55543 Expanding Housing Finance to the Underserved in South Asia Expanding Housing Finance to the Underserved in South Asia Market Review and Forward Agenda Tatiana Nenova Washington, DC © 2010 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 13 12 11 10 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8322-3 eISBN: 978-0-8213-8323-0 DOI: 10.1596/978-0-8213-8322-3 Cover photos: Hye Sakm Abdul and Lohita Karunasekera Cover design: Quantum Think, Inc. Library of Congress Cataloging-in-Publication Data Nenova, Tatiana. Expanding housing finance to the underserved in South Asia : market review and forward agenda / Tatiana Nenova. p. cm. Includes bibliographical references and index. ISBN 978-0-8213-8322-3 -- ISBN 978-0-8213-8323-0 (electronic) 1. Low-income housing--South Asia--Finance. 2. Housing policy--South Asia. I. Title. HD7359.3.A4N46 2011 363.5'820954--dc22 2010013120 Contents Foreword xiii Acknowledgments xvii Abbreviations xix 1 Introduction 1 Growth Potential of Housing Finance Markets 3 Summary of Key Challenges 5 Objectives of the Report 6 Scope of the Report 6 2 Housing Demand 9 3 Housing Supply 13 Current State of Housing 13 Development of Housing Stock 14 Construction Quality and Enforcement of Construction Standards 16 Housing Costs 16 Housing Price Index 17 Secondary Property Markets 17 v Contents 4 Housing Finance Institutions and Instruments 19 Traditional Mortgage Lending 20 State-Owned Specialized Housing Banks 22 Private Specialized Mortgage Lenders 24 Islamic Housing Finance 25 Large-Scale Project Finance 26 Informal Housing Finance 26 Homeowners' Insurance 27 Risk Management 27 Secondary Market for Mortgage Finance 29 Prudential Regulation and Oversight 30 A Word on the Global Financial Crisis 32 5 Low-Income Housing Finance 33 The Power of Microfinance 33 Affordability of Housing 34 Low-Cost Market Solutions 34 6 Regulatory Framework for Housing and Housing Finance 37 Land Administration 37 Foreclosure 43 Credit Information 45 7 Policy Options to Improve the Framework for Housing and Housing Finance 47 Efficient Land Administration 47 Housing and Housing Finance Information and Transparency 48 Long-Term Funding Availability through Mortgage Refinancing 48 Expanded Housing Finance to Low-Income Groups through Product Innovation 49 Appendix A Afghanistan 51 Overview 52 Housing Demand 56 Housing Supply 57 Housing Finance 60 vi Contents Low-Income and Microfinance Housing 67 Regulatory Framework and Mortgage Market Infrastructure 72 Policy Options for Improving the Framework for Housing Finance 76 Appendix B Bangladesh 83 Overview 84 Housing Demand 85 Housing Supply 86 Housing Finance 91 Low-Income and Microfinance Housing 105 Regulatory Framework and Mortgage Market Infrastructure 111 Policy Options for Improving the Framework for Housing Finance 116 Appendix C India 127 Overview 129 Housing Demand 133 Housing Supply 135 Housing Finance 141 Low-Income and Microfinance Housing 160 Regulatory Framework and Mortgage Market Infrastructure 175 Policy Options for Improving the Framework for Housing Finance 191 Appendix D Pakistan 209 Overview 211 Housing Demand 214 Housing Supply 216 Housing Finance 223 Low-Income and Microfinance Housing 247 Regulatory Framework and Mortgage Market Infrastructure 254 Policy Options for Improving the Framework for Housing Finance 262 Conclusion 275 Appendix E Sri Lanka 279 Overview 280 Housing Demand 282 Housing Supply 284 Housing Finance 289 vii Contents Low-Income and Microfinance Housing 304 Regulatory Framework and Mortgage Market Infrastructure 308 Policy Framework and Suggestions for Future Development 314 Index 323 Boxes 1.1 Twin Challenges: High Property Prices and Expensive Mortgages 5 4.1 Successful Financial Sector Development in India 20 4.2 Securitization in India 23 4.3 Dynamic Specialized Mortgage Lenders in Bangladesh 24 4.4 Malaysia's Experience with Housing Finance Reforms 29 4.5 India's Recent Prudential and Supervisory Strengthening for the Financial Sector 31 Appendix Boxes A.1 Meeting Housing Needs: Efforts from the Public Sector 58 A.2 A Housing Developer in Need of Finance: Shelter for Life 59 A.3 Finding Strong Demand for Upper-Middle-Class Housing in Kabul 67 A.4 Rehabilitating Land Records in Afghan Courts: A Success Story 73 A.5 Improving Ownership Security in Informal Settlements in Kabul 74 B.1 Market Failures in the Housing and Housing Finance Markets in Bangladesh 92 B.2 Role of NGOs in Housing-Related Social Services 109 B.3 Property Registration in Bangladesh 114 B.4 Contractual Savings Schemes for Housing: A Primer 122 C.1 Typical Low-End Housing Available in Urban Markets 140 C.2 Market-Based, Low-Income Housing Pilot Projects in India 171 D.1 The Central Bank's Prudential Regulations 237 D.2 Low-Income Housing Initiatives in Pakistan 251 D.3 Land Titling System Description and Shortcomings 255 E.1 Cagamas Berhad, the Malaysian National Mortgage Corporation 319 Figures 1.1 Ratio of Housing Debt to GDP, Selected Areas and Countries 3 1.2 GDP Growth Rate, 2007 and 2008 3 1.3 Urban Population, 2008 4 2.1 Housing Shortages in South Asia, 2008 10 4.1 Growth Rates of the South Asia Housing Finance Industry, 2007­08 20 6.1 Cost and Time to Register a Property in South Asia 39 6.2 Time to Foreclose on Property in South Asia 44 viii Contents Appendix Figures A.1 Growth of Nominal GDP, 2003­07 53 A.2 Rates of Inflation, 2003­07 53 A.3 Bank Assets, Liabilities, and Capital, 2004­09 54 A.4 Projected Demand for Housing, 2006­14 56 BB.1 Urban Households' Estimated Incomes and Mortgage Affordability, 2003 92 C.1 Population and Urbanization Trends in India, Selected Years, 1901­2008 132 C.2 Size of the Housing Finance Industry, 2008: Outstanding Mortgages as a Share of GDP 133 C.3 Residential Price Movements, by City, 2007­08 137 C.4 Housing Affordability Index, 1995­2006 138 C.5 Retail Portfolio of Banks, 2008 144 C.6 Housing Finance Disbursements, by Type of Institution 145 C.7 Performance Statistics of HFCs, 2004­08 148 C.8 Mechanism for Issuing Residential Mortgage-Backed Securities 151 C.9 Housing Finance System 157 C.10 Feasibility of Low-Income Housing 161 C.11 Potential Housing Demand of Urban Households 162 C.12 Current Market for Housing Loans Served by Financial Institutions 165 C.13 Estimated Construction Costs for Low-Income Housing 166 C.14 Agencies Involved in Land Administration 176 C.15 Time and Cost to Register Property in Indian States and Selected Countries 183 C.16 Lengthy Delays in Judgment across India 188 C.17 Cost of Contract Enforcement across India 189 C.18 International Comparison of Credit Information Availability, Selected Countries, 2009 191 D.1 Urban and Rural Populations in Pakistan, 1950­2030 212 D.2 Segmentation of the Housing Market by Income Group 215 D.3 Housing Finance Outstanding Loans, March 31, 2008 and 2009 224 D.4 Number of Housing Finance Borrowers with Outstanding Loans, March 31, 2008 and 2009 225 D.5 Market Shares of Housing Finance Lenders, March 31, 2009 226 D.6 Mortgage Products Offered by Housing Finance Lenders, 2009 227 D.7 Nonperforming Loans in the Housing Finance Sector, 2009 229 D.8 Commercial Banks Versus the HBFC--Outstanding Portfolios' Historical Trends, 2003­08 230 D.9 Competition between Banks and the HBFC--Market Shares, 2004­08 231 D.10 Commercial Banks Versus the HBFC--Number of Borrowers, 2003­08 231 ix Contents D.11 Herfindahl Concentration Index for the Banking Sector 233 D.12 Growth in Islamic Assets, Deposits, Financing, and Investment 235 D.13 Debt-to-GDP Ratio, 2003­09 247 E.1 Sri Lanka GDP Growth, 2000­08 281 E.2 Current and Potential Housing and Housing Finance Markets at a Glance 290 E.3 Projected Growth of Housing Finance Loans from Banks, 2006­20 301 Tables 2.1 Glum Reality in the Slums: Statistics 10 3.1 Prevalence of Temporary Housing in South Asia 14 4.1 Typical Mortgage Terms in South Asia 21 4.2 Total Mortgage Balances Outstanding and Market Shares 22 6.1 Typical Data Collected by Credit Bureaus in South Asia 46 Appendix Tables A.1 Financial Access, 2009 54 A.2 Bank Performance, 2009 55 A.3 Financial Stability, 2009 55 A.4 Cost and Affordability of Housing in Kabul 67 A.5 Advantages and Disadvantages of Potential Housing Finance Providers 69 A.6 Potential Target Groups for Housing Microfinance 70 A.7 Affordability Analysis for Housing Microfinance 71 A.8 Home Affordability in Afghanistan and Comparator Countries, 2009 71 B.1 Housing Sector Characteristics, 1998 86 B.2 Dhaka Urban Area Housing Subsystems, by Income Group and Land Coverage 87 B.3 Residential Land Prices in Dhaka, 2004 88 B.4 Rise in the Prices of Construction Materials, Selected Years, 1988­2007 90 B.5 Outstanding Housing Loans and Overview of the Housing Industry, 2006­08 94 B.6 BHBFC Loan Portfolio, 2008 96 B.7 Typical Lending Terms of NHFIL, 2009 99 B.8 Performance of Investments under the IBBL's Special Investment Schemes, 2008 101 B.9 Yields on Treasury Bills and Bonds, 2001­08 104 B.10 Housing Finance Rates, June 30, 2008 105 B.11 Property Registration Process and Costs 113 C.1 Growth Trends for the Urban and Rural Populations, 2000­09 132 C.2 Distribution of Units by Location and Type of Structure, 2002 136 C.3 Housing Completion, 1971­2001 136 x Contents C.4 Banking and Nonbanking Activity, 2001 and 2008 142 C.5 Indian Banks' Financial Soundness Indicators, 2000 and 2009 145 C.6 Key Performance Indicators, by Bank Group, Selected Years 146 C.7 Key Statistics on the 10 Largest HFCs, 2009 147 C.8 Financial Highlights of the HDFC 149 C.9 NHB Refinance Disbursements, Fiscal 2007/08 149 C.10 Liquidity through Securitization 151 C.11 Distribution of Commercial Banks According to Risk-Weighted Adequacy, Selected Years 155 C.12 Evolution of Housing Policies in India: A Succession of Five-Year Plans, 1951­2012 173 C.13 Reforms of Property Registration Procedures, 2008­09 184 D.1 Basic Housing Indicators for Pakistan, 1998 and 2009 214 D.2 Details of Katchi Abadis in Pakistan, by Province, 2008 216 D.3 Interest Rates and Maturities in the Housing Finance Sector 227 D.4 Loan-to-Value Ratios and Loan Processing Times 228 D.5 Concentration in the Mortgage Market, 2003­07 232 D.6 Banks and DFIs with Housing Finance Portfolios Exceeding PRs 1 Billion, March 31, 2008 234 DD.1 Mortgage Loan Classification 238 D.7 Action Plan of the 2001 National Housing Policy, Implementation Status 240 D.8 Housing Advisory Group Recommendations 243 D.9 Estimated Urban Mortgage Potential, 2009 247 D.10 Microfinance Industry at a Glance 249 DD.2 Land Allotment Agencies' Locations, Procedures, Rules, and Documents 256 D.11 Cost Structure of Land Titling and Registration, 2009 257 D.12 Average Per-Land-Record Cost to the Customer 258 E.1 Cost Indexes of Housing Construction, Selected Years 287 E.2 Distribution of Banks and Bank Branches, 2007 and 2008 292 E.3 Total Outstanding Housing Mortgage Loans in the Financial System 293 E.4 Mortgage Lending by the Banking System, 2000­06 294 E.5 Commercial Banks' Advances for Housing and Property Development, 2000­06 295 E.6 Comparison of Mortgage Rates with Other Consumer Lending Rates 296 E.7 Nonperforming Loans of the Banking System, Selected Years 297 E.8 Efficiency of State-Owned and Large Private Banks, 2006 297 E.9 Average Income by Population Decile, Fiscal 2006/07 308 E.10 Land Parcels and Registration Status, Urban and Rural Areas, 2003 309 xi Foreword The importance of housing around the world cannot be overstated. Housing is a critical necessity that in some countries is equated with a human right. Housing not only provides shelter, but is the cornerstone of community life, the center of family activities, and a reference point for societal esteem, as well as the most important financial asset most households will ever acquire. In South Asia, which is home to one in four people on the planet, every second person has an income below the poverty line, and housing shortages have become a pervasive problem that has defied resolution for decades. In India alone, more than 25 million families have no home. Compounding the housing shortages are high room-density figures (more than 3.5 people per room in India and Pakistan) and the pervasiveness of slum dwelling areas and squatter settlements, sometimes on footpaths or even in abandoned sewage pipes. Nearly half of the major metropolitan areas in the region are taken up by slum dwellings--Bangladesh alone has more than 2,100 slums, and India over 52,000. In Afghanistan, about 36 percent of people are not able to meet their basic needs, and in Kabul city alone, nearly two-thirds of the housing stock has been dam- aged or destroyed during the last three decades of war. Indian and Sri Lankan mort- gage markets, the biggest in the region, are still dwarfed by Chinese mortgage markets, at double the size relative to national income, and Malaysian markets, roughly four times larger. Several factors have contributed to the problem of inadequate housing. First, rapid population growth in the region has added a considerable number of families to the ranks of those in need of housing. Second, most population growth has occurred in urban areas, resulting in rapid urbanization. The developing world absorbs an average of 5 million new urban residents every month and is responsible for 95 percent of the world's urban population growth. In 1950, about 232 million people lived in urban areas, which represented about 17 percent of Asia's total population. In 2005, Asia's urban population had risen to 1.6 billion people, or about 40 percent of the region's total population. By 2030, one out of every two urban xiii Foreword residents in the world will be in Asia, and the vast majority of them will be urban poor, living in poor habitats and in slums. Third, economic growth has contributed to a significant increase in the size of the middle class in South Asia, where changing culture and shrinking household size have fuelled demand for homes even further. Fourth, household savings have been inadequate to fund housing for a consider- able share of South Asian families, and housing finance has been limited. Banks have been risk-averse in their exposure to lower-income clients, and slow to acquire mod- ernized processes of bulk lending to higher-risk clients. Also, innovative microfinance approaches have not spread as fast as needed. As a result, South Asia has the lowest penetration of mortgage finance in the world. The supply of low-cost houses has not been forthcoming on the part of developers, who face their own financing challenges and regulatory constraints, as well as record high land prices. Fifth, another reason for the housing shortage is that existing housing stock has been depleted due to poor maintenance, civil strife, natural disasters, and other factors. The governments in South Asia cannot be expected to provide a solution to a housing problem of such proportions. Efforts on the regulatory front to provide an enabling environment with an aim to promote housing development have in some cases been successful, but urban planning, serviced land availability, and transport links from outskirts to places of employment have been generally inadequate. The success stories from both the pri- vate and public sectors need to be shared across South Asia, since housing is a com- mon concern that can be better addressed by the exchange of knowledge and better access to information. The region's housing institutions and central banks, aware of the importance of housing, have taken important steps. In addition to the many ongoing national hous- ing initiatives, a regional initiative called the South Asia Housing Finance Forum (SAHF) is a key milestone, which serves as a more permanent platform for housing issues in the region, with its Secretariat in Delhi, India. The SAHF's purpose is to promote better understanding and cooperation among the countries of the region on issues relating to housing and housing finance. The first Advisory Board of SAHF is chaired by Abdul Qadeer Fitrat (Da Afghanistan Bank), and consists of Mr. Kam- ran Shahzad (State Bank of Pakistan), R.V. Verma (National Housing Bank India), Mohammad Abul Quassem (Bangladesh Bank), P.D.J. Fernando (Central Bank of Sri Lanka), and Zaigham Mahmood Rizvi (SAHF Secretary General). Mr. Fitrat is the chairman of the Advisory Board. The forum has functioned since its founding in 2009 as a platform for knowledge sharing and networking among countries, which will in turn help in policy design, strategic intervention, and the design of different instruments for low-income hous- ing and the housing finance market. The forum provides for the exchange of learn- ing and best practices to facilitate a coordinated approach toward "Housing for All" in the region. The forum promotes coalition and coordination among the South Asian countries in determining their housing and housing finance solutions. SAHF also tracks global housing market events and focuses on initiatives taken by the dif- ferent member countries in their own regional contexts. The users of the platform include institutions engaged in housing and housing finance--such as housing xiv Foreword finance institutions, builders and other construction agencies--co-operatives, government agencies, policy and regulatory bodies, and international financial insti- tutions. The forum uses both a virtual platform (e-Portal) and a physical platform (conferences, seminars, and publications). Steps have been initiated to expand the forum's coverage to the entire Asia-Pacific region. Expanding Housing Finance to the Underserved in South Asia, a first regional effort on the topic, examines housing needs and shortages in South Asia, and outlines shortcomings of the market for home mortgages. The primary purpose of this book is to pull together housing and housing finance information for the countries of South Asia. Housing and housing finance data have been scarce in most countries in the region, preventing policy makers, private sector businesses, and financial sector stakeholders from enabling and improving these markets. The book presents illus- trations of best practices and country-specific examples that can enable builder/ developers, mortgage lenders, and land administrators to create foreclosure and other relevant regulations that will strengthen home ownership in South Asia. Special emphasis is accorded to innovative solutions for low-income housing. Finally, the book draws attention to regional challenges to efficient and effective housing and housing finance markets and suggests an agenda for the future. His Excellency Abdul Qadeer Fitrat, Governor, Central Bank of Afghanistan and Chairman of SAHF Zaigham Rizvi, Secretary General of SAHF xv Acknowledgments This report owes much to our colleagues at Da Afghanistan Bank, Bangladesh Bank, Reserve Bank of India, State Bank of Pakistan, and the Central Bank of Sri Lanka. The update of individual country chapters would not have been possible without the guid- ance of several specialized housing institutions in-country, including the Bangladesh House Building Finance Corporation, the National Housing Bank of India, and the Pakistan House Building Finance Corporation. The cooperation, assistance, and joint work with the International Finance Corporation have been invaluable. Finally, the team thanks all South Asia banking associations, as well as many individual banks, other financial institutions, developers, real estate associations, other housing and housing finance stakeholders, and government agencies who took time to discuss and brainstorm on pertinent issues. The effort has benefited from able advisers, including Junaid Ahmad, Nazir Ahmad, Saddiq Ahmed, Dan Biller, Loic Chiquier, Songsu Choi, Richard Clifford, Kenroy Dowers, Syed Farhan Fasihuddin, Syed Ejaz Ghani, Douglas Grayson, Max Heimann, Sriyani Hulugalle, Reazul Islam, Lohita Karunasekera, Akbar Zaman Khan, Shabana Khawar, Colleen Mascenik, Ashna Mathema, Yibin Mu, Ary Naim, Richard Nash, Kyoo-Won Oh, Michael Pomerleano, Shah Nur Quayyum, Kiatchai Sophastien- phong, John Speakman, and Niraj Verma. The team owes particular appreciation to the peer reviewers Olivier Hassler, Narayanasamy Kokularupan, and Samuel Maimbo for their detailed and careful comments, overall guidance, review, and advice. The excellent support from Marjorie Espiritu, Sakm Abdul Hye, Sashikala Jeyaraj, Parwana Nasiri, Aza Rashid, Vinod Satpathy, Imtiaz Ahmad Sheikh, Sumriti Singh, and Tasnim Zohra, has made this work possible. Mehnaz Safavian contributed extensively to chapter 1, drafted appendix E, co-led the coordination work, and provided key input (including the book title) and much valued support. Valentina Saltane contributed to research for chapter 1. Zaigham Rizvi provided extensive background information in updating the appendixes and tirelessly helped with the overall research, report drafting, and review. xvii Abbreviations ABAD Association of Builders and Developers (Pakistan) APCA All Pakistan Contractor Association BHBFC Bangladesh House Building Finance Corporation CBSL Central Bank of Sri Lanka CIB Credit Information Bureau (Pakistan) CRAR capital-to-risk-weighted-assets ratio CRIB Credit Information Bureau (Sri Lanka) CRISIL Credit Rating Information Service of India, Limited CSH contractual savings scheme for housing DFI development financial institution DRT Debt Recovery Tribunal (India) EMI equated monthly installment EPF Employees' Provident Fund (Sri Lanka) ETF Employees' Trust Fund (Sri Lanka) GDP gross domestic product GNP gross national product HBFC House Building Finance Corporation (Pakistan) HDFC Housing Development Finance Corporation (Sri Lanka) HFC housing finance company HUDCO Housing and Urban Development Corporation (India) IBBL Islami Bank Bangladesh Limited LICHF Life Insurance Corporation Housing Finance Limited (India) LRIS land registration information system MFI microfinance institution MISFA Microfinance Investment Support Facility for Afghanistan NBFC nonbank financial company NBFI nonbank financial institution NCB nationalized commercial bank xix Abbreviations NGO nongovernmental organization NHA National Housing Authority (Bangladesh) NHB National Housing Bank (India) NHDA National Housing Development Authority (Sri Lanka) NHFIL National Housing Finance and Investments Limited (Bangladesh) NSB National Savings Bank (Sri Lanka) NSS National Savings System (Bangladesh) NSSO National Sample Survey Organization (India) RAJUK Rajdhani Unnayan Kartripakkha [Dhaka development authority] RBI Reserve Bank of India REIT real estate investment trust ROR Record of Rights SBP State Bank of Pakistan SCB scheduled commercial bank SEWA Self-Employed Women's Association (India) SMIB State Mortgage and Investment Bank (Sri Lanka) UDA Urban Development Authority (Sri Lanka) USAID U.S. Agency for International Development Currency conversion: Exchange rate at the time of the study (December 2009): Afghani: Af 1 $0.0205972 Bangladesh taka: Tk 1 $0.0144613 Indian rupee: Re 1 $0.0212517 Pakistan rupee: PRs 1 $0.0118484 Sri Lanka rupee: SL Re 1 $0.00875503 All amounts are presented in U.S. dollars, unless otherwise indicated. xx Chapter 1 Introduction Currently, more than a billion people around the globe live in inadequate housing. One in every four people on the planet lives in the South Asia region, and more than 14 percent of South Asians have no homes or live in such inadequate housing as urban slums and squatter settlements. South Asia's housing and housing finance markets are dynamic, but limited in their outreach, catering to upper-income groups. The impressively high growth rates of these markets give hope that housing and housing finance services have the potential to be expanded to middle- and even lower-income families. In some countries, such as India, this shift is already afoot. The challenge is that South Asia is home to about half of the world's poor people, and these people require more complex housing solutions because existing market terms are not directly affordable for them. The contribution of the housing and real estate sector to overall economic growth, social uplift, and employment is considerable in the context of the rapidly expanding economies. Personal residences account for 75­90 percent of household wealth in emerging-market countries, which amounts to three to six times their annual income. Housing represents 15­40 percent of the monthly expenditure of households worldwide. Similarly, economic development investment in housing accounts for 15­35 percent of aggregate investment, whereas housing construction and housing-related sectors constitute approximately 9 percent of the labor force worldwide. In South Asia, 40­50 industries are considered to be directly linked with housing construction. A unit increase in expenditure in this sector has a multiplier effect and the capacity to generate income as great as five times the cost of the unit. If the economy grows at the rate of 10 percent, the housing sector has the capacity to grow at 14 percent and generate 3.2 million new jobs over a decade (World Bank 2008a). In India, for example, for every rupee invested in housing, Re 0.78 gets added to the national GDP. The National Housing Bank of India 1 Expanding Housing Finance to the Underserved in South Asia 1 estimates that the construction sector provides direct employment to 16 percent of India's workforce; and the housing sector alone is the second-largest employment generator after agriculture, accounting for 58 percent of the workforce in the con- struction sector. Although the Pakistani construction sector has only a 2.3 percent share in GDP, its share of the employed labor force is disproportionately large (7 percent). Construction in Pakistan has been growing at an average annual rate of 10.2 percent over the last five years (data from Pakistan's House Building Finance Corporation). Housing finance plays a critical role in the development process by support- ing strong housing markets, while strengthening the financial sector and con- tributing to overall economic growth. With strong housing and housing finance markets come many economic and social benefits, such as greater consumer savings, more social and labor mobility, and increased investment. In addition, strong housing markets support job creation in construction and manufacturing, and they improve living conditions and basic infrastructure. The availability of housing finance also has social implications. Where the rapid flow of population to cities puts pressure on the housing supply, and where either new construction is unable to keep up with escalating demand or the lack of financing makes housing unaffordable, slums proliferate. Conversely, when mortgage financing is available, the market for housing grows, and a larger share of the population can become homeowners. Housing and housing finance build assets and livelihoods, and thus contribute toward poverty reduction. However, a housing finance system could provide benefits to the economy beyond development of a housing market and contributions to employment and growth. For example, instituting a sound property registration sys- tem would enable entrepreneurs to use their property as collateral for business loans. With transparent lending for housing, where risks can be measured and mitigated, financial institutions would be able to put capital that they might otherwise have held against residential mortgages to other productive uses, such as business lend- ing. In addition, housing finance represents an important asset class in the financial sector--an asset class that could help develop a long-term finance market for other industries, including infrastructure development. And private sector lending for housing would free scarce government resources for other social and economic needs. Mortgage finance contributes to financial sector growth and accounts for a sizable part of a liberalized financial sector. In the context of the global financial crisis, it became clear that housing finance is critical to economic stability; and the sector requires careful focus and prudent development. South Asian countries have low levels of mortgage debt outstanding (ranging from 0.1 percent to 7.0 percent), relative to all other regions globally (figure 1.1) (World Bank 2008a). The low penetration implies room for growth. A number of countries in the South Asia region are developing reform agendas that would strengthen their housing finance systems. As a result, additional data, more infor- mation on good practice in housing finance reforms, and country-specific examples are in high demand. 2 Introduction 1 Figure 1.1 Ratio of Housing Debt to GDP, Selected Areas and Countries South Asia 3.4 Afghanistan 0.1 Bangladesh 3.0 India 7.0 country/region Pakistan 1.0 Sri Lanka 6.0 China 15.0 European Union 42.0 Malaysia 26.0 Thailand 17.0 0 10 20 30 40 50 percent Source: World Bank 2008a. Growth Potential of Housing Finance Markets The world population is expected to reach 7.9­10.9 billion by 2050. Within the next two decades, 60 percent of the world's people will reside in urban areas. Urban growth rates are highest in the developing world, which absorbs an average of 5 mil- lion new urban residents every month and is responsible for 95 percent of the world's urban population growth (World Bank 2008a). Rapid economic growth in the South Asia region, coupled with significant population increases and growing urbanization trends (figures 1.2 and 1.3), make for a considerable potential for Figure 1.2 GDP Growth Rate, 2007 and 2008 20 16.2 16 12 percent 9.1 8 6.4 6.2 6.8 7.1 6.0 6.0 6.0 4 0 Bangladesh Sri Lanka India Pakistan Afghanistan country 2007 2008 Source: World Bank 2009. 3 Expanding Housing Finance to the Underserved in South Asia 1 Figure 1.3 Urban Population, 2008 40 percent of total country population 36.2 35 29.5 30 27.1 25 24.0 20 15.1 15 10 5 0 Afghanistan Bangladesh India Pakistan Sri Lanka country Source: World Bank 2009. housing and housing finance growth. The factors contributing to urban migration are greater economic growth, rising income levels, new employment opportunities in the cities, occupational shift from agriculture to manufacturing and services, and changing attitudes toward consumption and lifestyle. Household size is shrinking, and a significant middle class will be forming in the coming decades, mostly com- posed of young people who are eager to have their own, albeit modest, homes. This increased demand will add to already existing urban housing shortages. The atti- tudes to debt are changing, and a decreasing aversion to borrowing is observed, par- ticularly for home ownership. High economic growth and remittances have created a plentiful middle class, willing and able to pay for a home; but these factors also have pushed up urban property prices. These developments are propitious to a scaled-up outreach of housing and housing finance markets. It is difficult to make precise estimates of the growth potential of the housing and housing finance sectors in South Asia; however, it is possible to make some tentative projections. Afghanistan's potential financing needs amount to $276 million a year for home improvement and new construction in Kabul's informal housing sector alone, involving total fixed private capital investment of $2.5 billion (excluding land value). Microfinance clients in Afghanistan have demonstrated a large demand for housing loans, often under the guise of business microloans. It is estimated that 20­25 percent of business loans there may have been used for home construction and improvement (data from Da Afghanistan Bank). In Bangladesh, fully 3.5 million people in rural areas are without shelter. In urban areas, about 21 million people belong to the lower-income groups, residing in 1.9­3.0 million (mostly temporary) dwellings that require incremental construction and repair (data from Bangladesh House Building Finance Corporation). Thus, the potential housing finance demand is considerable. In India, the National Housing Bank estimates investment for meet- ing housing needs to 2012 at $108 billion. In urban Pakistan, for 2009 alone, pro- jected effective demand amounts to 470,000 housing units, at a total construction 4 Introduction 1 cost of $5.6 billion, or a requirement of $2.0 billion (1.26 percent of Pakistan's GDP) in financing. Mortgage financing requirements in Sri Lanka are projected at up to $8.8 billion by 2020 (data from the Central Bank of Sri Lanka). Summary of Key Challenges In South Asia, housing finance reaches the upper-income population groups; and the main challenge is to expand access to the middle- and lower-income groups, developing housing and housing finance markets that are both sound and accessible. The markets need balanced funding models and a diversified toolbox of instruments for different lenders. Complicating factors include the rapid increase of land prices in most countries over the past decade; poor legal infrastructure; deficient financial systems; a dearth of long-term funding at fixed rates; limited developer finance; and problematic access to housing finance for low-income, rural, and informal popula- tion groups (box 1.1). Sound and prudential regulations for housing finance, inno- vative housing finance products, and increased mortgage affordability will con- tribute to an expanded market for housing and to vibrant and fast-growing housing finance systems and products. Islamic finance is an area that could match demand in underserved market segments, and it needs to be developed. A sharper focus on sec- ondary market finance is appropriate; mortgage refinancing companies are key to the provision of long-term funding, standardization, and stabilization of mortgage markets. One of the most difficult obstacles that the housing and housing finance sectors are facing at the moment is the inefficiency of the overall regulatory regime, includ- ing foreclosure and land administration frameworks. Land and titling procedures, registration procedures and costs, and a poor regulatory framework for housing and real estate stifle housing markets, as does the lack of an organized database and key information on the housing and housing finance sectors. Other challenges include insufficient developed land and inappropriate land planning and urban development Box 1.1 Twin Challenges: High Property Prices and Expensive Mortgages In Bangladesh, improving the loan terms from 10- or 15-year maturity and 14­15 percent inter- est rates to 20-year maturity at 8 percent interest would double the population that can afford mortgage finance from 0.28 million to 0.56 million households. Lowering the average property price from a typical Tk 2.0 million ($29,000) to Tk 0.7 million ($10,000) quadruples the market to 2.24 million households, or 40 percent of the urban population. Source: Author's calculations. 5 Expanding Housing Finance to the Underserved in South Asia 1 policies driving up real estate prices, sluggish building supply, poor government success in addressing low-income housing needs, and lack of commercially viable housing microfinance lending. Public-private partnerships and creative cooperative solutions for low-cost housing are especially promising. Objectives of the Report The report reviews key issues affecting the state of development of the housing and housing finance systems in the South Asia region, including Afghanistan, Bangladesh, India, Pakistan, and Sri Lanka. Four key dimensions of housing finance that affect the market include housing availability, financial instruments and institu- tions, low-income housing, and mortgage market infrastructure. Policy and reform options to improve the framework for housing finance are discussed in the context of the South Asia region. The report specifically focuses on taking stock of the cur- rent state of housing finance in South Asian countries, identifying critical areas for long- and short-term reform and sharing lessons from reformers both within and outside the region. When discussing housing finance in South Asia, one is faced with a segmented market. The market caters to upper-income groups, where supply is ample from both developers and financial outfits. Prices in this market are not affordable to the middle- and lower-income groups. The housing finance market has exhibited fast growth, very healthy development, and considerable success and sophistication, driven by steady property price escalation in the past decades. The upper-income market has remained unaffordable even to middle-class families, however. This report necessarily focuses on the market shortcomings as they concern these middle- and lower-income housing and housing finance markets; and frequently it does not discuss in thorough detail the upper-income, smoothly functioning housing and housing finance markets in the region. Scope of the Report This report is based on five country-specific studies conducted by the World Bank Group, which were rewritten and updated as needed, and supplemented by further research on location.1 Looking forward, this regional effort hopes to extend its scope to all countries in South Asia, based on further housing and housing finance work carried out in the region, notably extending the work to Bhutan, Maldives, and Nepal. Note 1. The country-specific studies are Monitor Group (2007) and World Bank (2004, 2006, 2007, 2008b). 6 Introduction 1 References Monitor Group. 2007. "Expanding the Housing Finance Market to Cover Lower Middle Income Segments in India." Unpublished manuscript prepared for the World Bank and the Financial Sector Reform and Strengthening Initiative, Cambridge, MA. World Bank. 2004. "Housing Finance Reforms in Bangladesh: Unleashing Supply to Meet Growing Demand." South Asia Private and Financial Sector, Washington, DC. ------. 2006. "Improving the Performance of the Housing, Tourism, and Retail Sectors." Foreign Investment Advisory Service, Washington, DC. ------. 2007. "Housing Finance in Sri Lanka: Opportunities and Challenges." South Asia Private and Financial Sector, Washington, DC. ------. 2008a. "Financing Homes: Comparing Regulations in 42 Countries." Washington, DC. ------. 2008b. "Housing Finance in Afghanistan: Challenges and Opportunities." South Asia Private and Financial Sector, Washington, DC. ------. 2009. World Development Indicators 2009. Washington, DC. 7 Chapter 2 Housing Demand Housing shortages are not present at high-income levels in South Asia, but are a con- cern among middle- and low-income groups. This is the result of rapid population growth, increased urbanization, high income inequality, and displacement of people by natural disasters and conflict. Large-scale slums and informal settlements that prevail in almost all South Asian countries underline the urgency of meeting the demand for homes. There is a shortage of more than 38 million housing units in the region, not counting housing in need of repair or replacement (figure 2.1). Taking into account average household size, this translates to 212.5 million homeless people, 14 percent of a total region population of 1.5 billion. The current need for new housing units in Afghanistan is estimated at more than 1 million, with 70 percent of these units required in urban areas. In Bangladesh, housing needs extend to about 5 million units, with 1.0­1.5 million needed in urban areas. The urban housing backlog in India is estimated at 27 million, and a major part of it is in the low-income category. About 71 percent of urban households in India live in overcrowded dwellings. The housing backlog in Pakistan was at an estimated 7.57 million units in 2009-- 2.5 million of them in urban areas. Official estimates put the existing demand-supply gap at 350,000 housing units in Sri Lanka, and the overall shortage could exceed 650,000 units in 2010. In general terms, worse urban shortages are hiding behind squatter settlements and higher persons-per-room densities. The average persons- per-room count in urban areas is high: 3.5 in India and Pakistan. That density is comparable regionally, but worse than in developed countries: density is 0.5 persons per room in the United States and 1.1 in the European Union. Density in Sri Lanka is an outlier at 1.1 persons per room.1 South Asian slums and informal settlements are overcrowded and have unsani- tary living conditions: no proper electricity supply, lack of clean drinking water 9 Expanding Housing Finance to the Underserved in South Asia 2 Figure 2.1 Housing Shortages in South Asia, 2008 Share of total population 35 35 31 30 25 percent 20 18 15 13 11 10 5 0 Afghanistan Bangladesh India Pakistan Sri Lanka country Sources: World Bank (2009) and data from appendixes A through E of this volume. and/or heating. Moreover, because properties in shantytowns are devoid of legal titles, formal financing does not reach these informal settlements. The absence of legal titles and the constant fear of being expelled from current dwellings also discourage people from investing in house infrastructure and improving their homes. Informal settlements provide shelter for 80 percent of Kabul's population (2.44 million people) and cover 69 percent of Kabul's residential land, including 10,000 people living in tents and 5,000 living in destroyed buildings. In the infor- mal settlements of Kabul (both legal and illegal), it is estimated that more than two thirds of houses are occupied by more than one family. Almost 9 of every 10 fami- lies in informal housing live in only one or two rooms (table 2.1) (data from Da Afghanistan Bank). Table 2.1 Glum Reality in the Slums: Statistics Country Slums Statistics Afghanistan 80 percent of the Kabul population (2.44 million) lives in slums. Bangladesh There are 2,100 slums; more than 2 million people in Dhaka either live in slums or are without any proper shelter. India There are 52,000 slums holding 8 million urban households, representing about 14 percent of the total urban population. Pakistan Karachi alone has 600­800 slums, sheltering about 7.6 million people (or 1 million households) out of the total city population of 15.1 million people. Sri Lanka A considerable share of the population of Sri Lanka lives in plantations, slums, and shanties. Sources: Data from appendixes A through E of this volume. 10 Housing Demand 2 Bangladesh has more than 2,100 slums, and more than 2 million people in Dhaka either live in slums or are without any proper shelter. Of the 3 million urban low- income Bangladeshi households, roughly half are estimated to be located in slums (both owner occupied and rented), sometimes on squatted land. Formal slums (with full land ownership, but in bad maintenance condition and badly overcrowded because of urbanization) take another quarter of the dwellings in this group. Other categories of low-income housing include government-provided squatter resettle- ment camps, plots of land with basic services that are provided on a leasehold basis; employee housing consisting mostly of small apartments in high-rise complexes provided by the government; makeshift houses built by squatters on illegally occu- pied public or private land; and pavement dwellings (data from Bangladesh House Building Finance Corporation). The slums and squatter settlements in India number 52,000 and hold 8 million urban households, representing about 14 percent of the total urban population. Pakistan's lower-income groups expand into illegal habitats and squatter settlements in major cities. More than 50 percent of the urban population lives in slum areas in Pakistan. Karachi alone has between 600 and 800 shantytowns, sheltering more than 7.6 million people out of the total city population of 15.1 million people. Slums are not as prevalent in Sri Lanka (only 1.5 percent), although estate housing is charac- terized by abject living conditions. Note 1. Unless otherwise noted, all data provided in this chapter are from background papers prepared by Zaigham Rizvi in 2009. Reference World Bank. 2009. World Development Indicators 2009. Washington, DC. 11 Chapter 3 Housing Supply Current State of Housing Not all houses in South Asian countries are permanent dwellings (table 3.1). Afghanistan's housing stock is in poor condition, damaged, or destroyed, with a majority of housing units located in unplanned areas. In Bangladesh, only 23 percent of housing in urban areas and merely 2 percent in rural areas is permanent housing.1 Even among the housing defined as permanent, there are many types of dwelling places with some degree of permanency, ranging from construction with brick masonry and reinforced cement concrete pillars to tin-roofed and tin-walled houses. Close to half of all housing units in Bangladesh are constructed entirely of temporary materials and are predominantly occupied by the rural poor. Fully 3.5 million people in rural areas are without shelter. In urban areas where incomes are higher, the lowest- income groups are estimated at 70 percent of the population (about 21 million people in 3 million households). Urban dwelling units are estimated at between 3.5 and 4.0 million, of which 1.9­3.0 million are temporary. The heterogeneity of housing makes it difficult to estimate construction and housing materials costs, and compli- cates the measurement of real estate price indexes. The Indian housing market is highly fragmented, with the unorganized sector accounting for more than 70 percent of the housing units constructed. It is estimated that, at the beginning of 2006, the total housing stock in the country stood at 211.86 million units (153.03 million units in rural areas and 58.83 million units in urban areas), out of a total of 222.93 million households (156.63 million households in rural areas and 66.30 million households in urban areas). Of the total estimated housing stock, 117.63 million houses are pucca houses, 65.53 million units are semi- pucca houses, and the remaining 28.70 million units are kutcha houses that need replacement.2 13 Expanding Housing Finance to the Underserved in South Asia 3 Table 3.1 Prevalence of Temporary Housing in South Asia Country Temporary Housing Afghanistan No reliable data Bangladesh 50 percent of all housing India 45 percent of all housing Pakistan 39 percent of all housing Sri Lanka 20 percent of housing units have mud floors Source: Zaigham Rizvi, 2009. In Pakistan, about one third of available housing units are urban, and the rest are rural. Approximately 40 percent of houses are semi-pucca (semi-permanent) houses, most without planned sanitation or sewerage systems; 39 percent are kutcha (tem- porary) houses with minimal water supply and sanitation or drainage services; and only 21 percent are modern brick constructions. In Sri Lanka, building units are classified as housing units (including single house, attached house, flat, estate [line room] house, and slum/shanty--the last being 1.3 percent of all housing units) or other dwellings (including collective living quarters, institutions, and nonhousing units). Colombo buildings are about 82 per- cent housing units and 18 percent other dwellings. Estate housing is not present in Colombo, but is very prevalent in rural areas. For roofing, mostly tiles are used (58.6 percent), followed by asbestos and metal sheets. About 55 percent of the house- holds have brick walls and 20 percent have walls of cement block; 80 percent have cement floors. Nationwide, 20 percent of houses still have mud floors. Development of Housing Stock The sheer scale of housing shortages makes it impossible for government alone to provide sufficient housing for low-income groups. The considerable housing shortages require a large-scale solution that realistically can be provided only by a functioning market. In some countries, such as India, private developers have been stepping in as they discover that cheaper housing is actually capable of presenting good commercial opportunities. However, most of the industry in South Asia has remained focused on the upper-income end of the scale. That focus can be cost effec- tive based on traditional products and processes, without the need for costly inno- vation. Faster reorientation of the building industry to lower-income construction is predicated on two conditions: developer finance and efficient land administration. Sound governance, professional standards and norms, transparency, and good consumer protection will help as well. The current land development framework is inefficient in some countries,3 and it acts as a critical bottleneck to the availability of developed land in the market. In some countries, the government still has to reorient from housing provision to 14 Housing Supply 3 regulation. Land development is plagued with legal and administrative burdens, which increase the costs and time for delivery and reduce the availability of devel- oped land. The institutional framework for identifying, acquiring, and servicing land available for construction, and subsequently selling the constructed housing units, is underdeveloped. Some of the factors causing this shortcoming are poor master plan- ning, governance issues at national and local government levels, multiple institutions and administrative procedures, problematic zoning restrictions, restrictive building codes, and unreliable utility connections. In addition to regulatory inefficiencies, physical land scarcity is a constraint (for example, in Bangladesh). The conversion of rural land into urban land at the town periphery is a key means of increasing the supply of land, which should be most beneficial for the lower classes (including farmers). The process is made problematic by local and provincial governments' ill- inspired interventions, however. Overly restrictive building codes and laws on sub- division limit the efficient use of urban land and increase the price to consumers, especially in zones where land prices are high. Poor incentives and poor efficiency within government-controlled companies, especially in power and gas industries, can delay large-scale construction projects considerably. Scarcity of developed land at affordable prices beyond the highest-income groups has been the biggest challenge for policy makers and the housing sector. There are roughly 800 developers in Bangladesh, and the market is active and rel- atively competitive. The housing development process is lengthy and costly. That results from poor preparation of master plans and the dearth of planning profes- sionals in the Bangladesh public sector, and from inadequate infrastructure provi- sion, land acquisition, development, construction, and mortgage financing. In India, housing construction is managed mostly by the owners who engage the services of small contractors. In some larger metropolitan areas, builders and develop- ers have played a role in supplying housing by building high-rise, multifamily units and, more recently, by developing large-scale vertical and horizontal residential complexes. Pakistan is another example where the current land development framework is highly inefficient and acts as a critical bottleneck to the availability of developed land in the market. Private developers in Pakistan are relatively dynamic and have been engaging in considerable activity, including large infrastructure projects. In Sri Lanka, the developer market is not as active; and self-construction is the norm because owners prefer to act as their own general contractors. Virtually all developer activity in South Asia is focused on the upper-income sec- tion of the population. Although the high-end market is small in absolute numbers, it is large enough to take more than the supply of housing that builders are capable of producing, and to provide high margins and a quick turnover. That leaves a consid- erable gap in housing development and provision for middle- and lower-income groups. In India, however, some medium and small developers see low-income hous- ing as a good commercial opportunity, and they have developed viable and profitable models to serve lower-income groups. Their efforts are worthy of emulation across the region (appendix C, box C.1 and the "Housing Microfinance and Low-Income Housing" section). Public-private partnerships have also been extremely effective in 15 Expanding Housing Finance to the Underserved in South Asia 3 international experience in assisting and encouraging small-scale developers to build affordable housing. Construction Quality and Enforcement of Construction Standards Poor construction quality and noncompliance with construction standards resulting from poor enforcement are common features of the South Asian housing market. There is a lack of quality control in the materials sector, inadequate supervision, and limited enforcement of building codes. In the absence of oversight, most home con- struction follows neither the code nor approved blueprints. Moreover, most houses that are being built are not durable and require much maintenance. There is very limited application of robust, low-cost technologies in the region, although such construction is critically needed. Most housing developments are being done by local, poor-quality construction companies; except in India, no serious attempts have been made to establish technical collaboration with globally acknowledged low- cost construction technology providers. Housing Costs In the last decade, there has been an unprecedented rise in South Asian property prices. This rise has impaired affordability considerably, especially for low- and middle-income families. Increases in housing prices are driven mostly by the costs of land and construction materials, which in turn are increased by rising demand, urbanization, high economic growth, increasing remittances, limited availability of land, speculative trade in real estate, and rising international materials prices. In Afghanistan, the cost to construct low-income housing might range from $15 to $20 per square foot. High-income dwelling prices stand at $19 per square foot. Home improvement costs range between $100 and $5,000 per housing unit. In Bangladesh, prices have grown at 15­16 percent, much above the general cost-of- living level. The least expensive residential land in Dhaka costs approximately $27 per square foot, and upscale areas can be priced as high as $60 per square foot. In the case of India, although urban housing prices have risen by 30­40 percent annually, salaries have increased by an annual average of 20 percent over the last two years. In South Mumbai, the price per square foot increased from $215 in 2004 to $430 in 2006. In central Delhi, land price per square foot doubled from $193 in 2005 to $387 in 2006. For a typical property in Mumbai, affordability has declined from 4.4 times in 2004 to 5.5 times in 2006 (and prices declined in 2008, resuming their ascent thereafter). In Pakistan, costs range from $9.6 to $12.0 per square foot for low-cost housing, from $16.7 to $19.1 per square foot for middle-cost housing, and PRs 26.3 to PRs 35.8 per square foot for high-end housing. The cost of construction in Sri Lanka increased about threefold between 1990 and 2005. Building materials that registered substantial price increases since 1990 include sand (1,070 percent), timber (568 per- cent), and bricks (678 percent). Labor cost increased by nearly 250 percent during 16 Housing Supply 3 that period. Some countries (such as India) have successfully counteracted the land price increases by carefully crafted land-use policies and urban planning, giving a chance at housing affordability to middle- and lower-income households. Housing Price Index The importance of housing and real estate as major areas for creation of both phys- ical and financial assets and their contributions to overall national wealth suggest the need for setting up a mechanism that could track the movement of prices in the res- idential housing segment. Regular monitoring of house prices can be useful input for a variety of housing sector stakeholders--including policy makers--and can pro- mote secondary housing markets. India is the only country in South Asia with a housing price index, the RESIDEX, launched by the National Housing Bank in 2007 (see appendix C). Price indexes are useful as tools for both housing policy and finan- cial stability support. Secondary Property Markets Secondary markets for property are limited, potentially compounded by a speculative attitude toward property prices. Secondary markets in South Asia are sluggish, and suffer from a lack of transparency, from price speculation, and from sales involving legally imperfect documentation. There is a dual pricing system as a result of high property costs and in an effort to avoid taxes, which obstructs revenue collection, lim- its transparency, and acts as an obstacle to market liquidity. Weak market signals in the mortgage market and a lack of foreclosure enforcement do not help. Institutional mechanisms and the legal framework are not in place to stimulate a smoothly func- tioning real estate transactions market. Specifically, bottlenecks are created by the lack of secondary housing market infrastructure and by a high transfer tax. In Bangladesh, for example, the transfer tax rate is extremely high (12.5 percent of gross price), and it strongly discourages the transfer of assets. There is a dearth of real estate specialists, including brokers and appraisers, and information on market prices and valuation is poor. Secondary market links supporting both the housing and housing finance industries could substantially contribute to market growth. Notes 1. Unless otherwise noted, all data provided in this chapter are from background papers prepared by Zaigham Rizvi in 2009. 2. A kutcha house is one built with nondurable materials like unburned bricks, mud, thatches, leaves, and either timber or bamboo. A pucca house is one built with permanent materials like oven-burned bricks, concrete, iron, or other metal sheets and timber. A semi-pucca is a hybrid of kutcha and pucca construction materials. 3. India is an example where the framework does function well. 17 Chapter 4 Housing Finance Institutions and Instruments Financial markets, dominated by large banking sectors in South Asia, have seen a gradual liberalization in recent decades, and have been exhibiting remarkable growth as a result (box 4.1).1 Yet outreach remains limited, in view of extensive rural populations and predominating informality of finances for lower-income groups. Financial services have been limited to basic products, and have extended in a very modest manner into housing or related insurance products. Most banks avoid lend- ing for housing, which typically involves long-term commitments to clients. Among the main reasons for this are the lack of an adequate supporting legal framework, including poor foreclosure and eviction procedures and land titling regimes; absence of reliable property valuation; lack of a well-functioning collateral system; a missing yield curve; and absence of long-term treasury instruments in some countries. Fur- ther complications include weak competition in the financial sector, poor trans- parency, underdeveloped market structural features such as second-tier lenders, and the lack of a level playing field for financial institutions. Funding distortions tend to handicap the most dynamic and efficient actors; and in some countries, a legacy of a high level of nonperforming housing loans among state-owned banks plagues the system to this day. However, housing finance is growing at very high rates, albeit from a very low base, spurred on by high excess demand and market innovations to increase supply (figure 4.1). In Bangladesh, the sector as a whole stood at Tk 142.5 billion, or a little over 6 percent of total credit to the private sector. In India, mortgage growth was about 30.0 percent in 2003­07, before the impact of the global financial crisis was felt. In Pakistan, the housing market grew only by 1.3 percent from March 2008 to March 2009, but earlier growth rates were in double digits. In Sri Lanka, housing lending makes up about 3.4 percent of total financial assets. 19 Expanding Housing Finance to the Underserved in South Asia 4 Box 4.1 Successful Financial Sector Development in India The Indian financial sector has steadily increased its spread and outreach in the past several decades. The banking and nonbanking sectors experienced market-based reforms. Interest rates were liberalized. Capital markets were deregulated and opened to international flows. Private entry into banking and insurance increased the contestability of the domestic financial market. Important public sector financial institutions, such as ICICI Bank, were made private, thus putting competitive pressure on the system. As a result, financial depth has increased sub- stantially, and regulatory oversight and risk management have been strengthened. Outcomes on nonperforming assets and capital-to-risk-weighted assets have improved significantly, and compare favorably with international benchmarks. Banks have experienced strong balance sheet growth, improved financial health, increased competitiveness, and productivity gains through adoption of technology, wider outreach, and larger allocations of credit to important productive sectors. India's capital markets have deepened and matured over time to become the largest markets in developing countries, relative to GDP (155 percent in 2007, 53 percent in 2008 after the crisis). The size of the government bond market is also significant, at about 42 percent of GDP. India's ratio of financial assets to GDP, at approximately 170 percent, is higher than that in such middle-income countries as Brazil and Mexico. The ratio of domestic bank credit to GDP has been growing steadily, from 29.0 percent in 2000 to 54.5 percent in 2008. Source: World Bank 2009. Figure 4.1 Growth Rates of the South Asia Housing Finance Industry, 2007­08 50 40 40 30 30 percent 20 20 14 13 10 0 Afghanistan Bangladesh India Pakistan Sri Lanka country Source: Zaigham Rizvi, 2009. Traditional Mortgage Lending In South Asia, the banking sector generally dominates the financial system, and non- banking financial institutions play a small role. In India, for example, nonbanking assets amount to barely 0.2 percent of banking assets. Traditionally, banks have also 20 Housing Finance Institutions and Instruments 4 Table 4.1 Typical Mortgage Terms in South Asia Country Mortgage Terms Afghanistan $10,000­$20,000, 12%, 20-year maturity Bangladesh $36,300­$43,600, 14­15%, 10­15-year maturity; restrictive loan-to- cost maximums (capped at 70% generally, but averaging 50% in practice); interest rates mostly variable India $30,000 on average, 12%, 13-year maturity, "plain vanilla" home amortizing loans, on predominantly floating rates and with prepayment penalties; average loan-to-value ratio of 65% (maximum 85% at origination) Pakistan $21,000­$44,400 on average, 15­17%, 12.5-year maturity on average Sri Lanka $10,000­$40,000, 15­17%, 15­25-year maturity on average; generally fixed rates by state-owned banks and variable rates by private banks Source: Zaigham Rizvi, 2009. been the providers of housing finance. Their services have been limited to the upper- income, formally employed population groups. In some state-owned commercial banks, a legacy of inefficiency subsidies and resultant bad mortgage debt still per- sists. Private and foreign banks are more aggressive in new financial services, includ- ing housing finance, although foreign banks have specialized more in construction lending and larger housing finance, feeling a weaker local experience and higher transaction costs. Nevertheless, mortgage products remain inflexible and mostly unaffordable. Mortgage terms are geared to high-income borrowers, with minimum loan amounts of approximately $20,000­$30,000 in most countries and interest rates in the teens. Maximal maturities are low (10­20 years), and loan-to-value ratios are cautious (starting at about 50 percent). Fixed-rate mortgages are hard to get (table 4.1). The country appendixes in this volume provide a detailed review of housing finance market players and available statistics. The outreach of financial services in Afghanistan remains very limited, although they are expanding. The housing sector, in particular, has almost no access to bank credit. The absence of major laws, including the long-awaited mortgage law, is at the root of these developments. Two state-owned banks, Bank-e-Millie Afghan and Pashtany Bank, have made small forays into the housing finance market. Most com- mercial banks, along with microfinance institutions, are keen to launch housing finance products in the near future. Traditional mortgage finance in Bangladesh dominates the market as well (table 4.2), with nationalized commercial banks and private and foreign commercial banks holding 23.6 percent and 35.7 percent market shares, respectively. The housing finance market in India is serviced by a variety of financial institu- tions, including scheduled commercial banks, scheduled cooperative banks, regional rural banks, housing finance companies (HFCs), and state-level apex cooperative housing finance societies. The housing finance market is concentrated, with four 21 Expanding Housing Finance to the Underserved in South Asia 4 Table 4.2 Total Mortgage Balances Outstanding and Market Shares State Private Specialized Specialized Islamic Country Bank Mortgage Mortgage Mortgage Mortgage (Balance) Financing (%) Financing (%) Financing (%) Financing (%) Afghanistan (no 100 0 0 No separate estimate balance available) exists; amount is included in "Bank Mortgage Financing" column Bangladesh 64 17 12 7 (Tk 142.4 billion) India 66 4 30 No separate estimate (Rs 3,970.0 billion) exists; amount is included in "Bank Mortgage Financing" column Pakistan 76 13 0 11 (PRs 89.9 billion) Sri Lanka 75 25 0 0 (SL Rs 181.0 billion) Source: Zaigham Rizvi, 2009. financial institutions controlling more than two thirds of the market. In contrast to other South Asian countries, the central bank, Reserve Bank of India, shares regula- tory oversight of the housing sector with the National Housing Bank, the second-tier refinancing institution that is also responsible for regulating HFCs. Indian banks account for roughly two thirds of the market, having overtaken the HFCs, which had enjoyed a considerable lead. Lending for housing by commercial banks as a percent of gross credit outstanding was up from roughly 4 percent in 2001 to about 11 per- cent in 2008. In Pakistan, the share of commercial banks in housing finance grew from 10 percent in 2003 to more than 70 percent in 2009. However, the state-owned Bangladesh House Building Finance Corporation (BHBFC) is the only institution that caters to the financ- ing needs of middle- and low-income groups, servicing some 71 percent of housing finance borrowers. Banks dominate housing lending in Sri Lanka as well, where private specialized mortgage lenders are not well developed. In fact, banks and state-owned specialized mortgage lenders are the sole players in the country's market. State-Owned Specialized Housing Banks To fill in the gaps left by commercial bank mortgage lending, especially for lower- middle- and low-income, remote rural, or informally employed population groups, South Asian countries have established state-owned housing banks, which usually 22 Housing Finance Institutions and Instruments 4 enjoy privileged terms and funding. Generally, these banks have been inefficient in terms of operating costs and delays. They have failed to reach the lower-income groups or people who are informally employed. Afghanistan experimented unsuc- cessfully with a state housing bank prior to 2002, but its operating performance and loan portfolio were so weak that it failed to obtain a provisional license in 2003 under the new Banking Law. A new specialized institution is being planned, under a public- private partnership, with management control in the private sector. The proposed institution will have a business focus on the middle- and high-income end of the population. The BHBFC, the state housing bank, remains a dominant market player, despite its modest size (about 1,100 loans disbursed annually to a total of Tk 1.2 billion, with an overall 45,000 loans in stock and 550 employees). The institution is beset by a poor governance structure, misaligned operational incentives, high nonperforming loans, and ill-targeted government subsidies. The bank takes months to approve a loan, suffers prohibitive operating costs per loan, and offers a limited product range. In spite of these shortcomings, it is the only institution serving the low- and middle- income housing segment and a wider geographic area beyond Dhaka and a few larger cities. India's National Housing Bank is a second-tier housing finance institution that refinances primary lenders, including HFCs and banks. It also lends directly to proj- ects undertaken by public housing agencies for housing construction and develop- ment of housing-related infrastructure. This is the only second-tier institution in the South Asia region's housing finance market (box 4.2). Compared with the commercial banking sector, Pakistan's HBFC is remarkably focused on lower-income groups. The average loan size is PRs 1.08 million, com- pared with an average of PRs 2.29 million in the case of private sector banks, Box 4.2 Securitization in India There have been more than 500 issuances of asset-backed securities in India, amounting to Rs 900 billion as of mid-2009. Although asset-backed securitization has grown on the back of the retail boom, mortgage-backed securitization remained rare because regulations and liquid- ity issues make direct assignment preferable. There is no secondary market for these securities as yet, and cross-border issuances are not currently allowed. The residential mortgage-backed securities issues were based on conforming mortgages with standardized pool selection crite- ria. The requirements included a seasoning rule and consistent performance, high collection effi- ciencies, consistent payments to investors, servicer audit, special-purpose vehicle audit, and surveillance by rating agencies. The 14 deals made by the National Housing Bank were pre- dominantly direct assignment deals, asset-backed securities transactions with simple struc- tures, single tranche, and credit enhancement (primarily cash collateral). For more information, see the "National Housing Bank" and "Securitization" sections of appendix C. 23 Expanding Housing Finance to the Underserved in South Asia 4 PRs 1.75 million for public sector banks, and PRs 2.68 million for foreign banks. The institution has been reforming and modernizing in the past few years, and can show some impressive results in operating costs and shortened delays for loan approvals. The central bank, State Bank of Pakistan, is now working to set up a long-term liquidity institution (the Pakistan Mortgage Refinance Company). In Sri Lanka, state banks are similarly inefficient and poorly competitive. Their focus is more on lower- income clients, but their impact on the market is weak. Private Specialized Mortgage Lenders In response to remaining excess demand for home funding, many innovative private sector specialized institutions have appeared, extending the housing market down the income scale (box 4.3 and the "Power of Microfinance" section of chapter 5). The microfinance sector has expanded into housing from the lower end of the income spectrum, lending for home improvement instead of home purchase. These initia- tives have been sensitive to the specificity of low-income group needs--the lack of formal documentation or even title to the land; irregular earnings that cannot be documented; and cultural, religious, and gender constraints such as the reluctance of women in some regions to leave their houses. These institutions have not only diver- sified products but also segmented clients to handle risk better and process a larger number of smaller mortgages more cheaply and efficiently. The sector still struggles with attaining access to long-term, low-cost funding, in a market competition with banks that have access to cheap deposit funds. As a case in point, over the last decade, Indian HFCs have lost considerable mar- ket share to commercial banks that have the cheaper funding advantage--access to deposits. Some HFCs are public or use government funding to cope with the lack of Box 4.3 Dynamic Specialized Mortgage Lenders in Bangladesh Private institutions specializing in housing finance in Bangladesh have made remarkable strides in a short time. The management teams at Delta Brac Housing Finance Corporation Limited, National Housing Finance and Investments Limited, and IDLC Finance Limited, aspiring to the same professional standards as their counterparts in other areas of the world, have made sub- stantial progress in underwriting, loan servicing, streamlining local decisions, eliminating delays, and increasing transparency. With a national market share of 6.7 percent, Delta Brac has done particularly impressive work in putting together credit histories of potential borrowers and beginning to correlate credit history with loan performance. Still, these institutions focus on clients at the upper- and upper-middle-income levels. Specialized financial institutions have also made strides in loan servicing. A vigilant surveillance system, keeping careful track of delin- quencies and communicating with borrowers via cell phones and automated messaging at the first sign of trouble, maintains nonperforming loans at 0.5 percent. 24 Housing Finance Institutions and Instruments 4 a level playing field in funding costs. A funding advantage is also enjoyed by HFCs owned by commercial banks. For the rest, funding sources include commercial banks, bonds and debentures, refinancing from the National Housing Bank, term deposits (because HFCs are not allowed to have checking and savings accounts), loan assignment, and mortgage-backed securitization. But HFCs perform an extremely useful role in the housing finance market. They provide smaller loans to lower-income clients, and their performance is comparable to that of banks--a capital-to-risk-weighted-assets ratio well above the prescribed 12 percent and non- performing loans at 2.2 percent (banks maintain them around 1.0 percent). HDFC Bank, the largest HFC in India, has an asset size of more than Rs 850 billion and 3.3 million mortgages financed (compared with 4.5 million financed by the entire banking sector). It enjoys a widespread geographic presence, a very nominal share of nonperforming loans, the right to accept deposits, and a position in the financial market permitting it to float its own bonds and mortgage-backed securities. Its total loan write-offs since inception are only 4 basis points of cumulative disbursements, and its cost-to-income ratio is 8.8 percent--one of the lowest in financial services in Asia. The lack of a level playing field for funding is a pervasive factor in most South Asian countries. In Pakistan, funding disadvantages have eliminated the specialized housing finance sector, including companies like Asian Housing Finance Limited, Citibank Housing Finance Company, and International Housing Finance Limited. Islamic Housing Finance South Asia has a considerable potential to absorb Islamic financial opportunities and instruments. This mode of finance can play an important role because it often targets customers with lower-than-average incomes. One of the most popular housing prod- ucts in Islamic finance is diminishing musharakah, a declining-balance/shared-equity partnership. With this product, both parties share the risks and obligations related to the property. Ownership of the property is divided into equal units and is purchased regularly by the customer. Under this arrangement, leasing and selling contracts must be conducted separately. The main steps involved in diminishing musharakah require the consumer to select a house and obtain approval for financing. The financier and consumer then jointly buy the house and obtain joint-ownership title through a musharakah agreement. The consumer leases the house from the financier, and even- tually buys out the financier's equity, thus getting sole title to the property. A simplified version of home finance is a sort of lease-buyback arrangement. Under these loans, known as murabaha, the bank buys the property and resells it to the client at a markup, owning the property outright until final payment--an arrangement offering the bank a level of security not available with a traditional mortgage. The murabaha may enable banks to overcome some difficulties posed by poorly functioning foreclosure frameworks. Although Islamic banking is growing substantially and has massive untapped potential, there is a need to introduce more 25 Expanding Housing Finance to the Underserved in South Asia 4 Shar'ia-compliant products to meet the varied needs of customers. Although Afghan banks are only experimenting with such instruments, the Islami Bank Bangladesh Limited (Bangladesh) holds a 6.6 percent market share of housing finance in Bangladesh, and is the largest private commercial bank. In Pakistan, Islamic housing finance increased by 43 percent from January 2008 to March 2009. Islamic banks more generally have an 11 percent share in total outstanding loans in Pakistan, and they have showed the highest growth in the financial sector (21.5 percent). It is sur- prising that Islamic banks have not played any role in lower- and middle-income home financing. Large-Scale Project Finance One of the major stumbling blocks for large-scale construction market development, especially for low-income housing, has been the dearth of developer financing. Fund- ing for builders is limited because of the underlying risks involved. As a result, land acquisition for housing development is mostly funded by equity. That in turn con- fines development projects to an inefficiently small scale. Although construction loans are available, they have low loan-to-value ratios and carry such high rates of interest that developers consider them to have little value. Owing to the lack of robust construction financing, the ultimate buyer of the property usually finances the con- struction costs through installment payments. Even if used, construction loans can run into titling transfer problems in some countries. These loans also transfer risk from the developer to the homebuyer because the latter pays significant advances to the developer before construction is completed. The advances are paid entirely in cash, from savings. This also reduces affordability and further limits the market for new construction to clients who can pay fully up front--the upper-income groups. In most South Asian countries, such as Afghanistan, developers fund entirely out of own equity. Bangladesh developers fund mostly out of equity as well. Developer finance from both banks and capital markets is somewhat more avail- able in India, although still used exclusively for upper-end construction. In Pakistan, banks shy away from funding developers as well. There is little trans- parency in the ownership of land for development, which makes collateral insecure. The situation is similar in Sri Lanka, where private commercial banks are not overly eager to finance development because of the complications to obtaining a clean title to the land to be developed. As a result, builders focus on high-end condominiums, whose prices increase so fast that they sell for cash in installments during construction. Informal Housing Finance Low-income groups in South Asia, rural dwellers, and those with irregular or informal incomes have little chance of obtaining home funding from the formal financial sys- tems. Those population groups fund their homes through savings and borrowing from 26 Housing Finance Institutions and Instruments 4 the informal market, including family and friends. Afghan businesses and home- owners, for example, must rely almost entirely on friends, relatives, and moneylenders to meet their financing needs. The hawala system, an underground banking network based on traditional methods of paperless transactions, does not provide a significant source of housing finance. Despite the need, less than 10 percent of credit in urban areas is obtained from local moneylenders because of the high interest charged and onerous potential repercussions of delinquency. Informal credit is generally a short- term loan for consumption to cope with temporary shortfalls in cash flow. Similarly, much of the funding for house construction in India comes from informal sources that include households' cash savings and loans and gifts from relatives, mon- eylenders, shopkeepers, and landlords. This informal sourcing can be attributed to the inadequacies of the existing formal financing system and the still-timid steps of micro- finance institutions into the housing sector. In the absence of both formal and infor- mal home funding, low-income households resort to rental housing. Rental markets are very strongly developed in India, where 29 percent of urban low-income groups and 4 percent of rural ones are housed in urban rented units. The informal rental mar- ket in slum areas is characterized by slumlords who organize squatting or sell hutments to the urban poor, with some amount of administrative protection at the local level. Homeowners' Insurance In South Asia, life insurance and property insurance are confined to the upper- income population groups, covering a very small portion of homeowners. Mortgage insurance and title insurance are generally not available. Credit guarantee mecha- nisms may serve as a good intermediate step to provide risk mitigation to financial institutions and collect data that can be used to price the mortgage insurance. Risk Management The ability of South Asian financial institutions to identify and assess risks is crucial to their viability and sustainability. Mortgage markets in South Asia are exposed to several key risk aspects, including liquidity risk (maturity mismatch), market risk (interest rate risk), credit risk, and risk stemming from the poor legal framework. Risk management is sound for banks, but less so for their state-owned specialized cousins that are also frequently not regulated by commercial bank rules. The most serious risk in incomplete financial markets is liquidity risk resulting from term mismatch. Housing loans are long term, but both banks and nonbanking financial institutions fund these assets with short-term liabilities. Bank short-term assets are relatively stable, provided there is a steady core deposit base (which is not always the case in South Asia). The more risky approach is to fund housing finance liabilities with short-term resources from the capital market, as the global financial crisis amply demonstrated. Nonbanking financial institutions suffer more strongly from this fundamental term structure mismatch. Smaller housing finance institutions 27 Expanding Housing Finance to the Underserved in South Asia 4 mostly operating in the private sector are unable to raise long-term funds. This inabil- ity creates a situation where they are forced to lend over a shorter term, all the while bearing an exposure to their mismatched maturity position. Commercial banks actively engaged in providing mortgage financing are better able to cope with the liq- uidity risk, relying on demand and time liabilities; and are able to benefit from higher yields. However, this poorly mitigated exposure creates interest rate risk and maturity mismatches, and will prove increasingly costly in a tightening prudential regulatory environment. Financing long-term mortgages with short-term liabilities is not sus- tainable and could lead to credit shortages and instability of the financial system. In the long run, a viable funding structure for housing loans is required, such as a mort- gage refinancing company, mortgage guarantee facility, and eventually recourse to the capital markets. At present, however, the option of going for outright securitization is remote because of institutional inadequacies and cost constraints. The capital markets and market-based funding sources have not yet developed for long-term instruments of finance. A refinance facility could provide liquidity to the financial institutions with recourse. Challenges in creating such a structured funding mechanism include the low volume of long-term mortgage loans, compared with the total deposit base of all banks; the general use of cost-efficient funding for housing through demand and time liabilities; and the broad spread between the weighted average cost of deposits and the price of housing loans. In spite of a historically relatively stable macroeconomic environment in most of the region, market risk stemming from interest rate fluctuations is not well managed by housing finance outfits; instead, it is passed on to the clients. At present, banks are mitigating interest rate risk by offering adjustable rate mortgages, whereby interest rate risk associated with housing finance is borne by the borrower. This could fur- ther translate into a credit risk in a rising interest rate scenario so as not only to restrain housing demand but also to pose a systemic risk. The practice also limits the outreach of the mortgage market. Credit risk is not particularly problematic for housing lenders at this stage of housing finance market development, but will become increasingly important as mortgages are extended down the income scale to less affluent, middle-class, and eventually low-income borrowers. Furthermore, most commercial banks currently have limited their credit risk by maintaining a trivially small exposure to the hous- ing sector. Credit risk is being managed by increasingly better client information and a more efficient use of credit bureau data. Mortgage origination and servicing 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 to tackle term mismatch and interest rate and credit risks more successfully. The mortgage finance market is exposed to a further set of risks stemming from an underdeveloped legal framework. These risks are caused by a poorly defined basis of ownership and title, which creates a risk factor for the banks on the legality and enforceability of borrower property rights. Dysfunctional secured transactions laws 28 Housing Finance Institutions and Instruments 4 create the risk of failure of loan collateral enforceability or, at the very least, the risk of an exceedingly costly and lengthy foreclosure. Poor professional support for the real estate industry creates its own risks, including unreliable property valuation. Secondary Market for Mortgage Finance The first step toward a second-tier structure for mortgage finance is usually a mort- gage refinance/liquidity facility (box 4.4). This is a specialized second-tier institution that provides short-term liquidity, long-term funding, or guarantees to housing finance lenders. It acts as an intermediary between lenders and capital markets, and issues bonds to raise long-term finance while purchasing loans with recourse or receiving assignment of mortgages/loans. The mortgage refinance facility helps develop the primary mortgage market by providing financial resources to enable primary lenders to grant more loans at fixed rates and for longer tenures; helping primary lenders to narrow the gap between the maturity structure of the housing loans and the source of funds; promoting sound lending norms and eligibility crite- ria, enabling smaller lenders to access long-term funding and to foster competition; lowering the cost of long-term funding (scale effect, liquidity, prime standing, lim- ited intermediation cost); and acting as a catalytic tool for housing finance market Box 4.4 Malaysia's Experience with Housing Finance Reforms Since the 1980s, the Malaysian government's policies have been directed toward low-income group housing and toward encouraging private sector participation in low-cost housing provi- sion. As a result of enabling policies, the private sector has accounted for 66­75 percent of houses developed. By the 1970s, commercial banks and finance companies had replaced Malaysia Building Society Berhad in providing housing finance; and their current share is 82 per- cent. The government also promoted the sector, via savings promotion (for example, the employee provident fund, which can be used for education, housing, or medical needs). Various tax exemptions were also used to spur new housing purchases for low-income groups. The central bank also promoted the low-income housing market, via interest-rate ceilings. Most recently (in the 2008 budget), a guarantee fund was established to facilitate lending to house- holds in the informal sector. However, the major effect came with the setting up of the Malaysian national mortgage corporation Cagamas Berhad. As a result of competition among banks, mortgage rates dropped, maturities lengthened, and banks moved to hybrid mortgages (interest rates fixed for one to five years and floating thereafter). During the financial crisis, Cagamas successfully provided liquidity to the banks, thus avoiding a credit crunch. Finally, Cagamas deepened the bond market, linking housing finance with the capital markets. Lessons from the Cagamas success story suggest that political stability, government support, and coor- dination among government agencies are crucial for the success of low-cost housing finance market development. Private sector participation is also important. Product customization is of the essence. Legal and regulatory issues can make a considerable difference as well. 29 Expanding Housing Finance to the Underserved in South Asia 4 development. Unlike securitization, mortgage refinance facilities do not need a fully developed primary market with the history required to value securities. To develop the capital market for housing finance, it is necessary to provide more private debt securities (bonds) with different maturities and rates; to issue secured and simple instruments with risk that is easy to assess and with no valuation issues; and to create potentially liquid investment vehicles, as well as a yield curve. The intrinsic benefits of securitization remain, however, in spite of its potentially dangerous results if handled in an unstable manner. Capital markets can provide liquidity relief to commercial banks and nonbanking financial institutions; and they can help a bank cover an unexpected short-term deposit outflow (or other tempo- rary losses of funds), thereby avoiding potentially costly short-term borrowing or asset liquidation. The mechanism provides long-term investment opportunities to institutional investors, contributes to the deepening of the bond market, and can be further used to attain specific policy objectives (that is, it may set criteria for refi- nancing of loans to low-income groups). Securitization produces instruments that are well secured, simple to value, and potentially liquid--frequently in a poorly developed, illiquid, and risky market. To establish a mortgage refinancing facility, financing institutions must have effective funding needs, and a sufficient demand for and supply of housing and housing finance should exist. A functioning primary mortgage market, based on a sound infrastructure, is a very important precondition, as is the existence of a capital market and a relatively diversified investor base. The legal system should provide adequate enabling regulations, including allowing the effective assignment/transfer of mortgage loans. Commitment by the central bank and/or government to support the new institution in its initial phase and to help promote name recognition is of the essence. Prudential Regulation and Oversight Most countries in South Asia do not have specific prudential regulations for housing finance, although the level of sophistication ranges from low in Afghanistan (where only now are major laws being enacted, including the Mortgage Law 2009 and the Secured Transaction of Movable Property Law 2009), to high in India (where Basel II is being implemented and a number of rules strengthening the financial system have been introduced [box 4.5]). Specialized state-owned housing banks, such as Pakistan's HBFC, are not under the general banking supervision of the central bank, thus introducing a major potential instability risk in the financial system. In India, HFCs are not under central bank supervision, and that might introduce a lack of coordination and weaknesses in the prudential supervision of the financial system. In some countries, such as Bangladesh, housing finance issues are covered under the prudential regulations for consumer finance. These generalized arrangements miss out on important rules suited exclusively to housing finance activities, such as a reg- ulatory ceiling on the "mortgage debt-to-advances" ratio on the mortgage lending of commercial banks. 30 Housing Finance Institutions and Instruments 4 Box 4.5 India's Recent Prudential and Supervisory Strengthening for the Financial Sector In the past decade, India's central bank, the Reserve Bank of India (RBI), carried out a phased introduction of prudential norms on capital adequacy, income recognition, asset classification, and provisioning to bring Indian banks in line with international practices. It also strengthened the corporate governance of banks. A credit bureau was established, and significant changes were made to improve the execution of collateral. The RBI was strengthened substantially, in both its monetary policy functions and in its regulation of the banking system; and it is widely regarded as a credible, reputable, and expert institution. Higher levels and standards of disclo- sure were achieved to enhance market transparency. Micro prudential measures were insti- tuted, risk-based supervision was introduced, and the supervision process was streamlined with a combination of on-site and off-site surveillance along with external auditing. The owner- ship of public sector banks has been broadened through disinvestment up to 49 percent, and banks have been listed. A mechanism for greater regulatory coordination was instituted for reg- ulation and supervision of financial conglomerates. Finally, measures were taken to strengthen creditor rights. Specifically, a number of steps have been taken to improve banking, including the introduc- tion of core banking solutions, installation of automated teller machines, enhanced working hours, simplification of procedures and forms, guidelines on fair practices, and guidelines on credit card operations. The Enforcement of Security Interest and Recovery of Debts Laws (Amendment) Act 2004 and the Securitization and Reconstruction of Financial Assets and Enforcement of Security Interest Act 2002 have been enacted to facilitate quick and effective recovery of nonperforming assets. The Credit Information Companies (Regulation) Act 2005 has been enacted to facilitate setting up of credit information companies for collection, sharing, and dissemination of credit information, which will help significantly improve the quality of credit appraisals and decisions. To further accelerate the process of reforms in the banking sector, the government has amended the RBI Act of 1934. Amendments also have been proposed in the Banking Regulation Act of 1949. More recently, RBI has taken some further necessary actions to strengthen prudential norms for the banking sector. Among others, these actions include (1) issuing prudential guidelines regarding capital adequacy, pursuant to the Basel Committee on Banking Super- vision's International Convergence of Capital Measurement and Capital Standards: A Revised Framework; (2) adopting prudential norms on off­balance sheet exposures for scheduled commercial banks; and (3) strengthening prudential norms for valuation, classifi- cation, and operation of commercial banks' investment portfolios. Updated guidelines were issued in July 2008. The guidelines required all banks to migrate to the Basel Revised Frame- work on March 31, 2009, and to maintain a minimum capital-to-risk-weighted-assets ratio of 9 percent. New guidelines on prudential norms for off­balance sheet items were issued in August 2008. These guidelines included directives on provisioning norms for derivative exposures and on computing credit exposure on account of derivatives. The norms for val- uation, classification, and investment portfolios were updated and issued in July 2008; they covered issues such as internal controls and accounting. Stress testing also has been insti- tuted as a regular measure in the banking system. 31 Expanding Housing Finance to the Underserved in South Asia 4 More recent regulations, linked to the aftermath of the global financial crisis, have brought about a slew of liquidity and other support for housing finance institutions, not all of which is benign. In Bangladesh and India, subsidized credit lines were extended to mortgage lenders. If care is not taken, such schemes might distort the system and introduce a nonlevel playing field. A Word on the Global Financial Crisis The global financial crisis has been high on the economic priority list; and its impact has spread worldwide, starting in housing markets in developed countries and with the subprime crisis. The subprime crisis revealed that securitization markets could be dangerous, and thus produced a deep loss of confidence in the instrument. It is important to note that subprime issues are not typical in developing countries. The subprime crisis was characterized by features completely foreign to developing hous- ing markets--low interest rates, high loan-to-value ratios, an explosion of claims within the financial system, extension to clients with poor credit, and complexity of the securitization process and underlying rating and valuation models. In contrast, markets in South Asia are very limited and generally accessible only to people in the top income bracket, loan-to-value ratios are low, and market complexity is manage- able. Islamic banking is even less vulnerable to subprime pressures, on the face of it. The global financial crisis is a very relevant experience, however, in a scenario where housing finance markets do grow at a faster speed and expand successfully to a much larger portion of the population. In hindsight, the global financial crisis has had very little impact on South Asian countries, relative to expectations. Liquidity was down, and some dampening effect on economic growth was noted. Central banks in most countries stepped in to cush- ion the blow, providing liquidity to key sectors and complimenting fiscal stimulus solutions with monetary policy, to the extent allowed by national reserves. (RBI, in particular, was very successful in its policies). Note 1. Unless otherwise noted, all data provided in this chapter are from background papers prepared by Zaigham Rizvi in 2009. Reference World Bank. 2009. World Development Indicators 2009. Washington, DC. 32 Chapter 5 Low-Income Housing Finance The formal financial system is limited to upper-income groups in South Asia. However, market outreach can be stretched down the income scale through innova- tion and enabling regulation. Mortgage services could be offered to the middle-class population and even to lower-middle-income groups. But there will be some income groups at the bottom of the scale for whom mortgages and new home purchases will remain unaffordable, even in developed housing finance markets. In India, lower- income rental markets can be trusted to service some of the very-low-income demand, although in suboptimal living conditions. In all countries, microfinance institutions (MFIs) target some of these clients for home improvement, incremental construction, and repair loans as well. The Power of Microfinance Microfinance is most developed in Bangladesh, and is picking up with a fast pace in India and elsewhere. Since 1987, when Grameen introduced its first housing loan fol- lowing a devastating flood, MFIs have made progressive forays into servicing the housing finance market.1 Currently, the Grameen housing portfolio is at $3.3 mil- lion, with 89 percent repayment rates. So far, it has financed 674,435 homes. Loans are usually for very small amounts,2 maturity is only a few years, interest rates are relatively high, and repayment is guaranteed by a community-group approach. Collateral is typically not required. Repayments are made weekly. The challenges facing housing microfinance programs include affordability constraints, especially for rural households; high land prices in the case of urban clients; commercial viability of the microfinance lenders; and need for new 33 Expanding Housing Finance to the Underserved in South Asia 5 products, including savings-for-housing instruments, and for mechanisms limiting disaster and disability/death risk in housing lending. Careful blending of government policies, smart subsidies, planning, public-private partnerships, and technical assistance for housing microlending would be required to further develop the low-income housing and housing finance markets. Success stories, such as the new low-income housing lending mechanisms in India, would provide useful models as well. Afghanistan's MFIs are still gathering experience, detailed knowledge of the housing microfinance market, capacity in loan appraisal tech- niques, effective guarantee mechanisms to limit risks from mortgage lending, lending methodologies and policies designed for high-risk areas, and cost-effective ways to reach rural clients. In contrast, a small but well-developed network of microfinance lenders in Sri Lanka (amounting to 0.1 percent of financial institu- tion assets) is currently funding housing repair and upgrades. Pakistan's MFI market has a very limited outreach at barely 2 percent of the population, and generally is not engaged in housing finance (Nenova, Niang, and Ahmad 2009). The appendixes in this volume discuss in detail the microfinance industry for each South Asian country. Affordability of Housing The lack of affordable housing is pervasive because of high interest rates, high real estate prices, high costs of building materials, few world-standard and low-cost tech- nologies, inflexible financing products, informality, and costly regulations. In Afghanistan, estimates suggest that about 80 percent of the population cannot afford new low-cost home purchase. As a consequence, housing solutions include home additions, rent, and sharing with friends and relatives for free. These figures amount to about a third of the population in Bangladesh. In India, affordability is a barrier for some 30 million households. The lowest two income segments making up two thirds of the population cannot be profitably provided with a dwelling or a mortgage in Pakistan. Likewise, approximately 40 percent of the population in Sri Lanka can- not afford low-cost housing. Subsidies for housing have been used to help house- holds acquire formal sector housing, curb the expansion of informal settlements, and upgrade existing informal housing. Low-Cost Market Solutions India's Monitor Group, in an effort spearheaded and supported by the National Housing Bank, has developed a viable model for providing housing to both formally and informally employed low-income groups. Its studies show that even at prevail- ing real estate prices, a significant portion of lower-income households may be able to afford housing. Low-cost housing--units of 200­350 square feet priced at 34 Low-Income Housing Finance 5 approximately Rs 200,000­320,000 ($4,300­$6,900)--would be affordable without subsidies to 23­28 million Indian households. Those households already pay rents comparable to the mortgage payments they would otherwise shoulder; and they live in small, one-room units that are badly designed and maintained, with poor living conditions and sometimes with shared service facilities. The obstacles to their obtaining home financing are irregular incomes, the lack of credit histories, and coordination failures preventing developers from funding low-cost construction for low-income families. Several solutions have been pilot-tested and found effective, sustainable, profitable, and feasible in market conditions. Annex C2 to appendix C presents an approach that potentially could serve 5­6 million households just above the affordability threshold (those with monthly incomes of Rs 5,000­8,000) and a further 4­5 million households in the next-higher income bracket (Rs 8,000­11,000). For informally employed groups earning more than Rs 5,000 a month, viable solutions may be provided by certain business models (such as finan- cial institutions using intermediaries like MFIs or nonbanking financial institutions to target their clients) and products (like a savings-for-housing product, a group liability product, a lease-to-purchase product, and others) (Monitor Group 2007). The Indian case also provides useful lessons for market or near-market solutions for other income groups. For Indian households just below the affordability threshold (those earning between Rs 2,500 and Rs 5,000 a month), a comprehensive housing program is suggested. The program would involve a combination of market-based solutions to lower the cost of the units (such as low-cost construction technologies and smaller housing units), policy initiatives to make the houses more affordable (such as reduction in stamp duties), and some support from the govern- ment (such as up-front cash contributions to buy down the monthly installments and partially finance up-front costs for financial institutions). Suppliers may need additional policy measures (such as measures to provide developers with access to nonprime land, with a stipulated percentage to be used for housing this segment, and tax incentives on constructing and financing such housing) to stimulate them to serve this segment. Market and near-market solutions still will leave a large number of urban lower- income groups who would not be able to afford such housing. For very-low-income levels, there is a role for government subsidies. Government programs, however, are not as efficient as could be achieved under a commercially based approach; and some programs have been poorly targeted to households with incomes higher than the lower-income groups, or have been excessively directed at civil-force staff. A more efficient approach to government housing provision for lower-income groups would be to rely on public-private partnerships through which government- provided land is competitively awarded to developers of low-cost housing, with accompanying enabling regulation and basic infrastructure service provision. The country appendixes in this volume provide a detailed review of existing government and nongovernmental organization programs for housing, housing finance, and social services related to housing. 35 Expanding Housing Finance to the Underserved in South Asia 5 Notes 1. Unless otherwise noted, all data provided in this chapter are from background papers prepared by Zaigham Rizvi in 2009. 2. For Grameen, the average loan size is about Tk 13,847, or $277. References Monitor Group. 2007. "Expanding the Housing Finance Market to Cover Lower Middle Income Segments in India." Unpublished manuscript prepared for the World Bank and the Financial Sector Reform and Strengthening Initiative, Cambridge, MA. Nenova, Tatiana, Cecile Niang, and Anjum Ahmad. 2009. Bringing Finance to Pakistan's Poor: Access to Finance for Small Enterprises and the Underserved. Washington, DC: World Bank. 36 Chapter 6 Regulatory Framework for Housing and Housing Finance Systems of housing finance cannot function efficiently without supportive mortgage market infrastructure that provides efficient mechanisms for buyers and sellers to engage in arm's-length transactions. At the same time, lenders ideally will have suf- ficient credit information on prospective borrowers and legal recourse to enforce mortgage contracts to reduce risks associated with providing mortgage finance at reasonable terms and conditions. In practice, this means that mortgage finance will be more widely available when property rights are ensured; when transferring own- ership can be done quickly, securely, and at reasonable cost; when credit information on potential borrowers is available; and when enforcement of mortgage contracts is fair and efficient. In South Asia, however, many aspects of a functional mortgage market infrastructure are missing or incomplete. The main concerns include the inability to establish clear titles to property, high costs of registering liens, and lack of secured transactions and/or foreclosure laws. Credit information is poor, and there are few housing and housing finance professionals to support market func- tioning. Each country system of policies and institutions is unique and exhibits particular weaknesses, which make it difficult to generalize the policies and reforms required. Those unique systems are described in detail in each country appendix. Overarching issues are reviewed below. Land Administration Establishing landownership, land registration, and titling remains problematic in all South Asian countries. The structure of land administration and land record systems contains weaknesses and dysfunctional elements. Inefficiency in the process of 37 Expanding Housing Finance to the Underserved in South Asia 6 property registration and title transfer matters because it can create a bottleneck in the housing finance system and can substantially increase the costs of a housing transaction. Registration of real property rights affects secured housing finance at almost every step. It enables the creditor to determine that the borrower owns the property and has the right to pledge it. And it makes it possible to identify any third- party rights to the property that might interfere with enforcement of the mortgage right. Speed, low cost, and simplicity are fundamental to an efficient system for reg- istering property and mortgages and for transferring titles, and security of land rights is the most important prerequisite for robust housing markets. Establishing and registering property in South Asia can be challenging, and trans- ferring title is often fraught with difficulties because laws and supporting regulations are complicated and outdated and frequently lack the support of an effective legal framework. The main concerns include the inability to establish clear titles to prop- erty and the high costs of registering liens. In addition, poor recordkeeping, with paper records being the norm, often leads to errors or opportunities for corruption that undermine the security of ownership. A large proportion of transactions is not recorded because of the costs involved in property registration procedures. In most of South Asia, land registration remains a lengthy, cumbersome, and costly procedure. No coherent land registration law exists in Afghanistan, and 70 percent of residents live on property that has not been registered formally.1 Fur- thermore, the titling process can take more than 30 days. The Afghani government has committed to lowering the cost of land registration (which includes stamp duty and legal and other charges) from 10 percent of the property value to 2 percent. In Bangladesh, registering property involves eight procedures, takes 245 days, and costs 10.2 percent of the property value. Registration officers are not entitled to question the validity of documents brought for registration. Furthermore, their office does not store supporting documents for verification, so even if registration officers knowingly register false transfers, they cannot be prosecuted. An average of 83 days and eight procedures are required to register property in Sri Lanka, at a cost of 5.1 percent of the property value. Registration fees and stamp duties amount to approximately $1,846, compared with $113 in India and $1,300 in Pakistan. Stamp duties alone represent 3­5 percent of the transaction price. Property registration in India involves five procedures completed in 44 days at a cost of 7.4 percent of the property value. In Pakistan, it requires six procedures, takes 50 days, and costs 7.2 percent of the value to register the property (World Bank 2008). Figure 6.1 highlights some of the bottlenecks and costs associated with property registration in the region. In Organisation for Economic Co-operation and Develop- ment countries, property registration costs 4 percent of the property value and takes 25 days; but in Pakistan, India, and Bangladesh, costs are more than double that amount and the time required to register is almost double that in most countries (World Bank 2008). Even within one country, it is not uncommon for property registration processes to vary from region to region. In the case of India, time and cost to register property dif- fer from city to city. In Gurgaon, it takes only 26 days and 7.7 percent of the property 38 Regulatory Framework for Housing and Housing Finance 6 Figure 6.1 Cost and Time to Register a Property in South Asia a. cost of registering property b. days to register property cost of registering property (% of property value) number of days to register property 12 300 10.2 245.0 250.0 % of property value 10 250 number of days 8 7.2 7.4 200 6 5.1 5.6 150 4.6 4.8 4.0 105.9 4 100 83.0 44.0 50.0 2 50 25.0 5.0 0 0 Pa ia ut a ng ia an So nka Pa ia n ng ia sh Sr an h an CD CD Sr pal l k ta Af des pa d Ba As As d an de ist t ist In In kis kis OE OE a Ne Ne iL an iL la la an h h ut gh gh So Ba Af country/region country/region Source: World Bank 2008. Note: OECD = Organisation for Economic Co-operation and Development members. value to register property--easier than in Guwahati, where it takes 84 days and 15.4 percent of the property value to do so. If all cities were to adopt India's best prop- erty registration practices, India's global ranking would improve by 35 ranks (to 70 out of 181 economies) on the ease of registering property, overtaking Hong Kong, China, and Kuwait. Computerization and proper staffing are most responsible for shortening delays. If the registration system is unified across offices, as in Jaipur, the encumbrance search and sale deed registration can be done in any subregistrar's office, regardless of the office's jurisdiction. Offices that complement technological improvements with staff training and performance management reap the fruits of their efforts. Gurgaon, India's top performer for registering property, is a case in point. The Haryana Urban Development Authority, in charge of title transfer, hired a consulting company to develop customized software and train one employee from each unit. The authority now requires all its employees to be proficient in the new computer system. As a result, its offices take only 15 days to transfer a property title (World Bank 2008). The number of institutions and registration procedures required to execute property transactions in Pakistan also needs to be reduced. Multiple institutions and procedural differences between and within states create market distortions and inef- ficiencies in the property development market. The distortions result in improper registrations and a culture favoring informal property transactions--an obstacle to effective liens and asset securitization. Administrative Efficiency of Land Offices and Cadastral Records Jurisdiction over land matters is often diffused under several government agencies, making simplification and improvement of procedures difficult. In Bangladesh, the matters fall under the purview of both the Ministry of Land and the Ministry of Law, 39 Expanding Housing Finance to the Underserved in South Asia 6 with little coordination between them. The land administration system is obsolete in India; and duplication of functions among the land revenue department, the survey and land records department (now increasingly integrated with the land revenue department, at least at higher levels), local bodies, and the stamps and registration department increases transaction costs and introduces a major source of tenure inse- curity because of the potential inconsistency of records. The present system of land records in Pakistan is fragmented, being administered by various landowning agen- cies at the provincial, district, and city government levels, and by autonomous land development authorities. The multiplicity of land administrators makes land admin- istration complex, function duplication increases cost, and the lack of coordination creates ample space for errors. To make land disputes worse, cadastral maps may not exist or may not be reli- able in South Asia. In Afghanistan, cadastral maps exist for only about 800,000 properties. In all countries, land records could benefit from computerization. Out- dated processes, corruption, and irregularities characterize the Bangladesh land survey process. A Land Registration Act was put into force on July 1, 2005, to help reduce false and multiple registrations of land. A pilot project to computerize land records has been completed in the Demra region of the Dhaka district, and is planned for replication elsewhere. In India, spatial records were not regularly updated, have considerable coverage gaps in both urban and rural areas, and do not follow common standards across the country. Their physical condition is precari- ous. Unless alternative options, such as use of satellite imagery, are explored, recon- stituting village maps in a vast plurality of areas remains a daunting proposition. In a bid to streamline property registration, selected cities have computerized their land records, making it easier for entrepreneurs to search for encumbrances and register their sale deeds. In Pakistan, the governments of Punjab and Sindh have started computerizing files, and pilot tests are being implemented in Balochistan as well. A nationwide land record system would be the ultimate goal. Sri Lanka is in the pilot stage of implementing a title registration system and a computerized cadastre in a few selected jurisdictions.2 Property Taxation The property tax framework is weak in South Asia, where high taxes for land regis- tration and transfer discourage formal property declaration or promote underre- porting of property prices. Registration fees in Bangladesh recently were cut by 1.0 percent, although they remain high overall at 10.2 percent. Simple reforms can have a big impact. Bangladesh can encourage formal property transactions by fur- ther reducing the registration fees and stamp duties (3 percent and 5 percent of the property value, respectively). By contrast, the stamp duty in Pakistan is 3 percent and the registration fee is 1 percent of the property value. In Sri Lanka, there is no separate registration fee; it is subsumed within a stamp duty set at 4 percent of the property value. In India, the stamp duty in some states is as low as 5 percent of 40 Regulatory Framework for Housing and Housing Finance 6 property value and the registration fee can be only 1 percent. Reductions in fees typ- ically lead to more revenues because more properties are registered and there is less underreporting of property values. Property tax is the main source of local govern- ment revenues in most good-practice countries (World Bank 2008). The stamp duty to be paid for registering deeds in India ranges between 3.0 per- cent (New Delhi) or 4.0 percent (Ranchi) and 12.5 percent (Kochi) of the prop- erty's market value; it is often somewhat higher in cities and municipal corpora- tions than in rural areas. Several states have cut their stamp-duty rate: Delhi (from 8.0 percent to 6.0 percent for men and from 6.0 percent to 4.0 percent for women), Gujarat (reduced to 4.9 percent in 2006), Orissa (from 11.0 percent to 5.0 percent in 2008), and Uttar Pradesh (reduced to 5.0 percent in 2005). In addition, a regis- tration fee that ranges between 0.5 percent and 2.0 percent of the market value must be paid. Deed registration costs, charged by subregistrars' offices in most cities, also vary significantly: from 1.0 percent of the property value in Ahmedabad to 8.5 percent in Guwahati. Registration of mortgages requires payment of the reg- istration fee plus a share of the mortgage value (0.1­2.0 percent). High stamp duties discourage registration. Despite recent reductions of stamp duty by some states, rates remain very high by international standards (having an average of only slightly below 10 percent of the value of the transaction). This discourages formal land transactions and leads the parties involved to avoid registration through such means as powers of attorney and other side agreements, thereby reducing govern- ment revenue and threatening to undermine the completeness of the land admin- istration system (World Bank 2008). In Pakistan, in response to falling property tax revenues resulting from poor enforcement and the prevalence of cash transactions, some provinces have initiated reform in improving enforcement and lowering fees. The Punjab government expe- rienced a marked increase in revenues after decreasing property rates in 2003­04. Multiple Legal Regimes The coexistence of modern, Islamic, and traditional laws in South Asia also poses landownership problems. The inefficiency of modern laws discourages property reg- istration. In Pakistan, where an oral gift of immovable property under Islamic law commands widespread juristic acceptance, the courts often have insisted on rigorous evidence that such a gift has been made. Another feature of the law that adds to uncertainty of ownership rights is the acceptance of the concept of benami (absen- tee) ownership. This acceptance leads to the claim that the ostensible owner of a property is not, in fact, the real or beneficial owner. Such a claim inevitably results in complex litigation, with the so-called real owner seeking to avoid the transactions entered into with the ostensible owner. In Afghanistan, land title has been complicated by several incompatible legal regimes in place: customary law (Rawaj), civil law (Qanoon Madani), religious law (Shar'ia), and statutory or national state law. In addition, frequent regime change 41 Expanding Housing Finance to the Underserved in South Asia 6 has led to more than 60 different land laws and amendments to the state law. This lack of clarity regarding the legal structure makes the establishment of a clear title to the land a major problem, further complicated by the past quarter-century of conflict. Corruption and forgeries often lead to several competing claims to the same piece of property. In Kabul, an estimated 70 percent of residents live on prop- erty that has not been registered formally. Simple, affordable procedures for property transfer and registration would encourage people to use formal registra- tion, limiting the opportunity for corruption. A recent move toward formalization has been successful and quickly should help improve the transparency of the secondary property market. The government has committed to lowering the cost of land registration (stamp duty and legal and other charges) from 10 percent of the property value to 2 percent. In addition, an effort has been made to reconstruct and improve the land records, which nearly were eliminated during the past 25 years (World Bank 2008). Land Titling Without a regime ensuring security of landownership, banks are unwilling to lend in the housing sector; and other private enterprises, such as developers, are reluctant to do housing business. In Bangladesh, land title authentication is much easier in the case of government-owned land, whereas the verification of title for private land is rather cumbersome and time consuming (60 days). Titling is even more complicated in the case of new developments where property sales are made before construction is completed (World Bank 2008). In India, as elsewhere in South Asia, there is no entity that provides indemnity for its opinion on property titles, although plans are being taken forward in this regard. There is general difficulty in obtaining information, which is costly and time con- suming. The concept is widely accepted in the developed markets and would be a valuable tool for enhancing the credit quality in India's growing housing finance market. The land titling system in Pakistan is rendered ineffective by several problems, including the lack of a sole land allotment authority, poor coordination among dif- ferent recordkeeping agencies, divergent recordkeeping systems and methods in different departments, and manual documentation without safety and security of records. Both the lack of transparency and prevalent incentives to take bribes lead to avoidance of official land registration. Poor Capacity of Registration/Titling Professionals Poor capacity of registration and titling professionals or specialists impedes the improvement of South Asia's mortgage market infrastructure. Few personnel have the necessary skills, capacity, and training to carry out procedures related to land titling and registration. Even when trained professionals are available, they do not 42 Regulatory Framework for Housing and Housing Finance 6 have the necessary infrastructure to do the job well. In Bangladesh, for instance, there currently is no single repository responsible for gathering, processing, and dissemi- nating information relating to housing and housing finance. Bangladesh Bank gathers some data on housing finance from the country's banks and other financial institutions. However, crucial databases of housing, real estate and land pricing, construction costs, and housing finance information used by modern housing finance instructions are not available in any South Asian country. Foreclosure Lenders need adequate information on prospective borrowers to make informed decisions, and ensuring that creditors are protected in the event of nonpayment is an important factor in the health of a mortgage market (even more so when housing loans carry longer terms). Without these key features, mortgage lending can be extremely risky and mortgage markets may be thin or nonexistent. Credit informa- tion on borrowers and mortgage foreclosure laws and processes are key to the availability of housing finance products. An inefficient foreclosure law that translates into delays and outcome uncertainty increases lender costs for offering mortgages. When lenders adjust for these costs, the outcomes may be higher interest rates, larger downpayments, or credit rationing. These costly outcomes for borrowers can be avoided, in part, with more efficient foreclosure processes that balance the rights of borrowers and creditors. Faster fore- closure processes can mean greater certainty in realizing collateral rights, and there- fore may lower the risks of lost value for the creditor. More rapid foreclosure also means lower costs and thus greater proceeds from the sale of a home. Enhanced proceeds benefit not only the creditor and the debtor, but also the government through greater tax revenues. In many South Asian countries, foreclosure is stigmatized as an inhumane proce- dure that puts defaulting homeowners on the street. Developed rental markets might address such concerns. In the experience of Bangladesh and Pakistan, the mere threat of foreclosure is sufficient to motivate payment from potentially defaulting borrow- ers. In several countries (such as Bangladesh), the state-owned housing bank enjoys special foreclosure rights, which create an uneven playing field and distort the market. National foreclosure provisions are discussed in each of the country appen- dixes in this volume. Time required to foreclose on a property in selected South Asian economies can be fairly short in best-case scenarios. For example, assuming the borrower does not contest and voluntarily concedes the property, foreclosure can be accom- plished in up to 187 days in India, 131 days in Pakistan, and 71 days in Sri Lanka (figure 6.2), partly because these countries allow for a nonjudicial process that gives the creditor the power to sell pledged property without court intervention. Under a power-of-sale procedure, a creditor may proceed directly to the sale of a 43 Expanding Housing Finance to the Underserved in South Asia 6 Figure 6.2 Time to Foreclose on Property in South Asia Sri Lanka 15 56 country Pakistan 58 73 India 60 127 0 50 100 150 200 days time for notification time for enforcement Source: World Bank 2008. property by auction or private transaction after a notice to the debtor. All too often, however, disputes stretch interminably, paralyzing the efficient functioning of the foreclosure law. In South Asia, banks willing to provide a mortgage loan also may face discourag- ing and costly obstacles to enforcing the loan's security. Currently, in spite of recent improvements, foreclosure remains a complex, expensive, and uncertain procedure. Costs to foreclose on a property vary from country to country: from 2.2 percent of property value in India to 6.4 percent and 12.3 percent in Sri Lanka and Pakistan, respectively. Afghanistan did not have mortgage or securities laws until 2009, and the enforcement of both laws is still pending (World Bank 2009). In Afghanistan, it costs 3 percent of the value of the contract (the loan amount) to execute the collateral registration and another 3 percent to release the property from registration. The government recently committed to reducing the cost of lien registration from 6.00 percent to 0.50 percent (0.25 percent at registration and 0.25 percent at release), which should encourage financial institutions to formalize collateral registration. Afghani judges and lawyers do not have good training in property law. Tribal or informal means of dispute resolution are used in most cases (World Bank 2009). In Bangladesh, the Money Loan Court Act 2003 set up a special court to deal exclusively with loan defaults exceeding Tk 500,000 (about $7,250), prescribed time limits for granting judgments, imposed restrictions on appeals, and provided judiciary support to lenders when evicting convicted defaulters. As a result of weak enforcement of the foreclosure framework, the procedure remains complex, lengthy, and ultimately ineffective. The situation is very similar in Pakistan, where the Financial Institutions (Recovery of Finances) Ordinance, 2001 provided for a non- judicial foreclosure, but where the law is not being enforced uniformly because it has not been upheld in some state lower courts. Lawyer fees in Pakistan commonly 44 Regulatory Framework for Housing and Housing Finance 6 amount to almost 5.0 percent of property value, compared with 1.9 percent in Sri Lanka and 0.3 percent in India (World Bank 2008). In contrast, India successfully improved its foreclosure procedure in 2003 when an enabling law was passed to facilitate out-of-court settlement--the Securitiza- tion and Reconstruction of Financial Assets and Enforcement of Security Interest Act 2002. One of the objectives of the law was to remove recoveries from the courts' domain. A creditor is now allowed to acquire the mortgaged property if a defaulter fails to pay within 60 days of being informed of possible foreclosure/ auction. Appeals may be made only to a specialized Debt Recovery Tribunal within 45 days. After the new foreclosure law took effect, the time needed to enforce a security was considerably reduced: from worst-case scenarios averaging 10­20 years to roughly 7 years with full delays and to six months in best-case scenarios. A non- judicial foreclosure also was instituted in Sri Lanka by the Debt Recovery (Special Provisions) Act No. 2 of 1990 and the Recovery of Loans by Banks (Special Provi- sions) Act No. 4 of 1990. The nonjudicial foreclosure excludes smaller loans, which limits its effectiveness. Furthermore, the eviction powers of lenders remain weak. As a result, the nonjudicial foreclosure is used mostly as a threat--and an ineffective one at that. Despite a few improvements, lengthy and rather defective foreclosure and eviction processes represent a major hurdle in the growth of the mortgage busi- ness in South Asia. Foreclosure procedures require an overhaul, transforming them into a quicker, and more simplified process of collateral disposal in case of default. Credit Information Readily available and reliable credit information is essential for the growth of a healthy mortgage market. Credit registries allow lenders to acquire important data on the reliability of borrowers, thus ensuring sustainable business operations. Most countries in South Asia have a credit bureau collecting borrower information (although Afghanistan is a notable exception). Most of the data are not computer- ized, allowing the possibility of error as well as rendering information dissemina- tion rather cumbersome and fraught with delays (table 6.1). Most credit bureaus cover only larger loans, and they may not collect both positive and negative infor- mation about borrowers. Similarly, coverage does not include information from retailers, trade creditors, utility companies, and nonfinancial institutions. Not much historical data have accumulated because most South Asian credit bureaus have been overhauled only recently. The quality of information can be further enhanced by giving borrowers the legal right to inspect and correct their data. South Asian credit bureaus need to be improved significantly. They are of little use to banks and other financing institutions if the information recorded is inaccurate, unreliable, or insufficient. 45 Expanding Housing Finance to the Underserved in South Asia 6 Table 6.1 Typical Data Collected by Credit Bureaus in South Asia Country Credit Information Collected Afghanistan None; there is no credit bureau. Bangladesh The public credit registry collects data on loans exceeding $800 (excluding microfinance institution, utility, retailer, and trade creditor data). The data are kept manually and extend only two years back. India The private Credit Information Bureau of India Ltd. covers more than 78 million individual borrowers (10.2% of adults) and more than a million firms. Data collected are both positive and negative, without a loan or other size limit (excluding microfinance institution, utility, retailer, and trade creditor data). The data are computerized and go back several years. Pakistan The public Credit Information Bureau collects data on both consumers (at most, 12% of adults) and firms (estimated at up to 2 million), without a loan or other size limit (excluding microfinance institution, utility, retailer, and trade creditor data). The data are computerized and go back several years. Three private credit bureaus--DataCheck Ltd., News-VIS Credit Information Systems, and ICIL/PakBizInfo--have opened recently, with narrower coverage than the public bureau. Sri Lanka The public Credit Information Bureau covers 80 percent of all loans by value, on individuals and companies. Both positive and negative data are collected on performing loans exceeding SL Rs 500,000, on 90-days-overdue loans exceeding SL Rs 100,000, and on 90-days-overdue credit card balances exceeding SL Rs 5,000 (excluding microfinance institution, utility, retailer, and trade creditor data). The data are computerized, although manual reports are used, and they go back several years. Source: Zaigham Rizvi, 2009. Notes 1. Unless otherwise noted, all data provided in this chapter are from background papers prepared by Zaigham Rizvi in 2009. 2. In developed countries, a cadastre is widely used as a fundamental source of data in disputes and lawsuits between landowners. References World Bank. 2008. "Financing Homes: Comparing Regulations in 42 Countries." Washington, DC. ------. 2009. Doing Business 2010: Reforming Through Difficult Times. Washington, DC. http://www.doingbusiness.org/. 46 Chapter 7 Policy Options to Improve the Framework for Housing and Housing Finance Although economies in South Asia share a common need for expansion in the hous- ing and housing finance markets, each country is at a very different level of market development; and specific policy directions have been suggested in each country appendix to address carefully the specific areas of importance in each system. Four common priority topics emerge across the region and are noted below: (1) land administration, (2) market information provision, (3) access to long-term funding, and (4) servicing low-income groups through innovative housing finance solutions. Efficient Land Administration A number of steps remain to be taken to improve landownership, registration, and titling processes in South Asia. Clear and consistent laws and regulations on land reg- istration and titling must be put in place and enforced. In each country, a designated authority should be assigned to take charge of dealing with land-related regulations and titling procedures. The titling processes need to be made less costly and more efficient because cumbersome procedures involving many steps lead to corruption and inefficiency. Stamp duty and legal charges should be reduced significantly. Land registration should be carried out in a transparent manner with clear recordkeeping. Computerized information and registration systems, property and ownership data- bases, and cadastral surveying would increase transparency and the enforcement of creditor rights. Land transfer procedures also need to be revised, simplified, and homogenized across different country regions. As land administration systems become simpler, cheaper, faster, and more reliable, an increasing number of transac- tions will be formalized, thus eliminating uncertainty in property ownership and easing ambiguities resulting from application of traditional and Islamic laws. 47 Expanding Housing Finance to the Underserved in South Asia 7 Housing and Housing Finance Information and Transparency The availability of ready and easy-to-consult information on property prices and real estate data; housing supply and demand; construction; and various housing finance data and pricing information, including credit records, is indispensable for the smooth functioning of the housing and housing finance markets. It is important to provide periodic surveys to the market, including collection, analysis, interpretation, and publication of data on the performance of the housing and housing finance sec- tors. The data should be electronic and available by subscription to all paying par- ties. Relevant information to collect includes · housing supply information, such as new construction and sales, vacancies, qual- ity and condition of property, prices, and dwelling characteristics; · housing demand data, such as demographic trends and housing needs, charac- teristics of occupants of the existing stock (crowding indicators), affordability (availability of income indicators are of paramount importance for this), and buyer's intentions; · housing finance data, such as sensitivity of monetary indicators to residential and commercial property exposures of lenders and investors (for example, inflation rates, interest rates), fiscal indicators and degree of reliance on housing-related developments (such as property taxes, construction industry contribution to GDP and employment), lending volumes by category of financial institution, loan characteristics (loan-to-value ratios, debt service ratios, prices, locations, borrowers); and · construction input data, such as production capacity, price of materials, labor and land, and availability and quality of materials. Long-Term Funding Availability through Mortgage Refinancing South Asian housing markets lack long-term funds. This paucity of funds puts non- banking financial institutions on an unequal footing with banks, which can benefit from their deposits as a source of cheap funding. To address liquidity and risk of term mismatch, international best practice suggests the creation of a mortgage refi- nance facility, a type of housing fund that takes mortgages off the balance sheets of financial institutions and injects the required liquidity so further mortgage lending is possible. The same effect can be achieved by securitization in countries where the institutional and developmental prerequisites for securitization are present. A mort- gage refinance facility also prepares the market for securitization in the long run. A secondary mortgage facility is a financial institution designed to support long- term lending activities by primary mortgage lenders. Its core function is to act as an intermediary between mortgage lenders and the bond market, with the objective of providing long-term funds at better rates and under better terms and conditions 48 Policy Options to Improve the Framework for Housing and Housing Finance 7 than lenders might be able to obtain if acting alone. The liquidity facility would be able to issue its own private bonds, thus separating market long-term debt from interest rate subsidies and enabling more efficient and transparent targeting of the subsidy. A facility could refinance all eligible lenders (both banks and finance com- panies) that provide as collateral high-quality mortgage loans. The facility would issue low-risk private bonds among institutional investors. This mechanism would centralize funding requirements, catalyze the primary market for housing finance, provide matching funds at an acceptable cost, encourage discipline by requiring higher lending and servicing standards, and provide some form of institutional sup- port to help build the link to securitization in the long term.1 Expanded Housing Finance to Low-Income Groups through Product Innovation In view of the very low incomes and widespread inability to afford housing in South Asia, expanding access to housing finance to lower-income groups will remain the main challenge. The customer base will broaden naturally in response to reduced intermediation costs, more competition, more efficient institutions, fixed (and lower) interest rates, and longer loan maturities. Tax-deductible mortgage payments would make mortgages further affordable to the middle-income groups. Strength- ening the enforceability of mortgage rights would have the additional effect of extending mortgage finance to lower-income groups because financial institutions would become less sensitive to client income and would rely more on the value of the enforced collateral instead. Pure market-based solutions are likely to work for many currently underserved lower-income customers, as experience in India shows (annex C2 of appendix C). The "employer-facilitated" model is a fundamental business innovation that makes the provision of low-income housing attractive to the private sector. Alternative models are required for the informal sector (self-employed and salaried unorganized employees). Further innovative products and instruments include the contractual savings scheme, mortgage product customization, reengineering of lending processes to allow for better risk mitigation and client segmentation, an interest risk mitigation fund, and a financial guarantee facility (see the "Policy Options to Improve the Framework for Housing Finance" section of appendix C). Significantly scaling up South Asian housing finance markets crucially hinges on product innova- tion, which would bring down lending costs, customize loans to client needs, enable financial institutions to better manage risk, and help extend supply farther down the income scale. Innovations and rapid scaling-up of housing finance markets require active capacity-building efforts for financial institutions, developers, regulators, and other housing market participants. These efforts could include streamlining for small-mortgage processing, new product development, demand aggregation, risk assessment, loans servicing, consumer information and protection, and defined business standards. 49 Expanding Housing Finance to the Underserved in South Asia 7 Microfinance will be of central importance in extending the outreach of the housing finance market, particularly in rural areas and for house upgrades and maintenance in urban slums. To solidify the link between microfinance and housing, the policy maker would need to tackle three challenges: (1) improving the commer- cial sustainability of microfinance housing products, (2) strengthening the microfi- nance regulatory framework, and (3) extending microfinance for housing to urban households. Below a certain income and affordability level, pure market-based solutions will not be enough; and home ownership will need to be promoted through a combina- tion of market-based levers, policy initiatives, and direct support from the govern- ment. When even those measures are inadequate, rental housing remains the only solution. Detailed studies of the existing rental framework are required, with a view to streamlining efficiency and protecting both tenant and landlord. Some of the more effective levers that could be used to bridge the affordability gap are smaller- size units, low-cost technologies, standard construction norms, tax exemptions, credit guarantees, and land policies geared to low-income groups. Targeting the lower-income groups where unsubsidized housing is not affordable would require assistance programs that target bankable lower-income households and a scheme designed to leverage public spending by market resources. The most efficient subsidy schemes are those that are directly allocated to the beneficiaries and that include a preliminary savings requirement as an eligibility criterion to trigger a leverage effect with private resources. The assistance scheme could be linked to the savings-for-housing scheme, given the synergies between the two mechanisms. Savings-for-housing schemes would go a long way toward spreading housing finance outreach by helping lower-income savers in building up downpayments. Islamic finance can play a considerable role in enlarging housing finance access because it often targets customers with lower-than-average incomes. Note 1. A best-practice example is that of Malaysia's national mortgage corporation Cagamas Berhad. 50 Appendix A 1 Afghanistan Afghanistan's conflict-ridden past and rapid current growth in both income and popula- tion pose unique challenges for the housing and housing finance sectors. The population is expected to grow to 37 million by 2015. Urbanization is high at 24 percent, and returning refugees are exacerbating the housing gap. The estimated housing shortage presently stands at 1.0­1.5 million, 70 percent of which is in urban areas; and it is expected to reach 1.5 million by 2015. The rehabilitation and new construction in Kabul alone is estimated at $2.5 billion of fixed capital investment. A large part of existing housing is in need of an upgrade. Informal settlements shelter 80 percent of the popula- tion and cover 69 percent of the residential land in Kabul. The quality of housing stock is poor and construction standards are weakly enforced. As evidenced by a shortage of 1.0­1.5 million units, both public and private housing provision has been lagging. Limited government capacity and an insufficiently supportive legal framework for private sector involvement have been responsible for this shortcoming. Low-income housing cost is high because of costly imported materials, unskilled labor, and a lack of low-income developers jeopardizing affordability. Although the financial sector in Afghanistan has experienced considerable growth since 2004, long-term and housing finance has remained limited. Afghanistan's housing finance-to-GDP ratio is less than 1 percent, compared with 50­70 percent in developed countries and with 7 percent in India. Most of the secondary market transactions in housing are for cash. The banks' lending ability is restricted to short maturities of three to five years for housing finance. Whereas the banks have excess liquidity, the tenure of their liability book does not permit them to create long-term assets. The financial system has experimented timidly with traditional and Islamic housing finance products and with developer financing, but penetration and affordability are limited to the upper- income bracket. Diversified products, including home insurance, are rare or nonexistent. The vast majority of the population relies on informal sources of finance--friends, 51 Appendix A A relatives, and moneylenders--as does more than 85 percent of the economic activity in Afghanistan. These considerations suggest a microlending approach, mostly oriented toward home improvement loans. The microfinance sector is making initial expansion efforts into low-income housing, relying on its comparative advantage of experience and knowledge of the borrowers. The sector is also uniquely suited to serve women, who have limited access to formal finance because they lack guarantees and title deeds and because of various cultural factors. Potential demand is considerable. Estimates suggest a total financing need of $276 million a year for home improvement and new construction in Kabul's informal housing sector. Future growth of the microlending sector might create funding bottlenecks, pushing microfinance institutions toward market funding and commercial sustainability, and possibly toward a second-tier institution for provision of long-term funding. Lack of an appropriate legal and regulatory regime has prevented commercial banks and, to a lesser extent, microfinance institutions from entering the housing finance mar- ket. Inadequate enforcement of laws and structural weaknesses in the housing market have made housing finance an overly risky proposition. Among the reasons commercial banks cite for staying out of the housing market are the inability to establish clear title to property, the high cost of registering liens, and the lack of secured transaction or foreclosure laws. Much work is needed to promote an enabling environment for housing and housing finance supply and demand. Among the key policy measures for housing are those needed to improve the security of ownership, titling, and registration, and to build the capacity of secondary participants in the housing market. Innovative mortgage products better suited to different market segments and risk profiles are essential, as are capacity building in the financial sector, expanded access to finance, and a source of long-term funding such as a liquidity facility. A regulatory framework for property insurance, life insurance, and (in the medium to long term) mortgage guarantees and mortgage insur- ance should be developed. Microlending could be promoted through a joint public- private effort. Finally, the central bank's supervisory and regulatory framework needs to be strengthened. Overview The Afghan economy has sustained strong GDP growth in the past few years, with per capita annual income rising from $123 in fiscal 2001/02 to $300 in fiscal 2005/06 (fig- ure A.1).1 Inflation is in single digits after a sharp spike in fiscal 2007/08 (figure A.2). Interest rates on central bank securities have been in the single digits. Projected annual GDP growth for the medium term is 10­12 percent, which would lead to a steady rise in per capita annual income to about $482 by 2010. According to United Nations' esti- mates, between 2000 and 2015, the population is expected to increase by 14 million (from 24 million to roughly 37 million people). Rapid population growth and urban- ization compound the housing problem in Afghanistan, especially in such cities as Jalalabad, Kabul, Kandahar, and Maser-e Sharif. Of the country's total population-- 52 Appendix A A Figure A.1 Growth of Nominal GDP, 2003­07 20 15 percent 10 5 0 FY2003 FY2004 FY2005 FY2006 FY2007 year Source: Data from the Central Statistics Organization, Afghanistan. Figure A.2 Rates of Inflation, 2003­07 25 20 15 percent 10 5 0 FY2003 FY2004 FY2005 FY2006 FY2007 year Source: Data from the Central Statistics Organization, Afghanistan. 23.6 million, according to the population census of fiscal 2005/06--an estimated 24 percent lives in urban areas and 76 percent in rural areas (2008 estimate). More than half of this growth will be in urban areas, where the housing deficit is already severe (Bedforth 2007). The country also has the youngest population in the world, with an estimated 57 percent under the age of 18. As these young people start to seek jobs and places to live, pressure on the urban housing sector will increase. The rising income lev- els may provide opportunities for a growing housing market, as long as the right kinds of policies and interventions are put into place (World Bank 2008). A third of the country's population is concentrated in 5 of its 34 provinces: Kabul (3.07 million), Herat (1.55 million), Nangarhar (1.26 million), Balkh (1.07 million), and Kandahar (1.00 million). These provinces have large urban con- centrations and perennial housing shortages. Estimates of Kabul's population vary significantly--reports from the Afghanistan Research and Evaluation Unit put the 53 Appendix A A Kabul population in excess of 3.0 million in 2005. Population growth is now estimated at approximately 5 percent. This rapid growth and urbanization result in substantial migration into informal settlements, and they challenge urban plan- ning as the city expands in an ad hoc manner. In spite of strong financial sector growth, financial services in Afghanistan still have very limited outreach (figure A.3 and table A.1). Performance and efficiency indicators have been weak, although financial stability measures have remained strong (tables A.2 and A.3). The housing sector has almost no access to bank credit, Figure A.3 Bank Assets, Liabilities, and Capital, 2004­09 0.0030 0.0025 0.0020 US$ millions 0.0015 0.0010 0.0005 0 2004 2005 2006 2007 2008 2009 year assets liabilities capital Source: Data from Da Afghanistan Bank, 2009. Table A.1 Financial Access, 2009 Indicator Afghanistan Nepal Pakistan Sri Lanka Demographic branch penetration 0.6 6.0 15.0 29.0 (branches per 100,000 people) Demographic ATM penetration 0.09 13.00 16.00 30.00 (ATMs per 100,000 people) Deposit accounts per 1,000 people 26 6 12 30 Loan accounts per 1,000 people 2 7 11 30 Geographic branch penetration 0.031 6.000 12.000 18.000 (branches per 1,000 sq. km.) Geographic ATM penetration (ATMs 0.04 12.00 12.00 30.00 per 1,000 sq. km.) Source: Data from Da Afghanistan Bank, 2009. Note: ATM automated teller machine. 54 Appendix A A Table A.2 Bank Performance, 2009 Indicator Afghanistan Nepal Pakistan Sri Lanka Return on equity 10 6 15 29 Return on assets 2 14 25 20 Source: Data from Da Afghanistan Bank, 2009. Table A.3 Financial Stability, 2009 Indicator Afghanistan Nepal Pakistan Sri Lanka Capital adequacy ratio 32 6 25 20 Nonperforming loans 0.7 7.0 23.0 19.0 Liquid assets ratio 25 6 25 13 Source: Data from Da Afghanistan Bank, 2009. and mortgage finance is very limited. In Doing Business 2010: Reforming Through Difficult Times, the World Bank ranked Afghanistan 168th among 183 economies on the ease of getting credit, a measure that reflects the quality and coverage of the credit information system and the legal rights of borrowers and lenders.2 The country has 2 state-owned and 14 private commercial banks now in operation, and 15 microfi- nance institutions providing financial services to urban and rural borrowers. But most banks avoid lending for housing, which typically involves a long-term com- mitment to the client. Among the main reasons for this avoidance are the lack of an adequate supporting legal framework, including poor foreclosure and eviction pro- cedures; poor land titling regimes; absence of reliable property valuation/appraisals; lack of a well-functioning collateral system; and poorly trained judges. Despite its recent economic growth and its maintenance of relatively low infla- tion, Afghanistan continues to face acute poverty. Meeting the challenges of allevi- ating poverty, correcting poor social indicators, and improving the security situa- tion are the government's main fronts. Doing so will require sustainable economic growth, continued macroeconomic stability, and a more developed financial sector. Development of the housing and real estate sector is globally acknowledged to be a main contributor to industrial growth and employment and an important asset class in the financial sector. Housing and housing finance build assets and liveli- hoods, and thus contribute to poverty reduction. They also stimulate savings and can be used as collateral for loans by home entrepreneurs. The share of mortgage debt to GDP in developed economies like the United States (65­70 percent), the European Union (42 percent), Canada (40 percent), China (15 percent), and Malaysia (26 percent) is considered to be an important indicator of the roles of housing and housing finance in economic development. The figure for Afghanistan is insignificant at less than 1 percent. 55 Appendix A A Housing Demand Potential housing demand is considerable. A long history of conflict, high popula- tion growth, rapid urbanization, and returning refugees results in housing shortages and a large amount of illegal and overcrowded housing without access to basic city services. Housing Shortage More than 6 million displaced persons are putting upward pressure on demand for housing. Returnees find their houses badly damaged by war and repeated earth- quakes and their land in the hands of others--either legally appropriated under suc- cessive governments or illegally grabbed. In Kabul, an estimated 700,000 returnees have settled since 2001 (UNHCR, Pakistan Population Census Organization, and Pakistan Ministry of States and Frontier Regions 2005). The Ministry of Urban Development estimates the current need for new housing units in Afghanistan at more than 1 million, with 70 percent of these units required in urban areas. The shortfall is likely to reach 1.5 million units by 2015 (figure A.4).3 In Kabul alone, the rehabilitation and new construction may involve fixed private capital investment of $2.5 billion (excluding land value) (World Bank 2008, p. 16). Despite the housing shortages, only 0.5 percent of Kabul's population can be considered homeless in a stricter sense, including 10,000 people living in tents and 5,000 living in destroyed buildings (World Bank 2005a). Many people construct their own dwellings: in Kabul and Jalalabad, approximately 60 percent of households surveyed in an Afghanistan Research and Evaluation Unit study have built their own homes, 20 percent are rent- ing homes, and the rest are living rent free as beneficiaries of charity (Schütte 2006). People migrating to the cities--generally poor--also mostly construct their own houses: the United Nations Children's Fund estimates that approximately 80 percent of the new houses constructed in Kabul in the 1990s were self-built (Schütte 2006). Figure A.4 Projected Demand for Housing, 2006­14 2.0 housing units (millions) 1.5 1.0 0.5 0 2006 2008 2010 2012 2014 year Source: World Bank 2008. 56 Appendix A A Squatters and Informal Settlements Informal settlements provide shelter for 80 percent of Kabul's population (2.44 mil- lion people) and cover 69 percent of Kabul's residential land (World Bank 2005b). Insecure titles and poor service provision make conditions in these settlements diffi- cult. A World Bank study called for upgrading rather than demolishing the informal settlements and for regularizing the tenure of their households (World Bank 2008). Security of title also is linked to quality of housing. Insecure title may discourage households from improving the conditions of their housing--but investments in home improvement can serve as an explicit strategy to secure tenure (Schütte 2006). Even in illegal settlements, most inhabitants pay rent to someone to avoid evic- tion. Indeed, rising rents have led to the eviction of a large number of people from the slums, forcing them into even more meager shelter. In the informal settlements of Kabul (legal and illegal), it is estimated that more than two thirds of houses are occupied by more than one family.4 Almost 9 of every 10 families in informal hous- ing live in only one or two rooms (Aga Khan Trust for Culture 2006). Housing Supply Housing provision, both public and private, has been lagging. Limited government capacity and an insufficiently supportive legal framework for private sector involve- ment have been responsible for this shortcoming. Current State of Housing The condition of the housing stock in Afghanistan is very poor. Most housing units are located in unplanned areas. A considerable share is damaged or destroyed. Roads and drains are nonfunctional. A vast majority of the population has no access to safe water, sanitary toilets, and solid waste collection; and there are no landfills. Aside from some improvements to the roads and, to a lesser extent, the water system, the challenges in housing supply and infrastructure remain.5 The government is working on a national housing program to build 150,000 units. This project is in its planning stage. In the past, the government had distrib- uted 450,000 plots and expected those who received them to do construction on their own. Unfortunately, houses were built on only 15 percent of the plots; the rest either were sold or are still vacant. The Ministry of Urban Development is also plan- ning a project of 20,000 low-income developed plots for construction of housing units around the city of Kabul. Under this project (which is in an advanced stage), the government intends to provide fully developed/serviced plots, equipped with all residential infrastructure. The construction will be done under a public-private partnership model. The government intends to use a cross-subsidy model, whereby it will allocate 75 percent of the units (15,000) to the low-income segment. To cross- subsidize the low-income segment, the remaining 25 percent of the units will be sold later at the then-prevailing market price. The commercial areas will be 57 Appendix A A Box A.1 Meeting Housing Needs: Efforts from the Public Sector The Ministry of Urban Development, projecting substantial investment needs in the housing sector, has recommended designing an incentive system for developers to alleviate the urban housing shortages. It plans to build a large "new city" near Kabul (with 250,000 housing units costing $13,000 per home) to house 700,000 inhabitants. The government will donate the land to a developer, whose profit will be guaranteed by fixed prices. The project is expected to cost $250 million. Plans call for it to be completed by 2013, with a pilot site for 20,000 homes launched in the near future. In addition, the Ministry of Urban Development plans to sign an agreement with a U.S.-based housing company to construct 10,000 single-story houses for middle- and lower-middle-class inhabitants of Kabul. The land will be free of cost, and the sales price of each house will be $20,000. The ministry's efforts to promote such housing projects have been under way for the past three to four years, slowed mainly by difficulty in mobilizing financial resources. Both the Kabul and Jalalabad municipalities have implemented programs to transfer land to needy (unsheltered) families. This land is sold for prices below market value as long as the fam- ilies build houses in the short term. Kabul has distributed 22,000 plots at an average price of $5,200, with the title transferred after 40 percent of the construction is completed. Jalalabad has distributed 4,400 plots at an average cost of $5,000. Neither land transfer program has had much impact on the housing stock. One reason for that lack of effect is tenure issues: although the municipal authorities claim they are able to provide clear titles to the land, the title transfers took almost three years for many families because of squatters on the land. A second reason is real concern about ensuring that the land goes to those who are eligible rather than to spec- ulators. (The main requirement for eligibility is that the recipient owns no other land in Kabul.) The Ministry of Education plans to implement a housing incentive scheme that will encour- age female teachers to move to rural areas by giving them land. When Parliament passes leg- islation authorizing the distribution of land to teachers (which it is expected to do), the main tasks will be to identify land, review applications (from a potential pool of 181,000 eligible staff), distribute land, and provide households with access to mortgage loans to construct their homes. But availability of housing finance is a key missing piece. Housing initiatives that offer land free of cost raise some economic issues. Although the housing shortage is acute, housing development programs that provide substantial subsidy through free land are not targeted to those who most need housing support. Instead, they are directed to middle-class residents, most with businesses or salaried employment, who could be served profitably by the private sector. The subsidies could be better targeted to improving water, sanitation, access roads, and other infrastructure provision. Source: World Bank 2008. designed to cater not only to the population of the project (120,000­150,000 peo- ple), but also to the general public. The commercial areas and the community cen- ter will be sold at the going market rate to further cross-subsidize the low-income segment. The ministry estimates the subsidy amount of $12,000 per household, which is the estimated cost of developing the residential infrastructure and utilities 58 Appendix A A Box A.2 A Housing Developer in Need of Finance: Shelter for Life Working with nongovernmental organizations in Kabul, Shelter for Life, a private, voluntary organization based in the United States, has completed and distributed houses for 3,000 fami- lies. It also plans a project to provide affordable apartments for middle-class residents: a mid- rise complex (three to five stories) with 40­60 units. The Kabul municipality has offered the program land with clear title free of charge. The land is in Mikro-Rayon, an area with good exist- ing infrastructure. Shelter for Life is the lead contractor and developer for the project, and is interested in partnering with financial institutions to provide the mortgage finance to borrowers when construction is completed. Shelter for Life has community development as its overarching goal: it strives to integrate its reconstruction projects with community development programs designed to enable independ- ence and self-sufficiency. This process begins with the physical reconstruction of the commu- nity: shelters, schools, clinics, and infrastructure. The program equips people with the training, tools, and materials they need to permanently rebuild their communities. Beyond Afghanistan, Shelter for Life is working in Indonesia, Iraq, Pakistan, Sri Lanka, Sudan, Tajikistan, and the United States. Source: World Bank 2008. supply. At present, the housing finance market is not geared to provide long-term mortgages for 20­25 years, but that is essential for the success of the project. Several other public sector initiatives are in the planning and development stages (box A.1). Municipalities, the Ministry of Refugees and Repatriation, and the Ministry of Urban Development have undertaken housing construction for returnees from Pakistan and the Islamic Republic of Iran. Progress in new construction has been slow because of limited capacity, the scarcity of serviced land, and lack of funds for constructing infrastructure. Some private initiatives are under way (box A.2). Poor Construction Quality and Lack of Enforcement of Construction Standards The quality of construction has become a major source of concern in Afghanistan. The main reasons are lack of quality control in the materials sector, lack of supervi- sion, and corruption in the enforcement of construction standards. The Afghan National Standards Authority, established by the Ministry of Commerce in May 2004, is responsible for setting standards for both imported and locally sourced con- struction materials. Results will depend on enforcement of these standards, however, and the enforcement authority generally has been ineffective. The enforcement of building codes--the responsibility of municipalities--also has been poor. People building in zoned areas are required to get proper building per- mits, issued only after the municipality approves blueprints and title. In reality, there 59 Appendix A A is no oversight, and most home construction follows neither the code nor approved blueprints. City engineers often ignore code violations for a small payment. Housing Costs The cost to build a 200-square-meter apartment for upper-middle-class housing is estimated to be $41,000. The cost to construct low-income housing (estimated for a two-room dwelling) might range from $5,000 to $10,000. A microfinance institution involved in lending for housing estimated that the cost to upgrade a two-room house would be $500­$600. Overall estimates from a construction company, microfinance institutions, and commercial banks suggest costs of $100­$5,000 for home improve- ment, depending on the type of upgrading done (for example, adding an extra room, building a kitchen, building a bathroom, fixing the ceiling, fixing the flagstones, painting the home, replacing doors and windows, installing plumbing and a water supply connection, installing wiring for electricity); and they suggest costs of $10,000­$40,000 to build a new apartment or house, depending on the location (rural or urban) and the construction materials used. In comparison with home improvement and construction costs in the low-income category, these estimates appear to be on the higher side. The obvious reasons are unavailability of imported construction materials, lack of a skilled labor force, and absence of organized devel- opers for the low-income category. Again, when these costs are viewed from the per- spective of affordability, they are very high for micro- and low-income categories, and thus present a major challenge for the housing finance market. Housing Finance Most of the secondary market transactions in housing are for cash. The banks' lend- ing ability is restricted to short tenures of three to five years for housing finance. Although the banks have excess liquidity, the tenure of their liability book does not permit creating long-term assets. Because there is no developed financial market, the availability of long-term funds is the binding constraint on the banks to provide long-term housing finance. The Financial Sector Afghanistan's financial sector is developing, but lending remains very limited, con- fined to trade financing and short-term business loans to customers well known to the lender. Most banks have avoided lending for industrial financing or housing, which typically involves long-term commitments to the client. The outreach of financial services in Afghanistan remains quite limited. There has been significant growth in private commercial banks since 2004. In 2002, there were only 3 state-owned banks in Afghanistan. In 2009, there were 17 licensed banks, including 10 private and 2 state-owned relicensed commercial banks and 5 branches of foreign banks operating in Afghanistan. The banks are concentrated mostly in 60 Appendix A A Kabul; and they provide services primarily to international donors, large businesses, foreign nongovernmental organizations, and foreign government agencies. They do not have any visible role in housing finance. There is also a small but growing num- ber of nonbanking financial institutions, including a credit union and 15 microfi- nance institutions that provide financial services to urban and rural borrowers. A leasing finance company is also operating in Kabul, although with limited capital. In addition, there are 111 money service providers and 37 foreign exchange dealers licensed by Da Afghanistan Bank (DAB; the central bank).6 Although financial serv- ices are expanding, these institutions have yet to broaden their offerings to meet new demands, such as those for long-term loans. The housing sector, in particular, has almost no access to bank credit. Liquidity. At present, the banking sector in Afghanistan is composed of 10 private sector commercial banks, 2 state-owned relicensed banks, and 5 foreign banks. The banks have very nominal advances and are flushed with excess liquidity, which is either being placed interbank, in DAB capital notes, or in overseas markets. On average, the banks are two-thirds liquid; but some banks' advances stand at less than 10 percent. Creating a vibrant investment environment remains a major challenge for Afghan financial managers. On the liability side, a sizable portion of bank deposits (in some banks, up to 90 percent) is in the bank's current account to avoid interest. In some cases, this placement lowers the average cost of bank funds to below 1 percent. On the asset side, 15­25 percent of bank advances is in corporate assets--specifically not long- term productive assets, but under overdraft facilities. The excess liquidity of com- mercial banks and other financial institutions is mostly being invested in DAB capi- tal notes of up to six months' maturity, earning 13­15 percent interest. Some of these excess funds are invested in overseas markets. Because this situation has made bank- ing relatively easy and risk free, the banks are not proactive in developing the demand for funds. No commercial banks are involved in housing finance, mainly because of the unavailability of long-term funds, poor title and loan security, and lengthy court procedures for foreclosure. The banks plan to review their positions after the announcement of a long-awaited mortgage law. Long-Term Funds. At present, DAB issues no Treasury instruments, T-bills, or medium- or long-term bonds. Instead, it issues an instrument called a "capital note," which has up to a six-month maturity to help banks manage the statutory liquidity ratio and to invest surplus liquidity. Preparations to float Treasury instruments are under way, and DAB may start issuing short-term Treasury instruments by the year 2011. Therefore, it is going to be quite a few years before Afghanistan's financial mar- ket will have long-term Treasury instruments or will develop a long-term yield curve. Until the financial market has long-term Treasury instruments and a long-term yield curve, it will not be easy to raise long-term, market-based funds, which are essential for the housing finance industry. In the interim, DAB is considering promoting and supporting some capital market instruments that will be issued by the commercial banks. 61 Appendix A A Although the availability of long-term funding remains a challenge, the problem is largely not a financing issue: commercial banks held current deposits of about $1.20 billion against outstanding loans of $0.74 billion in January 2008, and they are seeking ways to safely lend their available funds. Instead, the problems are an ineffi- cient financial intermediary function for generating long-term funds and the lack of a legal regime that supports long-term loans (such as those for housing). The conse- quence is that houses are purchased on a cash basis, precluding all but the wealthy from participating in the market. In more developed markets, where the risk is clearer, the capital markets take some of the burden of long-term lending by creat- ing funding and guarantee instruments for the housing market. Efforts now to develop a secondary mortgage market in Afghanistan would be very premature. However, creating a sound primary mortgage market could help lay the groundwork for eventual secondary market activities. Ensuring clean titles and banks' ability to place liens on properties and to enforce those liens in the event of default would give eventual investors in mortgages or related financial assets the confidence that the ultimate collateral is sound and acces- sible. Similarly, standardizing mortgages would help make them more marketable or more readily securitized instruments in the long run. In Afghanistan, where a 48-month loan is considered a long-term loan with a marginal interest of 20 percent, banks will need to be able to access other funding or guarantee sources to develop a housing finance product. Traditional Mortgage Lending Housing finance, whether formal or informal, is not significant in Afghanistan. The country has no bank or financial institution providing long-term mortgage financing. Before 2002, a government-owned housing finance bank was operating in Afghanistan, but its operating performance and loan portfolio were so weak that it failed to obtain a provisional license in 2003 under the new banking law (the Law of Banking in Afghanistan). Two state-owned banks, Bank-e-Millie Afghan and Pashtany Bank, have made small forays into the housing finance market. Bank-e-Millie Afghan has no estab- lished mortgage loan product. Each of its housing loans is unique, but it generally charges interest of about 12 percent for a two-year loan, with the loan sizes ranging from $400 to $10,000. Another product designed by Bank-e-Millie Afghan would range from $10,000 to $20,000, with a monthly payment of more than $100 over a 20-year period. Most commercial banks, along with microfinance institutions, are keen to launch housing finance products in the near future. In the microfinance sector, one institution has extended loans for home improve- ments to clients in Kabul's informal settlements. The loans, mostly short-term ones ranging in size from $400 to $2,000, are intended primarily for home repairs. The financial sector needs innovative product diversification and customization. Because the client base in Afghanistan is quite varied in nature on all socioeconomic 62 Appendix A A parameters, there is a dire need to devise tailor-made housing finance products for different market segments. Traditional mortgage finance poorly suits Afghanistan's socioeconomic conditions because it requires regular monthly payments over long periods. Most low-income households are self-employed, and their incomes vary greatly. Mortgage lending requires a valid and safe title, usually not readily available, as well as standard documentation for income verification (such as a salary certificate). Banks cannot verify the income of self-employed clients. Small loans to low-income households are less profitable and they stunt banks' equity and asset growth. Banks have little capacity to make small loans to many low-income households because the term-risk problem (funding long-term assets with short-term liabilities) is exacer- bated. Amortization schedules in traditional mortgages require "tilting" real payments toward the first years of a loan--a serious problem for low-income households. Traditional mortgage finance will only reach a modest fraction of households, even when all reforms are in place. Therefore, the financial institutions, once motivated to go for housing finance, will have to design, develop, and market segment-specific products. Specialized Mortgage Finance Prior to fiscal 2001/02, one specialized housing finance institution was operating in the public sector. However, because of the institution's poor performance, it was not relicensed to resume its operations. Because the commercial banks are not proac- tively coming forward to meet the demand in housing finance, the need for special- ized housing finance is now being felt. DAB intends to set up such an institution under a public-private partnership, with management control in the private sector. The proposed institution will have a business focus on the middle and the high end of the population. With nearly 80 percent of the population poor and likely candi- dates for low-income and microfinance for housing, the possibility of setting up a separate housing finance institution to cater exclusively to the needs of the low end of the market must be explored. Islamic Mortgage Finance Being predominantly a Muslim country, Afghanistan has a great potential demand for Shar'ia-compatible housing finance products. The popular preference for Islamic banking is demonstrated also by the fact that 70­80 percent of banking deposits are kept in current accounts to avoid interest for the depositors. At the moment, some banks are operating Islamic banking windows, alongside their conventional banking operations. However, this dual practice has not proved to be successful. DAB is con- sidering inviting the private sector to set up exclusively Islamic commercial banks, and currently is drafting Islamic banking laws to allow issuance of Islamic banking licenses and to develop the regulatory regime. Shar'ia-compatible mortgages surely would provide a substantive boost for the housing finance market, as well as access to finance overall. 63 Appendix A A Not much work has been done by the banks in this area, and most Islamic finance is done on "lease-buyback" arrangements. To meet this demand, some banks provide home purchase loans that are structured not as traditional mortgages but as lease- buybacks. Under such loans, known as murabaha, the bank buys the property and resells it to the client at a markup. The bank owns the property outright until final payment is received--an arrangement that offers the bank a level of security not available with a traditional mortgage. The murabaha, which are Shar'ia compliant, may enable banks to overcome some of the problems with foreclosure by retaining title until final payment. Leasing may offer many of the same advantages, giving financial institutions a level of security unavailable with traditional mortgages in the present legal system. Large-Scale Project Finance Another opportunity is to provide construction finance to housing developers at the front end. Such finance would be fully repaid after completion of the construction through mortgage loans to the homebuyers, with the completed houses serving as collateral. So far, developers have not focused on residential projects, and are more inclined toward commercial complexes. Among the most notable construction proj- ects in Kabul are the Safi Landmark (a hotel and shopping complex) and Kabul Trade Center (a shopping and office complex). Each of these projects involved investments of roughly $20 million, and each was financed by sponsors using their own resources. Informal Housing Finance Afghan businesses and homeowners must rely almost entirely on friends, relatives, and moneylenders to meet their financing needs. Informal financing may play an important part in manufacturing and commercial activities, but such funds are often costly and therefore inappropriate for long-term private sector or housing development. The hawala system, an underground banking network based on traditional meth- ods of paperless transactions, includes informal moneylenders and could provide an efficient alternative source of housing finance. But its presence in the housing finance market is insignificant. Moreover, thanks to the recent speed and efficiency improvements in the financial system, Afghan banks are now better able to compete with the hawala system on price and service. Afghans looking for credit outside the formal system overwhelmingly turn to rel- atives and, to a lesser extent, local moneylenders. Informal credit is generally a short- term loan for consumption, to cope with temporary shortfalls in cash flow. Less than 10 percent of credit in urban areas is obtained from local moneylenders, despite the need. Clients of microfinance institutions have pointed to the high interest rates charged and the potential repercussions of delinquency as reasons that moneylenders are best avoided. 64 Appendix A A Homeowners' Insurance Afghan banks are reluctant to offer large loans without insurance on the collateral. But no company in Afghanistan currently offers homeowners' insurance protecting against fire or earthquake (a particular concern in Kabul and the surrounding region). Although the country's insurance law was revised in January 2006, only one private insurance company has started doing business. The Insurance Corporation of Afghanistan began operating in Kabul in October 2007, although it did not obtain its formal registration until February 2008. Meanwhile, the public sector Afghan National Insurance Company is barely functional and substantially undercapitalized. Until more private insurance companies establish operations and begin providing homeowners' insurance in Afghanistan, the risk of lending to homeowners may remain too high for banks. Even if the conventional insurance companies are func- tional and become active, a sizable population may still decline to go for such a cover, unless it is Shar'ia compatible. These preferences would require the introduction of general and family takaful (Islamic property and life insurance). Risk Management The mortgage finance market in Afghanistan is exposed to the following sources of risk: · Dysfunctional ownership system: Ownership and the basis of ownership are not well defined in Afghanistan. Records have been destroyed or damaged during long years of war. Multiple allotments are also an issue. This creates a risk factor for the banks on the legality and enforceability of borrower property rights. · Dysfunctional loan security: The poor state of the titles and liens registration sys- tem creates risk in enforceability of loan collateral. · Costly foreclosure: Because of the country's weak legal system, defective titles, and lengthy court proceedings, foreclosures involve delays. When a court verdict is available, its execution entails another lengthy procedure--thus making for an expensive and time-consuming process. · Unreliable property valuations: These valuations are not reliable for the purposes of bank financing. When the mortgage market develops a valuation system, the profession of valuer will become more disciplined. · Pending enactment of essential laws: With the help of international agencies, the enactment of different laws is in progress. When fully enacted, the mortgage mar- ket will get the much-needed confidence of safe and secure lending. · Capacity constraints: The current staff at banks and financial institutions has no training in the field of mortgage lending. The only specialized housing finance institution (which existed in Afghanistan prior to fiscal 2001/02) was not reli- censed to operate. · Maturity and rate mismatch: The nominal outstanding mortgage loans have maturities of two to three years. One of the many reasons cited by the banks for 65 Appendix A A not going aggressively into mortgage lending is the unavailability of long-term funding. Interest rate mismatch is also a bearable risk. However, until such time as long-term, fixed-rate funds are available, lenders will not offer long-term, fixed-rate mortgages. Prudential Regulation and Oversight The legal framework of the financial sector is inadequate in protecting banks and other financial institutions in the event of default. The laws do not support foreclo- sure on borrowers' property, a clear disincentive for banks to take the risk of long- term lending. But pending laws, once enacted, should increase banks' incentives to extend long-term loans. During fiscal 2008/09, some significant regulatory developments took place relat- ing to housing finance. Among them was the enactment of the Law of Mortgage on Immovable Property and of the Secured Transaction of Movable Property Law, both signed by President Hamid Karzai on May 31, 2009. However, it is important to educate and train both the judiciary and banking professionals involved in mortgage finance on the use and implications of these laws. DAB does not have any specific prudential regulations on housing finance, although such regulations are needed, apart from the respective rules for consumer finance. A capacity-building program is also required for financial institutions involved in housing finance to ensure that staff members are well versed in the rules of the business. Credit Information Collection Credit information is not collected or available, beyond the private records of finan- cial institutions. Work is ongoing to develop a credit bureau in Afghanistan. Enabling rules are being framed, and they are likely to be introduced during 2010. Affordability of Housing Loans Indicative estimates on housing loan affordability are provided by the financial sec- tor. Bank-e-Millie estimates that middle-income mortgage clients could pay up to $100 a month, and that it would be feasible for the bank to charge 20 percent in annual interest for a 20-year loan. For middle-income residents, the costs of housing range from $10,000 to $20,000; and for high-income residents, the average cost is estimated at $50,000. Another source of information in this regard is the Shadab Zafar Construction Company, a developer with 40 years' experience in the Afghan market (box A.3). A market investigation undertaken by the Ministry of Economy in 2005 links the cost of housing to buyers' ability to afford it, based on their monthly incomes (table A.4). A house costing $5,000, the typical price for a house owned by low-income 66 Appendix A A Box A.3 Finding Strong Demand for Upper-Middle-Class Housing in Kabul The Shadab Zafar Construction Company targets the high end of the market, building upper- middle-class apartment complexes in Kabul. The demand for such housing is high: the developer quickly presold all its apartments, which cost an average of $41,000 for a 200-square-meter flat. Each buyer must pay a fee of $1,200 to register as a purchaser. When 40 percent of the con- struction is completed, the buyer pays 30 percent of the purchase price. Thereafter, the buyer pays $500 a month until the full price of the flat is paid. No interest is charged; the price is the same whether the buyer pays in cash or in installments. The company claims high demand for its planned new 1,200-unit apartment complex in Kabul. Source: World Bank 2008. Table A.4 Cost and Affordability of Housing in Kabul Market Average Monthly Estimated Average Monthly Disposable Income After Segmenta Household Income Housing Cost Paymentb Housing Cost Low income $30 $5,000 $49 19 Middle income $177 $13,000 $127 50 High income -- $50,000 $483 -- Source: Afghanistan Ministry of Economy data. Note: -- not available. a. Low-income classification is based on a poverty level of less than $1 a day per capita and an estimated six contributors to the household income. b. Payment assumes a 20-year loan at 10 percent interest. residents of Kabul, would require a monthly payment of $49 for a 20-year loan at 10 percent interest. Therefore, even for a 20-year mortgage, the mortgage payments for a low-income house are hardly affordable for a majority of Afghan people. At realistic market interest rates, it is difficult for low-income groups to afford housing, and home improvement loans appear to be a more variable market-based option (as discussed in more detail in the following section). Low-Income and Microfinance Housing Given the low income levels in Afghanistan (42 percent of the population was below the poverty line in 2007), the low-income housing segment effectively represents microfinance housing, mostly for home improvement rather than purchase pur- poses (World Bank 2009). 67 Appendix A A The Microfinance Market Microfinance for housing remains underdeveloped in Afghanistan. But there is a clear rationale for microfinance as a low-cost source of housing finance: homes serve a dual purpose as both a place to live and a place to perform income-generating activities--a good fit with microfinance's aim to reduce poverty by supporting such activities. More- over, microfinance could provide an immediate solution to the need in Kabul for infor- mal sector housing, which is unlikely to attract commercial bank financing because of the lack of clear land titles. Housing microfinance also offers microfinance institutions an opportunity to design new financial products and diversify their portfolios. Microcredit activities in Afghanistan were initiated in 2003 under an apex body, the Microfinance Investment Support Facility for Afghanistan (MISFA). MISFA now has 15 microfinance institutions, operating in 23 provinces and 111 districts. By March 2008, these institutions had 436,000 registered and 373,000 active borrowers. They had disbursed a total of $420 million in loans to low- and moderate-income families, and had current outstanding loans of $108 million. Some borrowers occa- sionally use their microfinance loans for home repair or construction. Many institu- tions have expressed interest in housing microfinance and already have identified housing finance products targeted to different groups of clients. So far, one microfi- nance institution has extended loans for home improvements to clients in Kabul's informal settlements. The loans, at a monthly interest rate of 2 percent, are mostly short-term loans (5­12 months) at a rough average of $600 (ranging in size from $400 to $2,000). They are intended primarily for home repairs. Loans are made on an individual basis and typically guaranteed by a cosigner, which eliminates the need for a mortgage or similar guarantee. Potential Demand for Housing Microfinance Microfinance clients in Afghanistan have demonstrated a large demand for housing loans, often under the guise of business microloans. It is estimated that 20­25 per- cent of business loans outstanding in July 2007 ($18­22 million) may have been used for home construction and improvement. Estimates suggest enormous potential demand for microfinance housing loans among low-income Afghanis. As noted, an estimated 2.44 million people in Kabul live in informal sector housing. Given an average household size of 7.5 people, that sug- gests approximately 325,333 informal sector homes in Kabul. If 50 percent of these homes need improvement each year, at a cost of $750 per home, the potential financ- ing needs for home improvement could be $122 million annually. Similarly, estimates of the construction cost of Kabul's informal sector housing indicate that for 325,333 informal sector homes, about $2.5 billion ($7,684 per household) would be needed (World Bank 2005b). If 20,000 new houses are constructed annually at $7,684 per unit, financing needs for new construction would amount to $154 million a year. These esti- mates suggest a total annual financing need of $276 million for home improvement and new construction in Kabul's informal housing sector. Today, financing needs are met through relatives and friends, labor sharing, and family incomes. 68 Appendix A A Potential Advantages and Challenges in Housing Microfinance Among potential finance providers, microfinance institutions have several compar- ative advantages in serving low-income clients, including experience with and knowledge of their borrowers (table A.5). Even so, several conditions need to be met to ensure effectiveness in extending informal housing loans: detailed knowledge of the housing microfinance market, capacity building in loan appraisal techniques, effective guarantee mechanisms to limit risks from lending for housing, lending methodologies and policies designed for high-risk areas, and cost-effective ways to reach rural clients. In addition, microfinance institutions need to provide loans to women to promote their homeownership. Potential Funding Mechanisms Microfinance institutions have few funding mechanisms at their disposal, including their own funding and funding from MISFA (mobilized from donors). When hous- ing finance activity expands, however, there will be a need for commercial sources of Table A.5 Advantages and Disadvantages of Potential Housing Finance Providers Provider Advantages Disadvantages Microfinance institutions · Experience in lending to · Minimal experience in housing ^ low-income families microfinance · Countrywide presence · Desire to launch housing loan products Commercial banks · Access to capital · Risk aversion · Branch network in urban areas · Unwillingness to lend to low-income households · Lack of technologies for lending to entrepreneurs with irregular income Nongovernmental · Countrywide presence · Lack of experience in lending organizations · Ability to build strategic alliances · Perception that a loan would be a with microfinance institutions to grant and thus not repaid by clients provide technical assistance in construction practices and other nonfinancial services Moneylenders · Countrywide presence · Lack of experience in housing lending, except for short-term (12-month) loans · Costly source of loans Government · Countrywide presence · Public institution requiring resources and time to get started · Risk of politicization and eventual bankruptcy · Potentially negative signals to the microfinance market Source: ShoreBank International, based on review of various studies and discussions with MISFA and microfinance institutions. 69 Appendix A A longer-term (three- to five-year) funding so that microfinance institutions can make the larger loans needed for construction and major renovations, without causing asset-liability mismatches. Market-based funding sources should be encouraged, to promote the commercial sustainability of microfinance institutions. At that stage, there will be a need for second-tier financing, possibly with a role for a stand-alone housing bank.7 Ability to Extend Loans to Women Microfinance institutions may have advantages in extending credit to women entre- preneurs. Women in Afghanistan have limited access to formal finance because of their lack of guarantees and title deeds, as well as cultural factors. Microfinance insti- tutions in Kabul have started planning their expansion into the female client market segment--for example, by considering joint-liability group lending mechanisms. However, much more needs to be done, including providing title rights to women and managing cultural barriers to permit more openness for women. Potential Microfinance Target Groups and Affordability of Microlending About 85­90 percent of economic activities in Afghanistan are carried out in the informal sector (World Bank 2004). In addition, 80 percent of the population is rural, and the overwhelming majority of Afghan households have low incomes. These data suggest the large potential target population for nonmortgage loans for informal and low-cost housing. There are two groups of potential borrowers: the self-employed (entrepreneurs, farmers), found in both urban and rural areas; and wage earners (salaried workers), found mainly in urban areas (table A.6). Table A.6 Potential Target Groups for Housing Microfinance Group Characteristics Self-employed Family-based businesses with fewer than 10 workers, including the owner; most have been established for several years. Urban self-employed Engaged in a wide range of activities, including trade, services, transport, home-based industry, and food production and processing; women work from home, mainly in domestic service situations, beauty salons, and handicrafts production. Rural self-employed Focus on agriculture, livestock, silk, day labor, transport, carpet weaving, food processing, and shopkeeping. Wage earners Employed mainly in the public sector (women made up about 21% of public sector workforce in 2003, earning $60­$800 a month); other wage earners work for international agencies. Source: MISFA. 70 Appendix A A Table A.7 Affordability Analysis for Housing Microfinance Minor Major Indicator Improvement Improvement Amount ($) 500 1,000 Term (months) 12 12 Monthly interest rate (%, 2 2 declining balance) Monthly payment ($) 46 92 Monthly household income ($) 184 368 Market segment Poor Upper poor Guarantees Loan evaluation Loan evaluation, cosigner Requirements Regular income Regular income (business or salary), (business or salary), ownership of home ownership of home Source: ShoreBank International, based on review of various studies and discussions with real estate developers, commercial banks, MISFA, and microfinance institutions. An analysis of affordability--based on clients' monthly payments, household incomes, and the type of financing and guarantees system--indicates that poor bor- rowers could afford a 12-month loan of $500­$1,000 to pay for home improvements (table A.7). For low-income groups, home affordability is a major problem. The ratio of mort- gage payment to income for low-income households in Afghanistan is 163 percent-- well above the maximum 25­30 percent that is required for traditional mortgages internationally. Home affordability in Afghanistan is similar to that in low-income countries like Suriname (table A.8). Table A.8. Home Affordability in Afghanistan and Comparator Countries, 2009 Venezuela, Indicator Afghanistan Bolivia Suriname R.B. de Cost of a 40-m2 home ($) 5,000 6,000 13,000 5,000 Typical amount borrowed ($) 4,000 4,800 10,400 4,000 Annual interest rate (%) 10.0 14.5 39.0 33.0 Average term (years) 20 9 15 10 Monthly payment ($) 49 78 328 111 Typical monthly earnings 30 75 200 225 of low-income households ($) Payment-to-income ratio (%) 163 104 164 49 Source: Data from Da Afghanistan Bank, 2009. 71 Appendix A A Regulatory Framework and Mortgage Market Infrastructure Lack of an appropriate legal and regulatory regime has prevented commercial banks and, to a lesser extent, microfinance institutions from entering the housing finance market in Afghanistan. Inadequate enforcement of laws and structural weaknesses in the housing market have made housing finance an overly risky proposition. Landownership, Registration, and Titling Among the reasons commercial banks cite for staying out of the housing market are the inability to establish clear title to property, the high cost of registering liens, and the lack of secured transaction or foreclosure laws. Right of Landownership. Clear title to land is critical to expanding the housing sec- tor because financial institutions will be unwilling to provide the financing needed by developers and home buyers if clear title is not present. Establishing clear title to land is a major problem in Afghanistan. Land titling has been complicated by the several legal regimes in place: customary law (Rawaj), civil law (Qanoon Madani), religious law (Shar'ia), and statutory or national state law. Shar'ia is largely enshrined in the civil law, but it often differs from customary law. And both are applicable where state law does not apply. In addition, frequent regime change has led to more than 60 different land laws and amendments to the state law (Beall and Esser 2005). As a result, property may be legally held in several different ways, some of which may not involve a formal registration with the municipality. Instead, ownership is legit- imized through a customary document certified by the local shura or jirga.8 The already difficult land titling process has been further complicated during the past quarter-century of conflict, as land was appropriated by several stakeholder groups in turn and frequently legitimized by the relevant institutions at the time of transfer. In addition, squatters moved onto land--either government-owned or private--and built homes while the owners were abroad during a period of conflict. This complex situation makes for a significant number of land disputes that must be mitigated through stronger land tenure systems. Land Registration. Lack of confidence in property titles is a major hurdle in devel- oping housing finance in Afghanistan. There is no coherent land registration law. In addition to several traditional means of property transfer, conflicting registry schemes have been put into place over the decades. And corruption and forgeries often lead to several competing claims to the same piece of property. In Kabul, an estimated 70 percent of residents live on property that has not been registered for- mally. Although these residents may be legitimately entitled to the land, they are at risk of being evicted by others who might acquire a more formal title. In the past, many people avoided formal registration, preferring instead to use traditional means to establish ownership. The titling process took more than 30 steps, each one involv- ing much time and cost and therefore creating opportunities for corruption. 72 Appendix A A Box A.4 Rehabilitating Land Records in Afghan Courts: A Success Story Afghanistan's land administration infrastructure was nearly destroyed during 25 years of conflict. Land records, including property and transfer deeds, were burned, eaten by mice, or destroyed completely. Starting in late-2004, the Land Titling and Economic Restructuring in Afghanistan Project (funded by the U.S. Agency for International Development) began refurbishing more than half the provincial land title offices (makhzans) in the country. Teams have cleaned, restored, and reorganized 5.4 million legal documents, including 500,000 title deeds. Land administration offices now are able to provide proper documents on land transfers and proof of ownership, and they can prevent the registration of fraudulent deeds. Throughout the country, property docu- ments are being photographed and digitized to prevent future manipulation. Source: Interview with Greg Massen, chief of party, Land Titling and Economic Restructuring in Afghanistan Project, March 2009, Kabul. Simple, affordable procedures for property transfer and registration would encour- age people to use formal registration and thus help prevent further fragmentation of the land titling system. Accordingly, reforms have been introduced to reduce the steps involved in the titling process to four or five, limiting the chances of corrup- tion.9 The recent move toward formalization has been successful and should help quickly improve the transparency of the secondary property market. The govern- ment has committed to lowering the cost of land registration, which includes stamp duty and legal and other charges, from 10 percent of the property value to 2 percent; but this currently remains a problem. In addition, an effort has been made to recon- struct and improve land records, which were nearly eliminated during the past 25 years (box A.4). Cadastre. Lack of clear property boundaries has led to land disputes. Few proper- ties have been properly surveyed, and most customary deeds offer only vaguely defined boundaries. Plots often are described as abutting someone else's land, which itself has not been surveyed. Technology that would allow mapping to be done prop- erly is now available in Afghanistan. The Afghan Geodesy and Cartography Head Office has begun digitizing the existing cadastral maps of about 800,000 properties. Producing accurate maps will be key to establishing an accurate titling system in the country. Land Titling. Without a regime ensuring security of ownership, banks will be unwilling to lend in the housing sector; and other private enterprises, such as devel- opers, will be reluctant to do business. Ownership security may be less critical for microfinance institutions, which are less likely to require title for a mortgage guar- antee because they use alternative forms of collateral. But it is a minimum require- ment for commercial banks and other market participants. And it can have effects beyond the housing sector (box A.5). 73 Appendix A A Box A.5 Improving Ownership Security in Informal Settlements in Kabul In 2005, the U.S. Agency for International Development­funded Land Titling and Economic Restructuring in Afghanistan Project launched two pilot projects in districts 7 and 13 of Kabul. The aim was to develop and test ways to improve the tenure security of residents in informal settlements. The project used three guiding principles: First, the communities have the means to solve most land problems themselves. (Community elders are knowledgeable about resi- dents' property and generally are able to back ownership claims, resolve disputes, and weed out fraudulent documents.) Second, solutions must be integrated into the formal legal and land planning systems. And, third, the process must involve a partnership among authorities, com- munities, and donors. Formalizing title and property mapping in the two districts produced impressive results, reflected in large increases in housing construction, home improvement, business start-ups, and land prices. Source: World Bank 2008. Poor Capacity of Registration and Titling Professionals. Few people in Afghanistan have the skills to conduct a proper title search and declare the title free and clear-- and those who do have the skills are unwilling to guarantee the result. The people with adequate skills are based mainly in Kabul, and they are outnumbered by those who are unqualified and possibly engaged in fraudulent activities. Real estate agent licenses can be obtained for a small fee, and there is no mechanism for verifying the credentials of a person acting as an agent. The lack of transparency and enforcement in the title system fosters corruption, creating an environment in which banks are unwilling to operate. Beyond the insecurity of title, weaknesses in property appraisal leave banks unable to rely on the sales values of property. The Ministry of Justice issues real estate licenses for a small fee, but has no mechanism for verifying the credentials of those who apply for such licenses. These real estate agents are the same people who appraise houses for banks. They also verify titles. But because the real estate agents lack training and a code of conduct, banks can have no confidence in either their property appraisals or their title verifications. Moreover, the lack of public informa- tion in the real estate sector makes it difficult to develop a set of comparable prop- erty values against which a bank might lend. Constraints for Women. Women in Afghanistan face particular constraints in gain- ing ownership of or title to property. Women gain ownership of property only through inheritance, but this is difficult. And most of them must rely on a male rel- ative to hold the title to their property. Title to a home would improve access to bank credit for women, and that would be particularly important for entrepreneurs. A recent report showed that access to formal finance for women in Afghanistan is con- strained because they lack guarantees and title deeds (Narain 2006). 74 Appendix A A Foreclosure and Eviction A bank willing to provide a mortgage loan would face discouraging obstacles in cre- ating a lien against the property to secure the loan and, in the event of default, enforcing its security. The introduction of a mortgage law and its effective enforce- ment would largely tackle this issue. Procedure for Creating Liens. The procedure for creating liens is bureaucratic, time consuming, and overly expensive. It is handled through the primary courts and may take up to two months to complete. The costs are 3 percent of the value of the contract (loan amount) to execute the collateral registration and another 3 percent to release the property from registration--a total of 6 percent. But the transfer of property as collat- eral in the name of the bank takes approximately 6 percent of the property value, and the transfer back to the owner takes another 6 percent--a very high cost for the bor- rower. Borrowers are reluctant to go through this costly and onerous process. The gov- ernment recently committed to reducing the cost of lien registration from 6.0 percent to 0.5 percent (0.25 percent at registration and 0.25 percent at release), and that should help encourage financial institutions to formalize collateral registration. The introduc- tion of the Law of Mortgage on Immovable Property and the Secured Transaction of Movable Property Law in 2009 is likely to address these issues. Some banks and microfinance institutions take physical possession of the title when property is used as collateral for other than housing loans, returning it to the borrower when the loan is repaid. And some banks do go through the process of creating liens--but only for very large loans. Liens are not centrally registered, thus creating a risk of multiple liens against the same property if an inadequate title search is conducted. Formal Legal Framework, Quality of the Judiciary, and Enforcement. Even where a valid lien is perfected against the property, the lender has virtually no options avail- able for acquiring and disposing of the collateral in the event of delinquency or default on the loan. As noted, the Law of Mortgage on Immovable Property, the Secured Transaction of Movable Property Law, and laws on negotiable instruments and leasing were ratified by Parliament and enacted by the end of 2009. Regulations on the credit bureau, collateral registry, and appraisal system are being prepared and are likely to be introduced during 2010. Under Shar'ia, which encompasses much of the property law used by Afghan courts, foreclosure or eviction from a property that is the borrower's sole shelter is almost impossible. For this reason, many of the banks that use property as collateral first ensure that the borrower has a second home, and then they file a lien against that home. Afghanistan lacks judges with good training in property law. Property disputes are heard in civil court, where the judges generally are poorly versed in property issues. The Rule of Law Program, a year-long training course for judges funded by the U.S. Agency for International Development (USAID), focused very little on property law. Although there is now a property tribunal set up in Kabul to resolve 75 Appendix A A property disputes throughout Afghanistan, the lack of training and authority has produced poor results. Provincial-level courts have had little or no impact on the way disputes are set- tled. Instead, tribal or informal means of dispute resolution typically are used. This approach works well for members of the dominant ethnic group or tribe, but less so for members of a minority and for women. Fair outcomes have been possible, but in cases where one party has political or other influence, the result is often biased (UNHCR, Pakistan Population Census Organization, and Pakistan Ministry of States and Frontier Regions 2005). The Afghan civil code does recognize mediation and arbitration as legal forms of dispute resolution. When a judge hands down a decision in a property case, there is little or no enforcement of the decision. The lack of follow-through undermines the legitimacy and authority of the court and forces claimants to rely on informal means of enforcement. Responsibility for enforcement lies with the Afghan National Police, but this institution is poorly trained and equipped to enforce judicial decisions. Where there is political will, however, the government is ready to enforce court deci- sions and evict people from their homes. Policy Options for Improving the Framework for Housing Finance The housing sector in Afghanistan operates under few guidelines and with almost no mechanisms for enforcing those guidelines. Remedying this situation will require concrete actions. The Ministry of Urban Development and DAB are playing proac- tive roles in the development of housing and housing finance in Afghanistan. The national government has taken some policy measures to overcome the country's housing shortage. Its primary aims have been to create new serviced and unserviced land, improve land titling, support home construction and improvement, promote the conditions for housing finance, and establish a housing subsidy program for low- and middle-income communities (Afghanistan, Ministry of Urban Development 2007). Policy interventions that involve the private sector in helping alleviate the housing shortage through free land need serious scrutiny. Such policy interventions usually are not targeted to those who most need housing support; and the resources could be better used to improve living standards, such as through the provision of water, sanitation, and basic roads to serve low-income housing. Overall, the govern- ment has to step up its efforts even more in developing a regulatory framework and providing an enabling environment for the housing finance markets. For financial institutions, the situation is still too tenuous to begin large-scale lending in the housing sector. Donors, the central bank, and several ministries are working on issues that will help overcome some of the obstacles, although they are doing so with varying degrees of coordination. The Law of Mortgage on Immovable Property and the Secured Transaction of Movable Property Law 2009 have been enacted. Beyond these legislative initiatives, there also needs to be an overhaul of land policy. City plans developed in the 1970s 76 Appendix A A are outmoded. In Kabul, for example, the master plan developed in 1978 was designed for a maximum of 2 million people. Today's population is 50 percent greater than that number. The municipality's continued use of this plan is the reason that so many settlements are considered illegal. Actions are also needed to establish an enabling environment for housing supply and demand, especially at the lower-income end of the economic spectrum. The mortgage market would be promoted through the introduction of innovative mortgage products better suited to different market segments and risk profiles, as well as through capacity building in the financial sector and expanded access to finance. In the medium term, a liquidity facility or other source of long-term fund- ing is needed to enable mortgage lending. Among the key policy measures for housing are those needed to improve the security of ownership, titling, and regis- tration; and to build the capacity of secondary participants in the housing market. A regulatory framework for property insurance, life insurance, and mortgage guar- antees and mortgage insurance in the medium- to long-term should be developed to encourage housing and housing finance market development. Affordable hous- ing and housing finance are complex challenges, requiring input and effort from both the public and the private sectors. Such input and effort would address the development of serviced land and its provision at affordable prices; affordable mortgages; development of large-scale, low-cost housing; the proactive roles of banks and specialized financial institutions in housing finance; and some level of government subsidies. Key advances will be achieved through the development of microfinance for housing. Finally, the DAB's supervisory and regulatory frame- work needs to be strengthened, including banking sector reforms and implementa- tion of national accounting standards. Some of the requisite policy and regulatory changes to be made on a priority basis are the following: · Strengthened regulatory framework: Developing the housing sector requires a clear legal framework and institutional infrastructure for commercial and financial transactions in the housing market. That means creating a sound legal environ- ment for market participants--financial institutions, housing developers, property buyers and sellers--and enhancing the systems for land titling and for registration and enforcement of liens. The government of Afghanistan already has initiated different actions in policy and legal areas to tackle legal constraints on developing an effective mortgage system in the country. Most of these initiatives have technical and financial support from donors and multilateral agencies. Some of the actions have higher priority than others, and thus are being established in the short or medium term. These programs involve the Law of Mortgage on Immovable Property, the credit information bureau (now being implemented), and the Secured Transaction of Movable Property Law 2009. · Training for judges and financial institutions: Property law has not received enough attention in judiciary training. Judges need the knowledge to review the decisions of shuras and other informal bodies, and to make timely decisions on 77 Appendix A A property cases themselves. Similarly, financial institutions need training in mort- gage and nonmortgage home loans. · Improved enforcement of property law: Enforcement of this law tends to be spo- radic and unrelated to any judicial decision. Landgrabbers forcibly and illegally evict people to take their land. However, where evictions are approved by the court, the police do not carry them out. There still is no formal eviction law in place to protect tenants or banks in cases of foreclosure. For banks and property owners to feel secure, actions to enforce property law must follow formal proce- dures and be backed by a court order, not imposed arbitrarily for political reasons. The police force should have properly trained officers dedicated to enforcing property law in each district. · Efficient system for property titling, registration, transfers, and enforcement of liens: Other steps would involve the enactment of statutes to enhance the land titling and registration system, to review existing title records and establish rights of ownership, and to simplify and reduce the cost of property registration and transfer; the development of integrated formal and informal ownership transfer systems; the computerization of landownership records; the mapping of all property boundaries (land cadastre); the creation of an improved lien registra- tion system (foreclosure); and the training of the judiciary on land titling, regis- tration, and establishment of ownership. The USAID-funded Land Titling and Economic Restructuring in Afghanistan Project is aiming to address some of these issues on a pilot basis. However, many more such projects are needed to address these issues comprehensively and to develop a secure and smooth plat- form for mortgage lending. · Improved title search capacity: A group of people with the training to provide a reliable guarantee of title quality needs to be developed. Some lawyers are knowl- edgeable about title search, but there are far too few of them to meet the demand. Over time, computerization and centralization of title records will make this process more straightforward. · Eased burden of lien registration: A quick, inexpensive, and transparent system for registering liens against property would encourage banks to enter the mortgage market. A long process of lien registration and release exposes the bank, the bor- rower, and the seller of property to risk if the property loses value. A simpler and less-expensive method of lien registration is required. The processing time should be shortened from two months to two weeks and, if the Parliament approves the change, the costs should be reduced. A system of escrow also needs to be estab- lished, allowing banks to deposit funds safely with a disinterested third party who will transfer those funds if a lien is registered by a qualified individual or com- pany (or contingent on any other conditions being met). Such an escrow arrange- ment would lift the burden of risk from all parties to the transaction. · Review and overhaul of land titling: Property deeds take many forms and can confer different types of rights, depending on the historical basis for property 78 Appendix A A ownership and the governing legal framework. To ensure single ownership, all title deeds--from both formal and informal registrars--need to be centralized in the land title office. It is equally important that the registry system be transparent and readily accessible to all participants in the housing finance system. Standard- izing documents, ensuring that each document is filed in only one of a small number of registry offices, and computerizing the records could support efficient searching from remote locations. · Cadastre: The USAID-funded Land Titling and Economic Restructuring in Afghanistan Project aims to address this issue. Primarily restricted to a few main cities, this project can only be termed a test at present. When successfully devel- oped and implemented, however, its scope should be expanded to cover more areas and cities. · Capacity building: Development of a housing finance market requires a profes- sional cadre of participants in all areas of the market. The quality of profession- als in this field is generally low, and there are no formal associations or codes of conduct providing minimum standards. Nor are there penalties for unprofes- sional or unscrupulous conduct. The Afghanistan Institute of Banking and Finance should train banking professionals in mortgage origination and title searches. DAB is considering initiating a capacity-building program to address the new mortgage and secured transactions legislation. And the bank is working on a program to train mortgage professionals in the fields of credit appraisal, title search and assessment, mortgage law, and foreclosure and eviction. How- ever, a comprehensive approach to this training will require some external tech- nical assistance. · Reliable appraisals: To extend mortgage loans, financial institutions must have strong confidence in the accuracy of property appraisals. Currently, appraisals are conducted by property dealers or loan officers on the basis of inadequate market information. One way to improve access to reliable information on comparable values in the market is to make the property registration system simple and user friendly so that people begin to properly register their property with an accurate state of the prices that they paid. Professional appraisals will still be necessary, of course, to mitigate the risk of false valuation and corruption in the housing finance market. Professional standards could be achieved through licensing, reg- ulation, and self-governing associations with strict codes of conduct. · High standards across housing finance­related professions: The emergence of com- panies to provide appraisal, title search, lien and title registration, and possibly escrow services would help pave the way for financial institutions to enter the market. To ensure confidence in these services, the institutions and individuals who provide them must be trained to operate objectively and professionally; must be independent, so that they do not become captive to other market participants (such as lenders) or to other parties to a real estate transaction; and must be subject to licensing and review to ensure that they have the necessary 79 Appendix A A knowledge and skills and adhere to appropriate standards of conduct. Certifi- cation could be carried out by the government or by professional associations in accordance with standards set by the government or subject to regulatory oversight. · Innovative mortgage products better suited to client preference, income, and risk pro- file: To promote and expand the outreach of the housing finance industry to dif- ferent socioeconomic segments, the banks and financial institutions involved will have to innovate segment-specific housing finance products. In addition, a savings- for-housing scheme could be considered, as could international experience with sustainable low-income housing finance in public-private partnerships. · Technical support to microfinance housing lending, mostly for home improvement and incremental housing provision: In view of poor people's very low income and inability to afford costly housing, housing microfinance will remain the main challenge for urban planners and the Ministry of Urban Development on the supply side, and for financial planners including DAB and the financial sector. There are many microfinance institutions and nongovernmental organizations actively involved in the field, and their experiences and practices should be doc- umented. Operational manuals and guidelines should be prepared, possibly blending local circumstances with similar international experience. · Technical support for the credit information bureau: DAB is working on a program to develop a database on consumer finance, which will lead to setting up a con- sumer credit bureau in the country. In the absence of such a database, microfi- nance institutions, nongovernmental organizations, and records kept by utility companies could be good starting points for accumulating credit data. Notes 1. Afghanistan uses a solar calendar in which a year corresponds to the period March 20­March 19. 2. The report is available at http://www.doingbusiness.org/. 3. Information is from an interview with Mohammad Yousef Pashtun, minister of urban develop- ment, June 26, 2009. The housing shortage estimate falls anywhere between 1.0 and 1.5 million units, of which nearly 0.5 million make up the urban housing shortage because at least 50 per- cent of the shortage would be in the low-income category. 4. Interview with Mohammad Yousef Pashtun, minister of urban development, Kabul, June 25, 2009. 5. The Urban Water Supply and Sanitation Project, financed by the Afghanistan Reconstruction Trust Fund, provided $41 million to support the government's short-term goals in the sector. This initiative, which began in September 2006, has supported other water supply rehabilitation projects, which have increased the distribution network by 132 kilometers. 6. Interview with Governor Abdul Qadeer Fitrat, Office of the Governor, Kabul, June 25, 2009. 7. Currently, establishing a stand-alone microfinance bank specializing in housing microfinance is not a feasible option for several reasons. First, building a new institution for low-cost hous- ing finance would take time and capacity, and the costs are likely to outweigh the benefits. Existing microfinance institutions would more quickly and efficiently expand lending activities to the housing market, relying on their countrywide presence and access to existing clients. Sec- ond, a stand-alone housing bank would be viable only in the presence of a developed primary market with commercial volumes of business, a wider geographic coverage, and long-term and 80 fixed-rate products. Appendix A A 8. A shura (the Dari term) or jirga (the Pashtu term) is a committee in a village, community, or township whose members are selected by the residents. The people selected are typically elderly, wealthy, and educated; often, they are landlords. 9. The formal registration process now has the following steps: (1) register with the provincial land title office (makhzan), (2) register with the municipality, (3) pay taxes at the tax office (mustofiat), (4) complete the cadastral process (surveying the boundaries of the property) at the cadastre office, and (5) register the new title in the primary court. References Afghanistan, Ministry of Urban Development. 2007. "Urban Sector Strategy for the Afghanistan National Development Strategy." Kabul. Aga Khan Trust for Culture. 2006. "Bagh-e Babur and Gozar Gah Area Survey Initiative, Preliminary Report." Historic Cities Program, Kabul, Afghanistan. Beall, Jo, and Daniel Esser. 2005. Shaping Urban Futures: Challenges to Governing and Manag- ing Afghan Cities. Issues Paper Series. Kabul, Afghanistan: Afghanistan Research and Evaluation Unit. Bedforth, Garth. 2007. "Housing Finance in Afghanistan: Opportunities and Risks." Housing Finance International 21 (September): 41­44. Narain, Sushma. 2006. "A Gender Strategy for the First Microfinance Bank, Afghanistan." Gender Entrepreneurship Markets Initiative, International Finance Corporation, Washington, DC. Schütte, Stefan. 2006. "Poverty Amid Prosperity: Urban Livelihoods in Herat." Case Study Series, Afghanistan Research and Evaluation Unit, Kabul, Afghanistan. UNHCR (United Nations High Commissioner for Refugees), Pakistan Population Census Organization, and Pakistan Ministry of States and Frontier Regions. 2005. Census of Afghans in Pakistan 2005. Islamabad, Pakistan: UNHCR. http://www.unhcr.org/431c7b1a2.html. World Bank. 2004. "Afghanistan: State Building, Sustaining Growth, and Reducing Poverty. A Country Economic Report." Report 29551-AF, Poverty Reduction and Economic Manage- ment Sector Unit, South Asia Region, Washington, DC. ------. 2005a. "Kabul Urban Land Crisis: A Summary of Issues and Recommendations." Kabul Urban Policy Notes Series, no. 1, South Asia Region, Washington, DC. ------. 2005b. "Why and How Should Kabul Upgrade Its Informal Settlements?" Kabul Urban Policy Notes Series, no. 2, Energy and Infrastructure Unit, South Asia Region, Washington, DC. ------. 2008. "Housing Finance in Afghanistan: Challenges and Opportunities." South Asia Private and Financial Sector, Washington, DC. ------. 2009. World Development Indicators 2009. Washington, DC. 81 Appendix B 1 Bangladesh With 154 million people, Bangladesh is one of the most densely populated countries in the world. Land prices are high; natural disasters are frequent; and sturdy, permanent housing is rare--barely 2 percent in rural areas and 23 percent in urban centers. Esti- mates suggest a shortage of about 5 million houses in Bangladesh, with as many as 500,000 houses added annually in urban areas and 3.5 million added in rural regions. The need for housing upgrades is also high, given the temporary nature of most dwellings and the slum problem in Dhaka City. Housing supply is not forthcoming, however, except for the upper-income groups catered to by private developers. Some nongovernmental organizations (NGOs) and microfinance institutions (MFIs) fund self-construction for lower-income groups--but on a far smaller scale, relative to the existing need. Govern- ment housing provision has not managed to address the issue adequately. Moreover, housing affordability is being eroded by poor land administration policies, which have resulted in very high land prices that make urban housing prohibitive for lower-income groups; and in infrastructure that is inadequate for expansion into periurban and rural areas. There is no active secondary market for real estate, mainly because of the high transfer taxes and an uninterrupted long-term increase in land prices. The housing finance market is characterized by weak competition, poor transparency, underdeveloped structure (such as second-tier lenders), and lack of a level playing field for financial institutions. Bangladesh's ratio of housing finance to GDP is less than 3 per- cent, compared with 50­70 percent in developed countries and 7 percent in India. Mortgage lenders have to rely on short-term funding for financing the loans they make. As a result, formal mortgage finance from private banks and specialized lenders is avail- able only to urban households with high incomes (more than Tk 25,000 a month). Government-subsidized housing finance through the Bangladesh House Building Finance Corporation (BHBFC) does not succeed in targeting lower- and lower-middle- income groups either. That leaves a large lower-income housing market completely 83 Appendix B B unaddressed, despite currently high growth rates of housing finance. Bangladesh best practice MFIs hold some promise for bridging the housing finance gaps; but there are many challenges, including affordability constraints and the need for new products, such as savings-for-housing instruments. The legal, regulatory, and taxation frameworks are not enabling development of pri- mary and secondary housing finance markets. Doing Business 2010 (World Bank 2009a) rates property registration procedures in Bangladesh among the 10 slowest in the world. There is no systematic land survey and cadastre; land administration purview is shared by two poorly coordinated ministries, which increases inefficiencies; and land registers are not computerized, which causes delays and a considerable backlog of land disputes. Transparency and governance are weak. This is aggravated by a framework of very poor court enforcement, also assessed to be among the worst in the world. Titling regulations could be streamlined, especially those regarding new developments. The enforcement of property liens is largely annihilated by the length and cost of cumber- some foreclosure procedures. A public credit registry functions reasonably well in Bangladesh, but an expansion of its coverage is needed. Housing finance reforms will be ineffective unless accompanied by radical improve- ments in regulatory enforcement and property registration. Necessary regulatory reforms include prudential norms customized to housing finance, a more effective collateral and foreclosure regime, improved availability of housing and mortgage infor- mation, and an overhaul of land administration. Pressing policy actions are required on (1) the revamping of the BHBFC, (2) the promotion of long-term financing via a refi- nancing facility and a national savings-for-housing scheme, (3) the encouragement of a secondary housing market via a level playing field for all market participants, and (4) greater affordability of mortgage financing through promotion of microlending for housing and innovative housing finance instruments. Overview Bangladesh averaged a healthy growth rate of 6.0 percent over the last six years. Overall GDP growth in fiscal 2009/10 is projected at approximately 5.5 percent (World Bank 2009b). Minimal exposure to global financial markets has staved off most effects of the global financial crisis; secondary effects can be expected, however, mainly through lower growth in remittances and export demand. Price increases on construction materials in 2008 hampered the housing sector, raising the cost of new housing; at the same time, land prices continued their upward trend. Bangladesh also has made good progress in the banking sector in recent years, as reflected in the growth of credit and deposits. Nonperforming loans in the banking sector have declined from 13.00 percent in 2007 to 10.79 percent in 2008. At the same time, according to Bangladesh Bank, nonperforming loans of state-owned commer- cial banks decreased from 30 percent to 25 percent. The country of 154 million is densely populated with one of the lowest land-person ratios in the world: the density is 1,198 people per square kilometer. Bangladesh is 84 Appendix B B situated in a zone prone to natural disasters (river erosion, flooding, storms, cyclones, and fires) and constrained by the shortage of high land, making for a limited stock of permanent housing. Eighty percent of the country's population is rural; and 50 per- cent of its GDP comes from this sector, setting development (including on housing) in rural areas as a priority. The annual growth rate of the population is 1.75 percent, and there are about 20 million households. Some 44.3 percent of the people are literate, and the primary school enrollment rate is 86 percent. The youth in Bangladesh constitute 36 percent of the total civilian labor force (World Bank 2009b). Dhaka, the biggest city (population 8.5 million) is the capital of the country and the main economic hub. The urbanization of Bangladesh stands at about 20 percent; and the growth rate of the urban population, spurred by both population growth and urban migration, is high, at about 4.5 percent. This growth is creating pressure on the supply of housing stocks in urban areas (World Bank 2009b). Housing Demand Bangladesh's housing market is characterized by a surplus of upper-echelon housing stock and an acute shortage of affordable housing for the great majority of middle- and lower-income population groups. Estimates suggest a shortage of about 5 mil- lion houses in Bangladesh in 2009. In urban areas, the annual estimated demand amounts to 300,000­500,000 houses. In rural areas, with an assumed 2 percent new household formation annually, the new demand could be as much as 3.5 million a year. Of the larger cities and towns in Bangladesh, Dhaka is the hardest pressed in terms of unsatisfied housing demand (according to data from the BHBFC). The housing needs of lower- and lower-middle-income groups are considerable and remain largely unfulfilled. The BHBFC reports that there is great demand for houses priced moderately between Tk 600,000 and Tk 1,000,000; but the supply of such dwellings is almost nonexistent because of high land prices in downtown areas of metropolitan cities, particularly in Dhaka City. In areas on the city periphery, land prices are lower; however, there is no practical transportation method for dwellers in farther outskirts to get to the city and to their workplaces on a regular basis. Careful market-based demand studies are not available, and there is a need for further research on the housing demand of lower- and medium-income groups.1 Paradoxically, housing supply caters to upper-income groups, and the market is glutted with new luxury housing units. There are around 800 developers in Bangladesh, and the market is active and relatively competitive. According to the Real Estate and Housing Association of Bangladesh, private developers (focused mostly on Dhaka City) typically engage in the building of dwelling units covering 1,000­1,500 square feet. New formal construction accounts for about 3 percent of all housing (permanent and temporary) being built formally or informally in the coun- try. Such development is exclusively limited to the upper-income urban groups, and it is undertaken mostly for investment and rental purposes, not for use as primary residences. 85 Appendix B B Approximately one third of all houses in urban areas were constructed outside of the formal regulatory system, mostly on land to which the homeowner does not have a formal title. A 1991 Bangladesh national census counted 2,100 slums in Greater Dhaka. Currently, more than 2 million people in the capital city live either in slums or are without any proper shelter. Housing Supply Current State of Housing Bangladesh has a high population density--about 1,200 people per square kilometer-- and a limited area of 134,000 square kilometers, which is frequently flooded during the summer months. In such difficult conditions, only 23 percent of housing in urban areas and merely 2 percent in rural areas is permanent housing. Close to half of all housing units in the country (approximately 3.3 million) are made of temporary materials. This nondurable housing requires replacement within one to five years after construction. Even among the housing defined as permanent, there are many types of dwellings with only some degree of permanency, ranging from construction with brick masonry and reinforced concrete construction pillars to tin-roofed and tin-walled houses. The heterogeneity of housing makes it difficult to estimate con- struction and housing materials costs, and complicates the measurement of real estate price indexes. Table B.1 shows some characteristics of the housing sector. The higher- and middle-income groups are housed in either low-rise, single-family houses or, increasingly, in multifamily apartment buildings (table B.2). The lower- income households (approximately 70 percent of the urban households) are housed in a variety of house types. Approximately half of the low-income housing units are in bustees (slums), informal settlement areas that include both private rental and pri- vate ownership housing built either on privately owned land or on illegally occupied public land. Conventional tenement slums (rental and owner occupied) take up another quarter of the low-income sector. These multiunit buildings originally were Table B.1 Housing Sector Characteristics, 1998 Characteristic Bangladesh Overall Rural Areas Urban Areas Total number of dwelling units 19,020,489 15,474,566 3,545,923 Per capita floor space (sq. ft.) 54.9 53.5 62.3 Occupancy level (people per dwelling unit) 5.48 -- -- Proportion of permanent structures (%) -- 21 46 Proportion of rental units (%) -- (Dhaka 65) 5 40 Access to clean water (% of population) -- 78 42 Source: Center for Urban Studies, Dhaka, 1998. Note: -- not available. 86 Appendix B B Table B.2 Dhaka Urban Area Housing Subsystems, by Income Group and Land Coverage Approximate Approximate Income Group and Housing Proportion of City Coverage of City Subsystem Population (%) Residential Land (%) Upper-income group 2 15 Middle-income group 28 65 Source: Marja C. Hoek-Smit, United Nations Development Programme/United Nations Centre for Human Settlements, 1998. built in compliance with the building code, but now are seriously overcrowded and poorly maintained. Overcrowding in these buildings has increased during the last several years because of an influx of rural migrants coming to work in the expanding garment industry. Other categories of low-income housing include government- provided squatter resettlement camps, plots of land with basic services that are given on a leasehold basis; employee housing that consists mostly of small apartments in high-rise complexes provided by the government; makeshift houses built by squatters on illegally occupied public or private land; and pavement dwellings. Several organizations and rural innovators in Bangladesh developed and modi- fied various designs of rural low-cost housing after the devastating flood of 1987, and are doing the same for the people affected by cyclone SIDR in 2007. The houses vary in appearance throughout the country, but have similar basic structural components.2 Development of the Housing Stock The supply of housing is greatly affected by the land development process in Bangladesh. The process for the entire country is centrally controlled from Dhaka, with little autonomy at the local level. The housing development process is slow and costly because of the poor preparation of master plans; the dearth of planning pro- fessionals in the public sector; and inadequate infrastructure provision, land acqui- sition, development and construction financing, and mortgage financing. All of these challenges raise development costs and make affordability increasingly elusive for a large portion of the population. The problem is most acute in Dhaka, where the city development authority, Rajdhani Unnayan Kartripakkha (RAJUK), suffers from administrative inadequacies, impeding more than enabling the city planning process. But other regional planning authorities (such as those for Chittagong, Khulna, and Rajshahi) have even fewer resources and less power to implement plan- ning and development, and they are barely functional. The matter is further aggra- vated by the scarcity of information on housing supply and shortages for urban areas other than Dhaka, thus preventing adequate policy-making attention to these areas. A second but no less pressing issue is the high level of land prices for residential construction. Land prices have been particularly high in Dhaka, where the need for middle- and lower-income housing is the most severe. Town planning is difficult, given dense population, saturation of constructed areas, scarce city land, and a poor 87 Appendix B B transportation network to the farther outskirts (where housing infrastructure is also poor). According to the Bangladesh Institute of Planners in 2003, residential land in Dhanmondi--the most expensive area in Dhaka--costs Tk 3,500 per square foot (roughly $60). Prices have grown 15­16 percent in upscale areas such as Dhanmondi and Motijheel, much above the general cost of living. Prices are much more afford- able on the periphery of the city--Pallabi, Shyamali, and Uttara--but those areas involve a long and expensive commute to employment centers. The provision of planned, efficient, and inexpensive transportation facilities from Dhaka's outskirts to the heart of the city could shift the burden of housing needs from the center to the more affordable periphery. The Bangladesh University of Engineering and Technol- ogy planning faculty estimates that land costs in Khulna are similar to those in Dhaka; in Chittagong, they are about 15 percent lower (although the most expensive areas there match Dhaka), and are 30­40 percent lower in other cities. Even in the lowest-cost areas, however, prices are inordinately high, relative to GDP. For exam- ple, again according to the Bangladesh Institute of Planners, the least-expensive res- idential land in Dhaka costs approximately Tk 1,600 per square foot (about $27) (table B.3). If the density/price relationship were linear, this would mean that even with a floor-area ratio of 10 (which is extraordinarily high), the land cost of housing development in Dhaka would be at least $2.70 per square foot. The process of residential infrastructure development in Dhaka is fraught with market failures. RAJUK regulates city development and plays a planning role, charges itself with providing residential infrastructure, provides urban land for development, and carries out residential construction in its own right. Each of these Table B.3 Residential Land Prices in Dhaka, 2004 Location Land Value (Tk/sq. ft.) Dhanmondi 3,472 Gulshan 3,194 Baridhara 3,056 Banani 3,056 Elephant Road 2,778 Lalmatia 2,500 Santinagar 2,500 Siddeswari 2,500 Green Road 2,222 Segunbagicha 2,222 Mohammadpur 2,083 Niketon 1,944 Mirpur 1,667 Shyamali 1,667 Uttara 1,667 Source: World Bank 2004. 88 Appendix B B functions could benefit from an overhaul and increased effectiveness. RAJUK plays an important role in providing serviced land, but in a way that constrains rather than fosters the contribution of private developers and the widening of housing supply. RAJUK's role in development land allocation is distortive. The agency has obtained a vast share of Dhaka urban land at artificially low prices (through its ability to impose a sale on the land's owners--local government entities). Having this land gives RAJUK an unfair advantage against private developers, compounded by the fact that RAJUK developments benefit from municipal financing of residential service infrastructure, a cost that private developers bear themselves. Furthermore, RAJUK requires end clients to pay early installments during construction, thus restricting its subsidized new apart- ments to well-off customers. RAJUK's construction activities also are inefficient and vastly inadequate, given housing demand realities. Its inadequate advances in provid- ing residential serviced land globally restrict land supply. Since its formation in 1959, the agency has produced serviced parcels sufficient for 16,000 housing units, fewer than 400 units per year. Over that time, the area within RAJUK's jurisdiction has grown by at least 6 million people (more than a million households). Against this, RAJUK has been able to provide serviced land to meet between 1 and 2 percent of the demand for such parcels. Transparent land allocation at true market prices and a level playing field among developers are crucial for increasing the availability of urban land and meeting the growing housing demand of the middle-range income group. Where land to be developed is owned by a private party, the development process is inefficient as well. The developer does not acquire the property; rather, he or she receives from the owner a power of attorney that gives the developer permission to build on the property. Upon completion of construction, the land is transferred to the developer, in exchange for the ownership of a certain share of apartment units in the constructed property. Permits are required from at least eight different agencies, and acquiring each one involves delays, lack of transparency, and governance prob- lems. Additional delays are caused by an inadequate legal and financing framework. Developer financing is limited because of the underlying risks involved. As a result, land acquisition for housing development is funded mostly by equity from the developers themselves. Such capital constraints, in turn, limit development projects to an inefficiently small scale. Although construction loans are available, they have low loan-to-value ratios (typically 50 percent or less) and carry such high rates of interest (often more than 16 percent) that developers see little value in them. This sit- uation is in contrast to conditions in other countries, where construction costs very often will be financed at 100 percent, with an interest rate that is some margin above the London interbank offered rate. The reluctance of banks to fund developers also is rooted in strong preferences by developers to hoard land and time the market as land prices increase. Given the shortage of land for middle- and lower-income hous- ing, these practices further aggravate the supply problem. Owing to the lack of robust construction financing in Bangladesh, the ultimate buyer of the property usually finances the construction costs through installment payments. This system slows down construction as developers wait for installments before they can top off projects. In a typical case, although the actual construction 89 Appendix B B might take 14 months, delays arising from financing and other issues often add another 6­12 months to the process. These frictions are estimated to add 10­20 per- cent to the cost of construction. Even if used, construction loans run into an asset security problem. Specifically, until a developer fully repays a construction loan, end buyers of the apartment units cannot obtain financing secured by the same lien. There is no "release" structure in construc- tion financing in the case of leasehold land: title is not released to the home buyer until construction of the entire project is complete. Currently, this risk is mitigated by lenders working with reputable developers only, and by requiring third-party guarantees. A simplification of property transfer approval procedures could permit transfer during the construction. (Currently, the authorization of the public title owner is required.) These loans also transfer risk from the developer to home buyers because the buyers pay significant advances to developers prior to completion of the construc- tion, but are not protected against delays, project failure, developer bankruptcy, or failure to meet contractual specifications for price or quality. Given these risks, finan- cial institutions are reluctant to lend to home buyers for new developments. That reluctance, in turn, limits the market for new construction to clients who can pay fully up front--the upper-income groups. The builders/developers association--Real Estate and Housing Association of Bangladesh--was formed in 1991 to strengthen the private sector's role in housing development. It has a pivotal role in handling the issues of housing in an organized and modern manner through its 540 members. It is also playing an active role in framing most of the government's policies and laws relating to this sector, operating as a well-organized lobby defending developer interests. Rise in Cost of Construction Materials There has been a gradual rise in the cost of construction materials in Bangladesh (table B.4), with prices doubling over the past 10 years and quadrupling over the past 20. Construction material price increases still have lagged behind the phenomenal Table B.4 Rise in the Prices of Construction Materials, Selected Years, 1988­2007 Price of Sand, Price of Brick, per sq. ft. (Tk) Price of Cement, Price of Iron Year per piece (Tk) Local Sylhet per bag (Tk) Rod, per ton (Tk) 1988 1.00 3.60 5.25 105 11,000 1998 2.00 7.10 9.00 182 17,500 2000 2.50 7.80 10.50 193 21,300 2004 3.25 9.00 15.00 240 37,000 2005 3.15 8.00 22.00 283 39,700 2006 3.90 9.00 20.00 313 42,700 2007 4.20 9.50 21.50 337 49,200 Source: Authors' compilation. 90 Appendix B B property price escalation in the country, however. The rise in the cost of construc- tion materials has roughly followed the inflationary trend, and thus the rise in income levels. This fact indicates that it is the rise in land prices that really has pushed the property prices to unaffordable levels. The Secondary Property Market The flawed land development process has resulted in a very limited secondary market for property in Bangladesh. This is compounded by a speculative attitude toward property prices, built on the perception of unlimited future land price increases fueled by population and economic growth. Consequently, real estate is not sold, even if the cash flow from rents is not sufficient to cover the mortgage service. The weak market signals in the mortgage market, and lack of foreclosure enforcement, do not help. Institutional mechanisms and the legal framework are not in place to stimulate a smoothly functioning real estate transactions market. Specifically, bottlenecks are cre- ated by the lack of secondary housing market infrastructure and a high transfer tax. Bangladesh has a dearth of real estate brokers and appraisers, and poor informa- tion on market prices and valuation. The absence of a formal mechanism through which buyers and sellers can exchange information about the real estate market pre- vents the market from operating properly. The transfer tax rate is extremely high; at 12.5 percent of gross price, it strongly discourages the transfer of assets (or at least the "official" transfers of assets). Reduc- ing this rate would likely increase revenue from the tax and, even more important, stimulate sufficient sales to start making accurate determinations of market value-- determinations that are crucial to the development of a market for mortgages. Housing Finance The combination of high land prices and high interest rates makes housing unaf- fordable to middle- and lower-income groups, and simultaneously causes over- supply of housing for the very top income groups (box B.1). A large part of these inflated prices is the result of institutional factors that can be addressed by policy measures. To a large extent, the high land prices result from a faulty land develop- ment process, as noted above. As will be argued in this section, the high interest rates are a result of weak competition, poor transparency, underdeveloped market structural features (such as second-tier lenders), and the lack of a level playing field for financial institutions. But housing finance is growing at very high rates, spurred by high excess demand and market innovations to increase supply. During 2008, Delta Brac Housing Finance Corporation Limited's housing portfolio increased by 34 percent, while those of commercial banks increased by 16 percent. The sector as a whole grew by 14 percent in 2008, and stood at Tk 142.5 billion, or a little more than 6 percent of total credit to the private sector. The present housing finance system in Bangladesh is extremely small and highly segmented. Formal mortgage finance is available only to households with monthly 91 Appendix B B Box B.1 Market Failures in the Housing and Housing Finance Markets in Bangladesh Consider a typical Tk 2 million unit (see figure BB.1). Improving the loan conditions from maturity of 10­15 years and interest rates of 14­15 percent to terms of 20-year maturity and 8 percent interest would double the population who can afford mortgage finance, reaching the full top decile of the urban income distribution. Lowering the average property price from Tk 2.0 million to Tk 0.7 million expands the market to 40 percent of the urban population. Developers have started recognizing the potential depth of the market for moderately priced housing. But in the absence of fundamental improvements in infrastructure, land, and housing development, this nascent trend will not go far. But how to lower land prices and interest rates? In the case of Bangladesh, the argument is that neither amount reflects market conditions. For interest rates, competition, transparency, developed market structural features (such as second-tier lenders), and equal footing for financial institutions would go a long way toward lowering the cost of mortgage financing. Similarly, real estate prices reflect not only a physical scarcity of usable land in a high- density country; to a much larger extent, they also reflect institutional and regulatory inade- quacies that could be addressed by appropriate policy measures. India provides an eloquent example of how improving land use policy considerably can lower prices. Land prices started falling in the mid-1990s after several states repealed the Urban Land Ceiling Act, which was severely restricting land development. Prices fell by half from levels similar to those in Dhaka. Figure BB.1 Urban Households' Estimated Incomes and Mortgage Affordability, 2003 100 90 80 70 60 percent 50 40 30 20 10 0 0 3 4 5 6 8 9 12 15 25 54 lower-income threshold cumulative income distribution improved mortgage: 8%, 20 years present affordability range: Tk price = Tk 700,000, improved mortgage 2 million, 14.5%, 12 years Note: 10 percent equals 560,000 households. Lower-income threshold equals Tk 1,000 per month. Source: World Bank 2004. 92 Appendix B B incomes above Tk 25,000 (well above the 10th percentile of the urban income dis- tribution), and is restricted to selected housing submarkets in Dhaka. Within this section of the market, government-subsidized housing finance through the BHBFC is most prevalent; at the same time, the nationalized commercial banks (NCBs) are decreasing their housing loan portfolios. Expansion of the sector has to come through private sector mortgage lenders who recently have started to operate in this market. With existing capital constraints in the financial sector, the competition by government savings and lending institutions, and the weak debt market, new hous- ing finance institutions face a challenge in mobilizing funds; and it is unlikely that the cost of longer-term funds will come down in the short term. In addition, the housing finance sector has to convince these new housing finance institutions that it can be a profitable investment sector with long-term potential. Noncollateralized credit for house construction by MFIs is available to only a small proportion of poor rural households that have participated in income generation credit programs. The Housing Finance Market The financial sector in Bangladesh is fairly well-developed, though it is dominated by the banking sector. It comprises 4 NCBs, 5 government-owned specialized banks, 30 domestic and 9 foreign private commercial banks, and 29 private nonbank financial institutions (NBFIs), in addition to cooperative banks and insurance companies. Over- all banking sector assets amount to almost 60 percent of the GDP. The financial system has seen a gradual growth of Islamic banking in Bangladesh since 1983, with the estab- lishment of the first Islamic bank, the Islami Bank Bangladesh Limited (IBBL). In the housing finance market, an additional player is the BHBFC, a government entity that pioneered house building finance in the country and, until recently, was the principal source of housing finance there. In rural areas, NGOs make consider- able contributions to financing housing construction and improvements. The gov- ernment also funds lower-income housing via a special Tk 500 million fund called "Grihayan Tahabil." That fund was created in fiscal 1998/99, and monies are dis- bursed through NGOs. The multiplicity and diversity of players has not translated into a vibrant and competitive housing finance market because the primary mortgage market is char- acterized by (1) funding distortions that handicap the most dynamic and efficient actors, (2) the legacy of a high level of nonperforming loans among NCBs, and (3) an uneven playing field between the state-owned BHBFC and other lenders. All of those factors result in allocating the cheapest financial resources to the least efficient players and pushing up intermediation costs--in turn, driving interest rates beyond the reach of most Bangladeshis. Therefore, interest rates continue to show the frag- mentation of the market between the BHBFC (which lends at as little as 9 percent) and the other institutions (which lend at rates between 13 and 16 percent), without taking into account various add-ons.3 As of June 30, 2008, the total mortgage debt outstanding stood at Tk 142.5 bil- lion, a 14 percent increase over that of end-June 2007 (see table B.5). 93 Appendix B B Table B.5 Outstanding Housing Loans and Overview of the Housing Industry, 2006­08 a. Outstanding Housing Loans Amount Outstanding June 2006 June 2007 June 2008 Tk % of Tk % of Tk % of Lenders billions total billions total billions total SCBs 25.8 25.57 28.9 23.19 33.6 23.58 PCBs 35.1 34.79 43.7 35.07 50.8 35.65 Other banks 3.0 2.97 13.5 10.83 16.6 11.65 Subtotal 63.9 -- 86.1 -- 101.0 -- HBHFC 26.2 25.97 25.2 20.22 24.4 17.12 DBHFC 5.5 5.45 7.1 5.70 9.5 6.67 NHFIL 1.9 1.88 1.9 1.52 1.9 1.33 Other financial institutions 2.9 2.87 4.1 3.29 5.5 3.86 Subtotal 36.5 -- 38.3 -- 41.3 -- MCLs and GB 0.4 0.40 0.2 0.16 0.2 0.14 Total 100.9 100.00 124.6 100.00 142.5 100.00 Source: Bangladesh Bank Annual Report, July 2007­June 2008. Note: -- not available; DBHFC Delta Brac Housing Finance Corporation Limited; GB Grameen Bank; MCLs microcredit lenders; NHFIL National Housing Finance and Investments Limited; PCBs private commercial banks; SCBs = state-owned commercial banks. b. Housing Industry Overview Number Amount Average Loan Size Factor of Loans (Tk millions) (Tk millions) Housing companies 2,534 15,358 6.00 Urban housing 58,256 47,174 0.80 Rural housing 28,218 1,871 0.06 Renovation loans 2,611 3,150 1.21 BHBFC 41,573 14,336 0.34 Source: Data from Bangladesh Bank. Housing loans in Bangladesh have a maturity of 10­15 years. Lending criteria reflect a great cautiousness, particularly through restrictive loan-to-cost maximums (generally capped at 70 percent, but averaging 50 percent in practice) and relatively limited maximum taka amounts (Tk 2.5­3.0 million for many lenders). Interest rates are fixed, but loan agreements typically stipulate that the lenders can change the rates at their discretion during the life of the loan. BHBFC The BHBFC is a traditional specialized player, still dominant despite its modest size. According to data from the BHBFC Annual Report of 2006/07, the bank disbursed 94 Appendix B B about 1,100 loans that year (roughly Tk 1.2 billion), with 45,000 loans in stock amounting to Tk 33,558 billion outstanding; and it had 550 employees. The institu- tion is beset with a poor governance structure, misaligned operational incentives, high nonperforming loans, and ill-targeted government subsidies. In spite of these short- comings, the BHBFC is the only institution serving the low- and middle-income housing segment and a wider geographic area beyond Dhaka and a few larger cities. The BHBFC suffers from governance shortcomings that plague many institutions of its kind around the globe. It is closely controlled and wholly owned by the Min- istry of Finance, and is run in a manner that is more administrative than financial. These disadvantages stall flexibility and quick decision making, and they make it harder for the BHBFC to compete in an increasingly competitive housing finance market. Furthermore, from a financial sector stability point of view, the central bank's prudential regulation would be much more prudently imposed on the BHBFC than are the more lax Ministry of Finance rules. This situation is reflected in Bangladesh Bank's loan conditions, which are set by the board of directors with apparently no direct connection to the market. (As evidence, the conditions have not changed for several years.) Also, the institution is corseted into rigid business limitations--on the assets as well as on the liabilities side--that prevent it from adjusting to a changing environment or diversifying its risks. The BHBFC has become known for inefficient and lengthy loan processing: loan approvals have taken as long as a year in the not-too-distant past. Recent efficiency improvements at the BHBFC have reduced the loan processing time to three to six months. According to a recent office circular currently under implementation, the narrowly defined loan processing steps are restricted to a total processing time of 15­20 working days for loan approval, with respective deadlines for each process step. However, unless the workflow is redesigned and regrouped and certain independent functions are carried out simultaneously, it does not seem pos- sible to contain the processing time within 15­20 working days in practice. Specifi- cally, loan processing time is a function of three main procedures. Other aspects of the BHBFC's operational efficiency also are plainly apparent and calling for an upgrade. Administrative cost per loan is prohibitive, compared with private mort- gage providers. Other nonmarket operational practices exist in the BHBFC's functioning, such as sanctioning loans on the basis of property valuation instead of giving due weight to borrower repayment capacity. The BHBFC loan portfolio covers 18 cities; it shares 71 percent of the mortgage market in terms of number of loans, and 30 percent in terms of outstanding portfolio value (table B.6). The bulk of the BHBFC's portfolio is in Dhaka (69 percent), 9 per- cent is in Chittagong, and the remaining 22 percent is scattered elsewhere in the coun- try. The BHBFC has a limited product range that caters poorly to the variety of clients. It issues six types of loans: general, group, apartment, adjusted, semi-pucca house (combination of soil and concrete), and middle-/lower-income group. Despite a lend- ing limit of Tk 4 million, the corporation tends to lend to the higher-income house buyer: 76 percent of the borrowers are landowners who build multifamily houses for rental purposes. The BHBFC's lending rates are lower than the market (in June 2008, 95 Appendix B B Table B.6 BHBFC Loan Portfolio, 2008 Category Amount (Tk millions) Total outstanding portfolio 25,163 Principal 14,336 Interest 10,827a Source: Data from the BHBFC. a. Of this amount, Tk 931 million is suspended, Tk 2,436 million is deferred, and Tk 7,460 million is due or due but not received. they were 12 percent for Dhaka and Chittagong, and 10 percent elsewhere [effective rate 9.6 percent)]), which averages 14.5 percent. The maturity of BHBFC loans is 15­20 years, compared with approximately 10 years for market-offered mortgages. The BHBFC suffers from misaligned operational incentives. In view of its lower- income developmental focus, the institution has enjoyed various privileges regarding taxation, prudential obligations, and capital raising--chief among them, government support for its funding. Traditionally, the main source of funding has been subsidized debentures with government guarantees and statutory liquidity ratio eligibility, although the government has been reluctant to allow BHBFC recent issues (since 1998). Outstanding debentures amount to Tk 12.2 billion, at an average cost of capi- tal of approximately 6 percent. They are 75 percent subsidized by Bangladesh Bank and 25 percent subsidized by financial institutions. There is no active secondary mar- ket for these debentures. The BHBFC would rather benefit from aligning its funding and performance with the market; and receiving instead an explicit, transparent, and well-targeted subsidy for each low-income mortgage financed. The BHBFC traditionally has shown a poor ability to maintain payment discipline among its borrowers. Its poor performance in this field, added to the neg- ative legacy of past government programs in rural areas, has deeply affected the quality of its portfolio. Nonperforming loans amount to 18 percent of its gross loan portfolio--all the more worrisome in view of the Ministry of Finance's more lenient classification of loan arrears, compared with the standards used by Bangladesh Bank for commercial banks.4 The current recovery rate remains below 50 percent. What is worth mentioning, however, is the emphasis that management now puts on collection efforts. The BHBFC does enforce security rights: it grants only registered mortgages. Part of the low recovery rate reflects the length of court procedures; and the nonperforming loan rate has been decreasing in the past five years. The corporation is raising loan-loss provisions to the Bangladesh Bank's reg- ulatory level. But the provisioning capacity can be seen in a globally negative way, as a diversion of subsidies that enables the institution to afford the consequences of its former permissive policy at the expense of the borrowers to whom the subsidy should be passed. Because of the restrictions imposed on its funding by the gov- ernment, the BHBFC (which is not allowed to take deposits) has been forced to limit its new lending--essentially based on recoveries. New loan sanctions have been small: Tk 839 million in fiscal 2007/08 and Tk 2,366 million in fiscal 2008/09. 96 Appendix B B Instead, the BHBFC is the channel for a nontransparent and ill-targeted government subsidization mechanism. The rates charged on loans are significantly below those of all other lenders: 10 percent, 13 percent, or 15 percent nominally. The effective rates are even lower because of interest computation methods, rescheduling, and rewards to regular payers. The combination of these administrative rates with below-market refinancing makes measuring the precise subsidy volume almost impossible. According to its charter, the BHBFC should focus on lower-income groups; however, despite a lending limit of Tk 250,000, it tends to lend to higher-income house buyers. In spite of its shortcomings, the BHBFC has many positive aspects to harness. Tra- ditionally the oldest mortgage finance provider, it has accumulated institutional knowledge and experience; and it is well versed in housing, mortgage lending, titling, and documenting issues. As housing needs spiral upward, the shortfall in mass, bulk housing funding can be better addressed by this organization. It is, by far, the largest specialized financial institution, having widespread geographic presence and a very large borrower base. Because of its geographic spread, it is well placed to further any government programs aimed at providing housing solutions for lower-income groups. Traditional Mortgage Lending With the entry of commercial banks and financial institutions, the mortgage finance market in Bangladesh is rapidly growing in size and getting much more competitive in nature. The commercial banks, in both the public and private sectors, either have established or are in the process of setting up in-house mortgage finance depart- ments. There are a few NBFIs already very active in the housing finance market (described in the "Specialized Mortgage Lending" section below). Islamic banking, including mortgage lending, also is growing quickly in Bangladesh (see the "Islamic Mortgage Finance" section below). Although there are three or four Islamic banks now operating in Bangladesh, the IBBL is the leading Islamic commercial bank. The banking and financial sector regulator is Bangladesh Bank, the central bank, which is entrusted with supervision and regulatory oversight of the sector (discussed below in the "Prudential Framework for Housing Finance, and Recent Policies" section). NCBs are the second important source of housing finance in Bangladesh (after the BHBFC), reluctantly following government guidelines to provide housing finance. The NCBs have a low cost of funds because they receive the bulk of their deposits from government entities, which are statutorily required to deposit at least 75 percent of their funds with NCBs. Nonetheless, the NCBs suffer from a term- mismatch problem because they are constrained by the maturities of their deposits. Those maturities may extend to a maximum of three years in principle, and to less than a year on average. The NCBs thus give priority to corporate short-term lending, and they set a low ceiling on the amount of mortgage loans they extend (3­6 percent of their outstanding portfolios). Housing finance is only part of their mortgage portfolios, a significant portion of which is allocated to commercial real estate. The NCBs follow a policy of wide intermediation spreads for housing--about 7 percent--particularly as they seek to offset the poor profitability of their other 97 Appendix B B assets and make up for inadequate provisioning. Despite low loan-to-value ratios, the rate of delinquency is high. Partly as a result of insufficient capital bases, the NCBs tend to restrict their lending activity. Their growing liquidity means that large amounts of cash lie idle, while their widespread branch networks (a potential asset for large-scale retail lending) are underused. Private and foreign commercial banks are growing as providers of housing finance, with housing loans making up 5­10 percent of their loan portfolios. They use the same deposit-based funding as do the NCBs, but they are less handicapped by delinquencies: the ratio of classified real estate loans seems to be in the range of 5 percent. These banks do not enjoy the privileges of the NCBs; but they are more efficient and sport a better service quality, better product options, and expanded branch networks. Foreign banks are recent newcomers to this market; the credit performances of their mortgages have been good, but their impact on the market is likely to remain marginal because they target only upper-class borrowers. The weighted average rate on private commercial bank deposits in June 2008 was 6.85 percent, compared with a weighted average lend- ing rate of 12.09 percent. Foreign banks are in an even better position in this regard, with a weighted average deposit rate of 3.75 percent, against a weighted average lend- ing rate of 11.29 percent. Specialized Mortgage Lending Private institutions specializing in housing finance have made remarkable strides in a short time. The management teams at Delta Brac Housing Finance Corporation Limited, National Housing Finance and Investments Limited (NHFIL), and IDLC Finance Limited--aspiring to the same professional standards as their counterparts in other areas of the world--have made substantial progress in underwriting, loan servicing, streamlining local decisions, eliminating delays, and increasing trans- parency. Particularly impressive is Delta Brac's work in putting together credit histories of potential borrowers, and in beginning to correlate credit history with loan performance. However, these institutions operate at the level of upper- and upper-middle-income clientele, as evident from their average loan size per client (nearly three times the already high average loan size of the BHBFC). Specialized financial institutions also have made strides in loan servicing. Because they keep careful track of delinquencies and communicate with borrowers at the first sign of trouble, they have been able to reduce the number of serious nonperforming loans to less than 5 percent of their business. One thing that keeps nonperforming loans to a minimum is that lenders require owners to put up substantial equity (at least 25 percent and usually more) before getting a loan. Lenders also insist on obtaining and keeping title until the loan is paid off. Last, and perhaps most impor- tant, is lenders' ability to persuade borrowers that they are better-off with a property whose ownership interest is not disputed. The very presence of a foreclosure proceeding creates just such a dispute. Given the equity stake that owners have and the attitudes toward property in Bangladesh, the threat of uncertainty about property rights--even without actual foreclosure--remains a strong deterrent to loan 98 Appendix B B delinquency and default. These private institutions receive no indirect subsidies, but their share of the housing market is exploding at the expense of other institutions. Delta Brac Housing Finance Corporation Limited. Delta Brac started housing finance operations in 1998 as a joint venture between Delta Brac (an NGO), the Inter- national Finance Corporation, and the Housing Development Finance Corporation of India.5 It has a business focus of 80 percent in Dhaka and 20 percent in Chittagong. The portfolio comprises 60 percent new construction and 40 percent real estate pur- chases. Delta Brac has a diversified funding base, including credit from multilateral agencies and local commercial banks, the Bangladesh Bank refinance window, and public deposits. Its cumulative housing finance portfolio is Tk 8 billion. Delta Brac has a strong marketing liaison with the builders, and gets a major por- tion of its business from large-size development projects. With its vigilant loan sur- veillance system, it is maintaining a nonperforming-loans level of about 0.5 percent. The institution is using six months as a basis for loan classification and initiation of remedial actions, including legal action against a defaulting borrower. Delta Brac is a very profitable entity with a book value per share of Tk 250 (par value Tk 100). The earnings per share for fiscal 2006/07 were Tk 55. NHFIL. NHFIL started operations in 1998 and was registered under the Companies Act 1993 as an NBFI, regulated by Bangladesh Bank. It has two licenses (as a housing finance company and as a lease finance company), with the requirement to maintain housing finance as its core business (more than 51 percent). Its geographic focus is 80 percent in Dhaka and 20 percent in Chittagong. NHFIL funds its lending operations mainly from deposits, which results in a term mismatch of assets and liabilities. Its cumulative housing finance portfolio is Tk 4,380 million to 3,081 borrowers, resulting in an average loan size of Tk 1.4 million per borrower. The size of its outstanding portfolio is Tk 1,840 million, with 1,744 borrowers. The corporation's typical lending terms are presented in table B.7. Clients are charged a prepayment penalty of 2 percent. The income is assessed on the basis of present and future income, including rental income. The loan-to-value ratio is 50­70 percent. The value is based on construction cost only, excluding the cost of land. NHFIL also finances land up to 50 percent of the cost, with a condition that the house will be built within five years. In case of default, it has the right to repos- sess the property and recover its cost with penalties/profits. Table B.7 Typical Lending Terms of NHFIL, 2009 Terms Housing Leasing Small Ticket Lease Rate (% per annum) 15.5 18.0 15.0 Maturity (y) 15 3­5 3­5 a Average loan size 1.5 20 15 (Tk millions) Source: Data from NHFIL, 2009. a. The amount, as disclosed by the management, is based on recent financing. 99 Appendix B B NHFIL has a very effective loan surveillance system. It takes postdated checks for the subsequent 12 monthly installments, along with one undated blank check. The checks are deposited on the 1st, 10th, and 15th of every month, and encashment/default is closely monitored. A borrower must have a cell phone for effective personal contact. Computer-generated standard messaging system reminders also are sent to the borrower. If a default persists, the entire outstanding balance is called back through the earlier deposited blank check. Legal action against a defaulter is initiated within six months of default. The corporation's non- performing loans are stated to be less than 2 percent in 2009. Islamic Mortgage Finance The Bangladesh market has considerable potential to absorb Islamic financial opportunities and instruments. This mode of finance can play an important role in the country because it often targets customers with lower-than-average incomes. Presently, 6 fully fledged Islamic banks and 20 Islamic banking branches of conven- tional banks are operating in Bangladesh--most prominent among them, with respect to market share, product range, and balance sheet strength, is the IBBL. A premier and pioneer in the sector, IBBL began operations in 1983. Starting with a meager asset portfolio of Tk 56 million in 1989, its asset base reached Tk 174 billion by 2007. IBBL is the largest among the private commercial banks; and it is second only to Sonali Bank Limited, the largest public sector bank, with the widest network of branches. All of its 103 branches are connected to online banking. Whereas it caters to the need for Islamic commercial banking in a large segment of the population, it also is playing a prominent role in the social service sector through a poverty allevia- tion program covering about 500,000 people and by generating employment for more than 1 million people. Among its core financial businesses are real estate investment, housing investment, rural development, and household durable schemes. Although it is better placed in providing housing solutions to a predominantly Muslim lower- and middle-income population, its penetration down the income scale remains limited. The IBBL has a countrywide presence, maintaining about 196 branches (table B.8). Its investments in the real estate sector, including both retail and project finance, have been growing steadily--from Tk 4.1 million in 1983, to Tk 69.9 million by 1998, and to Tk 9.4 billion in 2007. IBBL's portfolio comprises 80 percent urban and 20 percent rural housing finance. The average loan size in urban housing is Tk 1.5 million. Under IBBL's rural housing finance program, it has an outstanding portfolio of Tk 1.7 billion for 12,337 clients. Prudential Framework for Housing Finance, and Recent Policies Bangladesh Bank, which enjoys a high degree of autonomy, supervises and regulates all banks and NBFIs in the country. Its regulatory role encompasses prudential regulations, including minimum capital requirements, insider borrowing, loan concentration, and asset classification. The Bangladesh Bank can enforce its authority through penalties for noncompliance and through intervention in the management of problem banks. In 100 Appendix B B Table B.8 Performance of Investments under the IBBL's Special Investment Schemes, 2008 Scheme Name Staff House Housing Building SI Investment Real Estate Investment Number Particular Aspects Scheme Program Scheme 1 Total clients 471 4,839 1,276 2 Average size of investment 0.0219 0.0252 0.0268 ($ millions) 3 Amount disbursed 10.30 122.00 34.20 ($ millions) 4 Percent of actual recovery 98 95 99 5 Percent of recovery including 109 102 105 advance recovery Source: Data from Bangladesh Bank, 2008. Note: SI special investment. continued efforts toward strengthening oversight of the banking sector, the Bangladesh Bank has issued risk management guidelines for commercial banks, has mandated that banks conduct systems audits, and has placed restrictions on the number and tenure of directors at banks. The existing regulatory and legal frameworks for NBFIs are weaker than bank prudential regulations. Rules also are poorly suited to the different nature of NBFIs. For example, there is no distinction between deposit-taking outfits and non- deposit-taking ones. There is no special housing finance prudential regulation, although all housing finance institutions are subject to Bangladesh Bank regulations. The notable exception is the BHBFC, which still remains outside the ambit of Bangladesh Bank and is solely reg- ulated by the much-weaker rules of the Ministry of Finance. The ministry does not have the expertise, framework, and dynamism needed to regulate a mortgage finance com- pany. Hence, the BHBFC suffers from a very weak regulatory regime that results in innu- merable administrative inefficiencies. As an example, private housing lenders are under the Bangladesh Bank loan classification, whereas the BHBFC uses the respective Ministry of Finance version; this discrepancy results in application of different standards of port- folio quality. If the BHBFC followed the Bangladesh Bank's prudential regulations for classifying its loans, it would have made a substantially higher amount of provisions. The Ministry of Finance's loan classification guidelines are very lenient, compared with the central bank's prudential regulations; do not reflect the true health of the balance sheet; and are not revised frequently with the changing needs of financial markets. Bangladesh Bank also has some other requirements that a bank or financial institution should meet, along with its loan classification regulations. For example, banks and financial institutions need to take tangible action against nonperforming loans on a basis of six-months' default; and if nonperforming loans exceed 10 percent, the bank or financial institution is restricted from further expansion. Under pruden- tial regulation, a loan is classified after six months of default. 101 Appendix B B At present, housing finance issues are covered under the prudential regulations for consumer finance. This generalized arrangement missed out on important rules suited exclusively to housing finance activities, such as a regulatory ceiling on the "mortgage debt-to-advances" ratio on the mortgage lending of commercial banks. Recent Regulatory Policies on Housing Finance In July 2007, Bangladesh Bank started a refinance scheme to help middle-income peo- ple buy homes in cities and to encourage banks and NBFIs to lower their interest rates on home loans. Initially, a fund of Tk 3 billion was created for this scheme; later, it was enhanced to Tk 5 billion. Middle-income people earning less than Tk 50,000 a month are eligible to get loans from any participating financial institutions (up to a maximum of Tk 20 lakh). The loan is for purchase of an apartment no larger than 1,250 square feet, located in one of six divisional cities or in Gazipur, Narayanganj, Savar, or Tongi. The loan period is up to 20 years (with a 1-year grace period) at an interest rate of 9 percent. The central bank makes 100 percent refinancing of the loans disbursed, meeting the given criteria, to the participating financial institutions; repayment of the loan to the central bank is the sole responsibility of that financial institution. As of March 2009, 12 banks and 20 NBFIs had signed up for this scheme, and Tk 2.43 bil- lion had been disbursed. The scheme terms are not as attractive to the NBFIs as they are to banks, however: NBFI cost of funds still ends up higher than that of banks. This scheme is untenable over the long term, but Bangladesh Bank cannot be expected to provide development finance at concessionary rates to further the creation and growth of a vibrant housing finance market. The central bank also is not equipped to monitor the efficiency and targeting of the concessionary credit line, and may not efficiently ensure that the liquidity goes to low-income groups. Instead, a specialized mortgage refinance company is a more sustainable vehicle for addressing the maturity mismatch on housing loans helping the growth of the country's mortgage industry. Homeowners' Insurance In Bangladesh, the loan portfolio risk is generally protected through insurance cov- erage of the property and through life insurance coverage of the borrower. How- ever, there is no institution providing mortgage insurance (or title insurance, for that matter). In view of the prevailing high loan-to-value ratios adversely affecting the loan affordability of low-income borrowers, the need for a mortgage insurance institution cannot be overemphasized. Risk Management Bangladesh financial institutions' ability to identify and assess risks is crucial to their viability and sustainability. In the Bangladesh financial market, several risks are of par- ticular relevance and present considerable challenges to the housing finance sector. Liquidity risk perceptions have heightened as a result of the global financial crisis. There is indication, however, that the country's banking system is flush with liquidity, 102 Appendix B B hoarded as insurance against the vagaries of international markets. That liquidity does not extend to specialized housing finance outfits, however. In the case of the BHBFC, whose ability to attract market funds is severely compromised, lower liq- uidity translates automatically into reduced lending. The BHBFC now relies entirely on recycling loan repayments. It also maintains large working capital funds (more than Tk 3.6 billion), which are invested in low-yield "bank placements." Independent of current events, the NBFIs suffer from a fundamental term structure mismatch, which is a source of considerable risk. NBFIs such as Delta Brac fund hous- ing lending with short- to medium-term deposits. The Bangladesh Bank's limited refinance scheme, introduced in the aftermath of the global financial crisis, has provided a somewhat temporary alleviation. Bangladesh has enjoyed a stable macroeconomic environment in the last 10 years. In spite of that, market risk is not well managed by housing finance outfits in Bangladesh; instead, it is passed on to the clients, thereby reducing the affordability of mortgages and limiting the outreach of the mortgage market. The mitigation of interest rate risk is hampered by the absence of a long-term yield curve. Only recently has the Bangladesh Bank floated long-term bonds with maturity longer than 10 years. These instruments do not have any active secondary market. Credit risk is not particularly problematic for private specialized housing insti- tutions, as revealed by their 5 percent nonperforming loan rate. Commercial banks have limited their credit risk by maintaining trivially small exposure to the housing sector. In contrast, the BHBFC has a large nonperforming loan portfolio (about 47 percent, if reclassified on the basis of international prudential loan clas- sification standards). It has more than Tk 10 billion as interest receivable from the problematic portfolio, partly because of the imprudent policy of interest deferment. This unrealized interest has a sizable opportunity cost (Tk 500­1,000 million). These portfolio weaknesses have not been adequately provisioned for, creating a sustainability risk for the BHBFC. Interest Rate Market Benchmarks The Bangladesh housing finance market benchmarks itself on the long-term yield curve created by government securities with maturity up to 20 years (table B.9), although the issue and trading of such securities (especially toward the outer matu- rity limit) is thin to non-existent. There is no interbank offered rate market to serve as a benchmark. The financial market is not yet developed for active, two-way quotes that would provide benchmarks. The 10-year Treasury bonds were first issued in March 2004, at the rate of 10.00 percent. By March 2005, the rate declined to 8.95 per- cent in successive issues, and then started rising. The auction in December 2005 was at 12.00 percent, and that rate has remained stable. In July 2007, Bangladesh Bank for the first time issued a 15-year Treasury bond at 14.00 percent. Since then, through successive issues, the 15-year Treasury bond rates have declined to 12.22 percent. In July 2007, the country had seen the first issue of a 20-year Treasury bond, issued at 15.95 percent; through successive issues, its rate has come down to 13.19 percent. 103 Appendix B B Table B.9 Yields on Treasury Bills and Bonds, 2001­08 Treasury Bills (%) Treasury Bonds (%) 30 182 362 10 15 20 Period days days days 2 years 5 years years years years January 2001 6.35 6.93 7.50 8.50 10.65 n.a. n.a. n.a. July 2001 6.10 7.05 7.25 8.08 10.50 n.a. n.a. n.a. June 2002 2.74 4.91 5.33 6.85 10.48 n.a. n.a. n.a. July 2002 4.98 5.67 5.94 6.90 8.63 n.a. n.a. n.a. January 2003 8.00 7.75 10.00 10.67 11.23 n.a. n.a. n.a. July 2003 6.92 9.25 9.81 10.44 11.00 n.a. n.a. n.a. January 2004 4.03 6.00 6.50 7.00 8.00 n.a. n.a. n.a. March 2004 4.05 6.00 6.40 7.00 7.94 10.00 n.a. n.a. July 2004 4.05 6.00 6.30 7.00 7.96 9.91 n.a. n.a. January 2005 4.04 5.75 6.18 6.75 7.87 9.21 n.a. n.a. July 2005 6.65 6.75 7.00 7.25 8.75 10.00 n.a. n.a. January 2006 7.00 7.20 7.60 8.25 10.50 12.00 n.a. n.a. July 2006 7.10 7.80 8.30 9.15 10.65 12.10 n.a. n.a. January 2007 7.35 7.85 8.45 Dsc September 10.88 12.50 n.a. n.a. 2006 July 2007 7.32 7.89 8.48 n.a. 10.79 12.17 14.00 15.95 February 2008 7.35 7.96 8.46 n.a. 10.60 11.72 12.22 13.19 Source: Data from Bangladesh Bank. Note: Dsc = discontinued; n.a. = not applicable. The fixed-income market in Bangladesh remains thin, not least because of a tra- ditionally low rate of formal domestic savings, the lack of institutions with long-term liabilities (such as insurance companies and pension funds), and the crowding-out effect of the public sector debt and the National Savings System (NSS). The NSS returns, in particular, are very high, despite being addressed by government since 2001. NSS schemes, although directed at the lowest-income groups, are used much more widely because of poor enforcement, creating a bias and crowding out private borrowers. However, the NSS demonstrates that untapped savings do exist (even among lower-income groups), and that they can be attracted by the formal market if there is sufficient outreach and if pricing distortions are removed. Housing finance lending institutions' weighted average deposit and lending rates (as of June 30, 2008) are given in table B.10. Growth Potential of the Mortgage Debt Industry Bangladesh, with a housing backlog of about 5 million units, offers a great potential for the housing supply market to grow. The access to finance, with a specific reference to housing finance, is quite limited in Bangladesh. Except for the BHBFC, which has 104 Appendix B B Table B.10 Housing Finance Rates, June 30, 2008 Institution Deposit Rate (%) Lending Rate (%) All banks 7.00 12.3 All housing finance companies (excluding BHBFC) 12.16 13.5­14.0 BHBFC n.a. 12.0 For Dhaka and Chittagong; 10.0 for others Sources: Data from the Bangladesh Bank Annual Report, July 2007­June 2008, and from the BHBFC. Note: n.a. not applicable. its presence and housing finance operations in 22 districts of Bangladesh, all financial institutions have practically confined their mortgage finance business to Dhaka and Chittagong. Even in those cities, mortgage finance is geared toward economic middle and high-end segments of the population. When measures are taken to boost hous- ing supply for low- and middle-income groups, the need to empower them for afford- ability through mortgage finance will enhance the mortgage industry. Low-Income and Microfinance Housing The great majority of the population is located in rural areas (80 percent, comprising 123 million people or approximately 20 million households), and almost all of these people are classified among the lower-income groups. In urban areas where incomes are higher, the lowest-income groups are estimated at 70 percent of the population (about 21 million people; 3 million households). Both groups are in dire need of affordable housing finance, mostly for housing improvement rather than new purchases. The Microfinance Market and Potential Demand for Housing Microfinance The rural poor live predominantly in temporary housing; and estimates of rural permanent housing vary between 0.4 and 3.3 million dwelling units. Fully 3.5 million people in rural areas are without shelter. With an assumed 2 percent annual new household formation, new housing demand each year could be as much as 3.5 million. Urban dwelling units are estimated at between 3.5 and 4.0 million, of which 1.9­3.0 million are temporary. Roughly half of the low-income dwellings are esti- mated to be located in slums (both owner-occupied and rented), sometimes on squatted land. Formal slums (with full landownership but in bad repair and greatly overcrowded by urbanization) comprise another quarter of the dwellings in this group. Other categories of low-income housing include government-provided squat- ter resettlement camps, plots of land with basic services that are provided on a lease- hold basis; employee housing, consisting mostly of small apartments in high-rise complexes provided by the government; makeshift houses built by squatters on illegally occupied public or private land; and pavement dwellings. 105 Appendix B B Among the lowest-income groups, housing needs are for incremental construc- tion and repair. The potential housing finance demand thus is considerable, pro- vided housing credit is made affordable. And it is exactly that affordable housing credit that Grameen Bank began to provide in the 1980s. Products, Funding Sources, and Challenges in Housing Microfinance The two main institutions serving the needs of microfinance for housing are Grameen Bank and IBBL. Grameen Bank has a special product to finance fisher- men's huts (chhappar 6 finance). These two institutions have wide access to funds. Whereas Grameen Bank provides conventional finance, IBBL caters to the needs of faith-based clients. As of June 2009, Grameen Bank's housing portfolio amounted to $3.3 million. The rate of repayment for all loans is 89 percent. The cumulative number of houses financed so far is 674,435.7 The bank introduced the "moderate housing loan" in 1984, with a current loan maximum of Tk 25,000. The microfinance for housing pro- gram got a considerable boost after the devastating floods of 1987 that destroyed 2 million houses in rural areas of Bangladesh. Those floods raised the popularity of Grameen's "basic housing loan" (which presently has a loan maximum of Tk 12,000). That loan targets the poorest rural households, similar to income-generating credit. This program has remained the most popular among its target population. There also is a loan program for the purchase of small parcels of land and one for the repair of houses. The maximum amount for a homestead purchase loan is Tk 10,000 ($145), and the house repair loan is Tk 5,000 ($72). The ratio of basic housing loans to orig- inal moderate housing loans is approximately 7:1. In addition to those loans, another kind of housing loan is the "pre-basic housing" loan, amounting to Tk 7,500­8,500. It has been introduced to meet the demand of house dwellers in the northern part of the country. The annual interest rate on housing loans is 8 percent. The lending is done via group loans, with insurance provisions against default, death, accident, and the like. Ninety-four percent of Grameen Bank borrowers are female, and women make up the large majority of housing loan borrowers. The loan application process usually takes three to four weeks, although members can receive money in less than 10 days when the need is urgent. The area manager must appeal to the zonal member for approval. Only 4­5 percent of loan applications are rejected, usually for lack of paperwork regarding landownership. The housing loan, like the other Grameen Bank loans, is provided without collateral; but the bor- rower must have a title to the land, must sign an individual pledge that includes a repayment obligation, and must obtain a pledge from all members of the group or the center that commits them to repay the loan in case the borrower fails to do so. Repayment is made in weekly installments, beginning five weeks after the start of house construction. The repayment period is calculated on the basis of repaying Tk 1,000 a year--the standard loan repayment for other loans made by the Grameen Bank. There is a maximum repayment period of 10 years, but faster repayment is encouraged. To date, the loan repayments have been excellent. As of September 1999, 106 Appendix B B a total amount of Tk 7.42 billion ($185.09 million) has been disbursed as housing loans for 510,310 houses. The average loan size was about Tk 13,847 (only $277). The basic house proposed by the bank's housing program measures 12 feet by 18 feet; it has a two-sided tin roof, four reinforced concrete construction pillars, one wooden door, and two windows; it can be extended and modernized if desired. The standard house measures 15 feet by 21 feet and has a four-sided tin roof, eight rein- forced concrete construction pillars, one wooden door, four windows, and a fence. For loans of TK 10,000 ($202) or less, members pay TK 1,000 ($20) a year; for loans greater than TK 10,000, they divide the amount over a 10-year period. There is a maximum repayment period of 10 years. Repayment is made weekly, usually at approximately Tk 20 (40 cents) a week. The borrower is responsible for the design of the house, but the bank ensures that basic health and safety requirements are met. The house must meet minimum Grameen standards, including having a pit latrine. (Since mid-1998, the bank has required members to install a latrine manufactured by the Grameen production facilities.) The homes built under the program represent a substantial improvement over traditional low-income housing, proving markedly sturdier to heavy rain or flooding and requiring less and cheaper annual upkeep. In addition to Grameen Bank, several other MFIs, such as Proshika, Delta Brac, and ASA Bangladesh, now provide long- and short-term credit for housing in Bangladesh.8 Loans are made on the basis of established membership in lending programs instead of collateral, and on a sound track record of repayments on pre- vious loans. Group pressure and mutual support are used as guarantees for loan repayment. The financial institution's only recourse in the case of nonpayment are the persuasive and legal ways to recover the loan and future exclusion of the bor- rower from the credit program. The NGO Proshika started a housing program for rural areas in 1988 and, to date, it has provided assistance for the construction of more than 30,000 houses. The lending is group based and subject to strict conditions: there can be only three hous- ing loans for any group in any year, and housing loans can be provided only to groups who have been in existence for at least three years. Only those who have a loan for income-generating activities are eligible for a housing loan, in an effort to ensure that they have repayment capacity. Delta Brac's housing loans program started just after the flood of 1988 and focused on the rural poor. Only members who have successfully repaid an income generation loan and have saved an amount equivalent to the monthly repayments for a housing loan are eligible for this loan. Delta Brac is considering establishing a moderate- income rural housing loan program, with loans of up to Tk 20,000. It is concerned about moving into this market in view of potential massive defaults in the housing portfolio that might be caused by major floods. ASA Bangladesh started a rural housing credit program in fiscal 1989/90. Its management considers housing loans to the lowest-income group not to be feasi- ble, and intends to target the rural middle-class farmers for lending activity. It is interesting to note that a sizable group of borrowers (close to 15 percent) use part of the income-generating loans they receive for other purposes, such as improving 107 Appendix B B their homes, even though that use is explicitly forbidden. This fact is an indication that the demand for housing loans is large. The challenges facing housing microfinance programs include affordability con- straints, especially for rural households; high land prices, in the case of urban clients; commercial viability of the microfinance lenders; and the need for new products, including savings-for-housing instruments and mechanisms limiting disaster and disability/death risk in housing lending. Careful blending of government policies, smart subsidies, planning, public-private partnerships, and technical assistance for housing microlending are required to further develop the low-income housing and housing finance market. Success stories, such as the new low-income housing lending mecha- nisms in India, would be useful in the endeavors in Bangladesh. Government and NGO Programs in Housing and Housing-Related Social Services There are several government programs in support of low-income housing. In 1998, the Bangladesh government started Grihayan Tahabil, a housing fund for homeless and low-income groups. This scheme is operating in 400 upazillas across 64 districts. The funds are disbursed through a little more than 400 NGOs. The current fund size is Tk 1.600 trillion, of which Tk 1.145 trillion had been disbursed by March 2009 for con- struction of 46,128 new houses. A similar government program in rural areas, Asrayan, is intended to fund and provide the construction of low-income barrack-type houses. Government programs, however, are not as efficient as could be achieved under a commercially based approach; and they have been known to be poorly targeted in certain cases to households with incomes higher than the lower-income groups, or to be excessively directed at civil-force staff. One relevant example is the National Housing Authority (NHA), which provides subsidized housing development. The institution manages a considerable supply of government land (at its current con- struction rates, the existing owned land would permit its continuous operation for the next 20 years). The NHA obtains the land at a multiple of its registry value (150 percent), which permits it to offer the housing at levels considerably below mar- ket prices. In the last two years, the NHA constructed only 1,000 units of 60­120 square meters; it distributed them by lottery (mostly to civil servants) among the 200,000 applicants who demanded the flats. The authority's housing is more popular than that provided by the private sector not only because of its subsidized pricing, but also because of the clean title the NHA can obtain (when private developers could not). A more efficient approach to government housing provision for lower-income groups would be to rely on public-private partnerships. The NHA could scale up its operations in partnership with private developers in arrangements whereby the NHA provides the land and requires that some fraction of the flats constructed will be pro- vided to low-income households. A low-cost flat construction plan that the NHA recently has proposed involves 22,000 units of 25­30 square meters priced at Tk 400,000 each (implying a monthly payment of Tk 2,500). An alternative (and not mutually exclusive) strategy to scale up involves cross-subsidies between the pricing 108 Appendix B B of middle-income units and low-income ones. That option would permit the NHA to increase its leverage and invest the additional funds in low-cost housing construc- tion, servicing the loans in part with profits from middle-income housing sold. (The middle-income housing prices would remain below market levels.) Such an approach also would enable the NHA to leverage the housing production by generating additional income that would permit repaying the financial expenditures of any debt contracted by the NHA. (To date, the authority has not issued any debt and therefore cannot scale up its activities for fear of a resulting increase in sale housing prices.) Other lower-income housing initiatives include public-private partnerships involving Bangladesh NGOs in the area of housing-related social activities (box B.2). Bangladesh has a long and enviable history of NGOs playing an active role in raising socioeconomic conditions for lower-income groups, particularly those living in rural areas. Some NGOs have quite successfully spread their network throughout Bangladesh, and are carrying out efficient operations. Box B.2 Role of NGOs in Housing-Related Social Services Housing Construction Financing There is no direct provision of housing by NGOs in Bangladesh's urban areas because of land availability constraints. New housing construction is also not funded in urban areas because most lower-income groups do not own land. Some NGOs are seeking funds to build houses for their project beneficiaries, and intend to recover the cost through an affordable schedule. Some are generating funds through members' savings and income-generating activities. A few large NGOs are planning to introduce urban housing credit, subject to finding viable methods to over- come the landownership problem. Service Provision The NGOs have undertaken limited projects to improve sanitary and drainage systems, provide potable water, regularize garbage clearance, and pave roads. The beneficiaries participate in some stages of the project, such as advocacy and identification, decision making, negotiation, construction, toll collection, monitoring, and maintenance. The NGOs are operating several successful and innovative urban water supply and sanitation projects in which they've mobi- lized the households; negotiated for them with the authorities who do not cover bustees; and given guarantee that the facilities would be maintained, costs would be met, and bills would be paid regularly. Some NGOs have promoted door-to-door garbage collection, and have taken charge of the general neighborhood cleanliness and street sweeping. A few self-sustaining waste-recycling plants have been set up as well. Housing Improvement Credit Credit has been provided in a few cases after clients' income-generating activities were ham- pered by natural disaster that left the beneficiaries homeless. Some NGOs motivated the (Continued) 109 Appendix B B Box B.2 Role of NGOs in Housing-Related Social Services (continued) landowners, negotiated bank loans, and initiated slum-upgrading projects involving physical con- struction. These housing improvement loans would not have been possible without the NGO intermediating with the financial institution, because the latter generally requires clean title of ownership--something rarely available to lower-income households. Special Housing Several NGOs are catering to the housing needs of specific groups, like garment workers, destitute and delinquent women and youth, street children, orphans, mentally retarded and physically disabled children, and working mothers. Their dormitory-type rental accommoda- tions are part of the empowerment or rehabilitation programs that comprise literacy and skills training; legal, social, and entrepreneurial support; health facilities; and mother and child care facilities, among other things. Some NGOs, run by women, have successfully acquired funds to construct their own buildings, which will cross-subsidize loan repayment costs, mainte- nance, and operation costs via their partly commercial rental use. Slum Development The government-initiated Slum Improvement Program has been undertaken by a few NGOs. These organizations focus on income generation; health and family planning; education; infra- structure; and services such as roads, drainage, water supply, and sanitation aimed at environ- mental upgrading and enhancement of affordability to improve the housing situation indirectly. Along with the NGOs, several clubs, youth organizations, and social welfare societies are engaged in various components of the overall programs. Management Support Several NGOs and community-based organizations have undertaken the housing maintenance and utilities management functions for the implementing housing agencies. The outsourced functions can include operation and maintenance of water supply, sanitation, drainage and garbage collection services, collection of fees, bills and maintenance expenditure, collection of technicians and management staffs' salaries, collection and deposition of loan repayments, liai- son with banks and completion of documents, monitoring of housing and utility services, and credit management for a large number of beneficiaries. Advocacy Some NGOs and community-based organizations offer legal aid to low-income households facing eviction. Additional advocacy programs include seminars, workshops, gatherings, street processions, celebration of important days, group meetings, trainings and symposiums, and exhibitions of building materials and demonstration houses. These efforts target both the urban poor and policy makers; and they cover housing and environmental rights, the role of savings and microcredit, human rights, and gender issues. Source: Study conducted by Mahbubur Rahman, North South University, Dhaka, Bangladesh. 110 Appendix B B Regulatory Framework and Mortgage Market Infrastructure Titling and registration, secured transaction and foreclosure rules, and the availabil- ity of credit information are important for the development of vibrant housing and housing finance markets. In Bangladesh, the legal framework suffers from certain shortcomings considered in detail below. Land Registration and Administration The structure of land administration and the land record system in Bangladesh con- tains weaknesses and dysfunctional elements. The present land laws and supporting regulations are complicated, and there is no effective legal framework. The absence of an updated, systematic, and universally accepted source of information on land resource availability and land rights is one of the principal barriers to the successful implementation of land reform programs leading to the effective adjudication of land disputes. Land surveys are an important tool for identifying land, documenting the names of the rightful owners, measuring the area, and other related matters. Outdated processes, corruption, and irregularities in the land survey deprive the illiterate, landless, and minority groups of ownership rights to their land. Even the govern- ment is being deprived of state-owned khas land.9 There also is a general lack of pub- lic awareness about land resources, procedures, and regulations. Land surveys have been conducted over different periods, beginning with the cadastral survey, the state acquisition survey, and the provisional settlement (still in use to this day).10 Cadastral and state acquisition survey records are kept at the Zila (district) Collector's office, and provisional settlement survey records remain with the Assistant Commissioner's (Land) office. The present structure of land administration is built around three core functions--land settlement or survey, land registration, and recordkeeping. Land settlement identifies the utilization purpose of land--that is, use for agriculture, housing, ponds, roads, waterways--and all other information related to the land. Registration deals with registering land transfers and preparing the deeds. Record- keeping documents ownership rights to the land. Registration and recordkeeping are based on information gathered from the land settlement. Three separate offices deal with the different functions of land administration. The Tahsil office is responsible for land recordkeeping and collection of land rev- enue; registration is entrusted to the subregistrar's office; and the settlement office handles land settlement. The Tahsil office and the settlement office have a chain of command extending to the Ministry of Land, whereas the subregistrar's office is under the purview of the Ministry of Law. The overlap in ministerial functions is problematic in the presence of conflicting land claims, where resolving the disputes would require proof of ownership. Diverging ownership proof from the Tahsil office, the subregistrar's office, and the settlement office render it extremely difficult to adjudicate ownership because each document seemingly bears the legitimacy of the 111 Appendix B B authority. This multiplicity of documents or records of rights is the central flaw in the system of land administration. Another flaw is that registration officers are not entitled to question the validity of documents brought for registration. Furthermore, their office does not store sup- porting documents for verification. Hence, even if registration officers knowingly register false transfers, they cannot be prosecuted. The breakdown of information and the lack of a system of checks and balances among the three offices allow regis- tration of numerous false land transfers. The bulk of civil and criminal litigation in the country arises from such conflicting claims of ownership. Adding to and compounding the problem is an inefficient system of dispute resolution. Doing Business 2010 (World Bank 2009a) ranks Bangladesh fourth among the worst coun- tries in the world in contract enforcement.11 As a result, land remains in a legal vacuum for many years; and it is difficult to develop, finance, or sell properties. These problems stunt the mobilization of land for urban needs, block creation of a market for existing housing, and considerably undermine the potential scope and the security of formal housing finance. The weak enforcement regime also stunts fore- closure (as discussed in the following section). A third flaw challenging land administration is the excessive backlog of land- related documents awaiting printing. Certain records also have been intentionally destroyed or were damaged during the Bangladesh independence war. The land reg- istration system also is greatly marred by corruption. According to Transparency International (2005), almost 40 percent of all bribes are paid in land administration departments. Doing Business 2010 (World Bank 2009a) rates property registration procedures in Bangladesh among the 10 slowest in the world (table B.11 and box B.3). Register- ing property involves eight procedures, takes 245 days, and costs 10.24 percent of the property value. In contrast, the procedures are five in India, and they are completed in 44 days at the cost of 7.43 percent of the property value. The figures in Pakistan and Sri Lanka also are much better than those in Bangladesh.12 The government owns large amounts of khas land in different parts of the coun- try. With no single conclusive record or document related to land, however, the dis- tribution of these lands among the poorest sections of the population is a daunting proposition. The government does not have firm statistics on the total area available, whether the land is in the control of the authorities, whether certain areas have been affected by river erosion, and so forth. There have been previous attempts to redis- tribute khas land, and different laws have been passed for that purpose. The lack of proper implementation, however, has resulted in the landless people invariably being left out. Land Titling Land title authentication is much easier in the case of government-owned land. Ver- ifying title for private land is rather cumbersome and time consuming. Assuming that the client already has completed all transfer formalities (mutations and so forth) 112 Appendix B B Table B.11 Property Registration Process and Costs Procedure Time Required to Complete Cost Required to Complete 1--Verify the record of rights 15­60 days (simultaneous with Tk 2,000 from the Land Office (also known procedures 2 and 3) as Land Revenue Office) 2--Conduct survey mutation on 60 days (simultaneous with Tk 6,000, approximately propertya procedures 1 and 3) 3--Obtain inspection for survey 60 days (simultaneous with Included in procedure 2 mutation procedures 1 and 2) 4--Obtain the nonencumbrance 2­3 days Tk 500­700 certificate from the relevant subregistry office 5--Prepare deed of transfer and 1 day 5% of property value pay stamp duty (stamp duty) 6--Pay capital gains tax, registration 1 day Capital gains tax (5.0%) VAT fee, VAT, and other taxes at a (1.5%) local government tax designated bank (2.0%) registration fee (2.5%) 7--Apply for registration at the 180 days (simultaneous with Paid in procedure 6 relevant subregistry office procedure 8) 8--Register the change in ownership 45­60 days (simultaneous Tk 5,000 at the Land Revenue Office with procedure 7) Source: Information available at http://www.doingbusiness.org/Exploretopics/registeringproperty/. Note: VAT value-added tax. a. This is a conversion (mutation) of the property entry in the relevant registration land survey, for updating purposes. The process results in the issuance of a "mutation certificate" from the assistant commissioner of land. and has paid all taxes and dues, it usually takes about 60 days.13 In contrast, the regulation requires the process to be completed in 2 to 3 days. Titling is even more complicated in the case of new developments where property sales are made before construction is completed, involving advance payments before the transfer of property rights. Buyers are not protected against the developers' insol- vency during the development process. The subdivision of new multiunit complexes tends to hamper the financing of the construction cycle and to favor buyers who can make large downpayments. The formal registration of the transfer of ownership is generally the responsibility of the developers, who tend to wait for the complete mar- keting of the project before transferring ownership, thus creating for the buyer a legal vacuum that can last several years. Some lenders provide loans during this time (some- times through a three-party agreement with the developer and the purchaser), but the financial mechanism is not secured other than by personal assessments of the borrow- ers and the developers. It would seem difficult for such a mechanism to be extended to large-scale lending, new developers, or moderate-income borrowers. Financial institu- tions, faced with such uncertainties, either abstain from lending or restrict their services to wealthy borrowers by relying on income creditworthiness alone instead of considering property collateral when making underwriting decisions. 113 Appendix B B Box B.3 Property Registration in Bangladesh The longest delay is in registering the property at the municipal deed registry office--a process that takes between 180 and 540 days. The buyer may obtain a certified registration document within a week, but obtaining the original certificate may require 6­18 months, or even as much as two years in some cases. In Sri Lanka, by contrast, it takes only 36 days to register at the land registry. In Pakistan, it takes 38 days. In Nepal, it takes only 1­2 days for registration of the deed at the land revenue office and issuance of a new title certificate. Obtaining from the municipality office (the RAJUK) the permission to transfer property ownership--one of the early steps in the process--adds 60 days in Bangladesh. Verification of the record of rights from the land revenue office adds another 15­60 days. The ease of registering property varies by city. Although there are seven common registra- tion procedures throughout the country, for land being developed by a city development authority (such as RAJUK's development of model towns in Banani, Baridhara, Gulshan, the Nikunja Residential Area, and Uttara), one additional step is required: obtaining permission from the city development authority to transfer ownership of the property. That one additional step makes the registration process in Dhaka lengthier and costlier than in the rest of the coun- try. This obligation also is a source of distortion because the BHBFC, contrary to other lenders, is exempt from it. In Bogra and Chittagong, it takes 391 days and costs 10.0 percent of prop- erty value; in Khulna, it takes 373 days and costs 9.0 percent of property value; and in Dhaka, it takes 425 days and costs 10.5 percent of property value. Reforms have started to make property registration easier. A Land Registration Act came into force on July 1, 2005. It is intended to help reduce false and multiple registrations of land. Registration fees were cut by 1.0 percent, although they remain high overall (at 10.5 percent). A pilot project to computerize land records has been completed in the Demra region of the Dhaka district. Title deed requirements are being simplified, with the title, the location of the land, and the map on a single page; and all other documents in a backup database. The government is moving ahead in replicating the pilot model in phases in other subdistricts of the country under a public-private partnership. Simple reforms also can have a big impact. Bangladesh can encour- age formal property transactions by further reducing the registration fees and stamp duties, which cost 3 percent and 5 percent of the property value, respectively. By contrast, the stamp duty in Pakistan is 3 percent and the registration fee is 1 percent of the property value. In Sri Lanka, there is no separate registration fee; it is subsumed within a stamp duty set at 4 percent of the property value. In India, although the stamp duty is 5 percent of the property value, the registration fee is only 1 percent. Typically reductions in fees lead to more revenues because more properties are registered and there is less underreporting of property values. Source: World Bank 2007. Capacity of Housing and Housing Finance Professionals There is a shortage of professionals with adequate training in security valuation, title search, property valuation, and credit appraisal. Bangladesh also suffers from a dearth of real estate brokers and appraisers. There is a need to develop and run training and capacity-building programs covering these areas of expertise in different parts of the 114 Appendix B B country.Assessment of property value is performed by the civil engineering department of BHBFC--a department present in all BHBFC loan processing centers. These experts perform site inspection, review drawings to assess covered area, review structural design/drawings, prepare cost estimates using approved construction rates, verify con- struction progress (in case of staged loan disbursements), review necessary approvals by competent authorities, and issue letters of permission to mortgage. Most of the housing finance institutions do not go into such engineering detail, and rely instead on approval of the relevant government departments. They use the covered area shown on the draw- ings and the approved construction rates to assess the construction value. Private mort- gage lenders also outsource part of the duties to approved "valuation" firms. At present, there is no single repository responsible for gathering, processing, and disseminating the data and information relating to housing and housing finance. Bangladesh Bank gathers some data on housing finance from the banks and finan- cial institutions. However, critical databases of housing, real estate, and land pricing; construction costs; and housing finance data used by modern housing finance insti- tutions are not available. It is further important to have a database that would be responsible for developing real estate sector policy suggestions to be considered by the relevant government institutions and regulators. Constraints for Women In Bangladesh, there are no gender issues for accessing finance in general, regardless of the size of the financing sought. This is even more evident for microfinance and housing microfinance. The living example of Grameen Bank proves that point: more than 90 percent of its clients are women. Foreclosure and Eviction The inability to enforce collateral weakens financial institutions' ability to manage credit risk efficiently. Enforcement of property liens is largely annihilated by the length and cost of cumbersome judicial procedures. Foreclosure is regulated by the Money Loan Court Act 2003. Purporting to expedite procedures and to establish a new method of dispute settlement under the supervision of magistrates, the act set up a special court to deal exclusively with loan defaults exceeding 5 lakhs, prescribed time limits for granting judgments, and imposed restrictions on appeals. In addition, a new provision was adopted in 2004 to enable lenders to request judiciary support when evicting convicted defaulters. It remains to be seen if these changes will remedy the inefficiencies of the enforcement process, which is plagued by delaying tactics, diffi- culties in vacating properties to be auctioned, and huge backlogs of cases despite the inception of Money Loan Courts in 1990. Other impediments remain--notably the cumbersome and sometimes dangerous auction procedures, and the weak demand for repossessed properties. Currently, in spite of recent improvements, foreclosure remains a complex, lengthy, and ultimately ineffective procedure. The weak enforce- ment of the legal framework in Bangladesh, as discussed above, further aggravates the 115 Appendix B B foreclosure regime to the extent that it encourages willful defaults among higher-end borrowers. The weakness of the formal procedures has pushed private financial insti- tutions to use alternative means of enforcement, such as postdated checks; in the meantime, the BHBFC has accumulated a considerable nonperforming portfolio. The BHBFC benefits from special foreclosure rights that might strengthen its unfair advantage against private mortgage lenders. Presidential Order 1973 regulates the BHBFC's power to enforce a mortgage security through the court, and the Direct Sales Rule 1977 allows the BHBFC a nonjudicial procedure of foreclosure. The court case is tried before the civil judge (district level), followed by a further case for the execution of the court decree. Appeal is at the level of the High Court, but appeal actions do not stay the BHBFC's recovery process. After a stage of due notification to the delinquent borrower, the direct sale option allows for an outright sale of the mortgaged property by the BHBFC, either through public auction or by calling a public tender.14 The direct sale procedure allows excessive appeals, challenging the direct sale of the property and causing delay. As a result, to the extent that foreclo- sure is exercised (which is extremely rare), the BHBFC would be better advised to use the safer option of a court case to establish default, followed by a direct sale. This excessively lengthens the procedure, in effect negating the advantages of a direct sale and rendering foreclosure ineffective. Credit Bureaus A public credit registry functions reasonably well in Bangladesh, providing satisfac- tory, if not rapid, information about potential borrowers. More effective information sharing can be further encouraged by computerizing the credit information; making available at least two years of historical data; expanding the registry's coverage to include information on loans under $800; and expanding the credit registry to cover information from MFIs and nonfinancial institutions, such as utilities, retailers, and trade creditors. Policy Options for Improving the Framework for Housing Finance Access to housing is important to meeting urbanization and demographics challenges and preventing slum proliferation. An active system of housing finance provides key real 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 an indirect positive impact on employment levels, the retire- ment system, fiscal returns, and consumption. Housing finance development plays a role in boosting equitable economic growth and reducing poverty by helping house- holds build assets, by improving living conditions, by empowering the middle- and lower-income populations, and by strengthening communities. Bangladesh requires a careful reform strategy to be addressed effectively, thereby spurring growth in home financing to a wider population group at a lower cost. 116 Appendix B B Housing finance in the country is limited to the top income group; titling is not foolproof; mortgage lenders have to rely on short-term funding to finance the loans they give out; and the legal, regulatory, and taxation frameworks are not enabling primary and secondary housing finance markets to develop. The timing for further reforms is appropriate because the production of accessible housing for a larger part of the urban population is a national priority of the authorities, and because both banks and private NBFIs are scaling up their mortgage production. However, housing finance reform work cannot be effective in a vacuum; without radical improvements in regulatory enforcement and property registration, housing finance development will not be effective. The policy options below include strengthening the regulatory framework and its enforcement, customizing prudential norms to housing finance, developing a more effective collateral and foreclosure regime, improving housing and mortgage infor- mation availability, and overhauling land administration. The policy options also address the revamping of the BHBFC, the promotion of long-term financing via a refinancing facility and a national savings-for-housing scheme, the encouragement of a secondary housing market via a level playing field among all market participants, and the improved affordability of mortgage financing by promoting microlending for housing and innovative housing finance instruments. Strengthening the Regulatory Framework and Improving Enforcement The most pressing elements for reform in the institutional and regulatory spheres are these: · To level the field among housing finance providers, eliminating all privileges enjoyed by the BHBFC or privileges enjoyed by banks over specialized housing institutions; and to prepare these entities to be better able to compete in a dynamic market. · To customize prudential regulations in a way that can improve the sustainability of housing finance operations and incentivize the efficient management of credit risk. Real estate finance involves specificities--property collateralization, price cycles, market and financial risks of property development, and balance sheet mismatches induced by long-term lending or fixed interest rates that require spe- cial prudential rules. The prudential framework could be customized to reflect these specificities.15 · To provide favorable tax incentives that can encourage desired behavior to stim- ulate the secondary mortgage market. Promising directions include tax relief on mortgage payments,16 limitation of the income tax exemptions of competing sav- ings instruments (such as NSS interest), the introduction of levies on speculative gains and idle properties, and the use of property taxes to spread the infrastruc- ture cost.17 · To streamline lending procedures. Age-old insurance and trust acts are not facili- tating lending by long-term investors in the housing sector. Developers sometimes 117 Appendix B B are accused of not properly following the procedures for borrowing money from financial organizations. The involvement of all three parties (the developer, the landowner from whom the land was leased, and the apartment buyer) in the process makes the borrowing cumbersome. Efficient System for Land Administration A functional registration/titling system is a vital precondition for mortgage lend- ing. Computerizing and modernizing the titling registries, scaling up the Dhaka pilot projects, is a priority reform in land administration to improve property rights and decrease mortgage lender risks regarding collateral values and to elimi- nate accumulated backlogs. That would also speed up title verifications, which cur- rently are cumbersome and time consuming. A systematic cadastral survey must be implemented to safeguard property values, provide information, and improve transparency and good governance practices. An updated, systematic, and univer- sally accepted source of information on land resource availability and land rights is needed to aid in the implementation of land reform programs and in the reso- lution of land disputes. Furthermore, streamlining of jurisdictions between the ministries of law and land would enable better checks and balances and improved coordination of land administration activities, and would diminish the scope for property disputes. The registration and transfer taxes require a downward revision. Bangladesh can encourage formal property transactions by further reducing stamp duties (currently at 5 percent of the property value). By contrast, the stamp duty in Pakistan is 3 per- cent and the registration fee is 1 percent of the property value. In Sri Lanka, there is no separate registration fee; it is subsumed within a stamp duty set at 4 percent of the property value. In India, although the stamp duty is 5 percent of property value, the registration fee is only 1 percent. Typically, reductions in fees lead to more rev- enues because more properties are registered and there is less underreporting of property values. The property registration procedure could be streamlined, using the one-stop- shop model.18 Reforms to make property registration easier have already started. The Land Registration Act that came into force on July 1, 2005, will help reduce false and multiple registrations of land. Registration fees were cut by 1 percent, although they remain high overall (at 10.5 percent). Title deed requirements are being simplified-- with the title, the location of the land, and the map on a single page and all other documents in a backup database. An urgent set of measures involves the overhaul and institutional beefing-up of the land policy framework, including urban planning, state-owned land provisions in underused urban spaces, incentives (particularly through specific taxation), and the building of related professional expertise. It is crucial to decentralize, empower, and revamp the resources and capacity of RAJUK and the other city development authorities. Capacity is required in terms of public sector planning professionals, overhauled master plans, faster processes, and improved efficiency. 118 Appendix B B Collateral/Lien Registry and Credit Information Bureau The inability to enforce collateral weakens financial institutions' ability to manage credit risk efficiently. The enforcement of property liens is largely annihilated by the length and cost of cumbersome judicial procedures. Foreclosure procedures require an overhaul, transforming them into a quicker and simpler process of collateral dis- posal in case of default. A recent reform with several important lessons learned in this aspect is the "parate" right in Sri Lanka, which permits mortgage lenders to fore- close and sell all defaulted property without going to court. Several improvements to the public credit registry are advisable. More effective information sharing can be further encouraged by computerizing credit informa- tion; making available at least two years of historical data; expanding the registry's coverage to include information on loans under $800; and broadening the credit reg- istry to cover information from microfinance institutions and nonfinancial institu- tions, such as utilities, retailers, and trade creditors. Housing Observatory There is a need to set up a Housing Observatory either at the Ministry of Housing, the Ministry of Planning, or the Bangladesh Bank, or as an independent entity sup- ported by all the stakeholders. The observatory would maintain the following databases: · database on changes in real estate prices, over time and between different localities; · database on costs of construction; · database on land price movements; · database on supply of housing in different price categories--low, medium, and high; · database on housing demand and the demand-supply gap; · database on affordable housing solutions; and · database on existing operational policies like loan-to-value ratios, mortgage rates, loan tenures, income, and affordability. Capacity Building for Financial Institutions and Housing-Related Professionals Capacity and institution building are strongly needed among all the main stake- holders actively engaged in housing finance-related activities. For many people, those activities are still relatively new or radically different from past policies and procedures. Capacity-building activities include training of both public and private sector agents and institutions: bank supervisors, members of the ministries of finance and housing, magistrates and other judicial communities, staff of financial institutions, developers and builders, brokers, appraisers, and lawyers. 119 Appendix B B Revamping the BHBFC National housing finance systems have failed to prove themselves competitive in mobilizing and allocating capital efficiently. The BHBFC has made only very limited progress in becoming self-sustaining or in reaching lower-income households. Instead, it currently is subsidized, is not competitive, crowds out private sector mortgage lending, and mistargets subsidies to groups other than those with lower incomes. It also suffers from improper procedures and systems, an inadequate credit underwriting and servicing policy, a high proportion of nonperforming loans, and weak debt recovery. The corporation could benefit considerably from overhauling its pricing and product mix, risk management, resources, corporate governance, legal status, and financial and business planning; and from identifying a viable recapital- ization strategy and an optimal capital structure. Last but not least, the BHBFC needs to be put under Bangladesh Bank regulations, like all other housing finance lenders. Strengthening Low-Income (Microfinance) Housing Lending Enhancing access to housing finance for lower-income groups is a priority in the Bangladesh context, where less than the top 10 percent of the population (by income) currently benefits from such access. The customer base will broaden natu- rally in response to reduced intermediation costs, more competition, more efficient institutions, fixed (and lower) interest rates, and longer loan maturities. Tax- deductible mortgage payments would make mortgages more affordable for the middle-income groups. Strengthening the enforceability of mortgage rights would have the additional effect of extending mortgage finance to lower-income groups because financial institutions would grow less sensitive to client income and rely more on the value of the enforced collateral. Stretching affordability farther down the income range would require assistance programs targeting bankable lower- income households and a scheme designed to leverage public spending by market resources. The most efficient subsidy schemes are those that are directly allocated to the beneficiaries and that include a preliminary savings requirement as an eligibility criterion to trigger a leverage effect with private resources. The assistance scheme could be linked to the savings-for-housing scheme (see the "Mortgage Product Innovation" section below), given the synergies between the two mechanisms. Savings-for-housing schemes would go a long way toward spreading housing finance outreach by helping lower-income savers build up their downpayments. Islamic finance can play a considerable role in enlarging housing finance access because it often targets customers of lower-than-average income. Microfinance, which is particularly developed in Bangladesh, can play a consid- erable role in developing the housing finance market. This is particularly so in rural areas, and for house upgrades and maintenance in urban slums. To solidify the link between microfinance and housing, policy makers would need to tackle two chal- lenges: (1) improve the commercial sustainability of microfinance housing products, 120 Appendix B B and (2) extend microfinance for housing to urban households (Irving Fisher Com- mittee 2009). To be viable and profitable, microfinance housing lenders cannot rely on internal cross-subsidization by other microfinance products. That reliance increases the cost of microcredit and limits the scale of MFI housing programs. MFI housing programs should be separated, customized to client needs, matched with appropriate longer- term funding sources, and defined by the institution's own lending terms and con- ditions such that it breaks even or turns a profit. A second consideration to make housing credit programs more sustainable is to target households that already have established a solid income base through previous income-generating credit pro- grams or microfinance loans. Third, a more stringent savings requirement could be built into the housing loan package--not only to cover up-front payment for part of the house, but also as a savings account that would be accessible in case of possible delinquency. Finally, the lending process should carefully assess housing quality and location from the perspective of disaster vulnerability. Extending microfinance for housing to urban households would prove more challenging, given high land prices caused by poorly designed land administration policy in Bangladesh. A housing improvement loan program for owners of informal housing may be a starting point for extending microcredit to urban households--for example, in conjunction with existing services improvement projects in bustees. Sec- ond, lending options for new urban lower-income housing should be analyzed care- fully to explore the feasibility of a public-private partnership with local government land allocation and development programs. In addition, NGOs and MFIs could experiment with joint land ownership forms for low-income housing projects, through which credit for ownership housing is provided for the unit only while the group retains ownership of the land. Finally, to serve urban households, MFI pro- grams would require a different set of underwriting criteria, equity and savings arrangements, and interest rate structures based on the actual risks to which they are exposed. Subsidies may have to shift more toward land cost than to the cost of finance. Mortgage Product Innovation Further housing finance market development crucially hinges on product innova- tion, which would bring down lending costs, customize loans to client needs, enable the financial institutions to better manage risk, and help extend supply farther down the income scale. A well-studied example of a successful product is the contractual savings scheme, or savings-for-housing (box B.4). The product involves several years of saving by the client, who thus establishes a good reputation with the financial institution and accumulates a downpayment for the future home. Following the sav- ings period, the client receives a proper mortgage loan and purchases the home. This instrument is very appropriate in developing countries where credit records are not widespread, the population is mainly young, and savings are not customary. In view of the savings habits in Bangladesh, as evidenced by the outreach of the NSS and its 121 Appendix B B Box B.4 Contractual Savings Schemes for Housing: A Primer Contractual savings schemes for housing (CSHs) have been a traditional means of raising hous- ing finance (for example, in France, Germany, and Mexico), and most recently have been adopted by Central and Eastern Europe, China, India, Morocco, Nicaragua, Peru, and Tunisia. Several public housing institutions in Africa, Asia, and Latin America also have started imple- menting CSHs as a means to attract low-cost deposits. The concept is simple, requiring the potential borrower to save money in fixed installments over a number of years, thus building up some equity while demonstrating to the bank his or her reliability and capacity to repay a debt. The savings can be withdrawn at any time; however, the bank does not give out a housing loan if funds are withdrawn early. This incentive structure converts an ostensibly short-term deposit into a de facto long-term deposit, thereby mitigating liquidity risk. When the saving period is over, the bank advances a loan proportional to the amount already saved; this provides an incentive for the client to save more. Interest rates for savings generally are fixed below the market rate and the rates on the loan are fixed below market, thus pro- tecting the client from interest rate risk in a manner that is sustainable and profitable for the bank. CSH loan pricing generally is identical for all savers because the savings behavior already reveals borrower quality and promise. This makes CSH administration cheap and easy. These schemes have been justified on the basis of six central arguments: 1. CSHs fill the void left by the lack of long-term funding instruments, hindering specifically the development of fixed-rate mortgage products. The aggregate liquidity management depends crucially on how credible the scheme is as a generator of loans. 2. CSHs address the problems of access to mortgage finance for young and low-income households that are caused by high downpayments and high credit risk management costs. 3. In countries where credit information on borrowers is scarce, CSHs will enable financial insti- tutions to accumulate credit information for a large number of individual borrowers--and thus lower credit risk. 4. CSHs act as a means to generate loan supply that is lacking in areas not covered by stan- dard mortgage finance and characterized by low loan volumes and high servicing costs, especially modernization and small transaction loans. 5. CSHs contribute to a greater mobilization of savings and, therefore, to economic investment. 6. CSHs have a comparative advantage in addressing risk of mortgage products, which is espe- cially high in emerging markets characterized by high levels of credit and inflation risk. These savings schemes can be structured in a way to provide considerable protection from interest rate risk for the home buyer, which is more valuable in countries with higher interest rate volatility. use as a source of pension supplements, a savings-for-housing scheme will prove popular. It also will boost national savings, provide a stable source for long-term funds in the financial system, and widen access to mortgage finance among lower- income groups. 122 Appendix B B Mortgage Refinancing Facility The Bangladesh housing market lacks long-term funds. This lack puts NBFIs on a unequal footing with banks, which can benefit from their deposits as a source of cheap funding. To address the twin issues of liquidity and term-mismatch risk, international best practice suggests the creation of a mortgage refinancing facility-- a type of housing fund serving to take mortgages off the balance sheet of financial institutions and to inject the required liquidity so further mortgage lending is pos- sible. The same effect can be achieved by securitization (in countries where the institutional and developmental prerequisites for such action are present). The cur- rent liquidity facility maintained by Bangladesh Bank somewhat serves such func- tions; but it is not sustainable over the long term, is not organized on a market basis, and is not equipped to monitor the precise targeting of subsidies to lower-income groups. The current Bangladesh Bank housing credit line therefore would benefit from being transformed, over the long term, into a liquidity facility capable of issuing its own private bonds. This transformation will separate market long-term debt from interest rate subsidies, enabling more efficient and transparent targeting of the sub- sidy. A facility could refinance all eligible lenders (both banks and finance compa- nies) who provide high-quality mortgage loans as collateral. The facility would issue low-risk private bonds among institutional investors. (A key success factor to keep the price attractive to lenders would be making these bonds eligible as statutory liq- uidity ratio assets.) This mechanism would centralize funding requirements, catalyze the primary market for housing finance, provide matching finance at an acceptable cost, encourage discipline by requiring higher lending and servicing standards, and provide some form of institutional support to help build the link to securitization in the long term (following the best practice of Malaysia's Cagamas). Enabling Developers Further work is needed to examine options for securing the advance payments made by clients to developers, including the possibility of introducing either escrow or insurance guarantee programs to protect advances and ensure the completion of projects. In particular, an insurance guarantee scheme could be devised under the management of the Real Estate and Housing Association of Bangladesh, with fees paid by developers. Both ways may imply additional costs and field inspectors. Solu- tions to improve the standards of their profession and scale up their activities should be discussed with the best developers through the association. The legal framework on titling in the case of new developments should be clari- fied and revised for robustness, avoiding a legal vacuum for the ownership title of people buying homes, and protecting them from the bankruptcy or underperfor- mance of specifications by the developer as well as from ownership uncertainty. A simplification of property transfer approval procedures could permit title transfer during construction. 123 Appendix B B Notes 1. A good example of such a study is the Monitor Group's research in India. 2. There are four reinforced concrete pillars on brick foundations at the corners of the house and six intermediary bamboo posts, with bamboo tie beams, wooden rafters, and purlins support- ing sheets of corrugated iron roofing. This design provides stability during floods and strong monsoon winds, and protection from the heavy rain during monsoons. Pillars and sanitary latrine are being provided by the concerned NGO. The housing program of the NGOs is usually linked with livelihood promotion activities to improve the income and food security of the fam- ilies. A range of skills is gained as a result of the integrated program, including skills in con- struction (carpentry, making of reinforced concrete construction pillars, masonry, brick/stone cutting, rope making, bamboo treatment, blacksmithing, plumbing, and so forth), forestry, live- stock raising, and home-based microenterprises. 3. Add-ons include processing and servicing fees and monthly interest computed on nonamortiz- ing balances within one year. 4. For "classification" of its nonperforming portfolio, the BHBFC follows the Ministry of Finance directives. Those directives are less strict than the prudential regulations of Bangladesh Bank's standard classification of nonperforming loans, which would more than double their amount. 5. Delta Brac is currently owned 25 percent by Brac, 15 percent by Housing Development Finance Corporation (India), 15 percent by the International Finance Corporation, 20 percent by insur- ance companies, and the remainder by the general public. 6. A chhappar is a hut or a shelter made of straw and bamboo. 7. Data are available at http://grameen-info.org/index.php?option=com_content&task= view&id= 453&Itemid=527. 8. The remaining part of this section is based on information found in Irving Fisher Committee (2009). 9. Khas land is state land under the administration of the Ministry of Land. Its state ownership arises either from seizure of land in excess of land reform legislation's 3.3-acre ceiling or from the rezoning of land. 10. Cadastral surveys were conducted between 1890 and 1940. State acquisition surveys were con- ducted from 1956 to 1962. Land surveys began in 1965. The khatian (the code number related to ownership of a particular land area) surveyed is known as the cadastral survey, or the Zila khatian. Cadastral survey records comprise landownership records of the Indian Emperor, Zamindari (landlords), and the general public. The state acquisition survey records consist of landownership by the general public and the government. After the abolition of the Zamindari system in 1950, lands belonging to them were converted into khas land and recorded under kha- tian 1. They consist of landownership records of the general public and the government. 11. Bangladesh ranks 180 out of 183 countries in 2010. It takes 41 procedures and 1,442 days to enforce a simple claim (a bounced check), and costs about 63 percent of the value of the claim. This is very similar to the region as a whole: the numbers in India, Pakistan, and Sri Lanka are 40­47 proce- dures and 976­1,420 days. Costs are significantly lower in the region, however, at 23­40 percent. 12. In Pakistan, the procedures are 6, completed in 50 days at the cost of 7.25 percent of property value. In Sri Lanka, the corresponding figures are 8 procedures, 83 days, and 5.11 percent. 13. The exception to that number of days occurs in certain relatively better-managed areas of Dhaka City. 14. If no bidder comes forward with a bid above the reserve price, even after three consecutive bids, then the mortgaged property is purchased by the BHBFC. 15. First, asset/liability management could be strengthened by requiring that long-term lenders be appropriately equipped to monitor, measure, and document interest rate and liquidity risks. Second, the framework should induce lenders to follow sound risk management practices. Two areas are worth mentioning in this regard: The lending criteria raise some concern insofar as they include without "haircut" rental income to assess the ability of borrowers to repay the loans, when this type of income is directly exposed to a (real estate) market risk. Valuation standards should be defined to 124 Appendix B B make collateralized lending rely more on market values--and particularly resale values, in the case of foreclosure--than on construction cost, as is often the case now. Such standards would also be needed for the provisioning of nonperforming loans. Prudential regulations should recognize the lower credit risk profile of residential mortgage lending. Specifically, the capital-adequacy risk weighting of residential mortgage loans can be lowered. Similarly, the risk weighting of credit lines granted by banks to specialized institutions also can be lowered. Explicit guidelines regarding the purchase of mortgage-related securities can be defined for institutional investors. 16. There is an opportunity to offer tax relief on mortgage payments that can be an efficient driver for increasing the real estate turnover without necessarily being socially regressive. The afford- ability issue leads one to wonder whether a larger part of the potential demand could be made effective by means of income tax alleviations for mortgage takers. This method has been applied forcefully in India and, more recently, in Pakistan. Its impact on mortgage lending is undeniable, but the question is whether its fiscal social cost and its redistributive effect are not too high a price to pay. 17. However, altering tax provisions may raise issues beyond the narrow approach that motivates advocates to suggest it. If attention were given to these recommendations, a comprehensive assessment of their impact on fiscal and economic points of view should be undertaken. 18. For example, in the case of government-owned leased land, the mortgage lender needs to obtain permission from the government to transfer usage rights to the new homeowner and to record the mortgage. This procedure has no legal rationale; it is justified currently by the need for good mortgage records, given the state of the property registry and its poor functioning. Nevertheless, limiting the process to recording the mortgage, without the delay created by the need for an express ex ante government permission of the property transfer, could help streamline the procedure significantly. References Irving Fisher Committee on Central Bank Statistics. 2009. Measuring Financial Innovation and Its Impact: Proceedings of the IFC Conference, Basel, 26­27 August 2008. IFC Bulletin No. 31. Basel, Switzerland: Bank for International Settlements. Available at http://www.bis.org/ifc/ publ/ifcb31.pdf. Transparency International. 2005. "National/Local Corruption Diagnostics and Measurement Tools in Asia." Berlin, Germany. World Bank. 2004. "Housing Finance Reforms in Bangladesh: Unleashing Supply to Meet Growing Demand." South Asia Private and Financial Sector, Washington, DC. World Bank. 2007. "India: Land Policies for Growth and Poverty Reduction." Agriculture and Rural Development Sector Unit, India Country, Management Unit, South Asia Region, Washington, DC. World Bank. 2009a. Doing Business 2010: Reforming Through Difficult Times. Washington, DC. http://www.doingbusiness.org/. World Bank. 2009b. World Development Indicators 2009. Washington, DC. 125 Appendix C 1 India India, one of the most populous and fastest-growing countries in the world, hosts promising conditions for a vibrant housing market with considerable growth potential. Steady macroeconomic management and a stable financial sector with solid perform- ance have underpinned India's growth. Although there have been continued deepening and broadening of the financial system, and a series of comprehensive financial reforms, housing finance remains underdeveloped at 7.25 percent of GDP, when compared with China (12 percent), Thailand (17 percent), and Malaysia (29 percent). In spite of its small size and confinement to upper-income clients, housing finance grew at an annual average of 30 percent during the period 2003­08. Population growth at 1.7 percent a year, urbanization projected to reach 50 percent in 2030, and rapid formation of a middle class contribute to the colossal growth potential of the housing sector--estimated at $108 billion by 2012 alone. The housing sector further contributes to national growth and employment through its forward and backward links to the economy: it is estimated that every Indian rupee invested in housing adds Re 0.78 to national GDP, and the hous- ing sector contributes more than 9 percent of national employment. According to one estimate, the urban housing backlog in India is 27 million units, and a major part of it is in the low-income category. With congestion and obsolescence included, this count rises to approximately 71.75 million units. Of the total housing stock of 211.86 million units (2006 data), 72 percent are in rural areas and 45 percent are nonpermanent structures. Despite significant growth in housing stock, the slum and squatter settlements are on the increase. There are 52,000 slums, holding 8 million urban households, representing about 14 percent of the total urban population. The real estate market in India has witnessed a sharp increase in prices (30­40 percent annually, prior to 2008), primarily driven by the cost of land (as measured by the housing price index RESIDEX launched in 2007). There is a vibrant developer market, mostly focused on upper- and upper-middle-income clients. 127 Appendix C C The housing finance market in India amounts to about Rs 4 billion, of which 66 per- cent is held by commercial banks and the balance by specialized housing finance compa- nies. The market offers mainly "plain vanilla­type" amortization home loans averaging $30,000, with an average maturity of 13 years, at predominantly floating rates, and with prepayment penalties. Currently, there is minimal speculative property demand. The typical customer belongs to the salaried class, is a first-time home buyer with an average age of 37 years; the average loan-to-value ratio is 65 percent (maximum 85 percent at loan origination). The housing finance market is concentrated, with four financial institutions controlling more than two thirds of the market: HDFC Ltd., Life Insurance Corporation (LIC) Housing Finance Ltd., ICICI Bank, and the State Bank of India. The National Housing Bank (NHB) regulates the specialized housing finance companies (HFCs), and acts as a second-tier lender to all mortgage originators. In contrast to other countries in the region, India has experimented with securitization, although the pene- tration of such deals has remained low because of regulations and liquidity issues. There is no secondary market for these securities as yet, and cross-border issuances are not allowed by the regulator. Developer finance for upper-end construction is easily obtained in India through a rich variety of possible sources, including bank finance, self-finance, and the capital markets. For lower-income housing, however, financing options are not as abundant. The NHB offers a limited scheme for affordable housing construction, but the private sector's involvement is crucial if such housing is to be scaled up adequately. Low affordability could be addressed by vibrant rental markets; but India's regulatory framework has been restrictive, and rental markets have failed to flourish. Slum areas, however, are characterized by popular rental arrangements that exist outside of the reg- ulated sector. Housing finance is not reaching lower-income segments in urban areas, amounting to more than 80 percent of the urban population. At the lower end of the market, there has been strong growth in microfinance, almost all of which has been focused on rural India. Estimates of the housing shortage range from 20 million to 70 million. Currently, local and national governments do not have the resources to provide housing for all these households. As much as half of this excess demand can be profitably serviced by the housing and housing finance markets. Specifically, it may be commercially viable to build housing for 23­28 million households (35­45 percent of urban India) whose monthly household incomes are in the range of Rs 5,000­11,000 ($110­$250). For example, a household earning Rs 6,000 a month might be able to afford--with financ- ing and without subsidies--a unit of 225­250 square feet priced at Rs 900­1,000 per square foot. Business models have been developed and tested, using actual market par- ticipants (financial institutions, developers, low-income households and their employ- ers); and such pilot projects have proved viable, easy to scale up, and very successful. The land administration system is obsolete, with multiple responsible institutions carrying some overlapping functions, inaccuracy of records, and the resulting high delays and transaction costs. Records are manual, recourse is inadequate, and staff lacks capac- ity and efficiency. Registration fees and stamp duties are high, discouraging formal registration. It takes five procedures, 44 days, and 7.43 percent of property value to register real estate, placing India 93rd (among 183 countries) in complexity of property 128 Appendix C C registration (World Bank 2009a). Foreclosure procedures in India have been improved considerably in the last five years. Doing Business 2010 also ranks India 138th world- wide, after China (65th) and Brazil (131st). In the last few years, India has considerably improved the environment for getting credit, bringing it to 30th place in the worldwide ranking by Doing Business 2010. Rapidly scaling up housing finance will require significant effort in promoting an enabling regulatory framework. A more effective land administration system requires improved efficiency of the subregistrar and transfer offices; and possible consideration of a central land office; reduced stamp-duty rates; expanded computerization, integration, and use of textual records to ensure full coverage; and removal of restrictions on land markets. Following recent improvements in the foreclosure regime, the framework implementation could be further strengthened, including improving the functioning of Debt Recovery Tribunals and creating a National Companies Law Tribunal, enhancing unsecured creditor rights, and building institutional capacity in the field. The credit bureau requires further strengthening, and a mortgage repository is needed. A unified record of property and housing data would enable the industry to price its products bet- ter, enhance transparency, and increase market and policy-maker efficiency. The NHB's supervisory and developmental roles could be further strengthened by modernizing its regulations and activities. Mortgage product innovation initiatives could contemplate more efficient funding support to the market through the NHB, an interest risk mitigation fund, a financial guarantee facility, and capacity building for the industry as it innovates its mechanisms and products. Strengthening low-income (microfinance) housing lending involves different market solutions for formally and informally employed households earning more than Rs 5,000 a month, which are sus- tainable without the need for subsidies. Pure market-based solutions will not be enough to bridge the gap and serve this segment; and home ownership would need to be pro- moted by a combination of market-based levers, policy initiatives, and direct support from the government. For the poorest households, rental markets could provide a hous- ing solution (despite low home ownership affordability) if they are properly enabled through regulatory and market means. Any sustainable scaling-up of the mortgage market would require a stable and sizable source of long-term funding, sources from pensions and insurance sectors, mortgage-refinancing, or a secondary market--all of which have potential for dynamic development in the Indian context. Overview India is the home of 1.125 billion people, nearly one sixth of the world's population, representing a galaxy of diversified cultures and languages. In terms of population, it is the largest country in the South Asia region, as well as the home of the world's largest population of very needy and very poor people. One third of India's popula- tion lives below the World Bank­defined poverty line of $2 a day. An estimated 27 percent of the population falls below the government of India's poverty line of 129 Appendix C C $10 a month in urban areas and $8 a month in rural areas.1 The high proportion of low-income households forms the major social challenge for urban planners. In the past few years, India has emerged as one of the world's fastest-growing economies. Since 1990, its economic growth rate has more than doubled, rising from 1.9 percent per capita during the period 1961­90 to 4.6 percent in the period 1991­2008. Growth accelerated to 7.4 percent over the five years preceding the 2008 global slowdown. As a result, India has gone from the world's 50th-ranked economy in nominal U.S. dollar terms in 1980 to the 10th-largest economy today (and the 4th-largest economy in terms of purchasing power parity). The real growth rate reached 9.0 percent in fiscal 2007/08, after peaking at 9.7 percent in the previous year; and real per capita annual income stood at $950, more than double the fiscal 1993/94 level. Stable macroeconomic management, outward orientation, high savings and invest- ment rates, decreasing inflation, and--since 1991--rapid increases in productivity that accompanied economic liberalization account for the good economic performance and more efficient use of its endowments. Inflation has been at 11 per- cent (consumer prices) and around 1 percent (wholesale prices). India has a very strong domestic consumers market. The investment rate increased to 38 percent in fis- cal 2007/08, up from 24 percent in fiscal 2001/02; and domestic savings rose to 36 per- cent of GDP, from 23 percent. On the government side, fiscal consolidation, enhance- ment in revenues, and restructuring of selected expenditures at both federal and state levels increased India's public savings from ­2.0 percent of GDP to 4.5 percent. The major engines of India's recent rapid growth have been manufacturing and services--contributing roughly 29 and 54 percent of the national output, respectively--supported by stable macroeconomic management, a young and increas- ingly well-educated workforce, and a dynamic entrepreneurial class. According to Doing Business 2010 (World Bank 2009a), India ranks 133 out of 183 countries on the friendliness of its investment environment. In terms of construction permits, it stands at 175 of 183; for registering property, it performs in 93rd place out of 183; and for getting credit, it ranks high at 30 of 183. India's rapid growth also reflects its increased integration with the world economy, and it is a widely recognized success in services exports. The emergence of export-oriented sectors, Indian firms' global investments and their increasing access to financing from international financial markets, the growing presence of foreign institutional investments in India, and increasing remittances and income from services exports are some examples of stronger links with the international economy. The share of goods trade increased from 17 percent of GDP in fiscal 1993/94 to 34 percent in fiscal 2007/08. Capital flows also benefited from robust foreign direct investment and increased portfolio investment in real estate, banking, and equities in the financial sector. India has a well-managed balance of payments, with Forex reserves exceeding $250 billion.2 Comprehensive reforms of India's financial systems and institutions and its access to the international financial market have supported this performance. The outcome of India's well-managed financial sector reforms and the rapid growth and diversifi- cation of the economy is a stable banking sector, with high profitability and adequate 130 Appendix C C capital, combined with an ability to expand credit at an average annual rate above 20 percent over the past three years. At the same time, nonperforming assets were reduced sharply to below 2 percent. The global financial crisis resulted in a sharp reversal of capital flows, tighter bank credit, and reduced trade and remittance flows; however, these effects were wearing out by late-2009. The economic growth rate fell sharply to 6.1 percent in 2009 (the estimate for fiscal 2010 is 6.5 percent). Despite a well-developed banking system and significant progress in banking sector reforms, performance, and stability (especially since 1991), improving access to finance still remains a challenge. Although India has a relatively wide bank-branch network, given the spread and size of the country's population, institutional factors within banks and the transactions costs and credit risk perception in dealing with smaller clients have meant that a significant part of the population is underserved by the for- mal banking sector. Policy makers are addressing the need for financial access expan- sion down the income scale to the underserved population. The most notable impact of financial sector reforms is clearly discernible in the development of various segments of financial markets in India. The reforms since the early 1990s have led to a regime characterized by market-determined interest and exchange rates, price-based instruments of monetary policy, current account con- vertibility, substantial capital account liberalization, and vibrant government securi- ties and capital markets. Derivative instruments have been cautiously introduced in a phased manner, both for product diversity and, more important, as a risk manage- ment tool. All these developments have facilitated the process of price discovery in various financial market segments. It is widely recognized that the Indian financial sector over the last decade has been transformed into a reasonably sophisticated, diverse, and resilient system. It delivers a wide variety of financial services efficiently and profitably, with a spectrum of financial market segments in which financial institutions are able to participate with operational and functional autonomy in an environment of increasing deregulation and international competition. The average annual population growth rate in India during the period 1990­2007 was 1.7 percent. The country has witnessed a rapid urbanization with projections reaching 50 percent by 2030. In 1990, 217 million people were living in urban areas; by 2008, the urban population of India had risen to 337 million (30 percent), show- ing an annual urbanization growth rate of 2.5 percent (figure C.1) (latest figures from World Bank [2009c]). The urban population of India is showing a net positive annual growth rate of around 2.73 percent during the period 2005­10. A consider- able middle class is forming, creating demand and potential for housing develop- ment. The population of rural India, which is 72 percent of the total population, is on the decline ( 0.43 percent growth during 2005­10), and is likely to have a net negative growth of 0.40 percent during 2025­30 (India, Office of the Registrar Gen- eral 2001; India, Ministry of Urban Employment 2005). Table C.1 shows the trend from 2000 to 2030. The factors contributing to this urban migration are greater eco- nomic growth, rising income levels, new employment opportunities in the cities, occupational shift from agriculture to manufacturing and services, and changing attitudes toward consumption and lifestyle. 131 Appendix C C Figure C.1 Population and Urbanization Trends in India, Selected Years, 1901­2008 1,600 35 1,400 30 29 29 29 30 28 28 28 28 26 1,200 23 25 population (millions) 1,000 20 20 percent 18 1,095 1,110 1,125 1,140 800 17 1,032 1,049 1,064 1,080 14 15 844 600 12 11 11 10 683 10 400 548 439 200 319 361 5 252 251 279 295 301 308 314 322 329 337 238 288 217 109 159 26 26 28 33 44 62 79 0 0 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2002 2003 2004 2005 2006 2007 2008 year urban population total population share of urban population Source: World Bank 2009c. Table C.1 Growth Trends for the Urban and Rural Populations, 2000­09 Years Urban (%) Rural (%) 2000­04 2.81 0.82 2005­09 2.73 0.43 2010­14 2.70 0.12 2015­19 2.74 0.09 2020­24 2.52 0.22 2025­29 2.25 0.40 Source: World Bank 2009c. By the year 2007, an estimated 11.49 percent of the population lived in urban conglomerations of more than 1 million, and 5.54 percent lived in the largest cities. In terms of social indicators, 89 percent of India's population enjoy access to safe water, and 28 percent have access to improved sanitation (18 percent in rural and 52 percent in urban areas). The contribution of the housing and real estate sector to overall economic growth, social uplift, and employment is considerable in the context of India's fast- expanding economy. It is estimated that out of every Indian rupee invested in housing, Re 0.78 is added to the national GDP. Nevertheless, and in spite of consid- erable growth, the housing sector remains small and confined to the upper-income population groups. The housing finance (mortgage debt) industry measures at 7.25 percent outstanding mortgage debt-to-GDP ratio, which is low in international 132 Appendix C C Figure C.2 Size of the Housing Finance Industry, 2008: Outstanding Mortgages as a Share of GDP 100 90 86 80 80 70 percent of GDP 60 50 39 41 40 29 32 30 26 20 17 12 10 7 1 2 0 p. a a sia na n sia e na m es nd di in ta or Re do at hi hi ay ne In Ch kis la ap ,C ,C ng St a, ai al do ng Pa re ng Th an Ki d M In ite Si Ko Ko iw d ite Un Ta ng Un Ho economy Source: World Bank 2008c. comparisons (figure C.2). The low penetration implies room for growth. An esti- mated investment of $108 billion is required to meet housing needs up to 2012. According to data from the NHB, India has a tremendous potential to grow mort- gage finance and thus to substantially enhance its share in GDP growth. Real estate is one of the fastest-growing sectors in India, with the housing sector growing at an average of 30 percent annually during 2003­08. The importance of the real estate sector as an engine of the nation's growth can be gauged from its considerable size-- $12 billion, or 5 percent of GDP. The real estate industry has significant links with several other sectors of the economy and more than 250 associated industries. In South Asia, 40­50 industries are considered to be directly linked with housing con- struction. A one-unit increase in expenditure in this sector has a multiplier effect and the capacity to generate income as high as five times. If the economy grows at the rate of 10 percent, the housing sector has the capacity to grow at 14 percent and generate 3.2 million new jobs over a decade. It is estimated that the construction sector pro- vides direct employment to 16 percent of India's workforce; and the housing sector alone is the second-largest employment generator after agriculture, accounting for 58 percent of the workforce in the construction sector. Housing Demand Urbanization and population growth have been pushing housing demand upward, especially for lower- and middle-income groups. Slums are numerous. In view of these factors, there is a very strong and growing demand for low-cost housing in India. 133 Appendix C C Housing Shortage There are no official estimates of the housing shortage in India. According to the National Building Organization, the housing shortage comprises 10.6 million units in urban areas and 14.1 million units in rural areas.3 These figures include the esti- mated housing need based on household size, congestion based on people per room, the impact of urbanization, and the depletion of existing housing stock. The NHB, HDFC Bank, and other sources more frequently quote the figure of 27 million units as the urban shortage only. It is further estimated that nearly the entire urban short- age is in economically weaker sections of the population and in low-income groups. According to one estimate, the overall housing shortage in India hovers around 71.75 million units,4 including congestion and obsolescence. The Planning Com- mission of India and the Ministry of Housing and Urban Development have recently set up the Kundu Panel to compile a fresh estimate of the housing shortage, using more realistic parameters.5 The presentation made by McKinsey of India before the Kundu Panel suggests that the urban housing shortage for middle-income groups would be between 8.6 and 13.8 million. The average number of people per room in urban areas is high at 3.5; it is region- ally comparable but worse than average in developed countries (density is 0.5 in the United States and 1.1 in the European Union). About 71 percent of urban house- holds live in overcrowded dwellings. In terms of number of rooms available per household, India's 2001 census revealed that 39 percent of households lived in one room, 30 percent in two rooms, 14 percent in three rooms, and 17 percent in more than three rooms; the median number of rooms was two. However, the number of households living in one room has declined over the years. Most dwelling units (81 percent) in urban areas are durable; and home ownership is at 71 percent for urban households. Using the figure of 5.3 as the household size in India, and a pop- ulation size of 1.25 billion, the country would need about 212 million housing units. Squatters and Informal Settlements The Indian housing market is highly fragmented, with the unorganized sector accounting for more than 70 percent of the housing units constructed. The unor- ganized sector of the housing market is characterized by local small builders and contractors. The organized sector accounting for the balance (30 percent) comprises larger developers as well as government and other parastatal entities involved in housing and construction activities. Various informal and sometimes even illegal methods of land acquisition and construction are used to gain shelter, specifically in slums and squatter settlements. This phenomenon is quite common in the large cities. Slum areas are characterized by poor living conditions, shared service facilities, and lack of (or scarce) access to safe water and sanitation. Rental accommodation has become common in informal settlements as well. The rent is paid to "slumlords" who own a number of hutments, or to individuals who reside elsewhere or share the unit with the tenant. 134 Appendix C C The growing slum population has been a major area of concern for governments at the central, state, and local levels. They have set up agencies and adopted policies for rehabilitation and resettlement of slum areas. Given the extent of the issue, how- ever, such efforts have been vastly inadequate. In rare cases, slum dwellers also have organized themselves into cooperatives and built shelter for themselves through community efforts. In general, slum communities in India are extremely heteroge- neous; characterized by many conflicting interests, they find it hard to self-organize. A market solution seems to be the sole viable and easily scalable method of address- ing the issue. Housing Supply The housing shortages result in a rising pressure on available housing stock, exerting upward pressure on rents, people-per-room density, and urban congestion. The Plan- ning Commission of India estimated the total requirement of housing during the 10th five-year plan (2002­07) at 22.4 million housing units in urban areas--raising the urban housing stock from 8.9 million units in 2002 to 13.5 million by 2007. Current State of Housing It is estimated that, at the beginning of 2006, the total housing stock in the country stood at 211.86 million units (153.03 million units in rural areas and 58.83 million units in urban areas), against s total 222.93 million households (156.63 million households in rural areas and 66.30 million households in urban areas). In 2001, there had been 187.1 million houses and 191.96 million households. The National Sample Survey Organisation (NSSO) uses the classifications of kutcha (primarily of mud/soil), semi-pucca (combination of soil and concrete), and pucca (primarily of concrete) construction. A kutcha house is built with nondurable materials like unburned bricks, mud, thatch, leaves, and either timber or bamboo. A pucca house is one built with permanent materials such as oven-burned bricks, concrete, iron, or other metal sheeting and timber. A semi-pucca house is a hybrid of kutcha and pucca house materials. According to the NSSO, of the total estimated housing stock in 2002, 117.63 million houses were pucca houses, 65.53 million units were semi-pucca houses, and the remaining 28.70 million units were kutcha houses needing replace- ment (table C.2). For housing other than pucca construction, formal financing is unavailable because of the added risks. Despite such a significant growth in housing stock, the slum and squatter set- tlements are on the increase, numbering 52,000 slums holding 8 million urban households--about 14 percent of the total urban population (India, Ministry of Statistics 2004; NHB 2006). "Housing completion" (defined as the absolute increase in housing stock during a particular period) was approximately 5 units per 1,000 people annually in India between 1971 and 2001 (table C.3). During those three decades, the average housing 135 Appendix C C Table C.2 Distribution of Units by Location and Type of Structure, 2002 Semi-pucca Kutcha Area Pucca House House House Rural (%) 38 22 40 Urban, including slum and 74 14 12 squatter areas (%) Source: National Sample Survey Organisation, 2004. Table C.3 Housing Completion, 1971­2001 Category 1971­80 1981­90 1991­2000 Urban Houses added during the period (millions) 11.55 16.55 19.53 Added households (millions) 10.00 11.64 12.95 Annual housing completions per 1,000 7.23 7.61 6.83 population Rural Houses added during the period (millions) 19.25 29.02 34.56 Added households (millions) 15.50 19.16 25.61 Annual housing completions per 1,000 3.66 4.62 4.65 population Total Houses added during the period (millions) 26.53 45.58 54.08 Added households (millions) 25.50 30.80 38.56 Annual housing completions per 1,000 3.87 5.39 5.26 population Sources: India, Office of the Registrar General 2001; NHB 2004. completion per 1,000 people in urban areas was roughly 7 units--lower than the United Nations' recommended minimum annual threshold of 8­10 housing units per 1,000 people for developing countries (NHB 2004). Housing Costs The real estate market in India has witnessed a sharp increase in prices, primarily driven by the cost of land; the increase in the cost of construction materials was in line with the rise of incomes. There has been an unprecedented rise in property prices during the fiscal years 2001/02 to 2006/07 as a result of rising demand, short- age of housing, limited availability of land, and speculative trade in land. The increases in home prices have been particularly noticeable in metropolitan areas-- more so in Mumbai and Delhi where prices increased at a compound annual growth 136 Appendix C C Figure C.3 Residential Price Movements, by City, 2007­08 160 150 140 130 RESIDEX index 120 110 100 90 80 70 60 e ta a ne Lu ur w t Ba elhi Ca i Hy pal Fa ad ad Ch d i i ra ba ch or tn na a ut no ip Pu b ab ab Pa Su al um o Ko en lc Ja D ra ck Bh ng rid ed de M m Ah city 2007 January­June 2008 July­December 2008 Source: NHB RESIDEX, 2009. rate of 30­40 percent. For instance, in South Mumbai, the price per square foot increased from Rs 10,000 in 2004 to Rs 20,000 in 2006. In central Delhi, land price per square foot doubled from Rs 9,000 in 2005 to Rs 18,000 in 2006. Figure C.3 pres- ents price increases in major cities in fiscal 2007/08, according tot RESIDEX. Whereas urban housing prices have risen by 30­40 percent annually, salaries have increased by an average of 20 percent a year over the period 2008­09. According to the Credit Rating Information Service of India, Limited (CRISIL), the average household income in urban areas has grown at a compound 10 percent in nominal terms over the years 1995­2006 (CRISIL 2006). For a typical property in Mumbai, affordability declined from 4.4 times in 2004 to 5.5 times in 2006 (figure C.4). Ana- lyzing over a longer period (from 1995 to 2006), however, the housing affordability has improved and looks rather satisfactory, dropping from 22 in 1995 to 5 in 2006 (Karnad 2008). More recently, property prices have started to decrease in some parts of India, perhaps because of some technical adjustment in speculative rise of the past, and because of the increased supply of housing. Housing Price Index Keeping in view the prominence of housing and real estate as a major area for creat- ing both physical and financial assets, and that area's contribution to overall national wealth, a need was felt to set up a mechanism that could track the movement of prices in the residential housing segment. Regular monitoring of house prices can provide useful input for a variety of housing sector stakeholders. In July 2007, the NHB launched RESIDEX for tracking prices of residential properties.6 RESIDEX 137 Appendix C C Figure C.4 Housing Affordability Index, 1995­2006 30 6 property value (Rs lakhs and affordability) 25 5 annual income (Rs lakhs) 20 4 15 3 10 2 5 1 0 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 year property cost affordability annual income Source: Karnad 2008. Note: Housing affordability is measured by ratio of the price of a residential property to the annual income of the borrower. In the Indian financial nomenclature, 1 lakh is roughly equivalent to 100,000. currently covers 15 cities, but is envisioned to cover all 63 major cities eventually. At present, the index is being developed only for the residential housing sector. How- ever, at a later date, the scope of the index could be expanded to develop separate indexes for commercial property and land; and those indexes could be combined to arrive at the real estate price index. With 2007 as a base, RESIDEX has been updated to December 2008, with two half-yearly updates during 2008. Prices have been stud- ied for various administrative zones/property tax zones constituting each city. The index has been constructed using the weighted average methodology with price rel- ative method (modified Laspeyres approach). Primary data on housing prices are being collected from real estate agents by commissioning the services of private consultancy/research organizations of national repute; and data on housing prices are being collected from housing finance companies and banks, based on housing loans contracted by these institutions. Development of a residential property price index for selected cities is envisioned, with a subsequent all-India composite index produced by suitably combining city- level indexes to capture the relative temporal changes in the prices of housing at different levels. Development of Housing Stock Housing construction is managed mostly by the owners engaging the services of small contractors. In some larger metropolitan areas, builders and developers have 138 Appendix C C played a role in supplying housing by building high-rise, multifamily units and, more recently, large-scale vertical and horizontal residential complexes. The Indian housing sector has witnessed the emergence of some very large builders undertaking mega projects. However, their business focus has been on commercially attractive dwellings for high-income families. Although the high-end market is small in absolute numbers, it is large enough to more than take the supply of housing that builders are capable of producing, while providing high margins and a quick turnover. The private sector developers are currently building housing at rates between Rs 800 and Rs 1,000 per square foot in vibrant neighborhoods within one hour of the city center in most major metropolitan areas. The smallest units they are building are apartments of 450­500 square feet, costing Rs 400,000­500,000 (Monitor Group 2007). Large developers, with a national or wider regional presence, are not interested in serving the low-income housing market as a com- mercial proposition. Some medium-size and small developers, however, see low-income housing as a good commercial opportunity. To serve the market, they need assurance that their low-income customers will receive housing loans. Low-income housing developers also need some help in demand aggregation--at least initially. Some of these medium- size and small builders could build cheaper apartments (perhaps 200­350 square feet) at costs up to Rs 320,000. The Monitor Group (2007) has studied the expan- sion of housing finance to cover the lower-middle-income market in India, and has generated a very comprehensive set of data on income and affordability. The opportunity to serve low-income customers on a commercially viable basis has res- onated with a cross-section of stakeholders interested in engaging in specific pilot projects. The Monitor Group, with the support of the NHB and the World Bank, is active in promoting such projects, starting in cities like Ahmedabad, Hyderabad, Mumbai, and Pune (box C.1).7 These projects are at different stages of negotiation and implementation. Developer Profiles Large, private developers with a pan-India presence--such as Unitech, DLF, Ansal Housing and Construction, and S.P. Constructions--focus on luxury townships and premium residential apartments for the high-income and premium segments. Those developers are leveraging the trend toward premium housing, retail, and informa- tion technology sector expansion into small cities and special economic zone devel- opment. Examples of premium residential townships enclosing apartment buildings or stand-alone premium flats are Unitech's Nirvana Country Gurgaon or DLF's Park Place, consisting of apartments measuring 1,000­2,800 square feet, with two to four bedrooms, hall, and kitchen; and luxury row houses. These developers own large land banks that range between 5,000 and 11,000 acres. Large, private regional developers--such as Anant Raj Industries, Prajay Syndicate, D. S. Kularni Developers, and Peninsula--focus on construction activity 139 Appendix C C Box C.1 Typical Low-End Housing Available in Urban Markets · Area of city: ° Within an hour of the city center ° Close to primary and secondary schools, health care center, and marketplace ° Well connected to city by bus or train · Typical complex: ° Three to five buildings with four floors and four to eight flats per floor ° Regular water and electricity ° No elevator and a single staircase ° Fenced by a compound with wall, shared open spaces, including garden and access to play area for children · Each flat: ° Built-up area of 450­550 square feet ° One bedroom, hall, kitchen, with an attached toilet and bathroom ° Well-painted walls and good interiors ° Monthly maintenance charge of Rs 400­500 · Cost: Rs 450,000­500,000 Mumbai: Ambivili Ahmedabad: Vatva Neptune Group 100 acres Taral Bakeri IBRD 37761 Phase 1: 1,800 units MAY 2010 Phase 1: 800 units Sector 1: 600 flats sold out in three days Construction Start: June 2009 1-BHK and 2-BHK IRs 0.5 million and IRs 0.84 million Price: IRs 0.33­0.56 million Project launched on March 27, 2009 Maharashtra: Karjat Ahmedabad: Kalol TMC-Matheran Realty Sintex Industries Phase 1: 3,000 units June 2009 Phase 1: 300 units 15,000 units by June 2011 6,000 flats at IRs 0.3 million Price: IRs 3­5 lakhs Possession June 2009 Maharashtra: Bolsar Tata Housing 67 acres Phase 1: 1,200 units for LIH Bangalore: Atibela Bangalore: 1-RMK and 1-BHK Janadhar 11 acres Value Budget Housing IRs 0.4 million and 0.7 million Phase 1: 1,500 units IRs 0.4 million and 0.6 million IRs 0.3­0.9 million minimum 10 acre plots Source: Monitor Group 2007. Note: BHK bedroom/hall/kitchen; LIH low-income housing; RMK room/kitchen. 140 Appendix C C in residential and commercial real estate; and have a strong regional presence, mainly in the premium, high-income, and middle-income segments. Those firms typically construct high-quality residential townships or stand-alone high-grade flats. Exam- ples include D. S. Kularni's Madhuban or Prajay Syndicate's Moonrock Apartments, consisting of apartments with two to four bedrooms, hall, and kitchen, measuring more than 800 square feet; and independent row houses. These developers also pos- sess large land banks concentrated in regional pockets. Medium-size developers--such as Bakeri Engineering and Infrastructure, Janapriya Engineers Syndicate, Naiknavare and Associates--focus most of their con- struction activities in one or two cities. (For example, Janapriya Engineers Syndicate concentrates its efforts in Hyderabad and Bangalore, limiting itself to the high- and middle-income segments.) The constructions are standard apartment complexes, such as Naiknavare's Trinity Court in Pune or Janapriya's Vilas in Hyderabad. They primarily comprise apartments of 700­1,600 square feet, featuring two or three bed- rooms, hall, and kitchen. These developers possess medium-size land banks, gener- ally within one city. Small, private developers--such as Libra Shelters in Hyderabad--have a single- city presence and focus on the middle-income segment. They construct standard apartments of 600­1,200 square feet, with one or two bedrooms, hall, and kitchen. The land banks commanded by these developers are of negligible size. Housing Finance The housing finance market in India is growing fast, but remains concentrated among a few dominant institutions. It is roughly two thirds in the hands of com- mercial banks and a third accounted for by specialized housing finance companies. The market offers mainly plain-vanilla amortization home loans of moderate amounts and with relatively short maturities, at predominantly floating rates and with prepayment penalties. That being said, the housing finance market in India is the most developed in the region. The NHB, a second-tier institution, provides mortgage refinance. In contrast to other countries in the region, India has experi- mented with securitization, though the penetration of such deals has remained low because of regulations and liquidity issues. Developer finance for upper-end construction is easily obtained in India. For lower- and middle-income (especially informally employed) groups, however, financing options are not as abundant. There has been some growth in microfinance, almost all of it focused on rural India. However, housing finance is not reaching lower-income segments in urban areas-- more than 80 percent of the urban population. Such customers can be served prof- itably and sustainably by the market. Specifically, business models have been devel- oped and tested, using actual market participants (that is, financial institutions, developers, low-income households, and their employers), and such pilot projects have proved viable, easy to scale up, and very successful. 141 Appendix C C Table C.4 Banking and Nonbanking Activity, 2001 and 2008 2001 2008 Total Assets Deposits Total Assets Deposits Institution Number (Rs billions) (Rs billions) Number (Rs billions) (Rs billions) Banks Commercial banks 97 12,954 10,533 79 43,264 33,200 Public 27 10,299 8,594 28 30,222 24,538 Private 30 1,633 1,366 23 9,401 6,750 Foreign 40 1,021 592 28 364,099 1,911 Nonbank financial companies Reporting HBFCs 981 538 180 364 947 243 Source: RBI 2009. The Financial Sector The Indian financial sector has steadily increased its spread and outreach. Since 1991, the government has reformed India's banking and nonbanking segments sig- nificantly, as part of its new market-based approach to development. Interest rates were liberalized progressively and banks' required holdings of public debt were reduced. Capital markets were deregulated and opened to international flows. Pri- vate entry into banking and insurance increased the contestability of the domestic financial market. Important public sector financial institutions, such as ICICI Bank, were made private, putting competitive pressure on the system. The public sector remains significant, at 70 percent of the banking system, 74 percent of aggregate deposits, and 85 percent of the total bank branches spread across urban and rural India. The banking sector vastly dominates the financial system, with nonbanking assets amounting to barely 0.2 percent of banking assets (table C.4). Over the past decade, financial depth has increased substantially, and regulatory oversight and risk management have been strengthened. Outcomes on nonperform- ing assets and capital-to-risk-weighted-assets have improved significantly and com- pare favorably with international benchmarks. Banks have experienced strong growth in their balance sheets, improved financial health, increased competitiveness, and productivity gains through adoption of technology, wider outreach, and larger allocations of credit to important productive sectors. Over time, India's capital mar- kets have deepened and matured to become the largest markets in developing coun- tries, relative to GDP (155 percent in 2007, 53 percent in 2008 after the crisis). The size of the government bond market is also significant at about 42 percent of GDP. India's ratio of financial assets to GDP (approximately 170 percent) is higher than the ratio in middle-income countries, such as Brazil, Indonesia, and Mexico. The ratio of domestic bank credit to GDP has been growing steadily, from 29.0 per- cent in 2000 to 54.5 percent in 2008. This growth was inevitable with a growing middle class having high disposable income, wider choice of consumer durables, 142 Appendix C C increased acceptance of credit cards, and increased demand for housing loans. Nonetheless, the share of banking assets to GDP is low when compared with middle- income countries and East Asia: in 2008, the figure was higher for Brazil (58.4 per- cent), China (125.3 percent), and the Republic of Korea (118.3 percent) (Fitch Ratings 2009). The Reserve Bank of India (RBI; the central bank) and the govern- ment recognize the need to improve levels of access to finance by underserved segments. Based on major policy documents underpinning recent reforms, the government is implementing policies that aim to provide stability and effective tools for managing liquidity, savings, and investment; and to promote financial inclusion. With the establishment of the Housing and Urban Development Corporation (HUDCO) in 1971, during the fourth five-year planning period, India saw the begin- ning of formal and institutional housing finance. HUDCO primarily catered to the low-income housing segment, and provided technical and financial assistance to the state housing boards, urban development institutions, and the cooperative sector (Garg 1998). This process paved the way for institutionalized housing finance and a shift in the role of government from that of finance provider to facilitator. The Hous- ing Development Finance Corporation (HDFC) gained prominence in 1994. The NHB was established in 1988 as an apex-level institution for housing finance (National Housing Bank Bill 91 of 1987, National Housing Policy 1988), and is wholly owned by the RBI. The housing finance market in India is serviced by a variety of financial institu- tions, including scheduled commercial banks (SCBs), scheduled cooperative banks, regional rural banks, HFCs, and state-level apex cooperative housing finance soci- eties. The housing finance market is concentrated, with four financial institutions controlling more than two thirds of the market: HDFC Ltd., LIC Housing Finance Ltd., ICICI, and the State Bank of India. Regulatory oversight of banks rests with the RBI and, for housing finance companies, with the NHB. As a whole, there are Rs 3,970 billion in housing loans outstanding. As of 2009, SCBs held Rs 2,632 billion and HFCs held Rs 1,200 billion. The market offers mainly plain-vanilla amortization home loans averaging $30,000, with an average maturity of 13 years, predominantly floating rates, and prepayment penalties. There is minimal speculative property demand. The typical customer belongs to the salaried class, is a first-time home buyer of an average age of 37 years, and has an average loan-to-value ratio of 65 percent (maximum 85 percent at origination). Traditional Mortgage Lending There are 97 SCBs (27 public, 30 private, and 40 foreign banks), with a total asset fig- ure of Rs 447 trillion. To encourage housing finance, the RBI initially directed the commercial banks to enter into housing finance by including lending for housing under a directed credit regime. The banks were required to lend, as concessional finance for housing, 1.5 percent of their incremental deposits over the previous year. In 1998, the RBI removed housing from the concessional finance regime, but 143 Appendix C C Figure C.5 Retail Portfolio of Banks, 2008 Percent other personal loans 35 housing auto loans 44 15 credit card consumer durables receivables 1 5 Source: NHB 2009. increased the allocation from 1.5 percent to 3.0 percent. As a result, SCBs now account for the major share of the housing finance market. Their outstanding housing loans grew to 11.0 percent of the gross bank credit in 2008 (9.4 percent in 2009) from 4.0 percent in 2000. SCB retail loans constitute about 25 percent of total loans and advances, and mortgages constitute 44 percent of the retail loans portfolio. Lending for housing by commercial banks as a percent of gross credit outstanding was up from about 4 percent in 2001 to about 11 percent in 2008. Figure C.5 shows the retail port- folio of banks, which represents about 25 percent of overall bank lending. After SCBs entered the housing finance market in the 1990s, they rapidly expanded their housing loan portfolios and overall market share. In 2000, the hous- ing finance companies accounted for 69 percent of yearly disbursements. By 2009, their share decreased to 38 percent, with SCBs accounting for more than 60 percent of all housing finance disbursements (figure C.6). The main funding sources of banks are savings and current accounts, loan assignments through interbank partic- ipation certificates, and refinance from the NHB. The SCBs had 4.52 million housing loan accounts in 2006 (up from 2.53 million in 2000), with average credit per account of Rs 463,000 (up from Rs 93,000 in 2000). Total outstanding housing loans increased from Rs 185 billion to Rs 2,632 billion between 2000 and 2009. Bank credit growth in the Indian economy was 17.3 percent in fiscal 2008/09, and banks were sound and profitable. As of March 2009, the 27 public scheduled banks reported capital adequacy ratios averaging 12.3 percent, a net nonperforming assets ratio of just 0.9 percent, and a 0.9 percent return on assets. The equivalent ratios for private banks were 15.1 percent, 1.4 percent, and 1.1 percent, respectively. Nonbank 144 Appendix C C Figure C.6 Housing Finance Disbursements, by Type of Institution 3,000 2,500 1,450 2,000 1,288 Rs billions 1,101 1,500 860 764 550 1,000 470 537 401 414 274 260 500 209 818 900 232 178 586 700 182 504 126 56 146 86 236 328 0 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09a fiscal year total HFCs banks Source: NHB 2009. a. Estimated. financial institutions grew as well--with total asset growth of 9 percent annually between 2001 and 2008. India's housing sector seized the opportunities provided by brisk growth and expanded rapidly. Even as various segments of the domestic finan- cial market develop, the potential long-term demand prospects for credit are high. For housing finance specifically, given the efficiency and effectiveness of the Foreclo- sure Law, the quality of collateral in India is among the best in the region. The ratio of nonperforming loans to total gross loans is 0.9 percent for public SCBs and 1.1 percent for private SCBs (2009 data from the RBI). India's banks are well capitalized, their asset quality is good, and they have recorded impressive levels of profitability (tables C.5 and C.6). Growth in 2009 was relatively strong, with credit expanding by more than 17 percent and deposits grow- ing approximately 20 percent. Without exception, all banks are above the RBI's min- imum capital adequacy level of 9 percent. As of March 31, 2009, the ratio of capital Table C.5 Indian Banks' Financial Soundness Indicators, 2000 and 2009 Performance Indicator 2000 2009 Capital-to-risk-weighted-assets ratio 11.1 13.2 Gross nonperforming assets to gross advances 13.1 2.3 Net nonperforming assets to net advances 7.1 1.1 Return on total assets 0.7 1.0 Return on equity 12.7 13.2 Efficiency (cost income) ratio 61.2 45.4 Sources: Government of India; RBI 2009. 145 C 146 Table C.6 Key Performance Indicators, by Bank Group, Selected Years March 2009 (Rs billions) March 2001 (Rs billions) March 2008 (Rs billions) Private Performance Private Private New Indicator Public New Old Foreign Public New Old Foreign Public and Old Foreign Deposits 8,594 629 737 592 23,567 5,010 1,656 1,912 30,033 7,239 2,154 (% of total) (81) (6) (6) (6) (73) (16) (5) (6) (76) (18) (5) Advances 4,146 301 379 429 16,963 3,593 1,130 1,629 21,018 5,203 1,697 (% of total) (79) (6) (7) (8) (73) (15) (5) (7) (75) (19) (6) CRAR (%) 11.2 11.5 11.9 12.6 12.5 14.4 14.1 13.1 12.3 15.1 14.9 ROA (%) 0.4 0.8 0.6 0.9 0.9 1.0 1.0 1.8 0.9 1.1 1.8 Gross NPA (%) 12.4 5.1 11.1 6.8 2.2 2.9 2.3 1.9 2.0 3.2 3.8 Net NPA (%) 6.7 3.1 7.3 1.9 1.0 1.4 0.7 0.8 0.9 1.4 1.7 Sources: Ministry of Finance; RBI 2009. Note: CRAR capital-to-risk-weighted-assets ratio; NPA nonperforming assets; ROA return on assets. Appendix C Appendix C C to risk-weighted assets (CRAR, or capital adequacy ratio) for the banking sector stood at 13.2 percent, and no banks had a CRAR of less than 10 percent. The Indian banking sector is well managed and efficient; and the cost-to-income ratio has fallen by 15.8 percentage points between 2000 and 2009, rendering the sector increasingly profitable. Indian banks have largely stable sources of funds, with deposits account- ing for 78 percent of total resources. Specialized Mortgage Finance The origin of HFCs is in the 1980s, when a few housing finance companies were set up either as private limited companies (for example, Dewan Housing Finance Corporation Ltd.), as joint ventures partnering with the state government (such as Gujarat Rural Housing Finance Corporation), or as bank-sponsored housing finance companies (Can Fin Homes, SBI Home Finance, and PNB Housing Finance). Other HFCs were set up by state-owned insurance companies: Life Insurance Corporation sponsored LIC Housing Finance Ltd., and General Insurance Corporation of India sponsored GIC Housing Finance Ltd. Until 1988, HFCs were regulated by India's central bank before the NHB was created. The 1980s saw a rapid growth of HFCs in the private sector as well. As of June 30, 2008, there were 43 HFCs registered with the NHB (table C.7). Twenty of those are permitted to accept deposits and therefore take advantage of cheaper funding. Six HFCs are public, and another three (HUDCO, Orissa Rural Housing and Development Corporation, and Rajiv Gandhi Rural Housing Corporation) are government sponsored. Three HFCs remain owned by financial institutions, which gives them potential access to the cheaper deposit funding of banks (GIC Housing Finance Ltd., LIC Housing Finance Ltd., and IDBI Home Finance Ltd.). The funding advantage is also shared by another eight HFCs owned by commercial banks, including ICIC Home Finance Ltd.--one of the largest HFCs. Table C.7 Key Statistics on the 10 Largest HFCs, 2009 Rs millions HFC Loans Portfolio Shareholders' Equity Profit after Tax HDFC 360.1 38.8 10.36 LIC Housing Finance Ltd. 122.4 11.9 1.44 Dewan Housing Finance 15.3 1.9 0.27 Corporation Ltd. GIC Housing Finance Ltd. 15.5 1.2 0.18 Can Fin Homes Ltd. 13.1 1.5 0.21 Sundaram Home Finance Ltd. 8.7 0.9 0.04 Gruh Finance Ltd. 8.2 0.7 0.17 Sources: Ministry of Finance; RBI 2009. 147 Appendix C C Figure C.7 Performance Statistics of HFCs, 2004­08 7 6.2 6 percent 5 4.4 4 3.6 3 2.2 2 2.0 1.6 1 0 March 2004 March 2005 March 2006 March 2007 March 2008 year gross NPLs/gross loans net NPLs/net loans Source: Financial Sector Self-Assessment, 2009. Note: NPL nonperforming loan. The main funding sources for HFCs are commercial banks, bonds and debentures, refinance from the NHB, term deposits,8 loan assignment, and mortgage-backed securitization. The aggregate outstanding housing loans of HFCs stood at Rs 1,220 billion in 2009. Loans with maturities longer than seven years made up 63 percent of the out- standing portfolio, whereas 23 percent of the loans had maturities of one to five years (2006 data). Size-wise, disbursements of housing loans by HFCs as of March 31, 2006, showed that 41 percent of the loan portfolio was in loans of Rs 1.0 million or more, and 37 percent was in the range of Rs 0.3 million to 1.0 million. HFCs have been able to maintain a CRAR well above the prescribed level of 12 percent. Portfolio risk for HFCs is moderate: the ratio of nonperforming loans to total gross loans is 2.2 percent (figure C.7). The HDFC is, by far, the largest HFC in India (with a yearly housing finance port- folio of more than Rs 200 billion), enjoying a widespread geographic presence, a very nominal share of nonperforming loans, the right to accept deposits, and a position in the financial market permitting it to float its own bonds and mortgage-backed securities. It started operations as a housing bank in 1977, with public shareholding of 90 percent; and it was privatized in 1994. In 2009, it was transformed into a finan- cial conglomerate with interests beyond mortgages, owned 75 percent by foreign investors. Its asset size is more than Rs 850 billion (table C.8). National Housing Bank The NHB primarily performs the functions of regulation, refinance, and promotion. In terms of the National Housing Bank Act 1987, the NHB regulates the housing 148 Appendix C C Table C.8 Financial Highlights of the HDFC Performance Indicator Data Mortgage loan assets $16.9 billion Cumulative loan disbursements $38 billion Cumulative housing units financed 3.3 million Cost-to-income ratio 8.8% (among the lowest in financial services in Asia) Nonperforming loans Less than 1% (total loan write-offs since inception are only 4 basis points of cumulative disbursements) Source: D'Souza 2009. finance system of the country, carrying the power to determine policy and give direc- tions to housing finance institutions and their auditors. One of the more important regulations of the NHB is the Housing Finance Companies Directions 2001 that guides HFC asset-liability management. These directions are updated periodically through the issuance of circulars and notifications. As part of the supervisory process, an entry-level regulation is sought to be achieved through a system of regis- tration for housing finance companies. The NHB supervises the sector through a system of on-site and off-site surveillance. The NHB supports the housing finance sector by extending funding to different primary lenders, refinancing eligible housing loans to individual beneficiaries, and making project loans to various implementing agencies (table C.9). All housing finance companies registered with the NHB under section 29A of the National Housing Bank Act 1987 and scheduled commercial/cooperative banks are eligible for refinance support, subject to terms and conditions laid down under the respective refinance programs. The NHB also lends directly to projects undertaken by public housing agencies for housing construction and development of housing-related infrastructure. The apex institution further guarantees the principal and interest on bonds issued by HFCs. Finally, it can act as a special-purpose vehicle for securitizing housing loan receivables. Table C.9 NHB Refinance Disbursements, Fiscal 2007/08 Golden Jubilee Rural Housing Refinance Institution Category Regular Scheme Total HFCs 8.36 3.53 11.89 Urban cooperative banks 0.66 0.04 0.70 Banks 38.29 34.99 73.28 Total 47.31 38.56 85.87 Source: NHB 2009. 149 Appendix C C The NHB raises short-term and long-term funds from banks and other financial institutions (such as Life Insurance Corporation of India). It issues commercial papers as well as mortgage-backed securities. During fiscal 2007/08, it had raised Rs 121 bil- lion from the market and repaid Rs 58 billion, with a net inflow of Rs 63 billion dur- ing the year. The NHB had issued 14 mortgage-backed security instruments, covered by 38,809 mortgages of six HFCs and one commercial bank, raising Rs 8.6 billion. This instrument is in its trial-and-testing phase to discover its regulatory, fiscal, and investor aspects. The NHB promotes housing finance institutions to improve and strengthen the credit delivery network for housing finance in the country. For this purpose, the bank has issued the Model Memorandum and Articles of Association, as well as guidelines for participating in the equity of housing finance companies. As a part of its promotional role, the bank also has formulated a program for guaranteeing HFC bonds. To address the need for trained personnel in the sector, the NHB has designed and conducted various training programs as well. Securitization There have been more than 500 issuances amounting to Rs 900 billion as of mid- 2009, with the first asset-backed securitization in India carried out in 1992 by Citibank, and the first mortgage-backed securitization carried out in 2000 (origina- tors were HDFC and LIC Housing Finance). Whereas asset-backed securitization has grown on the back of a retail boom in India, the penetration of mortgage-backed securities remained low because direct assignment is still preferred as a result of reg- ulations and liquidity issues. There is no secondary market for these securities yet, and cross-border issuances are not allowed by the regulator. The residential mortgage-backed securities issues were based on conforming mortgages with standardized pool selection criteria. The requirements included a seasoning rule and consistent performance, high collection efficiencies, consistent payments to investors, servicer audit, special-purpose vehicle audit, and surveillance by rating agencies. Figure C.8 illustrates the securitization mechanism. These 14 deals were predominantly direct assignment deals, asset-backed security transactions with simple structures, single tranche, and credit enhancement with primarily cash collateral. The deals amounted to Rs 8.62 billion, involving seven lending institutions. Nonetheless, mortgage-backed security deals are not a major source of liquidity (table C.10). Large-Scale Project Finance Developer finance for upper-end construction is easily obtained through a rich vari- ety of possible sources, including bank finance, self-finance, and the capital markets. For lower-income housing, however, financing options are not as abundant. Obtain- ing cheap financing for low-income developments is a big hurdle faced by develop- ers. The alternative to developer finance is to find working capital for a project, with 150 Appendix C C Figure C.8 Mechanism for Issuing Residential Mortgage-Backed Securities Escrow Guarantee a/c Cash Monthly pool payouts EMI loan Servicing and Borrowers Originator H paying agent F Mortgages Investors Custodian C Banks Unrated pool Custody Servicing and Rating FIs agreement paying agency agency agreement Insurance Rated MBS Declaration companies of trust Guarantee Mutual funds SPV N Trustee H Registrar and Payment of B PTCs transfer agent consideration by SPV Issue Pass-through certificates arrangers Source: NHB. Note: EMI equated monthly installment; FI financial institution; MBS mortgage-backed security; PTC pass-through certificate; SPV special-purpose vehicle. Table C.10 Liquidity through Securitization Rs billions Security FY2005 FY2006 FY2007 FY2008 FY2009 Asset-backed 222.9 178.5 234.2 31.32 135.7 securities Mortgage-backed 33.4 50.1 16.1 5.9 32.9 securities Corporate debt 25.8 21.0 119.0 318.2 351.2 obligations/loan sell-off Total 282.1 249.6 369.3 637.3 519.8 Source: ICRA. advance payments made by customers. However, this approach not only may put the purchasers' money at risk; it also constitutes a barrier to access for low-income buyers because it implies the need to build large downpayments and the ability to continue paying rent while making advance payments. The shortage of developer finance was exacerbated by the global financial crisis. Land price bubbles are now 151 Appendix C C correcting themselves, and many players in the real estate market have lost money. This is an outcome of speculative land banking (of overvalued land) and of financial projections based on real estate that suddenly has stopped appreciating. Banks, as a result, are freezing their lending activity, and developers are holding off their invest- ments in housing until the market stabilizes. The negative impact of this chain of events has been felt largely by the majority of real estate developers catering to upper-income groups. Even those building middle-income housing (units priced between Rs 10 lakhs and Rs 25 lakhs, or between $20,000 and $50,000)9 have put their plans on hold until the market stabilizes. However, at the low end of the mar- ket (houses costing less than Rs 10 lakhs, or $20,000), building activity already was minimal and hence has not led to much speculation. On the contrary, developers venturing or planning to venture into this market appear more resilient: they are not pulling out because they realize the huge profit-making potential of the low-income housing market, as a result of the sheer volume of demand. (They also understand the relative stability in the pricing structure and the market appreciation, and they see this as a possible strategy for diversifying risk.) The bank provides some refinance for construction of affordable housing. The scheme draws on the National Urban Housing and Habitat Policy 2007 (India, Min- istry of Housing 2007), which has emphasized measures to respond to housing needs in urban areas and the growing pressure on housing and related infrastructure facil- ities. The policy seeks to promote various types of public-private partnerships for realizing the goal of "affordable housing for all," with special emphasis on urban poor people. The goal is sustainable development of habitat in the country, with a view to ensuring equitable supply of land, shelter, and services at affordable prices to all sections of society, ultimately shifting to a demand-driven approach through proactive financial sector interventions. To encourage and support the banking sector's involvement in this endeavor, and as part of its developmental role, the NHB has focused on those population groups who may afford a moderately priced dwelling unit by way of refinance support. The bank will refinance lending for new construction, basic housing infrastructure, upgrades for economically weak/low-income housing, slum redevelopment, indus- trial worker housing, hostels for working women, old-age homes, housing projects financed under the Jawaharlal Nehru National Urban Renewal Mission, rental hous- ing programs for migrant laborers, community bulk loans (with tenancy rights ensured by the municipal authorities), and any other similar central government/ state government program that includes housing affected by natural disaster. SCBs, regional rural banks, HFCs, and microfinance institutions (MFIs) are eligible under this scheme.10 The end borrowers envisaged under the scheme are public housing agencies; private developers; public-private partnerships; cooperative societies; and employee-housing organizations, such as the Army Welfare Housing Organization, the Central Government Employees Welfare Housing Organization, the Indian Railway Welfare Organization, and the Air Force Naval Housing Board. Eligible areas are urban and rural centers, including metropolitan cities (such as Bangalore, Calcutta, Chennai, Delhi, Hyderabad, and Mumbai). Focus will be on Tier II and 152 Appendix C C Tier III cities and slum redevelopment projects in metropolitan areas. Funding maturity is coterminous with the original loan, up to a maximum of five years. Inter- est is fixed or floating (both options are available), linked to market rates. Conver- sions from fixed rate to floating rate and vice versa are permissible on payment of a conversion fee. Lower rates are available to public agencies, nongovernmental organ- izations, and community-based organizations. Informal Housing Finance Much of the funds for house construction come from informal sources that include cash savings by households; loans and gifts from relatives, moneylenders, shopkeep- ers, landlords; and so forth. This sourcing can be attributed to the inadequacies of the existing formal financing system. In the urban areas, MFIs currently provide small, short-maturity loans (Rs 5,000­25,000 loans for six months to two years) for livelihood generation and consumption; loans for housing are restricted to small amounts for incremental construction or home repair. Their current customer base comprises households with monthly incomes between Rs 3,000 and Rs 10,000. The ownership pattern of the assets and the nature of borrowing, particularly for the lower-asset and lower-income categories, make it clear that bringing much of hous- ing finance into the formal system would require a massive effort at bridging the present gap between the formal and informal institutions. As an alternative to housing purchase, low-income households resort to rental housing. Rental markets are very strongly developed in India, where 29 percent of urban low-income groups and 4 percent of rural ones are housed in rented units (2001 India census data). Formal rental markets in India are characterized by rent controls; harsh, tenant-favoring laws; and a restrictive regulatory environment. A person can rent a house from a private individual or agency at a predetermined rent, with or without making a fixed deposit. Evicting a tenant is extremely difficult because of various legal and administrative stipulations. As a result, the deposit money has risen to huge amounts in large cities--sometimes amounts equal to the market price of the house. Owners unwilling to take the risk of renting out their houses are extremely selective about their tenants; and they insist that tenants per- sonally or through their employer provide a legally enforceable commitment to vacate the premises after a stipulated period or on demand. All these conditions have hindered the smooth functioning of the rental market and have increased the num- ber of vacant houses in large cities. People working in government and semigovern- ment organizations and undertakings and a few large private companies are entitled to a certain percentage of their basic salary as a rent allowance. The amount in large cities is much less than the actual rent paid by the households for reasonable accom- modation rented on the open market. The informal rental market in slum areas has been flourishing as a viable alter- native for lower-income groups. In the informal sector, some slumlords organize squatting or sell hutments to the urban poor, with some amount of administrative protection at the local level. The rent is paid to the slumlords who own a number of 153 Appendix C C hutments, or to individuals who reside elsewhere or share the unit with the tenant. Frequently, those units, in turn, have been owned informally by squatting landlords for decades. Housing construction and upgrades are funded through savings or money given by family and friends. The current scale of activities for most MFIs in the urban areas is small, but growing. Some new MFIs are focused on the urban poor (for example, Swadhaar FinServe Pvt. Ltd. and Ujjivan) and are keen on quickly growing their presence in the urban market. Other traditionally rural MFIs also are entering the urban market (for example, Basix). Overall, urban microfinance is expected to grow rapidly in the near and medium terms. Homeowners' Insurance Life insurance and property insurance are confined to the upper-income population groups, covering a very small portion of India's homeowners. Mortgage insurance is not available in the market. The NHB and other key stakeholders recognize the potential for this product to really help scale up the market, and they are working on bringing it to market. In the meantime, a credit guarantee may be a good interme- diate step to provide risk mitigation for financial institutions and to collect data that can be used to price the mortgage insurance. Risk Management Risk management is sound for banks, but not for HFCs because the NHB does not match the strictness of RBI risk management guidelines. The successful implemen- tation of Basel II is likely to lead to an even sharper focus on risk management at the institutional level. SCBs in India began implementing Basel II in March 2008. In view of the fast pace of deregulation, liberalization, and the emergence of financial con- glomerates, the supervisory process is being fine-tuned constantly to ensure that adequate consideration is given to the complexities of organizational structures, business processes, and risk positions of the banks. The focus is on a risk-based supervisory framework and smooth migration to Basel II that will require appropri- ate capacity building at the RBI and in the banks. At the end of March 2009, all 81 SCBs operating in India maintained a CRAR at or above 9 percent, relative to the Basel I norm of 8 percent (table C.11). For the Indian banking system, the CRAR currently works out to 13 percent. Detailed guide- lines relating to Pillar 2 also were issued. In fiscal 2007/08, the RBI reviewed the exposure norms and risk weights for a few classes of loans extended by banks. In spite of active migration to Basel II, weaknesses in risk management remain (RBI 2009). Several initiatives were taken by the RBI to further strengthen risk man- agement systems in banks: modifications in the asset-liability management guide- lines; adoption of guidelines on managing risks in outsourced financial services; and strengthening of the guidelines for exposure norms for commercial bank credit to individuals/groups, industry, and capital markets. Banks also were advised to avoid equity funding of projects. Comprehensive guidelines were issued to banks to do a 154 Appendix C C Table C.11 Distribution of Commercial Banks According to Risk-Weighted Adequacy, Selected Years Year (end of Capital Adequacy March) Below 4% 4­9%a 9­10%b Above 10% Total 1996 8 9 33 42 92 2001 3 2 11 84 100 2007 -- -- 2 79 81 c 2009 -- -- -- 81 81 Source: RBI 2009. Note: -- not available. a. Relates to 4­8 percent before fiscal 1999/2000. b. Relates to 8­10 percent before fiscal 1999/2000. c. Per Basel II norms. proper risk assessment and accounting for the letters of credit issued by them. Guidelines on stress testing for commercial banks, know-your-customer norms, and anti-money-laundering initiatives also were issued. The NHB has a risk management system in place for its own use. It consists of committees for managing asset-liability risk, credit risk, and operational risk. Its board has also appointed a risk management advisory committee, with three exter- nal members who are experts in banking and finance, to review the NHB's risk management system. Through its regulatory and surveillance mechanism, the NHB further endeavors to ensure that HFCs also have similar risk management systems. Housing finance is an entirely different asset class than consumer finance, and requires a different set of prudential regulations. It involves surveillance and moni- toring of loans over a much longer period. Maintaining a healthy loan portfolio in housing is rather a challenge from the perspective of credit risk. It is difficult prima- rily because these loans are of much longer tenor--between 15 and 25 years, with an average retention period of 12­14 years. It is difficult to assess the income of the bor- rower in the first place, and equally or more challenging to ensure continuity and reliability of income over the period of the loan. Interest rate risk is managed by offering mostly adjustable-rate mortgages. Those mortgages lower affordability and could adversely affect credit risk when there is a drop in property prices. Banks, and deposit-collecting HFCs have better means to mitigate term-mismatch risk. The NHB refinancing function also serves in this regard. Prudential Regulation and Oversight The RBI is the banking sector regulator. For the small, nonbank financial institutions in specific sectors, the RBI has delegated oversight power. (HFCs are regulated by the NHB.) Banking legislation is extensive. The Banking Regulation Act of 1949 governs overall regulation of the banking sector. Other key pieces of legislation include the 155 Appendix C C Banking Companies (Acquisition and Transfer of Undertakings) Act of 1980 and the Legal Services Authority Act of 1987. In the past decade, the RBI carried out a phased introduction of prudential norms for capital adequacy, income recognition, asset classification, and provision- ing to bring Indian banks in line with international practices; and strengthened the corporate governance of banks. A credit bureau was established, and significant changes were made to improve the execution of collateral. The RBI was strengthened substantially, in both its monetary policy functions and its regulation of the banking system; and is widely regarded as a credible, reputable, and expert institution. Higher levels and standards of disclosure were achieved to enhance market transparency. Microprudential measures were instituted, risk-based supervision was introduced, and the supervision process was streamlined with a combination of on-site and off- site surveillance and external auditing. The ownership of public sector banks has been broadened through disinvestment up to 49 percent, and banks have been listed. A mechanism for greater regulatory coordination was instituted for regulation and supervision of financial conglomerates. Finally, measures were taken to strengthen creditor rights. Specifically, a number of steps have been taken to improve banking: introduction of core banking solutions, installation of automated teller machines, enhanced work- ing hours, simplification of procedures and forms, the adoption of guidelines by the Indian Banks' Association on a Fair Practices Code, the introduction of RBI guide- lines on credit card operations, and so forth. The Enforcement of Security Interest and Recovery of Debts Laws (Amendment) Act 2004 and the Securitization and Reconstruction of Financial Assets and Enforcement of Security Interest Act 2002 have been enacted to facilitate quick and effective recovery of nonperforming assets. The Credit Information Companies (Regulation) Act 2005 has been enacted to facil- itate the creation of credit information companies for collection, sharing, and dis- semination of credit information, which will help in arresting the fresh accretion of nonperforming assets and will significantly improve the quality of credit appraisals and decisions. To further accelerate the process of reforms in the banking sector, the government has amended the RBI Act of 1934. Amendments to the Banking Regu- lation Act of 1949 also have been proposed. HFCs are subject to guidelines regarding, among other things, minimum capi- tal, asset composition, composition of boards of directors, and appointment of auditors. Those HFCs permitted to take deposits are limited to deposits of one to seven years; the deposits cannot exceed five times the net owned funds and they require a credit rating; and a statutory liquidity ratio must be maintained. Furthermore, these institutions need to abide by the know-your-customer norms. Prudential norms for HFCs include asset classification, provisioning requirements on nonperforming loans and standard nonhousing assets, concentration of credit, and a capital adequacy ratio of 12 percent (with a granular approach with higher risk weight for higher loan-to-value ratios). The regulatory structure of housing finance in India is explained in figure C.9. The fact remains that there are some inherent differences, especially in the liability 156 Appendix C C Figure C.9 Housing Finance System Group Subgroup Channel NABARD Development financial institutions SIDBI Financial institutions HFCs Nonbanking finance companies Other NBFCs Private banks Scheduled commercial banks Public banks Public Banks Urban cooperative banks Cooperative banks District cooperative banks State cooperative banks Agriculture and rural development banks Primary land development Other banks institutions Apex cooperative Housing societies housing societies Source: Authors' illustration. Note: NABARD National Bank for Agriculture and Rural Development; SIDBI Small Industries Development Bank of India. structure of commercial banks and HFCs, that generate certain fundamental differ- ences in their regulatory requirements. The banks and HFCs differ in liability struc- ture, tax treatment, capital adequacy, liquidity requirements, deposit insurance, and disclosure requirements. Convergence on some of these differences is required to provide equal footing for all participants in the housing finance market. To go into further detail on regulatory divergence and its pros and cons, consider the following: The RBI requires commercial banks to allocate resources for priority sectors, including housing finance, and to devote to the sector a mandatory 3 percent of incremental deposits. But HFCs are not bound by lending limits. This restriction does not currently seem to be binding because banks are lending much more than this limit. Banks' capital adequacy requirements (9 percent) are lower than HFCs' require- ments (12 percent), which is justified by the higher risk of HFC assets fully invested in a single sector. Primarily to regulate overenthusiasm of the market for housing finance, the risk weight for individual loans recently was enhanced from 50 percent to 75 percent; and for commercial real estate, it was raised from 100 percent to 150 percent. The investments in mortgage-backed securities issued by SCBs carry a risk weighting of 75 percent, and those issued by HFCs have a 100 percent risk weight. Both banks and HFCs are required to report nonperforming loans of 157 Appendix C C 90 days. The banks are required to maintain provisioning at the rate of 1 percent of standard loans, which is not the requirement for HFCs. Banks are mandated to hold a statutory liquidity reserve ratio of 25 percent of all their borrowings, whereas HFCs are required to maintain a statutory liquidity reserve of 12 percent on their public deposits only. Although the regulatory distinctions cited above appear justifiable, the playing field is considerably skewed by funding cost differences. Most banks fund their retail port- folios through savings and current deposits. In contrast, the HFCs are allowed only to raise term deposits with maturities of one to seven years. Another discrepancy that has been created recently is that bank deposits of five years or more are tax deductible from total income of the depositor up to Rs 100,000. The same benefit has not been granted to the depositors of HFCs. On the other hand, NHB financing is available to HFCs, though at a higher cost than bank deposit funding. Large HFCs (such as DDFC) also can issue their own long-term instruments and tap the capital markets. Finally, taxation differences exist as well. Banks are allowed to deduct the provi- sioning amounts as an expense, but this is not permissible for HFCs. However, HFCs are permitted to transfer up to 40 percent of taxable profits from long-term housing loans to a special reserve, and to claim this amount as a deduction when computing tax liability. This provision helps HFCs bring down their effective interest rate. As a response, certain banks have floated their own HFCs. Growth Potential of the Mortgage Industry The demand for housing in India is very high, and it is expected to continue to increase. The natural population growth of 1.7 percent annually adds about 20 mil- lion people to the housing demand. The population of India will increase from its current level of 1.125 billion to 1.370 billion by 2023. The proportion of population in the working-age group (15­59 years) is 58 percent. Urbanization has accelerated. Household savings are strong at approximately 29 percent of disposable income, and incomes have increased with economic growth. The McKinsey Global Institute esti- mates that average annual real household disposable income in India will grow from Rs 113,744 in 2005 to Rs 318,896 by 2025--a compound growth rate of 5.3 percent (Narayanswamy and Zainulbhai 2007). As a result, the middle class is rapidly grow- ing in India. An estimated 54 percent of population with the average age of 24 years or less and with rising incomes and employment will create a very strong housing demand because the standard age for purchasing a home is between 29 and 39 years. Attitudes concerning debt are changing; and a decreasing aversion to borrowing is observed, particularly for home ownership. Tax incentives are available on mortgage interest and principal payments, and interest rates have been falling in the past decade. The issue remains whether the rapid growth of the mortgage industry will result in sufficient availability of housing finance. Those disbursements grew from Rs 232 billion in fiscal 2001/02 to Rs 3.97 trillion in 2009. The growth of housing finance was 30 percent between the years 2003 and 2008. 158 Appendix C C Recent Housing Finance Policies The global financial crisis has affected India to a larger extent than other countries in the region who have closed financial systems. Nevertheless, the effects were not considerably damaging, and the RBI and the government have succeeded in weath- ering the storm. To ensure uninterrupted and adequate credit to the productive sectors of the economy, the RBI and the government have taken quick, proactive measures to provide rupee and foreign exchange liquidity,11 lower key monetary rates, credit to the production and trade sectors, and three fiscal stimulus pack- ages.12 Although credit by the banking sector expanded for fiscal 2008/09 as a whole, the pace of credit flow decelerated rapidly from its peak in October 2008. The RBI also implemented a policy rate easing, as a result of which commercial banks have announced cuts in the lending rates. In the housing finance sector, there was no direct effect of the U.S. subprime cri- sis because most mortgage lenders in India offer plain-vanilla amortization home loans; and there are no mortgages of the more risky variety--such as interest-only loans, 2/28 adjustable-rate mortgages, piggy-back loans, and subprime or Alt A cat- egories. In general, Indian borrowers are cautious and averse to high leverage; and speculative housing demand is very small to nonexistent. Loan-to-value ratios are low and prepayments are common. Finally, the securitization market is at a nascent stage, and there is very limited exposure to structured products. The RBI further took preemptive measures, including early caution on rising asset prices, increases in provisioning requirements, and higher risk weights. Measures to boost housing finance also were taken, including an NHB refinance facility of Rs 40 billion, a spe- cial liquidity facility provided to banks to enable them to meet the liquidity require- ments of housing finance companies, and a reduction of risk weights on commercial real estate and provisioning requirements for banks. HFCs were temporarily allowed access to short-term foreign currency borrowings. More recently, the RBI has taken some further necessary actions to strengthen prudential norms for the banking sector, including, among others, (1) issuing pru- dential guidelines regarding capital adequacy pursuant to the Basel Committee on Banking Supervision's "International Convergence of Capital Measurement and Capital Standards: A Revised Framework"; (2) adopting prudential norms for off­balance sheet exposures for SCBs; and (3) strengthening prudential norms for valuation, classification, and operation of commercial banks' investment portfo- lios. Updated guidelines were issued in July 2008, requiring all banks to migrate to the Revised Framework on March 31, 2009, and to maintain a minimum CRAR of 9 percent. New guidelines on prudential norms for off­balance sheet items were issued in August 2008. These included directives on provisioning norms for deriv- ative exposures and on computation of credit exposure on account of derivatives. The norms for valuation, classification, and investment portfolios were updated and issued in July 2008; and they covered issues such as internal controls and accounting. Stress testing also has been instituted as a regular measure in the bank- ing system. 159 Appendix C C Keeping in view both international and domestic initiatives in the financial sec- tor, the RBI has proposed the following actions in its annual policy statement for fiscal 2009/10: · to constitute a task force to look into all the issues that have arisen with regard to the G-20 Working Groups and the report of the Committee on Financial Sector Assessment; and suggest follow-up actions relevant for the RBI in the domestic context on an ongoing basis, for every quarter; · in consultation with all regulators and the government, to consider setting up a working group to implement the recommendations of the Committee on Finan- cial Sector Assessment; and · to set up a financial stability unit in the RBI, drawing on interdisciplinary expert- ise from supervisory, regulatory, statistics, economics, and financial markets departments to carry out periodic stress testing and to prepare financial stability reports. Low-Income and Microfinance Housing Strong housing sector growth has been largely in urban areas, and has served the middle- and high-income segments. The financing has focused primarily on the for- mal sector with incomes above Rs 10,000­12,000. In the recent past, financing has started expanding to the informal sector, but it is not reaching lower-income seg- ments in urban areas (Monitor Group 2007). At the lower end of the market, there has been strong growth in microfinance, almost all of which has been focused on rural India. Furthermore, the product focus is on livelihoods and consumption. The loans that are given for housing have short maturities and relatively high interest rates; they are given for repairs or incremental construction. One of the key issues with the current housing finance market in India is that it is not reaching more than 80 percent of the urban population. Estimates of the housing shortage range from 20 million to 70 million. The current housing situation is quite poor for most of these households--small, single rooms with shared toilets, multiple families sharing dwelling units, poor ventilation and lighting, constantly increasing rents, and so forth. With the current shortage of housing and continued migration into urban centers, the situation is likely to get worse. Currently, local and national governments do not have the resources to provide housing for all these households. As much as half of this excess demand can be profitably serviced by the housing and housing finance markets. The Low-Cost Housing Market Initial data on urban housing prices suggested that even at prevailing real estate prices, a significant portion of lower-income segments may be able to afford housing. 160 Appendix C C Figure C.10 Feasibility of Low-Income Housing "We can provide housing at Rs 750­Rs 800 per sq. ft. IBRD 37838 However, the problem is who will give MAY 2010 these people housing finance." Jaipur-based developer Jaipur (population 1.0­2.5 million) Karishma Builders: 450-sq. ft. flats at Rs 700/sq. ft. Vadodara/Ahmedabad Calcutta (Kolkatta) (Chandkheda) within 7 km of city center; (population > 2.5 million) (population > 2.5 million) 45 minutes of travel Mumbai (population > 2.5 million) Hyderabad Casa Bella Developers: (population > 2.5 million) Shankar Haldar Developers: 350-500 sq. ft. flats at Rs 1,000­1,050 350­400 sq. ft. at (North Mumbai) Rs 1,000/sq. ft. approximately 30 km from city center; Area: Goria Jeel (South Calcutta), 1 hour travel by train proximity to the city center "1BHK flats (450­500 sq. ft.) are being sold at Rs 900­1,000 per sq. ft. in areas such as Uppal, L.B. Nagar, Kuthapet, Kukatpalli, etc., but not on a large scale." Hyderabad-based developer Source: Monitor Group 2009. Note: BHK bedroom/hall/kitchen. Currently, the private sector is building small, one-bedroom units in the suburbs adjacent to major towns. These units are approximately 500 square feet in size and sell for Rs 400,000­500,000, including land (figure C.10). They are well connected by public transportation (within an hour of the city center); and are close to markets, schools, and health care facilities. So if units of 200­350 square feet selling for approximately Rs 200,000­320,000 are constructed at the same property rates (in most metropolitan areas, Tier I and Tier II towns), a number of low-income customers could afford them, without subsidies, if they had access to finance. For example, using 3.5 times annual income as the affordability criterion for this income group, a household earning Rs 5,000 a month potentially could afford-- with financing--a 200-square foot unit at current property rates (Rs 800­1,000 per square foot). The numbers are considerable: There are 23­28 million middle- and lower- income households in urban India with monthly incomes between Rs 5,000 and Rs 11,000 (figure C.11). The financial and social impact for these households could be transformative--from enhancing economic prosperity by converting rent into 161 Appendix C C Figure C.11 Potential Housing Demand of Urban Households Monthly Estimated household monthly expenditure: income: 16% >Rs 9,625 (10 million) Rs 11,000 37% Rs 4,575­Rs 9,625 (~23 million) Rs 5,000 Rs 2,500­Rs 4,575 33% (~21 million) Rs 2,500 Rs 10,000 banks typically treat the large private and government banks, large HFCs unorganized salaried segment as part of the self-employed large private and population Rs 8,000­10,000 government banks, NBFCs select medium-size HFCs, target segments for the project cooperative banks, NBFCs Rs 5,000­8,000 government banks, medium-size HFCs, cooperative banks, NBFIs shaded portion represents (select players, small % of portfolio)a unserved customer segments Rs 3,000­5,000 limited MFIs (home repair/extension loans) Rs 2,500­3,000 Source: Monitor Group 2007. a. Certain large, private banks are willing to look at salaried customers with monthly income levels as low as Rs 5,000 if there is proper documentation and if sourcing/collection is done through third parties, thereby managing the overall cost to serve the segment. Few large, private banks have provided housing loans to contract employees or large companies on the basis of collaborations with the employer. 165 Appendix C C are the key barriers to serving the target segment. Market surveys (Monitor Group) suggest that financial institutions have a strong interest in serving salaried organ- ized sector15 customers in this income range if payroll deduction can be facilitated by employers. However, there is limited interest in serving the informal sector (those who are self-employed or salaried unorganized16). Even the financial institu- tions that are open to serving the informal sector will likely need help in develop- ing products and business models that reduce credit risk and transaction costs; and (at least initially) they may need some risk-sharing support. Across financial insti- tutions, there is very low interest in serving customers with household incomes of less than Rs 5,000 a month. Among developers, large developers with a pan-India or large regional presence are not interested in serving the low-income housing market as a commercial propo- sition. However, some small and medium-size developers see low-income housing as a good commercial opportunity. The sector can attain housing cost (land and con- struction) within a square-foot price range of Rs 900­1,000 and make reasonable margins (20­25 percent gross margins); some developers are concerned about the commercial feasibility of low-cost housing in metropolitan areas and Tier I cities. Construction costs are in the range of Rs 400­600 per square foot; and low-cost con- struction materials and technologies could make it possible to reduce construction cost by 10­20 percent (figure C.13). To serve the market, developers need assurance Figure C.13 Estimated Construction Costs for Low-Income Housing 1,200 1,000 980 rupees per square foot 80 800 160 830 500 25 600 450 400 200 200 15 100 0 un n t 5% r fo r st la d d os .2 e re pe ed tio of n an co st op it nd n ya ) ot sc ish tra ua ce fl (e el n tio ut o al io fin gis p ev sq pri ra d st ov ct ku d of re st p co g tru pr ar ss gi am llin d ap ns m gro an se st co nd ty al la du fin re ga p am le st cost factor lower bound of cost estimate higher bound of cost estimate Source: Monitor Group 2009. 166 Appendix C C that lower-income customers will receive housing loans, some help (at least initially) in demand aggregation, and reasonable cooperation from the public sector in pro- ducing basic infrastructure. Construction financing would be helpful in stimulating the market. Financial institutions are open to providing construction finance to small and medium-size developers who meet required credit norms. Large, private banks and HFCs charge higher interest than do government banks. Standard norms include a combination of a track record of executing projects and of repaying loans (if any), financial posi- tion, assessment of project salability, and market reputation. Financial institutions' willingness to provide construction finance to a builder would be further enhanced by the availability of confirmed buyers who have prearranged financing. Employers could play an important role in facilitating housing for their mid/lower-income employees by arranging direct payroll deductions of mortgage payments. Support for housing is strongly advocated by employers who believe housing is a significant need among their employee base; and they view facilitation of housing as an additional employee benefit that would help address attrition and improve motivation. Although employers would not be expected to guarantee pay- ments, they could add value by offering access to their employees, supporting them with documentation, and offering payroll deductions for EMI payments for their permanent workers.17 MFIs are uniquely suited in playing a role in facilitating a housing market for low- income households because they see this as their customers' "most important need"; and because it enables them to get more value from their existing capabilities, rela- tionships, and knowledge of target customers. MFIs would benefit from playing an agency role with financial institutions, whereby they would provide support for demand aggregation; credit evaluation (for example, by sharing the history of their customer base, information on surrogates for income assessment, and expense pro- file); and servicing (that is, collecting EMIs). MFIs' cost structures are much lower than those of financial institutions, and it appears that using MFIs may be econom- ically viable for a financial institution wishing to serve informal sector customers: collection costs may be about Rs 50 and data collection for credit assessments may cost Rs 1,000­1,500. MFIs cannot provide funding or risk sharing because they do not have the capital base or funding structure to do so but they may be able to help reduce risk by providing emergency loans and consumer education.18 Whereas market interest and commercial viability apply to potential mortgage customers with household income above Rs 5,000, it is very difficult to make a purely market-based case for households with incomes below Rs 5,000. Although house- holds earning between Rs 2,500 and Rs 5,000 a month are interested in owning homes, current land prices, construction rates, and lending rates are barriers to affordability. If home ownership in this segment is to be promoted, a combination of lower construction cost, policy initiatives, and government support may help bridge the affordability gap. Among private developers and financial institutions, there is very little interest in serving customers earning less than Rs 5,000 a month, and they are likely to need strong incentives to serve this segment. 167 Appendix C C Market Solutions Several solutions have been found effective, sustainable, profitable, and feasible in market conditions. For low-income salaried customers earning more than Rs 5,000 a month, employer-driven models that target leading employers will be useful in starting the market. Financial institutions are comfortable in lending to these cus- tomers. Developers also are interested because they see this as a way of getting large volumes of "prefinanced" customers. The developers also would like construction finance, and financial institutions are open to this because the presence of prefi- nanced customers provides take-out financing that significantly reduces risk. Because most of these small developers do not get customers to finance their con- struction, the construction finance would improve returns significantly. Scaling up this market may be possible by demonstrating commercial viability and implementation feasibility through pilot projects, by spreading the word over a broader cross-section of employers and developers, and by the government intro- ducing policy initiatives that could increase the supply of "affordable land" and encourage developers to serve this market. Introduction of low-cost construction technologies and potential policy changes (such as reduced stamp duty for small housing units) could enhance affordability and further facilitate serving this market. This approach may serve a large number of the 5­6 million target households with monthly incomes of Rs 5,000­8,000; and the 4­5 million households in the underserved group earning Rs 8,000­11,000 a month. For low-income self-employed and salaried unorganized customers earning more than Rs 5,000 a month, financial institutions are concerned about credit risks and high transaction costs involved in serving this segment; therefore, addressing these concerns is the key task. Certain business models (such as financial institutions using intermediaries like MFIs, NBFCs targeting their clients) and products (a savings product, a group liability product, a lease-to-purchase product, and so forth) may be able to address some of these concerns, but will require further testing. However, to provide momentum to the market and scale it up, financial institutions are likely to need risk-sharing support (such as a credit guarantee and perhaps mortgage insur- ance). The government may have to provide financial support to ensure the credit guarantee is affordable (especially in the initial stages when pricing is likely to be high as a result of a lack of data on the actual credit risk). Financial institutions may also require additional stimulus to serve this market--perhaps extra priority sector benefits for serving this segment, tax credits to lenders serving this market, and the like. Other initiatives mentioned in the solution for salaried households earning more than Rs 5,000 (such as the introduction of low-cost construction technologies, policy changes to enhance affordability) would help provide better housing for this segment. For households earning between Rs 2,500 and Rs 5,000 each month, pure market- based solutions will not work. If home ownership is to be promoted in this segment, it needs a comprehensive housing program that makes housing more affordable and prompts financial institutions and developers to finance and construct such housing. 168 Appendix C C To make housing affordable at current rates, the program may require a combina- tion of market-based solutions to lower the cost of the units (low-cost construction technologies, smaller housing units), policy initiatives to make the houses more affordable (reduced stamp duties), and some support from the government (upfront cash contributions to buy down EMIs and partially finance upfront costs for finan- cial institutions). Suppliers may need additional policy measures (such as providing developers access to nonprime land with a stipulated percentage for housing for this segment, tax incentives on constructing and financing such housing, and so forth) to stimulate them to serve this segment. The process for developing the low-income housing market could be an inte- grated and phased one. The first phase (near term) could involve stimulating the market using the "easiest segment" (the employer-facilitated model for salaried cus- tomers), working in parallel to develop products and models to serve the informal sector, and changing policies to facilitate market development. As housing for the salaried segment gains momentum (and establishes the supply from the "real" sector), the second phase would increase activity in the informal sector to make financial institutions comfortable, with the segment, leading to scaling up housing activity for this part of the market in the medium term. In parallel, if home ownership is to be promoted for customers earning between Rs 2,500 and Rs 5,000 a month, the government could develop a comprehensive package for them, while leveraging the market that is served by the private sector (for customers earning between Rs 5,000 and Rs 11,000) to implement the package (for example, by using the same developers to build this housing). There are three interlinked attributes of the above strategy that are noteworthy. First, the underlying "cluster approach" being suggested here is a fundamental busi- ness "innovation" that makes low-income housing attractive to the private sector. It addresses risks for all the key players and provides a business opportunity that makes the market fundamentally more attractive. For example, in the employer-led model, the presence of a large pool of customers with low risk (that is, employees of a good company) and low transaction costs (because of a homogeneous customer base, documentation support from the employer, and payroll deduction for payments) makes it attractive to the financial institution. The large pool of customers with financing reduces the developer's market risk and makes the project attractive. Prefinanced customers also make it possible for the financial institution to provide construction financing as the take-out risk is addressed. This makes it even more attractive to the developer because it reduces his capital needs and may allow him to speed up construction--both of which improve his return. The core elements are the same in the approach for the informal sector--the MFI (analogous to the employer) acts as a demand aggregator, reduces costs, and helps address risks for the financial institution. Given that the risk is perceived to be higher, a credit guarantee could be introduced to manage the institution's risk. The other elements are the same-- a prefinanced pool of customers and construction finance for the developer. Second, there must be clear recognition that market-based solutions have limits and that one may need to have "smart" support to extend them. For the salaried 169 Appendix C C customers earning more than Rs 5,000 a month, the economics of market-based solutions work, but government support on policy initiatives to increase the supply of affordable land or to standardize low-cost construction can really help scale up the market. For the informal sector, although the base economics work, active participa- tion by financial institutions would require risk sharing; and the government could stimulate this through an affordable credit guarantee. For households earning less than Rs 5,000 a month, although market-based solutions can make housing more affordable, if home ownership is to be promoted, two issues must be addressed: there is still an affordability gap that the government may have to help close; and because financial institutions are reluctant to serve this market, government may have to provide incentives to get them interested. Third, although there are numerous ways in which the government could facilitate housing for lower-income households, many of these ways do not have a financial cost. For example, for the 9­11 million households of salaried customers earning more than Rs 5,000 a month, some of the potential government interven- tions (such as increasing the supply of affordable land) may require only policy changes and no financial support. For the informal sector, there would be some cost for initiatives like an affordable credit guarantee; but many potential interventions have no cost. The only time that significant support may be required from the gov- ernment is when home ownership is promoted for households earning less than Rs 5,000; even for these, however, the needed support would be much less than it would be without market-based solutions and the suggested policy changes. Pilot Projects in Low-Income Housing The market solutions proposed above have been pilot-tested and proved potentially successful. The Monitor Group, in association with the NHB and with technical assistance from the World Bank, has initiated demonstration pilot projects purely as commercial initiatives with the private sector. The opportunity to serve low-income customers on a commercially viable basis has resonated with a cross-section of stake- holders who are interested in engaging in specific pilot projects spanning Ahmed- abad, Hyderabad, Mumbai, Pune, and other towns. A number of similar initiatives are in preparation. Box C.2 presents a brief review of some ongoing pilot projects. Housing Finance for Women Another unique player in providing housing microfinance in India, with a business focus on self-employed women, is the SEWA Bank (Center for Urban Development Studies 2000). The Self-Employed Women's Association (SEWA) was established in 1972 in Ahmedabad City as a trade union with the goal of organizing low-income women working in the informal sector. SEWA targeted what amounted to 96 percent of employed women in India who worked in the informal sector with no rights, security, or protection. SEWA categorizes informal sector workers into three cate- gories: (1) vendors/hawkers, (2) home-based workers, and (3) manual laborers and 170 Appendix C C Box C.2 Market-Based, Low-Income Housing Pilot Projects in India Mumbai Pilot Project The two "anchor" participants are UTI Bank and Lodha Group. Casabella is a recently completed, 65-unit project at Nalasopara, an upcoming suburb on the western railway line in Mumbai. Lodha Group has started another project at Virar, a suburban town also on the western railway line. UTI Bank recognizes that housing is a problem for its lower-income staff, and would like to help its staff buy apartments. There is strong backing from key senior managers in the organization, and the team working on the pilot project has received clearance to finance these customers. The tar- get group for this pilot project is the drivers of UTI officers. These individuals are not formal employees of UTI Bank (they are employed by the officers themselves), but the bank thinks of them as part of its extended staff. Drivers earn roughly Rs 7,500 a month and thus can afford the housing being proposed by Lodha Group. UTI Bank is willing to provide home loans to the drivers and, potentially, to provide finance to the developer (covering the number of units required for the bank employees). The pilot project is located at Nalasopara, within about a one-hour commute by train to the city center. The proposed development will have 300 apartments of 225 square feet each (in four-story walk-up buildings. Units will be priced at Rs 325,000. Ahmedabad Pilot Project The anchor here is SEWA Bank--a cooperative bank for women, which has a large presence among low-income households in Ahmedabad. The senior managers at SEWA and SEWA Bank are interested in facilitating housing for their employees and clients, and are willing to provide them with mortgages. Gruh Finance is a medium-size HFC with headquarters in Ahmedabad. It has low-income customers and is interested in expanding its portfolio. Gruh is interested in financing the SEWA Bank employees and clients, and is willing to provide construction finance to reliable developers. The project team has identified two developers, Makwana and Gajjar, who work in the areas suggested by the focus group participants and who have good reputa- tions (based on feedback from Gruh, which has financed the developers' customers). These developers have recently completed projects that are of good quality. Both have access to land and can construct apartments of 260 and 324 square feet at Rs 800 a square foot. Pune Pilot Project The anchor, Bajaj Auto, is one of the largest two-wheeler manufacturers in the country. It is interested in facilitating housing for its factory employees and for the employees of six of its vendors. Bajaj is willing to help with the documentation requirements and to do payroll deduc- tion for their permanent employees. Initial discussions with groups of both Bajaj Auto's and its vendors' employees have confirmed the employees' interest in buying housing facilitated by Bajaj Auto, and have helped the project team identify geographic preferences and affordability. ICICI Bank and Dewan Housing Finance have agreed to provide financing to the employees. Source: Monitor Group 2007. 171 Appendix C C service providers. Its borrowers either are self-employed or they work as casual laborers, they maintain little or no savings, and they hold no assets. The main goal of SEWA, articulated by its founder Ela Bhatt, is to empower "invisible" female infor- mal sector workers and help them become self-reliant, with employment security, income security, food security, and access to social services such as health care. Through SEWA, female members access many services (including capital from sav- ings and credit groups, and health and child care), which have evolved to become operationally and financially autonomous cooperatives. By the end of 1999, SEWA had a total membership of 220,000. Access to capital, one of SEWA members' most important needs, led to the establishment of the association's largest cooperative entity. In 1974, the Shri Mahila SEWA Sahakari Bank, known as SEWA Bank, came into existence by way of small deposits (Rs 10, or $0.23) from 4,000 self-employed women. These deposits made up most of the bank's initial working capital, Rs 60,000. Government Housing Policies and Programs Since independence in 1947, successive Indian governments have highlighted the priority of housing in government planning through different five-year plans (table C.12). The total investment proposed in housing has increased from Rs 11.5 billion under the First Plan to Rs 7,263 billion under the 10th Plan. However, the absolute investment in housing as a percentage of the total plan investment has declined because of the shift in the government's emphasis from provider to facilita- tor (Garg 1998). Government assistance for low-income housing is also imparted through some urban and rural support schemes. The Basic Services for the Urban Poor scheme is an urban poverty reform by local bodies. It seeks to provide security of tenure, affordable housing, water and sanitation, health, education, and social security in low-income settlements in 63 target cities. The Interest Subsidy Scheme for Housing the Urban Poor seeks to provide the aforementioned seven entitlements/services in towns and cities other than the target cities. Both programs were launched in 2005. The Basic Services for the Urban Poor project amount is Rs 500 billion ($10 billion), using financing shared among the central, state, and local agencies. The program tar- gets universal access to a minimum level of services, provision of basic services to urban poor people, introduction of a system of e-governance, reform of property tax, and rationalization of stamp duty. The Interest Subsidy Scheme addresses the need to provide affordable housing for the lower-income segments in urban areas, in a partnership among individual, state, and local bodies, and banks. It is beneficiary driven, and uses a central government subsidy of 5 percent for loan amounts of Rs 100,000 for a period of 15­20 years. The net present value of the subsidy over 15­20 years is deducted upfront from the principal amount. The Golden Jubilee Rural Housing Finance Scheme was launched in fiscal 1997/98, targeting the poorest population segments in rural areas. It is a market- based program implemented through lending institutions, with the NHB as the 172 Appendix C C Table C.12 Evolution of Housing Policies in India: A Succession of Five-Year Plans, 1951­2012 Plan Main Features Investment Targets First Plan, 1951­56 · Housing was introduced into policy frame- · Public: Rs 2.5 billion work at the national level. · Private: Rs 9.0 billion · To enhance affordability, government · Total: Rs 11.5 billion support through subsidies and loans was considered necessary. · Separate Ministry of Works and Housing was created. · National Building Organization was created. Second Plan, 1956­61 · The initiatives of the First Plan were · Public: Rs 3 billion strengthened. · Private: Rs 10 billion · Policy shift moved housing from central · Total: Rs 13 billion to state governments (housing as state subject). · Central government decided to provide assistance to the state governments for developing low-income housing (instead of direct intervention). · State housing boards came into existence. Third Plan, 1961­66; · Emphasis was placed on planned develop- · Public: Rs 4.25 billion Annual Plan, 1966­69 ment and land acquisition in urban areas. · Private: Rs 11.25 billion · More emphasis was put on low-income · Total: Rs 15.50 billion housing. · Resources of the state development boards were enhanced to enable them to address the issue of housing shortfall in their respective areas. Fourth Plan, 1969­74 · Decision was made to encourage private · Public: Rs 6.25 billion and cooperative housing schemes by pro- · Private: Rs 21.75 billion viding financial assistance. · Total: Rs 28.00 billion · In 1970, decision was made to set up HUDCO to provide housing finance to the low-income segment at below-market inter- est rates and for longer tenors. · HUDCO was to finance urban development activities and help decongest cities. · HUDCO was to finance state housing boards by funding their bonds and providing other financial assistance. Fifth Plan, 1974­79 · The Urban Land Ceiling and Regulation Act · Public: Rs 7.69 billion was introduced. It sought to prevent con- · Private: Rs 36.40 billion centration of landholding in urban areas and · Total: Rs 44.09 billion to make more land available for equitable distribution. However, it was not successful in achieving its targets. · The HDFC, a completely private initiative, was set up in 1977 as the first HFC. It sought to provide financial assistance to individuals, groups, cooperative societies, and the corporate sector for staff housing. (Continued ) 173 Appendix C C Table C.12 Evolution of Housing Policies in India: A Succession of Five-Year Plans, 1951­2012 (continued ) Plan Main Features Investment Targets Sixth Plan, 1980­85 · Aimed to increase housing in small and · Public: Rs 14.9 billion medium-size towns. · Private: Rs 180.0 billion · Intended to improve conditions of the slums · Total: Rs 194.9 billion and living conditions in such habitats, while emphasizing the need to increase support to private groups. · More housing companies from the private sector entered the market. Seventh Plan, 1985­90 · There was a radical reorientation of all · Public: Rs 24.6 billion policies relating to housing. · Private: Rs 290.0 billion · The plan argued that major responsibility for · Total: Rs 314.6 billion house construction would have to be left to the private sector. · Government was to promote and facilitate the housing sector and move away from its earlier role as "provider." · India subscribed to Global Shelter Strategy, passed by the United Nations General Assembly in 1988. · Based on that, India drafted its first national housing policy. · The NHB was set up in 1988 to promote and regulate HFCs and to mobilize addi- tional financial resources for housing. · The Building Material and Technology Promotion Council was constituted. · During this period, several HFCs were promoted. Eighth Plan, 1992­97 · The plan acknowledged that low-income · Public: Rs 315 billion housing requires state intervention. · Private: Rs 660 billion · The National Housing Policy was adopted by · Total: Rs 975 billion the Parliament in 1994. · Recognizing that urbanization was inevitable, it concentrated resources on upgrading urban centers. · Recommended need for reforms on financial and legal aspects to promote the mortgage market. · Laid special emphasis on government incentives to enhance flow of credit to the housing finance sector. Ninth Plan, 1997­2002 · Included further reforms to enable the gov- · Public: Rs 520 billion ernment to play its role as a facilitator in · Private: Rs 990 billion promoting the mortgage market. · Total: Rs 1.51 trillion · Included market-friendly reforms for improv- ing, both taxes and infrastructure to help increase investments in housing. · In 1999, the central government repealed the Urban Land Ceiling and Regulation Act. · Revised the National Housing Policy in 1998. · During the period, commercial banks entered the housing finance market. 174 (Continued ) Appendix C C Table C.12 Evolution of Housing Policies in India: A Succession of Five-Year Plans, 1951­2012 (continued ) Plan Main Features Investment Targets 10th Plan, 2003­07 · Maintained the spirit of the Ninth Plan. · Public: Rs 4.15 trillion · Commercial banks aggressively entered the · Private: Rs 3.11 trillion housing finance market. · Total: Rs 7.26 trillion 11th Plan, 2007­12 · There was an increase in capital assistance Total: Rs 5.10 trillion under the Indira Awaas Yojana pro-poor gov- ernment housing scheme · Launch of market-based composite housing loan schemes for rural borrowers just above the poverty line--economically weak and low-income segments of the population. · The National Rural Shelter Fund was established. · Introduction of a title indemnity scheme and an interest subsidy scheme for urban poor: 5% annual interest for a maximum loan amount of $2,000. · Affordable housing continued to be the major consideration behind the various pol- icy measures taken during fiscal 2005/06 by the government, the RBI, and the NHB. Source: NHB 2008. monitoring agency. Roughly 3.5 million dwelling units have been financed since the program's inception. Support from the NHB includes refinance assistance at conces- sional rates, as well as technical assistance and capacity building. The program is a centrally sponsored subsidy scheme, with costs shared between the central govern- ment and the states on a 75/25 basis, respectively. Other examples of government support programs include the Bharat Nirman Program, which involved the construction of 6 million houses over a four-year period beginning in fiscal 2005/06; and the Rural Housing Fund, administered by the NHB, which targets the shortfall in priority sector lending by SCBs. Regulatory Framework and Mortgage Market Infrastructure The land administration system is obsolete, with multiple responsible institutions carrying some overlapping functions, inaccuracy of records, and the resulting long delays and transaction costs. Records are manual, recourse is inadequate, and staff lacks capacity and efficiency. Registration fees and stamp duties are high, discourag- ing formal registration. In contrast, foreclosure procedures in India have been improved considerably in the last five years, as has the environment for getting credit. 175 Appendix C C Landownership, Registration, and Titling The land administration system is obsolete and has not been substantially modified in the last century.19 The Land Revenue and the Survey and Land Records depart- ments are the key institutions responsible for land administration in India. Several inefficiencies characterize the system. First, any land deals transacted after 1882 entered the land registry system and were recorded in the revenue and the stamps and registration departments. This duplication of institutions increases transaction costs for landowners--without providing commensurate benefits--and introduces a major source of tenure insecurity resulting from potential record inconsistencies. Second, the responsibility for maintaining an accurate spatial record of landowner- ship of rural areas at the urban fringe has been transferred from the survey depart- ment to municipal corporations who maintain the records for tax purposes only (in some cases rendering them unsuitable as a basis for ownership record), which results in outdated maps and high levels of land-related conflict in peri-urban areas (28 percent of parcels, according to one study [World Bank 2007b]). Third, land administration is not extended to unrecorded wastelands currently brought under agricultural cultivation. Extending administration to those lands will require clarifi- cation of the interface with the forest department and a broadening of tenure choices that can be accommodated within the land administration system to include, for example, forms of communal ownership.20 The basic structure of land administration in any Indian state comprises four main institutions (figure C.14). The land revenue department maintains the textual database for land records (as well as tax registers where they exist), and it continues to collect land revenue; however, it took on a host of administrative functions. The Department of Survey and Land Records is responsible for maintaining spatial data, mapping and demarcating boundaries, and executing surveys for subdivision on demand. The Office of Stamps and Registration is responsible for registering deeds and for collecting stamp duty due on these transactions. Although details and names vary across states, a discussion of the general principles can help identify some of the key problem areas that need to be confronted. Figure C.14 Agencies Involved in Land Administration Department of Land Revenue Office of Stamps Survey and Local Bodies Department and Registration Land Records · field measurement · Record of Rights · register of · town survey maps book/tippan register transactions and register · Village map · adangal/pahani · layout plans · Shetwar · khata register · property tax register Source: World Bank 2007b. 176 Appendix C C The key institution for managing land records is the department of revenue, which traditionally has constituted the main interface between the state and the local population in rural areas. Its importance is illustrated by the fact that at the district and lower levels, revenue officials also assume judicial functions. The revenue department continues to maintain an enormous field presence that dates back to colonial times. Functions at the district level are overseen by a collector. Below the districts are blocks (taluks, mandals), often subdivided into revenue circles compris- ing a number of revenue villages. Within each revenue village, an official (called pat- wari, talathi, village accountants, lekhap/kanungo, and the like) is responsible for maintaining textual land records through annual reconciliations (jamabandi), recording in the Record of Rights (ROR) any transfers of ownership that may have taken place, and issuing certified copies of the ROR for use by cultivators as proof of ownership. Several original documents are kept at the village level, with copies at the block or district level. The ROR is equivalent to a title to the land. It records the nature and extent of the respective rights and interests of all persons, including their credits and liabilities and, in some states, encumbrances such as mortgages existing on the land. The ROR constitutes prima facie proof of ownership, and all entries in it are presumed to be true until proved otherwise. It is thus an important document routinely used by farmers wishing to prove landownership to obtain loans from banks or to take out crop insurance. An authorized copy of the ROR can be obtained from the village accountant or, where these records have been computerized, from the local kiosk. The adangal/pahani contains crop details that were originally col- lected on a seasonal basis to compute land taxes. In modem times, the information contained in this record is expected to be transferred to the agricultural department to make supply projections and plan marketing campaigns. The frequency and accu- racy with which this information is collected at the local level differ widely across states; and the channels through which the information is transmitted to higher lev- els, as well as the use made of it, are not entirely clear. The khata is a tax record show- ing land revenue, land tax, water rates, and other government dues to be paid by a cultivator; and the amount paid in a given year. The khata register is prepared once a year, during the annual reconciliation of accounts by the village official. Land Revenue Department. The Land Revenue Department was starved of funds for improvement of service until the late 1980s. With a gradual decline in the impor- tance of land revenue as a source of government income, the department has increas- ingly been given other mandates that are unrelated to its original focus on land administration. These mandates made it difficult for officials to keep up the quality of service provision, and were often associated with a significant deterioration in the ability of the revenue department to focus on the maintenance of land records that had once been its core function. Until the launch of centrally sponsored programs to modernize land records in the late 1980s, states got neither funding nor technical advice on land matters. A further challenge was presented by the excessive discretionary power of local officials. Conditions in which demand for public services outstrips supply by a wide 177 Appendix C C margin are fertile ground for rent seeking by officials. Indeed, village officials have been reported to charge high fees (speed money) for routine services such as the pro- vision of certified copies of the ROR to farmers or to conduct property transfers in a timely manner. These fees further drive transactions into informality, in addition to marginalizing poor landholders for whom the direct and indirect costs of navi- gating around a complex and often nontransparent system may well be too high. The impact of this is exacerbated by the facts that land records are rarely in the public domain; that revenue codes and procedures are complex to comprehend; and that in a context of local power structures that are not at all in favor of the poor, even enforcement of court orders becomes a nontrivial issue. It is thus not surprising to find reports of local officials using their privileged access for tampering with records and extracting bribes. Department of Survey and Land Records. Even though it originally formed an independent department, the Department of Survey and Land Records in many states is now integrated with the Land Revenue Department, at least at the higher levels (as is true in Tamil Nadu). In many cases, however, integration is more limited lower down the hierarchy, and there are large variations across states in the extent to which municipal bodies are integrated into the department's structure. To provide a spatial record of landownership, the colonial administration car- ried out ground surveys with basic technology on a village-by-village basis, with- out a link to a national coordinate system. Transfers were recorded in the textual records on an annual basis by the local administrator. However, the main mecha- nism to update cadastral maps was through periodic resurveys to be undertaken every 30 years. By then, population growth and inheritances, construction of infra- structure (such as irrigation), and possibly expansion of cultivation into adjoining public land implied that a revisional survey, combined with reassessment of land revenue, was probably cost effective. The basic spatial records are field measure- ment books, village maps, city surveys, and the tax records of local authorities. In the process of the original surveys by the British, parcel boundaries and the village limits were demarcated on the ground with monuments. For each field, a field measurement book was prepared, normally on a 1:1,000 or 1:2,000 scale. It records the plot's relative coordinates and distance measurements (in chains, links, or other local units), thereby allowing computation of its area. Each field is assigned a survey number that is unique within the village, and that number provides the legal basis for identification of parcel boundaries. The information recorded in the field books is used to compile the village map, usually at a somewhat smaller scale (1:7,920). This document contains boundaries of individual plots according to the original settlement or any revisional survey that may have taken place in the interim. Main physical features as well as the location of fields relative to each other and prominent topographical marks are included in the village maps. The Land Revenue Act stipulates that for urban areas with a population of more than 5,000 (in some cases, 2,000), city survey maps may be prepared. Where this has been done, details of landownership are maintained in a simplified document 178 Appendix C C known as a property card. If city surveys do not exist, property tax receipts kept by the local authorities assume an important function. Although the maps are not meant to provide legal evidence of ownership, local authorities have put consider- able effort into establishing tax maps that allocate a unique assessment number to each property. Because of a lack of resources and staff, spatial records were not regularly updated, however. Furthermore, the lack of a common regulatory framework for surveys led to the proliferation of many map products that did not follow common standards. Third, there were no systematic efforts to expand the spatial database for land administration into areas that had not been settled at the time of the original survey, or to close existing gaps. Given their age, the physical condition of the field measurement books and village maps is precarious. Although overall figures are dif- ficult to obtain, investigations in Andhra Pradesh revealed that only about two thirds of plot maps were in good condition and 21 percent were completely missing. The level of preservation varied significantly; in areas of the state that were formerly under Zamindari rule, more than 30 percent of plot maps were missing and less than 50 percent were in good condition. In areas of the state where the ryotwari tax sys- tem was formerly used, 13 percent were missing and 79 percent were in good condi- tion. Also, identification of boundaries is complicated by the fact that a large share of the boundary stones put in place during the original survey is missing as well. Pilot project experiences suggest that splicing field measurement books together to obtain a substitute for village maps will be nearly impossible. Thus, unless alterna- tive options (such as use of satellite imagery) can be explored, reconstituting village maps in the 30­40 percent of cases where they are missing or not usable will require huge amounts of actual fieldwork. Given the cost involved, it is not surprising that this has rarely been done. Even in cases where original survey records are still available, intervening trans- actions that were not surveyed imply that their relationship to current patterns of ownership and occupation may be tenuous or nonexistent. In rural areas, a key rea- son for discrepancies lies in transfers and partitions--either through inheritance or sale--that were not recorded or not accompanied by a resurvey. Given the decline in land revenue, carrying out such resurveys at the required speed and scale is no longer an option, implying that the spatial framework for land administration even in rural areas has fallen into a state of disrepair. For example, in Andhra Pradesh, the average survey is overdue by about 60 years, which suggests that the last resurvey in the aver- age district was carried out around the turn of the 20th century--and a similar sit- uation applies in other states. Given its original focus, the spatial coverage of the land administration system has considerable gaps in both rural and urban areas. In rural areas, large stretches of land were not included in original surveys for various reasons, even though they may long have been under cultivation. In urban areas, many of the cities that would be obliged to do so have not conducted city surveys, or have failed to update the surveys and the system of property cards that is supposed to be based on those surveys. Moreover, there is no obligation to have any record of landownership, either spatial 179 Appendix C C or textual, in habitations with populations of less than 5,000. Such habitations are home to the vast majority of India's population. This implies that, at least as far as their place of living is concerned, most Indians do not have a spatial record. The lack of updating is most serious for urban areas that were previously part of a village--something that applies to the large majority of urban areas in India today. It is not too difficult to imagine that, in these cases, maps will bear little relationship to current settlement patterns. In the process of conversion, town development authorities, who normally acquire land for extension of towns, prepare layout plans that are then passed on to the municipal council. However, the council has no responsibility, incentive, or capacity to maintain detailed data on landownership. Developers or individuals, of course, can take the initiative to get their plots sur- veyed, but this is not a routine process; and no cadastral maps to provide physical descriptions of plot boundaries are normally used in the process of registration or kept at the registry. Even in cities or municipal corporations that maintain maps or some other type of spatial record, the quality and reliability of such information dif- fer widely. Clarification of responsibilities and establishment of clear standards, together with technical capacity to monitor their implementation, could greatly enhance the value of such maps even for land use and urban planning and for a sys- tem that could help define property rights in land. Office of Stamps and Registration. In most states, the Office of Stamps and Regis- tration, which deals with registration of commercial vehicles in a much broader sense, is headed by the Inspector-General of Revenue and Stamps or equivalent, and functions under the Finance Department or the Commercial Taxes Department. Even in cases where revenue and stamps are located on the same campus, links and contacts with the land records section of the Land Revenue Department are rather rare. Coordination is made more difficult by differing administrative structures of the two departments in some states, implying that subregistry offices (the lowest administrative unit for registration of documents) often are located in different towns and have different areas of responsibility than do the lowest revenue depart- ment units. According to the 1882 Transfer of Property Act, any sale of tangible, immovable property worth Rs 100 or more can be made only by registered instrument. Struc- ture and procedures for registration are governed by the 1908 Land Registration Act, with state amendments. The importance of the Office of Stamps and Registration derives from the facts that, upon registration, a fee proportional to the value of the property transacted has to be paid; and that in view of rather high rates, revenues from stamp duty have become a major source of revenue, amounting to about one third of the total revenue in a large number of states. Concerning real estate, the key function of the Office of Stamps and Registration is the registration of sales and mortgages, thus creating a public record of transac- tions. Documents to be registered must be presented in person at the subregistrar's office within four months of execution. The stamp duty to be paid for registering deeds ranges from 3.0 percent (New Delhi) or 4.0 percent (Ranchi) to 12.5 percent 180 Appendix C C (Kochi) of the property's market value, and is often somewhat higher in cities and municipal corporations than in rural areas. In addition, a registration fee that ranges between 0.5 percent and 2.0 percent of the market value must be paid. Deed regis- tration costs, charged by subregistry offices in most cities, also vary significantly: from 1.0 percent of the property value in Ahmedabad to 8.5 percent in Guwahati. Registration of mortgages requires payment of the registration fee plus a share of the mortgage value (from 0.1 percent to 2.0 percent). Thus, with the decline in the importance of land revenue, income from registration fees has evolved into one of the major sources of states' income. For a nominal processing fee, the subregistries also supply encumbrance certifi- cates that contain all registered dealings for a given parcel over a stipulated period. These certificates are usually prepared manually by going through the indexes of reg- istered transactions. Having an encumbrance certificate that covers a long period is important for purchasers of land to explore possible defects in the title that a poten- tial seller holds to a specific piece of immovable property, implying that the ease with which such documents can be made available will have a major bearing on tenure security. The encumbrance certificate is also relevant for financial institutions want- ing to use a property as collateral for a loan, to ensure that no prior and possibly superior liens have been registered on the property. It is thus not surprising that this service is in high demand by the public and the financial sector. At the same time, the fact that registration is not compulsory--and is evaded to avoid stamp duty even in cases where registration would be required by law--greatly constrains the value of evidence from the registration system, irrespective of whether it is generated manu- ally or by computer. In practice, the effectiveness of registration is reduced by four factors: (1) high stamp duty and often opaque methods of valuing properties prompt parties to keep their transactions informal, with negative consequences for the completeness or integrity of the system; (2) low efficiency and high cost of even routine operations act as further deterrents to formalizing transactions; (3) lack of links to other parts of the property system reduces the security that registration can provide, even under optimum conditions; and (4) the fact that registration neither ensures legal validity of a transaction nor provides parties with additional assurance as to the validity of the transaction discourages registration. Despite recent stamp-duty reductions by some states, rates remain very high by international standards, at an average of only slightly less than 10 percent of the transaction value. Such high rates discourage formal land transactions; and they lead the parties involved to avoid registration through various means, such as powers of attorney and other side agreements. Such avoidance reduces government revenues and undermines the completeness of the land registration system. In the case of mortgages, a common practice to avoid payment of stamp duty is to enter into an equitable mortgage by deposit of title deeds rather than a formally registered mort- gage. To avoid having to pay high stamp duties, many parties resort to reporting land values that are significantly below what actually was paid. The problem with explic- itly or implicitly condoning such an approach in manual systems is that land values 181 Appendix C C have to be determined or certified by a government official who will often demand a fee for "service." Deeds are often complex documents that may exceed 40 or 50 pages, and specialist knowledge will be required to decipher them and navigate the system. One reason for this is that the plethora of state-specific land reform laws and land use restrictions adds to the complexity. Simplification of these laws could help standardize deeds--something presently often not complied with even in situations where regulations exist--and thus help make the process of registering property more transparent and less costly. The registration system is governed by antiquated procedures, including manual copying and indexing of documents by person rather than by parcel (which increases the cost of obtaining information). The high price for information is particularly important in case of encumbrance certificates for mortgages where they could well result in a large number of otherwise creditworthy borrowers being rationed out of the credit market. The absence of performance standards made it possible for public servants to keep documents pending for registration or to not return original documents for years without good reason. Registration of real estate requires five procedures, 44 days, and 7.43 percent of property value. These figures compare favorably with the South Asia region, but there is ample space for improvement when compared with international best practices and with best-practice aspects of Indian states (figure C.15). Registra- tion delays of two to three years are not uncommon where extensive and often unnec- essary quality assurance procedures slow the process, or where many offices have to be contacted for official papers. Simple improvements in information management could have far-reaching effects on these performance parameters. To tackle these issues, the Ministry of Urban Development launched the Jawaharlal Nehru National Urban Renewal Mission, a nationwide program aimed at streamlining property registration. Under the program, a series of reforms is under way. Several states have cut their stamp-duty rates: Delhi (from 8 percent to 6 percent for men, and from 6 percent to 4 percent for women), Gujarat (reduced to 4.9 percent in 2006), Orissa (from 11 percent to 5 percent in 2008), and Uttar Pradesh (reduced to 5 percent in 2005). Selected cities have further computerized their land records, making it easier for entrepreneurs to search for encumbrances and register their sale deeds (table C.13). Registering property is not equally cumbersome everywhere. Although all states require the same main steps--searching for encumbrances, drafting a sale deed, registering the deed at the subregistry, and transferring the property title at the Circle Revenue Office--time and cost differ from city to city. In Gurgaon, it takes only 26 days and 7.7 percent of the property value to register property--easier than in Guwahati where it takes 84 days and costs 15.4 percent of the property value to do so. Well-performing cities like Gurgaon, Jaipur, and New Delhi provide good examples from which other cities can learn. If all cities were to adopt India's best practices in property registration, India's global ranking on ease of registering property would improve by 35 ranks (to 70 out of 181 economies), overtaking Hong Kong, China, and Kuwait. Computerization and proper staffing are mostly 182 Appendix C C Figure C.15 Time and Cost to Register Property in Indian States and Selected Countries Saudi Arabia 0 2 Jaipur 9.9 24 Gurgaon 7.7 26 Kochi 15.5 27 Bangalore 9.2 28 China 0 29 Hyderabad 10.5 37 Noida 25.4 37 Indore 10.7 39 Ahmedabad 7.0 42 Brazil 0 42 Mumbai 7.4 44 Chennai 10.1 48 New Delhi 7.5 55 Ranchi 5.4 56 Ludhiana 10.6 67 Guwahati 15.4 84 Patna 12.4 87 Calcutta 7.9 107 Bhubaneshwar 7.5 126 0 20 40 60 80 100 120 140 cost (% of the property value) 7.5 24 time (days) Source: World Bank 2008c. responsible for shortening the delays. If the registration system is unified across offices, as in Jaipur, the encumbrance search and registration of the sale deed can be done in any subregistry office, regardless of its jurisdiction. Offices that com- plement technological improvements with staff training and performance man- agement reap the fruits of their efforts. Gurgaon, India's top performer in property registration, is a case in point. The Haryana Urban Development Authority, in charge of title transfer, hired a consulting company to develop customized software and train one employee from each unit. The authority now requires all its employ- ees to be proficient in the new computer system. As a result, its offices take only 15 days to transfer a property title. 183 Appendix C C Table C.13 Reforms of Property Registration Procedures, 2008­09 Improved Reduced Improved Reduced Land Record Administrative Administrative Stamp Duty Computerization Fees Efficiency Ahmedabad, Gujarat Bangalore, Karnataka* Bhubaneshwar, Orissa* Chennai, Tamil Nadu* Gurgaon, Haryana Guwahati, Assam Hyderabad, Andhra Pradesh* Indore, Madhya Pradesh Jaipur, Rajasthan* Kochi, Kerala Calcutta, West Bengal* Ludhiana, Punjab Mumbai, Maharashtra* New Delhi, Delhi* Noida, Uttar Pradesh Patna, Bihar* Ranchi, Jharkhand* Sources: World Bank 2007a (locations marked with an asterisk), 2008a. Note: These reforms took place between April 2006 and February 2009. Many of the problems with land administration in India can be traced to dupli- cation induced by institutional overlap, which also greatly increases the cost of land administration. Local governments as well as the Department of Survey and Land Records are likely to use city surveys and tax maps, on the one hand, or village maps, on the other, for some of the same areas. Information on ownership of the same pieces of rural land may be available in the registry (if land was transacted or a mort- gage was taken out) or the ROR. However, the two sources will differ if land was transferred but the transfer was not effected, and neither source will be definitive. As a consequence, any person interested in entering into a transaction with land must conduct searches in the land record and the registry office. In each of them, time has to be spent and both official and unofficial fees must be paid. Such inconsistencies create potential for fraud and conflict. Estimates of the magnitude of discrepancies between the different systems are difficult to obtain. Cases where registered transac- tions are not followed by a transfer of the ROR and/or a survey is not done can have substantial cumulative effects, given that it has been 70­100 years since the last sur- vey was undertaken in many states. A considerable challenge is posed by the presumptive nature of registration and the lack of cross-checks in the system. Whereas the law states that registration of a 184 Appendix C C legally valid deed will result in transfer of title to a given property, the registrar has no obligation--and often lacks the ability--to check the validity of any deed that is presented for registration. Even in cases where rudimentary checking could have dis- covered a defect in the transaction, the registration department cannot currently be held accountable for having registered a deed in which title was unclear or disputed. In fact, given that the rationale for registration was the generation of revenue, the registrar (at least in principle) will register any document presented, as long as the formal procedures are met and appropriate stamp duty is paid. This means that the perceived benefits from having a transaction registered may be significantly below the cost of doing so, implying that landowners may opt out of the system. The registry's inability to access and easily refer to cadastral maps makes it very difficult for registry officials to perform even basic checks on either the validity or the legal- ity of instruments delivered to them, making it nearly impossible to enforce stricter standards of accountability. This is exacerbated by the fact that the descriptions of land parcels in many deeds do not provide accurate locations of the properties, par- ticularly in urban areas--something that greatly increases the possibility of future disputes. Thus, the system expends large amounts of effort to ensure the correct identification of parties to a deed (an issue that is regulated in minute detail), but vir- tually no regulation exists to provide for the accurate geographic description of the land in question. Title Insurance. Title indemnity is primarily a business of loss avoidance, and it can be availed by both lenders and borrowers and home buyers. It is incidental and directly related to the business of housing finance as a risk mitigation tool providing the necessary protection in respect to the underlying securities for extending hous- ing finance. It ensures transparency and clarity in the ownership status of underly- ing mortgaged properties at any given time, and enables safe transfers of property and mortgages so as to curb fraudulent transactions. The National Housing and Habitat Policy (India, Ministry of Urban Employ- ment 2005) has envisaged the introduction of title indemnity as a tool for mitigat- ing the risks pertaining to imperfections associated with property and mortgage titles. This policy will also facilitate the process of creating a central registry to aid transparency in the ownership status of residential properties and will enable their safe transfers, as envisioned in the Securitization and Reconstruction of Financial Assets and Enforcement of Securities Interest Act of 2002 (hereafter, Securitization and Reconstruction Act). Furthermore, the policy will bring about greater special- ization in the process of transferring property and mortgage transactions that will lead ultimately to enhanced quality of services benefiting house buyers, housing loan borrowers, housing finance intermediaries, and the overall housing finance system. Protection by way of title indemnity against loss resulting from defects in title helps both primary and secondary residential mortgage markets function more smoothly. With title guarantees, the primary mortgage lenders would be induced to extend credit to additional potential house buyers under more permissive financing 185 Appendix C C terms, thereby expanding the house-buying market. The product will immensely benefit all parties to and stakeholders in the overall housing finance system. Presently, India has no entity that provides indemnity for its opinion on property titles. The concept is widely accepted in the developed markets and would be a valu- able tool for enhancing the credit quality in India's growing housing finance market. Foreclosure and Eviction Until 2003, India was following the court route for foreclosure on a property from a loan defaulter.21 The courtroom procedures sometimes took as long as 20 years. In 2003, an enabling law was passed to facilitate out-of-court settlement--the 2002 Securitization and Reconstruction Act. One of the objectives of the law was to remove recoveries from the ambit of the courts. A creditor is now allowed to acquire the mortgaged property if a defaulter fails to pay within 60 days of being informed of possible foreclosure/auction. Appeals can be made only to a specialized Debt Recovery Tribunal within 45 days. After the enforcement of the new foreclosure law, the time needed to enforce a security has decreased from as many as 20 years to 7 years, and the amount recovered has increased from 11.6 percent of the claim to 15.1 percent. Doing Business 2010 (World Bank 2009a) ranks India at 138th place worldwide, far after China (65th) and slightly past Brazil (131st). Although India's bankruptcy laws are national, there are differences in the func- tioning of the pertinent judicial bodies throughout the states. Across the 17 Indian cities, the insolvency process takes an average of 7.9 years, costs 8.6 percent of the estate value, and recovers 13.7 cents on the dollar. The process is slower than in South Asian countries (an average of 5.0 years), East Asia and the Pacific (2.7 years), and the countries of the Organisation for Economic Co-operation and Development (1.7 years). India is also less friendly to lenders than are neighboring South Asian countries, where creditors can expect to recover an average of 19.9 cents on the dollar. In the economies of the Organisation for Economic Co-operation and Development, the recovery rate is almost five times higher, at 68.6 cents on the dol- lar. There are variations in the speed and cost of the insolvency process and in the recovery rates across Indian cities. It is easier to go through bankruptcy in Hyderabad and more difficult in Calcutta. Hyderabad has the highest recovery rate (15.9 cents on the dollar), and Calcutta has the lowest (9.1 cents). It costs between 7 and 10 per- cent of the estate value to go through the debt-recovery process, mostly because of attorney fees, newspaper publication costs, liquidator fees, and preservation costs. India's legal and regulatory framework in the area of bankruptcy has advanced in the last two decades. Among the various channels of recovery available to financial insti- tutions, the Debt Recovery Tribunals (DRTs) and the Securitization and Reconstruc- tion Act have been the most effective in terms of amounts recovered. The government introduced DRTs to speed up the resolution of debt-recovery claims due to banks and financial institutions in 1993. The Securitization and Recon- struction Act was enacted in 2002 and amended in 2004, with a view to making ade- quate provisions for the recovery of loans and to enforce security interests without 186 Appendix C C the intervention of the court--or at least with minimum intervention. Cases under the Securitization and Reconstruction Act are resolved faster than under the traditional Debt Recovery Act, and give the debtor a very limited scope of defenses. When the bank sends the notice taking possession of the asset under the Securiti- zation and Reconstruction Act, two things may happen: the debtor may agree to the seizure of the secured asset, or the debtor may file an objection or appeal with the tribunal. This appeal starts the procedure through which the debtor will try to prove that the creditor's taking possession of the asset is not justified or does not respect the terms of the agreement. Therefore, tribunals currently handle cases both under their original jurisdiction contemplated in the 1993 Debt Recovery Act and under their jurisdiction according to the Securitization and Reconstruction Act. Debt Recovery Appellate Tribunals are the appellate courts that review decisions made by the DRTs. These positive changes in the law have taken full effect in practice only in the past couple of years. Until recently, the validity of the Securitization and Reconstruction Act--and before that, the validity of the DRTs--was held up in court challenges. Now, however, both acts are used by creditors. In the past few years, case manage- ment at the DRTs also has advanced. Each tribunal in Chennai has a Web site provid- ing online information about the tribunal, its composition, and a list of upcoming cases. The tribunals of Mumbai have started working faster because more judges were assigned recently. Additionally, increasing use of the Securitization and Recon- struction Act has reduced resort to the courts, particularly now that most of the issues generated during the period of transition to this new law have been resolved, eliminating the jurisdiction disputes that arose with its enactment. The Code of Civil Procedure of 1908, last amended in 2002, is the applicable rule across all 17 Indian states benchmarked.22 Based on the Doing Business 2010 case study, it takes 46 procedural steps to enforce a contract in India, but only 21 in Singapore, 24 in Hong Kong, China (the highest-ranking economy for enforc- ing contracts), and 31 in the Organisation for Economic Co-operation and Development economies (World Bank 2009a). Some people would argue that more-formal procedures in dispute resolution ensure that due process is followed and justice is done. The evidence suggests otherwise. The more complex the procedure for resolving disputes, the less likely firms are to report that judges are impartial and court decisions fair. Resolving commercial disputes through the courts is time consuming in India-- across the 17 cities, it takes an average of 961 days to enforce a contract, faster than elsewhere in South Asia but slower than international best practices and such coun- tries as Brazil and Indonesia. Overall court costs and attorney fees across India add up to an average of 26.6 percent of the value of the claim, similar to South Asia's aver- age of 27.2 percent and much below the East Asia and Pacific average of 48.4 percent. Within India, it is easier to enforce a contract in Guwahati, Hyderabad, Patna, and (600 days) and more difficult in Ahmedabad (1,295 days) and Mumbai (1,420 days) (figure C.16). The ease of enforcing contracts also reflects the differences in volume and backlog of court cases. Two reasons it takes a long time to go through the courts 187 Appendix C C Figure C.16 Lengthy Delays in Judgment across India Guwahati Kochi Bhubaneshwar Hyderabad Patna Ludhiana Chennai New Delhi state Noida Ranchi Indore Jaipur Bangalore Gurgaon Calcutta Ahmedabad Mumbai 0 365 730 1,095 1,460 time (days) filing judgment enforcement Source: World Bank 2008a. across all states in India are the limited number of judges and the high incidence of unfilled vacancies, both of which create backlogs of pending cases. India's population- to-judge ratio is approximately 14 judges per 1 million people (by comparison, the ratio in the United Kingdom is 51 per 1 million). Nearly 4,000 cases are being handled by each judge. Even if the disposal rate at the courts increases, as seen in Ahmedabad and across other courts in India over the past few years, there are still a lot of pending and newly filed cases, adding to the total time it takes for a case to be resolved. Indian courts are increasingly taking steps to appoint more judges. For example, the High Court in Ranchi increased the number of judges in 2008 and 2009 by an additional five. The lengthiest delays occur during the litigation period. Under Order 17 Rule 1 of the Code of Civil Procedure, the court may give more time to the parties and may adjourn the hearing up to three times at any stage of the suit. In practice, judges often grant adjournments and litigants use dilatory tactics to slow 188 Appendix C C Figure C.17 Cost of Contract Enforcement across India Patna Hyderabad Jaipur Ludhiana Guwahati Noida Bhubaneshwar Chennai state Indore Calcutta Kochi Ranchi Ahmedabad Gurgaon Bangalore New Delhi Mumbai 0 10 20 30 40 percent of claim attorney fees court fees enforcement fees Source: World Bank 2008a. proceedings. These add to the long litigation period, showing the need to balance the power of judges and the responsibilities of litigants and their lawyers. Contract enforcement costs least in Patna (16.9 percent of the value of the claim) and Hyderabad (17.7 percent), and most in New Delhi (33.7 percent) and Mumbai (39.5 percent)--both large cities with high attorney fees (figure C.17). The main dif- ference in the court cost is attributed to the different ad valorem fees deposited at the time of filing the suit. These fees are regulated by the respective Court Fees Act of each state and calculated based on the value of the claim. At an average of 6 percent of the claim value, the ad valorem fee is the second-highest cost after attorney fees (on average, 16 percent of the claim value). Although court rules also provide for fee structures for attorneys, they are not mandatory; in practice, attorneys charge clients according to their own fee structure and expertise. Increasing competition among lawyers reduces attorney fees. Enforcement costs, which account for an average of 2.6 percent of the value of the claim, are similar across states. 189 Appendix C C In past years, India's judiciary has taken steps to improve case management with the use of information and communication technology. In 2008, the Supreme Court of India allowed for electronic filing of cases. E-filing systems are being planned for the various high courts in the near future and eventually for the district courts as well. Allowing lawyers to file and submit documents online requires a reliable net- work infrastructure and authentication framework. Indiancourts.nic.in is a Web site that provides single-point access to information related to the Supreme Court and all 21 high courts. Litigants can browse through judgments, case status, rules, and the list of judges of each court. Further reforms could be focused beneficially on setting up commercial divisions in existing courts, implementing case management sys- tems, strengthening judicial capacity, and advancing alternative dispute resolution mechanisms. Indian courts could benefit from the proposed establishment of com- mercial divisions within the high courts. In seven African countries that introduced commercial courts or sections in the past five years--Burkina Faso, the Democratic Republic of Congo, Ghana, Mauritania, Mozambique, Nigeria, and Rwanda--the average time to resolve a commercial dispute dropped by about 19 percent. Strength- ening administrative control and information technology over the litigation process and court operation will better equip judges and court staff to address backlogs and delays. In 2008, the former Yugoslav Republic of Macedonia continued to systemat- ically equip courts with electronic case-management systems, speeding contract enforcement in commercial matters. Tonga's newly appointed chief justice intro- duced case management, transferring 90 percent of all cases--criminal, commercial, and land--from paper to computer and cutting the time to enforce contracts from 510 days to 350. Statistical information helps assess court performance. Keeping statistics on the number of commercial cases being handled and disposed of, the clearance rate on the time to resolve cases, and the appeals rate allows an analysis of reform and helps guide future policies of judicial reform. This effort can also assist in optimal case assignment and promotion criteria for judges. Similar practices already exist in India. Indore has a system in place whereby every judge at district court and high court levels is given a mark for each disposed case, and those marks are taken into consideration for promotions. Because of the slow judicial process, there has been a big focus on noncourt dis- pute resolution in India, especially through the Lok Adalat system, constituted under the National Legal Services Authority Act of 1987. Although India has a strong alter- native dispute resolution tradition, its impact on dispute resolution is minimal. Strengthening its mechanisms can reduce court burdens and create cheaper and speedier options for resolving disputes. Credit Information In the past few years, India has considerably improved the environment for getting credit, bringing it to 30th place in the worldwide ranking created by Doing Business 2010 (figure C.18) (World Bank 2009a). The Credit Information Bureau of 190 Appendix C C Figure C.18 International Comparison of Credit Information Availability, Selected Countries, 2009 140 120 113 rank among 180 economies 100 87 87 80 61 61 60 40 30 20 15 1 0 Malaysia Japan India China Mexico Brazil Russian Indonesia Federation country Source: World Bank 2009a. India Ltd., a private partnership among several commercial banks and credit infor- mation service providers, has started to increase the amount of credit information available in the country. It has drastically increased coverage from roughly 20 million in 2004 and 44 million in 2005 to more than 78 million individual borrowers (10.2 percent of adults) and more than 1 million firms. The data collected include both positive and negative details, and go back historically over several years. The registry, however, does not distribute information from retailers, trade creditors, or utility companies. The quality of information could be enhanced by giving borrow- ers the legal right to inspect and correct their data. Policy Options for Improving the Framework for Housing Finance The housing and housing finance markets in India have seen remarkable progress in the recent past, but various challenges remain. The sectors are vibrant enough and have substantial growth potential. With a modicum of enabling activities, housing will become accessible to a much wider share of the population. Some promising policy directions are outlined below. Efficient System of Land Administration An effective housing market depends on efficient and reliable property transac- tions.23 Buyers and sellers need to be able to verify and confirm their rights to title 191 Appendix C C with the appropriate third-party authorities. Inordinate delays in verifying owner- ship, cases of fraud, and failure to secure legal redress undermine the confidence required for a vibrant housing market. To achieve this, India's land administration and titling system requires some timely initiatives, including computerization, better coordination and improved efficiency of agencies with land administration jurisdic- tion, and reductions in duties and fees. The validity and security of titles must be upheld as well. A title registry should be developed, and issues related to title trans- fer should be streamlined. Effective land policies will also increase the availability of serviced land and improve housing affordability. A strong real estate regulator, cur- rently absent, would serve as a single window for overseeing the affordable housing agenda, would promote real estate reforms, would ensure transparency, and would improve consumer protection. What follows are descriptions of specific actions to enhance the effectiveness and efficiency of the system. Improve Subregistry and Transfer Office Efficiency, and Consider a Central Land Registry. Improving the efficiency of the subregistry and land transfer offices would dramatically reduce registration time in all 17 cities. It would also reduce costs by encouraging entrepreneurs to conduct the procedures themselves, not through lawyers. To do so, local government could further computerize land records and train personnel to use the new system. Such reforms, under different names, have already taken place in a number of states:, such the Property Registration Informa- tion System Module (PRISM), initiated in 1997 in Punjab; the Computer-Aided Administration of Registration Department (CARD) program in Hyderabad; and the Karnataka Valuation and e-Registration Initiative (KAVERI) in Bangalore, set up in 2002. Local governments could also improve the transparency of the process by explaining the different steps on their Web sites and publishing fee schedules. The workflow at the offices could be improved to avoid queues and bottlenecks. In addi- tion, local governments could introduce statutory limits capping the time taken by the offices to register a deed and transfer the property title. So far, Calcutta, Gurgaon, and New Delhi have drafted citizen charters that set clear time objectives for title transfer; but the time taken in practice still far exceeds these objectives. To truly reduce time, a citizen charter needs to be accompanied by better enforcement and sanctions imposed if deadlines are not respected. Increasing the number of offices-- especially in cities where the real estate market is buoyant--and fast-track proce- dures could improve efficiency. For a higher fee, entrepreneurs would receive the option to have their application processed more quickly. Cases would be prioritized based on a set of transparent criteria. This system is already offered in a number of cities around the world, such as Buenos Aires (Argentina), Kiev (Ukraine), and Ulan Bator (Mongolia). All cities require the same main procedures to register property: check for encumbrances, draft a sale deed, register the sale deed, and transfer the property title. Nonetheless, entrepreneurs in every city must undergo one or two additional proce- dures, such as checking for secondary nonencumbrance claims, obtaining a tax clearance, or paying stamp duty separately. These procedures could be eliminated or 192 Appendix C C combined with other procedures conducted at the subregistry office. Secondary encumbrance claims should be checked along with primary ones, tax clearance should be an integral part of the verification process preceding registration of the final sale deed, and franking services should be offered at the subregistry office to spare entrepreneurs the inconvenience of buying stamp paper separately. Since the Telgi stamp paper scam in 2003, most cities offer franking services. New Delhi even introduced an e-stamping facility that enables entrepreneurs to pay stamp duty online by credit card. However, many businesses still resort to traditional stamp paper­based methods because they are not aware such facilities exist. Publicizing franking services would help reforms truly affect the number of procedures and time required for registration. Reduce Stamp-Duty Rates and Explore the Scope for Replacing Them with a Land Tax. There is little doubt that the high rates of stamp duty currently assessed when registering land transfers push people into informality and reduce government rev- enue. Lowering these rates, which are very high by international standards, is likely to be necessary to ensure the sustainability of any improvements made in land administration, and would encourage entrepreneurs to formally register their prop- erty. The rise in registrations could offset, if not exceed, the financial loss caused by the tax cut. In July 2004, after reducing the stamp duty from 10 percent to 5 percent, the state of Maharashtra saw its stamp-duty revenues jump by 20 percent. Munici- palities could also eliminate surcharges and registration fees. They impose an addi- tional burden on entrepreneurs wishing to register. Egypt successfully implemented a reduction in property registration fees: In 2006, registering a property in Cairo could cost as much as 5.9 percent of the property value. Ninety percent of properties were either not registered or undervalued. The Egyptian government reduced regis- tration fees from a complex formula of a fixed fee of 3 percent of the property value plus additional charges to a single, low, fixed fee, abolishing all other costs. It cost merely 1 percent of the property value to transfer property in 2008. The reform paid off because property registrations and concomitant revenues rose in the six months following the reform. Expand Computerization, Provide Statewide Spatial Coverage, and Improve Textual and Spatial Records. Even though some states have successfully computerized textual records, progress is uneven across states. Government efforts to accelerate this process could be improved by (1) clarifying the overall policy and using it to establish clear cri- teria and accountability mechanisms for allocation of central funds; (2) publicizing best-practice approaches on the technical and legal or regulatory side and promoting exchange and communication among technical staff across states; and (3) putting greater emphasis on full functional integration between records and registry, together with a systematic process to establish consistency and eliminate any ambiguities in the legal status of specific land parcels. Large amounts of money have been, and continue to be, spent on surveying pilot projects with ill-defined objectives, the results of which are rarely subject to a 193 Appendix C C rigorous evaluation. Drawing on India's capabilities to combine satellite imagery with existing village maps and other readily available spatial products to generate a basic cadastral index map would be a lower-cost option to provide a comprehensive framework, identify gaps, and establish criteria for ways to address spatial data prob- lems in an affordable manner. Pilot projects that can be scaled up quickly should be encouraged (possibly by subcontracting to the private sector) to improve the textual and spatial records. Target areas for improvement could include unrecorded subdi- visions, inconsistencies across records, decay and loss of maps, change of land use patterns, and unsettled lands. Private sector participation could be allowed in sur- veying, focusing government on a regulatory role. This will free up public resources, and is strongly consistent with international best practices and India's own experi- ence in computerizing textual records. Eliminate Restrictions on Land Markets. All over the world, land rental markets allow rural dwellers to join the rural nonfarm economy in a way that provides those who stay back with access to additional productive resources. Indian evidence shows that rental restrictions reduce equity as well as efficiency. It will thus be desirable to (1) make leasing legal where it is currently prohibited, and replace rent ceilings with regulations to facilitate rather than constrain rental markets; (2) allow transfer of land by land reform beneficiaries at least through lease, and explore options for making permanent the gains from such reform; (3) drop restrictions on sale of land to nonagriculturalists and on subdivision that has little economic justification; and (4) review legislation on compulsory land acquisition and, subject to the prevention of undesirable externalities, allow farmers or their representatives to negotiate with (and if desired, transfer land directly to) investors rather than forcing them to go through government and often receive only very limited compensation. Collateral/Lien Registry and Credit Information Bureau The following specific actions are suggested to improve information availability in the housing and housing finance markets. Strengthen the Functioning of DRTs and Create a National Companies Law Tribunal. The DRTs presently are handling only cases of financial institutions. The National Companies Law Tribunal constitutes a single forum for dealing with the rehabilitation of all types of companies, their liquidation, and the winding up of their operation. If established, this tribunal will ensure uniform and standardized insolvency and liquidation procedures for nonfinancial institutions. Even though it takes less time to go through debt enforcement proceedings at the DRTs than in the ordinary courts, the process from the moment of loan default until recovery of the debt is time consuming and burdensome. The government has been committed to strengthening DRT functioning over the last three years, but problems remain. The number of DRTs across India increased from 29 in 2005 to 33 in 2008. Still, not all cities have DRTs; and as a result of staffing problems, not all DRTs have 194 Appendix C C presiding officers to hear cases. Those officers are civil court judges appointed for five-year terms. Where no appointment has been made or where an appointment is pending in a particular city, the presiding officers of nearby DRTs rotate. Newly appointed officers also have to spend time in training and familiarize themselves with the recovery proceedings. When fully in the job and at the request of the par- ties, they may have to rehear the cases of the preceding officer. This repetition creates case backlogs and slows down the recovery process. Enhance Creditor Rights. The proposed Companies Bill (2008), which amends the current Companies Act, introduces many important reforms for India's insolvency framework. The government should focus on passing the proposed legislation in this area to deal with insolvency more efficiently. The Companies Bill (2008) contem- plates forming a creditors' committee, similar to other common law jurisdictions, and thus expanding credit rights (particularly to unsecured creditors). This will improve the liquidation and revival process within India. It also has been the inter- national trend in the past five years, with France and the Republic of Korea allowing the creditors' committee to vote on reorganization plans as well. Build Capacity. The government also has been working toward improving the legal and regulatory framework in the broader area of bankruptcy. Initiatives to improve the insolvency framework include guidelines for one-time settlement of outstanding dues. India could improve its insolvency and debt recovery practice by building insti- tutional capacity and enforcement among the key players in the regime--insolvency practitioners, judges, liquidators, creditors, and appraisers. Establishing public or private benchmarks for insolvency practitioners will strengthen the country's capac- ity to deal with insolvency cases. In India, the administrators are being chosen from a panel consisting of advocates, company secretaries, chartered accountants, and such other professionals as the government may deem fit. There is no requirement, however, that these professionals have any demonstrated experience or capacity in dealing with insolvent entities. The Indian Institute of Insolvency Practitioners is a private sector body might carry out the functions of an insolvency regulator, includ- ing licensing, training, and disciplining. The government could empower the insti- tute to facilitate the development of a licensing and training scheme in this area. Poland's new Law on Trustee Licensing tightened professional requirements for administrators to ensure they have the skills and education needed to oversee bank- ruptcy proceedings. Obtaining a trustee's license there now requires passing an exam covering economics, law, finance, and management. The reform also limits trustees' pay to 3 percent of the bankrupt estate's value (down from 5 percent). Strengthen the Credit Bureau and Create a Mortgage Repository. The Credit Infor- mation Bureau of India Ltd. could extend its coverage to include information from retailers, trade creditors, utility companies, and nonfinancial institutions generally. The quality of information can be further enhanced by giving borrowers the legal right to inspect and correct their data. The financial sector would be encouraged to 195 Appendix C C increase its reliance on data from the credit bureau, especially regarding housing finance. The Indian government is considering the proposed introduction of a mort- gage repository, which would improve the functioning of housing finance markets by easing a lender's risk as to loan security. Unified Record of Property and Housing Data A housing databank is designed to provide periodic property and housing data, including data on housing supply and demand, housing finance, and construction. Specifically, · housing supply data--new construction and sales, vacancies, quality and condi- tion of property, prices, and dwelling characteristics; · housing demand data--demographic trends and housing needs, characteristics of occupants of the existing stock (crowding indicators), affordability (availability of income indicators is of paramount importance for this), and buyer's intentions; · housing finance data--sensitivity of monetary indicators to residential and com- mercial property exposures of lenders and investors (inflation rates, interest rates), fiscal indicators and degree of reliance on housing-related developments (property taxes, construction industry contribution to GDP and employment), lending volumes by category of financial institution, loan characteristics (loan- to-value ratios, debt service ratios, prices, locations, borrowers, and so forth); and · construction inputs--production capacity; price of materials, labor, and land; availability and quality of materials, labor, and land. National Housing Bank, HFC Supervision and Regulation, and Housing Development Policies To strengthen the NHB's supervisory and developmental roles, the bank needs a strategic plan to foster market deepening, as well as a clearly defined role for the bank in the development of new market products and instruments. The NHB's institu- tional capacity requires scaling-up, including its capacity in licensing new market entrants and regulating the market in general. This could be achieved by sharing best practices and experiences from other countries; and highlighting the benefits and applicability of new instruments, such as credit risk mitigants (guarantee schemes for informal sector borrowers, mortgage payment protection systems), interest rate risk mitigants (variable-rate mortgages), standards and procedures, and the like. The bank would be the clearinghouse for these tools and instruments. A sustainable approach to enhanced lending operations should also be developed for the low- income segment in economically weak sections. The following items take priority: · Coordination improvement--This includes coordination and information exchange between home supervisors and the NHB, as well as a formalization of the relationship with foreign regulators. 196 Appendix C C · Clear definition of a housing finance company or housing finance institution in the National Housing Bank Act--Permissible activities should be delineated clearly so that, for example, builders and construction companies are not permitted to use the term "housing finance" in their names. The Ministry of Corporate Affairs needs to issue necessary guidelines to registrars of companies in this regard. · Guidelines on corporate governance, including fit and proper provisions--Boards and senior management should have clear responsibilities to ensure effective con- trol over an HFC's entire business. The NHB should be empowered to enforce such provisions effectively. · Definition of "significant ownership" or "controlling interest" in the case of HFCs-- The NHB must lay down norms for large exposure limits, exposure to related parties, and provisioning for off­balance sheet items; and must issue necessary guidelines to HFCs. · Monitoring of related-party exposures and conflicts of interest--The NHB needs to take steps to mitigate the risks arising from exposure to related parties, including strengthening its ability to monitor and review such information and issuing suit- able guidelines to mitigate these risks. In terms of risk management, these are key priorities: · NHB should issue guidelines on market risk, along the lines of commercial banks for HFCs.24 · NHB should issue detailed liquidity risk management guidelines, in addition to the existing asset-liability management guidelines for larger HFCs. The NHB could also consider issuing management of operational risk guidelines to HFCs. · NHB's capacity to monitor the HFCs' internal control and audit could be strengthened. · Know-your-customer norms for HFCs should include provisions for adequate HFC capacity to monitor its compliance risks, adequate screening policies on staff ethical and professional standards, and whistleblower policies. · Norms for HFC disclosure to the NHB should be created. HFCs should notify the NHB of any substantive changes in their activities, structure, and overall condi- tion, including material adverse developments and any breach of legal or pru- dential requirements. The NHB should be empowered regarding the rejection and rescinding of external auditors, if needed. HFCs should submit consolidated financial statements and consolidated prudential returns, and the NHB should be able to conduct consolidated supervision. Mortgage Product Innovation The Indian housing finance market could benefit from further market deepening and broadening. The "Long-Term Funding" section below presents mechanisms to 197 Appendix C C enhance long-term funding available for housing finance, including the develop- ment of secondary financing markets. The sections above provided priorities in cre- ating an enabling regulatory framework. This section focuses on promising housing finance mechanisms and products to capture a significant scaling-up in the housing market, reaching out to the middle class in a "leaps-and-bounds" fashion. The next section focuses on expanding the market to lower-income clients. Fund Support for the Market through the NHB. The NHB could channel market (nondistortive) funding to financial institutions when lending to lower-middle-class clients (creditworthy borrowers with monthly incomes below Rs 12,000). The fund- ing should be neutral with respect to primary lenders, who would have equal access to it; and will not involve any crowding-out through below-market interest rates. Also, there should be underwriting norms designed to ensure adequate lending dis- cipline as well as compliance with the intended purpose of a loan. Assuming that the funding cost would be more attractive to primary lenders than existing refinance loans by the NHB, one possibility is to tie this resource to an equivalent amount of loans for the target population or to the achievement of a minimum volume of activ- ity in this market segment. The borrowing conditions for the end borrowers should remain market based--improving the affordability of finance should be left to the government subsidy policy. Enhancing affordability through interest rates can be distortive and counterproductive: given the high sensitivity of real estate prices to interest rates, this approach may fuel price increases, and money will end up in the developers' pockets instead of helping low-income buyers. A much better option would be to enhance the willingness of lenders to be active in this segment by allow- ing wider financial intermediation margins that entry cost and credit risk consider- ations can justify economically. This approach would strike a balance between activ- ity commitments and financial incentives. Other options would be to use part of the funding to set up a cap mechanism to mitigate the interest rate risk borne by low- income mortgagors, or to use the differential with regular funding cost to partly fund a guarantee scheme to mitigate the credit risk incurred by lenders. Each option is dis- cussed below. Create an Interest Risk Mitigation Fund. Mortgage loans in India carry adjustable interest rates. The targeted population does not have the same capacity as wealthier borrowers to absorb interest rate increases during the life of the loans. In the absence of a mature and efficient interest rate swap market, the market could use a mechanism that protects borrowers against adverse market develop- ments. The shaping of the mechanism, and its calibration, would require an expert design. Such mechanisms exist in other countries (such as Colombia).25 An inter- est risk mitigation scheme should be seen as a transitional device until the Indian market provides efficient hedging instruments, and should be calibrated as such-- for instance, based on an estimated volume of low-income mortgages over a period of two or three years. 198 Appendix C C Establish a Financial Guarantee Facility. A financial guarantee facility could be used to mitigate the credit risk incurred by lenders. The guarantee fund could be set up to enhance the credit of low-income borrowers--mostly from the unorganized sector. Its mechanism should rely on risk-based premiums to ensure its sustainabil- ity. Also, setting it up should be coordinated with mechanisms that are being designed or are under consideration in similar areas: mortgage insurance (targeting high loan-to-value lending operations) and guarantee mechanisms to facilitate the financing of slum upgrading in particular. Build Capacity. Innovations and rapid scaling-up of housing finance markets require active capacity-building efforts for financial institutions, developers, regula- tors, and other housing market participants. This capacity building could include the following actions: · standardizing and disseminating operational processes for extending and servic- ing small mortgages; · developing new financial products and delivery mechanisms; · strengthening the ability of microfinance institutions to act as intermediaries between low-income groups, particularly slum dwellers, and mortgage lenders: demand aggregation, risk assessment, borrower qualification, loans servicing, and so forth; · developing a consumer information and protection framework suited to the spe- cific needs of targeted beneficiaries; and · defining business standards for secured lending to small or emergent developers. Low-Income (Microfinance) Housing Lending Pure market-based solutions are likely to work for many customers in this segment: as project findings demonstrate, affordability is not an issue for most of these customers; and there is high interest among stakeholders, especially in an employer-facilitated model (that is, a model in which the employer provides access to employees, administrative support in terms of documentation, and, most important, payroll deduction for EMI payments). Develop Market Solutions for Formally Employed Households Earning Rs 5,000­8,000 a Month. The employer-facilitated model is a fundamental busi- ness innovation that makes low-income housing attractive to the private sector. This solution could help stimulate the housing market for low-income salaried sector households earning Rs 5,000­8,000 per month (5­6 million households), and for the 4­5 million households in the Rs 8,000­11,000 range. The model both addresses the main risks for all the key players and enhances the economics so that the returns are attractive. For a financial institution, getting a pool of customers 199 Appendix C C from a large organization that facilitates the process and provides payroll deduc- tion is attractive because it reduces both risk and costs. To a small developer, the model improves market returns and provides a pool of "prefinanced" customers, thus reducing market risk. Policy initiatives may be very effective in helping scale up. Such measures could include increasing the supply of affordable land and encouraging a broader set of developers to serve this market. Additional initiatives that could enhance affordability include standardizing norms to lower construc- tion costs using alternate materials and techniques, and such incentives as lower stamp duties. Develop Market Solutions for Informally Employed Households Earning More Than Rs 5,000 a Month. Alternative models are required for the informal sector (self-employed and salaried unorganized) households earning greater than Rs 5,000 a month. Those households remain viable customers, but they are associated with greater credit risks for mortgage lenders. As a result, viable business models include using MFIs as intermediaries, using NBFCs to serve their premium customers, and using a direct-to-customer lease-to-purchase vehicle. The products include loan products (savings-linked loans, flexible EMI loans) and risk-sharing solutions (credit guarantees, mortgage insurance). The team has also identified "forward- thinking" financial institutions that are willing to pilot-test these models and prod- ucts. Intermediaries like MFIs could be used to play an agency role for mortgage lenders. In addition to being low cost, they can help reduce risk by providing emer- gency loans and consumer education. NBFCs could be used to target their premium customers, offering additional loans using the home as the underlying asset (that is, home equity loans). Market solutions for informal customers would involve a series of new products, including these: · A modified group liability product, through which the group provides a "rolling guarantee" for one year of EMI collections. After each year and with the mortgage lender's approval, defaulting members can be dropped by the group. The defaulting members then are served by the financial institution directly, at a higher interest rate. In other words, being part of the group gets the client a lower interest rate. The group may be maintained for five years, after which the client has built more equity and has established a track record of payments. · A savings-for-housing product, through which the customer deposits a certain percentage of his or her monthly income (for example, Rs 500 from an income of Rs 5,000) with a financial institution for a specific period (one to three years). If the customer can adhere to a regular pre-agreed deposit schedule, he or she is given a loan at the end of the period. The loan is typically a multiple of the total money deposited (for example, three times) plus an additional downpayment made by the customer. This product implies some initial regulatory difficulties, but it could be a very effective product over the medium term. 200 Appendix C C · A flexible product that provides flexibility in EMI payments for self-employed people who experience income fluctuations. It could include payment once a quarter (or minimum monthly payments ensuring that quarterly commitments are met). · A lease product, through which a customer pays an up-front deposit (equal to the downpayment) and a monthly rent (equal to the EMI) for a period equivalent to the loan tenor. More analysis is required to develop this product into a viable market alternative, though the potential value of such a product is considerable. This product could be used to go direct to the self-employed customer and really open up this market. · An EMI deduction via bank account would lower transaction costs. A variation on this involves the borrower putting some of the initial money (perhaps by paying a slightly smaller downpayment) into an escrow account so that if an EMI pay- ment is missed, the mortgage lender can use the money in escrow to settle one or two EMIs, while working with the customer to solve the repayment problem. · Mortgage insurance is a classic insurance product that would reduce the risk for lenders, thereby allowing them to serve the informal sector. · A credit guarantee, in the simple form of a partial guarantee with the first loss being borne by the lender. The level of first loss would have to be low (maybe con- sistent with the broader housing mortgage portfolio) to encourage lenders to par- ticipate in this market; but it would be good from a design perspective because it aligns incentives. The actual execution of the guarantee could be simple--when the portfolio losses cross the threshold, the guarantee fund would pay out the losses in proportion to the pre-agreed guarantee level.26 When the financial insti- tution is able to collect any funds from the customer, the same proportion as the pre-agreed guarantee level is redeposited in the guarantee fund. The advantage of this structure is that the guarantor only has to bring in funds when actually required and the redepositing of recoveries renews the fund. Develop Market Solutions for Households Earning Rs 2,500­5,000 a Month. Pure market-based solutions will not be enough to bridge the gap and serve this segment; and home ownership would need to be promoted through a combination of market-based levers, policy initiatives, and direct support from the government. Some of the more effective levers that could be used to bridge the affordability gap are these: · smaller housing units (for example, units with a built up area of 200 square feet); · an increased floor space index; · low-cost materials and building technologies; · exemption on stamp duty charged to the customer; and · up-front monetary contribution through lump-sum subsidy to partially fund EMI payments. 201 Appendix C C Although the above levers can be of use to bridge the affordability gap, an impor- tant practical barrier to serving this market is the potential lack of interest from financial institutions. They are concerned about the high transaction costs (as a result of small ticket size and disaggregated customer base) and the risk of default. Hence, there may be a need to supplement those levers with additional policy initia- tives, direct support, and a reasonably priced credit guarantee to encourage financial institutions to serve this segment. Here are some additional levers: · policy initiatives, such as income tax exemption on profit generated from low- income housing loans (for example, loans of less than Rs 300,000) to encourage credit flow to this segment; · support to financial institutions to cover their higher transaction costs for serving this segment; · reasonably priced credit guarantees providing risk-sharing support; · construction finance available to small and medium-size developers; · policy initiatives to increase the supply of affordable land and encourage devel- opers to serve the market; · norms standardized for construction using low-cost materials and techniques; and · initiatives to lower cost to the end consumer, such as reduced stamp duty. Furthermore, this income segment may require an element of financial support from the government. This support may have to be used both to affect the direct affordability of the dwelling unit and to cover the higher cost of service by mortgage lenders. Long-Term Funding Given that the demand for housing finance is so strong, in the current milieu, the greater challenge appears to be in raising low-cost, long-term resources. It has been a long-standing demand of housing finance players (especially housing finance com- panies) to be allowed access to pension funds. These funds occupy a prominent place in the financial savings of the household sector, are long-term in nature, and repre- sent long-term savings of the salaried segment of the population. Currently, how- ever, these funds follow rigid investment patterns set out by the government. Besides insurance companies and provident funds, there are no lenders with a long-term appetite. The NHB provides some long-term funds as a second-tier institution, but cannot support a significant scaling-up of the housing finance market to serve the middle class as a whole. Given that the primary mortgage market is now on a sure footing, more efforts have to be made to deepen the secondary mortgage market. Enabling regulation 202 Appendix C C could support the more active development of a secondary mortgage market. Fur- thermore, deepening the secondary market will enable more recycling of funds and will reduce capital adequacy requirements. The housing finance market has scope for mortgage-backed securitization development because it is a capital-deficient sector with a strongly growing primary market, with a considerable shift toward market- based resource mobilization. Securitization will match the needs of borrowers and investors, promote secondary market intermediation, and create a long-term debt market. This is also a unique opportunity for the housing finance sector to reposition in a changing financial environment. The National Housing Bank Act amendment in 2000 provided for securitization, as did the Securities Contracts (Regulation) Act, 1956 Amendment; the Insurance Regulatory and Development Authority and Securities and Exchange Board of India guidelines for investment; and the guidelines on securitization of standard assets issued by the RBI in February 2006. The lack of mortgage guarantee/insurance will be a hindrance and should be addressed. Enabled Rental Markets The rental market would be enabled by liberating it from the existing legislative and bureaucratic controls and activating the housing supply. The modification in the Rent Control Act, as adopted in Delhi and several states (enabling periodic rent revisions and getting the premises vacated when needed), would be an important step. Public agencies can play a role in rental housing for the poor in large cities. Cur- rently, a rental market of some sort exists for them, exploited by slumlords, blanket- renting merchants who provide a place to sleep or a blanket at a high monthly or daily rent. The night shelters built by the local governments or development author- ities have not made much of a dent in the problem because of their small scale and certain operational deficiencies. It would be important to build a large number of night shelters and family flats in different parts of the city, particularly in the prox- imity of business centers. Better use of the facilities and cost recovery can be ensured through better management. Creating this facility on a large scale would solve the housing problem for the floating population and recent migrants in the low-income brackets (at least for those above a certain income level). Industrial houses and voluntary organizations may be motivated to take up chains of such night shelters and family flats. Policies should be considered to make cities more friendly to rental housing-- especially for the poor. An enabling rental policy framework should acknowledge already existing rental arrangements and find flexible, realistic ways to regulate them. In most cities, rental housing arrangements contribute enormously to local liveli- hoods. Measures to govern and promote their further development should be for- mulated. Rental housing should be placed on the larger urban policy agenda. Policy makers can work with the local private and financial sectors, community organiza- tions, nongovernmental organizations, and civil society organizations to create a rental housing regulatory framework that promotes affordability and maximizes 203 Appendix C C shelter options for poor people. The framework should include means and incen- tives that encourage the production, exchange, and consumption of a variety of rental housing options. The framework should also include a system of checks and penalties to curb mismanagement or abusive practices within a city's rental housing market and prevent the deterioration of rental housing stock for lack of maintenance or repairs. This regulatory framework should be simple, practical, and easy to apply. For implementation and enforcement purposes, the policy could be integrated into a city's already-existing administrative procedures. Given the low-income settlement focus, the policy framework should include combining efforts to improve rental housing quality and basic infrastructural services. Two further focus areas for an enabling rental housing policy are the issue of mobilizing finance to improve and expand rental housing; and the task of encouraging large-scale rental housing invest- ments, as discussed elsewhere in this appendix. Notes 1. Unless otherwise indicated, the data in the "Overview" section come from the World Bank's World Development Indicators. 2. According to data from the Reserve Bank of India, reserves stood at $282 billion on October 9, 2009. 3. Data are taken from the National Building Organization's housing data tables, http://nbo.nic.in/ Images/PDF/housing_data_table.pdf. Specifically, see table 5, last column. 4. These estimates were made by the Technical Group on Housing for the 11th Plan Period, set up by India's Ministry of Housing and Urban Development and Ministry of Rural Development in fiscal 2005/06. 5. The Kundu Panel's report is expected in early 2010. 6. To launch RESIDEX, NHB (at the behest of the Ministry of Finance) undertook a pilot study to examine the feasibility of preparing such an index at the national level. The pilot study covered five cities--Bangalore, Bhopal, Calcutta, Delhi, and Mumbai. Additionally, a technical advisory group was constituted to deal with all the issues relating to methodology and data collection, and to guide the process of constructing an appropriate index. The groups was chaired by an adviser from the Ministry of Finance and comprised expert members form the RBI, the NSSO, the Cen- tral Statistics Office, the Labor Bureau, the NHB, and other market players. Actual transaction prices were used, so that the index reflects market trends. The base year 2001 was taken to make the study comparable with the wholesale and consumer price indexes. Year-to-year price move- ments during the period 2001­05 were captured in the study, which was subsequently updated for two more years (to 2007). RESIDEX has been expanded to cover Ahmedabad, Chennai, Faridabad, Hyderabad, Jaipur, Kochi, Lucknow, Patna, Pune, and Surat; and the base year was shifted from 2001 to 2007. In the first phase, RESIDEX will be expanded to cover 35 cities hav- ing million-plus populations. The proposal is to expand the index to 63 cities covered under the Jawaharlal Nehru National Urban Renewal Mission to make it a truly national index. To guide and oversee the construction of RESIDEX and the expansion of its coverage, a standing com- mittee of technical experts has been constituted under the chairmanship of the NHB, with rep- resentations from the government of India (Ministry of Finance, NSSO, Central Statistics Office, Labor Bureau), the RBI, and other prominent market players. 7. The TMC project caters to the very-low-income group and claims to have managed the cost by using Australian low-cost construction technology. The salient features of the TMC project are these: units ranging between 160 and 400 square feet; net project size of 8 million square feet; and a total of 25,000 units with an estimated total number of occupants under 100,000. The master planning was executed by Sinclair Knight Merz, Australia. The project boasts full-fledged 204 Appendix C C infrastructure, including schools, hospital, city center, and so forth; fully mechanized skywalk stretching for 2 kilometers across the colony. The estimated project delivery time is 18 months, and past output has been 10,000 square feet a day. 8. HFCs are not allowed to have checking and savings accounts. 9. Figures are based on a conversion rate of Rs 50 = $1. 10. Financing conditions are as follows: The developers' ratio of long-term debt to equity should not exceed 2:1, according to the latest annual audited accounts. There is a minimum outlay of 20 percent margin from the developer of a project. The loan should be secured by a mortgage of property or other security used as collateral. Minimum fixed asset coverage ratio should be 1.33:1, based on market value identified through an independent third-party valuation. Addi- tionally, the maximum selling price of each dwelling unit in the project to be financed may not exceed Rs 1.5 million, with a maximal built-in price escalation of 10 percent. The refinancing can cover up to 100 percent of the loan sanctioned by the eligible housing lender. Repayment is fixed on a case-by-case basis (depending on the project size), typically in installments with an initial moratorium of one year. Bullet repayment of refinance may be permitted in exceptional cases. 11. The actual/potential primary liquidity released was $85 billion. 12. Specifically, the cash reserve ratio was reduced from 9 percent (October 2008) to 5 percent (May 2009); the repo rate was brought down from 9.00 percent (October 2008) to 4.75 percent (May 2009); the reverse repo rate went down to 3.25 percent in the same period; the statutory liquidity reserve was reduced by 150 basis points; export credit refinance eligibility was enhanced to 50 percent of outstanding export; a special 14-day term repo was introduced; and a special refinance facility for banks was made available. Furthermore, special refinance facilities were put in place for the Small Industries Development Bank of India, the NHB, and the Export-Import Bank; a foreign exchange swap facility was made available to banks; the all-in-cost ceiling for external commercial borrowings was raised; the systemically important, nondeposit-taking, nonbanking financial companies were permitted to raise short-term foreign currency borrowings; a provision for liquidity was made for systemically important, nondeposit-taking, nonbanking financial companies by special-purpose vehicle; and nonresident deposits were allowed higher interest rates. Finally, the risk-weight and provisioning requirements were relaxed. 13. The current mortgage rate for the retail customer is 12 percent. Key elements include a complex of four to six buildings (four or five stories high, single staircase, no elevator), fenced by a com- pound wall with open spaces including a garden and playground. Regular water and electricity services are provided. The individual apartments would vary from 200 to 350 square feet, with the smaller ones being studios (a room with an attached toilet, bathroom, and a kitchen) and the larger ones having an additional room. The complex is located in a vibrant neighborhood, with convenient access to public transportation, schools, health care facilities, and a market. Typically, these locations are farther away from customers' workplaces than their current residences, but are within an hour's distance. 14. Savings range from 10 to 20 percent for households with family incomes of Rs 6,000­8,000, with the higher savings rate occurring in smaller towns. 15. The salaried organized sector includes customers working in firms with more than 100 employees. 16. The salaried unorganized sector includes customers working in firms with fewer than 100 employees. 17. For contract employees, there might be less willingness to offer payroll deduction because most employers do not engage directly with contract employees. All interactions (including payroll) typically are routed through contractors, and contract employees usually are rotated across com- panies every six months. 18. Consumer education has been shown to be one of the most effective techniques for reducing defaults. 19. This section is based on World Bank (2007a, 2008b). 20. Some of these decisions have come under debate in the discussion of the Forest Rights Bill. 21. This section is based on World Bank (2008b, 2009b). 205 Appendix C C 22. This section is based on World Bank (2008b). 23. This section is based on World Bank (2007a, 2008b). 24. These guidelines can be issued in a phased manner. At the outset, surrogate risk weights can be stipulated for instruments susceptible to market risk. In the second stage, the assets can be seg- regated into banking book and trading book. Capital charge on market risk for items in the trad- ing book may be considered. 25. In Colombia, a funded mechanism compensates mortgage lenders against mismatches between the conditions of mortgages (fixed rate in real terms) and the conditions of their funding base (variable nominal rates). Lenders choose a maximum threshold for their cost of funding, above which the protection is activated. The cost of the protection--incorporated in premiums that feed the mechanism--depends on the level of the threshold. 26. On an individual loan, the loan would be considered a "loss" at a predetermined stage--say 90 or 180 days of no payment--and the amount of the loss would be the unpaid interest and principal. References Center for Urban Development Studies, Graduate School of Design, Harvard University. 2000. "Housing Microfinance Initiatives--Synthesis and Regional Summary: Asia, Latin America and Sub-Saharan Africa with Selected Case Studies." Cambridge, MA. http://www .microlinks.org/ev_en.php?ID=7454_201&ID2=DO_TOPIC. CRISIL (Credit Rating Information Service of India, Limited). 2006. "Demand Fundamentals Support Mortgage Asset Quality." CRISIL Insight January. D'Souza, Conrad. 2009. "The Housing Finance Market in India." HDFC paper presented at the Workshop on Housing Finance in South Asia, Jakarta, Indonesia, May 27­29. Fitch Ratings. 2009. "Bank Systemic Risk Report." New York. Garg, Y. K. 1998. "India Case Study." Paper presented at the Irving Fisher Committee Confer- ence on New Directions in Asian Housing Finance, Bali, Indonesia, February 5­6. India, Ministry of Housing and Urban Poverty Alleviation. 2007. `National Urban Housing and Habitat Policy­2007." http://mhupa.gov.in/policies/duepa/HousingPolicy2007.pdf. India, Ministry of Statistics and Program Implementation, National Sample Survey Organiza- tion. 2004. "Housing Condition in India--Housing Stock and Constructions: NSS 558th Round (July­December 2002)." Report 488(58/1.2/1), New Delhi, India. India, Ministry of Urban Employment and Poverty Alleviation. 2005. "National Urban Hous- ing and Habitat Policy­2005." Draft. New Delhi, India. http://mhupa.gov.in/policies/ duepa/DraftNHHP2005-9.pdf. India, Office of the Registrar General and Census Commissioner. 2001. "Census of India 2001." New Delhi, India. http://censusindia.gov.in/. Karnad, Renu. 2008. Housing Finance Mechanisms in India. Nairobi, Kenya: United Nations Human Settlement Program. Monitor Group. 2007. "Expanding the Housing Finance Market to Cover Lower Middle Income Segments in India." Unpublished consultant report prepared for the World Bank and the Financial Sector Reform and Strengthening Initiative, Cambridge, MA. ------. 2009. "Market-Based Experiences in Low-Income Housing in India." Presentation to the Workshop on Housing Finance in South Asia, Jakarta, Indonesia, May 27­29. Narayanswamy, Subbu, and Adil Zainulbhai. 2007. "India's Consumer Evolution." McKinsey Global Institute, Mumbai, India. http://www.mckinsey.com/mgi/mginews/india consumerevolution.asp. 206 Appendix C C NHB (National Housing Bank). 2004. Report on Trend and Progress of Housing in India 2004. New Delhi, India. ------. 2006. Report on Trend and Progress of Housing in India 2006. New Delhi, India. http://nhb.org.in/Publications/Trends_Progress_Report2006.pdf. ------. 2008. Report on Trend and progress of Housing in India 2008. New Delhi, India. ------. 2009. "Low-Income Housing Finance." Presentation to the Workshop on Housing Finance in South Asia, Jakarta, Indonesia, May 27­29. RBI (Reserve Bank of India). 2009. India's Financial Sector: An Assessment--Committee on Financial Sector Assessment Reports, Volume 5. New Delhi, India: Cambridge University Press. World Bank. 2007a. Doing Business in South Asia 2007. South Asia Financial and Private Sector Department, Washington, DC. ------. 2007b. "India: Land Policies for Growth and Poverty Reduction." Agriculture and Rural Development Sector Unit, India Country, Management Unit, South Asia Region, Washington, DC. ------. 2008a. Doing Business 2009: Comparing Regulation in 181 Economies. Washington, DC. ------. 2008b. Doing Business in South Asia 2008. South Asia Financial and Private Sector Department, Washington, DC. ------. 2008c. "Financing Homes: Comparing Regulations in 42 Countries." Washington, DC. ------. 2009a. Doing Business 2010: Reforming Through Difficult Times. Washington, DC. http://www.doingbusiness.org/. ------. 2009b. Doing Business in South Asia 2009. South Asia Financial and Private Sector Department, Washington, DC. ------. 2009c. 2009 World Development Indicators. Washington, DC. 207 Appendix D 1 Pakistan Pakistan's housing finance-to-GDP ratio is below 1 percent. The ratio in developed countries is 50­70 percent, and 7 percent in India. In spite of active and robust financial sector reforms led by the State Bank of Pakistan [SBP] in the recent decade, the unwieldy land administration environment, unprecedented rises in land prices, and inadequate mortgage lender experience with lower-income housing have prevented the market from advancing in the provision of housing and housing finance solutions. Pakistan is facing unprecedented challenges of acute housing shortages, unhealthy living conditions, and nonexistent or dilapidated infrastructure across the country. Housing shortages stood at an estimated 7.57 million units in 2009--6 million of which are concentrated in the lower-middle-income and low-income population groups, which the market is incapable of servicing currently. Housing stock is composed mostly of nonpermanent dwellings--about 40 percent of houses are semi-pucca (semiperma- nent) houses, generally without planned sanitation or a sewerage system. Thirty-nine percent are kutcha (temporary) houses with minimal water supply and either sanita- tion or drainage services. Only 21 percent of houses are modern brick constructions. Pri- vate developers are relatively dynamic, but have focused exclusively on the upper-income population, constrained by developer industry fragmentation, unorganized brokers, the lack of developer finance, an unstructured framework for property development, skewed tenant laws, poor master planning, problematic zoning restrictions, unreliable utility connections, and impractical building codes. There is little research on local, viable, low- cost housing solutions in the country. The quality of construction is grossly compromised by high prices for construction material. The secondary market is active, but nontrans- parent and speculative. Currently, there are 25 commercial banks offering mortgage finance in Pakistan. The state-owned House Building Finance Corporation (HBFC) is the only institution that caters to the financing needs of middle- and low-income groups in the country. Only 209 Appendix D D 1­2 percent of all housing transactions are processed through housing finance credit lines; the remaining, mostly lower-income, clients are funded by personal savings and informal borrowing. More than two thirds of its population falls below the mortgage affordability threshold of PRs 200,000. The microfinance market has a very limited out- reach in Pakistan, at barely 2 percent of the population. Microfinancing for housing is even more limited. Growth is brisk, however, starting from a small base. Combined efforts of nongovernmental organizations (NGOs) and the government have proved inadequate to address the low-income housing and housing finance gaps. The role of private banks has grown significantly over the years, and they now hold 46 percent of outstanding mortgages--though the HBFC still handles 77 percent of mort- gage clients by number. Mortgage terms are typical for the region, though unaffordable for lower-income groups. The industry performance has suffered from global conditions, with a rising nonperforming portfolio above those of neighboring countries. Commercial banks have remained very conservative in their limited exposure to housing finance, and Islamic mortgage lending has not fulfilled its considerable potential in servicing lower- income borrowers in the housing finance sector. In spite of the small share of housing finance in banks' portfolios, those loans cause exposure to credit risk, as well as liquidity risk from the term-maturity mismatch, given short-term deposits and long-term lending. Credit risk is being managed by increasingly better client information, but will take on renewed importance as the market moves down the income scale. Interest rate risk is borne by the borrower because banks do not offer fixed-rate instruments. The palpable shift from public to private housing finance provision was achieved as a result of aggressive and enabling policy measures. The Financial Institutions (Recovery of Finances) Ordinance, 2001, established a new nonjudicial procedure for recovering secured claims. The National Housing Policy was published the same year. The regime of mortgage lending was liberalized and strengthened by lifting various regulatory restric- tions. An appropriate prudential framework for housing finance was designed. Credit information services were introduced. The SBP constituted the Housing Advisory Group in 2002 to analyze existing housing and housing finance policies and to address key chal- lenges, and it created a special department focused on promoting market-based credit in housing finance. Further measures included fiscal incentives in 2003­04 and several gov- ernment mega-housing projects in big cities to increase the housing supply. Important regulatory weaknesses remain, creating continued uncertainty on prop- erty and housing finance transactions. An inefficient legal framework; fragmented own- ership and titling procedures; inefficient land information systems; inaccessible, unused government land; high stamp duties and registration fees; a weak tax framework; and ineffective land dispute mechanisms dampen the willingness of banks to increase their mortgage portfolios. In Karachi, more than 17 different agencies are responsible for land titling and registration. Stamp duties and registration fees can go up to 9 percent in some provinces. Land registration is manual, nontransparent, and cumbersome, with several possibly applicable legal regimes. As a result, most deals are made for cash and without official title documents. The recovery, foreclosures, and eviction framework from 2001 has certain legal weaknesses, but its biggest challenge is the reluctance of courts to apply it in practice. 210 Appendix D D Overhauling housing development regulations; providing infrastructure for residen- tial development; improving the reliability of land administration, records, and titling; streamlining property taxes; rewriting urban rent laws; enforcing foreclosure laws; and promoting long-term funding will boost the overall demand for and supply of mortgage finance. And, in the medium to long term, those efforts will address the supply of serv- iced land. New financing mechanisms including real estate investment trusts, municipal bonds, and a secondary mortgage facility would help deepen the housing finance mar- ket. An overhaul of the HBFC and innovative instruments would promote microbor- rowing and first-time purchases. Overview With a population of more than 167 million people (as of May 2009), Pakistan is the sixth most populous country in the world; and has the third-largest Muslim popu- lation after Indonesia and India. The country is riddled with problems of terrorism and lagging rural regions, with a poverty rate of 24 percent and access to any type of formal finance at barely 14 percent. With inflation at 12.0 percent in fiscal 2007/08 and projected to rise to 20.8 in fiscal 2008/09, and with an unstable macroeconomy, GDP growth slowed from 6.8 percent in fiscal 2006/07 to 4.1 percent in fiscal 2007/08. Projection for fiscal 2008/09 is 2.5 percent.1 The high poverty level is coupled with an inequitable income distribution, lack of basic infrastructure, substandard living conditions and erosion of basic health and educational facilities--particularly for low-income groups. Despite hectic government measures, economic indicators are showing down- ward trends. A turbulent political situation, law and order instability, global supply shocks--the oil, food, and financial crises--and a softening of external demand have adversely affected key macroeconomic fundamentals. Given low population growth rates, however, per capita income has grown at an average annual rate above 13.0 per- cent during the last five years. Total investment declined to 19.7 percent of GDP in fiscal 2008/09, from its peak level of 22.9 percent in fiscal 2006/07; and is expected to experience further decline in 2010. In recent years, Pakistan has undertaken major financial sector reforms that have resulted in a more efficient banking system. During the global financial crisis, the Pakistan banking system was relatively resilient because it has had an expanding asset base fueled by healthy deposits and equity growth. In spite of the worsening macro- economic environment, the banking system appears to be generally well capitalized, profitable, and liquid. Consumer lending has taken off in recent years, with banks expanding aggressively into consumer, automobile, and housing loans. At the same time access to finance remains limited. Currently, less than 14 percent of the popu- lation uses formal financial services, compared with 32 percent in Bangladesh, 48 percent in India, and 59 percent in Sri Lanka. Microfinance institutions and microfinance banks are the formal providers of finance to poor households and microenterprises, whereas commercial banks are the source of credit and finance to 211 Appendix D D small and medium enterprises. In Pakistan, 1.7 million people use the microfinance sector--this is about 1 percent of the population and about 2 percent of adults-- because the country has the lowest outreach in the region (except for Afghanistan). Operational and financial sustainability of microfinance institutions is also the low- est in the region. Similarly, access to finance for small and medium enterprises is low on an absolute and comparative basis. The percentage of such enterprises with bank loans in Pakistan is 7 percent, compared with 32 percent in Bangladesh and 33 per- cent in India. Macroeconomic instability has had an adverse impact on financial system per- formance and growth. The heightened credit risk resulted in a significant increase in nonperforming loans and associated asset quality indicators. Furthermore, the slow- down in the economy dampened the growth rate of the banking system. During the first quarter of 2009, the asset base of the banking system passively grew by 1.6 per- cent. The net loan portfolio declined by 5.6 percent over the same quarter, while investments in government paper increased by 20 percent. Bank deposits remained stagnant, and banks had to resort to interbank borrowing to finance the marginal increase in their asset base (SBP 2009). The urban share of the total population is already high at 37 percent, and is projected to rise to 50 percent by 2030 (figure D.1). Annual city growth rates are high, ranging between 3.8 percent and 5.5 percent in the late 1990s. Karachi is said to be growing at 6.0 percent annually, requiring the city to accommodate 700,000 new residents each year. Pakistan now has nine urban centers close to or Figure D.1 Urban and Rural Populations in Pakistan, 1950­2030 160 60 projected 140 50 120 population (millions) 40 100 percent 80 30 60 20 40 10 20 0 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 year rural population (millions) urban population (millions) percentage urban projected rural population (millions) projected urban population (millions) Source: UN-HABITAT 2008. 212 Appendix D D exceeding 1 million people. High urbanization considerably contributes to chronic housing shortages. Pakistan is facing unprecedented challenges of acute housing shortages, unhealthy living conditions, and nonexistent or dilapidated infrastructure across the country. More than 50 percent of the urban population lives in slum areas. The lack of affordable housing is pervasive because of the exorbitant markup of interest rates, high real estate prices, and eroding incomes of middle- and low-income groups. The average national occupancy rate per dwelling is above six; density per room is 3.5 people, compared with the international standard of 1.1 per room. The mortgage- to-GDP ratio is below 1 percent. Currently there are 25 commercial banks offering mortgage finance in Pakistan. The HBFC is the only state-owned corporation that caters to the financing needs of middle- and low-income groups. The housing market has also witnessed an increasing participation in housing finance by Islamic banks since 2007. High economic growth and remittances have produced escalating real estate prices and a high ratio of urban property prices to purchasing power that is typical for densely populated cities, such as Mumbai and Hong Kong, China. The twin prob- lems of supply shortages and escalating prices underscore the underdevelopment and undercapitalization of the housing sector. Until recently, housing finance remained neglected--the supply of housing finance actually decreased from 1.5 per- cent of GDP in 1994 to approximately 0.5 percent in 2001. Although this trend has experienced a reversal in the 2000s, in response to government policies and increased bank interest in retail lending, mortgage finance in 2007 was still less than 1.0 per- cent of GDP. The government has taken numerous steps to encourage development of a more stable and vibrant housing sector. The Financial Institutions (Recovery of Finances) Ordinance, 2001, established a new (nonjudicial) procedure for recovering secured claims. The regime of mortgage lending was liberalized and strengthened by lifting various regulatory restrictions. An appropriate prudential framework was designed. Credit information services were introduced. The Housing Advisory Group was appointed in 2002 to iron out housing sector­related issues. A Develop- ment Finance Group was created within the SBP, with enhanced focus for creating an enabling environment to facilitate intermediation of market-based/high-priority credit (which included housing finance as a key area). The 2001 National Housing Policy was published. Finally, fiscal incentives were granted in the federal budget of fiscal 2003/04 to facilitate mortgagors and banks. The greatest challenges that the housing and housing finance sectors are fac- ing at the moment are the inefficiency of the overall regulatory regime, includ- ing land and titling procedures; a poor regulatory framework for housing and real estate; and the lack of an organized database and key information on the housing and housing finance sectors. Other challenges include insufficient devel- oped land and exorbitant land and housing finance prices; housing shortages; maturity mismatch and liquidity risk in mortgage lending; poor government success in addressing low-income housing needs; and lack of commercially viable housing microfinance lending. 213 Appendix D D Housing Demand Housing shortages have increased from 4.33 million units in 1998 (the latest census) to an estimated 7.57 million units in 2009 (table D.1). The factors responsible for substantial increase in the housing backlog are generic population growth, year-on- year short supply, and depletion of existing stocks. The yearly new housing units needed because of population growth is 0.7 million, whereas the addition to the existing stock is only 0.3 million a year. Considering depletion of stock at 1 percent or 0.2 million units a year, the new addition would practically have no impact on existing stock. Figure D.2 presents a breakdown by income group of various aspects of Pakistan's housing finance industry. The high- and very-high-income brackets, with annual earnings of more than PRs 51,000, represent the top 5 percent of the Pakistani popu- lation that is serviced by commercial banks. These people can afford a mortgage above PRs 2.5 million.2 This is the sole group that can be sustainably serviced on a commer- cial basis, given current financial development and mortgage know-how in Pakistan. The market can address this demand in spite of high land prices, scarce developed land, and low capacity for housing units supply because the segment can afford a considerable amount in mortgage (commensurate with property prices) and because of the modest size of the segment (estimated at a little over 1 million households). For the upper-middle- and lower-middle-income groups, incomes range from PRs 12,000 to PRs 50,000; and together they represent 35 percent of the population in the country. These groups are serviced by the HBFC, and the mortgage level they can afford falls in the range of PRs 0.5­2.5 million. The housing shortage estimates for this middle-income group of about 8.1 million households is 2.6 million units. This income group, although currently not serviced by commercial banks, could be Table D.1 Basic Housing Indicators for Pakistan, 1998 and 2009 Description 1998 2009 (estimated) Total housing (units) 19.30 million 20.48 million Housing backlog (units) 4.33 million 7.57 million New housing need resulting from 0.30 million 1.28 million population growth (units) Annual depletion of housing stock 1% 1% Total population 132.2 million 167.0 million Urban population 32.5% of population 37% of population Annual growth rate of housing 8% 8% demand in urban areas Urban housing units 32.30% 32.20% Average household size 6.0 people 6.6 people Sources: Pakistan's 1998 national census; authors' estimates. 214 Appendix D D Figure D.2 Segmentation of the Housing Market by Income Group Housing finance player Market segment Income distribution (%) High - High PRs 1 PRs 5 1 Lakh and Commercial and Banks 0.075 above above PRs 2.50 PRs 51,000 High End to 4 0.300 to PRs 5.00 Market PRs 100,000 Housing PRs 25,001 to PRs 1.25 ~ 2.50 Upper Middle 15 1.125 PRs 50,000 HBFC Lower Middle PRs 12,001 to 20 1.500 PRs 0.50 ~ 1.25 PRs 25,000 HBFC and PRs 6,001 to PRs 0.20 ~ PRs 0.50 Social Small 40 3.000 PRs 12,000 Housing Social Bank Housing Micro 20 PRs 0.20 up to PRs 6,000 1.500 Mortgage affordability Housing market Income distribution Housing shortage* (PRs millions) (millions) Source: Rizvi 2007. Note: Housing shortage is estimated from the total existing backlog of 7.3 million, proportionately, to the population shares for each income bracket. However, actual shortage is much higher in low-income segments than in higher-income segments. profitably addressed by the market if there were better technology, more competition, a level playing field, and further financial development within a stable macroeconomy. The lowest two income segments making up two thirds of the population cannot be profitably provided with a dwelling or a mortgage at this stage, and require recourse to some sort of subsidy. The group with income of PRs 6,000­12,000, rep- resenting 40 percent of the population, can afford a mortgage of PRs 0.2­0.5 million. These roughly 9.3 million households are exposed to an estimated housing shortage of 3 million units. The lowest 20 percent, with incomes below PRs 6,000 (roughly 4.6 million households), can afford nothing above PRs 200,000; and they face a housing shortage of 1.5 million units. As a result, lower-income groups expand into illegal habitats and squatter settle- ments (katchi abadis) in major cities. These are pockets of unregulated tracts of land scattered throughout the landscape of large cities; and over the years, they have sprouted to fulfill the housing needs of the low-income stratum of the society. Because property ownership in these shantytowns is devoid of a legal title, housing finance from the formal credit system could not trickle to these areas. Moreover, as a result of their unregulated nature, katchi abadis are a cause of revenue slippages for the government on a continuous basis. More than half of the major cities (like Faisal- abad, Karachi, Lahore, and Peshawar) have katchi abadis. Karachi alone has 600­800 such settlements, sheltering about 7.6 million people (or 1 million households, as of 2005) out of the total city population of 15.1 million people. Singh province, which comprises Karachi as well as Hyderabad, has 1,300 katchi abadis overall, located on 24,300 acres of government-owned land and 1,700 acres of privately owned land (table D.2). The situation is also difficult in Punjab, with 902 katchi abadis. 215 Appendix D D Table D.2 Details of Katchi Abadis in Pakistan, by Province, 2008 Government Privately Total Katchi Abadis Owned Owned Area Province (number) (number) (number) (acres) Balochistan 55 2,826 0 2,826 North-West Frontier 65 1,509 3,434.03 4,943 Punjab 902 8,875 501.16 9,376 Sindh 1,300 24,300 1,700.38 26,000 Total 2,322 3,7510 5,635.60 43,145 Source: HBFC country report, 2008. The province comprises the cities of Faisalabad, Gujranwala, Islamabad, Lahore, Multan, and Rawalpindi, all with populations exceeding 1 million. Housing Supply Pakistan is facing unprecedented challenges of acute housing shortages, unhealthy living conditions, and nonexistent or dilapidated infrastructure. Current State of Housing About one third of available housing units are urban, and the remainder is rural. Ownership in the country is at approximately 81 percent of dwellings, with about 9 percent of households living rent-free and 11 percent renting. Roughly 40 percent of houses are semi-pucca (permanent) houses, mostly without planned sanitation or a sewerage system; 39 percent are kutcha (temporary) houses with minimal water supply and sanitation or drainage services; and 21 percent are modern brick con- structions. In addition to rising land prices, the costs of construction materials have been increasing considerably, particularly those of cement and iron. Unavailability of developed land, coupled with the high prices of construction materials, has com- pounded the problem of housing shortages in Pakistan. Private developers are relatively dynamic and have been engaging in considerable activity, including large infrastructure projects. Their supply of housing has exclusively targeted the upper- income population. Unavailability of requisite bulk developer financing from com- mercial banks is another big hurdle in the way of housing supply in the country. The government has initiated some low-income housing projects; but these efforts have been vastly inadequate, relative to need. The brisk pace of activities in housing and high-rise building construction, along with large public sector spending on physical infrastructure and ongoing reconstruction activities in earthquake-affected areas, have contributed to the sharp pickup in construction value added. Construction has been growing at an average annual rate of 10.2 percent over the last five years. 216 Appendix D D Although the construction sector has only a 2.3 percent share in GDP, its share of the employed labor force was disproportionately large, at 7 percent. Development of Housing Stock The current land development framework is highly inefficient and acts as a critical bottleneck to the availability of developed land on the market. Procuring Land for Development. The land administration system requires that the development agencies acquire land, develop it, and then sell it to the public. This process is plagued with legal and administrative burdens, which increase the costs and time for delivery and reduce the availability of developed land. The development authorities are also faced with a lack of adequate funding to finance this cumber- some process, which also impedes their ability to meet demands of the market. Accordingly, developers are confined to procuring scarce developed urban land through the government. Landowning government agencies (such as city govern- ments) auction plots, which then give rise to speculative secondary trading until they are purchased by the highest-value client. This speculative element considerably impairs affordability of subsequent developments on auctioned land, so the real estate naturally is placed with only the highest-income groups. This can be remedied by transforming the land auctions into government contract awards for completed housing projects, thereby making it less attractive for the speculators to bid. More- over, such an undertaking would increase the supply of constructed housing units rather than that of plots (which usually remain vacant even many years after auc- tion). In sum, the institutional framework for identifying, acquiring, and servicing land that is available for construction and subsequently selling the constructed hous- ing units is underdeveloped. Some of the factors causing this shortcoming are poor master planning, governance issues at national and local levels, multiple institutions and administrative procedures, problematic zoning restrictions, unreliable utility connections, and restrictive building codes. National and local master plans are either inadequate or poorly enforced, and this has led to inefficient allocation of land and uncontrolled urban development. The root cause is the lack of clear responsibility among a multitude of government institutions (such as in Lahore, where responsibility is shared among the land devel- opment authority, the city district government, and the tehsil municipal adminis- trations), as well as a mismatch between expenditure and revenue responsibilities. The lack of transparency and accountability in the planning process also opens opportunities for land mafias to have valuable inside information about future infrastructure developments or to influence such plans so that mafia land increases in value. The number of institutions and registration procedures required to execute property transactions needs to be reduced. Multiple institutions and procedural dif- ferences between and within states create market distortions and inefficiencies in the property development market. The distortions result in improper registrations and 217 Appendix D D a culture favoring informal property transactions--an obstacle to effective liens and asset securitization. Increasing the supply of land has been problematic in Pakistan. According to the Association of Builders and Developers (ABAD), the government of Pakistan directly controls 40 percent of the land, and acquiring the land from the government is challenging. Therefore, a large share of the land directly controlled by the govern- ment is not being used for productive or strategic purposes, and the remainder is leased to the private sector. According to a Foreign Investment Advisory Services administrative barrier study, it takes an average of 135 days to acquire land from the government. Private land can be acquired in 63 days, but even this is not a good- practice benchmark. Despite imposition of a nonutilization penalty on plots, construction on a signif- icant number of residential plots--auctioned to the private sector by the various public sector landowning authorities--has not been initiated because of the allot- tees' tendency to reap speculative gains. The present land allocation policy encourages horizontal development, which adds to the land scarcity problem. At present, only a small portion of land is allo- cated for high-rise buildings: for example, in Karachi only 5 percent of the land is earmarked for apartment buildings, but 55 percent is assigned for individual houses. Because land is a major component of the total building cost, imbalances created under the present land allocation policy, among other things, have led to a real asset price spiral in the urban centers of Pakistan. Various landowning agencies currently release public land to the end users in an unplanned and inadequate manner. The quantity of auctioned land falls short of quenching the demand for housing facilities. Because a limited number of plots are offered through these public auctions, speculative gains abound. Speculators in large numbers participate in the auction process, which breeds "file-culture." Because of the gross involvement of the speculators, genuine recipients are able to purchase a piece of land only at a much higher price through the secondary market. At present, various government agencies (such as City District Government Karachi) are enjoy- ing the status of a natural land bank; they release public land to the private sector on a recurring basis but in a haphazard manner, devoid of any medium-to-long-term strategic focus. Although the conversion of rural land into urban land at the town periphery is a key means of increasing the supply of land--and one that should benefit the lower classes most (including farmers)--it is made problematic by the ill-inspired interven- tions of local and provincial governments. The government of Karachi, for example, has put a ban on the conversion of Barani land around Karachi following malprac- tices (specifically, converted land made available with the intention of developing social housing ended up being used for more profitable commercial ventures). The Sindh government has also cancelled the validity of 5,000 land allotments around Karachi, but the original culprits resold their land long ago and moved on. The government is slowly regularizing these cases. (Less than 200 cases have been reviewed so far.) 218 Appendix D D Overly restrictive building codes and laws on subdivision limit the efficient use of urban land and increase the price to consumers, especially in high-land-price zones. Lahore, for example, has a restrictive floor/area ratio of 1.0:1.5 (floor areas less than 1.5 are low density; plot ratios above 5.0 are very-high density). The trouble with such a low ratio is that the city cannot accommodate enough population to provide adequate demand or funding for the appropriate infrastructure. Similar issues pre- vail in Karachi, which has a very low limit on the height of apartment buildings. Fur- thermore, a large share of the city buildings cannot be put on the secondary market or used as collateral because their developers did not respect the limit on the num- ber of stories. Their legal status has been pending for many years because the devel- opers are well connected or no longer around. (There are 300 such buildings in Karachi alone.) Poor incentives and governance within government-controlled companies (espe- cially a problem in power and gas) typically can delay large-scale construction projects by as much as a year. Some developers are relying on gas-powered, self-generation power plants as an alternate and viable source of electric power. Poorly targeted subsi- dized pricing, widespread theft, and expensive mechanical failures considerably reduce the incentives and financial capacity of government-owned utility companies to con- nect new housing developments. For example, the Water and Sanitation Authority esti- mates that 40 percent of water is lost through mechanical failure and theft. By far, the scarcity of developed lands at prices that are affordable to other than the highest-income groups is the biggest challenge for policy makers and the hous- ing sector. Housing shortages and exorbitant real estate prices fundamentally are two by-products of developed land shortages in the country. Though land itself is not scarce, there is an acute dearth of serviced land along with the attendant residential infrastructure. In fact, the government already has devised a land policy to make developed lands available throughout the country, through the National Housing Authority. The policy provides for urban and rural settlements for housing develop- ments that reduce the concentration of metropolitan and major urban centers, and that enhance land availability through such innovative measures as land banking on a continuous basis to cater to longer-term needs.3 Further, the aim is to amend the land acquisition laws to make provision for a unified, transparent, and market value­oriented system that will minimize litigation. A comprehensive land informa- tion system would be created, including inventory and land classification, geographic information system mapping, settlement patterns, land values, land availability, and other data. The policy provides for federal and provincial funding for infrastructure, amenities, and other developments in the planned housing development areas, with the responsibility and cost fully shouldered by public utility agencies, not the devel- oper. Integrated development would be promoted, optimizing land use and man- dating the preparation of master structural plans. For all its redeeming features, this policy has not been implemented as yet. Low-Cost Construction Technologies. Developers use conventional construction technologies with gradual introduction of recent improvements for residential and 219 Appendix D D commercial construction. Few initiatives exist in low-cost housing development, and all of these are indigenous. No serious attempt has been made to have technical collaboration with a globally acknowledged, low-cost construction technology provider. These local initiatives include the use of prefabricated materials technology (such as prefabricated slabs, roofs, pillars, and so forth). The primary gain of the pre- fab material technology is the cost reduction resulting from production economies of scale and reduced labor costs. An alternative--the "8x8 construction technology" (also known as movable housing technology)--has been tried on a very limited scale. Under the 8x8 technology, all dimensions of the house are eight feet and the structure is placed over a movable frame with wheels fixed on the bottom at the four corners. To promote development of low-cost construction technologies and mate- rials, the HBFC in 2006 set up a Center for Promotion of Low-Cost Housing, in asso- ciation with NED University of Engineering and Technology (Karachi). However, this useful undertaking lost momentum after 2008. That leaves the country with no existing platform to evaluate, certify, and introduce the internationally proven low- cost housing technologies capable of matching local requirements. There is a need in Pakistan to initiate low-cost construction pilot projects on the pattern of four such projects implemented in India. The Developer Industry. The emergence of an active and competitive building industry is constrained by industry fragmentation, unorganized brokers, the lack of developer finance, an unstructured framework for property development, and skewed tenant laws. The property development industry suffers from low public confidence. A history of scams; financial weaknesses; and the absence of clear, uni- form, and fair business practices have tarnished its credibility. Financial institutions are reluctant to provide construction finance, and individuals are reluctant to apply for mortgages. The majority of real estate builders and developers are organized as sole proprietorships or partnerships, with limited capital and informal corporate governance structures. Without using a strong regulatory authority to enforce build- ing standards, the quality of housing still leaves much to be desired. The unstructured and unsupervised nature of real estate brokers is also a signifi- cant constraint to the provision of housing and housing finance. Only appraisers have professional conduct requirements that were established by the SBP and the Pakistan Banks' Association. The real estate agencies, which could be natural brokers or arrangers for the provision of financial services, remain unorganized and insuffi- ciently supervised. The protection of individual purchasers remains limited because the market is dominated by cash transactions and lacks transparency, and because no systematic information is available. Risk assessment and portfolio valuation are frag- ile as well--another factor prompting lenders' extreme caution. The absence of sound governance structure within the housing developer indus- try creates major deficiencies, such as lack of good practices, illegal construction, unreliable building permits, and legally unprotected advance purchase of units that are required to be built in the future. Consequently, it is difficult for financial insti- tutions to verify the character, capital, and capacity of potential clients. As a result, 220 Appendix D D financial institutions are reluctant to enter this market. Their reluctance, in turn, makes finance scarce and constrains the supply of housing. Developer Financing. At present, there is little external bank financing for property development. Land for a housing project is usually owned by many interconnected companies with complex relationships, understandings, and arrangements that are hard to assess. Further, transparency and accountability are not easily practiced in the presence of such complex ownership arrangements. Historically, the real estate business has been booking driven; in the recent past, however, the paradigm has shifted and now the main concern regarding the provision of funding to real estate builders and developers is that financing generated through the general public may not be sufficient to complete and hand over a housing project/scheme in accordance with the booking schedule. Therefore, there is a need to figure out the modus operandi for financing arrangements for builders and developers through the formal banking system that is viable, acceptable, and sustainable. Developer Associations. The All Pakistan Contractor Association (APCA) was created in 1998, under the Companies Ordinance 1984; and is affiliated with the Federation of Pakistan Chambers of Commerce and Industry as an "A" class association.4 APCA's efforts to address the problems facing the construction industry have met with sub- stantial success. On the initiative of APCA, the construction sector has been recognized as an industry in official government classification, entitling the sector to fiscal incen- tives, special treatment, and other benefits. That also helped association members win a number of government housing projects throughout the country. APCA was instru- mental in drafting the National Housing Policy for low-income groups. Finally, APCA achieved duty-free status for imported construction machinery. ABAD, created in 1972, is a national-level representative organization under the Companies Ordinance 1984, affiliated with the Federation of Pakistan Chambers of Commerce and Industry.5 Its mission is to unify and streamline the construction activ- ities of the private sector; to promote the housing industry; to provide assistance to the government in formulating housing policy and explore the potential for large-scale public projects for private implementation; to coordinate housing stakeholders; and to support capacity-building, awareness, and training efforts. ABAD has made a valuable contribution in the formulation of the National Housing Policy and the Sindh Building Control Ordinance. It has also supported the enhancement of the role of the HBFC. Rental Regulations. Tenant laws further increase the riskiness of investment in rental property, resulting in a very small rental market. (Private banks refuse to finance devel- opments destined for the rental market.) The letter of the law is highly skewed in favor of tenants because eviction is very difficult. Relevant laws act as an impediment to the provision of housing finance aimed at generating rental incomes. Consequently, large shares of land-related disputes are tenant related. This issue disproportionately affects the middle and lower classes for whom renting is the only way to move out of the slums, especially in the current high-price environment. This issue also affects the retail 221 Appendix D D sector, especially the development of professionally managed shopping malls. Tenant laws in Islamabad are more appropriate, and the rental market is significantly more developed (constituting 40 percent of the market, versus 20 percent in Lahore). The recent increase in rental taxes (Finance Act 2008; the prior tax was 5 percent) is con- siderably decreasing transparency in the leasing environment.6 There is no single statute that governs the landlord and tenant relationship; instead, each state has its own governing statute, and most of them are outdated. Although the law was designed to protect both the landlord and the tenant, imple- mentation has been fraught with poor governance and delays in enforcement. Specifically, a rent case takes about 24­42 months in the Court of the Rent Con- troller; about 8­10 months before the Appellate Court (district judge); and when a second appeal or a revision is entered before the High Court, it takes approximately seven to eight years before the final disposition. The position and powers of a Rent Controllers--basically, a civil judge--have become a major source of corruption and delays in the settlement of cases relating to tenancy. It is significant to note that nearly 40 percent of the total pending cases in the courts of law relate to land dis- putes; those relating to tenancy are roughly 23 percent. Construction Costs. The quality of construction is grossly compromised by high prices for construction materials. There are no formal data available on housing construction costs, and those vary considerably with the quality and luxury level of the construction plans and with the location of the property. Nonetheless, a casual regional comparison makes it plain that construction costs in Pakistan are much higher than those in Bangladesh, India, and Sri Lanka. An interview with the devel- opers puts current estimates at PRs 800­1,000 per square foot for low-cost housing, PRs 1,400­1,600 per square foot for middle-cost housing, and PRs 2,200­3,000 per square foot for high-end housing. Construction Quality and Enforcement of Construction Standards. Following the 2005 earthquake in the northern areas of Pakistan, building codes were redrafted in 2007 to include seismic provisions, with the technical support of the National Engi- neering Services Pakistan (Pvt.) Ltd. The 2007 seismic provision of the building code serves two purposes. First, for each tehsil (the smallest administrative unit in a province) of the country, it defines the earthquake level for which buildings must be designed. Second, it provides codified procedures and guidelines for planning, design, analysis, and detail of building structures. The topics covered include site considerations, soil and foundations, general structural design requirements, rein- forced concrete buildings, structural steel buildings, masonry buildings, architectural elements, and mechanical and electrical systems. Effective implementation of this code can lead to significant improvements in the seismic safety of buildings in the country. Compliance levels remain low, however. The Secondary Property Market. Real estate in Pakistan has an active secondary market; but it suffers from certain limitations including lack of transparency, price 222 Appendix D D speculation, and sales involving legally imperfect documentation. Real estate brokers are neither licensed nor registered. More than 80 percent of the mortgage finance being done by the commercial banks is for secondary market transactions. Housing Finance Overall, consumer lending in general, and housing finance in particular, are grow- ing. Banks, once conservative consumer-loan lenders, are now aggressively expand- ing into consumer loans (including those for housing). The share of commercial banks involved in housing finance grew from 10 percent in 2003 to 79 percent in 2009. Each bank is investing resources and skills to attract market share from tar- geted segments of the market. In this relatively new product development phase, the market is open for larger portfolios and for more institutions to enter. Currently, 25 commercial banks offer mortgage finance in Pakistan. The state- owned HBFC is the only institution that caters to the financing needs of middle- and low-income groups. Although the entry of commercial banks in the housing mort- gage market has occurred relatively late (in 2007, with the consumer boom), its evo- lution has remained promising. The market has also witnessed an increasing partici- pation in housing finance by Islamic banks since 2007 (annexes D3 and D4).7 Other sources of long-term funds include provident funds, various organizations, builder's credit, and friends and family. Although the current legal and macroeconomic environment is increasingly con- ducive to the growth of the housing finance market, financial institutions in general have yet to provide home financing to a great majority of the population--especially the urban working class that is most in need of mortgage financing. Primarily, banks are constrained in their lending by the lack of long-term, fixed-rate funding that would enable them to increase the volume and tenure of their housing portfolios. Thus, they have continued to rely on short-term deposits to fund long-term loans-- a practice that increases the risk for lenders and restricts the menu of long-term mort- gage products that can make housing more affordable. The current cost of short-term liabilities is significantly lower than the average markup charged by the banks on housing loans, so it yields a huge spread for the banks. However, this practice is creating asset-liability mismatching that may not be sustainable over the long term. Even if long-term funds were immediately available, however, the volume of mortgages would be restrained by broader challenges facing the sector as a whole-- challenges for which public intervention is required. The weak property rights and land development framework, the embryonic property development framework, and the emerging building industry pose significant problems for Pakistan's nascent mortgage finance market. The Housing Finance Market. On March 31, 2009, the total outstanding finance of all banks and development finance institutions (DFIs), including the HBFC, stood at PRs 80.87 billion; private banks had a significant lead, with PRs 38.6 billion--a 223 Appendix D D Figure D.3 Housing Finance Outstanding Loans, March 31, 2008 and 2009 43,325 50,000 amount outstanding (PRs millions) 38,679 45,000 40,000 35,000 30,000 16,400 25,000 14,591 20,000 10,498 10,392 8,986 15,000 7,395 5,910 5,594 10,000 633 497 5,000 0 public sector private Islamic foreign DFIs HBFC banks banks banks banks banking sectors March 2009 March 2008 Source: SBP 2009. 48 percent share in the total housing finance sector. Compared with the quarter ended March 31, 2008 (figure D.3), outstanding loans of all commercial banks and DFIs increased by 1.3 percent by March 31, 2009. Islamic banks have an 11 percent share in total outstanding loans with PRs 8.98 billion, followed by public sector banks with PRs 10.39 million (13 percent) and foreign banks with PRs 5.91 billion (7.3 percent). The cumulative outstanding loans of the HBFC stood at PRs 16.40 billion as of March 31, 2009, showing a 20 percent share in the overall stock of outstanding loans. In 2009, Islamic banks showed the highest growth, at 21.5 percent. The HBFC's outstanding lending grew by 12.3 percent. In contrast, private banks' lending decreased by 10.9 percent, while public sector and foreign bank lending remained virtually unchanged. During the quarter ended March 31, 2009, the banks and DFIs extended loans to 712 borrowers (compared with 3,673 borrowers served during the same quarter in the previous year). The HBFC served 46.6 percent of the total borrowers by extend- ing finance to 332 borrowers during the quarter. The total number of housing finance borrowers as of March 31, 2009, was 121,368 (down from 126,595 on March 31, 2008), a 4.1 percent decrease (figure D.4). That decline masks a growth in the clientele of foreign and Islamic banks (by 286 percent and 16 percent, respectively), compensated by a decline in the number of borrowers served by private banks, public sector banks, the HBFC, and other DFIs (declines of 14.6 percent, 2.0 percent, 5.0 percent, and 18.0 percent, respectively). No disburse- ments were made by DFIs during the quarter ended March 31, 2009. Average loan size for the HBFC is PRs 1.08 million, compared with those of private sector banks (PRs 2.29 million), public sector banks (PRs 1.75 million), and foreign banks (PRs 2.68 million). Surprisingly, Islamic banks have not played any role in middle-income home financing; their average loan size remains high at PRs 3.7 million. 224 Appendix D D Figure D.4 Number of Housing Finance Borrowers with Outstanding Loans, March 31, 2008 and 2009 120,000 97,642 92,854 100,000 number of borrowers 80,000 60,000 40,000 15,377 15,822 7,707 8,049 3,077 2,997 2,193 20,000 1,889 196 160 0 public sector private Islamic foreign DFIs HBFC banks banks banks banks banking sectors March 2009 March 2008 Source: SBP 2009. Private banks vastly dominate the housing finance market when outstanding loan balances are considered; however, the HBFC remains the largest lender by number of clients, serving some 71 percent of housing finance borrowers (figure D.5). The over- all market share of commercial banks (excluding DFIs) remained around 81 percent between March 2008 and March 20009, with rapid growth in the market share of Islamic banks compensating for a decrease in private bank share. The market share of the HBFC has risen from 17.8 percent to 20.3 percent (not accounting for nonperforming loans). Housing finance sector products range from loans for home purchase or renova- tion to financing for construction (figure D.6). The most popular product by far, with an outstanding portfolio of PRs 47.6 billion and a share of 59 percent, is the home purchase loan. Construction loans stand at PRs 23.8 billion (29 percent share), and renovation lending is at PRs 9.39 billion (12 percent of total).8 In a nascent mar- ket, the dominant share of home purchase loans with smaller exposure to construc- tion and renovation loans is not surprising. The need for verifiable collateral for the loan makes construction loans more difficult to finance; renovation loans tend to be smaller, of shorter maturity, and with lower profit margins for the banking institu- tion. In 2009, financing for construction increased by 6.0 percent, while outright purchase and renovation loans decreased by 4.0 percent and 7.3 percent, respectively. Private banks have taken a lead in financing for construction and outright purchase, with 41.5 percent and 55.4 percent shares in the respective sectors. Public sector banks enjoy the greatest share in the renovation sector--25 percent. The overall weighted average interest rate in the housing finance market for the quarter ended March 2009 was estimated to be 15.42 percent--up 2 percentage 225 Appendix D D Figure D.5 Market Shares of Housing Finance Lenders, March 31, 2009 a. by outstanding portfolio public sector HBFC, banks, 20.3% 12.9% DFIs, 0.6% foreign banks, 7.3% Islamic banks, private banks, 11.1% 47.8% b. by number of borrowers public sector banks, 6.30% Islamic banks, private banks, 1.80% 12.60% foreign banks, 2.50% DFIs, HBFC, 0.13% 71% Source: SBP 2009. points from March 2008 (table D.3). This underlines higher markups for Islamic and foreign banks (average interest rates of 17.8 percent and 17.5 percent, respectively), and more competitive pricing for public and private sector banks and the HBFC (15.0 percent, 14.6 percent, and 11.9 percent, respectively). Average maturity periods have declined marginally, from 13.2 years to 12.5 years in March 2009. DFIs and the HBFC extend loans with considerably longer maturities than those extended by commercial banks--14.5 years (DFIs) and 15.2 years (HBFC)--and so do Islamic banks (14.9 years). Public, private, and foreign banks offer home loans for 10.6, 11.9, and 12.0 years, respectively, on average. The loan-to-value ratios declined progressively over 2009 (down to an average of 53 percent for the housing finance market) as a result of increased caution and the impact of the global financial crisis (table D.4). Loan-to-value ratios remain high at 226 Appendix D D Figure D.6 Mortgage Products Offered by Housing Finance Lenders, 2009 23,119 25,000 20,000 PRs millions 15,000 7,436 7,548 10,000 6,248 4,507 4,383 3,252 3,005 1,999 1,956 1,749 5,000 674 640 551 397 307 86 17 0 public private Islamic foreign DFIs HBFC sector banks banks banks banks banking sectors construction outright purchase renovation Source: SBP 2009. Table D.3 Interest Rates and Maturities in the Housing Finance Sector Weighted Average Interest Rate (%) Average Maturity Period (years) Finance March December September June March March December September June March Institution 2009 2008 2008 2008 2008 2009 2008 2008 2008 2008 Public 14.96 14.30 13.33 13.13 12.78 10.6 10.8 9.4 11.1 11.0 sector banks Private 14.55 13.34 13.61 12.94 12.75 11.9 11.9 11.9 12.1 12.5 banks Islamic 17.81 18.23 14.86 13.22 13.84 14.9 15.3 15.0 16.6 16.6 banks Foreign 17.53 16.89 15.05 14.67 14.40 12.0 11.8 9.1 12.5 14.5 banks All banks 15.52 14.57 13.94 13.23 13.10 12.3 12.1 11.6 12.7 12.7 DFIs 17.60 16.84 16.01 14.84 13.68 14.5 14.7 12.4 14.7 14.8 All banks 15.60 14.65 14.02 13.29 13.12 12.4 12.2 11.6 12.8 13.0 and DFIs HBFC 11.88 11.35 13.71 13.71 13.71 15.2 15.1 15.1 16.7 16.7 Total 15.42 14.40 13.97 13.45 13.42 12.5 12.4 11.8 12.9 13.2 Source: SBP 2009. public banks (above 60 percent), followed by the HBFC, private banks, and Islamic banks (roughly 56 percent). The ratios evince a far more cautious approach by for- eign banks (with a average loan-to-value ratio around 36 percent). Average process- ing time has remained at just over 25 days for the housing finance industry overall, 227 Appendix D D Table D.4 Loan-to-Value Ratios and Loan Processing Times Average Loan-to-Value Ratio (%) Average Time for Loan Processing (days) Finance March December September June March March December September June March Institution 2009 2008 2008 2008 2008 2009 2008 2008 2008 2008 Public 60.5 63.1 58.3 63.7 58.2 31.7 38 38.6 39.2 38.0 sector banks Private 55.3 40.3 52.7 57.9 56.8 22.8 22 22.8 22.6 22.8 banks Islamic 55.3 39.8 56.7 59.1 55.9 30.4 36.9 33.6 21.5 20.9 banks Foreign 36.1 41.9 36.5 59.8 58.9 22.5 23.6 18.8 23.8 23.8 banks All banks 53.0 44.0 52.4 58.9 56.5 25.3 26.3 25.7 24.5 24.7 DFIs 43.1 44.2 49.1 45.4 47.5 30.0 30.0 30.0 30.0 30.0 All banks 52.6 44.0 52.3 58.3 56.5 25.5 26.4 25.9 24.7 24.9 and DFIs HBFC 55.8 67.5 55.0 55.0 55.0 30.0 30.0 30.0 30.0 30.0 Total 52.8 44.7 52.3 58.1 56.5 25.7 26.6 26.1 25 25.17 Source: SBP 2009. with private and foreign banks showing much more efficient processing and the HBFC being the slowest in loan administration. Foreclosure activities intensified considerably at the HBFC in 2009, whereas public banks remained unable or unwill- ing to collect the collateral of delinquent borrowers. Efficiency and Performance of the Mortgage Finance Industry Macroeconomic instability in Pakistan has had an adverse impact on financial system performance and growth, including in the housing finance sector. Accordingly, hous- ing finance nonperforming loans had increased to PRs 13.09 billion by the end of March 2009--a 45 percent increase over the previous quarter (figure D.7). Nonper- forming loans as a percentage of total loans outstanding for the entire housing finance sector (commercial banks and DFIs) are reported to be 16.2 percent. Commercial banks (excluding DFIs) collectively have nonperforming loans constituting 10.3 per- cent of their total outstanding portfolios. The HBFC reports nonperformers to be 39 percent of their total loans outstanding, whereas other DFIs stand at 17.1 percent. A 300 percent increase in nonperforming loans has been witnessed among Islamic banks (from PRs 88 million to PRs 336 million); that number is followed by an increase of 121 percent among public sector banks (from PRs 0.47 billion to PRs 1.50 billion). For foreign banks, such loans have increased from PRs 141 million to PRs 445 million--a 222 percent increase. Private banks have reported an increase 228 Appendix D D Figure D.7 Nonperforming Loans in the Housing Finance Sector, 2009 7,000 6,404 6,000 5,823 5,000 4,760 PRs millions 4,000 3,000 2,383 2,000 1,048 1,000 473 455 336 82 141 88 92 0 public private Islamic foreign DFIs HBFC sector banks banks banks banks banking sectors March 2009 March 2008 Source: SBP 2009. to PRs 4.76 billion (or 99 percent over the year), and the HBFC has reported an increase of 45 percent during the year (from PRs 5.8 billion to PRs 6.4 billion). Pakistan HBFC Historically, housing finance had been provided exclusively by the government- owned HBFC, established in 1952, because banks were not allowed to provide hous- ing finance. From 1952 to 1972, the HBFC disbursed loans only to homeowners for construction of homes in urban areas. During this period, the HBFC's disbursement activities were quite modest, averaging PRs 18 million annually. In 1972, the HBFC's mandate was expanded to housing projects, finance to housing authorities and hous- ing corporations, and rehabilitation loans; and it was permitted to undertake real estate development projects. Following this expansion of scope, its volume of origi- nations increased significantly, averaging around Rs 500 million annually between 1972 and 1979. In the 1980s and 1990s, the HBFC was averaging around Rs 1.5 bil- lion in annual disbursements/originations. Since fiscal 2002/03, the corporation has maintained a good pace in growth, and it disbursed PRs 4 billion in the last financial year. The HBFC is active in low- and middle-income housing finance. Commercial banks entered the mortgage finance business in 2002-03, with a business focus only on the upper-middle portion and the high end of the housing finance market. With their entry, the need for an institution to serve the low-and middle-income segments became more obvious, and the HBFC was and is the only institution responding to that need. The HBFC's outstanding portfolio has been hovering in the vicinity of PRs 16­20 billion for the last few years, in contrast to the fast portfolio growth of 229 Appendix D D Figure D.8 Commercial Banks Versus the HBFC--Outstanding Portfolios' Historical Trends, 2003­08 100,000 90,000 80,000 gross outstanding (PRs millions) 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 December December December December December December 2003 2004 2005 2006 2007 2008 month and year commercial banks HBFC total Source: SBP 2008b. commercial banks (over PRs 70 billion in 2008; figure D.8). Similarly, banks have outperformed the HBFC in terms of market share, as the HBFC's share fell from 50 percent in 2004 to less than 20 percent in 2008 (figure D.9). Conversely, commer- cial banks have taken a considerable lead--from virtually nothing in 2003, to just under 50 percent in 2004, and to 80 percent in 2008. During 2005­07, the HBFC (with technical assistance from the International Finance Corporation) availed the advisory services of the Canadian Mortgage Hous- ing Company to carry out a diagnostic study of the HBFC and make recommenda- tions for upgrading and reforming its operations, work environment, preparation of manuals, and computerization. The study also developed a business plan for the HBFC to make it a commercially viable and self-sustainable corporate entity. By the end of 2007, the HBFC had undergone a comprehensive reform program. That year, it was converted into a corporate entity under Companies Ordinance 1984, and thus was removed from the State Charter of 1952. Its financial restructuring and privati- zation were to follow, and they are ongoing. In terms of numbers of borrowers, the HBFC has shown negative growth over the years, while commercial banks have shown consistently positive growth (figure D.10). Although the total number of HBFC borrowers is still on the higher side, having the lowest average loan size makes its overall market share not com- mensurate with its high number of borrowers. 230 Appendix D D Figure D.9 Competition between Banks and the HBFC--Market Shares, 2004­08 90 80 70 60 percent 50 40 30 20 10 0 2004 2005 2006 2007 2008 year banks HBFC Source: SBP 2008b. Figure D.10 Commercial Banks Versus the HBFC--Number of Borrowers, 2003­08 180,000 160,000 140,000 120,000 borrowers 100,000 80,000 60,000 40,000 20,000 0 December December December December December December 2003 2004 2005 2006 2007 2008 month and year commercial banks HBFC total Source: SBP 2008b. 231 Appendix D D Table D.5 Concentration in the Mortgage Market, 2003­07 Total 80% Number Total Amount Number Share of Banks/ Herfindahl Year Outstanding of Banks/ (PRs DFIs with Concentration Concentration (June) (PRs millions) DFIs millions) 80% Share (%) Index 2003 2,372.3 14 1,898 3 21 0.259 2004 8,267.4 22 6,614 6 27 0.132 2005 26,699.2 24 21,359 10 42 0.095 2006 43,839.2 29 35,071 10 34 0.087 2007 57,541.8 29 46,033 9 31 0.094 Source: SBP. Note: This analysis is conducted only on commercial banks and DFIs; the HBFC is excluded. Commercial Bank Lending9 The mortgage market at inception can easily be characterized as concentrated. With the exception of the HBFC, only 3 of 14 commercial banks extended the bulk (80 percent) of housing loans during fiscal 2003/04 (table D.5). As more banks entered the mortgage market, the concentration of banks lessened somewhat: the number of banks participating increased from 24 to 29 during 2006, as Islamic banks entered the banking sector. In 2009, nine banks held an 80 percent share. The Herfindahl Concentration Index for the Pakistani banking sector--a meas- ure of the size of firms in relation to the industry in which they operate and an indicator of the amount of competition among them--has experienced a progres- sive improvement (figure D.11). The index, computed for five years, shows that competition in Pakistan among commercial banks and DFIs (excluding the HBFC) actually increased over time, with the index falling from 0.259 (highly concentrated) in 2003 to 0.132, 0.095, 0.087, and 0.094 in 2004, 2005, 2006, and 2007, respectively. The mortgage market, although in its embryonic stage, enjoys a fair degree of com- petition. This calls for a cautionary attitude to portfolio quality, however. Desire to capture competitive market share can obscure good lending practices and build up bad loans. Table D.6 lists banks and DFIs (excluding the HBFC) that loan portfolios exceed- ing PRs 1 billion in the housing finance market. These institutions collectively con- stitute 80 percent of the housing finance market. There is also a clear mix of local, public, and foreign banks, with Bank Alfalah, National Bank of Pakistan, and Stan- dard Chartered Bank being the top three dominating the market. Overall, statistics suggest that banks remain very cautious in their lending strat- egy to the housing sector. Housing-related risk is presently manageable because of the relatively small size of the housing finance portfolio. The limited exposure enables banks to manage liquidity risk by leveraging their stable deposit bases, but this is not sustainable in the long term because bank deposits have relatively short 232 Appendix D D Figure D.11 Herfindahl Concentration Index for the Banking Sector 0.30 0.25 Herfindahl Concentration Index 0.20 0.15 0.10 0.05 0 2003 2004 2005 2006 2007 2008 year Source: Kamran Shehzad, deputy governor of the SBP, at a 2009 workshop on housing finance in South Asia, Karachi. maturities (that is, the bulk of a banks' deposits are less than one year in duration). In addition, the active lenders are typically the smaller private commercial banks with limited branch networks that compete for deposits with large nationalized banks. As a result, banks are currently forced to restrict their housing loans to rela- tively shorter maturities. In part, the banks' cautious approach to lending is because they are still attract- ing the requisite expertise and resources to exploit this business area and manage its associated risks. The requisite expertise to undertake housing finance transactions has yet to become widely available among market players in the financial system, in related institutions and industries, and among regulatory or supervisory bodies. The financial sector lacks experienced sales and origination staff, especially in high- impact areas like construction and developer finance. The lack of experience and expertise impedes this reorientation of the financial sector approach to using mar- ket mechanisms and actively mobilizing private sector resources for investment in housing. Some bankers remain focused on making large loans to the manufacturing/ industrial sector, and they do not understand the centralization/standardization/ portfolio approach that is required for consumer finance. Specialized Mortgage Finance In Pakistan, the HBFC is the only specialized housing finance institution that pro- vides loans for middle- and low-income housing. In the past, some other housing finance companies were established--namely, Asian Housing Finance Limited, 233 Appendix D D Table D.6 Banks and DFIs with Housing Finance Portfolios Exceeding PRs 1 Billion, March 31, 2008 Number Weighted Average Amount of Loans Market Average Maturity Bank (PRs millions) Outstanding Share (%) Markup Period (years) Bank Alfalah Ltd. 10,559 4,644 15.9 12.50 10.6 National Bank of Pakistan 8,513 5,370 12.8 12.45 16.2 Standard Chartered Bank 7,210 1,764 10.9 13.38 12.2 United Bank Ltd. 5,513 1,212 8.3 13.68 15.8 Faysal Bank 5,364 1,079 8.1 13.90 14.9 Askari Bank 4,251 1,517 6.4 11.33 12.8 MCB 3,799 914 5.7 12.00 15.0 ABN AMRO 3,610 1,211 5.5 14.80 18.8 Meezan Bank 3,372 1,285 5.1 13.70 14.1 Dubai Islamic Bank 3,287 467 5.0 13.75 16.5 Habib Bank Ltd. 2,463 1,633 3.7 13.33 13.3 Citibank 1,753 718 2.6 15.67 13.0 NIB 1,406 1,209 2.1 13.00 11.1 Bank of Punjab 1,222 1,349 1.8 12.02 9.2 Other banks 4,059 2,361 6.1 n.a. n.a. Total (excluding HBFC) 66,380 26,733 100.0 n.a. n.a. Total outstanding 76,719 124,358 n.a. n.a. n.a. mortgage debt Source: SBP. Note: n.a. not applicable. Citibank Housing Finance Company, and International Housing Finance Limited. These specialized firms could not compete in the market, primarily because long- term funds were not available at attractive rates. Gradually, they moved out of the business. Because the SBP is now working to set up a long-term liquidity institution (Pakistan Mortgage Refinance Company) with technical and possibly financial assis- tance from the World Bank Group, it is expected that the market may see the emer- gence of some more specialized housing finance companies. Islamic Mortgage Finance Islamic banking in Pakistan is growing substantially and has massive untapped potential. In housing finance, specifically, loans increased from 1,675 in December 2007 to 2,144 in March 2009--in monetary terms, from PRs 6.0 billion to PRs 8.7 bil- lion (43 percent). But there is a need to introduce more Shar'ia-compliant products to satisfy customers' varied needs. Islamic banking remained resilient during the 2008 global crisis, despite negative shocks and pressure on the rupee experienced by the banking sector in October 2008. The portfolios of Islamic banks are performing bet- ter than those of conventional banks in terms of nonperforming loans. 234 Appendix D D Figure D.12 Growth in Islamic Assets, Deposits, Financing, and Investment a. Islamic banking, 2003­08 300 250 200 PRs 150 100 50 0 December 2003 December 2004 December 2005 December 2006 December 2007 December 2008 month and year total assets deposits financing and investments b. mode of financing, December 2008 istisna salam 2.90% 1.80% others 1.80% diminishing mushakata 30.50% murabaha 40.60% mudaraba ijarah 0.20% 20.50% mushakara 1.70% Source: SBP 2008b. Islamic assets, deposits, financing, and investment have shown increasing growth trends over the past several years (figure D.12). The number of branches of Islamic banking institutions has crossed the 500 mark, and Islamic banking grew at a yearly average of 37 percent. The respective shares of Islamic banking assets and deposits are 4.9 percent and 4.8 percent of the banking sector. On the face of it, these figures look modest; but given the short time frame and other countries' experiences, the 235 Appendix D D growth has been significant (SBP 2008). Islamic housing finance is growing very fast (121 percent over 2008) because of new entrants and because of increased consumer preferences for Islamic banking. One of the most popular housing products in Islamic finance is diminishing musharakah, a declining-balance/shared-equity partnership. With this product, both parties share the risks and obligations related to the property. Ownership of the property is divided into equal units and is purchased regularly by the customer. Under this arrangement, leasing and selling contracts must be conducted separately. The main steps involved in diminishing musharakah require the consumer to select a house and obtain approval for financing. The financier and consumer then jointly buy the house and obtain joint-ownership title through a musharakah agreement. The consumer leases the house from the financier, and eventually buys out the financier's equity, getting sole title to the property. Prudential Framework for Housing Finance, and Recent Policies Commercial banks and DFIs in Pakistan are regulated by the SBP under the Bank- ing Companies Ordinance 1962. Furthermore, banks and DFIs engage in housing finance under the auspices of prudential regulations. The promulgated prudential regulations provide specific rules for banks and DFIs engaging in housing finance activity. To encourage those institutions to increase their investments in the housing and construction sector, the SBP has issued the following new regulations for hous- ing loans (box D1): · Banks and DFIs are required to determine the housing finance lending limit, in both urban and rural areas, in accordance with their internal credit policy, client creditworthiness, and borrowers' loan repayment capacity. · Banks and DFIs are required to ensure that the total monthly amortization pay- ments of consumer loans, including housing loans, do not exceed 50 percent of the net disposable income of the prospective borrower. · Banks and DFIs may originate a housing loan at a debt-to-equity ratio of 85:15 for up to 20 years. · The maximum housing finance limit of PRs 10 million per owner or owning group has been removed. · To curb speculation, banks and DFIs are not allowed to provide housing finance purely for the purchase of land plots. · Banks and DFIs are allowed to provide housing loans amounting to no more than 10 percent of their net advances. · Banks and DFIs are encouraged to develop floating-rate housing finance products. · A housing unit financed by a bank or a DFI should be mortgaged in favor of the bank/DFI by way of an equitable or registered mortgage. 236 Appendix D D Box D.1 The Central Bank's Prudential Regulations Following are the up-to-date prudential regulations for housing finance (part of the consumer finance prudential regulations) pertaining to banks and DFIs. They are taken from the SBP Web site, http://www.sbp.org.pk/. Regulation R-15 Banks/DFIs shall determine the housing finance limit, both in urban and rural areas, in accor- dance with their internal credit policy and the creditworthiness and loan repayment capacity of the borrowers. At the same time, while determining the creditworthiness and repayment capac- ity of the prospective borrower, banks/DFIs shall ensure that the total monthly amortization payments of consumer loans, inclusive of a housing loan, should not exceed 50 percent of the net disposable income of the prospective borrower. Banks/DFIs will not allow housing finance purely for the purchase of land/plots; rather, such financing would be extended for the purchase of land/plot and for construction on it. Accord- ingly, the sanctioned loan limit--assessed on the basis of a borrower's repayment capacity, the value of the land/plot, and the cost of construction on it--should be disbursed in tranches; that is, up to a maximum of 50 percent of the loan limit can be disbursed for the purchase of land/plot, and the remaining amount can be disbursed for construction on the land/plot. Furthermore, the lending bank/DFI will take a realistic construction schedule from the borrower before allowing disbursement of the initial loan limit for the purchase of the land/plot. Banks/DFIs may allow housing finance facility for construction of houses against the security of land/plot already owned by their customers. However, the lending bank/DFI will ensure that the loan amount is used strictly for construction and that the loan is disbursed in tranches according to the construction schedule. Loans against the security of existing land/plot, or for the purchase of a new piece of land/plot, for commercial and industrial purposes may be allowed. But such loans will be treated as commercial loans, covered either under prudential regulations for corporate/commercial banking or prudential regulations for small and medium enterprise financing. Banks/DFIs may allow housing loans in rural areas, provided all relevant guidelines and regulations on the subject are complied with by them. Regulation R-16 The housing finance facility shall be provided at a maximum debt-equity ratio of 85:15. Regulation R-17 Banks/DFIs may extend mortgage loans for housing up to any tenure defined in the bank's/DFI's duly approved credit policy, keeping in view the maturities profile of their assets and liabilities. Regulation R-18 The house financed by the bank/DFI shall be mortgaged in bank's/DFI's favor by way of an equi- table or a registered mortgage. (Continued ) 237 Appendix D D Box D.1 The Central Bank's Prudential Regulations (continued ) Regulation R-19 Banks/DFIs shall either engage professional expertise or arrange sufficient training for their con- cerned officials to evaluate the property, assess the genuineness and integrity of the title doc- uments, and so forth. Table DD.1 Mortgage Loan Classification Classification Definition Interest Treatment Provisioning Substandard Where markup/ Unrealized markup/ Provision of 10% (25% from interest or principal interest to be kept December 31, 2006) of the is overdue by in Memorandum difference resulting from the 90 days or more Account and not outstanding balance of principal from the due date credited to Income less the amount of liquid assets Account, except when realizable without recourse to a realized in cash; court of law and adjusted forced unrealized markup/ sale value of mortgaged/ interest already taken pledged assets, as valued by to income account to valuers on the approved panel be reversed and kept of the Pakistan Banks' in Memorandum Association. Account Doubtful Where As above Provision of 50% of the markup/interest or difference resulting from the principal is overdue outstanding balance of principal by 180 days or less the amount of liquid assets more from the due realizable without recourse to a date court of law and adjusted forced sale value of mortgaged/ pledged assets, as valued by valuers on the approved panel of the Pakistan Banks' Association. Loss Where As above Provision of 100% of the markup/interest or difference resulting from the principal is overdue outstanding balance of principal by one year or less the amount of liquid assets more from the due realizable without recourse to date a court of law and adjusted forced sale value of mortgaged/pledged assets, as valued by valuers on the approved panel of the Pakistan Banks' Association. Source: Authors' compilation. (Continued ) 238 Appendix D D Box D.1 (continued ) It may be noted, however, that the requirements for full-scope and desktop evaluation (under R-8 and R-11 of Prudential Regulations for Corporate/Commercial Banking and Small And Medium Enterprise Financing, respectively) will not be applicable to housing finance. Regulation R-20 The bank's/DFI's management should put in place a mechanism to monitor conditions in the real estate market (or other product market) on at least a quarterly basis to ensure that its policies are aligned with current market conditions. Regulation R-21 Banks/DFIs are encouraged to develop floating-rate products for extending housing finance, thereby managing interest rate risk to avoid its adverse effects. Banks/DFIs are also encouraged to develop an in-house system to stress-test their housing portfolio against adverse movements in interest rates and maturity mismatches. Regulation R-22 Mortgage loans shall be classified and provided for in the manner described in table DD.1. There is a need to revamp exiting prudential regulations in line with the industry requirements. Specifically, it is expedient to introduce a comprehensive set of housing finance guidelines to boost the housing finance sector in Pakistan. These guidelines should be focused on key facets of the housing finance sector and should provide the necessary details and guidance to all key stakeholder to conduct prudent and supportive housing activities. The SBP is in the process of reviewing existing prudential regulations in order to launch a comprehensive set of housing finance guidelines and the much-needed mortgage finance­specific prudential regulations in the coming year. Recent Regulatory Policies on Housing Finance The government of Pakistan has taken active policy measures and practical initiatives to provide an enabling framework for housing and housing finance. The Financial Institutions (Recovery of Finances) Ordinance, 2001, established a new (nonjudicial) procedure for recovering secured claims. The National Housing Policy was published the same year (table D7). The regime of mortgage lending was liberalized and 239 Appendix D D Table D.7 Action Plan of the 2001 National Housing Policy, Implementation Status Policy Measures/Action Required Implementation Status (update) Financial institutions shall be The regulatory regime for the provision of housing finance by encouraged to give mortgage banks/DFIs to a cross-section of the population has been greatly loans for housing purposes at liberalized. Accordingly, the maximum debt-to-equity ratio for housing market rates. loans has been increased from 70:30 to 85:15, and the maximum loan tenure for housing finance has been increased from 15 years to 20 years. Moreover, as of March 19, 2005, the maximum per-party limit of PRs 10 million in respect to housing finance was removed; and banks/DFIs have been allowed to determine the housing finance limit in accordance with their internal credit policies and the credit- worthiness and loan repayment capacity of the borrowers. While determining creditworthiness and repayment capacity, banks/DFIs are required to ensure that the total monthly amortization payments of consumer loans, inclusive of housing loan, should not exceed 50 per- cent of the net disposable income of the prospective borrower. All commercial banks shall be The maximum exposure of banks/DFIs to housing finance has been motivated to advance loans enhanced from 5 percent to 10 percent of their net advances; how- for housing and housing proj- ever, it is up to banks to decide if and when they plan to offer housing ects by earmarking a substan- finance to various strata of society and to undertake financing of tial percentage of their loan housing projects. portfolios. Financial institutions and Banks/DFIs are allowed to securitize mortgage/construction/developer housing financial institutions finance through special-purpose vehicles. Moreover, the minimum shall be encouraged to float credit rating for banks/DFIs to make direct investment and for taking long-term bonds at market exposure (that is, undertaking lending and reverse repo) against listed rates. and unlisted asset-backed securities for mortgage/construction/ developer finance is reduced from "A" to "A­" (or equivalent). Housing refinance window The Finance Division approved the final scheme for setting up a shall be set up at the SBP for housing refinance window at the SBP, with the loans amounting to long-term funds from multilat- PRs 450 million from multilateral institutions for low-cost housing. eral agencies. Under the scheme, the SBP is acting only as an agent of the govern- ment of Pakistan for providing funding facility to housing finance companies. The scheme was circulated accordingly by the SBP in October 2001; however, housing finance companies failed to fulfill conditions under the scheme, so the funds have not been used. It is important to note that, in accordance with the instructions of the Finance Division on March 30, 1995, the loan under the housing refinance window of PRs 450 million was credited to the government of Pakistan. Therefore, utilization of the aforesaid amount requires permission from the government. Housing finance institutions The Microfinance Institutions Ordinance allows MFBs/MFIs to take shall be encouraged to deposits and to appoint agents, such as NGOs, to undertake activities promote savings and provide that are essential for proper discharge of their functions (these microloans for low-income functions include savings mobilization and provision of credit to the groups through NGOs marginalized strata of the society). Therefore, in the realm of microfi- and community-based nance, MFBs are allowed to develop products that not only help organizations. mobilize savings, but also have a link with the provision of housing 240 Appendix D D Table D.7 (continued ) Policy Measures/Action Required Implementation Status (update) finance. Although housing finance companies are regulated by the SECP, the HBFC comes under the regulatory purview of the SBP. Moreover, in accordance with Section 21 of the HBFC Act of 1952, the SBP has inti- mated to the Finance Division that the HBFC can undertake issuance of certificates of investments and float long-term bonds. At the same time, the SBP has already advised the government of Pakistan to declare special-purpose vehicles as financial institutions under the 2001 Recov- ery Ordinance, and to grant the HBFC permission to institute asset secu- ritization through special-purpose vehicles. HBFC and other financial institu- The regulatory regime has been liberalized for the promotion and tions shall formulate packages development of a market-based housing finance system in the coun- of preferential/concessional try, and it is up to banks to decide if and when they plan to offer hous- rates, with an affordable sys- ing finance to different strata of society. Although the banks have tem of installments for repay- increased their exposure to housing finance since March 2003, ment, to provide affordable banks/DFIs are facing a number of critical impediments to providing credit to low-income groups. housing finance to the general public. These impediments include, among other things, lack of proper property titling, a land registration information system, and implementation of foreclosure laws. More- over, in the case of low-income groups, the impediments are further compounded by difficulties in ascertaining monthly incomes because there are no mechanisms/documentation for determining the monthly incomes of low-income groups, and most low-income dwellings are situated in shantytowns and irregular settlements without legal/clear title documents. The aforementioned major constraints are not only impeding countrywide growth and development of housing finance to the general public; but also holding back the trickle-down of housing loans to the low-income groups for whom the cost of financial inter- mediation is higher than that for high- and middle-income groups. Foreclosure laws shall be intro- According to Section 15 (sale of mortgaged property) of the Financial duced to ensure effective Institutions (Recovery of Finances) Ordinance (promulgated by the recovery of loans and advances government of Pakistan on August 30, 2001), after service of notice of from the defaulters. demand by a financial institution, all the powers of the mortgagor with respect to recovery of rents and profits from the final mortgage property stand transferred to the financial institution. Accordingly, under the provisions of Section 15(4), the financial institution, after expiry of final notice, can sell the mortgaged property or any part thereof by public auction and appropriate the proceeds thereof toward total or partial satisfaction of the outstanding mortgaged money, without inter- vention of any court. In doing so, neither the banking court nor the high court shall grant an injunction restraining the sale or proposed sale of mortgaged property, unless courts are satisfied that no mortgage was created or the secured sum has been fully paid or the mortgagor has deposited cash in the banking court. The foregoing provision gives more powers and remedies to the lending institutions, which suggests that foreclosure laws have been strengthened and would help improve the recovery process of financial institutions' defaulted loans. Source: 2001 National Housing Policy, Pakistan Ministry of Housing and Works. Note: MFB microfinance bank; MFI microfinance institution; SECP Securities and Exchange Commission of Pakistan. 241 Appendix D D strengthened by lifting various regulatory restrictions. An appropriate prudential framework for housing finance was designed, and credit information services were introduced. The SBP took a vanguard position for providing the institutional arrangements and reforms to housing finance, to support the shift of housing funding from gov- ernment to the banking system. Since 2002, with an objective of enhancing access to market-based housing finance, the SBP has been playing a facilitating role in associ- ation with private and public sector stakeholders (1) to increase the number of chan- nels through which housing finance can be provided to the general public in the marketplace, (2) to enhance the general public's ability to afford housing finance products and services, and (3) to improve the service quality of housing finance. As part of this effort, the SBP constituted the Housing Advisory Group in 2002 to con- duct a thorough analysis of existing housing and housing finance policies and conditions and to formulate a concrete strategy to address key challenges (table D8). The group's specific activities have included providing assistance in developing long- term funding/liquidity mechanism for housing, conducting capacity-building pro- grams on housing issues for the banking and financial system, monitoring the flow of credit by banks and DFIs in the areas of infrastructure and housing finance, and analyzing emerging trends and developments in the housing and housing finance industries. The Housing Advisory Group, through its first set of recommendations, empha- sized the following actions: liberalization of the credit regime; provision of fiscal incentives to mortgagors and mortgagee financial institutions, streamlining of data related to banks' and DFIs' housing finance portfolios, reduction of the SBP's minimum threshold of loans on which information is collected by the Credit Infor- mation Bureau (CIB), implementation of foreclosure laws, and other essential meas- ures. The implementation of these recommendations and the keen interest of all public and private sector stakeholders resulted in significant growth in housing finance lending by banks and DFIs. The SBP established a separate department--the Infrastructure and Housing Finance Department--under the Development Finance Group of the SBP, and charged it with creating an enabling environment to facilitate intermediation of market-based credit in housing finance. In the federal budget for fiscal 2003/04, fis- cal incentives were granted to facilitate mortgagors and banks. The government further initiated several mega-housing projects in big cities to increase the housing supply (see the "Government Housing Programs" section below). The Housing Advisory Group has placed particular emphasis on the need for a well-structured and efficient legal framework and property development process. Therefore, there is a need not only to reform the present procedures and practices of the real estate sector, but also to have a supportive institutional framework for enhancing the general public's access to housing finance. The group has taken a long-term view of expanding the scope and outreach of housing finance by focus- ing on a few key, broad-based areas. An overhaul of the legal and regulatory frame- work for housing is required, coupled with the institutional reform of property 242 Appendix D D Table D.8 Housing Advisory Group Recommendations Recommendation Action Plan Implementing Agency Reforming legal frame- A) Implementing Recovery Ordinance · Ministry of Law, Justice, and work affecting foreclosure, 2001 Human Rights transfer, tenancy, rent con- B) Amending Recovery Ordinance · Provincial governments trol, and acquisition of 2001 immovable property C) Establishing an electronic public registry D) Transferring property only through a registered instrument E) Rewriting urban rent laws F) Amending the Land Acquisition Act Rationalizing transaction A) Rationalizing the cost of stamp · Provincial governments costs duty and registration fee on con- · City district governments veyance and mortgage deeds B) Rationalizing commercialization fee C) Automating land records and improving the tax collection system Establishing an integrated, A) Streamlining the role of the public · National Reconstruction nationwide system of land sector Bureau registration information B) Promoting area-specific private · National Database and sector initiatives Registration Authority C) Building the capacity of the · Provincial governments existing system · City district governments · Landowning agencies · Central Board of Revenue · Central Depository Company Increasing supply of land A) Private sector using existing · City district governments for affordable housing allotted land · Landowning agencies B) Reforming katchi abadis · Provincial governments C) Pursuing vertical land allotment · Ministry of Law, Justice, and policy to diversify Human Rights D) Government auctioning existing land E) Creating city-based land banks and satellite towns F) Amending the Land Acquisition Act Structuring and streamlin- A) Registering real estate builders · City district governments ing large-scale developer and developers with ABAD · Landowning agencies finance B) Introducing credit rating · Association of Builders and C) Government schemes on com- Developers pleted housing/project basis · Financial institutions D) Encouraging real estate invest- · Pakistan Banks' Association ment trusts · Credit rating companies E) Encouraging joint ventures with foreign developers (Continued ) 243 Appendix D D Table D.8 Housing Advisory Group Recommendations (continued ) Recommendation Action Plan Implementing Agency F) Simplifying the documentation · Mortgage Bankers Association required by banks G) Standardizing valuation methodologies Providing housing finance A) Developing innovative solutions · Financial institutions, such as for affordable and low- B) Establishing anonymous and housing finance companies, cost/low-income housing self-regulating structures microfinance banks, and NGOs facilities (Condominium Act) · Ministry of Law, Justice, and C) Amending the Land Acquisition Human Rights Act · Provincial governments · City district governments Promoting and developing A) Disseminating real estate invest- · Ministry of Finance real estate investment ment trust rules · Ministry of Law, Justice, and trusts B) Amending Rent Control Laws Human Rights C) Rationalizing transaction costs · Securities and Exchange D) Establishing anonymous and self- Commission of Pakistan regulating structures (Condo- · Provincial governments minium Act) · Islamabad Capital Territory E) Establishing a tax framework for real estate investment trusts Providing long-term fund- A) Establishing a mortgage refinance · Ministry of Finance ing for housing loans company · SBP · Securities and Exchange Commission of Pakistan · International financial institutions · Primary mortgage lenders Developing housing mar- A) Upgrading Web sites of primary · Pakistan Bureau of Statistics ket information systems mortgage lenders · Mortgage Bankers B) Mortgage Bankers Association Association supporting a dedicated Web site · Primary mortgage lenders for the housing market information system C) Setting up a housing observatory Developing a mortgage A) Building the capacity of the · Mortgage Bankers banking industry Mortgage Bankers Association Association · Primary mortgage lenders · Various private sector play- ers, such as real estate builders and developers, real estate brokers, insurance companies, and pension and provident funds Source: Authors' compilation. 244 Appendix D D acquisitions and transactions, the establishment of an integrated nationwide land registration information system, and the creation of a housing market information system. The framework for financing large-scale property development should be structured and streamlined,; new products and services related to housing develop- ment and finance are needed, including real estate investment trusts (REITs); there should be enhanced provision of long-term funding for housing loans; and active development of the mortgage banking industry should be encouraged. In the area of low-income housing, increased supply of land for affordable housing is essential, as is provision of housing finance for affordable and low-cost/low-income housing facilities. Homeowners' Insurance To secure the housing loans against loss of property or life, the insurance companies provide property insurance and life insurance. Mortgage insurance and title insur- ance are not currently available in the market. The lack of mortgage credit insurance facilities seems to result from two sources: the risky mortgage industry milieu and regulatory compulsions on one side, and borrowers' inability to provide documented evidence of income and ability to repay on the other side. Risk Management In Pakistan, banks are in the gradual process of implementing Basel II; and non- bank institutions must establish policies, procedures, and practices to define risks, stipulate responsibilities, specify security requirements, design internal controls, and ensure compliance.10 In spite of the small share of housing finance in banks' portfolios, those loans cause exposure to credit risk, and to liquidity risk resulting from the term-maturity mismatch (given short-term deposits and long-term lend- ing). Credit risk is being managed by gathering increasingly better client informa- tion (see the "Credit Information" section below), as well as by limiting mortgage finance to top-income clients. The issue will take on renewed importance as the market moves down the income scale, requiring innovative products and approaches to mitigate credit risk. At present, banks are mitigating interest rate risk by offering adjustable-rate mortgages through which the interest rate risk is borne by the borrower. Floating-rate mortgages, however, lower the affordability of mort- gage finance for middle- and lower-income borrowers--especially given a volatile environment with high interest rates. This could further translate into a credit risk in a rising-interest-rate scenario so as not only to restrain housing demand, but also to pose a systemic risk. Financing long-term mortgages with short term funds is not sustainable, and eventually would lead to a credit crunch and instability in the financial system. The lack of a viable funding structure for housing loans greatly constrains the activity of housing finance companies, and of the HBFC specifically. Smaller housing finance institutions operating mostly in the private sector are unable to raise long-term 245 Appendix D D funds. This creates a situation in which they are forced to lend for shorter terms, all the while bearing an exposure to their mismatched maturity positions. Commercial banks actively engaged in providing mortgage financing are better able to cope with the liquidity risk, relying on demand and time liabilities; and are able to benefit from higher yields. However, this poorly mitigated exposure creates interest rate risk and maturity mismatches, and will prove increasingly costly in a tightening prudential regulatory environment. At present, the option of going for outright securitization is remote because of information asymmetries and cost constraints. The capital markets and market- based funding sources have not yet developed for long-term instruments of finance. In a promising recent development, the HBFC went to the market for funding for the first time, floating a six-year sukuk11 at an attractive rate. For most private financial institutions, however, this would be difficult. A viable option, contemplated by policy makers in Pakistan, is establishing a refinance facility that could provide liquidity to the financial institutions with recourse. Challenges in creating such a structured funding mechanism include the low volume of long-term mortgage loans, compared with the total deposit base of all banks; the general use of cost-efficient funding for housing through demand and time liabilities; and the broad spread between weighted average cost of deposits and price of housing loans. The refinance facility would help in providing fixed-rate mortgages, improving the liquidity of the financial system, and enabling banks and housing finance com- panies to prudently match the maturity profile of their assets and liabilities. It also would facilitate development of a private debt market and open up investment opportunities for long-term institutional investors (such as pension and provident funds) and insurance companies, which are looking for better returns on their investment portfolios. The facility, in turn, would serve as a preparatory step in developing a secondary mortgage market in Pakistan. More important, setting up a mortgage refinance company would act as an "equalizer" in a present environment where banks with large balance sheets have an advantage over medium-size banks and housing finance companies when originating and funding housing loans. When a mortgage refinance company is established, medium-size banks and housing finance companies would be in a position to compete with the big commercial banks on the bases of price, product features, and service quality--thereby improving the competitiveness of the housing finance market and increasing the outreach of hous- ing finance, especially to the low- and middle-income groups. Growth Potential of the Mortgage Debt Industry Figure D.13 shows the historical mortgage debt-to-GDP ratio in Pakistan. The ratio has risen only slightly from 0.49 percent since 2003, reaching its peak at 0.98 percent during 2006 and following a decreasing trend through 2008. This highlights the huge untapped potential of the market. Projecting future effective demand for housing finance for 2009, it is clear that there is demand 470,000 housing units in urban Pakistan alone (table D.9). If one 246 Appendix D D Figure D.13 Debt-to-GDP Ratio, 2003­09 1.4 1.2 1.0 0.8 percent 0.6 0.4 0.2 0 2003 2004 2005 2006 2007 2008 2009 (potential) year Source: Authors' calculations based on GDP and economy data from the SBP. Table D.9 Estimated Urban Mortgage Potential, 2009 Urban housing needs (units) 470,000 At average cost of PRs 1.0 million/unit PRs 470 billion Mortgage finance potential at 35% of urban PRs 165 billion (1.26% of GDP)a housing needs Sources: Pakistan's 1998 national census; authors' calculations. a. Taking GDP at PRs 13,065 billion for fiscal 2007/08 (or $167.5 billion, where $1 PRs 78). housing unit costs, on average, PRs 1 million, then total construction cost will be PRs 470 billion for the needed units. Assuming only 35 percent of the units are paid for through mortgage finance, there will be a requirement of PRs 165 billion in financing (1.26 percent of GDP). The effective potential demand for housing financ- ing is even greater beyond 2009. Low-Income and Microfinance Housing Low-cost and low-income housing is, by far, the most neglected area of housing and housing finance in Pakistan. 247 Appendix D D Challenges for the Development of Low-Income Housing Finance The lack of finance is primarily a supply problem. Only 1­2 percent of all housing transactions in Pakistan are processed through housing finance credit lines; the rest are funded via personal savings and informal borrowing. Key constraints to housing are poor people's lack of financing capacity and a low penetration of housing finance, especially in the poorest segments of the population. In turn, the high and ever-increasing supply gap has yielded a very high room density--that is, 3.5 people per room. That density is alarming when compared with the international standard of 1.1 per room. The majority of commercial banks' loans are focused on the high-income group in major cities. The HBFC remains the only institution catering to groups lower down on the income scale: its average loan size is PRs 860,000, compared with PRs 2.35 million among commercial banks. Nonethe- less, most HBFC loans are focused on the middle-, not lower-, income groups. National per capita income is around PRs 7,300 a month, and the minimum wage is PRs 6,000 a month. More than two thirds of the country's population fall below the mortgage affordability threshold (that is, PRs 200,000). This, coupled with the massive backlog of housing (7.57 million units), has led to a situation in which nearly three quarters of the population is unable to access the services of the hous- ing and housing finance sectors. The present "retail" approach to providing affordable housing facilities has failed to address the burgeoning gap between housing unit demand and supply. Several factors impede the success of a market solution. First, the legal framework is not con- ducive to large-scale access to housing finance ("Mortgage Market Infrastructure [Regulatory Framework]" section). Property titling in low-income areas is compar- atively more difficult because a representative of a mortgagee finance institution must visit at least five different offices to verify title documents. Documentation is poor because the level of illiteracy is high. The procedure is bureaucratic and char- acterized by weak governance. Land records are not computerized, so it is a challenge for mortgagee institutions to ascertain who holds true title to the property at the time of loan origination. It is also problematic to foreclose on property subsequent to loan delinquency. Second, the recent speculative wave in real estate has further exacerbated the sit- uation because land prices have skyrocketed, thereby eroding the general public's ability to afford housing (especially low-income groups). Third, rental markets are not extended to lower-income groups because of the poor profitability of house rentals. High interest rates and low rental prices create a disparity between rental income and mortgage costs, thereby discouraging investors from placing their funds in rental properties. Generous tenancy laws worsen the situation. Providing fiscal incentives, implementing recovery laws, and instituting investor-friendly tenancy laws could improve the prospects of constructing housing units, especially for low-income groups. Fourth, the intermediation costs of providing housing loans to low-income groups are higher than those for middle- and upper-income groups because the loan sizes are smaller, the incidence of circumstantial defaults is higher, and there is a 248 Appendix D D negative propensity to save. Income stability is an issue for low-income groups, and the risk of losing the income-generating source in the longer term is compounded. This is especially true, given that the main income earner is rarely insured for acci- dent, prolonged illness, or death. Documented verification of income is also difficult because most low-income people are informally employed. Fifth, low-cost housing construction is not a viable proposition in the current legal and economic framework. Initiation and successful closure of low-income res- idential projects are also marked with barriers to site permits, environmental clear- ance, and the lack of infrastructure and utilities connections. The Microfinance Market and Informal Housing Lending The microfinance market has a very limited outreach in Pakistan: barely 2 percent of the population. Microlending for housing is even more limited. Growth is brisk, however, starting from a small base. The number of active borrowers at microfinance banks stood at 564,892 in the quarter ended September 2008, increasing by 26 per- cent over the preceding year. The number of branches was 252, up 36 percent over the previous year. Similarly, the number of depositors rose to 254,381 by December 2008. The market is concentrated, with the central role being played by the two largest microfinance banks (SBP 2008a). Extending this overview to all microfinance institutions, table D.10 shows 1.8 million microloan borrowers, 2.0 million deposi- tors, and 2.1 million holders of microinsurance policies. Although most microfinance institutions in Pakistan focus on credit delivery of loans for agriculture, livestock, trading, and consumption, a few are beginning to diversify and innovations have begun in enterprise loans and housing finance. Still, housing represents a negligible proportion of microfinance borrowing. Housing finance companies and microfinance institutions, alongside NGOs, need to take the lead in initiating housing projects for low-income groups, with autonomous and self-regulating structures within their jurisdictions/boundaries, and perhaps calling on the experiences and lessons of successful cases (such as India's low-income hous- ing models). Table D.10 Microfinance Industry at a Glance Microcredit Microsavings Microinsurance Active Value Active Value Sum Borrowers (PRs Savers (PRs Policyholders Insured (n) millions) (n) millions) (n) (PRs millions) 2008, Q4 1,732,879 18,752 1,876,802 5,419 2,241,552 34,340 2009, Q1 1,751,111 19,252 2,043,774 5,882 2,128,493 31,232 Change (net) 18,232 500 166,972 463 113,059 3,108 Change (%) 1 3 9 9 5 9 Source: MicroWATCH, first quarter 2009. 249 Appendix D D NGO Programs in Housing and Housing-Related Social Services NGOs, in collaboration with governments and other stakeholders, can play a pivotal role in accelerating the process of economic development, socioeconomic change, and improvement in the society to produce a sustainable livelihood and overall pros- perity. Historically, low-cost housing in Pakistan has been the realm of NGOs, the government, or the informal sector. Because of the heavy capital requirements and the level of skill involved in scaling up in the housing sector, NGOs have not gained much ground in this regard. A few initiatives from Saiban and from Aga Khan Plan- ning and Building Services are the only visible steps in this direction. Saiban is an internationally recognized NGO that provides urban squatters with access to plots of developed land at affordable rates (Rahman and Hassan 2009). In December 2005, the NGO received Acumen Fund financing of $302,500 to launch a low-income housing development. The selected location was 20 acres of flat and arid land, a 40-minute drive (20 kilometers) from Lahore. The project involved the con- struction of 500 homes. The model provides secure residential tenure and public util- ities in areas where squatters normally would have no access to legal property. The Saiban approach directly addresses a severe livelihood challenge for lower-income groups--those earning between PRs 5,000 and PRs 15,000 a month. Saiban offers the possibility of legal entitlement to land, giving the poor an important piece of economic collateral--an asset for a population with no access to capital for economic growth. The project, named Khuda-ki-Basti-4 ("God's Own Settlement"), is a good example of success in providing low-income housing on market principles (see box D.2). 12 Aga Khan Planning and Building Services, established in 1980, is an independent executive body with the mandate to plan and implement infrastructure and technology­related development initiatives to improve peoples' living conditions.13 The NGO assists local communities of the Northern Areas, Punjab, North-West Frontier Province, and Sindh in such development program areas as habitat risk reduction, energy-efficient building and construction improvement, water supply and sanitation, and natural resources conservation. The organization takes an integrated, community-based approach to sustainable development; and its devel- opment programs address not only the immediate needs of clean water, adequate sanitation, and safe housing, but also the overall impact that these initiatives have on economic, social, and environmental sustainability. Capable, proficient, vibrant, and self-reliant community-based organizations underpin all developmental interven- tions. Poverty alleviation, environmental sustainability, gender equality, and eco- nomic regeneration serve as cornerstones of these initiatives. Government Housing Programs Over the years, the role of government in the housing sector has evolved from that of predominant provider to that of a facilitator of the housing sector. The share of housing in public sector programs has decreased progressively, from 10.9 percent in the 1960s to 5.9 percent in the 1990s. The government's Medium-Term Develop- 250 ment Framework envisages only PRs 30 billion out of a total of PRs 950 billion of Appendix D D Box D.2 Low-Income Housing Initiatives in Pakistan There are several low-cost housing projects that can be studied and replicated across the coun- try, using an improved approach and technology. The first government housing project, implemented in 1959­60, was the Korangi Landhi Township in Karachi, built under the supervision of Gen. Azam Khan. The project was completed within a record time of six months. The settlement scheme is globally acknowledged as one of the best housing solutions to rehabilitate squatters and the poor. The cost of a house, including the land, was a token PRs 35, payable in three installments. The Sujrani Town low-cost housing project in Karachi was developed in 1970 by the now- defunct Karachi Development Authority, exclusively for low- and middle-income groups. With 47,736 residential plots, the project catered to the housing needs of approximately 800,000 people. The complex consisted of one- and two-room houses with toilet, bath, and kitchen. Units were built by construction companies whose designs were found suitable and economical. The total cost of a house of 80 square yards was between PRs 45,000 and PRs 80,000. Additionally, the HBFC provided housing finance of up to PRs 20,000­36,000. The scheme was self-financing, and all development work was done on a breakeven basis. It took more than 15 years follow- ing construction to provide transport, utilities, health, and education services and facilities to the locality. The Khuda-ki-Basti Incremental Housing Project in Karachi is ongoing, and is located in Phase-III of Taiser Town. Through this project, plots are provided by Saiban, an NGO, to homeless and poorly sheltered groups. The price of a plot of 80 square yards (including all development charges) is mod- est, at PRs 37,000. A prospective homeowner is required to provide a very low up-front deposit of PRs 4,000 to become eligible for a plot, and must prove his or her destitute housing situation. Until that is accomplished, the applicant is allotted a plot to start the construction work; and Saiban provides a temporary shelter. Upon completion, the plot is leased in the name of the applicant. The Taiser Town low-cost housing scheme in Karachi is a future joint project of the City District Government Karachi and a U.S.-based private company. The plan is to construct earthquake-resistant, low-cost housing on a breakeven basis to provide affordable living facili- ties. The plots are 80 square yards. The city government would help prospective homeowners get affordable loans from the HBFC or other banks. The project involves building times of less than a month for each house, at an estimated cost of PRs 600 per square foot. Khuda-ki-Basti-4 in Lahore is a 20-acre colony with 12 blocks and 460 houses. The project was initiated in June 2006, construction work started in September 2006, and the first families had moved in by February 2007. Each house is 816 square feet, and costs PRs 80,000. The downpayment is PRs 40,000, with the balance paid in monthly installments of PRs 1,000. The implicit interest rate amounts to 12­13 percent. The HBFC will provide 70 percent of the funding, with Saiban providing the balance of housing finance. Heritage Homes in Lahore is a project of 2,500 housing units launched by the private devel- oper M/s Sehar Associates (Pvt.) Ltd. The project, conceptualized as a "sustainable neighbor- hood," was marketed as respectable living at affordable prices. The plots are 100 square yards, and cost PRs 690,000. Prospective homeowners are required to live in the project for a mini- mum of five years. Source: Rizvi 2007. 251 Appendix D D public sector investment to go toward housing during the period 2005­10. This low share is attributed to the government's resolve to play a facilitating/regulating role in accordance with the 2001 National Housing Policy. However, recent preoccupations with the increasing housing shortage for middle- and lower-income groups has moved the government to action, on a public-private partnership basis. In the fiscal 2007/08 budget, the government committed to mak- ing 250,000 low-cost housing units available over the next five years. The Ministry of Finance entrusted the HBFC with the preparation of a road map for implementing 250,000 housing units for middle- and low-income groups, and the HBFC is doing that now. Ministry of Housing and Works. The Ministry of Housing and Works is responsi- ble for housing improvements and construction (along with other public agencies, including the National Housing Authority, the Pakistan Housing Authority, the Pakistan Public Works Department, the Federal Government Employees Housing Foundation, National Construction Ltd., and the Estate Office). The ministry is responsible for acquiring and developing sites and constructing and maintaining federal government buildings. It is actively involved in coordinating civil works, budget, fixation and recovery of rents from government-owned/hired and govern- ment-requisitioned buildings. Its activities include managing federal lodges; handling matters related to federal government lands; issuing licenses to various cooperative housing societies in Karachi (except those under the different ministries); adminis- tering to officers belonging to the engineering group; and registering with the Hous- ing Authority, National Construction Ltd., and the Federal Government Employees Housing Foundation. Since 2008, the Ministry of Housing and Works has undertaken the planning of a mega-scale housing project (known as "Housing for All") to address the acute shortage of housing units. Under the plan, 1 million low-cost houses will be constructed throughout the country. The scheme calls for units with two rooms plus a bath, and includes quotas for federal government employees and a 5 percent quota for the employees of print and electronic media. Satellite towns will be built in various cities throughout the country, on a total cumulative area of 70,000 acres of land. For example, one of the low-cost housing schemes, planned to launch in Sindh, involves 300 acres in Karachi and 10 acres each in the towns of Bin Qasim, Gadap, and Kemari. According to the HBFC, the government has earmarked PRs 10 billion as a revolving fund for this initiative. A government committee has been entrusted with the tasks of coordinating the scheme with the provision of affordable mortgage financing or easy- installment financial plans for the middle- and lower-middle groups, and ensuring housing for government employees on or before their retirements. National Housing Authority. The National Housing Authority was established in 1987 with the primary objectives of preparing short- and long-term programs and policy guidelines and mobilizing financing resources for the development of hous- ing for low-income groups. It was subsequently attached as a department to the Min- istry of Housing and Works. 252 Appendix D D Pakistan Housing Authority. The Pakistan Housing Authority was established under a cabinet resolution in May 1999 to provide shelter to low-income groups, the poor, and the needy. The scheme was thoroughly reviewed later that year, resulting in the 2000 approval of 18 commercially viable and marketable projects. The proj- ects involved 4,476 housing units in four major urban centers of the country (Islam- abad, Karachi, Lahore, and Peshawar), at an estimated total cost of PRs 5 billion. The Pakistan Housing Authority is also undertaking housing projects based on plot ownership for the officers of the federally constituted occupational groups and for low-paid federal government employees. Federal Government Employees Housing Foundation. The Federal Government Employees Housing Foundation was established in 1989 to achieve the following objectives: · to formulate policies, programs, and projects for providing shelter to serving and retired federal government employees and their dependents; · to initiate, launch, sponsor, and implement housing projects on an ownership basis for federal government employees in Islamabad, the provincial capitals, and other major cities of Pakistan; · to undertake redevelopment of FG Housing Estates to increase housing stock and improve the living and environmental conditions; and · to do whatever is incidental or conducive to attaining any or all of the above objectives. Pakistan Public Works Department. The objectives of the Public Works Depart- ment are to provide offices and residential accommodations for federal govern- ment employees, and to overcome the backlog of millions of square feet of office space and housing units presently located in rented office buildings and privately owned residences (within budgetary provisions of the federal government). The Estate Office. The Estate Office, a department attached to the Ministry of Housing and Works, was established in 1947 at Karachi. The government inherited certain units of housing property at the time of independence, and the primary objective for establishing this department was to allot these housing units to the fed- eral government officers and staff then posted at Karachi (at that time, the capital). The business of the department is being carried out under Pakistan Allocation Rules, 1993. The important functions are summarized here: · provision of government accommodations to the federal government officers or officials in basic pay scales 1 to 22; · provision of government accommodations to the ministers, minister of state, advisers, and other authorities, under the rules; · assessment of rents for privately hired/requisitioned houses, with the help of the Public Works Department and the Finance Division; 253 Appendix D D · maintenance of general waiting lists of federal government workers in basic pay scales 1 to 22; · ejection of unauthorized occupants from government accommodations; · maintenance of payment records (cash books, ledgers, checkbooks, and so forth) for privately rented houses and buildings; and · maintenance of records for house rent charges recovered from the federal gov- ernment workers who are provided with government or privately rented houses. In implementing "Housing for All," the Pakistan Housing Authority out- sourced construction to a number of competitively selected contractors from Canada, China, Egypt, Germany, Malaysia, the United Arab Emirates, and the United States. These contractors have also invested their own equity up to 80 per- cent in the venture. Outside of this program, and following its experience, contracts have been concluded with counterparts from Indonesia, the Islamic Republic of Iran, the Republic of Korea, Indonesia, and Malaysia for he con- struction of other housing projects in such towns as Hyderabad, Quetta, Rawalpindi, Sukkur. These initiatives are part of a 10-year construction plan that relies on foreign direct investment coupled with a policy and investment frame- work friendly to foreign construction ventures. Examples of other housing initiatives are offered in box D.2. Regulatory Framework and Mortgage Market Infrastructure Several aspects of the way the legal system operates at present generate pervasive uncertainty about property transactions in Pakistan. An inefficient legal frame- work, fragmented ownership and titling procedures, inefficient land information systems, inaccessible and unused government land, high stamp duties and registra- tion fees, a weak tax framework, and ineffective land dispute mechanisms dampen the willingness of banks to increase their mortgage portfolios. Poor master plan- ning, multiple housing administrative institutions and procedures, problematic zoning regulations, restrictive building codes, and unreliable utility connections further hinder the property development process; and a fragmented building indus- try, unorganized real estate agencies, the lack of developer finance, and skewed tenant laws also limit the interest and capacity of the private sector to increase the supply of affordable housing. Land Registration and Administration, and Land Titling The present system of land records is fragmented: it is administered by various landowning agencies at the provincial, district, and city government levels and by autonomous land development authorities (box D3).14 Even within a province, 254 Appendix D D Box D.3 Land Titling System Description and Shortcomings The land titling system in Pakistan is rendered ineffective by several problems, including the lack of a sole land allotment authority/agency of the country/city, no coordination among different recordkeeping agencies, divergent recordkeeping systems and methods at different departments, and manual documentation without safety and security of records. The system for ascertaining the chains of titles is cumbersome and lengthy. The lack of transparency and the incentives for graft provoke the avoidance of official land registration. The various land allotment agencies work- ing in different cities use divergent procedures, rules, and documents (detailed in table DD.2). The main bottlenecks involve title validity, encumbrances, and title informality. Title Validity Under the land registration system, lenders have developed policies and procedures regard- ing title validity. The chain of title relating to any given property is ascertained by the series of documents through which the property has been transferred from one buyer to another. To verify the chain of title, borrowers are required to provide a number of (original) title doc- uments for each transfer relating to the property from the time the property was conveyed by the government. Title verification is then performed solely on the basis of the gathered documentation; the subregistrar does not undertake an independent search to confirm the chain of title. The complexity of this process can cause a lender to deny financing to gen- uine borrowers. Lenders are further challenged in verifying title with respect to authorities and corporations, because of the nonstandard system of title documentation in these cases; and they are driven to employ specialized lawyers (which raises their costs). Encumbrances There is no practical and realistic system to ascertain encumbrances on land/property. Mort- gage deeds are registered; however, the information is not properly maintained, and lenders have resorted to requiring that the original title be deposited with them until the loan is repaid in full. This precludes property financing in cases where the original title documents are not available to the owner. Title Informality Informal/unregistered land is also prevalent, mainly in smaller cities and on the outskirts of major cities such as Karachi and Lahore. Informal title to these lands has passed for genera- tions without any conveyance document being registered. In these cases, property rights are transferred through bills of sale, powers of attorney, and other such informal documents. Infor- mality has been further exacerbated by urbanization, with squatters usually taking up unused government or corporate lands. Over time, the inhabitants without formal title do perceive an entitlement to the land, and they invest considerable resources in housing and better living conditions. (Continued ) 255 Appendix D D Box D.3 Land Titling System Description and Shortcomings (continued ) Table DD.2 Land Allotment Agencies' Locations, Procedures, Rules, and Documents Procedures, Rules, Agency Location and Documents Urban areas Lahore, Rawalpindi Sale deed Abbottabad, Faisalabad, Multa, Registered sale deed Rahim Yar Khan Quetta Fard Hyderabad, Sukkur Permanent transfer deed Islamabad Allotment letter/sale deed Jhelum Sale deed/gift deed Bahawalpur, Muzaffargarh Sale deed/Sanad/gift deed Dera Ghazi Khan, Vehari Registered sale deed/gift deed/ registered partition deed and permanent transfer deed Government Lahore Sale deed/registered sale agreement schemes (urban Karachi, Sukkur Lease deed area private) Jhelum, Multan, Rahim Yar Khan Registered allotment order Dera Ghazi Khan, Rawalpindi, Vehar Registered sale agreement/ Bahawalpur allotment order Muzaffargarh Registered allotment order/sale agreement Original allotment order Defense Housing Lahore Transfer letter Authority Karachi Sublicense Evacuee property Karachi, Lahore Permanent transfer deed/provisional transfer deed Rawalpindi Permanent transfer deed Cantonment board Faisalabad Sale agreement/lease deed/ and development Quetta allotment order authorities (urban Hyderabad, Jhelum, Rawalpindi Registered final transfer deed area) Karachi, Sukkur Lease deed Multan Sublease/lease Lease deed extract from the Land Register Municipal committee Quetta Lease deed Rahim Yar Khan Agreement for sale Rural areas Hyderabad, Sukkur Ownership certificate Freehold properties Bahawalpur, Bannu, Dear Mutation Ismail Khan, Gilgit, Kohat, Mardan, Peshawar, Swat Source: Authors' compilation. 256 Appendix D D land titling and registration practices are fragmented. An acute example of this problem is the situation in Karachi, where more than 17 different agencies are responsible for the process. Registration is at a comparable level of efficiency in the region as a whole--although it compared poorly with best-performing coun- tries like Georgia and Saudi Arabia, where registration takes two procedures, two or three days, and carries either no cost (Saudi Arabia) or the minimal cost of 0.02 percent of property value (Georgia). In Pakistan, six procedures are neces- sary, it takes 50 days, and it costs 5.29 percent of property value to complete the registration (World Bank 2008a). Elsewhere in the region: in India, five proce- dures completed in 44 days at a cost of 7.43 percent of the property value; in Sri Lanka, eight procedures involving 83 days at a cost of 5.11 percent. The 2001 National Housing Policy in Pakistan calls for a nationwide reduction in stamp duties and registration fees to an average of 0.0 percent and 1.0 percent for mort- gage and conveyance deeds, respectively. The provincial governments have not implemented this policy, however, and are charging between 3.5 percent and 9.0 percent in levies. As can be seen in table D.11, both registration fees and stamp duties are considerably above these suggested benchmarks. There are other levies on property, such as the commercialization fee on the purchase of properties that are converted to commercial use. This fee is considerable--20 percent in Punjab, for example.15 The logistics costs incurred by the customer to access the land transaction document are also not negligible (table D.12). By some estimates, the time delays involved, appropriately monetized, amount to $140 (in Punjab); and could be reduced to $25 with the establishment of an automated land record system (SBP 2007). The property tax framework is weak. Property tax is the main source of local government revenues in most good-practice countries, but property tax revenues in Pakistan have fallen way short of local government financing needs; they barely have been increasing in real terms, despite a marked increase in construction activities and a tripling in property prices. One of the key reasons is that property tax is based on rents, which are controlled. Other reasons include poor enforcement and Table D.11. Cost Structure of Land Titling and Registration, 2009 Registration Fee (%) Stamp Duty (%) Province Conveyance Mortgage Conveyance Mortgage Total (%) Balochistan 1.00 1.00 5.0 2.00 9.00 North-West Frontier 0.50 0.50 3.0 4.00 8.00 Punjab 1.00 0.25 2.0 0.25 3.50 Sindh 1.00 1.00 3.0 (plus 1.5 Ad valorem 5.00 plus tax in Karachi) PRs 2,500 for PRs 2,500 a loan of PRs 10 million or more Source: Land registration and titling regulations. 257 Appendix D D Table D.12. Average Per-Land-Record Cost to the Customer Rupees Direct Transaction Opportunity Scenario Costs, Including Graft Costs Total Without LRIS project 7,427 181 1,144 8,571 With LRIS project 641 32 841 1,514 Difference ­6,786 ­149 ­303 ­7,057 Source: World Bank's project appraisal document on a proposed credit (SDR 30.35 million; US$45.65 million equivalent) to Pakistan for a Punjab land records management and information systems project, December 21, 2006. Note: LRIS land registration information system; SDR Special Drawing Right. coverage. Many of the cities have neither the latest master plans for construction nor the revenues for implementation. Another major enforcement constraint is the prevalence of cash transactions, in which it is often easier to conceal the true sales price or understate the taxable property development income. There is a general consensus among fiscal experts and economists that revenue collection in lieu of property transactions is a function of documentation, valuation, collection efficiency, and overall economic activity. The Punjab government's "White Paper 2003-2004"16 also acknowledges the effect of rate reduction, improvements in administration, and stable economic conditions on revenue collection in lieu of stamp duties and registration fees. It is significant to emphasize that, after decreasing rates in fiscal 2003/04, the Punjab government experienced an increase in revenues. In spite of charging the lowest property levies, Punjab generates the highest revenues from stamp duty (PRs 7.50 billion in fiscal 2005/06), sale of stamps (PRs 6.31 bil- lion), and registration fees (PRs 1.80 billion).17 The issue of property tax revenue achieves strengthened significance because this category forms the single largest source of provincial tax receipts (for example, 29.12 percent in Punjab). Moreover, registration fees are the third-largest revenue source (6.98 percent in Punjab).18 The basic regulatory framework for land registration and transfer is sound, but nontransparent and cumbersome. Typically, patwaris19 manually maintain 15 sepa- rate books of land registries to which nobody else has access. Poor recordkeeping is pervasive. Issuances more than one property title document for the same plot are not uncommon. Land records are not a, and have numerous errors and omissions. Because of the costs involved in property registration procedures, many transactions are not recorded. Computerized information and registration systems, property and ownership databases, effective title regularization processes and cadastral surveying, and a land information service will increase transparency and the enforcement of creditor rights. The lack of an efficient, transparent, and reliable system for ascertaining who holds authentic titles to properties has forced banks to maintain a "negative list," thereby impeding access to housing finance in a number of urban localities within the urban centers. This situation is further aggravated by the law, which allows a claim of propri- etary interest in real estate based on an unregistered sale deed, instead of mandating 258 Appendix D D full registration. This leads to fraudulent claims and much litigation. Another major source of litigation is claims to title on the basis of an oral gift under Islamic law.20 The absence of an effective dispute mechanism compounds the problem. Courts have a considerable backlog of land-related disputes that take years to resolve. More than 40 percent of all court cases are related to land--that number exceeds 1 million. The weaknesses of the present land records in tandem with high transaction costs have led to exchanges of property on a cash basis and without any transparent and legally acceptable title documents in a majority of property transactions. The less- than-satisfactory state of real estate registration has not only deprived the property owners of an effective legal title, but is also a major cause of leaks in revenue collec- tion because stamp duty and registration fees on property transactions are not paid. Taking the directives of the Ministry of Housing and Works, the governments of Punjab and Sindh have already started projects to develop computer-aided, account- able, equitable, and secure land record management systems. These will secure the title of land by improving clarity and efficiency, controlling errors and fraud, and decreas- ing the transaction costs. In Punjab, work is under way in three districts--Gujrat, Lahore, and Rahim Yar Khan. Completion of the effort in an initial 18 districts is planned for no later than 2011, and projects will be completed in another 14 districts by 2013. The government of Sindh launched a pilot project in 2003­04 to computer- ize the revenue records in five districts--Hyderabad, Khairpur, Larkana, Mirpurkhas, and Sukkur. This was scaled up to the entire province in 2007­08. The computeriza- tion and establishment of a land administration and revenue management informa- tion system, at an estimated cost of PRs 921.812 million, will be finished in four years. The Support to Devolution Reforms in Balochistan developed a comprehensive model for land record computerization in Balochistan. The pilot test results (Mauza of District Loralai) proved that the model is user friendly; Urdu language­enabled; and offers many key features including live map generation, automatic error check- ing, and automated data processing. The Balochistan Land Record Management Information System was shared with the revenue department of the government of Balochistan, and a detailed training on surveying was provided for the staff of the revenue department. A nationwide land registration information system (LRIS) would be the ultimate goal, based on future similar projects instituted throughout the country. Imple- menting agencies of nationwide LRIS include the National Reconstruction Bureau, the National Database and Registration Authority, provincial governments, city and district governments, landowning agencies, the Central Board of Revenue, and the Central Depository Company. However, to expedite implementation there is also a need to encourage private sector­led initiatives to build the capacity of existing auto- mated property transaction­related systems. Capacity of Real Estate Professionals In Pakistan, "property valuers" have professional conduct requirements established by the SBP and the Pakistan Banks' Association.21 The majority of real estate builders 259 Appendix D D and developers do not have strong governance and disclosure standards, and that makes it difficult for financial institutions to assess them. The business of real estate brokers/agencies is unstructured and unsupervised as well. Property valuers are approved by the Pakistan Banks' Association; and they must undergo professional training in evaluating the property, assessing the genuineness and integrity of the title documents, and adequately performing all aspects of their work. Nevertheless, the lack of transparency in real estate valuation is a major bottle- neck. The present valuation methodologies are primitive and not standardized, and that leads to different valuers offering different proxy market price estimates for the same property. This is especially crucial for the promotion and development of REITs. Pursuant to launching the REIT product in February 2008, the Securities and Exchange Commission of Pakistan took major steps to promote the implementation of international best practices in the field of valuation, focusing particularly on Malaysia's experience. The commission has also applied for membership in the International Valuation Standards Committee--an organization responsible for set- ting worldwide standards for property valuations. And the commission is presently an associate member of the Asian Public Real Estate Association. Constraints for Women In Pakistan, there are no constraints or limitations on a woman's right to own assets, run a business, or hold a title in her own name. At the higher levels of income, there is no apparent gender-based credit rationing. In fact, there is a specific bank for women entrepreneurs--First Women's Bank Limited. However, lower-income groups, located mostly in rural areas, show considerable gender disparities regarding all forms of financing. These disparities exist for both cultural and informality reasons. Foreclosure and Eviction Recovery, foreclosure, and eviction laws in Pakistan are based mainly on the 2001 Financial Institutions (Recovery of Finances) Ordinance (sections 15[2] and 19[3]). The process empowers financial institutions to foreclose on a mortgage property without recourse to the court of law. The law is not being applied, how- ever. Some lower courts have granted stay-orders when financial institutions have tried to use the ordinance. Until lenders are confident that they can enforce their rights to collateral, they will remain reluctant to invest. Initial steps in implement- ing the law were taken by the Balochistan High Court in 2006, when it passed a judgment defending the right of a financial institution to foreclose. This precedent has been replicated successfully enough for recovery laws to be used more frequently. The ordinance also has weaknesses in the letter of the law. The foreclosed prop- erty auction, for example, does not require the prior objective valuation of the property, to set a reserve price. This often leads to recourse to a court of law, and adds to the cost and delay of the procedure. In addition, surveyors and property valuers 260 Appendix D D lack credibility in the absence of a trusted systematic transparency mechanism for the evaluation and certification and strict enforcement of professional ethics and standards. The ordinance also does not provide for adequately broad publicity/ advertisement of the proposed sale by the foreclosing financial institution. Advertis- ing in provincial newspapers alone is not likely to inspire judicial confidence. The law should provide for national and, in appropriate cases, international advertising.22 Poor notice is also given on attached properties that are subject to sale--typically, the notice is affixed near the plot location and it may be damaged or removed. This exposes potential buyers to the risk of paying in full for an attached property, and is the cause for numerous lawsuits. The ordinance compounds the problem by voiding property sales made after the banking court issues an order against the debtor, even if the execution of the foreclosure has not been requested at the time of the property sale. At present, there is no public registry of debtors against whom judgments passed by banking courts remain unsatisfied. Consequently, a purchaser can find his or her title challenged by a financial institution, even if the property was neither mortgaged nor attached at the time of the purchase. Credit Information The CIB was established by the SBP in December 1992.23 The scope and activities of the CIB are governed under provisions of the Banking Companies Ordinance 1962.24 The scope and administration of the CIB database were enhanced in April 2006 when credit information was transferred on an electronic platform, improving the speed, reliability, and security of data. The key features of eCIB include consumer and corporate reports and data input formats; records of credit inquiries made by the financial institutions; and consumer credit and default histories. The system pro- vides for online amendments and interim updates, and comes with a Web-based help desk and online technical support to the financial institutions. At present, only those banks and financial institutions that are either regulated by the Securities and Exchange Commission of Pakistan or the SBP have access to the eCIB database. The information there is being exchanged on a reciprocity basis; and every financial institution is required to submit its entire body of borrower records, irrespective of any limit to the CIB database. The member-institutions can access only limited information from the CIB database, however--for example, they do not know which financial institution is the source of the credit information about any given borrower. Individuals and corporate entities are not entitled to obtain their own credit information reports. The onus of correct reporting lies with the report- ing financial institutions. Consumer credit and default histories over the preceding 12 months are reported for all individuals and sole proprietors served by the participating financial institu- tions. Partnership concerns and corporate entities (both listed and unlisted) receive a unique borrower code, against which member-institutions report outstanding lia- bilities (fund and nonfund based); position of overdues; and details of litigation, write-offs, recoveries, and rescheduling and restructuring. The financial institutions 261 Appendix D D are not allowed to use credit reports for marketing their financial products, and are bound not to divulge information to a third party. There are also three privately owned and managed credit bureaus in Pakistan-- namely, Datacheck (Pvt.) Limited, News-VIS Credit Information Systems, and ICIL/PakBizInfo. These firms recently started operations, and mainly gather data on the credit histories of exiting borrowers. The information is obtained from various banks, utility companies, credit card issuers, and so forth. It is updated regularly and passed on to banks on demand. The information includes both positive and negative details on consumer credit. However, these credit bureaus remain less reliable and are less widely used than the CIB. The private services will become more useful when they have databases with longer credit histories. Policy Options for Improving the Framework for Housing Finance Housing finance is a fundamental pillar that supports development of the economy because the housing and construction sectors are labor intensive and have both back- ward and forward links with more than 40 industries. In most developed countries, housing finance is one of the main economic drivers and its share of the GDP may rise above 50 percent; in developing countries such as Thailand, however, its share is about 15 percent. Eliminating housing shortages enables lower-income groups to improve their living standards, accumulate assets, create livelihoods, and boost their savings. Vibrant housing and housing finance sectors can adequately accommodate population growth and urbanization pressures, thus minimizing slum problems. Reforming the real estate sector is a prerequisite for broadening and deepening housing finance. The Housing Advisory Group at the SBP has issued a detailed report on better enabling the housing and housing finance sectors. The following key recommendations are included there: 1. reform the legal framework affecting foreclosure, transfer, tenancy, rent control, and acquisition of immovable property; 2. rationalize transaction costs in real estate transactions; 3. establish an integrated, nationwide land registration information system; 4. increase the supply of land for affordable housing; 5. structure and streamline large-scale developer finance; 6. provide housing finance for affordable and low-cost/low-income housing facilities; 7. promote and develop REITs; 8. provide long-term funding for housing loans through the creation of a mortgage refinance company; 9. develop a housing market information system; and 10. develop the mortgage banking industry. Those recommended policy actions are discussed below in greater detail. Successful reform in these areas holds the promise of developed housing and housing finance 262 Appendix D D sectors, and an improved housing situation the population--especially for lower- income groups. Action also promises the economic benefits of housing and housing finance development: increased savings, investment, output, and employment. A strong housing market requires an effective legal and regulatory framework that instills confidence in the property rights of buyers and sellers and facilitates effi- cient and verifiable property transactions. Coupled with fragmented landownership and title registries, and with inefficient land information systems, a weak legal and regulatory framework hinders the development of a sound housing finance market. The following section discusses the overhaul of the land administration and prop- erty taxation systems. The "Housing Development Regulations" section tackles the rules, building codes, and regulations placed on property developers, as well as hous- ing finance prudential rules and faithful application of the existing foreclosure regime. As a first step, policy makers should provide an institutional framework for generating long-term plans for housing sector development in conjunction with the central planning agencies--the Ministry of Finance, Ministry of Housing and Works, and the SBP. This process should be conducted with the participation of the private sector, NGOs, community-based organizations, and the general public at both national and municipal levels. Regulating Land Administration, Titling, Property Transfer, the Land Registry, and Property Taxes To support a vibrant housing market, the property registration system in Pakistan needs to be overhauled, especially if banks are to gain greater confidence in loan documentation. Improvement in this system is particularly important for the sec- ondary mortgage market, which requires due diligence procedures to ascertain the security of underlying claims. This will be possible on a large scale only if the mech- anisms for determining this security are standardized and efficiently done. Specifically, the overhaul of the current systems should include coordinating land-use planning policies, coordinating titling and land administration, building internal technical capacity, and integrating the land registration and information systems nationally. Currently, there is no institutionalized policy framework or mechanism to determine the criteria and frequency of government land disposals, especially within and around major cities. And related to this, the policies and pro- cedures for ensuring that the land is appropriately serviced with infrastructure and utilities should be made more effective and transparent. Reducing the regulatory and bureaucratic complexity of the property ownership process is a key element to pro- moting a more vibrant and sustainable housing sector. Public sector institutions that are involved in the housing sector should not have redundancies and inconsistencies among their functions and they policies they implement. A well-coordinated insti- tutional framework will make it possible for the government (with its limited resources) to manage the housing sector in a manner that provides adequate and affordable housing for all people. 263 Appendix D D A key challenge is coordinating the titling and land administration process. Com- puterization is essential. Real estate transaction databases need to be linked to the registration entries. Specific steps to strengthen the titling and recording system include amending the land acquisition acts and the urban rent laws. The legal con- text should only consider formally registered transfers as generally binding, and should link registration to the authentication of deeds. A uniform framework of reg- istration standards is required, instituting a system for supervising and monitoring development agencies, designing quality control norms, and providing incentives for efficiency. Special procedures would be required to regularize (at the lowered tax rate) the stocks of unrecorded properties or those left without clear titles, with a pro- vision to include explicit criteria for the validation of rights and rules for dispute res- olution. Finally, and very important, effective dispute resolution mechanisms are required to deal with the many disputed properties pending resolution by the courts of law. Integrate the Land Registration and Information Systems Nationally. As part of the process of improving the land management system, the LRIS needs to be strengthened. Building on short-term initiatives to streamline and computerize the administrative bureaucracy responsible for the land titling and registration systems, efforts should be made to integrate all the provincial and interagency systems with one another. This initiative would provide policy makers with a better understand- ing of the housing situation on the national level, and would prevent variations in the degree of sophistication between the provinces. It would be useful--particularly regarding standards, procedures, and information technology systems--to launch a pilot program in a high-priority area that could then be replicated elsewhere. Some of these actions probably will have to be taken at the national level or in coordina- tion with the provincial governments (especially tax and legal issues and the institu- tion of uniform standards). Regularization actions, however, should be carried out selectively to prioritize regional areas according to their shortages of housing and the state of the local land or development agencies. A properly planned and efficient LRIS would not only secure the interest of mort- gagee financial institutions and investors (in the case of mortgage-backed securities), but would also help in consolidating a decentralized real property rights system that could help the general public in numerous ways, could lead to one-stop title documentation, and increase property registration--thereby enabling provincial governments to realize the revenue-generating potential of property transactions. Implementing an LRIS would help increase the outreach of housing loans and decrease the loan processing time for mortgagee financial institutions. To institute a nationwide LRIS within the next five to seven years, a three-pronged strategy is required. It should involve both public and private sectors. The three pillars of the strategy should be these: · Streamlining the role of the public sector--Implementing the LRIS in 102 districts (27 in Balochistan, 23 in North-West Frontier Province, 30 in Punjab, and 22 in Sindh) and 8 city districts (5 in Punjab and one each in Balochistan, North-West 264 Appendix D D Frontier, and Sindh) requires streamlining coordination among the National Database and Registration Authority, the National Reconstruction Bureau, and the respective district/city governments. At the same time, the eventual nation- wide integration of the LRIS requires standardization of the system across districts/cities; and under this public sector­led initiative, priority should be accorded to instituting the LRIS in rural and low-income urban areas so that residents may receive credit from the formal financial system. · Promoting area-specific private sector initiatives in cities--In the more affluent urban localities of major cities (Pakistan Employees Cooperative Housing Society neigh- borhood in Karachi, Gulbarg in Lahore), private sector organizations such as the Central Depository Company may be willing to automate land records. Therefore, these small-scale private sector initiatives need to be encouraged because they would act as a catalyst in automating the property database in the country. · Building the capacity of the existing system--To take advantage of the ongoing public or private sector initiatives relating to automation of property transac- tions, it is important to build their capacity. In this regard, a good example is the initiative of the Central Board of Revenue in automating land records in 18 Karachi towns. This initiative could be upgraded to an LRIS by enlarging its scope and scale. To weed out malpractice in the sale of attached properties, an electronic public reg- istry should be created forthwith. All people obtaining attachment orders or decrees would then be legally obligated to make a filing of fact concerning the attachment or the decree with the electronic public registry. Failure to do so should make the attachment or decree ineffective against any right acquired by a subsequent pur- chaser of property. The details of the public registry can be worked out. The registry may either be attached to the existing offices of the registrars of documents or be set up as an independent institution. Transfer Property Only through a Registered Instrument. No proprietary interest should be recognized unless the property has been registered. Initially, the recom- mendation contained in this paragraph may be effective only with respect to urban areas.25 Evidence of an oral gift should appear as a written memorandum attested by a judicial officer.26 A filing of acknowledgment of an oral gift with the electronic reg- istry may also be stipulated. Finally, claims of benami (absentee) ownership should not be recognized by law.27 Rewrite Urban Rent Laws. The Rent Control Act should be repealed, if that is polit- ically feasible. Repeal would eliminate the special status accorded to landlord-tenant relationships. Contracts between landlords and tenants would be treated as normal contracts; and dispute resolution could be instituted at higher tiers of the judiciary, thereby not only expediting the settlement process but also helping weed out ineffi- ciencies created at the level of the rent controller. At a minimum, the urban rent laws should be amended to equilibrate the rights of landlords and tenants. In the case of 265 Appendix D D a verbal lease, the landlord can evict the tenant within six months of notice, without any default by the tenant. Written leases should be encouraged. The landlord should not be able to evict on grounds of personal need when a written lease exists. A Urban rent laws should clearly state that written lease agreements containing a term less than 12 months may be registered at the option of either the landlord or the tenant.28 Implement a Tax Rationalization Program. An in-depth tax rationalization program needs to be implemented to develop a taxation scheme that promotes economic behavior consistent with the overall long-term economic policy objectives. There should be a reorientation of the government's taxation role in land-related transac- tions. In line with the spirit of the 2001 National Housing Policy, efforts need to be made to rationalize and standardize the rates of stamp duties and registration fees across the four provinces and the federal capital. Furthermore, by automating the tax collection system, the government will strengthen the revenue collection process. Stamp duties and registration fees need to be lowered. Keeping in view the posi- tive implications of rate reduction on documentation of property transactions and subsequently on provincial revenue receipts, the provincial governments should adopt the National Housing Policy in both letter and spirit by reducing transaction costs on residential properties in accord with the housing policy covenants--that is, for conveyance and mortgage deeds, the aggregate rate of stamp duty and registra- tion fee should be decreased to 1 percent and 0 percent, respectively. The commer- cialization fee should be rationalized as well. The present high transaction costs create an incentive to record transactions at a price much below market value (the D.C. Rate) because there exists a substantial disparity between the actual transaction prices and those recorded for tax purposes. This disparity puts corporate entities such as REITs at a serious disadvantage because these entities are required to disclose the actual purchase price (and pay fees based on it rather than on the D.C. Rate). Therefore, to encourage corporate entities in the construction sector, it is proposed that commercialization fees should be reduced to 10 percent. Instituting automated land records, in tandem with rate rationalization and improving the revenue collection system, has the potential to encourage docu- mentation of property transitions by reducing the direct transaction and opportu- nity costs, and would have a positive impact on the provincial governments' revenue collections. Moreover, housing tax revenues need to be directed toward managing the land management system in a commercially sustainable manner that enables the government to fulfill its social responsibilities toward middle- and low- income households, and its good regulatory governance responsibilities to the entire sector. Housing Development Regulations The current convoluted approach of development authorities is creating artificial shortages of serviced land for the public; this contributes, in turn, to the high cost of land that effectively makes landownership impossible for the middle class. 266 Appendix D D Provide Infrastructure for Residential Land Development. In view of the rising prices of land, the government of Pakistan needs to develop land and satellite towns on the outskirts of the cities. These should be adequately serviced with residential infrastructure such as water, electricity, sewerage, schools, clinics, transport service, and road networks. Data are available on availability of large pieces of land at the dis- posal of federal and provincial governments. In obtaining land for low-cost housing, preference would be given to those government lands that have lain idle for many decades. The land would be purchased at an official rate. Through this method, the government can directly share its subscription, so the land would be available at far cheaper rates than the market rates. Once developed, these plots could accommodate pro-poor housing projects under public-private partnership arrangements. Infrastructure agencies should devote greater attention to providing utilities and infrastructure rather than to focusing on landownership and resale. Funding for such a model should come from the public credit markets because such an undertaking may yield a number of direct benefits--namely, increased land availability, deepening of the credit markets, and required institutional reforms to make the development authorities more efficient. Any such strategy should also include a program to provide adequate infrastructure to squatter settlements. In addition, whenever possible, pro- grams for regularizing tenure should go hand in hand with infrastructure improve- ment in slum and squatter settlements and should seek to recover costs. Specific steps to providing infrastructure include the following: · The agencies responsible providing residential infrastructure (roads, drainage, water, sewerage, and electricity) should focus less on narrow physical objectives and more on opening up urban land for residential development. · There should be greater coordination in planning and possibly joint acquisition of rights of way, joint financing, and joint cost recovery. · Infrastructure agencies should review the impact of various regulations on the performance of the housing sector and propose new legislation to improve sector performance. · Existing communities should be encouraged to participate in the process of plan- ning and building of infrastructure projects to ensure accountability and smooth implementation. · Cost recovery mechanisms should be improved and opportunities for privatizing infrastructure provision and maintenance should be sought. · A clear statutory scheme for land acquisition for private housing schemes should be included in the land acquisition acts in force in the provinces and the capital territory. · The fee for nonutilization of land should be raised, especially in areas where infrastructure has been provided, so that speculation in land is avoided and con- struction is expanded. 267 Appendix D D Develop a Home Construction Industry and Financing. Consistent with the long- term policy objectives in achieving greater economic growth through the housing market, efforts should be made to help develop a sound and vibrant home con- struction industry. The government should seek to create greater competition in the building industry by eliminating regulatory barriers to entry, breaking up monopo- lies where present, facilitating equal access of small firms to markets and inputs, removing constraints to the development and use of local building materials and construction methods, and reducing trade barriers that apply to housing inputs. There is a need for a strong regulatory authority to enforce building standards to improve the quality of housing. Banks may be encouraged to establish business relationships with developers on the basis of past performance and the commercial and financial viability of the real estate projects, thereby making the credit rating of projects a viable alternative to credit rating of developers. However, where appropriate and possible, the credit rating of a builder should also be encouraged. It is important for a developer to be registered with ABAD, both to facilitate the builder's credit access to the formal financial system and as a tool to ensure builder compliance with the ABAD code of conduct. For streamlining the provision of credit to developers, documentation should be simplified and a standard set of documents should be instituted with the mutual consent of ABAD and the Pakistan Banks' Association. A REIT is an alternate model of developing real estate, and is more transparent and structured than the traditional real estate development process in Pakistan. Therefore, institution of REITs should be encouraged in both letter and spirit. To modernize and increase capitalization of the real estate sector, it is important that joint ventures with foreign developers be encouraged. Among other things, this would encourage the transfer of technology and management know-how. Promote Low-Cost Technologies. The industry of low-cost housing technology and materials should be strongly promoted, including through incentives. Bulk purchas- ing and reduced taxes can be a right mix of policies. The government should invite some international best-practice providers of low-cost housing technology and should facilitate technology transfer through joint ventures. These experts can also help in capacity building and trainings. There are many low-cost housing construction technologies being marketed by various international and local vendors. But there is no existing platform to eval- uate and certify which technologies are internationally proven low-cost housing technologies and which ones would satisfy local requirements. Technologies may also need some indigenization. The Center for Promotion of Low-Cost Housing should be encouraged to pick up its activities anew, with a focus on research and promotion of low-cost construction materials and technologies. More coordination is required among construction/development industries, builders, architecture/ design firms, and government agencies to promote affordable housing for the low- and middle-income population. 268 Appendix D D Organize Real Estate Agencies. The unstructured and unsupervised nature of real estate brokers is a significant constraint to the provision of housing and housing finance. Only property valuers have professional conduct requirements (established by the SBP and the Pakistan Banks' Association). The real estate agencies, which could be natural brokers or arrangers of financial services, remain unorganized and insufficiently supervised. The protection of individual purchasers remains limited because the market is dominated by cash transactions and lacks transparency and because no systematic information is available. Risk assessment and portfolio valua- tion are also fragile--another factor prompting lenders' extreme caution. The Pakistan Securities and Exchange Commission (in collaboration with the SBP) should assist the valuation profession to create an institute that would create a system of examinations, license valuers, and implement international best practices and standards. To standardize the valuation process, therefore, the Pakistan Banks' Association or the Mortgage Bankers Association needs to screen the various valua- tion methods that are used internationally, and adopt the one that best fits Pakistan's circumstances. To provide for an effective mechanism for sale by mortgagee finan- cial institutions without the intervention of the court, the law should require that SBP-approved surveyors fix a reserve price. For such surveyors to have credibility, the SBP or other reputable trade body/association should establish a mechanism for evaluating and certifying property valuers. Supporting Institutional Reforms for Financing Property Development Increasing the level of financing to the housing sector requires a double-edged approach: increasing the financing going into the construction of housing, and increasing the financing available for mortgages. To achieve these twin objectives, it is recommended that the government continue its efforts to introduce REITs, encourage the issuance of municipal bonds, and create a secondary mortgage mar- ket. Two other important policy focus areas are the creation of a housing observa- tory and an appropriate prudential environment for mortgage lenders. Finally, it is very important that the foreclosure law be adopted in practice if housing finance is to extend to the average consumer. As an interim measure, and as indicated by some recent experiences, raising medium-term funds (by floating debt papers of five to seven years) could be a good beginning for housing finance institutions. A recent example is the HBFC, which has successfully floated a six-year sukuk at a very attractive rate. Another option for rais- ing long-term funds would be to issue several medium-term (five-year) debt floats in succession, thereby raising funds for 15­20 years. As Pakistan extends housing finance to middle- and low-income segments of the population, the demand for long-term, fixed-rate mortgages increases because the low-income segment is hardly able to withstand the dynamics of interest rate volatil- ity. A refinancing facility would go a long way toward lowering financial institutions' cost to issue such instruments. 269 Appendix D D Apply the Foreclosure Law in Practice. Judge training and capacity building should be conducted throughout the country, especially in lower courts, to instill there the spirit of the foreclosure law. The decision made by the Balochistan High Court in 2006 in defense of the right of a financial institution to foreclose could be success- fully replicated throughout the country to expedite implementation of the recovery laws. The 2001 Financial Institutions (Recovery of Finances) Ordinance could also be amended to iron out some remaining weaknesses. Objective valuation to set a reserve price for a property prior to auction should be mandated. Certification and higher standards should be instituted for surveyors (with higher penalties for breach of standards) to promote credible valuations. Auction notices and the general trans- parency of the process should be improved. The issue of property sales made after the banking court issues an order against the debtor should be clarified, in conjunc- tion with other laws, to minimize title challenges. Ensure an Appropriate Prudential Environment. In the banking sector, calls have been made for regulatory adjustments that would make it easier for financial insti- tutions to invest in housing assets. It would be particularly advisable to adjust the prudential regime to reflect the low risk associated with secured housing lending-- through the risk weighting of the loans for capital adequacy requirements. The per- party limit of PRs 10 million and maximum 10 percent ceiling of housing finance with respect to a financial institution's total credit portfolio of have been waived recently by the SBP. Rather than imposing quantitative ceilings, the prudential regime could provide for asset-liability matching, which is critical for long-term loans because they are rate sensitive. Special attention should be given to ensure uniform treatment of all lenders, regardless of the regulator. Investment should be made in a wider-ranging credit information system in the private sector to include positive data on retail lending. Create a Housing Observatory. A housing observatory should be created to provide periodic census and survey activities, including collecting, analyzing, interpreting, and publishing data on the performance of the housing sector. The database would also address the issue of complete and reliable data not being available to stakehold- ers, consumers, and the housing industry to help facilitate property purchases, investments, or financing decisions. The housing observatory may include some or all of the following: · housing supply information (formal and informal)--new construction and sales, vacancies, quality and condition of property, prices, and dwelling characteristics; · housing demand data--demographic trends and housing needs, characteristics of occupants of the existing stock (crowding indicators), affordability (availability of income indicators are of paramount importance of this), and buyer's intentions; · housing finance data--sensitivity of monetary indicators to the residential and commercial property exposures of lenders and investors (for example, inflation 270 Appendix D D rates, interest rates), fiscal indicators and degree of reliance on housing-related developments (for example, property taxes, construction industry contribution to GDP and employment), lending volumes by category of financial institution; loan characteristics (loan-to-value ratios, debt service ratios, prices, locations, borrowers, and so forth); · construction inputs--production capacity; prices, quality, and availability of materials, labor, and land. The Pakistan government has shown commitment to establishing a housing obser- vatory. The effort is led by the SBP. In the meantime, after consulting the Pakistan Banks' Association, the SBP plans to launch a dedicated Web site for housing infor- mation systems and to require primary mortgage lenders to upgrade their Web sites. Introduce REITs. The government could help increase the volume of financing avail- able to the real estate business through the introduction of REITs, a company or trust that uses the pooled capital of many investors to purchase and manage rental prop- erty (an equity REIT) and/or mortgage loans (a mortgage REIT). Designed to be traded on stock exchanges, REITs enable small investors to invest in the real estate sec- tor in the same way they buy company stock. To increase their attractiveness, REITs are usually granted special tax considerations. Rather than own properties directly, REITs offer investors liquidity, which traditional real estate cannot offer. Second, REITs enable sharing in nonresidential properties, such as hotels, malls, and other commercial or industrial properties. Third, there is usually no minimum investment requirement, and that encourages small savers to invest in REITs and reap higher returns than would be possible with bank deposits and national savings schemes. The Securities and Exchange Commission of Pakistan has been working to estab- lish REITs in the country. In December 2005, draft legislation was sent to the Min- istry of Finance for vetting by the Ministry of Law. It has been proposed that REIT companies must have at least 70 percent of their assets invested in real restate, with minimum paid-up capital of PRs 50 million. To ensure that the REITs are held by a broad population, it also has been proposed that the REIT fund be held by at least 100 people, and that no 5 of them may hold more than 50 percent of the assets. The feasibility of these proposals is being discussed. However, there is a general consensus among policy makers in Pakistan that, if properly regulated and supervised, REITs would yield a multitude of benefits for promoting and developing the housing sector. Because REITs provide an alternative source for raising capital, they would help increase the liquidity of the real estate market. At the same time, the institution and proper functioning of REITs is expected to improve the governance structures of real estate companies in the indus- try, especially with regard to transparency and credibility. Encourage Issuance of Municipal Bonds. Institutional reforms in housing have to involve municipalities, which control access to land as well as the development of local infrastructure. Municipalities often cite their weak financial condition as a key constraint. At times, their only reliable source of revenue is taxation, and property 271 Appendix D D tax is the only major tax. Accordingly, municipal governments have increased the property tax rate on the premise that this will yield more tax revenues; however, empirical evidence suggests the contrary. A major step toward improving the availability of financial resources for infra- structure development is encouraging the issuance of municipal bonds (after appropriate governance and management reforms within the municipality). In developed economies, municipal bonds are issued by state, city, or local govern- ments (or their agencies) to raise funds for building schools, highways, hospitals, sewer systems, and other similar projects that benefit the general public. To make investment in municipal bonds attractive, the interest income is often exempt from national and local income taxes and from income tax levied by the state in which the bonds are issued. It is significant to elaborate that, for investors, tax-exempt municipal bonds are a popular type of investment because they offer a wide range of benefits, including the following: · attractive current income, free of federal and (in some cases) state and local taxes; · a high degree of safety with regard to payment of interest and repayment of principal; · a predictable stream of income; · a broad array of choices to fit investment objectives regarding investment quality, maturity, choice of issuer, type of bond, and geographic location; and · marketability, in the event the investor has to sell before maturity. In addition to administrative reforms to ensure that bond revenues are used for their intended purpose, the issuance of municipal bonds also requires legislative changes. Currently, section 120 of the Provincial Local Government Ordinance 2001 specifi- cally prohibits local governments from incurring debt. Create a Secondary Mortgage Facility. The ultimate objective of developing mort- gage finance is to create a secondary market that would help the market take off through a refinancing structure, which would not only alleviate the liquidity risk incurred by the primary lenders, but would also promote sound practices and stan- dards in lending activity through its refinancing requirements and have an overall catalytic effect on the development of lending activity. Such a structure could be seen as a first step toward a true secondary mortgage market in Pakistan. A secondary mortgage facility is a financial institution designed to support long- term lending activities by primary mortgage lenders. Its core function is to act as an intermediary between mortgage lenders and the bond market, with the objective of providing long-term funds at better rates and under better terms and conditions than lenders might be able to obtain if acting alone (Hassler and Walley 2007). The need for such an institution arises because of the mismatch between the liabilities and the assets of mortgage lenders. The facility would help financial institutions issue lower-cost, fixed-rate mortgages. 272 Appendix D D In the medium term, establishing a secondary mortgage facility could become viable when primary originations reach critical mass. This would be a way to provide external credit enhancement at a lower cost than would be possible if such enhance- ment were arranged on an individual basis. In the initial stage, a liquidity facility could be structured for lending against a bank's mortgage portfolio, with recourse to the lender, like Cagamas29 in Malaysia. In such a case, the second-tier institution acts as a credit enhancer of the originating bank itself (rather than the loan portfolios) by issuing bonds that are its own general obligation. In Pakistan, a secondary mortgage company needs to be established to purchase mortgage loans from mortgage originators (commercial banks and housing finance companies), with full recourse to the primary lenders at a fixed or floating rate ini- tially up to a period of 10 years. At the same time, the securities of the company could be eligible as liquid assets and subjected to concessional risk weights, that is, 10­20 percent, compared with a 50 percent risk weight for housing loans. Further, the SBP could support a repo market by accepting mortgage securities as collateral in its repo transactions. Finally, the government could provide incentives to investors by exempting mortgage security interest from taxes (subject to appropriate fiscal considerations). An alternative is to set up an institution that would issue bonds against securi- tized paper, with the originating bank taking the first credit losses, while multilateral and local institutions would underwrite the mezzanine and senior tranches. A key element in proposing a set of reforms for the secondary market is one of sequencing. A secondary market depends on the existence of a strong primary market in which mortgages are regarded as attractive assets with good and well-documented performance. A major prerequisite is a strong legal infrastructure supporting the reg- istration, enforcement, and eventual pledging or sale of mortgage loans. Finally, it is difficult (although not impossible) for a secondary market to lead to capital market development. The existence of a robust bond market is an important precursor to sec- ondary market development. To be sustainable, the secondary markets should be run as a profitable entity. However, such structures generally benefit from some kind of institutional support during the first years of their operation, if only for legal reasons (for instance, restrictions to bond issues by a newly created or an unlisted entity). Such support is vital in helping establish the entity's credibility and speed up the acclimation process among investors, thus avoiding the premium that a totally new issuer would normally pay. Introducing New Housing Development and Finance­Related Products and Services Overhauling the operational capacity of the industry, improving the reliability of the land records, and promoting long-term funding will boost the overall demand for and supply of mortgage finance and, in the medium to long terms, will address the supply of serviced land. In the face of pressure to address the housing finance needs of specific communities, however, the government may wish to consider some of the 273 Appendix D D programs for microborrowers and first-time buyers that recently have been imple- mented in other countries. While microfinance institutions in Pakistan continue to focus on credit delivery of loans for agriculture, livestock, trading, and consumption, a few are beginning to diversify and innovations have begun in enterprise loans and housing finance. The government may wish to begin with the strongest players as a way of reducing risk, getting to know the market, and helping adapt the selected financial institutions' internal systems and design their mortgage risk assessment databases. Although reaching scale would not be a central objective of a pilot project, efforts should be made to reach a scale sufficient to test the adequacy of the systems used to analyze and monitor the institutional system (of the financial institution) developed for lower- end borrowers. In time, other banks may consider investing in this end of the market. However, it must be recognized that, at best, this channel for delivering housing finance may remain small. As of June 2005, the microfinance network served only 718,000 borrowers and held $99 million in total loans outstanding, covering less than 12 percent of the potential market in terms of an estimated possible outreach to 6 million. The key challenge for initiating such housing finance programs is securing long- term funding to finance them. The first option could be to let banks secure their funding through their own individual issues. The instruments could either be term finance certificates or securitization, on or off balance sheet. There are obstacles to the use of both tools: individual financing requirements will stay low in a first phase, thus generating illiquidity premiums. Public offerings of term finance certificates are restricted to listed companies; the market is far from being ready for securitization, despite available regulation for trusts and special-purpose vehicles.30 Alternatively, the financing can be of different types: it could take the form of capital sponsorship (the idea of "seed" capital), a debt guarantee or backup line of credit provided by the government of Pakistan, the SBP, or the benefiting institutions, which would give a mutualized structure to the scheme. These various options are not exclusive of each other. In the United States, the Federal Home Loan Bank system enjoys both the support of its members and a quasi-government guarantee; in Jordan and Malaysia, the share capital of the sec- ondary market facilities is held by lending institutions and the central bank. Sup- porting tools--either a guarantee or a standby line of credit--could be imple- mented only during an initial phase, until the structure acquires full autonomy and maturity, and a sunset provision may be stipulated from the beginning. Otherwise, the institutional support could make the risk-mitigation scheme built into the structure meaningless, and could easily turn into a market distortion factor. Some proposed solutions to improve housing supply for low-income groups are as follows (Rizvi 2007). Mega-housing projects should be initiated in each province. Policy makers should create a land supply at low prices. The model for microlending for housing (Progressive Housing) should be implemented in every province, and construction sites should be located as near as possible to the developed cities so that the infrastructure for the new settlements can be created easily and cost effectively. 274 Appendix D D Policies should enable the supply of low-cost microlending and microfinance for housing. The design, layout, and construction materials should be standardized. Tax incentives on construction materials could be provided for low-cost housing proj- ects. Financing at low cost must be made available through smart subsidies. Cross- subsidization should be considered, possibly via an infrastructure and development fund dedicated to affordable housing developments. Prominent organizations hav- ing expertise in low-cost housing techniques and models could be used for technical assistance and capacity building on affordable housing solutions. Foreign direct investment in the construction sector should be encouraged. Finally, fiscal incentives for builders, banks, and developers should be considered for promoting low-cost housing schemes. Conclusion Despite the growing economic and social importance that development economists are attributing to housing finance, it remains largely underdeveloped in Pakistan. Private mortgages remain small and unaffordable, and are provided by only a lim- ited number of depository institutions. Housing finance is expensive and still rationed in favor of higher-income populations. Policy makers, in collaboration with the private sector, can strengthen property rights and the land administration process, strengthen the property development framework, support the building industry, facilitate the growth of the primary and secondary housing finance mar- kets, and introduce useful targeted housing finance programs. A fundamental challenge, however, is developing the institutional framework for managing the housing sector, including the property titling/land administration procedures, an overhaul of the legal provisions, and the standardization of processes and computerization that will not only bring efficiency gains, but also drastically improve the overall level of confidence in property transactions. The overhaul of registration systems is also a prerequisite to establishing mort- gage lending on a large scale; and it is likely to have additional benefits, such as facil- itating the upgrading of slums, improving tax collection, and enhancing the overall marketability of properties. Because of the federal structure in Pakistan, these reforms must be undertaken by the provincial governments. As all of these reforms are being implemented, a sectorwide capacity-building pro- gram will be required. It is important that all stakeholders--including policy makers, financial institutions, private sector developers, regulators, and consumers--become more aware of the detailed aspects of housing policy. Such capacity-building processes should include strengthening the governance structures and business orientation of housing development authorities, property market developers, and related institutions. Currently, some developers generate much mistrust and suspicion among both individual buyers and potential financiers because of weak governance structures and practices. The capacity-building methodology could be based on a "training-the- trainers" principle and the transfer of best practices and know-how from 275 Appendix D D experienced foreign institutions. Targeted training and conferences, particularly for decision makers, could help facilitate a better understanding of the issues and the responsibilities of each group of stakeholders in the housing market. They would also promote a more coordinated effort in accomplishing long-term objectives. It may be stressed that the existence of an intersectoral working group on housing finance (already used for the design of the 2001 National Housing Policy) could be instrumental in preparing measures and monitoring implementation. Specifically for the banking sector, banks currently are constrained in expanding their housing finance outreach because they do not have enough staff with the expertise to undertake housing finance for origination of housing loans in the pri- mary mortgage market. Banks' in-house capabilities for undertaking housing loans should be developed to mitigate any systemic crises arising from their ill-prepared entry into housing finance. More technical training should be provided on the processes of originating loans, servicing loans, managing risk, and funding mortgage portfolios. Such training should be aimed at bankers; developers; real estate brokers; appraisers; lawyers; and regulatory and supervisory staffs of the SBP, the Ministry of Finance, and the Ministry of Housing and Works. The training could include spe- cialized topics, such as the handling of borrowers with no documented incomes, developer finance, public-private partnerships and real estate project assessment, and Islamic finance products. Specific steps to develop training may include the following: · Periodical needs assessment for developing vocational skills should be conducted, for instance, by the Pakistan Banks' Association and other profes- sional organizations. · Depending on training outcomes, inducements could be considered to speed up and organize training programs on a large scale. · Strategic advisory services could be devised, with the view to making bank man- agement aware of the role that housing finance can play in the soundness and growth of their institutions. In developing tailor-made training courses, it would be important to engage foreign trainers who have experience with housing finance in developing countries, and who could provide both "academic" and "hands-on" training so that banks may learn effi- cient ways of doing mortgage business. Additionally, foreign training trips to coun- tries in the region (for example, India, Malaysia, Singapore, or Thailand) could show trainees case studies of ongoing and actual housing finance projects. Among the regional countries, Malaysia provides the most relevant housing finance system for study and adaptation in Pakistan. As with all reforms, however, training is only a supportive component. Success is determined by the adequacy of resources committed to the reforms, the consistency and follow-up of implementation actions, and the level of political will associated with the reforms. 276 Appendix D D Notes 1. Data presented here are from JS Global Capital Limited, Pakistan, July 2009. 2. The figure also notes that this income segment is experiencing at most a housing shortage of 375,000 units. (There is a bias in this calculation, because the housing shortage out of the total backlog is assumed proportionate to the share of population in each income bracket; more real- istically, the housing shortage for this income bracket is negligible.) 3. Information is available at http://www.pakistan.gov.pk/divisions. 4. Information is available at http://www.apca.org.pk. 5. Information is available at http://abad.com.pk. 6. Information comes from the Securities and Exchange Commission of Pakistan, Internal and External Communications Unit, July 2008. 7. Information is from the SBP housing finance reviews of 2007 and December 2008. 8. These figures are presented net of nonperforming loans, which accounts for the difference in total outstanding amounts when compared with figure D.4. 9. This section is adapted from World Bank (2008b). 10. This information comes from Kamran Shehzad, deputy governor of the SBP, in a speech at a 2009 workshop on housing finance in South Asia, Karachi. 11. A sukuk is a Shar'ia-compliant financial instrument, similar to a bond in non-Islamic finance. 12. Khuda-ki-Basti-4 in Lahore is a variation on the original Khuda-ki-Basti model of incremental housing for the poor in Karachi. The Khuda-ki-Basti-4 Lahore model uses a more commercial approach than does the original Karachi model. 13. Information is available at http://www.akdn.org/default.asp. 14. Kemal (2008) delves in considerable detail into various issues, including corruption in titling and registration process; and gives some suggestions for improving the situation. 15. The rate is levied on historical value, not the market value of the property. In the city of Karachi, the commercialization fee is levied by the City District Government Karachi on a per-square- foot basis. A proposal is in the offing to increase the commercialization fee up to three times so as to reflect the increase in prices of real estate during the last three or four years. 16. This annual provincial budget report is available at http://www.punjab-prmp.gov.pk/Evidences- 2009/2.4%20WhitePaper2009-10.pdf. 17. Balochistan generates the lowest revenues from stamps and registration fees: PRs 160.5 million (fiscal 2005/06) and PRs 17 million, respectively. For Sindh, stamps account for PRs 3.81 billion (fiscal 2004/05), but sale of stamps yields only PRs 1.94 billion; revenues from registration fees total PRs 0.5 billion. For North-West Frontier Province, stamps and registration fees yield PRs 240 million (fiscal 2004/05) and PRs 40 million in revenues, respectively. 18. The second major source is land revenue. 19. Patwaris hold and maintain the land records in revenue offices. Any land document, including mutations and "fard" (a part of revenue records), is prepared initially by a patwari; it is then ver- ified and signed by the gardawar (supervisor of the patwari) and presented to the subregistrar for final signature. 20. Although an oral gift of immovable property under Islamic law commands widespread juristic acceptance, the courts often have insisted on rigorous evidence of such gift having been made. Yet another feature of the law that adds to the uncertainty of ownership rights is the acceptance of the concept of benami ownership. This acceptance leads to the claim that the ostensible owner of a property is, in fact, not the real or beneficial owner. Such a claim inevitably results in involved litigation, with the so-called real owner seeking to avoid the transactions entered into with the ostensible owner. 21. This information is taken mainly from a 2008 press release issued by the Internal and External Communications Unit of Pakistan's Securities and Exchange Commission. 277 Appendix D D 22. Another weakness of the law is that the tenant rights with respect to the landlord in default cur- rently do not transfer to the creditor institution holding a mortgage over the leased property; rather, they require court interpretation. The law gives the Banking Court the power to termi- nate a bona fide lease on reasonable terms, and that undermines the sanctity of contractual engagements. 23. Much of the information in this section is based on material available at the SBP Web site, http://www.sbp.org.pk/. 24. Section 25(A) vests with the SBP the power to call for credit information in such a manner as it may deem necessary, and to make such information available to any banking company. 25. Section 53-A, Transfer of Property Act, 1882. 26. Amending the Transfer of Property Act and/or the Qanun-e-Shahadat Order, 1984. 27. Amending the Transfer of Property Act, 1882. 28. Lease agreements with a term of 12 months or more already are subject to this requirement (section 17, Registration Act, 1908). 29. Cagamas Berhad, the national mortgage corporation in Malaysia, is the major issuer of asset- backed securities. 30. Securitization requires a developed primary market (if only for cost efficiency considerations), trustworthy property collateralization, credit enhancement tools, specific prudential and accounting regulations, reliable historical information on credit and prepayment risks, and option pricing capabilities among investors. References Hassler, O., and S. Walley. 2007. "Mortgage Liquidity Facilities." Housing Finance International December 1. Kemal, Ali. 2008. "Property Rights and Corruption in Pakistan." Blog posting, January 5. http://makemal.blogspot.com/2008/01/property-rights-and-corruption-in.html. Rahman, Aun, and Batool Hassan. 2009 "The Challenges of Expanding Saiban: Scaling Afford- able Housing for Low-Income Communities in Pakistan." http://www.acumenfund.org/ uploads/assets/documents/Challenges%20of%20Expanding%20Saiban%20Scaling%20 Affordable%20Housing%20for%20Low-Income%20Communities%20in%20Pakistan_ YnSGfsNG.pdf. Rizvi, Zaigham Mahmood. 2007. "Affordable Housing for Low-Income Groups." Presentation prepared for the Third International Conference on Construction Industry, Karachi, Pak- istan, August 28. SBP (State Bank of Pakistan). 2007. "Recommendations for Nationwide Provision of Housing Finance." Housing Advisory Group, Karachi. ------. 2008a. "Development Finance Quarterly Review--December 2008." Karachi. ------. 2008b. "Quarterly Performance Review of the Banking System­December 2008." Karachi. ------. 2009. "Quarterly Performance Review of the Banking System­March 2009." Karachi. UN-HABITAT (United Nations Human Settlements Programme). 2008. State of the World's Cities 2008/2009: Harmonious Cities. Nairobi, Kenya: UN-HABITAT. http://www. unhabitat.org/pmss/getPage.asp?page=bookView&book=2562. World Bank. 2008a. Doing Business 2009: Comparing Regulation in 181 Economies. Country Profile for Pakistan. Washington, DC. ------. 2008b. "Expanding Housing Finance in Pakistan." Washington, DC. 278 Appendix E 1 Sri Lanka Economic growth averaged 5.2 percent for the period 1991­2000. The current Sri Lankan population is estimated at approximately 20.2 million, and its growth rate is one of the lowest in the developing world (1.0 percent annually). Nearly 15­20 percent of the Sri Lankan population is urban--and that population is expected to grow at 3­4 per- cent a year. However, demographics--and resulting housing needs--are quite uneven in the various regions of the country. Housing needs are estimated to range from 350,000 units to 1 million units. The cost of housing is being affected by the rapidly rising cost of land in major metropolitan areas, coupled with the rising cost of construction. Nationwide, 76 percent of the housing units are owner occupied. Nearly 90 percent of the houses are in the "single house" category. People-per-room density is 1.1, a number that mirrors the European average. Nearly 53 percent of the houses have brick walls, and another 20 percent have walls of cement block. Three fourths of the housing units have cement floors, and another 20 percent have mud floors. For roofing, tiles are used in most cases (58.6 percent), followed by asbestos and metal sheets. Much of the existing stock has been self-constructed. The developer market is inactive and focused on large-scale or top-end construction. The size of Sri Lanka's housing finance sector is 6 percent of GDP--much lower than the world average of 30 percent, but the highest in the region. 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 for 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. The mortgage lending market has been swelling in the past several years, at real annual rates of 10­30 percent. However, the mortgage market is 279 Appendix E E limited both in terms of players (only commercial banks and state-owned specialized banks are active, and in its range of products. 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 solu- tions (such as covered bonds or a liquidity facility) would make this possible. In the longer run, when a sizable primary mortgage market of a certain scale develops, securi- tization will become a viable option. Microfinance lenders are numerous in Sri Lanka, with varying degrees of com- mercial success. Much scope remains in the low-income housing and housing finance markets to expand both outreach and affordability. A major development in the microfinance sector was the progress made in formulating the proposed Microfinance Institutions Act by the Central Bank of Sri Lanka (CBSL). Housing policy focuses on improving government land use and maximizing the use of the existing housing stock by providing basic public services and upgrades. The share of state-owned housing institutions--such as the State Mortgage and Investment Bank (SMIB), the Housing Development Finance Corporation (HDFC), and the National Savings Bank (NSB)--has come down to about one third of the mortgage market share, as the pri- vate sector has displaced the government as the primary housing finance provider. The housing finance sector needs to have a supportive regulatory framework that does not impede its growth, but stimulates it. The potential constraints on the develop- ment of a market for home mortgages in Sri Lanka may be divided into three categories. The first category relates to the legal and regulatory environment. The Land Registry Office estimates that it will take at least 15 more years to fully implement the new cadas- tral and title system. Though parate is now more widely available, its effectiveness has been questioned. Registration charges remain high. Market data are scarce, precluding fast-response policy decisions and careful market analysis. The second category 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 category relates to the total supply of financial savings and the institutional arrange- ments to access those savings. Overview Economic growth in Sri Lanka averaged 5.2 percent over the years 1991­2000; and it has held steady around 5 percent for the past decade, in spite of conflict and the economic impact of the tsunami (figure E.1). Much of this resilience has to do with the country's relatively developed human capital and the continuation of market- friendly reforms enacted since 1970. The GDP at current market prices (which was SL Rs 322 billion in 1990) reached SL Rs 1.29 trillion in 2004, and is estimated at around SL Rs 4.31 trillion in 2008 (CBSL 2008). Based on the central bank's figures for the year 2008, the service sector is the major contributor to GDP (59.5 percent), followed by industry (28.4 percent), with transport and agriculture at 13.1 percent 280 Appendix E E Figure E.1 Sri Lanka GDP Growth, 2000­08 10 8 6 percent 4 2 0 2 2000 2001 2002 2003 2004 2005 2006 2007 2008 year Source: World Bank 2009. and 12.1 percent, respectively. Home remittances from more than 1 million Sri Lankans working abroad is another significant contributor to the economy. The total migrant population is 1.8 million, and female workers were nearly 65 percent of the total overseas workers. Foreign remittances during the year 2008 amounted to $2.9 billion (CBSL 2008, p. 98). The government priorities emphasize infrastruc- ture, including housing, in the imminent national growth agenda. The buildup of the housing and real estate sector calls for 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. The CBSL reports a GNP per capita at market prices of $1,969 for the year 2008, compared with $881 per capita for the year 2000 (CBSL 2008). In the few years preceding the global financial crisis, the country's inflation increased from less than 10 percent to more than 13 percent, and lending and mort- gage rates have mirrored the inflation patterns with a margin (except for those of the National Housing Development Authority [NHDA]). Real government bond rates have been low, especially in 2005 and 2006. Deposit rates have been close to zero in real terms. The national savings rate in Sri Lanka is comparatively low: 18.2 percent during 2008, a decline from 23.0 percent during 2007. One reason for a low savings rate is the rising inflation rate (13.7 percent in 2006) that results in a negative return on savings and a rising demand-supply gap. The 2008 Sri Lankan population is estimated at around 20.2 million, with one of the lowest population growth rates in the developing world (1.0 percent in 2008). Nearly 15­20 percent of the population is urban, an amount that is expected to grow at 3­4 percent annually. The population density is 322 people per square kilometer. 281 Appendix E E However, in 4 of the country's 25 districts, the population density is comparatively very high, resulting in massive urban housing issues. The people-per-square- kilometer density of Colombo (3,330) is followed by those of Gampaha (1,539), Kalutara (677), and Kandy (667) (Sarathchandra 2008, p. 11). About 46 percent of the country's population is below the age of 25 years, and 14 percent of the popula- tion is over the age of 54 years (CBSL 2008). The population is aging, however, with 69 percent in the 15­54 age group in 2008 (compared with 55 percent in 1990), and with 15 percent in the 55-plus 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 signifi- cantly, hovering just above 15 percent since 2000 (World Bank 2009). In the coming decades, with the urban population growing annually at 3­4 percent, 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 con- dominium growth. The size of the housing finance sector is 6 percent, much lower than the world aver- age of 30 percent. 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, with an average rate of 30 percent. Housing Demand The demand for housing and urban infrastructure has been growing with the increas- ing population, rising income levels, urbanization, and changes in socioeconomic lifestyle. The government's involvement in housing development is mostly limited to a facilitator role, rather than a direct provider role. With the increase in the income lev- els and changing lifestyles, the private sector has emerged as the major provider of houses for middle- and high-income people; the government continues to be involved in providing housing facilities for low-income households and specific groups. With a population of more than 20 million and an average household size of 4.1 people in 2008,1 there are about 5 million households in the country. Against a potential demand for 5 million housing units, Sri Lanka has an existing stock of 4.7 million units. Official estimates put the existing demand-supply gap at 350,000 housing units. According to the Ministry of Housing and Plantation Infrastructure, the overall shortage could reach 650,000 units by 2010 (World Bank 2007b); some other estimates place this figure at 1.5 million units by the same year.2 Sarathchandra (2008) believes the shortage had already reached 1.0 million units in 2007 (p. 11). The CBSL estimates the annual demand for new houses to be increasing at the rate of 50,000­100,000 units. However, the yearly incremental demand not being met is roughly two thirds, leaving an unmet demand adding to the backlog. The damages to housing as a result of the 2004 tsunami have further aggravated the 282 Appendix E E challenge to end the housing shortage. Because Sri Lankans appear to be moving toward peace and stability in the north, there will be a massive pressure on housing reconstruction and rehabilitation in that part of the country. The cost of housing is affected by the rapidly rising cost of land in major met- ropolitan areas, coupled with the rising cost of construction. The cost of building has increased about threefold since 1990, thus inhibiting the growth of housing construction. According to the cost index of housing construction in Sri Lanka, tak- ing 1990 as the base of the index (that is, 100), the index in 2000 was 184, and 304 in 2005. Market failures preventing the closure of the housing gap can be attributed to a faulty regulatory framework (land and titling, eviction rules, and so forth); an under- developed housing provision market; improperly targeted 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 being able to afford a house. Low penetration of banks and microfinance institutions into households from these income groups is another contributory factor. The hous- ing gap is predicated on weak credit information, the lack of technology, a perceived risk in lending to irregularly salaried employees, and problematic enforcement of eviction provisions. Conversely, drastic improvements in education and literacy have helped Sri Lankans feel more comfortable with banks and financial instruments, including mortgages. At 41.6 percent of the population, the poverty head count at $2 per day is high in Sri Lanka (World Bank 2009). Low savings rates through formal financial intermediaries have contributed little to the growth of mortgage financing, and household expenditure on housing is around 11 percent (CBSL 2006a). 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. An alterative view on housing cautions against overestimating housing needs. National income and output accounts do not capture owner-financed equity invest- ments in owner-occupied housing, nor the returns to such investments 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 invest- ment in the housing stock. Simply because of the failure to measure informal hous- ing construction and owner contributions to formal construction, 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. 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 options and the 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 neces- sary 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 283 Appendix E E housing units. Hence, the government might perceive a need to provide the "miss- ing" 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. Housing Supply The supply of luxury housing and home financing abounds. Simultaneously, many middle- and lower-income households have no access to home financing (some have no homes), and they face scarce market rentals. The efficiency of land use could be improved, and housing turnover is low. In rural areas, the housing supply is even weaker, and housing is self-constructed. Current State of Housing The Department of Census and Statistics put the country's estimated housing stock at 4.7 million units in 2001. Sri Lanka's five major provinces--namely, Central, North Western, Sabaragamuwa, Southern, and Western--together account for approximately 77 percent of the total housing stock. Four of the country's 25 districts--Colombo, Gampaha, Kurunegala, and Kandy--provide more than one third of the country's housing stock. Sri Lankans, by and large, live in rural areas: more than 80 percent of them. Except for the capital district, Colombo (which is 50 percent urban), the rest of the country follows the majority rural pattern. Nation- wide, 76.0 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 units are 8.1 percent and 6.4 percent, respectively. Huts and other types of housing units constitute only 1.7 percent. The capital city comprises 608,103 building units, of which 82 percent are classi- fied as housing units and the balance is composed of 8,477 collective living quarters, 2,599 institutions, and 89,349 nonhousing units. The majority of housing units (72.3 percent) are owner occupied, 17.5 percent of households live in rented or leased housing units, and 4.8 percent of households live rent-free. Another 1.5 per- cent of households live in encroached housing units. 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 hous- ing turnover also makes the initial appraisal more problematic. Turnover is a more acute problem in rural areas. 284 Appendix E E Furthermore, there is virtually no rental market for middle- and low-income renters. Rental units are especially scarce for households that cannot afford to pur- chase 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. Development of the Housing Stock 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 exist- ing stock of multifamily urban houses is in condominiums. The government con- verted 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 are rented. Some of the new subsi- dized condominium apartments allocated to middle-income buyers (including gov- ernment 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. Devel- opers frequently work with government housing institutions on lower-income hous- ing projects, 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 7­8 months to get a condominium permit. Side pay- ments 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 include major banks and DFCC Bank). Private commercial banks might not be overly eager to finance development because of the complications in 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. A buyer cannot bor- row against future title because the title is held by the developer until completion of the units. Even then, delays in titling are not uncommon--even delays of one to two years may 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 comple- tion. Condominiums are frequently resold for profit upon completion. The new legislation regulating condominium construction has some shortcom- ings.3 Development of high-rise buildings is restricted. Professional property man- agement companies have not 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 (by law, comprising the owners) have not had the ability to subcontract for professional management. Such subcontracting may occur if, for example, a developer spins off a 285 Appendix E E property management company, which then seeks to expand by bidding for work with independent condominium associations. Construction Quality In Sri Lanka, nearly 90 percent of the houses are in the "single house" category, with an average of four rooms each. One measure of the quality of housing would be the size of the house. Whereas more than one third (36 percent) of the houses are less than 527 square feet of covered area, nearly one fourth are less than 928 square feet (CBSL 2003). The average covered area of a house is comparatively much smaller. Unless these units are part of multistory, multifamily units, such small-size houses are mostly one-story dwellings and are not made of quality construction. The people-per-room density in Sri Lanka (1.1) mirrors the European average (1.1) and is lower than the average in India and Pakistan (3.5). Based on the construction material used, nearly 53 percent of the houses have brick walls and another 20 percent have walls of cement block. Three fourths of the housing units have cement floors, and another 20 percent have mud floors. Tiles are most frequently used for roofing (58.6 percent), followed by asbestos and metal sheeting (CBSL 2008). Both the initial quality of construction and the ongoing maintenance and repair are somewhat deficient. Most of the new housing stock being added will likely be maintained and repaired by condominium associations. Maintaining and repairing the stock formerly owned by the government and not fully converted remain the responsibility of the government. 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 pro- fessional contractors because people typically act as their own general contractors. 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 in the way that a general contractor would do so, nor can they evaluate material quality. People tend to stick to traditional construction methods (including concrete, clay, and brick construction) that are labor intensive and time consuming. The cost of construction increased about threefold between 1990 and the end of 2005 (Sarathchandra 2008, p. 14), inhibiting the growth of house construction (see table E.1). Building materials that registered substantial price increases since 1990 include sand (1,070 percent increase), timber (568 percent), and bricks (678 percent). Labor cost 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 has taken hold in domestic construction. More work on reducing con- struction costs is needed. 286 Appendix E E Table E.1 Cost Indexes of Housing Construction, Selected Years 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 government has been making progress in providing electricity, water, and sanitation services to households. Electricity is available to more than three fourths (83 percent) of households, whereas only 4 percent had the service at the time of independence (1948). The water sanitation and sources also have improved substan- tially, with 84.7 percent of households having access to an improved water source and 93.9 percent having separate toilets.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 for private commercial banks in the immediate term and a functioning secondary mortgage market in the long term. Past policies that centered on government-owned institutions, nationalized land, and housing provision programs diverted scarce fiscal resources from other poten- tially more productive investments and likely contributed to the less-than-desired rate of economic growth.5 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 hous- ing "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 com- ponents. The first component is to enhance and encourage private sector participation. The second is to put government land to higher and better uses, with the proceeds from the increased land value paying for housing for displaced dwellers and for low- and 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, 287 Appendix E E although inevitably it is involved indirectly through the land disposition and devel- opment 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 nonviable 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 the board's dictates. The state institutions do not have adequate governance, management, and opera- tional efficiency (although there is improvement in some state banks, such as the Bank of Ceylon and People's Bank) to play a sufficiently active role in providing lower-income housing beyond their current contributions, or to better use the government budget to the wider benefit of Sri Lankan society. 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). These executive organs facilitate and coordinate operation of the institutions that implement government housing policy. The imple- menting agencies and state housing banks are insufficiently focused on lower- income housing, failing to address market deficiencies in that sector, misdirecting (implicitly or explicitly) subsidized loans to middle-income markets, and crowding out potentially viable commercial alternatives. The NHDA is the national government entity that implements government hous- ing 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 develop- ments, including roads and water. In addition, in its early days, the 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 subsi- dies were curtailed in the mid-1990s, the NHDA had to redirect its operations to more viable middle-income population groups. The NHDA manages substantial assets of mostly urban land owned by the gov- ernment, using it for housing development projects. Such development is typically done in a joint venture with a private sector developer: the 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 for middle- income households.6 The deep subsidy and imprecise targeting of need-based housing provision enables buyers to abuse the system. The NHDA engages in urban renewal and joint ventures on private land. Its com- parative 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 288 Appendix E E property against the acquisition price. For such developments, average unit prices run from SL Rs 800,000 to SL Rs 1,200,000, with a downpayment of SL Rs 7,500. Financing is secured mostly through state-owned banks, but private financing is also available at market rates. The NHDA experimented in 2007 with direct subsidies to buyers financed from straightforward land sales to developers. The Urban Development Authority (UDA), much like the NHDA, has a primary responsibility for addressing low-income needs--but with 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 munici- pal governments (for which it develops municipal government land). UDA keeps a significant focus on housing needs of displaced households. As is true for the NHDA, most low-income programs have been curtailed. For example, past strategies of con- structing low-cost, high-rise (14-story) buildings for low-income groups have been abandoned because of high maintenance costs. Five-story walk-up structures are now favored. Several organizations are active in low-cost housing development. The National Building Research Organization 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 in upgrading urban shanties. The Plan- tation Human Development Trust upgrades houses in the estate sector, where about 6 percent of the population lives, and where housing facilities are mainly provided in line rooms. Housing Finance National housing needs are 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 demands a mortgage; that is, individuals totaling 3.0­4.5 percent of the population had taken out mortgages in 2006.7 To boost effective demand to match actual national housing needs, the availability of housing financing is essential. Another strategy to close the divide between effective demand and housing needs is to lower the cost of housing. Estimates suggest that a majority of the population could afford low-income housing at current prices if financing were available and credit screening were appropriate. Housing finance does not only involve mortgage lending on new construction or purchased property. Substantial funds are needed also for housing rehabilitation, although this is rarely provided through mortgage financing. Microfinance lenders tend to finance housing improvements directly through 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 activities. 289 Appendix E E Figure E.2 Current and Potential Housing and Housing Finance Markets at a Glance 100 90 Bank market 80 70 60 percent potential market 50 40 MFI market 30 20 10 0 serviced by bank serviceable clients serviceable clients (bank and mortgages (bank market growth) MFI market growth, new housing) serviced by serviceable clients not bankable micro-credit (MFI market growth, housing upgrades) Source: World Bank 2007b. Note: MFI microfinance institution. The upper portion of the current market includes the wealthier bankable house- holds with regular salaries (white rectangle at the top of figure E.2). The 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 (striped white rectangle) (World Bank 2007b). The lower half of the market consists of households that cannot afford to pur- chase a new home, but already own one (blue and dotted blue rectangles). Micro- finance companies currently are responsible for financing housing maintenance and upgrades for a small portion of homes and small business (blue rectangle). The gap between the population served by commercial banks and that financed by microlending is the housing finance gap (striped blue rectangle) and represents the focus of this report. This gap includes households that can afford to buy new, low-cost housing but are not considered bankable by commercial institutions. This gap also includes families who own their homes and need upkeep funds, but have not been reached by the limited scope of the microlenders. Downscaling by banks, new credit-scoring techniques not relying on collateral, and better credit infor- mation would push some of the striped blue-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. 290 Appendix E E For households in the lowest income deciles, improvements in basic housing con- struction technology would lower home costs and enable 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 black rectangle 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. The financial sector in Sri Lanka is quite developed (see table E.2). It has a wide range of financial institutions, including 22 commercial banks, 14 specialized banks, and 34 nonbank financial institutions (among them, 22 leasing companies). Out of 22 licensed commercial banks, 2 are state-owned banks, 9 are domestic private banks, and 11 are foreign banks. In addition, there are 5 licensed specialized banks, which are publicly owned. The two state-owned banks (the Bank of Ceylon and People's Bank), under their mandate, provide financial services throughout the country to target the government's priority sectors rather than to pursue purely commercial objectives. The state-owned participants in the housing finance market accounted for about one quarter (that is, SL Rs 39 billion) of the total outstanding housing loan portfolio of SL Rs 172 billion, as of 2006 (Sarathchandra 2008, p. 34). Adding the outstanding housing loan portfolio of SL Rs 8.8 billion held by registered finance companies, the total outstanding loan portfolio of the financial institutions stood at SL Rs 181 billion at the end of 2006. The private domestic commercial banks control nearly half of the market, state- owned banks share 40 percent, and foreign banks share about 10 percent. By far the largest of the specialized licensed banks is the NSB, which is the second- largest deposit-taking institution in the country (behind the Bank of Ceylon). It 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 the NSB keep operating costs and credit risk low because its systems for credit risk management and credit administration are weak. The bank has under- taken limited computerization, coupled with plans for automated teller machines and electronic funds transfer capabilities. It has begun to develop a housing finance business for its customer base; in particular, the NSB aims to assist savers who want to expand their homes as dwellings for the extended family or as home-based small businesses. The bank has 112 branches. The Housing Finance Market The housing finance market is shared between private commercial banks and state-owned banks, and is not characterized by a wide variety of financial insti- tutions. The main financial institutions engaged in housing mortgage lending are private commercial banks, with 75 percent of market share; the NSB with 291 Appendix E E Table E.2 Distribution of Banks and Bank Branches, 2007 and 2008 Category End-2007a End-2008b Licensed commercial banks (LCBs) Total number 23 22 Domestic banks 11 11 Foreign banks 12 11 Total number of LCB branches and other outlets 4,331 4,733 Branches 1,934 2,068 Domestic bank branches 1,758 1,894 Main branches 1,261 1,355 Extension/pay offices/service counters 497 539 Overseas branches 12 12 Foreign bank branches and other outlets 176 174 Branches 43 41 Other outlets 133 133 Student savings units 2,397 2,665 Licensed specialized banks (LSBs) Total number 15 14 Regional development banks 6 6 Regional savings banks 2 2 Long-term lending institutions 3 2 Housing finance institutions 2 2 Private savings and development banks 2 2 Total number of LSB branches and other outlets 620 636 Branches 409 423 Regional development banks 218 219 National savings banks 118 125 Long-term lending institutions 14 18 Housing finance institutions 28 28 Private savings and development banks 31 33 Other outlets 36 38 Student savings units 175 175 Total number of commercial and specialized bank branches 4,951 5,369 and other outlets Total number of automated teller machines 1,370 1,583 Total number of electronic funds transfer facilities at 19,227 23,447 point-of-sale machines Banking density Number of LCB branches per 100,000 people 9.7 10.2 Source: CBSL 2008. a. Revised. b. Provisional. 292 Appendix E E Table E.3 Total Outstanding Housing Mortgage Loans in the Financial System Total Mortgage Total Percent Percent of Loans, 2005 Loans of Total Total Financial Item (SL Rs millions) (advance) Loans System Assets Licensed commercial banks 76,223.0 647,900.0 11.8 2.42 NSB 11,120.0 25,578.0 43.5 0.35 SMIB 8,406.0 8,475.0 99.2 0.27 HDFC Bank 7,858.0 7,900.0 99.5 0.25 Regional development banks 3,483.0 16,587.0 21.0 0.11 Registered finance companies 7.7 62.2 12.4 .. Source: World Bank 2007b. Note: . . negligible. another 9 percent; and SMIB and the HDFC with 7 percent and 6 percent, respectively. The mortgage lending market, which grew from 9 percent to 14 percent annually in the early 2000s (2000­03), is rapidly expanding; and since 2004, the annual growth has been around 30 percent. With their aggressive entry into housing finance beginning in 2000, commercial banks now dominate the housing finance market.8 Table E.3 shows that the outstanding mortgage loans made by commercial banks have grown substantially since 2000, moving from SL Rs 41 billion that year to SL Rs 133 billion in 2006. The housing portfolio of the three specialized mortgage lend- ing institutions grew during the same period from SL Rs 14 billion to SL Rs 32 billion. The specialized banks with the major share of mortgage lending are SMIB9 and the HDFC.10 SMIB, "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). In view of the growing competition in the housing finance market, particularly from the commercial banks, SMIB has shifted its business focus to the low- and middle-income segments; and its current average loan size is SL Rs 500,000. Most of its housing finance is in long-term, fixed-rate mortgages. The HDFC engages in housing construction and house financing activities; SMIB is more involved in pro- viding grants and subsidized loans to lower-income groups. Commercial banks and the NSB are rapidly becoming active consumer lenders, including for home loans. A well-developed, albeit small, network of microfinance lenders (amounting to 0.1 per- cent of financial institution assets) currently funds housing repair and upgrades; but given their small current level of outreach, the lenders in this network 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. Three small private savings and development banks collectively account for about 3 percent of the assets of the licensed specialized banks. 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 and thus affiliated with Seylan Bank. The Ceylinco Group includes 293 Appendix E E Table E.4 Mortgage Lending by the Banking System, 2000­06a Commercial Bank State (Specialized) Mortgage Lending Bank Mortgage Lending Outstanding Percent of Percent of NSB SMIB HDFC Mortgage Loansb Total Total Financial (SL Rs (SL Rs (SL Rs Year (SL Rs millions) Loans System Assets millions) millions) millions) 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 8,110 2006 133,640.4 16 4 16,098 -- -- Source: World Bank 2007b. Note: -- not available. a. Data are not clearly separated between property loans for houses and other real estate, and between construction and per- manent (that is, home mortgage) financing. The table indicates the percentage of loans and total assets invested in mortgage loans. Finance companies and National Development Bank loans are not included. b. Includes the two state banks because they are categorized under licensed commercial banks. the largest building society, Ceylinco Credit and Investments, and is active in the housing sector. Table E.4 shows the position of outstanding mortgages by commer- cial and specialized banks. The commercial banks enjoy nearly 75 percent of the total market share; the total outstanding mortgages, which stood at SL Rs 56 billion in 2000, reached SL Rs 172 billion in 2006. Traditional Mortgage Lenders 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--low by compar- ison with other countries. Historically, lending to the government and state-owned enterprises has been more than 10 times the amount lent to households for housing. Commercial banks have been shifting toward an increased share of housing finance in their portfolios (table E.5). 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, in 2006. Housing loans generally have fixed rates set by the state lenders. These loans are variable at the licensed commercial banks, secured with a first mortgage on the house and supported by mortgage protection policies against the death or disability of the borrower. Policies can be obtained to protect against title deficiencies. Reha- bilitation and upgrading can be financed only with personal loans because second 294 Appendix E E Table E.5 Commercial Banks' Advances for Housing and Property Development, 2000­06 Loans for Loans for Loans for Personal Housing, Construction of Property Including Purchases Business Premises Development SL Rs Percent of SL Rs Percent of SL Rs Percent of Year millions Total Loans millions Total Loans millions 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: CBSL 2007. 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 because fixed rates have almost doubled in the past few years. The state banks use an escape clause to allow for rate adjustments (and the clause might be politically difficult to utilize). Because of these potential rate adjust- ments, fixed-rate lenders are faced with interest and funding risks. Maturities range from 15 to 25 years (or more), with prepayment options avail- able to the borrower. Microlending terms are between 2­4 and 10 years. Bank bor- rowing, 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 recently were 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 bank profitability remains weak, and profits have declined from prior levels. The NSB offers a fixed rate for 20 (sometimes 30) years, currently set at 14.5­15.5 per- cent, reflecting its cheaper deposits. To be competitive, SMIB and the 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 con- sumer lending rates (table E.6). A well-established yield curve for government bonds facilitates financial market development and helps price various financial instruments, including mortgages and mortgage-backed securities. Although treasuries do trade, a robust yield curve is not yet established in the country. The maximum payment-to-income ratio appears to be high by international standards. Most lenders will lend up to a maximum of 40 percent of the ratio of debt service to income for low- to middle-income bor- rowers, and up to 60 percent for higher-income borrowers (up to the lenders' maxi- mum loan limits). The maximum loan-to-value ratio is somewhat conservative--at 75 percent for housing (50­60 percent for land)--because it is based on "recovery" 295 Appendix E E Table E.6 Comparison of Mortgage Rates with Other Consumer Lending Rates Overdraft Credit Card Auto Leasing Pawning Mortgage Institution Rate (%) Rate (%) Rate (%) Rate (%) Rate (%) National Savings Bank Not offered Not offered Not offered 18 14.50­15.50 Bank of Ceylon 20­32 Effective 34 19­21 20 15.50­15.74a People's Bank 17­19 Effective 35 Effective 24 18 17.00 Hatton National Bank 20­25 Effective 35 18­20 18­20 17.75­18.75 Commercial Bank 25­32 Effective 35 19­21 Not offered Fixed 18.00, floating 14.00­20.00 Source: World Bank 2007b. a. The rate is 0.756 percent to 1.000 percent over the weighted prime lending rate (14.74 percent in 2006). rather than market value (and 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-downpayment loans. The average loan size is SL Rs 1 million at private domestic commercial banks and the NSB. The state housing corporations do not target clients much farther down the income scale. SMIB's average loan is SL Rs 0.6 million, and the HDFC's average loan is slightly lower at SL Rs 0.25­0.50 million. Microfinance lenders offer much smaller loans to small and medium-size enterprises (SL Rs 15,000­100,000), although they also originate loans of SL Rs 0.5 million. 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 2.0 percent. Overall, loan terms could be characterized as mildly liberal in the context of Sri Lanka's foreclosure and eviction environment and the absence of mortgage insurance. Efficiency and Performance of the Main Housing Finance Institutions The banking system in Sri Lanka is liquid, with private sector credit expanding since 2004 (although at a lower rate than in 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 sec- tor that is reluctant to expand mortgage lending to a wider middle-income group or to take private sector risks. This reluctance results perhaps from a perceived lack of bankable opportunities and high entry costs into a new market segment or from a lack of sufficient credit information and adequate credit-scoring mechanisms to manage risk effectively. The Asian Development Bank's financial sector assessment notes that high spreads result from rigid and inflated lending rates, not sticky deposit rates. This means banks have high operating costs and high levels of nonperforming loans, which push up lending rates. Private banks predictably have lower numbers of nonperforming loans than do their state-owned counterparts; but both measures 296 Appendix E E Table E.7 Nonperforming Loans of the Banking System, Selected Years 1998 (% of total 2003 (% of total 2006 (% of total Institution advances) advances) advances) Commercial banks -- -- 5.4 State-owned banks 19.5 17.9 -- Private domestic banks 13.4 12.0 -- Foreign banks 12.7 6.9 -- Specialized banks -- -- 8.7 Source: CBSL 2007. Note: -- not available. Table E.8 Efficiency of State-Owned and Large Private Banks, 2006 Average of Bank of Four Large Indicator Ceylon People's Bank Private Banks Noninterest expense (% of total income) 61.9 68.7 52.4 Average branch size (assets, SL Rs millions) 827.8 645.0 899.5 Assets per employee (SL Rs millions) 27.4 18.1 33.2 Personnel cost (% of average assets) 2.1 3.2 1.7 Source: Banks' annual reports for 2006. stand in stark contrast to the much smaller bad debt amounts held by foreign banks (table E.7). For 2006, however, the CBSL reported significantly lower nonperform- ing loan figures for the commercial banks, no doubt reflecting bank restructuring efforts. Licensed specialized banks carry a slightly higher level of nonperforming loans (8.7 percent). The averages belie the situation of individual lenders. For exam- ple, SMIB continues to record a high level of nonperforming loans (25.0 percent). Table E.8 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.11 Furthermore, the seg- mentation of the borrowing market sustains higher rates. State housing banks have poor performance the world over. They are not disci- plined by the market, so operating costs are usually high and credit losses are usually significant (table E.8). Banks are reluctant to enforce liens for social or political rea- sons, so 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 cited above indicate. State housing banks have higher nonperforming loan 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 the Bank 297 Appendix E E of Ceylon and People's Bank.) 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). Within the group of four relatively large private domestic banks, there are distinct differences in ownership and financial perform- ance. Hatton National Bank and Seylan Bank have financial performance that is gen- erally inferior to the more widely held Sampath Bank and Commercial Bank, with higher proportions of nonperforming loans. The Bank of Ceylon consistently has been the most profitable domestic bank, in large part because of its much lower non- interest expenses. Sampath Bank, despite having average branch size and business volume per employee similar to the Bank of Ceylon, has not achieved the same lev- els of efficiency; and thus has achieved profitability only marginally better than that reported by Hatton National and Seylan banks. Risk Management State-owned specialized lenders cannot offer adjustable-rate loans because they do not have a servicing system capable of making such adjustments. Mortgage origina- tion and servicing each requires specialized human skills, investment in infrastruc- ture, and information technology. Third-party specialists provide appraisal, legal, brokerage, and other services. Absent the technology and systems for efficient oper- ations, 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 such risk 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 100­300 basis points of the loan principal annually, substantially improving loan affordability. Some private commercial banks originate a mortgage 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. The ratio of nonperforming loans at specialized mortgage lenders has dropped over the last five years. Some lenders had ratios greater than 20 percent, but that is changing with the restructuring of the two state-owned commercial banks and improvement in the performance of private banks. Extreme moral hazard that may 298 Appendix E E have been present in collections was partially addressed by the introduction of good management practices. Loss provisions for home loans as a percent of nonper- forming loans 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 moni- toring 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 is not expected to represent a systemic issue in Sri Lanka because mort- gages are a small part of overall financial system assets. Nevertheless, the CBSL's "Financial System Stability Review" (CBSL 2006b) 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 assets. This problem has been addressed on the basis of international experience by reviewing the CBSL collateral eligibility rules at the dis- count window12 to maximize industry liquidity. An alternative strategy with success- ful application internationally is the creation of a mortgage liquidity facility. No mortgage lenders have long-term deposits on their balance sheets--that is, deposits 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 lend- ing duration is matched against the duration of capital and is small in proportion to these banks' balance sheets. This is not the case at 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 securitiza- tion. 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 mortgage- backed securities, 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 insti- tutions 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. Mortgage securitization is a sophisticated, complex, and costly 299 Appendix E E transaction that requires considerable regulatory and institutional infrastructure to develop and an 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 accept- ing an adjustable-rate loan. The increased borrower risk will inhibit mortgage lend- ing 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 risk inherent in these types of loans. Growth Potential of the Mortgage Debt Industry Home mortgages are funded by savings that flow 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. The CBSL does not publish a complete flow-of-funds account database, thus making it difficult to project funds potentially available to housing. 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 3,14 trillion in 2006, or about 133 percent of GDP (excluding CBSL assets). About 57.0 percent of this total was held by licensed commercial banks and 7.7 percent was held by the NSB. The two specialized, state-owned mortgage lenders each accounted for less than 0.1 percent. The two public provident funds (the Employees' Provident Fund [EPF] and the 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 20­50 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 licensed commercial banks are invested in the private sector, however. The NSB has about 80 percent invested in government securities; the EPF and ETF combined have more than 90 percent, as do the private provident funds and life insurance compa- nies.13 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 funding sources 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 compet- ing against business as well as government and state-owned enterprises for bank demand and savings deposits. A 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 2006b). Lending by foreign banks is limited by regulation to funds raised domestically--a negligible amount. Total mortgage lending by banks is approximately 6 percent of total bank assets. Domestic deposit growth is about 300 Appendix E E Figure E.3 Projected Growth of Housing Finance Loans from Banks, 2006­20 1,200 1,000 800 percent 600 400 200 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 year Source: World Bank 2007b. 15 percent annually. Under these assumptions, the mortgage financing funds from pri- vate banks will increase seven times by 2020 (figure E.3), 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 0.6­0.9 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 mort- gages.) With population growth at 1 percent, 1.1 million mortgages nationwide translate into one loan per five households (compared with the current ratio of one loan per seven households). So that raises a question about where the funds will come from to finance a significant expansion in home mortgage lending. At the upper end of the market, the potential for the commercial banks and the 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 0.1 percent of financial institution assets, so this potential is limited in the medium term. 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 enable the institutions to operate independently and to serve low-income housing finance needs. 301 Appendix E E 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 term? 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 lend- ing 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 the ETF (absorbing 23 percent of income, including employer contributions).14 These funds operate in a manner similar to savings accounts because households can withdraw up to 75 per- cent of their balance, for example, for a downpayment 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. The EPF does not require evidence that with- drawals 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 periods. Although the EPF and the ETF might not be expected to oper- ate mortgage lending efficiently, their investment strategy could be liberalized, allow- ing the current and past contributions to be reallocated to housing-related invest- ments 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 the ETF--although technically funded plans almost entirely invested in government securities--might warrant more careful examina- tion of 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.15 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 plans are carefully reviewed and firmly funded, however, they are not likely to live up to expec- tations as significant sources of long-term funding for housing in the future. Life insurance companies have been 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 premiums (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. 302 Appendix E E In practice, the figures are much higher--34 percent of nonlife and 73 percent of life insurance assets were held in government securities in 2003.16 The life and nonlife market is dominated by the same three players: Sri Lanka Insurance Corporation, Ceylinco Insurance Company, and Eagle Insurance Com- pany. Life insurance reserves are growing, and 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, of which there are some 200, cover private sec- tor employees who originally contracted out of the EPF. These funds 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, approved private provident funds 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 for mortgages.17 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. Sri Lanka has a relatively small but sophisticated stock exchange, regulated by a well-established Secu- rities and Exchange Commission. The Colombo Stock Exchange has had a paperless, automated trading system and a central depository since the late-1990s. As of June 2008, it had 235 listed companies and a market capitalization of SL Rs 489 billion (a decline from SL Rs 821 billion in 2007, as a result of the global financial crisis). The All Share Price Index fell from 2,541 in 2007 to 1,503 in 2008 (CBSL 2008, p. 180). 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 (listed debentures amount to 0.6 percent of GDP [World Bank 2009]). Commercial banks, leasing companies, and merchant banks issue bonds/debentures on a very lim- ited scale. Such issues are both listed and privately placed, and the volume of debt raised through private placements is much larger.18 With the issuance of Treasury bonds, the debt market is likely to develop further. Sri Lanka has two rating agencies--Fitch Ratings Lanka Ltd. and RAM Ratings (Lanka) Ltd. The credit rating is another prerequisite for development of a debt market, and Sri Lanka has a well- established norm of desiring credit rating for public issue. All five issues of mortgage- backed securities with a total value of SL Rs 1.4 million were originated by the HDFC Bank during the period 2004­06. The Construction Guarantee Fund helps construc- tion contractors increase their operations by issuing bonds. By September 2006, the fund had issued 5,256 bonds in favor of contractors--and nearly three fourths of these bonds were in favor of small and medium-size contractors. Another major source of long-term funding could be foreign investment. Funds would likely take the form of foreign direct investment because foreign investors 303 Appendix E E 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 foreign direct investment, but the amounts have been disappointingly small thus far--most likely reflecting internal security and infrastructure problems. Foreign direct investment 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 mort- gages are homogeneous and there are multiple market participants. For example, it can 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 mortgage-backed securities 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.19 A stamp tax is due on home sales, making these transac- tions 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, there will be few investors with long-term funds in the next 10 years. The underdeveloped bond market and the small size of the stock market might present further problems with the development of a liquid (actively traded) mortgage- backed securities market. Low-Income and Microfinance Housing As Sri Lanka makes the transition from government-directed to private sector­based 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 those in regions struck by conflict or disaster or in 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 be addressed successfully through 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 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 serving 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 304 Appendix E E protections against eviction, and in part because of a culture of home ownership. 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 titles and with disputable records, these income groups are not good candidates for formal mortgage finance. But 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 institu- tions have succeeded in profitably lending, including for housing purposes, without relying on the title to collateral. Frequently, loans to micro, small, and medium-size enterprises (or the profits made from them) 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 is unlikely to pro- vide a major part of housing finance in the medium or even the long term. Microfinance and Housing 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; SEEDS (Sarvodaya Enterprise Development Services), opened in 1988; and the National Development Trust Fund, in operation since 1991. SEEDS uses an interesting busi- ness 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 Habitat for Humanity to grow its operations. SEEDS raises money and lends through 3,000 soci- eties with 500,000 members. These members pick borrowers and service the loans, under SEEDS' supervision. SEEDS has a nonprofit charter under the Companies Act, and 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. Bor- rowers 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 per- cent of the loans are current (performing). Three types of loans are offered: (1) small loans of SL Rs 15,000­50,000 (on average, SL Rs 20,000); (2) enterprise loans of SL Rs 50,000­100,000 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. 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 305 Appendix E E rural areas. Availability of funds to microfinance lenders is not reported to be a con- straint on home lending. The lending through microfinance entities likely will remain segmented from formal home mortgage finance for the foreseeable future. But first attempts at bank downscaling 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. Government's Role in Housing and Housing Finance The National Housing Act 37 of 1954 provided the basic legal framework for hous- ing development. The government had also made attempts to address the low- income housing gap by setting up the NHDA in 1979 primarily to assist the devel- opment of homes for the rural poor through the provision of loans with a ceiling of SL Rs 50,000. The loan could be used either for upgrading an existing house or for constructing a new one. The tenor of the loans was 5­15 years, at an annual interest rate of 11 percent. The NHDA was responsible for developing necessary residential infrastructure around new communities (including roads and water). The govern- ment had also taken the initiative to address plantation workers by forming the Plan- tation Housing and Social Welfare Trust, which enabled plantation workers to access housing finance facilities from SMIB. Existing government programs directed at housing provision for the very poor include those offered by the NHDA, UDA, and the National Building Research Organization. Because these entities currently are not funded from the government budget, they have shifted their focus toward commercially viable income groups. The shift has produced 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. The NHDA offers loans to a maximum of SL Rs 50,000 (but more typically SL Rs 25,000) to enable beneficiaries to upgrade or build their homes. The loans usu- ally are accompanied by some technical assistance provided by a pool of advisers skilled in homebuilding techniques. The NHDA facilitates the development of the necessary infrastructure around new housing communities, including roads and water. Loans must be repaid over a period of 5­15 years, at an average interest rate of 11 percent (ADB 2005). Among existing low-income loans, low recovery has been a major problem. The NHDA further provides some direct building material assis- tance grants, such as roofing sheets. It implements urban redevelopment with its own funds (46,021 housing units provided in 2006). The Department of National Planning intends to rehabilitate urban shanties with free water supply and sanitation services. The Plantation Human Development Trust plans to improve 250,000 line rooms for plantation workers and to make other improvements for the internally displaced population. Real Estate Exchange Ltd. plans to upgrade 66,000 shanties in urban areas over the next 10 years. Under Colombo's Sustainable Township Pro- gram, slum dwellers have been given the opportunity to voluntarily exchange their dwellings and land space for rehousing in high-rise apartment blocks. 306 Appendix E E Various subsidy programs exist, most of them interest rate based. The government has introduced the Affordable Housing Finance loan scheme for low- and middle- income households. It offers loans at 3 percent interest. The housing loan scheme oper- ated for EPF members by state banks provides loans at 4 percent a 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. Various housing loan programs 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 estab- lished when housing markets were underdeveloped, and the need for some of the sub- sidies may 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 envi- ronment. The state subsidizes the debt rather than the housing itself, robbing house- holds 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 enhanced by making the subsidy for low-income housing also available to private lenders 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 people (443,000 from the tsunami disaster alone). For the housing sector in particular, the 2005­07 activity plan for public investment included initiatives on housing for public servants, urban development and low-income housing construc- tion, rural community development and estate infrastructure, and the North East Housing Reconstruction Program. These efforts fall short of meeting low-income housing needs. When the housing needs of the urban poor remain unmet, those people are forced into informal settlements with shelter conditions that violate land development regulations and housing construction codes. In many cities, urbanization, economic growth, and sometimes financial globalization have made land acquisition costs for building affordable housing prohibitive for the landless urban poor, including new migrants from poorer rural areas. Mortgage Affordability Table E.9 shows that 50 percent of the population falls below the average annual income of SL Rs 70,000, and that only the top 10 percent of people make an average annual income above SL Rs 200,000. Using a set of typical mortgage terms, a mortgage with 16 percent annual inter- est for 15 years would require a monthly mortgage payment of SL Rs 14,680 for a housing loan of SL Rs 1 million. In other words, using industry-standard policies, this mortgage requires a monthly income of SL Rs 24,467 (SL Rs 294,000 annually). 307 Appendix E E Table E.9 Average Income by Population Decile, Fiscal 2006/07 Population Decile Median Income (SL Rs) 10th (top) decile 66,500 9th decile 37,940 8th decile 27,429 7th decile 21,774 6th decile 18,038 5th decile 14,987 4th decile 12,463 3rd decile 10,166 2nd decile 7,802 1st decile 4,685 Source: Household Income and Expenditure Survey, 2006/07, Sri Lanka Department of Census and Statistics. This means that a house valued at SL Rs 1 million could be affordable only for the top 10 percent of the Sri Lankan population. Using an industry standard of mort- gage lending up to three times annual income, a low-cost house of SL Rs 175,000 could be affordable to an individual with an annual income of SL Rs 58,000. Accord- ingly, a 350-square-foot house with one bedroom, a living room, a kitchen, and a toilet that, by one estimate, would cost SL Rs 175,000 would not be affordable for nearly 45 percent of the Sri Lankan people. Regulatory Framework and Mortgage Market Infrastructure The potential constraints on the development of a market for home mortgages in Sri Lanka may be divided into three categories. The first category 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 category of constraint 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. Land Registration and Administration A registered title is unambiguous evidence of ownership that cannot be challenged 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 Land Registry Office estimates that it will take at least 15 more years to fully implement the new cadastral and title system. 308 Appendix E E Fifty-five percent of Sri Lanka's 1,500 professional land surveyors and all 2,600 lawyers are committed to cadastral matters.20 The Registration of Title Act of 1998, as amended, provides a sufficient legal framework but has not been implemented. Rather, ownership is still evidenced by deed registration. The record is often ambigu- ous 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 com- parative data are available)--442 per 1,000 population. That number is relatively high, compared with the region. In urban areas, 90 percent of parcels are legally reg- istered and surveyed, 5 percent are legally occupied but not registered or surveyed, and the remaining parcels 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 E.10). Property registration involves an evaluation, payment of the stamp tax, and nota- rization. The registration is then filed in paper form by notary number. Failing to register leaves a deed unrecorded and without court standing in the event of a dis- pute. The 42 branch offices of the registrar provide records and information, but they do not monitor and correct inconsistent records or multiple claims. Inefficien- cies in the registries for land and movable 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 (World Bank 2007a). The lack of computerized files Table E.10 Land Parcels and Registration Status, Urban and Rural Areas, 2003 Registration in Registration in Land Land Parcels Urban Areas Rural Areas Parcels per 1,000 Economy (millions) Population a b c a b c China 246.5 205.4 -- -- -- -- -- -- Hong Kong, China 0.3 44.8 99 0.1 0.9 100 0 0 India 210.0 205.5 99 1.0 0 99 1 0 Indonesia 84.5 361.1 40 50.0 10.0 20 20 60 Korea, Rep. of 35.8 756.5 100 0 0 100 0 0 Malaysia 7.2 288.9 98 0 2.0 99 0 1 Nepal 24.0 1,025.6 95 2.0 3.0 97 2 1 Philippines 50.0 714.3 70 5.0 25.0 75 20 5 Sri Lanka 8.5 441.8 90 5.0 5.0 63 25 12 Source: Cadastral Template, Department of Geomatics, University of Melbourne, Australia, 2007. Note: -- not available. a. Parcels that are legally registered and surveyed. b. Parcels that are legally occupied, but not registered or surveyed. c. Parcels that are informally occupied without legal title (including illegal occupation or squatting). 309 Appendix E E 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 com- parative advantage of the NHDA and UDA when they partner with private develop- ers stems from the state institutions' ability to convey to developers and subsequent buyers clear landownership records that are acceptable to lenders, and their ability to remove all other claims. Land and titling issues are more problematic in rural areas than in urban ones. Foreclosure and Eviction Debt Recovery (Special Provisions) Act No. 2 of 1990 gave power to the licensed commercial banks (power later extended to all companies registered under the Finance Companies Act) to adopt a special procedure to institute legal action to recover sums exceeding SL Rs 150,000. By adopting this procedure, the delay nor- mally 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 licensed commercial banks.21 Parate rights refer to the ability of a lender to foreclose and sell a defaulted property without going to court.22 With parate, the lending institutions should be in a position to recover what is owed them in a better and speedy manner. Parate execution can lead to an ultimate sale within a period of 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 the weak eviction powers of most lenders render parate powers rather moot. Foreclosed properties may be occupied by the defaulting owner for 6­10 years, while the eviction proceeding winds through the courts. Ten- ant protections under the Rent Act are strong. These difficulties account for the low use of parate thus far in practice: only 18 cases regarding parate have appeared before the courts, and 17 were ruled favorably.23 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. Polit- ical and social pressures, 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. 310 Appendix E E 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.24 The proceedings may take three to five years if uncontested; appeals take much longer. Proceedings tend to be costly.25 When 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 unsalable. 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 regarding 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. That said, a sound title system and a non- judicial system of foreclosure and eviction certainly would benefit market develop- ment and could well make the difference for individuals seeking a loan, especially middle-income borrowers. Taxation The government collects stamp tax on house sale transactions, in the amount of 3­5 percent of the transaction price.26 International experience and economic the- ory 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 may be only a third of the actual prices. Various other taxes, levies, and charges apply to housing and housing finance transactions and construction. Housing finance institutions have a corporate profit tax of 20 percent (the standard rate is 30 percent) and a value-added tax of 15 per- cent 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 1,500­2,500. Local government authority charges are levied for street line support, plan approval, UDA, and letter of conformity. Development investors receive Board of Investment tax incentives for construc- tion of 25 units or more, which come in the form of a tax holiday for the first five years, a lower tax of 10 percent for the following two years, a 20 percent tax reduc- tion thereafter, and a customs duty exemption on the import of capital goods. Devel- opers and builders receive tax deductions for mortgage interest payments and capital repayments. Property Valuation Standards All lenders perform in-house loan origination and servicing. 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, low levels of computerization, technology, and staff training require action. These problems appear to be quite manageable, however--especially at private banks. 311 Appendix E E 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), for taxes, and for rental value appraisals of government rental properties. Appraisers are contract appraisers for single-family housing. Structures are valued using implicit or explicit rental values. When replace- ment costs are used, appraisers look at hard construction costs only. 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 have increased by about 28 per- cent, with a growth range of 12­71 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 per- cent during the same period. This increase is comparable to 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 by 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. Of more relevance in Sri Lanka, increasing housing prices cause a surge in demand for mortgage loans and a demand-pulled inflation. The surge in inflation in fiscal 2005/06 was caused, in part, 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 sys- tem. To determine these prices, the first steps would be to ascertain price behavior in the land market and to 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. Prelim- inary work on the construction of housing and land price indexes is under way at the CBSL. In 2006, the central bank conducted a sample survey on factors affecting land prices.27 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 pur- chases for investment and business purposes was about 23 percent, whereas most land purchases were for residential and related uses (77 percent). The survey findings indi- cated that most land transfers were direct purchases from owners, and only approxi- mately 8 percent took place through a broker or from real estate companies. Further- more, 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 312 Appendix E E activities. Consequently, although this survey revealed no evidence of threats to the financial system from land transactions alone, a further survey of all financial institu- tional involvement in the entirety of real estate development activities is required to produce a comprehensive assessment. "Excessive credit" growth is sometimes associated with asset price bubbles. Using credit granted by licensed commercial banks 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 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 loan value being only 60­75 percent of the forced sale value of such proper- ties and existence of a higher haircut28 for the purpose of provisioning mitigate the potential risks to the banks. Furthermore, loan repayments in most cases (partic- ularly for residential houses) depend not on cash flows from such property but on other sources, such as monthly incomes. Nonetheless, constant vigilance and mon- itoring of trends in credit to these sectors is needed to identify potential risks and to take preventive measures. 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. The only credit bureau in the country, its purpose is to collect and provide credit information from all finan- cial institutions engaged in providing credit and registered with the CBSL. Specifi- cally, the bureau collects both positive (regular) and negative (irregular) information on performing loans greater than SL Rs 500,000 and on nonperforming loans (three to four months overdue) greater than SL Rs 100,000. It also gathers data 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 com- panies, 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; on companies, it is detailed. CRIB is owned 40 percent by the CBSL and 60 percent by six state-owned banks.29 CRIB has 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 during 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 such reports have become an essential part of lending institutions' credit appraisal procedure. 313 Appendix E E 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, the bureau must move to a Web-based information system to expand the scope and efficiency of its operation. Measures are being taken to modernize and widen its scope and improve the coverage of CRIB. A modernization project has begun to provide online access to its credit reports. In addition, proposed amend- ments to the CRIB Act will enable the bureau to serve diverse user groups and to pro- vide 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 small and medium-size enterprise lending and microlending, and a wider potential pool of users (such as insurance providers, tele- com providers, utilities, and superannuation funds). Policy Framework and Suggestions for Future Development 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-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. The 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 homes. One third live in semipermanent housing, and 6 percent live in line room estates and shanties. For those people, 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. Difficul- ties 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 house- holds and focusing instead on regular (mostly government) employees. Second, the playing field is not level for private mortgage lenders, so 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; instead, they chase middle-income house- holds because government housing institutions have lost budget funding. 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 314 Appendix E E expected to comfortably handle growing demand at 15 percent or more a year on the value of housing loans30 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 rest of this section provides an overview of suggested improvements. Further liberaliza- tion in the lending environment is needed to transit to a home mortgage market for middle-income households. Housing finance growth is nowhere near the pace necessary to close the housing gap over the next 10-year period. Two serious preconditions for mortgage market growth are the assumption of continued liberalization policies and a broad imple- mentation of the National Development Strategy as planned by the government. Another significant defining factor is lenders' ability to evict and realize collateral. With collateral functioning only as a threat, not as an asset of monetary value, mort- gage 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 lend- ing expansion downward from the top of the income scale depends largely on the overall supply of funds. That brings up another key issue: mortgage loan funding. How much more credit will flow from the economy into housing finance ultimately depends on the claims to domestic financial savings made by the government, state-owned enter- prises, and business. In particular, government debt might crowd out consumer and business lending, including funds for mortgage loans. The banking sector has a lim- ited potential to fuel a "housing finance boom." The funds necessary to close this gap are not likely to be found outside the banking sector in the short or medium term. In the long term, funding can be sourced from provident funds and the insur- ance sector. 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 lend- ing 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 but safe policy environment. New housing production that efficiently uses land will improve housing affordability to middle-income house- holds. Policy should encourage private lending to middle-income households to be integrated into the home mortgage market in time, as banks downscale. Finally, with further governance and efficiency reforms, existing state-owned market par- ticipants could be directed to address clear market failures among the very poor through transparent and well-targeted subsidies (on an equal footing with the pri- vate sector); or they could be stripped of any government guarantee or subsidy (implicit or explicit) and left to pursue market lending, again on a level playing field with the private sector. 315 Appendix E E 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 sup- portive regulatory framework that stimulates its growth. Two main legal prerequi- sites from international experience should be emphasized: The first necessity to develop primary mortgage markets is a functional land registration and titling sys- tem. This work is ongoing in Sri Lanka. The second prerequisite to develop active and efficient mortgage markets, as suggested by international efforts, is data avail- ability. Hence, we have the following recommendations: · Property and house price indexes need to be created, following up on the initial preparatory work going on at the CBSL. Further improvements in the function- ing of CRIB will contribute to protecting consumers and avoiding lengthy court disputes involving finance companies. The Asian Development Bank's financial sector assessment suggests privatizing CRIB. · 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 overreg- ulating anticipated problems, and should focus instead 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 man- ner. Above all, policies should be robust and parsimonious, not complex or costly, in keeping with existing financial institution capacity and the level of sophistica- tion of the current financial infrastructure. Complex regulations that are not commensurate with current housing finance system needs would stifle growth and efficiency gains, and might lead to regulatory arbitrage. How can the government assist in expanding housing finance while avoiding mar- ket distortions? As a priority, the state-owned mortgage banks need a structural reform. Considerable effort is required to further improve governance, management, efficiency, training, 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. State banks (SMIB and the HDFC) should be exposed to general market competition and should oper- ate on an equal footing 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 practices 316 Appendix E E suggest that integrating specialized banks into the general regulatory system could eliminate market segmentation and offer considerable gains to homeown- ers. The Asian Development Bank's financial sector assessment suggests adopting a single banking license rather than the current distinction between commercial and specialized banks. State-owned mortgage banks (SMIB and the 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 institu- tions might merit rethinking as well, especially because 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 the 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 that of lower-income groups. Our specific recommen- dations are these: · Strategic refocusing of SMIB and the HDFC should be considered. The banks could undertake two sets of activities that are not mutually exclusive: (1) fully competitive housing lending without any (explicit or implicit) subsidies or guar- antees on equal footing with the private sector; and (2) transparent, well-targeted subsidies channeled exclusively to the lower-income groups (and unavailable to middle- or upper-income groups), again on equal footing with the private sector. · Boards of directors should be depoliticized to function better. They could be in a stronger position to make management accountable and could enforce the requi- site modern internal controls. · Deeper efficiency improvements should require upgrades and human resources training, as well as modernization and computerization of existing operations. · Improved risk management and credit loss indicators should address the expo- sure to credit and market risks as well (perhaps via reductions in nonperforming loans and a transition to adjustable mortgage rate instruments). For state-owned mortgage banks, there is no clear policy rationale for maintaining their special, historically determined, privileged status--and perhaps their implicit subsidies and government backing. Instead, transparent and well-targeted subsi- dies should be offered to any institution in the sector, thereby fostering competi- tion. 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 programs. Many of the same recommendations apply to rethinking the state role as a player in the housing market. International experience points to a tendency away from using publicly owned construction companies to build low-cost housing; instead, it points to using private companies. Housing is not a public good; its construction does not enjoy economies of scale. The housing con- struction business in most countries is highly competitive. In Sri Lanka, the system 317 Appendix E E has increasingly become nonviable 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 households' access to the mortgage market is currently inhibited because of the inability of banks to lend without collateral, on one hand; and the difficulties in real- izing collateral value, on the other. Solutions to the foreclosure and eviction prob- lems would remove the bias against middle-income lending by retail mortgage lenders; but such solutions have been slow in coming in most transitioning coun- tries, and will likely be so in Sri Lanka. An alternative solution that has been used is mortgage default insurance. Because this insurance tends to be information depend- ent, potential is low for the development of private mortgage default insurance in Sri Lanka; and international experience points to disappointing results with state- owned mortgage default insurance.31 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 and (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--and currently that 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, a 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. 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 pri- mary mortgage markets to attain a sizable scale and strength in the short and medium terms. Centralized liquidity facilities are used to prevent liquidity crises caused by term mismatch. Beyond this function (which can also be performed by the central bank through a discounting facility), liquidity facilities have been used to spur sec- ondary mortgage market development. Cagamas Berhad 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 E.1). Most liquidity facili- ties have some sort of government backing and take on some of the risk of mortgage lending. Liquidity facilities are vulnerable to abuse as a means for lenders to offload 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. 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 macroeconomic and budgetary conditions for 318 Appendix E E Box E.1 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 make housing loans more accessible and affordable, particularly to lower-income groups. It borrows money by issuing debt securities; and it 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 are presented in Chiquier (1999). Source: Cagamas Berhad Web site, http://www.cagamas.com.my. 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. Notes 1. According to the Department of Census and Statistics' 2001 census of housing, members per household were 4.24 in 1981 and 3.70 in 2001. 2. The estimate comes from the Workshop on Housing Finance in South Asia, Jakarta, Indonesia, May 27­29, 2009. 3. See the Apartment Ownership Law Act No. 11 of 1973 and amendments. 4. Data are taken from the 2006/07 survey of the Department of Census and Statistics, available at http://www.statistics.gov.lk/home.asp. 5. The National Housing Act 37 of 1954 provided the basic legal framework for housing development. 6. Current borrower income averages SL Rs 15,000­20,000 a month or more. 7. Authors' rough estimates are calculated by counting the number of mortgage loans disbursed in the economy (between SL Rs 600,000 and SL Rs 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). 8. Until the year 2000, the three state-owned specialized housing finance institutions (namely, SMIB, the HDFC, and the NHDA) collectively controlled 65 percent of the housing finance market. 9. One of the oldest banks in Sri Lanka, the SMIB has its origins as the Ceylon State Mortgage Bank inaugurated in 1931; it was amalgamated with the Agricultural and Industrial Credit Corpora- tion in 1979 to arrive at its present structure. 10. The HDFC, established originally in 1984 as a building society, is 78 percent owned by the NHDA and 12 percent owned by the Employees' Trust Fund (ETF). It was licensed as a special- ized bank in 2003. 11. Limited competition exists, despite the multiplicity of financial institutions in the market, because the market remains dominated (by volume) by a few weak players who dictate the terms. 12. These are the rules applied by the CBSL to borrowing banks at the discount window (that is, bor- rowing directly from the CBSL). 319 Appendix E E 13. The government likely would capture a considerable share of financial system assets regardless of the rate paid because it requires most financial intermediaries to hold Treasury paper. 14. As of 2006, the EPF maintained an estimated 11.3 million member accounts comprising 2.0 mil- lion contributing accounts and 9.1 million noncontributing accounts, and SL Rs 31.7 billion in contributions. The investment portfolio of the fund stood at SL Rs 451 billion (16 percent of GDP) and accounted for more than 70 percent of total superannuation assets. The ETF was created in 1981 to promote stock ownership among employees, and is financed by a 3 percent employer contribution. Its total assets are SL Rs 65.7 billion (2.3 percent of GDP). 15. A February 2005 report by Maxwell Stamp noted that the combined budgets for the EPF and the 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. 16. The balance is mainly invested in listed and unlisted equities. Other assets include bank deposits and bank loans. 17. A number of noncontributory pension plans are operated by nongovernment public and private enterprises such as banks, insurance companies, utilities, and the larger nonfinancial corpora- tions. The providers of these plans are required to set aside reserves to meet their liabilities. Information on the size, types of available benefits, and asset allocation policies of these plans is not readily available. However, it is well known that although they are neither restricted in shap- ing their investment policies nor subject to government direction, these plans also favor invest- ing in fixed-income instruments. 18. The 10-, 15-, and 20-year Treasury bonds were first issued by the CBSL in 2003, for a total value of SL Rs 1.4 billion. 19. A limited-purpose entity serves as a pass-through conduit in creating securities backed by mortgages. 20. Data are taken from the Cadastral Template, Department of Geomatics, University of Melbourne, http://www.geo21.ch/cadastraltemplate/index.htm. 21. State banks such as the Bank of Ceylon, the People's Bank, and the DFCC have been established by separate acts of Parliament. Among other rights, these acts provide the right to "parate pow- ers." Specifically, current beneficiaries of parate powers include the Bank of Ceylon, the People's Bank, the banks established under the provisions of the Regional Rural Development Bank Act No. 15 of 1985, the NSB Credit Corporation, and commercial banks holding a license under the Banks Act No. 30 of 1988. 22. Parate allows the lending 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 man- ager to take charge of and manage the property mortgaged, and to sell the produce and recover rent and other profits. 23. These data are 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. 24. The lender registers his or her action at the Land Registry Office and the court issues notice and summons all concerned parties. Following a trial, judgment can be appealed at the Court of Appeal (as well as the Supreme Court, on questions of law only). The court can issue an order to sell, pending borrower appeal. 25. Costs are specified in Costs (Regulations) 1997 made under Section 214 of the Civil Procedure Code. 26. On gifts, the first SL Rs 50,000 is taxed at 4 percent; thereafter, the tax rate is 3 percent. On trans- fers, the first SL Rs 100,000 is taxed at 5 percent; thereafter, the rate is 4 percent. 27. The survey used direct information from a sample of 427 respondent buyers identified through title deeds of land sales in 2003, 2004, and 2005; deeds were available in a Colombo District Land Registry Office of the Registrar General's Department. 28. "Haircut" relates to the amount by which an advance is below the nominal value of the collateral. 320 Appendix E E 29. The state-owned banks are the Bank of Ceylon, the People's Bank, the National Development Bank, the NSB, the DFCC, and the SMIB. The CBSL is contemplating a reduction in its share. 30. The annual growth in the number of mortgages issued is estimated at 3 percent. 31. This insurance relieves mortgage originators of the risk of default. It helps widen 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 in the development of secondary markets. (See also Villani [2007].) References ADB (Asian Development Bank). 2005. "Sri Lanka: Financial Sector Assessment." October. Manila, Philippines. Calhoun, Charles. 2005. "Housing Finance Assessment for Thailand." December. World Bank, Washington, DC. CBSL (Central Bank of Sri Lanka). 2003. "Consumer Finances and Socio-economic Survey 2003/04." Colombo, Sri Lanka. http://www.cbsl.lk/cbsl/cfs03_04.html. ------. 2006a. "Annual Report." Colombo. ------. 2006b. "Financial System Stability Review 2006." Colombo. http://www.cbsl.lk/ cbsl/FSSR-2006.pdf. ------. 2007. "Road Map for Monetary and Financial Sector Policies in 2007 and Beyond." Colombo. ------. 2008. Annual Report of the Monetary Board to the Hon. Minister of Finance for the Year 2008. Colombo. Chiquier, Loic. 1999. "Secondary Mortgage Facilities: A Case Study of Malaysia's Cagamas Berhad." World Bank, Washington, DC. Kusmiarso, Bambang. 2006. "Housing and Mortgage Markets in the SACEAN Countries." SACEAN Research and Training Centre, Malaysia. Maxwell Stamp. 2005. "Report on EPF/ETF." February. London, UK. Renaud, Bertrand. 2007. "Mortgage Finance in Emerging Markets: Constraints on Feasible Development Paths." Renaud Advisors, McLean, VA. Sarathchandra, C. A. 2008. Housing Development Finance in Sri Lanka. Colombo. Villani, Kevin. 2007. "Mortgage Markets in Emerging Economies: The Use of State-Sponsored Insurance and Guarantees." April. World Bank. 2007a. Doing Business in South Asia 2007. Washington, DC. ------. 2007b. "Housing Finance in Sri Lanka: Opportunities and Challenges." South Asia Private and Financial Sector, Washington, DC. ------. 2009. World Development Indicators 2009. Washington, DC. 321 Index Figures, notes, and tables are indicated by f, n, and t, respectively. A informal housing finance, 27 Affordable Housing Finance, Sri Lanka, 307 land administration, 38, 39, 41 Afghanistan landownership, registration, and titling affordability of housing, 34, 66­67 cadastre, 40, 73 credit information collection, 66 capacity of registration/titling current need for housing units, 9 professionals, 74 current state of housing sector, 13, 51­52 constraints for women, 74 economic overview, 52­55 improving ownership security, 74 foreclosures and evictions land registration, 72­73, 81n9 cost of foreclosures, 44 land titling, 73 legal framework and enforcement, 75­76 right of landownership, 72, 81n8 procedures for creating liens, 75 large-scale project finance, 26 homeowners' insurance, 65 low-income and microfinance housing housing demand advantages and challenges in, 69 housing shortage, 56 affordability of microlending, 71 squatters and informal settlements, 57 demand for housing microfinance, 68 upper-middle-class housing, 67 extending loans to women, 70 housing finance funding mechanisms, 69­70, 80n7 informal housing finance, 64 microfinance capacity status, 34 Islamic mortgage finance, 63­64 microfinance overview, 68 large-scale project finance, 64 target groups, 70 liquidity, 61 mortgage lending, 21 long-term funds, 61­62 policy options potential financing needs for housing, 4 appraisal reliability improvement, 79 specialized mortgage finance, 63 cadastre, 79 state of development, 60­61 capacity building, 79 traditional mortgage lending, 62­63 enforcement of property law, 78 housing supply land titling overhaul, 78­79 construction quality and enforcement of lien registration process standards, 59­60 improvement, 78 costs of housing, 16, 60 microfinance support, 80 current state, 57­59, 80n3 mortgage product innovation, 80 public sector efforts, 58 overview, 76­77 323 Index Afghanistan (continued) RAJUK's role, 89 regulatory framework strengthening, 77 secondary property market, 91 setting and maintaining standards, housing supply, 86­87, 124n2 79­80 institutional framework for land system efficiency improvement, 78 development, 15 technical support for credit Islamic housing finance, 26 information, 80 land titling, 42, 112­13, 124n14 title search improvement, 78­79 large-scale project finance, 26 training in property law, 77­78 low-income and microfinance housing prudential regulation and oversight, 66 government and NGO programs, 108­09 risk management, 65­66 market overview, 105­06 state-owned housing banks, 23 microfinance capacity status, 33 Aga Khan Planning and Building Services, NGO's role, 109­10 Pakistan, 250 products, funding sources, and challenges Ahmedabad Pilot Project, India, 171 in, 106­08 All Pakistan Contractor Association mortgage lending, 21 (APCA), 221 policy framework and state intervention ASA Bangladesh, 107 BHBFC revamping, 120 capacity building, 119 B collateral/lien registry and credit Balochistan High Court, 260, 270 information centralization, 119 Bangladesh contractual savings schemes, 122 current need for housing units, 9, 11 enabling of developers, 123 current state of housing sector, 13, 83­84 housing observatory, 119 economic overview, 84­85 land administration efficiency financial challenges in housing, 5 improvement, 118, 125n19 foreclosures costs, 44, 124n15 microfinance lending strengthening, housing demand, 85­86 120­21 housing finance mortgage product innovation, 121­22 affordability of housing, 34, 91, 92f mortgage refinancing facility, 123 BHBFC's role, 94­97, 124n4 overview, 116­17 growth in, 19 strengthening, 117­18, growth potential of mortgage debt 124­25nn16-17-18 industry, 104­05 private specialized mortgage lenders, 24 homeowners' insurance, 102 property taxation, 40 interest rate benchmarks, 103­04 prudential regulation and oversight, 30 Islamic mortgage finance, 97, 100 regulatory framework market failures, 92 capacity building, 114­15 market overview, 93­94 constraints for women, 115 mortgage lending, 91, 93, 124n3 credit information, 116 nationalized commercial banks' role, 97­98 foreclosures and evictions, 115­16 policies on, 102 land administration inefficiencies, 38, potential financing needs for housing, 4 39­40 prudential regulation and oversight, land registration and administration, 100­02 111­12, 124nn9-10-11-12 risk management, 102­03 land titling, 112­13 specialized mortgage finance, 98­100 property registration, 114 traditional mortgage lending, 97­98 Bangladesh House Building Finance housing stock development/land Corporation (BHBFC), 22 development foreclosures and evictions rights, 116 centralized control issues, 87 lack of low-income subsidies, 83 cost of construction materials, 90­91 need for revamping, 120 costs of residential construction, 87­88 prudential regulation and oversight, 101 lack of construction financing, 89­90 role in country's housing sector, 94­97 private party development, 89 shortcomings, 23 process for residential infrastructure Bank-e-Millie Afghan, 21, 62 development, 88­89 Basel Committee on Banking, 31 324 Index Basic Services for the Urban Poor, H India, 172 Haryanan Urban Development Authority, 39 benami (absentee) ownership, 41, 277n20 hawala system, 27, 64 Bharat Nirman Program, India, 175 HDFC (Housing Development Finance Corporation), India, 209 C HDFC (Housing Development Finance cadastral maps and land titling, 38­39, 40, Corporation), Sri Lanka, 280, 293, 73, 79 319n10 Cagamas Berhad, 29, 278n29, 319 Herfindahl Concentration Index, 232 capital note, 61 Heritage Homes, Pakistan, 251 Central Bank of Sri Lanka (CBSL), 280, homeowners' insurance, 27 298­300, 319n12 Afghanistan, 65 chhappar, 106, 124n6 Bangladesh, 102 Colombia, 206n25 Pakistan, 245 contractual savings schemes for housing House Building Finance Corporation (CSHs), 122 (HBFC), 209 Credit Information Bureau (CRIB), housing Sri Lanka, 313­14 causes of shortages, 9 Credit Rating Information Service of India, country GDP and, 1­2, 3f Limited (CRISIL), 137 economic impact on a country, 1­2 CSHs (contractual savings schemes for finance markets growth potential, 3­5 housing), 122 financial sector and regulatory framework (see housing finance) D key challenges, 5­6 Da Afghanistan Bank (DAB), 61 living conditions in slums, 9­10 Debt Recovery Tribunals (DRTs), 186 objectives of report, 6 Delta Brac Housing Finance Corporation policy options recommendations Limited, Bangladesh, 91, 98, 99, accommodating low-income groups 107, 124n5 through product innovation, 49­50 Dhaka, 11 finance information and transparency, 48 land administration efficiency improve- F ment, 47 foreclosures and evictions microfinancing, 50 Afghanistan, 44, 75­76 mortgage refinancing provisions, 48­49 Bangladesh, 44, 115­16, 124n15 scope of the report, 6 India statistics on shortages, 9 bankruptcy laws, 186 Housing and Urban Development Corporation case management process (HUDCO), India, 143 improvements, 190 Housing Bank of Thailand, 296­98 contract enforcement costs, 189 Housing Development Finance Corporation cost of foreclosures, 45 (HDFC), India, 143 resolution of debt-recovery claims, Housing Development Finance Corporation 186­87 (HDFC), Sri Lanka, 280, 293, 319n10 Securitization and Reconstruction Act housing finance and, 186 capital market development, 30 time to resolve commercial disputes, challenges facing, 19 187­89 critical role of, 2 Pakistan, 44­45, 270, 278n22 global finance crisis and, 32 regulatory framework for housing finance, homeowners' insurance, 27 43­45, 260­61 industry growth rates, 20f Sri Lanka, 44, 310­11, 320nn21-25 informal, 26­27 Islamic mortgage finance, 25­26 G large-scale project finance, 26 Golden Jubilee Rural Housing Finance Scheme, low-income, 33­35 India, 172 private specialized mortgage lenders, Grameen Bank, Bangladesh, 106­08 24­25 Grihayan Tahabil, 108 prudential regulation and oversight, 30­32 325 Index housing finance (continued) housing finance regulatory framework financial sector, 142­43 capacity of registration/titling growth in, 19 professionals, 42­43 homeowners' insurance, 154 credit information, 45­46 informal, 153­54 foreclosures, 43­45 large-scale project finance, 150­53, land administration, 37­42 204n10 risk management, 27­29 mortgage debt industry growth secondary market for mortgage finance, 29­30 potential, 158 state-owned housing banks, 22­24 National Housing Bank, 148­50 traditional mortgage lending, 20­22 overview, 141 See also specific countries policies, 159­60, 205nn11-12 housing supply prudential regulation and oversight, construction quality and enforcement of 155­58 standards, 16 risk management, 154­55 costs of housing, 16­17 securitization in, 150 current state of housing, 13­14 specialized mortgage finance, 147­48 housing price index, 17 traditional mortgage lending, 143­47 institutional framework for land for women, 170, 172 development, 14­16 housing supply secondary property markets, 17 costs of housing, 136­37 See also specific countries current state, 135­36 HUDCO (Housing and Urban Development developer profiles, 139, 141 Corporation), India, 143 housing price index, 137­38, 204n6 housing stock development, 138­39, 140, I 204n7 IDLC Finance Limited, 98 institutional framework for land India development, 15 affordability of housing, 34, 163­65 land administration improvement costs of housing, 16, 17 options credit information, 190­91 computerization of textual records, current state of housing sector, 13, 127­29 193­94 economic overview efficiency focus, 191­92 financial sector reforms' impact, 130­31 land markets restrictions elimination, 194 growth rate, 130 stamp-duty rates reduction, 193 housing and economic growth, 132­33 subregistry and transfer office efficiency, population growth rate and urbanization, 192­93 131­32 land administration inefficiencies, 38­39, 40 poverty levels, 129­30 landownership, registration, and titling financial sector development, 20 basic structure, 176 foreclosures and evictions corruption, 178 bankruptcy laws, 186 department of revenue, 177 case management process Department of Survey and Land Records, improvements, 190 178­80 contract enforcement costs, 189 inefficiencies in the system, 176 foreclosures costs, 45 Land Revenue Department, 177­78 resolution of debt-recovery claims, 186­87 Office of Stamps and Registration, Securitization and Reconstruction Act 180­85 and, 186 spatial record of land ownership, 178­80 time to resolve commercial disputes, title insurance, 185­86 187­89 large-scale project finance, 26, 150­53, government housing policies and programs, 205n10 172­75 low-cost market solutions, 34­35 housing and GDP, 1­2 low-income and microfinance housing housing demand current housing situation, 160 current need for housing units, 9, 11 employer-facilitated model, 199­200 shortages, 134 informal sector models, 200­01 326 Index low-cost housing market, 160­62 Pakistan, 26, 234­36, 277n20 pilot projects, 170, 171 region-wide, 25­26 slum rehabilitation potential, 162 for very low-income households, 201­02 J for women, 170, 172 Jawaharlal Nehru National Urban Renewal market solutions for access to housing Mission, India, 182 cluster approach, 169 financial institutions and credit risk, 168 K government support on policy issues katchi abadis, 215­16 and, 170 khas land, 111, 124n9 for low-income salaried customers, 168 khata register, 177 process for developing the low-income Khuda-ki-Basti-4, Pakistan, 251, 277n12 housing market, 169 Khuda-ki-Basti Incremental Housing Project, for very low-income workers, 168­69 Pakistan, 251 mid/lower-income group housing Korangi Landhi Township, Pakistan, 251 demand and affordability, 162­65, 205nn13-14 L developers' interest in loans, 166­67 land registration and administration employers' role in facilitating housing, Afghanistan, 72­73, 81n9 167, 205n17 Bangladesh, 111­12, 124nn9-12 financial institutions interest in loans, India 165­66, 205nn15-16 basic structure, 176 mortgage finance market, 167, 205n18 computerization of textual records, savings-linked loan product, 165 193­94 mortgage lending, 21­22 corruption, 178 policy options department of revenue, 177 collateral/lien registry and credit Department of Survey and Land Records, information centralization, 194­96 178­80 land administration, 191­94 efficiency focus, 191­92 long-term funding, 202­03 inefficiencies in the system, 38­39, low-income and microfinance 40, 176 housing, 199­202 land markets restrictions elimination, 194 mortgage product innovation options, Land Revenue Department, 177­78 198­99 Office of Stamps and Registration, NHB supervisory role strengthening, 180­85 196­97 spatial record of land ownership, 178­80 property and housing data unification, 196 stamp-duty rates reduction, 193 rental markets, 203­04 subregistry and transfer office efficiency, risk management priorities, 197, 206n24 192­93 private specialized mortgage lenders, title insurance, 185­86 24­25 Pakistan property taxation, 40­41 basic structure, 258­59, 277nn19-20 prudential regulation and oversight, 30, 31 information system creation, 259 savings-linked loan product, 165 present system and costs, 254, 257, securitization in, 23 277nn14-15 squatters and informal settlements, 134­35 property taxation, 257­58, 277nn16-17-18 state-owned housing banks, 23 titling system description and shortcom- Infrastructure and Housing Finance ings, 255­56 Department, 242 regulatory framework for housing finance insurance, homeowners'. See homeowners' inefficiencies in the process, 37­39 insurance land offices and cadastral records, 38­39 Interest Subsidy Scheme for Housing the land titling, 42 Urban Poor, India, 172 problems due to multiple legal regimes, Islami Bank Bangladesh, 106 41­42 Islamic mortgage finance property taxation, 40­41 Afghanistan, 63­64 Sri Lanka, 308­10, 320n20 Bangladesh, 26, 97, 100, 106 titling systems (see land titling) 327 Index land registration information system (LRIS), products, funding sources, and challenges 259, 264 in, 106­08 land titling challenges faced, 33­34 Afghanistan India cadastre, 40, 73 current housing situation, 160 capacity of registration/titling profession- employer-facilitated model, 199­200 als, 74 informal sector models, 200­01, 206n26 improving ownership security, 74 low-cost housing market, 160­62 policy options, 78­79 pilot projects, 170, 171 Bangladesh, 42, 112­13, 124n14 slum rehabilitation potential, 162 capacity of registration/titling professionals, for very low-income households, 301­02 42­43 for women, 170, 172 India microfinance institutions, 33 absence of performance standards, 182 Pakistan basic structure, 176 challenges faced, 248­49 challenges posed by lack of cross-checks, government programs, 250­54 184­85 informal housing finance, 249 corruption, 178 microfinance capacity status, 34 deeds and, 182 NGO programs, 250 factors reducing effectiveness, 181 Sri Lanka inefficiencies in the system, 176 government's role in, 306­07 key functions, 180­81 microfinance and housing, 305­06 lack of coordination between offices, 180 microfinance sector, 34 rates for stamp duties, 181­82 mortgage affordability, 307­08 spatial record of land ownership, 178­80 segmentation in the mortgage market, variations in the cost of registering prop- 304­05 erty, 182­83 Pakistan, 255­56 M registration and administration (see land Malaysia registration and administration) housing finance reform, 29 regulatory framework for housing National Mortgage Company, 319 finance, 42 microfinance. See low-income and microfi- Sri Lanka, 308­10, 320n20 nance housing Land Titling and Economic Restructuring in Microfinance Investment Support Facility for Afghanistan Project, 74 Afghanistan (MISFA), 67 Law of Mortgage on Immovable Property, Monitor Group, India, 34­35, 139, 204n7 Afghanistan, 75 mortgage financing liquidity risk, 27­28 Afghanistan low-income and microfinance housing Islamic, 63­64 Afghanistan mortgage lending, 21 advantages and challenges in housing mortgage product innovation, 80 microfinance, 69 specialized, 63 affordability of microlending, 71 traditional, 62­63 demand for housing microfinance, 68 Bangladesh extending loans to women, 70 growth potential of mortgage debt indus- funding mechanisms, 69­70, 80n7 try, 104­05 microfinance capacity status, 34 Islamic, 97, 100 microfinance overview, 68 mortgage lending, 21, 91, 93 microfinance support, 80 mortgage product innovation, 121­22 target groups, 70 mortgage refinancing facility, 123 Bangladesh private specialized lenders, 24 government and NGO programs, 108­09 specialized, 98­100 market overview, 105­06 traditional, 97­98 microfinance capacity status, 33 India microfinance lending strengthening, mid/lower-income group housing and, 167 120­21 mortgage debt industry growth NGO's role, 109­10 potential, 158 328 Index mortgage lending, 21­22 National Housing Bank (NHB), India, 127­29, mortgage product innovation options, 148­50, 172, 175 198­99 National Housing Development Authority private specialized lenders, 24­25 (NHDA), Sri Lanka, 281, 288­89, 306, specialized, 147­48 319n10 traditional, 143­47 National Housing Finance and Investments Islamic Limited (NHFIL), Bangladesh, 98, Afghanistan, 63­64 99­100 Bangladesh, 26, 97, 100 nationalized commercial banks (NCBs), 93 Pakistan, 26, 234­36 National Savings Bank (NSB), Sri Lanka, 280, region-wide, 25­26 291, 299 Pakistan National Savings System (NSS), Bangladesh, growth potential of mortgage debt indus- 104 try, 246­47 National Urban Housing and Habitat Policy industry efficiency and performance, 2007, India, 152 228­29 NCBs (nationalized commercial banks), 93 mortgage lending, 22 NHA (National Housing Authority), secondary mortgage facility creation, Bangladesh, 108 272­73 NHB (National Housing Bank), India, 127­29, specialized, 233­34 148­50, 172, 175 policy options recommendations, 48­49 NHDA (National Housing Development private specialized mortgage lenders, 24­25 Authority), Sri Lanka, 281, 288­89, 306, risks in the market, 28­29 319n10 secondary market for, 29­30 NHFIL (National Housing Finance and Sri Lanka Investments Limited), 98, 99­100 capital markets as a source, 303, 320n18 NSB (National Savings Bank), Sri Lanka, 280, employee provident funds and, 302, 303, 291, 299 320n17, 320nn14-15 NSS (National Savings System), expansion from middle-income Bangladesh, 104 demand, 302 facilitating growth in primary mortgage O markets, 318­19 Office of Stamps and Registration, India foreign investment as a source, 303­04 absence of performance standards, 182 funding sources for housing finance, challenges posed by lack of cross-checks, 300­01 184­85 life insurance funds and, 302­03, 320n16 deeds and, 182 lower-end financing challenges, 301 factors reducing effectiveness, 181 mortgage affordability, 307­08 key functions, 180­81 mortgage lending, 22 lack of coordination between offices, 180 mortgage securitization and, 304, 320n19 rates for stamp duties, 181­82 segmentation in the mortgage market, variations in the cost of registering property, 304­05 182­83 state-owned mortgage banks, 317­18, 320n31 P total system assets, 300, 320n13 Pakistan traditional, 294­96 affordability of housing, 34 upper end financing potential, 301 costs of housing, 16 traditional mortgage lending, 20­22 current state of housing sector, 14, 209­11, Mumbai, India, 16 216­17 Mumbai Pilot Project, India, 171 economic overview, 211­13 murabaha loans, 64 foreclosures and evictions, 44­45, 260­61 musharakah, 25­26, 236 government housing policies and programs Estate Office, 253­54 N facilitator role, 250, 252 National Development Trust Fund, Federal Government Employees Housing Sri Lanka, 305 Foundation, 253 National Housing Authority (NHA), low-income housing initiatives, 251 Bangladesh, 108 Ministry of Housing and Works, 252 329 Index Pakistan (continued) low-income and microfinance housing National Housing Authority, 252 challenges faced, 248­49 Pakistan Housing Authority, 253 government programs, 250­54 Pakistan Public Works Department, 253 informal housing finance, 249 housing and GDP, 2 microfinance capacity status, 34 housing demand, 9, 11, 214­15, 277n2 NGO programs, 250 housing finance mortgage lending, 22 current challenges, 223 policy options entry of commercial banks, 223 challenges facing housing finance pro- growth in, 19 grams, 274, 278n30 growth potential of mortgage debt foreclosure laws application, 270 industry, 246­47 home construction industry and financ- homeowners' insurance, 245 ing development, 268 potential financing needs for housing observatory creation, 270­71 housing, 4­5 key recommendations, 262 risk management, 245­46 land development infrastructure provi- housing finance market sion, 267 commercial bank lending, 232­33 low-cost technologies promotion, 268 data on loans and borrowers, 223­25 municipal bonds encouragement, 271­72 HBFC, 229­30, 231f new housing and finance products devel- industry efficiency and performance, opment, 273­75 228­29 property registration system, 263­64 interest rates, 225­26 property transfer process, 265 Islamic housing finance, 26, 234­36, prudential environment 277n20 appropriateness, 270 loan-to-value ratios, 226­28 real estate agencies organization, 269 prudential regulation and oversight, registration and information systems 236­39 integration, 264­65 sector products, 225 REITs introduction, 271 specialized, 233­34 secondary mortgage facility creation, housing finance market regulatory policies 272­73 funding shift from government to bank- securitization, 274, 278n30 ing system, 242 solutions for low-income groups, 274­75 Housing Advisory Group recommenda- tax rationalization program implementa- tions, 243­44t tion, 266 Housing Advisory Group scope, 242, 245 training and capacity building, 275­76 measures and initiatives, 239, 242 urban rent laws rewriting, 265­66 policy action plan, 240­41t property taxation, 40 support for market-based credit, 242 prudential regulation and oversight, 30 housing stock development regulatory framework construction costs, 222 capacity of real estate professionals, construction quality and enforcement of 259­60 standards, 222 challenges for, 254 developer associations, 221 constraints for women, 260 developer financing, 221 credit information, 261­62 developer industry, 220­21 foreclosures and evictions, 260­61, land procurement, 217­19 278n22 low-cost construction technologies, regulatory framework, land registration 219­20 basic structure, 258­59, 277nn19-20 rental regulations, 221­22 information system creation, 259 secondary property markets, 222­23 present system and costs, 254, 257, housing supply, 216­17 277nn14-15 impact of funding disadvantages, 25 property taxation, 257­58, institutional framework for land 277nn16-17-18 development, 15 titling system description and shortcom- land administration, 38, 39, 40, 41 ings, 255­56 large-scale project finance, 26 state-owned housing banks, 23­24 330 Index parate rights, 310, 320nn22-25 Sri Lanka Pashtany Bank, 21, 62 affordability of housing, 34 patwaris, 258, 277n19 costs of housing, 16, 283 Poshika, 107 current need for housing units, 9, 11 private specialized mortgage lenders, 24­25 current state of housing sector, 14, 279­80 prudential regulation and oversight, 30­32 economic overview, 280­82 Pune Pilot Project, India, 171 foreclosures costs, 44 growth scenario, 314­15, 320n30 R housing demand, 282­84, 319nn1-2 RAJUK (Rajdhani Unnayan Kartripakkha), housing finance 87, 89 effective demand vs. housing needs, 289, RBI (Reserve Bank of India), 31, 155­58 319n7 Real Estate and Housing Association of efficiency and performance of institu- Bangladesh, 90 tions, 296­98, 319n11 Record of Rights (ROR), India, 177 financial institutions, 291, 292t registration, land. See land registration and funding for rehabilitation, 289 administration growth in, 19 regulatory framework for housing finance market, 290­91, 293­94, 319nn8-9 capacity of registration/titling professionals, risk management, 298­300, 319n12 42­43 traditional mortgage lending, 294­96 credit information, 45­46 housing supply foreclosures, 43­45 construction quality, 286­87 land administration current state of housing, 284­85 inefficiencies in the process, 37­39 housing stock development, 285­86 land offices and cadastral records, 38­39 policy environment for housing finance, land titling, 42 287­89, 319nn5-6 problems due to multiple legal regimes, institutional framework for land 41­42 development, 15 property taxation, 40­41 land administration inefficiencies, 38, 40 REITs, 260, 268, 271 low-income and microfinance housing Reserve Bank of India (RBI), 31, 155­58 government's role in, 306­07 RESIDEX, 137­38, 204n6 microfinance and housing, 305­06 risk management, 27­29 microfinance capacity status, 34 Rural Housing Fund, India, 175 mortgage affordability, 307­08 segmentation in the mortgage market, S 304­05 Saiban, Pakistan, 250 market failures, 283­84 Sanasa Development Bank, Sri Lanka, 305 mortgage debt industry Sarvodaya Enterprise Development Services capital markets as a source, 303, 320n18 (SEEDS), 305 employee provident funds and, 302, 303, SBP (State Bank of Pakistan), 209 320nn14-17 secondary property markets, 17, 29­30 expansion from middle-income Secured Transaction of Movable Property demand, 302 Law, 75 foreign investment as a source, 303­04 Securitization and Reconstruction Act, India, 185 funding sources for housing finance, SEEDS (Sarvodaya Enterprise Development 300­01 Services), 305 life insurance funds and, 302­03, 320n16 Self-Employed Women's Association (SEWA), lower end financing challenges, 301 India, 170, 172 mortgage securitization and, 304, 320n19 Shadab Zafar Construction Company, 67 total system assets, 300, 320n13 Shelter for Life, 59 upper end financing potential, 301 SMIB (State Mortgage and Investment Bank), mortgage lending, 22 Sri Lanka, 280, 293, 299, 319n9 property taxation, 40 squatters and informal settlements recommendations for housing finance Afghanistan, 57 system India, 134­35 facilitating growth in primary mortgage Pakistan, 215­16 markets, 318­19 331 Index Sri Lanka (continued) State Bank of Pakistan (SBP), 209 government's role, 316­17 State Mortgage and Investment Bank (SMIB), regulatory framework, 316 Sri Lanka, 280, 293, 299, 319n9 state-owned mortgage banks, 317­18, state-owned housing banks, 22­24 320n31 Sujrani Town, Pakistan, 251 regulatory framework sukuk, 246, 277n11 credit information, 313­14, 320n29 foreclosures and evictions, 310­11, 320nn21-25 T Taiser Town, Pakistan, 251 land and housing price movements, titling systems. See land titling 312­13, 320nn27-28 land registration and administration, 308­10, 320n20 U property valuation standards, 311­12 Urban Development Authority (UDA), taxation, 311, 320n26 Sri Lanka, 289 state-owned housing banks, 24 urban migration, 4 332 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The · 6 trees Office of the Publisher has chosen to print · 2 million BTUs of Expanding Housing Finance to the Underserved total energy in South Asia on recycled paper with 30 percent · 576 lb. of net green- postconsumer fiber in accordance with the rec- house gases ommended standards for paper usage set by the · 2,772 gal. of waste Green Press Initiative, a nonprofit program sup- water porting publishers in using fiber that is not · 168 lbs. of solid sourced from endangered forests. For more infor- waste mation, visit www.greenpressinitiative.org. A vibrant housing finance market is an important pillar of economic prosperity. A strong institutional framework is indispensable for a functioning housing finance market. Home affordability is a fundamental challenge in South Asia and can only be solved by a comprehensive approach involving developers, the traditional financial sector, microfinance, and other innovative approaches experienced in recent years. Abdul Fitrat Governor, Da Afghanistan Bank The government's policies for the housing sector, particularly in the South Asian countries, must focus on "inclusive housing" coupled with "stability of the market." The two objectives are mutually complementary, reinforcing each other. Stability inspires the stakeholders to enhance their participation and enables greater institutional and financial deepening of the sector, both of which are necessary conditions for inclusive housing. R.V. Verma Managing Director, National Housing Bank, India The global financial crisis demonstrated the importance of a strong, disciplined, and honest mortgage market. As we learn our lessons, it is critically important that those who have been closed out of the housing sector do not get closed out again. In South Asian and Pacific countries, we hope that a balance can be found between discipline and access. Ira Peppercorn Former Deputy Federal Housing Commissioner, United States Housing markets in South Asia are undergoing a dynamic intermediate stage of their development, ranging from an incipient institutional framework in Afghanistan to a dynamic developing housing finance system in India. International experiences--including such examples as Malaysia, Mexico, the Republic of Korea, Singapore, and Thailand--provide many useful lessons and examples in building institutions, enabling markets, and expanding housing finance outreach to the middle- and lower-income groups. Narayanasamy Kokularupan Former CEO of Cagamas Berhad, Malaysia ISBN 978-0-8213-8322-3 SKU 18322