Cost and benefits of rent control in Kumasi, Ghana

t?f5f7- 3 The World Bank Policy, Planning and Research Staff Infrastructure and Urban Development Department Report INU 51 COST AND BENEFITS OF RENT CONTROL IN KUMASI, GHANA by Stephen Malpezzi A. Graham Tipple Kenneth G. Willis October 1989 DISCUSSION PAPER This Is a document published informally by the World Bank. The views and interpretations herein are those of the author and should not be attributed to the World Bank, to its affiliated organizations, or to any individual acting on their behalf.  Copyright 1989 0 The WorLd Bank 1818 H Street, N.W. ALL Rights Reserved First Printing October 1989 This is a document pubLished informatly by the world Bank. In order that the information contained in it can be presented with the Least possibte delay, the typescript has not been prepared in accordance with the procedures appropriate to format printed texts, and the World Bank accepts no responsibitity for errors. The World Bank does not accept responsibitity for the views expressed herein, which are those of the author and should not be attributed to the Wortd Bank or to its affilieted organizations. The findings, interpretations, and concLusions are the results of research supported by the Bank; they do not necessariLy represent officiaL policy of the Bank. The designations enmployed, the presentation of material, and any maps used in this document are solety for the convenience of the reader and do not impty the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the Legal status of any country, territory, city, area, or of its authorities, or concerning the delimitations of its boundaries or national affitiation. Stephen Matpezzi is an economist in the Urban Development Division, PoLicy, PLanning and Research, The World Bank. Graham Tipple is Senior Research Officer at the Centre for Architectural Research and Devetopment Overseas, and Ken Willis is Senior Lecturer in the Economics of Town and Country Planning in the Department of Town and Country Planning, both at the University of Newcastle upon Tyne, United Kingdom. This paper has been produced under a Wortd Bank research project on Rent Control in Devetoping Countries (RPO 674-01), directed by Stephen MaLpezzi. The project and the methodotogy are described in the fotlowing papers: Stephen MaLpezzi and C. Peter Rydell, Rent Controts: A Framework for Analysis (Water Supply and Urban Development Department Discussion Paper No. 102, 1986). S.ep,en Malpezzi, Stephen K. Mayo, Ricardo Silveira and Carmeta Quintos, Measuring the Costs and Benefits of Rent Control: Case Study Design. (Infrastructure and Urban Development Department Discussion Paper No. INU 24, 1988). An introductory descriptive paper on Rent Control in Kumasi has aLso been produced: A. Graham TippLe, The History and Practice of Rent Controt in Kumasi. Ghana (WorLd Bank, Urban DeveLopment Division Working Paper No. 88-1, 1988). The authors are grateful to Esra Bennathan, Robert Buckley, Per Liung, Edgar Olsen, Stephen Mayo and others for their comments on the drafts of this paper. Many Ghanaians have also assisted with this research. We are particuLarLy gratefuL to Samuel Afrane, SamueL Boapeah, and Martha Tamakloe. They are not responsibte for remaining shortcomings.  The World Bank COSTS AND BENEFITS OF RENT CONTROL IN KUMASI, GHANA DISCUSSION PAPER  ABSTRACT Over the past forty years, rent control has been a feature of housing in Ghana. This study focusses on the housing market in Kumasi, the second largest city in Ghana. We examine the characteristics of the rent control regime in force there, and assess the costs and benefits of rent control, on landlords and on tenants, and its effects on the housing stock. Rent control has been successful in ensuring that housing is very inexpensive for most households, both in absolute terms and in the proportion of income devoted to rent. LanJlords, who are found to be little different from many renters, have been deprived of economic returns from their property. Their reaction has been to withdraw stock from renting to use for their own family members and to reduce maintenance. There has recently been an increase in payment of rent in advance for new lets, and even some existing lets. Rent control is not the only constraint on the housing market, in Kumasi or in Ghana. The paper also describes other supply side and regulatory constraints; including those affecting land, finance, and choice of building design and materials. Where net benefits from rent control exist they are found to be unrelated to need, having little distributional efficiency. There is some tendency for lower income tenants to receive larger net benefits, but they are still small relative to their cost. The strongest pattern is that long term tenants receive large benefits at the expense of recent movers as well as landlords. Our best estimates are that typical room rents in the tenement and indigenous housing sectors were 250 to 300 cedis in 1986; that long run equilibrium rents of these *-nits in the absence of controls would be roughly twice that (median of 574 cedis); but that households would demand more housing services in a well fur.ctioning market and hence live in a unit yielding about 80 percent more "housing services" (i.e., with a median rent of roughly 1040 cedis). In other words most tenants pay less for their units (a gain to them) but live in smaller and/or lower quality units than we estimate they would consume in a well functioning market (a loss to them). Not all of the loss from disequilibrium in consumption is necessarily due to rent control, however; constraints on land and finance, and other regulations, play a part. Overall, however, for many tenants the welfare loss from consuming less housing more or less cancels most of the benefit of lower rents. A number of options for relaxation/decontrol are studied with the aid of a simple present value model. Along with decontrol of new construction it is recommended that floating up and out of controls over a period of about five years should be considered, along with policy changes to ensure ready supplies of land, finance, and building materials. Such policies are essential, given that private housing investment has provided and will continue to provide the great majority of rooms in Ghanaian urban areas. Finally, while we emphasized "floating up and out" for existing units along with immediate decontrol for new construction and upgraded units, there are a wide range of options which can be explored in Lore detail with the aid of the present value model. Building a political consensus behind decontrol is not independent of but is more important than the technical means chosen for decontrol or relaxation. v  TABLE OF CONTENTS I. OVERVIEW . . . . . . . . . . . . . . . . . . . . 1 A. Housing and the Current Macro-Economic Environment . . . . . . . 1 B. Summary of Findings ... . . . . . . . . . . . . . . . . .. 4 Key Findings & their Generality . . . . . . . . . . . . . . . 4 Housing Market Conditions ... . . . . .. 6 Costs and Benefits of Controls . . . . . . . . . . . . . . . 7 Distributional Issues . . . . . . . . . . . . . . . . . . . . 8 Hous-:ng Supply . . . . . . . . . . . . . . . . . . . . . . 9 Rent Control and Profitability . . . . . . . . . . . . . . 9 Other Market Imperfections . . . . . . . . . . . . . . . . . 10 Decontrol Options . . . . . . . . . . . . . . . . . . . . . . 11 C. Suggestions for future work ..... . . . . . . . . . . .. 14 II. HOUSING MARKETS AND RENT CONTROL IN GHANA ..17 A. Introduction to Urban Housing in Ghana . . . . . . . . . . . . 17 Typical Designs .17 B. Overview of Ghanaian Housing Markets . . . . . . . . . . . . . 17 Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Land . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Infrastructure . . . . . . . . . . . . . . . . . . . . . . . 20 Finance . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Current Housing Finance Institutions . . . . . . . . . . . . 21 The Bank for Housing and Construction . . . . . . . . . 21 First Ghana Building Society . . . . . . . . . . . . . 22 Social Security and National Insurance Trust . . . . . 22 Other Financial Intermediaries . . . . . . . . . . . . 23 Informal Sector Housing Finance . . . . . . . . . . . . 23 The Regulatory Framework . . . . . . . . . . . . . . . . . . 24 C. Chronology of Rent Control in Ghana .24 Rent Control before 1963 .24 Rent Control between 1963 and 1979 .26 Rent Control post-1979 .29 Recent developments, 1986 to 1988 .32 D. Current Status of Controls . . . . . . . . . . . . . . . . . . 33 Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Advances and side payments ... . . . . . . . . . . . . . . 33 Eviction .... . . . . . . . . . . . . . . . . . . . . . . 35 Treatment of houses in government estates . . . . . . . . . . 35 Reletting After Renovations ... . . . . . . . . . . . . . . 37 E. The Political Economy of P.ent Control . . . . . . . . . . . . . 38 Cost of Housing built by the Government . . . . . . . . . . . 39 Rent Changes.. ............... 39 F. Summary. . ............... 40 vii  III. THE HOUSING MARKET OF KUMASI . . . . . . . . . . . . . . . . . . . 41 A. An Introduction to Kumasi . . . . . . . . . . . . . . . . . . . 41 B. Household Survey Data ...... .. .. .. .. .. . .. . . 43 C. Kumasi's Housing Stock . . . . . . . . . . . . . . . . . . . . 43 Access to Utilities and Services . . . . . . . . . . . . . . 45 Environmentel Conditions ........ .. ... .. .. . 46 Housing Supply . . . . . . . . . . . . . . . . . . . . . . . 46 D. Household Characteristics . . . . . . . . . . . . . . . . . . . 47 Tenure. . . . . . . . . . . . . . . . . . . . . . . . . . 47 House Ownership and "Family Houses". . . . . . . . . . . 47 Income and Cornsumption . . . . . . . . . . . . . . . . . . . 48 Household Size . . . . . . . . . . . . . . . . . . . . . . 50 Household size by tenure ........ .. ... .. .. . 52 Rooms occupied . . . . . . . . . . . . . . . . . . . . . . 52 Occupancy rates ...................... . 53 Rents by Consumption Quartile . . . . . . . . . . . . . . . . 53 Summary of Key Data by Tenure . . . . . . . . . . . . . . . . 54 IV. ECONOMIC ANALYSIS OF RENT CONTROL: THEORY . . . . . . . . . . . . . 57 A. Introduction . . . . . . . . . . . . . . . . . . . . . . . 57 B. Analysis of Rent Control as a Tax on Hou. . . . . . . . . . 57 Rent Control as Expenditure Control . . . . . . . . . . . . . 58 C. Dynamic Models of the Effect of Rent Control on Maintenance . 59 D. The Relationship Between Controlled and Uncontrolled Markets 59 E. Two Approaches to Measuring Costs and Benefits . . . . . . . . 60 Cost-Benefit Using a Cross-Country Model . . . . . . . . . . 60 Cost-Benefit Using a Model Similar to Edgar Olsen's . . . . . 61 V. ESTIMATES OF THE COSTS AND BENEFITS OF CONTROLS . . . . . . . . . . . 65 A. Introduction .......... ... ... .... ... . . 65 B. Choice of Reference Group . . . . . . . . . . . . . . . . . . . 66 Controlled and "Uncontrolled" Households Compared . . . . . . 70 Selectivity Bias ................. .... . 71 Net versus Gross Rent ........... ... .... . . 72 Anecdotal Recent Evidence on Increase in Advances . . . . . 73 C. PcOc: Current Rent Paid . . . . . . . . . . . . . . . . . . . 74 D. Estimating PmOm With a Cross Country Model of Housing Demand 75 "Uncontrolled" Rents Compared to Predictions from the Cross- Country Model . . . . . . . . . . . . . . . . . . . . . 81 E. Estimating PmQm With Kumasi Survey Data . . . . . . . . . . . . 82 F. Price Elasticity of Demand ..... . . .. . . . .. . . . . 83 G. Estimating PmOc With Hedonic Indexes . . . . . . . . . . . . . 83 Hedonic Estimates ................. .... . 84 H. Cost-Benefit Measures .85 Costs and Benefits from the Cross-Country Model . . . . . . . 85 Costs and Benefits Constructed from the Hedonic and Demand Equations .87 I. Distribution of Cos:s and Benefits . . . . . . . . . . . . . . 90 Distribution By Consumption .90 Benefits By Length of Tenure .91 J. Landlord and Tenant Incomes Compared . . . . . . . . . . . . . 92 viii  VI. RENT CONTROL, PROFITABILITY, AND SUPPLY . . . . . . . . . . . . . . 95 A. Housing Supply in Kumasi ... . . . . . . . . . . . . . . . . 95 New Construction . . . . . . . . . . . . . . . . . . . . . . 95 Rental. Housing Losses -- Demolitions, Conversions and Foregone Starts .... . . . . . . . . . . . . . . . . . . . . 56 Housing From the Existing Stock . . . . . . . . . . . . . . . 96 Net Effects on Supply ... . . . . . . . . . . . . . . . . . 97 B. Rent Control and Landlord Profitability . . . . . . . . . . . . 98 Present Values .... . . . . . . . . . . . . . . . . . . . 99 A Present Value Model of Housing Investment in Kumasi . . . . 99 Gains and Losses from Four Components and their Interaction . 103 Rents .... . . . . . . . . . . . . . . . . . . . . . 103 Taxation .... . . . . . . . . . . . . . . . . . . . 103 Maintenance .... . . . . . . . . . . . . . . . . . . 104 Capital Gains ... . . . . . . . . . . . . . . . . . . 104 Effects of Rent Control on Landlord Profitabi'ity . . . . . . 105 Alternative Investment Opportunities . . . . . . . . . . . . 106 Affordability . . . . . . . . . . . . . . . . . . . . . . . . 107 Profitability and Affordability for an Existing Unit . . . . 108 What's Missing From the Model, and Related Conjectures . . . 109 Risk .... . . . . . . . . . . . . . . . . . . . . . 109 Imperfections in input markets and from other regulations ... . . . . . . . . . . . . . . . . 109 Market-wide effects ... . . . . . . . . . . . . . . . 109 Summary of Effects of the Current Regime . . . . . . . . . . 110 C. Effects of Other Regulations ... . . . . . . . . . . . 111 D. Analysis of Decontrol Policy Options . . . . . . . . . . . . . 112 Do Nothing .... . . . . . . . . . . . . . . . . . . . . . 113 Blanket Decontrol .... . . . . . . . . i . . . . . . . . . 113 Decontrol New Construction and Upgraded Units . . . . . . . . 115 Decontrol for New Tenants ... . . . . . . . . . . . . . . . 115 Floating Up and Out . . . . . . . . . . . . . . . . . . . . . 116 Other Options . . . . . . . . . . . . . . . . . . . . . . . 117 E. Summary .... . . . . . . . . . . . . . . . . . . . . . . . . 117 VII. A FINAL SUMMING UP .... . . . . . . . . . . . . . . . . . . . . 119 ANNEX TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 ix  LIST OF TABLES 1.1: Housing Infrastructure Indicators, Selected Countries . . . . . . . 3 2.1: Rent Control Measures, 1963 to 1987 . . . . . . . . . . . . . . . . 27 2.2. Prescribed Rents for Certain Premises, 1973 . . . . . . . . . . . . 28 2.3: Controlled rents July 1979 (AFRCD 5, 1979) . . . . . . . . . . . . 30 2.4: Rental levels un6-r PNDC Law 5, 1982. . . . . . . . . . . . . . . . 31 2.5: Rental levels under L.I. 1318, January, 1986. . . . . . . . . . . . 31 2.6: Prices of Selected S.H.C. House Types . . . . . . . . . . . . . . . 39 3.1: Household Environmental Indicators by Tenure . . . . . . . . . . . 46 3.2: Household Size by Tenure . . . . . . . . . . . . . . . . . . . . . 52 3.3: Mean Persons per Room by Tenure, 1980 and 1986 . . . . . . . . . . 53 3.4: Rents by Consumption Quartile . . . . . . . . . . . . . . . . . . . 54 3.5: Summary Descriptive Statistics . . . . . . . . . . . . . . . . . . 55 5.1: Controlled and Uncontrolled Samples Compared . . . . . . . . . . . 70 5.2: Gross and Net Rents Compared . . . . . . . . . . . . . . . . . . . 74 5.3: Controlled and "Uncontrolled" Rents Paid by Sector . . . . . . . . 75 5.4: Malpezzi and Mayo Demand Results . . . . . . . . . . . . . . . I . 77 5.5: Demand Equations . . . . . . . . . . . . . . . . . . . . . . . . . 83 5.6: Hedonic Index . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.7: Summary Cost-Benefit Measures From Cross Country Model . . . . . . 86 5.8: Cost-Benefit Measures From Survey Demand and Hedonic Models . . . . 88 5.9: Summary Coat-Benefit Measures By Consumption . . . . . . . . . . . 91 5.10: Summary Cost-3enefit Measures By Length of Tenure . . . . . . . . . 92 5.11: Landlord and Renter Incomes . . . . . . . . . . . . . . . . . . . . 93 6.1: Cash Flow Model of Rental Investment . . . . . . . . . . . . . . . 101 6.2: Key Model Inputs . . . . . . . . . . . . . . . . . . . . . . . . . 102 xi  LIST OF TABLES, CONTINUED A.A: Type of House by Number of Rooms . . . . . . . . . . . . . . . . 123 A.2: Type of House, 1979 and 1986 (Percent) . . . . . . . . . . . . . . 123 A.3: Monthly rents paid per household and per room . . . . . . . . . . . 123 A.4: Distribution of Household Monthly Consumption by Tenure . . . . . . 124 A.5: Household size distributions, 1970, 1980 and 1986 (Percent) . . . . 124 A.6: Total household size distribution by sector, 1986 (Percent) . . . . 124 A.7: Rooms occupied per household, 1980 and 1986 . . . . . . . . . . . . 125 A.8: Rooms occupied by tenure . . . . . . . . . . . . . . . . . . . . . 125 A.9: Total and co-resident household sizes, 1980 and 1986 . . . . . . . 125 A.10: Co-resident Occupancy Rates, 1980 and 1986 . . . . . . . . . . . . 126 A.ll: Overcrowded Ho-iseholds, Measured Against Different Thresholds, by Tenure . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 A.12: Households without access to services . . . . . . . . . . . . . . . 126 A.13: Households with exclusive access to services, 1980 and 1986 . . . . 127 A.14: Households without Access to Services, by Tenure . . . . . . . . . 127 A.15: Households with exclusive access to services, by tenure (Percent) . 127 A.16: Length of stay in the house, by tenure . . . . . . . . . . . . . . 128 A.17: Selected Index Numbers (Rounded), 1970 to 1986 (1963 = 100). . 128 xii  LIST OF FIGURES 1.1: Real GNP Per Capita . . . . . . . . . . . . . . . . . . . . . . . 1 1.2: Change in Real GNP Per Capita . . . . . . . . . . . . . . . . . . 1 1.3: Housing Investment and Capital Formation . . . . . . . . . . . . . 2 2.1: How Housing Markets Work . . . . . . . . . . . . . . . . . . . . . 18 2.2: Nominal Interest Rates . . . . . . . . . . . . . . . . . . . . . . 20 2.3: Real Interest Rates . . . . . . . . . . . . . . . . . . . . . . . 20 2.4: Measures of Financial Depth ..... . .. . . . . .. . . . . . 21 2.5: Selected Prices, Logarithmic Scale . . . . . . . . . . . . . . . . 29 3.1: Map of Kumasi . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.2: Distribution of Households by Tenure, Sector, House Type and Number of Rooms . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.3. Tenure by Sector . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.4: Households Witho-.;_ Access to Services . . . . . . . . . . . . . . 46 3.5: Monthly Rents . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.6: Consumption by Tenure . . . . . . . . . . . . . . . . . . . . . . 49 3.7: Household Size, Selected Years . . . . . . . . . . . . . . . . . . 50 3.8: Size Distribution by Sector . . . . . . . . . . . . . . . . . . . 50 3.9: Household Size by Tenure . . . . . . . . . . . . . . . . . . . . . 52 3.10: Rooms Occupied . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.1: Rent Control as Effectiv_ Price Control . . . . . . . . . . . . . 57 4.2: Rent Control with Elastic Supply . . . . . . . . . . . . . . . . . 58 4.3: Rent Control as Expenditure Control . . . . . . . . . . . . . . . 59 4.4: Rent Control and Consumer's Surplus . . . . . . . . . . . . . . . 62 5.1: Effect of Selectivity Bias . . . . . . . . . . . . . . . . . . . . 71 5.2: Advance by Length of Tenure . . . . . . . . . . . . . . . . . . . 73 5.3: Rents by Sector . . . . . . . . . . . . . . . . . . . . . . . . . 75 5.4: Rent to Income Ratios, Cross Country Model . . . . . . . . . . . . 78 5.5: Rent-to-Income Comparisons, Various Models . . . . . . . . . . . . 81 5.6: Rents and Predictions from the Cross Country Model . . . . . . . . 86 5.7: Cost Benefit Measures by Sector . . . . . . . . . . . . . . . . . 89 5.8: Changes in Consumers Surplus for Typical Tenants . . . . . . . . . 89 5.9: Cost Benefit by Consumption . . . . . . . . . . . . . . . . . . . 91 5.10: Cost Benefit by Length of Tenure . . . . . . . . . . . . . . . . . 92 5.11: Income Distribution by Tenure ..... . . . . . . . . . . .. . 93 xiii  LIST OF FI(URES, CONTINUED 6.1: Number of Houses by Sector . . . . . . . . . . . . . . . . . . . 98 6.2: Population, Rooms and Number of Houses . . . . . . . . . . . . . . 98 6.3: Change in Components of Landlord's Present Value . . . . . . . . . 103 6.4: Summary Present Value, New Unit .... . . ...... . . . . . 105 6.5: Internal Rate of Return: New Unit . . . . . . . . . . . . . . . 105 6.6: Real Interest Rates . . . . . . . . . . . . . . . . . . . I . . 106 6.7: Returns to Financial And Other Investments ared . . . . . . . 106 6.8: Willingness to Pay: Decontrolled New Unit 1) . . . . . . . . 107 6.9: Affordabilicy Over Time . . . . . . . . . . . . . . . . . . . . . 107 6.10: Internal Rate of Return, Existing Units . . . . . . . . . . . . . 108 6.11: Affordability, Existing Units . . . . . . . . . . . . . . . . . . 108 6.12: Internal Rate of Return, Existing Unit, Rents Overshoot Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . 114 6.13: Willingness to Pay Over Time, Existing Unit, Rents Overshoot . . . 114 6.14: Internal Rate of Return, Floating Up and Out . . . . . . . . . . . 116 6.15: Willingness to Pay, Floating Up and Out . . . . . . . . . . . 116 xiv  I. OVERVIEW A. Housing and The Current Macro-Economic Environment The last twenty five years have seen Ghana decline from one of the richest cou. :ries in Sub-Saharan Africa to one in which the infrastructure and capital stock, including ho;'sing, are in very poor condition. During the period 1965 to 1986, real per capita gross national product (GNP) declined at an average annua' rate of 1.7 percent per year. Figure 1.1 presents data for levels from recent years; Figure 1.2 presents the corresponding changes.1, Recently, however, there have been signs of an upturn. An ambitious package of reforms has been instigated by the Provisional National Defence Council Government of Flt-Lt. Jerry Rawlings. The structural adjustment program includes a devaluation and eventual market determination of the exchange rate for the Cedi (from C2.75 - US$1 to C90 in 1986, when we collected the household data, and C350 by March 1989) and other policy changes including agricultural price reforms ar& reductions in government staff levels. As a result, real GNP increased by 5 percent in both 1985 and 1986 (Figures 1.1 and 1.2). Figure 1.1: Real GNP Per Capita Figure 1.2: Change in Real GNP Per Capita '700 0 110% -4'700 _ ........ .... I.... $490~~ ~ ~ ~ ~~ 10 ... . . .. . ..... ... ..... ....' | GNP Pe, IaI . . , . .,/-Pefcenl coange 80 72 74 76 78 88 82 84 86 88 70 72 74 76 78 80 82 84 86 Concern remains that housing cclocitions have deteriorated over the past decade; improvement in these conditions is required to ensure that the bent'fits of adjustment are widely shared. As the adjustment in Ghana's economy takes hold, demand for housing can be expected to increase at least as fast as real incomes. One important task for the government is to ensure that its actions do not impede or constrain potential suppliers of housing. The purpose of this pdper is to examine the constraint on the housing sector imposed by rent control in the secona largest city of Ghana, Kumasi. This concentration on rent control is not intended to imply that it is the only, or most important, constraint. ;./ National accounts and other aggregate data are from the World Bank's BESD database, updated by written reports. These data are in turn obtained by the Bank from the Government of Ghana.  2 RENT CONTROL IN KUMASI, GHANA As will be discussed below, other problems restricting the supply of rental (and owner occupied) housing must also be confronted, such as problems in land markets, provision of infrastructure, finance, and other regulations, such as planning and building regulations. Shelter is the largest single form of fixed capital investment in most economies. In developing countries, the shelter sector usually ranges between 10 and 30 percent of household expenditure or 6 to 20 percent of Gross National Product. Housing ranges from 10 to over 50 percent of gross fixed capital formation and from 3 to 8 percent of Gross Domestic Product. The share of housing investment in GDP r_.ses as economies develop (see, for example, Mayo, Malpezzi and Gross, 1986). In general, these patterns have not been evident in Ghana over the last 15 years. Figure 1.3 shows that since 1970 the published numbers indicate that the share of housing investment in GNP has been falling from about 5 percent in the early 1970s to only about 3 percant in the years since 1980. However, such data are only indicative. Ghosh and Mohan (1983) reports that Ghanaian housing investment data are derived from data on cement imports and estimates of population growth, so, for example, changes in housing investment ziven a partictular change in populacion and level of cement imports will not be reflected in the data. On the other hand, the trend is still clear and is consistent with observation and anecdotal evidence (and survey data presented later in this paper). Figure 1.3: Housing Investment and Capital Formati.,n While these data have to be interpreted with more than the usual 14% caution, taken at tace value, they indicate I_S9OP IGCGOP that while both investment and housing \ % investment fell during the 70s, at the same *o. - ... .. V - time housing's shar. of fixed capital formation rose, because other fixed capital 8s . . . did even worse. Gross Fixed Capital Formation (GFCF) has declined from roughly 10 percent of GDP at the beginning of the 4x ....... ......... ... period to 5 percent in the early 1980s. The picture is just as bleak when annual 2%