ReportNo. 13171 -LSO Poverty Assessment

Public Disclosure Authorized August18, 1995

Human ResourcesDivision Southern Africa Department Africa Regional Office Public Disclosure Authorized Public Disclosure Authorized

Public Disclosure Authorized 94 7g~~~~~~~~~~ CURRENCY EQUIVALENTS (June 1993)

Currency Unit = Maloti (M) M I= US$.31403 US$ 1.00 = M3.19

GOVERNMENT April I - March 31

ABBREVIATIONS AND ACRONYMS

ARI Acute Respiratory Infection BEDCO Basotho Enterprise Development Corporation CBO Community-based Organization CLS Community Learning Systems CMA DRIG Drought Relief Implementation Group DSW Department of Social Welfare EC European Commission FMU Food Management Unit FNCO Food and Nutrition Coordinating Unit FSSP Food Self-Sufficiency Program HSA Health Service Areas IMR Infant Mortality Rate LCN Lesotho Council of NGOs LCU Labour Construction Unit LDTC Lesotho Distance Teaching Center LHDA Lesotho Highlands Development Authority LHWP Lesotho Highlands Water Project LISP Local Initiatives Support Program LNDC Lesotho National Development Corporation LTB Lesotho Tourist Board MFEP Ministry of Finance and Economic Planning MMU Microprojects Management Unit MOA Ministry of Agriculture MOE Ministry of Education MOH Ministry of Health NGO Non-Governmental Organization PHAL Private Health Association of Lesotho SACU Southern Africa Customs Union SSE Small Scale Enterprises SSIP Small-Scale Industries Project UNDP United Nations Development Project UNICEF United National Children's Fund USAID Agency for International Development VDCs Village Development Councils VHF Vulnerable Household Feeding VHWs Village Health Worker VWSS Village Water Supply Service WFP World Food Program WHO World Health Organization WIB Women in Business Lesotho Most Samwregiorn4come gros Nex Latestsingleyear recent Sub- higher Unit of estimate Saharan Low- income Indicator measure 1970-75 198tg 5 1987-92 Africa Income group Priority Poverty Indicators

POVERTY Upper poverty line local curr. 70.48c Headcount index %of pop 49.1Sc 19 Lower poverty line local curr. 35.24c Headcount index % of pop 25.79c GNP per capita USS 230 390 590 520 390 SHORT TERM INCOME INDICATORS Unskilled urban wages local curr. Unskilled rural wages Rural terms of trade Consumer price index 1987=100 21 76 197 Lower income Fooda 1i3 72 45 Urban Rural SOCIAL INDICATORS Public expenditure on basic social services %ofGDP . . 11.5 Gross enrollment ratios Primary % school age pop 105 113 107 66 103 Male 87 101 97 79 113 Female 123 125 116 62 96 Mortality Infant mortality per thou. live binhs 130.0 83.9d .. 99.0 730 45.0 Under 5 mortality 67 3 169.0 108.0 59.0 Immunization Measles % age group .. 63.0 70 0 54.0 72.7 DPT " .. 69.0 75.0 54 6 80.6 Child malnutrition (under-5) .. 15.6 28.4 38.3 Stunting " 33 Oe Wasting 2.4e Underweight 15 8e Life expectancy Total years 51 56 60 52 62 68 Female advantage " 4 0 6.0 5.0 3.4 2.4 6.4 Total fertility rate births per woman 5 7 5 5 4.8 6 1 3.4 3.1 Matemal mortality rate per 100,000 live birth 220 Supplementary Poverty Indicators

Expenditures on social security % of total gov't exp Social security coverage % econ. active pop. Access to safe water: total % of pop. 170 360 70.52c 41.1 68.4 Urban 65.0 65.0 95.66c 77 8 78.9 Rural 14.0 30.0 64 97c 27 3 60.3 Access to health care . 50.0 80 0 Population growth rate GNP per capita growth rate Development diamondb _ 6+ (annual average, percent) + (annual average, percent) 410 Life expectancy

4 5

2 * _ ~~~~~~~~~~~~~~~~~~~~~~~~GNPGross per primary capit enrollment

-2 -10 Access to safe water 1970-75 1980-85 1987-92 1970-75 1980-85 1987-92

Li Lesotho _ Lesotho Low-income Low-income a. Food price index: The index are from the U.N Statistical Diviion. Since the overall CPI is more up-to-date than the food price index, the data presentedmay represent different years and therefore may not be comparable. b The development diamond portrays relationships among four socio-economic indicators for a given country and compares them with the average of the country's income group The bold line diamond for the country can thus be compared with the average reference diamond drawn in fine line Any point outside the reference diamond representsa value better than the group average, and any point inside, a value below the group average. Lesotho Most Same region4ncomegrou Next Latestsingle year recent Sub- higher Unit of estimate Saharan Low- Income Indicator measure 1970-75 1980-85 1987-92 Africa income group Resources and Expenditures HUMAN RESOURCES Population (mre-1992) thousands 1,187 1,545 1,856 546,390 3,194,535 942,547 Age dependencyratio ratio 0 83 0 86 1 29c 0 95 0 67 0 66 Urban %ofpop. 108 161 209 295 267 57.0 Populationgrowth rate annual% 2 2 2 9 2.4 2 9 1 8 1.4 Urban 6.4 6.4 60 5.1 34 4.8 Labor force (1I564) thousands 597 730 844 224,025 1,478,954 Agriculture %of labor force 88 86 Industry 4 4 Female 47 45 43 37 33 36 Femalesper 100males number 78f Urban 128 Rural 100 NATURAL RESOURCES Area thou.sq km 30 35 30.35 30 35 24,27403 38,40106 40,697.37 Density pop.per sq km 39 1 50 9 59 7 21 9 81 7 22.8 Agriculturalland % of landarea 78 2 75 9 77 1 52 7 50 9 Changein agriculturalland annual% 0 7 0 3 0.0 0 1 0 0 Agriculturalland under irrigation % 0 8 18.2 Forestsand woodland thou.sq km Deforestation(net) annual% INCOME Householdincome Shareof top 20%of households % of income 62 04c 42 Shareof bottom40% of households 8 13c 19 Shareof bottom20% of households " 2.63c 8 EXPENDITURE Food % of GDP Staples Meat,fish, milk, cheese,eggs Cerealimports thou, metrictonnes 56 78 140 20,311 46,537 74,924 Foodaid in cereals 14 71 29 4,303 9,008 4,054 Foodproduction per capita 1987= 100 106 87 67 90 123 Fertilizerconsumption kg/ha 06 1 5 25 42 61 9 Shareof agriculturein GDP % ofGDP 28.7 17 6 8 4 186 29.6 Housing %ofGDP Averagehousehold size personsper household 5.4 6 25c Urban 5 26c Fixedinvestment housing %ofGDP 2 5 1 8 Fuel and power %ofGDP Energyconsumption per capita kg of oil equiv 258 335 1,882 Householdswith electricity Urban % of households Rural Transport and eommunication % of GDP Fixed investment:transport equipment 3.9 7.3 Total roadlength thou. km 3 4 5 INVESTMENT IN HUMAN CAPITAL Health Populationper physician persons 30,247 14,306 Populationper nurse 3,857 Populationper hospitalbed 597 1,329 1,050 516 Oral rehydyrationtherapy (under-5) %of cases 68 36 39 Education Grossenrollment ratio Secondary %ofschool-agepop 13 23 25 18 41 Female 14 27 30 14 35 Pupil-teacherratio: primary pupilsper teacher 53 55 54 39 37 26 Pupil-teacherrtio secondary 26 21 19 19 Pupilsreaching grade 4 % of cohort 79 78 64 Repeaterrate: primary % of total enroll 6 23 22 Illiteracy %ofpop. (ageIS+) 26 51 39 Female %of fem (age15+) 16 62 52 Newspapercirculation perthou. pop. 1 32 if 14 100 Source.World BankInternational Economics Department, April 1994 c Bankstaffestimates from Sechaba1993 Household Survey d 1986Census, Volume IV e: 1992National Nutrition Survey : 1938 data,'Human DevelopmentReport 1990" UNDP, 1990 LESOTHO POVERTY ASSESSMENT

TABLE OF CONTENTS

Title Page

EXECUTIVE SUMMARY ...... i A. Introduction. i B. Country Context. iii C. Poverty Profile .iv D. Fostering Labor-Intensive Growth .xii E. Investing in Human Resources .xxvi F. Strengthening the Social Safety Net .xxxi G. Strengthening Institutional Capacity. xxxiv H. An Action Plan .xxxviii

1. INTRODUCTION AND BACKGROUND .1 A. Country Context. I B. How , Drought, and Other Factors Affected Poverty. 3 C Objective and Organization of, the Report. 7

2. POVERTY IN LESOTHO: A DETAILED PROFILE. 8 A. Introduction. 8 B. Methodology. 9 C. Lesotho's Ranking by Important Socio-economic Indicators .10 D. Three Poverty Indices .13 E. Poverty's Distribution by Income Source and Occupation .15 F. The Geographic Distribution of Poverty .20 G. Poverty and Access to Land .22 H. Poverty and Household Demographics.23 I. Basic Needs Indicators .28 J. Targeting the Poorest Households .36 K. Basotho Perspectives on Poverty .41 L. Policy Recommendations.45

3. FOSTERING LABOR-INTENSIVE GROWTH .46 A Macroeconomic Perspective .46 B. Fostering Labor-Intensive Agriculture and Rural Development .50 C. Fostering Labor-Intensive Small Business.65 D. Developing Labor-Intensive Industry .71 E. Fostering Labor-Intensive Construction of Infrastructure .76 F. Policy Recommendations .80

4. INVESTING IN HUMAN RESOURCES .83 A. Introduction .83 B. Intersectoral Priorities .85 C. Priorities in Health Care .89 D. Improving Access to Safe Water and Sanitation Services.95 E. mproving Basic Education .101 5. ESTABLISHING AN EFFECTIVE SAFETY NET ...... 107 A. Introduction ...... 107 B. Current Safety Net Programs ...... 109 C. Major Problems and Lessons Learned...... 116 D. Consolidating the Safety Net ...... 119

6. STRENGTHENING INSTITUTIONAL CAPACITY ...... 127 A. Introduction ...... 127 B. The Role of Government ...... 128 C. NGOs' Role in Poverty Reduction ...... 135 D. Donor Coordination ...... 138

7. AN ACTION PLAN FOR POVERTY REDUCTION IN LESOTHO...... 140 A. Introduction ...... 140 B. Foster Labor-Intensive Growth ...... 143 C. Invest in Human Resources...... 156 D. Strengthen the Social Safety Net ...... 159 E. Improve Institutional Capacity ...... 161

ANNEX NO. Page 1. Modeling Economic Linkages ...... 165 Appendix A - Construction of the 1987 Social Accounting Matrix ...... 183 Appendix B - The Formal SIO Model ...... 187 2. 1993 Survey Data, Procedures for Calculations of Expenditure Values ...... 200 3 Poverty Measurement ...... 208 4 Alternative Poverty Measurement Based on Minimum Per Adult Caloric Requirement ...... 210 5 Poverty Measurement Indices ...... 211 6 Determinants of Poverty ...... 213 7 1993 Survey Data. Correlation of Non-Explanatory Variables with Poverty ...... 214 8 Summary of Qualitative Poverty Assessment ...... 215 9 Efficiency Comparison between Lesotho Mills and Village Mills ...... 218 10 Welfare Impact of Maize Pricing, Trade and Marketing Policies ...... 219 I I Diversifying Crop Production ...... 223 12 Minimum Wages ...... 233 13 Organization of Health and Education Services in Lesotho ...... 234 14 Employment Prospects in the South African Mining Sector...... 237

TEXT TABLE NO. Page 1.1 Trends in Selected Sectoral Aggregates, 1983-1992 (Million Maloti)...... 4 1.2 How Four Factors Affect Household Incomes in Lesotho...... 6 2.1 Social Indicators for Countries of Comparable Income ...... I1 2.2 Lesotho's Inequitable Income Distribution (% of National Income)...... 12 2.3 Shares of Income (Consumption) by Decile 1986/87 and 1993 ...... 12 2.4 Lesotho's Worsening Income Inequality (Gini Coefficients, 1967-1993)...... 13 2.5 The Incidence, Depth and Severity Poverty (1986/87 and 1993) ...... 14 2.6 Poverty's Distribution by Main Source of Income (1993) ...... 17 2.7 Distribution of Poverty by Occupation of Household Head (1993) ...... 19 2.8 Rural and Urban Poverty ...... 20 2.9 The Topography of Poverty (1993) ...... 21 2.10 Poverty by District (1993) ...... 22 2.11 Poverty's Distribution by Size of Landholding (1993) ...... 23 2.12 Demographic Characteristics of Basotho Households ...... 24 2.13 The Distribution of Poverty by Gender of Household Head (1993) ...... 24 2.14 Nutritional Status of Basotho Children Under Five (1992) ...... 26 2.15 School Enrollment by Gender (% of Total, 1993) ...... 27 2.16 Distribution of Poverty (by Age of Household and by Access to Pensions) ...... 28 2.17 Basic Needs Indicators of Basotho Households (% of all Households) ...... 29 2.18 Malnutrition Rates by Region (As a % of Children Under 5, 1992).31 2.19 Life Expectancy and Mortality Differentials by Place of Residence (1986).33 2.20 Environmental Health Conditions by Place of Residence (1993) .34 2.21 Educational Attainment by People Older than 15, by Location (1993).35 2.22 Determination of the Probability of NQI being Poor, 1993 .37 2.23 Key Indicators for Urban Households (% Identified)...... 39 2.24 Key Indicators for Rural Households (%) ...... 40 2.25 Combination of Key Household Indicators for Urban Households ...... 40 2.26 Combinations of Key Household Indicators for Rural Households ...... 41 3.1 How Output Growth in Selected Sectors Affects Different Classes of Households . 50 3.2 Labor Used, and Returns From, Irrigated Nontraditional Crops .53 3.3 Implicit Producer Price of Maize (1992/1993) .57 3.4 Characteristics of Lesotho's Small Businesses (1990) .66 3.5 Lesotho's High Utility Rates (in Rand Equivalent, September 1992).75 3.6 Proportion of Basotho Employed in Lesotho Highland Water Project Construction . 79 3.7 Projected Revenues from the Lesotho Highland Water Project (1991-2040).79 4.1 Changing Trends in Public Spending (1982/83-1993/94) .86 4.2 Breakdown of Spending on Social Services (1987/88-1993/94) .87 4.3 Skewed Health Spending (by Geographic Area).90 4.4 Cost Recovery Among Districts .93 4.5 Combined Capital Spending on Water and Sanitation .97 4.6 Actual Population Served by VWSS (1990) .97 4.7 VIP Latrine Construction by District ...... 99 4.8 Per Pupil Allocation in Recurrent Education Budget (1991/92) .101 4.9 Primary School Characteristics by District (1991/92) .103 5.1 School Meals Provided 1990-1992 .111 5.2 Subsidy of Different Grades of Maize Meal (per Ton, 1992/93).114 5.3 Coverage of Vulnerable Household Feeding Program .115 5.4 Proposed Safety Net System.119 5.5 Labor Intensity of Selected Public Works Projects .122 6.1 NGO Activities in Lesotho .136 7.1 Prioritizing the Policy Actions.140

TEXT FIGURE NO. Page 1.I GDP and GNP Growth and Overall Fiscal Surplus/Deficit as a Percentage of GNP. 3 2.1 Budget Shares in Basotho Households (by Quintile, 1993).11 2.2 The Pattern of Inequality .13 2.3 Main Sources of Income of Basotho Households, 1993 .16 2.4 Occupation as an Indicator of Poverty .18 2.5 School Enrollment Levels of Children Aged 6-17, by Location (1993) .36 3. 1 Lesotho Labor Force.48 TEXT INFORMATION BOXES 3.1 A Policy for Decreasing Herd Size in Lesotho ...... 64 3.2 Women and Small Business...... 68 3.3 Minimum Wages and Employment ...... 73 3.4 East Asia's Deliberation Councils...... 74 6.1 The Functions of Development Councils...... 132 6.2 Indonesia's Use of Household Survey Data to Design Anti-Poverty Policies ...... 134 7.1 Action Plan in Brief ...... 141

ANNEX TABLES 1.1 Production Activities in Lesotho (1987) ...... 166 1.2 Household Expenditures, 1987 Survey Results...... 169 1.3 Lesotho: Value Added Multipliers ...... 174 1.4 Distributional Impacts of Increases in Sectoral Output ...... 176 1.5 Lesotho: Trends in Sectoral Aggregates, 1980-1992...... 178 1.6 Household Incomes in Lesotho: Policy Simulations (Base SIO Model) ...... 179 1.7 Household Incomes n Lesotho: Policy Simulation (Base SIO Model with Reduced Urban Linkages...... 180 2.1 Estimation for Determinants of Total Expenditures for 1993 Survey ...... 202 2.2 Predicted and Extrapolated Expenditure Levels in 1993...... 203 2.3 Percentage of Households Consuming Reported Commodities in 1993 ...... 204 2.4 Food and Non-food Budget Shares Per Adult Equivalent Quintile in 1993 ...... 205 2.5 Percentage of Household Population Consuming the Following Items...... 206 2.6 Percentage Share of Household Expenditures by Given Categories ...... 207 3.1 Engle Food Scales...... 209 3.2 Index of Caloric Requirements by Age and Gender: Eastern Africa ...... 209 4.1 Minimum Per Adult Caloric Requirement: 1993...... 210 7.1 Correlation of Non-explanatory Variables with Poverty ...... 214 9.1 Lesotho Maize Mills: Operating Costs; 53 Weeks Ending 03/04/05 ...... 218 9.2 Comparison Between Lesotho Maize Mills and Village Mills ...... 218 10.1 Lesotho: Welfare Analysis of Monopolistic Maize Trade and Processing ...... 221 12.1 Minimum Wages as of August, 1992...... 233 13.1 HSA's in Lesotho ...... 234 14.1 Basotho Miners in ...... 237

ANNEX FIGURES 1.1 Graphic Representation of Multiplier Effects .170

APPENDIX TABLES Al.l Activities in the 1987 SAM .186 B1.l Model Equations .195 B1.2 Input-Output Model .198 Bl.3 Semi-Input-Output Model .199

APPENDIX FIGURES Bl.l SAM Structure .188

BIBLIOGRAPHY .239

MAP Lesotho IBRD Map 27147 ACKNOWLEDGMENTS

This report is the product of a collaboration between the Government of Lesotho, the Lesotho Council of NGOs and the . It is based on the findings of a World Bank mission that visited Lesotho in June 1993 and two subsequent missions during which various drafts of the document were discussed with the Government of Lesotho and the NGO community.

The first World Bank mission had the following members, each of whom prepared a background paper for the report: Ms. Neeta Sirur (Human Resources Division, Southern Africa Department, Task Manager); Mr. Paul Dorosh (USAID consultant, Input-Output Analysis); Ms. Maria Hoek-Smit (World Bank consultant, Urban Services); Messrs. Mats Lundahl and Colin McCarthy (World Bank consultants, Economic Growth and Poverty); Mr. Paul Shaw (Africa Technical Department, Social Sector Expenditures); Mr. Emmanuel Skoufias (World Bank consultant, Employment and Microenterprises); and Mr. Roger Sullivan (Africa Technical Department, Institutional Analysis and Safety Nets). Other background inputs were provided by: Ms. Natasha Palmer ( Delegation , Gender Issues); Mr. Seth Vordzorgbe (USAID Maseru, Tax Incidence Analysis); and Sechaba Consultants (1993 Household Survey and Participatory Poverty Profile).

Rogier van den Brink was the task manager for the in-country discussions of the report, which were organized by the Ministry of Planning. The Lesotho Council of NGOs coordinated the discussions within the NGO community, which included a three-day workshop to formulate comments and recommendations. Particularly useful contributions were made by Mr. E.M. Matekane, Mr. T. Ncholu, Mr. C. F. Sello (and his staff at the Lesotho NGOs), Mr. S. Santho, Dr. D.R. Phororo, Mr. P. Petlane, Mrs. M.E. Rapolaki, Dr. M. Maema, Mr. J. Zuidberg, Mr. N. Reynolds, Mr. P. Molapo, Ms. L.A. Makhele, Ms. M. Lerotholi, Mr. N. Mahao, Mr. S.M. Phamotse, Mr. J. Gay (and his colleagues at Sechaba Consultants), Dr. Majoro, and Mr. K. Matlosa.

The report was written by Neeta Sirur and Rogier van den Brink, and data analysis for the poverty profile was carried out by Ms. Yisgedullish Amde and Mr. Carlo Del Ninno (all Human Resources Division, Southern Africa Department), and Mr. Robert Floyd (Country Operations 1, Southern Africa Department). Ms. Pat McNees of the American Writing Corporation edited the report. Ms. Rhodora Mendoza Paynor, Ms. Guadalupe Prado, Ms. Ann Klofkorn, and Mr. John Brooks provided administrative support and did the formatting of the report. Peer reviewers are Mr. Lionel Demery (Education and Social Policy Department) and Mr. Peter Miovic (Eastern Africa Department). Mr. Roger Grawe is the Managing Division Chief and Ms. Katherine Marshall the Department Director.

EXECUTIVE SUMMARY

A. INTRODUCTION

1. In 1993, the Kingdom of Lesotho returned to democracy after 23 years of authoritarian rule. The democratically elected Government of Lesotho assigns high priority to the reduction of poverty. The aim of this document, produced in collaboration with the Government of Lesotho, is to help develop a strategy and action plan for reducing poverty in an efficient and sustainable way. It should help in setting development priorities, and should also be useful to nongovernment organizations (NGOs) and donors, many of which are reorienting their programs to address poverty reduction goals. The Lesotho Council of NGOs and its associated NGOs, the European Union (EU) and the United States Agency for International Development (USAID) provided invaluable help in the preparation of this document. Most of the policy recommendations in this report came from Basotho sources -- NGOs, academicians, civil servants, entrepreneurs, and villagers. They are not meant to be prescriptive. Rather, the options presented should be debated by appropriate democratic institutions, both local and national. At this point in Lesotho's history, the process of formulating policy is as important as the outcome, if not more so. And that process should be both transparent and inclusive.

2. Lesotho faces a number of enormous challenges. The young democratic process has proven to be fragile, but has valiantly survived the turbulent first years due to the broad-based popular support for the principles of democracy. Across the border, the peaceful transition to democracy in South Africa -- one of the political miracles of this century -- has completely changed the regional environment. These changes bring Lesotho new opportunities, as well as risks. This report highlights a set of polices which can help Lesotho reduce poverty by seizing these opportunitities and minimizing the risks.

3. This report sliows that poverty is worst in the rural areas. Many observers have all but written off broad-based agricultural growth as a viable economic strategy for poverty reduction in Lesotho, pointing to its limited and increasingly degraded natural resources. But only if the urban economy were able to generate massive amounts of employment, and it is unlikely to do so in the medium term, could the rural economy be absolved from the need to create more sustainable rural livelihoods. In the mean time, more profitable part-time, small-scale farming and livestock activities must be developed to compensate for a declining stream of mineworker incomes on which many rural households currently depend for their subsistence. The rural challenge facing Lesotho is to make family farming more profitable by shifting away from extensive grain production, in which it holds no comparative advantage, to those labor-intensive crop and livestock activities in which it can compete in the region. The priority policy changes necessary to reduce rural poverty are the deregulation of agricultural markets in order to attract private traders and processors, improving tenure security, especially for women who are the main farmers, and increased investment in those types of rural infrastructure that make small-scale family farming more profitable. Moreover, rural infrastructure investment which makes more use of local knowledge, management, labor and private contractors, will have a poverty reducing effect all of its own.

4. Access to wage income, either directly or indirectly (through ), reduces poverty. Currently, domestic wage employment absorbs only a small fraction of the total labor force, but its momentum is picking up, driven by an boomlet in textiles, garments and light manufactures. And domestic wvageemployment should increase rapidly to compensate for a likely decline in the demand for unskilled, foreign wage labor in South Africa's mines. While the Government and the - j -

Lesotilo uniionisshould vigorously enforce health and safety standards, there can be no doubt that a continued and accelerated expansion of these industries will increase employment dramatically and constitutes a key element of a strategy to reduce poverty. What will benefit the poor most in the long run is increased competition between more and more firms for local labor. To maximize the poverty reduction impact of private sector-led export growthi, the Government should keep wages competitive, improve industrial relations, increase efficiency in utility supply, and joill hiands with the private sector in increasing the supply of skilled workers through training and education.

5. Broad-based investmenit in hulimanicapital is the single most important input in the creation of employmenlt-initenlsive,poverty-reducinlg growtil. In both health and education, the Government should increase geographic equity and lower the costs of education and healthi services to the poor. In healthi care, a free or low-fee package of basic health services can be provided, while childreni under five, the elderly and the disabled should be exempted fiom user fees. In education, primary school costs to parents should be reduced, and more attentionl should be given to techncial and vocational eductation. Particular care should be given to boys' educatiotn. Boys in Lesotho are less educated than girls. They are apparently the victims of a cultural expectation patternlthat designates themiifor careers as herdboys and minieworkers and makes parents discoullt investmlienitsin their education.

6. This report also argues for a stronger social safety net, largely based on self-targetting measures. First, lowering the cost to the poor of essential goods such as food, fuel, water, healthi care, and educatior will greatly, and ilinediately, alleviate poverty. These basics use up most of the houselhold budget of the poor. In particular, the artificially highi food prices, remnants of a failed strategy to attain nationial food self-sufficiency, can be broughitdown by liberalizing markets for food staples. This is an iilportanit poverty-reducilIg policy action in a country that imports more than half of its food requiremilenits,and wvhereabout 90 percent of rural houselholds are net consumiiersof maize, the maini staple. Second, a substanitially expanded employment-initensive public works program should be contemplated, in vihicihbelow-market level wages would be used to self-target the employmenlt supply to the poor. Third, particularly vulnerable groups (e.g. elderly, the disabled, childreniunder five) should benefit from a food-supplement programil.

7. The sale of water to South Africa will bring in an important flow of revenues. Part of these revenues is set aside in a development fuLid. As this fund grows in size, it is essential that these revenues contribuite to increasing the productive potential of the couLitry and reduce poverty, in particular in the rural areas. At the heart of rural developmenlt is a core of rather simple, small investimienitprojects, suchi as the constructioll of rural roads and bridges, the digging of wells, the developmenlt of small-scale irrigation, and the building and repair of schools and primary healthi posts. Wheln coordinationi and implemiienitationlof these activities is unldertakeni by a faraway central bureauicracy or project authority, rural development becomes overly complex, costly and ineffective. But locally planniniigand executing suchi activities usually does not involve complex and costly coordinationi or inforimiationlgatherinig. This report makes the case for developing more decentralized mechianlisimisfor making investimlenitstargeted at poverty reduction.

8. There can be no doubt that Lesotho faces an uncertainifuture. But with apartheid relegated to the history books, an economiiycharacterized by macroeconomic stability and hiigh growth rates, a relatively well-educated and productive labor force, a small but quickly expanding export base, an agrarian structure based on family farming, and a sizeable inflow of revenues from the imminent sale of water to South Africa, Lesotho should be able to put its economy on a poverty-reducing pattern of growth. - iii -

B. COUNTRY CONTEXT

9. Lesotlio is often described as an economic "province" of South Africa, from where 95 percent of its imports come and to whicih 40 percent of its go. Lesotho is part of the Rand zone; its Loti is tied to the at a fixed exchange rate of 1:1. It is a member of the South Africa Customs Union (SACU), so South Africa collects its trade taxes. But most important is Lesotho's export of human capital to Soutih Africa, from which it has gained its reputation as a labor pool for South Africa's mines. About 40 percent of the male labor force is employed in South Africa, and remittances account for a third of GNP.

10. Lesotho's emergence as a labor pool for South Africa closely followed the general pattern of the creation of native reserves in the Southern African region. In the mid-nineteenth century European settler regimes effectively lowered the potential profits in a free African farming sector to force peasants into farm or mine labor. European settlers seized most of Lesotho's arable land, confining independent Basotho farming to a remote and mountainous area that was to become the independent state of Lesothio, and adopting protectionist measures against agricultural exports from Lesotho. During the twentiethi century, Lesotho's dependence on the South African economy cast it in the role of exporter of labor and exposed it to the long-term consequences of South Africa's economic policies. The failure of the inward-looking and capital-intensive South African economy to use its plentiful supply of labor efficiently has meant a general decline in employment opportunities, with negative consequences for Lesotho. And the mine labor recruitment policies followed in South Africa have had devastating effects on family structure in the recruitment areas, including Lesotho.

11. Not surprisingly, Lesotho's economy is highly vulnerable to changes in South Africa. In the mid-1980s, structural declines in external sources of revenue (SACU receipts and remittances from mining workers) led to an economic downturn and aggravated domestic imbalances: increasing uniemploymenlt, deteriorating social indicators, and a growing fiscal deficit. The need for macroeconomic stability was underscored by the signing in 1986 of the Treaty which started the implementation of the first phase of the Lesotho Highlands Water Project'. In 1988, the Government, assisted by the IMF and the World Bank, embarked on a structural adjustment program aimed at reversing the downturn. Lesotho's membership in SACU and the Rand zone gave it little flexibility about trade and monetary policies. Hence, the adjustment program implemented by the Government aimed to: reduce the fiscal deficit; avoid "Dutch disease" by creating a series of fiscal surpluses to absorb the flows of higih import duties due to peak construction levels of the Lesotho Highlands Water Project (LHWP); encourage private sector development, including the attraction of foreign investment; and increase social spending (especially on health care and education).

12. The adjustment program was largely successful and triggered an almost "classic" sequence of chaniges in the main macroeconomic variables (Figure 1). The pattern of growth of GDP and GNP2 was characterized by a contractionary period from the start of the program in 1988 until 1992. This

X The Lesotho HlighlandsWater Project is a large infrastructureproject aimed at exporting potable water from the mountainsof Lesothoto the industrialheartland of South Africa and generating electricityfor domestic consumption. Its initialphase w ill cost aboutUS$2.5 billion at 1990prices. Full scale constructionbegan in 1990. Exportsof water will start in 1994 and are estimatedto yield revenuesof about$55 million per year.

2 In ihe caseof Lesotho,which reliesso heavilyon remittancesfrom Basothoworking in South Africa,the difference betweenGDP and GNPis important.The GNPestimate adds net factor payments(mainly migrant mineworkers' remittances) to the GDPestimate. In the last few years. GDPconstituted about 60 percentof GNP. - iv - contraction was amplified by two years of drought in 1991 and 1992 and declining mineworkers' remittances. However, the economy turned around after 1992, when economic growth resumed, driven by peak activity in LHWP construction and export growth in textiles, garments, footwear and light manufactures.3 The fiscal balance was restored quite rapidly in the first two years of the program, and continued to improve during the rest of the period, until a small fiscal surplus was attained.

Figure 1: GDP and GNP Growth and Overall Fiscal Surplus/Deficit as a Percent of GNP

20 -- - -- 15 - GDP Growth

10

GNP Growth 0 - Rate

-10 -10 - - - ~~~~~~~~~~~~~~~~~~~~~SurplusOverall Fiscalin % 1988 1989 1990 1991 1992 1993 1994

13. It is difficult to assess the impact of structural adjustment on the poor, mainly because of a lack of compatible, consistent hiouselholdsurvey data collected at regular intervals. But model simulations (based on a Social Accountitng Matrix and a Semi-Input-Output model developed for this report) shiow that structural adjustment, wlhiclh vas maijnlyassociated with a tiglhtening of fiscal policy, seems to have hiad a limited impact oni liouselhold incomes in Lesotlio, and seems not to hiavedisproportionately affected the poor. Of four factors analyzed, the drought had the worst impact on the incomes of the poor, and the decline in workers' remittanices the least impact. The decliine in workers' remittances mostly hiurtthe non-poor. Constructioni activities tinderLHWP and growth in labor-intensive industry -- also part of the Governmenit's structural adjustment strategy -- appear to hiave increased lhousehold incomes, largely offsetting the adverse shocks of drouglht, reduced migrant earnings, and fiscal coniservatismii.But these estimates are only indicative and slhould not lead to complacency. Poverty in Lesotlho is widespread, as the poverty profile below sliows.

C. POVERTY PROFILE

14. Developing and implemiienitiniga comprelhenisive strategy for sustainable poverty reduction requires a poverty profile, a systematic analysis of available information about the number and clharacteristics of the poor, as well as an assessment of the implications of existing policies and programs affecting tlle living standards of the poor. For the poverty profile, two sources of primary data were used: the Lesotllo Bureatu of Statistics' 1986/87 Houselhold Budget Survey and a 1993 hiouselhold survey undertakeni by Seclhaba Consultanlts of Maseru, commissioned by the Drought Relief Implemenitationi Group.

3 From 19888o 1994 . the annual growth rates of real GDP and GNP averaged 8.5 percent and 6 percent. - v -

Lesotho's Ranking by Important Socioeconomic Indicators

15. Lesotho's per capita GNP of US$580 (in 1991) ranks it as one of the world's 40 lowest-income economies -- 93rd among 127 countries. In sub-Saharan Africa, however, it fares better, ranking ninth richest out of 32 countries.4 Indeed, Lesotho's per capita GNP is four to five times greater than some of the poorest African countries, such as Mozambique and Tanzania. Lesotho is also ahead of many of its African neighbors on social indicators, including some with much higher average incomes. In sub- Saharan Africa, for example, it ranks third in life expectancy and net primary school enrollment and has the fourth lowest maternal mortality rate among countries for which data are available. Still, Lesotho's social indicators are much worse than those of non-African countries with similar per capita GNP levels. In other words, there remains considerable scope for improving the quality of life of the Basotho people even at present income levels.

Table 1: Social Indicators for Countries of Comparable Per Capita GNP Guinea Sri Honduras -Liesotha:- Indo- Zim- Senegal Lanka nesia babwe Per capita GNP 460 500 580 -580 . . 610 650 710 (US$) Life expectancy 44 71 65 60 60 48 (years) Infant mortality 136 18 49 -81: . 74 48 106 Rate (per 1000) - - Adult illiteracy 76 12 27 23 33 62 Rate (%) Under five mal- na 45 21 2- 14 12 22 nutrition rate (%) . . _ : _ - _ - _. Notc: Datain this table are for 1991and have beentaken from the 1993World DevelopmentReport (WDR) for consistencyof internationalcomparisons. Data in this tablemay not be consistentwith datacited elsewherein this chapterand in the text,which were drawn from more recent Lesotho-specific surveys and reports.

16. Because of the country's inequitable income distribution, absolute poverty in Lesotho may be more prevalent than average income figures suggest. Of 40 developing countries for which data are available, only nine -- including Lesotho -- have Gini coefficients above 0.55, indicating a severely skewed income distribution.5 Five of the nine highly inequitable countries are in Latin America and four are in sub-Saharan Africa. There has been a marked increase in income inequality in Lesotho, especially in rural areas, over the past 25 years. Gini coefficients, which show inequality, climbed from 0.23 in rural areas in 1967/69 to 0.55 in 1993. In 1993, the richest 10 percent of households received 44.06 percent of national income, more than four and a half times the income share of the bottom 40 percent, wvhichreceived only 8.13 percent.

4Data areavailable only for sub-SaharanAfrican countries with populationsover I million.

5 Chen.Datt andRavallion, 1993, and World Bankstaff estimatesfor Lesothobased on surveydata. TheGini coefficientmeasures the degree of (in)equality,with 0 signifyingabsolute equality and i indicatingabsolute inequality. - vi -

Patterns of Poverty

17. To distinguish poor houselholds from nonpoor or better-off households requires choosing a poverty line, a level of welfare below which poverty is said to exist. Lesotho has no current official absolute poverty line, so we set two relative poverty lines for the purposes of this report: We characterize as relatively poor, hiouseholds whose members consume less than half of the mean consumption level. We characterize as very poor, or ultra-poor, households that spend less than 25 percent of the mean consumption level. To give the reader an idea of what these poverty lines mean in terms of meeting daily needs: A household with a spending level at the first poverty line would probably need to spend about 51 percent of its total household budget to get a basic daily adult diet of about 2500 calories. A hiouseholdat the second, ultra-poverty line would need to spend its whole budget to get the same diet and caloric intake.

18. Certain patterns emerge from the poverty profile. Relative poverty levels in Lesotho are high. Nearly half of all Basotho hiouseholds are poor (with consumption levels below half of average) and more than a quarter are ultra-poor (with consumption levels below 25 percent of average). A severely poor fifth of the population spends nearly two-thirds of its resources just on food and another 25 percent or so on such necessities as water, fuel, healthi and education. And malnutrition levels among children under five are unacceptably higiheverywhere.

19. Nearly half of Lesotho's populace is poor and slightly over a quarter is ultra-poor, but the poverty gap -- the depth of poverty -- is relatively small. In 1993, using the lower poverty line, the poverty gap was only 9.12 percent, whichi means that if perfect targeting were possible, a transfer of only M38.52 per person a year (or 2.3 percent of average adult-equivalent spending) would be enough to eliminate ultra-poverty. Using the higher poverty line, M203.04 per person a year (or 12.0 percent of mean adult-equivalent spending) would be needed to eliminate poverty. That translates nationally into M44.2 million (US$13 million) a year at the lower poverty line and M232.9 million (US$68.5 million) at the upper poverty line.6

20. Poverty is disproportionately high in the nearly 40 percent of Basotho houselholds that rely on farming, herding, informal business, or casual labor for most of their income, as well as among tile rotglgly 7 percent of people living in houselholds headed by unemployed adults. By contrast, the incidenice of poverty is belowl half the national average in households relying mainly on mining remittances from South Africa. It is also lower than average in households headed by regular wage earners workinig in Lesotho or by retirees receiving formal pensions.

21. Poverty is greater in rural than in urban Lesotho. About 82 percent of Basotho houselholds are rural (Figure 2). About 54 percent of rural houselholds are poor, compared with a Maseru average of 28 percent and an other urban areas' average of 27 percent (Figure 3). These numbers show that the incidence of poverty is lowest in urban areas outside of Maseru; only 4.5 percent of the population lives in these smaller towns. Of the poor population, 90 percent live in rural households (Figure 4).

6 In comparison.official developmentassistance averaged USS104.7 million a ycar between1990 and 1993,which, if perfectly targeted.is much more than "hat is neededto eliminatepoverty in Lesotho. Of course,in Lcsotho. as elsewhere, perfect targeting is nearly impossible.It is difficult to identify the poor. let alone design strategiesthat transfer exactl) the amountsneeded to each household.But the above estimatessuggest that substantially reducing poverty in Lesotho is an achievablegoal in thc mediumterm. - vii -

Figure 2: Geographic Distribution of Population

Total Population - 1.9 million

0 Maseru 14% 81% *Other Urban _ Rural

Figure 3: Incidence of Poverty

% of PeopleLiving BelowPoverty Line

60 -54% so S4

40- 28% 27% 30 20 10 0 Niaseru Other Rural Urban

Figure 4: Where the Poor Live

Total Poor = 980,000

10%

0 URban 90%URua - viii -

22. Poverty is also much deeper and more severe in rural areas (Table 2). In ultra-poor rural households, average adult-equivalent spending amounts to just 22 percent of mean adult-equivalent spending. The pattern is similar in poor households. Poverty is also far more severe in rural areas -- more than 3 times worse than in Maseru.

Table 2: Rural and Urban Poverty (1993) Incidence(P0) Depth(P1) Severity(P2) Population Poor Ultra- Poor Ultra- Poor Ultra- share (%) poor poor poor

Rural 53.86 28.63 26.53 10.24 16.16 5.60 81.91 Other urban 26.84 11.00 12.06 4.86 7.12 2.39 4.48 Maseru,urban 28.14 13.63 12.80 3.73 7.09 1.80 13.61 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source:World Bank estimates based on the 1993Sechaba survey. Notes: The POmeasure (or head count index or incidenceof poverty) measuresthe percentageof individuals belowa givenpoverty line, relative to thetotal population. The PI index(or poverty gap index) measuresthe depthof povertyas the relativegap betweenaverage actual incomein poorhouseholds and the povertyline. It measuresthe percentageof the povertyline incomewhich is neededto bringeveryone below the poverty line upto the poverty line. TheP2 indexis a modificationof the PI index.It weighsthe shortfall of a household'sincome more heavily, the furtherit fallsbelow the poverty line, suggesting the severity of povertyfor thepoorest of thepoor.

23. Poverty is worst in the mountains. The incidence of poverty is deeper and far more severe in the mountain and Senqu valley zone than in the foothills and lowlands. The mountainous and predominantly rural districts of Thaba Tseka, Mokhotlong, and Quthing are far poorer than other parts of the country, while houselholds in Maseru, Leribe, Berea, and Mafeteng are better off than average. The incidence of ultra-poverty in Thaba Tseka and Mokhotlong is nearly double the national average, and severe poverty at the lower poverty liie is similarly high (especially in Mokhotlong).

24. The incidence, depth, and severity of poverty are also generally below the national average in the predominantly lowland/foothill districts of Leribe, Berea, and Mafeteng. By contrast, the incidence of poverty is substantially below average in Maseru district, whiere 37 percent of residents are poor and 18 percent are ultra-poor. But because 25 percent of the Basotho population lives in the capital, Maseru's contribution to national poverty (18 to 19 percent) is still high.

25. Malnutrition levels among children under five are unacceptably high everywhere, but urban children are considerably better off than children in mountain areas; children in the lowlands and foothills fall somewhere in between. One sees similar geographic patterns in indicators for mortality and morbidity, access to safe water and sanitation facilities, and school enrollment ratios. Moreover, the quality of school, and hence educational achievement, is far worse in the mountains than elsewhere. Because better schooling is associated with higher income, these findings are especially troubling.

26. Poor households often have an unfavorable demographic structure. Lesotho's rural households are somewhat larger thani urban ones and have significantly highierdependency ratios, which typically is associated withi greater poverty. Rural households average about one more child under 16 than hotisehloldsin Maseru, the capital, and tvice as many adults over 60. In many poor households, too few members of working age support too many young and elderly dependents. - ix-

27. Some types of households and individuals are more disadvantaged than others no matter where they are. Households headed by old people or dejure headed by women are much poorer on average than other households. The number of households dejure headed by women who are single, divorced, widowed, or abandoned by their spouses (25 percent) is higher than in many other Afri can countries, undoubtedly because of the long tradition of Basotho men leaving to work in South African mines. And women are de facto heads of households in another 30 percent of households, an unsurprising figure given the high proportion of Basotho men employed in South African mines. These characteristics of poor households suggest that there is a strong life-cycle component to poverty, often aggravated by the fragility of the family structure caused by male labor migrancy. Poor households contain a higher number of persons of retirement age, with not enough life savings built up during their productive lives.

28. Especially vulnerable within households are children under five (especially boys), who apparently suffer malnutrition severe enough to inhibit their cognitive development and boys six to 15 (whio have much lower school-enrollment ratios than girls). Boys suffer more than girls, partly because even young rural boys are sent to help tend livestock, and partly because job expectations for Basotho boys are often for unskilled work in the mines in South Africa (Box 1).

Targeting the Poor

29. Statistical analysis undertaken for this report suggests that it may be virtually impossible to identify a group of easily measurable, objective household characteristics to distinguish poor households from nonpoor households, for purposes of targeting direct assistance with an acceptable level of error. Education levels are lower in poor households, for example, but they are low in many nonpoor households, too. What this means from a policy standpoint is that directly targeted assistance programs that require administrations to identify the poor or ultra-poor will probably not be an efficient or effective means of reducing poverty in Lesotho. More emphasis must be placed on self-targeting approaches, such as public employment schemes. x -

Box 1: Gender Differences Gentler bias against boys. Evidenceof genderbias favoring boys has been found in much of Asia; evidencefor Africa suggestsmore equitablepatterns. In Lesotho,however, there seemsto be a genderbias favoring girls over boys, regardlessof the genderand other characteristicsof the head of houselhold.For example, for all measures of malnutrition -- stunting, wasting, and underweight -- boys under five fare substantiallyworse tlian girls under five. The most commonexplanation offered for this is the commonrural practiceof sendingeven very young boys(three to five) withi older brothiersto tend •arnily livestock,as a result of whichithe boyseat only sporadicor inadequatemeals.

And Basothioboys clearly get less primary schoolingthian girls. Lower schioolenrollment for boys is found in all income groups, regions,and ecological zones,and in hiouseholdsheaded by both meniand women. The greatestdisparity, hiowever,is found in herding-orientedmountain zor.es,whlere nearly 30 percent more girls than boys attend school, lendinig credenceto tile view that livestock- tendinigdepresses boys' schioolenrollment levels. Still, that would not explain sizable disparities in schioolenrollimenit levels in urban areas.Other factors must be at play. One possibility is tihat many parentsstill view working in SouthiAfrican minesas the most promisingjob prospectfor Basotlio men, and de-emphiasizeboys' educationias irrelevantfor uniskilledand semi-skilled mininigwork. Clearly the quality of Basothioboys' lives and their income earning potential are diminishiedby low education levels.

Childlreninifenale-headled hotuselholls. Some data from tile 1993survey suggestthat chilldren living in liouselholdshieaded by a male residenthiave somewhiat lower levels of welfare -- poor nutritionl and educationialstatus -- thianthose living in houselholdsheaded officially, i.e. de jure, by women, and considerablylower levelsof welfarethian chiildren living in hiouseholdsdefacto hieadedby women. It is difficult to explain why cilildren in hiouseholdsdejnre hieadedby women fare better thanithlose headed by residenitmeni, despite the fact tilat the hiouselholdsdlejure headedby women are generally slightly poorer. Basedon evidencefrom other African countries,it seemslikely that wihenincome is controlled by wvomenrathier tlian by men, proportionatelymore h1ouselholdresources are allocated to children's w"elfare,partly becausewomen traditionally bearmore childcare responsibilities.

Basotho Perspectives on Poverty

30. To learn about Basothioexperiences of, and attitudes toward, poverty, we consulted two qualitative studiesof poverty in Lesotlio, bothicondticted by Secliabaconsultants. The studiesused such participatory tecihiques as focus group discussionIs,case studies, and participantobservation to get the perspectiveson povertyof Basotliocitizens (urbanand rural).

31. Poverty, especially severe poverty, is viewed as widespread in Lesotho, affecting every communityvisited by the researchiteams. In discussionswith groups of individuals identified by researchersas botil poor and better-off, the most commoniresponse to the questioni"Wliat is it to be poor?" wvasthat poverty'meanit a lack of food and clotliing. Basotilocommunities seem to identify only the most destitute-- probably a subsetof the ultra-poor-- as poor. Peoplealso often defined poverty as including: lack of employment (moniey), livestock, and agricultural assets,poor menitalor physical - xi - healtlh; and insecure shelter. Lack of adequate water and education were sometimes mentioned as defining poverty, but only infrequently.'

32. In rural areas, participants in discussion groups often associated poverty with small households comprised of only adult women or women and children without support from an adult male, principally because of a spouse or grown-up son's death, desertion, or illness (especially a mental or physical disability or tuberculosis). This is similar to findings of the quantitative analysis.

33. By contrast, those identified as destitute in urban areas were themselves largely unemployed adult men, classified by the researchers as: (1) mine applicants -- queueing daily outside the mining employmenit office in the hope of recruitment; (2) circle men -- some of them former mining applicants, who sleep "rough" in and around Maseru's main traffic circle, having depleted both social and material resources and having lost hope of stable employment; (3) shebeen men -- who remain around informal liquor hotises hoping for drink or accommodation; and (4) beggars, scavengers, and street boys -- of whlom the researchers felt there were relatively few in Lesotho compared with other countries. To be sure, the predominantly rural character of Lesotho, based on a relatively equitable agrarian structure svlicli is small-scale in nature, has prevented the emergence of large periurban slums such as can be foulid in Soutil Africa and Latin America, where the rural poor typically have very limited access to land.

34. Caitses of poverty. Participants' responses about the causes of poverty differed slightly in the 1991 and 1993 studies, partly because interviewers in the two studies took different approaches.8 In the 1991 study, the four causes of poverty mentioned most often were:

* Unemployment. * Enviroilnental disaster (including drought, floods, and hailstorms as well as loss of soil fertility and productive pasture land). * The lack of adequate agricultural land and inputs. * Ill healtlt or disability.

In the 1993 study, the causes of poverty mentioned most often were:

* Alcoholism, whichi participants considered to be the most common cause of poverty. * Unemployment (which moved from first to second place). * The drought. * Personal failings such as laziness. * Political factors such as injustice, oppression, and corruption, especially in local government. * Social issues such as conflict, hatred, and witchcraft.

7 By and large,education was not seenas a major factorin bringingabout prosperity. Rather,prosperity was seen-- especiallyby rural dwellers-- as drivenlargely by theability to securemine employment, an occupation believed to require physicalcourage and endurancemore than skills in literacy andnumeracy. 8 The first study took a "needsassessment" approach to causesof poverty, trying to link poverty with the lack of accessto economicservices or opportunitiesor to social services. In the secondstudy, researchersused a list of 42 pre- selectedconcepts to elicit participants' rankingsof a rangeof issues-- including theft, adultery, alcoholism, and corruption -- that they felt contributedto or affectedpoverty. - xii -

Not all of these issues can be addressed directly through public policy, but some have policy implications. Alcoholism, for example, is a preventable, treatable health problem that can be addressed throughi public hiealth measures. Action in this area may be especially important because participants strongly associated alcohol abuse with adultery, suggesting that it may be an important factor in the spread of AIDS. Similarly, the shocks to households from natural disasters could be lessened through better monitoring of weather conditions, prompt design and implementation of assistance interventions, and investment in irrigation. Finally, the frequent mention of abuses of local power must be kept firmly in mind if these authiorities are to be involved in identifying the poor for fee exemptions (for health care, for example) or for direct assistance (for food relief rations, for example).

D. FOSTERINGLABOR-INTENSIVE GROWTH

35. The poor depend mainly on income from labor to meet their daily needs. Any national strategy to reduce poverty in Lesotlio must promote not only macroeconomic stability and rapid economic growthi, but more productive use of labor, the country's -- especially the poor's -- most abundant resource. Given Lesotho's higih unemploymenit rate and incidence of rural poverty, there can be little doubt that thle genieral medium- to long-term strategy for balanced, sustainable poverty reduction must emphasize labor-intensive growthi in agriculture, small enterprises, and labor-intensive, export-oriented maniufacturinig,underpininled by major investments in the development of hiuman capital. In a labor- surplus economy like Lesothio's, the recommended sector-specific policy chianges and program initiatives to reduce poverty are also fully consistent with achiieving rapid growth and maintaining a macroeconomliicbalance (Box 2).

Macroeconomic Perspective

36. Historically, Lesothio's development strategy hias unduly emphasized self-sufficiency in the prodtictionlof basic commodities, wivetlieror not Lesotlio had a comparative advantage in doing so. In both agriculture and industry, Lesotho's policies have tended to promote capital-intensive import substituitioni. The strategy hias been one of parastatals, state companies, explicit and implicit capital cost stibsidies, protectionist meastires, and restrictive business licensing. Attempts to create mechanized, large- scale maize, whieat, and vegetable farms using rainfed and irrigated agriculture largely failed, despite considerable governmienitsupport. Lesotho imports steadily increasing amounts of these foods. Attempts to create an industrial and agroindustrial base producinigfor the local market hiave produced a range of subsidy or monlopolistic rent-depenidenitptiblic companies and parastatals. These parastatals have produced precious few jobs and lharmiedthe poor in at least tilree ways: by distorting budget allocations (for example, subsidizinlg parastatals at the expense of investments in rural infrastructure that would help the poor), by raising the consumer price of basic goods that dominate the houselholdbucdgets of the poor, and by crowding out more labor-intenisive, small and medium-scale private enterprises. - xiii -

Box 2: Policy Recommendations for Fostering Labor-Intensive Growth

In Agriculture: * Remove price, marketing, and processing distortions. * Invest in rural infrastructure that supports intensified agriculture and the production of export crops. * Make land tenure (and women's rights to land) more secure and rental markets more flexible. * Promote a reduction in herd size. * Reorient agricultural research and extension.

In the Small Business Sector: * Make it easier for small entrepreneurs (and in particular women) to get credit and do business. * Shift the emphasis away from broad entrepreneurship training in toward training in skills needed in the export and service sectors. * Develop rural and urban marketplaces.

In Export-Oriented Light Manufacturing: * Keep minimum wages down to a competitive level. * Improve industrial relations. * Reduce utility costs by increasing competition, regulation, and decentralization in the utility sector. . Root export firms more firmly in the economy by selling them factory shells and leases and by supporting skills training that promotes the use of local skilled workers in the export industry.

In Tourism: * Work muchimore closely with South African travel agents. * Develop a special "Lesotho package."

Through Infrastructure Investment * Invest more using labor-intensive approaches, but decentralize implementation. * Use private contractors on public works.

37. The import-substitution strategy was partly inspired by the political economy of the region-- whiicihmeanit rendering the country as independent as possible from South Africa's apartheid regime. But expansion of a labor-intensive private sector cannot be based on an inward-looking strategy. Lesotho's domestic market is simply too small for import substitution. More important, in a small, open economy that is not isolated by unduly higih transportation costs from its major trade partner, any evidence that import substitution works implies an opportunity to increase exports and to expand quickly.

38. Emiiploynent.There are high levels of under- and unemployment and far too few employment opportunities, in both Lesotho and South Africa. Underemployment and its byproduct, poverty, are wvidespreadin rural Lesotho and open unemployment is growing in urban areas. The little data on employmenit that exists shows the labor force to be greatly under-used. Projections from the 1986 - xiv - census show the labor force to be 800,0009. And population growth rates were high (2.6 percent a year) in the period between censuses (1976 and 1986), so net annual increases to the potential labor force should be between 20,000 and 25,000 over the next few years.

39. In 1993, opportunities for formal employment fell far short of need (Figure 5). Altogether only 28 percent of the workforce, or an estimated 225,000 people held formal wage jobs, distributed roughly allows (based on estimates from various sources):

* 115,000 Basotho worked in South African mines. * 21,000 were civil servants (including both salaried and waged workers). * 20,000 were employed in domestic manufacturing. * 18,000 worked in retail trade. * 15,000 worked in constructioni. * 13,000 did other work in South Africa -- for example, in agriculture, private industry, domestic service, or a professional occupation. * 10,000 were employed as teachers. * 7,000 had jobs in hotels, restaurants, transportation, or financial services. * 4,000 were employed in military or police security. * 2,000 worked in other services, noniprofit organizations, or private healthl care.

That left roughly 72 percent of the labor force (about 575,000 workers) to work in agriculture and the informal sector, sectors in whichi returnis to labor are generally low. The unemployment rate in 1993 -- countinig both unemployment and undereinployinent -- was estimated to be an alarming 35 to 45 percent.

40. Muiltiplier effects. Poverty has been reduced somewhat as the result of recent initiatives to develop the labor-intensive, export-oriented light industry (about 5,500 jobs created in five years, mainily in textiles) and to start up the Lesotho Highlands Water Project (around 3,000 jobs in construction), but the poverty-reducinig potential of these initiatives has not yet been fully exploited. In both cases, the benefits to the economy only consist of their direct employment effects. What is missing is a more economiiy-widemtiltiplier effect, whiclh could greatly enhance the employment impact of both the export boomlet in textiles and the massive construction investments under LHWP.

9 The potentiallabor forcein 1993(that is peopie15 to 64) vas slightly over I million. out of a total populationof about 1.9 millioni. The actuallabor force is calculatedassuming labor forceparticipation rates ofabout 85 percentfor men and 75 percenttor w%omen. - xv -

Figure 5. Lesotho Labor Force

Formal Labor Force *RSAMines

Labor Force - Total 800,000 225,000 *CivilServants

% 6% 4% 3% * Dom. Manuf.

8% 1% ERetail

8% ~~~~~~~~~~~~~~~oConstruction

7 Other Work in RSA

7\ _9% ETenchers

\Olnrormal/Ag I ric | 51% OTounsm/Financial * MilltaryAPolice

* Formal Wage * Sen,ice,lHealthSrceMat

41. It is important that Lesotho's growth strategy produce both direct and indirect growth in jobs and income (multiplier effects). The question is, from a multiplier perspective, which is more efficient: to invest directly in agriculture, for instance, or to invest in other sectors that could indirectly increase demand for rural products? To answer this question, the World Bank mission developed a Semi Input- Output (S10) model of Lesotho's economy to analyze the direct and indirect (multiplier) effects of changes in output in all sectors of the economy relative to their impact on the average incomes of Basotho households. In terms of poverty reduction, the model produced three main conclusions:

* Development projects and policies aimed at stimulating the demand-side of the economy cannot count on significant multiplier effects. In Lesotho, links between sectors are generally weak. Lesotho's economy is tightly knit with that of South Africa, through relatively open borders, membership in the Common Monetary Area (CMA), and the flow of migrant labor from Lesotho to South Africa. This tightly knit relationship helps to stabilize prices, encourage trade, and raise household incomes from migrant remittances, but results in substantial leakage from the domestic economy and thus few multiplier effects in Lesotho. Most goods and some services are tradable, so increased demand is met largely through increased imports.

* Because links between sectors are weak, the direct effects of changes in supply generally outweigh the indirect or multiplier effects. So the investments most likely to reduce poverty are those in sectors that direct a large share of value added to the poor. Most poor Basotho households are in rural areas, so increases in agricultural income have the most general potential to reduce poverty.

* Increasing Lesotho's export orientation is likely to strengthen supply links. In the long run, more local investment in, and increased production of tradables (or the conversion of nontradables to tradables -- for example, by canning vegetables), could substantially increase multiplier effects. Next to overseas markets, Lesotho could exploit the proximity and potential of the large South African market by developing exports for which productivity in Lesotho is high because of special characteristics of its land, existing investments, or available skilled labor. - xvi -

Fostering Labor-Intensive Agriculture and Rural Development

42. Despite agriculture's declininig share of GDP (from 35 percent in the 1970s to less than 20 percent today), its contribution to the economy is critical. Nearly 85 percent of Basotho households live in rural areas and about 70 percent derive part of their income from agriculture. And by and large, the 35 to 40 percent of households that rely mainly on agriculture for income are disproportionately represented among the poor and ultra poor. Agricultural employment and incomes must be expanded to avert medium- to longer-tern increases in poverty resulting from rapid population growth combined with stagnating opportunities for mine employment and too little expansion of formal employment to absorb the annual supply of new entrants to the labor force. To be sure, agriculture alone is unlikely to generate enough income to support the present rural population, given Lesotho's poor natural resource endowment (mountainous topography, limited arable land, unreliable climate, and serious soil erosion). But agricultural performance (and rural incomes) could improve substantially if the right policies were implemented: diversifying toward high-value export crops, involving the private sector more in marketing and processing, and improving land use efficiency.

43. From food self-sufficiency to food seciurity. Since the late 1970s, Lesotho's agricultural policies have been shaped largely by the perceived political need to reduce dependence on food imports from South Africa and to protect domestic producers from unfair competition from the heavily subsidized, large-scale South African producers. With the of a number of donors, including the World Banlk,Lesothio promoted food self-sufficiency in agriculture, witlh a focus oni increasinig output for staple food crops (especially maize). The assumption was that maize self-sufficiency was vital in case South Africa suddenly closed the border witli Lesotho.

44. In spite of massive subsidies and project assistance, Lesotho's maize production has continued to decline. Currently, even in a year with good rainfall, maize grain must be imported to meet as much as 60 to 65 percent of nationial food requirements. Rougilly 90 percent of rural agricultural households, as well as the rural landless and nearly all urban houselholds,are net consumers of basic staples. Data from the 1986/87 houselhold survey sliov only about 14 percent of national food consumption being homie-grown; the figures for 1993 are even lower, because of the drought. Only I to 10 percent of the maize processed by Lesotho's three industrial mills is procured from local sources; everything else is purclhased from the Soutlh African Maize Board.

45. Lesotho failed to achieve food self-sufficiency largely because of the unsuitability of its agroclimate for maize and vlveat production, includinig the vagaries of local rainfall patterns and the poor quality of its soils. Local and temporary dry periods are standard any summer, making rainfed maize production a risky venture every year. And drought -- a prolonged dry period over an extended area -- is relatively common. Erosion is severe because little of the land is protected by forest, pasture, or crop cover, for reasons related to land tenure, overgrazing, and fuel shortage problems. On top of these climate-related problems, some of the price incenitives offered under the food self-sufficiency strategy have accelerated soil erosion, by encouraging farmers to plant unisuitable marginal land to maize. This is especially true in the lowlands and foothills wlhere extensive long-term cultivation of food crops -- mainly maize -- combined with heavy grazing has caused the crusting and erosion of surface soil.

46. Lesotho's drive for self-sufficiency in food crops -- especially maize -- has hiurtthe poor and does not exploit of the counitry's comparative advantage in agriculture. Lesotho's food prices, wlichi are set administratively, have consistently been kept above SouthiAfrican prices, which are also administratively - xvii - set and kept wvellabove world prices. Both countries have introduced price and marketing distortions to help producers and to achieve nationial food self-sufficiency, but instead they have hurt their poor in the process. Because the poor are more likely to suffer food deficits and spend proportionately more of their income on food than the rich, the current marketing system is effectively a tax on the poor, one that largely benefits South African producers and Lesotho's large-scale mills. A poverty action plan that did not redress this inequity would miss an immediate, unique opportunity to reduce poverty and increase food security. Initial estimates suggest that lifting the ban on flour imports would immediately reduce the price of maize flour at least 30 percent. Using the household expenditure analysis done for this report, this would imply a reduction in the incidence of poverty from 49 to 44 percent, and of ultra- poverty from 26 to 18 percent.

47. AMovingtoward crop dliversification. Nontraditional crops such as asparagus, potatoes, and cabbage are much more labor-intenisivethani such traditional crops as maize, sorghum, and beans and yield greater returns per hectare planted. But artifically high food prices send the wrong price signal to farmers who should be diversifying away from the traditional food-grain monocrops toward non- traditional, highi-value, labor-intenlsive agricultural activities in which Lesotho has a comparative advantage. Lesotho shares a unique regional agroclimate with the eastern Orange Free State and the northierinmoulitaini areas of the Eastern Cape. The areas bordering Lesotho suggest the agroclimate's potential. The area arounid Ficksburg in the eastern Orange Free State (50 km from Maseru) is the center of Southi Africa's chierry productioll for export. Areas in the Eastern Cape bordering Lesotho do very well producilng wool and mohair. A recent Bank study'o concludes that Lesotho's potential for agricultural diversification is considerable because the country seems to hold a comparative advantage in several higih-valueexport crops (including asparagus, saffron, and some nuts and berries) and in some vegetables and fruits (includinlg cabbage, potatoes, tomatoes, onions, apples, pears, and peaches). The Ministry of Agricuilture hias identified other crops for whichi Lesotho might have a comparative advantage: beans, peas, sLlnflower,soybeans, cherries, dairy products, wool, and mohair.

48. Lesotho's agriculture strategy should be firnly targeted to shifting the country's agricultural product Imix toward this comparative advantage. It shiould do so by (1) removing price, marketing, and processing distortions that favor crops for whicil Lesotho has no clear comparative advantage; (2) investilng in rural infrastructure tilat supports agricultural intensification and a new mix of agricultural products; (3) makilng land tenlure (and woomen'srighits to land) more secure and making land rental markets more flexible; (4) intenisifylingefforts to promote smaller herds; and (5) reorienting agricultural research and extension toward the most promising crops and farming systems.

49. RemJove price, niarAeting, arld processinJg (listortions. To reduce food prices and shift the relative prices from food to export crops would require a phased deregulation of food grain and flour imports. This measure would greatly alleviate poverty by cutting consumer prices, helping restrain the upwvardpressure on wages, wvhicihhelps the labor-intensive export industry, and contributing to the sustainied, lonig-termprofitability of agricultural production by inducing a shift in production away from maize toward crops more suitable for Lesotlio.

50. For Lesotho's agriculture to develop, both indigenous and foreign traders shiould be able to operate in a liberalized market free of uninecessary regulation and the uncertainty of ad hoc government intervenitioni. The Governmenit should consider actively promoting private sector involvement in agricultural trade as a way to encourage agricultural diversification and raise rural household incomes.

10'Lcsotho: Enmironimicintand Agricultural Diversirication Study," World BankReport No. I 1906-LSO. - xviii -

It is a common fallacy to assume that small farmers could be made to market their produce themselves. International experience clearly shows that this only applies to situations in which farmers can directly sell their produce at a nearby village market. As soon as the farmer wants to diversify into crops demanded by markets which are further away (say the fresh asparagus market in ), he or she will lack the relevant information on price, quality and quantity. Traders are the economic agents who bridge this information gap between producers and the market. To constrain their operations ultimately means constraining the incomes of farmers.

51. Increased private trade would help expand the production of asparagus, one of the most promising export crops. Asparagus is not only labor-intensive but involves activities traditionally considered suitable for women -- an important consideration in Lesotho, where men are often absent from the household. Asparagus has been grown in Lesotho for export since the mid-1970s. Asparagus has been moderately popular with growers: The area planted to asparagus has grown from about 30 hectares in mid-1970 to about 400 hectares at present. Much more rapid growth would be possible, were it not for the policy environment. Tight controls on competition and private involvement in asparagus marketing, bans on private exports to South Africa in order to protect the Lesotho National Development Corporation-owned cannery, and the use of inappropriate technology have restrained yields to less than half their potential.

52. Until recently, Co-op Lesotho (a mainly government-funded farmers' cooperative) had a virtual monopoly on buying agricultural produce from growers and supplying them withi agricultural inputs. Becauseof serious financial difficulties, Co-op Lesotho has now suspended operations, and some de facto liberalization has thus occurred. But trade and marketing continues to be excessively regulated by restrictive licensing practices. A more liberal marketing environment would attract private traders and undoubtedlydiversify the crop mix, raise producer prices and increase yields.

53. In vest in rural infrastructure that supports intensified agriculture anndtile productionl of export crops. It will be impossible to exploit the country's comparative advantage in crop production without essential infrastructure. Rural infrastructurecan widen the range of crops grown and reduce productioni costs at the same time that it can lower marketing and transportation costs, make agricultural activities more profitable, and lower the price of urban imports. Such infrastructure may be related to water harvesting (for example, roof catchmenits, small dams, and dam linings), soil conservation, or small-scale irrigation (provided production is organized on a houselhold basis), and may range from fencing for garden areas to the physical infrastructure needed to establish periodic markets.

54. Make l(andtenure (antn womtent's rights) mnoresecure and land rental nmarketsnmore flexible. Lesotho's relatively equitable distribution of land is an important asset, both for agricultural growth and as a social safety net. Despite what powerful agricultural lobbies the world over assert, small-scale farming is not less efficient than large-scale farming. Most important is the pattern of rural and economic growth associated with the development of a small-scale farm sector: it is unambiguously pro-poor. So the small average farm size in Lesotho should be seen as an asset, not as an impediment to future agricultural growth.

55. Secure land-use rights to even a little land give the poor an important social safety net, especially in the absence of a formal social security system. Lesotho should avoid the problem of landlessness, which is increasing, and which has grave implications for rural poverty and urban slums. The Government should resist any attempt to redistribute land rights away from the poor in the name of so-called agricultural development. The premature introduction of land markets will merely allow the rich to accumulate land, because they often have better access to credit and subsidies thani the poor. The - xix - decision to allow land sales and purchases -- especially to people outside the village but also to people withinl the village -- should be left to the communities themselves. Even then, equity can be ensured only if community government is basically democratic.

56. Agricultural efficiency is currently most constrained by the reluctance of resource-poor households to rent out their land. This reluctance stems from the household's perception that renting out land, in particular for longer periods, could lead to the gradual extinction of property rights in that land. To ensure a well-functioning rental market and agricultural efficiency, it is important to establish secure property (use or leasehold) rights and legally enforceable (but unregulated) rental contracts. The Land Act of 1979 goes a long way toward protecting property rights and making contracts enforceable. It gives communlities the option of allowing individuals to transform traditional use rights into 99-year leases that can be sold, subleased, or mortgaged. But the Land Act has not been thoroughly explained to communities, its implementation has been haphazard, and survey requirements are sometimes too lengthiyand costly for the poor. Legal uncertainty about property rights especially harms the poor, who do not have the means to establish defacto ownership through such extra-legal channels as personal contacts, bribes, and political clout. The Land Act should be vigorously propagated and implemented. At a minililumii,the current option of converting traditionally held use rights to leasehold rights should be widely publicized and the survey requirements should be relaxed and made less costly.

57. More thani half of all houselholds in Lesotlio are headed by women. Although the Land Act expressly allows womenl to hold leases," their continued status as minors under Lesotho's customary (and modern) law hiliders their ability to acquire land use rights from their chiefs in their own names, whiich makes it difficult for them to apply for government leases. There is an urgent need to review the conistrainitswomeni face in acquiring and keeping land rights and seek appropriate solutions. How will woomen'srights be interpreted in the Land Act and its amendments? That is still open for debate and womeni's righits to own landed property apparently remain vulnerable. Yet their right to acquire secure land riglhtsis important both for poverty reduction and for agricultural efficiency.

58. Promote a reduction in herd sizes. Replacing livestock, a savings asset with a low and risky finanlcial return, with a more attractive financial asset could both contribute to raising living standards anld help protect the environment. In Lesotho, livestock production is roughly as important as the productioni of field crops in terms of value added. The livestock subsector's contribution to GDP declined from about 8 percent of GDP in 1985 to about 6.5 percent in 1990. Yet during the same time the nuImber of animals actually rose. The number of cattle increased from about 525,000 in 1986 to about 641,000 in 1990, while there were about 1.5 million sheep and I million goats.

59. In Basotho tradition, cattle, sheep, goats, horses, and donkeys, are generally considered to be men's animals. Livestock belongs to the man in the sense that only he can sell it. And men value livestock for more than its potential to generate income. It is valued, among other reasons, because (I) animiialsare used to pay bohale or bridewealth at the time of marriage; (2) social prestige is accorded to livestock ownverswhen, as is common, tiley place some animals with other households, allowing those houselholds to benefit from products such as milk, wool and dung; (3) livestock visibly demonstrates to the comIllunity that an absentee husband is still providing for his family; and (4) and perhaps most importanit, livestock functionis as a store of wealth or relatively illiquid saving that may be drawn upon only in times of great need or as a pension for old age. This gives livestock an attractive quality: it is relatively illiquid. As a result, Basotho men have tended to increase herd sizes at every opportunity.

lI lhat is, to acquirethe rightsto landuse from a deceasedhusband. - xx -

60. Everyone knows overgrazing in Lesotho is a problem, but solving the problem is complicated because people hold cattle not only for production purposes but as a form of savings. For those who hold cattle for productive purposes, we recommend pursuing the standard economic solution: Bring individual financial costs in line with economic costs by introducing grazing fees, or restricting the number of animals per household. But rules should be set and enforced by communities themselves. For instance, fees should be collected by local voluntary grazing associations, which should be allowed to keep most of the fees collected. Grazing fees could be used to improve range management and to compensate households that do not own cattle -- usually the poor -- for the use of the range, whicih is, after all, common property'2.

61. But introducing grazing fees may not lead to a noticeable reduction of the herds hield mainly for savings purposes, because people holding cattle as a savings asset may care much less about their financial rate of return than about their cattle's function as an illiquid savings deposit. The logical complement to grazing fees, then, might be to introduce a new savings instrument that could replace cattle as a savings asset. One promising option is long-term contractual savings instrumenits (certificate- of-deposit type instruments, for example, for pension purposes).

62. Reorient agricultural research and extension. With a shift away from the past emphasis on food self-sufficiency, Lesotho's research and extension efforts should be redirected toward different crops and farming systems. Applied research and related extension services are critical to the developmenit of many new horticultural cash crops, iiicluding such vegetables as pinto beans, seed potatoes, sweet corn, baby corn, frenchi beans, and broccoli; such fruits as apples, pears, and peaches; a variety of nuts and berries; spices such as saffron; and some species of flowers that can be dried for export.'3 To alleviate the severe, extensive poverty that flourishes in mouLitaiindistricts, highiest priority should be given to developing crops (such as seed potatoes) that are likely to grow well in the higlhlands.

63. One potentially successful extension model has been initiated -- the Production through ConiservationiApproach, supported by the Swedisih International Development Agency -- and efforts should be made to improve, intensify and expand the program.'4 Lesotho has also produced a number of private farmer "chiampions" like Mr. 'Musi, who, supported by the German-funded Matelile Rural Development Project, successfully designed a soil conservation strategy, reclaiming gullies by planting shrubs and grass. Role models like Mr. 'Musi are important and should be actively used in Lesotho's agricultural extension program, because in farmers' eyes they are often more credible advocates of innovationithan are the extension agents. And several NGOs (such as the Basotho Mineworkers Labour Cooperative, Lesotho Home Makers, and the Seventh-Day Adventists) have already accumulated considerable knowledge about small-scale farm models whichi can be tapped to better adapt extension messages to local opportunities.

1- Whereasin theory the chiefs managegrazing rights on behalf of the community, in practice it seemsthat some have tendedto view certaingrazing rights as their personalproperty. The issueof grazing rights is thereforeoften politically sensitive,and unlikely to be solvedin the nearfuture.

13These crops, and issuesaffecting agricultural research and extension, are discussed more iully in the 1993World Bank "Environmentand Agricultural Diversification Study".

14But farmers' investmentsin soil conservationare unlikely to be financially profilable under the traditional crop mix of maize and sorghum(Convery, 1995). This constitutesyet anotherreason to shift towards the more profitable export crops. - xxi -

Fostering Labor-Intensive Small Businesses

64. A number of export-oriented firms have met with considerable success recently and efforts to promote their expansion should continue unabated. But Lesotho's private sector consists mainly of small businesses, which continue to be neglected and underserved. Developing the small business sector is essential to solving Lesotho's unemployment problem, especially among the poor. It is important to increase their (including women's) access to capital and to business advice, to provide suitable training, and to develop rural and urban market places where they can do business.

65. The main impetus for labor-intensive growth in the small business sector will come from an increased demand for services resulting in turn from increased formal employment in the large export firms. As the export sector grows, more workers will be hired and competitive pressures will cause wages to rise. Part of these wages will be spent on imported goods, but another part will be spent locally, especially on services. Setting up service businesses (such as construction, small restaurants, processed foods, and the repair of durable consumer goods) does not require the substantial capital outlays often needed for small-scale manufacturing. Many service businesses can start small and grow by reinvesting profits.

66. Some small firms may succeed in exporting to certain niche markets in South Africa. That market is likely to grow considerably in the medium term. In the past, excessive regulation often kept Basotho firmis out of the South African market, but gradual deregulation in South Africa should open up new opportunities to small businesses in Lesotho. But because small firms rarely possess sufficient marketing skills to undertake export marketing themselves, the key to developing the export orientation of small frimislies in linking them to larger firms whicihdo possess such skills.

67. Improve access to capital atndto business advice. Lesotho's domestic commercial banks, unlike some of their Southi African counterparts, seem reluctant to expand into the small business sector, partly because they are so risk averse and face so little competition. But there is substantial and growing demanld amonig small businesses for capital and practical business advice, judging from the thriving informal credit market and recent private initiatives in small business lending and advice.

68. Modern institutions generally fail to reach the neediest part of the population with the small amoulits of capital they need. They especially don't reach women or rural dwellers. And Basotho houselholds outside the formal sector either cannot grasp or cope with the basic demands of bank borrowing or they view such loans as simple subsidies, because over the years Lesotho's financial institutions have repeatedly failed to enforce repayment. By contrast, traditional savings organizations seem to work. These organizations are predominantly rural-based, serve both men and women, and seem to be able to maintain financial discipline among their members, mainly through peer pressure. Two traditional savings and credit organizations that seem to be effective in Lesotho are the stokvel and the bturial society.

69. Stokvels and burial societies have been so successful as traditional savings and credit clubs that effectively promoting their growth in numbers may be a sustainable way to give small business owners and micro-entreprenieurs more access to needed capital. The clubs directly provide capital to small busines5es or help finanicecertain "lumpy" houselholdexpenses (such as health care costs or school fees, - xxii - in addition to burial feasts) and thus free up household income for use in business.'5 Banks could promote the opening of joint accounts for these clubs, and extend loans to them whicil could be filily collateralized by savings. And membership of a group with an outstanding savings and repayment record could help qualify an individual for a personal loan, thereby linking the informal to the formal sector. To encourage more innovative approaches to the financing of small businesses, and to business support for project preparation and implementation, the Government should strongly promote competition in the financial sector by various means and help banks design new types of financial intermediation.

Box 3: Traditional Savings and Credit Organizations. The stokvel, which is relatively common in developing countries and known in the literature as rotating savings and credit association, functions in a simple, cost-effective way. A group of individuals with a common bond -- such as kinship, community, or employment -- agree to save together a fixed amount of cash over a fixed period. Then members of the group take turns using the funds accumulated. For example, ten people agree to save MI0 a month for ten months. Each monthia different member uses the combined savings of Ml00, until everyone has had a turn. The borrowing order is negotiated prior to the savings contract. The economic efficiency of this institution is simple and elegant for two main reasons. All but the last person in the cycle has access to a lump sUIIm earlier than by saving for it individually, and all but the last person receives an interest-free loan. As long as group discipline is maintained, the system is a highly efficient method to mobilize business capital or funds for lumpy houselhold expenditures. The amount of capital which can be raised depends on the capacity of the group to increase membership or individual contributions and still maintain strict (re)payment discipline.

Burial societies function much like insurance clubs, collecting a fixed monthly amount from members. The amount collected, usually by an elected treasurer, may be small -- for example, less thani MI a month. Part of the accumulated funds are used when needed to defray the costs of the burial feast for someone in a member's family. Some burial societies have developed into small banks and lend money, against interest, to both members and nonmembers. All surplus funds are maintainied as general reserves (rather than belonging to individual members). The treasurer often deposits them in the nearest bank. There are usually more members in a burial society than in a stokvel, often as many as 150. But as in stokvels, members usually have a common bond, such as residence in the same village.

70. Make it easierfor women to get credit and do business. The small business sector is dominated by women. But women in Lesotho can still not get credit without their husband's consent, even if they have been long-standing depositors, and even if their deposits have been sizable. These restrictions hamper small business development. Women-specific restrictions on business and credit should be identified and eliminated. Indeed, all legislation -- all customary and common law, including "personal" law -- should be systematically reviewed for provisions that erode women's legal status and prevent their equitable treatment under the law. The National Organization of Women Lawyers in Lesotho has already done extensive work in this regard.

15Stokvels may not beespecially useful for seasonalactivities such as agriculture, because members would all needcash at aboutthe sametime of year. But they do lendthemselves to periodicmicroenterprises, including thosethat purchasegoods for retail,garment manufacturing, and so on. - xxiii -

71. Chlaitge tlhefocis of smyiallbusiness tranintig. Lesotho is ahead of the region in general level of schoolilng, but to stimulate economic groxvth it must improve labor skills in large firms and the business skills of small-scale entreprenieurs in the service sector. Some argue for targeting much of the training toward people already working in the export industry, rather than to people with no formal job experience. There is considerable merit to this argument. Experience in both Lesotho and elsewhere suggests that local entrepreneurs will emerge from the pool of skilled workers in formal industry (maciniilists, technilcians, and supervisors), rather than from the pool of the unemployed and self- employed. To begin improving, training institutions should consult more with businesses to identify the skills theyi need. By financing joint government-industry training programs, the Government can encourage export firms to train more local skilled labor on-site. Some organizations offer business trainilng programs, but they are largely geared toward small and medium-scale manufacturing for the local market, whenl training for the service sector might be more relevant. And some of the managerial training offered seems too sophisticated for emerging businesses. Small new businesses are often in urgenlt need of shimple bookkeeping skills, for which no appropriate training is available. Perhaps Lesotlio's stokvels and burial societies, whiclh apparently have excellent bookkeeping systems, could serve as a model for locally developed bookkeeping systems, a model that could in time be upgraded to a more formialsystem.

72. Develop rural ami! irbarDnmzarketplaces. Small manufacturers often run into marketing probleis. The solItioLnproposed to remedy these problems is generally to provide training in marketing skills. However, small manluifacturers(and small farmers) generally do not market their products themselves, but rely on middlemeni and traders to convey to them information about the demand side of the market. To promote the marketing of small business products initiatives other thanl training in marketing technilques should be considered. A legacy of the colonial, settler-dominated history of Southern Africa is the near-total absence of physical marketplaces where small businesses and producers can market their goods and services. Even a city like Maseru provides mainly for up-scale businesses, and has onlv a rudimentary bazaar-type central market. In the rural areas, no system of periodic, rotatilg markets lhasdeveloped, as it has in other parts of Africa. Periodic rural markets would give local small busilness operators an opportunity to sell and advertise their merchandise to more clients and to get valuable information on demand, supply, and prices.

Fostering Labor-Intensive Growth in Export-Oriented Light Manufacturing

73. Lesothio has actively been developing an export-oriented light manufacturing sector. And the wisdom of a strategy of labor-intenisive, export-oriented firms has already been confirmed for large Lesotho-based firms. Since the late 1980s, foreign direct investment in the textile and garment industry has created a small base of labor-intenisive,export-oriented firms, attracted by Lesotho's low wages and preferential trade agreemenits with , North America, and South Africa. In the last five years this export boom created 5,500 jobs. (By contrast, after nearly three decades of government investment and subsidies, Lesotho's 51 parastatals currently support a total of 7,700 jobs). The textile and garment industry can still expand and dramatically increase employment.

74. Curr-enitlyLesotho ranks third out of ten African countries exporting apparel and textiles to the United States6 . So far, Lesotlio's strategy for creating more formal sector jobs and growth in exports of

16 In 1994, Lesothoexported 17.7 million square metres to the U.S., compared to 17.1 for Kenya, 23.8 for South Africa. and 5(0.4square metres for Mlauritius. - xxiv - manufactured goods has largely succeeded, but there has been some concern that external factors will slow down the rate of growth. For one thing, the lifting of sanctions against South Africa has made investments in Lesotlio less attractive for regional investors. At the same time, past U.S. quotas" on some Lesotho clothing exports, as well as problems (under the Lome Convention) with domestic value- added on exports to Europe, may have had a negative effect on the investment plans of some foreign investors. But international experience suggests that the U.S. quota system, although disruptive, has never prevented a particular country from continuing to export to the U.S. market. Continuous switching to non-quota items -- for instance, from jeans to shirts -- reduces much of the "bite" of the system.

75. Expanding labor-intensive light manufacturing is the most promising component of Lesotho's industrial strategy. It could create direct and indirect jobs on a large scale. Local businesses can emerge from the pool of skilled, experienced machinists, technicians, and supervisors already working in large export firms. Continued expansion of the export industry will also increase demand for services as varied as catering, machine repair, and building construction and maintenance. Local sourcing of more inputs (such as packaging and transport) would also give local business new opportunities. And some production (for example, embroidering) could be given out as piece work to local subcontractors. Hence, if export industries continue to expand, the ripple effects will be felt throughout the economy and will lead to reduced unemployment and highier wages -- both crucial to poverty reduction. To keep the cost of doing business down, Lesotlio should maintain the minimum wage at competitive levels, improve industrial relations, and regulate utility cost pricing.

76. Keep nminimumwages down1 to conmpetitivelevels. Although the Government should refrain from administratively setting private wage scales, some would argue that it has a social obligation to set the minimum wage. But it should avoid the temptation to set the minimum wage at or near South African levels. Doing so would merely result in job rationing, more unemployment or parallel labor markets, and a loss of foreign direct investment. Export industries were attracted to Lesotho because of its preferential trade agreements and the low cost of doing business in the country. The cost of business could rise, however, if increased union activity and the excessive salary demands now emanating from civil servants raise the wages of those employed in the formal employment to the levels prevailing in Soutih Africa. Lesotho would then mimic South Africa's inefficient growth path and associated social problems: higih unemployment for most, along with higher wages and better working conditions for those fortunate enough to find employment in the formal sector. It is important that Lesotho not fall into this trap.

77. Imiprove industrial relationts. The Government has engaged in few pro-active attempts to improve strained relations between management and organized labor. It behooves the Government to actively foster regular communications between management and labor on a range of issues, perhaps by organizing regular deliberation councils on the East Asian model.

78. Lower utility costs by increasingconmpetition, regulation, and decentralization in the utility industry. It will be easier to improve wages and working conditions if utility costs -- typically second in importance to labor costs -- can be kept tinder control. To date, there have been few attempts to regulate utility cost pricing or to allow competition in the sector. The result has been aggressive cost

1 The U.S. quota on textilc categories 338/339.638/639. and 347/348 wcrc suspendcdin latc 1994, after having filled about90% of the quota. At the writing of this report. no quota call has beenmadc by the Coinniittce for Implementation of Textile Agreementsor the U.S. Trade Reprcsentati%e's 01fice for consultationsconcerning new quotas. - xxv -

recovery by utility comilpaniies.with no market or regulatory counterforcesto push costs down. In telecommunications.for examiiple,the numberof employeesper 100lines increasedfrom 48 to 68 under structural adjustmenit.meaning that the utility becameless efficient while recovering costs. Similar trends exist in the other utility parastatals.Consumers and businesseswill suffer, and unemployment will increase,if the trend is not reversed.Several studies of utility costs are currently under way or planned;their recommendationsshould be implementedas soon as they are available.

79. Lesotho could probably significantly reduce its utility costs by increasingcompetition in the utility sector (from both domestic and South African sources)and by strictly regulating utility price increases.At the sametime, the country should rationalize the managementand sourcesof utilities. In telecommunicationsand electricity supply, for example, it should consider complete integration with South African utilities. Lesotho's export firms would benefit from South Africa's cellular phone companiesentering Lesotho'smarket, Lesotho'sindustrial estatescould benefit from direct hook-upsto South Africa's electricity grid, and many parts of the country would be more easily connectedto the South African electricity grid than by investing in costly transmissionlines emanatingfrom Maseru. At the sametime, decentralizationof some utility serviceswould increasecoverage by reducingcosts. For example, internationalexperience shows that per capita water production costs are four times higher in centralizedthan in fully decentralizedsystems and are lowest under centrally coordinateddecentralized systems.

80. Root export firms more firmly in the economy by selling them factory buildings and leases and by supporting skills training that promotes thleuse of skilled local workers in the export industry. There is concernthat Lesotho'sexport industry is footlooseand could easily pack up and leave. LNDC (which currently owns factory buildings, holds the leaseto the land, and chargesthe companiesa rental fee) could sell the buildings and transfer the leasesto the export firns. This would give LNDC the capital to build new shells. Government-supportedtraining programsthat elevate local workers into technical and supervisoryjobs in the export firms would also help root the industry in the local economy.

Foster Labor-Intensive Tourism

81. In the past, Lesotho followed the South African-inspired hotel and casino model for tourist development. This may not suit Lesotho's comparative advantage or goal of labor-intensive development.Rather than expandthe traditional hotel and casino package,the Governmentcould try to develop more small-scaletourism (small hotels, lodges,and cabins);this would be more labor-intensive and would provide more opportunitiesfor local entrepreneurs.A real opportunity existsto involve local communities in the designand managementof suchsmall-scale tourism.

82. Tourism is still an underdevelopedsector in Lesotho,despite the country's beautiful scenery. The most important tourist market may be South Africa itself, where interest in African culture and village life is growing. But with tourism booming in South Africa and international demand for ecotourism increasing rapidly, Lesotho should also develop a tourist product attractive to tourists visiting South Africa. Ratherthan go it alone, Lesotho should market a Lesotho option as part of an integrated tour package to South Africa. To begin with, Lesotho could start harmonizing its visa regulationswith SouthAfrica's to prevent borderhassles on side trips to Lesotho. - xxvi -

Investing in Labor-intensive Infrastructure

83. Investing in labor-intensive infrastructure can help reduce poverty in two ways: The poor would benefit from infrastructure development and maintenance directly, as wage laborers, and indirectly, as consumers of infrastructure services and as employees of private firms emerging in the improved environment. Unfortunately, Lesotho's infiastructure bureaucracy has not been adept at developing labor-intensive infrastructure, especially in rural areas. Priorities set centrally are not responsive to local concerns and technologies chosen centrally have tended to be capital-intensive. Without local involvement in the planninig and implementation of public works, maintenance suffers and infrastructure deteriorates rapidly. Lesotho needs to decentralize implementation and use more private contractors in public works. Future financing of a substantial expansion of labor-intensive public works programs can be secured from the Development Fund in which part of the revenues of the Lesotho Highlands Water Project are deposited.

84. Decentralize implementation. Public infrastructure projects can be made more efficient and equitable by decentralizing decision-miiakinigto the local level. Decentralized programs can quickly and cheaply create and maintaiin rural investment in stuchtechinically simple infrastructure as rural roads and bridges, school buildings, and small-scale water supply systems. A recent World Bank review of experience in 42 developing countries founid that where road maintenanice was decentralized, backlogs were considerably lower and road conditiows tenidedto be twice as good.

85. Use private contractors in public Kwrks. Public infrastructure development and maintenance has traditionally been a monopoly of thc ptiblic sector. But efficiency could be enhanced, output expanded, and more jobs created if the public sector restricted its role to planning and coordinating infrastructure investment, privatizing infrastructure development wherever possible, and leaving infrastructure maintenance to local communities and the private sector.

E. INVESTINGIN HtIMAN RESOURCES

86. Lesotho's record on human rcsource development has been more favorable than many other African countries at similar levels of development. For at least a decade, one government after another has recognized that human resource development is central to Lesotho's development prospects and allocations for social services have steadily increased. Over the last ten years, real per capita public spending on health care and education has nearly doubled, despite rapid population growth; allocations for those sectors climbed from under 20 percent of public spending in 1982/83 to nearly 30 percent in 1993/94. This financial commitment, combined with generally favorable sectoral policies, has enabled Lesotho to expand access to basic social services considerably. Lesotho has established a broad-based, decentralized health care system and has significantly expanded primary health care services targeted at major childhood and adult diseases. It now provides more potable water for urban and rural citizens. And school enrollment has increased dramatically: 30 percent in primary schools and 70 percent in secondary schools.

87. Still, the general level of human resource development is low, especially among the poor, and consistent, effective action is needed to stem declines in the quality and coverage of services. A comparatively high proportion of household spending goes to health and education, both in the population as a whole and especially in the three lower income quintiles. As experience in countries -xxvii - such as Honduras, Sri Lanka, and Zimbabwe shows, resources can be used more cost-effectively, thereby giving the poor better coverage and producing better sectoral outcomes generally. Human capital is, almost by definition, the only capital the poor possess. In the long run, investing in human capital is the only way to eradicate poverty in Lesotho.

88. Simultaneousaction is needed in several mutually reinforcing social services, includinghealth care, primary education, water and sanitation,and family planning(Box 4). Educatingwomen not only makes their worktime more productive, but is also associated with better health and nutrition for children. Iivestments in family planning services improve the health and incomes of women who choose to space or limit pregnanciesand also lower infant mortalityand morbidity rates. Lower fertility and (hence) population growth rates reduce pressure on limited health care and education resources. Healthy, well-fed children learn more.

Box 4: Policy RecommendationsFor Investing in Human Resources In Health Care: * Increasegeographic equity. * Exemptchildren under five, the elderly,and the disabledfrom user fees. * Introducea free or low-feepackage of basic healthservices. * Allow local facilitiesto retaincertain feesso peoplefeel they have localcontrol over services.

In Water and Sanitation Service Delivery: * Find ways to make water distributionmore equitable geographically. * Increase governmentallocations for maintenanceof water systems * Strengthen local staffing for central and district managementof water and sanitation programs, and consider subsidizingthe buildingof sanitary facilities,so poorer householdscan afford them. * Replace upfrontwater connectioncharges with loans.

In Education: * Spend proportionatelymore on primaryeducation and proportionatelyless on collegeand( * universityeducation. * Improvevocational and technicaleducation. * Investmore in educationin poorerregions. * Create incentivesfor qualifiedteachers to teach in remoteareas. * Lowerthe cost to parentsof primaryschooling. * Strengthennonformal education for out-of-schoolyouth.

Improve the Health Status of the Poor

89. Lesotho has a relatively well-developednetwork of health services, in which the churches and NGOsplay an active role. This network is an importantasset that can be improved by streamliningthe activities of service providers,especially through joint planning of health activities. In many respects, current health priorities, policies, and programs provide an appropriate framework within which to improve the health status of the poor. The Miniistryof Hiealthhas emphasizedthat broad-based primary health care is to be stressed over morc ,ophisticated care as the principal means of dealing with preventable (or easily treatable) inifectiotisdiseases, the primary cause of serious illness and death among the poor. Priority is also given to decentralization(through health service areas) and to fostering - xxviii -

a working partnershipwith private(often church-related NGO) health providers to expandcoverage into moreremote rural areaswhere the poor often live. The Ministryof Healthis emphasizinghealth, family planning,and nutrition interventionstargeted to childrenunder five andto mothersas well as the treatmentof debilitatingillnesses (such as tuberculosis)that affect adults' ability to work. These prioritiesare consistent with povertyreduction goals because they focus on someof the mostvulnerable individualsin society.

90. Still, the poorare not servedas well as thebetter-off; they do not havethe sameaccess to basic servicesand to good health care. By and large,health care facilities and outreachprograms serve proportionatelymore people in richerareas than poor areas, and rural mountainareas get lessattention thanurban centers. The following policy actionscould improve this situation.

91. Improve geographic equity. Poverty in Lesothois associatedwith geography,and rural mountainareas harbor the worstpoverty. It is importantto targethealth services increasingly toward thesepoor areas. To improvegeographic equity, the sector should:

* Reviewthe currentdistribution of hiealthservices, to providea benchmarkfor futureplanning. * Targetinvestments to the Senquvalley and rural mountains,where poverty is concentrated. - Createappropriate -- real, market-related-- incentivesfor health care workers to serve in remote areas:Make it payboth financially and professionally to beposted to remoterural areas. - Strengthencommunity-based approaches, so healthservices can expand cost-effectively into poorer regions.

92. Exempt children underfive, thleelderly, and the disabledfrom userfees. Increasedefforts to recovercosts under structuraladjustment have often meantacross-the-board fee increases.It is importantto exemptthe poorfrom suchcost increasesand to targetfree healthservices to vulnerable groups.

93. Introduce a free or low-fee package of basic health services. To improve universal accessto basicprimary health care, the public healthsystem should identify a packageof basichealth services, assignvery low fees for servicesin the basicpackage and aim for full cost recoveryon all other services.Most of the basic-packageservices are generally provided by low-endfacilities rather than by hospitals,so deliverywould be biasedtoward the poor,since low-end facilities service the poor,not the rich.

94. Decentralize by allowingfacilities to retain certainfees. Service delivery could improve if healthservices were decentralized. Local participationin the managementand maintenanceof health servicefacilities will improveif peoplefeel their payinguser fees allow healthcare services to function andfeel they have local control over the use of suchfees.

95. Pay more attention to important but underemphasized health problems. It is crucial to put moreemphais on AIDS prevention, giventhe potentialfor an epidemicand the fact thatAIDS tendsto strikeadults in their prime,leaving whole households without a meansof support.And consideringthe known links betweenpoverty and alcohol abuse, the sectorshould probably also study *vays to address problems of alcolholism. - xxix-

Improving Access to Safe Water and Sanitation Services

96. Improving health, especially where infectious diseases account for most illness and death, generally involves making safe water and sanitation facilities more widely available. One study credited improved waste disposal with a 36 percent reduction of diarrhea in Lesotho, prompting the conclusion that many sanitary interventions could have a bigger impact on health than improvements in health care services do, especially in highly contaminated environments. As a result, government policy emphasizes that all households have access to safe water and substantially increased access to improved latrine facilities.

97. To date, however, less than half of rural households in Lesotho have access to clean water, although the situation is better today (an estimated 35 to 45 percent coverage) than it was in 1984 (14 percent). Sanitation facilities are available to even fewer rural households (less than 20 percent). In urban areas, especially in Maseru, about 80 to 85 percent of households have access to safe water, and about 60 percent have access to sanitation facilities. Spending for urban water and sanitation (from both donor and government sources) is proportionately much higher than spending on rural services. Government policies seem to favor the richer urban population over poorer urban-dwellers, although this may be partly because of the greater risks of infection and epidemics in densely populated areas. More telling are coverage differentials within rural and urban areas, which do suggest that the poorest are neglected while better-off areas receive service.

98. The Government should consider the following measures to give the poor better access to water and sanitation services:

- Review program priorities in the water sector and find ways to make water distribution more equitable geographically. Safe water is far less available in the poorest rural districts (Thaba Tseka and Mokhotlong) and the newer areas of fast-growing towns such as Maseru and Maputsoe. * Steadily increase government allocations for maintenance of water systems and gradually reduce the current dependence on donor resources, to ensure that sectoral investments are sustainable. * Strengthen local staffing for central and district management of water and sanitation programs, to ensure a smooth transition when donor-funded expatriate staff are phased out. Priority should also be given to coordinating activities in now-separate rural water and rural sanitation programs, especially in such areas as hygiene education and promotion. * Consider subsidizing the building of sanitary facilities, so poorer households can afford them. The cost of building materials (M400-M500) is now well beyond the reach of most poor families. The high public returns from better sanitation could justify a public subsidy. * Replace upfront connection charges with loans. The main barrier to using infrastructure services is often that poor households cannot mobilize the heavy initial connection charges. In urban areas, utilities can be useful conduits for loans extended to finance connection costs because payment can be secured using regular billing procedures.

Improving Education

99. Massive investments in primary education are a common denominator in economic success stories all over the world. Top priority should be given to investing in primary education, for which the social returns are extraordinarily high. Lesotho should aggressively ensure primary education to as much of the population as possible. A good argument can be made for providing primary education virtually free of charge. - xxx -

100. Spend proportionately more public funds on primary education and proportionately less on college and university education. It is difficult to overstate the importance of primary education to a pro-poor economic growth pattern. Lesotho's general growth in spending on education is currently pegged at fiscally sustainablelevels, so increasing primary education's share of the sector's budget should be achieved mainly by redirecting expenditures within the sector. Some costs can be cut by continuing to decentralize the educational system, but most of the proposed increase in spending on primary education should come from reduced spending on tertiary education. Currently, the Government spends only about M200 per primary school student, compared with Ml ,000 per secondary school student and a staggering Ml 5,000 per university student.

101. About a quarter of the Ministry of Education's budget is allocated to the National University of Lesotho (NUL), which spends three-quarters of its budget on administrative overhead. Students at NUL receive a generous "all-inclusive" benefit package (which even includes laundry and student-room cleaning services). Students also receive loans, but repayments rates are very low and Government could show more diligence in recovering these loans. Students in Lesotho, as in most other countries, should get funding from various sources, including means-tested and merit-based scholarships and part- time work. But government spending per student is so high that it would probably be more efficient to send students to first-class South African universities than to educate them at NUL. The Government could save substantially by financing scholarships for studies abroad (in particular in South Africa), rather than try to offer comprehensive, first-rate tertiary education in Lesotho.

102. In any case, political changes in South Africa and Lesotho demand that tertiary education be evaluated more and more from a regional perspective, and achieving "self-sufficiency" in tertiary education may turn out to be too costly to be sustainable. NUL's academic reputation is not strong enough to warrant continued across-the-board public subsidies. Instead, NUL should try to establish a regional comparative advantage in pre-college and college-level education, possibly with an increased emphasis on technical, health, and education training. If NUL succeeded in providing bachelor-level college education at a regionally competitive level, it might increasingly attract paying students from South Africa, which would reduce Lesotho's fiscal burden. And the Government could direct those cost savings toward primary education.

103. Invest more in poorer regions. Investing more in basic education in the rural mountain zones would be unambiguously pro-poor. But as educational services in remote rural areas expand, the country must also develop more cost-effective ways to deliver educational services. Decentralization would both improve the delivery of education and reduce costs. School buildings could be maintained more cost-effectively (and more labor-intensively) by local communities than by a central bureaucracy.

104. Give qualifed teachers an incentive to teach in remote areas. If primary education is the cornerstone of Lesotho's human resource development strategy, teachers -- the first-line providers of education -- should be paid accordingly. Government should carefully monitor teachers' salaries and keep salary levels competitive. To get teachers to work in remote areas, the system must put together an attractive remuneration package that combines real financial and career incentives.

105. Lower the costs to parents of primary and secondary schooling. Basic primary school fees, currently about MI 00 a year, are already excessive and clearly anti-poor. In addition, parents must pay many extra costs (for books, school uniforms, and fees to hire extra teachers). And certain exam fees are prohibitively expensive, often several hundred Maloti. These costs are biased against the poor and result in high dropout rates. There have been several studies of what schooling actually costs parents. These studies should be used to formulate policies to cut family costs. General costs could also be lowered - xxxi - through decentralization -- for example, by creating local school boards with authority over local budgets.

106. Improve vocational and technical education. The general level of schooling in Lesotho puts it ahead of the rest of the region, but it is weak in vocational and technical education. Yet jobs and the economy cannot grow unless firms can hire enough workers with the right skills. The impact of LHWP on domestic employment has been below expectations partly because the Basotho labor pool lacked the right skills. Education planning in Lesotho seems to assume that students will one day be employed in the public sector; the idea of the private sector as a stakeholder in education is fairly new. But Lesotho's private sector should be encouraged to participate in curriculum design so that the skills mix of the labor pool improves.

107. The Government, in collaboration with the private sector, should undertake market research to determine what types of vocational and technical education are in high demand and short supply, and should develop the institutional framework appropriate for providing and financing the vocational and technical education needed. Such a framework might include, for example, a business training center operated as a joint venture between the Government and the private sector.

108. Strengthen nonformal education for out-of-school youth. The low level of education for boys is a serious social problem in Lesotho. In the past, it was realistic to expect no more for boys than a career of unskilled physical work in South African mines, but Basotho parents are still unwilling to invest in their boys' education, and instead of sending boys to school often use them as cheap labor for herding livestock. Declining mine employment will eventually change people's attitudes, but the Government can help by actively promoting education and changing public perceptions about future opportunities for mine work. Next to reducing the cost of primary education, the Government should consider imposing age restrictions on herding. It might also be more effective to allow local teachers to earn overtime by providing afternoon and evening classes to out-of-school youth. The overtime option might add to the allure of teaching in remote rural areas.

F. STRENGHTENTHE SOCIAL SAFETY NET

109. The first two strands of the poverty reduction strategy -- fostering labor-intensive growth and investing in human resources -- aim to reduce poverty indirectly and in the long run. An essential third strand -- short-term and direct -- is to strengthen the social safety net. The best way to begin is to lower the cost of staple foods, expand labor-intensive public works (self-targeted to the poor), consolidate existing safety net-type programs, and target special programs to vulnerable groups (Box 5).

Box 5: Strengthening The Safety Net * Lower the consumer price of maize meal, sorghum meal, sorghum, and cooking oil. Remove trade, marketing, and processing restrictions. Exempt basic foods from the general sales tax. * Expand labor-intensive public works. Decentralize planning and management to local communities for cost-reduction effectiveness. Pay in cash (at below market wages, to make projects self-targeting), rather than in-kind. * Provide free food supplements to vulnerable groups. - xxxii -

110. Several government agencies and NGOs carry out a range of saftey net programs in Lesotho, but these programs are often not reaching the poor in a cost-effective way. There is no framework of priorities (in terms of vulnerable households or individuals to be targeted), impact monitoring is virtually nonexistent, and the risk of duplication is high -- meaning that a single household or individual could draw benefits from more than one program while others with equal or greater needs are shut out. The consolidation of a number of existing fragmented and uncoordinated safety net-type programs can result in substantial efficiency gains.

111. Administrative targeting of safety-net programs will be difficult, however. The statistical analysis done in this report concluded that there may be no clear-cut set of household characteristics by which to identify poor households in Lesotho; poor households are not easily differentiated from nonpoor households. And the alternative -- to rely on village leaders to select program beneficiaries -- may invite nepotism or rent-seeking. Under the circumstances, the best way to target assistance to the poor may be self-targeted safety net approaches -- for example, lowering the price of goods consumed primarily by the poor, or paying below-normal wages in public works programs designed to strengthen the social safety net.

112. The safety net proposed here takes three roads to addressing the needs of special groups: first, making policy changes to lower prices for basic food commodities to indirectly help all poor households, rural and urban; second, substantially expanding labor-intensive rural public works projects that give income in- cash or kind to poor rural households with unemployed or underemployed able- bodied adults; third, giving food supplements to those with special needs (such as malnourished children and disabled adults), in particular during drought periods.

113. Lower the consumer price offourfood staples - maize meal, sorghum meal, sorghum, and cooking oil. These foods represent a major share of spending in poor households and a minor share in richer households. Lower prices would be achieved principally by abandoning trade, marketing, and processing restrictions (which tend to raise the prices of these products). Consumers suffer a per capita loss of M2 1 a year -- or about Ml 26 for an average household of six" -- because of trade restrictions that benefit only Lesotho's inefficiently operated industrial mills. In effect, consumers are taxed to benefit producers and the industrial mills."9 Another M26 per capita -- or about M156 per household -- would be saved if South Africa would remove its international trade restrictions, and stop protecting inefficient domestic producers of maize. And because poor households spend proportionately more on food than better-off households do, current policies are unambiguously regressive. The poorest 20 percent of the population spends roughly two-thirds of its income on food -- especially maize meal -- while the top 20 percent spends only 30 to 35 percent of its income on a much more varied diet, spending only a fraction of its food budget on maize). Lowering consumer prices for maize would also benefit labor-intensive industry, by reducing the pressure on wage bills.

Is Theper capita loss is calculatedassuming a population of about1.7 million for 1992/93.The total loss each year is an estimatedM35. I million,even with conservativeassumptions about costs. 19To the extentthat maize-consuminghouseholds also produce maize, consumption losses are partiallyoffset by productiongains (estimated at aboutMl9 millionin 1992/93).But the reduction in lossesis relativelysmall. Domestic production accountedfor lessthan 15percent of consumptionin 1992/93. - xxxiii -

114. Moreover, the Government could consider returning to the pre-1991 exemption of less refined food staples from the general sales tax. Poorer households tend to consume coarser qualities of meal, so better targeting might be possible by exempting from taxes only unsifted maize meal produced by industrial mills and all meal produced by village hammer mills, while retaining the tax on the more refined meals consumed by the better-off. This option would be feasible if the contemplated shift to a value-added tax system is implemented. Note, however, that Lesotho's value-added tax system will have to be in line with South Africa's to prevent smuggling. Because South Africa exempts food staples from the value-added tax, Lesotho would need to do the same anyway.

115. Slhift from free food distribution schemes toward a substantially expanded labor-intensive public works initiative. Targeted to the able-bodied but un- or underemployed poor, this initiative would build on existing food-for-work schemes, the labor-intensive rural roads program, and recent drought emergency relief efforts. It would employ a self-targeting mechanism (below-normal wages) as well as some geographic targeting to reach the poor; it would probably emphasize cash rather than in- kind wages; and it would focus on projects (such as small-scale irrigation, feeder roads. foot bridges, anti-erosion works, and afforestation) with high labor-to-capital ratios as well as high economic returns. Institutional constraints would be relieved partly by merging disparate government units, partly by establishing mechanisms to effect collaboration between line ministries, but mainly by involving NGOs, private contractors, and local communities in all aspects of program design and implementation. If adopted, this initiative could be financed by reallocating ineffective safety-net funds as well as through the Lesotho Highlands Water Project Development Fund.

116. Providefood supplements to vulnerable groups. The third element of the proposed safety net program is to replace the current school feeding program with a community-based supplementary feeding program targeted to the most vulnerable groups in Basotho society: malnourished pre-school children, pregnant and lactating mothers, and individuals unable to work because of severe mental or physical disability, chronic illness, or old age. This program could build on the recent drought emergency initiative, under which local communities (often the school feeding committee) weighed children under five and provided meals to those identified as malnourished. High rates of malnutrition prevail among Basotho pre-school children, which can seriously limit mental development, learning, and ability to work.

117. Objective indicators would be used to identify beneficiaries (for example, low weight for height) with health care facilities or village health workers certifying eligibility. If adopted, this vulnerable-groups supplementary feeding program could replace the poorly targeted school feeding program. Financing could come principally from money formerly used to fund the phased-out or better- targeted school feeding program. Logistical arrangements would rely heavily on NGO and community involvement, as school feeding now does. Lesotho could also investigate the feasibility of a food stamp program. Food stamps are like quasi-money, which beneficiaries can exchange for approved food items in any food store. The Government would guarantee payment of the cash value to the food supplier. We recommend trying this scheme on a pilot basis, with careful evaluation of the pilots before a final decision on whether to proceed nationwide.

118. Providing food to vulnerable programs is of life-saving importance during drought and other emergencies. Because Lesotho is so susceptible to climatic and man-made shocks, such as retrenchments in the employment of mine workers, the Government will need to build the capacity to respond quickly and effectively to sudden adversity. Lesotho's most recent experience with drought was ameliorated by the Government's productive collaboration with the NGO community in delivering drought relief. Rather than rely on ad hoc collaboration, the Lesotho Council of NGOs and the - xxxiv -

Government should formalize their recent working relationship so a clear institutional mechanism is in constant readiness.

G. STRENGTHENING INSTITUTIONAL CAPACITY

119. Making a sustained impact on poverty in Lesotho will require coordinated action in many sectors plus an integrated approach to special assistance programs aimed at reducing poverty. It is necessary to set clear povertv goals and targets, reorient policies and programs in many sectors, expand some programs and eliminate others, consolidate efforts where there is duplication and fragmentation, and continuously monitor and evaluate how development programs are effecting poverty. Moreover, all sectors must try harder to reach rural-dwellers, because the poorest members of Basotho society often times live in the country-side; often in scattered, remote communities.

120. Strengthening institutional capacity is crucial to poverty reduction. The resources now devoted to poverty reduction would be far more effective if there were less program fragmentation and duplication and better coordination of public and NGO resources. NGO donors and central and local governments have sometimes succeeded in cooperating on poverty-reduction programs, but usually in programs giving direct income transfers to households. Coordination of broader development programs has been much more limited.

121. Strengthening government's capacity to design and implement poverty reduction strategies would involve reforming the civil service, decentralizing authority, privatizing activities better suited to the private sector, improving planning and budgeting, merging fragmented program units, developing a transparent poverty reduction program from the Development Fund, and doing annual household surveys. (See Box 6)

Box 6: Policy Recommendations for Improving Institutional Capacity * Reform the civil service. * Decentralize political, fiscal, and administrative authority. * Privatize activities better performed by the private sector. * Improve planning and budgeting. * Merge fragmented program management units. * Design and implement a poverty reduction program of investments financed from the Development Fund. * Start doing annual household surveys.

Reform the Civil Service

122. A decline in civil service performance has contributed to Lesotho's current crisis in governance. In the near future, with the help of the UNDP, the Government will implement the civil service reform Lesotho urgently needs. Civil service reform will require making some politically difficult choices. First, the Government must take a hard look at which activities to undertake and which to shed. Many services currently provided by the public sector -- tractor services, for example, and the maintenance and cleaning of government buildings -- could be privatized with no net reduction in employment in the - xxxv - economy.If the Governmentwants a high-quality civil servicethat undertakesonly those activities that the government should be doing (for example, ensuring law and order, delivering social services, providing public infrastructure,and regulatingeconomic institutions), it must ensure its civil servants both good pay and rewarding careers.And that meansreducing the size of the civil service to avoid a fiscal crisis. A civil service reform programthat postponesthese difficult choiceswill fail even before it begins.

Decentralize Authority

123. Decentralizing political, fiscal, and administrative authority is the key to reducing the cost of, and henceexpanding, the delivery of infrastructure,utilities, and social services.Decentralized planning and executioncan yield substantialefficiency gains becausecoordination and information cost less. At the heartof rural development,for example,is a core of rather simple, small investmentprojects, such as the construction of rural roads and bridges, the digging of wells, the developmentof small-scale irrigation, and the building and repair of schoolsand maternities.Planning and executingsuch activities locally does not usually involve complex and costly coordination or information gathering. But when you coordinate and implement simple activities through a faraway central bureaucracy or project authority, rural developmentdoes becomeboth complex and costly.

124. To be effective, decentralization must proceed along all three paths: political, fiscal, and administrative. The key is public accountability, which is why leaders should be chosen locally, finances should be controlled locally, and administration should be the responsibility of local authorities. Assigning clear responsibilitiesto a particular governmentunit and a specific service is essential to successful decentralization. If local government does not have clear financial and managerialdiscretion, it cannotbe held accountablefor either successand failure.

125. Local election of councils, committees,and mayorscan greatly improve political accountability but other measuresare also needed,especially in Lesotho,which hasjust emergedfrom a long period of authoritarian rule characterizedby weak local institutions. Such supplementarymeasures typically include regular consultationwith constituents,strong monitoring of user satisfactionwith local services, and the linking of civil servants'career paths to their responsivenessto constituents.

126. To be held accountable,local leadersmust have fiscal and administrative control. This means control over revenues,which meansthere must be adequatelocal finance laws, covering budgeting, financial reporting (accountsand audits),taxation, contracting, and dispute settlement.Most important, fiscal accountabilityrequires being able to tax and to chargeuser fees.The key to making decentralized governmentmore sustainableis to allow more local revenueraising, which includes getting those who use public servicesthe most to pay the most for them. Transfersfrom the central governmentshould be widely publicized, for transparency.In Lesothothis means,among other things, widely publicizing the existenceand operating guidelinesof the DevelopmentFund (in which LHWP revenuesaccumulate) and all allocations madefrom it.

127. In Lesotho,decentralization may initially be hamperedby a shortageof the skills and technical expertise neededfor even simple village projects. Here, NGOs and associationsof private firms can help enormously,both to easethe transition from centralto local rule, and to help make the delivery of services more efficient. But local stakeholdersshould ultimately be responsiblefor deciding which particular NGO or firm to hire to help with project planning or execution, or whether to deliver a specific service. Local capabilitieswill not grow if mandatesand resourcesare simply transferedfrom the central bureaucracyto NGOs and private firms that bypass local stakeholders.Rather than fund - xxxvi -

NGOs directly, central government could allocate resources to the local administration and let it decide how it wants services to be delivered-- with what mix of line ministries, NGOs, and private firms.

128. Political, fiscal, and administrative decentralization can quickly and dramatically increase the number of successfully executed development projects. The key to success, based on international experience, is to stress the importance of all local stakeholders participating directly. Three principles are important to remember. First, beneficiaries should be involved directly, not through intermediaries who claim to represent the community. (Consulting with local officials and NGOs is not a substitute for the direct participation of beneficiaries through something like a regularly held town or village meeting). Second, early consensus should be sought on project proposals, so costs are kept low and project design is more appropriate in terms of technological complexity and labor intensity. Third, cash or in-kind contributions should be mobilized up front. This not only reduces costs to the government, but increases local commitment, ensures the true setting of priorities, and increases the likelihood of the community's subsequent maintenance of the project.

Privatize Activities Better Suited to the Private Sector

129. The civil service could be reduced considerably and made more efficient through privatization. Viewing itself as a necessary substitute for what it views as the weak private sector, the Government continues to involve itself in productive and trading activities. Ironically, loss-making parastatals are staffed by relatively well-paid civil servants who are unequipped to run companies, while the delivery of social services is declining because civil servants working in those sectors are underpaid and have no career prospects. The Government should stop trying to substitute for the private sector, and the people running crucial public services should be paid adequately. Careers in public health and education should be well-rewarded, attractive career options.

130. The privatization of parastatals and state companies could give the Government the resources it needs to focus aggressively on activities that address core unemployment and poverty problems. Because privatization is a politically sensitive issue -- mainly for fear of foreigners taking control of previously state-owned companies -- it is tempting to proceed at a snail's pace on privatization, but that would be both costly and counterproductive. Experience elsewhere dramatizes the many pitfalls of a gradualist approach to privatization and the greatly increasing cost to Government. The Govemment's Privatization and Private Sector Development Program, funded by the World Bank, adopted a process (rather than blueprint) approach, which should make possible the design of a privatization strategy that will suit conditions in Lesotho, one that should be able to achieve quick results.

Improve Planningand Budgeting

131. In Lesotho, annual planning and budgeting often merely replicates historical budget allocations. There seems to be neither an adjustment to new priorities or a phasing out of old ones. A rudimentary start was made with public investment planning by incorporating data on development aid flows from donor agencies into budget numbers under the Public Sector Investment Program (PSIP). But to date no efforts have been made to scrutinize the PSIP and to check for consistency with the Government's stated policy objectives. The Ministries of Economic Planning and Finance should take the lead in establishing a truly planned budget process; one that is reflected in the PSIP. Regular consultations with the Ministry of Employment ensure that the goal of increasing employment is considered in all project design and evaluation. - xxxvii -

Merge Fragmented Program Units and Consolidate Safety Net Programs

132. One problem with Lesotho's safety net programs, administered by several government agencies and NGOs, is that initiatives are highly fragmented. The most pernicious effect of multiple, uncoordinated programs may be their effect on Lesotho's institutional capacity to provide assistance. As various initiatives to help the poor were launched, it was apparently more convenient (and more attractive to external funders) for the Government to establish a new office or department in a ministry to supervise the program than to strengthen an existing government structure. As a result, there is a plethora of program offices, often with roughly similar mandates, each staffed by one or two national professional staff and varying numbers of expatriates. When expatriate staff are withdrawn, usually upon completion of donor projects, the weakened units are left behind, each with less than a critical mass of Basotho staff to implement the mandated program effectively. To transfer resources to the vulnerable groups more efficiently, and to ensure the sustainability of safety net programs, it is important to consolidate units, and the programs they manage, into larger entities geared to reaching clearly specified target populations. Some initiatives might be replaced and some might be expanded.

133. As mentioned earlier, the Government could establish a more consolidated safety net system targeted to the poorest individuals and households and focusing on: (I) Self-targeting mechanisms (where feasible) other than administrative targeting, because self-targeting costs less and is better at reaching the poor; (2) The need to consolidate disparate units to strengthen national capacity for program management; (3) Strengthening existing delivery mechanisms rather than creating new ones so to avoid straining Lesotho's already weak institutional capacity; (4) Developing partnerships with NGOs and the private sector; and (5) Making the best use of limited budgetary resources.

134. Duplication and lack of coordination among complementary operations characterizes other government and parastatal bureaucracies, too. The Ministry of Works has two units responsible for paved roads and intermediate roads, for example, while the Civil Works Section of the Ministry of Home Affairs is responsible for unpaved roads. These three sections should be merged into one roads department. Similarly with soil and water conservation: units responsible for donga reclamation, conservation works, and village water supply, which operate in different ministries, could cut costs substantially by merging. Water and sewerage are provided through three organizations: the Ministry of Natural Resources, the Water and Sanitation Authority, and the Ministry of Home Affairs. Reducing such fragmentation should increase increase cost-effectiveness.

135. Also problematic is the inadequate, fragmented attention paid to small-scale local entrepreneurs. LNDC treats foreign investors to a one-stop investment center but local entrepreneurs have to battle with various bureaucracies, with no support from a strong advocacy institution. A merger between LNDC, Basotho Enterprises Development Corporation (BEDCO), and the Business Advisory and Promotion Service (BAPS) should be considered.

Design and Implement a Transparent Poverty Reduction Program of Investments Financed from the Development Fund

136. The Development Fund collects revenues attributable to the Lesotho Highland Water Project. A transparent, decentralized mechanism to use part of these funds for poverty-reducing investments such as rural infrastructure would improve the LHWP's status. The pilot community survey program undertaken with UNICEF's assistance, seems a promising example to replicate nationally. From these community surveys, done by villagers themselves, can emerge priority investment activities that could be funded from the Development Fund. Currently, the mechanism through which village-level projects - xxxviii -

Start Doing Annual Household Surveys

137. Too much time and effort is wasted debating who the poor are and whether or not they are hurt by policy changes such as those advocated under structural adjustment. Computer technology reduces the financial and manpower demands of representative household surveys to a fraction of their former levels. These surveys should be conducted annually, with sections on income, spending, assets, and other indicators of social welfare, perhaps supplemented each year with a section with a special emphasis for that year (such as nutrition, gender issues, land tenure, and local government).

138 To stimulate applied research, the data should be put immediately into the public domain. And the Government could contract for a private institution (a consultancy firm or the National University) to annually analyze the state of the economy, especially in terms of poverty and employment. The survey analysis could serve as an independent assessment of how the Government is doing in important areas.

H. AN ACTION PLAN

139. The various policy recommendations discussed above will need to be prioritized. A first attempt is made to differentiate between policy actions which are likely to have a high or medium impact on poverty reduction, and between policy actions which are feasible to implement in the short or medium term (Table 3). The poverty impact of a particular policy action derives from the poverty profile. The feasibility derives from a rough assessment of implementation constraints and incremental fiscal costs.

Table 3. Prioritizing the Policy Actions Poverty reduction Feasibility Short term Mediumterm

High impact Polky I Policy 2 Medium impact Policy 2 Policy3

140. The resulting proritized summary matrix of policy recommendations made in this report is reprinted below (Box 7). It is suggested that this matrix can form an input into a more wider and inclusive process of public debate aimed at designing a concrete set of programs to reduce poverty in Lesotho. - xxxix -

Box 7: A Prioritized Action Plan

FOSTER LABOR-INTENSIVE GROWTH

In Agriculture: * Remove p.ice, marketing, and processing distortions. * Invest in rural infrastructure that supports intensified agriculture and the production of export crops. • Make land tenure a'andwomen 's rights to land) more secure and rental markets more flexible. * Promotea reductionin herdsize. * Reorientagricultural research and extension.

In the Small Business Sector: * Make it easier for small entrepreneurs (especially women) to get credit and do business. * Shifl the eniphasis awavfrom br oad entrepreneurship training in manufacturing toward training in skills needed in the export and service sectors. * Develop rural and urbanmarketplaces.

In Export-Oriented Light Manufacturing: * Keep minimum wages at a competitive level. * Improve industrial relations. * Reduceutility costsby increasingcompetition, regulation, and decentralization in the utility sector. * Rootexport firns morefirmly in theeconomy by sellingthem factoryshells and leases and by supportingskills training thatpromotes the useof local skilled workersin the export industry.

In Tourism: * Work much more closely with South African travel agents. * Developa special"Lesotho package."

Through Infrastructure Investments: * Use private contractors on public works. * Invest more using labor-intensive approaches, but decentralize implementation. - xl -

INVEST IN HUMAN RESOURCES

In Health Care: * Exempt children under five, the elderly, and the disabled from user fees. * Introduce a free or low-fee package of basic health services. * Increase geographic equity. * Allow local facilities to retain certain fees so people feel they have local control over services.

In Education: * Lower the cost to parents of primary schooling. * Strengthen nonformal education for out-of-school youth. * Improve vocational and technical education. * Invest more in education in poorer regions. * Create incentives for qualified teachers to teach in remote areas. * Spend proportionately more on primary education and proportionately less on college and university education.

STRENGTHEN THE SAFETY NET

* Lower the cost of staple foods. -Remove trade, marketing, and processing restrictions. -Exempt basic foods from the general sales tax. * Expand labor-intensive public works. -Decentralize planning and management to local communities for cost-reduction effectiveness. -Pay in cash (at below market wages, to make projects self-targeting), rather than in kind. * Target special programs (food-stamp-typeprograms, for example) to vulnerable groups, including an improved drought reliefprogram.

IMPROVE INSTITUTIONAL CAPACITY

* Design and implement a poverty reduction program of investments, financed from the Development Fund of the Lesotho Highlands Water Project. * Decentralize political, fiscal, and administrative authority. * Reform the civil service. * Privatize activities better performed by the private sector. * Improve planning and budgeting. * Merge fragmented program management units. * Start doing annual household surveys. 1. INTRODUCTION

A. COUNTRYCONTEXT

1.1 The Kingdom of Lesotho is a mountainous, landlocked country surrounded by South Africa, which is 150 times as large. More than 85 percent of the populace lives in rural areas, engaging mainly in informal economic activities. Agriculture is cited by a majority of rural inhabitants as its main occupation, but only accounts for a small share of rural household income. The Basotho who live in urban areas work mainly in the public sector, in industry. and in services.

1.2 About 80 percent of the populace lives in the lowlands and foothills that comprise 30 percent of the land area and contain most of Lesotho's scarce arable, productive land. Land in the highlands and the Senqu River Valley, which is rapidly eroding, is suitable only for grazing and for low population densities. All over the country, rainfall is sporadic and unreliable, and drought and hailstorms often wipe out entire crops. Winters are severe, with cold winds and snow flurries in the lowlands and heavy snowstorms in the mountains. Only 9 percent of Lesotho's land is arable, and mineral resources are few and difficult to exploit. The country does posses an abundance of scenic beauty -- a major asset for tourism -- and an exportable resource in the form of water, which is now being developed for sale to South Africa through the Lesotho Highlands Water Project (LHWP).

1.3 Lesotho is often described as an economic "province" of South Africa, from where 95 percent of its imports come and to which 40 percent of its exports go. Lesotho is part of the Rand zone; its Loti is tied to the South African Rand at a fixed exchange rate of 1:1. It is a member of the South Africa Customs Union (SACU), so South Africa collects its trade taxes, which accounted for half of Lesotho's budget revenues (excluding grants) between 1988 and 1993. Linked currencies and a relatively open border mean that prices of goods in Lesotho are effectively determined in South Africa. Most important is Lesotho's export of human capital to South Africa, hence, its reputation as a labor pool for South Africa's mines. The 1986 census found that nearly half of Lesotho's adult male workers were employed in South Africa. In the 1980s, remittances from Basotho laborers working in South African mines accounted for about half of the country's gross national product (GNP), and equaled 100 percent of its (GDP).20 Today, about 40 percent of the Basotho male labor force is employed in South Africa, and remittances account for a third of GNP.

1.4 Lesotho's emergence as a labor reserve for South Africa closely followed the general pattern of the creation of so-called native reserves in the Southem African region. In the mid- nineteenth century European settler regimes effectively lowered the potential profits in a free African farming sector to force peasants into farm or mine labor. European settlers seized most of Lesotho's arable land, confining independent Basotho farming to a remote and mountainous area that was to become the independent state of Lesotho, and adopting protectionist measures against agricultural exports from Lesotho.

20 In the caseof Lesotho,which relies so heavilyon remittancesfrom Basothoworking in SouthAfrica, the differencebetween GDP and GNP is important.The GNP estimateadds net factor payments(mainly migrant mineworkersremittances) to theGDP estimate. In the lastfew years, GDP constituted about 60 percentof GNP. 1.5 During the twentieth century, Lesotho'sdependence on the South African economy cast it in the role of exporter of labor and exposedit to the long-term consequencesof South Africa's econonic policies. The failure of the inward-looking and capital-intensive South African economy to use its plentiful supply of labor efficiently has meant a general decline in employment opportunities, with negative consequencesfor Lesotho. The high degree of protection granted to South Africa's parastatal and private white-owned industry against ilnterinationalanid domestic competition stifled the development in Lesotho of a diversified productive basegeared to supplying South African markets.The inefficiencies associatedwvith the overly capital-intensiveand uncompetitiveproduction of basic consumergoods and services in SouthAfrica hurt Basothoconsumers in generaland the poor in particular.And the mine labor recruitmentpolicies followed in SouthAfrica have had devastatingeffects on family structure in the recruitmentareas, including Lesotho.

1.6 Not surprisingly, Lesotho's economy is highly vulnerable to changes in the external environment. In the mid-1980s, structural declines in external sourcesof revenue (Southern Africa Customs Union (SACU) receipts and remittances from mining workers) led to an economic downturn and aggravated domestic imbalances: increasing unemployment, deteriorating social indicators, and a growing fiscal deficit. An imprudent rise in military expendituresand growing lossesin an inefficient parastatalsector causeda further decline in fiscal balances.In 1988,the Government,assisted by the IMF and the World Bank, embarkedon a structural adjustment program aimed at reversing the downturn. Lesotho's membership in SACU and the Rand zone gave it little flexibility in trade and monetary policies. Hence, the adjustmentprogram implementedby the Governmentaimed to:

* Reducethe fiscal deficit through reducedspending, measures to increaserevenues, and betterbudget control and monitoring;

* create a seriesof fiscal surplusesto smoothout the high inflows of SACU import revenuesdue to peakconstruction activities of LHWP, and avoid "Dutch disease";

* encourage the private sector by changing regulations and incentives and strengtheninginstitutions that promote private sector development,including the attractionof foreign investment;and

* increasesocial spending (especially on health careand education),to begin reversing the decline in social indicatorsthat had recently becomeapparent.

1.7 The adjustmentprogram triggered an almost "classic" sequenceof changesin the main macroeconomic variables (Figure 1.1). The pattern of growth of GDP and GNP was characterizedby a contractionaryperiod from the start of the program in 1988until 1992. This contraction was amplified by two years of drought in 1991and 1992and declining mineworkers remittances. However, the economy turned around after 1992, when economic growth resumed.2' In retrospect, the fiscal balance was restoredquite rapidly in the first two years of the

21 From1988 to 1994,the annual growth rates of realGDP and GNP averaged 8.5 percentand 6 percent. Thefiscal deficit steadily improved from 9 percentof GNPin 1988/89to a surplusof 2 percentof GNPin 1992/93; thebudget remains in surplusup to now. - 3 - program, and continued to improve during the rest of the period, until a small fiscal surplus was attained. This fiscal surplus will gradually be brought down. It is temporaryand caused by peak activity in the construction of the Lesotiho Highlands Water Project and the associated increase in revenues from import duties.

1.8 The Government's commitment to macroeconomic stability improved the climate for investors. Lesotho attracted more foreign and local investment and its exports made a strong recovery. Exports are currently increasing by more than 15 percent a year, driving the improvement in the current account balance.22 The sharp increase in exports came mainly from direct foreign investment in textiles, garnents, footwear and light manufacturing.

Figure 1.1: GDP and GNP Growth and Overall Fiscal Surplus/Deficit as a Percent of GNP

20 15 * jAv * ----- GDP Growth 10

O6. / _ *NP G Growth

-5- -. ~ ~ ~ ~ ~ ~ ~ ~ - -Rt

-10 Surplus in % 1988 1989 1990 1991 1992 1993 1994 GNP

B. How STRUCTURALADJUSTMENT, DROUGHT AND OTHER FACTORs AFFECTED POVERTY

1.9 Despite the favorable macroeconomic environment achieved under the adjustment program, the incidence of poverty in Lesotho is still high and some observers contend that it may be increasing. Several factors are believed to contribute to the problem. Using a Semi Input- Output (SIO) model (described in Annex I), we analyzed the effects on household income of the main domestic and external shocks in Lesotho since 1987. We estimated how household incomes in Lesotho have been affected by:

* Reduced fiscal spending under structural adjustment, begun in 1988; * the severe drought in 1991 and 1992; * declines in workers' remittances from South Africa; and * increased construction activities associated with LHWP.

1.10 Model simulations based on trends in selected sectoral aggregates (Table 1.1) show that structural adjustment, which was mainly associated with a tightening of fiscal policy, has had a relatively limited impact on household incomes in Lesotho, and seems not to have

22Foreign reservesrose fromUSS41.5 million in 1988/89to US$378.6million in 1994/95. - 4 - disproportionately affected the poor. Of the four factors analyzed, the drought had the worst impact on the incomes of the poor, and the decline in workers' remittances the least impact. The decline in workers' remittances mostly hurt the nonpoor. Construction activities under LHWP and growth in labor-intensive industry -- also part of the Government's structural adjustment strategy -- appear to have increased household incomes, largely offsetting the adverse shocks of drought, reduced migrant earnings, and fiscal conservatism.

Table 1.1: Trends in Selected Sectoral Aggregates, 1983-92 (Maloti, Millions) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Agriculture: Field Crops, Vegetables, Fruits Current prices 27.5 32.1 41.2 43.4 52.5 86.6 82.8 126.2 69.1 26.1 Constantprices 17.8 19.4 22.8 22.2 22.0 34.4 30.7 39.2 18.6 8.8

Agriculture: Livestock Currentprices 39.6 50.3 47.9 57.1 58.6 103.7 131.9 113.3 96.4 122.2 Constantprices 24.5 28.4 24.0 29.5 24.7 27.3 35.7 28.8 30.1 25.9

Factor Income from Abroad, Net (Remittances) Current prices 423.1 487.2 514.3 583.3 705.9 829.7 927.8 1101.1 1238.8 1266.4 Constantprices 285.1 289.1 265.8 254.8 271.7 285.3 264.9 266.1 258.5 224.9

Government Expenditure Current prices 91.7 98.6 135.9 159.3 171.5 215.8 222.2 234.6 314.2 380.5 Constant prices 76.7 79.7 84.5 93.2 94.5 97.0 91.3 92.7 98.5 103.8

Construction Currentprices 39.8 50.8 78.3 71.5 98.5 134.3 214.2 291.7 352.0 422.4 Constantprices 24.9 29.3 38.4 30.3 36.0 41.5 63.2 76.2 78.8 85.7

Gross Fixed Capital Formation: Building and Construction Current prices 84.7 109.9 171.2 158.1 215.7 317.2 634.2 957.7 1089.2 1359.6 Constantprices 52.7 62.0 84.8 64.6 77.3 102.2 180.8 239.6 239.6 262.2

Source:Lesotho Bureau of Statisticscomputer files.

1.11 Reduced government spending. Model simulations show poor and nonpoor households suffering roughly equal declines in income under structural adjustment (Table 1.2). Under adjustment policies aimed at reducing the deficit, real government spending was cut by 5.8 percent in 1989. Since then, it has risen 13.7 percent, but remains below pre-adjustment levels.23 The cut in government spending reduced real household incomes an average 1.58 percent. The brunt of the impact was felt by urban households, whose incomes fell an average 3.13 to 5.25 percent. Rural households, relatively unaffected, suffered declines in income of less than I percent.

23 Realgovernment spcnding rose 26.4 percentbetween 1983 and 1988,or an average4.8 percenta year. Theactual level of governmentspending in 1992was 13.1 percent below the earliertrend. 1.12 Drought. The poor were especially hard hit by the drought; their real incomes declined 5.4 percent from 1987 levels and probably even more from 1990 levels.24 In rural households with no income from worker remittances -- most poor households-- incomes declined nearly 8 percent (Table 1.2). The drought affected overall average household income only slightly more(- 1.85) than the cut in government spending -- mainly becausecrop production accountsfor less than 20 percent of GDP.

1.13 Reduced migrant incomes. Remittances from migrant workers fell 17.2percent in real terms between 1987 and 1992, sharply reducing national income. But this decline affected mostly nonpoor households,whose incomes fell an estimated 11.19 percent (Table 1.2). Simulations show incomes in rural householdswith remittance earnings falling dramatically, nearly 16 percent. The decline is smaller in urbanhouseholds, perhaps because other sourcesof income make those householdsless dependenton remittances. In rural householdswith no remittances,incomes decline lessthan 1.5percent.

1.14 Lesotho Highlands Water Project. Simulations show the combined effect of cuts in governmentspending, the drought, and declines in migrant remittancesto reduce income 12.75 percent for the poor and 14.24 percentfor the nonpoor. Fortunately,this is partially offset by constructionassociated with LHWP. Value addedgenerated in the constructionsector increased 138 percent; gross fixed capital formation rose by a similar proportion. The model simulation, which assumesthe sameproportion of domesticlabor as in 1985,shows that the project helped offset the income decline of the poor (from 12.75 to only 2.07 percent), probably through increaseddemand for labor, especially from householdswith no migrant remittancesor formal wagejobs. But if, as many observershave alleged,construction for the LesothoHighland Water Project has increased demand for non-Basothoworkers or for substantially more capital- intensive technologiesthan in 1985,the project may not have benefitedthe poor as much as the model simulation estimates.

1.15 The model did not simulate the rapid growth of some labor-intensiveindustries. Textiles and footwear, among other industries,grew an average 20 percent a year between 1986 and 1992, albeit from a tiny base. This growth createdan estimated8,000 new jobs between 1986 and 1992,which obviously improvedaverage household incomes. Poor and nonpoorhouseholds probably benefitedabout equally from theseincreases in output (Table 3.1).

24 Thedrought in 1991-92reduced the real output of fieldcrops 60 percent from 1987 levels -- evenmore fromlevels in 1990,when production was unusually high. By contrast,1992 livestock production was almost identicalto 1987.In simulatingthe drought's effect on income, therefore, only the 60 percent decline in fieldcrop productionwas modeled. - 6 -

Table 1.2: How Four Factors Affect Household Incomes in Lesotho (percentage change) Reduced Drought Reduced Subtotal LHWP Total Government Migrant Spending Incomes

Urban Male HH w/ migrant income - 5.78 - 1.06 - 3.26 - 10.10 5.90 - 4.20 w/o migrant income - 4.74 - 1.59 - 1.56 - 7.88 9.56 1.67 FemaleHH w/ migrant income - 5.32 - 1.19 - 3.13 - 9.63 6.82 - 2.81 w/o migrant income -2.09 - 2.18 - 1.64 - 5.91 13.91 8.00 Rural Lowlands Male HH w/ migrant income - 0.21 - 0.53 - 15.81 - 16.55 1.06 - 15.49 w/o migrant income - 2.42 - 7.66 - 1.27 - 11.35 14.53 3.18 FemaleHH w/ migrant income - 0.21 - 0.54 - 15.62 - 16.37 1.22 - 15.15 w/o migrant income - 2.07 - 6.93 - 1.38 - 10.38 15.31 4.93 Rural Mountains Male HH w/ migrant income -0.21 -0.54 - 15.66 -16.42 1.18 - 15.23 w/o migrant income - 2.36 - 7.54 - 1.29 - 11.20 14.66 3.46 FemaleHH w/ migrant income - 0.21 - 0.56 - 15.49 - 16.26 1.32 -14.94 w/o migrant income - 2.38 - 7.83 - 1.40 - 11.61 15.72 4.11

Urban Male HH - 5.25 - 1.33 - 2.39 - 8.96 7.77 - 1.19 FemaleHH - 3.13 - 1.86 -2.12 - 7.11 11.64 4.53 Rural Male HH -0.60 - 1.77 - 13.23 - 15.60 3.46 - 12.15 FemaleHH - 0.97 - 3.12 -10.15 -14.24 6.69 - 7.55

Poor - 1.69 - 5.39 - 5.67 -12.75 10.68 - 2.07 Nonpoor - 1.54 - 0.79 - 11.90 -14.24 3.23 - 11.00 Total - 1.58 - 1.85 - 10.46 - 13.89 4.95 - 8.94 Source:Results are for SIOmodel simulations with reducedurban links. Notes:I-IH = headof household;w/ = with; w/o = without.

1.16 All told, the simulation results show poverty in Lesotho as having changed little since 1987. But simulations and household surveys focus on household averages. And averages conceal catastrophic changes experienced by individual households, which may have lost an entire crop or a job and may thus have fallen far more in income. Knowledgeable observers of Lesotho's economy and participants in the qualitative studies have said they feel poverty has increased in recent years, partly because they are personally acquainted with one or more households that have experienced sudden major shocks. Moreover, households that suffered such shocks, and those that gained little from new jobs in textiles and construction, may have - 7 - been particularly harmed by increasesin food costs and by efforts to improve cost recovery in health care and other social services.

1.17 Structuraladjustment and external shockshave not increasedthe generallevel of poverty in Lesotho,but that does not justify complacency. Poverty,including severepoverty, still exists at high levels. It is imperativethat coordinatedinitiatives be put in place as soon as possibleto reducethe suffering in poor Basothohouseholds. The questionis, which initiatives will be most effective?

C. OBJECTIVE AND ORGANIZATION OF THE REPORT

1.18 In 1993, the country returned to democracyafter 23 years of authoritarian rule. The democratically electedGovernment of Lesothoassigns high priority to the reduction of poverty. The aim of this report is to help that governmentdevelop a strategyand action plan for reducing poverty in an efficient and sustainable way. This report, which was produced in direct collaboration with the Governmentof Lesotho,should help in setting developmentpriorities as will be reflected in the Government'snew 3-year, rolling PerspectivePlan. (This Plan will replacethe Five Year DevelopmentPlan exercises). It should also be useful to nongovemment organizations (NGOs) and donors, many of which are reorienting their programsto address poverty reduction goals. The Lesotho Council of NGOs, the European Union (EU) and the United StatesAgency for International Development(USAID) provided invaluable help in the preparationof this report.

1.19 The report draws on existing survey information, special studies and reports, and line ministry statistics. Chapter 2 profiles the extent, severity, and effects of poverty in Lesotho today. Chapters3 to 6 analyze the extent to which Lesotho's policies and programs already provide a supportiveenvironment for reducingpoverty in the mediumto longer term and suggest areasfor change.Chapter 7 is an action plan for reform, public spending.and the developmentof institutions to reduceand monitor poverty in Lesotho. - 8 -

2. A PROFILE OF POVERTY IN LESOTHO

A. INTRODUCTION

2.1 Developing and implementing a comprehensive strategy for sustainable poverty reduction requires a systematic analysis of available information about the number and characteristics of the poor as well as a thorough review of the implications of existing policies and programs affecting the living standards of the poor. The most comprehensive recent efforts to profile poverty in Lesotho were two poverty mapping exercises (in 1991 and 1993) undertaken by a local consulting firm, Sechaba Consultants, of Maseru. These exercises were commissioned by the Drought Relief Implementation Group (DRIG) and donors25 to identify areas of the country with high concentrations of poor households as a means of targeting drought relief assistance. Both reports provide a number of valuable insights into the nature and magnitude of poverty in Lesotho, but given their primary objective of guiding drought relief efforts they include only limited analysis of the relationship between poverty and broader socioeconomic policies and programs.

2.2 This chapter presents a reasonably detailed and comprehensive account of the extent and severity of poverty in Lesotho today. The poverty profile analyzes the economic activities, expenditure patterns, nutrition status, and assets of poor households; assesses the degree to which these households are covered by and use basic social services; and identifies the main obstacles the poor encounter in their efforts to improve their welfare.

2.3 Certain patterns emerge from the poverty profile presented in this chapter. For one, absolute levels of poverty in Lesotho are lower than in many sub-Saharan African countries: Only a fraction of Basotho households suffer extreme poverty or absolute deprivation (for example, chronic hunger) of the type seen fairly commonly in such countries as Mozambique and Tanzania. Still, relative poverty levels in Lesotho are high. Nearly half of all Basotho households are poor (with consumption levels below half of average) and more than a quarter are ultra-poor (with consumption levels below 25 percent of average). A severely poor fifth of the population spends nearly two-thirds of its resources just on food and another 25 percent or so on such necessities as water, fuel, health and education.

2.4 Poverty is disproportionately high in the nearly 40 percent of Basotho households that rely on farming, herding, informal business, or casual labor for most of their income, as well as among the roughly 7 percent of people living in households headed by unemployed adults. By contrast, the incidence of poverty is below half the national average in households relying mainly on mining remittances from South Africa. It is also lower than average in households headed by regular wage earners working in Lesotho or by retirees receiving formal pensions.

2.5 Poverty is greater in rural than in urban areas. Smaller landholdings are associated with poverty, but the relation between farm size and poverty is not strong and Basotho without any agricultural land are among the better-off, perhaps because their income base is more diversified

25 Principally the European Union (EU), UJNDPand USAID. and includes wage jobs. The incidence of poverty is deeper and far more severe in the mountain and Senqu valley zone than in the foothills and lowlands. The mountainous and predominantly rural districts of Thaba Tseka, Mokhotlong, and Quthing are far poorer than other parts of the country, while households in Maseru, Leribe, Berea, and Mafeteng are better off than average.

2.6 Poverty is worst in the mountains. These geographic disparities are reflected in data on the fulfillment of basic needs and access to social services. Malnutrition levels among children under five are unacceptably high everywhere, but urban children are considerably better off than children in the mountain areas; children in the lowlands and foothills fall somewhere in between. One sees similar geographic patterns in indicators for mortality and morbidity, access to safe water and sanitation facilities, and school enrollment ratios. Moreover, the quality of school, and hence educational achievement, is far worse in the mountains than elsewhere.

2.7 Some types of households and individuals are more disadvantaged than others no matter where they are. Households headed by old people or by women without absent spouses are much poorer on average than other households. Especially vulnerable within households are boys six to 15 (who have much lower school-enrollment ratios than girls) and children under five (again, especially boys), who apparently suffer malnutrition severe enough to inhibit their cognitive development. Boys suffer more than girls, partly because even young rural boys are sent to help tend livestock, and partly because job expectations for Basotho boys are for unskilled work in the mines in South Africa.

2.8 Statistical analysis failed to identify a single characteristic (or manageable group of characteristics) to distinguish poor households from nonpoor households, for purposes of targeting direct assistance. Education levels are lower in poor households, for example, but they are low in many nonpoor households, too.

B. METHODOLOGY

2.9 To assess the nature and prevalence of poverty in Lesotho, as well as its distribution by region, income source, and household composition, we developed a poverty profile using various sources of data. We relied especially on two: the Bureau of Statistics' 1986/87 Household Budget Survey (the "1986 survey") and a 1993 household survey by Sechaba Consultants commissioned by DRIG (the " 1993 survey"). The 1986 survey was a comprehensive national survey of income and spending, covering a random sample of 7,680 households. The 1993 survey, also national, covered a smaller sample -- slightly over 1,700 households -- and elicited less information about expenditures, so the 1986 survey was used to estimate spending for 1993. Also used to prepare Lesotho's poverty profile were two qualitative studies of poverty (1991 and 1993) undertaken to provide a "people's perspective" on poverty, coping strategies, and the effectiveness of existing poverty-reducing programs and services. Research for these studies included participant observation, case studies, and individual and group interviews in both rural and urban areas (Annex 2).

2.10 To rank households for poverty, economists have used various indicators as proxies for welfare or standard of living -- among them, income, consumption (or expenditures), caloric intake, and access to social services, as well as such "intangible" factors as freedom or political and human rights. In this report, we use total real consumption of goods and services as a proxy - 10- for household welfare.26 Householdsare ranked by level of real consumption divided by householdsize, modified by an "adult-equivalent" index to control for differences in household composition and needs.The adult-equivalentindex is usedbecause a householdof five with two children, for example,requires less food than a householdof five with five adults(Annex 3).

2.11 To distinguish poor householdsfrom nonpooror better-off householdsrequires choosing a poverty line -- a level of welfare below which poverty is said to exist. Poverty lines are arbitrary cut-off points, so two householdsclose to the poverty line -- one above and the other one below it -- may share more characteristicswith each other than with householdsat the extremes of poverty or wealth. Lesotho has no current official poverty line, so we set two relative poverty lines for the purposesof this report. We characterizeas poor householdsthat spend less than half of the average adult equiivalent -- householdswhose membersconsume less than half of the mean consumption level. We characterize as very poor, or ultra-poor, householdsthat spend lessthan 25 percentof the averageadult equivalent."

2.12 To give the reader an idea of what these poverty lines mean in terms of meeting daily needs:A householdwith a spendinglevel at the first poverty line would probably need to spend about 51 percent of its total householdbudget to get a basic daily adult diet of about 2500 calories (roughly 500 gramsof maize meal, 60 gramsof pulses,50 grams of sugar,and 40 grams of cooking fat). A householdat the second,ultra-poverty line would need to spend its whole budgetto get the samediet and caloric intake.(For details, Annex 4.)

C. LESOTHO'S RANKINGBY IMPORTANTSOCIOECONOMIC INDICATORS

2.13 Lesotho'sper capita GNP of US$580 (in 1991)ranks it as one of the world's 40 lowest- income economies-- 93rd among 127countries. In sub-SaharanAfrica, however, it fares better, ranking ninth richest out of 32 countries.28 Indeed,Lesotho's per capita GNP is four to five times greater than some of the poorest African countries, so the absolute level of its poverty is therefore lower than in countries such as Mozambiqueand Tanzania.Lesotho is also aheadof many of its African neighborson social indicators,including some with much higher average incomes (Table 2.1). In sub-SaharanAfrica, for example, it ranks third on life expectancyand net primary school enrollment and hasthe fourth lowestmaternal mortality rate amongcountries

26 Thismeasure includes all householdexpenditures aswell as the value of foodproduced by thehousehold, transfersin- cash and kind, and goods received in barter.Current consumption is used as a proxyfor incomebecause of thepoorer quality of informationabout income, because it is abetter measure of thepresent standard of living,and becauseit is areasonably good indicator of long-termwell-being (revealing information about past and future income). Fora thorough review of issuesin povertymeasurement, seeMartin Ravallion, Poverty Comparisons," World Bank 1993.

27 Onereason to usea relativerather than an absolutepoverty line is thatconsumption in Lesothois generallyunderestimated. (There appeared to be an even greater underestimate of household income.) Use of absolute povertylines would probably lead to a seriousoverestimate of the incidence of poverty.The use of relativepoverty linesis technicallysound as there is noreason to believethat the measurement error on surveyconsumption data is nonrandom(that is, relatedto thestandard of living). Rankinghouseholds in terms of consumptionand using that rankingto establishthe cut-off between poor and nonpoor should be valid even though all householdsmay seem to havea lowerstandard of livingthan is actuallythe case. Finally, the use of twopoverty lines provides a built-in sensitivitytest for thepoverty analysis.

28 Comparativedata are available only for sub-SaharanAfrican countries with populations over I million. - I I - for which data are available. Still, Lesotho's social indicators are much lower than those of non- African countries with similar per capita GNP levels. In other words, there remains considerable scope for improving the quality of life of the Basotho people even at present income levels.

Table 2.1: Social Indicators for Countries of Comparable Income Guinea Sri Honduras_ w1- Indonesia Zimbabwe Senegal Lanka Per capita GNP 460 500 580 - 610 650 710 (US$) Life expectancy 44 71 65 56 60 60 48 (years) Infant mortality 136 18 49 - 74 48 106 Rate (per 1000) Adult illiteracyrate 76 12 27 -38 23 33 62 (%) :-:-:: Under 5years na 45 21 17 14 12 22 malnutritionrate (%) -_'_._-_'._--'.: Note:Data in thistable are for 1991and have been taken from the 1993 World Development Report (WDR) for consistencyof intemational comparisons. Data in thistable may not be consistent with data cited elsewhere in this chapterand in the text, which were drawn from more recentLesotho-specific surveys and reports.

2.14 Although many Basotho households are poor, even severely poor compared with the average living standard in Lesotho, they probably do not face the type of absolute deprivation seen elsewhere in Africa or in South Asia. Spending on food as a share of household spending is a common indicator for the prevalence and magnitude of poverty. The average Basotho household spends nearly 51 percent of income on food which is not very high by international standards. The poorest 20 percent of the population spends about 66 percent of total household budget meeting food requirements -- a level typically associated with severe, but not desperate, poverty. A comparatively high proportion of spending goes to health and education, both in the population as a whole and especially in the three lower quintiles (Figure 2.1).

Figure 2.1: Budget Shares in Basotho Households (by quintile, 1993)

08 0 / in 3 the r 8 0% 60% 1- OH ealth & 40% - Education 20 % MW ater & Fuel'

Q I Q2 Q3 Q4 Q5 IFood E x p e n d itu re Q u in tiles Q 1 = I o w e s t t o Q 5 = h i g b e s t

Source:World Bank estimates, based on the 1993Sechaba survey. - 12 -

2.15 Absolute poverty in Lesotho may be greater than average income figures suggest, because of the country's inequitable income distribution (Table 2.2). Of 40 developing countries for which data are available, only nine -- including Lesotho -- have Gini coefficients above 0.55, indicating a severely skewed income distribution.29 Five of the nine highly inequitable countries are in Latin America and four are in sub-Saharan Africa.

Table 2.2: Lesotho's Inequitable Income Distribution (% of National Income) Year Lowest Second Third Fourth Highest Gini Country 20% Quintile Quintile Quintile 20% Index

______~~~~~~~~~~~~~~~(%) Thailand 1988 6.12 9.37 13.53 20.28 50.70 43.81 Indonesia 1990 8.65 12.13 15.85 21.12 42.25 34.15 Sri Lanka 1985 4.86 8.40 12.38 18.15 56.12 51.01 Colombia 1991 3.33 7.21 12.12 20.00 55.78 51.32 Panama 1989 2.00 6.30 11.60 20.30 59.80 56.57 Jamaica 1990 5.98 9.88 14.45 21.32 48.37 41.78 1993 2.63 5.$O~3 19.74 6A4 Cote d'lvoire 1988 7.28 11.87 16.34 22.34 42.17 34.55 Zimbabwe 1990 3.98 6.29 10.01 17.38 62.34 56.83 Morocco 1990/91 6.57 10.45 14.97 21.71 46.30 39.20 Jordan 1991 6.47 10.29 14.61 20.94 47.69 40.66 1989 10.87 14.75 17.98 21.97 34.43 23.34 Note: Data for all countriesexcept Lesotho were taken from Chen,Datt and Ravallion. "Is PovertyIncreasing in the DevelopingWorld?" World Bank Working Paper 31146,June 1993. Figures for Lesotho are Bank staff estimatesbased on surveydata.

2.16 Comparisons between 1986/87 and 1993 are problematic, due to the incompatibility of the two surveys. In 1993, the richest 10 percent of Lesotho's households received 44.06 percent of national income, more than four and a half times the income share of the bottom 40 percent, which received only 8.13 percent. The distribution seemed similarly skewed in 1986/87 (Table 2.3 and Figure 2.2).

Table 2.3: Shares of Income (Consumption) by Decile (as a Percentage of National Income, 1986/87 and 1993) Lowest Decile Decile Decile Decile Decile Decile Decile Decile Highest 10% 2 3 4 5 6 7 8 9 10% 1986/87 0.92 1.89 2.77 3.77 4.88 6.29 8.21 11.23 16.68 43.40 1993 0.86 1.77 2.43 3.07 4.32 5.81 7.93 11.77 17.98 44.06 Source: World Bankestimates.

29 The Gini coefficientmeasures the degree of (in)equality,with 0 signifying absolute equality and I indicatingabsolute inequality. When multipliedby 100 to derive a Gini Index, 0 signifies perfect equality and 100 signifies perfect inequality. - 13-

Figure 2.2: The Pattern of Inequality (1986/87 and 1993)

X 0.8 . _ / Equality

0.6 - - 1

0O.2 _ _ _ a. 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Proportion of Popuiadon

Note: The further awaythe Lorenz curve is from the diagonal.the greaterthe inequality.

2.17 But from a longer term perspective, there has been a marked increase in income inequality, especially in rural areas.In 1967/69,the lowest 60 percentof householdsearned 38 percent of income while the top 20 percenttook home 40 percent.In 1972/73,the bottom half receivedabout 15 percent of income,the top 10 percent, 35 percent.(The similarity of income distribution estimates from 1986/87 and 1993 suggeststhat the trend toward increasing inequality may have begunto level off. But, as notedearlier, the two householdsurveys are not strictly compatible, and the readeris advised not to place undue emphasison any comparisons betweenthe two.) Gini coefficients,which show inequality, climbed from 0.23 in rural areas in 1967/69to 0.55 in 1993(Table 2.4).3°

Table 2.4: Lesotho's Worsening Income Inequality (Gini Coefficients, 1967-93) 1967/69 1972/73 1986/87 1993

Urban NA 0.50 0.57 0.58 Rural 0.23 NA 0.55 0.55 All NA NA 0.57 0.57 Source: Gini coefficients for 1967/69 and 1972/73are cited in 'Options for a Dependent Economy", InternationalLabour Organization, 1979.Those for 1986/87and 1993 are Bank estimates.

D. THREE POVERTYINDICES

2.18 It is important to focus not only on the numberof poor householdsbut on the depth and severity of poverty. Not all householdscategorized as poor or ultra-poor suffer the samedegree

30 Data from the 1967/69Rural Household Consumption and Expenditure Survey yield a Gini coefficientof 0.23. Datafrom the 1972/73Urban HouseholdBudget Surveyyield an estimatedGini coefficientof 0.50. - 14- of deprivation. In our analysis, we use three common measures from the Foster-Greer- Thorbecke index of poverty.

* The PO measure (or head count index or incidence of poverty) measures the percentage of individuals below a given poverty line, relative to the total population.

* The Pi index (or poverty gap index) measures the depth of poverty as the relative gap between average actual income in poor households and the poverty line.3 It measures the percentage of the poverty line income which is needed to bring everyone below the poverty line upto the poverty line.

* The P2 index is a modification of the P1 index. It weighs the shortfall of a household's income more heavily, the further it falls below the poverty line, suggesting the severity of poverty for the poorest of the poor. (Annex 6 discusses these three measures in more detail.)

2.19 Based on the two poverty lines used in this report, 49.15 percent of Lesotho's population is poor, consuming less than half of average adult-equivalent consumption, and 25.79 percent is very poor or ultra-poor, consuming less than a quarter (Table 2.5).32

Table 2.5: The Incidence, Depth, and Severity of Poverty (1986/87 and 1993) PovertyLine Ultra-PovertyLine 1986/87 1993 1986/87 1993 Incidence (P0) 46.82 49.15 22.55 25.79 Depth (P1) 22.26 24.01 9.41 9.12 Severity (P2) 13.77 14.52 5.24 4.94 Source:World Bank staff estimates based on the 1986/87and the 1993Sechaba Surveys. Notes:The PO measure (or head count index or incidence of poverty) measuresthe percentage of individualsbelow a givenpoverty line, relative to the totalpopulation. ThePl index(or poverty gap index) measuresthe depthof povertyas the relativegap between averageactual income in poorhouseholds and the povertyline. TheP2 index is a modificationof the Pi index. It weighsthe shortfallof a household'sincome moreheavily, the furtherit falls belowthe povertyline, suggesting the severityof povertyfor the poorestof thepoor.

2.20 Overall, poverty measures in Lesotho may not to have changed significantly in the past six to seven years (Table 2.5). About 49 percent of the population was poor in 1993, up 2 percent from 1986. About 26 percent of the population was ultra-poor in 1993, up 3 percent from 1986. The expenditure level of the poor was 38 percent of the mean adult equivalent in 1993, down from 39 percent in 1986/87. But spending of the ultra-poor was 23 percent in 1993, up from 16

31 The povertygap index is not sensitiveto the severityof povertyin that it givesequal weight to each incrementalLoti of householdincome below the povertyline. The sumof all individualpoverty gaps provides a nationalpoverty gap, which suggests the minimumtransfers to the poor neededjust to bring individualsup to the povertyline.

32 Averageadult consumption is estimated(from 1993survey data) to beequivalent to M 140.96per month, somembers of poorhouseholds consume less than M70.48 per adult equivalent, and members of verypoor households consumeless than M35.24 per adult equivalent. Research suggests severe underreporting of consumption,however, so our anslysisrelies more on the relativeranking of households. - is - percent in 1986/87, suggesting that the relative poverty of the ultra-poor decreased slightly, a trend also reflected in the P2 measure. However, the two surveys were not strictly comparable, and the reader is advised not to place undue emphasis on these inferences.

2.21 Nearly half of Lesotho's populace is poor and slightly over a quarter is ultra-poor, but the poverty gap -- the depth of poverty -- is relatively small. In 1993, using the ultra-poverty Iine, the poverty gap was only 9.12 percent, which means that if perfect targeting were possible, a transfer of only M38.52 per person a year (or 2.3 percent of average adult-equivalent spending) would be enough to eliminate ultra-poverty. Using the other poverty line, M203.04 per person a year (or 12.0 percent of mean adult-equivalent spending) would be needed to eliminate poverty. That translates nationally to M44.2 million (US$13 million) a year at the lower poverty line and M232.9 million (US$68.5 million) at the upper poverty line.

2.22 Lesotho could afford these amounts. Official development assistance averaged US$104.7 million a year between 1990 and 1993, which, if perfectly targeted, is much more than what is needed to eliminate poverty in Lesotho. Of course, in Lesotho, as elsewhere, perfect targeting is nearly impossible. It is difficult to identify the poor, let alone design strategies that transfer exactly the amounts needed to each household. But the above estimates suggest that substantially reducing poverty in Lesotho is an achievable goal in the medium term.

E. POVERTY'S DISTRIBUTIONBy INCOME SOURCEAND OCCUPATION

2.23 Poverty's distribution by source of household income and by occupation is important from the standpoint of policy design. The principal way economic policies affect households is through the economic role their members play in society. The knowledge that many of the poor depend largely on agricultural income for their livelihoods, for example, suggests that development strategies that focus on making small-scale agriculture more productive are likely to help reduce poverty. Of course, individual shifts from lower- to higher-income sectors can and do occur and can reduce poverty in the long run, but the pace of such shifts is limited, among other things, by the growth rate of the higher-income sector and by the human resource base of the poor. In other words, short and medium-term strategies for poverty reduction must focus on raising the incomes of the poor in their present occupations and locations. Long-term strategies must focus on the fundamental, long-run comparative advantage of the various economic sectors.

Sources of Income

2.24 The principal sources of income are substantially different for the poor and the better off (Figure 2.3). In nonpoor households, income comes mainly from mine workers' remittances from South Africa (40 percent) and regular wage employment in Lesotho (24 percent). The main source of income of poor and ultra-poor households is farming and livestock-related activities (between 30 and 34 percent). The other main sources of income in poor households are wage work, mine work, informal business, gifts, and casual labor. - 16-

Figure 2.3: Main Sources of Income in BasothoHouseholds (1993) Ultra Poor Households *Wage Work 1' % 10% *Mine-RSA

*CasualLabor 14% GlFarm/Livestock

* Informal Business 10% OGifts 34% 12% 100 GOther

Poor Households * Wage Work 10% 10% *Mine-RSA

*Casual Labor 16% OFarm/Livestock

3 Informal Business o Gifts * Other 12%

Non-poor Households * Wage Work 10% 3% 5% 8% * Mine-RSA

0Casual Labor 10% _oFarm/Livestock 24% * Informal Business O Gifts * Other 40%

Source: Bank staff estimates from 1993 Sechaba Survey

2.25 The survey does not distinguish between types of wage employment. Poor households probably tend to work for lower wages, in occupations such as domestic service or farm employment, while the nonpoor tend to work in better-paid civil service or industrial jobs. That wage work and mine workers' remittances are important sources of income for the poor (as well as the better off) should not be surprising. Mine workers' remittances may be smaller when the - 17 - recipients are relatives other than the worker's spouse and children -- in which case, although the remittance may be the household's main source of income, the remittance may be too small to raise the household above the poverty line. Generally, however, the incidence of poverty (23 percent) and ultra-poverty (9 percent) is notably lower for individuals in households claiming mining remittances as the main income source than for any other main source of income (Table 2.6). Households claiming pensions and wage work also have a relatively low incidence of poverty. Participation in the formal sector (past or current) is a key factor in avoiding deprivation.

Table 2.6: Poverty's Distribution by Main Source of Income (1993) Income Source Incidence(P0) Depth (P1) Severity(P2) Population share (%) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor Reg. wage work 38.21 18.93 17.36 6.30 10.23 3.38 19.71 RSA mine work 23.39 9.30 10.34 3.33 5.89 1.65 26.88 RSA other work 54.55 26.15 26.16 11.22 16.29 6.54 3.01 Casual labor 71.24 37.83 37.92 16.98 24.37 10.54 5.92 Farm/livestock 73.94 43.47 37.71 13.97 22.87 7.42 20.00 Informnalbusiness 54.17 28.87 27.72 13.31 18.08 7.85 10.93 Pension/disability 35.64 12.24 15.29 1.18 6.90 0.15 0.84 Gifts 66.13 39.40 33.04 11.34 19.61 5.72 7.51 Other 54.44 32.50 28.09 11.34 17.33 5.73 5.21 All 48.47 25.85 23.88 9.18 14.51 5.00 100.00 Source:Bank staff estimates from 1993 Sechaba survey. Notes:The POmeasure (or head countindex or incidenceof poverty) measuresthe percentageof individuals belowa givenpoverty line, relative to thetotal population. The PI index(or povertygap index)measures the depthof povertyas the relativegap betweenaverage actualincome in poorhouseholds and the povertyline. The P2 index is a modificationof the PI index. It weighsthe shortfallof a household'sincome more heavily,the furtherit fallsbelow the povertyline, suggesting the severityof povertyfor the poorestof the poor.

2.26 Poverty is greatest among households whose main income source is farming or livestock production. Many of the poor (but only 20 percent of the total population) indicated farming/livestock as their main income source in 1993, a figure that might have been higher if not for two years of drought just before the survey. Moreover, a significant proportion of those listing casual labor as a main income source (6 percent of the population and about 9 percent of the poor) are likely to be employed mainly in agriculture. Not only is there a high incidence of poverty and ultra-poverty among farmers, herders, and casual laborers (more than 70 percent), but the depth of poverty is also greatest among these groups, approaching 38 percent compared with a national average of about 24 percent (Table 2.6). Similarly, the severity of poverty is 9 to 10 percent higher than the national figure of 15 percent.

2.27 Households that rely heavily on informal business and gifts for income are also disproportionately vulnerable to poverty. Over 54 percent of households engaged in informal business -- about 11 percent of Basotho households -- are poor and 29 percent are ultra-poor. Households dependent on gifts represent about 8 percent of the population but high percentages of the poor (66 percent) and ultra-poor (40 percent). - 18 -

Occupation

2.28 Most of the poor and ultra-poor live in households headed by self-employed workers or farmers -- rural dwellers for the most part (Figure 2.4). By contrast, well over half of the nonpoor live in households headed by someone who earns a regular wage or salary, usually in the formal sector or in mining.

Figure 2.4 Occupation as an Indicator of Poverty (1993)

Ultra Poor Households OSelf employed/farm 12%/ 4% 6% * Regular wage/salary

*Casual Laborer

15% _ Pensioner

* Unemployed

55% 0 Other

Poor Households o Self employed/farm

5% 10% 9% 6% U Regular wage/salary

,Casual Laborer

19% _ Pensioner * Unemployed 1% OOther

Non Poor Households O Self employed/farm 4%4° 3% 4% * Reg. wage salary _ CasualLaborer

0* Pensioner 30% 55% J * Unemployed

OOther

Source:Bank staff estimates from 1993 Sechaba Survey - 19-

Table 2.7: Distribution of Poverty by Occupation of Household Head (1993) Occupation Incidence(P0) Depth (P1) Severity(P2) Population share (%) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor Self-employed 62.27 34.92 31.48 12.70 19.46 6.84 40.69 workeror farmer Regularwageor 25.62 10.14 10.83 3.35 6.05 1.81 36.96 salaryearner Casual worker 69.20 35.27 35.17 12.81 21.27 7.77 6.11 Pensioner or 52.59 24.13 24.13 8.81 14.36 4.79 4.77 retired person Unemployed 73.60 45.42 39.61 15.90 24.78 8.84 6.57 Other 57.29 32.99 26.88 9.64 15.73 4.14 4.90 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source:Bank staff estimates based on the 1993Sechaba survey. Notes:The PO measure (or head count index or incidenceof poverty) measuresthe percentageof individuals belowa givenpoverty line, relative to thetotal population. ThePI index(or poverty gap index) measuresthe depthof povertyas the relativegap betweenaverage actual incomein poorhouseholds and the povertyline. It measuresthe percentageof the povertyline income which is neededto bringeveryone below the povertyline upto the povertyline. TheP2 index is a modificationof the PI index.It weighsthe shortfallof a household'sincome more heavily, the furtherit fallsbelow the povertyline, suggesting the severityof povertyfor the poorestof the poor.

2.29 For obvious reasons, the incidence of ultra-poverty is greatest among the unemployed: 45 percent of households headed by an unemployed person (7 percent of the population) are ultra-poor (Table 2.7). These households represent 12 percent of all ultra-poor households and suffer the deepest poverty, with average adult-equivalent spending just 21 percent of mean adult- equivalent spending. The next highest level of ultra-poverty, 35 percent, is found among households headed by casual laborers or self-employed workers and farmers who, because of their large weight in the population (nearly 47 percent), constitute about 63 percent of all ultra- poor households in Lesotho. Households headed by someone earning a wage or salary are least likely (10 percent) to be ultra-poor, but since nearly 37 percent of all household heads reported receiving regular wages or salaries, the proportion of wage-receiving households among ultra- poor households is relatively high, at about 14.5 percent.

2.30 The incidence of poverty in households headed by a pensioner or retired worker slightly exceeds that in the general population, although it is still lower than for any other group except households headed by a wage- or salary-earning worker (including mine workers). Apparently incomes in some households dependent on pensions lie between the two poverty lines.

2.31 The incidence of unemployment in the survey data appears to be low (7 percent), but this is largely because the data reported were limited to heads of households only. For the population as a whole, an estimated 25 to 30 percent of the workforce is unemployed. - 20 -

F. THE GEOGRAPHICDISTRIBUTION OF POVERTY

2.32 Poverty is greater in rural than in urban Lesotho. About 82 percent of Basotho households are rural, and 91 percent of the ultra-poor and 90 percent of the poor live in rural households. An average 29 percent of rural households are ultra-poor, compared with an average of 13 percent in Maseru and II percent in other urban areas. About 54 percent of rural households are poor, compared with a Maseru average of 28 percent and an other urban areas' average of 27 percent. These numbers show that the incidence of poverty is lowest in urban areas outside of Maseru; only 4.5 percent of the population lives in these smaller towns, and only 2 percent of ultra-poor households are found there. Using the upper poverty line, the incidence of poverty is only 27 percent in small towns, about 22 percent lower than the national average.

2.33 Poverty is also much deeper and more severe in rural areas. In ultra-poor rural households, average adult-equivalent spending amounts to just 22 percent of mean adult- equivalent spending. The pattern is similar in poor households. Poverty is also far more severe in rural areas -- more than 3 times worse than in Maseru (Table 2.8).

Table 2.8: Rural and Urban Poverty (1993) Incidence(P0) Depth(P1) Severity(P2) Population Poor Ultra- Poor Ultra- Poor Ultra- share (%) | poor I poor poor Rural 53.86 28.63 26.53 10.24 16.16 5.60 81.91 Other urban 26.84 11.00 12.06 4.86 7.12 2.39 4.48 Maseru,urban 28.14 13.63 12.80 3.73 7.09 1.80 13.61 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source:World Bank estimates based on the 1993Sechaba survey. Notes:The POmeasure (or head count index or incidence of poverty) measuresthe percentageof individuals below a given poverty line, relativeto the total population. The PI index (or poverty gap index) measuresthe depthof poverty asthe relativegap betweenaverage actual incomein poor householdsand the poverty line. It measuresthe percentageof the poverty line incomewhich is ncedcdto bring everyonebelow the poverty line upto the poverty line. The P2 index is a modificationof the PI index. It weighsthe shortfall of a household'sincome more heavily, the further it falls below the poverty line, suggestingthe severityof poverty for the poorestof the poor.

2.34 In predominantly rural economies such as Lesotho's, the natural resource base is often a key indicator of poverty. It is not surprising that both poverty and ultra-poverty are heavily concentrated in Lesotho's most poorly endowed zones, the Mountain and Senqu Valley regions (Table 2.9).

2.35 Poverty is also deeper and more severe among households in the mountains and the Senqu Valley. The P1 and P2 measures are more than twice as high in the mountains than in the lowlands, and over four times higher than in urban Maseru. Mountain and Senqu Valley residents (again, 26 percent of the national population) account for about 39 percent of the national incidence of ultra-poverty and 43 percent of both its depth and severity. They account for about 35 percent of the incidence of poverty above the upper poverty line, 38 percent of its depth, and 40 percent of its severity. - 21 -

Table 2.9: The Topography of Poverty (1993) Ecological Incidence(P0) Depth (PI) Severity(P2) Population Zone share (%) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor Lowland 44.08 21.68 20.51 7.54 12.16 3.99 38.93 Foothills 56.38 29.12 26.87 9.82 15.92 5.06 16.96 Mountain 71.17 41.36 37.70 15.47 23.86 8.94 14.20 Senqu Valley 61.70 35.58 32.39 13.49 20.45 7.69 11.82 Maseru,urban 28.14 13.63 12.80 3.73 7.09 1.80 13.61 Otherurban 26.84 11.00 12.06 4.86 7.12 2.39 4.48 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source:World Bank estimates based on the 1993Sechaba survey. Notes: The POmeasure (or head count index or incidenceof poverty) measuresthe percentageof individuals belowa givenpoverty line, relative to thetotal population. The P/ index(or poverty gap index) measuresthe depthof poverty as the relativegap betweenaverage actual incomein poorhouseholds and thepoverty line. It measuresthe percentageof the povertyline income which is neededto bringeveryone below the povertyline upto the povertyline. TheP2 indexis a modificationof the PI index.It weighsthe shortfall of a household'sincome more heavily, the furtherit fallsbelow the povertyline, suggesting the severityof povertyfor thepoorest of the poor.

2.36 The poorest districts are Thaba Tseka and Mokhotlong, followed closely by Quthing -- predominantly mountainous districts. And the seriousness of disparities is disturbing (Table 2.10). The incidence of ultra-poverty in Thaba Tseka and Mokhotlong is nearly double the national average, and severe poverty at the lower poverty line is similarly high (especially in Mokhotlong). The ultra-poor in Mokhotlong have consumption levels equal to just 20 percent of mean adult-equivalent consumption. The ultra-poor in Thaba Tseka fare only a little better. The figures are similar using the higher poverty line: 80 percent of Thaba Tseka residents and 74 percent of Mokhotlong residents are poor compared with a national average of 49 percent. By contrast, the incidence of poverty is substantially below average in Maseru district, where 37 percent of residents are poor and 18 percent are ultra-poor. But because 25 percent of the Basotho population lives in the capital, Maseru's contribution to national poverty (18 to 19 percent) is still high. The incidence, depth, and severity of poverty are also generally below the national average in the predominantly lowland/foothill districts of Leribe, Berea, and Mafeteng. - 22 -

Table 2.10: Poverty by District (1993) District Incidence (P0) Depth (PI) Severity (P2) Population share Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor Butha Buthe 54.95 37.66 29.84 11.95 18.64 5.76 4.68 Leribe 45.34 18.90 19.35 5.99 10.80 2.88 17.07 Berea 46.89 23.20 22.64 9.31 13.91 5.59 11.20 Maseru 36.88 17.97 16.79 5.16 9.46 2.43 24.81 Mafeteng 42.73 22.43 20.38 8.25 12.42 4.43 11.04 Mohale's Hoek 53.21 30.67 28.52 12.67 18.31 7.48 10.30 Quthing 63.93 39.11 33.79 14.41 21.68 8.17 6.61 Qacha's Nek 53.20 26.31 26.46 10.05 15.84 5.47 4.25 Mokhotlong 73.73 43.11 39.84 19.22 26.59 11.95 4.24 Thaba Tseka 79.89 44.58 41.00 13.84 24.41 7.11 5.79 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source: World Bank estimates based on the 1993 Sechaba survey. Notes: The PO measure (or head count index or incidence of poverty) measures the percentage of individuals below a given poverty line, relative to the total population. The PI index (or poverty gap index) measures the depth of poverty as the relative gap between average actual income in poor households and the poverty line. It measures the percentage of the poverty line income which is needed to bring everyone below the poverty line upto the poverty line. The P2 index is a modification of the PI index. It weighs the shortfall of a household's income more heavily, the further it falls below the poverty line, suggesting the severity of poverty for the poorest of the poor.

G. POVERTY AND ACCESS TO LAND

2.37 In a predominantly rural economy like Lesotho's, the ability to generate income is closely linked to the household's capacity to productively use the natural resource base. The only productive assets many poor households own are their own labor and some land.

2.38 The welfare of rural households varies significantly with the size of plots of agricultural land available to them (Table 2.11). The incidence of poverty is greatest among those with the smallest plots (less than 0.5 hectares): 61 percent are in poverty, compared with a rural average of 54 percent, and 36 percent are in ultra-poverty, compared with a rural average of 29 percent. The incidence of poverty declines as plots get larger -- except that severe poverty is greatest among those with 0.5 to 1.5 hectares of land. Perhaps poor households with this amount of land and more have less diversified sources of income, which makes them more vulnerable to shocks such as drought. Surprisingly, households without any agricultural land appear to be better off on all three measures of poverty. This appears to be attributable to this group's more diversified income base, including greater reliance on regular employment (especially mining, business, and wage work).

2.39 The counter-intuitive relation between farm size and poverty begs the question: "What prevents households in possession of a small landholding from making full use of its agricultural potential, either by cultivating the land with household resources or by renting it out to other households?" In the next chapter, we will suggest that the answer to this question can be found in an inappropriate agricultural policy framework, which depresses overall agricultural profits, and a poorly operating land rental market. - 23 -

Table 2.11: Poverty's Distribution by Size of Landholding (1993) Agricultural Land Incidence (P0) Depth (PI) Severity (P2) Population (hectares) share (%) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor 0 51.94 22.69 22.74 7.24 16.16 3.67 14.53 >0 and <0.5 61.23 35.77 32.40 13.70 12.97 8.05 23.34 >0.5 and <1.5 54.46 31.57 27.71 11.41 20.58 6.04 33.86 >1.5 48.06 22.25 22.20 7.53 17.24 4.05 28.27 All Rural 53.86 28.63 26.53 10.24 12.87 5.60 100.00 Source:World Bankestimates based on the 1993Sechaba survey. Notes:The POmeasure (or head count index or incidence of poverty) measuresthe percentageof individuals below a givenpoverty line,relative to the total population. The PI index(or poverty gap index) measuresthe depth of poverty as the relativegap betweenaverage actual incomein poorhouseholds and the povertyline. It measuresthe percentageof the povertyline incomewhich is neededto bringeveryone below the povertyline upto the povertyline. The P2 index is a modificationof the PI index. It weighsthe shortfallof a household'sincome more heavily, the furtherit falls belowthe povertyline, suggestingthe severityof povertyfor the poorestof the poor.

H. POVERTY AND HOUSEHOLD DEMOGRAPHICS

2.40 Some types of households are more disadvantaged than others no matter where they are. A large family size with an unfavorable ratio between the number of mouths to feed and the number of productive adults will make it difficult for the household to accumulate wealth. And sometimes characteristics of the household head may affect the whole household's well-being in a manner not fully explainable in terms of general characteristics. For example, in many countries (especially in Asia), girls are often more poorly nourished and educated than boys. Similarly, studies in several countries have shown that households headed by women or older adults tend to be poorer than otherwise comparable households headed by men or younger people. We now examine the degree to which such factors are relevant to Lesotho.

2.41 Lesotho's rural households are somewhat larger than urban ones and have significantly higher dependency ratios, which typically means greater poverty (Table 2.12). In many poor households, too few members of working age support too many young and elderly dependents. Hence, many poor households actually lack sufficient labor and capital to productively use the natural resource base. Rural households average about one more child under 16 than households in Maseru, the capital, and twice as many adults over 60. According to 1993 survey results, the number of households officially headed by women who are single, divorced, widowed, or abandoned by their spouses (25 percent) is higher than in many other African countries, because of the long tradition of Basotho men leaving to work in South African mines. And women are effectively heads of households in another 30 percent of households, an unsurprising figure given the high proportion of Basotho men employed in South African mines. These characteristics of poor households suggest that there is a strong life-cycle component to poverty, often aggravated by the fragility of the family structure caused by male labor migrancy. Poor households contain a higher number of persons of retirement age, with not enough life savings built up during their productive lives. When such households are subsequently deprived of adult men, their situation often becomes extremely precarious. - 24 -

Table 2.12: Demographic Characteristics of Basotho Households All Income Group Location Lesotho Non- Poor Ultra- Rural Urban Maseru poor Poor Household size 6.25 6.34 6.17 6.40 6.47 5.81 5.08 Dependency ratio"' 1.29 1.25 1.34 1.39 1.35 1.20 0.99 Avg of household head 50 47 53 54 51 46 45 % of households headed by 24.81 19.84 29.95 33.36 24.52 29.18 25.12 women Avg no. of children <16 yrs. 2.94 2.97 2.88 2.99 3.08 2.60 2.10 Avg no of adults > 59 yrs. 0.45 0.37 0.53 0.57 0.49 0.32 0.27 Source:World Bank estimates based on the 1993 SechabaHousehold Survey. Notes:a/ Thenumber of dependents(aged <16 & >64 years)as a ratioof the numberof householdmembers aged 16 to 64. b/ Refersonly to householdsofficially headedby women(not includinghouseholds headed by absentmen).

Households Headed by Women

2.42 Under apartheid in South Africa, women (including migrants from Lesotho) had only very limited access to wage work in the urban areas. Combined with the exclusively single male labor migrancy system operated by the South African mines, these past labor policies have profoundly affected the African family structure in the region. In Lesotho today, about 54 percent of all households are headed by women. About 25 percent are formally (de jure) headed by women (who are single, widowed, divorced. or abandoned by their spouses); and about 29 percent are effectively (de facto) headed by women, because the male head of household is absent. The other 46 percent of households are headed by men who live at home. With the head of household thus classified by gender and marital circumstances, significant welfare differences between households are apparent (Table 2.13).

Table 2.13: The Distribution of Poverty by Gender of Household Head (1993) Household Head Incidence (P0) Depth (PI) Severity (P2) Population Share (%) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor defacto headed 25.68 12.57 11.86 4.24 7.04 2.30 29.35 by women dejure headed by 59.33 34.68 30.42 11.96 18.83 6.07 24.81 women Headed by men 58.67 29.45 28.31 10.69 17.08 6.02 45.84 All 49.15 25.79 24.01 9.12 14.52 4.94 100.00 Source:World Bankcstimate based on 1993 Sechabasurvey. Notes: ThePO measure (or head count index or incidenceof poverty)measures the percentageof individuals belowa givenpoverty line, relativeto thetotal population. The PI index (or povertygap index) measuresthe depth of povertyas the relativegap betwecn average actual incomein poor householdsand the povertyline. It measuresthe percentageof the povertyline incomewhich is neededto bringeveryone below the povertyline upto the poverty line. The P2 index is a modificationof the PI index. It weighsthe shortfallof a household'sincome more heavilv. the furtherit fallsbelow the povertyline, suggestingthe severityof povertyfor the poorestof the poor. - 25 -

2.43 By all three poverty measures, members of households defacto headed by women (when the male head of household is absent) are far better off than households de jure headed by women or those headed by men. The incidence of poverty among households defacto headed by women is about half that of the population generally, and the depth and severity of poverty are equally low. At the other extreme, households dejure headed by women are disproportionately worse off: about 59 percent of them are poor, 35 percent are ultra-poor, and the depth and severity of their poverty are significantly worse than national averages. Poverty measures for households headed by men are only marginally better than measures for households de jure headed by women.

2.44 Once again, these observations suggest that access to income from regular wage employment, especially mining employment, is perhaps the single most important factor in determining welfare levels in Lesotho. Households headed by men or dejure headed by women have little or no access to remittances from mine workers in South Africa, and are generally poorer. By contrast, 74 percent of households effectively headed by women cite remittances from absent mine-worker husbands as their main source of income -- and those households tend to be better off. Only about 9 percent of households officially headed by women and only about 6 percent of households headed by men cite mining remittances as their main source of income.

2.45 The reason households headed by men have a slight advantage in welfare over households dejure headed by women is that men tend to have more access to wage (excluding mining) income. Wage income is the main source of income for 25.3 percent of male-headed (and for only 18.2 percent of households officially headed by women). Labor scarcity may also play a role in explaining the relative poverty of households dejure headed by women, as does the fact that the average age in such a household is 58.4 years, compared with an average age of 50.9 in households headed by men suggesting less capacity for work. The need to perform time- consuming household tasks such as fetching water and collecting fuel material could also limit labor available for income-generating work. (In a household headed by a man, the spouse may do such chores -- but by definition, in households headed by women, no spouse is present). This may not be an important limitation, however, since households headed by men and de jure by women have the same average number (2.71) of adults aged 16-55 years present in the household.

2.46 Some data from the 1993 survey suggest that children living in households headed by a male resident have somewhat lower levels of welfare -- poor nutrition and educational status -- than those living in households headed officially by women, and considerably lower levels of welfare than children living in households effectively headed by women, because the male head is absent. There are slightly better consumption levels in the households headed by men, but there is also more stunting among children under five (62 percent, compared with only 55 percent in households headed by women). Similarly, slightly more children aged 6 to 17 living in households officially headed by women are enrolled in school (79 percent, compared with 75 percent for children in households headed by resident men). Households effectively, i.e. defacto, headed by women have the best child welfare indicators overall.

2.47 In terms of material welfare, there is little intrinsic difference between households headed by men and de jure headed by women that cannot be explained by such factors as occupation, age, or main source of income. It is more difficult to explain why children in households officially headed by women fare better than those headed by resident men, despite the fact that the households headed by women are generally slightly poorer. Based on evidence - 26 - from other African countries, it seems likely that when income is controlled by women rather than by men, proportionately more household resources are allocated to children's welfare, partly because women traditionally bear more childcare responsibilities. This issue warrants further investigation.

Gender Bias Against Boys

2.48 Evidence of gender bias favoring boys has been found in much of Asia; evidence for Africa suggests more equitable patterns. In Lesotho, however, there seems to be a gender bias favoring girls over boys, regardless of the gender and other characteristics of the head of household. For example, for all measures of malnutrition -- stunting, wasting, and underweight -- boys under five fare substantially worse than girls under five. The most common explanation offered for this is the common rural practice of sending even very young boys (three to five) with older brothers to tend family livestock. As a result, these boys eat only sporadic or inadequate meals (Table 2.14).

Table 2.14: Nutritional Status of Basotho Children Under Five (1992) % stunted % wasted % underweight boys 36.1 2.8 17.1 Girls 30.1 2.0 14.5 13oth331241. Source:National Nutrition Survey, 1992.

2.49 The tradition of Basotho boys tending livestock is often given as a main reason why boys show lower school enrollment levels than girls in Lesotho. Basotho boys clearly get less primary schooling than girls. As Table 2.20 shows, among young people 6 to 17, 87 percent of girls attend school compared with only 72 percent of boys. Lower school enrollment for boys is found in all income groups, regions, and ecological zones, and in households headed by both men and women. The greatest disparity, however, is found in herding-oriented mountain zones, where nearly 30 percent more girls than boys attend school, lending credence to the view that livestock- tending depresses boys' school enrollment levels. Still, it would not explain sizable disparities in school enrollment levels in urban areas. Other factors must be at play. One possibility is that many parents still view working in South African mines as the most promising job prospect for Basotho men, and deemphasize boys' education as irrelevant for unskilled and semi-skilled mining work. Clearly the quality of Basotho boys' lives and their income earning potential are diminished by low education levels. - 27 -

Table 2.15: School Enrollment by Gender (% of Total, 1993) Males Females All By Poverty Level Nonpoor 80.69 91.04 86.06 Poor 64.30 82.30 78.06 Ultra-poor 58.43 78.06 67.75 By Region Rural 70.78 86.62 78.71 Other urban 82.55 92.80 87.88 Maseru, 84.40 87.73 86.08 By Ecological Zone Lowlands 77.93 87.42 82.79 Foothills 72.19 90.41 81.30 Mountains 53.88 83.36 67.60 SenquValley 66.45 82.35 74.56 By Genderof HouseholdHead/ Male 72.44 87.15 79.82 Female 72.48 86.39 79.48 Source:Bank estimates based on 1993Sechaba survey. Notes: a/ The categoryof householdsheaded by men includeshouseholds with both residentand absent men as heads.

The Age Factor

2.50 Households headed by an older person are especially vulnerable to poverty unless they have access to pension income from past employment in the mines or other formal sector jobs in Lesotho. On the whole, households headed by individuals over 50 are substantially poorer than those headed by younger people. Generally, the older the head of household, the poorer the household. Table 2.12 shows a notable progression in the average ages of household heads: heads of nonpoor households are an average 47 years old, heads of poor households, 53, and heads of ultra-poor households, 54. Households headed by Basotho over 50 who have access to pensions fare better than who don't have such access (Table 2.16). Lack of access to pensions increases the likelihood of poverty, and the likelihood that poverty will be deeper and more severe. For all three measures of poverty and ultra-poverty, households headed by older individuals are represented much more heavily. But when households headed by people over 50 are poor, if they have access to pensions, they tend not to be ultra-poor. - 28 -

Table 2.16: Distribution of Poverty by Age of Household Head and by Access to Pensions (1993) Incidence(PO) Depth (P1) Severity(P2) Poor Ultra- Poor Ultra- Poor Ultra- poor poor poor By age of householdhead <50 yrs 48.52 19.52 19.07 6.64 11.17 3.59 >50 yrs 58.52 32.25 29.02 11.51 17.82 6.17 All 49.15 25.79 24.01 9.12 14.52 4.94 By age and accessto pensions >50 yrs. w/ pension 41.96 14.41 18.01 1.39 8.12 0.18 >50 years w/o pension 57.72 31.73 28.63 11.41 17.57 6.09 Source:World Bank estimatesbased on 1993Sechaba survey. Notes: The PO measure (or head count indexor incidence of poverty) measuresthe percentageof individuals below a given poverty line, relative to the total population. The PI index (or poverty gap index) measuresthe depth of poverty as the relative gap betweenaverage actual incomein poor householdsand the poverty line. It measuresthe percentageof the poverty line income which is neededto bring everyonebelow the poverty line upto the poverty line. The P2 index is a modification of the PI index. It weighs the shortfall of a household'sincome more heavily, the further it falls below the poverty line, suggestingthe severityof poverty for the poorestof the poor.

2.51 Finally, although survey data do not clearly demonstrate it, there is little doubt that children five years and under are an important risk group. Very young children are at risk partly becauseof their higher representation among rural households, which tend to have poorer living standards. The high levels of malnutrition found among young children could seriously impair their ability to learn as children and then to earn a decent income as adults. The alarming levels of stunting among children in the 1993 survey -- nearly twice the levels found in the 1992 National Nutrition Survey -- suggest an urgent need first to independently confirm the results and then, if they are accurate, to take action to address the causes of food deprivation.

I. BASIC NEEDS INDICATORS

2.52 An analysis of income sources, remittances, and occupations is helpful in determining which economic activities are most likely to result in, or lead the way out of, poverty. But poverty is also associated with outcomes (such as malnutrition) that result both from low incomes and from poor access to key social services such as health care, education, safe water, and sanitation. Better nutrition, health care, education, and skills training -- that is, human capital -- is fundamental to individuals' efforts to lift themselves out of poverty. This section, a discussion of the human resources aspect of poverty, focuses on access to social services, especially those linked to nutrition, health care, water, sanitation, and education.

2.53 And as Table 2.17 shows, there is a systematic correlation between many common indicators of quality of life and the level of household consumption. Households characterized as poor and ultra-poor by consumption measures also show much lower levels of education, less healthy living environments, lower nutritional status, and more energy spent getting water and fuel. - 29 -

Table 2.17: Basic Needs Indicators of Basotho Households (% of All Households) All Non-poor Poor Ultra - Poor Education Level Householdhead w/ no formaleducation 31.90 21.50 42.50 46.40 Household head w/ grades 1-6 40.50 40.00 40.90 39.10 Householdhead w/ grade 7 10.80 14.40 7.10 6.90 Householdhead w/ more than grade 7 16.90 24.10 9.50 7.60 Schoolattendance of children6-17 80.00 86.06 72.99 67.75 Environmental Health Householdswith accessto safe water 70.52 78.74 62.01 58.25 Householdswith accessto a latrine 42.73 58.59 26.34 22.13 No. of residents per room 1.95 1.72 2.25 2.31 Nutritional Status of Children (under 5) A child is stunted 58.14 56.88 59.49 62.29 A child is wasted 19.97 17.70 22.41 23.54 Accessto Basic Necessities Time to fetch water over 30 mins. 22.16 20.22 24.10 25.38 Time to collect fuel over 30 mins. 51.31 41.12 61.40 65.92 Source: WorldBank estimates from Sechaba1993 household survey.

Nutrition

2.54 Poor nutritional status, especially among children, normally correlates highly with both income poverty and low human resource development. Nutritional status is normally measured using outcomes such as height and weight. Thus we may measure someone's weight for height, weight for age, or height for age, and compare the outcomes with a healthy reference standard population.33 A cutoff has to be deternined to establish when an individual is suffering from malnutrition. One of the most common approaches is to compute Z scores.34 The standard cutoff is then a Z score below -2.

2.55 Stunting (chronic malnutrition) occurs when an individual is at least two Z-scores below expected height for a given age. It represents the cumulative or long-run effects of poor diet and related factors. Wasting (acute malnutrition) occurs when an individual is two Z-scores below an expected weight for a given height. This in essence is the short-ternn effect of poor diet and related factors. The term underweight (global malnutrition) is used when an individual is below an expected weight for age.

2.56 The long-run indicators of nutritional status are poor among the Basotho population, especially children. According to the 1993 household survey, stunting affects 51 percent of all Basotho children under five. In a shocking 58 percent of households, at least one child is stunted

33There is somecontroversy about which referencestandard to use for comparison:healthy individualsfrom developedcountries, healthy individualsfrom the country under study, or healthy individualsfrom an ethnicallyand culturallysimilar country. We use an intemationalreference standard developed by the WorldHealth Organization.

34 The Z score of weightfor height, for example,is the differencebetween the weight of the individualand the medianweight of the healthypopulation divided by the standarddeviation of weightfor individualsof that height in the healthypopulation. - 30 -

(Table 2.17). Stunting usually reflects a persistently below-minimum-requirement food intake combined with such factors as inadequate health care, childcare, and sanitation -- which in turn are often correlated with limited access to health, water, and sanitation services and limited education for the parents, especially the mother. The incidence of wasting is lower, but 20 percent of households have at least one wasted child.

2.57 However, the 1993 survey results suggest such unusual patterns of stunting, wasting, and underweight that the data may not be reliable. For example, the prevalence of stunting is very high for children under 6 months, an improbable finding in a country where most children are breast-fed up to that age and are therefore unlikely to suffer chronic food deprivation. Data reported in the rest of this section are drawn largely from the National Nutrition Survey report compiled in 1992 by the Lesotho Ministries of Health and Agriculture with assistance from UNICEF, WHO, and UNDP. That survey found that for children under five, the general incidence of stunting was 33 percent, the incidence of wasting was 2.4 percent, and underweight prevalence was 15.8 percent -- figures lower than those from the 1993 survey, but still unacceptably high. Most disturbing, perhaps, is the fact that levels of malnutrition appear now to be substantially higher than in the past. The 1976 Nutrition Survey, for example, found 27 percent of children to be stunted. That figure had risen to 33 percent by 1992.

2.58 The prevalence of malnutrition is higher in some ecological zones and districts than in others (Table 2.18). Often, malnutrition rates are higher in areas disproportionately affected by income poverty. Notably, rates of stunting are above average in Mokhotlong and Thaba Tseka and in the mountainous zone generally. But the pattern of malnutrition prevalence does not always coincide with the incidence of poverty. In the mountainous Senqu Valley in Quthing district, the incidence of childhood malnutrition is relatively low -- comparable to that in Butha Buthe. This may be the result of a pilot nutrition initiative that involved training district nutrition committees and nutrition assistants, the early identification of and care for children at risk (through growth monitoring) and the establishment of income generating activities for women, including community gardens and commercial milling. - 3 1 -

Table 2.18: Malnutrition Rates by Region as a Percentage of Children Under 5 (1992) % stunted % wasted % underweight Ecological Zone Urban 28.3 1.5 12.5 LowlandsiFoothills 32.8 2.8 15.1 Mountains 36.4 2.5 18.6 District Butha Buthe 28.6 2.2 15.2 Leribe 37.4 1.2 14.4 Berea 24.7 4.8 13.7 Maseru 36.4 2.2 13.5 Mafeteng 33.1 1.7 14.7 Mohale's Hoek 31.1 2.3 16.3 Quthing 28.4 2.5 16.3 Qacha's Nek 39.8 2.5 19.1 Mokhotlong 36.5 1.9 20.6 Thaba Tseka 35.5 1.9 17.9 All Regions 33.0 2.4 15.8 Sourcc:National Nutrition Survey, 1992.

2.59 On the other hand, some districts with a lower incidence of income poverty display high rates of malnutrition. Most striking are the relatively high levels of stunting in rural Maseru and Leribe, districts with relatively low incidence, depth, and severity of poverty. Poor nutrition can be the result of inadequate access to social services (rather than poverty-related dietary inadequacies), but that is unlikely in these two districts, which are comparatively well served by basic social services.

2.60 The above results are extremely worrying. Even though Lesotho has gone through a severe period of drought, the malnutrition patterns are indicative of longer-term trends. Since the majority of rural households are net buyers of food, the price and availability of food in Lesotho will greatly affect household food security. In Chapter 3, we will show that inappropriate food policies -- pursuing national food self-sufficiency by artifically raising food prices and reducing competition in marketing and processing in both South Africa and Lesotho -- have substantially increased consumer prices of food in the region. The resulting high cost of food goes a long way toward explaining Lesotho's severe malnutrition rates.

Health Care, Safe Water, and Sanitation Services

2.61 Lesotho's health indicators are much poorer than those of comparable countries elsewhere, and in such neighboring countries as , Swaziland, and Zimbabwe. Despite considerable progress in the past decade, preventable (or easily treatable) health problems and diseases associated with poverty remain widespread in Lesotho. Among children under five, about 70 percent of deaths are attributable to gastro-intestinal illnesses, acute respiratory infections, malnutrition and peri-natal complications. Diarrhea and vomiting, acute respiratory infections, and measles accounted for over 80 percent of reported illnesses among children under five. The incidence of protein calorie malnutrition among children is high, and micronutrient deficiencies -- especially iodine deficiency, which can cause goiter and cretinism in severe cases -- are prevalent among both children and adults. Tuberculosis and complications related to - 32 - pregnancy are the most frequent causes of death among adults, with pregnancy-related problems accounting for nearly 30 percent of the deaths of women in thieir child-bearing years. Most common health and clinical nutritional problems are more prevalent in areas with higher proportions of poor households (in the rural mountains, for example).

2.62 The recent exponential increase in AIDS cases (from 25 cases in 1991 to 100 in 1992 and 300 in 1993) indicates that sexually transmitted diseases and AIDS may become an increasingly severe problem. Some estimates suggest that up to 20,000 Basotho may already be infected with the AIDS virus.35 AIDS and HIV are most often an urban problem, but in Lesotho they are also associated with the concentration of prostitutes around the mountain construction sites of the Lesotho Highlands Water Project.

2.63 Recent data on Lesotho's health indicators are not available at a sufficient level of disaggregation for use in this poverty profile. Because data from a 1992 Demographic and Health Survey carried out by the Ministry of Health have not been released yet, we have had to use estimates from the most recent census. As Table 2.19 shows, life expectancy at birth in 1986 was about 56 years, but with considerable variation across ecological zones and districts. Life expectancy was greater in urban areas (58.5 years) than in rural areas (55.0), and it was lower in mountain regions (51.9) than in lowland areas (57.4). Quthing (Senqu Valley and mountain areas) had the lowest life expectancy (50.7), but Butlha-Buthe, although partially mountainouts, had the highest (60.2). As expected, patterns of infant and child mortality closely parallel life expectancy patterns, mainly because they are major determinants of life expectancy.

2.64 Regional disparities in health status are likely to be related, among other things, to a household's access to health services and to environmental conditions, especially access to safe water and sanitary facilities. But it is extremely difficult -- given limited data -- to map the relationship between the incidence of poverty and available health services, partly because health care service statistics are maintained on a "Health Service Area" basis that does not correspond to other classifications used in this analysis.

35A detailedanalysis of this potential epidemicis beyondthe scopeof this report, but an AIDS epidemic could severelyharm growvth.poverty reduction,and the quality of life of the Basothopeople. For a fuller discussionof the potentialAIDS epidemic.see Lesothto Popuilation Sector Review. World Bank ReportNo. 12553-LSO,November 14, 1993. - 33 -

Table 2.19: Life Expectancy and Mortality Differentials by Place of Residence (1986) Area Infantmortality Childmortality Life expectancyat rate per 1000 rate per 1000 birth Rural-Urban Residence Urban 71 45 58.5 Rural 86 58 55.0 Residenceby Zone Lowland 76 48 57.4 Foothill 82 54 55.9 Mountain 99 71 51.9 Senqu River 94 66 53.0 Residence by District ButhaButhe 65 39 60.2 Leribe 84 57 55.3 Berea 79 52 56.5 Maseru 74 47 57.9 Mafeteng 77 50 57.0 Mohale'sHoek 94 66 53.1 Quthing 104 76 50.7 Qacha'sNek 86 59 54.9 Mokhotlong 100 72 51.8 Thaba Tseka 91 63 53.9 Source:Govenment of Lesotho(1986), Bureau of Statistics, Population Census, Analysis Report: Volume IV, Table 3.14.

2.65 What is known is that since the 1986 census Lesotho's health care system has launched important initiatives to extend coverage for basic health care services, especially preventive care. There have been campaigns for immunization; for the prevention, detection, and treatment of acute respiratory infections; for maternal health care; and for the management of diarrheal diseases. Coverage has ranged from 70 to 80 percent. These initiatives have probably improved the general health status of the Basotho people, but scattered evidence suggests that the households still to be reached by the initiatives are disproportionately the poorer ones. As with most social and economic services in Lesotho, health campaigns are often most limited in coverage in the rural mountainous areas where poverty is most severe and widespread. A review of health spending patterns shows that per capita spending in mountainous districts is lower than elsewhere. Finally, there is some evidence that even in lowland areas stringent new requirements for cost recovery have reduced poor people's use of health care services. These issues are discussed further in Chapter 4.

2.66 As for environmental health conditions -- principally access to safe water and sanitation services and freedom from residential crowding -- the 1993 survey results show clearly that conditions are much worse for the poor and ultra-poor than for the nonpoor (Table 2.17). Nearly 80 percent of the nonpoor consume water from safe sources, whereas only 62 percent of the poor and 58 percent of the ultra-poor do. Sanitation indicators are even more skewed: nearly 60 percent of the nonpoor have access to a latrine, compared with only 26 percent of the poor and 22 percent of the ultra-poor. The same pattern is apparent in patterns of housing density: members of poor households typically live an average 2.25 to a room, and the ultra-poor live 2.32 to a room, while members of better-off (nonpoor) households live 1.72 to a room. It is - 34 - reasonable to assume that poor access to safe water and sanitation facilities, combined withi more crowded households, increases the risk of ill healthi in poorer households.

2.67 Environmental health conditions are also worse for houIseholds in the mounitain and Senqu Valley zones (Table 2.20). Only 12 to 15 percent of mounitainiand Senqu Vallev households have access to a latrine, compared with 48 percent in the lowlands and 84 percent in the city of Maseru.As for residential crowding, Basotho in the mountains live an average 2.43 to a room, compared with 1.52 to a room in urban Maseru.

Table 2.20: Environmental Health Conditions by Place of Residence (1993) Area Residential Accessto Accessto crowding latrine safe water (persons/room) (%) (%) Rural-Urban Residence Urban 1.54 95.66 81.08 Rural 2.04 64.97 34.27 Residence by Zone Lowland 1.94 76.95 47.73 Foothill 1.89 67.83 35.10 MLountain 2.43 43.15 12.27 Senqu Valley 2.26 47.63 15.13 Source: World Bank estimates based on the 1993 Sechaba surveN.

2.68 The same pattern prevails for access to safe water and sanitation services. By and large, the primarily mountainous disiricts of Quthing, Qacha's Nek, Mokhotlong, and Thaba Tseka have the worst indicators. Of these districts, Thaba Tseka is the worst off, as only about 39 percent of its households have access to safe water and less than 6 percent have access to a latrine. In Qacha's Nek, the best-served of those four districts, just over half of its residents have access to safe water and about 28 percent have access to latrines. The best served districts are Maseru and Mafeteng, where respectively 82 and 85 percent of residents have access to safe water and 61 and 54 percent have access to a latrine.

Education

2.69 Educational status and poverty are negatively correlated: people with less education are much more likely to be poor. About 43 percent of the heads of poor households and 46 percent of the heads of ultra-poor households have no formal schooling, compared withi less than 23 percent of the heads of nonpoor households. Conversely, over 40 percent of the heads of nonpoor houselholds have completed primary school (or beyond) as compared to 17 percent in poor households and 13 percent in ultra-poor households. A clear pattern is also discernable in current school enrollment of school-aged children (6 to 17 years of age): children from nonpoor households are far more likely to attend school (86 percent) than those from poor (73 percent) and, especially, ultra-poor (68 percent) households.

2.70 A sinilar pattern is apparent in geographic disaggregations of education data. Nearly three-quarters of mountain-zone residents 16 years old or older have either no formal education or incomplete primary education, compared with a lowlands figure of about 59 percent (Table 2.21). Naturally, urban areas have the highest levels of education, with over 60 percent of - 35 -

Maseru residents and about 65 percent of other urban residents having completed primary school or beyond. Among districts, Maseru and Mafeteng fare best, followed closely by Leribe and Berea. Thaba Tseka has the worst education indicators; only about 21 percent of its residents have completed primary school or received further education.

Table 2.21: Educational Attainment of People Older Than 15, By Location (in percentages, 1993) No formal Grades1-6 Grade 7 Grades8- 1 Over Grade I I education Lowlands 15.19 43.75 22.45 14.85 3.76 Foothills 19.83 44.48 18.51 13.29 3.88 Mountains 33.96 40.20 16.98 7.28 1.57 Senqu Valley 31.34 37.01 17.48 11.49 2.68 Maseru 12.97 26.30 24.14 24.22 12.37 Other urban 8.68 25.84 20.32 30.72 14.44 Source: World Bank estimates, based on the 1993 Sechaba survey.

2.71 Everywhere in Lesotho, the younger generation is probably attending school more than its parents and grandparents did. Disparities in school enrollment among regions have narrowed over time, but the overall pattern has remained essentially unchanged (Figure 2.5). The pattern is one of lower school enrollment levels in the mountain areas, where poverty is most widespread and severe.

2.72 Of course, years of school attendance is one important factor; equally important is the quality of schooling. In this respect, the mountain areas also appear to be disadvantaged. Generally, 17 percent of teachers are formally unqualified, but in Mokhotlong that figure rises to 25.8 percent. As a result of having poorly trained teachers, together with higher pupil-teacher ratios (59:1), Mokhotlong has the highest repetition rate in Lesotho (27.5 percent). - 36 -

Figure 2.5: School Enrollment Levels of Children Aged 6-17, by Location (1993

By Ecological Zones By District

N 100 "0 100

50 *%0 f 3 3 280 i%of

60 Children 40 ~~~~~~~~~~~20 4280 0 Dsrc Z 20 123 5 7 1 e 0 0 1 2 3 4 5 6 District Zone

Source:World Bank estimates,based on 1993Sechaba Survey Notes: The numberingof zonesand districts follows the samenumbering as the map on the last pageof this report.

Ecologicalzones: I =Lowlands, 2 = Foothills,3 = Mountains,4 = SenquValley, 5 = Urban Maseru,6 =Other Urban Areas.

Districts: I =Butha Buthe, 2 = Leribe, 3 = Berea,4 =Maseru, 5 = Mafeteng,6 = Mohale's H-oek, 7 =Quthing, 8= Qacha's Ne!, 9 = Mokhotlong, 10 = Thaba Tseka.

J. TARGETINGTHE POOREST HOUSEHOLDS

2.73 Understanding the intra-household aspects of poverty helps identify households vulnerable to poverty, and can be helpful in designing targeted income transfer programs. Examples of such programs include relief food assistance schemes, food stamp programs, and old age or disability pension programs, which are usually intended as "safety nets" to keep the most vulnerable groups (or subset of groups) from becoming absolutely destitute. Efficient targeting of such transfer programs generally entails identifying a single characteristic or a manageably small (and easily measurable) combination of characteristics that will (1) identify as many poor individuals or households as possible, so they may receive the benefit offered, and (2) systematically exclude nonpoor individuals or households, so the transfer program remains affordable.

2.74 To identify key household characteristics that would help separate the poor (especially the ultra-poor, for who safety net schemes would most likely be intended) from the rest of the population, we developed a model to determine which factors would predict the probability of not being poor (Table 2.22 below, plus details in Annex 6). - 37 -

Table 2.22: Determining the Probability of NOT Being Poor (1993) Variable Coefficient T Statistic Odds Ratio' Mean Impact on probability of NOT being poor Characteristics of Household Head Female (0,1) 0.602 1.510 1.830 0.230 14.510 Age -0.015 -2.270 0.980 49.820 -0.038 Education (years of school) 0.106 4.400 1.110 4.140 2.640 Resides in household (0,1) -0.895 -4.840 0.410 0.670 -21.450 Marital status (0,1) 0.637 1.860 1.890 0.730 15.800 Foreign remittance income (0,1) 1.020 5.940 2.770 0.330 24.220 Household Composition Children aged 7 or less -0.044 -0.710 0.960 1.310 -1.080 Children aged 8-15 0.012 0.260 1.010 1.680 0.300 Members between 16-34 -0.140 -1.290 0.870 1.810 -3.470 Men between 35-59 years 0.022 0.120 1.020 0.490 0.540 Men 60 years and older 0.292 1.140 1.340 0.180 7.230 Women 35-59 years old 0.038 0.250 1.040 0.530 0.940 Women 60 years and older 0.172 0.890 1.190 0.260 4.270 Education of Household Members Men with no education 0.036 0.240 1.040 0.430 0.890 Men with grades 1-6 -0.034 -0.250 0.970 0.560 -0.840 Men with grade 7 0.238 1.360 1.270 0.190 5.890 Men with grades 8-11 0.142 0.760 1.150 0.170 3.520 Men with 12th grade and above 0.531 1.700 1.700 0.060 13.170 Women with no education -0.414 -0.209 0.660 0.140 -10.270 Women with grades 1-6 -0.073 -0.570 0.930 0.630 -1.820 Women with grade 7 0.102 0.750 1.110 0.400 2.520 Women with grades 8-11 0.671 4.190 1.960 0.280 16.630 Women with 12th grade and above 0.946 3.540 2.570 0.080 23.450 TotalHectaresofFarmland 0.093 2.210 1.100 1.110 2.310 EcologicalZones Foothill -0.426 -2.600 0.650 0.180 -10.600 Mountains -0.593 -3.070 0.550 0.140 -14.710 Senqu River Valley -0.546 -2.630 0.580 0.100 -13.570 Maseru 0.847 3.800 2.330 0.110 18.730 Other urban locations 0.783 2.680 2.190 0.060 17.520 Constant 0.004 0.010 Number of observations = 1610 Chi2(29) = 488.470 2 Prob > Chi = 0.000 Log Liklihood = -869.740 Pseudo R2 = 0.219 Source:Bank staff estimatesbased on 1993Sechaba Survey. Note: The impacton the probabilityhas been calculateddifferently for continuousand dichotomousvariables. For continuous variablesit is P(Y=I)*(I-P(Y=I))*B,where P(Y=I) is evaluated at the mean values and B is the estimated value of the coefficient on thc particular variable. For dichotomousvariables or a set of them, the impact has been calculated as P(Y=lIX=O) - P(Y=11X=l), whereall the other valuesare evaluatedat the mean values.At these mean values, P(Y=l)=54.47 and the odds are 1.20. a/ The odds ratio explainsthe impactof the unit changein a dependentvariable on the odds of not being poor. b/ The impacton the probabilityof not being poorevaluates the changein the probabilityof not being poor for a unit change in one of the variableswhen all other variablesare set at their mean ratios. - 38 -

2.75 Our first step was to identify the key predictive variables for poverty. The main findings of the probability estimate derived from the regression analysis, and reported in Table 2.22, are as follows:

* Thle probability of being poor decreases substantially when the major source of income is mining remittances or when the head of the household is married. When the household is clejure or de facto headed by a woman, the probability of being poor also decreases, but taken together with other factors (marital status and residence of the household head in the household) this effect can disappear easily. The probability of being poor also decreases when the head of the household has more education. It increases, but only slowly, with age.

* The probability of being poor increases slightly in households where there are childreni under 7. It decreases somewhat when there are children 8 to 15 (probably because they contribute to productive activities). But the more persons between 16 to 34 years old present in the household, the more likely it is to be poor, perhaps because of high unemployment levels. Poverty is a bit less likely when there are people 35 and over in the house, especially when houselhold members (as distinct from household heads) are 60 and over. This suggests that older individuals residing in a houselhold headed by someone else contribute some resources -- probably from lifelong accumulation or from pension income -- that substantially improve the household's welfare, although the same resources would be insufficient if the older person were the main or sole provider of household income.

e Poverty is less likely to be found in houselholdswith educated people (especially if they have completed primary school throuigihgrade 7); and in Maseru and other urban areas. Poverty is more likely when there are women with no schooling in the houselhold(a girl is generally sent to school tinless the household is ultra-poor or she herself suffers some disability); or when the houselhold resides in higher altitude regions. Although there is a negative, and statistically significant, assccation between size of land holdings and the probability of being poor, its economic significance is rather weak (one extra hectare would only lower the chances of' being poor by a few percentage points).

2.76 Our second step in isolating the characteristics of poor households was to evaluate objective correlates with poverty. Some variables -- for example, crowded living quarters and the absence of a latrine or radio -- tend to be relatively highly correlated with poverty and could be useful indicators for identifying the poor (Annex 7).

2.77 Our third and final step was to identify the most suitable predictors and correlates, either singly or in combination, to identify poor households toward which direct assistance might be targeted. We did this separately for rural and urban populations (Tables 2.23 through 2.26).

2.78 The ideal targening variable (or combination of variables) would be one that:

* would identify 100percent of the target group, the poor; * would exclude 100 percent of the nonpoor; * could be easily and accurately measured; and - 39 -

* entails no risk of "moral hazard" -- that is, selection criteria would not unintentionally give potential beneficiaries an incentive to do something contrary to their long-term interests (for example, keep boys home from school because they knew assistance would be targeted to households with low school enrollment for boys).

2.79 It is unlikely that one could ever find a combination of variables which would perfectly identify all the poor and exclude all the nonpoor. But in many countries it is possible to find a set of characteristics that limits both types of identification error to acceptable levels -- about 20 to 25 percent for each type of error.

2.80 Unfortunately, even this practical standard may not be feasible in Lesotho, where none of the single or combined characteristics considered yielded both high coverage of the poor and low leakage to the nonpoor. Generally, characteristics, or combinations of them, that identified a high proportion of the poor also (mis)identified a high proportion of the nonpoor, so using them for targeting purposes would waste limited resources, on the one hand, and yet fail to reach enough of the needy, on the other. Using combinations of key indicators (Tables 2.25 and 2.26) would exclude 25 to 30 percent of the target population and would include roughly 35 to 45 percent of non-target households in both urban and rural areas. Using single characteristics (Tables 2.23 and 2.24) yielded even less satisfactory results.

Table 2.23: Key Indicators for Urban Households (% Identified) % Identified Poor Nonpoor Ultra-poor Non-ultra- poor Householdsheaded by a woman 34.05 23.07 45.07 23.31 Head not marriedand HH size 2 2 37.50 24.37 44.95 25.50 Major incomenot from remittance 88.25 75.32 85.67 77.91 Headolderthan65years 12.11 8.35 11.13 9.14 Head with no schooling 30.62 9.70 30.77 13.25 Head residesat home 93.54 71.68 91.57 75.70 More than 3 membersaged 16-35 17.19 6.80 24.02 7.55 Householdsnot residingin optaka'& 97.59 89.29 97.64 90.70 No accessto piped water 83.52 62.85 82.51 66.53 No accessto a latrine 34.02 13.11 26.30 17.83 Does not own a radio 29.22 17.82 39.45 18.24 At least5 personsto a room 19.72 7.42 11.99 10.67 Source:Estimates based on the 1993Sechaba household survey. Notes:HH = household. a/ Anoptaka is a modern,multiple-room house with a peakedcorrugated iron or tileroof. - 40 -

Table 2.24: Key Indicators for Rural Households (%) % Identified Poor Nonpoor Ultra-poor Non- ultra-poor Head not married and HH size 2 2 32.10 18.73 35.39 22.14 Major income not from remittances 73.13 46.10 76.87 54.15 Head olderthan 60 years 33.96 19.12 37.01 23.15 Head with no schooling 42.41 24.29 46.64 29.00 Head resides at home 83.14 50.99 84.72 61.73 Owns <0.5 Ha of agricultural land 26.54 19.62 29.18 21.01 Some members reside in rondavel 74.04 57.41 78.05 61.68 Water source from spring 53.28 38.42 55.14 42.92 No access to a latrine 78.19 50.13 83.04 58.10 Does not own a radio 47.26 76.55 45.01 67.10 At least 5 persons to a room 17.14 11.20 20.62 11.90 Source: World Bankestimates based on the 1993Sechaba household survey. Note: HH = household.

Table 2.25: Combinations of Key Indicators for Urban Households % Identified Poor Non-poor Ultra-poor Non- Ultra-poor A. Predictors I. Unmarried female head 33.51 23.07 43.89 23.31 2. Head older than 65 12.11 8.35 11.13 9.14 3. Head with no schooling 30.62 9.70 30.77 13.25 4. More than 3 members aged 16-35 17.19 6.80 24.02 7.55 Households with at least one of the above characteristics 66.67 37.16 75.20 40.93 B. Objective Correlates 1. Not residing in optakaa', no access to piped water, no latrine 30.62 10.37 23.94 14.82 C. Households with at least one of the characteristic listed in A and B 68.42 37.92 74.57 46.27 above Source:World Bankestimates based on the 1993 Sechabahousehold survey. Note: HII= household. a/ An optaka is a modern,multiple room house with a peakedcorrugated iron or tile roof - 41 -

Table 2.26: Combinations of Key Indicators for Rural Households % Identified Poor Nonpoor Ultra-poor Non- Ultra-poor A. Predictors I. Unmarried head older than 60 years 17.15 6.93 20.83 9.07 2. Unmarried head with no schooling 8.47 2.53 10.75 3.71 3. Unmarried head with <0.5 Ha of agricultural land 7.38 3.15 8.97 4.01 4. Head older than 60 years with no schooling 15.06 4.63 17.53 7.32 5. Head older than 60 years old with <0.5 Ha of agricultural land 7.38 2.68 8.96 3.71 6. Resident head with no schooling 36.13 11.43 40.71 18.33 7. Resident head with <0.5 Ha of agricultural land 20.94 8.30 23.10 11.90 Households with at least one of the above seven characteristics 57.02 24.65 64.51 33.09 B. Objective Correlates 1. Resides in rondavel, with no access to piped water, no latrines, and no radio 23.92 6.71 27.33 11.43 2. At least 5 members per room 17.14 11.20 20.62 11.90 Household with at least one of the above two characteristics 32.93 14.02 37.28 18.96 C. Households with at least one of the characteristiclisted in A and B above 71.33 36.14 77.39 46.15 Source: World Bank estimates based on the 1993 Sechaba household survey. Note: Single-personhouseholds have been excludedfrom the sample.

2.81 What this means from a policy standpoint is that directly targeted assistance programs that require administrative procedures to identify the poor or ultra-poor will probably not be an efficient or effective means of reducing poverty in Lesotho. More emphasis must be placed on self-targeting approaches, such as public employment schemes.

K. BASOTHO PERSPECTIVES ON POVERTY

2.82 To learn about Basotho experiences of and attitudes toward poverty, we consulted two qualitative studies of poverty in Lesotho, undertaken by the Sechaba consulting group. The studies used such participatory techniques as focus group discussions, case studies, and participant observation to get the perspectives on poverty of Basotho citizens (urban and rural).3 6 These studies explored, among other things, people's experience and depiction of poverty; their

36 Both studies were undertakenby Sechaba Consultants,a local consultantinggroup, the first with EU financing and the second with combined EU and World Bank financing. Complete findings are available in Poverty in Lesotho: A Mapping Exercise, Sechaba Consultants, June 1991, and in Poverty in Lesotho: A Mapping Exercise, Volume 11,(draft) Sechaba Consultants, December 1993. - 42 - view of the principal causes of poverty and effective strategies for coping with it; their understanding of the relationship between basic material needs and poverty; and their views about the relationships between poverty and sociopolitical factors or cultural traditions. The studies provide a rich and vivid portrait of people's experience of, and attitudes toward, poverty as well as its causes and effects (summarized in Annex 8). The findings presented here confirm or modify the results of the quantitative analysis, and seem relevant for policy formulation, although participatory studies have obvious limitations as a basis for policy analysis and design.3"

2.83 General observations. Poverty, especially severe poverty, is viewed as widespread in Lesotho, affecting every community visited by the Sechaba research teams. In discussions with groups of individuals identified by researchers as both poor and better-off, the most common response to the question "What is it to be poor?" was that poverty meant a lack of food and clothing. Basotho communities seem to identify only the most destitute -- probably a subset of the ultra-poor -- as poor. People also often defined poverty as lack of employment (money), livestock, and agricultural assets; poor mental or physical health; and insecure shelter. Lack of adequate water and education were sometimes mentioned as defining poverty, but only infrequently.

2.84 By and large, education was not seen as a major factor in bringing about prosperity. Rather, prosperity was seen -- especially by rural dwellers -- as driven largely by the ability to secure mine employment, an occupation believed to require physical courage and endurance more than skills in literacy and numeracy. This view of the road out of poverty may help explain the earlier finding that boys have lower school enrollments than girls (for whom mining is not an option).

2.85 In rural areas, participants in discussion groups often associated poverty with small households comprised of only adult women or women and children without support from an adult male, principally because of a spouse or grown-up son's death, desertion, or illiess (especially a mental or physical disability or tuberculosis). This is similar to findings of the quantitative analysis.

2.86 By contrast, those identified as destitute in urban areas were themselves largely unemployed adult men, classified by the researchers as: (I) mine applicants -- queueing daily outside the mining employment office in the hope of recruitment; (2) circle men -- some of them former mining applicants, who sleep "rough" in and around Maseru's main traffic circle, having depleted both social and material resources and having lost hope of stable employment; (3) shebeen men -- who remain around informal liquor houses hoping for drink or accommodation; and (4) beggars, scavengers, and street boys -- of whom the researchers felt there were relatively few in Lesotho compared with other countries. This classification of the urban destitute suggests that the survey data used for the quantitative analysis probably underestimates the incidence, and especially the severity, of urban poverty in Lesotho. Since many of the urban destitute sleep "rough" -- have no fixed abode -- they were unlikely to have been sampled in either the 1986/87 or 1993 surveys. On the other hand, the predominantly rural character of Lesotho, based on a relatively equitable agrarian structure which is small-scale in nature, has prevented the

37 Partly becausethe sample is not statistically representativeof the population as a whole. and partly becauseopen-ended responses are difficult to usein quanhitativeanalysis. - 43 - emergence of large periurban slums such as can be found in South Africa and Latin America, where the rural poor have been deprived of access to land.

2.87 Causes of poverty. Participants' responses about the causes of poverty differed slightly in the 1991 and 1993 studies, partly because interviewers in the two studies took different approaches.38 In the 1991 study, the four causes of poverty mentioned most often were:

* Unemployment. * Environmental disaster (including drought, floods, and hailstorms as well as loss of soil fertility and productive pasture land). * The lack of adequate agricultural land and inputs. * III health or disability.

In the 1993 study, the causes of poverty mentioned most often were:

* Alcoholism, which participants considered to be the most common cause of poverty. * Unemployment (which moved from first to second place). * The drought. * Personal failings such as laziness. - Political factors such as injustice, oppression, and corruption, especially in local government. - Social issues such as conflict, hatred, and witchcraft.

2.88 Not all of these issues can be addressed directly through public policy, but some have policy implications. Alcoholism, for example, is a preventable, treatable health problem that can be addressed through public health measures. Action in this area may be especially important because participants strongly associated alcohol abuse with adultery, suggesting that it may be an important factor in the spread of AIDS. Similarly, the shocks to households from natural disasters could be lessened through better monitoring of weather conditions and prompt design and implementation of assistance interventions. Finally, the frequent mention of abuses of local power must be kept firmly in mind if these authorities are to be involved in identifying the poor for fee exemptions (for health care, for example) or for direct assistance (for food relief rations, for example).

2.89 Coping strategies. Both qualitative studies paid considerable attention to the coping strategies of the poor. The picture that emerges is this: by and large both the rural and urban destitute manage to find enough food (principally maize porridge and vegetables) to meet subjective minimum nutritional requirements, but many experience great uncertainty and anxiety about the source of their next meal. The high levels of malnutrition observed above, however, suggest that objective minimum nutritional requirements are often not met.

2.90 In rural areas, many of the poor provide their own food by farming their land, renting land out to others under share-cropping arrangements, or through casual labor on neighbors' property (for example, weeding, ploughing, mending fences, or guarding the crop or houses).

38 The first study took a "needs assessment"approach to causes of poverty,trying to link poverty with the lack of accessto economicservices or opportunitiesor to socialservices. In the secondstudy, researchers used a list of 42 concepts to elicit participants'reactions on a rangeof issues -- includingtheft, adultery,alcoholism, and corruption -- that they felt contributedto or affectedpoverty. - 44 -

They undertake many of these activities in exchange for food, clothing, or other in-kind payments, which are sometimes viewed as "gifts" rather than as payments for service rendered. They also earn income through informal retail activities or by making and selling handicrafts or common household objects, such as brooms. An important source of income for many of the rural destitute is the natural environment, where they find fuel wood and shrubs, thatching grass (when available), and medicinal herbs. Many of these activities, as well as subsistence farming, were seriously affected by the recent drought.

2.91 In urban areas, many of the destitute earn money by collecting cardboard boxes and selling them to liquor houses for fuel. They also make sporadic income doing "piece jobs" for near-by businesses or private citizens. Finally, gifts from relatives and neighbors are important to the survival of the poor, especially those that researchers classified as destitute. The poor also rely greatly on free meals provided by church missions and other organizations and on income from begging. Many urban participants cite petty theft as an important survival strategy, one they turn to only as a last resort.

2.92 The most striking impression one gets from the discussion of coping strategies is the strong sense of community that apparently exists in rural villages (despite concerns about political and social conflict). Not only did the poor and the better-off report much giving and receiving of gifts, but a variety of traditional village feasts are open to all members of the community, regardless of social status. These feasts are often the only chance the ultra-poor have to eat foods like meat. More than 80 percent of the better-off households report helping others in need and that assistance is almost as likely to be provided to a non-relative as to a relative. This strong sense of community is an important starting point for the development of community-based poverty reduction strategies.

2.93 Poverty Reduction. Interviewers for the 1993 study asked people how they thought the poverty problem could be solved in their communities. Among the most common solutions suggested:

* Provide jobs (36 percent). * Provide counseling and/or ban alcohol and tobacco (21 percent).39 * Provide food and clothing, especially for those suffering from ill-health or disability ( 18 percent). * Ensure good local leadership as a means of combatting injustice and corruption and fostering community cooperation (9 percent). * Provide direct assistance to the aged and to children (5 percent).

2.94 Clearly the qualitative research turned up information not readily obtained thlrough quantitative household surveys. For instance, two new policy implications stemming from the Basotho people's view of ways to reduce poverty were the need to eliminate alcohol abuse and (in the view of a significant minority) provide good local leadership. Prominent in its absence was any mention of basic economic and social services as a path out of poverty. This may be because the study was designed to get at poverty issues of personal and emotional importance to

39The Sechabastudy reports counseling(16 percent)and the banningof tobaccoand alcohol(5 percent)as separateissues. We lump them together here because the later write-up implies that most of thc participants mentioningcounseling in connectionwith alcoholabuse, although it was also seen as necessaryto motivatethose with other personalfailings (such as "laziness")to work. - 45 -

participants rather than at economic issues per se. And because the focus was on the poorest members of their communities, it is not surprising that participants focused on factors that would alleviate the plight of these destitute people in the shortrun, rather than on factors such as education, child health, feeder roads, water supplies, and the like which, though essential to sustainable poverty reduction, could have only a medium- to long-term impact.

L. POLICYRECOMMENDATIONS

2.95 Given Lesotho's poverty profile, certain policies are more likely to result in sustainable medium-term reduction of poverty. These policies will be elaborated in the remainder of the report. But given the poverty profile established in this chapter, the following general themes can already be emphasized. High priority should be given to:

* Increasing returns to rural labor, because of the high incidence of poverty in rural areas and among households engaged primarily in farming, herding, and informal business. While many observers have all but written off broad-based agricultural growth as a viable economic strategy for poverty reduction in Lesotho, both the quantitative and participatory analysis of poverty emphasize the need for a strategy aimed at creating sustainable rural livelihoods. More profitable part-time, small- scale farming and livestock activities must be developed to supplement remittance income.

- Rapidly increasing formal wage jobs in domestic industry and services, to offset declining opportunities for mine work in South Africa. Workers' remittances from mine work have in the past played a key role in keeping households out of poverty.

- Lowering the cost to the poor of essential goods such as food, fuel, water, health care, and education. These basics use up most of the household budget of the poor, whose burden would be lightened if markets for food staples were liberalized and fees for social services rationalized.

- Making the provision of social services more equitable geographically, in both quantity and quality, so rural Basotho households have more equal access to decent basic health care, education, safe water, and minimal sanitation facilities.

* Instituting a well-targeted safety net, to protect vulnerable populations -- including the sick or disabled, the elderly, and those living in resource-poor regions -- against severe deprivation. Self-targeting approaches should be used, when possible.

* Studying the feasibility of introducing special initiatives aimed at children under five and school-age boys, whose malnutrition problems are apparently severe enough to affect their cognitive functioning and hence their ability to be economically productive later in life. - 46 -

3. FOSTERING LABOR-INTENSIVE GROWTH

3.1 The poor depend mainly on income from labor to meet their daily needs. Any national strategy to reduce poverty in Lesotho must promote not only macroeconomic stability and rapid economic growth, but more productive use of labor, the country's -- especially the poor's -- most abundant resource. Chapters 3 to 6 analyze the extent to which Lesotho's policies and programs already provide a supportive environment for poverty reduction in the medium- to longer-term and identify areas where changes may be necessary if the country's poverty reduction focus is to be enhanced. Key topics that are covered in this regard include:

* The adequacy of existing macroeconomic and sectoral policies for encouraging sustained, labor-intensive growth and hence affording the poor the opportunities to make productive use of their most abundant asset -- their labor;

* the efficacy of existing efforts to provide basic social services to the poor to enable them to accumulate the human capital (good health and nutrition, education and skills) necessary to take advantage of income-earning opportunities arising from economic growth; and

* the efficiency and effectiveness of "safety net" measures and programs for those segments of the population (e.g., the aged or the disabled) that are temporarily or permanently unable to take advantage of income-earning opportunities. A brief analysis of institutional capacity for regular poverty monitoring and coordination of poverty reduction efforts is also presented.

3.2 Given Lesotho's high unemployment rate and incidence of rural poverty, there can be little doubt that the general medium- to long-term strategy for balanced, sustainable poverty reduction must emphasize labor-intensive growth in agriculture, small enterprises, and labor- intensive, export-oriented manufacturing, underpinned by major investments in the development of human capital. In a labor-surplus economy like Lesotho's, the recommended sector-specific policy changes and program initiatives to reduce poverty are also fully consistent with achieving rapid growth and maintaining a macroeconomic balance. The rationale for these recommendations is discussed in this chapter.

A. MACROECONOMIC PERSPECTIVE

3.3 Historically, Lesotho's development strategy has unduly emphasized self-sufficiency in the production of basic commodities, whether or not Lesotho had a comparative advantage in doing so. In both agriculture and industry, Lesotho's policies have tended to promote capital- intensive import substitution. The strategy has been one of parastatals, state companies, explicit and implicit capital cost subsidies, protectionist measures, and restrictive business licensing. Attempts to create mechanized, large-scale maize, , and vegetable farms using rainfed and irrigated agriculture largely failed, despite considerable support. Lesotho imports steadily increasing amounts of these foods. Attempts to create an industrial and agroindustrial base producing for the local market have produced a range of subsidy or monopolistic rent-dependent public companies and parastatals. These parastatals have produced precious few jobs and harmed the poor in at least three ways: by distorting budget allocations (for example, subsidizing - 47 - parastatals at the expense of investments in rural infrastructure that would help the poor), by raising the consumer price of basic goods that dominate the household budgets of the poor, and by crowding out more labor-intensive, small and medium-scale private enterprises.

3.4 The import-substitution strategy was partly inspired by the then prevailing paradigm of economiiic development, and partly by the political economy of the region -- which meant rendering the coun1tryas independent as possible from South Africa's apartheid regime. Recent efforts to reduce or eliminate domestic macroeconomic imbalances have succeeded and an export-orientation is beginninig to take root, but not enough advantage has been taken of econoinic stability and the emerging opportunities presented by the fundamental transition under way in the Southern African region to increase employment opportunities for the poor or to enhance productivity in the activities in which the poor commonly engage.

Employment

3.5 There are high levels of under- and unemployment and far too few employment opportunities, in both Lesotho and South Africa. The little data on employment that exists shows the labor force to be greatly underused. Projections from the 1986 census show the potential labor force in 1993 (that is people aged 15 to 64) to be slightly over I million, out of a total population of about 1.9 million. But the actual labor force is only 800,000, assuming a labor force participation rates of about 85 percent for men and 75 percent for women. And population growth rates were highl (2.6 percent a year) in the period between censuses (1976 and 1986), so net annual increases to the potential labor force should be between 20,000 and 25,000 over the next few years.

3.6 In 1993, opportunities for formal employment fell far short of need. Altogether only 28 percent of the workforce, or an estimated 225,000 people held formal wage jobs, distributed roughly as follows (based on estimates from various sources):

* I 5,000 Basotho worked in SoutlhAfrican mines. * 21,000 were civil servants (including both salaried and waged workers). * 20,000 were employed in domestic manufacturing. * 18,000 worked in retail trade. * 15,000 worked in construction. * 13,000 did other work in South Africa -- for example, in agriculture, private industry, domestic service, or a professional occupation. * 10,000 were employed as teachers. * 7,000 had jobs in hotels, restaurants, transportation, or financial services. * 4,000 were employed in military or police security. * 2,000 worked in other services, nonprofit organizations, or private health care.

That left roughly 72 percent of the labor force (about 575,000 workers) to work in agriculture and the informal sector; sectors in which returns to labor are generally low. Underemployment and its byproduct, poverty, are widespread in rural Lesotho and open unemployment is growing in urban areas. The unemiployment rate in 1993 -- counting both unemployment and underemploymenit -- was estimated to be an alarming 35 to 45 percent. Growth strategies that produce jobs are urgently needed. - 48 -

Figure 3.1. Lesotho Labor Force Formal Labor Force *RSAMines

225,000 E*CivilServants

Labor Force - Total 800,000 6% 4% 3% *Dom. Manur.

8% 1% ERetail OConstruction

28_ 9-Y. N Other Work in RSA

72% 9% _ Teachers

51% OTourismnFinanciaI

* Miliiary/Police

* ServicestHealth * Formal Woge

Multiplier Effects

3.7 Poverty has been reduced somewhat as the result of recent initiatives to develop the labor-intensive, export-oriented light industry (about 5,500 jobs created in five years, mainly in textiles) and to start up the Lesotho Highlands Water Project (around 3,000 jobs in construction), but the poverty-reducing potential of these initiatives has not yet been fully exploited. In both cases, the benefits to the economy only consist of their direct employment effects. What is missing is a more economy-wide multiplier effect, which could greatly enhance the employment impact of both the export boomlet in textiles and the massive construction investments under LHWP. And poverty could have been reduced considerably by initiatives to increase output in the activities in which the poor are disproportionately engaged: horticulture, livestock production, and small service enterprises. Not that a focus on the development of agriculture and small enterprises would be a panacea for Lesotho's poverty problems, but development efforts to date have been lopsided. More attention should have been paid to the labor-intensive economic sectors. - 49 -

3.8 It is important that Lesotho'sgrowth strategyproduce both direct and indirect growth in jobs and income (multiplier effects). Growth in manufacturing output, for example, would directly increaseformal employmentat the same time that it boostedincome and profits in the service sector and domesticagriculture, by raising demand for servicesand agricultural goods. Similarly, investmentand growth in agricultural production, by raising farmers' incomes,could also generate more jobs in manufacturing (in agro-industries,for example) and services by increasingdemand both for farming inputsand for processingand marketingservices.

3.9 The questionis, which is moreefficient: to investdirectly in agriculture, for instance,or to invest in other sectorsthat could indirectly increasedemand for rural products? To answer this question,the World Bank missiondeveloped a Semi Input-Output(S1O) model of Lesotho's economy (describedin Annex 1). The SIO model analyzesthe direct and indirect (multiplier) effects of changesin output in all sectorsof the economyrelative to their impact on the average incomesof Basothohouseholds (Table 3.1). In terms of poverty reduction,the model produced three main conclusions:

Developmentprojects and policies aimed at stimulatingthe demand-sideof the economy cannot count on significant multiplier effects. In Lesotho, links betweensectors are generally weak. Lesotho'seconomy is tightly knit with that of South Africa, through relatively open borders, membershipin the Common Monetary Area (CMA), and the flow of migrant labor from Lesothoto South Africa. This tightly knit relationshiphelps to stabilize prices, encourage trade, and raise household incomes from migrant remittances,but results in substantialleakage from the domesticeconomy and thus few multiplier effects in Lesotho. Most goods and some servicesare tradable, so increased demandis met largely through increasedimports.

* Becauselinks betweensectors are weak (Annex I, Table 4), the direct effects of changes in supply generallyoutweigh the indirect or multiplier effects. So the investmentsmost likely to reducepoverty are those in sectorsthat direct a large shareof value addedto the poor. Most poor Basotho householdsare in rural areas, so increasesin agricultural incomehave the most generalpotential to reducepoverty.

* IncreasingLesotho's export orientation is likely to strengthensupply links. In the long run, more local investmentin, and increasedproduction of tradables(or the conversion of nontradablesto tradables-- for example,by canning vegetables),could substantially increasemultiplier effects. Next to the overseasmarkets, Lesotho could exploit the proximity and potential of the large South African market by developing exports for which productivity in Lesotho is high becauseof special characteristicsof its land, existing investments,or availableskilled labor. - 50 -

Table 3.1: How Output Growth in Selected Sectors Affects Different Classes of Households: Household Income Effect of I Loti Increase in Value-Added Field Crops Livestock Handicrafts Textiles Other Construction Migrant Manufacturing Remittances

Effects on Household Income Urban 0.150 0.154 0.582 0.571 0.385 0.287 0.052 Male HH 0.112 0.110 0.480 0.452 0.280 0.210 0.043 FemaleHH 0.038 0.044 0.102 0.119 0.105 0.077 0.009 Rural 0.725 0.713 0.776 0.972 0.239 0.464 1.008 Male HH 0.557 0.546 0.590 0.736 0.169 0.348 0.890 FemaleHH 0.169 0.168 0.186 0.236 0.070 0.116 0.117

Poor 0.589 0.579 0.645 0.806 0.193 0.373 0.133 Nonpoor 0.287 0.288 0.714 0.736 0.431 0.377 0.928 Total 0.876 0.867 1.358 1.543 0.624 0.750 1.060

Share of Income to Households Poor 0.672 0.667 0.475 0.523 0.310 0.498 0.125 Nonpoor 0.328 0.333 0.525 0.477 0.690 0.502 0.875 Source:World Bank staff estimates. Notes: HH = households.

3.10 In income gains, poor households benefit most from a I Loti increase in value-added from handicrafts, textiles (including apparel and clothing), agriculture, and livestock. They benefit slightly less from a similar increase in construction output. They benefit little from a similar increase in other manufacturing output or from an increase in the wage rate paid to workers in South Africa, because such remittances do not generally go to poor households, and generate few multiplier effects. To put it another way, about two-thirds of the income generated from increases in field crops and livestock accrue to the poor, and half of the income generated from increased output in handicrafts, construction, and textiles. The nonpoor get most of the increased income from other manufacturing (69 percent) and migrant remittances (88 percent).

3.11 The implications are clear: to reduce poverty, Lesotho's growth strategy should emphasize growth through labor-intensive development of agriculture, the informal sector, and labor-intensive industries (such as textiles). To improve domestic multiplier effects, Lesotho should also emphasize the production of tradables and the conversion of nontradables (such as water, vegetables, and fruits) to tradables (such as water-for-export and canned goods).

B. FOSTERING LABOR-INTENSIVEAGRICULTURE AND RURAL DEVELOPMENT

3.12 In recent years, agriculture has accounted for less than 20 percent of Lesotho's GDP, down from about 30 to 35 percent in the 1970s. In normal years, field crops and livestock are roughly equal in terms of value added, but livestock production is more resistant to such ravages of climate as drought, so its contribution to GDP is more stable. Maize, wheat, and sorghum dominate field crops; cattle, other livestock,and wool and mohair production are roughly equal in importance in livestockproduction. Although the declinein agriculture'scontribution to GDP is partly the result of rapid growth in other sectors such as light manufacturingand construction,it is also attributableto absolute declinesin agriculturaloutput. These declinesoccurred, among other - 51 -

reasons, because of environmentalfactors (such as soil erosion),and because of an inappropriate agriculturalpolicy framework,both in SouthAfrica and in Lesotho.

3.13 Despite agriculture's declining share of GDP, its contribution to the economy is critical. Nearly 85 percent of Basotho households live in rural areas and about 70 percent derive part of their income from agriculture. And by and large, the 35 to 40 percent of households that rely mainly on agriculture for income are disproportionately represented among the poor and ultra poor (Chapter 2). Agricultural employment and incomes must be expanded to avert medium- to longer-term increases in poverty resulting from rapid population growth combined with stagnating opportunities for mine employment and too little expansion of formal employment to absorb the annual supply of new entrants to the labor force.

3.14 To be sure, agriculture alone is unlikely to generate enough income to support the present rural population, given Lesotho's poor natural resource endowment, characterized by a mountainous topography, limited arable land, unreliable climate, and serious soil erosion. It is vital that Lesotho continue expanding other sectors of the economy, especially labor-intensive light manufacturing and the service sector. But agricultural performance and rural incomes could improve substantially if the right policies were implemented: diversifying toward high- value export crops, involving the private sector more in marketing and processing, and improving land use efficiency. The policy issues are discussed below.

Food Self-Sufficiency

3.15 Since the late 1970s, Lesotho's agricultural policies have been shaped largely by the perceived political need to reduce dependence on food imports from South Africa and to protect domestic producers from unfair competition from the heavily subsidized, large-scale South African producers. With the aid of a number of donors, including the World Bank, Lesotho promoted food self-sufficiency in agriculture, with a focus on increasing output for the dominant staple food crop, maize. The assumption was that maize self-sufficiency was vital in case South Africa suddenly closed the border with Lesotho.

3.16 The Food Self-Sufficiency Program (FSSP) initiated in 1980/81 involved, among other things: (1) setting fixed producer prices for maize grain well above world prices and even substantially higher than South African producer prices; (2) heavily subsidizing government- provided tractor services for field preparation and planting40; (3) supplying seed and subsidized fertilizer for participating farmers through a marketing parastatal, Coop Lesotho; and (4) providing loans through local banks (Lesotho Bank and Lesotho Agriculture Development Bank) to pay for these services and inputs. The recovery rate on these loans was well below 10 percent, so most loans (totally about M2 million a year at their peak) were effectively subsidies to maize producers. Price incentives and subsidies heavily favored maize production.

3.17 Lesotho was expected to be self-sufficient in maize production five years after FSSP was initiated, but maize production continued to decline slowly. Currently, even in a year with normal rainfall, maize must be imported to meet as much as 60 to 65 percent of national food

40 MOA recordsshow that 1988J89revenues for its TechnicalOperations Unit (which is responsiblefor tractor services)covered only 26 percentof currentoperating costs, including salaries and wages direct;y related to machinery operations. - 52 - requirements. Roughly 90 percent of rural agricultural households, as well as the rural landless and nearly all urban households, are net consumers of basic staples. Data from the 1986/87 household survey show only about 14 percent of national food consumption being home-grown; the figures for 1993 are even lower, because of the drought.

3.18 Lesotho failed to achieve food self-sufficiency largely because of the unsuitability of its agroclimate for maize and wheat production, including the vagaries of local rainfall patterns and the poor quality of its soils. Local and temporary dry periods are standard any summer, making rainfed maize production a risky venture every year. Crop failure is frequent in areas where soils do not retain moisture or when dry spells occur before root systems have developed enough to extract moisture from the soil. And drought -- a prolonged dry period over an extended area -- is relatively common. The 1991-92 drought nearly halved maize production (from 110,000 mt in a normal year to 55,500 mt that season) and hurt sorghum and wheat production even more. Frequent drought and local dry spells make soil in many parts of the country highly vulnerable to erosion; when the rains return, the soil has less capacity to absorb rainwater. Topsoil erosion is severe because little of the land is protected by forest, pasture, or crop cover, for reasons related to land tenure, overgrazing, and fuel shortage problems. On top of these climate-related problems, some of the price incentives offered under the food self-sufficiency program may have accelerated soil erosion, by encouraging farmers to plant unsuitable marginal land to maize. This was especially true in the lowlands and foothills where extensive long-term cultivation of food crops -- mainly maize -- combined with heavy grazing has caused the crusting and erosion of surface soil.

Potential for Crop Diversification

3.19 In recent years, the Government has recognized that this strategy of food self-sufficiency is not working, not only because of the country's limited resource base but because the food self- sufficiency program costs too much, creates few jobs, and is damaging the environment. Instead, there is a new emphasis on moving away from food-grain monocrops and toward high- value, labor-intensive agricultural activities. In particular, the Government (supported by bilateral donors) is trying to promote a variety of nontraditional crops, including asparagus, potatoes, and sunflowers. This diversification, implemented on a large scale, could substantially increase both national and rural household income and employment. A recent World Bank study4' concludes that Lesotho's potential for agricultural diversification is considerable because the country seems to hold a comparative advantage in several high-value export crops (including asparagus, saffron, and some nuts and berries) and in some vegetables and fruits (including cabbage, potatoes, tomatoes, onions, apples, pears, and peaches). Moreover, Lesotho is located near a good market, South Africa; has good (often preferential) trade agreements with South Africa and Europe; and also has a convertible currency, a good airport, and one of the highest education participation rates in Africa, especially for women who are the main agricultural producers.

3.20 A move towards crop-diversification would certainly make sense from a poverty- reduction perspective, because nontraditional crops such as asparagus, potatoes, and cabbage are much more labor-intensive than such traditional crops as maize, sorghum, and beans and yield greater returns per hectare planted (Table 3.2). Asparagus, in particular, is not only labor-

4 "Lesotho:Environment and Agricultural Diversification Study," World Bank Report No. 11906-LSO. - 53 -

intensive but involves activities traditionally considered suitable for women -- an important consideration in Lesotho, where men are often absent from the household.

Table 3.2: Labor Used, and Returns from, Irrigated Nontraditional Crops X2 LaborUse Returns(Net of Costs) Returns/LaborDay ( Days/ Hectare (Maloti/Hect=)e {MWot" Asparagus 240 2,000 8.33 Cabbage(Summer) 100 1,210 12.10 Cabbage(Winter) 80 1,050 13.13 Potatoes 80 880 11.00 Lucerne 25 490 19.60 Beans 17 160 9.41 Maize 25 80 3.20 Sorghum 10 30 3.00 Source:LASA Report No. IO,MOA. 1982 and Bankestimates. Notes:Labor use includesboth field labor and laborfor processingand marketing.Data on returnsper hectarepredate price incentives for maize.

3.21 Although crop diversification would create jobs and reduce environmental damage, efforts to diversify away from maize monocropping have had only limited success. Two factors in particular appear to be inhibiting growth in output of nontraditional crops: lack or distortion of private sector incentives, and inadequate infrastructure and research and extension services.

Price, Marketing and Processing Distortions

3.22 Substantial distortions exist in agricultural marketing and processing. These distortions inhibit the development of private sector-led trade, marketing and processing (which ultimately keeps producer prices down), skew relative prices towards the production of traditional food crops for which Lesotho has little comparative advantage, and raise consumer prices of staples above their import parity price.

3.23 Until recently, Co-op Lesotho (a mainly government-funded farmers' cooperative) had a monopoly on buying agricultural produce from growers and supplying them with agricultural inputs. And as a result of Co-op's inefficient operations (including frequent shortages of agricultural inputs at critical times and inadequate funds to purchase growers' produce), producers' yields have suffered and produce could not always be marketed when it was harvested, which depressed producer prices. More important, the uncertainties created by such an inefficient marketing system have made farmers reluctant to experiment with nontraditional crops, especially those that cannot be consumed by the household or sold or bartered locally. As a result, domestic production of commonly consumed vegetables such as cabbage, potatoes, tomatoes, and onions has been low, and Lesotho imports 30,000 to 35,000 tons of these vegetables a year to meet domestic demand. A more liberal marketing environment would attract private traders and undoubtedly raise producer prices and increase yields, given the network of private traders who engaged fairly effectively in agricultural trade before Co-op Lesotho was created. Because of serious financial difficulties, Co-op Lesotho suspended operations in about mid-1993 and a commission was established to review its status and future.

3.24 Under the circumstances, the Government should consider actively promoting private sector involvement in agricultural trade as a way to encourage agricultural diversification and - 54 - raise rural household incomes. It is a common fallacy to assume that small farmers could be made to market their produce themselves. International experience clearly shows that this only applies to situations in which farmers can directly sell their produce on a near-by village market. As soon as the farmer wants to diversify into crops demanded by markets which are further away (say the fresh asparagus market in Johannesburg), he or she will lack the relevant information on price, quality and quantity. Traders are the economic agents who bridge this information gap between producers and the market. To constrain their operations ultimately means constraining the incomes of farmers. And should Co-op Lesotho be revived, care should be taken to keep the playing field level for the Co-op and private traders, so producers can reap maximum benefits.

3.25 Traditional food crops. One factor in households' apparent reluctance to shift from the traditional foodcrop (maize) to nontraditional crops is the presence of continued distortions in producer incentives, often holdovers from the food self-sufficiency era. Maize pricing, trade, marketing, and processing policies still heavily skew incentives toward maize production and away from nontraditional crops. Part of the price distortion originates in South Africa. In 1993/94 and 1994/95, for instance, South Africa's Maize Marketing Board set producer prices of R417 and R330 per metric ton respectively, well above the export parity price of about R220 and R240. Because 1993/94 and 1994/95 were "surplus" years42, the Maize Board exported part of the harvest at a loss. The concept of "surplus" was a relative one. Because the system was operated as a single-channel, legal monopoly, the Maize Board determined the extent of the "surplus," because it manipulated the price of maize. Given the powerful lobbying of producers, the Maize Board removed part of the harvest in order to artificially reduce local supply. It sold the removed surplus on the international market at a loss, since South African producer prices were higher than the world market price. These export losses were not borne by the producer (in whose interest the scheme is operated), but were passed onto the mills, and hence the consumer. In 1994/95, for instance, export losses of the South African Maize Board amounted to an astounding R600 million. In 1993/94 and 1994/95, the Maize Board charged R103 and RI 85 per ton back to the consumer to finance its export losses.

3.26 What happens in Lesotho? The producer miligate or gazetted price (the price at which Basotho farmers can sell to the mills in Lesotho) is set by the Marketing Division of Lesotho's Ministry of Agriculture. This price is not set as the import parity price of the South African producer price, but as the import parity price of the Maize Board's into-mill price (the price the Maize Board charges to millers when they buy grain from the Board). Lesotho's consumers are asked to pay for a system which benefits South African farmers, because the into-mill price includes the mark-up needed to finance the export losses. In 1993/94, the Marketing Division in Lesotho set the price at M627 per metric ton -- half again as high as the South Africa producer price.43 The price differential of M210 also includes the cost of rail transport (about R42-4S per/mt in 1993/94) and the Lesotho border levy on imported maize grain (Ml per metric ton).

3.27 In other words, Lesotho's miligate producer prices are higher than they would be if there were producer parity pricing across borders. However, even with these high domestic producer

42 Basedon pasttrends, about 6 out of 7 yearsare "surplus" years. 43TheM627 producer price is thegazetted price for GradeI whitemaize. Prices for GradeI yellow maizeand Grade11 and Grade Ill whiteand yellow maize grain are slightly lower. Industrialmillers indicated that, given the small amountsof grainbought from domestic farmers (between 1.5 and 5 percentof all maizeprocessed, depending on the mill), theytended to payfarmers the highest price - generallythe gazetted price for GradeI whitemaize. - 55 -

prices (which should induce Basotho producers to increase supply), the Lesotho mills typically only procure 1.5 to 5 percent of their grain from Basotho producers. The rest is procured from the South African Maize Board, and, at times, from donors' drought relief programs. And despite the implicit subsidy to domestic producers, farners' groups in Lesotho all agree that the prices do not cover costs and also allow for a normal profit. In short, Lesotho seems to have no comparative advantage in maize production, except perhaps in the north, where rainfall is more reliable.

3.28 In fact, open market producer prices for maize grain are even higher than official gazetted producer prices. Most Basotho households produce far less than they consume, so the effective value of home-grown maize grain to producer households is not the producer price set by the Ministry of Agriculture's marketing division, but rather the farm gate price. With so little local production, this price will approach the consumer price of maize meal (sold by the three industrial millers that control local marketing of maize meal) minus the cost of transport and milling at village hammer mills. From the perspective of an average rural household, each additional unit of maize grain produced and converted to meal at village hammer mills substitutes for the expense of purchasing meal from local retailers. According to Bank estimations, in 1992/93, the implicit producer price of maize grain was at least M693 -- more than 16 percent higher than the gazetted producer price for Grade I white maize (M597) (Table 3.3) 44

3.29 Efforts to reduce price distortions that unduly favor maize production over nontraditional crops must not only bring gazetted producer prices in line with border parity prices, but must also reduce consumer prices for maize meal, to make the implicit producer price of maize grain more competitive. How can this be done? Making producer prices more competitive means re- thinking policies on imports of maize meal and grain and on their processing and marketing in Lesotho. Imports of milled maize are currently banned (a restriction sometimes lifted in times of drought). This ban eliminates foreign competition for domestic millers. In principle, the ban benefits both village mills and industrial mills, but the main beneficiaries are Lesotho's three industrial mills (one parastatal, Lesotho Flour Mills, and two joint ventures of Lesotho National Development Corporation, or LNDC, with private parties). The industrial millers handled more than 95 percent of maize grain imports in recent years, and account for the lion's share of maize meal sold commercially in Lesotho. Since little domestic maize grain is marketed, the benefits of protection are effectively available only to millers with substantial access to maize grain imports, and the existing licensing requirements for importing maize grain have prevented the emergence of small-scale traders and millers.45

44Estimates of the implicitproducer price of maizegrain werc devcloped using information on operatingcosts of villagemills. Theanalysis focused on thecosts of unsiftedindustrial maize meal, the product most comparable to that producedby the villagehammer mills. Dataon wholesalerand retailermargins and incrementaltransport costs were limitedso conclusions are tentative. Estimates of wholesaleand retail mark-ups were relatively conservative (10 percent at eachstage). Implicit producer prices might be even higher using actual data. 'Proceedingsof a Seminaron the Informal GrainMarket in Lesotho,"organized by thePlanning Division, Ministry of Agriculture,September II, 1992. 45 Officialsin MOA's MarketingDivision told the Bank missionthat maizegrain importlicenses are now grantedrelatively easily to anyonerequesting them. But the threeindustrial millers continue to monopolizemaize imports (97percent of total importsin 1993).Either many traders and village mill operatorsare unaware of the relaxedlicensing environmentor theyare unable to get maizefrom the SouthAfrican Maize Marketing Board, the solesource of maizein SouthAfrica. Apparentlytraders are alsoconcerned that importingmaize grain for directsale to consumescould jeopardizetheir relationship with thebig millersand hence their access to maizemeal. - 56 -

3.30 The oligopoly of the three industrial millers need not increase consumer maize meal prices if economies of scale and competition among the three mills lowered their operating costs and if they took normal profits. But the March 1993 profit and loss statement for Lesotho Flour Mills shows the reverse to be true: it had operating costs of M]15.60 plus oligopolistic profits of M95.84 per metric ton (combined total: M21 1.40). By contrast, operating costs plus profits for village mills averaged M94.00 (Annex 9). It is reasonable to assume that the other two industrial millers have similarly high operating costs and profits; lower operating costs would mean only higher profits, given the pre-set sales price under the current system of administered pricing. The higher operating costs and profits of industrial millers substantially raise maize meal costs to consumers and hence the implicit producer prices of maize grain. Clearly, liberalized maize grain imports and maize meal consumer prices would enable fair competition between industrial and village millers, bringing down consumer prices from their current levels and, in turn, lowering implicit producer prices for maize grain.

3.31 So far we have argued for more liberal policies because distorted policies on maize pricing, trade, marketing, and processing discourage producers from diversifying away from maize monocropping. More important, perhaps, such liberalization would reduce basic food costs for Basotho consumers, who now pay artificially high prices for maize meal. Consumers suffer a per capita loss of M21 a year -- or about M126 for an average household of six46 -- because of trade restrictions that benefit only Lesotho's inefficiently operated industrial mills. In effect, consumers are taxed to benefit producers and the industrial mills.47 (Another M26 per capita -- or about Ml 56 per household -- would be saved if South Africa would remove its international trade restrictions, and stop protecting inefficient domestic producers of maize. See Annex 10.) And because poor households spend proportionately more on food than better-off households do, current policies are unambiguously regressive. The poorest 20 percent of the population spends roughly two-thirds of its income on food -- especially maize meal -- while the top 20 percent spends only 30 to 35 percent of its income on a much more varied diet, spending only a fraction of its food budget on maize. Lowering consumer prices for maize would also benefit industry, by reducing the pressure on wage bills (by raising workers' living standards).

46 Theper capita loss is calculatedassuming a populationof about1.7 million for 1992/93.The total loss each year is an estimatedM35.11 million, even with conservative assumptions about costs. See Annex10 for a moredetailed analysisof thewelfare impact of current maize policies. 47To theextent that maize-consuming households also produce maize, consumption losses are partially offset by productiongains (estimated at about M19 million in 1992/93).But thereduction in lossesis relativelysmall. Domestic productionaccounted for less than 15percent of consumption in 1992/93. - 57 -

Table 3.3: Implicit Producer Price of Maize (1992/93)

Maloti per metric ton

Gazetted milidoor price for unsifted maize meal"" 80 kg polypropylenepacking 757.63 50 kg polypropylenepacking 762.80 50 kg cotton packing 798.60

Gazetted milidoor price minus costsof packing materialb 80 kg polypropylenepacking 733.50 50 kg polypropylenepacking 733.60 50 kg cottonpacking 733.60

Sub-total: Retail or consumer pricec deductions 887.66 Milling costat village milld' -94.00 Net milling loss" -70.00 Incrementaltransport costsf/ -30.00

Implicit producer price for maize grain 693.33 Notes: a/ Gazettedprices for unsiftedmaize were convertedto metric ton equivalentsthrough simple multiplication. b/ Costsfor packingmaterials, provided to the Bank missionby the industrialmills, wereMl.93 per 80kgpolypropylene bag; M 1.46per 50kg polypropylene bag; and M3.25 per 50kg cotton bag. c/ Assumes10-percent mark-up at wholesaleand a further10 percent at retaillevel. d/ Averagecost derived from milling costsreported in the MOA surveyof "The InformalGrain Market in Lesotho,"September 1992. e/ Accordingto the industrialmills, milling produces86-87 percent maize meal (unsifted) and 13-14 percentbran or offal (usedfor animalfeed). The valueof the branwas M250per metricton in 1992/93,much lower thanthe estimatedM887.66 retail valueof maizemeal. The net milling loss from convertingmaize grain to maizemeal (and bran) at a village mill is M70 (subtractedby subtractingfrom the value of maize meal (M887.66x0.135=103.10)the value of the bran (M250.OOxO.135= 33.10). f/ Accordingto the "InformalGrain Market in Lesotho"survey, farmers report finding villagemills an average3km away. This may be fartherthan the distanceto a retailerof maizemeal so an incrementaltransport cost of M30 permetric ton wasassumed.

3.32 Livestock products. Some policies hurt livestock producers by restricting channels for marketing products and purchasing inputs: this tends to depress the profitability of livestock production. Marketing monopsonies(where only one buyer makes all purchases) have affected producer prices in egg productionand dairy farming, for example. The productionof fresh eggs is a good example. Until recently, egg importswere banned in Lesotho and the Lesotho Poultry Cooperative Society (LPCS) had a monopoly on egg sales to domestic retailers, which meant they also had a monopsonyon purchasesfrom egg producers. Eggs that producerscould not sell informally had to be transported to LPCS depots, which caused many producers to incur transport costs they wouldn't have had if they could sell direct to retailers. And because of financial difficulties within LPCS, producersoften had to wait a long time to get paid for their eggs. Consumer prices for eggs were often unnecessarilyhigh because of improper storage conditionsand high spoilage rates at LPCS,especially in the summer. LPCS has now suspended operations pending investigation of its management, and producers are being given licenses to - 58- sell direct to retailers. If this temporary system is expanded and if full liberalization ensues, both producers and consumers will probably benefit considerably.

3.33 Similarly, Lesotho Dairy Products (a Maseru-based parastatal) was established to sell raw milk, which means frequent delays in payments and higher transport costs for producers who cannot sell on the local informal market. So producers located at a distance from Maseru have little incentive to produce more milk than they can sell informally, although milk production could be very profitable.4" An import ban on fresh milk exists to protect this parastatal. The import ban and the inefficient operations of the dairy plant have resulted in a contraction of the national consumption of fresh milk and an increase in imports of long-life milk, a much more expensive product, but one for which no import ban exists. Again, deregulation of the marketing environment will most likely benefit both producers and consumers.

3.34 In Maseru, the National Abattoir -- owned and operated by the Ministry of Agriculture -- holds a monopoly on the slaughter of animals. This prevents the emergence of small-scale abattoirs (also known as "slabattoirs"), which could create new entrepreneurs and employment, and lower the price of meat. And because the abattoir also operates a feedlot, butchers bringing in live animals for slaughter are often pressured into parking their animals in the feedlot at a fee.

3.35 Export crops. The policies that keep producer prices of maize artificially high have tended to depress relative prices of high-value nontraditional crops, making them less attractive to growers. Asparagus, for example, has been grown in Lesotho for export since the mid-1970s. Asparagus has been moderately popular with growers: the area planted which asparagus has grown from about 30 hectares in mid-1970 to about 400 hectares at present. Agronomists believe much more rapid growth would be possible, were it not for the policy environment. Tight controls on competition and private involvement in asparagus marketing, bans on private exports to South Africa in order to protect the LNDC-owned cannery, and the use of inappropriate technology have restrained yields to less than half their potential.49

3.36 Asparagus is an especially good example of the problems in crop diversification. Asparagus growing, canning, and export are vertically integrated in Lesotho. Asparagus growers must sell their output to a domestic parastatal, Basotho Canners, which theoretically provides growers with inputs, extension, and credit and creates additional employment through its canning operations. Most of the canned asparagus is then exported to Europe. In practice, poor management at Basotho Canners has meant low capacity utilization (the factory is idle more than half the year), poor marketing, and inappropriate extension advice to growers -- factors that have depressed not only yields, exports and foreign exchange earnings, but also the prices paid to producers. In 1992/93, Basotho Canners paid Basotho producers about M2.16 per kilo of asparagus (excluding transport costs), less than half the price asparagus fanners in South Africa receive. South African entrepreneurs seeking to purchase asparagus in Lesotho were refused

48 Returnsto investmentin dairyfarming are about 21 to 37 percentafter all costsare considered, including the cost of feed and of farmers'labor. (Mochebelele and Ranko,"Rural Dairy Development and the Roleof Agricultural Development Projects in Lesotho" Agricultural Marketing Research Project Report No. 26, Institute of Southem African Studies,National University of Lesotho,1990.)

49 In 1990,a yearof "normal"rainfall, yields averaged about 4.5 tonsper hectare, although the best growers reportedyields of 10.5tons per hectare, and yields on somefarms in SouthAfrica are reportedlyas highas 20 tonsper hectare. - 59 - permission to do so. This has also meant that Basotho producers have had to forego the much more lucrative fresh asparagus market in South Africa's main cities. The Lesotho Government is considering privatizing Basotho Canners to improve its performance. Privatization may solve some problems but growers (generally small landholders) will benefit more from such privatization if it is accompanied by liberalized trade and marketing, which allows both domestic and South African buyers to bid producer prices up to more competitive levels.

Irrigation, Research and Extension

3.37 In addition to chianging policies to improve producer incentives, it is also important to improve irrigation and to reorient agricultural research and extension toward the production of nontraditional export crops. It is especially important to improve irrigation, given Lesotho's variable rainfall patterns and the dependence on regular watering of most horticultural crops (asparagus being an exception). Estimates vary about how much of Lesotho's arable land is irrigable (from 5 to 15 percent), but experts agree that the minuscule area that now has access to controlled water represents only a fraction of the potentially irrigable area. The Ministry of Agriculture recently estimated that meeting domestic demand for vegetables would require only 400 to 500 hectares of irrigated land; any further expansion could be used for export crops. So far, Lesotho's efforts to expand irrigation have met with only limited success, partly because efforts to date -- mainly large, capital-intensive efforts (such as the Bauer scheme) that require extensive community cooperation -- have been impractical. Smaller projects should be explored; ones administered in such a way that farmers can operate individual plots, choose their own crop mix, and would pay only for the water they use. One promising technical option is the gravity- fed irrigation system initiated under the UNDP-supported Local Initiatives Support Program (LISP), which could be replicated elsewhere in the country, especially in the mountains and foothills. Surface water storage schemes should also be considered, as they could help reduce soil erosion.

3.38 Applied research and related extension services are critical to the development of many new horticultural cash crops, including such vegetables as pinto beans, seed potatoes, sweet corn, baby corn, French beans, and broccoli; such fruits as apples, pears, and peaches; a variety of nuts and berries; spices such as saffron; and some species of flowers that can be dried for export.50 To alleviate the severe, extensive poverty that flourishes in mountain districts, highest priority should be given to developing crops (such as seed potatoes) that are likely to grow well in the Highlands. Agricultural extension should focus not only on promoting nontraditional crops, but on reducing the risks from periodic dry spells for poor rural households that continue to grow traditional food crops, fodder, thatching, and so on. Among strategies that should be promoted are intercropping, relay planting, strip planting, and more use of drought-resistant cultivars (which, although not the highest-yielding under optimal conditions, better withstand dry weather). Lesotho should explore collaborating with South Africa in agricultural research, to reduce its research costs.

3.39 Apart from marketing and pricing problems, several technical factors affect the development of dairy farming. These include (I) limited access to credit; (2) the unavailability of timely veterinary services; (3) inadequate training and extension services to compensate for

50 Thesecrops, and issues affecting agricultural research and extension,are discussedmore fully in the 1993 WorldBank "Environment and Agricultural Diversification Study," part of whichis reproducedin Annex 11. - 60 - the limited experience of Basotho farmers with the relatively complex activities involved in dairy farming; and (4) the frequent undernourishment and the variable genetic quality of milk cows -- the latter the result of poor breeding practices.

3.40 Similar constraints affect wool and mohair production, domestic output from which per livestock unit is considerably lower than in South Africa. Wool fleece weights in Lesotho are about 60 pet-cent of South African weights, for example, and mohair weights are even lower. Lesotho products also tend to be inferior in quality to those from South Africa, so the price to Lesotho producers is lower.

3.41 Efforts to improve agricultural extension and other services affecting dairy farming and wool and mohair production (as well as the management of cattle production) could help considerably to raise the incomes of poor Basotho households. The Government should also consider privatizing some services, such as veterinary care, so that more timely service could be provided to farmers. Government provision of tractor services seems an equally inefficient, unnecessary and private sector-replacing activity. Finally, there is a need to explore ways of providing assistance for intensive livestock activities involving pigs, poultry (eggs and broilers), angora rabbits, and fish ponds -- ways that do not put pressure on the degraded rangeland. These activities are generally undertaken by women, so strengthening them could raise incomes in households officially headed by women, which are often poor.

Land Tenure

3.42 Lesotho's small-scale agrarian structure and relatively equitable distribution of land is an important asset, both for agricultural growth and as a social safety net. Despite what is often asserted, small-scale farming is not less efficient than large-scale farming. Decades of international agricultural research have failed to demonstrate economies of scale beyond what a family can manage comfortably on its own. And this research clearly demonstrates that small- scale farming is not only more labor-intensive, it is also often more efficient -- as measured by profits per Dollar invested -- than large-scale farming. The most productive agricultural sectors in the world are dominated by small-scale, part-time, family farming. Most important is the pattern of rural and economic growth associated with the development of a small-scale farm sector: it is unambiguously pro-poor. So the small average farm size in Lesotho should be seen as an asset, not as an impediment to future agricultural growth. And secure rights to land are one of the most important social safety nets imaginable.

3.43 But farming households have little incentive to invest in irrigation or other productivity- enhancing land improvements unless they can be sure of occupying the land for a reasonable length of time. Such uncertainty also constrains rental markets for land, and resource-poor households will be unwilling to rent out their land to other households. Land tenure is uncertain under traditional law, whereby the rights to land use (allocated by the local chief) lapse at the end of a farming season, with no assurance to households that they will be allocated the same plot(s) of land for the next season. Since passage of the Land Act in 1979 -- and of amendments to it in the late 1980s and early 1990s -- this situation has changed, in theory. It is now possible for both male and female holder of customary land use rights to apply to the Government for a lease enabling them to keep the customary landholding for up to 90 years, which helps make land investments relatively secure. Lessees are responsible to the State (through the District Administrative Officer) for the maintenance of land and can use the lease as collateral for loans, - 61 - forfeiting their land use rights in case of default. Sub-leasingis also permissible, so those without the resources or desire to farm can sublet the land to someone who can make productive use of it.

3.44 However, there is considerable concern that in practice many farmers, especially the poorest, are unable to take advantage of the opportunities the Land Act affords. First of all, many farmers are not well-informed about the Act's provisions and the Government has made little systematic effort to disseminate easily understood information to farmers. This omission should be rectified as soon as possible through several communications channels -- for example, through the mass media, through village meetings or pitsos (public meetings at which district and ward officials could explain the Act), through agricultural extension workers, and through NGOs and community-based organizations active at the grassroots level. Second, the cost of preparing documentation for lease applications is an estimated MIl 0 for the first hectare, which is high enough to keep many of the poor from trying to establish secure land tenure. The Government should consider subsidizing application costs for the poor, so more of them can take advantage of the Act. Finally, although the Land Act expressly allows women to hold leases,5' their continued status as minors under Lesotho's customary and modern law hinders their ability to acquire land use rights from their chiefs in their own names, which, in turn makes it difficult for them to apply for government leases. Since more than half of all households in Lesotho are headed by women, these constraints seriously impede overall agricultural land use efficiency.

Livestock and Range Management

3.45 Replacing livestock, a savings asset with a low and risky financial return, with a more attractive financial asset could both contribute to raising living standards and help protect the environment. In Lesotho, livestock production is roughly as important as the production of field crops in terms of value-added. The livestock subsector's contribution to GDP declined from about 8 percent of GDP in 1985 to about 6.5 percent in 1990.52 Yet during the same time the number of animals actually rose. The number of cattle increased from about 525,000 in 1986 to about 641,000 in 1990, while there were about 1.5 million sheep and I million goats.

3.46 In Basotho tradition, cattle, sheep, goats, horses, and donkeys, are generally considered to be men's animals. Basotho households distinguish between men's, women's, and household property. Whatever a Mosotho woman makes using her own skills -- for example, through baking, brewing, or raising pigs or fowl (women's animals) -- is hers to spend. Men's earnings -- such as remittances -- belong to the household and can be spent by the woman on daily family needs even when the man is not present. But livestock belongs to the man in the sense that only he can sell it. Should he do so, however, the money he receives belongs to the family as a whole. So men have an incentive to invest in livestock, while women often have "modern attitudes" toward livestock, preferring assets over which they have more influence.

3.47 Men's livestock is valued for more than its potential to generate income. It is valued, among other reasons, because (1) animals are used to pay bohale or bridewealth at the time of marriage; (2) social prestige is accorded to livestock owners when, as is common, they place

s That is, to acquirethe rightsto landuse from a deceasedhusbank.

52 Lesotho:Statistical Yearbook, 1992, Bureau of Statistics,Maseru. - 62 - some animals with other households,allowing those householdsto benefit from productssuch as milk, wool and dung; (3) livestock visibly demonstratesto the community that an absentee husbandis still providing for his family; and (4) and perhapsmost important, livestock functions as a store of wealth or relatively illiquid saving that may be drawn upon only in times of great needor as a pension for old age. As a result, Basothomen have tendedto increaseherd sizes at every opportunity. Under the current system of communalownership of grazing land, the costs of overgrazingare low to the individual but high to society in terms of soil erosion and declining soil fertility.

3.48 Discourage overstocking. Over the years, Lesotho's government,assisted by donors (including FAO, , the ,, USAID, and the World Bank), has tried to reduce overstocking. Among other approaches,it has tried (I) increasingcommunity awareness about environment and land conservation; (2) instituting a system of grazing fees whereby livestock owners were chargeda nominal fee per head of livestock53 and part of the proceeds were allocatedto village developmentprojects underthe stewardshipof an electedseven-person Village Development Committee, with the local chief serving as ex officio chairperson; (3) supporting small-scale community soil and water conservation projects such as pasture improvementand tree planting; and(4) instituting culling programsfor unproductivelivestock.5'

3.49 Thesedestocking efforts have beenlargely unsuccessfulto date. Programweakness has been attributed to underfundingand problems in program management;"to the resistanceby many chiefs of the partial loss (to a Village DevelopmentCommittee) of their formerly sole authority to allocate grazing rights and make decisions on community projects; to the slow implementation of programs to train Village DevelopmentCommittees; to the low levels of grazing fees; and, paradoxically,to the perceptionin many communitiesthat grazing fees are a Government ploy to tax a traditionally "free" activity. As a result, the Government has temporarily suspendedthe collection of grazing fees and is reviewing all policy and program options for addressingthe relatedproblems of overstockingand soil erosion.

3.50 It is essentialthat a comprehensivepolicy be institutedto reduceoverstocking and hence retard soil degradation. Such a policy must raise the private cost of livestock ownership to something approaching the social cost, whether through grazing fees or another means. If grazing fees are reinstituted and the funds are used for village developmentprojects, recent experiencesuggests that conservationprojects such as tree planting may be more successfulif implementation shifts away from communal approaches,which have often led to abusesof power or (becauseof neglect)have proved unsustainableafter the investmentphase. Successis more likely if individual householdsbenefit directly from investments.For example,the broader social goal of controlling erosion could still be achievedif funds collected from grazing fees were used to pay householdsto plant trees on their own land, a practice that has worked in

531in 1991 the recommendedfee levelswere MO.5a head for sheep and goats,M3.0 for cattle,and MS.0 for horses,ponies and donkeys.

54 These effortshave been reviewedin reasonabledetail in a numberof reports,including the previouslycited 1993 WorldBank Environment and AgriculturalDiversification study.

55 At the level of the Ministry of Interior, ChieftainshipAffairs, and Rural Development,which has responsibilityfor overseeinggrazing fees, Village Development Committees, and small-scaleprojects. - 63 -

Lesotho in the past.56 Individuals are more likely to acceptfees from which they reap tangible benefits.

3.51 Introduction of alternative savings instruments. But the introduction of grazing fees may not noticeably reduceherds held mainly for savingspurposes, because people holding cattle as an assetmay care much lessabout their rate of return than about their cattle's function as an asset in which to invest savings. There is considerableevidence to suggestthat in Lesotho livestock may be held mainly for illiquidity, not income,and savingsbehavior is generally price inelastic.(Banks are familiar with this phenomenon:Changes in interestrates do not causelarge fluctuations in customers'savings accounts. Customersmay be more interestedin the type of savings instruImenta bank offers -- whetherit is secureand illiquid, for example,and helps them maintain savings discipline.) The logical complement to grazing fees, then, might be to introducea new savingsinstrument that mimics and could replacecattle as a savingsasset (Box 3.1). There is so little competition in Lesotho'sfinancial sector that banks are unlikely to be interested in experimenting with new types of rural savings schemes. In the short run, the Central Bank could use its influence with the commercial banks, especially the Lesotho Agricultural Development Bank (LADB), which has a large number of rural branches, to experimentwith alternative savingsinstruments. One promising option is long-term contractual savingsinstruments (certificate-of-deposit type instruments,for example, for pensionpurposes). In the long run, increasedcompetition is neededto encouragemore innovation in the financial sector.

56 Of course,private woodlots cannotbe promotedunless households have more secureland tenure. The successof forestryschemes is vital not only becausethey reduce erosion but becausemore access to fuelfrom woodlots would enablefarmers to useanimal dungas manureinstead of fuel andwould reducethe time poor householdsnow spend gatheringshrubs for cookingand heating (well over an houra day in manyhouseholds). - 64 -

Box 3.1: A Policy for Decreasing Herd Size in Lesotho The most important function cattle fulfill in Lesotho is as an asset in which to store household savings. The attractiveness of that asset is its relative illiquidity, next to its financial rate of return. Developing rural savings alternatives that "mimic" cattle as a savings asset, and pay a (moderately) positive real interest rate would make livestock less attractive as a capital asset, and would probably lead to reduced stocking rates.

Most households save mainly for a rainy day -- to ensure against the problems of fluctuating incomes and income shocks -- rather than as an investment from which to make profits, trading off risk against the expectation of high returns. When consumption-smoothing and insurance are the main reason for saving, saving will be relatively price-inelastic and risk aversion will be of primary importance. Attempts to make livestock production less profitable as a way of reducing herd size (e.g., by introducing grazing fees) is unlikely to work when livestock are owned as a hedge against income changes.

Where does the preference for keeping cattle as a store for savings come from? Households do not like to keep cash (for example, from miners' remittances) around the house for fear that they may spend it on minor or short-term emergencies or because friends and relatives make legitimate or not-so-legitimate demands on them. One advantage of having cattle as an asset for storing wealth is that cultural norTns against converting cattle to cash provide a built-in informal social insurance system, particularly useful when cash is in demand. Those same cultural norms facilitate the accumulation of cattle to provide retirement income. Adding to cattle's attractiveness as a fairly illiquid asset is the inelasticity of demand for cattle: When many households face the same income shock (such as a drought) and livestock sales are up, cattle prices fall. This makes selling cattle less attractive, and increases the incentive to keep the cattle until prices rise again (after the need to sell diminishes).

If one reason for keeping cattle is for financial discipline, another important reason is to get a good financial rate of return. In Lesotho, the rate of return on cattle assets has often been better than the rate of return on more traditional financial assets, which increases the attractiveness of cattle ownership. Between 1978 and 1984, for example, the average real prime interest rate in Lesotho was zero and the average real interest rate on savings was minus 10 percent, whereas the average real rate of return on livestock assets was 8.3 percent for cattle, 7.2 percent for sheep, and 10.1 percent for goats.' But today, interest rates on savings held in financial institutions are positive in real terms. And a recent study estimates rather low financial rates of return for holding cattle (3.3 percent), sheep (7.6 percent), and goats (-6.1 percent), even if the opportunity cost of labor is calculated as zero.2 Currently, then, the finanical returns on cattle holding seem low, and work against holding cattle for profit.

In this situation, the most obvious way to reduce demand for cattle owned as a savings asset is to offer an attractive savings alternative, such as medium or long-term deposits. High interest rates on longer- term savings deposits would be desirable but not essential, since the price elasticity of savings (when the motive is consumption-smoothing) is probably not high. More important is the security of the deposit and the difficulty of withdrawing it before the term expires. Swallow,et al.,(1987), Livestock Development and Range Utilization in Lesotho. ' Cole,P. (1993), Gross Margins for AgriculturalEnterprises in theMatelile Rural Development Project.. - 65 -

C. FOSTERING LABOR-INTENSIVE SMALL BUSINESS

3.52 Small and informal businessesare an important sourceof income and employment for many Basotho,especially in poorer households. Roughly 30 percentof all Basothohouseholds engage in some informal or small-scalenonfarm businessactivity and for about half of those who do, the businessis their main sourceof income. Nearly 103,000small businesses(including microenterprises)were operating in Lesotho in 1990,employing an estimated 161,000people. 5" Small businessesin Lesothotend to be very small but thereare so many of them that they are an important source of employment, providing more work to the Basotho than either the South African mines or the domesticformal sector.

3.53 Lack of start-up and working capital is a serious problem for small businessowners. Most small businessowners in Lesothouse their own savingsfor start-upcapital, partly because credit is not readily available. Most small businessescater almost exclusively to the domestic market, with a few exceptions:carpets, tapestry, leather goods, and other decorativehandicrafts. The majority specializesin manufacturing. Many of them seeweak product marketsand lack of marketing know-how as a problem. But they also have supply problems. Most raw materials must be imported; even the mohair made into sweatersand tapestries is usually processed outside the country, from locally producedmaterials. Small businessesin Lesothoare typically run by people with little formal educationand few managerialor technical skills -- even such basic businessskills as bookkeeping. The skills they do have are usually acquiredon the job, through "learning by doing." Other characteristicsof Lesotho's small business sector are summarizedin Table 3.4.

3.54 SuccessiveLesotho governmentshave generally neglecteddevelopment of the small businesssector, choosing insteadto develop local industry mainly through medium- and large- scale parastatals. Despite the preferential treatment parastatalshave received -- and maybe partly becauseof it -- many of them have operatedinefficiently and at great social cost, either through the undue pressurethey put on budget resourcesor through higher consumerprices for many products and services(the flour mills and the dairy plant are a good example). Most importantly, perhaps, the protection given to parastatalshas crowded out small businesses. Recently, stepshave beentaken to reducegovemment involvement in medium- and large-scale enterprisesand create more room for private sector development. A privatization program is scheduledto start mid-1995, with the assistanceof the World Bank and other donors. And growing recognition that more must be done to support the growth of the small business (including microenterprise)sector led to the establishmentin 1992of the BusinessAdvisory and Promotion Services(BAPS). BAPS has a mandateto coordinate initiatives to support small businessesand to identify new ways to encouragetheir development.

57 Informationhere is from a 1990survey of smallbusinesses commissioned by USAID through the GEMINI project:"Small-Scale Enterprises in Lesotho:Summary of a CountrywideSurvey," GEMINI (Growth and Equity through MicroenterpriseInvestments and Institutions) Project, Technical Report No. 14. This sectionalso draws on datafrom the 1993Sechaba household survey. - 66 -

Table 3.4: Characteristics of Lesotho's Small Businesses (1990) Characteristics Maseru Small Towns Rural RuralAreas Total Towns

1.Small-Scale Enterprises (SSEs) a. Shareof totalSSEs (%) 12.0 6.8 1.6 79.9 100.0 b. One-person-operatedSSEs (%) 76.0 74.0 72.0 81.0 79.6 c. SSEsw/ laborforce >5 (%) 6.5 4.1 1.8 1.4 2.2 e. Ownedby soleproprietors (°) 93.3 95.3 94.7 97.2 96.6

2. SSE'sLabor Force a. Numberof laborersper SSE 2.1 1.7 1.5 1.4 1.6 b. Shareof proprietors/family 87.4 83.0 81.2 84.5 85.2 members(%) c. Shareof women(%) 80.1 79.3 77.1 75.4 76.2

3. Distributionof SSEs a. Manufacturing(%) 36.1 40.8 52.5 62.8 58.0 b. Tradeor commerce(%) 40.7 43.7 34.5 26.9 29.8 c. Service(%) 23.2 15.5 13.0 10.3 12.2

4. Female-OwnedSSEs a. As % of all SSEs 75.6 76.3 74.8 71.2 73.2 b. Shareof employmentcreated (%) 50.4 58.1 56.5 59.9 58.6 Source:"Small-Scale Enterprises in Lesotho,"GEMINI TechnicalReport No. 14.

3.55 As a way to help reduce poverty, the Government's new focus on small businesses is welcome for several reasons. First, small businesses tend to create more jobs at less cost than medium- and large-scale industries, which are generally more capital-intensive. Second, growth in the output of microenterprises is likely to benefit the most vulnerable groups, who engage in small business when there are no other opportunities for higher-paying formal sector work. (Many of those who mention microenterprise or informal business as their main source of income are poor.) Third, many microenterprises allow flexible hours, which makes it easier for women to participate. Most women must juggle competing demands for their time, including farming, child-care responsibilities, and time-consuming household tasks. The flexibility of the microenterprise especially helps households headed by women. That work in small enterprises is feasible for women is apparent; many women already work in the sector (Table 3.4). Given the important role that small businesses can (and do) play in creating employment for the poor, the rest of this section focuses on ways to alleviate key constraints on small business growth and profitability.

Access to Capital

3.56 Small businessowners often assert that the main constrainton developmentof this sector is the unavailability of start-up and working capital, which in turn is the result of a paucity of - 67 -

formal credit. Married women who are entrepreneurs may have special difficulty getting loans through formal banking channels because under current Lesotho law they cannot legally enter into contracts without the express consent of their husbands (Box 3.2). So far, donors and the Government have failed to successfully channel funds to small-scale and microentrepreneurs; the two organizations designed to do so are in serious trouble. In 1980, the Government established the first of these two organizations: a special parastatal, the Basotho Enterprise Development Corporation (BEDCO). BEDCO was to channel funds to small local enterprises either as loans or as investments in the enterprises. At the same time it was to provide many of them with support in the form of technical and managerial assistance, machinery, and workspace. Considerable financial resources were available (from the Government and donors), but because of poor management BEDCO helped few enterprises in the course of the decade. Of those few, about 30 percent failed altogether, and most of the rest of them are in arrears on loan payments. So most donors have withdrawn support from BEDCO. Part of the problem lies in BEDCO's focus on attempting to create small manufacturing businesses which target the domestic market. This market is small and firms face strong competition from South African imports.

3.57 A second, newer effort, the Small-Scale Industries Project (SSIP), supported by UNDP, is trying to channel loans to small businesses through Lesotho Bank. Under SSIP, training is provided to intermediary agencies (BEDCO, the Ministry of Trade and Industry, and some NGOs) to help entrepreneurs make loan applications to Lesotho Bank, which is responsible for evaluating loan applications and monitoring loans. Loans are guaranteed up to 90 to 95 percent by SSIP and require no collateral. Interest rates are subsidized: SSIP charges 15 percent, compared to a commercial rate of 20. SSIP is probably helping to meet capital needs for some small businesses, but there is some concern that not enough rural entrepreneurs are being reached. And whether SSIP is sustainable over the long run is doubtful: it has been in effect for only four years and more than half of its 135 outstanding loans is already in default.

3.58 What are we to conclude from all this? First, modern institutions generally fail to reach the neediest part of the population with the small amounts of capital they need. They especially don't reach women or rural dwellers. Second, Basotho households outside the formal sector either cannot grasp or cope with the basic demands of bank borrowing (for example, taking out individual loans, paying interest, and making regular payments) or they view such loans as simple subsidies, because over the years Lesotho's financial institutions have repeatedly failed to enforce repayment. By contrast, traditional savings organizations seem to work. These organizations are predominantly rural-based, serve both men and women, and seem to be able to maintain financial discipline among their members, mainly through peer pressure. Two traditional savings organizations that seem to be effective in Lesotho are the stokvel and the burial society.

3.59 The stolkvel, which is relatively common in developing countries and known in the literature as rotating savings and credit association, functions in a simple, cost-effective way. A group of individuals with a common bond -- such as kinship, community, or employment -- agree to save together a fixed amount of cash over a fixed period. Then members of the group take turns using the funds accumulated. For example, ten people agree to save MI 0 a month for ten months. Each month a different member uses the combined savings of M 100, until everyone has had a turn. The borrowing order is negotiated prior to the savings contract, usually on the basis of greater need. The economic efficiency of this institution is simple and elegant for two main reasons. All but the last person in the cycle has access to a lump sum earlier than by saving - 68 -

for it individually, and all but the last person receives an interest-free loan. As long as group discipline is maintained, the system is a highly efficient method to mobilize business capital. The amount of capital which can be raised depends on the capacity of the group to increase membershiip or individual contributions and still maintain strict repayment discipline. Stokvels also often help members withi fundraising for, and marketing of, home-made products by hosting "parties" at which members may sell refreshments or other merchandise.

Box 3.2: Women and Small Business Women in Lesotho have limited legal rights. Under Roman Dutch Law, which is the general civil and commercial law in Lesotho, women under 21 are considered minors and cannot enter into a contract unless assisted by their guardian. Women who are married are regarded as adults but, since marriage involves "community property" and the husband is regarded as head of the family, he exercises "marital power" over his wife. This means she has no legal power without her husband's explicit consent. The husband must authorize all contracts entered into by his wife with the following exceptions. She:

* May purchase household necessities on her own. * May enter into contracts connected with public trade if her husband gives initial approval for her to enter into business. * May have a bank account in her own name and under her own control. * Can receive her own wages, if she is employed.

Under customary law, the head of household (the husband of a married woman) has ultimate responsibility for and authority over his family, including control of any household member's property. Contracts are understood to be made between whole families and another party, rather than between individuals, so they require the consent of the household head. And because many men are absent for long periods as migrant workers, their wives cannot easily get the consent needed for business transactions.

A woman cannot legally start a business or enter into a contract (including a loan) without the consent or signature of her husband or male "guardian." Even if a lender makes an exception about consent (as sometimes happens), the woman is still limited in her ability to get credit, as the only collateral she can typically offer is legally her husband's property. Moreover, a husband has the legal right to take his wife from a job or take her business from her at will. This seldom happens, but the prolonged physical separation of many wives and husbands sometimes fosters unstable marriages, which puts many women in a vulnerable position. If women succeed in business, they have few rights with which to safeguard their investments or to ensure their economic future. Women have a strong incentive to develop small-scale businesses that allow them to conceal their profits and develop informal working relationships.

3.60 Burial societies function much like insurance clubs, collecting a fixed monthly amount from members. The amount collected, usually by an elected treasurer, may be small -- for example, less than MI a month. Part of the accumulated funds are used when needed to defray the costs of the burial feast for someone in a member's family. Burial societies also organize periodic fundraising events such as raffles or generate funds by selling food or beer. All surplus funds are maintained as general reserves (rather than belonging to individual members). The treasurer often deposits them in the nearest bank; it is easier for one person to travel to town for a - 69 - group deposit than for severalindividuals to do so. Thereare more membersin a burial society than in a stokvel, often as many as 150. But as in stokvels,members usually have a common bond, suchas residencein the samevillage.

3.61 Stokvels and burial societies have been so successfulas traditional savings clubs that effectively promoting their growth in numbersmay be a sustainableway to give small business owners and micro-entrepreneursmore accessto neededcapital. The clubs directly provide capital to small businessesor help finance certain "lumpy" householdexpenses (such as health care costsor school fees, in addition to burial feasts)and thus free up householdincome for use in business.58 In promoting such expansion,it is important to ensurethat there is some bond between group members, as this bond is the main vehicle for financial discipline in the traditional savings clubs. BAPS is currently considering how to support the expansion of traditional savingsorganizations as a way to easecapital constraintsfor small businesses.Banks could promote the opening of joint accountsfor these clubs, and extend loans to them which could be fully collateralized by savings. And membershipof a group with an outstanding savingsand repaymentrecord could help qualify an individual for a personalloan. Theseefforts should be given high priority.

3.62 The Governmentshould also consider systematicallyinvolving NGOs and community- based organizations (CBOs) in assessingthe feasibility of options for involving traditional savings clubs more in the support of small business development.NGOs and CBOs have extensivegrassroots experience; their involvement in the designand implementationof savings schemesshould increasethe probability of success.In 1992,Women in Business(WIB), a local NGO, helped establisha small stokvel-like revolving fund for women entrepreneursusing dues collected from its roughly 190 members. It has been relatively successful so far; most beneficiariesmake paymentsroughly on schedule. WIB is now exploring (with assistancefrom CARE International) possibilities for developing further group lending options, under which WIB members, in groups of five to seven, would receive loans for productive enterprise investment. The proposedscheme is expectedto be modeled on such group schemesas the GrameenBank in Bangladeshand CARE's Maradi Project in Niger, which have done quite well both at providing capital for small businessesand at getting peopleto pay on time.

Training

3.63 Governmentand donor efforts to promotesmall businesshave so far focusedmainly on credit schemesand training. Skills training predominantlyfocused on small-scalemanufacturing and general "entrepreneurshiptraining" aimed at teaching basic businessskills and motivating participants to start small businesses.Most participants in both types of training -- provided through government technical and vocational training institutions59 and by NGOs -- lack experience in small businessand, judging from anecdotalevidence, most do not actually start small businesseswhen they finish training. Internationalexperience shows that pre-employment training is not an effective way to promotethe growth of small businesses,as traineesmay lack

58 Stokvelsmay not be especially useful for seasonal activities such as agriculture, because members would all needcash at aboutthe same time of year.But they do lend themselves to periodic microenterprises, including those that purchasegoods for retail,garment manufacturing, and so on.

59 Includingthe MOE-runLerotholi Polytechnic and the ThabaTseka Skills TrainingCenter, the MTI- supervisedLOIC and BEDCO, and several smaller institutions. - 70 - the entrepreneurialspirit neededto risk businessstart-ups, may not have accessto enough start- up capital, or may simply lack a good product idea. Usually, entrepreneursemerge from the ranksof the skilled labor pool alreadyemployed in existing firms.

3.64 More promising are smaller efforts by the Institute of Extra-Mural Studies([EMS) at the National University of Lesotho, by the Lesotho Distance Teaching Center (LDTC), and by business-orientedNGOs (such as the LesothoChamber of Commerceand Women in Business) to target shorter, more focused seminarsat owner-operatorsof existing businessesin specific skills such as accountingor marketing -- skills with which a numberof small businessowners say they needassistance. The opportunity cost of taking time away from a businessto attend a seminar is so higihthat small businessoperators are likely to attend theseseminars only if they think the traininig will directly improve their ability to generateincome. There has been no comprehensiveevaluation of small business-orientedtraining in Lesotho,so it is difficult to assessthe relative effectivenessof various approachesthere. The evaluationsthat do exist -- usually of individual training courses-- usually review the training itself, not its impact on small business development. Before more investmentsare made in training, there should be a thoroughassessment of how training hasaffected the small businesssector.

3.65 International experience suggests,however, that training should shift emphasisaway from its currentexclusive focuson small-scalemanufacturing to also include skills development in the service sectors. Training programsshould also be developedin collaboration with the large,export-oriented firms and include on-the-jobtraining aimed to createand further developa critical massof skilled, supervisoryemployees in existing large firms.

Markets

3.66 Small manufacturersoften run into marketing problems. The solution proposed to remedy these problems is generally to provide training in marketing skills. However, small manufacturers(and small farmers)generally do not market their productsthemselves, but rely on middlemen andtraders for information aboutthe demandside of the market. Hence,to promote the marketingof small businessproducts other initiatives than training in marketing techniques should be considered.

3.67 Policymakersshould seriously considerupgrading urbanmarkets to ensurethat there is physical infrastructurethat can accommodatea wider rangeof entrepreneursthan is currently the case. The Governmentcould also assistin organizing a rural periodic market system, common in other African countries but virtually absentin Lesotho. Periodic rural marketswould give local small businessoperators an opportunity to sell and advertise their merchandiseto more clients and to get valuable information on demand,supply, and prices. The existenceof rural markets would encourage suppliers of inputs to set up stalls, thus reducing the high transportationand other costs rural businessowners now incur traveling to urban centers(or sometimesto South Africa) to purchaseinputs. Such markets would especially help female entrepreneurswho must often pay for child-care or other serviceswhen they make long trips to buy supplies. Efforts to promoterural marketsfor small businessesare unlikely to encouragethe growth of small businessesif they are coupledwith efforts to increasecompliance with licensing requirementsor other regulationsgoverning small businesses.Several such regulations exist, but because enforcement is lax, the regulations may not constitute a major constraint on developmentof the small businesssector. - 71 -

D. DEVELOPING LABOR-INTENSIVE INDUSTRY

3.68 Growth of mediumand large-scale labor-intensive industries in the formal sectorcould greatlyimprove the incomeand welfare of Lesotho'spoor. As the SIO analysisshows, a I Loti increasein value-addedin thetextiles industry (including clothing) produces an MO.806increase in the incomesof the poor. Growth in the formal sector will come through growth in manufacturingand tourism.

Manufacturing

3.69 Sinceindependence, growth in formal industry,especially large-scale manufacturing, hasbeen a principalelement of Govemmentefforts to createjobs for the Basothopeople. These effortsfocused initially on developingrelatively capital-intensive industries for thehome market (for example,a brewery,a bakery,and flour mills) as part of the emphasison nationalself- sufficiency,and on the exportof suchproducts as mohair. In the early 1980sthe focusshifted toward promotingdirect foreign investmentin labor-intensiveindustries -- an appropriate emphasisgiven the country'sshortage of capitalbut abundanceof largelyunskilled but literate (hencetrainable) workers. In the first yearsof the newapproach, foreign investment (especially by Far Easterninvestors) increased significantly, thanks to a strategyof foreign promotional visits,an incentivepackage featuring a tax holiday,and (especially) Lesotho's preferential access to Europeanand American markets.' This growthcontinued and was boosted after 1985,when sanctionsagainst South Africa promptedSouth African investorsto cometo Lesotho.Between 1980and 1994,real growth in manufacturingaveraged about 12 percenta year,and growth in labor-intensive,foreign-owned industries accounted for much of the increase. The new investmentswere mostly in clothingand footwear, which in July 1993accounted for 12,151(76 percent)of the 15,972jobs in firms assistedby the LesothoNational Development Corporation (LNDC).

3.70 CurrentlyLesotho ranks third out of ten Africancountries exporting apparel and textiles to the UnitedStates. 6' Sofar, Lesotho'sstrategy for creatingmore formal sector jobs andgrowth in exportsof manufacturedgoods has largely succeeded, but therehas been some concern that externalfactors will slowdown the rateof growth. For onething, the lifting of sanctionsagainst SouthAfrica hasmade investments in Lesotholess attractive for regionalinvestors. At the same time,past U.S. quotas 62 on someLesotho clothing exports, as well as problems(under the Lome Convention)with domesticvalue-added on exportsto Europe,may have had a negativeeffect on

60 Lesothomanufacturers enjoy preferentialaccess to Europeanmarkets under the terms of the Lome Conventionand to the Americanmarket under the GeneralSystem of Preferences,which appliesto low-income countries. 61 In 1994,Lesotho exported 17.7 million squaremetres to the U.S.,compared to 17.1for Kenya,23.8 for SouthAfrica, and 50.4 for Mauritius.

62 The U.S. quotaon textilecategories 338/339, 638/639, and 347/348were suspended in late 1994,after havingfilled about 90%of the quota. At thewriting of this report,no quota call hasbeen made by the Committeefor Implementationof Textile Agreementsor the U.S. TradeRepresentatives Office for consultationsconceming new quotas. - 72 - the investment plans of some foreign investors. But international experience suggests that the U.S. quota system, although disruptive, has never prevented a particular country from continuing to export to the U.S. market. Continuous switching to non-quota items -- for instance, from jeans to shirts -- reduces much of the "bite" of the system. Hence, domestic issues seem to be the most important factors affecting growth in manufacturing in coming years.

3.71 Expanding and integrating labor-intensive light manufacturing is the most promising component of Lesotho's industrial strategy. It could create direct and indirect jobs on a large scale. Local businesses can emerge from the pool of skilled, experienced machinists, technicians, and supervisors already working in large export firms. Continued expansion of the export industry will also increase demand for services as varied as catering, machine repair, and building construction and maintenance. Local sourcing of more inputs (such as packaging and transport) would also give local business new opportunities. And some production (for example, embroidering) could be given out as piece work to local subcontractors. Hence, if export industries continue to expand, the ripple effects will be felt throughout the economy and will lead to reduced unemployment and higher wages -- both crucial to poverty reduction. To keep the cost of doing business down, Lesotho should maintain the minimum wage at competitive levels, improve industrial relations, and regulate utility cost pricing.

3.72 Wages. There appears to have been some erosion of Lesotho's comparative advantage (for foreign investors) of lower wages than South Africa and more labor stability. Wages for assembly workers in Lesotho average about R300-350 a month, which is 50 to 75 percent lower than in South Africa. Until recently, government involvement in the labor market was limited largely to setting minimum wages for various types of work (Annex 12) and minimum wage policies were by and large pragmatic and hence no impediment either to industrial growth or to job growth, a key factor in poverty reduction. From 1989 to August 1992, however, minimum wages rose 25 percent, and many export manufacturers have complained that wages are too high for productivity performance. Of course, employers always complain that wages are too high, just as workers always complain they are too low, but if wages are too high, the implications for poverty reduction could be serious (Box 3.3).

3.73 Although the Government should refrain from administratively setting private wage scales, some would argue that it has a social obligation to set the minimum wage. But it should avoid the temptation to set the minimum wage at or near South African levels. Doing so would merely result in job rationing, more unemployment or parallel labor markets, and a loss of foreign direct investment. Export industries were attracted to Lesotho because of its preferential trade agreements and the low cost of doing business in the country. The cost of business could rise, however, if increased union activity and the excessive salary demands now emanating from civil servants raise the wages of those employed in the formal employment to the levels prevailing in South Africa. Lesotho would then mimic South Africa's inefficient growth path and associated social problems: high unemployment for most, along with higher wages and better working conditions for those fortunate enough to find employment in the formal sector. It is important that Lesotho not fall into this trap. - 73 -

Box 3.3: Minimum Wages and Employment In a labor-surplus economy such as Lesotho's, reducing poverty through industrial development means creating more jobs rather than increasing wages for the people already employed, who tend to be better-off. Increasing the minimum wage, however well-intentioned, can work against the goal of reducing poverty. For one thing, employers may decide to make their operations less labor-intensive and more capital-intensive. For another, if labor costs grow faster than labor productivity, firms become less competitive, which depresses output. Either way, fewer jobs are likely to be created, which hurts the poor.

Industrialists in Lesotho (especially exporters) have strongly voiced their objections to the 25 percent increase (over 1989 levels) in minimum wages announced in 1992, even though the increase was lower than the increase in domestic , because it was not related to productivity. If the claims of export firms are true, the increase in minimum wages could have made them less competitive internationally, because the Dollar value of the Loti fell only 8.7 percent between 1989 and 1992. It is hard to know if their claims are true, because there are no reliable data on factor productivity and unit labor costs. More recently, public sector employees have started putting pressure on the Government to raise their wages to South African levels. Such a move would spill over into the private sector and greatly undermine Lesotho's competitiveness. The Government should pay serious attention to these issues, as most export manufacturers in Lesotho do not have strong links to the domestic economy and they could relocate easily -- at a huge cost to Lesotho in terms of domestic jobs.

3.74 Industrial relations. Other labor market regulations, by indirectly raising labor costs, could also limit the creation of new jobs. There is some concern, for example, that the labor code instituted in April 1993, if enforced strictly, could sharply increase the cost of doing business in Lesotho. As it stands, the code (I) establishes and outlines the functions of new institutions, including the Wages Advisory Board, the National Advisory Committee on Labor, the National Advisory Council for Occupational Safety, and a Labor Court for settling disputes between employers and employees; (2) specifies details about employment contracts, termination, dismissal and severance pay, leave, and standards for the employment of women; (3) limits and regulates the activities of Labor Agents; (4) regulates contracts for service abroad and for the employment of foreigners in Lesotho; (5) tries to define "unfair labor practices"; and (6) establishes procedures and regulations for the establishment, registration, and rights and responsibilities of trade unions (which have begun gaining strength in Lesotho). Some aspects of this labor code can and should be enforced -- especially considering the difficult labor relations in some export manufacturing concerns -- but the Government should be careful that any government interventions seem even-handed to both workers and employers and that costs to employers not rise to the point that firms shilt toward more capital-intensive production technologies or leave Lesotho. Moreover, Government should also actively promote an improvement in industrial relations, possibly along the lines of East Asia's experience with deliberation councils (Box 3.4). - 74 -

Box 3.4: EastAsia's DeliberationCouncils Theexperience of thehigh-performing East Asian economies in managingindustrial relations through a type of institution known as "deliberation councils" seemsgenerally positive. Japan's efforts to establish deliberation councils are the most widely recognizedand have beenthe most thorough. Since the beginning of the postwar period, the Japanesegovernment and private industry have engagedin seriouspolicy deliberations through deliberation councils of two types. The first is organizedalong functional or thematic lines, such as pollution or finance; the second is organizedaccording to industry, for example, automobiles or chemicals. Governmentministries establishthe councils, which are generally associatedwith a specific bureauwithin the establishingministry.

In Korea, from the mid 1960sto the early 1980s,government and private sector relations were also close and cooperative,although some in the private sector arguethat governmentwas too strong-handedand dictatorial. Governmentand businessleaders met often and regularly although less formally than in Japan. Government solicited business' views and included them as a critical policy component. Similar consultative mechanismsappear to be evolving in Malaysiaand Thailand.

Public-private consultation is most explicit in Singapore. The private sector participates in policy making in many ways. For example,private citizens serve as directorson governmentstatutory boardsand as membersof ad hoc governmentadvisory committees. In both capacities,they review policies and programs and make recommendationsfor official consideration. In addition, the Governmentregularly invites chambers of commerce,trade associations,and professionalsocieties to submit their views on specific issues.

The Singaporegovernment's efforts to managethe economy are most fully apparentin the National WagesCouncil (NWC), which includesrepresentatives from government,business, and labor. Becauseof this tripartite structure, the NWC fulfills several coordination functions, simultaneously furthering the Government'sguidance of businessand of labor, as well as business-laborcooperation. Having rejectedmore typical democratic forms as too adversarial,the Government has also establisheda wide variety of citizen associationsto provide channelsof communicationwith the Government.

Deliberation councils naturally reflect the history and culture of the society in which they operate. Even so, the experience in the high-performing East Asian economiessuggest that their applicability is not limited, as someanalysts may suggest,to the Confuciancultures of NortheastAsia.

Source: TheEast Asian Miracle: Economic Grovth and Public Policy. World Bank, 1993. - 75 -

Table 3.5: Lesotho's High Utility Rates(in Rand Equivalent,September 1992) South Africa Swaziland Lesotho Botswana Durban Cape Town Electricity (Rand/kwh) 0.05 0.05 0.23 0.20 0.17 Water (Rand/kiloliter) 1.36 na 1.45 3.00 5.33 Telephonecall to (Rand/minute) USA 6.2 6.2 5.65 10.80 11.40 UK/Ireland 6.2 6.2 5.65 9.00 9.39 Lesotho 5.4 5.4 2.10 local 2.81 Botswana 5.4 5.4 2.10 .80 local Swaziland 5.4 5.4 local 1.80 3.81 South Africa local local 2.10 1.80 2.07 Source:LNDC

3.75 Utility costs. The fact that utility (electricity, water and telecommunications) rates are higher in Lesotho than in South Africa, and occasionally than in Botswana and Swaziland, could limit industrial growth and employment in Lesotho. Lesotho's high utility rates (Table 3.5) are a tremendous disincentive to potential investors who might otherwise consider investing in Lesotho because of its lower wages. If Lesotho continues to maintain fully self-sufficient utility companies -- a policy justified in the past by the fear of dependence on hostile, apartheid- oriented South Africa -- high utility rates are likely to continue, because Lesotho's small market does not allow for economies of scale. But recent political changes in South Africa now make it more pragmatic for Lesotho to consider integrating with South African utility services to lower costs. At a minimum, govemment should regulate utility cost pricing. Several utilities, including telecommunications, have increased their number of employees without increases in productivity, while at the same time pursuing full cost recovery from consumers. Cutting utility costswould benefit not only large-scale industry but also farmers (who use electricity to power irrigation pumps) and many small-scale enterprises.

3.76 The manufacturing sector has been responsible for most of the jobs created in the formal sector in Lesotho in the past 10 to 15 years, but maintaining that job growth requires protecting and perhaps improving the country's competitive position. Many export-oriented manufacturing concerns do not have strong links with the domestic economy, which makes the option of relocation easier. Some risk factors lie beyond government control, but the Government does determine policy on the minimum wage, on labor regulation, and on utility costs. And decisions on those issues could affect export growth, job growth, and hence the country's ability to reduce poverty.

Tourism

3.77 Lesotho could attract far more tourists than it does, because of its location and its scenic beauty, but tourism has been largely neglected so far. Unlike other tertiary economic activities, tourism is an autonomous (rather than derived) source of growth, so it can generate substantial employment for skilled, semi-skilled, and unskilled workers in a range of sectors, including commerce, transport, catering, accommodations, handicrafts, horticulture and personal services. Developing Lesotho's tourist industry, by producing both jobs and foreign exchange, could certainly help reduce poverty. - 76 -

3.78 So far, efforts to promote tourism in Lesotho have focused largely on attracting South African visitors on short, border-hopping gambling excursions.63 In recent years, however, Lesotho has encountered increased competition from new casino-operating hotels in South Africa's former homelands (one of which is near Lesotho's border). Since 1980, tourist receipts - - tourist spending on hotel rooms, hotel food, handicrafts, car rentals, Lesotho Tourist Board tours, and poniytrekking -- as a share of GDP has been falling, from an already low 3.2 percent of GDP in 1980 to 2.1 percent in 1991. Bed occupancy rates in Lesotho's hotels were only 29.9 percent in 1992, which indicates substantial excess capacity. In marketing Lesotho as a holiday destination and encouraging longer stays, more emphasis could be placed on promoting (1) the country's temperate climate, dramatic topography, and special scenic attractions, such as Africa's highest waterfall (at Semonkong) and the huge mountain lake to be created on the Malibamats'o river by the Katse Dam (now accessible by a good tarred road); (2) the scope for "adventure" activities such as ponly trekking, hiking, trout fishing, rock- and mountain-climbing, and bird- and wildlife-watching; and (3) Lesotho's unique cultural and historical traditions and its relatively wide range of handicrafts.

3.79 Efforts to expand tourism in Lesotho should -- with the cooperation of regional and local entrepreneurs experienced in travel promotion -- aim both to increase the flow of regional holiday visitors and to capitalize on the increased international tourism to South Africa that recent political changes make likely. The place to start may be a review of operations at the Lesotho Tourist Board, because some observers believe that agency's weak capacity has been a brake on the development of tourism. It is also important to develop adequate infrastructure (including roads that give easier access to mountain areas) as well as hotels, restaurants, and entertainment facilities. It will be necessary to simplify visa and other border requirements and to harmonize them with South Africa's; Lesotho should certainly not require more than South Africa does, if it wants to become a key side trip for visitors to South Africa. More airline activity to and from Maseru airport (from South African tourist hubs) should be encouraged. To achieve better integration with the now booming South African tourist sector, the Lesotho Tourist Board should attempt to have Lesotho be incorporated as part of the packages offered by South African tourist agencies. Finally, environmental factors should be carefully assessed on every development so that tourism does not develop in a way that damages the country's natural beauty in the long run.

E. FOSTERING LABOR-INTENSIVE CONSTRUCTIONOF INFRASTRUCTURE

3.80 In terms of the budget, creating jobs in the private sector is generally better than adding public sector jobs, but when poverty reduction is the goal, increasing employment in some public services or on temporary public work programs can be worthwhile. Especially important are projects to develop and maintain infrastructure of social and economic value -- roads, bridges, dams, schools, health clinics, etc. -- that are generally considered the responsibility of the public sector. Public works projects can be an especially important source of work for poorer households, as most public works jobs can be done by people with little or no education and few vocational skills. Priority should be given to projects that require labor-intensive construction

63 Nearly 85 percentof all holiday visitors in 1987-89were South African. - 77 -

(rural road construction using manual labor, rather than road construction with heavy machinery).

Types of Infrastucture

3.81 Ranking projects in priority is beyond the scope of this report. The following criteria should be used in identifying, selecting, and developing a public works project: Does it provide work for the most vulnerable (for example, those who live in poor rural regions)? Is it socially or economically valuable? Is the ratio of labor to capital high? Are administrative and management costs low? Does it create permanent jobs? Is it replicable? The options for increasing public works discussed here are mainly those that can be done in the existing institutional and policy framework in Lesotho. Chapter 5 suggests others that can be undertaken if the Government initiates a major public works initiative as a way to strengthen Lesotho's safety net for the poor.

3.82 An example of a public works activity which could probably be expanded relatively quickly is the Labour Construction Unit (LCU). The unit is responsible for building and maintaining secondary and feeder roads using labor-intensive construction methods. Public works projects can generally create jobs more quickly than agriculture, small businesses, or formal industry can, but in this case implementation has been slow, largely because of budget constraints. In 1992/93, for example, LCU achieved less than half its target for rural road maintenance (about 750 km), regravelling and upgrading (about 110 km), and new construction (about 80 km), mainly because of budget shortfalls."I Had that target been met, some 2,500 to 3,000 additional jobs might have been created, benefiting mainly poor rural households. The indirect costs to poor households of this failure to meet these targets may have been equally high. Growth in the output of agriculture and off-farm enterprises depends heavily on access to markets, access that is greatly facilitated by a well-maintained network of rural roads.

3.83 Other labor-intensive works schemes that warrant greater coverage include water and sewerage services. The water supply system for Maseru urgently needs rehabilitation and expansion if supply failures in the near future are to be avoided. And less than 35 percent of rural residents have access to safe water, so there is room for substantial additional public investment in the rural water supply. Responsibility for the rehabilitation of schools, clinics, and other public buildings, which is labor-intensive, also lies with the Ministry of Public Works.

Privatization

3.84 Budget shortfalls inhibit the acceleration of public works projects but so does a shortage of skilled technical, managerial, and supervisory workers in government to implement projects. This bottleneck could be alleviated by making more use of the private sector. There is a precedent for this in the schools construction program, in which relatively few long-term government employees manage a relatively extensive program of school construction and expansion by contracting construction out to local private enterprises, who work from designs approved by the Ministry of Public Works. Efforts are also being initiated under the World Bank-supported Infrastructure Engineering Project to involve local private contractors in road

64The targets(in parentheses)are among those presented in theLCU/Ministry of Workssubmission to donors in September 1991. - 78 - construction and maintenance, by giving about 25 companies training both in road construction and in the use of labor-intensive techniques. Similar initiatives are possible in other areas, for example, water supply, and should receive high priority. The Government should also consider involving (after training) NGOs and district governments in the implementation of public works schemes.

Lesotho Highlands Water Project

3.85 New jobs were created in Phase IA of the Lesotho Highland Water Project. At the time of project appraisal, an estimated 4,700 jobs were to be created in construction activities directly related to the project and another 1,800 in spin-off activities, including road and housing construction, transport, trade and food production.65 About 70 percent of these jobs were expected to use unskilled and semi-skilled workers and hence to be suitable for poor Basotho. Many observers in Lesotho believe that Lesotho Highland Water Project contractors have brought in unskilled and semi-skilled expatriate workers from South Africa rather than recruit locally, so that the number of Basotho actually employed by the project is considerably lower than expectations. But this view is not borne out by employment figures compiled by the Lesotho Highlands Development Authority (LHDA). LHDA data for 1993/94 on employment in activities directly related to the project show that workers are overwhelmingly from Lesotho (Table 3.6).

3.86 The project has largely failed, however, to use a significant number of local entrepreneurs as contractors or subcontractors in its main construction activities. Partly, this is due to the complexity and speed requirements of some of the project's components. Few Basotho entrepreneurs would have the capacity to successfully complete the job. Partly, however, this is due to the specific formal requirements for contractors' bids on the main construction contracts. The labor costs reflected in these bids needed to be pegged at Orange Free State wage scales. This requirement neutralized the only competitive advantage Basotho entrepreneurs held over their South African counterparts.

3.87 Data are still being collected on employment in spin-off activities but a similar proportion of Basotho and foreign workers is probable: spin-off activities are generally less complex than direct construction activities and hence even less likely to require foreign expertise or skills. Still, it is unclear whether the project has generated the level of spin-off activities originally projected. LHDA intends to commission a study of spin-off effects in late 1994, with a view to determining what can be done to increase spin-offs in Phase IB. This should be given high priority.

65 Source:Coopers and Lybrand, "Lesotho Highlands Economic Impact Study," 1989. - 79 -

Table 3.6: Proportion of Basotho Employed in Lesotho Highland Water Project Construction (March 1994)

LHWPComponent Lesotho ExpatriateWorkers Totalemployment workers SouthAfrica Other Sub-Total

Deliverytunnel south 353 28 6 34 381 Consultants 123 na na 94 217 KatseDam 1,923 89 134 223 2,146 Transfertunnel 96 274 136 400 496 LHDA staff 305 na na 37 342 Total 2,820 276 364 877 3,697 Source:LHDA. Note:Figures cited include skilled, semi-skilled. and unskilled workers.

3.88 The Lesotho Highland Water Project is a capital-intensive project. Its importance to poverty reduction in Lesotholies more in the revenuesLesotho will get from increasedSACU receipts and water royalties than in the jobs it will generate. In the next several years, Fund revenuesare expectedto grow rapidly as constructionactivities (and associatedSACU revenues) intensify and as water royalties begin to be collected (Table 3.7). The Governmentshould be giving high priority to designing a viable, poverty-reducing program of investments to be financed from Lesotho Highland Water Project revenues, a major portion of which are maintained in a specialDevelopment Fund intended to finance local developmentprojects. Over the next 50 years, current allocation rules imply that about M120 million per year would flow into this fund. By June 1, 1994,roughly Ml S0 million had accumulatedin the LesothoHighland Water Project DevelopmentFund -- principally from increasedSACU receiptsassociated with imports for Lesotho Highland Water Project construction activities. Disbursementson local development projects have started, but the mechanismsby which projects are identified, managed and monitored require further elaboration. Labor-intensive public works projects should receivehigh priority.

Table 3.7: Projected Revenues from the Lesotho Highland Water Project (1991-2040)

_ 1991-20401 1991 2000 I 2020 I 2040 In 1991Maloti millions: 167 47 127 206 202 As a proportionof GDP: 5% 4% 8% 7% 4% Source:World Bank Staff Appraisal Report: Lesotho Highlands Water Project (Phase IA), July 2, 1991. Note:Assumes all projectphases are implemented. a/annual average. - 80 -

F. POLICYRECOMMENDATIONS

3.89 The preceding analysis suggests the following policy recommendations.

3.90 Agriculture and rural development. A shift from traditional food crops to more labor- intensive and high-value nontraditional horticulture and livestock products would benefit the poor through increased output and employment, higher returns to labor, and lower basic food prices. Such a shift would require that the Government:

* Remove price distortions and restrictions on marketing, processing, and trade. These distortions now favor the production of traditional food crops (especially maize) at the expense of higher-value horticultural crops, other nontraditional crops, and livestock products for domestic consumption and export. The existing restrictions also prevent a stronger private trading and processing sector from developing--a key ingredient of agricultural development. These reforms would also substantially reduce consumer prices for maize, an important consideration in a country where more than 90 percent of rural and all urban households are net consumers of maize.

* Reorient agricultural extension and research (in collaboration with South Africa) geared to support nontraditional and drought-resistant crops and intensive livestock activities, especially those suitable for women and mountain areas.

e Invest in rural infrastructure that supports agricultural intensification. Examples include roads and small-scale irrigation schemes designed to operate on a household rather than community basis.

e Make land tenure secure and land rental markets more flexible. Publicize and actively implement the provisions of the Land Act (including provisions about women's rights to hold land) and consider options for targeted subsidies to reduce initial costs of entering into lease arrangements (e.g., relax survey requirements for plots below a certain size).

* Promote smaller herds (as a way to stem declining soil fertility and reorient savings behavior), partly by reintroducing a modified system of grazing fees and partly by exploring savings alternatives so that cattle need not be a household's main savings assets.

3.91 Snall business sector. More attention is required to foster efficient growth in this sector, given its high labor-to-capital ratio, the large share (over 20 percent) of the workforce employed in small enterprises, and their disproportionate importance among the poor. Short- and medium-term actions likely to encourage development of small enterprises include the following:

* Improve access to capital, especially in rural areas and for women, principally by shifting away from bank loans administered by the modern sector (unsustainable in the past) toward expansion of traditional savings clubs such as stokvels, burial - 81 -

societies, or other revolving-credit schemes that rely on a common bond among members and peer pressure to maintain financial discipline.

* Shift the emphasis in business training and advice away from broad entrepreneurship training in small-scale manufacturing toward skills needed in the export and service sectors. Target the training to individuals already engaged as workers in large enterprises or already engaged in small businesses.

* Develop a systenmof periodic rural markets -- now virtually nonexistent in Lesotho - - to attract traders and to give operators of small enterprises the opportunity to sell and advertise their merchandise to a wider clientele.

3.92 Export-oriented light manufacturing. Past emphasis on developing relatively capital- intensive and inefficient parastatals has given way in recent years to systematic efforts to attract labor-intensive light manufacturing enterprises. Bolstered by favorable external factors (such as sanctions against South Africa and preferential access to U.S. and EU markets), about 5,500 new jobs have been created in the last five years. The challenge is to maintain the country's competitive edge in a changed domestic and international environment, to keep the capital that exists, and to attract new investments. Areas for action include the following:

* Ensure that public sector wage demands (which will spill over into the private sector), and minimum-wage and other labor legislation enacted does not raise labor costs to manufacturers so much that they cannot compete and are encouraged to move out of Lesotho or shift to more capital-intensive modes of production, thereby reducing job opportunities for the poor.

* lmprove industrial relations by establishing institutional channels for regular communication between entrepreneurs, organized labor, and government, and maintain an even-handed approach to labor and management in disputes brought before the newly established labor court and other labor institutions.

e Lower utility costs. Consider integrating certain utility services (e.g. telecommunications, electricity) with those in South Africa -- an approach made more feasible by recent political changes -- as a way of lowering utility costs, which are several times as high in Lesotho as in South Africa. Combinations of increased regulation, privatization and decentralization may also help increase the efficiency of utility services.

3.93 Promote tourisn. Lesotho's ecology and mountaneous character enjoy increased popularity in international tourist markets. Collaboration with South African travel agents rather than by marketing Lesotho overseas as an independent destination seems the appropriate policy. Tourism could generate many job opportunities for unskilled and semi-skilled workers now that tourism in South Africa is rapidly growing.

3.94 Invest in labor-intensive infrastructure. Public works projects could benefit the poor directly (by creating jobs) and indirectly (by stimulating economic activities and improving living standards). But budgetary resources for labor-intensive infrastructure construction - 82 - activities have been limited, while substantial funding has been targeted to activities that create few jobs (for example, through subsidies to parastatals). The following actions would improve the poverty-reducing potential of the infrastructure sector:

* Increase budget allocations for labor-intensive public works projects targeted io rural development. Criteria for project selection would include the following: an ability to create jobs for the poor, especially permanent jobs; output with high economic or social value; a high labor/capital ratio; low management costs; and replicability.

* Design and management of such public works should be decentralized in order to ensure efficiency and a high labor-intensity. Cooperating more with private contractors and training them in labor-intensive construction techniques can also improve the efficiency of public works programs.

* Design and inmplemnenta decentralized, labor intensive il?fraslruclure investnmenl progranmto be funded by SACU receipts and water royalties accumulated in the Lesotho Highlands Water Project Development Fund. - 83 -

4. INVESTING IN HUMAN RESOURCES

A. INTRODUCTION

4.1 To break the cycle of poverty, Lesotho must develop its human resources. Lesotho's record on human resource development has been more favorable than many other African countries at a similar level of development. For a least a decade, one Government after another has recognized that human resource development is central to Lesotho's development prospects and allocations for social services have steadily increased. Over the last ten years, real per capita public spending on health care and education has increased dramatically, despite rapid population growth; allocations for those sectors climbed from under 20 percent of public spending in 1982/83 to nearly 30 percent in 1993/94. This financial commitment, combined with generally favorable sectoral policies, has enabled Lesotho to expand access to basic social services considerably. Lesotho has established a broad-based, decentralized health care system and has significantly expanded primary health care services targeted at major childhood and adult diseases. It now provides more potable water for urban and rural citizens. And school enrollment has increased substantially: 30 percent in primary schools and 70 percent in secondary schools. Without question, the quality of life in Lesotho has improved.

4.2 Still, the general level of human resource development is low, especially among the poor, and consistent, effective action is needed to stem declines in the quality and coverage of services. As experience in countries such as Honduras, Sri Lanka, and Zimbabwe shows, resources can be used more cost-effectively, thereby giving the poor better coverage and producing better sectoral outcomes generally. Outlined in this chapter are key areas for change in health care, water and sanitation services, and education.

4.3 In heal/ti care, the Government could focus on three priorities to help reduce poverty. It could:

* Make the geographical distribution of financial, physical, and human resources more equitable. This can be done by (1) focusing new investments (and related recurrent allocations) in underserved areas -- principally the rural mountain and Senqu valley regions; (2) reviewing the distribution of government, donor, and NGO resources among levels of care in order to increase allocations to primary care; (3) developing an appropriate system of monetary and professional incentives to encourage health personnel to work in underserved areas; and (4) strengthening community-based approaches to primary health care in areas where health-care staff shortages are likely to persist.

* Make the cost-recovery system in health care more equitable: make it easier for the poor and other vulnerable groups, such as children, to use health care services. The Government can do this by (1) reviewing current inequitable and overly stringent exemption policies with an eye to replacing the current exemption for health workers and their families with an exemption for children under five, elderly people, and the physically or mentally disabled; (2) correcting the current bias favoring better-off urban residents by increasing cost recovery at hospitals, which the poor rarely use, and reducing cost recovery for basic services at lower-level facilities (such as - 84 -

clinics) used more often by the poor; and (3) exploring options to allow fees paid at a facility to be used by that facility, so fee payers benefit from the improved quality of care their fees make possible.

Pay mioreattention to important but underemphasized health problems. It is crucial to put more emphasis on AIDS prevention, given the potential for an epidemic and the fact that AIDS tends to strike adults in their prime, leaving whole households without a means of support. And considering the known links between poverty and alcohol abuse, the Government should probably also study ways to address problems of alcoholism.

4.4 The Government could consider the following measures to give the poor better access to water and sanitation services:

* Review program priorities in the water sector and find ways to make water distribution more equitable geographically. Safe water is far less available in the poorest rural districts, Thaba Tseka and Mokhotlong, and the newer areas of fast- growing towns such as Maseru and Maputsoe.

* Steadily increase government allocations for maintenance of wvater systenms and gradually reduce the current dependence on donor resources, to ensure that sectoral investments are sustainable.

* Strengthen local staffingfor central and district management of water and sanitation programs to ensure a smooth transition when donor-funded expatriate staff are phased out. Priority should also be given to coordinating activities in now-separate rural water and rural sanitation programs, especially in such areas as hygiene education and promotion.

* Consider subsidizing the building of sanitary facilities, so poorer households can afford thenm. The cost of building materials (M400-M500) is well beyond the reach of most poor families. The high public returns from better sanitation could justify a public subsidy.

4.5 The Government could do the following to give the poor better access to education:

* Increase proporlionate public recurrent spending per primary school pupil and spend proportionately less on tertiary schooling, in accordance with targets laid out in the Ministry of Education's 1991/92-1995/96 Education Sector Development Plan.

* Give a much higher priority to the poor geographic regions in investment plans for classroom furniture, facilities, and equipment. Encourage qualified teachers to work in remote areas by giving them better monetary and professional incentives to do so, and by cooperating with church-school proprietors to see that primary teaching posts are allocated to schools in the mountain and Senqu valley zones.

* Inmprove cost recovery equitably after reviewing the real costs of schooling to parents. Creating school boards, which could help set appropriate fee levels, and - 85 -

recovering University student loans will also contribute to more equitable and efficient cost recovery efforts.

* Evaluate and strengthen efforts to provide nonformal primary and secondary equivalertce education to out-of-school youtlh such as herdboys in mountain areas.

4.6 Simultaneous action is needed in several mutually reinforcing social services, including health care, primary education, water and sanitation, and family planning. Educating women not only makes their work time more productive, but is also associated with better health and nutrition for children. Investments in family planning services improve the health and incomes of women who choose to space or limit pregnancies and also lower infant mortality and morbidity rates. Lower fertility and hence population growth rates reduce pressure on limited health care and education resources. Healthy, well-fed children learn more.

4.7 Taking into account the close links between human resource development and poverty and the interaction between investments in different social services, the remainder of this chapter examines Lesotho's current policies and programs for providing social services in order to determine how they could more effectively reach the poor. First we look briefly at the relative priority accorded to social service provision compared with other sectors of the economy. Then we look in some detail at three critical sectors -- health care, water and sanitation services, and education.

B. INTERSECTORALPRIORITIES

4.8 Recent Lesotho governments have made the provision of quality social services a central element of the country's development strategy. Successive National Development Plans have stressed investing in human resources, and many political leaders have publicly confirmed their commitment to good schoolinig, healthicare, and other basic services for the Basotho people.

Public Expenditures

4.9 Total expenditures. Public spending -- capital and recurrent -- on "health and welfare" increased as a share of GDP from about 2.9 percent in the early 1980s to between 5 and 6 percent in the early 1990s (Table 4.1).66 In the same period, public spending on education and community services grew from about 7 percent to 10 percent of GDP. Lesotho is devoting more public resources to social services than most countries in Sub-Saharan Africa, and more than low-income countries elsewhere in the world.

4.10 Real per capita public spending on health and education has steadily increased over the past decade, despite high rates of inflation and rapid population growth. In 1982/83, about M6

66 Assessmentsof public expendituresare based on summarytables preparedby the Ministryof Financefor the years 1987/88to 1993/94. Up to 1991/92,MOF figures are for actual expendituresas reportedby the Treasury and adjusted to incorporatesupplemental funding and underreporting(according to MOF sources). Figures for 1992/93arc revisedbudget estimates. Those for 1993/94are from the budgetpresented to Parliamentin June 1993. Intrasectoralassessments of public expendituresdraw on MOF tables for specificministries that do not contain MOF adjustmentsfor underreporting.A problemwith assessingthe budgetaryimplications of socialsector spendingis that the recurrentbudget is preparedby the MOF,but the capitalbudget is preparedby the Ministryof Planning. - 86 -

per capita was spent on health, rising to about M 15 by 1991/92 (constant 1980 Maloti). Over the same period, per capita spending on education increased from about M16 to M30 (constant 1980 Maloti).

Table 4.1: Changing Trends in Public Spending (1982/83 - 1993/94) Year Health & Education & Economic Defense& Other Total welfare community services publicorder services Expendituresas % of GDP 1982/83 2.9 7.4 20.5 7.4 12.6 50.9 1985/86 3.3 8.2 16.0 8.8 21.9 58.2 1990/91 5.3 10.2 21.8 6.2 11.8 55.3 1991/92 6.2 12.4 20.0 7.0 16.9 58.6 1992/93 6.3 10.3 17.7 5.4 12.0 51.3 (rev.)a 1993/94 5.3 10.4 20.8 5.2 11.0 52.7 (budget)I Expendituresper capita in constant1980 Maloti 1982/83 6.2 15.9 43.9 15.6 26.3 107.7 1985/86 6.7 16.3 32.0 17.2 44.0 116.2 1990/91 13.9 25.6 53.4 15.2 28.7 135.3 1991/92 15.1 30.2 48.8 17.0 31.7 142.9 Expendituresas a % of total governmentspending 1982/83 5.7 14.5 40.6 24.2 24.2 100.0 1985/86 5.8 14.0 27.7 14.8 37.7 100.0 1990/91 9.7 18.5 39.4 11.2 21.2 100.0 1991/92 10.6 21.1 34.1 11.9 22.3 100.0 1992/93 12.3 20.1 34.6 10.6 10.8 100.0 (rev.)a" 1993/94 10.0 19.7 39.5 9.8 21.0 100.0 (budget)bl Source: Bank staff calculationsbascd on Governmcntof Lesothobudget documents. Notes: Percentages are calculated on basis of data provided by Ministry of Finance. Unless otherwise indicated, spendingdata corrcspondto "actualexpenditures" as reported by the Treasury, and adjusted by the Ministry of Finance to incorporate supplementary funding. a/ Revised budget expenditures. bl Program budgeted expenditurcs presentcd to Parliamcnt in June 1993. Interscctoral assessments of public expenditures in this report are based on summary tables preparedby the Ministry of Finance. Up to 1991/92, MOF figures pertain to actual expcndituresas reported by the Treasury and adjusted to incorporate supplemental funding andunderreporting.

4.11 Government spending on health and education grew as a share of total government spending over the decade -- a good sign, since governments control the distribution of resources among competing sectors. Health and welfare received about 6 percent of total government expenditures in 1982/83, rising to about 10 percent in the early 1990s. Lesotho is on the "high" - 87 - end of the health spending continuum in Sub-Saharan Africa.6" And over the same period, the share of total expenditures devoted to education increased from about 15 percent to 20 percent, which is also high compared with similar African and non-African countries.

4.12 Higher government allocations to health and education were partially offset by reduced spending in less productive sectors, such as defense. Spending on defense and public order declined from about 24 percent of total expenditures in 1982/83 to about 10 to 12 percent in the early 1990s. But spending on defense is still equivalent to spending on health, with few discernible benefits to the economy or the poor. Public expenditures on economic services has varied but remained steady over the decade. This includes spending on agriculture and rural development; commerce, tourism, and industry; water, energy, and mining; and roads, transportation, and communication.

Table 4.2: Breakdown of Spending on Social Services (1987/88 - 1993/94) Item | Recurrent Capital Total 1987- 1990- 1993- 1987- 1990- 1993- 1987- 1990- 1993- 1988 1991 1994 1988 1991 1994 1988 1991 1994 Budg. Budg. Budg. Expenditures as % of GDP Health & welfare 4.3 3.1 3.5 2.3 2.2 1.7 6.6 5.3 5.3 Education& other community services. 8.2 6.5 7.5 1.9 3.7 2.8 10.0 10.2 10.4 Water, energy & mininga/ 0.3 0.3 0.4 7.0 6.3 4.9 7.3 6.6 5.3 Expendituresper capita in constant 1987/88Maloti Health & welfare 19.5 17.4 20.9 10.4 12.8 10.1 29.9 30.2 31.0 Education& other community services 37.0 36.7 44.4 8.4 20.9 16.7 45.4 57.7 61.1 Water,energy & miningA! 1.5 1.7 2.2 31.9 35.5 28.7 33.4 37.2 30.9 Expendituresas a % of total governmentspending Health & welfare 10.7 10.1 10.7 7.8 9.1 8.8 9.5 9.7 10.0 Education& other community services 20.3 21.0 22.7 6.3 14.9 14.5 14.4 18.5 19.7 Water, energy & miningal 8.0 1.0 1.1 24.1 25.2 25.0 10.6 11.9 10.0 Source:Bank staff calculationsbased on Governmentof Lesothobudget documents. Notes:a/ includesonly expenditures on water.

4.13 Recurrent and capital expenditures. Providing social services requires adequate levels of recurrent spending -- on teachers' and health workers' salaries, on instructional materials, and on pharmaceuticals and other medical supplies. Breaking sectoral expenditures down to compare recurrent and capital spending suggests the efficiency of resource use (Table 4.2). Recurrent spending on education and water shows a steady, if not rising, trend over the last six years, whetiler suchiexpenditures are expressed as a share of all government spending or in real per capita terms. Growth of recurrent expenditures in the health sector seems to be lagging

67 Other countries on the "high" end of the health spending continuum include Botswana, Mauritius, Swaziland,and Zimbabwe. In contrast, in eight countries in Sub-SaharanAfrica, representingabout 220 million people,the shareof centralgovernment expenditures going to health is only around 2.5-3.0percent. In a World Bank study.Better Healthlin Africa,these have been classifiedas on the "low"end of the healthspending continuum - 88 - behind growth in education, although major salary increases were awarded in both sectors between 1991 and 1994.

4.14 Part of the reason education performs somewhat better than health may be that recurrent expenditures in education have been more adequately pegged to growing capital expenditures as part of Lesotho's Education Sector Development Plan for 1991/92 to 1995/96, which allows for the recurrent implications of donor-supported projects. With no similar process in the health sector, where 80 to 90 percent of capital investments are donor-funded, shortfalls may be incurred in providing for salaries, commodities, and maintenance. This matters for the poor because much of the ongoing donor-financed capital expansion of the health sector is to improve health clinics and expand district hospitals close to rural populations -- for example, in Leribe, Mafeteng, Mohale's Hoek and Quthing. Although some resources should probably be shifted from capital to recurrent spending (or recurrent spending should simply be increased, if fiscal ceilings permit) in all three sectors, the present balance is considerably better than in many comparable developing countries.

4.15 Contributions from church groups. Church groups own and operate about half of the hospitals and clinics in Lesotho as well as virtually all primary schools and about 86 percent of secondary schools (Annex 13). User fees paid to these church groups are presumably used to cover incremental expenditures not provided by the public sector68 -- for example, some portion of initial building costs and subsequent maintenance, salaries for any staff beyond the government-funded complement, and general management costs.69 Systematic data are not available on user fees collected in health and education or on nonfee contributions of church groups, so it is difficult to estimate total funds available for health care and schooling in Lesotho. Rough estimates of private outlays by parents on education and health were extrapolated from the household survey data for 1993. Parents spent roughly M75 million to M80 million (or about 50 percent of public sector expenditures) a year on school fees and another M35 million (or about 40 percent of public sector allocations) a year on user fees at health care facilities. These amounts probably do not include such extras as school uniforms, payments in kind to traditional healers, or some purchases of drugs and medical supplies from private pharmacies.

Achievements

4.16 Combined public and private spending on education, health, and water amount to between 16 and 17 percent of GDP'° -- whichi is high for a developing country. Moreover, real public spending on social services has continued to rise even under structural adjustment, which helps explain the limited negative impact on poverty of adjustment-related fiscal policies. Lesotho has achieved improvements in the provision of social services that have improved the quality of life and human capital for the Basotho people.

68 For roughlytwo decades,the governmenthas assumedresponsibility for virtuallyall currentcosts (includingmost salaries and materials costs) as wellas for suchcapital investments as buildings,equipment, furniture, andmanpower training for church-owned health and educationfacilities.

69 However, in the education sector at Icast, therc is concern that some portion of student feesare being used for purposes other than improving educational services, since most recurrent and capital expenditures arc financed from budgetary sources.

70 Excluding donor contributions, mainly for capital investments, estimatcd at about M20-22 million a year. - 89 -

4.17 Major strides have been made in providing health care services since the early 1980s, when efforts were begun to establish a broad-based primary health care-focused system, with supporting referral facilities, based on the concept of Health Service Areas (HSAs)." There is now a network of community-based village health workers (VHWs), 162 health centers, and 18 secondary or general hospitals, plus a single (tertiary-level) referral hospital located in Maseru. Between 1976 and 1989, the population/doctor ratio dropped from about 20,600 to below 10,000 and the population/nurse ratio fell from 3,400 to below 3,000. Moreover, a number of special programs were launched, targeted to major mortality and morbidity factors. These included an expanded immunization program, which has reached about 80 percent of children under five; an acute respiratory infections program that has significantly reduced case fatality rates in children; a diarrheal disease program aimed at preventing dehydration (the major cause of death) among afflicted children, which helped reduce case fatality rates from 9.5 percent in 1984 to 2.3 percent in 1990; a tuberculosis control program that reduced the incidence of TB from 300 per 100,000 in the late 1970s to 150 per 100,000 in 1991; and an intensive maternal health and family planning initiative under which about 85 percent of pregnant women now receive ante-natal care, under which 60 percent of deliveries occur in health facilities, and under which contraceptive use for child-spacing or limiting has risen from under 10 percent in the early 1980s to about 23 percent in the early 1990s. Important progress has also been made in the provision of safe water, a key factor in preventing ill health. About 70 percent of households today have access to safe water, compared with less than 25 percent in 1984.

4.18 Progress has been substantial in education as well. Most notable was the dramatic rise in student enrollment in primary schools (30 percent) and secondary schools (70 percent) between 1982 and 1991. The gender gap shrank slightly, too. The proportion of school-aged boys attending primary school increased from 64 percent in 1982 to about 69 percent in 1991, while the proportion of girls attending school (85 percent) remained steady. Commendable efforts were made to increase the teaching staffs and physical facilities proportionate with the increases in enrollment, but these indicators deteriorated somewhat nationally, because of the rapid increase in enrollment. Some initiatives to improve the quality of primary education were launched successfully: (1) a revolving textbook distribution scheme now supplies primary school children with textbooks in each core subject on payment of a minimal book levy; (2) from 600 to 650 uncertified teachers participated in an in-service course upgrading skills and more recruits were given pre-service teacher training; and (3) primary school teachers' classroom techniques were improved somewhat through the production and dissemination of instructional materials and the establishment of a corps of 47 experienced itinerant district resource teachers who are to provide in-service guidance.

C. PRIORITIESIN HEALTH CARE

4.19 The substantial financial resources spent on the social sectors has greatly improved human resource development in Lesotho. But experience in other countries (e.g., such as

71 Eachof 19Health Service Areas has a general(secondary level) hospital that functions as the first point of referral.A supervisoryteam guides health workers at lowerlevels of thesystem, including staff atHSA healthcenters aswell as community-basedvillage health workers (VHWs). VHWsand health centers provide basic preventive and curativeservices and serve as the primary point of contactfor peopleseeking health services. - 90 -

Honduras, Sri Lanka and Zimbabwe) suggests that more cost-effective use of resources could hielp bring the quality and coverage of social services up to the levels needed to substantially reduce poverty.

4.20 In many respects, current health priorities, policies, and programs provide an appropriate framework within which to improve the health status of the poor. The Ministry of Health has emphasized that broad-based primary health care is to be stressed over more sophisticated care as the principal means of dealing with preventable or easily treatable infectious diseases, the primary cause of serious illness and death among the poor. Priority is also given to decentralization through health service areas and to fostering a working partnership with private (often church-related NGO) health providers to expand coverage into more remote rural areas where the poor often live. The Ministry of Health is emphasizing health, family planning, and nutrition interventions targeted to children under five and to mothers as well as the treatment of debilitating illnesses, such as tuberculosis, that affect adults' ability to work. These priorities are consistent witlh poverty reduction goals because they focus on some of the most vulnerable individuals in society.

Geographic Distribution of Resources

4.21 Re(listributingfirnancial resources niore equitably. Still, the poor are not served as well as the better-off; they do not have the same access to basic services and to good health care. By and large, healtlh care facilities and outreach programs serve proportionately more people in richer areas than poorer areas, and rural mountain areas get less attention than rural lowlands and far less attention than urban centers (Table 4.3).

Table 4.3: Skewed Health Spending (by Geographic Area)a/

Infant % of recurrenthealth expendituresspent on District mortality Share of the 9 health service areas served by public rate total facilities" (per 1,000 population 1991/92 1992/93 1993/94 live births) (%) [actual] (%) [revised] (%) [budget] %) MaseruD' 74 18.6 24.8 55.1 42.4 Non-Maseru 86 81.4 75.2 44.9 47.6 LoMlands 74 45.4 44.4 27.2 32.4 Mountain& other 95 36.0 30.8 17.7 24.9 Notes: a/ The expenidituresreported herc do not includegeneral MOH expenditureson administration,mental health.public health, leprosycontrol, laboratoryand research,or the NationallHealth Training Center. b/ Expenditureson Maseru exclude the two special hospitals (psychiatricand leprosy)and the National HealthTraining Center. c/ Does not includeexpenditures for the FlyingDoctor Program.

4.22 The inequitable distribution of the financial resources for health care shown in Table 4.3 is reflected in different life expectancies at birth. The life expectancy for the richer, urban areas (58.5 years) is nearly seven years greater than in the rural mountains (51.9 years), where poor houselholds abound. Similarly, in the mountain zone, 170 children out of 1000 die before their fiftIl birthday compared with only 116 per thousand in urban areas and 124 per thousand in the lowlands (Table 2.18). Morbidity levels are also generally higher in poorer regions. The 1992 - 91 -

Nutrition Survey found that among children under five in the two poorest districts of the country -- Thaba Tseka and Mokhotlong -- incidences of most common preventable childhood illnesses, including gastro-intestinal problems, measles, and acute respiratory infections, was much higher than average. Differences in morbidity and mortality probably partially reflect variations in environmental health conditions and nutrition, but increased financial allocations for basic health services (especially preventive care) in poor regions would go a long way toward reducing inequities in health care.

4.23 Increasing allocations to primary health care. Spending on basic health services, especially preventive care, is generally far more cost-effective (in improving health) than spending it on curative (secondary and tertiary) care, provided through hospitals. And the poor tend to benefit far more from public spending on preventive care because, for lack of good referral systems, hospital services tend to cater almost exclusively to better-off urban residents. MOH budgeting norms make it difficult to break down how much funding is devoted to primary, secondary, and tertiary facilities, but there is concern that public funds devoted to primary health care (PHC) activities remain at very low levels. Only M4.75 million (5 percent of the MOH recurrent budget) was allocated to primary health care in 1990/91. That amount represents only a portion of total expenditures on PHC -- since major expenditure items such as the salaries of staff engaged in PHC are included in the general salary category and substantial donor outlays are off-budget -- there nonethieless appears to be a mismatch between this low allocation and the higih priority (appropriately) placed on PHC in MOH policy statements. MOH should consider increasing allocations to primary healthi care even at the cost of reducing allocations to hospital facilities. That would require systematically analyzing the distribution of government and donor resources among the various levels of care, which cannot be done under the current system.

4.24 Encouraging health personnel to work in underserved areas. Incentives must be developed that encourage trained personnel to accept assignments in what are generally considered to be "hardshiip" locations. Staff shortages are acute in poorer regions, so it would be difficult to expand health care services, preventive campaigns, or other initiatives there even if funds were available. Trained medical and paramedical personnel are reluctant to accept assignments in remote areas partly because of the higher costs of basic necessities, the difficulty of finding housing, transport problems, reduced social and cultural opportunities, and regular contact with professional colleagues (because so few supervision visits are made to remote facilities).

4.25 So far, the only incentive to encourage personnel to locate in remote areas is a small monthly stipend ranging from M15 for nurses to a maximum of M33 for married doctors with dependents. This stipend, which is less than 5 percent of salaries (adjusted only irregularly), bears little relationship to actual differentials in cost of living. The consumer price of common staples such as maize meal is roughly 10 to 15 percent higher in mountain areas than in the lowlands, because of transport. Monetary incentives to personnel in rural areas should at least reflect actual cost-of-living differences. But other incentives should also be considered. As much as possible, emphasis should be placed on incentives with low recurrent cost implications - - providing a simple staff house, for example. Finally, to address problems of professional isolation in remote areas, MOH should consider increasing the frequency of supervision visits by HSA hospital teams to remote health centers, or the introduction of regular short training rotations to HSA hospitals by staff posted to remote facilities, or both. Efforts should be made to - 92 - recruit students from poorer regions into pre-service nursing as well, as these students would presumably be less reluctant to return to their home areas after training.

4.26 Strentgtlhening conrntunity links. More effective use could also be made of community- based health workers, including village health workers (VHWs), traditional healers (THs), and traditional birth attendants (TBAs). This would mean greater efforts to train and devise appropriate incentives for VHWs and to increase cooperation between THs, TBAs, and the modern health care delivery system. Such cooperation might include training THs/TBAs in such basic health practices as hygiene and referring high-risk cases to formal health facilities. Levels of prenatal and maternal morbidity and mortality are high in Lesotho, so one priority should be to expand existing (small-scale) initiatives to train TBAs. Effective cooperation with trained THs/TBAs could have a high pay-off in health outcomes; roughly two-thirds of rural Basotho live within a kilometer of a TH/TBA and often seek them out at the first sign of illness.

Cost Recovery

4.27 Reviewing current exemption policies. The poor may have less access to health care services as a result of the policy to enforce user fees, which MOH adopted in 1988 to raise revenues and to encourage more equitable use of private and public facilities (by bringing fees at MOH facilities more in line with fees at private clinics and hospitals). To encourage the use of primary facilities as the first point of contact, MOF introduced a differential between fees paid at health centers and those paid at hospitals. Unfortunately for the poor, the new scheme involved a doubling of fees from Ml to M2 for initial adult outpatient consultations at health centers, and a tripling to M3 for initial outpatient consultations at hospitals. (The consultation fee at hospitals for an adult referred by a health center was only Ml, to keep the total user charge (M2+MI) the same for those going directly to hospitals and those referred to hospitals.) In addition, varying fees were charged for special services, such as X-rays, and for any medicines, dressings, or medical equipment provided. Consultation charges for children were half the adult charge but special charges were the same. Only two types of individual were exempt from charges: (I) all health workers (including VHWs) and their families, or about 30,000 people -- most of whom tend be relatively better off; and (2) people certified by their chiefs to be "paupers" (defined, perhaps too severely, as those with no means of income, land, livestock or other belongings). Preventive services -- many maternal and child health services (including family planning and immunizations); tests for syphilis, typhoid, and brucellosis; and treatment of outbreaks of contagious diseases -- are still provided free of charge.

4.28 The public health exemption for preventive services and control of infectious diseases is eminently justifiable but the exemptions for health care workers and paupers do not adequately protect the poor, who now use health care less. Attendance at MOH facilities dropped 51 percent in Mokhotlonig district -- one of the country's poorest districts -- after introduction of the new fee schedule.72 This decline was partly offset by increased attendance at private facilities, but there was still a net 30 percent reduction in outpatient attendance at all facilities, which suggests that some who needed care may have foregone treatment because of the costs. Small children (plus adults with cardiovascular and genitourinary conditions) were at greatest risk of having reduced

72 Daniel Byrmeand PaulGertler surveyedthe impact of userfees. Seetheir "User Fee Impacton Outpatient Utilization in Lesotho--,March 1990. - 93 - access to care. This is a problem since pre-school children are vulnerable to malnutrition and associated health problems.

4.29 Correcting urban bias. Implementation practices for cost recovery exacerbate inequities in the provision of health care. In general, hospitals recover a much lower proportion of their costs through user fees than do the lower level facilities frequented by the poor. Revenues as a share of costs in better-off Maseru district (which includes the country's most sophisticated hospital) are only one-third to one-quarter those of health service areas in the non-Maseru districts (Table 4.4).'3 Since all revenues go to the central treasury, non-Maseru districts are implicitly subsidizing health care among relatively well-off Maseru residents. Efforts to step up cost recovery at the central hospital are clearly in order.

Table 4.4: Cost Recovery Among Districts Revenueas % of expenditures Year Maseru OtherDistricts

1986/87act. 5.9 7.6 1987/88act. 12.6 16.2 1988/89act. 15.8 15.6 1989/90act. 7.8 14.9 1990/91act. 10.0 14.0 1991/92act. 4.5 13.2 1992/93rev. 5.3 22.0 1993/94budget 8.7 24.3 Source:World Bank staff estimates.

4.30 Allowing localfees to be used locally. The practice of requiring that user-fee revenues be returned to the treasury, rather than kept at the facility or HSA, may inhibit improvements in the quality of services. International evidence on cost recovery schemes suggests that the quality of health care improves substantially where revenues collected from clients are decentralized to the district or facility level, usually with broad guidelines about the priority use of funds. Patients wlho benefit directly from the payment of health care fees are thus more willing to pay them. Adopting such a system in Lesotho would help ameliorate present inequities, because although the poor might pay proportionately more for services than the better off, they would at least benefit directly from improved quality of care.

4.31 In the medium term, however, it is essential to review the entire fee schedule,"4 to raise the fees charged for hospital services, and perhaps to reduce charges for health services provided through facilities in remote areas. This would help encourage people to use health centers as

73 The ratio of cost recoveryto expendituresis higher for QEII if expenditureson servicesthat benefit all districts are removed,such as expendituresfor the specializedpsychiatric and leprosy hospitals,and the National Health TrainingCenter. Even so, the ratio of revenuesto expendituresin QEII is less than half those of the other publicly servedHSAs combined.

74 Fee scheduleswere revisedin 1992/93but the goal was to simplifyrevenue administrationby charginga flat fee for consultationplus feesfor specialservices and medicines.This simplificationdid not address fee inequities. In fact, to the extent that the upfrontcost of an outpatientvisit could now be higher in all facilities,the problemsthe poor have getting healthservices may have been exacerbated. - 94 - their first point of contact with the health system, rather than proceed directly to hospitals.75 MOH shiould also consider removing the exemptions for health workers and their families and adding an exemption for vulnerable population groups, especially children under five, elderly people over 65, and the physically or mentally disabled people of all ages.

Underemphasized Health Problems

4.32 Provide AIDS prevention services. Two special health problems should be emphasized to enhance the sector's impact on poverty. The first of these is AIDS prevention. The incidence of AIDS now appears to be minimal in Lesotho. Only 300 AIDS-related deaths have been recorded, and just over 2 percent of the population is believed to be HIV-positive. But according to a recent population sector study undertaken by the World Bank,"6 Lesotho could have one of the worst AIDS epidemics in Africa in terms of proportion of the population infected. The number of new cases increased from 25 new cases in 1991 to over 300 by 1993. HIV seroprevalence rates are now believed to be over 5 percent in Maseru, 2 percent in other lowland towns, and 0.5 percent in the mountain regions. These levels are still relatively low, but Lesotho combines several high-risk factors that could lead to rapid increases in seroprevalence: unusually highl rates of sexually transmitted diseases (STDs), including those thought to facilitate HIV transmissioni; an unusually large proportion of married couples separated for extended periods; early initiation into sexual activity and high rates of poly-partnerism among the sexually active; and poor knowledgeor acceptance of practices that could reduce the risk of infection with AIDS. A rapid increase in the incidenceof AIDS would probably increase poverty substantially because those affected are usually adults in their prime, who tend to be the main income earners of their houselholds. An increase in the incidence of AIDS would deprive entire households of their principal means of support, pushing themiiinto poverty or destitution. Moreover, the burden on the healthi care system of caring for more AIDS patients would undoubtedly require reallocating resources away from pro-poor programs such as maternal and child health care, immunization, and the like.

4.33 The National AIDS Prevention and Control Program, established in Lesotho in mid- 1987, appears to have done little to actually lower transmission risks and discourage an epidemic. A participatory study on healthi by Sechaba Consultants found that most ordinary Basotho believe AIDS to be a relatively minor issue, even a non-issue, because their experience of it is still limited and education efforts have not reached them. A survey on sexual practices77 founid that wlile condoms are understood by some to be a protection against AIDS, there is still considerable reluctance to use themilas they are believed to interferewith pleasure. Most miners who received free condoms said they rarely use them, and those who did use them said that cost would have kept them from purchasing their own. Aggressive education campaigns are needed to begin bringing about the behavioral changes needed to avert an AIDS epidemic. It makes sense to combine, in a single MOH department, AIDS-prevention activities witil better- established interventions to deal with sexually transmitted disease (STD), given the commonality of many issues.

75 The impactstudy foundthat the presentdifferential betwveen health center and hospitaloutpatient consultationfees is too smallto keepclients from going directly to thehospital.

76 Lesotlo PopirlatioiiSector Review. ReportNo. 12553-LSO,November 14, 1993.

77 BasothoMineworkers Sexual Practices. T. Niatobo,National Universityof Lesotho, 1992. - 95 -

4.34 Provide treatment for alcoholisniL The second area deserving special attention from health authorities is the problem of alcoholism, which many participants in the Sechaba qualitative studies felt was a major problem deserving public attention. Both rural and urban survey participants saw alcoholism as a major contributing factor to the impoverishment of households and the spread of disease (STDs associated with alcohol-related adultery; mental breakdown; pellagra; asthma; and injuries). Alcoholism has been treated mainly as a social problem affecting a minority of the population, so it has not been emphasized either in health plans or in public social welfare activities. The temptation to deal with alcoholism by taking the route of heavily restricting the production, purchase, and use of alcohol should probably be avoided. Experience in other countries suggests this route is often counterproductive. Rather the focus should probably be on mass education programs aimed at preventing alcohol abuse and at encouraging alcoholics to seek treatment. This would help increase the use of the few existing, underused NGO programs aimed at helping people with alcohol abuse problems -- for example, the Community Alcohol Rehabilitation program at Scott Hospital and the Blue Cross Center. What is needed is a thorough analysis of the problem and options for addressing it, before deciding on the appropriate role of the public sector in the prevention and treatment of alcohol abuse.

D. INIPROVINGACCESS To SAFE WATER AND SANITATIONSERVICES

4.35 Improving health, especially where infectious diseases account for most illness and death, generally involves making safe water and sanitation facilities more widely available. One study credited improved waste disposal with a 36 percent reduction in diarrhea in Lesotho, prompting the conclusion that many sanitary interventions could have a bigger impact on health than improvements in health care services do, especially in highly contaminated environments.78 As a result, Government policy emphasizes that all households have access to safe water and substantially increased access to improved latrine facilities. To date, however, less than half of rural houselholds in Lesotho have access to clean water, although the situation is better today (an estimated 35 to 45 percent coverage) than it was in 1984 (14 percent). Sanitation facilities are available to even fewer rural households (less than 20 percent). In urban areas, especially in Maseru, about 80 to 85 percent of households have access to safe water, and about 60 percent have access to sanitation facilities. Spending for urban water and sanitation, from both donor and government sources, is proportionately much higher than spending on rural services. Government policies seem to favor the richer urban population over poorer urban-dwellers, although this may be partly because of the greater risks of infection and epidemics in densely populated areas. More telling are coverage differentials within rural and urban areas, which do suggest that the poorest areas are neglected while better-off areas receive much higher levels of service.

78 Daniels,D.L., L.N. Makoae,and R.G. Feachem, 1990,"A Case-ControlStudy of the Impactof Improved Sanitationon DiarrheaMorbidity in Lesotho,"Bulletin of the WVHO.Vol. 68, pp. 455-63. - 96 -

Rural Water Supply

4.36 Geographic distribution of resources. Intensive development of village water supplies in Lesotho began in 1979 with a USAID-supported program that later evolved into the Village Water Supply Service (VWSS). Before 1979, the Government had been continually involved in improving rural water supplies but limited resources had prevented the program from having a major impact. The VWSS, however, has exceeded the targets it set for itself and presently reaches 75,000 to 80,000 people a year, producing about 175 additional systems. Since its initiation the program has received major donor and capital assistance -- a total of about US$40 million -- from international organizations; Lesotho had contributed about US$4 million, if one counts the estimated value of voluntary labor. The Government pays for local staff and overhead and local communities contribute labor to install the new systems, which amounts to about 30 percent of construction costs. Villagers also have to pay for the system's running costs, such as diesel oil. In the early years the program focused on districts that were fairly accessible. This strategy was probably inappropriate: the poorest districts, which are generally the most difficult to reach, were neglected. The current workplan calls for extending the system to more remote, poorer areas of the country.79 Current variations among districts are highly inequitable. In the two poorest districts, Thaba Tseka and Mokhotlong, the Village Water Supply Service (VWSS) program serves only 10.5 and 22.5 percent of the population, respectively. In the lowland areas, 50 to 60 percent of residents have received program attention (Table 4.5).

79 Table 4.5 includesonly VWSS's own estimatesof working systems(86 systemsin needof major repair were excluded)installed by the VWSS and thoseinstalled by NGOs for which the VWSS had acccptedmaintenance responsibility(Moran 1990.P.1). Dataon total coverageby organizationsother than VWSS arenot available,nor is there a record of the actualcondition of the systems. Hodgkin (USAID 1993)assumes that about 25 percentof the systemsneed repair. - 97 -

Table 4.5: Combined Capital Spending on Water and Sanitation Government and donor funded capital expenditures (Millions of Maloti) Accumulated to: Revised Approved Estimated total Clean water March 31, 1992 1992/93 1993/94 to completion Ministry of Interior - rural water supplies 5.6 5.8 7.0 23.5 Ministry of Water, Energy, Mines [Water Affairs Dept.] - 4 towns 9.6 11.0 10.0 34.6 - Maseru 6.1 4.0 3.0 13.1 - Ground water resources 1.3 0.5 0.5 2.8 Central Planning - Advisory support to water 0.3 0.3 0.3 1.0 Other 1.4 3.3 4.4 12.4 - Rural water as % of total 28.4 25.3 29.8 26.7

Sanitation Ministry of Health [Public Health Dept.] - Rural sanitation 1.6 1.3 1.5 6.0 Ministry of Water, Energy, Mines [Water Affairs Dept.] - 13 towns sanitation 4.0 0.8 0.6 5.3 - Maseru sewerage 9.6 9.0 3.0 22.6 - Sanitation study 0.0 0.0 0.3 0.7 Ministry of Works - Water and sewerage 0.1 0.1 0.1 0.3 Ministry of Interior - Construction of toilets 0.0 0.0 0.0 0.1 Rural sanitation as % of total 10.5 12.5 9.7 17.4 Source:Staff estimates.

Table 4.6: Actual Population Served By VWSS, (1990) District No. of Pop. Adtion Rural % projects served population Covered Qacha's Nek 53 25,450 65,810 38.7 Quthing 77 29,520 117,540 25.1 Mohale's Hoek 190 91,690 173,660 52.8 Mafeteng 223 112,030 203,250 55.1 Maseru 388 128,210 247,400 51.8 Thaba-Tseka 42 11,860 113,160 10.5 Berea 248 120,470 196,990 61.2 Leribe 138 116,930 264,770 44.2 Butha-Buthe 101 67,030 108,990 61.5 Mokhotlong 64 17,800 80,100 22.2 TOTAL 1,524 720,990 1,571,670 45.9 Source:The VillageWater Supply Maintenance Study, Mark Moran,1990. - 98 -

4.37 By and large, the VWSS has maintained high technical standards and has chiosen appropriate technologies (gravity systems where they are feasible and robust handpumps in lowland areas where they are not). Moreover, the standards for accessibility are highi as well: a maximum distance of 150 meters to the water source for each household in the village, about 50 persons per standpipe, and a volume averaging 30 liters per person/day (VWSS, 1991). One problem, which the VWSS is beginning to address, is no regular testing of the quality of the water produced by the systems. However, recent opinion surveys about the quality of the water revealed that most households believe their water to be clean.90

4.38 Maintenance. More difficult is the problem of system maintenance. The VWSS has focused more on producing new systems than on developing maintenance structures. Village Water Communities (VWCs) are formed at the time a system is installed and instruction is given on simple system maintenance, but this is a one-time event with no follow-up to strengthen village management. As a result, a VWC often does not understand its role in preventive maintenance, main maintenance, and the reporting of breakdowns. Maintenance issues are now getting more attention in VWSS. Now that the systems are older and will need more repair, it is time to focus more on operations and maintenance. Within VWSS, special maintenance units are being strengthened and an attempt is being made to develop institutional support systems in collaboration with district or regional rural development departments. Coordination among government agencies has been weak. For example, although the VWSS has sought closer collaboration with the National Rural Sanitation Program (NRSP) under the MOH -- because NRSP is better equipped to handle hygiene education and promotion -- results have been disappointing because of differences in financial policy, lack of donor funds, and an alarming shortage of personnel in NRSP.

4.39 In 1989 a policy of cost recovery was approved for the maintenance of existing systems. Introduced at the national level in 1991, it required communities to cover the costs of minor repairs and maintenance with their own funds; major repairs would still be paid for by the VWSS. Roughly half of invoiced maintenance and repair costs are being paid for by the village water committees, whichi collect maintenance funds when the systems are built. The VWSS bills the community for each repair job, and the community meets the cost of labor and spare parts; VWSS covers transport, logistics, and overhead. The maximum contribution for a repair job is M I per capita; excess cost is subsidized by VWSS. But there is some concern that maintenance will be reduced when donor funds are withdrawn. Even withi good village management systems in place, villagers still depend on VWSS for complex or expensive maintenance tasks. That subsidy would not be possible with present government funding levels alone, so there is concern that the program may not be sustainable in the long run unless the Government gradually increases its contribution in anticipation of donors wvithdrawingtheir resources.

Rural Sanitation

4.40 Geographic distribution. Efforts to improve rural sanitation began in 1983, when the Environmental Health Section of the Department of Public Health, in the Ministry of Health, started a program to promote the construction of Ventilation Improved Pits (VIPs) through hygiene education and the training of local builders in Mohale's Hoek. The program, sponsored by UNDP, UNICEF, and the Technical Advisory Group of the World Bank, supports the

gO Scchaba. 1992. - 99 - voluntary, unsubsidized household construction of latrines. In 1987, a National Rural Sanitation Program was launched with support from several international agencies. At the district level, the program is implemented by district sanitation coordinators (senior health assistants) assisted by technical officers and other staff of the Ministry of Health. Households are responsible for maintenance and, later, for emptying the systems. Presently, the program is operating in all but one of the districts (Thaba-Tseka is not operational because of staff problems) but has built only about 8000 VIP's on homesteads and school sites (Table 4.7). Rural households have a generally positive attitude toward VIPs. They offer privacy and comfort and are viewed as status symbols and as having a positive impact on the environment (Evans, 1986).

Table 4.7: VIP Latrine Construction by District District End of'92 Coverage Mohale'sHoek 2,800 Leribe 1,165 Berea 1,562 Butha-Buthe 1,251 Maseru 123 Quthing 795 Qacha'sNek 43 Mokhotlong 278 Mafeteng new Total 8.027 Source: National Rural Sanitation Programme Annual Report- 1992.

4.41 Tlhe casefor loans and subsidies. Government contributions are low. Different donor agencies contribute to programs in different districts, making funds available for transport and district office equipment, health education equipment and materials, VIP latrine kits, the training of local builders, and salaries for extension workers and technical supervisors. Households pay for construction materials and provide labor. If locally available materials are used, a latrine costsabout M400 to M500 to build; if commercial building materials are used, the costs double. Costs are relatively high, so in 1987 a scheme was introduced to make loans available through the Lesotho Cooperative Credit Union League (LCCUL). However, the LCCUL, like most other formal financial institutions, found it impossible to make collections in rural areas cost-effective and stopped extending credit to program participants. Rather than stopping, LCCUL and other organizations should attempt to improve the credit system, because international experience with infrastructure credit to the poor is quite good (e.g., the Grameen bank). But affordability remains a serious concern. Without some form of public subsidy, only the higher-income rural households or those with remittance incomes can carry the cost of building a latrine. Perhaps subsidies are the grounds for equity in the poorest areas of the country.

4.42 Staffing. The long-term sustainability of the program is further jeopardized by the low priority the Government assigns it. In two districts where international agencies have handed the program over to the Government, program performance has declined dramatically. The MOH, which devotes less than 7 percent of its recurrent budget to all public, environmental health programs combined, gives the program low priority. Moreover, staff attrition has left the program severely understaffed and the staff has no logistical support. The many problems - 100- limiting the sanitation program have prevented integration with the rural water supply program. A major review of the program is in order.

Urban Water and Sanitation Services

4.43 Urban residents generally have better access to safe water and sanitation services than rural dwellers do, except in some poor areas. In Maseru, close to 9000 individual plots (40 percent of the total) are directly connected to the water main, but there are great disparities in service. In better-off Central Maseru, the Old Reserve area, close to 75 percent of plots have individual connections; in Maseru North, only 15 percent; in Maseru South, 20 percent; and in Maseru East, 24 percent. Households living on the unserviced plots depend on communal standpipes (which serve more than 50 percent of the plots and an even larger proportion of the households), on taps in neighbors' plots, or on vendors. Only a small proportion of households in Maseru do not have access to any piped water.8'

4.44 However, while the standpipe network has not expanded since 1988, the peri-urban areas have grown rapidly. Distances from water points have increased and water vendors have taken advantage of the situation, charging urban households high prices (25 to 75 cents per 20 liter bucket) for water deliveries. There is a growing disparity between service levels in different parts of town that is not necessarily related to income levels, but high prices for vendor-supplied water have 'lit the low-income houses the hardest. The poor are also unable to pay to have the water main privately extended to their plot, something higher-income households do increasingly. Priority should probably be given to expanding the urban standpipe network, targeting service first to areas with more poor households.

4.45 The picture is similar for sanitation services. Rapid urban population growth is outstripping sanitation facilities. In Maseru, the only urban area with a sewer system, fewer than 20 percent of plots are connected to the sewer system, whicihcovers only parts of Central Maseru and Maseru Northi. About 40 percent of the rest of the plots have VIP toilets and 30 percent have pit latrines; some governnenit units (4 percent of plots) still have bucket latrines, and about 9 percent of plots have no latrines at all. Since VIP toilets are constructed by individual plot- owners, more higiler-income households have a VIP toilet than do lower-income households. On plots used for room-renting, the level of sanitation is lower. Most room-renters have access only to a pit latrine.

4.46 In other urban areas for which recent data exist, general services have failed even more than Maseru to keep pace withi fast urban growth. In rapidly expanding Maputsoe, only 19 percent of plots have metered water connections and more than 60 percent of plots depend on communal standpipes. More than a fifth of plots have access only to spring water, bore-holes, or other unpiped water sources. Nearly two-thirds of the plots have pit latrines, 25 percent have VIPs, and II percent have no toilet facilities at all. Most households share a latrine with other households. Mohale's Hoek, on the other hand, has not experienced such rapid population

S1 Indicatorsfor Lcsotho's 45.000 to 50.000 urbandwelling units were last collected in the 1986census. More recentdata areavailable only for Maseruand Mapulsoe,two rapidly expandingurban areas, and Mohales' Hoek and Leribe, two medium-groxwthareas. Unfortunately.the samplingmethods used do not allow inferencesto be made for the whole urbanarea and are usedwith caution here. Sources:WASA 1993,Physical Planning Department 1990, Working PaperNo. 10, DOH 1992 - 101 -

growth, has the greatest proportion of plots connected to the water main (38 percent), and only 30 percent of its plots are serviced by standpipe. On an additional 9 percent of plots, households buy water from neighbors or vendors. Households on other plots (8 percent) depend on unpiped water sources. As the result of a successful campaign to install VIP toilets, 40 percent of plots have VIP toilets. Another 25 percent still depend on the unsanitary bucket system. The other plots either have pits or no latrines at all and most households share latrines with others.

E. IMPROVING BASIC EDUCATION

4.47 From the standpoint of poverty reduction, many problems in education are similar to those in health care. Primary education (like primary health care) tends to be underemphasized compared with higher levels of education; the quality of basic education appears to be worse in poorer geographic regions; and cost recovery is better in the lower levels of the system than in higher levels. These issues and problems in nonformal adult education are discussed briefly here. We make no attempt to cover all problems in education in Lesotho, but focus on the issues that have the most bearing on poverty.

Expenditures on Different Levels of Education

4.48 The Ministry of Education (MOE) has rightly emphasized improving primary education in several policy documents, including its recent Education Sector Development Plan (1991/92- 1995/96). This policy emphasis has to some extent been reflected in primary education's increasing share of public resources (from about 30-35 percent of total education spending in the 1970s and early 1980s to about 52 percent of the recurrent budget by 1991/92). But to the extent that enrollment in primary school has also increased, significant differentials still exist between per-pupil expenditures for primary education and higher levels (Table 4.8).

Table 4.8: Per Pupil Allocation in Recurrent Education Budget (1991/92) Subsector Allocation As % of total Per pupil (M) allocation Primarya' 75,022,748 52.00 M208 Secondaryb/ 42,908,391 30.00 M921 Technical and vocational education 3,417,474 2.40 M3,513 Tertiary(NUL) 21,260,080 15.00 M 15,024 Adultnonformal education 220,936 0.01 na Overhead (admin.) 2,145,972 1.00 Total 144,975,600 100.00c' Source: Adaptedfrom the EducationSector Development Plan, 1991/92-1995/6,MOE. Notes: a/ primaryincludes all MOE activitiesrelating to primary education including teachertraining. b/ Secondaryincludes all MOE activities relating to secondaryeducation, including secondary teacher training at the National Teacher Training College (NTTC) but excludingsecondary teacher training at the NationalUniv. of Lesotho(NUL). c/ Numbersmay not add up exactlydue to rounding. - 102-

4.49 As Table 4.8 shows, MOE spends 70 to 75 times more per university student (on average) than per primary pupil, although expenditures on primary education are well-known to produce much higher social returns than expenditures on higher levels of education and are thus more deserving of government subsidy. And direct government support to the University, and to the households of university students, is probably considerably higher than these estimates show, but more precise data were not available to the Bank mission. Moreover, to the extent that few of the poor avail themselves of education services beyond the primary level, the poor receive at best a minuscule share of public spending on higher education. Most public resources support students from better-off households. Per-pupil spending on higher education is even more disproportionate when bursaries to university students through the National Manpower Development Secretariat (NMDS) are considered. Many of these bursaries (for living costs and tuition) are provided as loans, but recovery rates are abysmally low, seemingly due to a lack of real effort by the Government to recover the loans.

4.50 MOE's Education Sector Plan shows that the Ministry is well aware of these disparities. The ministry stated its intention to gradually increase relative allocations to primary education. This was to have been accomplished by devoting at least 70 percent of increased sectoral allocations (an expected 4 percent a year in real terms over the Plan period) to primary education while permitting no real growth in allocations to the University.82 But a review of intra-sectoral allocationszover the first two years of the Plan period shows that allocations to primary education have not grown at the rate projected and that efforts to contain allocations to higher tiers of education have been less than successful. Inequities have probably not increased, but they have certainly not lessened. The Government needs to renew its commitment to a more equitable allocation of public resources for education.

Geographic Distribution of Resources

4.51 Increasing investments in poorer areas. Net primary school enrollments are considerably lower in the poorer mountainous districts of the country. This may be true partially because of the high opportunity costs of sending children to school (especially boys needed for herding), but the generally higher repetition rates and lower primary school completion rates suggest that lower enrollments may be partly attributable to the poorer quality of schooling in these areas. Although there is no 1:1 correspondence, the poorer -- Mokhotlong, Quthing, Qacha's Nek, and Thaba Tseka -- tend to have proportionately more uncertified teachers, higher repetition rates, and lower completion rates for primary school (Table 4.9).

4.52 Differences in geographic terrain, roads, and transportation facilities among districts greatly affect access to schools. In lowland areas, especially towns, transportation and travel are easier, so schools can be expected to have a larger catchment area than schools in mountain districts, whiere the difficult terrain makes it harder for young (primary-aged) children to travel long distances. Probably more, and smaller, schools and classes are needed in mountainous districts to reduce the distance students must travel to get to school. This issue is explicitly recognized by the MOE, wliose standards theoretically provide for lower student:teacher ratios in

82This would allow for a gradual reduction in public spending per university student, thus allowing NUL time to develop cost containment programs and increase reliance on allemative sources of funds (such as cost recovery and endowment income). - 103 -

(poorer) mountainous areas than in the lowlands. In practice, however, the student:teacher ratio districts are not significantly lower -- and in some cases are higher -- then in the lowlands and foothill districts. The lower primary school attendance and completion rates in the mountain districts may be related to problems of both access and quality.

Table 4.9: Primary School Characteristics by District (1991/92) District No. of No. of Number of % teachers Repetition % completing schools students students per unqualified rates primary (000's) teacher school Butha-Buthe 67 24.8 56.1 18.5 22.8 73.1 Leribe 151 61.4 56.3 14.0 23.5 82.1 Berea 105 51.1 60.7 17.6 21.7 88.6 Maseru 196 71.6 48.8 15.1 19.4 77.6 Mafeteng 135 43.0 52.3 18.1 22.0 93.3 Mohale's Hoek 140 33.9 52.6 18.9 22.4 71.4 Quthing III 22.6 57.2 21.3 21.4 53.2 Qacha's Nek 83 13.9 51.9 18.4 22.3 42.2 Mokhotlong 95 19.5 59.2 25.8 27.5 67.9 Thaba-Tseka 115 19.3 50.0 19.0 23.3 47.8 Source:Ministry of Education.

4.53 There is some evidence that physical facilities also tend to be poorer in the mountain areas. MOE's five-year education sector development plan shows pupil:classroom ratios to be worse in the highlands. A single room (such as a church hall) is sometimes used to teach over a hundred children in several grades. It is not usually possible to hold classes outdoors in the mountains because of the harsh weather conditions, and classroom overcrowding probably makes it difficult for pupils to concentrate and learn. The lack of appropriate classrooms also makes it more difficult to install appropriate furniture and equipment, which further affects children's learning environment. The MOE plan emphasizes improving physical facilities for primary education, especially in the mountain zone. Implementation targets set out in the plan -- building and furnishing 850 new classrooms and refurbishing 200 existing ones over five years -- should be a high priority.

4.54 Increasing access to teaclhers. Improving education in poorer geographic regions (such as the mountain zone) is impossible without improved incentives for qualified staff to relocate there. Priority should be given to recruiting teacher trainees from poorer regions, to paying adequate cost-of-living adjustments for service in remote areas, providing housing or other facilities where appropriate, and increasing opportunities for teachers to interact with their peers, possibly by building on existing initiatives such as the district resource teacher program.

4.55 Getting more teachers to work in disadvantaged geographic regions will require coordinating efforts with the church groups that operate primary schools in Lesotho. Under the current system, MOE allocates teacher positions to the main church school proprietors, who have considerable discretion as to how to allocate posts among schools. To the extent that positions in lowland schools are easier to fill -- and perhaps also because parishioners in the lowlands tend to be more vocal -- proprietors tend to allocate new positions to the lowlands despite the greater need in mountain areas. One way for MOE to counteract this tendency would be to allocate positions by school or parish, relying on its considerable database of primary school statistics to - 104- identify schools with especially high student:teacher ratios. Because of the sometimes politically sensitive relationship between MOE and the church school proprietors, it makes sense to engage the church education secretaries in a dialogue about equity in an effort to persuade them to take poverty considerations into account in allocating teacher posts. If such a dialogue has little effect, the MOE may decide whether more direct measures are needed.

Cost Recovery

4.56 In addition to eliminating the geographic bias in the distribution of human, physical, and financial education resources, giving the poor better access to basic education is also likely to require changing current fee requirements for education. In theory, the current fee structure is appropriate. Primary school is supposed to be free (except for a textbook levy of M4 in the first grade, rising to M9 by grade 7) and user fees charged for higher education get progressively higher. In practice, however, government bursaries to Basotho university students cover most fees. These are theoretically loans, but recovery rates are so low that they are effectively indirect grants83 to the University, which retains fees to cover administrative and educational costs. In practice, few fees are actually recovered from the families of University students.

4.57 By contrast, fees paid at the primary level tend to be higher than mandated. MOE regulations prohibit charging for tuition at the primary level, but a 1992 survey84 of a partial sample of T.esotho's schools revealed that many other fees are collected by virtually all primary schools. The mandated book levy is the only fee passed back to the MOE, but other fees are commonly collected for registration (from M4 to M64 a child each year in the sample surveyed) for examinations (sometimes in lower and middle primary as well as for the Primary School Leaving Examination in grade 7); for building maintenance; for school meals; for special activities; and sometimes for recruitment of private teachers (over and above the MOE-funded complement). And most schools require uniforms for pupils, for which parents must pay an estimated MI 00 to M120. The school survey found that the average cost to parents of maintaining a child in primary sclhool range from about M150 a year for a first grade pupil to about M200 a year for a seventh grade pupil, although the range is wide. Secondary school fees are even higher. Mean registration fees were M126 for the sample of secondary schools surveyed, examination fees average between M78 and M248, depending on the year of study, book fees can range from MIOOto M260, and miscellaneous fees average M182, but can be as high as M416 per student per year.

4.58 Households without secure access to cash income are unlikely to be able to send their children even to primary school (let alone secondary school), which effectively excludes many of the poorest children from education. The 1993 Sechaba household survey showed that even with lower than average enrollment, the poorest households devoted 20 to 25 percent of total expenditures to school costs. And because the completion of primary school is essential for breaking the vicious circle of poverty, MOE should more systematically study the household

83A directannual grant is alsomade to the University. 84Reporton Institutional Reformsand Policy Developmentin the Provisionof Textbooks,Supplementary Materialsand Journals/Periodical, Vol. 2 LesothoSchools Survey, 1992. The survey, undertaken principally to review textbookand instructionalmaterials needs, covered relatively few schoolsand there was little analysisof the cost informationcollected. But it doessuggest the range and magnitude of feescollected in primaryand secondary schools. - 105- costs of primary and secondary schooling and identify ways to lower those costs, especially for primary school.

4.59 There should be more awareness of school costs when MOE's Planning Department completes its school mapping exercise, which includes questions about fees. A quick review of the school mapping questionnaire should reveal whether the data being collected are adequate to formulate a fee policy or whether supplementary data needs to be collected. Once MOE has a clear picture of fee structures operating at the primary level, it must begin negotiating with the church school proprietors, to find ways to lower the costs to parents of enrolling children in school. One strategy that has been adopted successfully in many countries is to stop requiring school uniforms, which represent a substantial financial burden to parents and add little to learning. The value of other fees should also be examined. Finally, it may be necessary to introduce basic cost accounting procedures to ensure that any fees collected by primary and secondary schools are actually used to improve the quality of the schools rather than for other purposes. To the extent that MOE covers most costs for primary education, there is some question whether many of the fees charged are justified educationally.

4.60 More fundamentally, the schools need appropriate management. MOE efforts to decentralize school inspection responsibilities to the district level should be given high priority as this would allow closer, systematic supervision of the schools. Initiatives to create school boards with parents, teachers, and school proprietors represented should also be given strong support, as these boards could help determine appropriate fee levels as well as monitor the use of revenues.

Nonformal Education

4.61 So far, this discussion has focused on access to formal schooling. But current levels of adult illiteracy and innumeracy are high (38% by some estimates) and some children (such as herd boys) are likely to be kept out of school no matter how accessible it is. MOE and some donors are concerned that more attention be paid to nonformal education (NFE) initiatives. Nonformal education is currently provided mainly through MOE's Lesotho Distance Teaching Centre (LDTC) and by the University's Institute for Extra-Mural Studies (IEMS), although several smaller initiatives are run by NGOs and community organizations. By and large, the impact of the two main NFE institutions has been limited mainly by a lack of focus on objectives and the dissipation of limited institutional capacity on a multitude of small-scale programs -- among them, (1) distance primary teacher upgrading; (2) correspondence school alternatives for secondary schooling; (3) entrepreneurship training and support for income-generating schemes; (4) basic rural education on a range of development topics; and (5) a number of basic literacy and numeracy programs.

4.62 In its five-year development plan, MOE correctly notes that this lack of focus inhibits the impact of all initiatives, and states its intention to focus on two main programs: (1) literacy and numeracy programs (including post-literacy support through reading materials and income- generating activities) targeted principally to out-of-school youth and secondarily to adults; and (2) alternative secondary schooling. In terms of reducing poverty, these emphases seem to be justified and the decision to concentrate on key areas seems reasonable. But without systematic evaluation of current programs in both areas it is difficult to assess their impact to date or their potential for sustainable expansion. We do know that current programs provide little coverage. - 106-

The current learning post program, for example, which is aimed primarily at illiterate herdboys, had reached only 2,000 individuals by 1991/92 (after more than 15 years in operation) and the attainment of enrollees was relatively disappointing. Very few NFE initiatives have succeeded anywhere in the world, so the relatively poor performance of Lesotho's NFE programs is not especially surprising. But MOE should continue to explore ways to improve the programs' coverage and impact. Priority should be given to implementing a proposed comprehensive evaluation of programs. Innovations suggested as a result of the evaluation should be carefully pilot-tested and closely monitored before significant resources are committed to program expansion. - 107-

5. ESTABLISHING AN EFFECTIVE SAFETY NET

A. INTRODUCTION

5.1 The main focus of economic policy should be to increase productive employment opportunities for the poor and to invest in human resources so that the poor can take advantage of such opportunities. But it would be unrealistic to expect the quick elimination of poverty. Many poor households will not benefit from strategies designed to stimulate economic growth and employment, either because implementation is delayed (given Lesotho's weak institutional capacity) or because it takes time to acquire the new skills and technologies needed for jobs or improved labor productivity. Moreover, frequent natural disasters make many rural households vulnerable to temporary but severe declines in income. And some labor-scarce households (including households with high dependency ratios, households headed by women or the elderly, or adults who are physically disabled or mentally incapacitated) will always be unable to take full advantage of income-earning opportunities. These vulnerable households need special assistance -- sometimes temporary, sometimes permanent -- to ensure their access to such basic necessities as food, clothing, shelter, and health care.

5.2 Strong informal and community mechanisms provide some assistance to those in need, but nearly 26 percent of the population lives in severe poverty, so some income transfers from public sources need to be targeted to those most in need. In general, except in labor-scarce houselholds, these transfers should be only temporary, and should be reduced as economic growth begins to make an impact on poverty. It is vital to avoid creating a permanent dependency on government assistance; transfers should, without destroying incentives to earn income, augment low incomes for extremely vulnerable households. Transfer schemes should help everyone in a defined beneficiary group without helping better-off individuals and households. And safety net or income transfer schemes should neither create unsustainable budgetary imbalances nor unduly strain administrative capacity.

5.3 Several government agencies and NGOs carry out a range of safety net programs in Lesotho, but often these programs are not reaching the poor in a cost-effective way. One reason current efforts are most often not cost-effective is that program designers and implementors have failed to target benefits to those in need. Most current programs have not explicitly defined the population groups to be assisted, which makes it difficult to target benefits. Some major projects -- including the school feeding program and the maize subsidy scheme -- make no explicit effort to target benefits to the poor, with the unintentional result that more funds are being transferred to better-off households than to the poorest Basotho households. Programs (such as the emergency feeding program) that have tried to identify and assist the poor on the basis of household or individual characteristics have often chosen unduly broad criteria, thereby identifying too large a population as legitimate beneficiaries. This put the program's financial viability at risk. However, there may be no clear-cut set of household characteristics by which to identify poor households in Lesotho; poor households are not easily differentiated from nonpoor houselholds. And the alternative -- to rely on village leaders to select program beneficiaries -- may invite nepotism or rent-seeking. Under the circumstances, the best way to target assistance to tl-.e poor may be self-targeted safety net approaches -- for example, lowering the price of goocls consumed primarily by the poor, or paying below-normal wages in public works programs desigiled to strengthen the social safety net. - 108-

5.4 Another inefficiency problem with Lesotho's safety net programs is that initiatives are highly fragmented. There is no framework of priorities (in terms of types of vulnerable households or individuals to be targeted), impact monitoring is virtually nonexistent, and the risk of duplication is high -- meaning that a single household or individual could draw benefits from more than one program while others with equal or greater need are shut out.

5.5 But the most pernicious effect of multiple, uncoordinated programs may be their effect on Lesotho's institutional capacity to provide assistance. As various initiatives to help the poor were launched, it was apparently more convenient (and more attractive to external funders) for the Government to establish a new office or department in a ministry to supervise the program than to strengthen an existing government structure. As a result, there is a plethora of program offices, often with roughly similar mandates, each staffed by one or two national professional staff and varying numbers of expatriates. When expatriate staff are withdrawn, usually upon completion of donor projects, the weakened units are left behind, each with less than a critical mass of Basotho staff to implement the mandated program effectively.

5.6 To transfer resources to the vulnerable more efficiently, and to ensure the sustainability of safety net programs, it is important to consolidate units, and the programs they manage, into larger entities geared to reaching clearly specified target populations. Some initiatives might be replaced and some might be expanded. The consolidated safety net system we recommend would do three things: (1) lower the consumer price of the main food staples; (2) shift from free food distribution schemes toward a substantially expanded labor-intensive public works initiative; and (3) provide malnourished pre-school children, pregnant and lactating women, the elderly, the chronically ill, and the disabled with supplementary food provided through community-based distribution points.

5.7 Lower the consunier price offourfood staples -- maize meal, sorghum meal, sorghum, and cooking oil -- that represent a major share of spending in poor households (and a minor share in richer households). Lower prices would be achieved principally by abandoning trade, marketing, and processing restrictions (which tend to raise the prices of these products). Consumers suffer a per capita loss of M21 a year -- or about M126 for an average household of six85 -- because of trade restrictions that benefit only Lesotho's inefficiently operated industrial mills. In effect, consumers are taxed to benefit the industrial mills and the few surplus producers existing in the country. Another M26 per capita -- or about M156 per household -- would be saved if South Africa would remove its international trade restrictions, and stop protecting inefficient domestic producers of maize. And because poor households spend proportionately more on food than better-off households do, current policies are unambiguously regressive. The poorest 20 percent of the population spends roughly two-thirds of its income on food -- especially maize meal -- while the top 20 percent spends only 30 to 35 percent of its income on a much more varied diet (spending only a fraction of its food budget on maize). Trade liberalization with South Africa would reduce consumer prices of maize by about 30 percent in the short run. Using the poverty lines established in Chapter 2, the implied reduction of food

95 The percapita loss is calculatedassuming a populationof about 1.7 million for 1992/93.The total losseach year is an estimatedM35.11 million, even with conservativeassumptions about costs. SeeAnnex 10 for a more detailed analysisof the wvelfareimpact of currentmaize policies. -109- expenditures would lower the incidence of national poverty from 49 to 44 percent, and the incidence of ultra-poverty from 26 to 18 percent.

5.8 The Government could also consider returning to the pre-1991 exemption of less refined food staples from the general sales tax. Poorer households tend to consume coarser qualities of meal, so better targeting might be possible by exempting from taxes only unsifted maize meal produced by industrial mills and all meal produced by village hammer mills, while retaining the tax on the more refined meals consumed by the better-off. This option would be feasible if the contemplated shift to a value-added tax system is implemented. Note, however, that Lesotho's value-added tax system will have to be in line with South Africa's to prevent smuggling. Because South Africa exempts food staples from the value-added tax, Lesotho would need to do the same anyway.

5.9 Slift from free food distribution schemes toward a substantially expanded labor- intensive public works initiative. Targeted to the able-bodied but un- or underemployed poor, this initiative would build on existing food-for-work schemes, the labor-intensive rural roads program, and recent drought emergency relief efforts. It would employ a self-targeting mechanism (below-normal wages) as well as some geographic targeting to reach the poor; it would probably emphasize cash rather than in-kind wages; and it would focus on projects (such as anti-erosion works, afforestation, and rural infrastructure) with high labor-to-capital ratios as well as high economic returns. Institutional constraints would be relieved partly by merging disparate government units, partly by establishing mechanisms to effect collaboration between line ministries, but mainly by involving NGOs, private contractors, and local communities in all aspects of program design and implementation. If adopted, this initiative could be financed by reallocating ineffective safety-net funds as well as through the Lesotho Highlands Water Project Development Fund.

5.10 Provide mnialnourishedpre-school children, pregnant and lactating women, thle elderly, thle chronically ill, and the disabled with supplementary food provided tilrough community- based distribution points. Objective indicators would be used to identify beneficiaries (for example, low weight for height) with health care facilities or village health workers certifying eligibility. If adopted, this vulnerable-groups supplementary feeding program would replace the poorly targeted school feeding program. Financing could come principally from money formerly used to fund the phased-out school feeding program. Logistical arrangements would rely heavily on NGO and community involvement, as school feeding now does. We recommend trying this scheme on a pilot basis, with careful evaluation of the pilots before a final decision on whether to proceed nationwide.

B. Current Safety Net Programs

5.11 Safety net schemes in Lesotho -- more extensive than in many comparable developing countries -- are implemented by numerous government agencies, usually with assistance from donors and local NGO's. Most programs are ongoing but some are created to deal with emergencies such as the recent drought. Some of the larger, formal efforts are described below. In addition, NGOs provide many specialized forms of relief. - 110 -

Non-Emergency Income Transfer Schemes

5.12 Safety net fund. The safety net fund introduced in conjunctioni with the structural adjustment program -- and financed by some of the counterpart funds genierated by IMF structural adjustment assistance -- is one of the larger non-emergency Government efforts targeted at poverty alleviation. Managed by the Ministry of Finance and Economic Planning (MFEP), it was designed both to reduce the transitory pressures of adjustment on househiolds and to support small projects to reduce poverty. The fund received M22 million over thlethlree years: M5 million in 1991/92; M7 million in 1992/93; and M IO million in 1993/94.

5.13 The use of funds is overseen by an MFEP-appointed committee headed by a senior economic planner, with representatives from several ministries and NGOs. The committee was set up too late to commit the initial M5 million, so funds from the first tranche reverted to the general budget. Half of the funds were approved for seven projects for an average M500,000 each. Major grants went to the Matuoesase orphanage (MI30,194), an extension to tile ltjareng Rehabilitation Center (M359,000), and the Maseru Children's Village (M375,000), amlonig others. The Government directed the committee to allocate the remaining M3.5 million to finance the initial operations of the Drought Relief Implementation Group (DRAG). No doubt these projects assisted vulnerable populations groups, but the commiittee (individuals withi other full-time jobs) did little to ensure that they targeted groups effectively -- that thiey reachied the needy. For lack of time, the committee has done little more than vet proposals and allocate grants. Uniform criteria have not been established to ensure targeting to thle mnostvulnlerable population groups and there have been no site inspection visits. The only mlonlitoringof beneficiaries' use of funds has been thiroughprogress reports, usually received onla regular basis. It would be difficult to know if the funds met the needs of the poor and vulnerable.

5.14 Microprojects programt. Thle MFEP is also the institutional hiome of a mnicroproj'ects program supported by EU grants, including funds from a small monetized food aid schieme. Run by a special Microprojects Management Unit (MMU) -- two expatriates funded by thle EU -- thle program supports small-scale, community-initiated grassroots operations. A grant covers up to 75 percent of the cost of each microproject, so the scheme provides direct income transfers to the individuals requesting support, albeit generally for development purposes. The project began operating in early 1992 with a MS million annual budget. Of the 100 applications received by mid-1993, 57 were approved. Projects were approved for borehiole drilling, classroom rehabilitation, training for bricklayers, and the purchase of knitting machlines, among others. Because the projects are concentrated in rural com-munities,they are broadly targeted to the poor, but no explicit poverty criteria are used to judge proposals. Moreover, a program so dependent on expatriate technical staff and funding is probably not sustainable.

5.15 Non-emergency food distribution progranms. The World Food Program-me (WFP) provides extensive food aid to Lesotlio in the form of assistance specifically targeted to poor adults (mostly women engaged in food for work in rural infrastructure schemes) and to children (who get food supplements at primary and secondary schools). The total cost to WFP is $28.0 million for the school feeding program for 1989-94 and $15.4 million for thie food part of the rural infrastructure program. WFP provided extra assistance in 1992/93 in response to the drought. In addition to the 20,000 metric tons of corn meal, maize meals, pulses, anid vegetable oil imported under the regular program that year, WFP imported an additional 15,800 mietric funds of food, which were channeled through emergency relief programs. 5.16 The school feeding program has provided meals for most of the children attending primary school and many of those attending secondary school (Table 5.1).

Table 5.1: School Meals Provided (1990-92) 1990 1991 1992 Primarystudents 342,455 307,287 263,917 Secondarystudents 5,033 5,197 5,300 Total 347,488 312,484 269,217 Source: World Food Program, Maseru.

5.17 WFP is phasing out its program in the relatively well-off lowland areas, by promoting self-reliance at those schools (gardens, livestock activities, and school water supply schemes), whlich explains the declining number of children fed. The school feeding project is run through the Ministry of Education, but a multi-ministry committee provides policy advice. Several other donors and NGOs (especially Save the Children) are active in implementation. The only apparent targeting mechanism is the decision to phase out the better-off lowland schools in favor of those in the mountain areas, where poverty is more concentrated. Parents contribute about M5 a child each year to partially offset transport costs for food.

5.18 Because this program is effectively a direct income transfer to the families of children receiving free school meals, it may encourage school attendance. But to the extent that the poorest households are least likely to send their children to school (possibly because of the up- front costs for uniforms and registration), they are also the least likely to benefit from the transfer. In this sense, the program is probably transferring more income to better-off hotiselholdsthan to poorer ones. Also, the meals are not targeted toward pre-school children who are far more vulnerable to undernutrition than are school-aged children. It is difficult to assess the nitritionial effects of the school feeding program within households without more study, but a platisible scenario is that when older children eat meals at school, they get less food at home, whiichifrees up food for younger siblings and adults.

5.19 The WFP-supported non-emergency program best targeted to the poor is the rural food- for-work infrastructure program. Implemented through the Civil Works section of the Ministry of Home Affairs, this program is self-targeting: By paying relatively low wages, it tries to attract only the people wlio most need money or food. Casual workers employed in WFP schemes receive a daily wage equivalent of M8 for a five- hour day, M2 in cash, and a WFP package of maize meal, vegetable oil, and pulses with a market value of about M6. The Government's minimum wage for unskilled labor working on road construction and maintenance projects is M 17 for an eight-hour day (equivalent to about Ml I for a five-hour day). About 80 percent of the workers are women, who work five hours a day and spend up to three hours walking to and from work. The program provides 20 days of work a year for 5,000 people who receive enough food rations at the end of the month to feed five people the following month. So altogether about 30,000 people a year benefit from the program. Demand for the program exceeds supply, and as so many villages and villagers want to participate, no worker participates for more than three to five montlhsa year.

5.20 The program's inability to meet demand weakens the program's self-targeting meclianlisimi(the low wage rate). Final choices of workers are generally made by village - 112- development councils -- rather than through self-targeting -- on the basis of physical fitness and social criteria (whether the household is landless, is headed by a woman, or has no other means of support). Many participants have complained that some local authorities use unfair practices in selecting beneficiaries, but as long as demand for jobs in food-for-work projects exceeds supply, some method of administrative rationing is necessary, and relying on local authorities is one feasible approach. An alternative would be to use people on a "first-come, first-served" basis, but this would means that some participants might have to wait in long queues in hopes of a job -- and not always get one. Another problem is that some groups of poor -- the physically disabled, the infirm, and many mothers with young children -- might not be able to benefit. Many countries provide creches for young children in order to make working easier for mothers, but a WFP evaluation report indicated that a pilot effort to do so in Lesotho was not effective. The reasons for this failure deserve careful analysis. Perhaps changes in the creche operation could make it more acceptable to Basotho women.

5.21 The food-for-work construction projects have been successful in transferring income to the poor, but there is some concern about whether they are realizing their full economic potential. In general, the 125 km of food-for-work rural roads built annually under the auspices of the Ministry of Home Affairs are inadequately maintained after construction. Responsibility for maintenance lies with the Ministry of Works' Labor Construction Unit (LCU) -- which also oversees construction of "normal" (budget-financed) rural roads which are also built using labor- intensive methods, but for cash wages. Apparently the Ministry of Home Affairs got involved in food-for-work road construction partly because of the slow approval process in the LCU Ministry of Works. Rather than try to speed up LCU approvals, the Government decided to bypass the Ministry of Works by setting up a new unit, the Civil Works Section (CWS), in the Ministry of Home Affairs. Under the present food-for-work scheme, Village Development Councils decide which roads or tracks are to be constructed or upgraded; these choices are submitted to the District Development Councils, which make final recommendations to the Ministry of Home Affairs, which authorizes projects without consulting the Ministry of Works. Not only does this system duplicate and fragment resources (two government units, CWS and LCU, have virtually the same mandate), but many of the roads selected by CWS which has limited technical and supervisory expertise are ultimately difficult to maintain. And the Ministry of Works has little commitment to the maintenance of roads about whose construction it was not consulted. Finally, CWS's limited capacity may act as a brake on the size of the food-for-work program. Ironically, informed observers believe to have some excess capacity but insufficient funds.

5.22 Governmentassistance to "paupers," orphans, and the disabled. Government assistance to paupers, the disabled, and orphans (or abandoned children) is the responsibility of the Department of Social Welfare (DSW). DSW, founded in 1976, was initially located in the Ministry of Health (MOH), but belonged to various ministries over the years before returning to MOH in the 1993 reorganization. DSW has little funding and has staffing problems (only one of eight authorized social worker positions are filled), but its three programs are part of the official government safety net. One program supports the Vocational Rehabilitation Center which provides training in practical skills for the disabled. A second program (pauper support) gives each person who qualifies M50 a month. To qualify, a person must present a letter from the village chief certifying that he or she is destitute. This is then verified (with difficulty) by a social worker and, if approved, the person becomes eligible for the pauper stipend, which is paid by the District's Secretary's office. Only 500 people, most of them in Maseru, are registered as - 113 - paupers. DSW's third program provides support for abandoned children. DSW works with the Lesotho Save the Children Fund to find homes for the children, as there is only one official orphanage in the country (although churches and some NGOs run small homes for temporarily abandoned children). DSW, with a budget of under Ml million and a staff of 32 (out of 52 authorized positions) has a marginal role in poverty alleviation. Given so many other programs of direct assistance to the rural poor, there is little apparent reason to expand DSW's program of pauper assistance -- given the difficulty of verifying an individual's status as a pauper, the possibly unreliability of chiefs certificates, and the program's heavy urban bias. But DSW's vocational rehabilitation and abandoned children programs may deserve expansion with the support of NGOs.

5.23 NGOprograms. It is difficult to assess the role NGOs play in providing non-emergency assistance to vulnerable population groups without fuller information about NGO projects. Avail-able information suggests that the many NGOs and church missions active in Lesotho -- 169 registered NGOs and many more unregistered NGOs operating in rural areas -- are providing support in-cash and -kind to the especially vulnerable, including abandoned children, orphans, the elderly, the handicapped, the homeless, and poor households headed by women. NGOs are an important part of Lesotho's safety net for the destitute, contributing perhaps 20 percent of their estimated annual expenditures of M75 to MIOO million86 to non-emergency safety net activities. A review of a sample of programs reveals that some NGO income transfer schemes could be improved. For example, NGO-supported programs (World Vision, Lesotho Save the (Children, and The International Save the Children Fund) provide tuition assistance and ensure that distribution is fairly equal between primary and secondary education, although a case could be made for concentrating on the primary level.

5.24 Maize meal subsidy. In addition to these direct income transfers to the poor,87 the Government provides indirect income transfers to all consumers in subsidizing maize meal channeled through the three main industrial maize millers. That subsidy totaled about M12 million in 1992/93 but was expected to fall to about half that level in fiscal 1993/94, and currently no subsidy exists. The subsidy was introduced primarily to offset the effects of the drought. Its effect was to lower the consumer price of maize meal, effectively subsidizing its purchase. In a sense, the maize meal subsidy was an important part of Lesotho's official safety net, albeit one poorly targeted to the poor.

5.25 The maize meal subsidy went through certain steps as it passed from the Government to the purchaser. First, the Government gave the three industrial millers a subsidy (M104.10 per 1992-93 per metric ton of maize grain processed). After each ton of maize grain was converted to meal and offal (used as animal feed), the millers applied the full subsidy to the wholesale price of maize meal (applying none of it to offal). As a result, different levels of subsidy were applied to different qualities of meal; the more refined the meal, the less the offal (and the more the subsidy). The assumption was that wholesalers and retailers would pass the reduced price on to consumers.

86 Excludingchurch contributions to healthand education.

87 A direct income transfer occurs when consumers are given money or in-kind gifts. An indirect income transfer occurs when consumersare able to buy goods or services (for example, subsidizedfood commoditiesat belownmarket prices). - 114-

5.26 There were two main problems with the system in terms of poverty reduction. First, the Government's monitoring of the system to ensure that the reduced price of maize meal was passed on to the consumer was minimal. All monitoring occurred at the industrial mills where invoices for sales to wholesalers explicitly reflected the subsidy. There was no monitoring further down the chain to ensure that the lower price was passed along to consumers, even though the limited numbers of maize meal wholesalers made oligopolistic pricing possible. Competition among retailers reduced the risk of unfair pricing, but some abuse may have occurred in remote areas (where the poorest households are) served by one or only a few retailers.

5.27 The second problem is that the subsidy did not chiefly benefit the poor. Consumers of the coarser meal -- mountain dwellers and the rural poor -- got a lower subsidy per unit of meal consumed than did consumers of the more refined maize meal products (who tend to be better off). Regional sales data from the industrial mills show that demand for unsifted maize meal is higher in the (poorer) mountain districts than in the lowlands and urban areas. The amount of subsidy was higher when the proportion of meal per metric ton of grain is higher, so the subsidy on premium quality "special sifted" maize meal (M137.50) was about 26 percent highlerthan the subsidy on unsifted maize meal (M 108.78) (Table 5.2).

Table 5.2: Subsidy of Different Grades of Maize Meal (Per Ton, 1992/93) Meal/ton of grain Offal/ton of grain Subsidy/ton of meal Specialsifted meal 78% 22% M137.50 Sifted meal 83% 17% M 121.47 Unsifted meal 86-87% 14-13% M108.78 Source:Maseru Roller Mills.

Emergency Drought Relief Programs

5.28 The largest government program addressing short-term problems of deprivation related to Lesotho's 1992 drought was run by the Drought Relief Implementation Group (DRIG), located in the Prime Minister's office. DRIG was established in July 1992 specifically to deal with the drought. It operates with money from the Safety Net Fund and grants from several international institutions. DRIG's proposed budget for 1992/93 totaled about MIO million; 80 percent was approved. With a staff of 10 Basotho and expatriates, it implemented a nationwide feeding and disaster mitigation program in the 12-month period from July 1992 to June 1993. DRIG accomplished this with food aid, primarily from the World Food Program, and by relying on four NGOs -- the Lesotho Council of NGOs (LCN), World Vision, the Christian Council of Lesotho, and the Lesotho Red Cross -- to identify and register some 330,000 vulnerable individuals in conjunction with village Development Councils (VDCs) (Annex 13). Registered individuals received free food rations of maize meal, pulses, and oil under DRIG's Vulnerable Household Feeding (VHF) Program. - 115-

Table 5.3: Coverage of Vulnerable Household Feeding Program (by District) District Participating Numberof Populationcovered NGOs beneficiaries (%) Berea LCNw 39,001 18.2 ButhaButhe LCN' 22,430 24.5 Leribe LCN" 44,790 13.7 Mafeteng CCLb/ 2,242 18.4 WVC/ 41,976 Maseru LCNA/ 53,272 11.7 Mohale'sHoek WVC/ 34,798 16.8 Mokhotlong LRCd 18,286 21.1 Qacha'sNek LRCdI 20,655 22.7 Quthing CLb 19,940 14.5 Thaba Tseka LRCd/ 25,147 24.9 Total -- 322,547 Source:DRIG. Note:Registration had not beencompleted in threesub-areas of Maserudistrict when this table wasprepared. It wasanticipated that an additional7-8,000 individuals would be registered there. a/ LesothoCouncil of NGOs; b/ ChristianCouncil of Lesotho;cl WorldVision; d/ LesothoRed Cross.

5.29 The VHF program played an important role in providing direct income transfers (mainly food supplies) to many poor rural families throughout the long drought to help ensure their survival and health. Some 11,000 metric tons of food (mainly maize) were distributed in the ten months ending April 1993. Especially impressive was the speed with which DRIG established and implemented the feeding program, using existing institutional resources. Most important was DRIG's early decision to involve NGOs fully in the program planning and management and then to rely on their extensive local knowledge and networks of volunteers to oversee the program locally. Despite the absence of a formal link with the Food Management Unit (FMU), which has traditionally handled logistics for non-emergency food-for-work schemes, DRIG coordinated its efforts with FMU's in the movement of relief food supplies. FMU established 190 new food distribution points (in addition to the normal 35) from which NGOs could distribute food to vulnerable households. Some abuses occurred, as might be expected in a large-scale program, especially one launched rapidly in emergency circumstances. In some cases, for example, church leaders tried to get free food for all parishioners, regardless of need. However, serious problems were detected and dealt with effectively.

5.30 DRIG coordinated three other programs to increase transfers to the rural poor during the drought: (I) additional non-emergency food-for-work schemes, so an additional 50,000 people could earn cash for food; (2) a supplementary feeding program targeted to pre-school children, pregnant and lactating women, and the chronically ill (for example, tuberculosis patients); and (3) a therapeutic feeding program aimed at rehabilitating children admitted to hospitals for severe malnutrition. These programs all built upon existing non-emergency programs. Expansion of the food-for-work road construction program was made possible by DRIG's insistence on lower standards of supervision and road selection in the Civil Works section of the Ministry of Home Affairs, whose limited capacity has long constrained food-for-work - 116- schemes.88 The supplemental feeding program, supported technically by the Ministry of Health, was administered through existing village school feeding committees that provided either pre-set rations of cooked food at a suitable community site or take-home "dry" rations to beneficiaries whose weight was low for their age.89 Therapeutic feeding, accompanied by daily weight measurement, was done in health facilities, where malnourished children were fed about 840ml of a mixture of dried skimmed milk, oil, sugar, and water in five to seven feedings a day.

5.31 Data from the Sechaba 1993 household survey90 indicate that the DRIG relief food distribution programs (the vulnerable households program and special programs) probably did not experience serious problems of leakage to better-off households. In the 31 constituencies where questions were asked about food aid, about 17-18 percent of nonpoor hiouseholds received free food, which is perhaps an acceptable level of mistargeting, under the circumstances. A more serious targeting problem was that over half of poor households were not reached by the program. About 45 percent of the poor households and 49 percent of ultra-poor houselholds received food aid which means that many poor people suffered prolonged undernutrition. The apparent rapid increase between 1992 and 1993 of stunting caused by chronic undernutrition among children under five may partly reflect this failure of program targeting.

5.32 There was also no mechanism to prevent households from benefiting from two or more emergency and non-emergency safety net programs. In theory, for example, a household could receive relief food from the VHF program, one household member could get work on a food-for- work project, the household could get rations for pre-school children and mothlers from the supplemental feeding program, and school children could participate in the school feeding program! Such abuses may not have been widespread because local leaders -- presumably knowledgeable about local households -- helped identify beneficiaries for many programs. But duplication could have occurred where local leaders were not vigilant or were unfamiliar with community residents.

C. MAJOR PROBLEMSAND LESSONSLEARNED

5.33 How effective and efficient were Lesotho's various safety net programs? What common lessons did we learn from our analysis that might improve the design of future initiatives? The chief problems were weak targeting, fragmented programs, and weak administrative capacity.

88 Thesetemporary measures were justified by the drought emergency,but a quality review should probably be undertakento ascertainwhether the works projectsbuilt underthe relaxedguidelines met standardssimilar to those met on regular food-for-work activities. If they did, perhapsexisting guidelines are over-stringentand should be modified; if not, retumsto the emergencyfood-for-work projectsare probably relatively low.

89 Rations consistedof maize, pulses,and oil, providing for a daily caloric intake of slightly over 1,850 caloriesand 25.8 gramsof protein pertwo meals.

90 In the Scchabasurvey, questionsabout relief food wvereasked of householdsin 31 of 60 constituencies. Resultsare for thoseconstituencies only, but probably suggesthow the whole programoperated. Unfortunately,the 29 constituenciesnot coveredwere those the Early WarningSystem identified as bcing morevulnerable to drought! - 117-

Weak Targeting

5.34 Virtually all safety net programs in Lesotho are poor at targeting help to the intended beneficiaries. Targeting problems take at least four forms:

5.35 First, some programs, such as the school feeding program and the maize subsidy, make no attempt to direct resources to the most vulnerable groups in Basotho society, so providing a safety net for the poorest households is needlessly costly. Worse yet, by distributing benefits through institutions that may not reach the poorest households, such as schools, the safety net programs may sometimes miss the poorest households altogether.

5.36 Second, some programs, such as the safety net fund and the microprojects scheme, try to target the vulnerable but fail to precisely define the target group, do not accurately identify beneficiaries, and do not evaluate the coverage and efficiency of resource transfers. As a result, it is hard to assess how such programs affect poverty and there may be a tendency to support programs with disproportionately highi administrative cost for the number of beneficiaries assisted.

5.37 Third, even when criteria are defined to identify beneficiaries (as in the VHF program), they may be so wide that too many people are eligible for the benefit -- exacerbating shortages and thus excluding those most in need. And some criteria may be difficult to apply because the data are hard to verify. For example, if rural households were included in a feeding scheme on the basis of whether or not they produced 180 kg of cereal a year per household member, over 70 percent of rural households would qualify as beneficiaries in a normal agricultural year and virtually all households in a disastrous year. But measuring household cereal production is difficult and time-consuming, especially if accuracy is expected. This type of problem has sometimes been mitigated by including additional objective criteria (such as markers for the undernutrition of young children). Precise targeting through easily measurable, objective criteria may not be feasible in Lesotho because of so many overlapping characteristics in ultra-poor, poor, and non-poor houselholds.

5.38 Finally, Lesotho's rationing of benefits currently relies heavily on local leaders (village chiefs and development councils or other grassroots organizations) selecting program beneficiaries -- especially for food-for-work construction projects and the VHF program. This approach has three main drawbacks: (I) It is costly and diverts conscientious leaders from their routine and important responsibilities; (2) many leaders may be unable to identify the most marginalized -- and least visible -- poor members of a village; and (3) leaders responsible for identifying beneficiaries are at high risk for nepotism -- especially if the standards of living of the poor and the nonpoor are not dramatically different, as is true in Lesotho.

5.39 Much more emphasis should be placed on self-selection or self-targeting mechanisms. This would reduce program cost for each legitimate beneficiary and increase the proportion of legitimate beneficiaries reached. Shifting from administrative targeting to self-targeting would require redesigning some programs because self-targeting generally would require making the benefit offered (or means of gaining access to the benefit) less attractive to the nonpoor than to the poor. The only large scale safety net program that uses self-targeting is the food-for-work construction program, where the nature of the work and the low wages paid tend to discourage - 118 - individuals from better-off households for whom the opportunity costs of participating are likely to be higher than the benefits.

Program Fragmentation

5.40 The high degree of institutional fragmentation in Lesotho makes it difficult to set priorities or targets for safety net programs. The lack of clear measurable goals makes it difficult to assess various initiatives' impact on poverty or the efficiency with which resources are used. The fragmentation of programs and the limited sharing of information among them also increases the risk of duplication, increasing the chances for some households to benefit from more than one program while others fall between the cracks.

5.41 Programs have become fragmented partly because donors have insisted that separate project offices be set up to implement specific externally funded programs, rather than take on the more difficult task of strengthening or revitalizing existing offices. Nowhere is this better illustrated than in the response to the recent drought when donors and the Government bypassed an understaffed agency, the Disaster Management Unit (DMU) of the Ministry of Home Affairs established during drought in the 1980s and closely involved with the FAO-supported early warning system, and set up a new Drought Relief Task Force and its executing arm, DRIG. DRIG's efforts provided an extensive safety net, supplementing existing targeted poverty programs, while DMU -- with offices in each district -- remained for all practical purposes outside this relief effort. Similarly, CWS was created within the Ministry of Home Affairs to implement WFP-supported food-for-work road construction projects -- despite LCU's substantial experience in rural road construction, within the Ministry of Public Works.

Weak Administrative Capacity

5.42 Closely related to the problem of institutional fragmentation is that of weak administrative capacity. Typically, each new initiative is initially managed by a team of one or two professional Basotho staff and three or more expatriates funded by the donor or an NGO involved in the initiative. After a time, the expatriate staff is reduced or withdrawn, leaving a weakened unit with too few Basotho professionals or support staff to carry on the program effectively. When the next initiative is launched, the unit is considered to be too weak to run the new program and another unit is set up, starting roughly the same cycle over again. As a result, skilled, national staff are scattered among too many different government agencies and units to be effective. Consolidating initiatives into larger programs would help alleviate the problem of weak institutional capacity and would make programs more sustainable and facilitate the setting and monitoring of goals.

5.43 Improving collaboration between organizations in the design and implementation of safety net programs would reduce some capacity restraints, as DRIG's recent experience demonstrates. DRIG's success in coordinating activities was facilitated by DRIG's visibility and by everyone's recognition of the urgency of the situation, but valuable lessons about the organization of Lesotho's safety net programs can be drawn from DRIG's operations. First, considerable institutional capacity exists through NGOs, which are willing to cooperate with Government in sound poverty reduction programs. Second, effective collaboration is possible among government agencies with different objectives if there is clear evidence of high-level commitment to an initiative. And finally, the key to successful collaboration is to involve NGOs - 119 - early (preferably in goal-setting and program design) and to systematically consult them throughout implementation.

D. CONSOLIDATINGTHE SAFETY NET

5.44 In this section, we propose a framework for a more consolidated safety net system targeted to the poorest individuals and households by focusing mainly on self-targeting mechanisms, because self-targeting costs less and is better at reaching the poor. Institutional capacity strengthening is a pre-condition for implementing these safety nets (Chapter 6), and can greatly benefit from (I) consolidating disparate units; (2) strengthening existing delivery mechanisms rather than creating new ones; and (3) developing partnerships with NGOs and the private sector. In addition, because Lesotho is so susceptible to climatic and other exogenous shocks, such as retrenchments in the employment of mine workers, the Government should try to build the flexibility to respond quickly and effectively to sudden adversity.

5.45 The safety net proposed here takes three roads to addressing the needs of the poor: first, making policy changes to lower prices for basic food commodities to indirectly help all poor households, rural and urban; second, substantially expanding labor-intensive rural public works projects that give income in cash or kind to poor rural households with unemployed or underemployed able-bodied adults; and third, giving food supplements to those with special needs (such as malnourished children and disabled adults), both rural and urban. Table 5.4 summarizes the purposes and nature of, and possible delivery mechanisms for, the three main activities of the proposed safety net.

Table 5.4: Proposed Safety Net System Proposedaction Targetpopulation Type of transfer Delivery mechanism The price of basic food Market-dependentpoor Self-targeted, Markets commodities households(urban & rural) indirectincome that may buy food staples transfer Expandlabor-intensive Poor rural householdswith Self-targeted,direct Government public works projects unemployedor incometransfer agencies,NGOs, underemployedadults; privatecontractors householdssuffering transitoryshocks Providefree food Malnourishedunder-5 Administratively Healthfacilities, supplements children;pregnant and targeted,direct local leaders lactatingwomen; elderly incometransfer or disabled adults

Lower Food Prices

5.46 Poor consumers spend so much of their income on food, especially on basic staples, that policies that increase basic food prices are extremely regressive, as we explained in chapter 2. Reducing the cost of food consumed by the poor would greatly benefit Lesotho's poorest groups, because 80 percent of households (70 to 75 percent of rural households) are net consumers of food. Rural-dwellers produce less than 20 percent of the food they consume and depend heavily on food purchases. Lowering consumer prices for basic foodstuffs would free up household resources that could be used to pay school fees or buy more or better food or other essentials - 120- such as clothing. Lowering food prices would be self-targeting because proportionately more income in poor households goes to basic food purchases, requiring virtually no new institutional capacity for delivery. The prices of basic foodstuffs can be reduced by: (1) adjusting commodity market regulations that exert upward pressure on food prices; (2) exempting key food items from sales or other taxes; and (3) applying subsidies to selected foodstuffs. We propose emphasizing the first two approaches initially because of the greater risk of market distortions and leakages with subsidies.

5.47 Maize meal. Clearly, the cost to consumers of current restrictions on imported grain processing, and marketing -- M21 per capita a year or M126 a year a household of six -- is a tax on households that benefits the industrial mills. All consumers of maize meal, rich and poor, pay this tax, but it most hurts poor and ultra-poor households. For a household right on the ultra- poverty line (M35.24 a person a month), the average savings from liberalizing trade in maize grain would roughly equal about 6 percent of 1992/93 income. And with South Africa's recent move to free the price of maize grain, the price differential between maize and maize products in Lesotho and South Africa is even greater -- so the potential savings would be even higher now. Liberalizing maize pricing, trade, and processing policies would not only increase the purchasing power of poor households, but would also promote a shift toward the production of nontraditional crops (a strategy for increasing export earnings).

5.48 Consumers of maize meal would also benefit from an exemption from the general sales tax for maize products. Such an exemption existed until December 1991, when it was removed. The rationale for this action was to simplify revenue collection. There has been no systematic monitoring, but officials in MFEP said that removing the exemption has not appreciably improved revenue collection. When village mills' production and sales of maize meal increases -- as can be expected after liberalization -- the task of collecting the general sales tax will becomemore complicated. All maize meal products could once again be made exempt from the sales tax, and the subsidy ended, at the same time that maize grain imports are liberalized.

5.49 A shift from the general sales tax to a value-added tax (VAT) system,9 ' an option the Government is considering for 1995, would make it possible to target the tax exemption to the poor by exempting only the course meal produced by village mill and the unsifted maize meal producedby industrial millers -- products consumed mainly by the poor. A VAT could still be collected on the more refined maize meal products consumed by the better-off. Such a strategy would improve government revenues but may be politically difficult to implement. Similar targeting is theoretically possible with a general sales tax, but the practical difficulties of monitoring retail sales all over the country for tax collection (and sales reports of types of maize meal product) would probably make fair implementation extremely difficult.

5.50 Other food items. Maize meal is the single most important product consumed by the poor, but the Government should also consider lowering the cost of other basic food items consumed by the poor, especially sorghum, sorghum meal, and cooking oil. As soon as possible, the Government should eliminate anv marketing, trade, or other restrictions that could increase basic food costs to consumers and should consider exempting food staples from the general sales tax (after analyzing the budgetary implications of such an action).

9 The VAT is collectedat (he point of production;the general sales tax is collectcdat the retail level. - 121 -

Expand Labor-intensive Public Works

5.51 The second -- and most important -- element of the proposed safety net system is to shift away from free food distribution (for example, the VHF and school feeding programs) and other poorly-targeted efforts toward more reliance on labor-intensive public works, drawing on experience from LCU's ongoing road projects, the WFP-supported food-for-work road construction program, and some emergency programs. Most poor households in Lesotho -- including most of those headed by women -- include able-bodied but unemployed and underemployed adults, and it makes sense to give them employment opportunities on temporary public work schemes that provide infrastructure and public goods (including environmental conservation) in rural areas. A major public works initiative -- expanding on the current framework -- would increase short-term employment opportunities for those most in need of additional income; it would provide vital infrastructure needed to increase rural output and services provision; and (if flexibly designed) it could be expanded in times of emergency, such as mine retrenchment. Unlike special emergency programs activated at the time of major national disasters, a regular but flexibly designed public works program could help households during both major emergencies and smaller, local disasters that occur routinely because of Lesotho's variable climate.

5.52 Not only would public works programs be easier to efficiently implement than free food distribution schemes, they might be more useful psychologically. Employment itself seems to offer social benefits that are difficult to quantify. Prolonged unemployment damages morale and self-esteem, and can lead to attitudes and habits that make people less productive and employable. Certainly it is better for society to maintain a potentially productive labor force than to create a population dependent on handouts. Developing a work ethic is especially important with unemployment rates so high among Lesotho's young adults who face diminishing job prospects in the South African mines (Annex 14).

5.53 A new public works initiative should be designed to reach Lesotho's poorest households and to exclude better-off households. Self-targeting approaches, combined with broad geographic targeting, are most likely to achieve this objective. Lesotho has had some experience with self-targeting in the WFP rural infrastructure program, which pays below-normal wages in cash and food to only poor individuals, for whom the opportunity costs of participation is relatively low. The level and type of payment need further evaluation before expanding the public works program. Excess demand for jobs probably reflects the small scale of the WFP/CWS scheme, but it might be necessary to reduce the level of pay so that only the very poor apply and there is no need to choose among applicants.

5.54 The ratio of cash to food should also be reconsidered. By and large, except in emergencies that disrupt normal food marketing, cash payments are much cheaper and easier to administer than payments in food, which require a special distribution system, including transportation and storage facilities. Cash payments are also more likely to stimulate demand for locally produced food and other commodities as well as rural retail and wholesale trade. Another factor to consider in that the ratio of cash and in-kind payment affects the gender mix of participants. In some African countries, cash payments tend to attract more men, who apparently find it less acceptable to work for in-kind payments. Finally, it may be difficult to implement an - 122 - expanded works schemes nation wide; if so, the initial priority should perhaps be on those regions and districts where the incidence of poverty is higher.92

5.55 What types of projects should be supported under an expanded public works program? Clearly, labor-intensive projects with potentially high economic or social returns should be considered. For example, such projects could include (l) rural road construction, rehabilitation and maintenance; (2) erosion control, afforestation, soil improvement, the furrowing of steep slopes, small-scale irrigation schemes and water storage schemes; and (3) the rehabilitation or construction of schools, clinics and safe water systems. One way to choose among possible projects is to select those that employ the most labor per Loti spent, although those are often the projects whose asset-building effects on income and employment mature only in the long term. We do not have information in Lesotho on the relative labor-intensity of various types of public work schemes, but we do have such information for public works projects in comparable African countries and one South Asian country (Nepal) with topography similar to Lesotho's (Table 5.5).

Table 5.5: Labor Intensity of Selected Public Works Projectsa/ Public works program Labor intensity coefficients0 (rangeof observations) Anti-erosion works 66.9-98.7 Afforestation 70.0-74.0 Wells, reservoirs, small irrigation systems 41.0-76.0 Dirt roads 45.6-62.0 Buildingconstruction (schools, clinics etc.) 8.2 Source: Joachim Von Braun et al, Labor-(ntensive Public Works for Food Security: Experience in Africa, International Food Policy Research Institute, Working Papers on Food Subsidies #6, 1991. Notes: a/ Basedon data from Burkina Faso, Burundi, Rwanda, Nepal and Tanzania. b/ Labor costs as a percent of total costs.

5.56 As Table 5.5 shows, soil conservation activities and afforestation -- both essential to increasing agricultural production in Lesotho -- have extremely high labor-to-capital ratios and should therefore be given high priority in public works programs, but past efforts by WFP/CWS to implement food-for-work activities in these areas encountered serious problems. Described in a recent evaluation report93 a lack of integrated guidelines for land use planning and soil conservation led to fragmented interventions so farmers had little incentive to maintain erosion control activities. Afforestation efforts similarly suffered for lack of coherent forestry policies and programs. Villagers also resented an MOA decision to convert village woodlots into official forestry reserves that provided few benefits for the villages. The subsequent inability of MOA's Forestry Division to manage marketing from 400 dispersed woodlots further reduced returns to the afforestation program, so food-for-work activities in forestry and soil conservation were phased out beginning in 1992. Any renewed activities in these areas -- highly desirable from the standpoint of poverty reduction -- must include the design of viable, integrated operational strategies based on priorities outlined in the National Environmental Action Plan (1988), The

92 But tradeoffs must be weighed to maximize each project's economic returns. The poorer areas (in mountain districts) tend to have less economic potential, so investments in infrastructure there are likely to yield lower retums.

9' WFP/FAO/ILO Interim Evaluation Summary Repon on Project Lesotho 0352 (Exp. 7) -Development of Rural Infrastructure with Appraisal of Ncw PhaseApril/May 1993. - 123 -

Soil and Water Conservation Policy (1988), the Conservation Incentive Package (1992), and the 1993 Congress Party Manifesto.

5.57 As much as possible, specific projects should be selected within the framework of general development plans (for conservation and rural infrastructure) and with the full participation of the communities affected. As a general rule, communities should be encouraged to take the initiative and identify the most desirable projects for their locality, possibly in collaboration with NGOs or other grassroots organizations active in the area. WFP/CWS has already adopted this practice in food-for-work schemes and should review it to determine if any modifications are needed. After a technical evaluation of proposed options, projects selected should also be implemented by workers from the beneficiary community. A strategy of decentralization will increase the chances of projects being sustainable by getting communities and households to assume responsibility for, and reap direct benefits from, the completed projects.

5.58 Is its crucial that Lesotho develop the institutional capacity to implement and manage the expanded public works program. First, it should consolidate all units with experience managing labor-intensive public works or designing safety net programs, especially LCU, CWS, DRIG, DMU, and FMU, and perhaps certain staff members or scattered groups within such line ministries as MOA or MOH. A single entity should have responsibility for the design and implementation of public works. That entity, which might include specialized divisions, should be responsible for:

* Liaison with relevant line ministries to develop broad operational plans for soil conservation, water storage, afforestation, and rural roads. * Designing phased implementation plans, including specific annual targets for numbers of projects by geographical region, jobs to be generated, and so on. * Evaluating and approving project proposals and providing technical supervision for projects (sometimes in conjunction with line ministries), through clearly established mechanisms. * Continuously evaluating and monitoring projects and programs so that lessons learned are incorporated in project designs and so that the appropriate target groups benefit. * Providing logistical support for food-for-work projects * Maintaining and using the early warning system to target needed program activities. * Coordinating with MFEP and donors to ensure appropriate financing. * Coordinating with VDCs, district agencies, NGOs, other community-based organizations, and private contractors involved in project implementation.

5.59 Project implementation should be as decentralized as possible (within specified guidelines), building on recent VHF successes. Probably the only way to implement a meaningful safety net program in Lesotho is to involve NGOs, private contractors, and locally based organizations (following training, if necessary).

5.60 Other lessons from the DRIG model can be applied in the public works initiative. It is clear that the highest levels of government must visibly show strong commitment to the initiative; a case could be made for locating it in the Prime Minister's office. Possibly a committee of senior line ministry officials and NGO and donor representatives should oversee the initiative, both to tap into their collective experience and to facilitate technical coordination. - 124-

For all initiatives, it is crucial that all agencies essential to an effort be involved at an early stage and consulted throughout implementation.

5.61 Finally, an expanded, consolidated public works initiative will require far more financial resources than are now devoted to the WFP/CWS rural infrastructure scheme and the LCU road construction activities. Between M12 million and M14 million annually has provided temporary or permanent jobs for approximately 6,500 workers, and about 39,000 beneficiaries, including households of the employed persons. Expanding the program over three to five years to reach a target of, say, 30,000 jobs (or 180,000 beneficiaries) in nion-emergencyyears requires at least quadrupling the current level of financial resources.94 Some funds should be reallocated from current safety net programs of doubtful efficiency and sustainability -- for example, the safety net fund, pauper assistance, and possibly the microprojects scheme. Another major source of funds that should be tapped is the LHWP Development Fund, part of which is intended for rural development activities. Finally, budgetary savings from activities unrelated to the safety net (for example, subsidies to parastatals are no longer needed because of privatization) could also be allocated to public works, as appropriate.

Provide Food to Supplement Vulnerable Groups

5.62 The third element of the proposed safety net program is to replace the current school feeding program with a community-based supplementary feeding program targeted to the most vulnerable groups in Basotho society: malnourished pre-school children, pregnant and lactating mothers, and individuals unable to work because of severe mental or physical disability, chronic illness, or old age. This program could build on the recent drought emergency initiative, under whicih local communities (often the school feeding committee) weighed children under five and provided meals to those identified as malnourished. High rates of malnutrition prevail among Basotho pre-school children, which could seriously limit mental development, learning, and ability to work. Extending the drought initiative (possibly in a modified form) is far more justifiable -- in terms of reducing poverty -- than the current practice of providing free food to children enrolled in school. We propose an accelerated phasing out of the school feeding program.95 The money currently spent on school feeding -- between M17 million and M18 million annually, excluding parental contributions -- should be reallocated to a vulnerable groups supplementary feeding program. We suggest only the broadest outlines of program design below.

94 Some savingsarc clearly possiblein existing programs. For example,LCU pays workers almost 40 percentmore than WFP/CWS does. This level should be reexamined,at leastfor unskilled workers,with an eye to establishingequity and uniformity among programs.

95 Except perhapsin the mountainzone. - 125 -

5.63 Poverty reduction through lowering staple food prices and expanding the rural public works program rely on self-selection or self-targeting. The vulnerable groups feeding program would depend on administrative targeting of beneficiaries. The identification of malnourished children, using specific anthropometric measures, should be carried our either through health facilities or by specially trained women from the community, accompanied by the village health worker. 6 If feasible, there should be an emphasis on growth monitoring rather than on one-time measures (such as weight or height for age), and food should be given to children who show evidence of growth faltering (as defined by MOH), with benefits provided for three to six months, during which period they should continue to be weighed. Other potential beneficiaries, including pregnant or lactating women, the disabled, the chronically ill, and the elderly (perhaps those over the age of 65) should probably be identified by health clinics, who could provide a letter certifying eligibility for supplementary food. The Ministry of Health should carefully choose criteria for certifying individuals as disabled or chronically ill, because some types of disability and illness do not necessarily prevent an individual from working.

5.64 Benefits may need to be tailored to the specific group being assisted. The program should probably not try to meet a person's full nutritional requirements, as this would simply displace food now being provided from other sources, including strong informal community mechanisms. Moreover, keeping the level of benefits low would limit the potential for abuse, as only those truly in need would be willing to take the time and trouble to receive the benefit. Experience withi feeding programs for children and pregnant or lactating women in several countries shows that a nutritionally rich "snack" (an enriched bun, for example), rather than a recognizable meal, is least likely to replace food provided in the home. Obviously, the nutritional impact of the supplementary feeding scheme would be wasted if the intake of food from other sources is reduced in proportion to the food provided by the program. It is likely to be enhanced, however, if mothers are also educated about child nutrition, including breast- feeding, weaning practices, and the critical importance of feeding young children several small meals a day. If agreement is reached that an appropriate snack be provided, the MOH should specify its nutritional content and make arrangements for its preparation. For disabled, chronically ill, and elderly person, it would probably be feasible to provide a take-home ration of food commodities such as maize meal, pulses, and oil equivalent to about half of daily caloric requirements (that is, about 1,000 calories).

5.65 The food should probably be delivered to distribution sites roughly the same way as the school feeding program or the supplemental feeding program during the drought -- that is, making substantial use of NGOs. Food can be distributed monthly to the chronically ill, and elderly (if they are provided with take-home rations). For malnourished children and pregnant or lactating women, however, options for daily distribution should be considered, with participants weighed once a month. Responsibility for distribution to all beneficiaries should be assumed by an appropriately trained community-based group (for example, a reorganized school feeding committee), supervised by health staff or trained NGO volunteers.

96 Using village health workers to identify beneficiarieswould probably enhancetheir standing in the community, which would make themmore effective in their other duties. - 126-

5.66 Whatever the program design and delivery mechanism, carefully monitored pilot initiatives should begin before a full-fledged program is launched. Results from pilot efforts should be used to design the national program and to estimate probable coverage of (and impact on) target beneficiaries and it probable costs. Only if the pilots work, should a nationwide program be launched. If a national program is initiated, start-up should be accompanied by a mass education campaign explaining the reasons for the program, its goals, how it works, and who is entitled to what. After communities understand how the program works, the focus in mass information, education, and communication (EIC) should probably shift toward nutritional education. - 127-

6. STRENGTHENING INSTITUTIONAL CAPACITY

A. INTRODUCTION

6.1 Making a sustainedimpact on poverty in Lesotho will require coordinatedaction in many sectors plus an integrated approachto special assistanceprograms aimed at reducing poverty. It will be necessaryto set clear poverty goals and targets, reorient policies and programs in many sectors,expand some programs and eliminate others, consolidate efforts where there is duplication and fragmentation,and continuously monitor and evaluate how developmentprograms are affecting poverty. Moreover,all sectorsmust focusmore on reaching rural-dwellers,because the poorestmembers of Basothosociety live in the country-side;often in scattered,remote communities.

6.2 Strengtheninginstitutional capacityis crucial to poverty reduction. The resourcesnow devoted to poverty reduction would be far more effective if there were less program fragmentation and duplication and better coordination of public and NGO resources. NGO donors andcentral and local governmentshave sometimes succeeded in cooperatingon poverty- reduction programs, but usually in programs giving direct income transfers to households. Coordinationof broaderdevelopment programs has beenmuch morelimited.

6.3 Strengthening Government's capacity to design and implement poverty reduction strategieswould involve:

* Reaffirmingthe role of the Planningwing of the Ministry of Financeand Economic Planning (MFEP) as the focal point for setting poverty policy and coordinating programs. This means(1) reintegratingpolicy functions for multisectoral poverty initiatives (such as nutrition and household food security) within Planning; (2) initiating regular, substantive coordination between central Planning staff and Planning staff outpostedto line ministries; and (3) creating a small team of high- level centralPlanning staff to work exclusivelyon povertyissues. * As part of civil service reform, consolidating program managementunits with overlappingmandates, to reducecurrent fragmentation. * Developinga clear action plan for strengtheninglocal governmentso there is better coordination with community leadersin the implementationof poverty-oriented developmentprojects. * Reorientingstatistical data collection toward periodic, demand-driven,multisubject household surveys that would allow cross-sectoraland time series analysis of poverty issues. - 128 -

B. THE ROLE OF GOVERNMENT

Setting and Coordinating Policy

6.4 Setting poverty policy and coordinating NGO donor efforts in support of national policy is (in theory at least) the responsibility of MFEP's Planning section. It involves both the trained economists in MFEP and the planning units in each government ministry. But officials inside the ministry and out admit that MFEP has been largely unable to fulfill its central role in poverty alleviation. The most recent Five-Year Development Plan shows that MFEP is unable to articulate a clear strategy for alleviating poverty. The plan says its main objectives include poverty alleviation and the promotion and generation of productive employment opportunities, but it fails to articulate who the poor are, to provide a coherent strategy for addressing their needs, or to show how the programs it advocates will specifically improve their well-being.9"

6.5 Three factors contribute to MFEP's difficulties coordinating Lesotho's many, but fragmented, poverty reduction efforts.

* First, MFEP has an acute shortage of technical and professional staff. More than 30 percent of authorized positions in the Planning wing were vacant in 1993/94 and the vacancy rate was even higher for planning positions in line ministries. Available staff members are overextended and unable to spend enough time on the technical analyses needed to design and monitor effective poverty reduction policies and programs. Moreover, althougih planning staff posted to line ministries are theoretically responsible to MFEP, they interact only sporadically with central planning staff, whichi makes coordinating policy among ministries extremely di fficuIt.

* Second, Planning's ability to carry out its mandate has been diluted. Over time, various pieces of the Government's approach to poverty issues have been given to other ministries, especially the Prime Minister's Office (PMO) or its equivalent unider the previous military regime. For example, policy setting and program implementation for emergency drought relief were coordinated by the Prime Minister's Office, which also had initiated household food security policy, a critical (multisectoral) component of poverty relief (later taken over by the Ministry of Agriculture).

* Third, as in othier parts of tile civil service, many MFEP staff have little motivation, partly because of unclear job descriptions, work goals, and priorities; overcentralized decision making; limited opportunities for upgrading technical skills; little recognition of merit; and real or perceived inequities in job grading. These issues are beginning to be addressed as part of civil service reform now being undertaken wvithassistance from UNDP and the British Government.

97The Kingdom of Lesotho.Fifth Five-YearDevelopment Plan, 1991/92- 1995/96.Ministr) of Planning, Economicand Manpower Development. Maseru, December 1992. - 129-

6.6 It will take some time for civil service reform to show results, but the setting and coordination of poverty policy could be improved if there were high-level commitment to doing so. A review of plausible alternatives revealed MFEP to be the most logical place to plan for poverty alleviation and to coordinate all government efforts, if action is taken to enable it to better fulfill its mandate. The Government should consider integrating within MFEP all responsibility for setting policy on multisectoral poverty reduction issues (such as nutrition and household food security). This implies reorganizing existing units -- such as the Food and Nutrition Coordination Office (FNCO) -- to consolidate policy setting and coordination within Planning. In addition, to ensure adequate coordination among sectors, regular meetings should be held between high-level central planning officials (such as the Principle Secretary or Deputy Principle Secretary) and outposted principle and senior planners, to review line ministry priorities and ensure their consistency with each other and with MFEP-determined poverty reduction goals. Systematic consultation with outposted planning staff experienced in specific sectoral issues should give Planning input vital for designing feasible poverty-reduction strategies.

6.7 Some reorganization will also be needed within Planning. One option is to establish a small poverty analysis and coordination team or unit, reporting directly to the Principle Secretary of Planning, with responsibility for: (1) formulating the Government's poverty alleviation strategy in consultation with planning sectoral staff and planning staff posted to line ministries; (2) ensuring that the strategy gets incorporated in the five-year plans and annual capital and recurrent budgets; (c) coordinating donors' and NGOs' roles in poverty alleviation programs; (3) planning for the continuation of the most important safety net and poverty reduction activities when donor funding diminishes; and (4) continually monitoring the impact of poverty initiatives. This team should be headed by a principle economic planner and should include three or four senior economic planners taken either from the current central planning staff (of 88) or from senior planning positions in the line ministries (staffs for which will be reduced when consolidation of government shrinks the number of ministries from 17 to 12).

Streamlining Priority Programs

6.8 One important task is to eliminate duplication and the overlapping of priority programs and to coordinate related efforts. Many agencies have overlapping responsibility for emergency relief activities, for example. Rural health and hygiene education initiatives need coordinating, as do the institutions and efforts associated with microenterprise development and labor- intensive public work schemes. The responsibilities for child nutrition of the FNCO and the Ministry of Health are blurred, and at least three institutions have overlapping mandates for water and soil conservation.

6.9 There is an urgent need to audit institutions working on essential poverty reduction programs. The audit should determine where in government specific functions should be located, shiould suggest how to reorganize and merge units with overlapping responsibilities into larger units with enough staff and resources to discharge their duties, and should see that the consolidated units have clear, monitorable workplans, reviewed by senior officials at least once a year. Government is already generally committed to these goals of civil service reform, but has been slow in implementing structural changes. It also charged the Management Services Unit (MSU) of the PMO (the Prime Minister is also the Minister for Public Service) with looking for ways to consolidate responsibilities and streamline programs. UNDP is supporting this - 130- program98 whicih has three components: civil service reform, improving economic and social planning, and support for the personiel management and information system. The result should be a comprehensive review of all government structures. Poverty reduction should be emphasized as a national priority.

Piloting an Integrated Sector Approach

6.10 The operational capacity of a typical line ministry in Lesotho is often greatly undermined by a proliferation of projects within the ministry. Government officials routinely complain about the inordinate amount of time they spend on organizing and receiving donor missions, and complying to specific reporting, procurement, disbursement, accounting and auditing requirements from the various donors. Most projects also set up separate management and accounting uniits,whichi again contributes to the dilution of management capacity within the ministry. A more efficient method would consist of Government and donors jointly designinig, appraising and supervising an integrated sector program of assistance and investments. Donors would finance "time slices" of such a program -- whichiwould effectively be a national sectoral program. Whereas the preparation of such a program would be quite resource-intensive, its subsequent operation would typically require less management resources than under the old methiods. The Government may want to consider piloting this approach in one or more sectors.

Decentraiizing

6.11 Since the mid-1980s, the Government has stepped up efforts to decentralize development and involve communities more actively in the management of local resources. This is important because much1 of Lesotho's population lives in mountainous, often hard-to-reacih, villages; commilunitiesare williig and able to be more active in mobilizing and managing resources for developmenit; and the planning and administration of government has been linked, through a network of 10 District Development Councils (DDCs), to 24 Ward Development Councils (WDCs), and to an estimated 2,500 to 3,000 Village Development Councils (VDCs).99

6.12 Decentralized rural development efforts were placed under the District Councils to promote intra-ministerial coordination at the district and local levels. It is the DDC's responsibility to coordinate the extension activities of various line ministries in a given district, to reduce duplication of effort. DDCs are to monitor extension workers in healthi, nutrition, and agriculture, and are charged with getting extension workers to cooperate with rural communities (througlh VDC representatives). The VDCs are the liaisons between national development policy and grassroots initiatives.

6.13 This structure, established in 1986 as a result of Development Councils Order No. 9, received added clout in 1991 through Development Councils Order No. 18, which formalized the functionis of VDCs and other councils. Seven elected individuals from each village serve on a VDC withi the village chief or official headinan serving as the ex-officio member and chairman

9qThroughi its program, "Improvements of the Economic and Social Planning Process and Civil Service Improvement and Reform," starting early in 1994.

99Estimates vary widely on the number of VDCs. There are roughly 2,500 to 3,00 villages in Lesotho, and eachcould form a VDC, with a chief or headman as ex-offlicio chairperson. Currently. there are about 1.200 chiefs in the country. - 131 -

(Box 6.1). In a significant departure from tradition, the chief can no longer veto decisions the elected officials have agreed upon. VDCs have been empowered, legislatively, to mobilize village resources for local development projects. The Ministry of Home Affairs is charged with enforcing District Council orders, a formalization (though not effective) of institutional links between the center and the grassroots.

6.14 People in government and university circles are aware of the Development Councils Orders but realize that decentralization has been more symbolic than real. For one thing, traditional chiefs, as might be expected, have been reluctant to yield power to the District Councils. As ex-officio members, chiefs are officially constrained in their authority over District Councils, but reputedly have enough sway over elected officials that "agendas for action" all too frequently often become "business as usual." Another problem is that the tenure period for elected officials is only three years. In that short time, they must acclimatize themselves, begin planning and executing projects, get training (if it is available), and prepare for re-election. And having a clear understanding of the powers and functions of VDCs seems to be the exception rather than the rule. Elected officials seldom have the training needed to carry out even the minimal functions of accounting, monitoring, and evaluation. Because effective local government is so important for poverty reduction initiatives, the Govemment should develop a comprehensive plan for strengthening VDCs and clearly defining their role in development. This is especially urgent because recent disagreements between Government and some key donors (including USAID and the World Bank) about grazing range and other resource management issues have led to the suspension of support for VDC training.

6.15 Government needs not only to strengthen village governments, but to make sure that enough qualified line ministry staff are available at the district level to support and monitor poverty-reducing development programs. The health sector already has a strategy to ensure the presence of trained medical and public health staff at the district level. The Ministry of Education is trying to strengthen its field inspectorate by deploying more senior district inspectors and by placing teaching service officers in district capitals. The efforts of the Ministry of Agriculture and other agencies to strengthen their technical presence at the district level should receive strong support, both in suitably graded professional positions and in other resources (such as transportation) that would enable more regular site visits from district headquarters. - 132-

Box 6.1: The Functions of Development Councils

Village Development Councils (VDCs) The chief or headman of the village serves as ex-officio member and chairperson along with seven members who are elected by villagers. The functions of the VDC are to:

Consult with the Government and implement and maintain local development activities and social services, Represent and guide the local community in its efforts to identify village development needs, Raise funds for local development, Stimulate local participation in development activities, and Inform the Government (through the District Development Council) of local development priorities.

Ward Development Councils (WDCs) The ward or principle chief of twenty-four wards serves as ex-officio member and chairperson of the WDC along with twelve other members or, in the case of a district where there is only one ward, seventeen other members. Functions of the WDC are (among others) to:

* Collate development proposals from VDCs for scrutiny anidsubmission or reconsideration, and * Monitor the implementation of village development projects.

District Development Councils (DDCs) The ward or principle chief of the district serves as ex-officio member and chairman along with fifteen members elected at a meeting of the various ward DCs in that district. The functions of the DDC are (among other things) to:

* Promote socioeconomic development at the district level, * Formulate and implement district development projects, * Ensure that the district projects comply with the national development plan, * Monitor the implementationof development projects, * Raise funds for implementing development projects, * Consult, through the DDC's secretary, with appropriate government ministries on matters relating to * Development and planning, and * Coordinate the activities of WDCs. - 133 -

Monitoring Poverty

6.16 Finally, the Government must pay much more attention to building an integrated statistical database and systems for regularly monitoring changes in living standards in Lesotho, analyzing the nature and extent of poverty over time, and evaluating the effectiveness of poverty reduction policies (Box 6.2). The Government should consider periodic multisubject integrated surveys that would provide the data to analyze the relationship between various social and economic factors, living standards, and poverty. Integrated surveys could be designed to address core policy issues by providing realistic data on poverty levels (household income and spending), household demographics, women's role in the economy, factors influencing school enrollment, levels of malnutrition, access to basic services, labor force participation and levels of employment, access to credit, general savings behavior, and the effectiveness of safety net programs. Special modules could be added to elicit data on literacy, use of health services, and similar topics helpful for policy and planning.

6.1 7 Lesotho's National Bureau of Statistics (BOS) showed substantial capacity for houselhold survey work in the 1986/87 Household Budget Survey, which is scheduled for re- execution in 1996. It also conducted several single-topic surveys, such as the 1985/86 Labor Force Survey and the 1988/89 Agricultural Survey, full-fledged censuses, and selected demographic surveys.'°° These single-subject surveys have been useful, but they do not allow the cross-sectoral and time series analysis essential for poverty monitoring. And because end users of the surveys were not consulted, the surveys have had limited influence on policies and program design. Some line ministries have conducted surveys independently of BOS, sometimes with relatively useful results (such as the 1992 Nutrition Survey). Others were less successful, including the Ministry of Health's demographic and health survey and the National Employment Service's employment survey, because of the ministries' relative inexperience with survey work. Government and donor agencies have sometimes turned to private consultants (such as Sechaba) for survey work. Because these surveys are more timely and demand-driven, they are more useful for policy design, but continuing to depend on the use of private consultants would deprive Lesotho of a single comprehensive ad hoc database.

6.18 A strong case can be made for a major review by BOS, in conjunction with central planning and certain other data users, of BOS's program of data collection. To make the analysis of poverty more efficient, there should be a shift from single-topic surveys every ten years to more frequent, integrated, multisubject surveys. BOS should review its approach to data collection right away, because planning is already under way for the 1996 budget survey, the next major household level survey BOS has planned. If BOS decides to move to a multisubject format, the budget survey should be modified to cover such topics as well as standard household demographics, income, expenditures, and employment. In the medium term, BOS should also consider doing qualitative surveys using anthropological and sociological techniques to gain a deeper understanding of issues not accessible through the standard questionnaire used in quantitatively oriented surveys. A well-designed qualitative survey could help find out what ordinary people think about development programs and might be useful in fine-tuning programs to help the poor.

'°0 External assistance (both financial and technical) was used in the design, conduction, and analysis of these surveys (and may be required again future survey work), but local profession3l and field staff were fully involved and gained valuable experience. - 134-

Box 6.2: Indonesia's Use of Household Survey Data to Design Anti-Poverty Policies

Policymakers in the Indonesian national planning agency, BAPPENAS, recognized that anti-poverty policies required better, more up-to-date knowledge about the poor and about public spending on poverty reduction programs. Understanding that the usefulness of poverty analysis for policy depends on the quality and timeliness of the underlying data, BAPPENAS dramatically improved its ability to collect and analyze household survey poverty data regularly.

For instance, better knowledge of who benefits from public spending programs was needed to assess the possibility of redirecting public subsidies to the poor. Analyses of the poverty data showed that much could be done to improve the distribution of subsidies by reallocating spending to more efficiently targeted programs. More public resources were spent on poorly targeted programs -- the kerosene subsidy, hospitals, and post-primary education -- than on three programs found to especially benefit the poor -- health centers and subcenters and primary education. These findings pointed to the need to rethink public spending policies. The inequities implicit in public spending on hospitals and post-primary education had not been apparent, and the pro-poor impact of health centers and subcenters had been overlooked.

The data also showed that, even with subsidies, the poor faced significant costs in using public services. The analysis pointed to four options to help shift public spending toward the poor:

* Strengthen the system of individual means testing -- through letters of exemption by local authorities, for example; * Use geographic price discrimination to keep prices of social services lower in poorer areas -- reducing fees in the poorest 30 percent of subdistricts, for example -- could work; * Encourage self-selection by charging different prices for different levels of care; * Charge higlherprices to groups of the better off, such as civil servants enrolled in the government healtlh insurance scheme;

The large improvements in the content and regularity of the national houselhold survey, together with the demonstration effect of analytical insights generated by the information base, have laid a strong foundation for better poverty analysis and policy formulation in the future.

Source:The WorldBank, Development Brief, No. 56, May 1995. - 135 -

C. NGOS' ROLE IN POVERTYREDUCTION

6.19 Nongovernmental organizations (NGOs) play a significant role in Lesotho's poverty reduction. Early in Lesotho's history, church groups were the main source of education and health care, especially in rural areas. After independence, the Government gradually took over responsibilities in these areas, but churches still provide significant assistance in education and health care. International NGOs responded to Lesotho's difficult circumstances in the 1960s by setting up offices to provide emergency relief. Lesotho's governments have generally been receptive to support for economic development programs in general and poverty reduction programs in particular. NGO assistance will be of even greater value in a major national poverty reduction initiative.

Current Activities

6.20 The number of NGOs operating in Lesotho has grown significantly. More than 120 were registered as of early 1993, and many unregistered NGOs operate informally in rural areas. Neither the Government nor the NGOs track the size of annual operations, but a reasonable estimate of annual national spending on programs is M75 million to MIOO million (not counting church spending on health and education).

6.21 The range of NGO activities in Lesotho is considerable (Table 6.1). NGOs provide assistance during emergencies and in other cases of urgent need. NGOs are involved in natural resource management, rural infrastructure, human resource management, agricultural development, gender-specific development, small enterprise sponsorship, and the provision of credit. They are working for policy change at the district and national levels, including support for programs that introduce principles of democracy.

Lesotho Council of NGOs

6.22 The decision to establish the Lesotho Council of Non-Governmental Organizations (LCN) came in November 1989, when a group of 50 NGOs held a workshop with representatives from Government and international donors to discuss the possibility of establishing an umbrella organization or national council for all NGOs operating in Lesotho. Funding for the workshop and for subsequent efforts to establish LCN were provided by various donors. For example, Canada Fund sponsored the 1989 workshop, USAID financed the preparation and circulation of the LCN constitution, and UNDP sponsored the initial programs and establishment of the LCN Secretariat through its partnership development program. Moreover, the core funds for LCN's activities for the first three years of operation (1990-93), were provided by two Canadian NGOs: the Partnership Africa Canada and the Unitarian Service Committee of Canada. The establishment of LCN is a major positive step toward better coordination between NGOs, Government, and donors in the implementation of development programs, as there is now a potentially efficient vehicle for soliciting the support of NGO institutional resources and grassroots experience for national programs. - 136-

TABLE 6.1: NGO Activities in Lesotho Type of Activities No. of Comments NGOs

Agriculture Agriculture-relatedactivities include irrigation, credit, incomegeneration & coop - farming& livestock,gardening storage 6 formation.Grant fund, and some membership fees. Mostmedium sized, but some - rural enterprise,extension services marketing 4 NGOsare small (7 members).

Business& commercialdevelopment marketing 6 Activities includetrade, exports, tourism and finance. Also trainingand counseling.Some charge fees. Most aremedium sized but some have as many as 1,500members. Environmentand natural resources 4 Most with communitybased development angle. Also includedare education and research.Reforestation is a majorcomponent. Membership reaches 2,000. Outside fundingfor feworganizations HumanResources - nutrition & healthcare (include PHC, AIDS) 6 Sometarget specific groups (children), or issues(e.g., AIDS) and include - education,training and counseling 20 scientific/technicalassistance (also from abroad). Researchand grants to students.A few religiousorganizations are involved. Some incomegeneration activities. Mediumsized (up to 100). Shortterm relief - emergency,food aid, clothingdistribution 6 Organizationsthat work with specialgroups (aged, youth and children) give out emergencyrelief suppliesin somecases Specialgroups (long-term) * aged:self-help, training, aid andcare 5 Mostmedium-sized organizations (up to 300members). - youth: incomegeneration, education 4 Includecultural and religious aspects. 300 to 15,000members. - women:training and incomegeneration, health I I Sizevaries from 20 to 10,000.Some charge fees, and focus on specialgroups (e.g., - childrenand family wvelfare,day-care, school feeding 13 disabled),and issues(breast-feeding). - disabled,handicapped 8 Focusis on abuse,neglect, children's rights andorphans. Health, the disabled, and family planningare also incorporated.Membership reaches 1,000. Prevention& rehabilitationof alcohol& drugabuse 2 Mostare small-sized and activities include income generation, social clubs, and promotionof rights. Someminimal feescharged. Jobcreation, income generation, poverty alleviation 20 Participationby religiousorganization, community health care department. Membersreach up to 5,000. Sntallscale credit program& unions,revolving funds 4 Somefocus on specificgroups (e.g., retrenched miners, weavers). Financing from feesand grants. Often includedare training and marketing of produces.Sizes rangefrom smallto medium. Activities includecredit union promotion Religious/Cultural 21 Mostof thiereligious organizations are involved in communitydevelopments activities(nutrition, aid, waterprojects), with someaffiliation to organizations abroad.Memberships vary from 15 to 50,000. Minimal feesare charged by some. Communit)/participatory developments 14 Culturalactivities also include sports, and a few targetspecific groups (e.g., youths). Somewith grants. Water,Lesotho Highland Water Project (LHWP) 6 Thesecomprehensive programs include infrastructure (including road), environmentalissues, income generation and most of the activitieslisted in the table. Small feesare sometimes clarged, and some have grants and funding. Others Membershipreach up to 32.000. - legalaid andeducation 2 Includesground water research; some specifically focus on LHWP. - humanrights, political 4 - mediaand information 1 Somefocus on specialgroups (women, needy). - research:funding and teaching 4 Promotionof issues,membership reaches 2.000. - unions 12 Disseminateinfonmation and news. Someoverseas affiliation: few includespecial issues (e.g., education.) - profcssionalgroups and associations, alumni 27 Bothemployers and employees of variousdisciplines. Most chargeminimal fees. Membershipranges from 75 to 35,000. Professionsvary from taxi driversand mcdicine men to architects,alumni, and "%omengroups. Fewwith incomegenerating activities. Somefee charged. ______o ___a ______Nu_____i____ber______o______NG_____s__ _69__ SlnMembershipsize vary from 20 to over 5.000. Total Numberof NGOs 169 Note: Some NGOs participate in more than one actis ,ty - 137 -

6.23 Objective. LCN's main objective in the early 1990s has been to organize the NGO community and to improve government-NGO coordination. It has addressed its mandate in two ways. It has organized and carried out a management and technical needs assessment study of NGOs, and it has organized a number of sectoral and technical commissions comprising government ministries and NGOs with related programs. The assessment of management and technical needs revealed a wide range of needs, especially among the linked and local NGOs, for management training and the development of budgeting and financial accounting systems. The study recommended a multiyear program of skills training and technical assistance, to be coordinated by LCN but provided by external sources, because LCN is still developing its own management and organizational structure. LCN is currently looking for external funding and technical assistance to implement the proposed program. This effort should receive support from donors and Government, because of the important role strengthened NGOs could play in implementing poverty reduction programs in partnership with other agencies.

6.24 Sectoral conmntissions. LCN's second initiative has been to establish six sectoral commissions whose purpose is to provide a forum for the staff of sectoral ministries and agencies to meet with the staff of relevant NGOs. The idea is that the commissions would permit a consensus on general approaches in the sectors, would provide a clearing-house for information on1who was doing what, and would ultimately reduce duplication and gaps in coverage.X° The commissioni concept is good in theory, but in practice it will require a good deal of organization, persistence, and follow-up by LCN to make the commissions meaningful. LCN will not, at least initially, be able to rely on Government to take the initiative to participate regularly in commission meetings given the staffing problems in the higher grades of the civil service. Establishing a poverty planning team or unit within central Planning should over time facilitate systematic contact between the Government and the commissions.

6.25 Futture role. LCN needs to develop a strategic plan over the medium term to guide the development of the NGO sector. The plan should spell out the role of the NGO sector in Lesotho's development efforts (especially efforts at poverty alleviation), the functions LCN should assume in coordinating and directing NGO efforts, and the development effort LCN must undertake (including its financial and staff training requirements), to carry out its mandate.

6.26 LCN's decision to join DRIG as an active player in the current emergency drought relief effort precipitated a major turn of events for LCN. In departing from its role as a coordinator and spokesperson for all NGOs, LCN accepted responsibility for managing the practical operations of the relief effort in seven of Lesotho's 10 districts (the other three are run by the Lesotho Red Cross). LCN, whose annual budget amounts to M250,000, received M2.5 million from DRIG to register vulnerable households and run the food distribution effort. LCN added 200 staff to its payroll for this effort, in addition to its regular complement of ten. The drought relief effort has dominated the thinking and time of LCN staff to the detriment of its traditional role as NGO coordinator. In setting up a system to deliver food to vulnerable households, LCN established 190 distribution points, which it now wishes to use for developmental purposes -- to develop forestry, orchards, and other income-earning projects.

101The commissions are: Disaster and Humanitarian Relief, Democracy and Human Rights; Agriculture, Environment and Natural Resources; Human Resources Development; Economic, Social and Political Concerns; and Business and Commercial Concerns. - 138-

6.27 The risk LCN now faces is that its original much-needed mandate -- to coordinate NGO activities and link up NGO activities with those of Government and donors -- will be submerged under an aggressive development program. In a sense, LCN will be competing with other NGOs for Government's business. Yet LCN is obviously vulnerable to external support until its membership -- primarily through dues -- can support LCN operations. In 1992, LCN received only M10,000 from dues towards its M250,000 budget. LCN must improve its local resource base to maintain its independence, as it cannot expect the Canadian NGO community to provide support indefinitely. But, if LCN becomes a super-NGO, it will lose its credibility as the umbrella group leader and spokesperson for all NGOs. If the decision to operate development programs out of some or all of the 190 distribution points currently managed by LCN for the drought relief program proves to be sound -- and this issue needs to be addressed as part of a comprehensive review of current safety net programs -- LCN should consider bringing in other local NGOs to manage these operations and should redirect its energies and focus to its role as coordinator for Lesotho's NGOs.

D. DONORCOORDINATION

6.28 External donors contribute significant resources to Lesotho's poverty alleviation efforts. Most formally targeted projects -- including the emergency relief efforts, the food-for-work schemes and the labor-intensive construction projects -- have been largely donor-funded. And specific sectoral projects with a strong poverty focus, such as the nutrition and household food security program and the village water supply schemes, are financed largely with donor funds. A detailed review of all donor programs is clearly outside the mandate of this report, but because donors are likely to continue helping to finance many of the programs in a national poverty initiative, we briefly discuss here mechanisms that could help the Government, NGOs and donors cooperate.

6.29 By and large, the Government, NGOs and donors in Lesotho, have cooperated rather effectively, especially in direct assistance programs such as the school feeding program and the recent drought emergency. In both these programs, food supplies were financed and imported by the World Food Programme (WFP), were transported to depots around the country by the Government (the Food Management Unit), and were distributed to beneficiaries by NGOs working in cooperation with village committees. The experience gained in these and other successful cooperative endeavors should be replicated as much as possible in other programs, including broader development programs.

6.30 In the past two years or so, donor coordination has improved substantially, with UNDP coordinating monthly donor meetings to discuss development policies and programs. This improved donor coordination appears to be minimizing duplication and serious divergence among programs. As a next step, UNDP should consider helping MFEP organize the coordination structure needed between Government, donors, and NGOs, perhaps beginning with programs aimed at poverty alleviation. This might be accomplished by expanding every third donor meeting to include MFEP (and/or other government representatives, as appropriate) as well as LCN or other NGOs to discuss issues of mutual interest. This would enable NGOs -- who are now more or less in the dark about trends in donor assistance -- to gear up for roles they might fill in development programs. Serious consideration should also be given to organizing - 139- the quarterly joint discussions around LCN's proposed sectoral commissions (or some variation on that theme).

6.31 Finally, should Lesotho's government decide to launch a major new initiative on poverty, based on the analysis contained in this and other reports, consideration should be given to organizing a major donor conference to lay out goals and strategies. Such a conference could help ensure that donor programs explicitly reflect government priorities in poverty reduction and might help secure needed funding for development programs. If feasible, the conference, perhaps co-sponsored by UNDP, should be held in Lesotho to facilitate attendance by NGO representatives and district and local government officials. Sufficient advance notice of the conference should be provided that donors without a continuous presence in Lesotho have time to arrange to send to send representatives. - 140-

7. AN ACTION PLAN

A. INTRODUCTION

7.1 Drawing together the findings and recommendations from the report, this last chapter presents a strategy of policy reforms, public expenditure re-allocation, and institutional development actions to reduce and monitor poverty in Lesotho. It also suggests a sequence of priorities, taking into account each action's potential impact on poverty, ease of implementation, and affordability.

7.2 Although a number of concrete suggestions came from Basotho sources -- NGOs, academicians, civil servants, entrepreneurs, and villagers -- the policy suggestions presented in this final chapter are not meant to be prescriptive. Rather, the options presented should be debated by appropriate democratic institutions, both local and national. At this point in Lesotho's history, the process of formulating policy is as important as the outcome, if not more so. And that process should be both transparent and inclusive.

7.3 The policy recommendations are discussed in groups, representing broad recommendations to: foster labor-intensive growth (in agriculture, small businesses, export- oriented light manufacturing, tourism, and infrastructure investments), invest in human resources (health care and education), strengthen the social safety net, and improve institutional capacity (Box 7.1).

7.4 The various policy recommendations discussed below will need to be prioritized. A first attempt is made to differentiate between policy actions which are likely to have a high or medium impact on poverty reduction, and between policy actions which are feasible to implement in the short or medium term (Table 7.1). The poverty impact of a particular policy action derives from the poverty profile. The feasibility derives from a rough assessment of implementation constraints and incremental fiscal costs.

Table 7.1. Prioritizing the Policy Actions Povertyreduction Feasibility Short term Mediumterm

High impact Policy I Policy 2 Medium Policy 2 Policy3 impact

7.5 The resulting proritized summary matrix of policy recommendations made in this report is reprinted below (Box 7.1). It is suggested that this matrix can form an input into a more wider and inclusive process of public debate aimed at designing a concrete set of programs to reduce poverty in Lesotho. - 141 -

Box 7.1: A Prioritized Action Plan FOSTER LABOR-INTENSIVE GROWTH

In Agriculture: * Remove price, marketing, and processing distortions. * Invest in rural infrastructure that supports intensified agriculture and the piroductionof export crops. * Make land tenure (and women 's rights to land) more secure and rental markets more flexible. * Promote a reduction in herd size. * Reorient agricultural research and extension.

In the Small Business Sector: * Make it easier for small entrepreneurs (especially women) to get credit and do business. * Shift the emphasis awayfrom broad entrepreneurship training in manufacturing toward training in skills needed in the export and service sectors. * Develop rural and urban marketplaces.

In Export-Oriented Light Manufacturing: * Keel) minimum wages at a competitive level. * Imtproveindustrial relations. * Reduce utility costs by increasing competition, regulation, and decentralization in the utility sector. * Root export firms more firmly in the economy by selling them factory shells and leases and by supporting skills training that promotes the use of local skilled workers in the export industry.

In Tourism: * Work much more closely with South African travel agents. * Develop a special "Lesotho package."

Through Infrastructure Investments: * Use private contractors on public works. * Invest more using labor-intensive approaches, but decentralize implementation. - 142 -

INVESTINHUMANRESOURCES

In Health Care: * Exempt children under five, the elderly, and the disabled from user fees. * Introduce a free or low-fee package of basic health services. * Increase geographic equity. * Allow local facilities to retain certain fees so people feel they have local control over services.

In Education: * Lower the cost to parents of primary schooling. * Strengthen nonformal education for out-of-school youth. * Improve vocational and technical education. * Invest more in education in poorer regions. * Create incentives for qualified teachers to teach in remote areas. * Spend proportionately more on primary education and proportionately less on college and university education.

STRENGTHEN THE SAFETY NET

* Lower the consumer price of maize meal, sorghum meal, sorghum, and cooking oil. -Remove trade, marketing. and processing restrictions. -Excmpt basic foods fiom the genieralsales tax. * Expand labor-intensivepublic uorks. -Decentralize planning and management to local communities for cost-reduction effectiveness. -Pay in cash (at below market wages, to make projects self-targeting), rather than in kind. * Target special progranis (food-stamnp-typeprograms, for example) to vulnerable groups, including an improved drought relief program.

IMPROVE INSTITUTIONAL CAPACITY

* Design and implement a poverty reduction program of investments, financedlfrom the Development Fund of the Lesotho Highlands Water Project. * Decentralize political, fiscal, and administrative authority. * Refonn the civil service. * Privatize activities better performed by the private sector. * Improve planning and budgeting. * Merge fragmented program management units. * Start doing annual household surveys. - 143 -

B. FOSTER LABOR-INTENSIVE GROWTH

7.6 In both agriculture and industry, Lesotho's policies have tended to promote capital- intensive import substitution. The strategy has been one of explicit and implicit capital cost subsidies, protectionist measures, and state involvement in the creation of parastatals and state companies. This strategy has not created jobs or reduced poverty. Attempts to create mechanized, large-scale maize, wheat, and vegetable farms using rainfed and irrigated agriculture largely failed, despite considerable support. Attempts to create an industrial and agroindustrial base producing for the local market have more often than not produced inefficient, rent-seeking, or subsidy-dependent parastatals. These parastatals produced precious few jobs yet raised consumer prices and government costs because inefficiencies, monopoly profits, and protection from international markets went unchallenged.

7.7 Expansion of a labor-intensive private sector cannot be based on an inward-looking strategy. Lesotho's domestic market is simply too small for import substitution, a strategy that has failed so far. More important, in a small, open economy that is not isolated by unduly high transportation costs from its major trade partner, any evidence that import substitution works implies an opportunity to increase exports and to expand quickly.

7.8 The wisdom of a labor-intensive, export-oriented strategy has already been confirmed in the case of Lesotho. Since the late 1980s, foreign direct investment in the textile and garment industry has created a small base of labor-intensive, export-oriented firms, attracted by Lesotho's low wages and preferential trade agreements with Europe, North America, and South Africa. In the last five years this export boom created 5,500 jobs. (By contrast, after nearly three decades of governmenit investment and subsidies, Lesotho's 51 parastatals currently support a total of 7,700 jobs). The textile and garment industry can still expand and dramatically increase employment. But this export boomnlethas had no "ripple effect" so far, mainly for lack of a critical mass of Basothiomachiniists, technicians, and supervisors who could eventually venture out on their own and set up garnent businesses to subcontract or otherwise provide goods or services to the large export firms.

7.9 The main export market for small firms should be in South Africa. That market is likely to grow considerably in the medium term. In the past, excessive regulation often kept Basotho firms out of the South African market, but gradual deregulation in South Africa should open up new opportunities to small businesses in Lesotho. Because small firms rarely possess sufficient marketing skills to undertake export marketing themselves, the key to developing the export orientation of small firms lies in linking them to larger firms which do possess such skills.

7.10 The main impetus for labor-intensive growth in the small business sector, however, will come from an increased demand for services resulting from increased formal employment in exports. As the export sector grows, more workers will be hired and competitive pressures will cause wages to rise. A portion of these wages will be spent locally, especially on services. Setting up service businesses (such as construction, small restaurants, processed foods, and the repair of durable consumer goods) does not require the substantial capital outlays often needed for small-scale manufacturing. Many service businesses can start small and grow by reinvesting profits. - 144 -

7.11 Promoting export-oriented, labor-intensive production in agriculture and industry is a key macroeconomic strategy for reducing poverty in Lesotho. Growth in the number of private firms and increased employment will lead to a tightening of the labor market and the bidding up of wages, judging from experience in Southeast Asia and, closer to home, in Mauritius. Small service-oriented businesses will grow in size and number as exports expand.

Agriculture

7.12 Lesotho's drive for self-sufficiency in food crops -- especially maize -- has hurt the poor and does not exploit the country's comparative advantage. Lesotho's food prices, which are set administratively, have consistently been kept above South African prices, which are also administratively set and kept well above world prices. Both countries have introduced price and marketing distortions to help producers and to achieve national food self-sufficiency, but they have hurt their poor in the process. South Africa has maintained food self-sufficiency for thirty years, and in six out of seven years exports a surplus, but at a financial loss that is passed on to consumers. Lesotho's drive for food self-sufficiency has failed. Lesotho imports more than half of its food grain requirements and imports constitute an ever larger share of domestic consumption, even though maize occupies over 60 percent of the area cultivated.

7.13 Most Basotho households are net consumers of basic staples,'0 2 and most of the maize consumed in Lesotho is imported. Naturally, nearly all urban households are net consumers of maize, but even in the agricultural areas most suited to maize production -- the northern lowlands and foothills -- an estimated 90 percent of rural households are net consumers of maize. More than 90 percent of the western foothills suffer a structural food deficit, and in other regions (the western lowlands, the mountains, and the south) more than 95 percent of rural households are net consumers of maize.'03 Only I to 10 percent of the maize processed by Lesotho's three industrial mills is procured from local sources; everything else is purchased from the South African Maize Board. Because the poor are more likely to suffer food deficits and spend proportionately more of their income on food than the rich, the current marketing system is effectively a tax on the poor, one that largely benefits South African producers and large-scale mills. A poverty action plan that did not redress this inequity would miss an immediate, unique opportunity to reduce poverty. Initial estimates suggest that lifting the ban on flour imports would immediately reduce the price of maize flour by at least 30 percent.

7.14 Moreover, in the long run Basotho producers could increase their profits by shifting gradually to crops for which Lesotho holds the comparative advantage to produce at a lower cost than other countries. Certainly Lesotho has no comparative advantage in maize production, except for a limited area in the northern lowlands. Despite massive direct subsidies and price supports, the country has repeatedly failed to increase maize production enough to meet domestic demand. And given the high transportation costs for such a bulky crop, Basotho maize producers are unlikely ever to be able to compete with a deregulated South African market. Food prices are likely to fall in South Africa as a result of both deregulation -- which will substantially reduce marketing margins -- and the gradual shift from large-scale farming to a more diverse and

102 The averagehousehold spends about 40 percentof cash incomeon food; the poor spend proportionately more.

103Brendan Bailey, 1994. "AgriculturalMarketing in Lesothoand Price Transmissionto the VillageLevel: Implicationsand Optionsfor Policy." Photocopy.Drought Relief Implementation Group. - 145 - efficient agrarian structure, with small-scale farming promoted through land restitution and redistribution efforts.

7.15 Lesotho shares a unique regional agroclimate with the eastern Orange Free State and the northern mountain areas of the Eastern Cape. The Ministry of Agriculture has identified a wide variety of crops for which Lesotho might have a comparative advantage: beans, peas, sunflower, soybeans, potatoes and seed potatoes, asparagus, peaches, apples, cherries, dairy products, wool, and mohair. The areas bordering Lesotho suggest the agroclimate's potential. The area around Ficksburg in the eastern Orange Free State (50 km from Maseru) is the center of South Africa's cherry production for export. Areas in the Eastern Cape bordering Lesotho do very well producing wool and mohair.

7.16 Lesotho's agriculture strategy should be firmly targeted to shifting the country's agricultural product mix toward this comparative advantage. It should do so by (I) removing price, marketing, and processing distortions that favor crops for which Lesotho has no clear comparative advantage; (2) investing in rural infrastructure that supports agricultural intensification and a new mix of agricultural products; (3) making land tenure (and women's rights to land) more secure and making land rental markets more flexible; (4) intensifying efforts to promote smaller herds; and (5) reorienting agricultural research and extension toward the most promising crops and farming systems.

7.17 Actively pursuing this strategy will be a daunting challenge for Lesotho's agricultural institutions, especially the Ministry of Agriculture and Livestock Development. But successfully implementing it would solve many of the country's development problems in rural areas, where most poverty is concentrated. To alleviate capacity constraints, the Ministry should develop a comprehensive, consolidated national program that would rally and coordinate donor assistance around a single, sectorwide strategy. The current plethora of often fragmented donor projects dilutes Ministry efforts and unduly taxes its already strained capacity.

7.18 Rentove price, marketing, and processing distortions. Consumer food prices (and hence implicit producer prices) in Lesotho are higher than South African prices, which are higher than world prices. These artificially high prices are the cumulative effect of the pricing policies of South Africa's Maize Marketing Board; of Lesotho's ban on flour imports; and of the oligopoly in grain imports of three large Lesotho mills (one of them a parastatal). Other grain imports are regulated by monthly permits that may or may not be granted by the Ministry of Agriculture, depending on its subjective assessment of local availability. Such a system creates opportunities for corruption, needless bureaucratic hassles and uncertainty among traders.

7.19 To reduce food prices and shift the relative prices from food to export crops would require a phased deregulation of food grain and flour imports. This measure would greatly alleviate poverty by cutting consumer prices, helping restrain the upward pressure on wages - which in turn helps the labor-intensive export industry -- and contributing to the sustained, long- term profitability of agricultural production by inducing a shift in production away from maize toward crops more suitable for Lesotho.

7.20 As a first step, food grain imports should be deregulated and all permit requirements abolished. Consumer food prices in Lesotho would fall immediately fall if the large mills lost their oligopoly -- especially when South Africa's maize marketing margins decrease as a result of deregulation (beginning with the abolition of pan-territorial pricing and the replacement by May - 146-

1, 1995, of the Maize Board's import monopoly by a system). For Lesotho's agriculture to develop, both indigenous and foreign traders should be able to operate in a liberalized market free of unnecessary regulation and the uncertainty of ad hoc government intervention. The liberalization of grain imports should also stimulate local small-scale milling in areas close to South African maize production. Producer prices in Lesotho should generally decline, but will vary by region; price changes will depend on how much marketing margins are reduced in the shift from monopolistic large-scale to competitive small-scale milling.

7.21 As a second step, imports of maize and wheat flour should be deregulated. Many of the country's poorest regions structurally suffer a deficit in basic foods and are isolated by high transportation costs. Local milling of imported grains in these regions is unlikely to be cheaper than importing flour, either from Lesotho's lowlands or from South Africa. If the deregulation of grain imports is not followed quickly by the deregulation of flour imports, there is a risk of establishing a local milling capacity that would not be competitive in the long run and would gradually be undermined by smuggling. Instead, regions isolated by high transportation costs should specialize in food crops for which there is immediate local demand and in export products such as peas, seed potatoes, wool, and mohair. Ironically, the potential for exports has already been demonstrated by the substantial dagga production in some of the more mountainous areas.

7.22 Small-scale farm production, processing, and marketing of export crops can be highly profitable and is usually labor-intensive. International experience shows that developing even a single successful export crop can greatly alleviate rural poverty. However, doing so can be a complex undertaking and requires substantial management skills, especially in processing and marketing -- skills not currently abundant in Lesotho or even in South Africa, except in horticulture. Private sector development of processing and marketing capacity would best be promoted through foreign direct investment and the attraction of traders with intimate knowledge of destination markets.'04 South African firms and private traders already offer immediate opportunities for Lesotho's marketing and processing of horticultural products (such as asparagus and strawberries) for export to South African markets. Firms in Zimbabwe and Kenya, wilich have made substantial inroads into European markets, could also provide marketing opportunities for Basotho producers. The LNDC Investment Promotion Center would need to conduct an export promotion drive to attract such firms and private traders, and any successes could then be used to attract other foreign investors.

7.23 The Ministry of Planning, together with the Ministry of Employment, should critically evaluate proposals to create monopolistic agroprocessing firms in which local producers' cooperatives or the Government hold a majority share. It is in the producers' interests to get the best price possible, from whatever source. Tying producers to export and agroprocessing firms in which they hold a share produces an immediate conflict of interest, and, when the firims are inefficient, might well lead to "infant industry" arguments for trade protection and subsidies. As past experience shows, inefficiency would creep in and widen marketing margins, keep producer prices down, inhibit the rapid expansion of the industry, and hence greatly constrain its ability to alleviate poverty. However enthusiastically the idea has been received initially, the Government should refrain from promoting and protecting this type of enterprise; introducing such distortions will compromise the long-term development of the agricultural industry.

104Exotic Foods,a British company,is currentlyexpanding its activities in asparagusand other horticultural crops, for example. More such investmentby privatefirms should beactively promoted. - 147 -

7.24 Invest in rural infrastructure thlat supports intensified agriculture and the production of export crops. It is impossible to exploit a country's comparative advantage in crop production without essential infrastructure. Rural infrastructure can widen the range of crops grown and reduce production costs at the same time that it can lower marketing and transportation costs, make agricultural activities more profitable, and lower the price of urban imports. Such infrastructure may be related to water harvesting (for example, roof catchments, small dams, and dam linings), soil conservation, or small-scale irrigation (provided production is organized on a household basis), and may range from fencing for garden areas to the physical infrastructure needed to establish periodic village markets. Investments in labor-intensive agricultural infrastructure can both enhance the rural capital base and alleviate rural poverty. One fringe benefit of developing labor-intensive rural infrastructure is the creation of jobs, which should be targeted toward the poor.

7.25 Lesotho's irrigation potential may not be great, and it may be costly to realize in many areas, but even a few hundred hectares of irrigated area can increase national production dramatically. International experience shows that the most efficient model is small-scale irrigation in which individual farmers own and operate areas they can comfortably manage themselves, in which they themselves choose the crop mix, and in which they coordinate water supply and maintenance through farmers' committees. The Small Scale Irrigated Vegetable Production Scheme (with assistance from the Irish Government and the EU) in the north of Lesotho is one promising example. In vegetable production, half a hectare per farmer family is usually more than can be managed and operated by family labor. To maximize employment, the Government should resist allocating parcels larger than a farm family can manage itself. Allocating larger parcels will result either in underuse of the land or in increased rental arrangements. Rental arrangements are not necessarily inefficient, but they can lead to underinvestment and inequity when linked markets (credit, capital, foreign exchange, labor, inputs, and outputs) favor the rich.

7.26 Make land tenure (and women's riglhtsto land) more secure and rental markets more flexible. Lesotho's relatively equitable distribution of land is an important asset, both for agricultural growth and as a social safety net. Despite what powerful agricultural lobbies the world over assert, small-scale farming is not less efficient than large-scale farming. Decades of international agricultural research have failed to demonstrate economies of scale beyond what a family can manage comfortably on its own. And this research clearly demonstrates that small- scale farming is often more efficient -- as measured by profits per dollar invested -- than large- scale farming. The most productive agricultural sectors in the world are dominated by small- scale, part-time, family farming. So the small average size and the high degree of fragmention of farms in Lesotho should be seen as an asset, not as an impediment to future agricultural growth.

7.27 Moreover, secure land-use rights to even a little land give the poor an important social safety net, especially in the absence of a formal social security system. The problem of land lessness (currently estimated at 15 percent) seems to be on the rise in Lesotho, and the Government should prevent it from accelerating, because of its grave implications for rural poverty and urban slums. The decision to allow land sales and purchases -- especially to people outside the village but also to people within the village -- should be left to the communities themselves. (Even then, equity can be ensured only if community government is basically democratic.) Land markets are not necessarily efficiency-enhancing, if other major markets (for - 148 - example, credit markets) remain distorted. The premature introduction of land markets will merely allow the rich to accumulate land, because they often have better access to credit and subsidies than the poor. And the Government should resist any attempt to redistribute land rights away from the poor under the guise of "agricultural development". The adverse impact of large- scale farming on rural poverty can be seen right across the border, in South Africa.

7.28 To ensure a well-functioning rental market and agricultural efficiency, it is important to establish secure property (use or leasehold) rights and legally enforceable (but unregulated) rental contracts. The poverty profile discovered a strong demographic life-cycle component to poverty. Hence, at certain times, a household may not have the resources to use the land productively and may want to rent its land out to those who do have the resources to use it. If households feel secure about their property rights, they will see no problem in renting their land out, even on a long-term basis. If they do not feel secure about their property rights, they will not enter into rental contracts or will prematurely end rentals, fearing to weaken their claim either by renting out the land or by renting it out for too long. A rental contract will normally reflect the relative bargaining positions of the two parties that enter into the contract. Such bargaining positions generally differ from place to place and from time to time. Once the free bargaining process has led to a particular type of rental contract, it should be legally enforceable by both parties.

7.29 The Government should resist directly intervening in the definition of conditions in rental agreements. Governments often undertake to control rentals, to protect vulnerable parties from terms deemed socially unacceptable. But when governments intervene in rental arrangements to protect the weaker party (usually the tenant), what results more often than not is that the rental agreement does not take place at all, or that existing tenants are evicted. This is why it is often counter-productive to regulate rental contracts. Lesotho is unique in that the weaker party is usually the land holder. In Lesotho, the poor often want to sublease their land to the rich, so it is generally unnecessary to protect tenants' rights. It is essential, however, to protect the land rights of the poor, so they are guaranteed rental income when they do not have the resources to use the land productively themselves.

7.30 The Land Act of 1979 goes a long way toward protecting property rights and making contracts enforceable. It gives communities the option of allowing individuals to transform traditional use rights into 99-year leases that can be sold, subleased, or mortgaged. But the Land Act has not been thoroughly explained to communities, its implementation has been haphazard, and survey requirements are sometimes too lengthy and costly for the poor. The resulting uncertainty impedes owner-operators' agricultural and other investment in the land and keeps the rental market, which could increase productive use of the land, from functioning well. Legal uncertainty about property rights especially harms the poor, who do not have the means to establish de facto ownership through such extra-legal channels as personal contacts, bribes, and political clout.

7.31 The Land Act gives the Government the legal authority to declare certain areas "development areas" and allocate them to individuals, parastatals, or the Government. The Act declares the State the owner of all land, which could create problems if the principle is not carefully applied. For one thing, it may generate unnecessary apprehension and resistance to the Act among the chiefs, who traditionally have held land allocation rights. For another, if not used wisely and sparingly, it may expropriate common property without adequate compensation for communities, most of which are poor. It also opens the door to arbitrary allocations to - 149-

individuals well-connected to the state bureaucracy, which almost certainly does not include the poor.

7.32 The Land Act should be vigorously propagated and implemented. At a minimum, the current option to convert traditionally held use rights to leasehold rights should be widely publicized and the survey requirements should be relaxed and made less costly.

7.33 More than half of all households in Lesotho are headed by women. Although the Land Act expressly allows women to hold leases.'05 their continued status as minors under Lesotho's customary (and modern) law hinders their ability to acquire land use rights from their chiefs in their own names, which makes it difficult for them to apply for govemment leases. There is an urgent need to review the constraints women face in acquiring and keeping land rights and seek appropriate solutions. How will women's rights be interpreted in the Land Act and its amendments? That is still open for debate and women's rights to own landed property apparently remain vulnerable. Yet their right to acquire secure land rights is important both for poverty reduction and for agricultural efficiency.

7.34 Promote a reduction in herd sizes. Everyone knows overgrazing in Lesotho is a problem, but solving the problem is complicated because people hold cattle not only for production purposes but as a form of savings.

7.35 Introduce grazing fees. For those who hold cattle for productive purposes, we recommend pursuing the standard economic solution: Bring individual financial costs in line with economic costs by introducing grazing fees. The higher the fee, the more likely individuals will be to reduce their herd. But fees should be collected by local voluntary grazing associations, which should be allowed to keep most of the fees collected. If part of the fee is transferred to the Govermment, there should be a clear quid pro quo: The associations should be confident that the Government will use the fees to provide a service to the community (for example, fund research and extension). Locally, grazing fees could be used to improve range management and to compensate households that do not own cattle -- usually the poor -- for the use of the range, which is, after all, common property. Such compensation could take the form of general village improvement. The pilot experiment with Range Management Areas, initiated with USAID assistance, was successful enough that it should probably be tried on a national scale.

7.36 Introduce alternative savings instruments. Introducing grazing fees may not lead to a noticeable reduction of the herds held mainly for savings purposes, because people holding cattle as an asset may care much less about their financial rate of return than about their cattle's function as an illiquid asset in which to invest savings. A recent study'06 suggests that in Lesotho livestock may indeed be held mainly for illiquidity, not income, and savings behavior is generally price inelastic. (Banks are familiar with this phenomenon: Changes in interest rates do not cause large fluctuations in customers' savings accounts. Apparently, customers are more interested in the type of savings instrument a bank offers -- whether it is secure and illiquid, for example, and helps them maintain savings discipline.) The logical complement to grazing fees,

105 That is, to acquirethe rightsto landuse froma deceasedhusband.

106 The study confirmslow financialrates of returnfor holdingcattle (3.3 percent),sheep (7.6 percent),and goats(-6.1 percent),even if the opportunitycost of laboris calculatedas zero. P. Cole,"Gross Marginsfor AgriculturalEnterprises in the MatelileRural Development Project, 1993. - 150- then, might be to introduce a new savings instrument that could replace cattle as a savings asset. Unfortunately, there is so little competition in Lesotho's financial sector that banks are unlikely to be interested in experimenting with new types of rural savings schemes. In the short run, the Central Bank could use its influence with the commercial banks, especially the LADB (which has rural branches), to experiment with such alternative savings instruments. One promising option is long-term contractual savings instruments (certificate-of-deposit type instruments, for example, for pension purposes). In the long run, increased competition is needed to encourage more innovation in the financial sector.

7.37 Reorient agricultural research and extension. With a shift away from the past emphasis on food self-sufficiency, Lesotho's research and extension efforts should be redirected toward different crops and farming systems. Several NGOs (such as the Basotho Mineworkers' Labour Cooperative, Lesotho Home Makers, and the Seventh-Day Adventists) have already accumulated considerable knowledge about small-scale farm models. Agricultural research costs so much that Lesotho should try to latch on to new developments through South African research. South Africa's basic agricultural research ranks among the best in the world; Lesotho could cut its costs considerably by linking up with South African research institutions. South African researchers have virtually no experience with small-scale fanning models (although this is likely to change soon) and might welcome a joint venture with Lesotho research and extension institutions on, for instance, peri-urban and small-scale farming. Such cooperation could be a win-win proposition for both sides, with South Africa bringing scientific research capabilities and Lesotho bringing experience to the venture. On crops unique to Lesotho's agroclimate, in which South Africa's research establishment shows no interest, Lesotho would need to develop special research, probably using experts from private international firms. One very cost-efficient way to make such expertise available to Basotho farmers and processors is through foreign direct investment in agricultural export crop production and processing.

7.38 As countries elsewhere have learned, returns to agricultural extension can be dramatic, once appropriate extension methods and messages are identified. Agricultural extension in Lesotho, unlike its South African counterpart, has a long history of working with small-scale farmers. One potentially successful extension model has been initiated -- the Production through Conservation Approach, supported by SIDA -- and efforts should be made to improve, intensify and expand the program'°'. Lesotho has also produced a number of private farmer "champions" like Mr. 'Musi, who, supported by the German-funded Matelile Rural Development Project, successfully designed a soil conservation strategy, reclaiming gullies by planting shrubs and grass. Role models like Mr. 'Musi are important and should be actively used in Lesotho's agricultural extension program, because in farmers' eyes they are often more credible advocates of innovation than the extension agents are.

107 Butfarmers' investments in soilconservation are unlikely to befinancially profitable under the traditionalcrop mix of maize and sorghum (Convery, 1995). This constitutes yet another reason to shifttowards the moreprofitable export crops. - 151-

Small Businesses

7.39 A number of export-oriented firms have met with considerable success recently and efforts to promote their expansion should continue unabated. But Lesotho's private sector consists mainly of small businesses, which continue to be neglected and underserved. Developing the small business sector is essential to solving Lesotho's unemployment problem, especially among the poor. It is important to increase their (including women's) access to capital and to business advice, to provide suitable training, and to develop rural and urban market places where they can do business.

7.40 Make it easier for small entrepreneurs (especially women) to get credit and do business. Women in Lesotho can still not get credit without their husband's consent, even if they have been long-standing depositors, and even if their deposits have been sizable. These restrictions make no sense because the small business sector is dominated by women. All women-specific restrictions on business and credit should be identified and eliminated. Indeed, all legislation -- all customary and common law, including "personal" law -- should be systematically reviewed for provisions that erode women's legal status and prevent their equitable treatment under the law. The National Organization of Women Lawyers in Lesotho has already done extensive work in this regard.

7.41 There is substantial and growing demand among small businesses for capital and practical business advice, judging from the thriving informal credit market and recent private initiatives in small business lending and advice. The flurry of activity and high interest rates in the informal credit market reflects the real opportunity cost of capital. Lesotho's domestic commercial banks, unlike their South African counterparts, seem reluctant to expand into the small business sector, partly because they are so risk averse and face so little competition. Their traditional lending and savings operations yield comfortable profit margins, so they have no immediate need to venture into uncharted financial waters.

7.42 South Africa's more competitive financial market has produced more innovative financial intermediation. Currently, for example, South Africa's First National Bank provides busiless credit to a number of small Lesotho businesses through a private Lesotho association specializing in business advice and financial services. It is too soon to tell if this operation is financially sustainable, but Lesotho's financial sector needs more of the competitiveness that produced it. To encourage more innovative, project-based approaches to the financing of small businesses, and to business support for project preparation and implementation, the Central Bank should strongly promote competition in the financial sector by various means, such as attracting new entrants, helping banks design new types of financial intermediation, and withdrawing the Lesotho Bank's monopoly on Miners' Deferred Pay.

7.43 Sltift the emphasis away from broad entrepreneurship training in manufacturing toward training in skills needed in the export and service sectors. Lesotho is ahead of the region in general levels of schooling, but to stimulate economic growth it must improve labor skills in large firms and the business skills of small-scale entrepreneurs in the service sector. Local businesses can emerge from the pool of skilled, experienced machinists, technicians, and supervisors already working in large export firms. Continued expansion of the export industry will also increase demand for services as varied as catering, machine repair, and building construction and maintenance. Local sourcing of more inputs (such as packaging and transport) - 152- would also give local business new opportunities. And some production (for example, embroidering) could be given out as piece work to local subcontractors.

7.44 Some organizations offer business training programs, but they are largely geared toward small and medium-scale manufacturing for the local market, when training for the service sector might be more relevant. Recently, there have also been attempts to train small entrepreneurs in marketing, but experience in other countries shows that small manufacturers (and small farmers) rarely market their products themselves. Rather, they rely on traders and middlemen, who have reliable, up-to-date information on destination markets.

7.45 To begin improving, training institutions should consult more with businesses to identify the skills they need. By financing joint government-industry training programs, the Government can encourage export firms to train more local skilled labor on-site. Such training programs can be discussed and designed in "deliberation councils" -- patterned after East Asian models -- that provide a platform for better communications between the private sector, labor unions, and the Government. Training institutions should consult with current export industries to identify opportunities to create links between the export industry and local businesses, links that do not now exist. Then skills training can be targeted to current needs and opportunities. One institution that could develop a curriculum more appropriate to current needs and opportunities is the Institute of Extra-mural Studies. Perhaps, too, the Cooperative College could be transformed into a Business Training Center as a joint venture between the Government, the cooperatives, the private sector, and the NGO community. The recently created Human Resource Development Commission of the Lesotho Council of NGOs could be asked to develop proposals in this direction.

7.46 Some argue for targeting much of the training toward people already working in the export industry, rather than to people with no formal job experience. Experience in both Lesotho and elsewhere suggests that local entrepreneurs will emerge from the pool of skilled workers in formal industry (machinists, technicians, and supervisors), rather than from the pool of the unemployed and self-employed.

7.47 Some of the managerial training offered seems too sophisticated for emerging businesses.Small new businesses are often in urgent need of simple bookkeeping skills, for which no appropriate training is available. Perhaps Lesotho's burial societies and stokvels associations, which apparently have excellent bookkeeping systems, could serve as a model for locally developed bookkeeping systems, a model that could in time be upgraded to a more formal system.

7.48 Develop rural and urban marketplaces. An unfortunate legacy of the colonial, settler- dominated history of Southern Africa is the near-total absence of physical marketplaces where small businesses and producers can market their goods and services. Even a city like Maseru provides only for up-scale businesses and street vendors, and has only a rudimentary bazaar-type central market. In the rural areas, no system of periodic, rotating markets has developed, as it has in other parts of Africa. The need to provide a time and place where buyers and sellers can meet to trade may seem obvious but "the market" is an economic institution that must be purposefully created and maintained by economic agents; its emergence is not automatic and costless. Lesotho's government should actively help develop such market systems, even if they are slow to materialize at first. A private association (the Market Council) is currently pioneering the concept of rotating markets in South Africa, which also has few village and township market - 153 -

places. It is important that Lesotho markets become part of the rotation schemes developing in South Africa. Integration of Lesotho and South African markets will open up larger markets to Basotho producers and encourage them to specialize more.

Export-Oriented Light Manufacturing

7.49 Expanding and integrating labor-intensive light manufacturing is the most promising component of Lesotho's industrial strategy. It could create direct and indirect jobs on a large scale. If export industries continue to expand, the ripple effects will be felt throughout the economy and will lead to reduced unemployment and higher wages -- both crucial to poverty reduction. To keep the cost of doing business down, Lesotho should maintain the minimum wage at competitive levels, improve industrial relations, and regulate utility cost pricing.

7.50 Keep minimum wages at a competitive level. Although the Government should refrain from administratively setting private wage scales, some would argue that it has a social obligation to set the minimum wage. But it should avoid the temptation to administer large increases, such as setting the minimum wage at or near South African levels. Doing so would merely result in job rationing, more unemployment or parallel labor markets, and a loss of foreign direct investment -- all of these effects will hurt the poor. For instance, export industries were attracted to Lesotho because of its preferential trade agreements and the low cost of doing business in the country. The cost of business could rise, however, if increased union activity and excessive salary demands by the public sector raise the wages of those employed in the formal employment to the levels prevailing in South Africa. Lesotho would then mimic South Africa's inefficient growth path and associated social problems: high unemployment for most, along with higher wages and better working conditions for those fortunate enough to find employment in a stagnant formal sector. It is important that Lesotho not fall into this trap.

7.51 Improve industrial relations. The Government has engaged in few pro-active attempts to improve strained relations between management and organized labor. It behooves the Government to actively foster regular communications between management and labor on a range of issues, perhaps by organizing regular deliberation councils on the East Asian model.

7.52 Reduce utility costs by increasing competition, regulation, and decentralization in tlhe utility industry. It will be easier to improve wages and working conditions if utility costs -- typically as important as labor costs -- can be kept under control. To date, there have been insufficient attempts to regulate utility cost pricing or to allow competition in the sector. The result has been aggressive cost recovery by utility companies, with no market or regulatory counterforces to push costs down. In telecommunications, for example, the number of employees per 100 lines increased from 48 to 68 under structural adjustment, meaning that the utility became less efficient while recovering costs. Other parastatal utilities seem to exhibit a similar trend. Consumers and businesses will suffer, and unemployment will increase, if the trend is not reversed. Several companies have already foregone investment or expansion in Lesotho because of high utility costs. But several studies of utility costs are currently under way or planned; their recommendations should be implemented as soon as they are available.

7.53 Lesotho could probably significantly reduce its utility costs by increasing competition in the utility sector (from both domestic and South African sources) and by strictly regulating utility price increases. At the same time, the country should rationalize the management and sources of utilities. In telecommunications and electricity supply, for example, it should consider - 154- complete integration with South African utilities. Lesotho's export firms would benefit from South Africa's cellular phone companies entering Lesotho's market, and Lesotho's industrial estates could benefit from direct hook-ups to South Africa's electricity grid. At the same time, decentralization of some utility services would increase coverage by reducing costs. For example, international experience shows that per capita water production costs are four times higher in centralized than in fully decentralized systems and are lowest under centrally coordinated decentralized systems.

7.54 Root exportfi-rms more firmly in the economy by selling themfactory buildings and leases and by supporting skills training that promotes the use of skilled local workers in the export industry. There is concern that Lesotho's export industry is footloose and could easily pack up and leave. This concern may be exaggerated, since the industry's continued expansion seems to be solidly based on low costs and a productive labor force, but to give export firms stronger roots, LNDC (which currently owns factory buildings, holds the lease to the land, and charges the companies a rental fee) could sell the buildings and transfer the leases to the export firms. (This would give LNDC the capital to build new shells.) Government-supported training programs that elevate local workers into technical and supervisory jobs in the export firms would also help root the industry in the local economy.

Tourism

7.55 Work much more closely witlh Soutl African travel agents. Tourism is still an underdeveloped sector in Lesotho, despite the country's beautiful scenery. The most important tourist market may be South Africa itself, where interest in African culture and village life is growing. But with tourism booming in South Africa and international demand for ecotourism increasing rapidly, Lesotho should also develop a tourist product attractive to tourists visiting South Africa. Rather than go it alone, Lesotho should market a Lesotho option as part of an integrated tour package to South Africa. That would mean establishing closer ties to the South African tourist industry, ratlher than trying to market Lesotho independently in overseas markets. To begin with, Lesotho could start harmonizing its visa regulations with South Africa's to prevent bureaucratic hassles on side trips to Lesotho.

7.56 Develop a special "Lesotho package". In the past, Lesotho followed the South African- inspired hotel and casino model for tourist development. This may not suit Lesotho's comparative advantage or goal of labor-intensive development. Rather than expand the traditional hotel and casino package, the Government could try to develop more small-scale tourism (small hotels, lodges, and cabins) around a special "Lesotho package". This would be more labor-intensive and would provide more opportunities for local entrepreneurs. A real opportunity exists to involve local communities in the design and management of such small- scale tourism.

Infrastructure Investments

7.57 Investing in infrastructure built labor-intensively could help reduce poverty in two ways: The poor would benefit from infrastructure development and maintenance directly, as wage laborers, and indirectly, as consumers of infrastructure services and as employees of private firms emerging in the improved environment. Unfortunately, Lesotho's infrastructure bureaucracy has been inept at developing labor-intensive infrastructure, especially in rural areas. Priorities set centrally are not responsive to local concerns and technologies chosen centrally - 155-

have tended to be capital-intensive. And without local involvement in the planning and implementation of public works, maintenance suffers and infrastructure deteriorates rapidly. Lesotho needs to use more private contractors in public works and invest more using labor- intensive techniques, but decentralize implementation.

7.58 Use private contractors in public works. Public infrastructure development and maintenance has traditionally been a monopoly of the public sector. But efficiency could be enhanced, output expanded, and more jobs created if the public sector restricted its role to planning and coordinating infrastructure investment, privatizing infrastructure development wherever possible, and leaving infrastructure maintenance to local communities and the private sector.

7.59 Decentralize implementation and chioose labor-intensive techniques. In its highly successfully municipal fund program, Mexico was able to finance and execute 75,000 infrastructure projects at an average cost of $11,000 each in a mere four years. This is only one of many examples of successful public infrastructure projects that were made more efficient and equitable by decentralizing decision-making to the local level. Decentralized programs can quickly and cheaply create and maintain rural investment in such technically simple infrastructure as rural roads and bridges, school buildings, and small-scale water supply systems. A recent World Bank review of experience in 42 developing countries found that where road maintenance was decentralized, backlogs were considerably lower and road conditions tended to be twice as good.

7.60 The core strategy is to allocate locally managed grant funds for infrastructure development that requires community participation in project selection and execution. Under this strategy, every year each community receives an allocation to finance the projects it has selected. Execution is managed by a community committee that purchases materials and hires and supervises skilled workers. Communities must also contribute part of the cost -- say, 20 percent - - usually in the form of "sweat equity" (unskilled labor and local materials), to help ensure that the only projects selected are those to which the community is really committed. Studies of Mexico's public infrastructure program show that decentralization not only expanded output greatly but reduced costs by from half to two-thirds.'08 Mexican emigrants returned to Mexico, both as managers on community committees and as skilled laborers. The same thing could happen with Lesotho's workers in South Africa.

7.61 To help give the poor access to infrastructure, it is often better to subsidize access than to subsidize user fees. The main barrier to using infrastructure services is often that poor households cannot mobilize the heavy initial connection charges. In urban areas, utilities can be useful conduits for loans extended to finance connection costs because payment can be secured using regular billing procedures. In Bangladesh, the Grameen Bank provides credit to about 2 million poor and landless persons, most of them women, through group lending (which allows the poor to use social collateral, based on peer pressure, instead of financial collateral). The bank has combined group lending with financial mechanisms to extend credit for tubewells and sanitary latrines. In 1993, it lent $18 million for this purpose; since 1992 it has provided loans for about 70,000 suction tubewells.

108See p. 73, box 4. 1, World DevelopmentReport, 1994,Infrastructurefor Development, World Bank, 1994. - 156 -

C. INVEST IN HUMAN RESOURCES

7.62 Economic growth attentive to the needs of the poor historically values broad-based investment in human resources. Human capital is, almost by definition, the only capital the poor possess. In the long run, investing in health care and education is the only way to eradicate poverty in Lesotho.

Health Care

7.63 Lesotho has a relatively well-developed network of health services, in which the churches and NGOs play an active role. This network is an important asset. It can be improved by streamlining the activities of service providers, especially through joint planning of health activities.

7.64 Exenmpt chlildren under five, the elderly, and the disabledfrom user fees. Increased efforts to recover costs under structural adjustment have often meant across-the-board fee increases. It is important to exempt the poor from such cost increases and to target free health services to vulnerable groups.

7.65 Introduce a free or low-fee package of basic healith services. To improve universal access to basic primary health care, the public health system should identify a package of basic health services, assign very low fees for services in the basic package and aim for full cost recovery on all other services. Most of the basic-package services are generally provided by low- end facilities rather than by hospitals, so delivery would be biased toward the poor, since low- end facilities service the poor, not the rich.

7.66 Intprove geographic equity. Poverty in Lesotho is associated with geography, and rural mountain areas harbor the worst poverty. It is important to target health services increasingly toward these poor areas. To improve geographic equity, the sector should:

* Review the current distribution of health services, to provide a benchmark for future planning.

* Target investments to the Senqu Valley and rural mountains, where poverty is concentrated.

* Create appropriate -- real, market-related -- incentives for health care workers to serve in remote areas: Make it pay both financially and professionally to be posted to remote rural areas.

* Strengthen community-based approaches, so health services can expand cost-effectively into poorer regions.

7.67 Allow localfacilities to retain certain userfees. Service delivery could improve if health services were decentralized. Local participation in the management and maintenance of - 157- health service facilities will improve if people feel their paying user fees allow health care services to function and feel they have local control over the use of such fees.

Education

7.68 Massive investments in primary education are a common denominator in economic success stories all over the world. Top priority should be given to investing in primary education, the social returns from which are extraordinarily high. Lesotho should aggressively ensure primary education to as much of the population as possible. A good argument can be made for providing primary education virtually free of charge.

7.69 Lower the costs to parents of primary and secondary schooling. Basic primary school fees, currently about MIOO a year, are already excessive and clearly anti-poor. In addition, parents must pay many extra costs (for books, school uniforms, and fees to hire extra teachers). And certain exam fees are prohibitively expensive, often several hundred Maluti.'09 These costs are biased against the poor and result in high dropout rates. There have been several studies of what schooling actually costs parents. These studies should be used to formulate policies to cut family costs. General costs could also be lowered through decentralization -- for example, by creating local school boards with authority over local budgets.

7.70 Strengthen nonformal education for out-of-school youtlh. The low level of education for boys is a serious social problem in Lesotho. In the past, it was realistic to expect no more for boys than a career of unskilled physical work in South African mines, but Basotho parents are still unwilling to invest in their boys' education, and instead of sending boys to school often use them as cheap labor for herding livestock. Declining mine employment will eventually change people's attitudes, but the Government can help by actively promoting education and changing public perceptions about future opportunities for mine work. When the cost of primary education has been substantially reduced, the Government might also consider imposing age restrictions on herding.

7.71 The Lesotho Distance Teaching Centre has a poor track record on reaching out-of-school youth. It might be more effective to allow local teachers to earn overtime by providing afternoon and evening classes to out-of-school youth. The overtime option might add to the allure of teaching in remote rural areas.

7.72 Improve vocational and techlnical education. The general level of schooling in Lesotho puts it ahead of the rest of the region, but it is weak in vocational and technical education. Yet jobs and the economy cannot grow unless firms can hire enough workers with the right skills. The impact of the Lesotho Highlands Water Project on domestic employment has been below expectations partly because the Basotho labor pool lacked the right skills. Education planning in Lesotho seems to assume that students will one day be employed in the public sector; the idea of the private sector as a stakeholder in education is fairly new. Lesotho's private sector should be encouraged to participate in curriculum design so that the skills mix of the labor pool improves. One important reason for the success of education in East Asia and West Germany is the private sector's extensive involvement in technical education.

109The CambridgeOverseas School exam costs about M700 per attempt;the Junior Certificate, between M200 and M300 per attempt. - 158-

7.73 The Government, in collaboration with the private sector, should undertake market research to determine what types of vocational and technical education are in high demand and short supply, and should develop the institutional framework appropriate for providing and financing the vocational and technical education needed. Such a framework might include, for example, a business training center operated as a joint venture between the Government and the private sector.

7.74 Invest more in education in poorer regions. Investing more in basic education in the rural mountain zones would be unambiguously pro-poor. But as educational services in remote rural areas expand, the country must also develop more cost-effective ways to deliver educational services. Decentralization would both improve the delivery of education and reduce costs. School buildings could be maintained more cost-effectively (and more labor-intensively) by local communities than by a central bureaucracy.

7.75 Create incentives for qualified teachers to teach in remote areas. If primary education is the cornerstone of Lesotho's human resource development strategy, teachers -- the first-line providers of education -- should be paid accordingly. Government should carefully monitor teachers' salaries and keep salary levels competitive. To get teachers to work in remote areas, the system must put together an attractive remuneration package that combines real financial and career incentives.

7.76 Spend proportionately more public funds on primary education and proportionately less on college and university education. It is difficult to overstate the importance of primary education to a pro-poor economic growth pattern. A common denominator in the success stories of the East Asian countries, which successfully boosted economic growth and reduced poverty, was and remains the high priority they assigned to primary education and the lower priority assigned to public investment in tertiary (college and university) education. East Asian governments routinely argued that it would be cheaper to send their university students abroad than to attempt to create centers of academic excellence in their own countries.

7.77 Lesotho's general growth in spending on education is currently pegged at fiscally sustainable levels, so increasing primary education's share of the sector's budget should be achieved mainly by redirecting expenditures within the sector. Some costs can be cut by continuing to decentralize the educational system, but most of the proposed increase in spending on primary education should come from reduced spending on tertiary education. Currently, the Government spends only about M200 per primary school student, compared with Ml,000 per secondary school student and a staggering M 15,000 per university student.

7.78 About a quarter of the Ministry of Education's budget is allocated to the National University of Lesotho (NUL), which spends three-quarters of its budget on administrative overhead. Students at NUL receive a generous "all-inclusive" benefit package (which even includes laundry and student-room cleaning services). Students in Lesotho, as in most other countries, should get funding from various sources, including means-tested and merit-based scholarships and part-time work. But government spending per student is so high that it would be more efficient to send students to first-class South African universities than to educate them at NUL. The Government could save substantially by financing scholarships for studies abroad (in particular in South Africa), rather than try to offer comprehensive, first-rate tertiary education in Lesotho. - 159-

7.79 In any case, political changes in South Africa and Lesotho demand that tertiary education be evaluated more and more from a regional perspective, and achieving "self- sufficiency" in tertiary education may turn out to be too costly to be sustainable. NUL's academic reputation is not strong enough to warrant continued across-the-board public subsidies. Instead, NUL should try to establish a regional comparative advantage in pre-college and college-level education, possibly with an increased emphasis on technical, health, and education training. If NUL succeeded in providing bachelor-level college education at a regionally competitive level, it might increasingly attract paying students from South Africa, which would reduce Lesotho's fiscal burden. And the Government could direct those cost savings toward primary education.

D. STRENGTHEN THE SOCIAL SAFETY NET

7.80 The first two strands of the poverty action plan -- fostering labor-intensive growth and investing in human resources -- aim to reduce poverty indirectly and in the long run. An essential third strand -- short-term and direct -- is to strengthen the social safety net. The best way to begin is to lower the cost of staple foods, expand labor-intensive public works (self-targeted to the poor), and target special programs to vulnerable groups.

Lower The Cost Of Staple Foods

7.81 The current pricing and marketing system are clearly regressive -- extremely anti-poor -- with grave social and nutritional consequences. The quest for food self-sufficiency and the protection of capital-intensive parastatals have artificially raised the price of such basic goods as maize. And since food is a basic good, its high price has influenced wage levels. Lower food prices would reduce pressure on the wage bill (both private and public) and make Lesotho's products more competitive internationally.

7.82 Remove trade, marketing, and processing restrictions. To protect poor consumers, bans should be lifted on imports of fresh milk and processed products such as maize and wheat flour. Removing trade restrictions on maize meal could reduce its price by 30 percent. Using the household expenditure analysis done for this report, this would imply a reduction in the incidence of poverty from 49 to 44 percent, and of ultra-poverty from 26 to 18 percent. If the Government wants to continue protecting Basotho grain producers against competition from South African producers, it should at least do so by subsidizing them more directly -- for instance, by subsidizing rural and agricultural infrastructure -- instead of punishing Basotho consumers, especially the poor. And instead of trying to stimulate agricultural production of crops for which Lesotho has little comparative advantage (and which would therefore continue to need protection), agricultural policy should encourage specialization in agricultural crops for which Lesotho would have a comparative advantage.

7.83 Exempt basic foods from tlhegeneral sales tax. South Africa has already moved to a value-added tax system, from which it has exempted maize. In the near future, Lesotho will also probably adopt a VAT system. After removing trade and price distortions, Lesotho should harmonize its tax regime and VAT exemptions with South Africa's, to prevent the loss of tax revenues through smuggling. - 160-

Expand Labor-Intensive Public Works

7.84 Public works programs can indirectly become part of the safety net for the poor, if labor- intensive methods of construction are used. In other countries, decentralizing the design and implementation of public works often results in the use of labor-intensive methods, which provide more jobs for the poor.

7.85 Decentralize planning and management. Decentralizing the construction of infrastructure tends to expand geographic coverage, reduce costs, create jobs on labor-intensive projects, and improve maintenance. Small, simple infrastructure should especially be provided locally.

7.86 Pay in cash (at below-market wages, to make project self-targeting), rather than in- kind. In-kind payments unnecessarily restrict the expenditure choices of the poor. Below- market wages avoid competing with the private sector for labor and to ensure that only the truly poor work on public works schemes.

Target Special Programs To Vulnerable Groups

7.87 It will be difficult to identify vulnerable groups at first, but with proper monitoring, targeting would improve over time. One practical suggestion for identifying beneficiaries is to have local chiefs select vulnerable households initially, using three criteria: (1) the household should have a member who is medically disabled (blind, deaf, crippled, or severely malnourished), a medical disability the local hospital would confirm annually; (2) no household member should be formally employed; and (3) the value of property owned by the household should be below a certain level. The Ministry of Social Welfare could administer the scheme. It would be important to conduct random checks on the appropriateness of selections made by the chiefs.

7.88 Lesotho could investigate the feasibility of a food stamp program to reduce malnutrition, whiclh has reached alarming levels. Food stamps are like quasi-money, which beneficiaries can exchange for approved food items in any food store. The Government would guarantee payment of the cash value to the food supplier. International experience with such systems is not good, however. Apart from imposing artificial constraints on spending choices by the poor, these programs often invite corruption.

7.89 Recurrent drought is almost a certainty in Lesotho, and the Government should prepare for it. Lesotho's most recent experience with drought was ameliorated by the Government's productive collaboration with the NGO community in delivering drought relief. Rather than rely on ad hoc collaboration, the Lesotho Council of NGOs and the Government should formalize their recent working relationship so a clear institutional mechanism is in constant readiness.

7.90 Regional arrangements have failed to address the problem of drought with long-term solutions. A concerted effort should be made by regional governments, major donors, the private sector, and the NGO community to develop a comprehensive long-range strategy for dealing with drought in Southern Africa.

7.91 A number of countries are testing innovative commercial drought insurance schemes. One promising option is a commercial voucher scheme that pays the holder the face value on the - 161 - voucher when drought is officially declared to exist by some objective standard. District rainfall stations might declare drought in a certain district on the basis of a certain cumulative rainfall deficit, for example, which would then trigger payment of the voucher in that district. The poor could be given subsidized, or even free, vouchers; others would pay for the vouchers, which could be organized as a commercial savings and insurance scheme.

E. IMPROVEINSTITUTIONAL CAPACITY

Design And Implement A Poverty Reduction Program Of Investments Financed From The Development Fund

7.92 The Development Fund collects revenues attributable to the Lesotho Highland Water Project. A transparent, decentralized mechanism to use part of these funds for poverty-reducing investments such as rural infrastructure would improve the LHWP's status. The pilot community survey program undertaken with UNICEF's assistance seems a promising example to replicate nationally. From these community surveys, done by villagers themselves, can emerge village development plans that could be funded from the Development Fund. Currently, the mechanism through which village-level projects are proposed for funding from the Development Fund is unclear. This project must become more transparent. Even in the civil service there are doubts about who the project's ultimate beneficiary is.

Decentralize Authority

7.93 Decentralizing political, fiscal, and administrative authority is the key to reducing the cost of, and hence expanding, the delivery of infrastructure, utilities, and social services. Decentralized planning and execution can yield substantial efficiency gains because coordination and information cost less. At the heart of rural development, for example, is a core of rather simple, small investment projects, such as the construction of rural roads and bridges, the digging of wells, the development of small-scale irrigation, and the building and repair of schools and matemities. Planning and executing such activities locally does not usually involve complex and costly coordination or information gathering. But when coordination and implementation of simple activities is undertaken by a faraway central bureaucracy or project authority, rural development does become both complex and costly.

7.94 To be effective, decentralization must proceed along all three paths: political, fiscal, and administrative. The key is public accountability, which is why leaders should be chosen locally, finances should be controlled locally, and administration should be the responsibility of local authorities. Assigning clear responsibilities to a particular government unit and a specific service is essential to successful decentralization. If local government does not have clear financial and managerial discretion, it cannot be held accountable for either success and failure.

7.95 Local elections of councils, committees, and mayors can greatly improve political accountability but other measures are also needed, especially in Lesotho, which has just emerged from a long period of authoritarian rule characterized by weak local institutions. Such supplementary measures typically include regular consultation with constituents, strong monitoring of user satisfaction with local services, and the linking of civil servants' career paths to their responsiveness to constituents. - 162-

7.96 To be held accountable, local leaders must have fiscal and administrative control. This implies control over revenues, which is regulated by adequate local finance laws, covering budgeting, financial reporting (accounts and audits), taxation, contracting, and dispute settlement. Most important, fiscal accountability requires being able to tax and to charge user fees. The key to making decentralized government more sustainable is to allow more local revenue raising, which includes getting those who use public services the most to pay the most for them. Transfers from the central government should be widely publicized, for transparency. In Lesotho this means, among other things, widely publicizing the existence and operating guidelines of the Development Fund (in which LHWP revenues accumulate) and all allocations made from it.

7.97 Generally, centralized regulation should be relaxed wherever the impact of local government is largely localized, and wherever regulation requires detailed knowledge of local conditions and priorities. But it is still clearly needed in areas where the behavior of local governments can affect national welfare -- for example, macroeconomic stability or the environment.

7.98 In Lesotho, decentralization may initially be hampered by a shortage of the skills and technical expertise needed for even simple village projects. Here, NGOs and associations of private firi.s can help enormously, both to ease the transition from central to local rule, and to help make the delivery of services more efficient. But local stakeholders should ultimately be responsible for deciding which particular NGO or firm to hire to help with project planning or execution, or whether to deliver a specific service. Local capabilities will not grow if mandates and resources are simply transferred from the central bureaucracy to NGOs and private firms that bypass local stakeholders. Rather than fund NGOs directly, central government could allocate resources to the local administration and let it decide how it wants services to be delivered -- with what mix of line ministries, NGOs, and private firms.

7.99 Political, fiscal, and administrative decentralization can quickly and dramatically increase the number of successfully executed development projects. The key to success, based on international experience, is to stress the importance of all local stakeholders participating directly. Three principles are important to remember. First, beneficiaries should be involved directly, not throughi intermediaries who claim to represent the community. (Consulting with local officials and NGOs is not a substitute for the direct participation of beneficiaries through something like a regularly held town or village meeting). Second, early consensus should be sought on project proposals, so costs are kept low and project design is more appropriate in terms of technological complexity and labor intensity. Third, cash or in-kind contributions should be mobilized up front. This not only reduces costs to the Government, but increases local commitment, ensures the true setting of priorities, and increases the likelihood of the community's subsequent maintenance of the project.

Reform The Civil Service

7.100 A decline in civil service performance has contributed to Lesotho's current crisis in governance. In the near future, with the help of the UNDP, the Government will implement the civil service reform Lesotho urgently needs. Civil service reform will require making some politically difficult choices. First, the Government must take a hard look at which activities to undertake and which to shed. Many services currently provided by the public sector -- tractor - 163 - services, for example, and the maintenance and cleaning of government buildings -- could be privatized with no net reduction in employment in the economy. If the Government wants a high- quality civil service that undertakes only those activities that the Government should be doing (for example, ensuring law and order, delivering social services, providing public infrastructure, and regulating economic institutions), it must ensure its civil servants both good pay and rewarding careers. And given current fiscal constraints that means reducing the size of the civil service. A civil service reform program that postpones these difficult choices will fail even before it begins.

Privatize

7.101 The civil service could be reduced considerably and made more efficient through privatization. Viewing itself as a necessary substitute for what it views as the weak private sector, the Government continues to involve itself in productive and trading activities. The Government is still creating and contemplating new state companies and parastatals, even though its record substituting for the private sector is abysmal. Ironically, loss-making parastatals are staffed by relatively well-paid civil servants who are unequipped to run companies, while the delivery of social services is declining because civil servants working in those sectors are underpaid and have no career prospects. The Government should stop trying to substitute for a private-sector, and the people running crucial public services should start being paid adequately or better. Careers in public health and education should be well-rewarded, attractive career options.

7.102 The privatization of parastatals and state companies could give the Government the resources it needs to focus aggressively on activities that address core unemployment and poverty problems. Because privatization is a politically sensitive issue -- mainly for fear of foreigners taking control of previously state-owned companies -- it is tempting to proceed at a snail's pace on privatization, but that would be both costly and counterproductive. Experience elsewhere dramatizes the many pitfalls of a gradualist approach to privatization and the greatly increasing cost to the Government. More radical approaches, such as mass-voucher privatization in some East European countries, generally have a much better record of success. Allowing broad popular participation in the privatization of companies, by distributing shares on a grant basis to the citizenry, seems to achieve privatization more quickly and with broader political support, effectively neutralizing the lobbying against privatization by parastatal managers and employees. The Government's Privatization and Private Sector Development Program, funded by the World Bank, adopted a process (rather than blueprint) approach, which should make possible the design of a privatization strategy that will suit conditions in Lesotho, one that should be able to achieve quick results.

7.103 In making the difficult choices needed for meaningful civil service reform, the Government will face enormous, but short-sighted, lobbying from the parties most affected. But the alternative is accelerating decline in civil service performance, increasing corruption, declining morale, and an increase in the brain drain to South Africa. Ultimately, such trends call into question the legitimacy of the state. Unfortunately, there is no discernible consensus on the need to trim government and to privatize many of its activities. If anything, the Government is still creating state companies at great expense and with only limited potential for increasing employment. Most recently, it created a ceramics tile factory that has demanded millions in investments but will yield only a few dozen jobs. -164-

Improve Planning And Budgeting

7.104 In Lesotho, annual planning and budgeting often merely replicates historical budget allocations. There seems to be neither an adjustment to new priorities or a phasing out of old ones. A rudimentary start was made with public investment planning by incorporating data on development aid flows from donor agencies into budget numbers under the Public Sector Investment Program. But to date no efforts have been made to scrutinize the PSIP and to check for consistency with the Government's stated policy objectives. The Ministries of Economic Planning and Finance should take the lead in establishing a truly planned budget process, one that is reflected in the PSIP. Regular consultations with the Ministry of Employment would ensure that the goal of increasing employment is considered in all project design and evaluation. Too many projects are excessively capital-intensive.

Merge Fragmented Program Units

7.105 There is much duplication and little coordination among complementary operations in the Government and parastatal bureaucracy. The Ministry of Works has two units responsible for paved roads and intermediate roads, for example, while the Civil Works Section of the Ministry of Home Affairs is responsible for unpaved roads. These three sections should be merged in one roads department. Similarly with soil and water conservation: donga reclamation, conservation works, and Village Water Supply, which operate in different ministries, could cut costs substantially by merging. Water and sewerage are provided through three organizations: the Ministry of Natural Resources, the Water and Sanitation Authority, and the Ministry of Home Affairs. Reducing such fragmentation should increase cost-effectiveness. Particularly problematic is the inadequate, fragmented attention paid to small-scale local entrepreneurs. LNDC treats foreign investors to a one-stop investment center but local entrepreneurs have to battle with various bureaucracies, with no support from a strong advocacy institution. A merger between LNDC, BEDCO, and BAPS should be considered.

Start Doing Annual Household Surveys

7.106 Too much time and effort is wasted debating who the poor are and whether or not they are hurt by policy changes such as those advocated under structural adjustment. This report, for instance, could not draw any firm conclusions as to whether or not the poor were hurt by structural adjustment, due to the lack of a time series of comparable household surveys. But computer technology has now reduced the financial and manpower demands of representative household surveys to a fraction of their former levels. These surveys should be conducted annually, with sections on income, spending, assets, and other indicators of social welfare, perhaps supplemented each year with a section with a special emphasis for that year (such as nutrition, gender issues, land tenure, or local government).

7.107 To stimulate applied research, put the data immediately into the public domain. The Government could also contract with a private institution (a consultancy firm or the National University) to analyze the state of the economy, especially in terms of poverty and employment. The survey analysis could serve as an independent assessment of how the Government is doing in important areas. - 165-

Annex I

LESOTHO POVERTY ASSESSMENT STUDY

Modeling Economic Linkages

1. Production, payments to labor and capital, household incomes and expenditures, and government spending and revenues are inter-related parts of an economy. Changes in output imply variations in wages and returns to land and capital which in turn determine household incomes. With their earnings, households purchase commodities and services for final consumption. Demand for intermediate goods also varies with output. Government expenditures help determine total demand in the economy and include wage payments to households. Taxes reduce real incomes for households while raising revenues for the Government.

2. These relationships between production, incomes and expenditures are captured in a social accounting matrix (SAM), a consistent set of accounts describing economic flows in an economy during a fixed period of time. Adding behavioral equations linking changes in production, incomes and expenditures defines a model of the economy with which the impacts of policy changes and exogenous shocks can be analyzed. In particular, by disaggregating households according to their sources and levels of income, it is possible to determine the effects of policies and other shocks on poor households.

3. In Chapter 1, a social accounting matrix for Lesotho is presented, highlighting salient features of the Lesotho economy, particularly the characteristics of lower income household groups. The methodology for constructing the SAM is described in broad terms below. A discussion of the semi- input-output (SIO) model follows.

The Social Accounting Matrix

4. The social accounting matrix used in the multiplier analysis (Appendix 1) derives from the Lesotho national accounts sectoral aggregates for 1987, data from the 1986/87 household budget survey,I and an earlier SAM for 1984/85.2 The 1987 social accounting matrix for Lesotho includes 24 activities producing the same number of commodities, 14 factors of production, 12 household groups, 5 accounts for non-government institutions (both private firms and various parastatals), government recurrent expenditures and separate accounts for the customs union, indirect taxes and direct taxes, a combined capital account, and two foreign accounts, showing trade and capital flows with the Republic of South Africa and with the rest of the world.

5. Annex Table I shows the production activities included in the SAM. Agriculture and livestock together account for only Ml I I billion, 17.8 percent of value added, a low figure by sub-Saharan African standards. The construction sector is almost as large, M99 billion. Public administration is in fact the largest sector in the economy, with a value-added of M143 billion, 22.8 percent of the total. Not shown in the table are workers' remittances, equal to M706 million, greater than all production activities

l Data on householdcharacteristics and expendituresfrom the householdbudget survey weregenerated by Emmanuel Skoufias,a memberof the PovertyAssessment team.

2Carvahlo, 1988 - 166- combined. With the large inflow of earnings from overseas, there is a corresponding large component of imports in total supply of commodities.

Annex 1, Table 1: Production Activities in Lesotho 1987 (Million Maloti) Production Value Added 1. Field crops 59.987 40.158 2. Vegetables 8.795 5.400 3. Fruits 7.476 6.930 Subtotal: Agriculture 76.258 52.488

4. Cattle 24.920 19.752 5. Other livestock 33.029 16.253 6. Wool, mohair production 23.653 22.610 Subtotal: Livestock 81.602 58.615

7. Mining 2.725 1.881 8. Meat processing 26.210 26.210 9. Dairy 0.000 0.000 10. Milling 172.145 22.168 11. Other food processing 46.401 22.637 12. Weaving, leather 11.047 3.900 13. Handicrafts 6.556 1.600 14. Textiles 47.713 19.000 15. Agricultural tools 4.057 1.474 16. Other manufactures 14.466 10.826

17. Electricity, water 5.678 4.747 18. Building,construction 135.375 98.542 19. Trade 106.499 60.000 20. Hotels, restaurants 21.807 11.839 21. Transport 26.970 13.658 22. Private services 76.575 53.758 23. Housing 47.268 38.300

24. Public administration 142.697 142.697

Total 1052.049 625.556

Source:Bureau of Statisticscomputer files, Carvahlo (1988) and World Bank calculations.

6. The 12 household groups in the 1987 Lesotho SAM are classified according to three major criteria: location (urban/rural), access to workers' remittances, and gender of the head of household. Only 12.6 percent of the population resides in urban areas, where household expenditures per capita are more than twice those of the average rural household. On average, households in the rural uplands regions (including the Senqu River Valley) are 20 percent poorer than their lowland neighbors in terms - 167- of per capita expenditures. Households with migrant income are generally significantly better off than those without migrant incomes. The exceptions are male-headed households in urban areas. Finally, female headed households are significantly poorer than male-headed households in all cases except for rural uplands households without migrant income.

Overview of the Model3

7. Investment in productive sectors of the economy, changes in export demand and variations in workers' remittances all have effects which reverberate throughout the economy. As sectoral output increases, demand for intermediate production inputs rise. Wage incomes and returns to capital also increase, leading to greater demand for consumer goods. Changes in export demand can also result in increases in output of sectors where excess capacity exists. Similarly, variations in workers' remittances affect household incomes, their demand for goods and services, domestic production and imports.

8. Where domestic supply is inelastic in the short run, increased demand leads to reduced exports or increased imports. Where excess capacity exists, an increase in demand, either intermediate demand or final demand, leads to increased production. This in turn leads to an additional increase in domestic incomes, another increase in demand and further increases in production.

9. The measurement of these indirect effects requires a model that relates sectoral output, household income, consumer demand, and inter-industry input linkages. A key determinant of the magnitude of these multiplier effects is the extent to which domestic productive sectors are able to increase domestic output when demand rises.

10. One option, the input-output model, embodies the classic approach to this question. It sets total supply in each sector (Z) equal to the two sources of demand, interindustry input demand (AZ) and final consumption demand (F). Final demand includes consumption by households (BY) and exogenous sources of demand such as exports (E). The value added share (v) in gross commodity output (Z) determines income (Y).

Z = AZ+ F (1) =AZ+ PY+ E =AZ+ PvZ+ E

I. Presuming supply to be perfectly elastic in all sectors, total output and incomes become determined by the level of exogenous demand (E).

Z = (I-M)-'E (2)

12. Because they assume perfectly elastic supply in all sectors, input-output models over-estimate output responses following from any intervention or exogenous shock. Yet in reality, in most developing

'This section is based on Dorosh and Haggblade (1992). - 168- countries some sectors face supply constraints. This is especially true for agriculture, where land, labor, rainfall and technology frequently limit output, and formal manufacturing where physical capital and skilled labor are limiting factors. By ignoring supply constraints altogether, input-output models typically overstate multiplier effects. For example, agricultural growth multipliers are, on average, overstated by a factor of two to ten (Haggblade, Hammer and Hazell, 1991).

13. A more realistic alternative is to use a semi-input-output (SIO) model. While retaining many of the basic assumptions of the IO approach, the SIO model differs in that it introduces supply rigidities in some sectors. The following two equations, contrasted with (I) and (2) above, capture the SIO model's essential distinction. By classifying all economic sectors as either supply-constrained (Z1) or perfectly elastic in supply (Z), the SIO model permits output responses only in some sectors (Z2)- In supply-constrained sectors (Z1), increases in domestic demand merely reduce net exports (El), which then become endogenous to the system.

Zi = A,Z + P,v,Z + E, (3)

Z2 = A2 Z + P2V2Z + E2

E,(I- |Z'K (4)

For a formal exposition of the SIO model, see Appendix B.

14. The semi-input-output (SIO) model is described graphically in Annex Figure 1. For simplicity of exposition, the figure shows only three aggregated categories: (ZI) textile factories; (2) other supply-constrained commodities: tradables such as foodgrains, livestock, other formal manufacturing; and (Z3) commodities highly elastic in supply: nontradables such as services, informal industries, perishable agriculture. - 169-

Annex 1, Table 2: Household Expenditures, 1987 Results Number of Share of Population Average Expenditures households households (000) Share HH Size per capita per HH (thousands) (Maloti/month) Urban 46.3 14.1 183.5 12.6 3.962 91.9 364.0 Female HH 14.6 4.4 53.0 3.7 3.639 68.0 247.4 w/ Migrant income 2.0 0.6 9.7 0.7 4.753 59.1 281.0 w/o Migrant income 12.5 3.8 43.3 3.0 3.457 70.0 241.9 Male HfH 31.7 9.7 130.5 9.0 4.111 101.6 417.5 w/ Migrant income 9.5 2.9 41.8 2.9 4.415 86.1 380.3 w/o Migrant income 22.3 6.8 88.6 6.1 3.981 108.9 433.4 Rural Lowland 191.2 58.2 855.5 58.9 4.473 47.6 213.1 Female HH 50.9 15.5 198.7 13.7 3.907 34.1 133.1 w/ Migrant income 14.3 4.3 74.1 5.1 5.197 40.5 210.6 w/o Migrant income 36.6 11.1 124.6 8.6 3.405 30.2 102.9 Male HH 140.4 42.8 656.8 45.3 4.679 51.7 242.1 w/ Migrant income 75.4 23.0 368.2 25.4 4.881 63.2 308.5 w/o Migrant income 64.9 19.8 288.5 19.9 4.443 37.1 164.8 Rural Uplands 90.8 27.6 412.4 28.4 4.543 37.6 170.9 Female HI-I 23.6 7.2 84.4 5.8 3.580 36.4 130.1 w/ Migrant income 5.0 1.5 23.9 1.6 4.821 36.2 174.4 w/o Migrant income 18.6 5.7 60.5 4.2 3.250 36.4 118.4 Male HH 67.2 20.5 328.0 22.6 4.881 38.0 185.3 w/ Migrant income 27.3 8.3 138.8 9.6 5.085 47.0 238.8 w/o Migrant income 39.9 12.2 189.2 13.0 4.741 31.4 148.7 All Lesotho 328.3 100.0 1451.3 100.0 4.421 50.3 222.2 Source: Lesotho 1987 Household Budget Survey computer data files and World Bank calculations. - 170-

Annex 1, Figure 1: Graphic Representation of Multiplier Effects Field Crops and Other Sectors with Perfectly Textiles Supply-Constrained Elastic Supply Tradables (Inelastic supply, (Inelastic supply, (Nontraded goods, imported good) exported good) perfectly elastic supply)

Round I Invest in Textiles - increase in supply - increase in farmer income

Round 2 Increased demand (final and Increased demand (final Increased demand intermediate goods); supply and intermediate goods); (final and fixed at Q' supply fixed at Q intermediate goods) - exports decrease - exports decrease - increased output - income unchanged - domestic output and - increased income income unchanged

Rounds 3 and 4 Increased demand (final and Increased demand (final Increased demand intermediate goods); supply and intermediate goods); (final and fixed at Q supply fixed at Q intermediate goods) - exports decrease - exports decrease - increased output - domestic output and - domestic output and - increased income income unchanged income unchanged FieldCrops nd Other Su4ply-Constrained Sectors with Perfectly Text le Trodables Elastic Suppiy

O 2 3 D4 S0 SI

'. \ ~ %

p2~~~~~~~~~~~~~~

'10 (t2 34 3 q1 04n ql qo - 172-

15. Consider first the consequences of public investments in production of textiles, e.g. through the construction of a new factory. The immediate impact of the intervention is to increase output of textiles.

16. In Round 1, this directly raises national income by 0.398 Loti for every one Loti of increased textile output. This direct injection triggers a series of responses that increase income even more.

17. In Round 2, the economy registers increased demand for the inputs used in textile production plus increased household spending on consumer goods. These twin channels increase domestic demand for textiles, other supply-constrained tradables and the highly elastic supply of nontradable domestic services, informal manufactures and perishable agricultural commodities. For textiles and other supply-constrained tradables, this increase in demand does not stimulate further domestic production. It merely decreases net exports. In contrast, because of the elastic supply of nontradables, increased demand leads to higher output and higher domestic incomes in these other sectors.

18. The increased production of commodities with elastic supply (Z3) once again raises demand for production inputs and consumer goods. In Round 3, this increases demand in all three sectors. As before, production of textiles and other supply-constrained commodities does not increase. Instead, net exports decrease still more. For this reason, as equation (4) indicates, exports in these supply-constrained sectors become endogenous to the model. Yet once again, output and incomes rise in the activities with highly elastic supply. This induces further rounds of successively dampening demand increases.

19. In total, the indirect effects of the investment in textile production, from Rounds 2 on, stimulate another 0.457 Maloti in national income. Thus the total increase resulting from textile investments equals 0.398 + 0.457 = 0.855.

20. An increase in worker remittances will generate a similar sequence of events. They differ mainly in that the shock, the increase in household incomes due to worker remittances, does not entail an initial increase in domestic supply. The second round demand shifts thus are based only on increased consumption, with no gains due to increased demand for intermediate inputs. Also because the change in incomes across households is different from that resulting the gain in textile production, the composition and magnitude of the demand effects will differ. Ultimately, the total income gain will also be different.

Underlying Premises

21. Two key assumptions underlie the semi-input-output (SIO) model: linearity and fixed prices. As with many kinds of economic models, the SIO requires that all relationships be expressed as linear functions. For intermediate inputs, this standard assumption suggests that increases in output require additional inputs in fixed proportions. For household consumption, it requires that consumption expenditures rise in tandem with income. Although marginal expenditures or input demands may differ from the average, the increments must be expressed as linear functions of output and income.

22. In general, this simplification does not pose great problems. Non-linear systems can be approximated by linear functions in the short-run. And they offer considerable conveniences in computing model solutions.

23. Fixed prices likewise vastly simplify computational requirements by side-stepping cumbersome issues of substitution in production and consumption. Input-output coefficients and marginal budget - 173 -

shares, which remain fixed in a fixed-price world, become endogenous variables in a model where relative prices vary. While computational convenience is not a strong rationale for imposing fixed prices, it does offer a strong incentive to investigate the plausibility of such an assumption.

24. Lesotho is an almost ideal case for a fixed price model because of its open borders and monetary union with the Republic of South Africa (RSA). For tradable goods, the markets in the RSA detennine the price level of these goods in Lesotho to a large extent. For nontraded goods, such as services, informal manufactured goods and many perishable agricultural commodities, fixed prices depend on firns' ability to increase output at constant cost. Formally, this requires a perfectly elastic output supply. Because of the considerable unemployment and excess capacity present in Lesotho, the constant cost assumption appears to be a reasonable approximation of reality. Moreover, given the monetary union, there is little scope for changes in the general price level of non-tradables in Lesotho relative to those in the RSA.

25. The proximity of Maseru and other major urban centers to the border with the Republic of South Africa poses other problems -- almost no goods or services are truly non-tradable for urban consumers. Urban Basotho purchase fresh vegetables imported from South Africa in local markets. Households also cross the border to make purchases of goods and services. To account for these purchases, which in terms of the multiplier analysis, represent a leakage from Lesotho's economy, a version of the SIO model is used in which a fixed fraction (70 percent) of urban households' marginal consumption expenditures are imported.

26. At some point, as expansion and liberalization proceed, supply constraints may develop in some of Lesotiho's non-tradable sectors. If this occurs, the SIO model will overstate multipliers resulting from changes in investment and demand. Consequently, some sort of adjustment will be required to capture the income-dampening effects of the price increases that will follow. Recent experiments suggest that in the face of upward-sloping nontradable supply, SIO models overstate true income multipliers by 10 to 25 percent.4 So the simplest accommodation would involve rule-of-thumb discounting based on these results.

27. A much more involved alternative would be to create a computable general equilibrium model with endogenous prices. This type of model would also enable analysis of changes in the real exchange rate vis-a-vis countries other than the RSA, which could affect the real price of certain tradable goods. Given the heavy data requirements, the small size of non-RSA trade, and the relatively small role of price changes for economy policy in Lesotho, this latter approach is probably not cost-effective.

Sectoral Linkages in Lesotho's Economy

28. Annex Table 3 provides measures of sectoral linkages in Lesotho's economy under various assumptions regarding investment and elasticity of supply. In column 1, multipliers using the basic SIO model are given. Here, investment and government spending are held fixed as is the production of the 11 sectors assumed to have inelastic supplies. Leakages from the economic system occur in three ways: through the share of demand of goods with elastic supplies allocated to imports, through increases in net imports of goods with inelastic supplies, and through savings and tax revenues which accumulate but do not lead to increased investment or government spending.

4Haggblade, Hammer and Hazel], 1991. - 174-

29. Because these leakages from the domestic economy are very large, the value added multipliers, (the additional value added generated after the initial shock per unit of value added from the shock itself), for Lesotho are quite small. Using value added by activity as weights, the average value added multiplier is only 0.527. Value-added multipliers in other countries typically range from 1.3 to 1.5 in other African countries.5

30. Only three sectors: mining, milling and weaving/leather have value-added multipliers greater than 1.1. Value-added multipliers are particularly low in agriculture, ranging from 0.09 to 0.17. The multiplier for workers' remittances is also very low, 0.159, given that there are no backward linkages involved in generating this income. Only demand-side forward linkages are generated with increases in workers' remittances.

Annex 1, Table 3: Lesotho: Value Added Multipliers Investment Inelastic SIO Model SIO Model SIO Model Supply with with (*) Reduced Endogenous Urban Investment Demand 1. Field crops * 0.149 0.136 0.232 2.Vegetables 0.091 0.062 0.225 3.Fruits 0.165 0.157 0.228 4. Caftle * 0.149 0.135 0.229 5. Other livestock * 0.141 0.125 0.228 6. Wool, Mohair production * 0.131 0.113 0.227 7. Mining * 1.955 1.890 2.269 8. Meat processing 0.736 0.686 0.941 9.Dairy 0.000 0.000 0.000 10. Milling 1.155 1.092 1.480 11. Other food processing * 0.652 0.591 0.903 12. Weaving, leather 1.021 0.957 1.201 13. Handicrafts 0.832 0.780 0.978 14. Textiles * 1.148 1.098 1.325 15. Agricultural tools * 0.667 0.615 0.909 16.Othermanufacturers * 0.209 0.174 0.357 17. Electricity, water * 0.155 0.106 0.281 18. Building, construction 0.811 0.785 1.148 19. Trade 0.764 0.717 1.012 20. Hotels, restaurants 0.793 0.725 1.014 21. Transport 0.670 0.620 0.911 22. Private services 1.034 0.969 1.264 23. Housing * 0.264 0.210 0.326 24. Public administration 0.173 0.111 0.229 Weighted average 0.527 0.482 0.699 Workers' remittances 0.159 0.155 0.219 Sourcc: Model simulations

SHaggblade, Hazell and Brown, 1989; Haggblade, Hazel], 1990;Lewis and Thorbecke, 1990. - 175 -

31. The multipliers are only somewhat lower if greater leakages from urban spending are modeled (column 2). Under the assumption that 70 percent of urban consumption of "non-tradables" is spent on goods and services produced in the RSA, the average multiplier falls to 0.482, a decrease of 8.5 percent. The decline in the multiplier is rather small due to the fact that large leakages are already modeled in the SIO framework, given that demand for goods that are inelastically supplied is met through an increase in imports.

32. Allowing savings generated to augment aggregate demand by allowing investment to be endogenous increases the multipliers slightly, to an average of 0.699. These multipliers capture only the demand-side effects of investment, not any gain in productive capacity that leads directly to increased production and incomes. Savings generated in the Lesotho economy do not necessarily lead to investment in Lesotho, of course. Capital flight may also occur, in which case there would be little, if any, gain in investment demand.

33. In Annex Tables 4 and 5 (a summary version of which is presented in the main text, Table 3.1), the breakdown in value-added paid to households is presented for multipliers for key commodities and for changes in workers' remittances. Increases in value added from handicrafts, textiles and agriculture (e.g., field crops, cattle) generate the largest gains in income for poor households (0.591-0.817). Benefits to the poor from increases in construction output are slightly less (0.379). By contrast, a I Loti increase in output of manufacturing benefits poor households little (0.200 M). Likewise, a change in the wage rate paid to workers employed in South Africa has only a small effect on poor households, since households receiving remittances are not considered poor (by definition), and the multiplier effects are small. The magnitudes of the multipliers are only slightly reduced with the reduced urban linkages version of the SIO model (Annex Table 5).

34. The share of income that accrues to poor households as a result of an equal I Maloti increase in output or demand of the sectors shown varies. About two-thirds of the income generated from increases in field crops and cattle accrue to the poor. Incomes generated from increases in demand for handicrafts and construction or and increase in textiles output are split evenly between poor and non-poor. For other manufacturing and migrant remittances, however, the non-poor accumulate the largest share of increased incomes, (69 and 87 percent, respectively, in the base SIO model runs). - 176-

Annex 1, Table 4: Distributional Impacts of Increases in Sectoral Output (SIO Base Model) Remittances 1* 4* 13 14* 16* 18 Field Crops Cattle Handicraft Tex- Other Manu- Construc- Migrant Household Income tiles facturing tion Urban Male HH: w/ Migrant income 0.045 0.039 0.239 0.208 0.106 0.080 0.029 w/o Migrant income 0.070 0.074 0.254 0.256 0.183 0.136 0.015 Female HH: w/ Migrant income 0.008 0.008 0.037 0.034 0.020 0.015 0.005 w/o Migrant income 0.031 0.037 0.068 0.088 0.087 0.063 0.005 Rural lowlands Male HH: w/ Migrant income 0.108 0.106 0.105 0.131 0.035 0.071 0.694 w/o Migrant income 0.271 0.265 0.299 0.370 0.080 0.166 0.010 Female HH: w/ Migrant income 0.014 0.014 0.014 0.018 0.006 0.010 0.087 w/o Migrant income 0.089 0.089 0.101 0.129 0.042 0.064 0.004 Rural uplands Male HH: w/ Migrant income 0.029 0.029 0.029 0.037 0.012 0.021 0.181 w/o Migrant income 0.151 0.148 0.167 0.208 0.048 0.095 0.006 Female HH: w/ Migrant income 0.004 0.004 0.004 0.005 0.002 0.003 0.025 w/o Migrant income 0.062 0.061 0.070 0.087 0.022 0.040 0.002 Urban 0.154 0.158 0.598 0.586 0.396 0.294 0.054 Male HH 0.115 0.113 0.492 0.464 0.289 0.216 0.044 Female HH 0.039 0.045 0.105 0.122 0.107 0.078 0.010 Rural 0.729 0.171 0.789 0.985 0.248 0.470 0.009 Male HH 0.559 0.549 0.600 0.746 0.175 0.353 0.891 Female HH 0.170 0.168 0.189 0.239 0.072 0.117 0.118 Poor 0.591 0.582 0.655 0.817 0.200 0.379 0.134 Non-Poor 0.292 0.293 0.735 0.754 0.443 0.386 0.929 Total 0.883 0.875 1.387 1.571 0.643 0.764 1.063 Share of income to households: Poor 0.670 0.665 0.472 0.520 0.312 0.495 0.126 Non-Poor 0.330 0.335 0.528 0.480 0.688 0.505 0.874 Note: ' indicates sectors with inelastic supply - 177 -

Impacts of Structural Adjustment, Drought and External Factors on Poverty

35. In this section, key aspects of policy changes and external shocks on Lesotho's economy in recent years are simulated in order to shed light on the likely impacts on poor households. The analysis focuses on the impact of reduced fiscal expenditures as part of the structural adjustment program initiated in the late eighties, drought, changes in workers' remittances and increases in construction activities associated with the Lesotho Highland Water Project (LHWP).

36. Each simulation presented uses the SIO model developed in Chapter 2 to analyze the effects of changes in policies and external shocks. Government spending is exogenous in these simulations as is the level of output of the public administration. It is important to keep in mind that the results represent averages for household groups and there may be a wide variance in experience of individual households within groups.

Reduced Government Expenditures

37. As part of efforts to reduce the government budget deficit, government consumption was cut by 5.8 percent in real terms in 1989 (Annex Table 6). Since then, real government consumption has risen by 13.7 percent. Despite this recent increase, government spending is below the trend levels prior to the structural adjustment program when it rose by 26.5 percent in real terms between 1983 and 1988 (an average of 4.8 percent per year). The actual level of government consumption in 1992 is 13.1 percent below the earlier trend.

38. The first simulation shows the impact of a 13.1 percent cut in real government output. This reduction in spending strikes government workers the most as wages and salary payments were reduced. Overall, the reduction in spending reduces real household incomes by 1.63 percent, using the base SIO model (Annex Table 7), with the reduced urban demand linkage version of the model, the reduction in real household incomes is 1.58 percent (Annex Table 8). The brunt of these policies is borne by urban groups, whose incomes fall by 2.25 to 5.90 percent with the base SIO model (Annex Table 7). Rural households with migrant income are essentially untouched by the policy since multiplier effects are small. Because earnings from rural government employment are spread across rural households, especially those without migrant incomes, some rural households suffer a decline in income.

Drought

39. The drought in 1991 and 1992 led to 77.6 percent reduced real output of field crops compared to 1990 output (Annex Table 6). The value of production in 1990 was unusually high, however, 78 percent greater than in 1987. Compared to 1987, output in 1992 was down by 60 percent. In contrast, production of livestock in 1992 was almost identical to that of 1987. In simulation 2, only a 60 percent decline in output of field crops is modeled.

40. Because agricultural crop production accounts for less than 20 percent of GDP, (less than 10 percent of national income), the drought has only a slightly larger impact on household incomes (-1.87 percent) in aggregate than the reduction in government spending, (Annex Table 7). The poor are especially hard hit, however, suffering a decline in real incomes of 5.42 percent. Rural households without migrant income see a decline of 6.97 to 7.87 percent. Annex 1, Table 5: Lesotho: Trends in Sectoral Aggregates, 1980-1992 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Agriculture: crops, vegetables,fruits Currentprices 28.8 33.7 30.6 27.5 32.1 41.2 43.4 52.5 86.6 82.8 126.2 69.1 26.1 Constant prices 28.8 27.3 22.0 17.8 19.4 22.8 22.2 22.0 34.4 30.7 39.2 18.6 8.8

Agriculture: livestock Currentprices 23.9 35.3 30.3 39.6 50.3 47.9 57.1 58.6 103.7 131.9 113.3 96.4 122.2 Constantprices 3.9 27.5 21.5 24.5 28.4 24.0 29.5 24.7 27.3 35.7 28.8 30.1 25.9

Factor Income from abroad, net Currentprices 205.0 254.8 372.9 423.1 487.2 514.3 583.3 705.9 829.7 927.8 1101.8 1238.8 1266.4 Constantprices 205.0 227.3 284.9 285.1 289.1 265.8 254.8 271.7 285.3 264.9 266.1 258.5 224.9

Govemment - consumption S Currentprices 74.0 81.0 89.9 91.7 98.6 135.9 159.3 171.5 215.8 222.2 234.6 314.2 380.5 Constantprices 4.0 74.3 78.4 76.7 79.7 84.5 93.2 94.5 97.0 91.3 92.7 98.5 103.8

Construction Currentprices 33.1 36.0 56.2 39.8 50.8 78.3 71.5 98.5 134.3 214.2 291.7 352.0 422.4 Constantprices 3.1 31.4 41.6 24.9 29.3 38.4 30.3 36.0 41.5 63.2 76.2 78.8 85.7

Gross fixed capital formation, building and construction Currentprices 74.7 79.9 125.8 84.7 109.9 171.2 158.1 215.7 317.2 634.2 957.7 1089.2 1359.6 Constantprices 4.7 69.7 93.1 52.7 62.0 84.8 64.6 77.3 102.2 180.8 239.6 239.6 262.2 Source:Lesotho Bureau of Statisticscomputer files. - 179-

Annex 1, Table 6: Household Incomes in Lesotho: Policy Simulations (Base SIO Model) (1) (2) (3) (4) (5) (6) Reduced Drought Reduced Sub- LHWP Total Govemement Migrant Total Spending Incomes

Urban Male HH w/ Migrant income -5.90 -1.09 -3.31 -10.30 6.10 -4.20 w/o Migrant income -4.88 -1.63 -1.63 -8.13 9.80 1.67 Female HH w/ Migrant income -5.44 -1.22 -3.19 -9.85 7.03 -2.81 w/o Migrant income -2.25 -2.22 -1.72 -6.19 14.19 8.00

Rural lowlands Male HH w/ Migrant income -0.22 -0.53 -15.81 -16.56 1.08 -15.49 w/o Migrant income* -2.53 -7.69 -1.33 -11.55 14.73 3.18 Female HH w/ Migrant income -0.22 -0.55 -15.62 -16.39 1.24 -15.15 w/o Migrant income* -2.20 -6.97 -1.44 -10.60 15.53 4.93

Rural uplands Male HH w/ Migrant income -0.22 -0.54 -15.67 -16.43 1.20 -15.23 w/o Migrant income* -2.48 -7.58 -1.35 -11.40 14.86 3.46 Female HH w/ Migrant income* -0.22 -0.56 -15.50 -16.28 1.34 -14.94 w/o Migrant income* -2.51 -7.87 -1.46 -11.83 15.94 4.11

Urban Male HH -5.37 -1.37 -2.45 -9.19 8.00 -1.19 Female HH -3.27 -1.90 -2.19 -7.36 11.89 4.53 Rural Male HH -0.62 -1.78 -13.25 -15.65 3.50 -12.15 Female HH -1.02 -3.14 -10.17 -14.34 6.79 -7.55

Poor -1.78 -5.42 -5.71 -12.90 10.82 -2.08 Non-Poor -1.59 -0.80 -11.92 -14.31 3.31 -11.00

Total -1.63 -1.87 -10.49 -13.99 5.05 -8.94 Note:* incicates"poor" households. - 180-

Annex 1, Table 7: Household Incomes in Lesotho: Policy Simulations (Base SIO Model with Reduced Urban Linkages) (1) (2) (3) (4) (5) (6) Reduced Drought Reduced Sub- LHWP Total Governement Migrant Total Spending Incomes

Urban Male HH w/ Migrant income -5.78 -1.06 -3.26 -10.10 5.90 -4.20 w/o Migrant income -4.74 -1.59 -1.56 -7.88 9.56 1.67 Female HH w/ Migrant income -5.32 -1.19 -3.13 -9.63 6.82 -2.81 w/o Migrant income -2.09 -2.18 -1.64 -5.91 13.91 8.00

Rural lowlands Male HH w/ Migrant income -0.21 -0.53 -15.81 -16.55 1.06 -15.49 w/o Migrant income* -2.42 -7.66 -1.27 -11.35 14.53 3.18 Female HH w/ Migrant income* -0.21 -0.54 -15.62 -16.37 1.22 -15.15 w/o Migrant income* -2.07 -6.93 -1.38 -10.38 15.31 4.93

Rural uplands Male HH w/ Migrant income -0.21 -0.54 -15.66 -16.42 1.18 -15.23 w/o Migrant income -2.36 -7.54 -1.29 -11.20 14.66 3.46 Female HH w/ Migrant income* -0.21 -0.56 -15.49 -16.26 1.32 -14.94 w/o Migrant income -2.38 -7.83 -1.40 -11.61 15.72 4.11

Urban MaleHH -5.25 -1.33 -2.39 -8.96 7.77 -1.19 Female HH -3.13 -1.86 -2.12 -7.11 11.64 4.53 Rural Male HH -0.60 -1.77 -13.23 -15.60 3.46 -12.15 Female HH -0.97 -3.12 -10.15 -14.24 6.69 -7.55

Poor -1.69 -5.39 -5.67 -12.75 10.68 -2.07 Non-Poor -1.54 -0.79 -11.90 -14.24 3.23 -11.00

Total -1.58 -1.85 -10.46 -13.89 4.95 -8.94 Note:e incicates"poor" households. - 181 -

Migrant Remittances

41. Migrant worker remittances fell by 17.2 percent in real terms between 1987 and 1992, sharply reducing national income over the period. In simulation 3, incomes of rural households with remittance earnings fall by 15.5 to 15.8 percent. Urban households with migrant incomes see a smaller percentage decline in household income other earnings sources make their households less dependent on remittances. Overall, the decline in remittance earnings affects mostly the non-poor households, whose incomes fall by 11.9 percent. Rural households without remittances see a decline in incomes of less than 1.5 percent.

42. Combining the effects of reduced government spending, drought and lower migrant remittances, simulation 4 shows an average decline of 14.0 percent in household income. Hardest hit are households with migrant incomes, but non-migrant rural households also see declines of 10.7 to 11.8 percent.

Lesotho Highlands Water Project

43. Fortunately, investment activity associated with the Lesotho Highlands Water Project began during this period. Value added generated in the construction sector increased by 138 percent as gross fixed capital formation rose by a similar magnitude. Assuming that this construction spending followed the same structure as that of 1985 in terms of domestic labor,6 the increase in construction generates demand for labor and sharply raises the incomes of households without migrant workers' earnings.7 Urban households benefit more than do rural households, though the gain in incomes of female headed households may be overstated if they are unable to obtain greater wage earnings from construction sector activities.8

44. Combining the effects of all four shocks, (column 6 of Annex Table 7), the importance of construction activities of the LHWP in preventing a substantial decline in incomes of the poor is highlighted. Without the increased earnings emanating from the increased investment spending, average incomes of the poor fall by 12.9 percent as a result of policy changes and external shocks. Incomes of tile poor fall only by 2.1 percent once the impact of the LHWP spending is included.

Conclusions

45. Linkages across sectors in Lesotho's economy are generally very small. Open borders, monetary arrangements and labor flows knit the economy of Lesotho tightly together with South Africa. These arrangements help to stabilize prices, encourage trade and raise household incomes from migrant remittances, but they also result in large leakages from the domestic economy and low multiplier effects within Lesotho. One implication is that development projects cannot count on large spread effects through second-round multipliers. The large neighboring market offers the potential for exports of goods

6 To the extent that this construction demand resulted in greater demand for non-Basotho workers, the positive effects on domestic households are reduced. ' Historically, as reflected in the 1987 SAM, households with migrant workers remittances generally eamed little income from other sources. Since the S10 model allocates shares of labor income according to the pattern observed in 1987, migrant worker households receive only a small portion of the gains in domestic labor incomes in this simulation. 8 Since many construction activities involve manual labor, female headed households without males of working age may not have gained wage earnings directly from the construction sector. Further disaggregation of the structure of households would shed some light on this issue, but without further disaggregation of labor in the SAM (for which little data exists), only limited refinements of the model analysis are possible. - 182 - for which productivity in Lesotho is high because of special land characteristics, existing investments or skilled labor.

46. The impacts of structural adjustment in Lesotho on household incomes have been relatively small given the limited policy measures undertaken (no devaluations, only limited tightening of fiscal policy). The poor, who largely reside in rural areas, do not appear to have been affected greatly by cuts in government spending, in part because of the linkage effects across sectors are small. Of greater importance has been the drought which reduced incomes of the poor by 5.4 percent relative to 1987 (and by a greater magnitude relative to 1990). The decline in workers' remittances has affected mostly the non-poor on average, reducing their incomes by 11.9 percent. However, among those households who lost remittance earnings entirely, the impacts are likely to have been much larger.

47. Fortunately, the direct impact of construction activities appear to have raised household incomes to offset to a large extent the adverse shocks of drought, reduced migrant earnings and fiscal conservatism. Greater spinoffs from the LHWP are possible if future government revenues are used to help establish permanent enterprises.

48. Finally, a word of caution. The analysis of impacts of policy changes and external shocks on households focus on household averages, using patterns of expenditures and sources of revenue estimated for 1987. The rather optimistic picture suggested by the simulations for average household incomes hides large adverse changes experienced by individual households, who unlike the average household, may have lost a crop or job and thus suffered a far greater income decline. For these households and others who gained little from the spinoffs of increased construction activities, targeted poverty al leviation measures may be justified. - 183 -

APPENDIX A

Construction Of The 1987 Social Accounting Matrix (Sam)

49. The original social accounting matrix (SAM) for Lesotho, hereafter referred to as the 1985 SAM, was constructed as part of a Ph.D. thesis by Joe Carvalho in 1988 (Carvalho, 1988). This SAM is not consistent with the current national accounts figures for 1985 for two reasons. First, national accounts figures are now reported on a calendar year basis rather than on a fiscal year (April - March) basis, while the 1985 SAM was constructed to be consistent with the 1984/85 national accounts.9 Second, the base by which output and value added of several sectors are calculated has changed as data from the October 1986 - September 1987 Household Budget Survey (HBS) have been incorporated in the methodology.' 0

50. Lesotho's national accounts are calculated without an input-output (10) table. Value-added is estimated based on production, known inputs and for some sectors, assumed value-added/output coefficients (Bureau of Statistics, 1990). The IO table constructed by Carvalho (1988) is not used.

51. Consistency of thte national accounts and the 1986/87 Houselhold Budget Survey An attempt has been made by the Bureau of Statistics to reconcile the data on household consumption in the national accounts for 1987 with the total expenditure data from the 1986/87 Household Budget Survey. A numberof adjustments to the survey data for cash and in-kind household consumption (equal to M828.9 million) are made including:'I

1. Under-reporting of consumption of alcoholic beverages and tobacco (M70.0 million), meat and chicken from the household's own production (M35.0 million), and food in-kind from the Work for Food Programme (M25.0 million).'2

2. Subtractions for expenditures on building materials (M51.2 million), fodder and animals (together M26.0 million) which are considered intermediate consumption in own-account conistruction, intermediate consumption in the production of livestock and investment in livestock, respectively.

3. Additions of expenditures on legal aid charges and funerals reported in the survey as domestic non-consumption expenditures (M3.1 million).

These adjustments total a net of M55.9 million (6.7 percent of HBS total).

52. In addition, the total for cash and in-kind production is increased by the amount of direct purchases abroad and imports of migrant workers (M125.3 million). Finally, an adjustment to transform the data to a calendar year basis is made (estimating fourth quarter consumption of 1987 based on the

9The Central StatisticsOffice began producingnational accountson a calendar year basis in 1987 and has since produced a revised calendar year series for 1980 to the present (1992). '>Te sectors for which production and value added estimates are in part derived from the 1986/87Household Budget Survey include vegetables, fruits, small-scale manufacturing, construction, wholesale and retail trade, restaurants and hotels, transport, real estate and business services, health and other services. "These adjustments are described in even greater detail in Central Bureauof Statistics (1990). pp.94-99. 2The estimates of under-reporting are based on national accounts estimates of total supply available for these commodities less reported consumption in the household budget survey. - 184 - survey results for the fourth quarter of 1986 using the estimated growth rate of total consumption between 1986 and 1987 from the national accounts, 5.9 percent, and changes in prices.)13

53. Given these manipulations, the final result (M1051.8 million) is only M4.2 million, (0.4 percent), from the national accounts figure of M1056.0 million. The closeness of the result speaks to the degree to which results from the Household Budget Survey are incorporated into the methodology for constructing the national accounts.

The 1987 Lesotho SAM

54. The 1987 Lesotho SAM takes as its starting point the production and total income data from the national accounts. The input-output coefficients implied in the table of intermediate consumption from Carvahlo (1988) are used in the new 1987 SAM. The mapping between the 24 activities in the 1987 SAM and the 48 activities of the 1984/85 SAM are given in Appendix A, Table 1.

55. A few adjustments are made in aggregating the 1984/85 SAM. Dummy accounts for sheep and goats, skins and hides, and livestock feeds are collapsed into the accounts for activities and uses of commodities. Accounts for government purchases by category (education, health, other government expenditures and public debt) and investment by institution (public investment, private domestic investment, private investment (RSA) and private investment by the rest of world) are aggregated into summary columns which correspond to the other government and capital accounts rows. Negative entries in expenditure columns in Carvahilo's SAM which represent sources of funds are replaced by positive entries in the corresponding rows.

56. In constructing the 1987 SAM, the input-output coefficients from the aggregate 1984/85 SAM are used to generate intermediate consumption and total production given the value added figures from the national accounts.'4 No import data by commodity type is available for 1987, (the latest published figures are for 1986), so the share of imports for each commodity from the 1984/85 SAM is used to break down the total imports figure for 1987. Similarly, import tariffs and indirect taxes on commodities from 1984/85 are scaled up by the increase in overall import tariffs and indirect taxes between 1984/85 and 1987, using data from the IMF, (IMF, GFS 1992). Exports by commodity are taken from the Lesotho Statistical Yearbook 1992. Total investment and government recurrent expenditures on commodities and services are from the national accounts; the approximate breakdown of expenditures by commodities from 1984/85 again provide the information on shares of expenditure by commodities. An adjustment is made to consumption of services and manufactured goods by the Government and as investment goods so that private consumption of services (the residual) is not negative. Accounts for the payments of factor incomes to institutions and transfers by institutions are calculated using the shares from the 1984/85 SAM.

"Not all expenditureitems were adjusted. The actualcalendar year adjustment amounted to only 4.7 percent. '4For themeat sector, a different methodologyis required,given the very high value addedper unit of Outputimplied in the nationalaccounts worksheets (43.3 percent)compared with the 1984/85SAM (10.7 percent). Intermediateinputs per unit of value addedfrom the 1984/85SAM are used for all commoditiesexcept the raw inputs of cattle, sheepand goatsand other animals,where the valuesfrom the nationalaccounts worksheets for the livestocksector are used instead. - 185 -

57. The amount of food aid (donated food) is given in the national accounts tables in the Lesotho Statistical Yearbook for 1992. Like in the 1984/85 SAM, this is shown as a transfer from the Rest of World directly to households.'5

58. Direct taxes are assumed to be paid only by urban households, in proportion to their total incomes. Sales taxes are allocated to households in proportion to total consumption expenditures. Savings rates of rural households without remittances are assumed to be 5 percent. Savings rates of all other households are equal to 11 percent, and are derived as a residual.

59. Expenditures and income sources of each household group were estimated as part of the final balancing of the SAM. Expenditures on individual commodities by each household group are based on data from the 1986/87 household budget survey. The survey data on total expenditures of each household group was adjusted to account for unrecorded consumption of imported goods by households with remittance incomes (see above) and to achieve consistency with data on income sources. Household incomes were calculated using the mapping of factor payments to households in the 1984/85 SAM and then splitting factor payments related to agricultural activities and remittances so as to produce income differentials approximately in line with the observed expenditure differentials across households observed in the household budget survey. Both the household expenditure matrix and the factor payments to households matrix were balanced using the RSA technique.

60. Allocating factor payments to households as above, however, resulted in inconsistencies between the household expenditure data and estimated incomes. Incomes of both urban households and households receiving remittances were too large relative to their expenditures, indicating either a relative under-reporting of expenditures by these households in the survey, a substantial change in the distribution of factor payments as compared with the 1984/85 SAM, or substantial inter-household transfers. In order to achieve a better consistency between incomes and expenditures by households, total consumption by urban households and all households with remittance incomes were each increased by 20 percent. These figures were chosen so that the share of non-remittance incomes in total income for rural households receiving remittances (94 percent) approximated the share in the 1984/85 SAM (91 percent).

"There is also food aid which is donatedby the rest of world and milled in Lesothobefore being sold on the open market. Carvahlo(1988) included this as a purchaseof grain by parastatals,but this does not then end up being milled. In the currentSAM donatedgrain whichis later milledenters as part of food imports. - 186-

Annex 1, Appendix A, Tablel: Activities in the 1987 Lesotho SAM

1987 SAM 1984/85 SAM 1. Field Crops 3. Field Crops (traditional) 4. Field Crops (improved) 5. FSSP Products 2. Vegetables I. Horticultural Crops 3. Fruits 2. Fruit 4. Cattle 9. Cattle Production 10. Cattle (feedlot) 5. Other Livestock 6. Poultry Layers 7. Poultry Broilers 6. Wool, Mohair Production 8. Wool, Mohair Production 7. Mining 46. Mining and Quarrying Production 8. Meat Processing 12. Butchery Products 9. Dairy 13. Dairy Products 10. Milling 18. Milling Products I1. Other Food Processing 17. Processed Fruit and Vegetables 22. Brewing Products (traditional) 23. Brewing Products (modem) 24. Other Beverages 12. Weaving, Leather 14. Processed Wool and Mohair 15. Weaving with Wool and Mohair 16. Leather Goods 21. Footwear 13. Handicrafts 25. Jewelry 26. Pottery 27. Other Handicrafts 14. Textiles 28. Other Clothing Goods 15. Agricultural Tools 19. Agricultural rools and Equipment 16. Other Mlanufacturers 29. Wood Products 30. Bricks 31. Steel or Metal Products 32. Other Manufactured Products 33. Fuel 17. Electricity, Water 45. Electricity and Water 18. Building, Construction 37. Building and Construction 19. Trade 20. Wholesale and Retail 20. Hotels, Restaurants 38. Catering (Hotels and Restaurants) 21. Transport 35. Air Transport 36. Road and Rail Transport 22. Private Services 34. Sales and Repairs of Autos 39. Financial and Insurance Services 40. Printing Services 41. Real Estate and Business Services 42. Commercial, Social and Personal Services 44. Communications and Posts 47. Domestic Services 23. Housing 43. Ownership of Dwellings 24. Public Administration 48. Government Services Source: 1987 LesothoSAM and Carvahlo(1988) - 187-

APPENDIX B

The Formal Sio Model'6

OBJECTIVES

61. This model aims to predict how all SAM accounts will change in response to a series of exogenous shocks. To do this requires several steps. First, all SAM row accounts must be written as linear functions of one another. Second, the modeler must make some decisions about what variables are endogenous to the system.

Appendix B Figure 2 describes the organization of the SAM and the variable labels adopted in this formulation.

Linearity Assumptions Production Accounts

(Xj). All production activities (Xj) produce a series of commodities (Dji) as outputs:

xi = EDp. (1)

Initially, assume that each activity produces a fixed share of each commodity's output. This assumption can be relaxed later:

Di, = d y Zi. (2)

'6Adaptedfrom Doroshand Haggblade, 1992. Annex 1, Appendix B, Figure 1 - SAM Structure

Activities Commodities Households Institutions Government Rest of Capital Row j h n World Sums

Activities Dji xj j

Commodities INTERMEDIATES FINAL CONSUMPTION Exports Invest.

Aij Cij Cih Cin | Cig Ei Ii

Households TRANSFERS TO HOUSEHOLDS h Vhj Rh Yh Bhh Bhn Bhg

00 Institutions TRANSFERS TO INSTITUTIONS n Vnj Rn Nn Bnh | Bnn | Bng

Government Tj Ti Th Tn Rg G

TRANSFERS TO R.O.W. Rest of World Mi R Brh Brn Brg

SAVINGS Capital SAVINGS Rk S Sh Sn Sg

Column Sums Xj Zi Yh Nn G R S - 189-

This leads to the following linear relationship between activity (Xj) and commodity (Z1) accounts:

Xj= Di= ,diZ (3) i f

62. Commodity accounts (Zn). Commodities supply their wares as inputs in production activities (A1j); add commercial margins to other commodity accounts (C33); supply households (Cih), institutions (Ci,), and government (Cig) with final consumption goods; and sell exports (E1) to the rest of the world and investment goods (I) to capital accounts:

Zi = Au + E Ci + E C,h + E C,n + Ci + E, + i. (4) j i h n

Assume that intermediates (Aj) remain a fixed share of gross output in each sector (Z1), the classic Leontief assumption:

Atf = au Xj. (5)

and that commercial margins (Ci) likewise remain a fixed share of final commodity value (Z1).

C i = Ci Z'* (6)

Assume also that household consumption (Cih) is a linear function of household expenditure (household income [Yh]less transfers paid [B.h] less direct taxes paid by the household [Th] less household savings [Sh]): 7

Cih = clh + flih(Yh - B-h Th - Sh), (7)

7 ' The notation*h indicatestransfers paid by households(h) to all sources(*), includingother households(h h), nongovernmentinstilutions (nh), andthe restof world(rh). - 190-

where

B.h Bh-h + B + B h ~~~~n that institutional (Cjn) and government (Cig) consumption of final goods and services remain a fixed proportion of final output,

Cin = chN. (8) Cig = cigG; and, initially, that investment remains fixed exogenously:

ii = I-L (9)

Substituting and combining like terms yields the following summation of the commodity accounts:

(J-Zcii)Zi Xi aXj + Zj,,,(YiThT Sh)+ Xaih (10) i j hI h + ci, N+ C,8G+ Ei + -i.

63. Houselhold accounts (Yh). Households (Yh) earn income from value-added payments by activities (Vhj)and from transfers they receive from other households (Bhh.), from institutions (Bhn),from government(Bhg) and from the rest of the world (Bhr):

Yh = Vi + , B.- + Z Bliw + Bhg + BAr (11)

As with intermediates, value-added payments to households (Vhj)remain a fixed proportion of activity output (Xj), while all transfers are taken as fixed:

Vhj = khjXji (12)

(13)

BA,, = Bh,h

Bi,m= Bhr - 191 -

Substituting Equations 12 and 13 into Equation 11 leads to the following relationship between household income (Yh) and activity output (Xj):

Y = h xhjXj + B,-,' (14)

where

h- ZBhh- + Bh + Bhg + Bh,- h

64. Institution (N). Institutions (N), like households, earn income from value-added payments by activities (Vnj) and from transfers they receive from households (Bnh), from institutions (Brn-), from government (Bng),and from the rest of the world (Bnf):

N = E + Bnh, + , A.h + Brig + B+nr (15) j h n.

As with intermediates, value-added payments to households (Vnj)remain a fixed proportion of activity output (Xj), while all transfers are taken as fixed:

V,= Xj, (16)

(17) B,,h Bh

B,ig Bug Btr= B,ir - 192-

Substituting Equations 20 and 21 into Equation 19 leads to the following relationship between institutions' income (N) and activity output (Xj):

N = ,X,,X. X + B,,h + B,,,, + Bng (18) J h

65. Government G. Government receives indirect taxes from activities (Tj) and commodities (T1) as well as income taxes from households (Th) and institutions (Ta) and transfers from rest of the world (Rg):

G = +±, , + ETh + ±T,,+ Rg. (19) J i h ?,

Assume that indirect taxes remain a fixed proportion of activity and commodity output,

Tj = tjXj, (22) Ti= t Zi; (23) that income and profits taxes increase linearly with household and institutions' earnings,

T = th + th, Yh,X (20) T= t) + t,,N; (21) and that transfers from the rest of the world remained fixed:

Rg = Rg (24)

Substitution among these relationships into (23) yields the following relationship between government revenuesand the remaining SAM accounts:

(25) G = JXj + Z ti i J +in + t n,,N+ Rg X i IF

= tj Xj + ti Zi + , th l h + t,, N + E ±t+ t, + R

66. Rest of the world (R). The rest of the world account receives payments for commodity imports (Mj) and receives transfers from households (Brh), institutions (Bm), and government (Brg):

R = XMi + B, + EBmI + B, . (26) i Ah - 193 -

Assume that imports (Mi) remain a fixed proportion of commodity output (Z1) and, as before, that transfers remain fixed:

M = m; Zi, (27)

(28) Brh' Brh 28 Bn= Bm

B,g =Brg

Total rest of the world earnings (R) then become

R = E>ms Zi + I r^ + Bm + Bog (29) i ,h

67. Capital (S). Total savings (S) equals the sum of savings by households (Sh), institutions (Sn), government (Sg), and capital transfers from the rest of the world (Rk):

S = LSh+ S,n + Sg + Rk- (30)

Assume that transfers remain fixed but that all domestic sources savings goes up linearly with their earnings:

Sh OhYhC + Sh' (31) S= a,,N + s,, (32)

Sg = gG + S-' (33) Rk = Rk* (34)

Total savings can then be expressed as a function of household income (Yh),institutional earnings (N), and government revenue (G):

S = cahYh + 01r3 N + cygG + Sh + Sn Sg +Rk (35) h h

68. Parameter restrictions. For the SAM to remain balanced, increased revenue must generate an equivalent increase in expenditure. For expenditures to equal revenue, the sum of all column coefficients in the M matrix must equal 1. - 194 -

In particular,

Eay+ Ekh, + X + Jj , for allj, (36) i h dji + Cii + ti + mi = 1 foralli, (37) i j

3 Z ih = 1 for allh, (38)

in + tn +

ECig + g 1,l;g 1 E Cig* (40) J i

69. System summary. Appendix B Table 1 summarizes this system of equations in matrix form. In doing so, it divides the commodity accounts into two groups, (ZI) and (Z2), depending on their supply elasticity. The first group, (Z1), represents all sectors with upward-sloping supply curves. In the Lesotho model, 11 sectors are modeled with inelastic supply, including field crops, livestock, textiles, and other manufacturing (Appendix B Table 3). For the remaining commodities (Z2), producers can increase output at constant unit cost. Hence, their supply is perfectly elastic.

Since the Z, commodities are tradable, the world price imposes the fixed-price requirement of tiheSlO model. The (Z2) commodities achieve fixed prices by virtue of their perfectly elastic supply.

Solving the Basic Model

70. The relationships defined above translate the SAM row accounts into a set of equations, one for each of the activities, commodities, houselholds,and other institutions in the SAM. Annex 1, Appendix B, Table 1: Model Equations

x- I -D, -D2 0 0 0 0 0 x 0

-A, I-Ci -C12 - j1[i-th-Gh] -Cloi, C1 g 0 0 z, Ei + I, + [a,-J3,K]i

-A2 -C21 I-C2 2 -fi21i-th-ah,] C2n -C2g 0 ° Z2 E2 + 12 + [a 2 -J32 K]i -AF, 0 0 1 0 0 0 y B. +Rh

-xn 0 0 0 1 0 00 N _AH ° ° ° I ° ° ° N ~~~~~~~~~~~~~~~~~~~BN+ R. tj' -tl' t2 -th' tn' 1 0 0 G T + T O -m, -m2 0 0 0 1 0 R _ _ _

o 0 0 -ah' -an -ag ° IS B, + B0n + B

Shi + So + Ss + Rk

Where K is a diagonal matrix with 7A + S%+ B,, as each diagonal element, i is a column vector of l's, I is an identity matrix. - 196-

71. In the standard input-output (10) model, exports, investments, and all transfers are exogenous, and the model solves for production, income and consumption as described in the equations above. In this formulation, where transfers and investiments are taken as exogenous, commodity export demand becomes the prime determinant of system change. Appendix Table B.2 displays the resulting input-output (10) model after differentiation.

72. The semi-input-output (SIO) model differs only in its choice of exogenous variables. Since (Z1) commodities are supply constrained, their output is fixed in a fixed-price world. So supply, not exports, become exogenous in these equations. With output fixed and domestic demand determined by changes in incomes, exports (El), rather than supply (Z), become endogenous in the equations. Appendix B Table 3 rewrites the SAM model in S1O form.

Extensions

73. Two extensions of the basic S10 model in Appendix B Table 3 are used in this paper: reducing urban demand linkages, and making investment endogenous.

Reduced Urban Demand Linkagec

74. Given the proximity of urban centers in Lesotho to the border, many "non-tradable" goods and services purchiased by urban households in Lesotho originate from the Republic of South Africa. These expenditures represent a leakage from Lesotho's national economy, necessitating a modification of the base S1O model. To account for these purchases, the marginal propensities to consume domestic non-tradables (bih's) are reduced by 70 percent for urban households. In order to maintain the adding- up condition for household expenditures (equation 38), the marginal propensity to consume imports from the RSA is increased to offset the change in the other bih's.

''In the modelsimulations presented in thispaper, governmentspending is alsoheld fixed (dG = 0). Thuschanges in tax revenuesdo not leadto changesin spending. - 197-

Endogenous Investment

75. To make investment endogenous requires adding additional equations to explain aggregate investment and distribution across commodities.

76. Using the accounting relationship that aggregate investment (1) equals aggregate savings, or equivalently, that changes in investment (dl) are equal to changes in savings (dS),

ZdI, = dI = dS., (41)

77. To explain how investors allocate their purchases across commodity groups, assume proportionality constant according to past investment shares (Cik):

d Ii = CikdI = CikdS. (42)

Since dI = dS, these relationships can be substituted directly into the commodity equations as described in Appendix B Table 3. Annex 1, AppendixB, Table2: Input-OutputModel

1 -D, -D2 0 0 0 00 x dX = dE, + 0 dE2

-A, 1-C,, -C12 -[l]-th- ah] - Cln C}g 0 0 dZ, 1 0 1 -A2 -C2, 1- C22 [1-th- ] - C2n -C2g 0 0 dZ2 0 1 xi,. O O 1 0 0 0 0 dY 0 0

-An 0 0 0 1 0 0 0 dN 0 0

-11, t2 -th1 to, 1 0 0 dG 0 0

0 -mi, -1r 0 0 0 1 0 dR 0 0

0 0 0 -Ch _n -Us 01 dS 0 0 Annex 1, Appendix B, Table 3: Semi-Input-Output Model

0° -D2 0 0 0 00 x dZI + O

-Ai -1 -C/2 -fifl -th-bh] -CIn -C/5 0 0 dE, -(I-C,,) 0

-A 2 0 1 C22 f82[I-Ithh] -C2n -C2g 0 0 dZ2 C2 1 I

-Xs 0 0 1 0 00 0 dY 0 0

-~k 0 0 0 1 0I0 0 dN 0 0

tjf 0 12' -th' -in I 0 0 dG tp 0 0 0 -mr 0 0 01 0 dR ml, 0

0 0 0 -ah -a,n -Tg 0 1 dS 0 0

To make investment endogenous, replace "0"with cl k + C2 k. - 200 -

Annex 2

LESOTHO POVERTY ASSESSMENT

1993 SURVEYDATA Procedures For Calculations of Expenditure Values

1. The 1993 household data survey contained information on both income and expenditure. Household income consisted of each member's revenues in the household from various sources, including wage work, farm products and so on. Expenditure was based on a few staple foods and some necessity non-food items (phoofo, moroho, sugar, fuel, water, travel) purchased in the last week, and expenditure on other non-food items (school, health, building, furniture, clothing) purchased in the last six months.

2. The data from both sources presented some problems. The expenditure data was incomplete and covered only a limited number of expenditure items, and some households did not report any expenditure at all. The income data was also incomplete, and the average household income was much lower than expected. Because the expenditure data appeared to be the more reliable (though still incomplete), we decided to use it to derive the level of total monthly per adult equivalent expenditure to be used as a basis for the measurement of welfare level.

3. The basic methodology that has been used to estimate the total household expenditLirein 1993 consists of two steps. First, we estimated the value of total household expenditure for those 186 households out of a total of 1719 for which the total expenditure level is less than M5 a month and/or food expenditure is zero. Next, we extrapolated the total expenditures values using the budget shares of the commodities reported in 1993 as they appeared for the same group of commodities in the 1986 expenditure data by quintiles. The methodology is briefly described below.

4. In order to estimate the value of expenditure for those 186 households (with missing/incomplete information), we calculated the predicted values from the results of an expenditure determination function. The results of the,equation are reported in Annex Table 1. The dependent variable is the log of total reported expenditure, which represents a measurement of household welfare level, and the explanatory variables are the characteristics of the household head, household composition variables, school level of the males and females and location specific dummy variables for ecological zones. Even though the purpose of the welfare determination function is simply to predict the value of missing observations, the ensuing results are as expected and confirm the validity of the prediction estimates and the quality of the expenditure data. In particular looking at Table 1, we find that most of the coefficients of the variable are significant at 5% level and the R-square is equal to .32. Among the variables relating to the household head we found that marriage has a strong positive impact on welfare. In addition, the non-residency status of the household head increases the probability for the household receiving remittances, and hence being better off. The household composition variables show that, as expected, welfare is positively correlated to household size. The education variables confirrn a strong influence of school level on the level of welfare, especially for female education. Finally, the analysis by ecological zones confirms that with respect to the lowland area (zone 1), households living in foothill, mountain, and Senqu River Valley areas (zones 2, 3 and 4 respectively) are in general poorer than households in Maseru (zone 5) and other urban areas (zone 6). - 201 -

5. The 1993 predicted monthly expenditure is reported in Annex Table 2. These expenditures only accounted for roughly half of the total household expenditure because of the limited commodities covered in the survey. The percentage of households consuming the given commodities are reported in Annex Table 3 and the budget shares are reported in Annex Table 4 (all for 1993). Even though the data is limited to a few commodities, it is evident that food captures the most important consumption element; in fact the budget shares of food represents 65.71 percent of total reported expenditure for the bottom quintile and decreases for higher expenditure quintiles to 34.93 percent for the top quintile (Annex Table 4).

6. To figure out the relationship between the commodities given in the 1993 survey and the total comprehensive expenditure of 1986, we first compared the expenditure level of 1993 with the same commodities reported in 1986. The percentage of households consuming the commodities, and the budget shares for the 1986 data are given in Annex Tables 5 and 6, respectively. Except for the first quintile, the food shares of the 1993 commodities are very similar to the data reported in 1986. Therefore it is fair to assume that the allocation of the expenditure between the commodities included in the 1993 questionnaire and the 1986 total did not change substantially in seven years for the same expenditure quintiles, thus enabling us to multiply the value of the total expenditure reported in 1993 by the inverse of the share of those commodities for each corresponding quintile. In Annex Table 6 the commodities in the 1993 survey represented 45.12 percent of the total for the bottom quintile, 55.38 percent for the second quintile and so on. The results of the extrapolations are reported in Annex Table 2. On the average total monthly household expenditure is M485.09 per month, while the per adult equivalent total expenditure that forms the basis of the determination of the poverty line is M140.96. - 202 -

Annex 2, Table 1: Estimation for Determinants of Total Expenditure for 1993 Survey Variable Coefficient t - values Female headed household (0,1) 0.037 1.70 Age of head of household -0.007 -2.17 Years of school of head of household 0.050 4.85 Head of household resides at home (0,1) -0.484 -7.28 Head of household is married (0,1) 0.380 2.14 Head of household is widowed (0,1) -0.037 -0.24

Number of males aged less than 7 -0.001 -0.03 Number of males aged between 7 and 15 0.147 5.02 Number of males aged between 16 and 34 0.094 1.19 Number of males aged between 35 and 59 0.259 2.69 Number of males aged over 59 0.326 2.65

Number of females aged less than 7 0.068 1.90 Number of females aged between 7 and 15 0.166 5.62 Number of females aged between 16 and 34 0.144 2.28 Number of females aged between 35 and 59 0.181 2.40 Number of females aged over 59 0.106 1.17

Number of males over age 15 with no formal education -0.009 -0.10 Number of males over age 15 with grades 1-6 -0.017 -0.21 Number of males over age 15 with grades 7 0.096 1.02 Number of males over age 15 with grades 8-1 1 0.157 1.60 Number of males over age 15 with grades over I1 0.236 1.82

Number of females over age IS with no formal education -0.151 -1.70 Number of females over age 15 with grades 1-6 -0.041 -0.62 Number of females over age 15 with grades 7 0.015 0.22 Number of females over age 15 with grades 8-1 1 0.164 2.17 Number of females over age 15 with grades over 11 0.346 3.26

Ecological Zone I - Lowland Ecological Zone 2 - Foothill -0.226 -3.06 Ecological Zone 3 - Mountain -0.321 -3.80 Ecological Zone 4 - S. River Valley -0.273 -2.97 Ecological Zone 5 - Maseru urban 0.175 1.80 Ecological Zone 6 - Other urban 0.224 1.87

Numberof observations=1532 R-squarc=0.3169Adjusted R-square=0.3028 Note: Includesonly householdsthat havetotal monthlyexpenditure of more than MS and valid food expenditure(greater than MO). - 203 -

Annex 2, Table 2: Predicted and Extrapolated Expenditure Levels in 1993 Quintile Quintile Quintile Quintile Quintile Poor Ultra- All 1 2 3 4 5 Poor PredictedMontly 35.20 79.36 155.04 290.89 708.09 70.34 41.97 254.04 Expenditure Extrapolatedtotal 78.02 143.30 269.31 544.33 1387.32 131.26 86.96 485.09 Expenditure Extrapolatedper 15.60 29.36 52.42 105.19 328.65 26.39 17.37 106.35 Capita Expenditure ExtrapolatedAdult 20.24 40.44 72.34 144.85 425.92 36.07 22.79 140.96 Equivalent Expenditure

Frequencies 62,209 62,284 62,313 62,224 62,480 153,197 80,398 311,510 Households 344 339 339 350 346 837 440 1718 Note:The sampleincludes all houscholdswith reportedor predictedexpenditure - 204 -

Annex 2, Table 3: Percentage of Households Consuming Reported Commodities in 1993 Quintile Quintile Quintile Quintile Quintile Poor Ultra- All I 2 3 4 5 Poor Commodity Cereals 74.94 65.78 64.88 75.83 78.22 68.91 73.91 72.06 Moroho 34.37 41.77 53.64 55.83 70.03 40.21 34.91 51.25 Sugar 28.81 33.21 56.43 62.68 78.11 34.84 28.49 52.06 Food_3 98.47 85.97 84.23 89.28 98.65 90.27 97.42 91.44 Clothing 16.89 30.60 43.02 61.79 70.38 25.90 19.38 44.77 Water 1.48 3.95 5.37 4.98 10.39 3.08 1.77 5.26 Fuel 24.03 30.91 46.63 60.04 75.99 29.68 24.77 47.72 Furnit 0.33 0.98 4.56 7.93 29.73 0.91 0.25 8.82 Build 4.59 17.76 35.07 46.83 65.60 14.67 6.51 34.17 Travel 4.59 17.76 35.07 46.83 65.60 14.67 6.51 34.17 Health 48.04 57.73 73.22 75.24 79.97 57.12 49.20 66.99 School 49.48 61.32 66.73 71.12 77.89 57.72 51.70 65.42 Other 79.14 91.89 97.42 99.08 99.71 87.44 79.81 93.53 Total 61,417 59,002 61,028 61,511 62,143 148,199 79,100 305,102 Population Total 341 321 331 345 344 811 434 1682 Households Note: Thesample includes only householdsthat report any expenditureat all. - 205 -

Annex2, Table 4: Food and Non-FoodBudget Shares Per Adult EquivalentQuintile in 1993 Commodity Quintile Quintile Quintile Quintile Quintile Poor Ultra- ALL 1 2 3 4 5 Poor Cereals 44.38 39.99 28.65 35.07 25.29 40.43 44.86 34.65 Moroho 11.17 11.10 11.08 6.03 4.61 11.42 11.05 8.67 Sugar 10.16 7.34 9.29 6.56 5.03 9.16 9.35 7.65 Food_3 65.71 58.43 49.02 47.66 34.93 61.01 65.26 50.97 Clothing 4.18 7.87 10.11 11.71 9.12 6.44 4.79 8.51 Water 0.22 0.50 0.83 0.46 1.67 0.40 0.20 0.75 Fuel 5.56 5.61 8.09 8.85 8.80 5.77 5.49 7.40 Furnit 0.13 0.08 1.09 2.46 9.86 0.21 0.11 2.91 Build 1.92 6.62 9.95 8.47 15.12 4.57 2.54 8.45 Travel 1.92 6.62 9.95 8.47 15.12 4.57 2.54 8.45 Health 7.13 5.86 6.92 5.20 3.88 6.72 6.87 5.76 School 15.06 14.40 11.56 11.68 7.78 14.31 14.60 12.02 Other 34.29 41.57 50.98 52.34 65.07 38.99 34.74 49.03

Total 100 100 99.999999 100 100 100 100 100

Total 60,476 49,826 50,727 54,675 61,301 132,209 76,883 277,005 Population Total 337 270 278 309 338 727.00 424.00 1,532 Households Note: The sample includes only households that have total expenditure of more than M5 per month and positive food expenditure. - 206-

Annex 2, Table 5: Percentage of Household Population Consuming the Following Items Item Category All Quintile Quintile Quintile Quintile Quintile Poor Ultra- 1 2 3 4 5 Poor All food & tobacco 99.13 95.92 100.00 99.80 99.96 99.98 98.24 96.39 Cereals 89.88 66.87 90.26 95.85 97.72 98.67 81.05 69.02 Phofoo 81.95 54.80 80.22 90.15 91.34 93.21 70.36 57.21 Other bread/cereal 64.88 36.39 56.23 67.97 78.54 85.26 48.99 37.37 Animal products 73.21 40.85 65.53 78.32 87.32 94.01 56.34 43.41 All fruit & vegetables 81.57 59.71 80.49 81.66 92.19 93.79 71.64 61.76 Fruit & vegetables 65.71 35.82 58.77 68.32 79.08 86.52 50.32 37.50 Moroho 69.38 48.09 68.53 69.05 79.28 81.92 59.91 49.98 Alcohol & tobacco 64.42 59.80 61.67 64.70 66.56 69.36 61.83 60.64 All other food 94.67 83.81 94.48 97.36 98.49 99.22 90.16 84.50 Sugar 79.33 60.44 77.31 82.46 87.63 88.80 70.55 61.89 Other food 91.59 76.65 90.45 94.47 97.73 98.64 84.97 77.52 Cloth & footwear 46.80 13.49 29.97 50.73 60.68 79.05 25.77 14.22 Blankets 10.40 1.66 4.26 8.74 15.01 22.30 3.38 1.83 Footwear 25.96 4.76 13.01 24.70 36.27 50.99 11.19 4.94 Other clothes 37.86 9.18 22.46 39.29 49.19 69.11 19.23 9.62 Rent and fuel 99.18 98.84 99.37 99.06 99.45 99.19 99.09 98.97 Gross rent 91.84 91.63 91.73 93.64 91.37 90.82 91.94 91.83 Water 1.30 0.21 0.53 0.97 1.61 3.20 0.41 0.20 Total rent 91.57 91.55 91.52 93.51 91.08 90.16 91.79 91.77 Misc. rent 0.45 0.12 0.03 0.37 0.34 1.40 0.15 0.11 Other rent 0.25 0.18 0.02 0.28 0.40 0.35 0.12 0.17 Fuel /power 93.18 86.35 93.59 94.39 95.92 95.66 90.71 87.15 Fuel 92 .99 86.35 93.59 94.39 95.83 94.79 90.71 87.15 Electric 0.53 0.00 0.00 0.07 0.49 2.09 0.00 0.00 Furniture 84.12 65.80 79.99 87.66 92.17 94.98 75.12 66.82 Transportation 37.49 15.40 28.19 37.34 48.49 58.00 23.47 16.33 Other cons. expend. 82.25 64.87 77.41 85.33 90.88 92.73 72.98 65.34 Medical 41.46 23.52 32.88 42.97 52.78 55.15 29.84 23.43 Education 28.81 18.10 23.67 31.44 33.89 36.95 22.29 18.25 Entertainment 7.84 2.02 4.86 6.71 9.21 16.37 3.89 2.05 Other educ/entertain't 0.89 0.50 0.23 0.56 1.15 2.02 0.41 0.44 Misc. goods 72.50 50.31 65.68 75.00 82.80 88.70 60.27 51.15 In-kind other cons. exp. 8.10 5.48 9.42 9.18 8.44 7.99 7.62 6.07 Cons. 86-93 99.24 96.47 100.00 99.83 99.93 99.97 98.48 96.87 Input expenditure 0.61 0.01 0.84 0.76 0.66 0.79 0.45 0.11 Animal Feed 6.75 1.90 3.48 5.28 9.03 14.06 2.92 2.03 other expenditure 4.59 1.99 2.29 4.03 5.21 9.41 2.42 1.86 Sample size 316311 63232 63265 63245 63222 63346 148090 71315 Note: Per adult equivalentmean expenditure=M81.58.Poverty lines dcfined as: for poor =50% of themean (M40.79), Ultra-poor=25% of mean (M20.40) ' For householdswith valid total consumptionvalue. - 207 -

Annex 2,Table 6: Percentage Share of Household Expenditure by Given Categories Item Category All Quintile I Quintile 2 Quintile 3 Quintile 4 Quintile 5 Poor Ultra- Il l l lI | | Poor All consumption expend. All food & tobacco 48.59 44.42 55.89 54.75 50.21 37.69 51.02 45.65 Cereals 22.39 18.17 27.65 28.55 22.72 14.87 23.75 19.09 Phoofo 20.23 16.06 25.29 26.27 20.45 13.11 21.52 16.95 Other bread/cereal 2.16 2.11 2.36 2.29 2.27 1.77 2.23 2.14 Animal products 6.05 4.08 5.87 6.28 7.38 6.66 5.18 4.33 All fruits & vegetab. 4.63 5.11 5.34 4.60 4.63 3.48 5.15 5.18 Fruit & vegetables 1.99 1.57 1.85 2.05 2.34 2.13 1.75 1.61 Moroho 2.64 3.55 3.49 2.55 2.29 1.35 3.40 3.57 Alcohol & tobacco 4.10 5.49 4.47 3.21 3.80 3.52 4.76 5.51 All other food 11.41 11.57 12.56 12.11 11.68 9.16 12.18 11.55 Sugar 3.83 5.16 4.40 3.93 3.45 2.21 4.70 5.09 Other food 7.58 6.41 8.16 8.18 8.23 6.94 7.48 6.46 Cloth & Footwear 11.67 3.60 7.45 12.33 15.59 19.38 6.57 3.78 Blankets 2.17 0.41 1.10 1.86 3.40 4.08 0.86 0.49 Footwear 3.66 1.46 2.66 4.06 4.80 5.33 2.45 1.52 Other clothes 5.84 1.73 3.68 6.41 7.38 9.97 3.26 1.78 Rent and fuel 18.54 35.49 19.75 14.77 13.26 9.43 25.77 34.17 Gross rent 12.38 27.33 12.29 8.82 7.06 4.91 18.69 26.51 Water 0.03 0.02 0.02 0.02 0.05 0.07 0.02 0.01 Total rent 12.32 27.80 12.83 8.84 7.27 4.88 18.66 26.49 Misc. rent 0.02 0.00 0.00 0.02 0.00 0.06 0.00 0.00 Other rent 0.00 0.00 0.00 0.02 0.00 0.01 0.00 0.00 Fuel /power 6.15 7.66 6.89 5.86 5.91 4.41 7.08 7.65 Fuel 6.08 7.66 6.89 5.85 5.83 4.16 7.08 7.65 Electric 0.07 0.00 0.00 0.00 0.09 0.26 0.00 0.00 Furniture 7.75 6.84 6.17 6.35 7.36 12.01 6.46 6.71 Transportation 3.24 1.64 2.86 3.29 3.63 4.78 2.34 1.72 Other cons. expend. 10.23 8.02 7.88 8.53 9.95 16.73 7.85 7.98 Medical 1.56 1.58 1.52 1.39 1.58 1.71 1.50 1.51 Education 1.31 0.88 1.06 1.51 1.42 1.67 1.01 0.84 Entertainment 0.49 0.13 0.25 0.31 0.49 1.25 0.20 0.14 Other ed/entertain't 0.03 0.02 0.00 0.00 0.03 0.09 0.01 0.01 Misc. goods 6.16 4.38 4.05 4.65 5.98 11.71 4.15 4.42 In-kind other cons. exp. 0.69 1.04 0.98 0.66 0.46 0.29 0.98 1.06 86 (93) food exp 26.70 24.77 33.18 32.75 26.19 16.67 29.62 25.61 86 (93) other exp 25.81 20.35 22.20 24.82 27.25 34.37 24.98 22.22 86 (93) all exp 52.51 45.12 55.38 57.57 53.44 51.04 54.60 47.83 86(93) Food / All 54.90 59.91 56.89 49.01 32.66 Sample Size 316,311 63,232 63,265 63,245 63,222 63,346 148,090 71,315 Note: Per adult equivalent mean expenditure=81.58 M. Poverty lines defined as: for poor= 50% of the mean (M40.79), Ultra=25% of mean (M20.40). * For households with valid consumption value 86 (93) food expenditure to match withm93, includes only phoofo, moroho, and sugar 86 (93) other expenditure to match with 93, includes cloth fuel, water, tran., educ., medicine, and furniture 86 (93) all exp, is food+other exp (animals and building exp in 93 unclear) - 208 -

Annex 3

LESOTHO POVERTY ASSESSMENT

POVERTY MEASUREMENT Methodology for Adult Equivalence Calculations

I. Having chosen total consumption (expenditure) to measure the welfare of Basotho households, there remains a need to adjust for household size and composition. The adjustment is done by converting the household size to an "adult equivalent" size. The methodology for calculating the adult equivalent and the justification for using it is given below.

2. We have chosen to calculate the adult equivalence measure as described by Collier et al. (1986).1 This methodology is used in Tanzania poverty profile study as well, and as compared to some other measures utilized in various studies, we find Collier's procedure to be the most robust one.

3. There are two adjustments to take into account when computing the adult equivalence. The first is the economies of scale. This is illustrated by considering for instance that food expenditure does not triple when household size increases by three-fold; the increase in food expenditure is less than the increase in family size. This is corrected for by using the Engel food scales estimated by Deaton (1980)2 and presented in Annex Table 1. The second adjustment involves family composition vis-a-vis gender and age,3 which is accounted for by using age and gender differentiated caloric requirements for (Eastern) Africa given in Annex Table 2.

4. The actual adult equivalence calculation involves the following: each household member is assigned a caloric index based on its gender and age using Annex Table 2, and this is summed within households. These values are then multiplied by the respective average cost factor (Annex Table I) for household size adjustment to get the adult equivalent.

5. We have used the adult equivalent measures to deflate household consumption, on which we have based our poverty measure assessment. Since the per adult equivalent consumption takes into consideration the above mentioned adjustments, it compares the welfare of households better than the per capita consumption (Muellbauer 1974),4 which takes into consideration head counts only.

' Collier, P., S. Radwan,and S. Wangwe,with A. Wagner 1986, Laborand Povertyin Rural Tanzania: Ujamaaand Rural Developmentin the UnitedRepublic of Tanzania, A Studyprepared for the InternationalLabor Office within the frameworkof the World EmploymentProgram, (ClarendonPress, Oxford).

2 Deacon,A. 1980, Economicsof ConsumerBehavior, (Cambridge University Press, Cambridge). For example, a teenage boy's caloric requirement is different from that of a 60 year old man. Muellbauer, 1974, "Inequality Measures, Prices, and Household Composition," Review of Economic Studies, 41:493-504. - 209 -

Annex 3, Table 1: Engel Food Scalles Cost Householdsize Marginal Average (numberof adults) 1 1.000 1.000 2 0.892 0.946 3 0.798 0.897 4 0.713 0.851 5 0.632 0.807 6 0.632 0.778 7 0.632 0.757 8 0.632 0.741 9 0.632 0.729 10and up 0.632 0.719 Source:Deaton( 1980)

Annex 3, Table 2: Index of Caloric Requirements by Age and Gender:Eastern Africa Gender Age group Male Female

0 - 2 0.40 3 - 4 0.48 5 - 6 0.56 7-8 0.64 9- 10 0.76 11- 12 0.80 0.88 13- 14 1.00 1.00 15- 18 1.20 1.00 19- 59 1.00 0.88 60 + 0.88 0.72 Source:Collier et al. (1986). - 210 -

Annex 4

LESOTHO POVERTY ASSESSMENT

Alternative Poverty MeasuremientBased on Minimum Per Adult Caloric Requirement

1. An alteniative povertymeasure can beattained using a calculationbased on a minimum caloric requirement, 2,500 calories, that one needs to lead a normal healthy life. We have chosen some basic food items that comprise of a staple food basket for Lesotho. The other staple, green leafy vegetable (moroho), is usually collected or home grown, and thus is excluded from the list (Annex Table 1).

Annex 4, Table 1: Minimum Per Adult Caloric Requirement (1993) Food Type Adult equiv. Adult equiv. Price/1S. Price/ks, Monthly daily monthly 1990 1993 Expenditure requirement requirement(kg) (M)

Cereals 500 15.00 0.76 1.16 17.43 Pulses 60 1.80 2.18 3.32 5.97 Sugar 50 1.50 1.67 2.54 3.81 Oil 40 1.20 4.96 7.54 9.05 Total monthlyexpenditure to meet 2,500caloric requirement 36.26 Notes:a/ ConsumerPrice Index (CPI). April 1990.Bureau of Statistics,Lesotho. b/ Inflationadjustment of 150/o/annum.

2. Based on the above calculation, one needs to spend M36.26 to acquire the necessary caloric requirement. For the report, we have used poverty lines of M70.48 for the poor, and M35.24 for the ultra-poor (50% and 25% of the mean adult equivalent expenditure respectively).

3. Considering the underestimation of expenditure data in the survey (discussed in the report), the ultra poor do not spend their total budget on food, as may have been implied above. On the other hand, the poor (unlike the ultra-poor who consume only the staple), usually vary their diets (may include meat during festivities and social ceremonies); thus, the budget share for food is higher than what is given above. Therefore, poverty line cut-offs used in the report are applicable, if one prefers to base the poverty lines on caloric requirements. - 211-

Annex5

LESOTHO POVERTYASSESSMENT

PovertyMeasurement Indices

I. Oncea povertyline is established,a set of measuresfor the degreeof povertycan be calculated. These shouldinclude the incidenceof poverty(number of peopleliving belowthe povertyline), and the intensityof poverty(the extentto whichthe pooresincome lie belowthe povertyline). In addition,the povertymeasure should be decomposableto allow for analysisof sub-groupsof the populationand their contributionto poverty.

2. An indexthat is widelyused and that incorporates the issues raised above is givenby Foster,Greer andThorbecke (FGT) (1984). The general formula for theFGT poverty index is:

P =_ * ( Y ) q Z

Where: Yi= Incomeof personi Z = Povertyline n = Totalnumber of individuals q = Numberof individualsearning an income below the poverty line

The indextakes the poverty gap of eachpoor person as a fractionof the povertyline, raises it to a powera. It is thensummed over poor people, and then averaged out for theentire population. One useful property of the FGT indexis that it allowsfor severalinterpretations as the parametera is varied,to reflectpoverty "aversion".

* for a=O: P0 is thehead count index (q/n),which measures the percentageof peoplebelow the povertyline, i.e. the poor.Note, this measuredoes not takeinto considerationthe extent of the person'spoverty.

* for a=I : P, is the average poverty gap, or averageincome deficit of the populationas a proportionof the povertyline. This measuretakes into accounthow pooron averagethe poor are, andreflects both the incidenceof povertyand its intensity.But Pa=l, is insensitiveto incomedistribution within the poor;if for instancethere had been an incometransfer from the poorestto thericher household within the poor, Pa=1 does not change.

* for a=2: P2 is theseverity index, whichweights households according to theirshortfall below the povertyline; that is, largeconsumption shortfalls, are weightedmore heavily than small ones. - 212 -

Using all three indices gives a fuller view of poverty and the variabilityin well-beingamong the poor householdsthan if one were limitedto singlealternative.

3. Anotherattractive featureof the FGT index is that it is additivelydecomposable (Kanbur, 1987). Thus, it is possibleto generateboth overallnational indices, and individualcomparable measures attributed to various(regional, and socio-economic)sub-groups of the population.

Inequality Measure: The Gini Coefficient

4. The Gini Coefficient is an inequality index that measures the degree of equity in a given population. In the case where everyone has an equal income, the Gini coefficient is zero (absolute equality), and on the other hand, if only one person has all the income,the Gini coefficientis 1 (absolute inequality).

The Gini coefficientis derivedfrom the followingmathematical formula:

Gini(G) I 2NY X Y,, - Y.

i 2 where, N = number of persons (or households) Y = incomeof individuali.

The Gini Index is calculablewhen the coefficientis multipliedby lOO.Thegraphic representationfor the coefficientis given by the Lorenzcurve line, wherea 45 degree line signifiesperfect equality. -213 -

Annex6

LESOTHO POVERTY ASSESSMENT

Determinants Of Poverty

I. A logistic regressionmodel hasbeen used in chapter11 to determinethe main characteristic of poverty. In particular,the purposeof the model is to determinethe factors that explain the probability (i.e. prevalence) of nQtbeing poor. The model used can be specified as follows:

Ln (P(Y=l)/(I-P(Y=I)) = Z = a + Bx where the outcomeis equal to one (Y= I) if the householdper adult equivalentexpenditure level is abovethe poverty line and zero (Y=O) otherwise. The vector of explanatoryvariables, x, includes characteristicsof the householdhead, the numberof people in eachage andeducational level, size of agricultural land and locationin any of the ecologicalzones.

2. To facilitate the interpretation of the results of the model the odds ratios and the impact on the probability of w=tbeing poor were calculated.The odds ratio explains the impactof the change in a dependent variable on the odds of not being poor,while the impacton probabilityevaluates the changein the probability of nQt being poor for a unit changein one of the variables when all the other variables have beenevaluated at their meanvalue. - 214 -

Annex 7

LESOTHO POVERTY ASSESSMENT

1993 SurveyData

Annex 7, Table 1: Correlation of Non-Explanatory Variables with Poverty Variable Rural Urban

Poor Ultra-poor Poor Ultra-poor Household accommodation Rondavels 0.1676 0.1486 0.1100 0.0815 Flats -0.2501 - 0.1962 0.0363 0.0145 Heisi 0.0212 0.0136 0.0280 - 0.0267

Optakas - 0.0964 - 0.0675 - 0.1104 - 0.0464 Source of water Pipedon site 0.0086 0.0127 - 0.1716 - 0.0724

Private borehole - 0.0526 - 0.0295 - 0.0806 - 0.0945

Communal pipe - 0.0741 - 0.0833 0.1125 0.0365 Communal hand pump -0.1150 - 0.0742 0.1061 0.0836 Covered spring 0.0443 0.0158 - 0.0231 - 0.0300 Other spring 0.1211 0.1012 0.0241 0.0514

Rainwater tank - 0.0033 - 0.0038 - 0.0359 - 0.0222 River 0.0105 0.0274 Dam - 0.0359 - 0.0222 No toilet available 0.2550 0.2119 0.2366 0.0923

Radio - 0.2886 - 0.2006 - 0.1633 - 0.1766 Crowding more than 4.5 people/room 0.0912 0.1050 0.1320 - 0.0102 - 215 -

Annex8

LESOTHO POVERTY ASSESSMENT

SunmmaryOf Qualitative Poverty Assessment Classification, Identification and Coping Strategies

I. Two surveys on perception of poverty were carried out by Sechaba Consultants during 1991 and 1993. The 1991 survey included Basotho perception on poverty. In addition, 40 cases of in-depth interviews with people identified as poor (by other community members, or by housing conditions), in both rural and urban areas were conducted separately. The survey provides a view of the people's perception of who the poor are, why they are poor, and the coping mechanisms utilized.

2. For the 1993 study, a qualitative (participatory) study of poverty was undertaken to gain people's perspective on poverty, coping strategies, and alleviation of poverty. Poverty alleviation was analyzed at various levels, including from the perspective of state and NGO's services and projects (including health, agriculture,and education), wage income (both from South Africa through remittances, and Lesotho), village based self-help groups, as well as inter-household and individual assistance and efforts.

3. The 1993 survey first attempted to find people's perception about poverty and aspects of local poverty by using a list of 42 factors that were identified through discussion among researchers (including local), and some villagers through field tests. The factors included such terms such as hunger, unemployment, destitution, drought, alcoholism, theft, misery, poverty, conflict, and disease. Each factor was thought to be related to poverty either as a cause or a result. Sixteen villages were selected at random from those where the quantitative survey was carried out. Households were selected in these 16 villages which, by design, contained the four ecological zones: urban, lowland, foothill, and mountains. In these villages, households who participated in the quantitative survey were selected whenever possible. The study was carried out 85 times.

4. The identified terms were written in Sesotho on index cards, which were then spread out on a table in no particular order in front of the subject(s) to be interviewed. Once each card was read out loud and digested, the participants were asked to group the cards in what they consider a logical sequence, from their experience. They would then relate the significance of each factor to their lives, and why they grouped the cards a certain way. Onlookers ofeten joined the discussion, though the subject ultimetely chose the answer.

5. Usually, 6-10 groupings would be formed out of the 42 cards. Normally, a leading card would be picked up, and cards that are seen as the effects of the leading card would be added to the pile. More ususally came come up, when the effects were, in turn described as causes of other factors. There were some cards that were omitted since the subjects could not identify how they might fit with other cards, while some cards did not strike the participants as being related to any of the formed groups (omissions comprised less than 20%). Some cards were selected very often, and in some cases, more than once in that they fit into more than one group (hunger and unemployment for instance). Next, solutions to poverty based on tracing cause and effect relations among the cards, as identified by the participants, were discussed. The last stage involved the ranking of the cards in order of importance and significance. - 216 -

6. To measure the relative importanceof the factors, threemeasures were considered; the number of times each card was chosen as part of a group, number of times the card is ranked, and the mean rank of the card. Each measure for individual factors score ranged from 0 to 100. Looking at the combined rank; Hunger, unemployment, destitution, and lack of clothing were the four top scorers, while low-class jobs, disaster, stupidity, and smoking/marijuana ranked last.

7. A cluster analysis was then carried out, based on the numbers of occasions in which people placed two cards in the same group. The process identifies the most closely linked pairs in the list of factors. For instance, the most closely linked pairs included injustice and oppression, and alcoholism and adultery. The cluster analysis summarizes how the participants view poverty into two broad groups; (I) Lack of basic needs (employment, drought etc.), and (2) pre-conditions of poverty, which is further sub- divided into subjective personal failings and objective structural problems (included here are politically- based poverty, disasters, and homelessness). The 1991 survey, conducted differently, nonetheless identifies lack of basic essentials, food and clothing, as the top concern, mentioned by 85% and 76% of the interviewees respectively, followed by (physical and mental) health, shelter, lack of livestock and employment.

8. Finally, causes and effects of poverty wvereclassified. Alcoholism, unemployment, and drought topped the list as causes of poverty-related factors in 1993, while, poverty itself, hunger, and theft were given as the leading effects. For the 1991 survey, unemployment topped the causality list (with 70.90/o mentioning it), followed by environmental disaster, and lack of fields/agriculture inputs. Thus unemployment has remained the leading cause of poverty according to this survey. Due to lack of mining jobs, which are the major income sources for Lesotho, the problem is intensified. Loss of income earning capacity, whether due to the loss or disablement of the income generating individual, and lack of social security system, causes a household to go from a relatively well-off position to poverty. The 1993 survey distinctly shows that households with migrant income are considerably well off in general, signifying the importance of mining in the economy vis-a-vis poverty.

9. The second major issue that the 1993 qualitative survey identified concerned coping mechanisms. For the 1993 study, 201 households from the same 16 villages where the cards (poverty perception) study was carried out, were interviewed on coping strategies. Of those households, interviewers assessed 118 as non-destitutes and 83 as destitutes based on housing conditions, furniture, clothing, food availability, and lack of paid employment.'

10. Households identified as destitutes were asked a number of detailed qualitative questions to learn about coping strategies. Sources of help were classified under four headings; relatives, neighbors, self, and institutions (government, church, etc.). They were also asked what kind of help was supplied (for instance, food, cloth, furniture, house, money and hope!). In addition, other sources of cash from informal sources were asked about (infonnal business, share cropping, odd jobs, etc.). On the other hand, the non-destitute households were asked who they help (relatives, neighbors), why they help (poverty, disability), type of assistance given (food, money, employment), for how long (from a year to over 10 years), and how often (ranging from rarely to weekly) they give help. They were also asked if there is an alternative in case they stop the support, and what suggestions they would offer to the needy households. In both destitute and non-destitute households, the principal solution was perceived to be the provision of jobs (36%).

' Onlya fewhouseholds overlapped between the cards study and the copingstrategies study (37 households),although the villages selected were the same. The uniformity is not an issue here, since the responses to the cards were not different for the different regions or wealth-based subgroups, but represented a broad consensus. - 217 -

11. For the 1991 study, coping mechanismswere identified for the 40 cases where in-depth interviews were conducted. The rural poor subsist by collecting shrubs for sale, cutting grass for thatching,doing piecejobs (weeding),participating in food-for-work projects,and obtaining gifts of food or materialsfrom neighbors. The urban poor, who mostly consistof malejob seekers(some for mining recruitnients) from the rural areaslive off their savings,do odd jobs when available. and some affiliate themselvesto the Shebeens,iiiformal liquor housesprimarily run be women, as watchmen or collect cardboardfor fuel, in return for shelterand food.

12. The next study componentin the 1993survey, which also included some commentsfrom other surveys carried-out by SechabaConsultants, addressed the impact/applicationof outside interventions. The projects discussed included, IFAD supported Local Initiative Support Project (LISP), rural employment generation(through work on roads), Thaba-TsekaRural DevelopmentProject, Renewable Energy Technology Project, population control, environmental protection, health, and modern technologyprojects.

13. From the above topics, key issuesexplored were identified as:

Food production/agriculture: The questionsraised here included farmers' reasonsfor switching crops,constraints faced in diversifying, why maizeis the staplecrop. and risks faced by them. In conjunction with the above, efficiency of the marketing/supplyof inputs including view of coops,and traders,and the adequacyof servicesand prices in general,was addressed.Another point specified was people'sview of food aid, and whether it has efficiently targetedthe needy. Finally, in order to learn from the successstories, we looked at what factors have contributed to these achievements,and whether it is successfullyswitching to cashcrops or adopting modem technology.

Education: This coveredissues ranging from the managementof schools/school boardsto why some studentshave problems in the system resulting, for instance in dropouts, repeatersand over-agedstudents.

Health: Services' performance,and the successof family planning efforts (includingattitudes were someof the points raised.

Local government: People's attitudes about government's effectiveness and the effects of training were addressed.

Social services: The main focus was on the existing infrastructure,especially in the urban and peri-urbanareas. - 218 -

Annex 9 LESOTHO POVERTY ASSESSMENT

Efficiency Comparison Betveen Lesotho Maize Mills (LMlM:)and Village Mills

Annex 9, Table 1: Lesotho Maize Mills: Operating Costs: 53 Weeks Ending 03/04/93 Production(tonnes) 64,028 Maize meal 1,228 Samp.and grits 20,328 By-products M/Tonne Al Maize mealonly products M/Tonne M Sales 696.49 59.608,821 814.82 52,171,419 Cost sharesby product 0.88 Cost adjustment* 7.35 789,108 690,651 Salesless cost adjustment 689.14 58,819,713 804.04 51,480,768 Lesscost of maizegrain 461.46 39,520,254 161.46 34,589,306 Lesspacking material 21.29 1,816,745 24.83 1,590,069 Lesscost of delivery 41.20 3,526,214 48.20 3,086,248 Grossprocessing margin 165.19 13,956.500 190.78 12,215,145 Lessoperating expenses 99.05 8,453,787 115.56 7,399,007 Profit before interest 66.14 5,502,713 75.22 4,816.139 Add interest 17.67 1,508,266 20.62 1,320,079 Profits 83.81 7,010,979 95.84 6,136,218 Notc: I/ Costadjustment include rebates, cash discounts allowed. transp. recovery, and stock adjustment.

Annex 9, Table 2: Comparison Between Lesotho Maize Mills and Village Mills aJ LMM Village Mills M/Tonne

Maize grain 461.46 461.46 Packingmaterial 24.83 24.83 Delivery/distribution 48.20 48.20 Operatingcosts b/ 115.56 94.00 Profits 95.84 Total 745.89 628.49 Notes:W For purposes of comparison,costs of maizegrain, packing material and delivery/distributionhave been assumed to bethe same for villagemills asfor theindustrial mills. b/ Operatingcosts of villagemills includeprofits. - 219 -

Annex 10

LESOTHO POVERTY ASSESSMENT

Welfare Inpact of Maize Pricing, Trade and Marketing Policies

1. Under the current trade and marketing regime, a limited number of large-scale, commercial agents import maize into Lesotho directly from the South African Maize Board. No import tariff or subsidy is applied, which implies that the licensed importers acquire maize at border prices plus transportation costs. Hence, one can distinguish two distortions which may affect consumer prices for maize in Lesotho:

1. inefficiencies embodied in the border price; and 2. inefficiencies originating in domestic marketing and processing.

2. South African distortions. South African agricultural policies have substantially subsidized white, commercial producers at the expense of consumers. Large-scale, capital-intensive farming although economically inefficient given the increasing labor surplus in the economy was heavily subsidized by the South African state in order to guarantee sufficiently high incomes to full-time, white farmers.

3. An important mechanism through which such implicit subsidies are transferred from consumers to producers is constituted by single-channel marketing system, such as the one in place for maize. Under this system, the South African Maize Board is the sole buyer, seller and exporter of South African maize. In this monopoly situation, South African maize producers are protected from international markets by import quantity restrictions, which, given Lesotho's membership in SACU, also applies to Lesotho. The Maize Board dictates high domestic prices (i.e. above international prices). Given the inelasticity of demand for maize, such price distortions hurt consumers substantially, in particular the poor (the black population consumes 94% of maize meal and 47% of this group is below the poverty line). In good years, the system artificially creates domestic surpluses, because domestic markets are not allowed to clear through a market price mechanism, and exports are only realized at a loss. However, Lesotho does not benefit from such export dumping, since it is a member of SACU. Hence, Lesotho's consumers are unambiguously hurt by the South African monopoly since the artificially high domestic prices are directly transferred to Lesotho.

4. An impression of the magnitude of the South African consumer losses incurred as a result of the marketing and pricing policies of the South African state is given in a recent analysis by Wright and Nieuwoudt (1993). Maize (white and yellow) producer benefits totalled R384 million or an average of R23,000 per farmer per year. In 1990, consumer losses (white maize only) amounted to R555 million per annum or R26 per year per consumer.

5. Given the trade regime in place between South Africa and Lesotho, these consumer losses are directly transferred to Lesotho consumers and should be added to any consumer losses originating from Lesotho's domestic policies. However, since local production accounts for about 30-35% of total consumption in Lesotho, open market prices in Lesotho at which all domestic production should be

' P.D. Wrightand W.L. Nieuwoudt, 1993. "Price Distortions in theSouth African Maize Economy: A ComparativePolitical Analysis", Agrekon 32(2):5 1-59). - 220- valued, will temper this loss somewhat,assuming that they are generally below the retail prices of imported maize.

6. Domestic distortions. In Lesotho,distortions around the competitive equilibrium arise because of the import ban on milled maize, a restriction which is sometimeslifted in times of drought, the monopoly on maize grain importation and pricing, and the oligopoly and subsidizationof large retailers. In the following, only the welfare implications of the monopoly situation in which the mills find themselvesare analyzed. The inefficiencies arise from the mill's monopolistic profits and inefficient operations, as compared to the profits and operating costs incurred by private millers in Lesotho's villages.

7. The domestic price representsthe actual, nominal price at which Lesotho Maize Mills sells maize meal to its clients. In 1992 this was around M745.89/tonne(Annex 1). The implied "tariff" is calculated by comparingthe processingprofits and operating expensesof the LesothoMaize Mills with the observedprocessing profits and operating expensesof the private village mills -- the Roller mills. Even though one would assumethat maize processingshould exhibit economiesof scale,the calculations suggestthe reverse:the small-scaleRoller mills seem to operatemore efficiently than the large-scale mills. The existence of these differences suggestsan efficiency loss, or an implicit "tariff" on consumers. Using the observedmilling feesof the Roller mills and applying them to the cost of maize grain purchasesfrom the South African Maize Board, one obtains the implicit "border" price, which is lower than the current domesticprice.

8. Given this implicit tariff imposition, consumersreduce their consumption, while production presumably expands due to the higher domestic price. Hence, there is a transfer of some of the consumers'surplus to producersand the mills. Taking all such transfers into account,one obtains the country's net loss of imposingthe implicit tariff.

9. Domestic maize producersbenefit from the distortion, which actually functionsas an instrument of protection. The Nominal ProtectionCoefficient (NPC) (Annex I) is the ratio of the domesticprice that producersface in the presenceof the distortion and the price they would have faced in the absenceof the distortion. The NPC is found to be 1.19. Recall, however,that the marketingand processingsystem certainly exerts an upward pressureon open market producer prices, but that it would be erroneousto value all domestic productionat the distorted domesticprice. In other words, the NPC calculatedhere should be seen as an upper bound. In the long run, one should question the economicefficiency of devoting agricultural resourcesto maize production,instead of morecompetitive crops. - 221 -

Annex 10, Table 1: Lesotho:Welfare Analysis of MonopolisticMaize Trade and Processing Variable Label Formula Value

Parameters Pd domesticprice of maizemeal (mill 745.89 gate) Pb "border"price of maizemeal (mill 628.49 gate) nd priceelasticity of -0.513 es priceelasticity of supply 0.151 NPC NominalProtection Coefficient Pd/pb 1.1868 t implicittariff NPC-I 0.1868 t' tPb/Pd 0 1574 V. valueof domesticproduction at Pd*Dom.prod. 126.4 domesticprice W. valueof domesticconsumption at Pd*TotalSupply 232.4 domesticprice Analys__ NELp deadweightloss in production 0.5*estV'2*V' 0.24 NELc deadweightloss in consumption 0.5*nd*t'A2*W 1.48 WGp changein producersurplus t'V'-NELp 19.66 WGc changein consumersurplus -(t'W+NELc) -35.11 "GR changein mill revenue t'(W'-V') 16.68

Source:World Bank statf calculations, Annex 9. Notes:Prices are per tonne. Priceelasticities are taken from Wright and Nieuwoudt, 1993. Valuesin M million.

10. Deadweight losses are estimated at MO.27 million in production and MI .48 million in consumption,yielding a total deadweightefficiency loss of M 1.75million. l1. The following welfare gains and losseswere figured as follows:

* producersgained less than M19.66 million (recall remark madeabove); * consumerslost M35. I1 million by consumingless maizeat a higher price--this amountsto a loss of M2 I per capita;and * mills gainedas profits or absorbedas inefficiercies M16.68 million. - 222 -

12. Environmental impact. As mentioned above, high consumer prices for basic foodcrops in South Africa were passedonto Lesotho consumersthrough food imports from South Africa. However, in Lesotho, inefficient marketing and processingstructures also addedextra costs, and Basothoconsumers were thus doubly hurt. Not surprisingly, given such high prices, the implicit value to the Basotho household of own-producedand consumedbasic foodstuffs, in particular maize, was made artificially high.

13. The fact that maize occupiesabout half of scarce,arable land is thus easily explained. Maize is even planted in the mountainareas, although maize is agronomically unsuited for this area. At the same time, Basotho householdsface a labor constraint due to the high incidence of labor-migration and adopted labor-extensive methods to produce basic food grains. These methods have harmful environmental effects on already over-crowded and scarce arable lands. The extensive methods in crop- cultivation employed have been a major contributing factor to soil-erosion in the country.

14. The environmental impact of low-input, extensive methods of maize production, as seen in increased soil depletion, may be one of the causesof the observed long-tern decline in grain yields. Other factors play a role, suchas increasedscarcity of labor, increaseduse of crop residuesas fodder for livestock, and the collection of cow-dung to meet householdenergy-requirements.

Policy Implications

15. Consumers. Increasing competition in the trade, marketing and processingenvironment will lead to a decreaseof consumerprices. This constitutesan unambiguouspro-poor policy.

16. Producers. Normally, a reduction of real consumerprices will exert a downward pressureon producer prices, thereby reducing producer incentives for the production of maize. The decreased attractivenessof maize production may representsubstantial benefits to producers in the long run, by allowing the reallocation of productive resourcesto agricultural activities for which Lesotho holds a comparativeadvantage.

17. Employment. An increasein small-scalemilling activities will result in a substantialincrease in employment given the increase in small-scaleentrepreneurs and jobs for mill operators. Moreover, indirect linkagesof suchmilling to the restof the local economyare also relatively high, and the indirect employmenteffect will be significant. - 223 -

Annex II LESOTHO POVERTY ASSESSMENT

Diversifiying Crop Production

1. An important way of dealing with Lesotho'senvironmental problems is to relieve pressure on the land by diversifying cropping agriculture. The aim would be to facilitate a significant expansionof exportablecommodities (both "raw" and processed)which would raise export earningsfrom agriculture and the incomes of farmers. The Governmentof Lesotho has since 1974 been supporting asparagus production and canning. Many observersregard asparagus as a crop with great production potential for Lesotho. The main constrainsseem to be excessivegovernment intervention in production, processing and marketing, and barriers to entry for the private sector. The first section which follows therefore reviews the production, canning and marketing of different asparagusproducts. Subsequentsections identify new crops for Lesotho and estimate farm budgetsfor these crops. Developmentof irrigation, crop researchand extension,and financing of the diversification activities are also discussed. Additional information on the major subjectsis containedin the annexes.

Asparagus Growing and Processing

2. Asparagusofficinalis belongsto the Lily family (Liliaceae), and is grown for the young shootsor "spears" that emerge from the roots. They are marketedas either white or green shoots. To produce white shoots,the soil is ridged up, so that the shootsgrow and elongatewithout being exposedto light and then they are harvestedas soon as the tips penetratethe soil. To producegreen asparagus the shoots are allowed to grow above the soil before they are harvested;upon exposureto sunlight, chlorophyll developsin the shoots, which changecolor from white to green. Almost all of the asparagusgrown in Lesothoat presentis harvestedwhite, but producinggreen shoots in addition would make it possible to extendthe harvestingseason beyond its current Septemberto mid December.

3. Asparagus growing in Lesotho. Asparaguswas first plantedcommercially in Lesothoin 1974, under the World Bank-financedThaba Bosiu Rural DevelopmentProject. The canning factory was financed by a grant from the United Nations Capital DevelopmentFund, arrange under the sameproject. The first harvest took place in 1976 and since then production has increasedevery year. Annex 3 contains a detailed description of the present production of asparagus,and the following outline summarizesthat description.

(a) Growers. There were 1,700registered suppliers to the factory in 1991,mostly smallholders, including numerous women. Each grower is supposed to have 2,000-2,500 sqm in production. Some have more and are still expanding. The current production area is 400 hectaresand the total production in 1991 was about 920 tons. The target for the future is 2,000 growersand a total of 500 hectaresin production,which appearsexcessively modest in view of much more rapid expansiontaking placeacross the border in the RSA.

(b) Yield. In 1991the averageyield for all growerswas 3.5 tons per hectare,which was modest as a result of cold weatherand rain during the harvest. The previous season,the yield was 4.5 tons per hectare. The best producerin Lesothohas achieved 10.5tons per hectare,while in South Africa some growersreport close to 20 tons per hectare. It is possible,therefore, with bettercrop management,to substantiallyincrease the averageyield attainedby growers. - 224 -

For example, if by 1993/94farmers were to have improved their production,and there were 500 hectares in production yielding 6.0 tons per hectare, all of which is achievable, production would increaseto 3,000 tons, or threetimes that for 1991.

(c) Prices. The price farmers receive after deducting the cost of transport to the cannery is M2.16 per kg, less that half of that reportedlyrealized by farmers in the South Africa. South African companiesare reportedto want to procureasparagus in Lesotho,but have not been permitted to do so.

(d) Canning. Asparagusshoots must be cannedthe same day they are harvested,because they deteriorate in storage. The capacity of the factory is therefore very important. During the 1991/92season, when total productionwas 920 tons, the presentfactory is estimatedto have attained a processingrate of 20 tons per day, at least for short periods (say, 2-3 days consecutively). Unfortunately, the proportion of waste also increasedwith this rate of processing,in part becauseof lower quality of work from employeesworking two shifts during the peak season. Without substantialimprovement in managementof processing,a production of 3,000 tons, which is a plausible target for 1993/94, will be impossible to handle,because it implies peak output of at least60 tons per day.

(e) Marketing. The current asparaguscannery also lacks a marketing policy, and therefore its marketing performanceis very poor, raising seriousquestions about whether it would be possible,with presentmanagement, to sell an expandedvolume of asparaguseven if it could be processedsuccessfully. There has not been any seriouseffort to use the facility more efficiently and therebydiversify the processingplant into other lines of produce.

(f) Technical knowledge. Local horticulturalistsin Lesotholack experiencein commercial production,and are unable to advisefarmers in financialand management matters. There is also evidencethat farmersare not receivingcorrect technical advice regardingsuch matters as the use of fertilizers. If the industry is to expand,there is a need to engageprofessional assistanceto train local horticulturalists. Such assistancecould come from South Africa whereappropriate expertise is availableat a costlower than from Europeor the U.S.A.

Market Possibilities

4. In 1990,the EU importedan estimated93,000 metric tons of cannedasparagus, of which Germanyaccounted for about57 percent.Lesotho's total productionin 1990(940 mt) was lessthan 2 percentof Germanimports alone. Morethan half of Europeanconsumption is suppliedby ,which hastaken over the majoritysupply position from ,as the latter'sproduction has fallen with the increasingvalue of its land,and the subsequentconversion of landto higher-valuedindustrial uses. In additionto cannedproduce, the four main consumersin Europe(Germany, , the UK and France)imported more than 40,000 mt of freshasparagus during 1990. Again,Germany accounted for morethan three quarters of theseimports. Thereis alsoa goodmarket for asparagusin SouthAfrica, whereproducer prices are substantiallyhigher than thosereceived by producersin Lesotho;attempts havebeen made by canningfactories in SouthAfrica to procureasparagus from Lesotho,but they have beenrefused. With carefulattention to quality,and good management of marketingin Lesotho,should be few problemsin expandingproduction, either of cannedor freshasparagus, at leastfrom the pointof viewof demand. - 225 -

Constraints to Expansion, and How to Address Them

5. There are no significant technical, infrastructural, labor, land tenure, marketing, financial or economic constrains to a substantial increase of production and export of asparagus. Such an increase would, however, require that the present factory be diversified and expanded. The current cash flow of the factory is a serious problem, since the business is being run on an expensive overdraft with interest rates between 17 and 23 percent per annum. There are serious management constrains in the present business, manifested in failure to resolve marketing problems, to diversify into other lines of processing in order to use better the capacity of the present plant, and to deal adequately with labor problems. If these were merely problems with the local management of the processing plant itself, they could be solved by replacing that management. But such a step has not been taken, and the Lesotho National Development Corporation (under whose auspices the factory is run) has refused to countenance privatization of the factory. The EU, which had been financing the processing operation. decided in 1992 to provide no further financial support until these management problems were addressed.

6. It is therefore of paramount importance that the above management problems be dealt with, and the entire asparagus industry liberalized if the right kind of private sector financing is to be attracted to secure a viable future for the asparagus industry and its expansion in Lesotho. The essential elements of this include: (I) privatization of the present canning factory and removal of constrains on establishment of competitive, private sector financed factories throughout the lowlands; (2) elimination of government intervention in asparagus production, processing and marketing; and (3) removal of constrains on the export of asparagus to South Africa. Attention is also needed to improve the technical knowledge of horticulturalists in Lesotho to ensure that farmers receive the correct technical advice. This requires minimal government intervention, much of which could be done by the industry if allowed to develop.

Identification of Other Horticultural Market Niches

7. In seeking to diversify its agriculture, the best option for Lesotho would be to begin increasing the crops already produced locally for which domestic demand outstrips domestic supply. In the discussion which follows for different groups of crops, those already grown in the country receive attention first, followed by a consideration of new ones with potential. This discussion is followed by an outline of growing requirements for all such crops, and possible constrains to expansion.

8. Vegetables. Attention should be directed first to increasing production of the following large vegetable species in which growers already have experience and where they also have a market in Lesotho. Each of these crops is described in a separate annex. In order of importance they are: cabbage (imports 18,000 tons in 1991), potatoes (imports 9,000 tons), tomatoes (imports 4,000 tons) and onions (imports 1,600 tons). New and smaller crops for which commercial production should be considered in Lesotho include: (1) pinto beans (experiments show these grow well in Lesotho, and there is a large market for the canned product in specific parts of SA; they could be an important addition to asparagus in the production line at Basotho Fruit and Vegetable Canners (Pty) Ltd.); (2) sweet com/baby corn (although it could be sold fresh or canned, the first attempts should be at canning; German imports of the canned product were almost 36,000 tons in 1990); (3) French beans (although this grows readily in Lesotho, irrigation would be necessary for commercial production); (4) broccoli (produced under irrigation and sold either fresh or frozen, this could be an attractive crop); and (5) seed potato (good potential exists in the highlands to grow all the seed potatoes Lesotho currently imports from South Africa, as well as providing scope for exports). - 226 -

9. Fruits. The climate in Lesothois suitable for most fruits and berries,as shown in several local experimentsand by the existenceof a small local production. Again, attention should be given first to increasing large specieswhere there is experienceand a market in Lesotho. Each of the crops is described in the annexes. In order of importancethey are: apples(imports in 1991 5,000 tons), pears (imports 1,600 tons), and peaches(imports 330 toiis). In addition to these, it is recommendedthat Lesothostart it own production of fruit trees. In order to get a good fruit tree of a variety that can sell both locally and for export, a specialproduction technique has to be adapted. The LesothoAgricultural College hiasits own nursery and is familiar with the technique, although it needs to be improved, especially in the areaof phytosanitation.

10. Berries. Experimentscarried out severalyears ago showedthat many types of soft berries would grow well in Lesotho. Listed below are the "new" and smaller crops that should be taken into consideration for commercial production. There are still many uncertaintiessurrounding their culture and marketing. Productiontechniques should be tested in field researchand when fruit bearing begins, market surveysand post harvesttrials shouldbe started. The main specieswhich should be investigated are: strawberries (although it is possible to grow them in Lesotho, post harvest handling is difficult, suggestingthat the way to start would be with small scale production for canning, extended later to exports); red and black currants (the key to their successwould be a canning factory interested in processingthem for jam or juice); blueberries(new varietieswould be needed,and it would be advisable to begin with field-scale experiments,on the order of 1,000sqm); Raspberries(more perishable than strawberries,these would be best for jam production,although this yields less incomethan for the fresh market; trials should be started,but they are not recommendedat this stagefor commercial production); cherries,(these grow well in Lesotho, but rainfall during the harvestperiod can causecracking of the berries; the problem can be dealt with, but requires rather expensivestructures to allow the trees to be coveredwith plastic during the harvestseason); table grapes(although the climate is suitable for grape growing in Lesotho, it is doubtful that sufficiently high quality can be attained to competewith table grapesgrown in the Bloemfonteinarea of RSA).

11. Nuts. There are indications that severalspecies of nuts will grow well in Lesotho. The main problem with nut production is the long time it takes from planting to harvesting. Investmentsin trees and infrastructureare high andthere is no knowledgeof nut production in the country. If it is decidedto start trial production,the following speciesshould be tried: almonds,pecan, walnuts and hazelnuts.

12. Flowers. There are similar possibilities to grow flowers in Lesotho as in RSA. Most of the current production in RSA is in expensivegreenhouses and requires a great deal of knowledge not available in Lesotho. There are types of flowers that can be grown outside without cover. Again, the lack of knowledge would be the biggest constraint. Three types of flowers have been suggestedfor possible production in Lesotho: proteas(currently grown in Zimbabwe, Israel, Australia, and the Cape Province of RSA, these needa period of cold climate, and would do well in the highlandsof Lesotho; they could also be sold as dried flowers); dried flowers (the market for dry flowers is controlled by the Dutch and they need large quantities of each flower. It might be possible to get smaller customers elsewhere,but in order to serve those customers,a wide assortmentof flowers is needed);and lavender (used both as a fresh cut flower and as a dried flower with a famousscent. Production in Lesothowould face competitionfrom Kenyaand Zimbabwe, whose produce can, when the marketis low, be easilyand cheaply dried for sale).

13. Spices and other crops. Thereis a demandfor certainspices and other cropson the world market, but it is difficult to find reliable information about the economicsof their production. Those which have beenmentioned as possibilitiesin Lesothoinclude: saffron (this very expensivedying - 227- material used in food preparation comes from the stamen of a flower belonging to the Crocus family. It will grow well in Lesotho if it is possible to get enough bulbs to start with); ginger (it grows in Lesotho, but competition with India, the world's biggest producer, will be difficult. Expansion in Lesotho should be contemplated only after studying varieties and production techniques in other countries); ginseng (this is an expensive "spice", mostly coming from Asia. A description of production techniques is not readily available, and it is not recommended at the present time).

Growing Requirements for Horticultural Crops

14. Most of the crops mentioned in this report can be grown in many parts of Lesotho, the main limiting factors to increased production being not soils or climatic conditions but entrepreneurs to initiate production. Because of that, this report does not attempt to identify specific production areas for the different crops. Following are some general guidelines.

(a) Temperature. At higher altitudes growth is slowed by the decrease in temperatures. Farmers in the lowland have the earli'est produce on the market, followed by farmers in the foothills. The latest crops are harvested in the highlands. It is thus possible to spread horticultural production in Lesotho over a lengthy period of time. The only areas which have very limited use on temperature grounds are those at high altitude, where the crops are susceptible to frosts.

(b) Rain and hail. Some crops do not tolerate too much rain, while other are easily damaged by hail. Although there are several meteorological stations in Lesotho which record hail, their records are sent to Pretoria in RSA, and not retained in Lesotho. The rainfall and possibility for hail must be investigated prior to establishment of a horticultural enterprise. In RSA it is possible to purchase insurance cover against hail damages, and the insurance companies handling such business have information about hail problems in the whole of Southern Africa.

(c) Irrigation. Apart from asparagus, it is not advisable to start any commercial horticultural production without having the possibility for irrigation throughout the growing season. All compromises will result in large problems.

(d) Distancefrom the market. The further the farm is from the market, the more expensive the transport of produce. In choosing the locations for new horticultural enterprises, the quality of the roads as well as the distances involved are important, since most horticulture produce will not withstand being carried over bumpy roads. - 228 -

Farm Budgets for Horticultural Crops

15. Over the last several years, the Agricultural ResearchDivision and the Lesotho Agricultural Productionand Institutional Support(LAPIS) Projecthave producedseveral "Production Guidelines" for horticulture crops. Researchersfrom Lesothoand the USA preparedthese guidelines, which consist of both production data and farm budgets, using data collected from farmers. The Directorate of Agricultural Economicsin RSA has also compiled "EnterpriseBudgets" based on data from commercial farmers in the Orange Free State. Since future commercial production in Lesotho must be able to competewith that in RSA, these two sourcesof farm budget information have beencompared, and used as the basis for budgetsestimated for this report, with supplementaryinformation from other sourcesin some cases. There are large differencesamong the budgets,resulting from different sizes of operation and estimatesof prices. Their value is mostly in the relative level of return when comparingcrops. Most show an acceptablelevel of return, but in spite of this, adoption of the crops concerned is at an astonishing low level in Lesotho,strengthening the argumentmade here that institutional problems are more important than technical problems in explaining the modestdevelopment of horticulture in the country. The farm budgetsare presentedin the annexesfor eachmajor crop.

Recommendations

16. There are many constrainsto developinga horticultural export industry in Lesotho,but with the right government policies a major expansion could be achieved. The priority goal should be to acceleratethe expansionof asparagus. The following section outlines a development program. It assumesgood linkages between Lesotho and RSA, under which special skills in agriculture and horticulture in RSA could be drawn upon.

Private Sector Development

17. The most important actions the Government can take in support of developing a dynamic horticultural industry in Lesothowould be to establisha good investmentclimate for the private sector. Fostering the private sector should be the "organizing principle" shapingthe Government'sinitiatives in policies, infrastructure(roads, water supply, electricity, telecoms, irrigation, and marketing), services (research,extension, veterinary, soil testing, seedcertification, market intelligence,weather information, education, health), farm credit, agroprocessing,rural finance. Paradoxically,strong, rather than weak governmentis neededfor successfulprivate sector development.

18. The Governmentneeds to make surethat:

* benefitsreach small farmers;

* policies are aimed at establishingand liberalizing as much as possible (markets for outputs, inputs, technologies, land, labor, transport, services,credit, domestic and foreign trade); all policies are consideredtogether to minimize conflicts; and that incentives and constrainsare understoodthoroughly in orderto changethem successfully;

* infrastructure and other public investmentsare undertakento the extent possible by private contractors;

* services are provided on a competitive basis by private contractors(especially veterinary and market intelligence),or that "privatization" principles are applied, such as cost recovery,targets - 229 -

for internal performance, so that essential public services are as efficient as possible, not diverting morecapital than neededfrom the privatesector; and that servicesin support of private sector development,such as legal,accounting and managementservices, are well established;

* credit is delivered in the context of a well-functioning, competitive market, by privately owned institutions;

* a legal and regulatory system is establishedwhich fosters private sector growth by removing obstacles,setting fair rules for competition,making them known freely to the businesssector and wider community, and enforcing them rigorously, protecting property rights and enforcing contracts,resolving businessdisputes, putting in placea framework for orderly formation and (if need be) dissolution of privateenterprises, correctly identifying "public goods" and the presence of "externalities,' and dealing with them equally, and protecting,but not overprotecting,against exploitation; this includesreformulating taxation systemsconsonant with a more indirect role for Governmentin the economyand the needfor private sector incentives;

* private ownership of land and other assets; or stable, alienable usufruct rights/tenure are espousedthroughout the economy(with realistic treatmentof "public goods");

* training to increasemanagement skills is widely encouraged;

* welfare is providedas income-supportsafety nets,which do not distort the operation of markets.

19. This is not easy to achieve, requiring as it does a comprehensiverealignment of public sector focus in order to increasethe public sector'seffectiveness, and removebarriers to private sectoractivity. Among other things, it requiresthat: the Governmentshould concentrate mainly on education,health and infrastructureplanning (with private contractingfor serviceswhere possible,and specialattention to the economic climate for transport services), while ceasing to be involved in business investment decision-making, price and margin setting, or other areas where private enterprise would be more appropriateand effective; that present restrictive regulations should be replacedwith transparentand helpful ones, from imports all the way through the production and marketing systems. and these regulations are made widely available to businessenterprises and the wider community; that efficient contractual services are established for soil laboratories,specialized adaptive research, and market intelligence; that ways are found to mobilize savings and channel them effectively to potential borrowers;that a crop insurancescheme to cover hail damageshould be established;that the DC should be radically transformed,or abolished;and that privatesector monopoliesshould also be avoided.

20. Under a proposedPrivatization and Private Sector Development Project, it is planned that a comprehensivereview of legislation affecting the private sector in Lesothobe undertaken. This report strongly recommendsthat that review should include all legislation and regulations relevant to

I Forexample, when private veterinary services are set up. it is importantto havein placea stronglegal system regulatingthe sale and use of toxicdrugs and chemicals, inspection and certification of meatand dairy products, their handling, storageand marketing, animal quarantine and control of livestockmovement where needed, certification and registration of veterinarypractitioners, conditions under which persons of variousqualifications may practice, market drugs etc. Sharingof functionsand costs between the public and private sectors when mandatory. disease prevention or other"campaigns" must be carriedout, possibly the control of feesand prices to preventmonopolies, and arrangements for training various parties as wellas settingup livestock producer groups and associations which will exercisecountervailing power in themarketplace. - 230 - agriculture, with a view to setting the stage for a dynamic private sector in intensive agricultural production, agro-processing,marketing, and foreign trade. Although a significant proportion of the legislation governingthe generaleconomy also has relevancefor agriculture.particular attention should be paid to the Land Act 1979,Land SurveyAct 1980.Agricultural Marketing Act 1967(Amendment Act 1977 and Regulatioiis 1988). Abbatoir Regulations 1972, Import and Export Control Act 1984, Importation and Exportation of Livestock and Livestock ProductsAmendment Act 1984, Exchange Control Regulations 1989, the DC Act 1967and AmendnmentAct 1983, and all regulations regarding PriceControls, and Useof Water Resources.

Research and Extension

21. Over the past 20 years,several African countrieshave made large investmentsin the production of horticultural crops for export. Most of this developmenthas been done by private entrepreneurs,with little help from local researchand extension. In many cases,the investorshave brought in their own farm managers,often expatriates. More investmentin human capital among the local people and a reorientationof the researchand extensionservice to applied, commercialcrop production is required to changethis situation.

22. Although there is no higher education in horticulture available in Lesotho. there are horticulturalists with good qualifications, including Mastersdegrees obtained abroad. Work conditions for such trained professionalsare poor, however. In particular they lack funds for operating costs (research and teaching materials, fuel for cars, running costs of trial greenhouses),and the communicationilinks betweenextension and research workers are poorly developed. The following steps are recommendedfor addressingthe situation:

- Increasethe number of M.Sc. studentssent out of the country for training, mainly to RSA, which has similar growing condition to those in Lesotho and will, in general, provide training more cheaply and betterthan countries further afield. The University of the Orange FreeState in Bloemfonteinshould be approachedfor cooperation.

* Strengthenresearch management and hold it responsiblefor obtaining results. Also secure enoughfunding to run researchfor at least4-5 years(the University of the OrangeFree State in Bloemfontein has an office working solely on fund-raising and public relations, and a similar office should be at LesothoAgricultural College).

* Start to work with crops that are useful for the development of horticulture in Lesotho, selected from the list of those discussed in this report. The work should begin with productionof young plantsfor farmers(plug plants).

* Concentrateduring the first 5-10 years on practical experiments,with close monitoring of economicresults, and a strict emphasison benefiting the country and the growers. For the time being, basic researchshould be left to countries better able to afford it, and, where necessary,postgraduate students should be sentto stationsabroad to do their research.

* Extensionworkers should spendas much time as possiblein the fields with the growers,and this should be achievedby making the staff responsiblefor results,and providing them with strong management,adequate means of transportation,and training courses for extension workers in both technical subjectsand budgeting/control. When extension workers have begunto provetheir worth, fees for their servicesshould be considered. - 231 -

Irrigation

23. The production of high quality vegetablesrequires irrigation becauseof the uneven rainfall experiencedin Lesotho,and without irrigationcommercial production cannot be recommended.According to reportsavailable (FAO) and informationgiven to the mission,Lesotho is well stockedwith sophisticated irrigation equipmentfor largescale farming. A high proportionof it is not in use. One Europeancompany, in cooperationwith the Government,is experiencingsome difficulties in a large irrigation scheme. The problemsare attributedto poor planning(resulting, among other things,in an incorrectlylocated irrigation pipe), too great a capital intensity,and too large a scaleof farms, with the farmersthemselves little more than laborers. Individual entrepreneursshould be encouragedto establishtheir own irrigation systems.The current market in and outside of Lesotho should justify considerableinvestments in this sector, if entrepreneurscan obtain controlof enoughland. Irrigation for smallholdersis alsowell suitedto collective action. Farmerslocated in the samearea should organize and buy their own irrigation equipmentwhich is simpleto run andwell suitedfor the site. The cooperation/jointuse must be institutionalized,with clear and identified responsibilities.Prior to purchasing,they shouldask for quotesfrom companiesin Lesotho,SA and abroad, and take care to establish good arrangementsfor spare parts and service. Purchaseof apparentlycheap second-hand equipment may end up beingmuch moreexpensive than beginningwith new equipment.

Financing Diversification Activities

24. Private enterprises intending to establish a company for the production and export of horticultural crops will normally require some financing. It is very often the casethat local banks lack knowledgeof such projectsand are not willing to give loans without extremely high security and a very high interest rate. The land tenure system in Lesothomakes it difficult for one entrepreneurto amass enough land for borrowing and horticultural production,but if severalfarmers join forces, they will be able to get enough land to createa good production unit. Banks in Lesothoare not eagerto lend money for agricultural projects. Export-oriented horticulture requires large amounts of money to begin operating, and the best way to begin is to find partnersabroad interestedin selling the produce in their own countries. Joint venture companieswith such partnersare commonly supportedby the EU and the U.S.A. through grants and soft loans. It is also possibleto apply for loans from organizationssuch as the African Project DevelopmentFacility (APDF), which has recently providedsoft loansto horticultural enterprisesin Africa (among 26 projectscompleted November 1, 1989 - December31, 1990, were II related to agriculture, among them 3 horticultural). Another possibility would be to send project documentsout for internationaltenders through some of the large internationalbanks.

Airfreighting From Lesotho

25. In the areaof Ladybrand(RSA), lessthan 30 minutes from Maseru,there is a largeconcentration of horticultural production,some of which productsis exportedto Europe. The climate in the areaand in the lowlands of Lesotho is the same. The farmnsin RSA are producing both fruits and vegetables, irrigating from the Caledon river (which forms the border between RSA and Lesotho). During the season,the group of farmers operatingthere produceenough for weekly export of 100tons of fruits and vegetables(equivalent to 3 full planesof the DC8/Boeing 707 type). The best way to export appearsto be flying cargo aircraft out of Maseruairport, and somecargo airlines have shown keen interestin sucha project. This may be worthwhile for the Governmentto consider as a way of laying a foundation for production within Lesothoitself, and airfreighting of the country'sown produce. The main consideration should be lowering transportcosts, as this will benefitproducers and increaseprofitability. - 232 -

Commercial Forestry Plantations

26. The needfor wood productsis visible throughoutLesotho, where production is much lower than local demand. Wood is used for construction, fuelwood, fence posts, and for coffins and furniture. At present most of the demand is met by imports from the Republic of South Africa. Lack of woody trees is a limiting factor in erosion control throughout the country. At present tree planting schemes are undertaken by the Government in specific problem areas such as watersheds, gullies and to meet the wood requests of local communities. Some of these projects are assisted through food for work projects of the World Food Programme (WFP). Under current arrangements. 20 percent of the revenue from the sale of wood products goes to the community and the remainder to the Forestry Division of the Ministry of Agriculture. This program has to address the same problems facing other communally owned resources such as rangelands, namely that trees need continuous care such as watering, weeding and protection from livestock, and because they are common property no one takes responsibility for these tasks. Thus failure of community plantations is a major problem. Furthermore, most community members assume that community forests are under the disposal of the chiefs who control the way in which the proceeds from the sale of wood are handled. To overcome this problem, it is proposed that small scale private voluntary plantations are promoted in Lesotho since forestry production can compete well with other agricultural production enterprises in rural areas. - 233 -

Annex 12 LESOTHO POVERTY ASSESSMENT

MininmunmWages

Annex 12, Table 1: Minimum Wages As of August 1992 (Maloti) Basic monthly Minimum Daily wage contract weekly contract contract (per month) (per week) (per day) Copy typist 301.51 75.38 15.87 Driver (car and light van) 350.04 87.51 18.42 Driver (medium size vehicle) 383.69 95.92 20.19 Driver (heavy vehicles) 491.33 122.83 25.86 Operator (hammer mill) 255.38 63.85 13.44 Junior Clerk 301.51 75.38 15.87 Machine operator 350.04 87.51 18.42 Messenger 255.38 63.85 13.44 Machine attendant 301.51 75.38 15.87 Receptionist 301.51 75.38 15.87 Shop assistant 281.49 70.37 14.82 Telephone operator 301.51 75.38 15.87 Ungranted artisan 325.30 81.33 17.12 Unskilled worker (heavy physical work) 281.49 70.37 14.82 Unskilled worker (light physical work) 255.38 63.85 13.44 Waiter 289.50 72.38 15.24 Watchman 266.94 66.74 14.05 Weaver (a) In-Training (6 months) 255.38 63.85 13.44 (b) Trained 267.80 66.95 14.09 Sewing machine operator (a) in-training (6 months) 255.38 63.85 13.44 (b) trained 267.80 66.95 14.09 Sourcc: LesothoGovernment Gazette, Vol. XXXVII,No. 91, Monday, 17th August 1992. - 234 -

Annex 13 LESOTHO POVERTY ASSESSMENT

Organization of Health anid Education Services in Lesotho

Health

I. Central to the performance of health care delivery in Lesotho is the concept of the Health Service Area (HSA). HSA's serve from 38,000 to 250,000 people, each with a first referral hospital. At the next level are Health Centers, each with a catchment population of 6,000 to 10,000 persons. The third basic level of services is provided at the village level, through a network of Village Health Workers (VHWs). Tertiary level health care is provided in the capital city, Maseru, through the services of QEIL.

2. Each HSA has a supervisory team associated with the HSA hospital. The HSA hospital and supervisory team is responsible for the training of all non-stipendiary staff, the implementation of primary health care and all other government policies. and the provision of basic hospital services. Once of the great strengths of the Lesotho system is that HSA hospitals are made responsible for the health not only for their users but also of the people who live in their area.

3. The Ministry of Health operates 9 general and 2 specialist hospitals in 9 HSAs, as well as 45 health centers in those HSAs. The MOH also provides a program called the Lesotho Flying Doctor Servicewhich serves 12 very remote clinics and uses QEII Hospital in Maseru as its referral center. The Lesotho Flying Doctor Service comprises one of Lesotho's 19 HSAs.

Annex 13, Table 1: HSAs in Lesotho MOH PHAL HSAs' HSAs"' CatchmentPopulation' 1.196,000 642,000 Averagenumber of peopleper

Doctor 17,582 8,233 Nurse 3,214 2,581 VHWC' 401 298 TBAC/ 2,544 1,194 HospitalBedsc/ 846 729 Notes: a/basedon EstablishmentLists (MOH and PHAL). b/Ralitapoleand Adu-Boahene, 1992, is citedin Sechaba(1993). c/basedon EstablishmentLists and data provided in Sechaba(1993).

4. The remaining 9 HSAs are served by 9 hospitals owned by various churches, under an umbrella organization called the Private Health Association of Lesotho (PHAL). PHAL has no direct control over its largely autonomous member hospitals but, rather, serves as a means of communication between PHAL member hospitals and the MOH. It is represented on most MOH committees, with the result that nearly all MOH programs apply equally to PHAL as they do to MOH hospitals. -235 -

5. PHAL sees itself as supporting the work of the hospitals and HSAs in four ways:

I) support for the HSA concept and decentralization; 2) health education for staff and, through them, for their communities: 3) improvement of hospital administration including manpower planning and staff development; 4) logistic support in terms of management of drugs, financing, etc.

6. Collaboration between the MOH and PHAL is reinforced by a Government of Lesotho subvention to PHAL for partial payment of salaries, financing of clinic improvements, and the procurement of drugs and other medical supplies from the Lesotho Pharmaceutical Cooperation. This is in recognition of PHAL's role as a health care provider in 9 HSAs, as well as for compliance with regulations and public health programs developed by MOH. Between 1991/92 to 1993/94, the MOH grant to PHAL has been about M6.5 million, representing between 5-6 percent of the MOH budget.

7. Comparisons of "catchment" populations served by MOH facilities and PHAL facilities in their respective HSAs are summarized in the table above. PHAL, which recovers large shares of costs, clearly provides a superior ratio of health personnel to its catchment population, though beds per capita are similar between the two categories of providers.

8. In addition to PHAL, there are various independent clinics in Lesotho. These are operated by private doctors, the Red Cross, and the Lesotho Highlands Development Authority. According to records of the Lesotho Medical, Dental and PharmnacyCouncil, there are some 39 doctors in private practice, of which 24 are Basotho. Of these, eight are based in Maseru, four are with the Lesotho Highlands Development Authority, one is in Butha-Buthe, and many of the remainder have Lesotho addresses but likely come from South Africa on certain days each week to clinics in Lesotho. Private practice is thought to serve largely the needs of urban-based, relatively well-to-do clients.

9. Finally, traditional practitioners are the most common medical practitioners in Lesotho, most operating out of their homes and providing medicines based on herbal remedies. In 1991, there were about 8,500 licensed traditional practitioners in Lesotho, and probably many more who are not licensed. In areas such Mohale's Hoek and Mokhotlong, some traditional practitioners have been working alongside HSA personnel in the distribution of condoms for family planning and AIDS prevention, as well as in the follow-up of TB patients.

Education

10. Unlike the situation in many Sub-Saharan African countries where education is the direct responsibility of the state, Lesotho's education system remains largely a joint venture between the Government and the churches (mainly Anglican, Evangelical and Roman Catholic), with the latter owning and operating 97 percent of primary and 86 percent of secondary schools. In addition, most technical and vocational training institutions, with the exception of the MOE-run Lerotholi Polytechnic (LP) and the Thaba Tseka Skills Training Center (TSTC), are church owned. The National University of Lesotho (NUL) is administered by an independent council but, in view of its heavy reliance on Government for both investment and recurrent expenditure includes MOE representation on its council. Specialized training institutions and programs are operated by the Ministries of Agriculture, Trade and Industry, Rural Development, Health, the Prime Minister's office, and in the industrial sector by the Lesotho National Development Corporation. - 236 -

I. Currently, the structure of Lesotho's education system consists of seven years of primary education, five years of secondary education (three years of junior secondary and two years of senior secondary) and four to six years of higher education. Primary education is conducted in Sesotho (the national language) for the first three years and mainly in English in Standards 4 through 7. Post-primary education is entirely in English. Non-formal education courses are provided by the MOE through the Lesotho Distance Teaching Centre (LDTC) and by the NUL Institute of Extra-Mural Studies (IEMS). Primary and junior secondary teacher education is provided by the National Teacher Training College (NTTC), while senior secondary teacher education is the responsibility of the Faculty of Education at NUL. Besides education, NUL offers courses in humanities, natural sciences, agriculture, social sciences and law. As noted earlier, enrollments have risen rapidly over the past decade, with over 350,000 children enrolled in primary and over 45,000 enrolled in secondary schools as of March 1990. Moreover, female enrollment levels are especially impressive with over 94 percent of girls aged 7-14 years enrolled in school. As discussed previously, more girls than boys are enrolled at all levels of the education system so that female:male ratios are currently about 1.2:1 in primary, 1.5:1 in secondary and 1.5:1 on average in post-secondary institutions.

12. The administrative, financial and academic control of the formal education and training system is vested in the MOE, which, inter alia, has responsibility for the training and payment of teachers, formal approval of teachers' appointments, dismissals, etc., administration of exams, review and authorization of curricula, school inspection, and regulation of the opening and closing of schools. While there are differences in the organization of the three main churches with respect to education management, these are essentially variations of the same structure. Common to all is a single line organization of three levels of hierarchy comprising an executive body at the top, a single education secretary at the middle and parish-appointed school managers at the floor.

13. Traditionally, financing of education in Lesotho has been shared by the Government, the local private sector (mainly through students' fees and community labor for construction projects) and direct contributions from abroad, whether in cash or kind, made by the churches. Although it is not possible to quantify the latter (church) contributions, there is no doubt that they have declined steadily in recent years and have become only a marginal source of financing. Over the same period, there have been substantial increases in funding of capital expenditures by bilateral and multilateral donor agencies, which are channelled through the government budget. In 1987, total expenditures for education in Lesotho (including the value of school feeding programs and parents' contributions) amounted to about Ml 13 million or about 3.7 percent of the country's GNP. This translates to about M68 or slightly over US$ 27 per capita at current exchange rates. The Government's contribution (including foreign aid) -- covering MOE administration, teachers' salaries at all levels, most NUL costs (including student loans) and a high share of sectoral investment -- made up nearly 70 percent of expenditures, with private sources, mainly parents, providing the balance of 30 percent. Domestically-generated resources covered 71 percent of education expenditures and foreign funds the remaining 29 percent. - 237 -

Annex 14

LESOTHO POVERTY ASSESSMENT

Employment Prospects in the South African Mining Sector

1. The employment of foreign migrant labor in South Africa has been declining in recent years. After reaching a peak in 1990, the number of migrant Basotho miners employed in South Africa declined by approximately 5,198 in 1991 and by another 6,226 in 1992. This is due to a combination of factors: (1) increased mechanization of mines (coal mines, in particular); (2) closing down of marginal mines; (3) relatively high mine wages which now attract indigenous (i.e. South African) blacks in larger numbers; and (4) a change in South African policy regarding relocation of South African black mine workers at the end of their contracts in the so-called "homelands" which now make it seem worthwhile for them to seek extension of their contracts as a way of avoiding forced repatriation to the "homelands". It seems, however, that the 1993 employment levels have stabilized around the 1992 level.

Annex 14, Table 1: Basotho Miners in South Africa' Period Average Avera§ee 4 number wage- DeferredPay' RemittancePayments employed Maloti % change Maloti % change '000 from a year '000 from a year ago ago 1983 115,327 3,436 103,568 35 74,207 45 1984 114,041 3,927 123,550 19 82,923 12 1985 116,223 4,452 141,546 15 92,675 12 1986 121,450 5,136 178,594 26 104,776 13 1987 125,934 7,160 188,568 6 124,354 19 1988 124,781 7,598 210,932 12 166,127 34 1989 126,264 8,679 242,139 15 190,445 15 1990 127,386 10,069 239,934 -1 233,509 23 1991 122,188 11,327 160,308 107,987 1992 115,962 137,693 91,877 1993:Ql* 110,014 33,975 24,650 Notes: Ccover all mines,with gold having the largest share. 2 Thesefigures are average wages, including overtime payments, carned by blackworkers in themines. Figures are suppliedby the South African Chamber of Mines. In effectfrom April Is, 1991,Compulsory Deferred Pay has been reduced from 60 percent of miners'wages to 30 percent.Deferred pay as shown in thistable, represents miners withdrawals from the fund. 4 Theseare part of miners'wages (remaining 70 percent including other employment benefits) transferred toLesotho throughrecruiting agencies. 5 Preliminaryestimate provided by National Employment Service.

2. There is reason to believe that Basotho migrants will not be a big part of any future general growth in employment in South African mines. Between 1970 and 1983 employment in South African mines rose by 20 percent, and the share of South African citizens in the total mine work-force rose from 24 percent to 60 percent. The Basotho share rose marginally from 17.7 percent to 21.3 percent mainly due to the withdrawal of migrant workers from Mozambique and Malawi. In the future, Basotho migrant workers may not benefit from mine employment increases because a combination of high labor force - 238 - growth and slow economicgrowth in South Africa is likely to result in increasedcompetition from South African labor. In addition, mechanization in coal mines is likely to continue, limiting growth opportunitiesand thus making a further fall in coal mine employmentmore likely.

3. These labor demand trends, in combinationwith the recent droughtand the increasingnumber of entrants into the domestic labor force (24,000 per year) is likely to intensify Lesotho youth employment problems. The policy of South African mines is to favor experienced and career miners. When the present generation of Basotho career miners retire in the 1990s, it is likely that they will be replaced by SouthAfrican miners instead of young miners from Lesotho. - 239 -

LESOTHO POVERTY ASSESSMENT

BIBLIOGRAPHY

Baffoe and Associates(1990), Direct Foreign Investment in Lesotho - Past and Present Forces Underlying Patterns of Investment Flws. Mimeo. Maseru.

Baffoe, Frank (1989), Skills Profile Survey. Reportprepared for the Governmentof Lesotho,Department of Labor, Ministry of Employment,Social Welfare and Pensions.

_ (1991),Constraints on the Developmentof Small Scaleand Informal Sector in Lesotho. Report prepared for the Southern Africa Team for Employment Promotion(SATEP).

Baholo, Salometsi(1993), Statementon Budget 1993/94,Deputy Prime Minister and Minister of Finance, Planning Economic and Manpower Development, i7 June, 1993. Mimeo, Ministry of Finance.Maseru.

Baily, Brendan(1994), Agricultural Marketing in Lesothoand PriceTransmission to the Village Level: Implicationsand Options for Policy, DRIG. Maseru.

Bojo, Jan(1991), The Economics of Land Degradation: Theory and Application to Lesotho. Stockholm:Stockholm School of Economics.

Brautigam,Deborah (1989), ComparativeAdvantage in LesothoAgriculture and Agroindustry, Mimeo. DevelopmentAlternatives, Inc. Washington,D.C.

Bruce, John W. (1984), A Layman'sGuide to the Land Act 1979,in: Governmentof Lesotho, Ministry of Agriculture, Proceedings,Land Act Seminar,Quithing, Lesotho, March 18-22,1984. Maseru.

Byrne, D. and P. Gertler (1990), An Analysis of the Impact of User Fees on Outpatient Utilization in Lesotho. Mimeo, RandCorporation. SantaMonica, LA.

Convery, FrankJ. (1995), Applying EnvironmentalEconomics in Africa. World Bank Technical PaperNumber 277. Africa TechnicalSeries. WashingtonD.C.

Carvalho,Joe William (1988), Agriculture and EconomicDevelopment in Lesotho:Analysis Using a Social Accounting Matrix, PhD thesis, Departmentof Agriculture Economics, WashingtonState University.

Chen, S., G. Datt, and M. Ravallion (June 1993),Is PovertyIncreasing in the Developing World? World Bank working paper# 1146. Washington,D.C.

Cole, P., (1993), GrossMargins for Agricultural Enterprisesin the Matelile Rural DevelopmentProject. Matelile Rural DevelopmentProject. Maseru. - 240 -

Collier, P., S Radwan,S. Wangwe,and A. Wagner(1986), Labor and Poverty in Rural Tanzania. Ujamaa and Rural Development in the United Republic of Tanzania. InternationalLabor Office/ClarendonPress. Oxford.

Coopersand Lybrand (1989), LesothoHighlands Economic Impact Study. Maseru.

Daniels, D.L., L.N. Makoae,and R.G. Feachem(1990), A Case-controlStudy of the Impact of Improved Sanitationon DiarrheaMorbidity in Lesotho,Bulletin of the WHO, 68: 455-63.

Deaton,A. (1980), Economics of Consumer Behavior. CambridgeUniversity Press. Cambridge.

DevelopmentAssociates, Inc. (1986), Assistanceto Lesothoon the Mining Repatriation Issues. Final report submittedto the United Nations DevelopmentProgram. Maseru.

Dororsh, Paul A. and S. Haggblade(1992), Agricultural Growth Linkagesin Madagascar,CFNPP Working Paper 22., Cornell Food and Nutrition Policy Program. Ithaca.

Ekl6f, Jan A. and M. Peete,Distribution of Incomesand Consumptionin Lesotho, a study basedon the 1986-87Household Budget Survey Data, Central Bank Occasional Paper No. 4. Maseru.

Evans,P. (1987), Planning Self-sustainingProgrammes for Sanitation:The Lesotho Experience,Waterlines, 6(2), 5-8.

Ferguson,James (1990), The Anti-Politics Machine: 'Development', Depoliticization, and Bureaucratic Power in Lesotho. CambridgeUniversity Press. Cambridge.

Foster, J.E., J. Greer and E. Thorbecke(1984), A Class of DecomposablePoverty Measures, Econometrica. Vol. 52, 3:751-66.

Foulo, T. (1991), Migrant Workers'Retrenchments: Implications for LesothoEconomy. Central Bank of Lesotho,Staff OccasionalPaper. Maseru.

GEMINI (1990), LesothoSmall and Macro EnterpriseStrategy Phase II: Sub-Sector Analysis. GEMINI Technical Report No. 9.

(1990), Small-ScaleEnterprises in Lesotho:Summary of a Country Wide Survey. GEMINI Technical Report No. 14.

Gill, D. (1992), Lesotho:A GenderAnalysis. Sechaba Consultants. Maseru.

Government of Lesotho, Ministry of the Public Service, The EstablishmentList, 1993/94. Maseru. - 241 -

(1967/69), Bureau of Statistics, Rural Household Consumption and Expenditure Survey. Maseru.

(1972/73), Bureau of Statistics, Urban Household Budget Survey. Maseru.

(1976), Ministry of Health and Ministry of Agriculture, Nutritional Survey. Maseru.

(1983-1992), Bureau of Statistics, Various computer files and government databases. Maseru.

(March 18-22, 1984), Ministry of Agriculture, Proceedings, Land Act Seminar. Quithing, Lesotho.

(1986), Bureau of Statistics, Population Census, Analysis Report: Vol III. Maseru.

(1986/87), Bureau of Statistics, Household Budget Survey. Maseru.

(1987), Bureau of Statistics, 1987 Yearbook of Industrial Statistics. Maseru.

(1988), Bureau of Statistics, Incomes, Expenditures and Consumption of Basotho Households, Two Volumes. Maseru.

(1988), Ministry of Finance. Lesotho: Policy Framework Paper, 1988-89-1990-91. April 11. Maseru.

(1988), Sectoral Study on Rural Development: Prospects and Challenges. Maseru.

(1988/89), Ministry of Agriculture, Technical Operations Unit records. Maseru.

(1989/90), Bureau of Statistics, Agricultural Census, Analysis Report (Draft). Maseru.

(1989), Central Bank, Lesotho's Foreign Trade in Services in the 1980s. Staff Occasional Paper No. 2. Maseru.

(1990), Ministry of Finance. Lesotho: Policy Framework Paper, 1990-91 to 1994-95. May 21. Maseru.

Ministry of Education, Education Sector Development Plan 1991/92-1995/96. Maseru

(1991), Ministry of Planning, Economic and Manpower Development, The Situation of Women and Children in Lesotho. Maseru.

(September, 1991), Labor Construction Unit/Ministry of Public Works, labor intensive construction submissions to donors. Maseru.

(September 11, 1992), Ministry of Agriculture, Planning Division. Proceedings of a Seminar on the Informal Grain Market in Lesotho. Maseru. - 242 -

(1992), Ministry of Health and Ministry of Agriculture, National Nutrition Survey. Maseru.

(1992), Government Gazette, Vol. XXXVII, No. 91, Legal Notice No. 193. Maseru.

(1992), Government Gazette, Vol. XXXVII, No. 118, Labor Code Order 8, No. 24. Maseru.

(1992), Government Gazette, Legal Notice No. 193, page 971. Maseru.

(1992), Report on Institutional Reforms and Policy Development in the Provision of Textbooks,Supplementary Materials and Journals/Periodicals,Vol. 2. Lesotho Schools Survey. Maseru.

(1992) Ministry of Planning, Economic and Manpower Development, Fifth Five-Year DevelopmentPlan 1991/1992-1995/1996.Maseru.

(1992), Ministry of Agriculture, Cooperatives and Marketing, The Lesotho National Fruit and Vegetable Production Survey. December, 1992. Maseru.

(1992),_ Ministry of Finance. Lesotho: Policy Framework Paper, 1992/93 to 1994/95. May 21. Maseru.

(1992), Bureau of Statistics, Lesotho NationalAccounts 1980-1991. Maseru.

(1992), Bureau of Statistics, Statistical Report, No. 1: Exports of 1986-1990. Maseru.

(1992), Bureau of Statistics, Prisons and Tourism Statistics in Lesotho, 1988/89, Annual Report. Maseru.

(992),(I Bureau of Statistics, Statistical Reports, No. 13: National Accounts 1980-1991. Maseru.

(1992), Bureau of Statistics, Lesotho Statistical Yearbook 1992. Maseru.

(1992), Central Bank, Annual Reportfor 1991. Maseru.

(1993), Ministry of Finance. Lesotho: Enhanced Structural Adjustment Facility. Policy FrameworkPaper, 1993/94-1995/96.Maseru.

(1993), Bureau of Statistics, Basotho Women and Their Men. Maseru.

(1993), Bureau of Statistics, Lesotho Statistical Yearbook 1993. Maseru.

(1993), Bureau of Statistics, Statistical Report, No. 2: 1986 Economically Active Population:Some Highlights. Maseru.

(1993), Bureau of Statistics, Consumer Price Index from January 1980. Mimeo. Maseru. - 243 -

(1993), Department of Agriculture, Lesotho Agricultural Situation Report: 1976/77- 1991/92.(Draft). Maseru

Haggblade, S., P. Hazell,and J. Brown (1989), Farm-NonFarmLinkages in Rural Sub- Saharan Africa. WorldDevelopment. (UK) 17: 1173-201,August 1989.

Haggblade, S., J. Hammer, and P. Hazell (1991), Modeling Agricultural Growth Multipliers. AmericanJournal ofAgricultural Economics,(U.S.) 73: 361-74, May 1991.

Haggblade,S., and P. Hazell(1990), Rural-UrbanGrowth Linkagesin India. Indian Journal ofAgricultural Economics.(India) 46: 515-29, October/December1991.

Hunter,John P. (1987), The Economicsof Wool and Mohair Productionand Marketing in Lesotho, Research Report No. 16, Institute of Southern African Studies. National Universityof Lesotho Roma.

Hunter, John P. and Makitimi,None L. (1990),The Developmentof Lesotho'sWool and Mohair Marketing System: Options for Continued Institutional Change and Policy Reform,Research Report No. 2S, March. Instituteof SouthernAfrican Studies,National Universityof Lesotho. Roma.

InternationalFund for AgriculturalDevelopment, Africa Division,Project Management Department (1992), Kingdom of Lesotho General Identification Report, Volume II, WorkingPaper 1-5.

InternationalLabor Office/JASPA(International Labour Office, Jobs and Skills Programme for Africa) (1979), Options for a Dependent Economy: Development, Employment and EquityProblems in Lesotho. Adis Ababa.

InternationalLabor Organization/SouthernAfrica Team for EmploymentPromotion (SATEP) (1985), Planning for Employment and Contingency Employment Programmes for RepatriatedMigrant Labour in Lesotho.

InternationalMonetary Fund (1989), Lesotho - RecentEconomic Developments. SM/89/1 1, January23. Washington,D.C.

(1992), Lesotho, Staff Report for the Midterm Review of the First Annual Arrangementunder the EnhancedStructural Adjustment Facility, February 4. Washington,D.C.

Katona, Emil (1992), Incomesand their Distributionin the Urban Informal Sector of Lesotho.Maseru.

Khabeleand Shale (1986), Some Subjectiveand ObjectiveExperiences of Rural Basotho Women in Food Production:a paper presented to the annual conference of Southern African UniversitiesSocial ScienceConference. Tanzania.

Letlala,M. (1987), Consequencesof Food-For-WorkProjects in Lesotho,National Universityof Lesotho. - 244 -

Lewis, Blane and T. Erik (1992), District-Level Economic Linkages in Kenya: Evidence Based on a Small Region Social Accounting Matrix. World Development, (UK) 20: 881-97, June 1992.

Lundahl, Mats and L. Peterson, (1991), The Dependent Economy: Lesotho and the Southern African Customs Union. Boulder, CO: Westview Press.

Matekane, E.M. (1992), Lesotho Consumer Price Index: An Evaluation of Its Reliability and a Review of Issues in Inflation and Domestic Policy Responses, Staff Occasional Paper, No. 8, Central Bank of Lesotho. Maseru.

Matobo, T. (1992), Basotho Mineworkers' Sexual Practices. National University of Lesotho. Maseru.

Mochebelele, Motsamai T. and L. Ranko, (1990), Rural Dairy Development and the Role of Agricultural Development Projects in Lesotho, Agricultural Marketing Research Project, Research Report No. 26. Institute of Southern African Studies, National University of Lesotho. Roma.

Moran, Mark (1990), The Village Water Supply Maintenance Study. VWSS, Government of Lesotho. Maseru.

Muelbauer (1974), Inequality Measures, Prices and Household Composition, Review of Economic Studies, 41: 493-504.

Murray, C. (1981), Families Divided, Cambridge University Press. Cambridge.

National Rural Sanitation Program (NRSP), (1992), Annual Report 1992. Maseru.

(1993), Annual Report 1993. Maseru.

National Workshop on Gender Research Priorities and Methods in Development (1992), National University of Lesotho.

Peterson, Lennart (1990), Lesotho's Economic Policy and Performance under the Structural Adjustment Program. The External Dependence, SIDA Macroeconomic Studies, No. 13. Stockholm.

_ (1991), Lesotho's Structural Adjustment Programs and Economic Performance, SIDA Macroeconomic Studies, No. 23. Stockholm.

(1992), SIDA Macroeconomic Studies, No. 35. Lesotho-Adjustment and Liberalization in Lesotho. Institute for Economic Research, Lund University. Stockholm.

Price Waterhouse (1990), Financial Intermediaries and Practice in Lesotho: Implications for the Agricultural Sector. - 245 -

Raditapole, D.K. and Y. Adu-Boahere (1991), Development of Framework for the Rationalizationof Lesotho GovemmentFinancial Support to the PHAL. Ministry of Health. Maseru.

Ravallion,Martin (1993), Poverty Comparisons,World Bank Report No. LSM-88. Washington, D.C.

Rubey, Lawrence(1993), ConsumerMaize Meal Preferencesin Zimbabwe:Survey Results and Policy Implications.Mimeo. Reportto USAID/Zimbabwe,Harare.

Sandier and Travis Trade AdvisoryServices (1995), Textile QuotaAnalysis Service. Washington,D.C.

Santho, S. (1988), Public Policy Issues and Contextof ContractFarming in Lesotho:the Case of Asparagus Production.Mimeo, Institute of Southern African Studies, National Universityof Lesotho. Roma.

Sechaba Consultants(1991). Povertyin Lesotho: A MappingExercise. Maseru.

(1991), Poverty in Lesotho: A Mapping Exercise, Volume II. Maseru.

(1993), Draft Baseline Survey for Small Scale Intensive Agricultural ProductionProject, Report preparedfor U.S. Peace Corps. Mimeo. Maseru.

(1994), Poverty in Lesotho, 1994: A Mapping Exercise. Government of Lesotho WorkingDocument. Maseru.

Senaoana,M.P., S.D Tumer., G.J. van Apeldoorn(1984), Research on RuralNon-Farm Employmentin Lesotho:Results of a BaselineSurvey. Maseru.

Senaoana,Moeketsi P. (1988), ContractFanning and OutgrowersScheme in Lesotho:the Case of BasothoFruits and VegetablesCanning Factory, interim report. Mimeo. Institute of SouthernAfrican Studies,National University of Lesotho. Roma.

Swallow,Brent M.; Brokken,Ray F.; Motsamai,Mabaitsi; Sopeng, Limpho and Storey, Gary G. (1987), Livestock Developmentand Range Utilization in Lesotho, Research Division Report RD-R-82, ISAS Research Report No. 18, May. Farming Systems Research Division,Ministry of Agriculture. Maseru.

Tangri, Roger (1993), ForeignBusiness and PoliticalUnrest in Lesotho,African Affairs, Vol. 92:223-38,April 1993.

Van der Wiel, Aric C.A. (1977), Migration Labour: Its Role in the Economy of Lesotho. MazenodBook Centre, Mazenod,Lesotho.

Von Braun, J., et al. (1991), Labor Intensive PublicWorks for Food Security:Experiences in African InternationalFood Policy ResearchInstitute (FPRI), WorkingPaper No. 6.

Vordzorgbe,Seth (1992), MacroEffects of the 1991/92Drought in Lesotho. Maseru:USAID. - 246 -

_ (1993), The Macroeconomic Environment for Poverty Alleviation in Lesotho: Tax Incidence Analysis. Maseru: USAID.

World Bank (1974), The Economy of Lesotho. Report No. 331a-LSO. Washington, D.C.

_ (1987) Lesotho Country Economic Memorandum. Report No. 6671 -LSO. Washington, D.C.

_ (1988), WorldDevelopment Report 1988. New York: Oxford University Press.

(1990), Lesotho - A Strategy for Industrial Sector Development, Mimeo. Washington, D.C.

(1990), Staff Appraisal Report, Lesotho- IndustrialandAgroindustries Development Project, Report No. 8990-LSO. Washington, D.C.

(1991), Staff Appraisal Report, Lesotho Highlands Water Project (Phase LA), Report No. P5533-LSO. Washington, D.C.

( 1(993)World Development Report 1993. Oxford: Oxford University Press

(1993), Agriculture Operations Division, Southem Africa Department, Lesotho: Environment and Agricultural Diversification Study, Report No. 11906-LSO. Washington, D.C.

_ (1993), The EastAsian Miracle - Economic Growth and Public Policy, New York, Oxford University Press.

(1994), Lesotho- PopulationSector Review, Report No. 12553-LSO. Washington, D.C.

(1995), Map # 27147.

(1995), World Bank Development BriefNo. 56, May 1995. Washington, D.C.

Better Health in Africa. Report No. 12577-AFR. Washington, D.C.

World Food Program (WFP/FAO/ILO) (April/May, 1993). Interim Evaluation Summary Report on Project Lesotho 0352 (Exp. 7) - Development of Rural Infrastructure with Appraisal of New Phase. Maseru.

Wright, P.D., and W.L. Nieuwoudt (1993), Price Distortions in the South African Maize Economy: A Comparative Analysis, Agrekon,32(2): 51-59. MAP SECTION

LESOTHO DISTRICTS AND AGRO-ECOLOGICALZONES

Buth he t Mt4OUNTAIN -ZONE

`^s Z BU.THABUTHE . \ SENQU- RVALLEY

FOOTHILLS

ERIBE ~~~~~~~~~~~~~LOWLANDS

XTeyateyonenq (

A 4BRA) W

4 g~~~~~~~~,~',,NSERUm MOKHOLAON-COGCCA ZNE

fTHA BA-TSE KA

\ - o'. MAFETENG ,4J&s F 9 j ~~~~~~~~~~~~INTERNATIONALBOUNDARIES

\ - n9 ~~~~~~~~~~~~~~~~~~~~AMROHA-= CG ECOL GC'AL ZNES

GRAVEL ROADS

S Mohohe'sHoek sj * NATIONAl CAPITA, /$4 j _QUTHING 0 AC m DSTRICT CAPITALS

rieso., Aiii9 ,,, , L Fi IF~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~F-air

z~~~~~1 11 C 2 i.yrotetit Fr ioi,,tc(~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ctn)1c)- I,,dqttteii rit5zel..go. u2 I

~~~~~ * | I _I - I | Nll_ W~~~~~~~~~~~~~.;

I 111 I .:

21~~~~~~~~~~~~~i ig I