign banks, which instead focus on trade finance, servicing the local operations of multinationals, and providing correspondent services for domestic banks. No specific barriers prevent foreign banks from entering the Sri Lankan market, although the more broadly applicable restrictions on foreign investment require that any foreign investor first obtain the approval of the Board of Investment (BOI). Banks may lend in foreign currency only to approved BOI companies. Commercial banks cannot by law provide funding using sources external to Sri Lanka. Capital account restrictions limit the ability of foreign investors to repatriate earnings. Since 2000, however, the departure of four banks (including ABN AMRO), and the absence of new foreign bank entrants, is indicative of the difficulties foreign banks face earning their required return on invested capital, rather than evidence of barriers to entry or obstacles to foreign bank operations. The Sri Lankan market is small, particularly for most foreign banks that do not target small and medium enterprises (SMEs) or retail customers. The 14 licensed specialized banks (LSBs), which all compete with the commercial banks for some products, were established to provide specific services, often to address perceived market failures. NSB is the largest LSB with 68 percent of total LSB assets, followed by Development Finance Credit Corporation (DFCC) Bank with 15 percent, Regional Development Banks with 7 percent, and SMIB with 3 percent. NSB and DFCC account for 31 percent of total LSB lending each, whereas Regional Development Banks and SMIB handle 15 percent and 7 percent of lending, respectively (CBSL 2007). Penetration and accessibility, including regionally, is good (see table 6). Table 6. Distribution of LSB Banks and Bank Branches, 2006 Total Number of LSB Branches and Other Outlets 415 Branches 376 Regional Development Banks 201 National Savings Bank 114 Long-term Lending Institutions 11 Housing Finance Institutions 28 Private Savings and Development Banks 22 Other Outlets 39 Source: Central Bank of Sri Lanka 2007. Note: LCB = licensed commercial bank. NSB, which is the second-largest deposit-taking institution in the country,12 invests about 80 percent of its assets in government debt. By law, 60 percent of NSB’s assets must be invested in Treasury securities. Traditionally, operating as a narrow bank has helped NSB to keep operating costs and credit risk low, as its systems for credit risk management and credit administration are weak. NSB has undertaken limited computerization, coupled with plans for ATM and electronic transfer capabilities. The bank has begun to develop a housing finance business for its customer base; in particular, NCB aims to assist savers who want to expand their homes as dwellings for the wider family or as home-based small businesses. The specialized banks with the major share of mortgage lending are SMIB13 and HDFC.14 SMIB, the “Housing Bank to the Nation,� is licensed as a specialized bank, is state owned, and has an SL A+ national rating from Fitch Ratings, the nation’s main rating agency. HDFC engages in housing 19 construction and house financing activities; SMIB is more involved in the provision of grants and subsidized loans to lower-income groups. Commercial banks and NSB are rapidly becoming active consumer lenders, including for home loans. A well-developed albeit small network of microfinance lenders (amounting to 0.1 percent of financial institution assets) currently fund housing repair and upgrades, but given their small current level of outreach, these lenders have limited potential in the medium term to expand into home loans for the poor. Hence, there is some competition for borrowers of all but the lowest income levels. Table 7 presents the main financial institutions engaged in housing mortgage lending in the country, by type and importance to housing finance. Table 7. Total Outstanding Housing Mortgage Loans in the Financial System, 2005 Item Total mortgage Total loans As a As a percent of loans 2005 (advances) percent of total assets of (SL Rs total loans the financial millions) system Licensed Commercial Banks 76,223 647,900 11.8% 2.42% National Savings Bank 11,120 25,578 43.5% 0.35% State Mortgage and Investment Bank 8,406 8,475 99.2% 0.27% Housing Development Finance Corp. 7,858 7,900 99.5% 0.25% Bank Regional Development Banks 3,483 16,587 21% 0.11% Registered Finance Companies 7.72 62.2 12.4% Insignificant. Sources: Central Bank of Sri Lanka 2006b; individual financial institution annual reports 2006. Note: SL Rs = Sri Lankan rupees. Three small private savings and development banks collectively account for about 3 percent of the assets of the LSBs. These include (1) Sanasa, established in 1997 by thrift and cooperative societies; (2) Sanasa’s district unions and the Federation of Thrifts and Cooperatives; and (3) Ceylinco, part of the Ceylinco group thus affiliated with Seylan Bank. The Ceylinco group includes the largest building society, Ceylinco Credit and Investments, also active in the housing sector. Table 8. Mortgage Lending by the Sri Lanka Banking Systema State (Specialized) Bank Mortgage Commercial Bank Mortgage Lending Lending Outstanding As a percent mortgage of total NSB SMIB HDFC loansb As a financial (SL Rs (SL Rs (SL Rs (SL Rs percent of system assets millions) millions) millions) Year millions) total loans 2000 41,281.4 13% -- 3,874 7,135 2,894 2001 45,767.3 14% -- 4,879 7,480 3,591 2002 52,804.1 14% -- 5,872 7,757 4,750 2003 59,983.5 14% 3% 7,168 7,931 5,826 2004 74,409.0 14% 3% 9,271 8,079 7,118 2005 94,444.4 14% 3% 11,179 8,475 8,100 2006 133,640.4 16% 4% 16,098 -- -- Source: Central Bank of Sri Lanka 2007. Note: HDFC = Housing Development Finance Corporation; NSB = National Savings Bank; SL Rs = Sri Lankan rupees; SMIB = State Mortgage and Investment Bank. a. Data are not clearly separated between property loans for houses and other real estate, and between construction and permanent (that is, home mortgage) financing. The table indicates the percentage of loans and of total assets invested in mortgage loans. Finance companies and NDB loans are not included. b. Includes the two state banks, BOC and PB, because they are categorized under LCBs. 20 The mortgage lending market, which grew at 9 to 14 percent in the early 2000s, has been swelling by leaps and bounds in the past three years, at rates hovering around 30 percent (see table 8). In part, housing prices have been increasing, fueling construction and speculative or investment housing purchases. The balance of such high growth results from the government’s liberalization policies and a price adjustment in transition to market equilibrium. Private commercial banks currently account for more than two-thirds of mortgage lending, NBS for another 9 percent, and SMIB and HDFC for 7 percent and 6 percent, respectively (table 8). 3.2. Current Sources of Funds for Mortgage Lending The amount of funds available for home mortgage finance ultimately reflects the overall savings rate in the economy, the relative demand of other borrowers, and the institutional arrangements available for homebuyers to access sources of savings. The home mortgage market is currently young and small in Sri Lanka, so modest growth will reflect the demand for credit and the legal and regulatory environment for lending. Whether the market can expand more rapidly will depend on the overall macroeconomic environment and the relative availability of funds. Deposits are one source of funding for mortgage lending. Commercial bank deposits make up about half of the assets of the financial system and about three-quarters of the investable funds in the economy. Specialized banks account for about 10 percent of the financial savings of Sri Lanka. NSB is the largest, with a significant share of household deposits. The two state-owned specialized mortgage lenders only recently were authorized to take in household deposits. Mortgage lending is not reported to be limited by the supply of funds at the private lenders or NSB. The two state-owned specialized mortgage lenders report that funds for housing lending are in short supply, because they do not have a large stable base of core deposits, and secondary market links have not yet been established. Microfinance lending is financed out of donor funds or through creative membership schemes. Annex table A6, column 6, presents the sources of funds for selected mortgage lenders. Secondary markets for securities could provide additional mortgage funding. Secondary financing for mortgages has been considered actively by policy makers, and some transactions were attempted. The state-sponsored HDFC’s SL Rs 722.2 million securitized paper is seen as the deal closest to securitization carried out so far. Although innovation and progress are heartening, more careful examination of the transaction reveals that the mortgage-backed securities (MBS) placement did not constitute a true sale. Cash collateral of 12.5 percent on pool balance and subordinated papers of SL Rs 99.8 million transformed the deal into a de facto two-tranche senior-subordinated security, which appropriately would be classified as debt financing. Nonetheless, the instruments were required to be rated, underwritten, and trusteed. This process is costly, complex, and time-consuming, and largely unnecessary for the subsequent Figure 6 private placement to a single, also state-owned Asset Share of Financial Companies, by Type owned, buyer. The HDFC deal could be justified as an effort to kick-start an active secondary capital market. As this report argues in Chapter 5, secondary private placements possibly offer most of the features that the mortgage market will need as it grows over the next decade. Policy makers should consider the need for additional hefty regulatory overhead, which would make securitization possible, on its merit as sized up against the risk of overregulation stifling the nascent market. In either case, strong primary mortgage markets are a necessary precondition for secondary transactions, and the latter would not take off without active and scaled- up mortgage lending. Source: Central Bank of Sri Lanka 2007. 21 3.3. Growth Potential of Mortgage Lending Funds Home mortgages are funded by savings that flows into the formal financial sector. Hence, the way to assess the potential for this market is to determine the total sources of funds to the financial sector and identify the competing demands to use these funds. CBSL does not publish a complete Flow of Funds account database, making projections of funds potentially available to housing difficult. A few relevant observations can be made from the data compiled with the help of the CBSL. Total assets of the financial system were SL Rs 3,144.7 billion in 2006, or about 133 percent of GDP (excluding CBSL assets). About 57 percent of this total was held by LCBs and 7.7 percent with the NSB (see figure 6). The two specialized state-owned mortgage lenders each accounted for less than a tenth of one percent. The two public provident funds, EPF and ETF, accounted for 18 percent. Private provident funds and insurance companies each accounted for about 4 percent. The average stock of outstanding mortgages in developed countries is about 20 to 50 percent of GDP. As the financial system in Sri Lanka grows to attain such size, the most promising avenues for housing finance investment, at first blush, could be the banking and provident fund sectors. Only about 55 percent of the funds at the LCB are invested in the private sector, however. The NSB has about 80 percent invested in government securities, and the EPF and ETF combined have more than 90 percent, as do the private provident funds and life insurance companies.15 Government deficit financing has historically left little for domestic corporate or household financing. Future plans for public infrastructure investment might exacerbate these figures. On the positive side, the National Government Strategy includes plans to reduce the budget deficit. The long run offers a more diversified group of sources of funding for housing finance, but in the short term, it seems that the commercial and specialized bank deposits would be the only sources of mortgage finance. Households will be competing against business as well as government and state-owned enterprises (SOEs) for bank demand and savings deposits. A recent CBSL survey indicates that 16 percent of commercial bank lending was for housing and property loans, an increase of 18 percent over the preceding year (CBSL 2006). Lending by foreign banks is limited by regulation to funds raised domestically, which is a negligible amount. Total mortgage lending by banks is approximately 6 percent of total bank assets. Domestic deposit growth is about 15 percent annually. Under these assumptions, the mortgage financing funds from private banks will increase seven times by 2020 (see figure 7), keeping the average mortgage size constant. A back-of-the-envelope calculation shows that this increase would imply an expansion of the number of mortgages from the current range of 0.6–0.9 million to about 1.1 million, assuming that housing prices increase with inflation at 12 percent (using the actual 18 percent growth rate implies a decrease in the number of mortgages). With population growth at 1 percent, 1.1 million mortgages nationwide translate into one loan per five households (as compared with the current ratio of one loan per seven households). So the question arises as to where the funds will come from to finance a significant expansion in home mortgage lending. Figure 7 Projected Housing Finance Funds Growth, Banks Projected growth of housing finance loans (Rs billion) 1200 1000 800 600 400 200 0 Source: Authors’ calculations. 22 At the upper end of the market, the potential for the commercial banks and NSB to expand consumer finance over the next 10 years is sufficient to meet the effective demand for mortgage credit. The banking system is expanding rapidly, as new participants enter the market and existing players branch out at a rapid pace. The number of households that have a banking relationship can be expected to increase as incomes rise, as well. At the lower end of the market, the existing microfinance lending channels will continue to fund housing upgrades, but total deposits of the entire industry are only about a tenth of one percent of financial institution assets, so this potential is limited in the medium run. These efforts could be complemented, at the lower end of the market, by the two state-owned specialized mortgage lenders; currently, however, these lenders can lend in a sustainable manner only to middle-income households. Efficiency improvements, new lending technologies, transparent and well-targeted subsidy policies, and the discipline of self-finding might allow these institutions to operate independently and to service low- income housing finance needs. Finally, the major expansion in the housing finance market is typically expected to come from middle- income demand. Can middle-income demand be funded in the medium run? Commercial banks might adapt their lending policies to these income groups; but with weak eviction powers and borrowers lacking regular or salaried income, banks may not move aggressively to meet this demand. Bank lending to middle-income groups could be boosted by further regulatory improvements, including the use of mechanisms such as liquidity facilities. Secondary mortgage markets significantly expand the possibilities for mortgage finance supply and can offer the needed long-term funds, provided that primary markets function well. About two-thirds of household financial assets are in the form of contributions into the two employee public provident funds, the EPF and ETF (absorbing 23 percent of income, including employer contributions).16 These funds operate similarly to savings accounts, because households can withdraw up to 75 percent of their balance, for example, for a down-payment on a housing loan. Nearly 19,000 members took advantage of this facility in 2003, often using it as a device to take out funds early and then default. (Members in the Northern Province claim that this facility was not open to them). The EPF does not require evidence that the withdrawals are used for housing facilities or other approved purposes, such as health care and education, nor are members required to replenish withdrawn funds over fixed agreed- on periods. Although EPF and ETF might not be expected to operate mortgage lending efficiency, their investment strategy could be liberalized, allowing the current and past contributions to be reallocated to housing-related investments, via existing mortgage lenders. For this reallocation to have a significant impact, a considerable divestment of government securities would be involved. Liberalizing EPF and ETF investment policies might not be a straightforward solution, however. The EPF and ETF, while technically funded plans almost entirely invested in government securities, might warrant more careful examination on the liquidity of their assets and their actual ability to cash out (some of) these assets. In other words, reallocating the existing portfolio to alternative investments requires the government to refinance the debt, and to do so in a financial system already saturated with Treasury bills. Furthermore, the projected contributions to the provident funds are expected to decline because of the aging population.17 The funds are still in surplus, and the government claim on provident fund collections may well decline faster than the net provident fund surplus. Unless these schemes are carefully reviewed and firmly funded, however, they are not likely to live up to expectations as significant sources of long- term funding for housing in the future. Life insurance companies were recently reprivatized. In 2003, the total combined gross premium written for long-term insurance and general insurance was SL Rs 24,147 million, an annual increase of 19 percent, representing 1.37 percent of GDP (CBSL 2007). In China, for comparison purposes, the total combined gross premium represented 3.4 percent of GDP (an increase of 30 percent from 2002). Life insurance accounts for 44 percent of total premia (nonlife insurance accounts for 56 percent). Total assets represent SL Rs 69 billion or 2.8 percent of GDP. By regulation, 20 percent of the technical reserves of nonlife insurance and 30 percent of the long-term life insurance fund are invested in government securities. In practice, the figures are much higher—73 percent of life insurance assets and 34 percent of nonlife assets were held in government securities in 2003.18 The life and nonlife market is dominated by the same three players: Sri Lanka Insurance Corporation, Ceylinco Insurance Company, and Eagle Insurance Company. Life insurance reserves are growing, and 23 the limitations on investment strategy are expected to be liberalized. The draft rule would allow up to 10 percent of assets to be held in mortgages and mortgage-related assets. But life insurance reserves make up only about a third of insurance reserves, and 1 percent of the total financial assets of the financial system, so the potential increase in mortgage investments will have only a modest impact. Approved Private Provident Funds (APPFs), of which there are some 200, cover private sector employees that originally contracted out of the EPF. The APPFs have 150,000 account holders with assets totaling SL Rs 200 million. The investment policies of provident funds are not subject to government direction, but these funds heavily favor fixed income investments. Moreover, APPFs have little fiduciary incentive to change, because most of the funds are held by state-owned companies. If they did have such incentives, these funds could have a greater impact than life insurers on the availability of capital market funds available for mortgages.19 The market for investment funds amounts to only SL Rs 3.8 billion, but recent surges in the activity of unit trusts promise considerable future growth potential. Capital markets are another potential long-term source of funds; however, the Colombo stock exchange currently has a market capitalization of SL Rs 850 billion. The Sri Lanka bond market is relatively large at 54 percent of GDP, but it is composed mostly of government securities, and there is virtually no corporate bond market to speak of (listed debentures amount to 0.6 percent of GDP, World Bank 2008). Another major source of long-term funding could be foreign investment. Funds would likely take the form of foreign direct investment (FDI), because foreign investors generally are not interested in long- term fixed rate debentures denominated in Sri Lankan rupees, especially given the possible inflation risk. Sri Lanka has been open to FDI, but the amounts have been disappointingly small thus far, most likely reflecting the internal security and infrastructure problems. FDI is expected to more than double, to around 2 percent of GDP, in the medium term. Mortgage securitization provides significant benefits in markets in which mortgages are homogeneous and market participants are multiple. Securitization can, for example, mitigate liquidity risk, free up lender balance sheets, and spur the development of a series of specialized market intermediaries. In Sri Lanka, these benefits might not be easily realized in the medium term. Issuance of MBS presents problems as well. Parate rights (that is, the ability of a lender to foreclose and sell a defaulted property without going to court) don't transfer to investors or special purpose vehicles (SPVs).20 A stamp tax is due on home sales, making these transactions even more expensive. The securitization act, which is judged to be of fairly high quality, is pending; however, it is likely to generate little market response. Underwriting, rating, trustee services, and the like remain expensive. But more important, in the next 10 years, investors with long-term funds will be severely lacking. The underdeveloped bond market and the small size of the stock market might present further problems with the development of a liquid (actively traded) MBS market. 3.4. Banking and Home Mortgage Sector Efficiency The banking system in Sri Lanka is liquid, with private sector credit expanding since 2004, although at a lower rate relative to the mid-1990s. As bank deposits increase more rapidly than loans, the surplus liquidity is absorbed primarily by government securities. In addition to crowding-out concerns, these trends suggest a banking sector that is reluctant to expand mortgage lending to a wider middle-income group or take private sector risks. This reluctance perhaps is due to a perceived lack of bankable opportunities and high entry costs into new market segment or to a lack of sufficient credit information and adequate credit-scoring mechanisms to manage risk effectively. The Asian Development Bank (ADB) Financial Sector Assessment notes that high spreads are due to rigid and inflated lending rates, not sticky deposit rates. This means banks have high operating costs and high levels of nonperforming loans (NPLs), which push up lending rates. Private banks predictably have lower NPLs than their state-owned counterparts; however, both measures stand in stark contrast to the much smaller bad debt amounts held by foreign banks (table 9). For 2006, however, CBSL reported significantly lower NPL figures for the commercial banks, no doubt reflecting bank restructuring efforts. Licensed specialized banks carry a slightly higher level of NPLs at 8.7 percent. The averages belie the situation of individual lenders—for example, SMIB continues to record a high level of NPLs (25 percent). Table 9. Nonperforming Loans of the Banking System 24 Nonperforming loans/total advances 1998 2003 2006 (percent) Commercial banks n.a. n.a. 5.4 State-owned banks 19.5 17.9 n.a. Private domestic banks 13.4 12.0 n.a. Foreign banks 12.7 6.9 n.a. Specialized banks n.a. n.a. 8.7 Source: Central Bank of Sri Lanka 2007. Note: n.a. = not available. Table 10 provides comparative operating efficiency parameters for commercial banks. Legal and regulatory obstacles and delays contribute to bank costs, which are passed on to the consumer because of limited competition.21 Furthermore, the segmentation of the borrowing market sustains higher rates. Table 10. Efficiency of State-Owned and Large Private Banks Bank of People’s Bank Average of Ceylon four large private banks Noninterest expense/total income ( percent) 61.9 68.7 52.4 Average branch size (assets, SL Rs millions) 827.8 645.0 899.5 Assets/employee (SL Rs millions) 27.4 18.1 33.2 Personnel cost ( percent of average assets) 2.1 3.2 1.7 Source: Bank annual reports 2006. Note: SL Rs = Sri Lankan rupees. State housing banks have poor performance the world over. They are not disciplined by the market, so operating costs are usually high and credit losses are usually significant (table 10). Banks are reluctant to enforce liens for social or political reasons, therefore borrowers often view loans from housing banks as grants. The allocation of subsidized credit is often regressive and politicized, and it may crowd out other lenders. Sri Lanka is no exception, as the tables above indicate. State housing banks have higher NPL ratios, operating costs, and personnel costs, and smaller assets per employee and branch sizes (as measured by assets per branch). The two state specialized lenders leave substantial room for efficiency improvement (some success in that area has been achieved by other state banks such as BOC and PB) (see section 3.1). One of the few successful examples of an efficiently operated housing bank is the Housing Bank of Thailand (Calhoun 2005). That said, CBSL financial sector reports generally indicate operating efficiency that is admirable for a country at this stage of development (although this efficiency remains below levels found in member countries of the Organisation for Economic Co-operation and Development [OECD]). Within the group of four relatively large private domestic banks, there are distinct differences in ownership and financial performance. Hatton National Bank and Seylan have generally inferior financial performance than the more widely held Sampath and Commercial Bank, with higher proportions of NPLs. BOC has been consistently the most profitable domestic bank, in large part because of its much lower noninterest expenses. Sampath Bank, despite having average branch size and business volume per employee similar to BOC, has not achieved the same levels of efficiency, and thus has achieved profitability only marginally better than that reported by Hatton National Bank and Seylan. 3.5. Uses of Funds for Mortgage Loans Home mortgage lending has been increasing at commercial banks. But it still only accounts for about 4 percent of total assets in the banking system, which is low by comparison with other countries. Historically, lending to the government and SOEs has been more than 10 times the amount lent to households for housing. 25 Commercial banks have been shifting toward an increased share of housing finance in their portfolios (table 11). The bulk of commercial bank housing lending is for personal housing construction and purchases (11.3 percent of total commercial bank loans). Lending for construction of business premises or property development accounts for a much smaller (though growing) share of overall loans, standing at 2.3 percent and 2.1 percent, respectively, for 2006. Table 11. Commercial Banks Advances for Housing and Property Development (SL Rs millions) Loans by LCB for Loans by LCB for Loans by LCB for property personal housing construction of business development including purchases premises (SL Rs million/percent of (SL Rs million/percent (SL Rs millions/percent of total loans) of total loans) total loans) 2000 32,649.2 10.3% 3,595.3 1.1% 5,036.9 1.6% 2001 36,788.5 11.3% 3,059.0 0.9% 5,919.8 1.8% 2002 39,934.4 10.6% 3,598.6 1.0% 9,271.1 2.5% 2003 45,672.9 10.7% 4,694.1 1.1% 9,616.5 2.3% 2004 59,106.8 11.3% 5,384.5 1.0% 9,917.7 1.9% 2005 76,222.9 11.6% 6,879.1 1.0% 11,342.4 1.7% 2006 86,122.1 11.3% 17,264.8 2.3% 15,717.7 2.1% Source: Central Bank of Sri Lanka 2007. Note: LCB = licensed commercial bank. 3.6. Mortgage Products and Loan Terms Housing loans generally have fixed rates set by the state lenders. These loans are variable at the LCBs, secured with a first mortgage on the house, and supported by mortgage protection policies against death or disability of the borrower. Policies can be obtained to protect against title deficiencies. Rehabilitation and upgrading can be financed only with personal loans, because second mortgages do not yet exist (but are allowed). The state lenders historically have offered only fixed rate mortgage loans, whereas private banks are experimenting with adjustable rate loans, as fixed rates have almost doubled in the past few years (see annex table A6, column 7). The state banks use an escape clause to allow for rate adjustments, which might be politically difficult to utilize. Because of these potential rate adjustments, fixed rate lenders are faced with interest and funding risks. Figure 8 Selected Interest Rates, 2002–06 25 20 LCB secured lending SMIB lending rate 15 DFCC Bank lending rate NSB lending rate 10 NHDA lending rate NSB deposit 24mo fixed 5 LCB deposit 24mo fixed 0 2002 2003 2004 2005 2006 Source: Central Bank of Sri Lanka 2007. 26 Maturities range from 15 to 25 years (or more), with prepayment options available to the borrower (see annex table A6, column 8). Microlending terms are between 2–4 and 10 years. Bank borrowing, on the margin, is all short term. Private lenders set the mortgage rate to reflect the cost of funds plus the cost of operations. State-owned lenders, which were recently reorganized to lend from their own resources, are also supposed to include these costs in their mortgage rates, although this is not achieved in practice. Instead, mortgage offer rates are set to match the rate offered by the competition. State banks remained profitable last year, but these profits declined from prior levels. NSB offers a fixed rate for 20 years, sometimes 30, currently set at 14.5 to 15.5 percent, reflecting their cheaper deposits. To be competitive, SMIB and HDFC offer a fixed rate of 16 percent. This rate is only 200 basis points above the rates offered for the longest maturity government paper. Mortgage rates are predictably below other consumer lending rates (table 12). Table 12. Comparison of Mortgage Rates to Other Consumer Lending Rates Institution Overdraft Credit cards Auto leasing Pawning Mortgages National Savings Not offered Not offered Not offered 18% 14.5–15.5% Bank Bank of Ceylon 20–32 % Effective 34% 19–21% 20% 15.5–15.74%a People’s Bank 17–19% Effective 35% Effective 24% 18% 17% Hatton National 20–25% Effective 35% 18–20% 18–20% 17.75–18.75% Bank Commercial Bank 25–32% Effective 35% 19–21% Not offered Fixed 18% Floating 14–20% Source: Individual bank annual reports and proprietary data. a. The rate is 0.756 percent to 1 percent over the Weighted Prime Lending Rate (WPLR). We take the 2006 WPLR or 14.74 percent. A well-established yield curve for government bonds facilitates financial market development and helps price various financial instruments, including mortgages and MBS. Although treasuries do trade, a robust yield curve is not yet established in the country (figure 9). Figure 9 Implied Yield curve, Treasury Bond Monthly Data, 2006 Yield curve 12 interest rate 11 10 9 8 3 6 12 24 36 48 60 72 maturity (months) Source: Central Bank of Sri Lanka 2006. The maximum payment-to-income (PTI) ratio appears to be high by international standards. Most lenders will lend up to a maximum of 40 percent of the debt service-to-income (DTI) ratio for low- to middle- income borrowers, and up to 60 percent for higher-income borrowers (up to their maximum loan limit). The maximum loan-to-value ratio is somewhat conservative at 75 percent for housing (50 to 60 percent for land), because it is based on “recovery� rather than market value (recovery is estimated to be about 10 percent lower than the market rate). Households do not need to establish a prior pattern of saving with the lender. There is no mortgage insurance for lower down-payment loans. Details can be found in annex table A6, columns 4 and 5. The average loan size is SL Rs 1 million at private domestic commercial banks and NSB. The state housing corporations do not target clients much further down the income scale. SMIB’s average loan is 27 SL Rs 0.6 million, and HDFC’s average loan is slightly lower at SL Rs 0.25–0.5 million. Microfinance lenders offer much smaller loans to SMEs of SL Rs 15,000–100,000, although they also originate loans of SL Rs 0.5 million (annex table A6, column 3). Some commercial banks are imposing (or considering) prepayment penalties on mortgage loans. These penalties typically take the shape of a percentage of the remaining loan value, usually between 0.5 and 2 percent. Overall, loan terms could be characterized as mildly liberal in the context of the Sri Lanka foreclosure and eviction environment and the absence of mortgage insurance. 3.7. Residential Development—the Multifamily Market Because most of the demand for new housing results from urbanization, and poor transportation infrastructure discourages suburbanization, rising land prices in urban areas will dictate the construction of multifamily structures. Most of the existing stock of multifamily urban houses in Sri Lanka is in condominiums. The government converted most of its stock of housing units to condominiums in 1998. Conversion was voluntary and about 60 percent of the existing tenants converted. All of the new apartments being built at the high end of the market are condominiums, with high investor ownership, and a considerable number rented. Some of the new subsidized condominium apartments allocated to middle-income buyers, including government employees, are rented as well. The rental market is developing, albeit slowly and only for high-income renters. The developer industry is active in Sri Lanka, though not highly competitive. Developers frequently work with government housing institutions on lower-income housing, and build on their own for the high-end market. Regulation is not overly onerous, although some issues need further improvement. It takes 18 months to get a building permit and another 7 to 8 months to get a condominium permit. Side payments are not uncommon. High-end condominiums typically are financed by future buyers in installments made ahead of construction (there is no escrowing) or by bank borrowing (lenders including major banks as well as DFCC). Private commercial banks might not be overly eager to finance development because of the complications to obtaining a clean title to the land to be developed. Condominium prices increase so fast that they sell for cash in installments during construction. Thus, lender exposure is limited and development loans are paid off before completion. Buyers cannot borrow against future title, because the title is held by the developer until completion of the units. Even then, delays with title are not uncommon (delays up to one to two years can occur). As a result of these delays, the market is constrained to buyers who can pay cash from savings or obtain personal loans until the title can be assigned on completion. Condominiums are frequently resold for profit upon completion. The new legislation regulating condominium construction has some shortcomings.22 Development of high-rise buildings is restricted. Professional property management companies have not yet been established in the country, and owners are not keen to act as property managers. Developers are forming professional property management companies for their own projects, but condominium associations— which by law are made up of the owners—have not yet had the ability to subcontract for professional management. This subcontracting may occur if, for example, a developer spins off a property management company, which then seeks to expand by bidding for work at independent condominium associations. 3.8. Lending by Microfinance and Informal Institutions As Sri Lanka transitions from government-directed to a private sector–based housing provision and financing, it is important to include the middle- and lower-income groups. Commercial solutions in developing countries frequently cannot meet the housing needs of the lowest income groups, such as regions struck by conflict or disaster or areas of urban poverty. In time, as the system of housing finance develops, the mortgage needs of an ever-widening share of income groups can successfully be addressed via market solutions. Initially, the gap between housing finance provision and effective demand is wide enough to include middle- and low-income households. As banks respond to incentives inducing them to 28 go down-market, they bridge the gap between upper-income groups and riskier middle- and low-income households. Microcredit operates from the bottom up, in turn servicing groups that the formal banking system regards as nonbankable. Finally, international experience points to the development of a low-cost rental market as a housing solution for very low-income groups for whom mortgage financing is not an option. In Sri Lanka, however, the mortgage market remains segmented by income groups, and middle-income households that would be viable mortgage borrowers under more advanced lending techniques do not have access to housing finance. Low-cost rental markets have not developed, in part because of strong tenant protections against eviction, and in part because of a culture of homeownership. This lack of rentals hurts the lowest-income households, especially in large cities. The relative dearth of housing finance options for middle- and low-income households means that these households often build their own housing over time, creating a form of equity finance. Lower-income groups have housing that might not meet common standards. Many live in shanties (that is, improvised houses) with no public services such as sanitation. Lacking title and with disputable records, these income groups are not good candidates for formal mortgage finance. Yet as long as households have an income, they can obtain some form of commercially viable financing. Microfinance and informal lending thus step in to fill this gap. Informal sources of finance include friends and relatives, moneylenders, shopkeepers, and Chitos (collective schemes among individuals). More formal microfinance institutions have succeeded in profitably lending, including for housing purposes, without relying on the title to collateral. Frequently, loans to micro, small, and medium enterprises (MSMEs) (or the profits made thereof) indirectly fund home upgrades, repair, and maintenance, because many of these businesses are run out of the home. Microlending remains confined to a modest scale, however, and therefore it is unlikely to provide a major part of housing finance in the medium or even the long term. Microfinance lenders are numerous in Sri Lanka, with varying degree of commercial success. Some of the more successful examples include Sanasa Development Bank, established in 1998, which caters to the upper end of the microcredit sector, as well as SEEDS (Savordaya Enterprise Development Society) and the National Development Trust Fund, which have operated since 1988 and 1991, respectively. SEEDS uses an interesting business model for microlending. It has 170,000 active borrowers and has expanded 20 percent each year for the last four years. It is currently working with Housing for Humanity on further growth of operations. SEEDS raises money and lends through 3,000 societies with 500,000 members. These members pick borrowers and service the loans, under SEEDS supervision. SEEDS has a nonprofit charter under the Companies Act, which must retain earnings as a donor condition, but it cannot raise equity capital or take deposits, thus constraining it from obtaining creative forms of financing. Borrowers are members and must provide two cosigners and maintain a compulsory savings account at about 10 percent of the loan balance. The lending terms, which have ensured a profitable operation, are 18 percent for two to four years, and 92 percent of the loans are current (performing). Three types of loans are offered: (1) small loans of SL Rs 15,000–50,000 (on average SL Rs 20,000); (2) enterprise loans of SL Rs 50,000–100,000 for home improvement; and (3) larger loans of up to SL Rs 500,000, which are made to existing clients for house purchases. All loans include a capacity building service, such as business plans, assistance in locating suppliers, and so on. To further improve efficiency, SEEDS has computerized two of its six offices. A major development in the microfinance sector was the progress made in formulating the proposed Micro Finance Institutions Act by the CBSL. Taking into account the concerns of the stakeholders, the draft act underwent several revisions and is expected to be finalized in 2007. Once enacted, this law will establish a sound regulatory regime for microfinance firms. The resultant greater transparency and accountability of the sector inevitably will create more development partner confidence in the microfinance sector and promote the flow of funds to that sector. An expansion of microlending through the traditional societies is the best way to expand the access and volume of lending for housing to the poor, particularly in rural areas. Availability of funds to microfinance lenders is not reported to be a constraint on home lending. The lending through microfinance entities likely will remain segmented from formal home mortgage finance for the foreseeable future. Yet, first attempts for bank downsizing exist in the country. Hatton National bank, the largest private commercial bank, has 120 microfinance offices for small business loans in addition to its retail bank branches. 29 Chapter 4 Constraints on the Development of the Home Mortgage Market The potential constraints on the development of a market for home mortgages in Sri Lanka may be divided into three categories. The first relates to the legal and regulatory environment. The second relates to risk management—that is, the ability of financial intermediaries to manage the risks of mortgage lending, and the ability of the CBSL to regulate these intermediaries. The third relates to the total supply of financial savings and the institutional arrangements to access these savings. Who this nascent home mortgage market serves and how well it serves them will reflect government policies in these three areas, as well as the state’s ability to provide an enabling environment and a limited presence. The savings aspect is discussed in Chapter 1, this Chapter inquires into the former two facets of the housing finance market. Legal and Regulatory Framework 4.1. Land Registration and Titling A registered title is unambiguous evidence of ownership that is not challengeable in the courts. Sri Lanka is in the pilot stage of implementing a title registration system and a cadastre with World Bank assistance, in a few selected jurisdictions, to curtail boundary and ownership disputes. The new cadastral and title system is estimated by the Land Registry Office to take at least 15 more years to fully implement. Sri Lanka has 1,500 professional land surveyors, 55 percent of which are committed to cadastral matters, as well as 2,600 lawyers committed to cadastral matters (Cadastral Template, Department of Geomatics of the University of Melbourne) http://www.geo21.ch/cadastraltemplate/index.htm). The Registration of Title Act 1998 as amended provides a sufficient legal framework but is yet to be implemented. Rather, ownership is still evidenced by deed registrations. The record is often ambiguous and incomplete. Disputes, especially regarding boundaries, are common and costly, taking an average of more than 10 years for courts to resolve. Sri Lanka had 8.5 million land parcels as of 2003 (the last year for which comparative data are available), or 442 per 1,000 population, which is relatively high compared with the region. In urban areas, 90 percent of parcels are legally registered and surveyed, 5 percent are legally occupied but not registered or surveyed, and the balance are informally occupied without any legal title. In rural areas, only 63 percent of parcels are fully legal, 25 percent are not registered, and 12 percent are occupied by squatters. Those figures are on the low side compared with other countries in Asia (table 13). Table 13. Land Parcels and Registration Status, Urban and Rural Areas, 2003 Land parcels/ Registration in Registration in Land parcels 1,000 urban areas rural areas Country (millions) population (1) (2) (3) (1) (2) (3) China 246.5 205.4 n.a. n.a. n.a. n.a. n.a. n.a. Hong Kong (China) 0.3 44.8 99 0.1 0.9 100 0 0 India 210.0 205.5 99 1 0 99 1 0 Indonesia 84.5 361.1 40 50 10 20 20 60 Korea, Rep. of 35.8 756.5 100 0 0 100 0 0 Malaysia 7.2 288.9 98 0 2 99 0 1 Nepal 24.0 1,025.6 95 2 3 97 2 1 Philippines 50.0 714.3 70 5 25 75 20 5 Sri Lanka 8.5 441.8 90 5 5 63 25 12 Source: “Cadastral Template,� Department of Geomatics of the University of Melbourne, 2007. Note: n.a. = not available. 30 (1) parcels that are legally registered and surveyed; (2) parcels that are legally occupied but not registered or surveyed; (3) parcels that are informally occupied without legal title (including illegal occupation or squatting). Property registration is voluntary and involves an evaluation, payment of the stamp tax, and notarization. The registrations is then filed in paper form by notary number. Failing to register leaves a deed unrecorded and without courtroom standing in the event of a dispute. The 42 branch offices of the registrar provide records and information, but they do not monitor and correct against inconsistent records or multiple claims. Inefficiencies in the registries for land and moveable property are evident in the average of 63 days and eight procedures required to register property in Sri Lanka, versus a regional average of 119 days and six procedures, and an industrial country average of 32 days and five procedures (Doing Business Report 2007). The lack of computerized files makes searching difficult. There is no cadastral survey in most jurisdictions, which complicates and perpetrates boundary disputes. Lenders currently require a 30- year record search to prove ownership, but often must go back 80 years or more. Developers encounter title problems as they apply for a deed on the development because of unreliable land records. In part, the comparative advantage of NHDA and UDA when they partner with private developers stems from the state institutions’ ability to convey to developers and subsequent buyers clear land ownership records that are acceptable to lenders and to remove all other claims. Land and titling issues are more problematic in rural areas relative to urban areas. 4.2. Foreclosure and Eviction Debt Recovery (Special Provisions) Act No. 2 of 1990 gave power to the LCBs (later expended to all companies registered under the Finance Companies Act) to adopt a special procedure to institute legal action to recover sums due over SL Rs 150,000. By adopting this procedure, the delay normally experienced in recovery actions in civil courts was overcome and the banks were able to speedily obtain writs against defaulters. A second important legal reform was the Recovery of Loans by Banks (Special Provisions) Act No. 4 of 1990 whereby the power of parate execution enjoyed previously by state banks was extended to the other LCBs.23 Parate rights refer to the ability of a lender to foreclose and sell a defaulted property without going to court.24 With parate, the lending institutions should be in a position to recover their dues in a better and speedy manner. Parate execution can lead to an ultimate sale within a period of about four to five months unless the mortgagor succeeds in obtaining an injunction or writ to stay the sale. Though parate is now more widely available, its effectiveness has been questioned by market participants, because of the weak eviction powers of most lenders, rendering parate powers in a sense mute. Foreclosed properties may be occupied by the defaulted owner from six years to a decade later, while the eviction proceeding winds through the courts. Tenant protections under the Rent Act are strong. These difficulties account for the low use of parate thus far in practice—only 18 cases have appeared before the courts regarding parate, of which 17 were ruled favorably.25 Most lenders agree that the foreclosure sale is a useful threat of embarrassment and inconvenience, but the expense of eviction often isn't worth the effort. Given that parate in practice has little bite other than its “threat� value, it is surprising that banks set average loan-to-collateral limits as high as 75 percent. Political and social pressure, especially on state- owned banks, prevent these lenders from auctioning collateral. Banks selling foreclosed property are required by the Banking Act (section 49) to obtain Monetary Board approval of the sale price. This regulation prevents banks from selling at a price that is lower than the market value set by the authorities and acts as another considerable impediment to the realization of collateral. For mortgage lenders without recourse to parate powers, the civil court process is long and complex, as provided by the Mortgage Act and the Civil Procedure Code.26 The proceedings may take about three to five years if uncontested; appeals take much longer. Proceedings tend to be costly.27 Once the order to sell is issued, whether pending appeal or whether no appeal is made, the auctioneer will be entitled to advertise and sell the property. Any defects in title might then render the property unsellable. 31 Title insurance is available in Sri Lanka, but it rarely is used (less than 1 percent of mortgage borrowers have title insurance). Further clarity is needed as to the true beneficiary of such policies (borrowers versus lenders). In sum, the weak collateral realization framework is not a fundamental barrier to the expansion of home mortgage lending, albeit that parate powers function as threats. That said, a sound title system and a nonjudicial system of foreclosure and eviction certainly would benefit market development and could well make the difference for individuals being granted a loan or not, specifically for middle-income borrowers. 4.3. Taxation The government collects stamp tax on house sale transactions, in the amount of 3 to 5 percent of the transaction price.28 International experience and economic theory suggest that other forms of tax may be less distorting. Stamp tax discourages sales and encourages underreporting of prices. As a result, recorded prices for private transactions are possibly for only a third of the actual price. Various other taxes, levies, and charges apply to housing and housing finance transactions and constriction. Housing finance institutions have a corporate profit tax of 20 percent (the standard rate is 30 percent) and a value added tax of 15 percent on profit (the capital gains tax was abolished in 2002). Legal fees range between 1 and 3 percent of the value of a property transfer or gift, and include a title report preparation fee of SL Rs 1,500–2,500. Local government authority charges are levied for street line support, plan approval, UDA, and letter of conformity. Development investors receive BOI tax incentives for construction of 25 units or more, which comes in the form of a tax holiday for the first five years, a lower tax of 10 percent for the following two years, and a 20 percent tax reduction thereafter, as well as a customs duty exemption on the import of capital goods. Developers and builders receive tax deductions for mortgage interest payments and capital repayments. 4.4. Property Valuation Standards All lenders perform loan origination and servicing in house. Annex C provides a discussion of some related mortgage system basics and terminology. State banks use in-house appraisers for property valuation, others use contract appraisers. Borrower data are limited and fragmentary, as are records of real estate prices. Property and house price indexes do not presently exist. In addition to data, shortcomings with available computerization levels, technology, and staff training require action. These problems appear to be quite manageable, however, especially at private banks. Appraisers complete a four-year college degree program and become associates or fellows in the Institute of Valuers. Valuations are in demand mostly for compulsory acquisitions (almost always for raw land), taxes, and other purposes, as well as for rental value appraisals for government rental properties. Appraisers are contract appraisers for single-family housing. Structures are valued using implicit or explicit rental values. When replacement costs are used, appraisers look at hard construction costs only. 4.5. Land and Housing Price Movements Information on land prices for 2003 and 2006 obtained from finance companies showed that average land prices in 18 selected areas, on average, have increased by about 28 percent with a growth range of 12 to 71 percent annually. In the meantime, the minimum and maximum prices had grown by about 23 and 31 percent, respectively. The real average land prices (GDP deflator adjusted) recorded an annual growth rate of 18 percent during the same period. This increase is comparable with alternative investments in equity markets, which posted an annual real rate of increase of 21 percent in the All Share Price Index during the same period. Land and house prices may be driven up by both perceived and projected shortages of urban land because of the government’s significant holdings (a virtual monopoly). Speculative buying may be fueled by a real borrowing rate close to zero and the dearth of alternative investments. Increases in housing prices could have direct and indirect impacts on inflation. Rent is a component of the commodity basket of the Colombo Consumers Price Index (CCPI).29 Of more relevance in the Sri Lanka case, increasing housing prices cause a surge in demand for mortgage loans and a demand-pulled 32 inflation. The surge in inflation in 2005–06 in part was caused by a growth in private credit, not the least of which was caused by mortgage loan growth. Studies on land prices are limited because of the lack of a suitable information system. To determine these prices, the first step would be to ascertain price behavior in the land market and identify factors that affect land prices to establish the baseline for a land price index. Such an index could then be used to identify the buildup of asset price bubbles and other potential related threats to the financial system. Preliminary work on the construction of housing and land price indexes is currently under way at the CBSL. In 2006, CBSL conducted a sample survey on factors affecting land prices.30 The findings indicated that price behavior in the land market was rational and strongly linked to distance from Colombo, convenience of access to goods and services, and degree of urbanization. The survey showed that the share of land purchases for investment and business purposes was about 23 percent, whereas the majority of land purchases were for residential and related purposes (77 percent). The survey findings indicated that most land transfers were direct purchases from owners and only around 8 percent took place through a broker or from real estate companies. Furthermore, it was found that institutional funds were used for only 18 percent of real estate purchases, whereas buyers used their own funds in nearly 79 percent of transactions. Information provided by this survey was limited to land transactions only, however, and it did not cover total financial institutional support to real estate development activities. Consequently, although there was no evidence from this survey of threats to the financial system from land transactions alone, a further survey of all financial institutional involvement in the entirety of real estate development activities is required for a complete assessment. “Excessive credit� growth is sometimes associated with asset price bubbles. Using credit granted by LCBs against the collateral of immovable property as a proxy for excessive credit, the share of such credit increased from 27 percent in 2000 to 32 percent in 2006. Loans granted for personal and commercial construction of houses showed a similar growth trend during the first half of 2006. Such growth rates are not indicative of a bubble. The increase in average loan size moreover can be explained by the rising cost of construction materials and the general price level. This growth is further substantiated by the relatively stable numbers of loans outstanding during the period. Factors such as the loan value being only 60 to 75 percent of the forced sale value of such properties and the existence of a higher haircut 31 for the purpose of provisioning mitigate the potential risks to the banks. Furthermore, repayments of loans in most cases, particularly residential houses, do not depend on cash flows from such property but from other sources, such as monthly incomes. Nonetheless, constant vigilance and monitoring of trends in credit to these sectors is needed to identify potential risks and to take preventive measures. 4.6. The Credit Information Bureau The Credit Information Bureau (CRIB) was created under the terms of the Credit Bureau of Sri Lanka Act No. 18 of 1990 as amended in 1995 as the only credit bureau in the country. Its purpose is to collect and provide credit information on all financial institutions engaged in providing credit as registered with the CBSL. Specifically, the bureau collects both positive (regular) and negative (irregular) information on performing loans greater than SL Rs 500,000 and on NPLs (three to four months overdue) greater than SL Rs 100,000, as well as information on credit cards with outstanding balances of SL Rs 5,000 that are more than 90 days overdue. Data are collected from commercial and specialized banks, registered finance companies, and specialized leasing companies. By its own estimates, CRIB has been able to cover approximately 80 percent of the value of total loans made. Information on individuals is comprehensive and on companies is detailed. CRIB is owned 40 percent by CBSL and 60 percent by six state-owned banks.32 CRIB recorded a considerable increase in the number of reports supplied to lending institutions. It issued 801,508 credit reports on corporate and individual borrowers in 2006 compared with 607,990 reports in 2005. The number of positive (regular) advances reported to CRIB increased by 38 percent over 2006 and the number of negative (irregular) advances reported to CRIB declined by 56 percent in 2006. With the expansion of credit to the private sector, the demand for credit reports has increased, and it has become an essential part of the credit appraisal procedure of lending institutions. 33 Although comprehensive, the work of the bureau is constrained by technology limitations. Most of its output is produced manually, including more than 2,000 reports that it produces annually on request. CRIB is profitable (SL Rs 33 million); however, it is essential that it moves to a Web-based information system to expand the scope and efficiency of the operation. Measures are being taken to modernize and widen the scope and improve the coverage of CRIB. A modernization project has commenced to provide online access to CRIB credit information reports. In addition, amendments to the CRIB Act have been proposed that will enable CRIB to service diverse user groups and to provide value added services, such as credit scoring, fraud prevention, and consumer protection. CRIB is expanding the scope of its operation to include any credit provider, thereby including SME and microlending as well as a wider potential pool of users, such as insurance providers, telecom providers, utilities, and superannuation funds. Risk Management 4.7. Operational Risk33 State-owned specialized lenders cannot offer adjustable rate loans, because they do not have a servicing system capable of making such adjustments. Mortgage origination and servicing each require specialized human skills, investment in infrastructure, and information technology. Third-party specialists provide appraisal, legal, brokerage, and other services. Absent the technology and systems for efficient operations, banks find it hard to profitably offer a rich range of relevant products. Some private lenders are in the process of implementing off-the-shelf systems with adjustable rate capability. Other operational risk mitigation measures include the requirement for commercial banks to develop Business Continuity Plans to deal with emergencies. The state-owned specialized lenders have not formulated detailed operating plans. In general, operational risk does not raise significant concerns for the housing financing system in Sri Lanka at this time, but these risks may cause a bottleneck to growth in the long run. Mortgage rates reflect a markup over the cost of funds plus the cost of operations. Operational efficiency gains can save Sri Lankan borrowers 100 to 300 basis points of the loan principle annually, substantially improving loan affordability. Some private commercial banks originate a mortgage as fast as within 10 days on average. Other borrowers are subjected to a difficult, stressful, and prolonged process to receive mortgage credit. Increased private sector competition is likely to provide the impetus for improved efficiency, as well as for mortgage innovation and product diversification as conditions change and borrowers with different needs enter the market. Even greater scope exists for service innovation. 4.8. Investment Risk and Return Credit Risk. The ratio of NPLs at specialized mortgage lenders has dropped over the last five years. Some lenders had ratios greater than 20 percent; however, that is changing with the restructuring of the two state-owned commercial banks, as well as with the improvement in performance of private banks (see NPL discussion in section 3.4). Extreme moral hazard may have been present in collections, which was partially addressed by the introduction of good management practices. Loss provisions for home loans as a percent of NPLs have trended down over the past few years. Whether these provisions are sufficient depends on the extent to which the threat of foreclosure and court proceedings is successful in ensuring that loans are made current. The CBSL sample survey indicates that price behavior in the land market is rational and there is no housing bubble. The situation will require constant monitoring as mortgage lending expands and affordability of mortgage credit increases the use of leverage to acquire houses. This increased monitoring will be essential because it is difficult to distinguish an asset price bubble from a price adjustment to market equilibrium. Liquidity Risk. Liquidity is not expected to represent a systemic issue in Sri Lanka, because mortgages are a small part of overall financial system assets. Nevertheless, the CBSL “Financial System Stability Review� (2006a) notes regulatory concern with funding long maturity loans and investments with "very short term borrowings in the call money market and reverse repo window of the Central Bank," sometimes referred to as "hot money." This warrants increased monitoring as mortgage lending expands as a percentage of 34 assets. This problem has been addressed based on international experience by reviewing the CBSL collateral eligibility rules at the discount window34 to maximize industry liquidity. An alternative strategy with successful application internationally is the creation of a mortgage liquidity facility (see Chapter 5). Interest Rate Risk. No mortgage lenders have long-term deposits on their balance sheets, that is, beyond five years. But interest rate risk is not expected to be a significant problem for private commercial banks or the NSB in the near term. Mortgage lending duration is matched against the duration of capital and is small in proportion to these banks’ balance sheets. This is not the case at the SMIB or the HDFC. Interest rate swaps with the life insurance companies could be used to solve the exposure to risk, provided that existing laws and regulations do not contain potential obstacles to such transactions. Interest rate risk of fixed rate mortgage lending is usually reduced by funding mortgage lending with long- term debt instruments, such as privately placed general obligation bonds and mortgage-backed bonds, or alternatively mortgage securitization. Experiments with private placements have occurred and can realistically develop into a secondary mortgage market in the long term. More complicated structures that provide for a sale of assets are not likely to be used by mortgage originators unless third-party mortgage credit insurance is offered because investors are unlikely to accept the credit risk. Unnecessary costs, such as withholding tax on MBS, rules for registration of public offerings, or ratings for which this adds little value, will stem secondary mortgage market growth. Other obstructions are likely to exist on pledge-of- collateral regulations and CBSL regulations for institutions issuing collateralized debt. At this stage and in the medium term, covered bonds are probably sufficient to meet the lending and risk management needs of retail deposit-funded home mortgage lenders in the foreseeable future. Mortgage securitization is a sophisticated, complex, and costly transaction, requiring considerable regulatory and institutional infrastructure to develop, as well as effective demand for long-term investment. In the event of a dearth of willing long-term investors and swap counterparties, home mortgage borrowers will have to bear more of the risk of volatility by accepting an adjustable rate loan. The increased borrower risk will inhibit mortgage lending growth in the next decade if the fiscal, and hence monetary, conditions cause rates to fluctuate excessively. The recent rise in the interest rate on adjustable rate mortgage loans demonstrated to borrowers the inherent risk of these types of loans. 4.9. Political Risk Political risk is often the most problematic for nascent home mortgage markets to deal with because market tools to manage it are not available, and hence the typical response is aversion. Sri Lanka is still in the early stages of a transition to market for houses, with spillover effects affecting housing finance. Increasing numbers of privately owned commercial buildings are being produced. Residential buildings are being constructed to serve the high-income end of the spectrum. Yet there is a dearth of larger multifamily buildings or new private developments for middle-income households. The most common reason for the absence of this type of housing is the real or perceived political risk of achieving a market return on investment because of rental and eviction protections. This problem then results in proposals for the state to build, and then finance, dwellings for the affected parts of the population. As subsidies are inevitably provided, either directly or indirectly, the result is that this part of the market remains segmented from the liberalizing financial markets. The greater the segmentation, the deeper the problem becomes. Sri Lanka has taken numerous steps to end this cycle. It has stopped central government production programs, although pressure for government production continues through a variety of avenues and the HDFC continues to build houses. It has required state-owned or controlled entities to fend for themselves, although in the case of the UDA and the NHDA, the state has provided a massive endowment of public land for development. The government has restructured the state-directed mortgage lenders to rely on self-funding and trained them in market management techniques. It is too early to predict whether the steps taken thus far will be sufficient. But early evidence shows that more work is needed (see Chapter 5). The state-sponsored lenders could do more to expand their lending to low-income groups. These lenders are involved simultaneously in production and sales. The mortgage lending terms and conditions do not reflect their cost of funds, cost of operations, or risk. These and other factors create distortions that will likely grow and become more problematic for financial system regulators and private competitors over time. 35 Chapter 5 Policy Options: Elements of a Housing Finance Strategy 5.1. A Housing Finance Policy Paradox Developers take out magnificent advertisements for big condominium projects and single-family attached dwellings at the luxury end of the housing spectrum. Buyers are numerous among upper-level-income groups, often including expatriates. Flats, including those on subsidized financing, are bought and immediately resold for a profit, and many are rented out for profit to high-profile businesses and individuals. Housing supply is growing at an unprecedented speed, as are prices. DFCC and banks are more than ready to fund these new developments. Newspapers are full of real estate classified ads for both rental and purchase properties. At the same time, more than 80 percent of the households in Sri Lanka have no access to home financing, and about 7 percent are lacking a home. One-third live in semipermanent housing, and 6 percent live in line room estates and shanties. For those, market rentals are scarce to nonexistent. Housing turnover is extremely low. Land does not appear to be used as efficiently as market forces would predict. What would explain this seeming paradox? First, markets are segmented. Difficulties recovering money from collateral cause banks to rely on regular salary and other secure means of repayment for their lending, shying away from middle-income households and focusing instead on regular, mostly government, employees. Second, the playing field is not level for private mortgage lenders, and as a result, primary markets are weak. State-owned banks compete in sectors in which the private financial firms are active, instead of addressing failed markets that cannot be commercially viable. Subsidies are not precisely targeted at vulnerable groups and are instead chasing middle-income households, because government housing institutions have lost budget funding. 5.2. A Growth Scenario The home mortgage market is growing and deepening in Sri Lanka. The current effective demand for home mortgage credit is being met adequately by the private commercial banks and the NSB. Furthermore, these financial institutions can be expected to comfortably handle growing demand at 15 percent or more a year on the value of housing loans35 in the coming 10 years. Both theory and experience strongly indicate that commercial banks are the logical intermediaries to make mortgage loans in the Sri Lankan context (Renaud 2007). Furthermore, the pace of growth in home mortgage credit should more than keep up with the pace of new construction. The institutional capacity for a rapid expansion remains to be strengthened. The remainder of this section provides an overview of suggested improvements. Further liberalization in the lending environment is needed to transit to a market for home mortgages to middle-income households. Housing finance growth is nowhere near the pace necessary to close the housing gap over the next 10- year period. One serious precondition for mortgage market growth is the assumption of continued liberalization policies and a broad implementation of the National Development Strategy as planned by the government. A second significant defining factor for development of the housing finance market is the ability to evict and realize collateral. With collateral functioning only as a threat, and not as an asset of monetary value, mortgage lending will remain confined to the top income groups, and it will continue to co-rely on regular salary income and other assurances of palpable monetary value so that lenders can ensure repayment. Failing those preconditions, the extent of lending expansion from the top of the income scale down depends largely on the overall supply of funds. Which brings us to the third key issue––that is, mortgage loan funding. How much more credit will flow from the economy into housing finance ultimately depends on the claims to domestic financial savings by the government, SOEs, and business. In particular, government debt might crowd out consumer and business lending, including funds for mortgage loans. The banking sector has a limited potential to fuel a “housing finance boom.� Having examined the entire financial system for possible sources of funds for mortgages, Chapter 3 concluded that the funds necessary to close this gap are not likely to be found outside the banking sector, either, in the short or the medium term. In the long term, funding can be 36 sourced from provident funds and the insurance sector. Successful cases from international experience with mortgage market development are provided in annex E and an overview of international best practice is provided in annex F. Neither macroeconomic resource constraints nor lender capacity constraints necessarily preclude efforts to improve the housing conditions of the poor. As emphasized by the National Development Strategy, some house improvement lending should be provided with growth in microlending, as infrastructure becomes the main focus of government policy. Microfinance institutional growth should be aided by an enabling yet safe policy environment. New housing production that efficiently uses land will improve housing affordability to middle-income households. Policy should encourage private lending to middle- income households to be integrated into the home mortgage market in time, as banks downscale. Finally, existing state-owned market participants, with further governance and efficiency reforms, could be directed to address clear market failures among the very poor, via transparent and well-targeted subsidies (on an equal footing with the private sector), or could be stripped of any government guarantee or subsidy (implicit or explicit) and be left to pursue market lending, again on a level playing field with the private sector. 5.3. An Inclusive and Sustainable Housing Finance System Macroeconomic stability is a required sound foundation for a well-developing housing finance system. In Sri Lanka, the inflation risk and the government debt crowding-out are particular concerns. The National Development Strategy has planned to tackle these problems. A workable legal and regulatory system is central to providing an enabling policy environment for housing finance. The housing finance sector needs to have a supportive regulatory framework that does not impede but stimulates its growth. A detailed list of longer-term legal changes that should be brought to the attention of policy makers is provided in annex G. Two main legal prerequisites from international experience should be emphasized. The first necessity to develop primary mortgage markets is a functional land registration and titling system. This work is ongoing in Sri Lanka. The second prerequisite to develop active and efficient mortgage markets, as suggested by international efforts, is data availability. Hence, we have the following recommendations: Recommendation 1: Property and House Price Indexes need to be created, following up on the initial preparatory work ongoing at the Central Bank of Sri Lanka. Recommendation 2: Further improvements in the functioning of the Credit Information Bureau (CRIB) will contribute to protecting consumers and avoiding lengthy disputes in court involving finance companies. The Asian Development Bank Financial Sector Assessment suggests CRIB privatization. The enabling role of the state should be strengthened. The government should forge ahead in its current strategy of housing finance liberalization and continue to reduce government interventions in the market. The government could play a catalyst role to expand accessibility. Policy makers should shy away from overregulating anticipated problems and instead focus on a simple but enabling policy. Regulations should be implemented to address specific problems as they arise with the functioning of markets in the least prescriptive and intrusive manner. Above all, policies should be robust and parsimonious, not complex or costly, in keeping with existing financial institution capacity and the level of sophistication of the current financial infrastructure. Complex regulations incommensurate with current housing finance system needs would stifle (not encourage) growth and efficiency gains, and might lead to regulatory arbitrage. How can the government assist in expanding housing finance while avoiding market distortions? As a priority, the state-owned mortgage banks need a structural reform. Considerable effort is required in further improvement in the areas of governance, management, and efficiency improvements, training, and modernization and computerization. These improvements should be made within the rules and risk management measures introduced by the CBSL. Revamping state-owned lenders is a significant challenge. 37 State banks (SMIB and HDFC) should be exposed to general market competition and should operate at a level playing field with other market participants. The policy choice between generic commercial banks and specialized mortgage lenders is also on the agenda in Sri Lanka at this juncture. Economic theory tells us that when mortgage credit risk is “borrower specific,� the generic bank model spreads the risk better than a system of specialized lenders. Furthermore, best practice suggests that integrating specialized banks into the general regulatory system could eliminate market segmentation and offer considerable gains to homeowners. The ADB Financial Sector Assessment suggests adopting a single banking license rather than the current distinction between commercial and specialized banks. Risk management and operating reforms suggested in annex G for mortgage lenders apply to state-owned banks as well. Recommendation 3: State-owned mortgage banks (the State Mortgage and Investment Bank [SMIB] and Housing Development Finance Corporation [HDFC]) would benefit from a rethinking of their strategy and mission, as well as further improvements in their governance and operational efficiency. The rationale for having two separate institutions might merit rethinking as well, especially as they compete with each other and with the private sector. These institutions could better use government budget funds to the wider benefit of Sri Lankan society. SMIB and HDFC were created at a stage when Sri Lanka’s mortgage markets were just developing; however, the current robust growth shows that the private sector is fully capable of absorbing existing demand, except for lower-income groups. Specifically:  Strategic refocusing of SMIB and HDFC should be considered. The banks could undertake two sets of activities, which are not mutually exclusive: (1) fully competitive housing lending without any (explicit or implicit) subsidies or guarantees at a level playing field with the private sector; and (2) again on an equal footing as the private sector, be allowed to channel transparent, well-targeted subsidies exclusively to the lower-income groups (and unavailable to middle- or upper-income groups).  Depoliticized boards of directors would function better. They could be in a stronger position to make management accountable and could enforce the requisite modern internal controls.  Deeper efficiency improvements would require upgrades and human resources training, as well as modernization and computerization of existing operations.  Improved risk management and credit loss indicators would need to address the exposure to credit and market risks as well (perhaps via reductions in nonperforming loans and a transition to adjustable mortgage rate instruments). In the case of state-owned mortgage banks, there is no clear policy rationale for maintaining their special historically determined privileged status, and perhaps implicit subsidies and government backing as well. Instead, transparent and well-targeted subsidies should be offered to any institution in the sector, fostering competition (see detailed recommendations on subsidies in annex G). A detailed study of existing subsidy programs would identify potential budget savings as well as opportunities for better targeting and efficiency improvements in the subsidy schemes. Much of the same recommendations apply to rethinking the state role as a market player on the housing market. International experience points to a tendency away from using publicly owned construction companies for building low-cost housing, and instead points to using private companies. Housing is not a public good; its construction does not enjoy economies of scale. The housing construction business in most countries is highly competitive. In Sri Lanka, the system has increasingly become unviable because of competing government priorities and fiscal pressures. More aggressive growth in primary mortgage markets would require addressing the mismatch between short-term deposits and long-term lending. Middle-income household access to the mortgage 38 market is currently inhibited because of the inability of banks to lend without collateral, on the one hand, and the difficulties in realizing collateral value, on the other. Solutions to the foreclosure and eviction problems would remove the bias against middle-income lending by retail mortgage lenders, but such solutions have been slow in coming in most transitioning countries and will likely be so in Sri Lanka as well. An alternative solution that has been used is mortgage default insurance (MDI). Because MDI tends to be information dependent, potential is low for the development of private MDI in Sri Lanka, and international experience points to disappointing results with state-owned MDI.36 To deepen and expand the mortgage market to middle-income households, a couple of possible mechanisms can be used: (1) private placements of covered bonds by mortgage lenders or (2) a liquidity facility. Secondary mortgage markets are an option, in theory, but they will not develop in a vacuum. These markets need to base their activities on a strong and active primary mortgage market of a significant scale, which currently does not exist in Sri Lanka. Whichever instruments are used, the policy success will be predicated on the ability to ensure a willing and able counterparty to the transaction that would invest long-term funds. Banks cannot finance from abroad. Whether it is a foreign investor or domestic institution (such as the EPF, ETF, private provident funds, and insurance companies) or the government (in the case of international success models of liquidity facilities), a counterparty is necessary for the market to develop. Recommendation 4: Viable options for deepening the mortgage market should be carefully considered, including covered bonds by mortgage lenders or a liquidity facility. Securitization is not suggested as an option at this early stage because of the limited potential of primary mortgage markets in Sri Lanka to attain a sizeable scale and strength in the short to medium term. Centralized liquidity facilities are used to prevent liquidity crises caused my mismatch between long- term lending and short-term borrowing. Beyond this function (which can also be performed by the central bank via a discounting facility), liquidity facilities have been used to spur secondary mortgage market development. Cagamas in Malaysia is an example of a successfully implemented liquidity facility. It makes advances with “full recourse� as a lender of first resort with relatively cheap funds (see box 2). Most liquidity facilities have government backing of some sort and take on some of the risk of mortgage lending. Liquidity facilities are vulnerable to abuse, as a means for lenders to offload other risks in addition to liquidity risk, not the least of which is political risk. As a result, in most industrial economies, the role of a guardian against systemic liquidity risk is performed by central bank “lender of last resort� discount facilities. Box 2. Cagamas Berhad, the Malaysian National Mortgage Corporation Cagamas was established in 1986 to promote the secondary mortgage market in Malaysia. Its corporate mission is to provide financial products that would make housing loans more accessible and affordable, particularly to lower-income groups. It borrows money by issuing debt securities and uses the funds to finance the purchase of housing loans from financial institutions, selected corporations, and the government. The provision of liquidity at a reasonable cost to the primary lenders of housing loans encourages further financing of housing at an affordable cost. Besides housing loans, Cagamas’ other products include industrial property loans, hire purchase and leasing debts (conventional and Islamic), and credit card receivables. The case study of Cagamas and an assessment of impact and lessons learned is presented in Loic Chiquier, “Secondary Mortgage Facilities: A Case Study of Malaysia’s Cagamas Berhad, World Bank, Washington, DC, 1999. Source: Cagamas Web site, www.cagamas.com.my 2007. Sri Lanka has successfully undertaken significant steps to liberalize the housing finance system provide housing finance infrastructure such as parate, titling, and data requirements; and ensure stable macroeconomy and budgetary conditions for the propitious development of home financing. Much work remains to be done, but private mortgage market growth is encouraging and suggests a significant potential for mortgage market development in the country. 39 Annexes Annex A: Statistical Annexure Annex Table A1. Population, Income, and Price Levels Year GDP (SL Rs GDP GDP GDP CCPI Inflation Population Population million) at Growth per per Index Moving Growth Current Rate Capita Capita Average Rate Market Prices (SL Rs) (US$) (December) 2000 1,257,636 6.0 68,102 810 2,540 6.2% 18,467 1.3 2001 1,407,398 −1.5 75,133 840 2,899 14.2% 18,732 1.2 2002 1,581,885 4.0 83,267 870 3,176 9.6% 19,007 1.2 2003 1,761,161 6.0 91,479 948 3,377 6.3% 19,252 1.2 2004 2,029,365 5.4 104,273 1,030 3,633 7.6% 19,462 1.2 2005 2,365,593 6.0 120,276 1,197 3,056 11.6% 19,668 1.0 2006 2,801,828a 7.4a 140,894 1,355 3,611 13.7% 19,886 — Sources: Department of Census and Statistics 2007. Note: Base: 1952 = 100. CCPI = Colombo Consumers Price Index; GDP = gross domestic product; SL Rs = Sri Lanka rupees; — = not available a. provisional. 40 Annex Table A2. Housing Stock in Sri Lanka by District: Type of Housing Unit, Type of Structure, and Tenure Row Total Semi- Single Attached House House / Hut / Housing Permanent permanent Improvised Houses Flats / Annex Line Room Shanty Owned Rented Encroached Division Population Units (percent) (percent) (percent) (percent) (percent) (percent) (percent) (percent) (percent) (percent) (percent) Nationwide 18,331,748 4,687,157 64.6% 26.5% 1.7% 80.1% 5.4% 3.3% 5.0% 0.7% 76.3% 5.4% 1.3% Colombo 2,234,289 507,678 86.8% 11.2% 0.3% 71.6% 9.2% 8.1% 6.4% 1.5% 72.3% 17.5% 1.5% Gampaha 2,066,096 511,621 80.0% 17.7% 1.1% 87.7% 1.6% 4.2% 2.7% 0.7% 77.8% 11.4% 2.4% Kalutara 1,060,800 270,147 78.5% 20.1% 0.4% 90.4% 1.0% 3.0% 2.9% 0.3% 84.4% 4.9% 0.7% Kandy 1,272,463 316,610 73.2% 25.8% 0.2% 83.9% 1.8% 3.9% 8.4% 0.3% 82.1% 5.8% 1.0% Matale 442,427 120,335 56.5% 41.7% 1.4% 88.1% 0.9% 2.5% 5.9% 1.1% 83.9% 3.7% 2.3% Nuwara Eliya 700,083 175,423 64.8% 34.0% 0.2% 49.9% 0.4% 9.9% 35.7% 0.9% 53.9% 4.5% 0.3% Galle 990,539 253,697 72.3% 26.3% 0.9% 93.9% 0.9% 1.7% 1.8% 0.3% 88.6% 3.7% — Matara 761,236 197,552 71.7% 27.3% 0.6% 93.5% 1.3% 1.9% 1.8% 0.2% 89.5% 2.0% 1.0% Hambantota 525,370 148,516 58.8% 40.1% 0.4% 95.6% 0.2% 1.7% 0.3% 1.1% 92.2% 1.6% 1.9% Jaffna 490,621 110,603 — — — — — — — — — — — Vavuniya 149,835 31,849 — — — — — — — — — — — Batticaloa 486,447 112,778 — — — — — — — — — — — Ampara 589,344 145,184 67.5% 28.7% 3.1% 91.0% 0.0% 1.9% 0.4% 3.1% 87.7% 2.2% 0.6% Trincomalee 340,158 81,379 — — — — — — — — — — — Kurunegala 1,452,369 411,545 63.5% 34.6% 1.2% 95.3% 0.3% 1.1% 1.0% 0.5% 90.1% 2.6% 1.2% Puttalam 705,342 192,295 61.1% 31.9% 6.2% 92.2% 0.2% 1.7% 1.4% 2.3% 82.8% 3.5% 1.8% Anuradhapura 746,466 206,078 53.2% 45.5% 0.4% 95.0% 0.5% 1.3% 0.2% 0.3% 89.5% 2.2% 1.1% Polonnaruwa 359,197 98,491 61.4% 37.4% 0.3% 95.9% 0.1% 1.4% 0.2% 0.6% 91.3% 2.2% 1.8% Badulla 774,555 200,076 63.6% 35.0% 20.0% 77.1% 1.1% 6.2% 13.2% 0.4% 72.9% 3.1% 2.0% Moneragala 396,173 110,063 50.3% 48.1% 0.6% 93.1% 0.2% 1.2% 2.0% 1.1% 87.6% 3.9% 5.1% Ratnapura 1,008,164 273,447 60.3% 39.1% 0.1% 86.0% 0.7% 2.6% 9.2% 0.2% 84.6% 1.7% 1.0% Kegalle 779,774 211,790 68.7% 30.1% 0.5% 89.5% 0.4% 1.3% 5.8% 0.6% 87.7% 2.0% 0.6% Source: Department of Census and Statistics 2003. Note: The data for Mannar, Mullaitivu, and Killinochchi Districts are not available. — = not available. Annex Table A3. Building Units by District Division Housing Units % Collective Living Quarters % Institutions % Nonhousing Units % Total Number of Units Total 4,687,157 87.1 36,980 0.7 24,097 0.4 632,469 11.8 5,380,703 Colombo 507,678 83.5 8,477 1.4 2,599 0.4 89,349 14.7 608,103 Gampaha 511,621 85.4 16,404 2.7 1,877 0.3 68,996 11.5 598,898 Kalutara 270,147 88.8 757 0.2 1,332 0.4 32,139 10.6 304,375 Kandy 316,610 87.4 705 0.2 1,937 0.5 42,940 11.9 362,192 Matale 120,335 87.1 464 0.3 639 0.5 16,659 12.1 138,097 Nuwara Eliya 175,423 90.1 526 0.3 722 0.4 18,072 9.3 194,743 Galle 253,697 88.3 1,079 0.4 1,422 0.5 31,032 10.8 287,230 Matara 197,552 87.3 518 0.2 968 0.4 27,176 12.0 226,214 Hambantota 148,516 88.3 569 0.3 885 0.5 18,221 10.8 168,191 Jaffna 110,603 84.0 284 0.2 471 0.4 20,260 15.4 131,618 Vavuniya 31,849 88.2 96 0.3 208 0.6 3,957 11.0 36,110 Batticaloa 112,778 89.2 277 0.2 377 0.3 12,963 10.3 126,395 Ampara 145,184 86.1 610 0.4 871 0.5 22,047 13.1 168,712 Trincomalee 81,379 87.7 530 0.6 654 0.7 10,178 11.0 92,741 Kurunegala 411,545 87.7 680 0.1 2,299 0.5 54,519 11.6 469,043 Puttalam 192,295 86.8 1,213 0.5 961 0.4 27,080 12.2 221,549 Anuradhapura 206,078 87.1 806 0.3 1,524 0.6 28,107 11.9 236,515 Polonnaruwa 98,491 86.1 716 0.6 523 0.5 14,657 12.8 114,387 Badulla 200,076 89.0 648 0.3 816 0.4 23,247 10.3 224,787 Moneragala 110,063 87.5 690 0.5 703 0.6 14,278 11.4 125,734 Ratnapura 273,447 89.7 689 0.2 1,262 0.4 29,329 9.6 304,727 Kegalle 211,790 88.1 242 0.1 1,047 0.4 27,263 11.3 240,342 Source: Department of Census and Statistics 2003. Note: The data for Mannar, Mullaitivu, and Killinochchi Districts are not available. Annex Table A4. Total Assets of the Financial System Financial Systems Number Assets Assets (SL Rs (percent of a billion) GDP) 2006 a 2006 Institutions supervised by the CBSL Deposit-taking institutions Licensed commercial banks 23 1,774.30 63% State banks 2 798.44 28% Domestic private banks 9 709.72 25% Foreign banks 12 88.72 3% Licensed specialized banks 14 356.4 13% DFCC Bank 1 53.5 2% National Development Bank 1 n.a. n.a. National Savings Bank 1 242.4 9% Housing Development Finance Corporation 1 n.a. n.a. State Mortgage and Investment Bank 1 10.7 0% Regional development banks 6 24.9 1% Private savings and development banks 3 n.a. n.a. Registered finance companies 26 113.1 4% Other institutions Employee Provident Fund 1 492.1 18% Institutions not supervised by the CBSL Deposit-taking institutions Cooperative rural banks 268 27.8 1% Thrift and credit cooperative societies 8,500 4.9 0% Contractual savings institutions Private provident funds 190 112.6 4% Employee Trust Fund 1 66.8 2% Insurance companies 12 107.3 4% Source: Central Bank of Sri Lanka 2006b. Note: CBSL = Central Bank of Sri Lanka; DFCC = Development Finance Credit Corporation; GDP = gross domestic product; n.a. = not available. a. 2006 figures are projected. Annex Table A5. Lending and Deposit Rates 2002 2003 2004 2005 2006 CBSL Rates CBSL bank rate 18 15 15 15 15 CBSL overnight Repo rate 9.75 7 7.5 8.75 10 CBSL reverse Repo rate 11.75 8.5 9 10.25 11.5 Deposit Rates LCB deposit 3-month fixed 6.50–10.50 4.00–7.13 4.00–9.25 5.00–13.00 5.75–14.00 LCB deposit 6-month fixed 7.00–10.50 5.00–7.25 4.50–9.50 5.50–11.25 6.00–13.75 NSB deposit 6-month fixed 9.80 6.50 7.80 8.80 11.25 LCB deposit 12-month fixed 7.50–11.00 5.00–7.75 5.50–9.75 5.50–11.50 5.50–14.00 NSB deposit 12-month fixeda 9.60 6.80 7.80 6.70 10.50 LCB deposit 24-month fixed 7.44–11.50 5.41–7.50 6.80–10.25 8.50–11.50 9.00–14.25 NSB deposit 24-month fixed 10.50 7.30 8.40 9.40 10.50 LCB savings deposit 3.50–11.00 2.10–7.25 3.00–7.75 3.00–10.25 3.00–10.50 NSB savings deposit 6.00 5.00 5.00 5.00 5.00 NSB savings certificates 11.00 11.00 11.00 11.00 11.00 LCB AWOR 7.47 5.27 5.31 6.24 7.60 Lending Rates LCB lending against inventory 12.00– 7.00–23.00 9.00–23.00 9.00–23.00 8.36–27.00 25.00 LCB lending against immovable 10.00– 7.00–29.00 8.00–22.00 10.00– 7.85–27.00 29.00 22.50 Other LCB-secured lending 12.00– 7.00–30.00 6.00–26.00 6.00–33.00 6.00–33.00 30.00 LCB unsecured lending 8.00–30.00 7.00–30.00 7.00–30.00 7.00–33.00 6.00–33.00 Bills purchased and discounted 10.00– 7.00–19.00 7.50–20.75 7.00–31.00 7.10–20.50 23.00 Weighted average prime lending 12.17 8.95 10.17 12.14 14.74 rate Overnight SLIBOR 10.67 7.73 9.66 10.93 14.73 Call market rate (average) 10.39 7.59 9.73 10.75 14.47 SMIB lending rate 15.00– 12.00–13.25 12.00– 12.00– 15.00– 16.50 13.25 13.25 16.50 DFCC lending rate 11.50– 9.25–16.00 9.25–16.00 5.00–17.00 15.00– 19.00 18.00 NHDA lending rate 10.00– 11.00 11.00 11.00 11.00 16.00 NSB lending rate 14.00– 10.00–12.00 10.00– 10.00– 12.00– 16.50 12.00 12.00 13.00 NDB lending rate 10.81– 7.75–15.20 9.60–16.00 –– –– 18.35 Source: Central Bank of Sri Lanka 2007. Note: AWOR = Average Weighted Offer Rate CBSL = Central Bank of Sri Lanka; DFCC = Development Finance Credit Corporation; LCB = licensed commercial bank; NDB = National Development Bank; NHDA = National Housing Development Authority; NSB = National Savings Bank; SLIBOR = Sri Lanka Inter Bank Offer Rate; SMIB = State Mortgage and Investment Bank. a. Monthly interest payments. 44 Annex Table A6. Mortgage Products Offered in the Home Lending Market, Major Players Institution Description Average Loan Debt Average Sources of Funds Fixed/ Average Efficiency Lending Prepayment (1) (2) Size Service-to- Loan to (6) Floating Loan (9) Rates Penalty (3) Income Value Rate Maturity (10) (11) Maximum (5) (7) (8) (4) State Mortgage State bank specializing SL Rs 600,000 60% for any 75% 29% of funds from All fixed. 15 years. Delinquency 16% n.a. and Investment in mortgage financing income deposits; deposits rate 25%. Bank (SMIB) for middle-income from ETF / EPF and 16% is less than population. other government their cost of agencies; some NGO funds plus sources (Cooperative administrative Society). expenses Housing Specialized bank––the SL Rs 500,000 40% for low 80% USAID and the Virtually 20 years Delinquency 16% n.a. Development biggest mortgage (SL Rs 250,000 income, 60% ADB, some private all - fixed rate 6%. Finance lender by loan volume, minimum) for high bond placements Mark up of 4%. Corporation as of 2004. Listed. NIM 1%, admin (HDFC) costs 2% National Private commercial SL Rs 1 million 45% - mid- 75% EPF line of credit, Yes, up to 15 Mark up of 2% 15.5–16.5% n.a. Development bank. income; 60% some private bond before years; 3 to 3% Bank/NDB high income placements October years for Housing Bank 2006. current loans National State-owned bank, only Larger than other n.a. 75% deposits Yes 15–20, NPL 4% for Up to SL Rs 0.5% of Remaining Savings Bank consumer lending by banks (land up to 30 mortgages, 2% 750,000– value (NSB) charter, must lend at 60%) years other assets 14.5%. least 60% of deposits Above SL Rs to the government. 750,000– 15.5% Bank of State bank. Started SL Rs 1–1.5 40–60% of 75% Deposits Floating Max 25 NPL 4% AWPLR+1 2% of remaining Ceylon (BOC) focusing on housing million income (50% years fixed for 1 value financing very recently land; 90% year up to 3 construc.) years or AWPLR+0.75 6 quarterly People’s Bank State bank. SL Rs 1 million 54% of 60% Deposits Virtually 5–20 NPL 5% 17% contemplating (PB) Gross (max fixed years introduction of Income 75%) prepayment penalty Hatton One of the leading SL Rs 1 million 40% of net n.a. n.a. Virtually 5–25 3.5 % NPL 17.75–18.75% 1% of out standing or National Bank home loan providers Income fixed years a max of SL Rs from the private sector 50,000 Institution Description Average Loan Size Debt Service– Average Sources of Funds Fixed/ Average Efficiency Lending Rates Prepayment (1) (2) (3) to-Income Loan to (6) Floating Loan (9) (10) Penalty Maximum Value Rate Maturity (11) (4) (5) (7) (8) Seylan Bank Private commercial SL Rs 1 million 60% of gross 75% Deposits Fixed 25 years 10% NPLs 17–18% <5 years: 2% bank. (max.) 5–10 years: 1% >10 years: 0% of remaining Commercial The leading private n.a. Salary >SL 75% Deposits Fixed and 5–20 <3% NPLs Fixed 18% For prepayment Bank sector bank in terms Rs 15,000 floating years Floating 14–20% of 25% of the of efficiency. and 60% of loan–3% of income remaining balance; for 50%–2%; for 75%–1%. Development Lends to real estate SL Rs 80–100 million 40% equity n.a. Deposits and market Floating <4 years No bad loans at WAPLR+1% to No penalty Finance developers (land/ or 50% sale borrowings the moment 2.5% Credit structure), syndicates value from Corporation loans. No individual the bank (DFCC) housing loans. Savordaya Nonprofit charter SME loans: SL Rs n.a. n.a. Can take deposits yes 2–4 years 7% 18% n.a. Enterprise under the companies 15,000–50,000). through membership administrative Development act, must retain Housing upgrades: SL societies costs, Society earnings as a donor Rs 50,000–100,000. 9% cost of funds (SEEDS) condition (no Existing clients: SL Rs shareholders). 500,000 (max.) Sanasa Bank Licensed specialized SL Rs 50,000–100,000 60% for low n.a. Capital; for yes 10 years 4.5% NPLs 16–17% No penalty bank with a unique income (SL microlending, special (max.), 8% cost of charter. Rs 15,000– lines (donor, CBSL, some 5 or funds; 5% 20,000/ and other banks) 7 years administrative month) costs Amana Merchant bank under n.a. n.a. n.a. Deposit financing is n.a. n.a. NPL is about 16% No penalty (a Investments the CBSL license. about 30–40% of 2.5% overall, few defaults, no Three business lines: total assets 0.3% for foreclosures) financial services, mortgages insurance, and stock (depositors brokerage. Shariah share the risk) compliant. Source: Interviews with individual banks 2007. Note: CBSL = Central Bank of Sri Lanka; NPL = nonperforming loan; SME = small and medium enterprise; AWPLR = Average Weighted Prime Lendign Rate; n.a. = not available. 46 Annex B: Strategies to Assist the Development of Low-Income Housing Governments have at their disposal a number of policy instruments to address demand-side and supply- side constraints that limit access to decent housing and constrain housing improvement. On the demand side, those instruments include tenure security provisions; laws and regulations protecting property and inheritance rights, in particular for women; well-targeted and transparent subsidy schemes that are affordable; expansion of healthy and competitive mortgage lending institutions; and establishment or strengthening of innovative arrangements to improve access of the urban poor to finance. On the supply side, instruments may include removing overly restrictive land development and homebuilding regulations; improving construction standards and building codes; promoting the production and use of local construction technologies and building materials; and creating a competitive environment in the housing construction sector. Fiscal Incentives. Although government priorities vary, it is useful to address both supply- and demand-side constraints at the same time. In China, for example, to facilitate access to affordable housing for families with limited income and savings, authorities have combined fiscal incentives for real estate developers, including tax reductions or exemptions to provide housing within a negotiated price range and once-in-a-lifetime equity grants for people living in substandard housing to facilitate homeownership. Grants are based on the market value of the existing housing of these people. Evidence suggests that policy instruments work best when supported by an adequate institutional framework. Proactive Policies. A proactive policy to acquire low-cost land on the urban periphery for future development of affordable housing for low-income households is an option governments may wish to consider. Private Land Markets. Serviced land to shelter the urban poor may be secured from private land markets. Builders could be allowed to develop a portion of the land more intensively than permitted by normal regulations in exchange for allocating other parts of the land to shelter slum dwellers residing in the area. Private landowners in an urban fringe area can be engaged in a partnership to develop and service their lands in an integrated manner, thus reducing development costs. Governments can acquire open land from the land market well in advance of assessed future housing demands, and hence at a relatively low price, and hold it for low-income housing as demands grow. Shelter Improvement. Self-help shelter improvement is an important option in slum communities to fulfill the housing needs and preferences of their residents within the limitations of their budgets. Community housing initiatives, often organized through associations of slum dwellers, were most successful when explicitly recognized by national governments and municipalities in their slum-upgrading programs and poverty reduction schemes and supported by investment in infrastructure and the provision of basic services. Community banks that provide construction materials have played a useful role in community-based self-help shelter improvement (for example, in some Latin American countries). Rental Housing. The rental housing market has an important role to play as a form of affordable housing for the poor. Such housing may be especially well suited to the needs of newly formed households with small children, whose accumulated savings and current savings capacity may be limited. These strategies require the mobilization of financial resources. Public and private domestic capital remains the main source to be tapped, in spite of continuing problems in mobilizing such resources in poor countries. Remittances from overseas workers can provide a significant source of finance for housing. Promoting community-based approaches to the financing of infrastructure and housing is one suggested strategy to mobilize resources. Savings and loan activities are a central function of many such organizations, drawing people together and pooling funds on a regular and continuous basis. Promoting and scaling up microfinance schemes is another approach to the issue of resource mobilization. In recent years, some diversification of microfinance has supported the incremental investments of the poor in home improvement, in part reflecting the dual function of many poor people’s homes as places of business. Various students have identified more than 40 programs in Asia, Latin America and the Caribbean, and Africa. Source: Excerpts from the United Nations, Economic and Social Council, Commission on Sustainable Development, “Human Settlements: Policy Options and Possible Actions to Expedite Implementation,� December 2004. Annex C: Mortgage Operations and Operational Risk Management: Theory Basics Mortgage operations consist of origination and servicing. Origination requires retail access to potential borrowers, and hence there are few scale economies. The primary function is to mitigate subsequent risk of default by underwriting both the borrower and the property. Loan underwriters determine the borrower’s ability and willingness to pay by verifying income and employment level and stability, credit history, sources of down-payment and wealth in the event of interruptions to income, and other factors. Appraisers verify property value by analyzing comparable sales, rental value, and replacement cost. All mortgage lenders underwrite the borrower’s willingness to pay and the value of collateral in the event of default. In cases in which foreclosure and eviction are cheap and easy, lenders rely more heavily on the housing collateral. In cases in which it is expensive and difficult, they rely more heavily on the personal income and wealth characteristics of the borrower. Lenders need to know the customer better for this type of lending. Hence, banking relationships, third-party credit bureaus, credit references, and other sources of information tend to be of paramount importance. This information is usually more readily available for higher-income households who have long-term banking relationships. Hence personal-backed mortgage markets tend to allocate credit from the top of the income strata down. Collateral-backed mortgage lenders, in contrast, have no need to segment borrowers by income. As long as borrower income is sufficient to pay the mortgage, and the collateral is easily marketable, the size and cost of the dwelling and income strata of the borrower is largely irrelevant. Servicing includes collecting monthly payments over the life of the loan, monitoring that insurance premiums and tax liabilities are paid, and foreclosing and evicting in the event of an incurable default. Collateral-backed mortgages require little ongoing monitoring, and hence are amenable to huge scale economies. Personal-backed loans require much more ongoing monitoring and hence are typically serviced by the originator. The investment attributes of home mortgage lending reflect risk and return. Investment risks include credit, liquidity, and interest rate risk. Credit risk refers to the difference between the promised contractual payments and the actual timing and amount of loan repayments. Liquidity risk refers to the inability to roll over short-term funding to finance loans and investments with a longer maturity. Interest rate risk refers to the difference between the contractual rates and maturities of mortgages (more technically, duration) and the liabilities that fund them. Credit Risk. An important determinant of home mortgage market growth is the ease and efficiency of foreclosure and eviction. There are two systems of foreclosure, judicial and nonjudicial. In the latter, the title is held by a trustee, who assigns it to the lender in the event of a default. The law permits the lender to immediately foreclose, evict, and dispose of the property. For judicial foreclosure, the borrower retains ownership throughout the legal proceedings, until a court recognizes the mortgage default and assigns ownership to the lender. This typically takes much more time than nonjudicial foreclosure. The initial borrower underwriting and collateral valuation attempts to mitigate as much risk as possible. The remaining risk is diversified through the "law of large numbers,", that is, it is “insured,� either internally or by a third party. Lenders charge a premium embedded in the mortgage rate sufficient to cover the cost of this credit insurance. The risk of default is typically retained by the mortgage originator or is insured by a mortgage insurance company. Mortgage insurers are involved in the loan underwriting at origination, but leave the loan servicing with the originator. Insurers typically retain the right to take over the servicing in the event of default, paying the investor and conducting the foreclosure, eviction, and sale. Liquidity Risk. Mortgages have long maturities, whereas deposit liabilities have a much shorter term. Hence the possibility exists that depositors will demand their funds at a time mortgage borrowings are not yet fully repaid. Commercial banks have a base of "core" deposits that allows a percentage of assets to be safely invested in adjustable rate home mortgage loans without raising liquidity concerns. The safe amount may vary from as low as 10 percent of assets to 50 percent or more, depending on the volatility in financial markets and inflation. Typically, this safe amount is constrained to be invested in a liquid asset, such as government securities. Some caveats need be considered, however. First, Treasury paper with longer maturities trades at less-than-par value. Second, secondary trading is often minimal and can be 48 nonexistent in the event of a systemic crisis. Hence, the only protection against a systemic liquidity crisis is the ability and willingness of the central bank to discount these securities. Liquidity risk is small so long as prices remain relatively stable and financial markets remain liquid. A liquidity crisis (such as that in the United States in the 1930s) can be prevented by a centralized liquidity facility. Central bank “lender-of-last-resort� discount facilities perform the same function, and currently, they are the preferred mechanisms of use. Interest rate swaps match the needs of deposit and institutional investors when deposit lenders face an excess demand for long duration loans, and institutional investors face an excess supply of long duration liabilities, but neither faces an imbalance between sources and uses of funds. The mortgage lender pays the short rate and earns the long rate, and the counterparty pays and earns the opposite. The market for swaps started in the 1980s and is now international in scope and is valued at more than a trillion dollars. These swaps can be directly negotiated between any two counterparties, although secondary markets now exist as well. Interest Rate Risk. The lender bears interest rate risk that the rate charged on the loan might not sufficient to cover the cost of borrowings plus the cost of operations. The borrower’s risk is that the monthly payment may rise relative to income. Literally hundreds of mortgage designs have been tried in the marketplace to mediate these concerns. Which works best largely depends on the source of the problem. If short-term interest rates are volatile but revert to a mean, reflecting financial market instability, then lenders can best absorb the volatility by offering borrowers a fixed rate loan. But if interest rates rise in response to a secular shift in inflation, which ultimately will be reflected in borrower incomes, then borrowers can better absorb the change in rates than lenders by accepting an adjustable rate loan. Of course, markets may experience both risks over a 15- to 30-year loan maturity. Hence most borrowers have revealed a preference for fixed rate loans in many countries. Political Attributes to Home Mortgage Lending. Political risk refers to the potential for political decisions to affect the risk of and return to mortgage lending. Governments influence the treasury yield curve through regulations favoring government securities and state ownership of financial companies. They may influence deposit rates or mortgage rates for political reasons. They may impose ceilings on deposit rates or mortgage rates in an attempt to control costs. State ownership of intermediaries tends to stifle the competition and create an uneven playing field. This is a particular problem for countries that liberalize the financial system while maintaining their state-owned credit allocation mechanisms from the previous system. These entities are not subject to market discipline, and they may distort market pricing and financial system stability. Taxes play a significant role in the development of the financial system, often distorting markets. 49 Annex D: Secondary Mortgage Markets, Mortgage Capital Markets, and Securitization The secondary market bridges the gap between a primary lender and a secondary source of funds, that is, usually an entity without retail branches that cannot originate mortgage loans. Secondary buyers most often do not want to originate or service the loan, preferring to leave that to the primary originator. As a consequence, wholesale funders also almost never take the credit risk of the loan. This is especially true when the mortgage loan is originated on borrower income, that is, is "personal," and the originator is expected to be in a better position to evaluate and monitor the borrower. Third-party mortgage insurers are often used, especially when a collateral-backed loan is used. The generic economic impetus to secondary mortgage markets is provided by either a funding or a duration mismatch. The term "mortgage capital market" evolved in the United States in the 1980s. Capital markets typically refer to a place where securities are underwritten and traded, as opposed to secondary markets that incorporate a wide variety of relationships and generally to not imply ongoing trading activity. Capital market security underwriting has high fixed costs, requiring a large volume of homogeneous securities to trade, and a deep pool of investors to trade among themselves. A secondary market trade may take no more than a phone call to arrange the sale of an individual mortgage loan. Almost all nongovernment bonds were privately placed in the United States during the first hundred years of capital market development and were never traded. Mortgage-backed securities (MBS) allow the assets to be "sold" to the investor (or to a special purpose vehicle [SPV] in which the investor owns shares), while the originator retains the servicing. The major advantage to lenders of issuing a pure pass-through security is that the assets come off their books, as if sold. But this requires that all the risk be transferred to the investor or a third party, typically an insurer. This technique is typically employed by high-volume mortgage bankers in the United States who may originate US$10 billion or more a year in new loans, and who have no balance sheet on which to fund these loans. Virtually all the loans have either government or private mortgage insurance. Mortgage lenders do not typically issue private MBS, because it is extremely difficult to transfer the credit risk and get sales treatment for the underlying securities. Around the world, nongovernment MBS are not widely traded, because doing so requires high volume and total loan homogeneity and standardization. Even in the United States, with more than a trillion dollars in government-backed MBS outstanding, the benefit to active MBS trading has diminished. Many other countries have passed legislation allowing the U.S. model to be mimicked, but few MBS have been issued to date. Hence, in the absence of huge economies, there is generally little benefit to securitized capital markets over the simpler secondary mortgage market approach. The policy choice of relying on secondary mortgage markets or even capital markets versus retail deposit intermediaries entails several considerations. First, secondary markets have a direct implication of increasing mortgage market growth rates. Second, when mortgages are backed by housing collateral and hence are more homogeneous—that is, more of a “commodity� and deep pools of long-term savings exist—secondary markets become feasible. Third, active trading of pass-through mortgage securities in capital markets developed only in the United States, and to a lesser extent in Europe, informing the policy choice between a secondary private placement market versus full-blown securitization.a A concern was widely expressed in Sri Lanka that the yield curve for government securities is distorted because of the pervasive presence of the issues and the central bank activities. Some countries look to an active secondary market in MBS as an alternative to the Treasury yield curve, although this report disagrees that mortgage securitization can credibly substitute for a robust Treasury securities yield curve. Other considerations in favor of securitization, should it be feasible, include capital market and trading stimulation, and the promotion of the development of a slew of technical financial intermediaries, including underwriters, rating agencies, trustees, and so on. A significant consideration involves the trade-off between the costs of operationalizing each mechanism, as borne by the market and the budget. The strict and heavy set of regulations necessitated by a well-functioning secondary mortgage capital market may stifle nascent market development. Requirements for registration and ratings may add substantial expense with little economic value. 50 Regulations requiring diversification and reducing exposure are problematic in the tiny domestic financial market. The ongoing liberalization, with ambitions relating to diversification and marketability of assets, might consider shying away from a capital market focus more appropriate to huge actively traded fixed income capital markets. The handful of mortgage lenders and potential institutional investors may function better in a framework that allows them to negotiate prudent and simple private transactions, unencumbered by underwriting costs, trustee fees, rating costs, multiple taxes, and so on. A second concern is that such regulations may well result in a situation of “regulatory arbitrage,� in which transactions are structured not to reflect the underlying economic benefits to issuer and investor, but to get around the intention and spirit of regulations, particularly those for capital requirements. The most logical explanation of the Housing Development Finance Corporation transaction is a failed attempt by the issuer to avoid holding capital against the credit risk, encouraged by policy makers out of a desire to facilitate secondary trading. Existing regulations should be carefully reexamined to ensure that there are no loopholes or incentives, including references to third-party ratings in the current regulations that govern regulated financial institutions, which allow the total capital required to back mortgage credit risk to be reduced by investment bankers using seemingly complicated SPV techniques. a. Even in the United States, where annual mortgage originations now exceed US$10 trillion and most mortgages are insured, the trend has been away from actively traded mortgage-backed securities. 51 Annex E: Successful Housing Finance Development—Examples from Malaysia, Korea, and India The experience in Malaysia, the Republic of Korea, and India warrant deeper study, which may come of use when detailed housing finance policies are being formulated and implemented in Sri Lanka. Malaysia By many criteria, Malaysia is an example of a very successful integration of housing finance into the broader financial system. Malaysia’s housing finance infrastructure is sound and well developed, which is reflected in a high ratio of outstanding housing finance assets to gross domestic product (GDP) of more than 20 percent. Four features of the Malaysian housing system are worth noting:  A solid legal infrastructure with associated registries that supports secured lending, permits the timely enforcement of mortgage foreclosures, and therefore permits an effective management of retail credit risks.  A secondary mortgage market institution, Cagamas Berhad (Malaysian Mortgage Corporation), which provides liquidity to the housing finance system, promotes lending to low-income groups, and has fueled the secondary mortgage market for the past 15 years.  A fully functional set of land development, real estate development, and construction codes that has permitted the evolution of a deep real estate market infrastructure and a skilled real estate industry. The Malaysian National Valuation Institute (INSPEN) continuously monitors real estate prices for a variety of purposes, including real estate risk monitoring of financial assets.  The pragmatic strategy followed by Bank Negara Malaysia toward the development of then- untested Islamic housing finance instruments and of specialized Islamic finance institutions, with their associated regulatory and supervisory framework. The Malaysian experience reflects converging world trends in evidence across a growing number of financial systems. First and foremost, Malaysia has built strong financial sector foundations in terms of macroeconomic management, accounting practices, and banking regulation and supervision of international standards. In the late 1970s, Malaysia was among the first developing countries to move away from directed credit policies, regulated interest rates, as well as specialized state-owned housing lenders. This permitted the proliferation of a broad range of retail mortgage lenders competing in a growing mortgage market, and eliminated market segmentation. The market share of the specialists housing credit institutions, which were the dominant lenders until the 1970s in Malaysia, had shrunk to less than 3.5 percent of outstanding housing loans by the end of the 1990s. The commercial banks hold 52.5 percent of the mortgage market. Finance companies make about 13 percent of outstanding loans. The share of Islamic finance from Bank Rakiat was less than 2 percent in the late 1990s. Outstanding social loans at preferential rates from the Treasury Housing Loan Division remained quite important and represented 28 percent of the total system.a To permit these competing retail financial institutions to manage their funding risks and to develop the bond market at the same time, the Mortgage Corporation of Malaysia (commonly known as Cagamas) was created in 1987. A decade later, Cagamas securities represented 40 percent of the domestic bond market.b Malaysia has been developing Islamic housing finance services since 1983, in parallel to standard housing finance products. This has permitted the Malaysian housing finance industry to draw on existing international legal and financial experience, skills, and information technology systems as needed––and to develop the deepest housing finance market in developing countries. The market response has been interesting because at the end of the 1990s, standard financial products still represented 98 percent of outstanding assets, and Islamic products about 2 percent of this very deep housing finance market. 52 Republic of Korea The Republic of Korea is a different Asian case of housing finance reforms. Important milestones have been the progression from a segregated, state-controlled housing finance system in the 1970s to an open and competitive housing finance system with the privatization of the Korea Housing Bank in 1997, its new charter as a commercial bank in September 2000 on the New York Stock Exchange, to the eventual Korea H&CB merger with Kookmin Bank in the fall of 2001 to become the largest commercial bank in Korea. One of the policy decisions that made the successful growth privatization of KHB possible was the separation of subsidized social housing from the commercial balance sheet of KHB in the early 1980s. Social loans were separately funded by the government and managed by KHB for a fee. The transition toward a competitive retail housing finance market in this new decade is likely to be enhanced by the creation of Komoco (The Korea Mortgage Corporation) as a secondary mortgage market institution. India India’s housing finance system has accumulated significant experience in managing credit risk both in terms of collateral substitutes and complements to mortgages when loans are made, and in terms of foreclosure alternatives in case of default. India’s Housing Development Finance Corporation (HDFC, unrelated to the same name institution in Sri Lanka) has applied its rich experience and a variety of techniques to successfully control its retail credit risks, both in its initial takeoff phase and its current predominant role as market leader. HDFC has developed a successful corporate culture and profitable operations that are driven by a strong consumer focus, in sharply contrast with Indian state-owned banks. At a time when no housing finance system existed in India, a critical contribution of HDFC was to demonstrate that successful and profitable commercial mortgage lending was possible. In the process, HDFC contributed to changing the backward-looking and puzzling conventional wisdom of many local macroeconomists about the housing sector being mere “consumption.� Sources:: Maimbo, Samuel, Olivier Hassler, and Isfandyar Khan, “Housing Finance Reforms in Pakistan: A Strategy for Strengthening the Real Estate Development Process,� World Bank, Washington, DC, June 2006. Foreign Investment Advisory Service. 2005. “Pakistan: Improving the Performance of the Housing Tourism, and Retail Sectors.� August. International Finance Corporation, Washington, DC. Chiquier, Loic, Olivier Hassler, and Michael Lea. 2004. “Mortgage Securities in Emerging Markets.� Policy Research Working Paper No. 3370, August. World Bank, Washington, DC. Maimbo, Samuel, Olivier Hassler, and Richard Green. 2004. “Housing Finance Reforms in Bangladesh: Unleashing Supply to meet Growing Demand.� November. World Bank, Washington, DC. Bank for International Settlements. 2006. “Housing Finance in the Global Financial Market.� Committee on the Global Financial System Papers No. 26, January. Basel, Switzerland. Hoek-Smit, Maria. 2005. “The Housing Finance Sector in Indonesia.� World Bank, Washington, DC. a. See, BANK NEGARA MALAYSIA [1999], "Housing Credit Institutions", in The Central Bank and the Financial System in Malaysia, in celebration of the 40th Anniversary of BNM 1959-1999. b. A case study of Cagamas is posted in the housing finance section of the financial sector Web site of the World Bank, http://www1.worldbank.org/finance/. 53 Annex F: International Best Practice in Housing Policy A sound housing policy in a rapidly urbanizing economy includes seven main components: three on the demand side, three on the supply side, and an institutional framework to deliver these policies. To strengthen economic housing demand—as opposed to formulaic and merely notional planning “needs�—the key areas for actions are as follows: 1. Improve property rights: bring cadastres up to date, ensure effective property titling, and provide prompt and fair procedures for conflict resolution of property disputes; 2. Develop market priced mortgage credit: move away from specialized housing vehicles (especially state- owned), remove obstacles to commercial lending for housing, and link up to secondary markets; and 3. Implement a system of well-targeted subsidies: identify groups in need of support, finance from the budget, separate credit from subsidies, link these subsidies directly to beneficiaries, and avoid subsidizing inefficiency and stagnation in privileged special institutions. Most of the policy levers to strengthen housing demand are at the central government level. In contrast, local governments play a critical operational role in shaping housing supply. The legal framework and the “rules of the game� may be written at the national level, but housing production is always local in nature.a To lower the cost of housing and make housing supply responsive to demand, the three main policy areas are as follows: 4. Expand the supply of serviced urban land by clarifying and improving the division of labor between public and private sector operators; 5. Provide the regulatory framework for land development and real estate by means of clear and predictable land development rules, consistent construction and real estate codes, and sound and stable national and local taxation; and 6. Promote the emergence of competitive construction and real estate industries by upgrading codes, taxation, public procurement and tendering, contracting, technology, and so on. In terms of institutional framework, there is a consistent pattern of decentralization of powers in the three key areas of urban policy: planning, housing, and finance: 7. Develop a transparent and stable institutional framework for national housing policy:  at the national level, provide a stable legal, regulatory, and enabling framework;  at the local level, support autonomous and innovative local governments to ensure the implementation of good housing policies; and  at both levels, develop timely and reliable real estate information systems. Annex figure F1 outlines the basic components of a well-developed housing finance system. Annex table F1 presents lessons learned from international experience with housing finance reforms, “the Do’s and Don’ts of Housing Finance.� Source: “Housing: Enabling Markets to Work,� World Bank, Washington, DC, 1993. a. Local governments play a key role in the day-to-day performance of private real estate markets. The worst- functioning cities around the world are those where the effective land supply drops to only a fraction of total potential supply and serviced land is scarce. Various reasons for this include the following: land tilting systems may be poor and the registration of property transfers costly; large land areas may be under the control of a few traditional landowners or may have been nationalized and be under the control of agencies that prove unable or unwilling to coordinate investment or regulatory decisions with local governments. In such cases, high housing and real estate prices prevail, and housing crowding is particularly severe for the low-income groups. 54 Annex Figure F1 World Trends in Housing Finance C:/PAK/World Trends Chart BR 011219 WORLD TRENDS IN HOUSING FINANCE 1. TREASURY Government Grants 2. CENTRAL BANK Annual Budget Funding INTERNATIONAL FINANCE INSTITUTIONS (WB, Others) Initial share equity with private buyback arrangements Initial long-term loans SOCIAL HOUSING Securities CAPITAL MARKETS SOCIAL HOUSING SECONDARY CAPITAL MARKETS FUND Pension funds - - Pension funds FUND MORTGAGE Life insurance companies - - Life insurance companies FACILITY Funds Mutual funds - - Mutual funds (Various versions) 0ther long-term investors - - 0ther long-term investors Upfront Mortgage loan Loans and advances Subsidies collateral Loan purchases Equity Securitization CONTRACTUAL SAVINGS: CONTRACTUAL SAVINGS: Various forms Various forms - - mandatory private pensions participation mandatory private pensions of funding for of funding for - - voluntary private pensions voluntary private pensions social housing social housing and insurance Primary Market Lenders: and insurance Mutual/private 1. RETAIL DEPOSIT BANKS Social housing mortgage insurance 2. FINANCE COMPANIES Private developers Developers All - Co-signature/salary Households - Mortgage lien Mortgage Loans - Mortgage insurance LOW INCOME HOUSEHOLDS WITH ALL OTHER HOUSEHOLDS HOUSEHOLDS TARGETED SUBSIDIES Source: “Housing: Enabling Markets to Work,� World Bank, Washington, DC, 1993. Annex Table F1. The Do’s and Don’ts of Housing Finance Instrument Do Don’t Developing property √ Regularize land tenure x Engage in mass evictions rights √ Expand land registration x Institute costly titling systems √ Privatize public housing stock x Nationalize land √ Establish property taxation x Discourage land transactions Developing mortgage √ Allow private sector to lend x Allow interest rate subsidies finance √ Lend at positive/market rates x Discriminate against rental housing √ Enforce foreclosure laws investment √ Ensure prudential regulation x Neglect resource mobilization √ Introduce better loan product Allow high default rates Rationalizing subsidies √ Make subsidies transparent x Build subsidized public housing √ Target subsidies to the poor x Allow for hidden subsidies √ Subsidize people, not houses x Let subsidies distort prices √ Subject subsidies to review x Use rent control as a subsidy Providing √ Coordinate land development x Allow bias against infrastructure infrastructure √ Emphasize cost recovery investments √ Base provision on demand x Use environmental concerns as reason √ Improve slum infrastructure for slum clearance Regulating land and √ Reduce regulatory complexity x Impose unaffordable standards housing development √ Assess costs of regulation x Maintain unenforceable rules √ Remove price distortions x Design projects without link to √ Remove artificial shortages regulatory/institutional reform Organizing the √ Eliminate monopoly practices x Allow long permit delays building industry √ Encourage small-firm entry x Institute regulations inhibiting √ Reduce import controls x competition √ Support building research Continue public monopolies Developing a policy √ Balance public/private sector roles x Engage in direct public housing and institutional √ Create a forum for managing the x delivery framework housing sector as a whole x Neglect local government role √ Develop enabling strategies Retain financially unsustainable √ Monitor sector performance institutions Source: “Housing: Enabling Markets to Work,� World Bank, Washington, DC, 1993. Annex G: Longer-Term Focus Areas for Policy Makers Further improving the legal and regulatory environment for mortgage markets  Efficient registration and transfer of title deeds is an ongoing policy goal, and the current cadastral survey and land registry work should be pressed on. Records and filing procedures should be computerized in registry offices. Consider nonjudicial options for property dispute resolution.  Improve the realization of collateral security and the eviction process. The use of parate powers should be extended to all loans (not just large ones) and all housing finance lenders to ensure a level playing field between institutions and fair competition. Eliminate monetary board approvals of collateral sale price.  Review sale taxes on financial mortgage-related instruments to avoid discouraging the development of a secondary market.  Create Property and House Price Indexes, following up on the initial preparatory work ongoing at CBSL.  Further improve the functioning of the Credit Information Bureau to protect consumers and avoid lengthy disputes in court involving finance companies. The Asian Development Bank Financial Sector Assessment suggests CRIB privatization.  Facilitate further the environment for multifamily development, including titling issues with buyer preconstruction financing and a professional property management regulatory environment.  Foster affordable land use and enabling urban development rules. Improve urban land use. Examine land development rules for excessive restrictiveness. As India repealed its restricting act on land development, prices went down, and the middle-income market for housing attained a fast pace of development in response. Policy makers should consider the following:  The introduction of a less distorting tax in place of the property sale stamp tax.  A government promotion or sponsorship of risk-sharing to make title insurance more universally available as an interim "second-best" solution to the current lack of a government title registry evidencing unambiguous ownership.  A nationwide savings-for-housing scheme. Strengthening the enabling role of the state  Commercialization and privatization can help government banks improve efficiency and performance. Special privileges and funding harm commercialization, whereas separating good bank and bad bank helps.  Credit-linked subsidies to borrowers can be designed to be nondistortionary. Some examples include the following:  Subsidies on savings for down-payment: an upfront subsidy, avoiding subsidized savings and lending rates  Pledge default account: a required deposit to be used in case of late payment  Buy-down-mortgage: initial subsidy on part of the monthly payment on a market rate loan  Upfront grants/allowances as part of a housing financing package, toward deposit or closing costs or the loan  Payment for mortgage insurance or guarantee premiums to ensure the top part of the loan or other risks  Payments for community mobilization and household counseling programs  Appropriate subsidy for each country would depend on expected inflation, growth, and income distribution, expected house value movements, prevailing borrower constraints, and the budgetary requirements of the state  The government can provide an enabling environment for housing finance by ensuring regulatory, institutional, or financial support to decrease risks and costs of down-market lending. One group of such initiatives, focusing on support of credit risk management, includes the facilitation of an effective credit bureau, improvement of property registration; establishment of real estate information systems; investment in infrastructure/services; improvement of foreclosure and eviction 57 laws; a decrease in political risk, and others. Governments can further help the development of housing finance by instituting various mechanisms for interest rate risk reduction. Other government actions in support of an enabling mortgage market environment center around improved liquidity position of down-market lenders. Specifically, actions here include allowing mutual and microfinancial institutions to take deposits, improving the microfinance institutions regulatory system and providing capacity building (for example, the South African success case); removing legal constraints for lending by institutional investors to housing finance institutions; and providing secondary market legal infrastructure.  Distorting taxes and regulations should be altered to remove market distortions that inhibit financial market deepening.  A review of tax and regulatory policy regarding the financial sector, especially those affecting the development of a secondary market for mortgage finance, should be conducted from the perspective that every approach is legal and encouraged unless specifically prohibited.  Liberalize further the portfolio composition of key long-term investments, to lower their dependency on Treasury paper.  Sharpen and crystallize the mission and direction of each implementing institution for the government housing policy, in view of changing market conditions. Refocusing attention on the lowest-income population is one alternative, another is a goal of commercial viability and profitability. These alternative objectives are mutually exclusive.  Improve governance of housing policy implementing institutions, depoliticize boards of directors, and make management accountable. Improve internal controls.  Improve management and efficiency, upgrade human resources, and modernize operations.  Invest in the further modernization of the state-owned housing banks, including computerization, full branch networking, and electronic transactions.  Improve operational efficiency, market discipline, operating costs, credit loss indicators, and risk management. Encouraging mortgage market deepening  Policies should facilitate the establishment of a competitive and competing mortgage lending market. This in turn will encourage the following:  The development of new credit products (term, rates, amortization) to better suit client needs  Training and innovation with a view of adoption of better risk management practices  Broader distribution channels  More transparency and comparability of products, and better consumer ability to make informed housing finance choices  The issue of home financing for lower-income groups should be addressed. The general strategy, as defined by the government, is to focus on home upkeep as opposed to new construction. In this case, it becomes particularly relevant to enable and encourage microlending for house financing purposes. A further initiative of importance is to continue working on cheaper housing technologies to produce low-cost affordable housing. 58 References ADB (Asian Development Bank). 2005. “Sri Lanka: Financial Sector Assessment.� October. Manila, Philippines. Bank for International Settlements. 2006. “Housing Finance in the Global Financial Market.� Committee on the Global Financial System Papers No. 26, January. Basel, Switzerland. Bank Negara Malaysia. 1999. "Housing Credit Institutions." In The Central Bank and the Financial System in Malaysia. In celebration of the 40th Anniversary of BNM 1959–1999. Kuala Lumpur, Malaysia. “Cadastral Template.� 2007. Department of Geomatics of the University of Melbourne. http://www.geo21.ch/cadastraltemplate/index.htm. Melbourne, Australia. Cagamas Web site. 2007. www.cagamas.com.my. Kuala Lumpur, Malaysia. Calhoun, Charles. 2005. “Housing Finance Assessment for Thailand.� December. World Bank: Washington, DC. Central Bank of Sri Lanka. 2006a. “Financial System Stability Review.� Colombo, Sri Lanka. Central Bank of Sri Lanka. 2006b. “Annual Report.� Colombo, Sri Lanka. Central Bank of Sri Lanka. 2007. “Road Map for Monetary and Financial Sector Policies in 2007 and Beyond.� Colombo, Sri Lanka. Chiquier, Loic, Olivier Hassler, and Michael Lea. 2004. “Mortgage Securities in Emerging Markets.� Policy Research Working Paper No. 3370, August. World Bank, Washington, DC. Department of Census and Statistics. 2003. “Census of Population and Housing–2001: Population and Housing Data: Various Districts.� May. Colombo, Sri Lanka. Department of Census and Statistics. 2007. “National Accounts.� Colombo, Sri Lanka. Doing Business Report. 2007. International Finance Corporation, Washington, DC. Foreign Investment Advisory Service. 2005. “Pakistan: Improving the Performance of the Housing Tourism, and Retail Sectors.� August. International Finance Corporation, Washington, DC. Government of Sri Lanka. 2005. “Sri Lanka New Development Strategy: Framework for Economic Growth and Poverty Reduction.� Colombo, Sri Lanka. Hoek-Smit, Maria. 2005. “The Housing Finance Sector in Indonesia.� World Bank, Washington, DC. Kusmiarso, Bambang. 2006. “Housing and Mortgage Markets in the SACEAN Countries.� SACEAN Research and Training Centre, Malaysia. Maimbo, Samuel, Olivier Hassler, and Richard Green. 2004. “Housing Finance Reforms in Bangladesh: Unleashing Supply to meet Growing Demand.� November. World Bank, Washington, DC. Maimbo, Samuel, Olivier Hassler, and Isfandyar Khan. 2006. “Housing Finance Reforms in Pakistan: A Strategy for Strengthening the Real Estate Development Process.� June. World Bank, Washington, DC. Maxwell tamp PLC 2005. Report on EPF/ETF, February. London, UK. 59 National Development Strategy. 2006. Mahinda Chintana: Vision for a New Sri Lanka. November. Colombo, Sri Lanka. Renaud, Bertrand. 2007. “Mortgage Finance in Emerging Markets: Constraints on Feasible Development Paths.� Renaud Advisors, McLean, VA. ShoreBank International. 2006. “Afghanistan Housing Sector Assessment.� December. Colombo, Sri Lanka. Task Force for Reconstruction of the Nation (TARFEN). 2005. “Rebuilding Sri Lanka, Post Tsunami Recovery and Reconstruction Strategy.� May. Colombo, Sri Lanka. United Nations. 2004. “Human Settlements: Policy Options and Possible Actions to Expedite Implementation.� December. Economic and Social Council, Commission on Sustainable Development, United Nation, New York. Villani, Kevin. 2007. “Mortgage Markets in Emerging Economies: the Use of State Sponsored Insurance and Guarantees.� April. World Bank. 1993. “Housing: Enabling Markets to Work.� World Bank, Washington, DC. World Bank. 2004. “Development Policy Review." December. World Bank, Washington, DC. World Bank. 2005a. “Sri Lanka Development Forum: The economy, the Tsunami and Poverty Reduction.� Report No. 32221-LK. South Asia Department, World Bank, Washington, DC. World Bank. 2005b. “Bangladesh Housing Report.� South Asia Private and Financial Sector Report, World Bank, Washington, DC. World Bank. 2006a. “Poland Housing Finance Policy Note.� World Bank, Washington, DC. World Bank. 2006b. “Sri Lanka: Improving Access to Financial Services: Selected Issues.� World Bank, Washington, DC. World Bank. 2007a. World Development Indicators. Washington, DC: World Bank. World Bank. 2007b. “North East Housing Reconstruction Project.� Selected project documentation. World Bank, Washington, DC. World Bank. 2007c. “Puttalam Housing Project.� Selected project documentation. World Bank, Washington, DC. World Bank. 2007d. “Tsunami Emergency Recovery Program.� Selected project documentation. World Bank, Washington, DC. World Bank 20087 (forthcoming), “Regional Bond Markets Study�, World Bank, Washington. DC. 60 Acronyms and Abbreviations ADB Asian Development Bank APPFs Approved Private Provident Funds ATM automated teller machine AWOR Average Weighted Offer Rate BOC Bank of Ceylon BOI Board of Investment CBSL Central Bank of Sri Lanka CCPI Colombo Consumers Price Index CRIB Credit Information Bureau DFCC Development Finance Credit Corporation DTI debt service-to-income ratio EPF Employee Provident Fund ETF Employee Trust Fund FDI foreign direct investment GDP gross domestic product HDFC Housing Development Finance Corporation IDA International Development Agency LCBs licensed commercial banks LSBs licensed specialized banks MBS mortgage-backed securities MDI mortgage default insurance MSMEs Micro, Small, and Medium Enterprises NBRO National Building Research Organization NDB National Development Bank NEHRP North East Housing Reconstruction Project NHDA National Housing Development Authority NPLs nonperforming loans NSB National Savings Bank OECD Organisation for Economic Co-operation and Development PB People’s Bank PTI payment-to-income ratio SEEDS Savordaya Enterprise Development Society SEMA Strategic Enterprises Management Agency SLIBOR Sri Lanka Inter Bank Offer Rate SMEs Small and Medium Enterprises SMIB State Mortgage and Investment Bank SOEs state-owned enterprises SPV special purpose vehicle UDA Urban Development Authority WPLR Weighted Prime Lending Rate Currency SL Rs Sri Lankan rupees US$ U.S. dollar 61 Notes The team gratefully acknowledges the invaluable comments and advice of Loïc Chiquier, Sriyani Hulugalle, Narayanasamy Kokularupan, Samuel Munzele Maimbo, and Peter Ellis, and the continued support of Prameela Namasivayam and Sashikala Krishani Jay under the leadership of Simon Bell. 1 2005 estimates. 2 As stated in the 2004 Economic Policy Framework, “Creating our Future, Building our Nation 2004,� and further articulated in the 2005 budget, the new development strategy is premised on pro-poor, growth income improvement and redistribution policies with complementary participation of a socially responsible private sector and a strong public sector. 3 The National Housing Act 37/1954 provided the basic legal framework for housing development. 4 Current borrower income averages SL Rs 15,000–20,000 per month or more. 5 Authors’ rough estimates are calculated by counting the number of mortgage loans disbursed in the economy (between SL Rs 600,000 and 900,000) and taking into account the excess liquidity of banks and elastic mortgage supply to eligible customers in the past years. The data on number of mortgage loans are taken from Kusmiarso (2006) and CBSL 2007. 6 As a low-cost housing price benchmark, we use the actual technology applied in Sri Lanka post-tsunami by the nongovernmental organization Practical Action: “A house with one bedroom, living room, kitchen and toilet with an extent of 350 square feet can be built with SL Rs 175,000� (see http://practicalaction.org/?id=tsunami_housing). 7 This crude estimate is based on a starting figure of SL Rs 600,000–900,000 existing mortgages and a 10 percent growth rate in mortgage lending; it also takes into account population growth and rising housing needs. 8 The figure is calculated as the estimated difference between the number of families and the number of separate dwellings. 9 At US$1 a day the ratio is low at 5.6 percent. 10 As a low-cost housing price benchmark, we use the actual technology applied in Sri Lanka post-tsunami by the nongovernmental organization Practical Action: “A house with one bedroom, living room, kitchen and toilet with an extent of 350 square feet can be built with SL Rs 175,000� (see http://practicalaction.org/?id=tsunami_housing). 11 BOC and PB have increased their operational efficiency under the measures taken by Strategic Enterprises Management Agency (SEMA). Periodic reviews of the banks have indicated performance improvements, as evidenced by BOC’s better capital adequacy and NPL ratios, as well as its AAlka credit rating from Fitch Ratings. PB’s net worth has turned positive and its capital adequacy ratio has improved. 12 Ranking behind BOC, but with more deposits than PB. 13 One of the oldest banks in Sri Lanka, SMIB has its origins as the Ceylon State Mortgage Bank inaugurated in 1931; this was eventually amalgamated with the Agricultural and Industrial Credit Corporation in 1979 to arrive at its present structure. 14 HDFC was established originally in 1984 as a Building Society, which is 78 percent owned by the NHDA and 12 percent owned by the ETF. It was licensed as a specialized bank in 2003. 15 The government likely would capture a considerable share of financial system assets regardless of the rate paid, as it requires most financial intermediaries to hold Treasury paper. 16 The EPF now maintains an estimated 11.3 million member accounts comprising 2 million contributing accounts and 9.1 million noncontributing accounts, and SL Rs 31.7 billion in contributions (EPF 2006). The investment portfolio of the Fund stood at SL Rs 451 billion (16 percent of GDP) and accounts for more than 70 percent of total superannuation assets. ETF was created in 1981 to promote stock ownership among employees and is financed by a 3 percent employer contribution. ETF’s total assets are SL Rs 65.7 billion (2.3 percent of GDP). 17 A February 2005 report by Maxwell Stamp noted that the combined budgets for EPF and ETF showed a ratio of outflows to contributions of 105 percent by 2004, up from 64 percent in 2000, but this appears to be an anomaly. 18 The balance is mainly invested in listed and unlisted equities. Other assets included bank deposits and bank loans. 62 19 A number of noncontributory pension schemes are operated by nongovernment public and private enterprises such as banks, insurance companies, utilities, and the larger nonfinancial corporations. The providers of these schemes are required to set aside reserves to meet their liabilities. Information on the size, types of available benefits, and asset allocation policies of these schemes is not readily available. It is well known, however, that although they are not restricted in shaping their investment policies, nor are they subject to government direction, they too favor investing in fixed income instruments. 20 A limited-purpose entity serves as a pass-through conduit in creating securities backed by mortgages. 21 Limited competition exists despite the multiplicity of financial institutions in the market, because the market remains dominated (by volume) by few weak players who dictate the terms. 22 Apartment Ownership Law Act No. 11 of 1973 and amendments. 23 State Banks such as BOC, PB, and DFCC have been established by separate Acts of Parliament, which provide, among other rights, the right to “parate powers.� Specifically, current beneficiaries of parate powers include the BOC, the PB, the banks established under the provisions of the Regional Rural Development Bank Act No. 15 of 1985, NSB Credit Corporation, and commercial banks holding a license under the Banks Act No. 30 of 1988. 24 Parate allows the lender bank board to pass a resolution to sell by public auction property mortgaged to the bank to recover sums loaned and in default. It allows a bank to appoint a manager to take charge of the property mortgaged and manage it and to sell the produce, recover rent, and other profits. 25 Data as of end 2006. In general, parate executions are extremely rare. The NSB, with a mortgage portfolio of more than SL Rs 6 billion, reports having sold only about 10 properties last year. 26 The lender registers his or her action at the Land Registry and court issues notice and summons all concerned. Following a trial, judgment can be appealed at the Court of Appeal (as well as the Supreme Court on questions of law only). The count can issue an order to sell pending borrower appeal. 27 Costs are specified in Costs (Regulations) 1997 made under Section 214 of the Civil Procedure Code. 28 On gifts, first SL Rs 50,000 at 4 percent, 3 percent thereafter; on transfers, first SL Rs 100,000 at 5 percent, 4 percent thereafter. 29 CCPI values for the recent past are provided in annex A1. The rent component of CCPI has not been updated in the recent past. 30 The survey used direct information from a sample of 427 respondent buyers identified through title deeds of land sales in 2003, 2004, and 2005 available in a Colombo District Land Registry of the Registrar General’s Department. 31 “Haircut� relates to the amount by which an advance is below the nominal value of the collateral. 32 The BOC, PB, NDB, NSB, DFCC, and the SMIB. CBSL is contemplating a reduction in its share. 33 Operational risks are those faced by management in implementing the mortgage lending program. Annex C provides a discussion of risk basics and terminology. 34 These are the rules applied by CBSL to borrowing banks at the discount window (i.e. directly from CBSL). 35The growth of the number of mortgages issued is estimated at 3 percent annually. 36MDI relieves mortgage originators of the risk of default. MDI helps widen the household access to housing finance by isolating the credit risk and encouraging private sector expansion into lower-income mortgage lending. It is a key ingredient to the development of secondary markets. See also Villani (2007). 63