Nepal - Relieving poverty in a resource-scarce economy (Vol. 1 of 2) : Main report

Report No. 8635-NEP Nepal Relieving Poverty in a Resource-Scarce Economy (in Two Volumes) Volume 1: Main Report August 1S, 1990 A Joint World BanklUND)I' Studv Country Department I Asia Region FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performnance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Currency Equivalents Year US$1.00 Equivalent (Average) 1984/85 Rs. 17.8 1988/89 Rs. 25.6 Fiscal Year July 16 - July 15 List of Abbreviations and Acronyms Used ADB Asia Development Bank ADB/N Agricultural Development Bank of Nepal AIC Agricultural Inputs Corporation CSI Cottage and Small Industries EGS Employment Guarantee Scheme ERL Environmental Resources Ltd. ha. hectare HMG Hia Majesty's Government of Nepal IBP Intensive Banking Program IRDP Integrated Rural Development Project kg. kilogram LF Labour Force MPHBS Multi Purpose Household Budget Survey mt. metric tons NFC National Food Corporation NGO Non-Governmental Organization NPC National Planning Commission p.a. per annum PCRW Production Credit for Rural Women Rs. Rupees SFDP Small Farmers' Development Program SSSR Social Sector Strategy Review UNICEF United Nations Children's Fund WDR World Development Report WFP World Food Program  FOR OFFICIAL USE ONLY ACKNOULEDGEEN HS This report was prepared on the basis of a mission to Nepal in November 1989 consisting of W. James Smith (Mission Leader), John H. Duloy (Consultant - Poverty and Incomes), Christopher Gibbs (Agricultural Economist), Lynn Bennett (Anthropologist), John Elder (Consultant - Agriculture, Food Security), Sanjay Sinha (Consultant - Informal Sector) and Neil Walton (Consultant - NGO and Income-Generating Programs). The analysis is based on a substantial amount of preparatory work that was done in Nepal; background studies were managed by Meena Acharya, Kul Shekhar Sharma, Madhup Dhungana and Harka Gurung. Thanks are due to the Nepal Rastra Bank for allowing us access to the Multi-Purpose Household Budget Survey Data, and particular recognition is due to Baskar Risal, who managed the reprocessing of the survey data. The report draws on numerous background papers and analyses, to Which the following persons contributed: Messrs./Mss. Ram Nath Acharya, Hamid Ansari, Aswasthama, Kishor Kumar Guru Gharana, Ana Maria Jeria, Chaitanya Mishra. Yogendra Prasai, Khem Raj Sharma, Azealia Ranjitkar, Prayograj Sharma, Ramesh Sharma, Salikram Sharma, Gajendra Man Shrestha, Subrama Laxmi Singh, Ganesh Bahadur Thapa, Indra Jung Thapa, and William Thiesenhusen. Irajen Appasamy provided computational support, and Deborah Ricks processed the report. Financial support for the study was provided by the United Nations Development Program. The report also benefited from the findings of a parallel study of poverty undertaken by the Canadian International Development Agency. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  NPAL BELEVIG POVERTY IN A IESOUE-SCAI UCONS Table of Contents Page go. Executive Summary ...................................... CEAPTER Is Introduction ............ 1 A. Objectives and Scope . ............................ . 1 B. The Nature of the Problem ......................... 3 CRAPTER IIs The Structure of Incomes and Poverty ......... .5 A. Levels and Distribution of Income in Nepal........ 5 B. Composition of Income. .................... . ... 9 C. Profile of the Poor. .................. 13 D. Conditions of the Pooroor.. ....... 16 CUAPTER Ills The Determinants of Poverty.............. 22 A. The Dynamics of Poverty- Output and Growth Issues............. 22 Population and the Resource Base................ 22 Economic Performance...... ......... 24 Macroeconomic Conditionstition.. ... ..... 26 Structural Adjustment .... .. .. . 26 The Trade and Transit Disputei........ 29 Emerging Trends....................... 29 B. The Mechanics of Poverty - Distributional and Equity Issuess.. ............... 32 The Social Context - Inequality at the Village Levl......e l . 32 Land Tenure and Tenancyn....... 33 Labour Force Issues ..... ... . 3? Migration ..............4 Debt and Indebtedness ........... 45 Gender and Poverty ......... 41 Social and Political Constraints................ 49 CHAPTER IVs Poverty and the Productive Sectors .......351 A. Agricultural Incomes andPoverty.................. 51 Poverty and the Environmenti....... 62 B. Off-Farm zmployment and lncotes...... 63 The Formal Sectore ........... 63 The Informal Sectore .........67  CHAPTER Vs The Social Dimensions of Poverty ...................... 71 A. Nutrition and Access to Food...................... 71 B. Education and the Poor ......................... ... 76 C. Poverty, Population and Health.................... 83 CHAPTER VI: Poverty Related Programs and Policiesi......8.......... 88 A. HMG's Policies and P rogra 8..... General Policies. ...........e........ ..... 88 The Composition of Public Expenditure........... 88 The Basic Needs Program and the Eighth Plal................... 90 B. Poverty Alleviation Programs ... 91 Subsidies and Transsr.. 91 Integrated Rural Development Projects.. 94 Food and Feeding Progras . .. 96 Targetted Credit ......... . . ..... 98 Employment Creation Projects .................... 102 Income Generating Projects .................... . 103 C. Institutional Issues ..... .. ..... 104 Civil Service and Service Delivery Issues ........................................................ 104 Decentralization and Popular Participation ...... 105 Non-Government Organizations . .............. 106 CHAPTER VII: Country Strategy Implications ........ ......... ......... 108 CHAPTER VIII: Policy Conclusions . . . 115 APPENDIX TABLE: Summary of Detailed Recommendations ................ 122 VOLINE It - ANNEXES Annex I - Sectoral Background Papers Annex II - Poverty Program Reviews Annex III - Statistical Appendices Annex IV - Bibliography  List of Tables Table 1.1 Indicators of Relative Underdevelopment-Selected Countries .... ........................................ 3 Table 1.2s Structural Economic Indicators ......................... 4 Table 2.1: Average Per Capita Monthly Income by Income Deciles.... 5 Table 2.2s Distribution of Total Household Income in Nepal and Selected Countries ......... . . 6 Table 2.3s Per Capita Household Income Distribution in Nepal ...... 7 Table 2.4: Incidence of Poverty Under Different Poverty Lines..... 8 Table 2.5: Composition of Household Income-1984185 . . 10 Table 2.6: Some Survey Estimates of Income Composition ............ 11 Table 2.7s The Effect of Farm Size on Incomes ..................... 12 Table 2.8: Returns to Family Labor Per Day Worked By Farm Size.... 13 Table 2.9: Nepal - Poverty Profile .. 14 Table 2.10: Monthly Consumption Pattern of Poor Households... 17 Table 2.11: Food Sources and Consumption Among the Poor-Rural Terai 8......... Table 3.1: GDP Composition and Growth ............................. 25 Table 3.2: Nepal - Structural Economic Indicators ................. 27 Table 3.3: Size Distribution of Operational Holdings . . 33 Table 3.4: Estimated Current Daily Wage Rates (1989) . . 39 Table 4.1: Characteristics of Farm Household Groups in the Hills ................................ . 55 Table 4.2: Potential Monthly Per Capita Income From Field Crops and Percentage of Poverty Line Income For Poor Households, 1988189 ................................ .. . 57 Table 4.3s Household Income and Landholdings in the Terai by Level of Poverty .... .. ...... ....... . 58 Table 4.4s Potential Income Per Hectare from Field Crops in the Terai and Area Needed for Poverty Line Income, 1988T89 and.Pro.ecte..Forml.Sector.................... 59 Table 4.5s Current and Projected Formal Sector Employment ................................ 66  Table 4.6: Estimated Informal Sector Employment - 1990 ............ 67 Table 5.1s Regional Per Capita Production of Grains and Potatoes . . *. . . . . .4 . . . . . . . .0 . . . . . .* . *. . **....... n 7 Table 5.2: Population Growth Rates (1971-1981) and Growth Rates of Agricultural Production (1967188-1987188) ......................... .. ........ 74 Table 5.3S Projected Daily Per Capita Availability of Calories from Grains and Potatoes ................................... * 75 Table 6.1: Composition of Public Expenditures - 1988189 ........... 89 Table 6.2: Nepal - Poverty Alleviation Programs at a Glance ....... 91 Table 6.3s Targetted Credit Programs ....................... 99 Table 7.1: Possible Labour Absorption - 2010 . ...............t 110 Table 7.2: Potential Impact of Various Developments on the Incidence of Povertyo...e rt...*.....t.*.*. . 111 List of Figures Figure 1: Composition of Income Among the Rural Poor ............. 9 Figure 2: Effect of Population Growth on Income and Food Availability .......................... .. ............. 23 Figure St Actual and Projected Population Growth 1910-2030 ....... 29 Figure 4s The Changing Spatial Distribution of Population ........ 30 Figure 5: Annual and Trend Variation in Food Production .......... 74 Figure 6: Income, Gender, and Education Participation ............ 77  - i - NEPAL LV POVERTY IN A REUR-SCARCE ECO Bxecutive S _ rr 1. This study seeks to build a better understanding of the nature of incomes and poverty in Nepal, and to propose an affordable set of measures to reduce the incidence of poverty. It consists of three main elements: - a quantitative analysis of the conditions of the poor and the non-poor, their sources and levels of income; - an analysis of the micro-economic factors influencing incomes, the constraints facing the poor, and the potential contribution of various sectors to raising personal incomes; and, - an appraisal of the effectiveness of existing poverty alleviation programs and projects. It concludes by outlining the elements of a medium- to long-term poverty alleviation strategy for Nepal. 2. By the most conservative definition, between 7 and 8 million of Nepal's population of 19 million live in absolute poverty, defined as having incomes below the level required to support a minimum daily calorie intake (about US$100 p.a. per capita). The poor are overwhelmingly rural subsistence farmers. They earn about half their incomes from their own agricultural production (almost none of which is marketed), they earn less than 302 from employment - mostly on-farm, and the remainder of their incomes is made up from miscellaneous subsistence activities. Although there has been some shift towards off-farm activities, the limited resource base, and lack of alternative opportunities resulting from Nepal's landlocked position beside an economically dominant neighbor, have limLted non-agricultural employment growth. 3. Population growth of 2.7Z p.a. has eroded the limited gains that have been made in GDP and agricultural output. The population has doubled sLnce 1960, and is projected to double again over the next 25 years. Furthermore, the demographic profile is such that within ten years the labour force will be growing at about 0.4 million persons per year - twice the average rate during the 1980's. The major challenge in avoiding a deterioration in the poverty situation will be managing the absorption of this massive labour force growth. In the absence of an effective program to slow population growth, all other measures will be meaningless. 4. At the very low level of average GDP, raising personal ineomes for most Nepalese depends on overall economic growth. Given the limited cultivable land base, agriculture alone cannot ultimately be counted on to provide the solution to poverty in Nepal. The basis for long-term growth, if it is to come at all, must thus eventually be sought ln the expansion of services, energy, and industry. For these it seems nevitable that Nepal  - iI - will have to look to a large extent to greater participation in a growing Indian economy. However even with the best policies and most robust external environment, industrial sector growth will be a very long term proposition. In the medium-term increasing incomes will have to rely in large part on agricultural intensification, and agriculturally-led growth in the informal sector. S. The agricultural land base is rapidly approaching saturation. There is, however, scope for increases in both labour absorption and agricultural productivity - largely through improved irrigation. About half of the agricultural poor could rise out of poverty on the basis of increased productivity, however to reach them will require a more subtle blend of agricultural interventions than has been tried to date. For the balance, their holdings are too small to ever be viable economic units - they need off-farm income-earning opportunities - either where they currently are, or elsewhere. 6. Even under the most optimistic assumptions, the formal sector will not absorb more than about 15-201 of the labor force by 2010. The informal sector holds more promise, since there is room for a 'catch-up, effect to compensate for the low levels of physical access and monetization In the past; although ultimately it must follow growth led by the other sectors - especially agriculture. However, wages are close to a subsistence minimum, and at existing wage levels the poor will not rise out of poverty on the basis of employment alone - that will require a tightening of the labour market. While a successful population program can contribute, in the long run the scope for effecting labour market factors may be limited by the free flow of labour to and from India, which will tend to lead to wages equalizing at Indian levels. Given the constraints on effective labour market interventions, Government can best contribute to off-farm employment growth through enabling mechanisms: providing transport and communications infrastructure, education, and to a lesser extent skills training and credit. 7. In the absence of any obvious source of rapid economic expansion, curbing population growth is central to relieving poverty. However, at any expected levels of populatior. and GDP growth, average incomes will not rise enough to have a major impact on poverty levels unless such growth is focussed to benefit the poor. Furthermore, forces are in place - through growing landlessness, monetization, and urbanization - for a deterioration in the distribution of income, unless policies are followed which consciously guard against it. The type of growth pursued must be balanced in such a way as to generate incomes for the poor (in the medium term this means largely increased agricultural productivity and labor absorption). 8. Poverty in Nepal is chronic, basically rooted in the insufficiency of the resource base vis. a vis. excessive population. Its solution will lie in productivity growth coupled with population control, but this will take a long time. In the meantime, there will remain a large number of absolute poor. It is therefore legitimate to consider a sustained program of support to the poor - some of it production-oriented and some of it welfare-oriented.  - iii - 9. There are too many poor (and too few resources) to realistically consider large-scale transfer or subsidy programs. Therefore it is important to design cost-effective transfers. Both Government and the donors, under a wide-range of projects, are already putting substantial resources into poverty-type programs - to relatively little effect. Better targetting and institutional strengthening are needed to improve the efficiency with which those resources are used. In addition, some significant welfare improvements for the poot can be achieved without major resource transfers or income increases (through, for example, improved hygiene and nutrient retention). 10. The report concludes that there is no easy poverty alleviation strategy for Nepal, but that significant gains can be made through a combination of measures - mostly involving increased labour absorption in agriculture (which could increase by about 50%), coupled with productivity gains in low-input farming systems; informal sector growth; and some redistributive measures, if tightly focussed. 11. There are also some areas (eg., tenancy, labor contracting) where policy reforms are required. In addition, limited gains can be made at the micro level, by strengthening the capacity of the poor to undertake self- reliant income generation - although these efforts require labor-intensive inputs. Given the weakness of service delivery mechanisms these are probably best delivered through non-governmental organizations. The distributional benefits of some aid can be improved without much efficiency loss, by channelling it down to lower 'evels, for example through NGO's and targetted credit programs. 12. While there is no easy solution to poverty in Nepal, the potential contribution of public policy is large. As an illustration - if the Government is able to achieve the best reasonable expected results from an effective family planning program and in sustaining economic growth, then per capita GDP could rise to about $270 equivalent per capita by 2010 - probably holding the number of poor to below 5 million. On the other hand, if population growth continues unabated, and GDP growth is no better than the average achieved over the last twenty years, then per capita incomes would stagnate at around $180 per annum - probably increasing the number of poor to over 20 million by 2010. In short, the cost of not getting public policy right - especially with respect to curbing population growth - is probably in the order of 15 million additional absolute poor. 13. The priority elements of a poverty alleviation strategy in Nepal need to consist of: (i) an effective program to curb population growth - through the promotion of temporary methods of contraception; (ii) an agricultural program that includes: small farmer irrigation (and measures to get all irrigation working in the terai), input supply deregulation, development of a technical package free of purchased inputs for inaccessible hill farmers; horticultural support for accessible small farms in the hills;  - iv - and design of an outreach cum extension program that takes account of the constraints facing poor farmers; (iii) a program of increased rural access in the terai and selected areas of the hills; (iv) intensification of basic education, including: (a) a national literacy campaign; (b) revamping of curriculum and examination systems to improve relevance and quality; (c) measures to lift constraints on attendance by girls and by the poor; and (d) agreement on coimitment of adequate financing for expansion of the primary school system to provide almost universal coverageg and, (v) in addition, there is scope for a package of low-cost measures to help the large numbers of those who will remain absolutely poor for the foreseeable future - through selectcd off-farm income gen-erating activities and improvements in health, nutrition, and access to food (A program of such poverty alleviation measures is described in paragraph 16, below). 14. Conversely, there are a number of areas in which the Government should probably exercise caution, and a further sub-set of areas in which action is justified, but where further analysis is needed first. The following are areas in which, on the basis of the analysis, it appears that public interventions would not be a cost-effective means of helping the poor: - large-scale infusions of targetted credit; - price controls or commodity subsidies; and - excessive intervention in the industrial sector. Those areas in which further analysis is needed to formulate effective programs includes - partitioning the hills for poverty alleviation purposes into areas for varying degrees of support, depending on the efficiency of providing access at reasonable cost; - cost-effective measures to improve access to food in remote areas; - analytical work to develop a strategy for managing the transition of population from the hills to the terai and to urban areas; - an assessment of the scope for effective measures to promote income-earning opportunities in the informal sector; - investigation of the scope for, and implications of reforms to th-e land tenure system, to improve both efficiency and equity.  - improvements in the design and implementation of income- generation and skills training projects. 15. Pervasive poverty in Nepal stems from four factors: a limited resource base, its physical location between two large countries - both of which are also poor, rapid population growth, and poor economic performance - with GDP growth averaging under 31 p.a. over the last 25 years. Little can be done about the first two factors, however effective policies can reduce population growth and accelerate economic expansion. Unfortunately Government initiatives in these areas to date have been partial and ineffective. The Government now has an opportunity to play a more aggressive role - by pursuing a very active population program, and laying the basis for more rapid growth, through, inter alia widespread public service reforms, accelerating education and skills training, further liberalization of trade and industrial restrictions, a more aggressive agricultural grow-h strategy, and reform of the financial sector. Additional measures are required to ensure that the poor specifically benefit, and these are discussed in the following sections. 16. The report concludes that even with the best expected performance in economic growth and family planning, there are likely to remain some 5-10 million absolute poor over the next twenty years. It therefore proposes, in addition to the broad priorities for creating equitable growth outlined above, a set of measures for Government and donor support specifically to assist those who will inevitably remain poor, consisting of the six elements outlined below: (1) Expansion of Selected Transfer Programs. These are large-scale rural employment programs which, in the absence of sufficient growth in the medium-term, can provide cost-effective transfers while at the same time creating productive rural assets. Two such programs exist in Pepal - Food for Work, and the Special Public Works Program - with some strengthening they could be amenable to substantial expansion. The next step would be to determine how much they could expand, their financial requirements, to identify measures to improve targetting and efficiency, and agree on a financing package for them. (2) Incremental Improvements to Existing Income-Generating Programs. There are a number of smaller programs which have had some success in raising incomes of the poor (including the Small Farmers Development Program, Production Credit for Rural Women, and localized NGO income- generating projects). 'While these can never have sufficiently wide coverage to make a major reduction in the incidence of poverty, they should be expanded as staffing constraints allow. There is also room for incremental improvements to their operations, including more intensive training of staff and better sub-project evaluation. (3) Food and Food Aid Programs. Enough areas have sufficiently large food deficits, which cannot be solved by out-migration or agricultural productivity increases in the near term, that food aid and food distribution programs will be an important continuing part of any poverty alleviation strategy. The objective now should be to identify the most cost-effective combination of interventions (eg. relying on increased local  - vi - production, improved feeding practices, etc. whenever possible), and to rationalize existing programs, which often do not reach the intended beneficiaries. This would includet - agreement on the appropriate level and form of food aid, its financing, and how best to distribute it; - reform of the National Food Corporation's program - eliminating mis-targetted subsidies, identifying which distribution measures reach those suffering food deficits, and how they should be expanded and financed; and, - identification of a program of effective interventions - including evaluating the relative roles of food distribution, agricultural productivity measures, food storage, vulnerable group feeding, and promotion of effective practices (e.g., weaning foods, feeding during pregnancy, etc.); followed by preparation of specific projects and agreement on financing for them. (4) Strengthening Welfare Measures. These are interventions which could substantially improve the welfare of the poor even in the absence of increased incomes. They include: - a national hygiene campaign - which would combine intensive hygiene-awareness education with the provision of rural water supplies; - a range of particularly cost-effective health measures - (immiunization, iodine supplementation, and oral rehydration salts) - these are existing programs which could be expanded at modest cost without awaiting major improvements in the (institutionally weak) health service; and, - (possibly) targetted nutrition and/or feeding programs - these can have a dramatic effect on reducing the long-term effects of malnutrition at affordable cost. Such programs usually rely on an administrative capacity which may not be present in Nepal; however, it is worth exploring the possibilities. (5) A Package of Policy Reforms. Tenancy Reform - a redistributive land reform would have some impact in the terai, but may not be feasible. However, the existing tenancy system provides many disincentives to maximizing output. It is possible to design a reform which at a minimum would regularize land tenure and remove barriers to efficient use of land. Thib, in the terai, could have a large impact on agricultural productivity. Labor Contracting Arrangements - current arrangements are exploitative beyond what would be expected on the basis of low average wage rates alone; some improvements would be possible by introducing revised procedures for public works (eg. the use of small local contractors) and increased monitoring and supervision.  - vii - Decentralization and Civil Service Reform - effective implementation of the decentralization process can potentially help the poor by improving the delivery of social and development services to peripheral areas (by devolving staff and resources to them). NGO Regulation - the current system constrains NGO activity and probably adds to their costs. Government should explore the scope for relaxing the regulatory framework to facilitate NGO operations - along with measures to improve the technical support provided to them (for instance, to improve project selection and design). (6) Analytical Work On Priority Areas. This includes assembling data on incomes, wages and employment, and further analysis needed to design policy and program reforms.  NEPAL - RELIEVIN POVERTY IN A RBSOUC-SCARCE EONOS I. Introduction A. Obiectives and Scope 1.1 This study is intended to deepen our understanding of the nature of poverty in Nepal; of its causes, and of the constraints which prevent the poor from improving their conditions. It investigates the effect of development policies and strategies on personal incomes, and seeks to identify the most promising areas for raising incomes of the poor. The objectives are to propose the outlines of a long-term country strategy to reduce poverty. as well as to recommend specific measures for Government and donor support. 1.2 The study came about as the result of a number of perceptions: Firstly, that extreme poverty is a widespread problem in Nepal, but that little analytical work had been done on income levels and trends, or the dynamics at work among the poor. Secondly, that there exists a desire within the donor community, and within the World Bank, to do more to support poverty alleviation efforts in Nepal, but that it has been difficult to identify which measures might be amenable to large-scale financial support. Thirdly, Nepal is a very poor country in which the Bank and the donor community have made substantial investments (presumably with the objective of raising living standards), but with little understanding of how our programs affect the incomes of the poor. The study seeks to better inform overall country assistance by investigating the links between incomes and various agricultural, industrial, and employment activities. Fourthly, the Government is already putting substantial resources into programs notionally justified on poverty-alleviation grounds - the analysis looks at ways these resources could be used most cost-effectively. Finally, in Nepal there have been many good village-level and sectoral studies, each of which tells an important part of the poverty story, but the isolated findings have never been pulled together. 1.3 The study has sought to assimilate the findings in various sectors, and build them into a wider analytical framework. The objectives are: (i) to outline the bounds of what might be possible (and expected) in the evolution of incomes over the next 20 years, and to evaluate the relative contributions of agriculture, industry and commercial development, and social services;  - 2 - (ii) to suggest sectoral interventions, or changes in approach, which may have a greater impact on incomes. especially among the poor; and, (iii) toi propose a menu of cost-effective poverty alleviation projects for possible support. 1.4 The report is organized in two volumes. The first presents a summary of findings, and attempts to fit them into a cohesive framework. The second contains: (i) more detailed papers on each of the main issues addressed (eg. agriculture, land tenure, food security, etc.); (ii) a review of the main poverty alleviation programs in Nepal, their successes and failures; and (iii) the statistical appendices. 1.5 In the main report, Chapter II analyzes the levels, distribution and sources of income. It presents for the first time in quantitative terms a profile of poverty in Nepal, characteristics of the poor and non- poor, and the conditions under which they live. This analysis 4s based on special tabulations of the Nepal Rastra Bank's household survey data undertaken as part of this study. Chapter III explores the causes of poverty - both those related to aggregate output and growth, and the more subtle mechanics of land tenure, employment, and social structures. Chapter IV looks at linkages between the productive sectors (agriculture, industry, etc.) and incomes, and the role of non-formal employment. Chapter V examines the social dimensions of poverty: food security, education, and health and population issues. Chapter VI briefly reviews His Majesty's Government's (HMG's) general policies, and summarizes the assessment of poverty-related programs contained in Volume II. The final two chapters suggest implications for country strategy and donor support. 1.6 Of necessity the report takes a longer-term perspective (20-30 years) than is usual in World Bank documents. It also covers a wide range of sectors. In not all of these were we equipped to bring our conclusions to closure; in some we could only highlight the issues involved and point to possible solutions, or areas for further work. 1.7 Other reports on poverty have tended to focus on social services and their delivery. These issues have been explored at length in an earlier Bank study.l/ We have concentrated instead on the objective of raising personal incomes and consumption.2/ While this may provide a somewhat one-dimensional picture, it was necessary to limit the scope of the work to something manageable, and furthermore raising personal incomes (broadly defined) is, after all, the single best measure of reducing poverty. 1/ The World Bank; Nepal - Social Sector Strategy Review, 1989. 2/ Incomes are taken throughout to include production of food for own consumption - which constitutes by far the largest share of personal income in Nepal.  - 3 - B. The Nature of the Problem - Poverty in Nepal in a Comparative Context 1.8 Average personal incomes are about Rs. 3,340 per capita (1984/85, expressed in 1988/89 prices) - equivalent to US$130 annually. Approximately 40Z of the population live in absolute poverty, defined as having less than the income required to consume a minimum bundle of calories on a daily basis. There are also chronic and seasonal food deficits, vhich probably affect half of the population. Large scale starvation is avoided because of the absence of catastrophic droughts, although slow physical deterioration through malnutrition is widespread. 1.9 In comparative terms, Nepal is one of the world's poorest countries. It ranks 115th in per capita GNP out of 120 countries listed in the 1989 World Development Report (WDR). With respect to life expectancy, Nepal ranks 103rd out of 118 reporting countries. Food availability, measured in calories per capita, is worse only in Bangladesh. Haiti, and a handful of African countries. Table 1.1 presents key development indicators for Nepal, relative to some of the world's poorest countries, and relative to neighboring states in Asia. Table 1.1: Indicators of Relatlve Underdovelopment - Selected Countries 9/ Calorie Life Por Capita Infant Population AvailabiIlty Expoctancy GNP Mortality Growth (Calories per (Years) (US8) (per 000) Rato capita per day) Ethiopia 47 180 164 S.1X 1749 Mall 47 210 169 a.O% 2074 Nepl 51 160 128 2.SX 2052 Bangladesh 51 180 119 2.4X 1927 ZaIre 52 1S0 98 8.1X 2163 India So 800 99 1.8% 2238 Thailand 64 8S0 39 1.51 2331 Sri Lanka 70 400 83 1.1X 2401 j/ According to the 1989 World Development Report (WDR). For consistency of inter- country comparisons all data are drawn from the 1989 EDR; these may not be consistent with estimates for Nepal cited elsehere in the text, which were drawn from more recont surveys and reports. 1.10 Nepal is one of a constellation of countries characterized by rapid population increases, low or negative per capita GDP growth, and a slow transition out of a subsistence agricultural economy. It exhibits many of the characteristics of similar sub-Saharan African economies, including a limited productive land base, a land-locked location, and a very low level of exports. Table 1.2 illustrates some of the structural features of Nepal, relative to other countries.  - 4 - Table 1.2: Structural Economic Indicator. Por Capita Come_rcel QNP Shere of 'or Capita Energy Consump- Per Capita Growth Rat. Arculture Merchandise tion (kg. oil GNP 1986-87 In (P Exporte equivalent) Very Poor Countries Ethiopia 8130 0.1% 425 89.00 21 kg. Moll 8210 0.1X 545 827.70 24 kg. Zaire 8150 -2.45 32t 848.90 73 kg. NepaI 8160 0.65 67X 88.60 28 kg. Neighboring Asian Countrioe Oangladbeh 8160 0.38 45X 810.10 47 kg. India 8300 1.8X 380 816.70 208 kg. Thailand S860 3.9X l8 8217.60 330 kg. Sri Lanka 8400 3.01 27X 889.90 160 kg. All date are for 1987 unless otheraise indicated. 1.11 Nepal's share of agriculture in GDP, one of the best summary indicators of the level of development and structural change, is, at 57?, higher than all but three of the very poorest countries (Uganda, Somalia and Tanzania). Similarly, the share of the labor force dependent upon agriculture (93Z) is the highest of any country listed in the 1989 WDR. Nepal is also one of the two lowest consumers per capita of commercial energy - a sensitive indicator of industrial and urban development. Nepalese consume 23 kg. of oil equivalent annually compared with 297 kg. for low income countries as a whole. 1.12 There is a very high cost in human suffering and wasted potential associated with Nepal's degree of poverty. For example, about 6 1/22 of the population is estimated to suffer some degree of mental retardation, due mainly to malnutrition and iodine deficiency. Again, by the Indian Academy of Pediatrics standard, about 29Z of the population exhibit 2nd or 3rd degree malnutrition. Associated with nutritional deficiency and diseases, there are estimated to be about half a million cases of blindness in Nepal (out of a population of about 18 million in 1988), a rate far in excess of international standards.  II. THE STRUCTURE OF INCOMES AND POVERTY IN NEPAL A. Levels and Distribution of Income 2.1 Average per capita incomes are estimated at about Rs. 260 per month in 1988/89 prices (US$122 p.a.) in rural areas; and about Rs. 426 (US$200 p.a.) in urban areas. In rural areas, the mean income in the hills is about 202 lower than in the terai;l/ in urban areas, reflecting in particular the relative wealth of Kathmandu, the order is reversed, with average incomes in the hills about one quarter higher than in the terai. 2.2 Within Nepal, the best source of data relating to incomes is the Multi-Purpose Household Budget Survey (MPHBS) conducted in 1984-85 by the Nepal Rastra Bank.2/ The analysis in this chapter is based on a re- processing of the KPHBS data undertaken by the World Bank specifically for this study to identify the characteristics of poor and non-poor households. Average per capita monthly income levels by decile are presented in Table 2.1. These data are expressed in terai-equivalent prices, so that they can Table 2.1: Average Per Capita Monthly Income by Income Deciles (1988/89 Rs per month) RURAL URBAN Decile Terai Hills Terai Hills 1 123 85 162 171 2 164 116 203 229 3 188 140 235 281 4 207 163 262 334 5 232 188 294 395 6 262 217 343 453 7 295 253 396 543 8 348 299 493 644 9 435 368 675 830 10 719 564 1196 1394 All Families 292 231 378 475 Source: Mission estimates derived from MPHBS. 1/ The terai is a narrow lowland belt along the Indian border containing about half of Nepal's population. Incomes here are adjusted for differences in purchasing power; consumer goods flow mainly from the terai to hills and transport costs are very high. Mission estimates of the price differentials, based upon a basket of eight staple foods, suggest that prices are about 32Z higher in the hills than terai. 2I The survey is described in Annex III. While the MPHBS is an exceptionally well-designed and well conducted survey, it suffers from the inability of all such surveys to capture fully the exceptionally rich and the exceptionally poor. For this reason, the actual income distribution is more skewed than that revealed in the tabulations.  - 6 - be compared across hills and terai. Incomes are uniformly low except in the top decile - and are even then only about two and a half times the average. 2.3 Like other very poor countries, Nepal's predominantly rural economy sustains a relatively even distribution of income, corresponding to a stage of development before the income and wealth-concentrating process of urban-based development has had much impact. From Table 2.2 it can be seen that Nepal has c; degree of income concentration somewhat more favorable to the poor than Sri Lanka, somewhat less so than, but comparable with, Sweden, and markedly more egalitarian than the distribution of income in Brazil. Table 2.2: Distribution of Total Household Income in Nepal and Selected Countries (Percentage of Share of Income) Household Group Nepal Sri Lanka Brazil Sweden Bottom 40S 182 162 7? 21? Middle 502 /a 542 49? 422 51? Top 102 282 35X 512 28Z a/ That is, lying between the 41st and 89th percentiles. Source: Nepal, MPHBS. Other countries, World Development Report, 1989. 2.4 These comparative data suggest that the scope for improving the income position of the poor in Nepal by redistribution from the upper income groups is limited. For example, in Brazil, transferring 52 of the income of the upper decile to the lowest 40? would involve an income increase for the latter of about 362, an amount which would have a significant impact in reducing the level of poverty. In Nepal, by contrast, transferring 5? of income from the richest 102 to the bottom 40? would increase the average incomes of the latter group by less than 8?, a welcome improvement no doubt, but not one which would produce a major impact upon poverty levels. At the same time it should be noted that the position of the poor in Nepal could be worsened greatly by a deterioration of the distribution of income. As discussed later in this report, there are reasons to believe that the mechanisms exist for such a worsening, which can only be countered by sustained off-setting policy interventions. 2.5 The distribution of per capita household income is presented in Table 2.3. As expected, it is less skewed than that of total household income (Table 2.2). (This is because household size increases consistently with total income, from less than 4 persons per household in the lowest income class to over 13 in the highest). The scope for poverty alleviation by redistributive measures alone is, therefore, even less than that  - 7 - suggested by the cross-country data contained in the previous table.3/ Table 2.3: Per Capita Household Income Distribution in Nepal (percentage share of income) Rural Urban Income Group All Nepal Terai Hills Mountains Terai Hills Bottom 40? 232 241 232 33Z 27? 242 Middle 50? 54S 532 56Z 542 522 562 Top 102 23? 231 212 13? 212 20? Source: Multi-Purpose Household Budget Survey. 2.6 The distribution of income is essentially the same in all regions except the mountains where it is remarkably flat, with poverty being shared by almost the entire population. However, the proportion of Nepal's population which lives in the mountains is very small, and mountain people are included within the hills for much of the subsequent discussion. The Incidence of Poverty 2.7 Table 2.4 shows the proportion of the population below the poverty line under three different sets of assumptions. The first is the poverty line defined by the National Planning Commission (NPC) on the basis of an income needed to supply minimum caloric requirements. In 1988/89 Rupees, this translates into Rs. 210 per person per month in the hills and Rs. 197 in the terai, or the equivalent of US$8.24 and $7.73 respectively per person. With this poverty line, the data suggest that about 40? of the population is in absolute poverty.4/ That this is a very conservative definition of poverty can be seen by considering the implications of a poverty line equivalent to US$150 per capita per annum.5/ The proportion of the population with incomes below even this modest level is about 70? nation-wide, and almost 80? in the hills. This calculation suggests that calorie-based poverty lines may not be appropriate for Nepal, a matter which is given additional support by the fact that the usual relationship between food intake and measured under-nutrition in Nepal is weakened by the widespread prevalence of intestinal disorders (see Chapter 5j. 3/ Note, however, the reservation contained in the earlier footnotes the data, like those from other such surveys, under-represent households at the highest income levels. 4/ The incidence data are approximations only. Note that the imputed rental value of owned housing was not included in total expenditures in these calculations. 5/ An accepted international definition of absolute poverty. (1988-89 rupee equivalent used; average exchange rate Rs. 25.5/US$l.00).  - 8 - 2.8 A third possible definition is one suggested by Lipton.6/ He has defined the poor as those whose food expenditures absorb 70Z or more of total expenditures. This criterion underlies the third poverty line in Table 2.4, in which the Nepal-wide incidence of poverty is about 66?. similar to the incidence implied by a poverty line of US$150 a year. Table 2.4: Incidence of Poverty Under Different Poverty Lines (Percentage of Population Below Poverty Lines Terai Hills All Nepal Poverty Line Rural Urban Total Rural Urban Total Rural Urban Total NPC 29X 172 28Z 55Z 132 52Z 42X 15Z 40Z US$150 692 51 68? 78X 32? 75? 74X 42Z 71X Lipton 70? 50? 68Z 65? 52Z 64t 68? 51? 66? 2.9 The sensitivity of the number of poor to the choice of poverty line results from the flat distribution of income around a very low average. Thus moving the line up or down only slightly results in large changes in the share of the population above or below it. Within the limitations of the data it is thus not very meaningful to talk of the precise number of poor. The best one can conclude is that the incidence of absolute poverty is very high - in the neighborhood of 50-601. In subsequent discussion we will use the (conservative) NPC poverty line.71 By this measure about 8 million Nepalese live in absolute poverty - out of an estimated total population of 19.2 million. 2.10 While in the following sections reference is made to the "poor" and the *non-poor", it is important to recognize that by any reasonable International standard everlone in Nepal is poor, except for a few professionals and businessmen, and perhaps some large farmers. The average income in the second decile (i.e. the second richest 10 of houlseholds) for instance is only about Rs. 500 per capita per month (US$20 per month) - in most countries these families would be among the very poorest. 2.11 It is often useful to distinguish between the poor and a subgroup of "ultra-poor, - at whom one may want to target assistance or development. In Nepal such a distinction is not particularly useful - both because there are so many very poor (eg. about half the population), and because most of the poor exhibit characteristics usually associated with the ultra poor (eg. high proportions of income spent on basic cereals). However, while they are frequently hungry, they are not ultra-poor in the longer-term sense that 6/ M. Liptons Poverty, Undernutrition, and Hunger; World Bank Staff Working Paper No. 597, 1983. 7/ In purchasing power parity terms this is estimated to be equivalent to about US$328 per capita p.a. - in the middle of the range of poverty lines cited in the 1990 World Development Report ($270-$375).  - 9 - they are an under-class caught in a chronic trap, out of which they could not raise themselves with access to technology and sufficient improvements in productivity. There is, however, a growing group of such ultra-poor emerging among the urban population, and among the landless in the terai. B. Composition of Income 2.12 Figure 1 illustrates the overall composition of income emerging from the MPHBS. Of note are the overwhelming importance of subsistence activities and of agriculture, and the relative unimportance of wage incomes. Figure 1. Composition of Income Among the Rural Poor Vm s(3296l -m Wages (24%) (;;z*1 % 17SE Otwt h4ndW t Coh I8% 6% WgS~~~~~~~~0 -'a'vS Coe Om f Ca1fl5% 1% Nonftwa 4 (C~~~~~~~~~) 4% ~ ~ ~ ~ ~ (40)F C4oh 9% Agdue (50%) b-4Gnd 41% AQIjAM (54%) - aI0 46% a Moslv emmltaoe perps. etC. bl Mo innea dws&o and cottae psdufon for own coamPm. * C0h woorm 2.13 Table 2.5 shows the composition of income for different groups. The following particular characteristics emerge: - among the rural poor, only 35Z of incomes arp in cash (32S in the terai) - coming mostly from wages; - the poor get a larger share of their incomes from wages and salaries than do the non-poor (who make much more from agriculture, especially in the terai); - even then wages and salaries account for only a quarter of income among the rural poor in the hills (a bit less than a third in the terai);  Tab:e 2.S: Coemo ltion of Household Income - 1984-85 (Re. per household per month) R UR AL URB9A N Toral Hills Tersi UHills Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Agrlcultu1 ' Cash 64 (S) 896 (27X) S6 ( 6) 282 (183) 20 (43) 90 (8s) S2 () 83 (43) Kind 28S (413) 884 (481) S80 (481) 828 (43X) 180 (23%) 271 (243) 197 (24%) 220 (12X) Subtotal 849 (603) 1082 (73X) 383 (64X) iWO (813) 188 (26%) iii (268) 248 (313) i3 (183) Non-Agricultural 1 Enterprises 29 (43) 60 (53) 28 (43) 121 (63) 106 (1SrJ 877 (273) 7 (IO0) 472 (25X) Waes" nd Salaries Cash 106 (13) 107 (7X) 121 (163) 209 (143) 300 (483) 417 (803) 860 (483) 721 (883) Kind 120 (173) 62 (143) a9 (63) 19 (13) 84 (5X) 21 (2X) 12 (1X) 19 (13) Suwotao I (82X) TR (113) 11 (243) H (16X) 3I (4814) 41 (32X) in (443) 7MU (393) Other Caoh Inco 24 (43) 88 (43) 80 (53) 96 (7X) 80 (4%) 148 (10%) 4e (6X) 839 (16X) Other Incom, In-kind / 7 71 (10O) 108 (M%) 91 (14X} 119 (8X. 47 t7X) 59 t4X) 83 6lo) 8 es c Totel 7e 7 TM03 4iai (103) iTH (IO0X) ii42 T1003)70(100X) !i79 (1003) 1R1 (1003) i1ii T10o3) Of Which: In Cash: 224 (823) 648 (433) 286 (853) 679 (47%) 464 (8S5) 1028 (74%) 628 (843) 1808 (843) In Kind: 47t (6X) 645 (5711) 487 (O65 778 (58X) 245 (83) 857 (286) 292 (386) 816 (16X) !I Almost all of which Is In cash. / Hos production for own conswmption plus rental value of rent-free dwellngs. / Excludes rental value of own how.  - 11 - - even in urban areas wages account for less than half of the incomes of the poor; - a significant portion of wage incomes of the poor is in kind rather than cash (over balf in the terai); _ cash incomes from agriculture (representing sales of foodgrains, cash crops or livestock products) are insignificant among the poor - less than US$3.50 equivalent per household per month (8-9Z of income); and, - off-farm enterprises (eg. small businesses, etc.) are also insignificant - contributing less than 5Z of household incomes amcng the rural poor ($1.50/household/month) - and even in urban areas agriculture remains three times more important than off-fanm enterprises as a source of income. 2.14 Localized surveys tell much the same story as the MPHBS, but provide some additional insight into sources of income (see Table 2.6). Table 2.6: Some Survey Estimates of Income Composition (Share of household income) Rural Rural 8 Villages - Acharya and Bennett Average Terai Hills Subsistence: Agriculture 48S 612 45Z Kitchen Gardening 2Z 42 22 Livestock 82 62 102 Hunting and Gathering 5? 2? 6? Food Processing 15? 172 12Z Household Manufacture 12 1? 22 Subtotal Subsistence 812 922 791 Non-Subsistence: Trading 72 12 3Z Wages and Salaries 12Z 72 182 Two Hill Districts - Rasuwa Nuwakot IRDP Low High Income Income Agriculture 692 892 Of Which: Livestock (42) (34S) Cottage Industry - 12 Trade 32 2Z Wages 15? 2? Portering 92 2S Remittances, Pensions, etc. 4? 4? Subtotal - Off-Farm Income 31? 112  - 12 - 2.15 In the sample surveys livestock contribute about 5-10 of income (except in the case of high income families in the hills sample - where they contributed a third). In no cases do wages and salaries account for more than about a quarter of household income. Cottage industries and trade are insignificant as sources of income, while portering was important only in one of the hill districts studied (where it accounted for about a third of off-farm income among the poor). 2.16 The relative importance of off-farm income among the poor (about 402), does not reflect their active participation in a robust informal sector. It seems to be generated mostly from a range of distress activities and very low paying employment. The fact that it constitutes a relatively large proportion of income is a function of the very low total incomes among the poor, and the very low levels of own-farm production, rather than an indication that the poor are doing well out of off-farm employment. 2.17 While the size of landholdings is a major determinant of rural incomes, the influence is not straightforward (see Table 2.7). Firstly, family size increases with the size of farm, so that while per cap ta income increases with farm size, it does so less than proportionately. Sec.ondly, returns to family labour are higher on large farms, due to diminish-.tg returns to labour, and a more favorable land/person ratio. Thirdly, -while small landholders are much more dependent on off-farm income, they earn substantially less per day worked off-farm - less than half as muc.h as those from larger farms (Table 2.8). Table 2.7: The Effect of Parm Size on Incomes (1984/85 Rs.) Non- Large Medium Small Marginal Cultivator Terai Farm Size (ha.) >5.4 2.7-5.4 1.0-2.7 <1.0 _ Percent of Households 3Z 9? 232 38? 27Z Average Household Size 12.2 9.0 7.0 5.5 4.7 Total Monthly Income 3380 1822 1210 787 633 Per Capita Monthly Income 277 202 174 143 136 Hills Farm Size (ha.) >1.05 0.5-1.05 0.2-0.5 <0.2 - Percent of Households 16Z 242 50? l1 - Average Household Size 7.9 6.1 5.0 4.3 3.8 Total Monthly Income 1284 841 635 462 559 Per Capita Monthly Income 162 139 128 108 150  - 13 - Table 2.8: Returns to Family Labor Per Day Worked By Farm Size (1984185 Rs./day) Non- LarRe Medium Small Marginal Cultivator Terai Agriculture on Farm 58 42 36 34 - Other 103 62 25 15 14 Hills Agriculture on Farm 40 34 35 31 - Other 56 31 26 20 20 Percent of Family Labor Time Spent in Off-Farm Work /a Terai 5? 82 22Z 65S 100? Hills 13? 242 39? 67Z 100o a/ Share of employed days 2.18 This decline in off-farm earning rates, combined with the increased dependence upon off-farm earnings on the part of small farmers, greatly reinforces the poverty-inducing impact of lower earnings from work on the farm itself. One likely explanation of the lower earning rates lies in the far lower educational attainments of family members on small farms. Breaking this relationship must be an important element of poverty alleviation efforts in Nepal. C. Profile of the Poor 2.19 The poor are concentrated overwhelmingly in rural areas (95?) and more in the hills than in the terai (because food costs are significantly higher in the hills; nominal incomes unadjusted for price differentials are in fact lower in the terai).8I The principal characteristics of the poor (and the non-poor) are shown in Table 2.9. SI Here, and in subsequent discussion, the poor are defined in accordance with the NPC poverty line as having per capita household incomes below Rs. 197/month in the terai and Rs. 210Imonth in the hills (1988189 prices). The profiles draw on special tabulations of the HPHBS data done for this study - many of vhich are presented in Annex III). While this is a very conservative poverty line which may in fact select only a sub-group of the poor, it does not materially change the profile that emerges.  - 14 _ Table 2.9: Nepal - Poverty Profile (Corresponding figures for the non-poor are presented in parentheses) RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills Estimated No. of Poor 2.9 m. 4.6 m. 0.3 m. 0.3 m. Economic Characteristics Average per capita Income Rs./day 3.4 (7.2) 3.9 (8.8) 3.4 (8.2) 4.6 (12.3) US$ p.a. $69 ($148) $81 ($181) $69 ($168) $95 ($253) Operated landholding (ha.) 1.1 (3.1) 0.3 (0.5) - - S of Family Labour Days - Worked Off-Farm 31? (142) 17? (10X) 362 (20Z) 39? (21?) Occupation /a Agriculture Self-Employed 54Z (67Z) 74? (74Z) 42? (31?) 46? (29?) Agricultural Labourer 28? (12?) 7? (2?) 4Z (3?) 32 (-) General Labourer 8S (3Z) llS (6?) 22? (9X) 30? (72) Other /b 10? (18?) 8? (18?) 32? (43Z) 21? (64?) Demographic Characteristics Household Size 7.1 (6.8) 6.1 (5.4) 6.9 (5.5) 6.3 (5.1) No. of Children 3.5 (2.7) 2.9 (2.1) 3.5 (2.2) 3.2 (2.0) Dependency Ratio /c 2.3 (2.1) 1.9 (1.7) 2.7 (2.7) 2.7 (2.7) 2 of Female-Headed Households 5? (4.5?) 12? (11.5?) 5? (6?) 15? (15?) Literacy Rate 22? (402) 39? (51?) 352 (59?) 49? (72?) a/ Proportion of economically active household members by main occupation. b/ Sales, Services, Production, Clerical and Professional. cl Number of household members per employed member. Source: MPHBS special tabulations, incomes in 1984/85 Rs. 2.20 Occupations. The poor, like everyone else in Nepal. are engaged mostly in agriculture on their own (or rented) land. The only notable exceptions are in the terai - where 28? of the economically active poor are employed by others as agricultural wage labourers, and in urban areas where about 25? are engaged as general labourers, and another 25? in services or other miscellaneous activities. Even in urban areas the largest proportion of the poor are engaged in agriculture as their principal pursuit (more so than among non-poor urban dwellers). Only about 5? of the active poor are  - 15 - employed in production or manufacturing jobs of any kind - including rural cottage industries. 2.21 Most people, and especially the poor, are employed only sporadically, even in their main occupation. Although the poor on average work twice as much time off-farm as the non-poor, they are still working only 212 of available labour days off-farm (higher in the terai, lower in the hills). There is correspondingly high underemployment - reportedly in the range of 35-45? of available labour days, with little distinction between the poor and non-poor (although this represents some under-counting of subsistence activities - see discussion of labour force issues in Chapter 3). 2.22 In rural areas, the main income-producing asset is of course land, and access to land largely determines income levels. The area of land operated by a poor household is about 60 percent less than that operated by a non-poor one in the terai, and about 40 percent less in the hills. In both regions, the proportion of the land which is irrigated is somewhat less for the poor than for the non-poor household, although the differential access to irrigated land reported in the MPHBS is rather less than is commonly supposed.91 2.23 Household Composition. The average size of poor hot '.holds is slightly greater than that of non-poor ones. Consistentl', 30, there are more children in poor households; and a tendency for more cl dren per household in the terai than in the hills. The relative concontration of children in poor families means that the children of Nepal are proportionately more exposed than are adults to all the disadvantages of poverty - including limited access to food, education, health services and sanitation. At the same time, they are more vulnerable than are adults to the consequences of poverty. 2.24 The proportion of female-headed households in the hills is quite high (about 11-15Z.) due to temporary and long-term out-migration. These proportions seem to differ little by income level or degree of urbanization. Reflecting the same basic phenomenon, the ratio of working-age adult males to females i-s m"ch lower for poor families than for non-poor ones, although it is not clear it this is cause or effect (see para. 3.90). 2.25 Finally, the aspendency ratio (or the ratio of total bou3chaid members to earning household members) shows some unusual patterns. Usually, a higher dependency ratio (many dependents per earner) characterizes porr families. This pattern shows up weaklv in Nepal, reflecting mainly the larger number of children in poor families d1scussed above. However. the least poor category (the urban areas) exhibits higher (ratbe. th~an the lower) average dependency ratios. The expla.ation lies i-. JE fact d.-n in the poorer areas, more of the family, and particularly chiidren and wt:..en, are pressed into becoming earners than is the case in less-poor areas. 9/ See Annex I.4; this may be due in par. to the reporting of all khet land as irrigated - where as in fact much is only monsoon-irrigated.  - 16 - D. Conditions of the Poor 2.26 This section attempts to convey a sense of the conditions under which the poor live in Nepal. It draws largely on the MPHBS and can therefore be regarded as providing profiles which are broadly representative of large groups in Nepal. They are in this regard to be distinguished from accounts of living conditions among particular communities in particular localities - accounts which are often strikingly vivid, but which by their nature may lack general applicability. Expenditure and Consumption 2.27 The poor are consuming a minimal bundle of goods, the cost of which is often not covered by their available incomes. Food (much of it self- produced or received as wages in kind) accounts for 70-80Z of the income of the poor, leaving very little for expenditure on other essentials, let alone to finance diversification. The following table summarizes key aspects of discretionary income. Conditions of the Poor - Summary Table - Expenditure (1984/85 Rs. per capita per month) RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills Monthly Consumption Expenditure 105 (185) 121 (234) 110 (202) 155 (323) US$ Equivalent $5.90($10.40) S6.80tS13.15) S6.20(S11.35) $8.70($18.15) Discretionary Income a/ 23 (110) 38 (128) 22 (147) 30 (200) ;Discretionary Cash Income bl 14 (77) 23 (98) 22 (123) 23 (189) US$ Equivalent $0.79($4.32) $1.29t$5.50) $1.24($6.90) $2.19($10.60) NOTES: Corresponding figures for non-poor shown in parentheses. Not adjusted for price differentials - hills purchasing power about 302 lower than terai. a/ After food; approximately 402 lower after fuel and water purchases. b/ After food purchases. 2.28 The table provides two somewhat different indications of upper bounds upon the capacity of the poor to pay for public services under the "user pays' principle. The first is the level of discretionary income, or the income remaining after the food (but not fuel or other) needs of the  - 17 - family are met. The per capita discretionary income among poor households ranges from about Rs. 20 to Rs. 40 monthly, or about $1.25 to $2.50 - and is about 40Z lower aiter fuel and water costs. The second is the availability of cash incomes. Cash incomes of poor families vary between about Rs. 30 and Rs. 45 per capita monthly in rural areas, and somewhat higher in urban areas. After essential food purchases (which even then do not bring them up to minimum caloric intake levels) available cash incomes range for Re. 14 to Re. 23 per capita monthly (equivalent to US$0.80 to US$1.30). The capacity of rural poor households to contribute significantly to the costs of public services is thus severely limited, as is their ability to purchase cash inputs, or finance investments which could ultimately raise their earning power (for instance in land improvement or education). As an example, the cost of applying fertilizer to only half of the land operated by a typical poor terai household would be in the neighborhood of Rs. 500 - compared with discretionary cash incomes of only about Rs. 100 per month, a large share of which is likely to be required for cooking fuel purchases. This absolute lack of discretionary income, coupled with the variability of incomes, means that the poor are extremely reluctant to take on new debt or diversify (eg. into cash crops or other enterprises) because of uncertainty regarding their capacity to repay loans or withstand losses. 2.29 Table 2.10 shows the typical consumption pattern of a poor family in the rural terai, although it does not differ markedly from that of other poor households (see Annex III). Table 2.10: Monthly Consumption Pattern of Poor Households (1984185 Rs. per month) Household Per Capita Per Capita la (Rs.) (Rs.) (USS Equivalent) Grains and Pulses 414 58 $3.25 Spices, Vegetables, etc. 87 12 $0.68 All Other Foodstuffs 53 7 $0.42 Fuel, Water, etc. 52 7 $0.41 Clothing 56 8 $0.44 Housing Costs lb 18 2.5 $0.14 Education and Health 24 3 $0.19 Transport 7 1 $0.06 All Other 1 38 5 $0.30 Total 749 105 $5.89 a/ Converted at Rs. 17.8/US$ (1984/85 exchange rate). b/ Excludes imputed rental value of own home.  - 18 - 2.30 Of note are the overwhelming preponderance of food and fuel (81S of the total) and the almost insignificant level of consumption of any other goods and services, except clothing. The lack of almost any discretionary purchasing power is reflected in the very low absolute levels of expenditure for other items - even those often considered essential (eg. US 19 cents per capita per month on health and education, six cents on transport). Food Security Conditions of the Poor - Summary Table - Food Security RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills Calorie Consumption la -52(+25X) -72(+15X) -52(+72) -112(+3Z) Share of Cereals in Total 86X(80Z) 812(76Z) 862(78Z) 82Z(69Z) Share of Calories Self- 622(812) 712(802) 432(46Z) 42Z(312) Produced NOTE: Corresponding figures for the non-poor shown in parentheses. al Relative to estimated minimum daily requirement (terai 2,140 kcal, hills 2,340 kcal). 2.31 The poor consume, on average, just below the estimated minimum caloric requirements, although there are quite wide variations among the poor and across geographic areas. More disaggregated data show the bottom third of the poor consuming substantially (20-25Z) below the required minimum. The problem is accentuated by poor retention of available calories due to parasites and other infections (see Chapter 5 for a fuller discussion of food security issues). 7able 2.11: Food Sources and Consumption Among the Poor-Rural Teral (Kcal/person/day) Market Home Purchased Pro'ced* Bartered Total Grains, Cereals and Pulses 200 1,233 406 1,838 (902) Oils and Fats 24 13 1 38 (22) Fruits and Vegetables 35 45 0.5 80 (4Z) Meat, Fish and Eggs 4 5 0.1 10 (0.52) All Other Foods 31 44 2 77 (42) Total 294 1,340 409 2,043 (14Z) (662) (202) * Includes food received free.  - 19 - 2.32 Table 2.11 shows the composition of food consumption for a poor terai family; again it is not significantly different from that of other poor households. Such a family subsists almost entirely on rice and dal - a lentil-based gruel - they consume less than 5? of their calories in the form of vegetables and fruit, and almost never eat meat, fish or eggs. During a substantial portion of the year (generally the three or four month pre- harvest period) they will eat only one meal a day. 2.33 Rural families are strongly motivated to grow their own subsistence needs, particularly basic cereal crops. It is likely that this tradition arose in the hills and mountains where isolation and poor access produce very high transport costs which, in turn, make self-sufficiency in food grains more economically rational than in an economic environment in which goods can move more freely. However, the poor rural families discussed in this section do not have sufficient land to produce their own calorie needs, although growing food crops has 'irst claim upon family land and labour resources. Note that such poor families sell only about Re. 60 (about US$3.50) worth of agricultural produce per month on average (Table 2.5). Non-poor families, with more land resources can not only produce a larger share of their own subsistence needs, but can also supply their needs by purchases, given that they generate cash income from agriculture of about Re. 300 - Rs. 400 per month. In summary, poor rural families are not only more dependent upon non-farm sources for food, but also for earning the income (in cash or kind) to supplement what they can grow themselves. Even so, their food intake tends to fall short of their food needs, not to speak of their appetites. Conditions of the Poor - Summary Table - Living Conditions RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills Access to Latrine 1? (7?) 6? (19X) 112 (38Z) 46? (742) Access to Water 91? (94Z) 60? (69?) 93Z (971) 772 (882) Persons per Sleeping Room 3.7 (3.0) 3.7 (2.8) 3.8 (2.8) 3.5 (2.4) Notet Corresponding figures for non-poor shown in parentheses. 2.34 Housing Conditions. The vast majority of households, including very poor ones, own a homestead plot together with a dwelling. There is thus no evidence of mass homelessness in Nepal, even among the poor - although housing conditions tend, of course, to be primitive throughout. For all poor families, the average number of persons per sleeping room is just below four.  - 20 - 2.35 The main problem associated with housing conditions is lack of access to water and sanitation. Lack of access to any nearby source of water is mainly a problem in the rural hills, with 40 percent of poor families having to fetch water from long distances. In addition, the drinking water source for all households is usually contaminated. There is almost complete lack of sanitation in rural areas, and very incomplete coverage in urban areas. In the rural terai and hills 99 and 942 respectively lack effective access to a latrine. Almost 902. of poor urban families in the terai have no access to a latrine (542 in the hills). The absence of toilet facilities for the poor produces generally low standards of sanitation, which affect non-poor and poor alike. 2.36 Time Use. The patterns of time use for men and women as well as boys and girls, are presented in Annex III and summarized below. The main impression which these data give is that the people of Nepal live lives of largely unrelenting work.l0/ They are not poor because they work an insufficient number of hours but, rather, because the rate of remuneration per unit of time worked is very low. Conditions of the Poor - Summary Table - Time Use (Average work burden - hours per day) RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills Adults - Male 8.0 (7.6) 7.9 (8.0) 8.4 (7.6) 7.6 (6.9) - Female 9.6 (9.1) 10.5(11.0) 9.5 (9.0) 9.9 (8.6) Children Age 10-14 - Male 3.8 (2.8) 3.8 (4.4) 2.7 (2.2) 2.21 (1.5) - Female 6.5 (5.4) 6.8 (7.7) 5.7 (4.4) 5.6 (3.6) Notes Corresponding figures for the non-poor shown in parentheses. 2.37 Overall, both poor men and poor women tend to work longer hiours than their non-poor counterparts. Their work burden is consistently higher, except in the rural hills, where the poor tend to work for slightly shorter hours than the non-poor. One reason for this might be the lack of work opportunities on the miniscule farms of the poor in the hills, combined with the lack of access to nearby off-farm employment opportunities. 10/ The apparent contradiction with high underemployment rates (para. 2.21) is explained by the fact that there is substantial seasonal variation in workloads, by the under-reporting of subsistence activities, and by the fact that in the absence of productive alternatives, many subsistence tasks are probably carried out in a time consuming fashion.  - 21 - 2.38 Women spend much more time than men on subsistence activities ant domestic work. As a consequence the work burden of adult women exceeds that of adult men by about 25 percent. This gender discrimination is common to all household categories, poor and non-poor alike, rural and urban, hills and terai. Gender patterns are established early in life. Girls 10 to 14 years old have a work burden about double that of boys in the same age group, and the pattern seems not to be dependent upon the poverty or otherwise of the family. Partly in consequence, girls spend less than 60 percent of the time that boys spend in education and reading. As noted below, however, girls participation in education is related to poverty, because in all four geographic categories, girls from poor families spend less time in education than girls from non-poor ones. 2.39 Education. The poor lag behind the non-poor in all categories of school enrollment, with the disparity increasing at each higher level. In the rural terai, for example, there is less than one chance in three that a poor child will be enrolled in primary school; for a child from a non-poor household the chance is rather better than one in two. At the same time, children of school age are disportionately concentrated in poor families. Systematically, educational chances are worse for girls than for boys, in the terai than in the hills, in rural areas than in urban and, from poor families than from non-poor ones. The relationships between education and poverty are explored more fully in Chapter 5. Conditions of the Poor - Summary Table - Access to Education RURAL NEPAL URBAN NEPAL Terai Hills Terai Hills School Enrollment Primary 30? (532) 53Z (67?) 38Z (63?) 642 (69?) Secondary 13? (30Z) 12? (30Z) 22? (42Z) 25Z (49Z) Ratio of Girls to Boys in Primary School 0.3 (0.6) 0.5 (0.7) 0.5 (0.7) 0.7 (0.9) Note: Corresponding figures for non-poor shown in parentheses. 2.40 Health. Among poor families there is continuing malnutrition and illness, and it is almost a certainty that at least one, and often two, children will die as a consequence of ill health aggravated by malnutrition. Health status is universally poor in Nepal - due to the poor coverage of health services, contaminated water supplies, overpopulation, and poor hygiene - it is not therefore primarily income-related. However, the poor exhibit a number of characteristics which make them, on average, less likely to be healthy - including poorer nutrition, larger family sizes, and less access to water supplies, sanitation, and health services. Issues of health care and the poor are discussed in Chapter 5.  - 22 - III. THE DE?ER)nI!WTS op POVERTY A. The Dynamics of Poverty - Output and Growth Issues 3.1 The combination of very low average income levels and a fairly uniform distribution of income suggests that the basic cause of poverty in Nepal is excessive population concentrated on an insufficient economic base, rather than the inegalitarian distribution of available wealth. The weakness of the base stems from natural disadvantages which hinder modern sector growth (eg., a difficult topography and location); the early stage of development, manifested ir. an unskilled workforce and weak managerial capacity; and a failure, thus far, to transform agricultural productivity in the way that the green revolution has done in most developing countries. All of these factors have been exacerbated by sustained rapid population growth. 3.2 Nepal is by nature a high-cost economy. It is land-locked, and suffers from difficult (and expensive) internal communications as a result of extremely mountainous terrain. Among the world's poorest countries are most of those which are land-locked. They suffer from higher border prices for imports, and are at a competitive disadvantage with respect to exports, due to both higher transport costs, and the higher costs of imported intermediate goods. There are of course counter-examples (eg., Switzerland, Austria), but these tend to be located at the center of efficient transport networks, surrounded by diversified high-income economies, and have also compensated by specializing in high value-added and skill-intensive goods and services. Nepal, at the moment, has none of these compensating advantages. 3.3 Approximately a third of the population live in areas in the hills and mountains which are inaccessible by road, where the costs of physical inputs are prohibitive, and marketed outputs are uncompetitive; and where it is infeasible to deliver developmental or social services at reasonable cost. The costs of road-building in these areas is many times that of constructing them in the neighboring plains of India. UTnlike other developing countries, Nepal does not have significant reserves of cultivable virgin land to be settled. Net cultivated land represents only 18Z of land area, with very limited scope for expansion. Finally, the long open border with India severely limits the scope for pursuing independent strategies to maximize growth (see para. 3.13). These constraints all limit the prospects for alleviating poverty through rapid growth. P02ulation and the Resource Base 3.4 The most fundamental factor contributing to poverty in Nepal has been the rapid increase in population - which approximately doubled in the last 20 years, and will double again by 2020. Population growth of 2.7? p.a. has eroded the limited gains which have been made in GDP and food production. If population growth had been contained to 1.5Z p.a. over th's period (instead of 2.7Z), then real per capita GDP would have risen 45S, rather than the 14? which was achieved.  - 23 - Figure: 2 Effed 1o Population Growth on Income and Food Availabrity IndtKles: 197WO-100 0 -1 / 120- -',_ 0oo so_o 40 -I 1*0 1S55 9So 1068 1070 197 1060 1S95 1900 MeWO_ GowmadIW4lmS The further doubling of population is an almost inevitable result of the momentum of past growth. Even if effective birth control were instituted today, it would take about 15 years until the population started to stabilize, since there are 50Z more girls below the age of 14 already born than there are total women currently of reproductive age. There is little effective family planning (only 172 of women practice contraception), and the total fertility rate, at about 6 births per woman, is the highest in Asia.l/ The two most critical consequences of this population explosion are unsustainable labour force growth, and saturation of the agricultural land base. 3.5 Population density is nov about 6.2 persons per hectare of cultivated land - ranging up to 10 persons per ha. in parts of the hills. Similar levels are found in the fertile Aseatic deltas - whereas in Nepal these densities obtain mostly on drylands of mediocre quality. The density per unit agricultural land has increased by nearly two-and-a-half times over the last two generations - resulting in rapid land subdivision and fragmentation. Availability of agricultural land has declined from 0.6 ha. per person in 1954 to 0.24 per person in 1990. The average farm size 11 Population programs, and issues related to the poor, are discussed in Chapter 5.  - 24 - in the hills in now below 1 ha. (and less than 0.5 ha. for 502 of holdings) which is insufficient to support the average farm household of 6 persons. 3.6 As a consequence of this pressure cultivation has been expanded up hill slopes and on to poorer land on plateaus and ridges. Marginal productivity is thus falling, and the land base itself is deteriorating - as soils are depleted, and erosion accelerates with encroachment onto more fragile slopelands. Furthermore, forest cover has been depleted - reducing the availability of fodder, natural fertilizer, and fuel - further decreasing the support capacity, and increasing erosion. A recent study estimates the probable costs of resource degradation at between $320-560 million per annum by 2010, in the absence of remedial measures.2/ 3.7 The response to saturation of rural economies elsewhere has been rapid urbanization. In Nepal the terai has provided a temporary safety valve. Following the eradication of malaria and subsequent land colonization, an estimated 1.2 million persons migrated from the hills to the terai in the 1960's and 70's.31 However the terai too is now approaching saturation, and further growth will need to be absorbed in urban areas and in non-agricultural employment (see para. 3.20). 3.8 It must be recognized that the terai is a strip of land only 30 miles wide at its widest. There are at most a further 400,000 ha.4/ of forest land which might potentially be converted for agriculture. Even if all of this were converted to farmland (a proposition of dubious ecological soundness), then at existing rates of natural increase and immigration, the rural terai would be saturated within 20 years (with densities of 8 persons per ha. - equivalent to China).5/ HMG's current policy is to restrict further settlement and protect remaining forests. This does not appear to be strictly necessary, although the implications of such a strategy are not well understood, and the Government needs to undertake an analysis of the economic and environmental costs and benefits of conversion of terai forests to agriculture (see para. 3.22). Economic Performance 3.9 Overall GDP growth has averaged about 3.42 p.a. over the last 20 years6/ - these gains have been largely eroded by population increases, 2/ Environmental Resources Ltd. (ERL); Natural Resource Management for Sustainable Development, 1989. Costs are estimated in deficits of food, fuel and fodder. 3/ 1961-1981 based on differentials between natural and observed increases. 4/ Equivalent to 202 of currently cultivated area. 5/ See ERL: even with full intensification of agricultural productivity, the carrying capacity would not expand beyond another 5-10 years' worth of population growth. 6/ National income data should be treated with caution; the data base is weak and estimates are not consistent across years.  - 25 - with real per capita GDP growing at only 0.752 p.a.7f The distinguishing characteristics have been sluggish growth in the agricultural sector, and limited but steady improvement in the formal sector (average real growth of 4? over the last 20 years), fueled by an acceleration in manufacturing output in the late 1980's. This high rate is, however, on a very low base, and manufacturing still accounts for only 62 of GDP and 2? of employment. 3.10 The changing structure of the economy reflects the early stages of transition out of subsistence agriculture, coupled with rapid growth of the Government sector (Table 3.1). The share of agriculture in GPP has declined from 652 to 52Z - but this change is accounted for mostly by the growing share of indirect taxation rather than a shift into services and non-manufacturing industries - which have maintained fairly constant sectoral shares (25? and 92 respectively). Table 3.1 GDP Composition and Growth Sectoral Shares Average Rate of Growth 1965 1975/76 1985186 1987/88 1965-1985 1980-87 /a Agriculture 65? jb 65Z 53? 52? 1.6? 3.9? Manufacturing 32 41 5? 6? 4.4? 5.7Z Other Industry 8Z 62 92 9? 3.0? 5.5? Services 232 23? 27? 25? 3.12 4.7? Indirect Taxes 1? 2? 7Z 7? 2.7Z 4.42 Total /C 13.5 17.3 24.6 27.7 2.71 4.4? US$ per capita - $150 $151 $172 Population 10.4 13.0 16.9 17.8 2.41 2.7? (millions) a/ Real growth rates outside of agricultural based on a composite GDP deflator, which may disguise inter-sectoral variations. Revised data for the last two years suggest a large increase in 1987/88 (about 82) followed by a decline in 1988/89 (as a result of the Trade and Transit dispute). 1988/89 per capita GDP is estimated at $164 equivalent. b/ Proportions based on current price data. cl Constant price data including indirect taxes, 1974/75 Rs. billions. 1965 data may not be strictly comparable with later years. 71 1970-1989.  - 26 - 3.11 Performance in the agricultural sector - which almost exclusively determines the incomes of the poor, has been even worse. Between 1975 and 1988 foodgrain production increased by about 20Z, while population increased by about 402. Nepal has gone from being a food exporter to a net importer, and cash crops remain of minor importance - accounting for only about 15Z of crop production. Yields of almost all key crops have been declining, with the limited increases in output accounted for mostly by ey;pansion of cropped area into more marginal lands. The reasons for this failure have been well documented elsewhere8l , and include: - inadequate availability of key inputs (fertilizer use, for example, is the lowest in Asia); - dependence on rainfed agriculture and the lack of effective irrigation; - the lack of developed marketing networks; - the lack of location-specific technologies - especially for the hills; and, - generally ineffective support and extension systems. The failure to expand agricultural output has a triple effect on the poor, since: - they are dependent for the majority of their income on their own production of foodgrains. - it reduces overall food availability, to which they are most vulnerable9/ ; and, - it reduces potential labour absorption on the farms of others. The relationships between agriculture and the poor are explored in Chapter 4. Macroeconomic Conditions 3.12 Poverty in Nepal is not, for the most part, a consequence of macroeconomic imbalances. The key features of the economy are illustrated in table 3.2. The balance of payments situation is structurally weak, with a heavy reliance on foreign assistance to finance essential imports. For a small economy (and thus one more likely to be trade dependent) Nepal has low trade/GDP ratios: exports are only 6? of GDP and imports 20S.10/ Despite 8/ See for example: Nepal Agricultural Sector Review; The World Bank, 1990. 9/ Because they are consuming only the minimum requirements; and because it drives up food prices. 101 1987/1988.  - 27 - some recent expansion of manufactured exports (mostly carpets and garments), the value of total merchandise exports amounts to only $10 per capita; the 1989 WDR shows only four countries with a lower level. Table 3.2: Nepal - Structural Economic Indicators 1982/83 1986/87 1987188* 1988/89* GDP Growth /a -3.0 2.7' 9.72 1.5% Exports/GDP 3.4Z 5.2? 6.12 5.62 Imports/GDP 18.8% 18.9% 20.52 21.8% Current Account (US$ m) -$217 m. -$202 m. -$267 m. -$296 m. Government Revenue/GDP 8.42 10.32 10.82 10.42 Government Expenditure/ 20.72 19.92 20.82 24.32 GDP Budget Deficit/GDP -12.32 -9.6% -10.02 -13.92 Debt Service Ratio -3.9? -6.9Z -8.8? nla Inflation 14.22 13.3? 11.0? 8.92 Exchange Rate /a -4.12 -7.92 -2.72 -13.62 Interest Rate /b 14X 172 17? 17? * Revised Preliminary Estimates. a/ Change from previous year. b/ Average of commercial lending rates. Sources: HMG; Economic Survey of Nepal; and World Bank Country Economic Memorandum, various tables. 3.13 The principal macroeconomic distortions arise from Nepal's situation vis-a-vis India. Being locked behind India's highly protectionist trade barriers, and integrated into a market dominated by administered prices, Nepal is dealing with second-best solu_ions in which distorted Indian prices are the effective border prices for most goods. The main effect of this is that the rates of return on deflecting third country goods to India are sufficiently high, and risk-free, that there is little incentive for entrepreneurs to invest in real-sector productive enterprises. Furthermore, Nepal's capacity to follow independent macroeconomic policies is extremely limited. Any attempt to liberalize without adopting the Indian structure of protection would result in large-scale smuggling from third countries. Even supposing such a scenario were politically feasible, the resulting appreciation against the Indian rupee would hurt the tradeable sectors much more than any gains from liberalization. The main effect of these distortions on the poor is the dampening impact on overall economic growth, and especially on employment-creating growth.  - 28 - 3.14 Within this framework of second-best solutions, the main economy- wide prices, the exchange rate and the interest rate, are not badly out of alignment. Nor does Uepal face an insupportable debt-service burden, having generally followed prudent borrowing practices, and enjoying sustained inflows of concessional resources. Structural Adjustment 3.15 Unlike many other countries which have initiated structural adjustment programs, Nepal at the time it did so was not greatly out of macro-economic adjustment, given the inflow of concessional resources it experienced and anticipated. The main outstanding distortions were removed under a stabilization program initiated in 1985 with support from an IMF stand-by arrangement. The structural adjustment program and two structural adjustment Credits in 1987 and 1989 were therefore concerned primarily with accelerating growth through better economy-wide and sector management. 3.16 As was analyzed in Chapter 2, the rural poor (who constitute about 952 of the poor in Nepal) participate only partially in the monetary economy. They are, therefore, largely insulated in the short-term from changes of the kind envisaged in the structural adjustment program. In the longer term, to the extent that the program is successful in accelerating growth and, in particular, in improving the delivery effectiveness of government programs, the poor will benefit from it. There are some components, however, which may have a direct impact upon the poor. These include the following: (a) Improvements in the procurement and distribution of fertilizer. These changes are unlikely to adversely affect small farmers; however, most of them are not in a position to effectively use fertilizer, and the reforms have not been sufficiently far-reaching for poor farmers, particularly in the hills, to gain significantly. (b) Measures to improve the financial viability and targetting of subsidies by the National Food Corporation (see Chapter 6), which, if implemented, could potentially help the poor. (c) The Program involves a more participatory approach to forestry and irrigation management and proposals for the privatization of smaller irrigation schemes. These are changes in directions likely to benefit small farmers. 3.17 The first two adjustment credits have, correctly, concentrated mainly upon improving general economic and development management. As such, they have had, or are likely to have, limited direct impact on the mass of the poor. Certainly adjustment in Nepal has not included the kind of cuts in public services or increases in the prices of staples which are often associated with adverse impacts on the poor under structural adjustment.  - 29 - The Trade and Transit Dispute 3.18 The short-term economic environment changed dramatically in March 1989 when the trade and transit (T&T) treaties with India expired. The most immediate effect was the lack of inputs and shortage of fuel which shut down many industries and disrupted transport. It seems likely that the effect on the poor of the T&T dispute has not been particularly pronounced, since they exist largely outside the cash economy and the urban areas where the shocks were felt most strongly. The longer term impact has been to slow the momentum of economic growth, driving up prices and costs, delaying construction of productive infrastructure, and further eroding confidence of potential investors in the formal sector. While these constitute a medium-term set-back, the impact on poverty in Nepal over a horizon of twenty years or more is unlikely to be significant. The T&T dispute is the most recent illustration of the inherent vulnerability of the Nepalese economy as a result of its position vis-a-vis India - from the perspective of poverty alleviation, it is this which is the significant factor, rather than the dispute itself. Emerging Trends 3.19 The fundamental problem is the pressure of continuing, rapid population growth illustrated in Figure 3. The population is expected to grow to about 24 million by 2000, and to 32 million by 2010. In the absence of a very aggressive population program it will expand to about 47 million by 2030. Figure: 3 Actual and Projected Popuaion Growth 1910 - 2030 60 65- Hlgh 50-i EXP/Cted 45- a40- LO S ASc ~~~~~~~~~~~~~~~~~~~~~~~~VerLO I202-6 1~~ t 10 5 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 EKlW4tl  - 30 - The demographic profile is particularly worrying - there are for example, 40Z more children in the 0-4 year-old cohort than there are in the 10-14 year-old cohort. This has disturbing implications for future incomes, employment trends, and saturation of the land base - especially in the terai. 3.20 Figure 4 shows the projected spatial distribution of population over the next 20 years. Figure: 4 The Changing Spatial Distribution of Population (Population in Millions) 1985 2010 j -4) 4-<-n >40) (14 Even with a conservative limit on population density, the rural population of the terai will double from 7 million to 14 million by 2010; and the urban population will increase more than five-fold - to almost 8 million people. This means, among other things, that cultivated land availability in the terai declines from an average farm size of 1.35 ha. per household to just under 0.75 of a hectare per household - implying the need for a doubling in agricultural productivity if current household incomes are to be maintained. 3.21 With respect to the massive increase in urbanization the concern is that with only limited industrialization, the agricultural base may not expand rapidly enough to create sufficient secondary off-farm employment opportunities for this new landless population. The likelihood is that there will emerge in Nepal a large underclass of urban poor who will not have the social fabric (however tenuous) which currently provides some support to the rural poor.  - 31 - 3.22 In the 20 years or so that it will take to effectively curb population growth, the task confronting hMG will largely be one of managing this transition - from the hills to the terai, and into urban areas. In the case of the terai, it will require a detailed analysis of the scope for conversion of remaining forest lands, along with a program to manage their conversion; and taking steps to expand agricultural productivity to increase the carrying capacity of the terai, including tenancy reform and irrigation improvements (see Chapter 4). It is recommended that HMG undertake analytical work on how best to handle the movement of incremental population growth to the terai and to urban areas - including the investment and institutional requirements, and the costs and benefits of opening up further terai lands. 3.23 The labor force will double over the next twenty years (from 6.7 million in 1985 to 13.6 million by 2010). 11/ In the past the economy has absorbed marginal population growth though a combination of smaller farms, the opening of the terai, and labor market expansion. With the saturation of arable land almost all of these new entrants will have to be absorbed in off-farm employment activities - a prospect of staggering proportions. By the year 2005 they will be joining the work force at the rate of 8.000 per week, posing a tremendous challenge to the Government and the economy. The scope for expanding sectoral labor absorption and the measures needed to do so are explored in subsequent chapters. 3.24 Even under the best expected conditions (i.e. sustained high GDP growth and low population growth) per capita income could only rise to about $270 p.a. by 2010 12/ This sort of growth will not be sufficient to alleviate poverty on a significant scale unless such growth is tightly focussed in areas which will benefit the poor. (a strategy for such *balanced' growth is explored in Chapter 7). Furthermore, the current flat distribution of income may not obtain in the longer run, due to increasing urbanization and monetization, both of which are usually accompanied by a deterioration in the distribution of income, and because with rising labour/land ratios, an increasing share of agricultural income can be expected to go to landowners. 3.25 Projections show that under expected conditions of growth the poorest 402 would remain below the poverty line by 2010 - even if the current distribution of income can be preserved. If the distribution of income were to deteriorate to a level more typical of developing countries, then even with moderately strong (3.3Z) GDP growth the lowest 40S of the population would drop well below levels necessary for survival (e.g. US$40 per capita annually for the lowest quintile, US$65 for the second lowest).13/ Even under a scenario of rapidly rising per capita GDP, and no erosion in the distribution of income, the poorest 202 would remain in absolute poverty for at least the next 20 years, leaving a minimum of 6 million absolute poor by 2010. 11/ Labor force growth over the next 20 years is largely deterministic, since new entrants have mostly already been born. See Social Sector Strategy Review (SSSR) for underlying assumptions. 12/ Constant 1987/88 Rs. equivalent - Source: SSSR. Assumes sustained real GDP growth of 4.31 p.a. 13/ See SSSR.  - 32 - 3.26 The policy conclusions that emerge are two-fold: (i) that under all conditions there will be a need to design programs to provide relief to the poor o0 a continuing basis for at least the next 20 or 30 years; and, (ii) for growth to have a substantial impact on reducing the number of poor, policies must be followed which guard against a deterioration in income distribution, by pursuing growth strategies which favor the poorer groups in society. B. The Mechanics of Poverty - Distributional and Equity Issues The Social Context - Inequality at the Village Level 3.27 While by international standards most Nepalis are uniformly poor, localized studies point to large perceived disparities between poor and wealthy at the village level. This is so because the distribution of assets is more obviously skewed than that of income, and because at the margins of survival small absolute differences in income (say $50 per capita p.a.)141 mean the difference between near-constant hunger, and a moderately comfortable surplus. 3.28 Distinct discontinuities exist between groups at the local level. In the hills the distinction is often between those producing surplus food, and those producing less than about six months' supply. (Indeed poverty in Nepal is popularly defined in terms of the number of months' food a family produces). In the terai it is between, at the top, landlords, landowners (who may or may not be poor), tenants, and at the bottom, the landless. With respect to location, those in the valleys are the well off, those on slopes and ridges less so, and those in more remote areas are generally the absolute poorest, with the exception of some mountain groups which have traditionally depended on trading. 3.29 Across all ethnic groups and regions there has been a tendency on the part of the poor to see the constellation of civil servants, local political leaders, and landlords as being a homogeneous group, who have at best little in common with the rural poor. These perceptions, the poor reach of public services to rural areas, and the difficulty in communicating between officials and illiterate peasants, have often resulted in a wide gulf between the rural poor, on the one hand, and the local political process and Government services on the other. 3.30 At the village level a range of microeconomic and social factors affect the incomes of the poor, including land tenure, labour market and debt relations, and the political environment. The following sections investigate the role of various constraints, the extent to which they prevent the poor from increasing their incomes, and measures which might be needed to lift these constraints. 14/ The difference between being 202 below the poverty line associated with minimum caloric requirements, and 202 above - and thus firmly in the 'non-poor' group.  - 33 - Land Tenure and Tenancy 3.31 Nepal does not exhibit the concentration of landholdings typical of many developing countries. This is largely because farm sizes have been reduced to a low average level by population growth and land fragmentation. However, data suggest that the top 52 of owners control about 402 of cultivated land, while the bottom 602 control about 202. This distribution appears skewed, particularly from the perspective of a marginal farmer, because access to a small plot (say less than 1 ha.) - especially of poor quality - is not generally enough to produce an above-poverty incomelS/ whereas a "large* landowner in Nepal (say 5 ha. - which would be small elsewhere in the world) produces a substantial surplus. The operational distribution of holdings is summarized in Table 3.3. Table 3.3: Size Distribution of Operational Holdings Size of Terai Hills Holding Average Percenrt of Average Percent of Hect2res Size Holdings Land Size Holdings Land <0.5 0.09 45.4Z 2.8Z 0.18 54.12 11.62 0.5-1.0 0.74 13.OZ 6.52 0.75 18.72 16.32 1.0-3.0 1.77 27.72 33.12 1.49 22.12 38.3Z >3.0 6.15 13.8Z 57.7? 5.67 5.12 33.8? Total 1.17 100.0? 100.02 0.86 100.0 2 100.0? Source: IDS, The Land Tenure Systems of Nepal, 1986. As with all such data, the proportion of land in the largest holdings is probably under-reported. 3.32 Tenancy and share-cropping arrangements are widespread, although it is particularly difficult to obtain accurate information concerning the extent of tenancies, partly because many households both rent-in and rent- out land, and partly because many tenancies are unreported because landlords have reacted defensively to the threats to tb, ir interests contained in the 1964 Land Reform Act (see discussion below). While officially reported tenancy dropped after the 1964 land reform, (to 82 of farm households in the 1981 agricultural census) informal tenancy went underground and flourished. 15/ This is because on unirrigated land cropping choices are limited and the poor cannot take advantage of their high person-land ratio to get high returns per unit of land. See M. Lipton, Land, Assets and Rural Poverty. World Bank, 1985.  - 34 - A sample survey found that somewhat over 21% of farm families in the hills, and 392 in the terai were tenants; for an average of 31X.16/ 3.33 Almost all tenancies are arranged on a share-cropping basis. Rent payable to the landlord is usually half of gross produce.171 In general there is no sharing of inputs, _< least on registered tenancies. Informal tenancy is apparently more participatory than certified tenancy, with landlords making some contributions to farm and entrepreneurial inputs. 3.34 Assessing the degree of landlessness is difficult because of inconsistent reporting and definitions. The census reports less than 1% of households being landless, but other sources report figures of between 10 and 20Z. It seems clear that the 45 percent of holdings in the smallest size class in the terai (Table 3.3) with an average size of 0.09 ha., include very many effectively landless households. These have a homestead block rather than a "farm'. As an approximation which is consistent with the data, it would be safe to assume that almost all rural hill families have some agricultural land (although by no means enough to produce their subsistence requirements), and about 20 percert of rural terai households are effectively landless. Land Reform Efforts 3.35 Under the 1964 reform ceilings were placed on landholdings (ranging from 3 ha. in the Kathmandu valley to 18.4 ha. in the terai), 'excess* land was to be purchased from landlords and redistributed - mostly to sitting tenants or neighboring farmers. In the end about only 1.5Z of cultivated land was distributed among 10,000 peasant families. 3.36 The reform was accompanied by a tenant registration drive (although only an estimated 10-20% of tenants were certified). An enormous problem was that the law provided for 'voluntary" surrenders by tenants; this provision was exploited to ease out tenants. In addition, land offices demanded payment of enormous extra-legal fees, and record-keeping did not allow easy identification of large landowners' holdings. Those tenants who did get registered, and who learned their rights and stuck to them (a very small minority) probably benefited from the reform. However, to the extent that the reform achieved the objective of greater tenancy security, the agricultural system was rendered less flexible, and the land market less viable. The Current Situation 3.37 Even though little land was in fact redistributed, there remains a strong sense of uncertainty. Under the current system those with "registered" tenancies can potentially claim title to the land they till. 16/ IDS, The Land Tenure System in Nepal. 17/ Ibid, p. 53.  - 35 - Because of the fear of possible transfers to tenants (either in completion of the 1964 reform, or under some new reform initiative) landlords, tenants, and would-be landlords are all using land inefficiently and inflexibly. 3.38 The current tenancy system is inequitable, and perhaps more significantly, it is inefficient. One commentator has remarked that it would be difficult to design a system less conducive to maximizing production. As a result of uncertainty and the perceived threat of transfer to tenants: (i) landlords are unwilling to make productive fixed investments (eg. in irrigation); (ii) tenants are rotated frequently (removing the incentive for them to make improvements); (iii) output is shared, but not inputs - so that tenants equate private returns with one-half of marginal output, and there are thus incentives to produce below the optimal level; (iv) small plots are farmed less intensively than is optimal using insufficient family labour, rather than risk using tenants; (v) landlords prefer to use Indian labourers who will not make claims on the land; and, (vi) tenants and smallholders lack land titles (due to the difficulties of registration) which they need for loan collateral or to participate in irrigation programs. Conclusions 3.39 In answer to the question: Is the current tenancy system inequitable? The answer is probably yes. Can much be done about it? Probably not. Because the problem is one of excess labour and a shortage of land the net terms of payment to landholders will, one way or another, be determined by the scarcity of land - whatever regulatory framework is put in place. Furthermore, the incentives to evade tenancy reforms are great, and where the peasantry is not organized, and the administrative capacity of the government is weak in its reach to the community level, the ability to abolish tenure types or regulate factor shares through policy is ineffectual at best. 3.40 There is some scope for physical redistribution in the terai,18/ and the distribution of operated land can potentially have more direct 18/ A redistribution of the 20Z of terai land in the largest holdings, if broken into small parcels (0.5 ha.) and effectively targetted, could possibly raise some 400,00 households out of poverty.  - 36 - equity effects than tenancy reform, but: (i) redistribution would provide only a temporary respite, and (ii) there has to date been little sign of political support for a redistributive reform. An arithmetic example illustrates the first point: if all land is excess of 5.4 ha. 191 were redistributed, this would bring 47Z of those in the lowest landholding-size class up to 1 ha., and would not benefit the landless at all. Furthermore, there are few reliable income-producing assets other than land. The absence of alternative investment instruments and social security measures act as a powerful disincentive both for landlords to sell land, or for middle-class support for a land reform. 3.41 With respect to efficiency issues, it is evident that there are many problems with the present agrarian structure. First, there is unfinished business left from the 1964 Land Reform Act, with registered tenants operating land at low efficiency levels. Second, the perceived threat of agrarian reform and the possibility of giving security to tenants has led to defensive action on the part of landowners. Third, there is already a large and growing class of landless, many of whom might become tenants if the laws were changed. Fourth, there are innovations, such as shallow tube-wells, which hold the potential for substantial production increases, but which are not reaching small farmers on account of tenure and collateral difficulties.20/ 3.42 For the reasons discussed above, a redistributive land reform does not seem to currently be a prospect in Nepal. However, some steps can be taken to regularize land tenure and tenancy arrangements, to reduce uncertainty and thus to remove some of the barriers to efficient use of land, especially in the terai. The key is de-linking tenancy reform from land redistribution. Elements of such a reform are outlined in Annex T, they might include: (i) bringing the 1964 reform to closure; (ii) giving legal status to alternative tenure arrangements which do not necessary involve full security of tenure; and (iii) streamlining procedures for land registration and transactions. 19/ The ceiling proposed in the 1986 IDS Land Tenure Report. 20/ There are also some special problems of land management in the vicinity of rapidly growing urban areas, such as in the Kathmandu valley. However, it is recommended that these land policy issues on urban fringes be treated as a separate problem from policy towards rural land proper. In this way, issues of agrarian reform can be considered in their own right and separate from issues of urbanization policy.  - 37 - 3.43 While it is clear that these type of changes are needed, substantial thought needs to go into designing a workable reform given the complexity of the issues, and the incentives for non-compliance. The tendency in Nepal is to frame reforms in terms of abolishing tenancy and transferring ownership to tenants. We do not know the extent of tenancies, the reasons for which they exist, (including the extent to which they may give the system economic flexibility it would not otherwise have), and the extent to which the country's rural social safety net is to some extent embedded within tenancy. We therefore do not know whether the existing tenancy system in fact serves some useful purpose for which other policies would have to be devised if tenancy were to disappear. It is recommended that analytical work be undertaken to (a) collect data on the extent and type of non-formal tenancies; (b) assess their effect on equity and efficiency; and (c) thereafter design a tenancy and agrarian reform package which at a minimum removes the barriers to efficient use of land contained in the current system. Labour Force Issues 3.44 The central features of the labour supply and demand landscape in Nepal are rapid labour force (LF) growth on one hand, and on the other the absence of the sort of agricultural transformation (eg. through irrigation) which would substantially affect labour absorption in the traditional sectors; coupled with relatively slow growth of employment in other sectors.21/ Between 1971 and 1981 the labour force grew by 41Z, whereas the total population grew by only 28Z. This is the result of growth being concentrated in the lower cohorts from which new LF entrants are drawn, and of a 5? p.a. growth in female participants - fueled by changes in the terai - where female LF participation rates doubled from 16t to 31? in ten years. 3.45 The labour market is overwhelmingly dominated by agriculture and by self-employed subsistence activities. According to the 1981 census, about 86? of the work force were self-employed, of whom 97Z were engaged in agricultural pursuits. To the extent that they are employed outside of subsistence agriculture, we have seen that the poor are engaged overwhelmingly as paid agricultural labourers (30Z of the economically active poor in the terai), or as casual labourers - mostly in construction (about 10? in the hills). 3.46 While most households participate in the labour market, the proportion of time hired out is low (19? for men, and only 6? for women, according to one survey). Estimates of open unemployment are low (about 6Z), however, time-use estimates of underemployment are generally in the range of 45Z-65Z. There is naturally substantial seasonal underemployment in rainfed agriculture (even adjusting for the high subsistence work burden), and except in some areas, at some times of the year, there is very 21/ The expected response would be to bid down wage rates, although it is difficult to see the potential for this in Nepal where wages appear to be at (or close to) a subsistence level.  - 38 - little in the way of off-farm work. MPHBS data estimates about 40? of available family labour days go unutilized, and one survey found, for example, that casual agricultural labourers are able to find employment only 682 of the time that they seek it. Employment Arrangements 3.47 In many parts of the hills, where the lack of a cash economy is combined with low mobility of labour (due to difficult access), one-to-one labour exchange between families functions in place of a wage labour market. In the terai, on the other hand, labourers are employed for payment in grain, usually under long-term contract, but also on a daily wage basis. There are two types of contracts: attached labourers work exclusively for one landlord, while semi-attached labourers must work for the contracting landlord when he needs them, but are free to work elsewhere at other times. 3.48 Both are usually contracted for a period of one year at an agreed price. Labourers are paid a subsistence wage or slightly below in grain and foodstuffs, but they also get access to credit, sometimes use of livestock, and often use of a very small piece of land on a sharecropping basis. These arrangements assure the farmer an adequate supply of labour during the peak season, and a fixed income (however meagre) to the employee. Such employees are often provided with a loan sufficient to carry them through the agricultural year. In the western terai there exists a form of bondage whereby such loans may bind the poor in a form of serfdom - in some cases indistinguishable from slavery - to a given landlord. 3.49 Non-agricultural employment of the poor is generally in construction or, to a much lesser extent, cottage industries. There is little documentation of the labour arrangements under which such labour is engaged. However anecdotal evidence suggests that there are a number of exploitative practices, especially in construction. 3.50 These reportedly include not paying wages, advancing pay in the form of meals at inflated prices, and charging excessive contractors margins (see Annex I-2). While these practices reflect in part the reality of low real wage rates; they are exacerbated by existing contracting arrangements and the lack of power of the poor to effectively demand the wages due to them. Wage Levels 3.51 Average wage levels are currently about Rs. 23 per day in the hills, and about Rs. 17 per day in the terai.22/ The following table provides a good indication of current wage rates: 22/ These are an average of daily wage rates for agricultural labour and unskilled construction work - the relevant rates for the poor.  - 39 - Table 3.4: Estimated Current Daily Wage Rates (1989) (Rs. per day) Men (Range) Women (Range) Agricultural Wage: Terai: Rs. 15 Rs. 13 Hills: Rs. 23 Rs. 20 District Wage: /8 Terai: Rs. 20 (18-22) Rs. 17 (15-20) Hills: Rs. 31 (22-42) Rs. 26 (16-42) a/ Supposedly paid on all public works Source: Mission Interviews (November, 1989) 3.52 Sufficiency. There is a remarkable uniformity across surveys of daily wage rates at around 4.25 kg. of paddy equivalent. Adjusting for days worked, and dependency ratios, this comes to about 1850 calories daily per capita - about 15Z below the estimated minimum requirement. 23/ In terms of cash income, these rates are equivalent to per capita incomesZ4/ of Rs. 2130 per annum in the hills ($76 p.a. in 1989 US dollars) and Rs. 1575 in the terai ($56 p.a.); well below the NPC defined poverty line. 3.53 This suggests two things: that wages have been bid down fairly uniformly to close to a subsistence minimum; and that at prevailing wage rates the poor are not going to rise out of poverty by getting access to employment alone. 3.54 Wage Trends and Variations. Several studies document declines in real wages of about 302 over the 1970's, however there are no consistent time series, so it is difficult to accurately assess movements over time. Compiling data from a variety of sources suggests that the average wage was equivalent to about Rs. 22/day in the early to mid-70's, and about Rs. 15/day in the mid-1980's (in constant 1987/88 prices) compared to about Rs. 18-20 today. The best one can conclude from the inconsistent data is that 23/ See Annex I-2 for the assumption regarding conversion to per capita incomes. 24/ Making the same adjustment for days worked and dependency ratios. (see Annex I-2).  - 40 - real wages have fallen during the last 15 years, recovered somewhat, and are currently either stagnant or declining. 3.55 There are marked regional and local variations in wage rates. Nominal wages in the hills seem to be consistently about a third higher than in the terai (consistent with a subsistence minimum, adjusted for grain price differentials). There are in addition very wide local variations in unskilled wage rates (e.g., ranging from 20 to 50 rupees per day in the hills) which result largely from remoteness (with higher cost structures and lower mobility of labour), and closeness to urban areas, where wages may be higher due to the level of agricultural development, proximity to markets, and the availability of alternative employment opportunities. The nominal daily rate for female workers is 15-20X below that for men, although some of this differential is accounted for by differences in hours and type of work. BOX: TYPICAL INCOME OF A WAGE EARNING FAMILY - Terai Agricultural Labourers - The family is contracted to a small landowner, the father works 7:00 a.m. until 6:00 p.m. about 15 days per month for which he is paid 3 kgs. of paddy per day, in addition to a breakfast and lunch equivalent to 1.5 kgs of rice. The mother might work about 10 days per month from 11:00 a.m. till 5 or 6:00 p.m. for 3 kgs. of rice. In addition, they have the use of a small garden plot (about 0.08 ha.) belonging to their employer, which may yield about 180 kgs. of rice annually in income. There is likely to be an adolescent child or older relative who works about the same amount as the wife. In addition, the family may get about 7 days a month of off-farm work at Rs. 15 per day. One child may tend livestock - either their own or the employer's. In the off-season, or on unemployed days, they may catch mice or crabs in the fields of landholders, or fish in local streams. Total Family Income: Rs. 683 per month Family Size: 7 Per Capita Income: Rs. 98 per month (US$3.45 equivalent) Food Consumption: 1700 kcal/person/day Rice and other cereals and pulses account for 90? Labour SupplyvDemand Issues 3.56 The high estimates of underemployment, the reported declines in wage rates, and village-level anthropological studies, all suggest that labour is available in surplus. In the hills out-migration has largely  - 41 - balanced labour force growth; but in the terai a class of landless agricultural labourers has emerged which exceeds the demand for labour on average - although not necessarily during periods of peak demand. 3.57 The labour force is growing at about 3.32 p.a. on average (5.5Z in the terai). Some of this labour is being absorbed in agriculture; however a number of contradictory forces influence the rate of absorption: Firstly, there is a farm-size effect. The reduction in farm-size due to subdivision increases labour intensity per unit of land, but the smaller farms tend to be worked exclusively by family labour - so the hiring in of landless (and marginal) labourers - who tend to make up the poor - is reduced. Secondly, there is a cropping intensity effect. The shift to irrigation and high yielding varieties, and the use of fertilizer, all result in more intensive production and use of labour - especially in the terai, and especially among large farmers. Thirdly, there is a land tenure effect. With the threat of transfer to tenants, landholders have reverted to self-cultivation, thus requiring more labour supply, while at the same time previous or potential tenants have become wage labourers. At the same time current tenancy arrangements discourage optimal production, and thus reduce the absolute quantum of labour used. 3.58 The increased demand for labour as a result of intensified cropping and the return to self-cultivation has not pressured wages upwards. Instead as a result of population pressure, all (or almost all) of the increased rents due to agricultural improvements have gone to landowners. Barring further major advances in technology (eg.the spread of effective irrigation), with LF growth of over 5? p.a. in the terai we should expect real wages to be pushed down further. 3.59 In the non-farm sector, the effect of urbanization is to open up additional employment opportunit,es, although the rate of growth of the urban labour force (about 202 p.a.) probably more than offsets the income effect for the poor. The monetization of the economy also has contradictory labour demand effects. It creates opportunities and increases the volume of off-farm activities (dlthough often not those which the poor engage in); and at the same time it has reduced demand for the skills of traditional occupational castes, whose products are substituted for by cheaper imported goods. It appears that the availability of off-farm employment opportunities has, so far, had little impact on labour force factors, (or on wage rates) - because in aggregate there exist relatively few such opportunities. This is however changing slowly in areas with greatly improved access, where studies suggest that the proximity to markets and alternative sources of income have had a positive impact on agricultural wage rates. 3.60 A number of things can be done to improve labour absorption. Perhaps the most important step would be to get irrigation working in the terai. This, with subsequent cropping intensification, would result in demand for about another 500,000 person-years of labour annually, about a  - 42 - 502 increase.25/ Opening up remaining terai forest land - if feasible - could contribute about another 500,000 full-time equivalent jobs,26/ and regularizing the land tenure situation would also have a significant impact. Improving access in the hills would enhance labour mobility, as well as lifting constraints on non-formal growth and cash-cropping which could contribute to substantial absorption. Employment and the Poor 3.61 Because they do not have sufficient land, the poor are disportionately dependent on labour market incomes. The level of such incomes depends on four factors, which may be subject to varying degrees of policy intervention: (i) family composition; (ii) participation rates; (iii) employment levels (the number of hours (or days) worked per participant); and, (iv) wage levels. The poor in Nepal are adversely affected by each of these factors, both singly and in combination with one another. 3.62 Analysis of family composition (Chapter 2) shows that the poor have larger average family sizes and higher dependency ratios than the non-poor, so that those who work in poor families have to support more non-working family members. 3.63 There is downward pressure on participation rates among the poor, not just because of family composition but because of a number of compounding factors. Firstly due to their lack of land, the costs of participation are higher (including in particular the costs of temporary migration). Secondly, illness and disability limit their capacity to participate actively. Thirdly, they have to spend more time in gathering fuel and fodder, and to a lesser extent water. Finally, females, of whom there are a greater proportion in poor families, exhibit lower participation rates due to cultural prohibitions, as well as frequent pregnancy and childbirth. 3.64 Un/Underemployment. Available data shows unemployment tends to be higher among the poor (by as much as 50X) in rural areas. The fact that the poor have less (or no) land to work with implies that in the absence of a robust demand for labour they are also more likely to be underemployed. Micro-level studies in the terai indicate that it is the landless labourers who work the least number of days per year. Finally, caste and ethnic group membership still have a major influence on access to employment - both of which militate against employment of the poor. 3.65 Earning rates for the poor are also lower than average, both for on-farm and off-farm activities (see Chapter 2). This is due to: (i) the lack of assets (eg. land) with which to work; (ii) lower levels of education and skills; (iii) the relative lack of political leverage to influence wage 25/ Assuming full double-cropping of rice and wheat, and a shift of about 352 of currently rainfed land to irrigation, see Appendix table 1.4-12. 26/ Assuming full irrigation and intensive cropping - see Appendix table I.4-12.  - 43 - rates; (iv) the fact that they tend to live in more remote areas where agricultural productivity is lower and wage rates are less influenced by the presence of off-farm opportunities; and (v) the higher proportion of children and female labour force participants, who are employed in lower wage rate occupations. 3.66 The fundamental facts of the labour market are excessive population on a limited economic base - which is keeping labour incomes at the margin close to survival levels. Within this framework the scope for increasing labour incomes is necessarily limited. However, there are a number of areas in which imprc-vements are possible. 3.67 Labour Absorption First, there is a need to expand employment creation in off-farm activities. An increased program of labour-intensive public works is one promising area (see Chapter 4); however given the scale of the problem the Government's contribution will be limited mainly to enabling mechanisms, by improving physical access and infrastructure, removing petty restrictions, providing credit and other incentives for employment-generating activities. Careful design is needed to ensure that these measures open up opportunities for the poor rather than the non-poor (eg. by focussing on the non-formal sector, construction, etc.). 3.68 Secondly, measures to improve agricultural productivity can potentially have a major impact on labour absorption, probably increasing it by about 502 in aggregate (para. 3.60). Equally importantly, agricultural employment growth of these dimensions would have a major impact on the demand for off-farm services and labour. 3.69 The scope for policy interventions to affect wage rates is limited - because they are bound to be driven by marginal agricultural productivity, and heavily influenced by the free flow of labour from India. However, the following would have some impact at the margin: - enforcement of District Wages for public works would serve not just to transfer more resources to the poor, but also - if the volume of works is large enough and their regularity assured - to bid up the reservation wage in selected areas; - improved physical access tends to increase the range of off- farm options, and thus bid up local wages - although this may affect the poor less than others; - programs involving group fotmation, credit to the poor, and other empowerment measures have been shown elsewhere to improve the bargaining power of the poor vis-a-vis employers, allowing some bidding up of wages. All of these measures will be likely have more effect in the hills than in the terai, where wages will tend more to be determined by the Indian level. 3.70 With respect to agricultural employment contracts, in answer to the question: Are they exploitative? the answer is probably yes. Can much be done about it? probably not, given the shortage of land and surplus of labour; although measures which empower the poor (eg. education, group  - 44 - formation, and decentralization, if effective) can serve to improve their influence on the terms of contracts. Construction contracts may be another matter. Monitoring and enforcement has shown to have some impact on the payment of wages and reduction of exploitative arrangements. Similarly the use of smaller local contractors and community-organized works has tended to lead to more wage income for the local poor. Both of these areas deserve further investigation to determine what does and does not work (see Annex I- 5). Migration 3.71 There has always been temporary migration in Nepal, however it has taken on a new importance and complexity in the last 25 years, with the population explosion, and opening up of the terai. Migration is basically of three types: (1) Seasonal (3-6 months), during the agricultural slack season in the hills; (2) Medium-term (up to 3 years), mostly to India, to supplement incomes and savings; and, (3) Permanent - as the result of land loss, indebtedness, and lack of employment opportunities. 3.72 Seasonal migrants generally go to the terai, and medium-term migrants to India. Internal migration is almost exclusively from the rural hills to the terai and urban areas, while international migration is more complex, involving a network of flows both into and out of India. 3.73 Recently short-term migration has taken on much more of a distress nature - as remaining plots of land cannot provide adequate food to support families, and thus the males outmigrate in the slack season so that both they and the family members left behind can adequately feed themselves. Anecdotal evidence suggest that remittances from such migrants are very small or non-existent. 3.74 Long-term temporary migrants (over 6 months) go overwhelmingly to India, where they work in the cities as watchmen, or in Assam and the border areas as construction labourers, porters at coal depots, or in forestry. They tend to stay about two years, and do bring back some cash savings. A substantial proportion (perhaps 30-50X) of hill males undertake some form of temporary migration - although the relative importance of long vs. short- term migration is not known. 3.75 There exists no good aggregate data on the importance of remittances in incomes of the poor. In financial terms it seems that migration is more of a coping strategy than one for poverty-alleviation. The Rastra Bank estimated that only 8X of the cash income of rural hills households came from remittances; and one survey suggests that the relevance of remittances is lower for the poor than the non-poor. Temporary migration is none the less an important coping strategy - especially in some areas, and for some families. One sample of three Districts found 25-30S of households to be 'solely dependent on seasonal migration for supplementary  - 45 - income' (i.e. presumably meaning that they had no other source of off-farm income). 3.76 Permanent Migration. It is difficult to accurately quantify the scale of internal permanent migration, but estimates suggest it accounts for about half of the marginal population growth in the hills. It is not clear how the income effects of permanent migration are distributed. Studies suggest (not surprisingly) that it is the relatively poor who out-migrate, and that on arrival in the terai most end up as marginal landless labourers. Thus they go from a completely untenable situation (in the hills), to a not much better one in the terai - an improvement, but certainly not a strategy which will raise them out of poverty.27/ 3.77 Data on the scale of international migration is inconclusive, although there is no doubt that the labour market in the terai (and elsewhere) is affected by substantial inflows of Indian labour. The only quantitative analysis (based on census data) suggests that net international migration is approximately zero, but there is probably very substantial under-reporting of both in- and out-migrants. 3.78 The pattern of international migration is complex. The 1964 land reform inadvertently encourages landholders to use Indian labourers and tenants rather than hill people. At the same time wages in some Indian border areas may be lower than in the terai, at least seasonally. This has resulted in a paradox - that hill people go to India, while Indians migrate into the terai. 3.79 Nepal has attempted to reduce the inflow, but to little effect - both because restrictions are physically difficult to enforce, and because there exist strong economic incentives for non-compliance. It is difficult to see what Nepal can do in this regard. The use of obvious restrictive measures which are successful in other countries would invite counter- vailing measures from India. The regularization of tenancy arrangements may go some ways towards removing incentives, but would be unlikely to have a dramatic impact. Debt and Indebtedness 3.80 Ai.:ost everyone in Nepal is in debt. The average size of borrowing appears to be in the range of Rs. 1,150 p.a. 28/ representing about 62 of annual income on average - but the absolute level of borrowing is fairly uniform across income classes - so that the poor are borrowing up to 20 of their incomes annually. The volume of borrowing from institutional sources has been growing rapidly, but this growth has been concentrated almost exclusively among the non-poor; less than 102 of institutional credit is estimated to go to the poor. 27/ The income effects of migration are not well documented, it is recommended that further research by undertaken on them. 28/ 1984/85 from MPHBS expressed in 1988/89 prices.  - 46 - 3.81 Institutional credit still accounts for only about 202 of total borrowing (and almost none among the poor). Private moneylenders appear to account for another 20-25?, while the majority of borrowing is from other non-formal sources. These range from friends and neighbors to large informal credit societies, and cover a spectrum of arrangements from small interest-free loans, through debt-bondage to landlords or employers, to large-scale commercial financing (up to Rs. 50,000) by highly structured local credit groups, which function in lieu of formal financial institutions. (Debt relations are described in detail in Annex 1-3). 3.82 The majority of borrowing by the poor is not for investment purposes, but rather to sustain consumption. Surveys show that between 602 and 90? of borrowing by small farmers is for consumption (including marriages, catastrophic health care expenses, etc.), and that less than a third is for capital investments. It is worth noting that small farmers are not using credit on a seasonal basis to finance agricultural operations in the way that would normally be expected (only 10-20? of borrowing among small farmers). This is probably because (a) they are not in a position to use cash inputs, and (b) they are averse to risking additional debt on the basis of outcomes which depend on (highly uncertain) rainfall. 3.83 Interest rates vary from 10 to 1502 per annum. On most informal borrowings they lie in the range of 30-40?. In addition there is often a transaction charge, which is reportedly higher for poor and marginal farmers (10-25t of the principal) than for large farmers (5-10Z). Institutional interest rates are in the range of 15-202 (as low as 7.5? on subsidized agricultural credit for selected investments). However numerous sources report that the actual costs of using institutional credit may be substantially higher, due to high transaction costs in terms of time and extra payments (often 10? of the amount borrowed), which are usually higher for the poor than the non-poor. One study estimates that the effective rate is close to the 35? p.a. charged by non-formal sources. 3.84 Village studies all report that chronic indebtedness contributes significantly to impoverishment of the poor. However, because debt relations are often combined with employment and tenancy arrangements, and repayment is often in kind (in either labour or grain), it is difficult to quantify the extent to which debt service at high interest rates erodes the incomes of the poor. Calculations suggest, however, that debt service on typical amounts borrowed might amount to a minimum of Rs. 400 p.a. 29/ - or 15? of discretionary income among the poor. More importantly, the MPHBS data show that those in the lowest deciles are accumulating debt at a staggering rate, and repaying at only about a fifth the rate that they are borrowing. This is clearly an unsustainable situation. While it almost certainly reflects some under-reporting of repayments, it is reinforced by reported net decreases in assets among the lower deciles. Other surveys also confirm that the poor are consistently borrowing to cover their consumption shortfall. 291 In 1988/89 Rs. per household, based on MPHBS data.  - 47 - 3.85 The exploitiveness of debt relations varies greatly across communities - influenced as much by degree of social cohesion as by economic factors. They range from: (i) systems which involve temporary debtor/creditor relationships between social and economic equals -- a kind of 'revolving' indebtedness in a relatively closed and socially homogeneous community. (ii) systems which express and maintain economic (and sometimes social/ritual) inequality and may involve expropriation of labor, grain and sometimes the use of land -- as well as the expectation of political support from the debtor -- but which are essentially equilibrium systems. In other words, the dependence of the debtor upon the creditor is also a source of security for the debtor and the relationship does not involve permanent transfer of assets to the creditor. (iii) systems which result in a permanent net transfer of assets (e.g. land) sometimes leading to bonded labor or forced migration out of the community for the debtor. There is no conclusive evidence on the importance of the various types of indebtedness, but local studies suggest that there is both loss of land (and especially usufructuary rights) as a consequence of debt, as well as continuing instances of bonded labour, in which the debtors' body (or that of his wife or offspring) is effectively seized in payment of debt. 3.86 In answer to the question: Are current debt relations exploitative?, the answer is: It depends on the type of debt. Can anything be done about it? Among the poor, only partially. The main problem is insufficient incomes to meet consumption needs. In this context massive debt is a symptom - not a cause - of poverty, and the solution to the debt trap lies not in providing more or cheaper credit but in raising incomes. However, some welfare improvements are possible by increasing access to credit on better terms than moneylenders (provided t .? credit is not captured by the non-poor), and some programs have hao success in doing this (see Chapter 6). In addition providing alternative credit sources to the poor may in some instances break the cycle of dependence on landlords and employers - allowing them to negotiate better employment or tenancy conditions. Gender and Poverty 3.87 Women are at the center of poverty alleviation strategies in Nepal - because they undertake a very large share of household agricultural production; because they are disproportionately not participating in education, and their lack of participation in education is the main contributor to low overall educational status among the poor; and finally because they are central to changes in fertility and hygiene behavior. Furthermore, in Nepal and elsewhere experience has shown that some of the greatest family welfare impacts among the poor can be achieved by placing more income in the hands of women.  - 48 - 3.88 Women predominate slightly in poor households, although it is not clear if this is cause or effect (i.e., because a family is poor for some other reason (like lack of land) it has more females in it (due eg. to out- migration of the males), or the family is poor because it has a higher proportion of females in it, eg. due to lower participation rates). The proportion of female-headed households, on the other hand, is no higher among the poor than the non-poor (para. 2.24). 3.89 Women's earn_ngs are substantially lower than men's - mostly due to lower participation rates, because of the household division of labour, which assigns domestic work mostly to women. Cultural and social customs also limit female participation in work outside the home - especially in the terai and especially among certain ethnic groups/castes. Wage levels for the same work - adjusted for hours and type of work - appear be slightly lower for women - however, there are substantial pressures with respect to health, education, mobility,and job-entry which keep women in lower wage work. 3.90 Women as a group are generally poor because they are in poor households, rather than because of gender-specific discontinuities in the distribution of income. However within poor households, which are subsisting on average at just about survival levels of food consumption, they are consuming somewhat less than equal shares, and they are thus more vulnerable to declines in incomes. Although anthropological studies report that that boys receive preferential care and feeding when resources are scarce, a recent survey of intra-household food allocations showed a fairly even distribution between male and female children. In all families, however, daughters-in-law reportedly receive a lower food share, and cultural practices prescribe less food for women in pregnancy. 3.91 The security of women's income is less than that of men - both because they lack individual ownership of assets (especially land), and because their LF participation is more tenuous. Furthermore, there exist a number of ways in which the welfare of women is probably substantially below that of males in families at the same income level, viz.: - women work significantly more hours per day, especially fetching water, fodder, etc; - their caloric requirements may be higher due to the longer work day, pregnancy, and lactation; - they suffer more illness as a result of frequent childbirths; - they are less likely to receive medical care or education; and, - they are the ones who are left behind to survive off a non- producing farm in the off-season (although there is no data on the consumption effects of this).  - 49 - There is scope for substantial productivity increases by: (a) increasing the labour force participation of females - mostly through education and fewer childbirths; (b) freeing up time spent on domestic tasks - particularly fetching water, fuel and fodder, by providing water supplies, reforestation, and trail improvement; (c) ensuring that agricultural services address the needs of women farmers (eg. by training of extension workers, who otherwise seldom deal with female farmers; the Lumle research station has had some success in doing this); and, (d) by pursuing programs which support on-farm income-earning activities which put more income in the hands of women (for instance, the Production Credit for Rural Women program). Social and Political Constraints 3.92 Nepal is in a state of transition, mirroring the evolution of the economy from subsistence agriculture to a more monetized one. The traditional political system, which prevailed until the middle of this century, was highly paternalistic. It involved a complex web of feudal patron-client relations; the majority (and certainly the poor) were not expected to participate actively in it, nor p&rticularly to benefit from it. 3.93 Traditional patron-client relations are breaking down as a result of increased monetization, off-farm employment opportunities, and education. However, in many cases they are being replaced with revised versions, which substitute cash payments for kind; multiple ties with one patron (eg. a landlord) with single ties to several people (eg. civil servants, businessmen, moneylenders). The modern political system has generally mirrored the traditional power structure; representing largely the same interests, although there is a complex intermingling of traditional political forces with economic interests, and cultural and caste groupings. 3.94 There is a widespread perception that at least until recently the political system was designed to operate for the benefit of a very small minority. The system had employed a range of formal and non-formal pressures to discourage unmanaged change or participation by the rural peasantry. The moves toward decentralization and more participatory approaches over the last five years represent changes in the right direction. The more dramatic recent changes (4n early 1990) appear to reflect an acceleration of this trend, although it is too early to tell how deep rooted the changes are, and how quickly the system as a whole can evolve.  - 50 - 3.95 As in most countries the whole array of social and political pressures is such as to prevent the poor from effectively taking advantage of services or developments which could benefit them. These include an aversion among the poor to approaching official institutions, the tendency of officials to prefer to work with the better off, and an unwillingness or inability to deal with illiterate peasants. In addition, the private costs to the poor of using such services (in terms of time, and extra-legal payments) can be very high. 3.96 The process of decentralization thus far does not appear to have been seriously geared to addressing this problem. The poor can more effectively demand and utilize services when they form into groups, and some programs have had some success in doing this. However there has been a tendency to confuse such groups with emerging political forces, (which had been banned), or as a threat to the established distribution of power. In fact some NGO and donor groups report incidents of physical intimidation when such group formation activities are undertaken. The implication is that measures designed to empower the poor (eg. in project design) will only succeed to the extent that the larger body politic is prepared to accept such change. Indications are that while the environment may be more open than it has been, there may remain serious constraints which limit the likely effe(tiveness of such measures.  - 51 - IV. POVERTY AND THE PRODUCTIm SECTORS A. Agricultural Incomes and Poverty Impact of Agricultural Growth on the Poor to Date 4.1 The three major crops in Nepal are rice, wheat and maize. Rice is the preferred staple and grown throughout the terai and in some areas in the hills. Maize is the dominant crop in the hills. Wheat production, in the hills and terai, has increased rapidly both in terms of area and yield. Millet, barley and potatoes are grown on poorer soils and at higher elevations. Most farmers grow small amounts of pulses and oilseeds (mustard, linseed and rape) for their own consumption. Sugarcane, cotton and jute are grown as cash crops in the terai - although the poor grow almost no cash crops. 4.2 Production of staples has increased only 1.42 per year over the past 20 years despite the expansion of cropped area at the rate of 1.9? per year. Yields of most crops have been stagnant or have declined with expansion of cropping onto marginal land and decreasing fertility due to increased cropping intensity and low fertilizer use. This slow growth has occurred despite an increase since 1975 of 980% in area irrigated, 1402 in area covered by high yielding varieties, and 240% in annual sales of fertilizer (Annex I.4, Table I.4-1). 4.3 Local improvements in agricultural productivity have occurred for rice, wheat and maize, but only in areas with better soils and/or irrigation and access to chemical inputs. Case studies show that the poor tend to be located in places with less fertile soils,l/ this limits their capacity to increase output, or to adopt productivity improvements which may be feasible for other farmers. (In many cases, of course, their poverty is at least in part a consequence of their poor quality landholdings). In addition, the poor have less access to irrigation and less ability to pay for chemical inputs. While the poor do grow rice, wheat and maize, because of geographic location and soil quality they are more likely than wealthier farmers to grow millet, barley and potatoes, all crops that have not even experienced localized increases in productiviLy. Effectiveness of Existing Programs in Reaching the Poor 4.4 It must be stressed at the outset that the Government's general agricultural programs are intended more to promote overall growth than to benefit the poor specifically. The following sections, however, focus in particular on their impact on the poor, in keeping with the objectives of this report. An assessment of their effectiveness more generally has been provided in the World Bank's recent Agricultural Sector Review. Government agricultural programs have faced many difficulties and have had limited impact, but to the extent they have benefited any farmers, the poor have received fewer benefits. This is in part because many of the poor have so little land that they cannot be effectively reached by general programs. It 1/ See, for example, Pachico (1980), and McDougal (1968).  - 52 - also reflects some inherent bias in the structure of general programs, which have not adequately taken account of the constraints faced by poor farmers, as well as a natural tendency (largely correct) to focus efforts on larger farmers where the greatest productivity gains can be made. Moreover, agricultural programs have with some justification concentrated on the terai, where potential productivity is higher and constraints fewer. However, the poor are disproportionately located in the hills and mountains. The following sections briefly describe the impact of the general agriculture programs on the poor; programs with a more specifically poverty focus are discussed in Chapter 6. 4.5 Irrigation. Government irrigation programs have historically concentrated on building large-scale, gravity-feed systems in the terai. The poor have benefited when they have owned land in the command area and when water delivery has been efficient. There is little evidence that the poor are systematically excluded - except to the extent that they are less likely to have title to land. Large-scale irrigation projects are difficult to target at the poor except by giving priority to poor areas. This has not been done explicitly, although large irrigation systems have been built in poor sections of the western terai. Most importantly, however, the large systems have severe water delivery problems, limiting benefits to poor and non-poor alike. 4.6 The focus of the government irrigation program has recently shifted from large-scale schemes to smaller-scale farmer managed systems. The government is in the process of handing over control of small and medium- scale irrigation systems to farmers in order to improve performance. In addition, it is trying to improve the operation and maintenance of existing public systems and, once water delivery is improved, improve the collection of water user fees. Under the new irrigation sector program, groups of farmers can receive assistance in constructing irrigation systems, or rehabilitating and/or expanding existing systems, provided they meet a set of technical criteria. The current irrigation effort could be tilted towards the poor by giving priority to groups consisting of poor farmers or by concentrating on poorer areas of the country. There is evidence of very high economic returns to investments in shallow tubewells in the terai, but the poor do not have sufficient land, in most cases, to justify the investment based on their own production potential. Work is being done to organize groups of farmers for purchasing and operating small tubewells which would make irrigation more accessible to poor farmers. 4.7 Input Supply. The supply of fertilizers has been controlled by the Agricultural Inputs Corporation (AIC), a government corporation with a monopoly on importing and wholesaling fertilizer. AIC is also involved in production and wholesaling of seeds. The AIC has been criticized for inefficiency, and failure to ensure timely delivery of inputs. Actual demand for chemical fertilizer from poor farmers is probably quite low, given their production systems and lack of irrigation, but supply is inadequate nevertheless. Demand for improved seed is strong, but the AIC is only able to produce s fraction of the improved seed being used. Recent changes in legislation and the establishment of the National Seed Board provide the opportunity to establish a system of private seed production. A pilot project on farmer-operated seed production and distribution has been  - 53 - successful in the hills and could be expanded, allowing AIC to withdraw from seed production altogether. More timely supply and wider availability of inputs would probably be possible through a system of small private traders more responsive to the demands of local farmers. 4.8 Fertilizer use by poor farmers will remain low even if fertilizer is available. In the absence of irrigation fertilizer use is a high-risk strategy. For poor farmers, the risk is compounded by the large cash costs involved (perhaps half of their available cash incomes). Poor households can only finance these costs by taking on further debt which they are ill- equipped to do, especially on the basis of highly uncertain returns under rainfed conditions. Therefore, in the absence of irrigation, poor farmers will continue to use minimal amounts of purchased inputs. 4.9 Research and Extension. The poor, in general, have less access to irrigation and chemical inputs than the non-poor, and lower quality soils. They are also more likely than the non-poor to grow rough cereals such as millet and barley. Poor farmers want crops with stable yields over a variety of conditions, and crops that require few chemical inputs. Until recently, however the focus of the research system has been on varieties of rice, wheat and maize that require irrigation and high levels of purchased inputs. This means that for most of the poor the work done by the research system has been largely irrelevant, either because they do not have irrigation, cannot afford inputs, or grow different crops. There have been some exceptions: the Lumle and Pakribas research centers have undertaken farmer-centered research that seeks to minimize chemical input use. In addition, recent work sponsored by USAID has started to focus on green manuring and bio-fertility. These efforts should be expanded since they are directly relevant to the needs of poor farmers. 4.10 The extension service has been unable to reach poor farmers because its message is irrelevant for many of them, it is not structured to reach them directly, and the socio-economic gulf between poor farmer and extension agent often inhibits contact. The model farmer system which was followed until recently by the extension service used wealthier and more educated farmers to demonstrate new techniques. This system allowed a wider dissemination of knowledge using fewer agents, but it meant that the poor were missed. Evidence from local level studies suggests that the farming systems of wealthier farmers are significantly different from those of poorer farmers, that poorer farmers know this, and therefore give less weight to the demonstration pilots. Extension agents rarely visit poor farmers 2/ , and when there is direct contact the difference in status often prohibits real communication; this is especially true with women farmers. 4.11 The research and extension system has recently been restructured to use a farming systems approach, so that information moves in both directions. The extension system has adopted a 'contact group method to increase the number of farmers directly communicating with extension services. In addition, the research system has recently shifted to a outreach approach which stresses on-farm crop testing and increases the 2/ See, for example, APROSC-Baseline Study for Crop Intensification Program (1985).  - 54 - contact between farmers and researchers. This increased contact has shifted the focus of research towards low input systems and should eventually increase the amount of research done on livestock, which are an integral part of the farming system, and on horticulture, which has the potential to raise incomes substantially in accessible areas. These changes will help the poor provided they are included in, and participate in, the contact groups and provided the research outreach program uses poor farmers to test crops. 4.12 Credit. Increased credit is crucial for providing small farmers access to shallow tubewell irrigation and for providing production loans once irrigation is in place - although it is only the medium-poor who are likely to be in a position to utilize agricultural credit. General institutional credit programs by the commercial banks and ADB/N are not used by poor farmers because, among other reasons, collateral requirements are too high and application procedures too complex. Targeted credit programs for the poor such as the Small Farmers' Development Program (see Chapter 6) have had some success in lending to the medium poor, but transaction costs are high and even where they are in place many poor farmers still do not participate in them. Potential and Constraints in Hill Agriculture 4.13 What then is the best that can be expected from agriculture in raising in the incomes of the poor? For the following sections we have calculated the maximum incomes the agricultural poor could possibly generate - given their typical landholdings, household sizes, and likely cropping patterns. The analysis is based on the potential yields of each of the major crops under varying conditions (eg. whether irrigated or not, with or without modern inputs, on valley floors or slopes, etc.). It draws on farm budgets and agronomic data in the recent Irrigation Master Plan, and from the World Bank's Agricultural Sector Review (See Annex 1.4 for details). 4.14 Recognizing that there is enormous diversity among farm households in the hills, they may still be divided into four groups with different agricultural potential and constraints (Table 4.1). Large and medium farmers (27Z of hill farm households) derive most of their income from agriculture, but also rely on other sources in order to produce above the poverty-line incomes. Relatively small improvements in agricultural production would allow these farmers to produce above poverty-line incomes from agriculture alone. Almost three-quarters of hill farm households, comprising about 7 million people, have holdings of less than 1.0 ha. Of these households, about one-quarter are small farmers (about 1.7 million people) with holdings that average 0.75 ha. in size; the remaining three- quarters are marginal farmers (more than 5 million people) with holdings that average 0.18 ha. - approximately equal in extent to two or three house lots in many western countries. Even a doubling of agricultural productivity would not raise them above the poverty line.  - 55 - Table 4.1: Characteristics of Farm Household Groups in the Hills LarRe Medium Small Marginal (Poor)/a (Very Poor) Average family size 8 6 5 4 Holding size (ha.) >3.0 1.0-3.0 0.5-1.0 <0.5 Percent of households 5? 22? 1) 54? Ave. holding size (ha) 5.7 1.7 0.75 0.18 Share of Household income from agriculture 76Z 701 58? 47? Share of 1985 poverty line income derived from agriculture 76Z 61S 46Z 34? a/ Landholding and household sizes for the poor and very poor drawn from MPHBS data - Household landholding distribution from IDS includes some non-poor households in 'small' landholding class. Source: MPHBS (1989) and IDS (1986), The Land Tenure System in Nepal. 4.15 Two spatial factors are important in understanding agricultural potential and constraints: topographical location, (i.e., whether on valley floors, slopes or hilltops); and location with respect to access, i.e., to transport infrastructure and markets. Crop yields throughout Nepal are low, but yields from valley floors are significantly higher than for hill slopes under both rainfed and irrigated conditions. The importance of the relationship between topography and productivity for this study lies in the fact that the majority of the poor in the hills do not farm valley floor land, nor do they have frequent access to irrigation. To be poor in the hills means not only having very small holdings, but also having holdings of lower inherent productivity and more subject to stress or shocks, eg. droughts and landslides, than the non-poor. 4.16 This picture of a general relationship is complicated by the complexity of mountain agroecosystems. Elevation, slope, aspect, and aeason all combine to determine the thermal regime (length of the growing season), potential evapo-transpiration (water balance), and photosynthetic rate (dry weight increase, of crops. As a result, there is a lot of variation in the potential of small plots, since microclimates can be found in the hills which both pose constraints on production technology and also offer opportunities for specialization, particularly in horticulture.  - 56 - 4.17 Accessibility affacts cost and availability of chemical inputs, credit, and information. Reliable data on the proportion of the hill population that is accessible are not available, but it is estimated that up to two-thirds live beyond one day's walk of a motorable road. Access difficulties limit the agricultural productivity of farmers by raising input costs and restricting access to markets. Where markets are poor, farmers reduce risk by using farming systems that maximize stability of production and minimize reliance on external inputs. 4.18 With the introduction of irrigation, and using currently available technologies, farmers can earn more than three times what they now earn from rainfed land. However, even with these substantial increases, the small landholdings of households in the small and marginal categories limit their ability to become self-sufficient from agriculture. Marginal farmers have an average of only 0.18 hectares of land. Even with all land irrigated and access to chemical inputs, they could earn only half the poverty line income from field crops if located on valley floors, and only 34Z on slopes (Table 4.2). Small farmers with an average of 0.75 hectares of irrigated land can do significantly better, producing well above-poverty incomes on the basis of field crops alone. 4.19 Potential incomes from switching to tree crops are quite high, between Rs. 25,000 and Rs. 50,000 per hectare, if markets are accessible. With these levels of production even the small holdings of marginal farmers are capable of producing between 50Z and 95Z of a 1988/89 poverty line income. This represents a significant increase over the income possible from an equivalent area in foodgrains or oilseeds. However, these levels of income for tree fruits are built up only slowly, usually over a 12-year period, and require capital investments equivalent to Rs. 10,000 - 14,000 per holding. For these reasons it is unlikely that the poor in the hills, with their limited assets and high levels of debt, will be able to invest in these types of enterprises without a very high level of assistance. Nevertheless, a horticulture program which provided the necessary extension advice coupled with credit could help a proportion of poor households in accessible areas. Furthermore, many of them could benefit from a partial shift to tree crops and vegetables, which could be introduced gradually with only limited reduction in subsistence grain output (which they are unwilling to give up). 4.20 While potential improvements in productivity are large, most poor farmers will be unable to achieve them. Few poor households have access to irrigation, and even with projected expansion of irrigation in the hills, most poor households will still be dependent on rainfed crops. Improvements of up to 20%-30X in the productivity of rainfed crops may be possible through improved management practices, but, while these are significant increases, they are not enough to raise the small and marginal farmers above the poverty line. These low input farming systems will require careful management of soil fertility using natural fertilizers such as mulch and manure. Even if irrigation is available, increases in productivity depend on increases in chemical inputs which are unavailable in remote areas and only sporadically available in accessible areas. Clearly, agricultural production alone will never be able to raise most poor and very poor hill households above the poverty line, although it can make a significant contribution.  - 57 - Table 4.2: Potential Monthly Per Capita Income From Field Crops and Percentage of PovertZ Line Income For Poor Households, 1988/89 3/ Hills 4/ Terai Valley Hill Floor Slopes (1) Small (0.75 ha.) (0.75 ha.) (0.75 ha.) (1.1 ha.) Rainfed (Actual) Rs. 92 53 Rs. 76 (441) (25Z) (38Z) Current year-round irrigation Rs. 298 151 Rs. 228 (Potential) (1422) (722) (115?) Future year-round irrigation Rs. 341 239 Rs. 335 (Potential) (1622) (1142) (1702) (2) Marginal (0.18 ha.) (0.18 ha.) (0.18 ha.) (0.5 ha.) Rainfed (Actual) Rs. 27 16 Rs. 39 (13?) (8?) (202) Current year-round irrigation (Potential) Rs. 89 45 Rs. 117 (43?) (212) (59?) Future yeer-round irrigation (Potential) Rs. 102 72 Rs. 172 (492) (34X) (87?) Source: Based on Master Plan for Irrigation Development in Nepal 1989, See Annex I.4 for yield and cropping pattern assumptions. Family Size Assumptions from MPHBSs Small Marginal Hills 5.0 4.0 Terai 7.1 6.3 3/ The calculation of potential income from current year-round irrigation assumes that farmers have reliable access to water and chemical inputs all year. The calculation of potential income from future year-round irrigation assumes that irrigation water control improves and that farmers increase their input use accordingly. 4/ The estimates of income possible from rainfed agriculture differ from current incomes in Table 4.1 because: (i) Table 4.1 includes income from livestock and horticulture while Table 4.2 considers only income from field crops; (ii) estimates in Table 4.2 are based on average cropping intensity per hectare while small farmers often crop more intensively; (iii) some small' and umarginalg farms in Table 4.1 have irrigation and are therefore more productive than purely rainfed farms.  - 58 - Potential and Constraints in Terai Agriculture 4.21 The constraints to agricultural production in the terai are fewer than in the hills, but growth in agricultural output has been only marginally better, 1.5Z per annum versus 1.42 per annum. As in the hills, agriculture is characterized by many very small holdings; 582 of households in the terai operate 1 hectare or less. Unlike the hills, about 20t of households in the terai are completely landless and agricultural wage labor is an important component of income for poor households. 4.22 The typical poor household has about one hectare of land and produces 402 of the poverty line income from agriculture (Table 4.3). This household will plant all of its khet 51 land with rice during the monsoon season, using several different varieties, one of which may be a modern variety. It will use little or no inorganic fertilizer and no insecticide on the rice crop because it is too risky given the lack of water control. The household will depend primarily on family labour except during periods of peak demand such as transplanting and harvesting. During the monsoon season the household will plant non-khet land to pulses. During the dry season they will plant half the khet land with maize and leave the rest fallow. For maize they will use no purchased inputs and no hired labour because the maize yields are too low and too variable, depending entirely on residual soil moisture and very sporadic dry season rains. Animal manure is used for fertilizer, but is not sufficient for the amount of land, and in areas where forests are completely gone, animal manure is used for fuel instead of fertilizer. Table 4.3: Household Income and Landholdings in the Terai by Level of Poverty Very Poor Poor Non-poor Family Size 6.3 7.1 6.8 Land Operated (ha) 0.5 1.1 2.8 Household Income from Agriculture (2) 372 502 732 Share of 1988/89 Poverty Line Income derived from Agriculture (Z) 152 402 1272 Source: MPHBS (1988/89 rupees). 5f Khet: land suitable for growing rice.  - 59 - 4.23 The very poor farmer has an operational holding of half a hectare or less, and generates only 152 of the poverty line income from agricultural production. The farming system of the very poor farmer is similar to that of the poor farmer with several important exceptions. The very poor farmer will probably not own a team of plow animals because they are too expensive to buy and feed. This will raise the cash demands of production considerably and may reduce the amount of land the very poor farmer is willing to plant in the dry season when the risks of crop failure are higher. The very poor farmers will not use even the small amounts of chemical fertilizer used by poor farmers because they have very little cash income. 4.24 With the introduction of irrigation, earnings per hectare could reasonably be expected to triple through improved crop yields and more intensive cultivation (Table 4.4). Households with irrigated land using Table 4.4: Potential Income Per Hectare from Field Crops in the Terai and Area Needed for Poverty Line Income, 1988/89 Estimated Area Needed for Maximum Income Poverty Line Income Per Hectare (family of 7) Rainfed (Actual) 5,865 2.8 ha Current year-round irrigation 17,670 0.9 ha (Potential) Future year-round irrigation 25,945 0.6 ha (Potential) Source: Master Plan for Irrigation (1989) and MPHBS (1989). Poverty Line (1988189) for the Terai- Rs.197 per capita. See Table 4.2 for assumptions concerning current and future production with irrigation. currently available technelogy could earn a poverty line income from 0.9 hectares and using foreseeable future technology could earn a poverty line income from 0.6 hectares. Recall, however that 45? of rural terai households own less than 0.5 ha., and 202 are completely landless. Households with access to urban markets could earn a poverty line income from less land, if irrigated, by growing fruits and vegetables. Even so a large number of households would still be unable to survive on own production alone. 4.25 In addition to benefiting from increases in their own production, poor households may benefit from increases in the demand for agricultural wage labor. Presently, poor farmers earn only half their income from self- production, and very poor farmers earn only a third of their income from self-production. Both groups earn an additional 30-502 of their income from agricultural wage labor. Labor inputs increase substantially with  - 60 - irrigation in the absence of mechanization, and irrigated farms tend to use more hired labor. With the projected increases in irrigated area, demand for labor should also increase by as much as 501. Aggregate Potential for Agriculture to Contribute to Poverty Alleviation 4.26 About a third of the agricultural poor in the hills have enough land to potentially raise themselves above the poverty line through agricultural improvements - but only if their land is irrigable - otherwise they may be able to produce about half of the poverty-level income on their own land. For the balance - who make up about half the population of the rural hills - their holdings are too small to ever produce more than a third to a half of the poverty-level income from field crops, even under the best of circumstances (eg. with year-round irrigation and full use of modern technologies). While irrigation in the hills could possibly be expanded by as much as 300,000 ha. (60X), most of the poor are located on slopes and/or land that cannot be irrigated - for them agriculture can never generate more than a small share (20-302) of a minimum family income. The possible exceptions to this would be households in accessible areas that shift completely to horticulture. However, few very poor households will be able to make this shift. 4.27 In the terai, about 15? of the poor have large enough holdings (say about 0.6 ha.) to produce above-poverty household incomes, if fully irrigated. Another 252 have sufficient land to achieve perhaps half of poverty-level incomes from their own production of grains. As in the hills, some of these very smallholders could produce sufficient income by switching eventually to fruit and vegetable production. About half of the rural terai poor have insufficient land to raise themselves out of poverty through their own agricultural production.61 4.28 A shift from rainfed agriculture to irrigated agriculture would increase labor demand by about 180 days per hectare per year, more if households shift from field crops to horticulture. It is estimated that 500,000 hectares of land in the terai not currently irrigated are irrigable. If all irrigable area were irrigated, including area now forested, then yearly demand for agricultural labor could be as high as 554 million days in the terai and 315 million in the hills. Increases in agricultural production in the hills will not lead to significant increases in the demand for agricultural wage labor. Family labor and labor exchange arrangements should provide the increased labor required. 4.29 Recommendations. While technical recommendations in the field of agriculture are beyond the scope of this paper, it is clear that some general directions are worth pursuing. Firstly, the greatest impact on the poor will come from getting irrigation working in the terai, by 61 Possibly more - there is no accurate breakdown of those owing less than 0.5 ha. in the terai by income level; some, with irrigated land, will be non-poor, many however are effectively landless.  - 61 - rehabilitating and Improving operations of the large public schemes and handing over management of medium-sized schemes to farmers, but most importantly by increasing the spread of small scale irrigation, (especially shallow tubewells) to small farmers. This requires: (a) efforts to form them into groups to achieve economies of scale and overcome fragmentation problems; (b) regularization of land tenure and registration, combined with (c) a package of credit and technical support, such as is being provided under the recent irrigation sector program. In the hills slow continuing expansion of small irrigation works will be helpful, although the beneficiaries are often the non-poor.7/ One option may be to use the selection criteria under the irrigation sector program to give priority to poor farmers in receiving assistance, provided they meet the established technical requirements. Reforestation and community forestry projects also need to be pursued for their impact on the availability of fodder and hence of organic fertilizer. 4.30 Secondly, the current input distribution system is unlikely to reach small farmers effectively. Steps have been taken to remove the remaining Government monopolies and restrictions on inputs, but careful follow-through is required to ensure implementation at the local level. Consideration should be given to actively promoting a network of small private traders who would: (a) have financial incentives to seek out and travel to small farmers; (b) be willing to provide the very small quantities of inputs which they require, and (c) have an interest in providing them with technical advice. Experience elsewhere has shown that such traders provide both inputs and a network for marketing agricultural produce; although they will only serve the accessible areas. 4.31 Thirdly, the recent shift to a farming systems research and extension system should be strongly supported. To benefit the poor in particular the new research outreach program should: (a) place the focus of research on low input systems with stable yields; (b) promote more work on livestock, which are an integral part of poor farmers' farming systems; (c) encourage more research on horticulture as a part of the overall cropping system, particularly in accessible areas where production of fruits and vegetables could provide considerable income. The success of the recent changes in the research and extension system depends on the ability of the system to exchange information with the farmers. This means that in order to produce information useful to poor farmers, extension and research staff will have to be trained to communicate with poor farmers in general, and particularly with female and low caste farmers. 4.32 While these steps will be effective in reaching the moderately poor, it has to be recognized that many of the poor have too little land for it to be efficient for general agricultural services to focus on them - such general programs should instead concentrate mostly on the non-poor and upper ranges of the poor, where returns will be greater. 7/ See, for example, B. Martens; Economic Development That Lasts - Labour Intensive Irrigation in Nepal; ILO; 1989.  - 62 - Poverty and the Environment 4.33 Poverty is both a cause and effect of environmental degradation. Lacking assets and choices, open access to Nepal's natural resources has provided the poor with the only feasible short-term mechanism for their survival, albeit at declining rates in both the Hills and Terai. At the same time, at a micro-level, environmental degradation imposes significant direct costs on them. The most important of these are low and declining farm productivity, and rising production and household maintenance costs. 4.34 Nepal's agriculture suffers from a chronic nutrient deficit. In the Hills this is the direct result of the human and livestock populations overtaking the productivity of forests and pastures. The average hill family maintains four cattle and two buffalo, and requires about 3.5 ha. of uncultivated forest or pasture land to sustain each hectare of land under cultivation. Fodder shortages translate directly into a lack of manure, declining soil fertility and declining crop yields. Restoring fertility will have a major effect on crop productivity in the hills, although for the poorest families this will have only a limited effect on household incomes because their holdings are so small. 4.35 Increasing production costs stemming from environmental degradation take several forms. Most important for the poor are the increasing costs of fodder and fuelwood provision borne by family labor, women and children in particular. As forests and pasture become degraded, fodder and fuelwood must be fetched from increasing distances and from less productive sources, using time that could otherwise be spent in income-earning activity, or in the case of children, in education. As a further result of forest depletion, dung is increasingly burnt as fuel instead of being returned as fertilizer to the soil. It is estimated that about 8 million tons of dung are burnt each year, equivalent to one million tons of foregone grain production. Alternative fuels, such as kerosene, electricity and biogas will generally be beyond the financial reach of the poor. 4.36 The high Himalaya is an extremely young and unstable area. Massive erosion is a natural phenomenon under these conditions, and the macro effects of erosion and hydrology - both within the hills, in the Terai, and beyond - would occur regardless of the actions of man. In these circumstances it is important to distinguish carefully between that which is inevitable and that which is potentially controllable. Those effects directly attributable to human activity include deforestation; loss of soil productivity by over-use; and increased erosion due to extension of agriculture and roads onto fragile slopes. Many of these are a consequence of over-population, rather than environmental mis-management as such. The solutions include: (i) population control; (ii) greater user management of resources (especially forests); (iii) increasing the supply of inorganic fertilizer; and (iv) improving the administrative capacity to manage public resources. The steps required have been outlined elsewhere8/ and many are already being pursued by HMG (eg. the formation of forestry user groups). The poor will, in general, be least able to protect themselves from the 8/ See for example the ERL report cited earlier.  - 63 - consequences of environrental degradation, and lacking alternatives, will be forced to contribute to a worsening situation. It is also important to be aware of the costs to farmers of environmental measures (eg. erosion control, medium term abstinence from using forests while they regenerate, etc.). and to be prepared to compensate them if necessary. B. Off-Farm Employment and Incomes The Formal Sector Sectoral Employment and Issues 4.37 The formal se-tor (including tourism and the construction industry) contributes about 25? of GDP, and employs approximately 10? of the labour force.9/ It is still in the very early stages of development; is concentrated almost exclusively in urban areas, and exists in virtual isolation from the rural subsistence economy. For most Nepalese - and almost all the poor, their involvement with the formal sector extends only as far as infrequent bus trips or kerosene purchases, and perhaps construction employment or contact with Government services on an exceptional basis. Manufacturing 4.38 The manufacturing sector employs about 150,000 persons - mostly in small grain mills and brick works. Impressive rates of employment growth have been achieved over the last ten years (10 p.a.), but this has been on a small base, and in absolute terms is still insignificant. Manufacturing has been adding jobs at the rate of 9,000 per year - compared with labour force growth of about 250,000 p.a.. 4.39 It is unlikely that manufacturing jobs absorb the poor - because only about 202 of jobs are unskilled, and because plants are located in urban areas where the poor are not. (Although the moderately poor are probably absorbed in textiles and brickmaking.) The wages paid for unskilled work are slightly, but not much, higher than agricultural wage rates (in the neighborhood of Rs. 25-50 per day). Employment at these wage levels would not be sufficient to generate above-poverty household incomes, unless all adult family members were employed. 4.40 Industrial expansion is hindered by the factors described earlier (poor access, high costs, and the effects of India's trade regime). To the extent that rapid growth can be sustained, employment growth will probably be slower, because the labour-output ratio is falling. This is consistent with the experience in other developing coinmtries, where industrial growth has resulted in less employment generation than had been hoped for. 9/ Note that except for government and manufacturing activities data on the formal sector is very poor, and estimates throughout should be treated with caution.  - 64 - Construction 4.41 Construction is the single most important source of off-farm employment of the poor. We estimate it provides about 350,000 jobslO/ - perhaps two-thirds of them unskilled. These jobs almost certainly employ the poor - the wage levels are low enough to be self-targeting, and in many cases they are located in areas where the poor live. 4.42 Typically only 302 of construction expenditure translates into wage incomes - due to high materials transport costs, high contractors margins (often 30-402 of costs) and the use of urban-based contractors which results in as little as 202 of expenditure remaining in the project area. There is scope for increasing the employment-generation effect of expenditure on public works, by adjusting contractual arrangements (eg. using local petty contractors), by shifting in some instances to more labour-intensive methods, and when feasible, selecting smaller rural works. Various programs have had success in raising the proportion of labour to SOX without raising costs - thus creating two-thirds more employment at the same level of construction expenditure. One promising area is in increased maintenance of roads - existing roads are often under-maintained, and it is estimated that 80? of expenditure translates into local labour incomes. 4.43 Public works and construction are not, in the long run, going to be a major solution to poverty in Nepal or anywhere else. However: - intensified expenditure on rural works can be an important stop-gap measure to support the rural poor until population programs and agricultural improvements take effect; - some of the adjustments outlined above can increase the poverty alleviation impact per unit expenditure; - these expenditures also create productive infrastructure in rural areas which, if there is careful project selection, will also benefit the rural poor; and, - there is room for a substantial increase in expenditure on maintenance, which could have a particularly beneficial impact on the rural poor. 4.44 There are clearly trade-offs involved in maximizing construction empl8yment. It is recommended that HMG undertake analytical work to investigate the scope for increasing the labour-intensity of public works at acceptable cost, to frame a program for donor support. This should include (i) an analysis of existing labour-intensive initiatives in Nepal and elsewhere; (ii) an assessment of efficiency trade-offs; (iii) ar. examination of contractual arrangements and employment conditions; and (iv) criteria for maximizing employment plus project benefits. An outline of the proposed analytical work is included in Annex I.5. 10/ Full-time equivalent jobs. Data on the construction industry is sketchy, total employment may be lower than these estimates.  - 65 - Hydroelectricity 4.45 Hydroelectricity is one of the few obvious natural resources which Nepal can potentially exploit to generate growth. Hydro exports to India could generate in the order of US$300 million equivalent in annual revenue within the next 20 years.ll/ This represents a 1O increase in GDP, potentially an important source of financing for investments and services which reduce poverty. 4.46 Government should be all means proceed with the development of hydro power both for export and for the beneficial effects of rural electrification, however some cautions are in order. There is very little direct employment benefit from hydro. The benefits thus depend on (a) the distributional effect of added foreign exchange earnings, and (b) the propensity of Government to spend on services which benefit the poor. The impact of resource-based expansion in other countries (especially oil- revenues)12/ has been a currency appreciation which shifts the terms of trade away from the tradeables and in favor of non-tradeables. Thus agriculture, manufacturing, and tourism lose, and the non-traded sectors (eg. especially real estate) gain. It is thus possible that without strong and effective counter-measures hydro development may not benefit the poor - and could possibly make them worse off. Finally, care should be taken that during the development phase resources are not diverted from other priority sectors. Government, Trade, and Services 4.47 The public sector (including parastatals) employs about 130,000, 13/ By the nature of its activities the Government tends not to employ the poor or uneducated, although there are about 25,000 peon-level jobs which may employ the medium-poor. The capacity to expand public employment is limited, at any rate, by the rate of growth of government revenue. 4.48 The main effect of Government expenditure on the poor may be through increased economic activity in small towns and road-head bazaars, where the salaries of growing numbers of public servants have fueled phenomenal local growth. It is not clear, however, that the poor have benefited. It is claimed that the businesses, teashops, and services catering to the public servants are not operated by the poor, and purchase their inputs largely from formal sector sources and Kathmandu. When the mission met with poor farmers living within a couple of kilometers of such a road-head bazaar, it appeared that this transformation had had no effect on them whatsoever. 4.49 There is no reliable data on formal sector employment in trade and services (eg. large retail and wholesale firms, financial and consulting services, trucking and airline companies). We estimate it to be about 180,000 (with another 500,000 in self-employment and the informal sector - 111 See Annex I.5 for details. 12/ See A. Gelb; Oil Windfalls, Blessing or Curse?; 1989. 13/ Data is incomplete, and may understate total public sector employment.  - 66 - see following section). It is unlikely that these firms employ the poor - for the same reasons that the government and industrial plants do not. It needs to be recognized that any formal sector job - even at low wage levels - is keenly sought after for the security of cash income it provides. Access is heavily influenced by connections of caste and family, and the ability to pay bribes, making it unlikely that the poor will be hired. Tourism 4.50 Although tourism is often cited as a potentially important source of incomes - especially in the hills, total tourist receipts contribute only 2? of GDP, and the effect on personal incomes is limited. A large proportion of tourist expenditures go on imports and the services of urban- based firms. ERL estimates that only 6? was spent in rural areas in 1988. 75? of tourists never leave Kathmandu or Pokhara, and when they do they only spend an average of $3-4 equivalent per day in villages. Even this tends to be concentrated along two or three narrow trekking routes, and among very specific groups (eg. Sherpa porters). While the effect on individual villages is of course great, the impact on poverty alleviation in aggregate is very small. We estimate that tourism employs about 10,000 persons in the formal sector (eg. hotels, tour companies, etc.), and generates another 10,000 full-time equivalent jobs in rural areas. Potential Contribution 4.51 Table 4.5 illustrates the importance of formal sector employment, and the potential level of employment under moderately optimistic assumptions (3.5Z p.a. sustained GDP growth, and 6-72 p.a. growth in manufacturing and services). Table 4.5: Current and Projected Formal Sector Employment Employment Annual 1990 Growth Projected New Jobs Sector Employment Rate 2010 (in 2010) Manufacturing 150,000 6.5X 528,500 34,400 Construction 350,000 3.0 631,000 19,000 Tourism 10,000 6.52 35,000 2,300 Public Sector 130,000 3.0? 235,000 7,000 Trade 60,000 6.5? 211,000 13,700 Transport 18,000 3.5? 36,000 1,250 Services 100,000 7.0? 387,000 27,100 Total Formal Sector: with Construction 818,000 2,063,500 104,750 without Construction 468,000 1,432,000 85,750 4.52 The formal sector might be expected to provide sbout 2.1 million jobs by 2010 (1.4 million excluding construction) - employing 15? of the expected labour force. New employment growth, if sustained at the  _ 67 - relatively high proportional levels, could absorb perhaps a quarter of the marginal labour force growth of 380,000 p.a., a useful, but not dramatic contribution to absorption of the poor in off-farm activities. Even this would involve adding jobs at about two-and-a-half times the current rate. The implication is that while HMG should certainly continue to pursue measures which imiprove the environment for formal sector growth - the most important efforts to raise incomes will have to be focussed elsewhere. The Informal Sector 4.53 Informal sector activities are the principal occupation for very few households - although they contribute small amounts to the incomes of many. This is because most households still have some landholdings, and because opportunities have been limited by the virtual absence of a cash economy. On the basis of MPHBS data it is estimated that some 112 of the rural labour force are engaged in off-farm activities in the non-formal sector as their primary occupation, and 32? of the urban labour force. This represents some 1.2 million persons, broken down as shown in Table 4.6. Table 4.6: Estimated Informal Sector Employment - 1990 (By main occupation in thousands of persons) Proportion of Rural Urban Total Informal Sector Sales 231 78 308 267 Service 121 57 178 157 Production 254 88 343 29? Transport 12 18 29 2? General Labourers 285 51 336 28? Total 903 291 1,194 100X Proportion of Regional Labour Force 11? 32Z 13? Source: Annex I.6. Many rural households are engaged in multiple occupations on part-time basis - correcting for this implies that about 15? of the LF may be engaged in informal sector activities, broadly defined (1.4 million full-time equivalent jobs). 4.54 There is substantial household cottage industry production in rural areas, but most is not marketed. Surveys indicate that probably only 20? of output is sold, and that cottage industry and trade accounts for only 42 of incomes among sample households (and less among the poor).14/ With respect to more structured small industries, it is estimated that some 3,500 cottage and small industry (CSI) units have opened over the last five years, but 14/ See Chapter 2.  - 68 - many fail and there is no reliable estimate of how many are currently producing, or the numbers they employ. They tend to be owned and operated by the non-poor, who can take advantage of subsidy and incentive programs. Many utilize only family labour, although some larger ones (especially in urban areas) do generate employment for the poor. 4.55 Small businesses consist mostly of small shops, tea stalls, eating houses and lodges. They generally provide a supplementary source of income for agricultural families, and are seldom profitable enough to sustain a household on their own. There is no data on the scale of such businesses. They have grown very rapidly along roads, and anecdotal evidence suggests they are also failing very rapidly as a result of over-supply. In urban areas many of the poor are absorbed in hawking and street vending (mostly textiles and vegetables) and services (eg. shoe repair, hair-cutting, etc.). 4.56 Portering provides incremental cash income for many in the hills, although usually only for a few days a month. (The vast majority of the loads in the hills are self-carried). It pays better than agricultural work, but is strenuous and intermittent, and involves significant employment costs (in terms of waiting and meals). Expansion of the road network has, if anything, increased portering opportunities, as more goods are distributed from roadheads. 4.57 Occupational castes include blacksmiths, tailors, cobblers, metalworkers, etc. Census data and surveys show a dramatic decline in the numbers engaged in these activities. The traditional patron-client relations which supported them have broken down, and demand for their products has been eroded by manufactured goods. Some have successfully made the transition to urban craftsmen or vendors, but most have become landless labourers. They face particular difficulties because, as untouchables, many forms of employment are not in practice open to them. Informal Sector Incomes 4.58 Marketed non-formal activity contributes relatively little to incomes, especially among the poor, although there is substantial production for home consumption (eg. household food processing - which constitutes the largest share). 4.59 We estimate that roughly 182 of total household income may be generated by non-formal sector activities (up to 342 in urban areas) - although of course there is wide variation between villages and households. Average non-formal sector incomes are estimated (very roughly) at about Rs. 250 monthly per household in rural areas (US$1.62 equivalent per capita), and at Rs. 850 in urban areas (US$6.20 per capita).15/ There is of course a marked discontinuity in returns between better off families who run small businesses, and returns to the subsistence non-formal activities in which the poor usually engage. ie have seen (Chapter 2) that daily returns to off-farm labour are both lower than those to on-farm labour, as well as being much lower among the poor than the non-poor. 15/ 1984/85, expressed in 1988/89 prices, see Annex 1.6.  - 69 - 4.60 The incidence of poverty among informal sector workers follows much the same pattern as for the population as a whole, except in the urlban hills, where 732 are below the poverty line (as opposed to 152 on average) - highlighting the marginal existence of those in activities like street vending, carpet and handloom production, and laundry and other services for which particularly high levels of poverty are reported. Dynamics and Potential Contribution 4.61 The informal sector in Nepal is in transition - traditional off- farm activities are dying out, and are being replaced by more cash-economy- oriented opportunities in urban areas and along roadheads. Informal activities are still seen by the majority as incremental sources of income - the transition to conceiving of them as potential principal occupations is only taking place slowly. Given the underdeveloped state of the informal sector (compared for instance to other countries in Asia and Latin America), there is still substantial scope for expansion. In aggregate, of course, such expansion can at best follow demand led by growth in agriculture and the formal sector. However, there is a backlog of untapped potential due to the very low level of monetization and access in the past - there thus is scope for a catch-up effect which may allow informal sector growth to exceed growth of the economy as a whole for some time to come. In addition, the existence of isolated the micro-economies also creates opportunities, (eg. for food processing, for localized import-substitution, and for production by poor households to supplement their own consumption). 4.62 There are a number of constraints - none of them overwhelming. (i) The most confining is probably locational. Non-formal sector opportunities for the most part do not exist where the poor currently are (i.e. in areas with poor access and limited cash economies) - to take advantage of them they will have to migrate. Although at the expected high rates of urban LF growth migration in pursuit of non-formal opportunities represents an opportunity for some, not a solution for many. (ii) It is reported that lack of information about markets and techniques (more than lack of marketing networks) often prevents the poor from taking advantage of opportunities. (iii) Nepal does not have the sort of regulatory framework that constrains non-formal sector activity in other countries - although the power of local officials and police reportedly results in a lot of unofficial taxation of small enterprises. (iv) There are few explicit barriers to entry - rather the non- poor will be better positioned (because of their asset base, education, and connections) to take advantage of commercial opportunities - the poor will almor- inevitably be limited to the more marginal informal activities.  - 70 - 4.63 One can identify a number of potential areas of opportunity mostly in retailing, tourism, food processing, service and repair works. On the basis of fairly conservative assumptions, it looks like the informal sector could create about 3 million jobs by the year 2010, accounting for 172 of the labour force. (see Annex 1.6). 4.64 The policy measures available to facilitate non-formal sector growth are limited. Perhaps the most effective will be to open up access through road building. Basic education can also equip the poor to better take advantage of opportunities as they arise. The effectiveness of credit and entrepreneurship interventions is questionable - there is little evidence that these are binding constraints, and experience with the CSI project suggests that they: (i) will be to-en advantage of by the non-poor; and, (ii) will result in few long-term productive investments which would not otherwise take place. Programs with a very small-scale focus (eg. SFDP) have had some success in getting the poor involved in incremental non-formal activities (eg. purchasing a goat, poultry raising), mostly through group formation and intensive support; although their effectiveness in raising incomes has been mixed.  - 71 - V. TEE SOCIAL DI1ENSIONS OF POVERTY A. Nutrition and Access to Food 5.1 Thirty-six percent of the population of Nepal consumes less than the estimated minimum calorie requirements. Nationally, more than 50? of children under six are stunted,l/ with local studies finding rates as high as 802. These rates of malnutrition are as high as anywhere in the world and comparable to those found in the African Sahel during periods of drought. Food security is complicated by high levels of disease and parasitic infestations so that consumption of adequate calories is a necessary, but not sufficient, condition to prevent malnutrition. 5.2 Access to food is not, of course, necessarily linked to local food self-sufficiency; in Nepal however the two are closely tied. Firstly, because in many areas the cost of bringing in foodstuffs is sufficiently high that they will not be affordable at any expected income levels in the near future. Secondly, production of foodgrains for own consumption constitutes the largest share of income among the poor, so in the absence of substantial off-farm employment opportunities, raising incomes (the correct solution to food insecurity) will consist largely of raising foodgrain production. Nutrition 5.3 The last national nutrition survey in 1975 found 511 of children under six were stunted. Rates of stunting are somewhat higher in the hills than in the terai, but nowhere are they below 45?. More recent local surveys have found similar rates, with some areas having 8O0 stunting among children under five. 5.4 Inadequate access to food is clearly not the sole cause of malnutrition in Nepal. About 30X of households have inadequate access to food, while more than 502 of the children are severely malnourished. Moreover, perhaps only 10? of the households have access to so few calories that the children would be expected to be severely malnourished. Children in Nepal suffer from a combination of inadequate access to calories and very high rates of disease (the two week morbidity rate from diarrhea for children under 5 is 312). These form a cycle in which disease leads to malnutrition and malnourished children are more susceptible to disease. In addition, most children also suffer from gastro-intestinal infections and parasites which can cause the loss of 201 or more of the calories consumed and weaken children further. Local studies show prevalence rates of 86-911 for intestinal pathogens. 5.5 Case studies comparing nutritional and economic status in Nepal show that there is some relationship, i.e. wealthier households have somewhat fewer malnourished children. But the change in economic status necessary to make a significant difference in nutritional status is quite large: with 50-100? increases in income needed before measurable declines in stunting are observed (see Annex 1.7). It would be unrealistic to I/ Stunting: less than 902 of the National Academy of Science median height-for-age. Stunting at this level represents a serious health risk and is associated with increased rates of morbidity and mortality.  - 72 - expect that income generating efforts alone could produce a change in income large enough to affect nutritional status for many of the poor. 5.6 Reducing malnutrition requires insuring access to sufficient calories and reducing the prevalence of parasites and disease. Specific interventions are possible to improve hygiene, promote appropriate weaning practices, manage disease episodes more effectively, improve sanitation and water supply and promote breastfeeding. In addition, effective health programs such as immunization will help break the disease-malnutrition cycle (see subsequent discussion of health and the poor). Adequacy of Food Supplies 5.7 The percentage of households consuming less than recommended levels of food is highest in the rural hills, 47Z, followed by urban areas, about 40Z, and the mountains, 31Z. 2/ Even in the rural terai, which is a food surplus producing region, 23Z of households do not consume enough food to meet the requirements.3/ 5.8 The proportion of income spent on food is a good indicator of poverty and of food insecurity. Typically, at very low levels of income, expenditures on food remain constant as a proportion of income until caloric requirements are met, or are close to being met. Then as income increases, the proportion of income spent on food declines. This is true in Nepal, but significant declines in the proportion of expenditure on food occur only at higher income deciles. 5.9 In rural areas, food expenditures remain over 602 of total4/ expenditures through the 6th or 7th income decile, suggesting that until such levels of expenditure, households do not feel they are meeting their food consumption needs. It appears that a per capita expenditure level of roughly Rs. 200/month (1984/85 prices) is needed before households greatly decrease the proportion of expenditure on food. Only about 352 of the households in rural areas have this level of income. Expenditure patterns for most of the population (see Chapter 2) are similar to those of very poor people in other countries who spend high percentages of their income on food, and specifically on foods that maximize calories for each unit of money spent. 2/ The National Planning Commission calculates per capita consumption requirements of 2140 kcals/personlday in the terai and 2340 kcals/person/day in the hills and mountains, based on WHO guidelines adjusted for climate and household composition. 3/ Based on actual expenditures for food, including the value of own production. 4/ Total expenditure includes the imputed value of self-produced food, non-purchased fuel and the rental value of owned homes. The inclusion of the latter two makes it appear as if the poor in Nepal are spending relatively less on food than the poor in other countries where the lowest income groups often have 80Z of expenditures on food.  - 73 - Regional Food Balances 5.10 With few substantial sources of foreign exchange, Nepal is primarily dependent on its own agricultural production to meet consumption needs. A net exporter of food grains throughout the 1960s and 1970s, Nepal now produces less than its own consumption requirements; total current production of grains meets only 90Z of requirements and is declining. While foodstuffs can in theory be imported to make up the shortfall, in fact they are not, because of limited purchasing power and high transport costs. In fact, it appears that the net flow of foodgrains is still out of Nepal to India. 5.11 Some inaccessible hill and mountain districts produce less than half of the food they need. Caloric availability resulting from regional production of grains and potatoes is shown in Table 5.1. Due to low incomes and poor transportation, local food availability is synonymous with local production in most mountain and many hill districts. 5.12 All the terai regions produce a surplus except for the eastern terai. The surplus produced in the terai does not necessarily go to deficit areas in the hills. Anecdotal evidence suggests that much of it goes to India. In the hills and mountains only the western mountains produce a surplus. Table 5.1: Regional Per Capita Production of Grains and Potatoes (kilocalories /person/ day) Development Region Ecological Far West Mid-West West Central Eastern Nepal Zone Mountains 848 795 2656* 930 1413 1035 Hills 846 1201 1219 1407 1741 1330 Terai 2197* 2140* 2253* 2262* 1659 2046 Nepal 1339 1499 1562 1795 1664 1635 * Indicates regions producing adequate or surplus food. Estimated minimum requirement (assuming 80Z of calories provided by grains and potatoes). Hills and Mountains: 1872 Terai : 1712 Source: DFAMS and CBS.  - 74 - 5.13 Population growth has been outstripping food production in most areas of the country and most importantly in the terai, the major surplus area (Table 5.2). Given the dependence of incomes on agriculture, the decline in per capita production of calories suggests a similar decline in per capita income and a decrease in food security. While population growth may not continue at the same high levels, it is unlikely that food production can continue to increase in the same manner as it has done over the last two decades - through expansion onto marginal lands (see para. 5.17). Table 5.2: Population Growth Rates (1971-1981) and Growth Rates of Agricultural Production (1967/68-1987/88) Annual Growth Rate of: Mountains Hills Terai Nepal Population 1.4Z 1.7Z 4.1Z 2.6? Agriculture 0.22 1.4? 1.5Z 1.4Z Source: DFAMS and CBS. 5.14 In addition to regional food shortages, there are wide seasonal variations in food supply. In the absence of reliable water supplies, foodgrain production varies by as much 402 from year to year. Thus every four or five years there are major seasonal shortages (see Figure 5) - which can be catastrophic for the poor. Figure 5 Annualand crKJ tauiueon m rwd Prodtcticn _(c cmeucladn WOd Pok to 1.900 - 1.8M - - - - - - - lAW 1.300 - . . . . . . 19S8 1970 1975 1980 1985 1988 ttw.w&.taxr >b'- W  - 75 - Food Prices 5.15 The rural poor in Nepal produce only about 65Z of the food they consume, so food prices have a significant impact on their food security. In the inaccessible hills, local supply and demand largely determine prices because transport costs make imported foods prohibitively expensive. For example, transportation costs for grain from the terai to a hill town 2 days from the roadhead doubles the price. In the terai, prices are largely determined by the Indian market. This dominance of the terai markets by Indian markets and the inaccessibility of the hill markets means there is very little scope for government price policy to improve food security. 5.16 Inaccessibility and transport difficulties also exacerbate seasonal price fluctuations. In remote districts, fluctuations may be 40-60? compared to a national average of 20-25X. The poor are hurt twice by the wide price differences because, in order to pay debts, they are often forced to sell crops immediately after harvest when prices are low, and buy food during the dry season when prices are high. Food Balance Projections 5.17 Projections show that with current agricultural growth rates, only a drastic reduction in population growth will improve per capita food production (column 1 in Table 5.3) and this is unlikely to occur. Improved agricultural performance (column 3) and moderate decreases in population growth will improve per capita food production somewhat, but domestic food supplies will remain below levels necessary to meet minimum needs. All these projections are dependent on continuing increases in cropped area and cropping intensity which are not sustainable at current rates. If, as is likely, limits to the expansion of farming land are reached tefore 2010, then in the absence of major yield improvements sharp decreases in per capita production will occur (column 4). 5/ Table 5.3: Projected Daily Per Capita Availability of Calories from Grains and Potatoes (Kcals/person/day) Agricultural Growth and Population Projection Current Ag. Current Ag. Improved Restricted Growth and Growth and Agriculture & Land Expansion Low Pop. Medium Pop. Medium Pop. & Medium Pop. (1) (2) (3) (4) 1990 1,499 1,499 1,499 1,499 2010 1,621 1,485 1,730 1,211 Corresponding Grain Deficit in 2010: (000 mt.) -567 -1,040 -232 -1,950 Source: Annex 1.7. 5/ Long-term agricultural productivity is dependent on maintenance of some forest land for fodder and fuel resources, so reserving land for forest is economically rational although politically difficult.  - 76 - Improving Access to Food 5.18 For the foreseeable future it appears that only the agriculture sector has the potential for substantially improving incomes among the poor, both directly through increased production and indirectly through increased demand for labor and decreased food prices (see Chapter 4.A). Until the transportation system improves dramatically, producing food locally will be much more efficient than importing food, even though local production costs are often higher and yields lower. In addition, local production largely determines local demand for labor and therefore local wages. 5.19 After improvements in agricultural productivity, improvements in infrastructure may have the most significant long-term impact on food security through decreased prices for food and agricultural inputs. Comparison of prices in adjacent hill districts show that districts without motorable road linkages to the terai have food prices 50-100 higher than districts with such road linkages. 5.20 A third method of improving long-term food security involves improvements in the utilization of current household resources. Fifteen to twenty percent of the population has access to more than the minimum recommended calorie levels, but is malnourished nevertheless due to disease and parasites. Improved health services, focusing on prevention - better water supplies, hygiene, improved feeding practices, could all have a substantial impact on the number of people malnourished. 5.21 Improved food storage facilities combined with a credit program would also potentially improve utilization of current household resources by allowing poor families to hold crops until prices were higher rather than selling them immediately after the harvest. Poor households would benefit through lower post-harvest storage losses, higher prices for the crops they sell and/or less need to buy food when prices are high. 5.22 The programs most frequently associated specifically with food security in Nepal are food subsidy and distribution programs. These programs provide what armount to short-term increases in incomes through lower food costs or free food. These and other food security programs are discussed in more detail in Chapter 6. B. Education and the Poor 5.23 As was pointed out in Chapter 2, the poor participate much less in education than do the non-poor. Official enrollment rates have grown from about 22Z in 1971 to about 85? today.6/ But this growth has been fueled largely by increases among the non-poor, enrollment by the poor - and especially poor girls - is substantially lower. 6/ Gross primary school enrollments according to Ministry of Education statistics - actual rates are probably substantially lower.  - 77 - P1gw 6. ncome, Gender and Education Potcipoton 70 90~ U9 7. _ so - 4- i 20 N 0n i MPHBS data show enrollment rates among the poor of about 302 in the terai (compared to 50Z among the non-poor), and 552 in the hills (732 among the non-poor). At secondary levels and among girls the differentials are even more marked (see Annex III, Table 2.6). 5.24 Access by the poor is hindered because they are less able to pay the direct costs of education; because they are less able to bear the indirect costs of labour forgone; and because they tend' (at the margin) to live in areas which are less well served with schools. 5.25 Children's labour time is a significant source of family income - LP participation rates are higher among 10-14 year olds than for the labour forc as a whole. They also provide a large proportion of domestic labour - tending livestock, caring for siblings, cooking, and carrying water and fuel - freeing older family members to engage in essential off-farm employment. Time-use data shows that girls above the age of 10 are engaged almost full- time in such pursuits, and about half time (3-4 hours per day) below the age of 10. There are marked differences between poor and non-poor female children's work hours in the terai - less so in the hills (perhaps due to the universal work burden imposed by poor access). Children also play a key role in meeting peak agricultural demand during harvesting and transplanting - at which times the family cannot free them for school.  - 78 - 5.26 The direct costs of school attendance may be in the range of Rs. 10-20 per month (much higher in secondary school)7/ - relative to discretionary household cash incomes of Rs. 100-140 per m3nth among the poor. Such families, with between three and six children, would likely be able to afford to send only one to school. 5.27 There are other, less tangible, disincentives to educational participation which relate to caste and ethnic group membership, and the subtle perception that the school is not a place where lower social groups are welcome. Finally, there is a perception among the poor that education is unlikely to make any material difference to their economic standing. This is based in part on accurate perceptions of limited income-earning opportunities, and of the rigid caste structure, which limits mobility. 5.28 Furthermore, some characteristics inherent in the school system discourage participation - particularly by the poor. The primary curriculum is excessively complex. Children are expected to join with a basic knowledge of the alphabet - an infeasible proposition in the case of the poor who come mostly from illiterate households. Similarly, education is provided exclusively in Nepali, which is less likely to be the mother tongue of the poor. The Education System 5.29 The education system as a whole suffers from poor facilities, inadequate teacher training, a curriculum that is too academic and often irrelevant, and excessive reliance on rote learning and examinations. Coverage (although much improved) remains insufficient, and many children are still an hour or more's walk from school. The internal efficiency of the system is very low (taking an average of 13 person-years of instruction to produce an elementary graduate), due to very high drop-out and repetition rates.8/ There is thus substantial scope for improved effectiveness within the existing system. However the two overriding problems are growing financial constraints and non-attendance of children (particularly girls) who are within range of a school. 5.30 Expenditure on education is about 1.92 of GDP - currently equivalent to US$3.10 per capita. This puts Nepal among the lowest spenders on education in the world, both as a proportion of GDP, and in absolute terms. In aggregate it is clearly insufficient, since it only covers the costs of the 65? of (primary school-aged) children who are in in primary school, and the 20? of those (secondary-aged) who are in secondary school. The allocation of funds could be improved (for example the university has 2? of the students but 23Z of the budget) - but this would not obviate the need 7/ While primary schooling is in theory free, costs are incurred on supplies, examination fees, and often other charges levied to support school budgets. 8/ For example 62? of those who start the first grade do not pass to the second grade.  - 79 - for major additional financing if educational capacity (even at the primary level) is to catch up with potential demand. 5.31 Attempts at alternative financing arrangements have had undesirable equity effects. HMG has consciously under-funded secondary expansion, providing only about 501 of total costs, and passing the rest on to the community. As a consequence there has been rapid growth of private secondary schools - which tend to serve only urban areas and the relatively well off. The danger is entrenchment of a two-tier system, which provides highly unequal access to, and quality of, education. Nor are user fees a very useful solution in the Nepalese context. Given the high value of social benefits resulting from primary education, it is doubtful that one would want to use financial instruments which would further discourage participation. Furthermore, with the very low level of cash incomes among the poor, there is likely to be a serious ability-to-pay constraint. 5.32 Insufficient capacity may not, however, be the binding constraint. Many children who live within range of a school are not attending. The reasons for non-attendance are complex, and include: - other demands on the use of childrens' time; - the low private returns of providing education (especially to girls, who are lost to the family at marriage), - inability to pay the private costs of education, - cultural factors, and, - perceived irrelevance and poor quality. The determinants of school attendance are not well understood, and require more analytical work to determine what the most cost-effective interven- tions may be to get children into school. However the time-use data suggest that increasing enrollments (which amounts largely to increasing female enrollments) may depend on designing measures which reduce the work burden on girls. The Role of Education in Poverty Alleviation 5.33 What then is the relative importance of education as a poverty alleviation strategy in Nepal - both for the country as a whole (i.e. in contributing to economic growth in general), and for poor households specifically? The corollary to this question is: What is the optimal level (and type) of education to provide to have an impact on the incomes of the poor? 5.34 There has been substantial research on the productivity effects of education in developing countries. Most indicates that agricultural output increases (by about 1O0) as the result of primary education, that the main impact is on the use of modern inputs, that the impact is higher in lower  - 80 - technology/poorer env!ronment conditions, and that basic education is generally adequate for traditional farming systems. 5.35 Studies in Nepal estimate the increase in farm output as the result of primary education at between 102 and 202 (the latter in conjunction with complementary inputs),9/ with higher gains on modernizing farms than on traditional ones. Jamison and MoocklO/ found education effects significant only in the case of wheat (an introduced crop), confirming other studies which found education to be most important when adapting to changed circumstances. Two studies also found definite evidence of a threshold level of education of about six or seven years - below which the impact on incomes was small. (This may however reflect the inefficiency of the education provided - requiring seven years to produce a primary education - rather than implying a need for lower secondary education to achieve an income effect). 5.36 The effectiveness of basic education in raising personal incomes thus appears Lo depend on having a constellation of supporting factors - perhaps good land, an accessible location, access to modern inputs - which are unlikely to be available to the poor. 5.37 The poor may in fact be following a rational strategy in not sending their children to school - if there is an absence of off-farm employment opportunities, if the perceived relevance of both the curriculum and quality is low, or if the opportunity cost in labour forgone is higher than the returns to family income (or all of these). And especially if they are correct in their perception that a fairly high level of education (eg. secondary or above - which they have little hope of achieving) is needed to make a marked difference in expected incomes. 5.38 On the other hand, no country has made the transformation to the sort of growth path which is necessary to eliminate poverty in Nepal without widespread general education to at least the primary level. Analysis by USAID suggests that providing universal primary education in Nepal would raise aggregate agricultural output by about 62 - paying for itself in the first year. Furthermore, there are a number of indirect but very important effects, including: - the overwhelming impact of female education on fertility; - the effect of general (and especially female) education on hygiene behavior (and thus on nutritional levels); - the role female education plays in breaking down the traditional cultural seclusion of women, leading to the higher female LF participation rates needed by the poor (especially in the terai); and, 9/ See Annex IV for references on tne returns to education in Nepal. 10/ D. Jamison and P. Moock Farmer Education and Farm Efficiency in Nepal, The World Bank, 1984.  - 81 - - the fact that education of the poor builds their confidence, improving their capacity to deal with creditors, landlords, and official services; and to form themselves into self- reliant groups. 5.39 There are thus substantial externalities involved which result in the social returns to basic education exceeding the private returns by some substantial margin. HMG should be willing to consider measures to compensate individuals to overcome this differential - by, for example, investing in improved primary schooling capacity, curriculum reform, and even schol-rships in some instances (eg. to encourage female participation). 5.40 With respect to the optimal level and type of education, the evidence is unclear, but analysis in Nepal and elsewhere suggests that the rate of return to primary education is consistently higher than to higher levels (mostly as the result of lower unit costs). There is no doubt that basic literacy and numeracy alone can have a transformational effect on the poor. These car. be imparted through non-formal training or about three years of primary schooling. Beyond that level, we do not know what the aggregate impact of further schooling would be on the incidence of poverty; although some analysis suggests that a level of lower secondary is required to make a substantial difference to personal incomes and fertility behaviour. 5.41 Higher level (eg. secondary) education can have a substantial benefit for the poor individuals who receive it, but is unlikely to have much effect on an aggregate scale in the absence of a transformation of the economy that would create substantial formal sector employment andlor the conditions for modern agriculture. However in the long-term, given the limited resource base, any development strategy for Nepal must rely on developing a skilled workforce. 5.42 With respect to other training and skills acquisition programs designed specifically to raise the incomes of the poor, the experience has not been encouraging (see Chapter 6 on poverty programs). While developing a skilled workforce is an important element of overall development, our analysis of non-formal sector employment suggests that the binding constraints on the poor are often not skills (which at the low levels involved could probably be most cost-effectively acquired informally or on the job), but rather access to markets and inputs, and the absence of a cash economy. Strategies 5.43 HMG should continue to upgrade general education for the beneficial impact it can have on aggregate economic growth. However to the extent that the objective is poverty alleviation, it should concentrate its efforts on: - literacy and numeracy, - lower-to-middle level primary education, and, - education of girls - for the effect this can have on fertility, hygiene, and nutrition.  - 82 - A package of specific measures should include those outlined belowt 5x 4 Allocate Sufficient Resources to Primary Education - if Nepal were to achieve universal primary education by the end of the century, it would requi,e a sustained level of about US$50 million equivalent per year (compared to a current level of $24 million equivalent). There is room for cost-effectiveness improvements (by getting more students into existing schools, and improving internal efficiency - by for example reducing repetition rates) - but in aggregate these would not substantially reduce the magnitude of the requirements, which are driven mostly by the growing numbers of school-aged children. HMG and the aid community should be prepared to provide this level of funding. 5.45 A National Literacy Campaign. An existing program has developed excellent materials and approaches - but reaches only 70,000 per year out of an estimated 12 million illiterates. It could be expanded to a national campaign at modest incremental cost. 5.46 Measures to Improve Participation by the Poor. Design of these measures - which may rely on non-educational interventions - will have to await deeper analysis of the determinants of school participation. Some measures need to be undertaken anyway to improve the utility, and attractiveness, of primary education, including: - improving the relevance of the curriculum; - shifting away from rote learning, and reducing reliance on examinations, at least in the early years; - considering automatic promotion in the first three grades (because studies show that if students stay through the first year or two they are much more likely to complete primary education); and, - considering instruction in languages other than Nepali. Based on experience in other countries the following measures are probably also part of the solution: - siting schoc.ls closer to families (which allows families readier access to their children's labour); - expanding programs which allow flexible school hours; - recruiting teachers from the local area; and, - adult literacy, because literate parents are more likely to send their children to school. 5.47 Measures to Improve Female Participation - again, the most cost- effective solutions may involve non-educational interventions, related mostly to measures to reduce household demand for female child labour time. Two specific educational interventions would include:  - 83 - (i) increasing the number of female teachers (in this regard HMG needs to postpone the requirement that all teachers hold an SLC pass - which severely limits the pool of female candidates - especially in remote areas.); and, (ii) considering programs for female scholarships and the provision of school uniforms for girls. C. Poverty, Population and Health Population 5.48 The'e is no prospect of seriously increasing average incomes if the population continues to double every 25 years. Cur1-ing population growth is thus the central, single most important poverty alleviation strategy for Nepal. To achieve it will require a drastic, national campaign-style solution - such as has been implemented ia Indonesia or China. The current incremental approach cannot yield sufficient results to have any impact on the incidence of poverty. To achieve some stabilization of population growth (say 1.5? p.a. by 2010) will require total fertility falling to about 3 (children per woman) from its current level of 5.8 - this in turn requires a contraceptive prevalence rate of about 50Z.11/ Only 15Z of married women of reproductive age presently use any form of contraception - and most of these are accounted for by sterilizations performed after having had four to six children, and thus have little or no effect on populatlon growth. 5.49 The Government's population efforts expanded in the early 1980's through a major sterilization program. However Nepal has not been able to develop the sort of broad-based national population program which has been successful in other countries. Essential temporary methods are delivered through a moribund health service, which lacks outreach capacity, and is staffed largely by male workers, who cannot effectively deliver family planning services to female clients. Despite a number of encouraging policy statements, empirical observations suggest that population enjoys nowhere near the level of priority that is necessary. Nepal, for example, spends about $0.10 per capita on population activities, compared to about $0.50 per capita in India and Bangladesh; similarly, the recent 8th Plan Approach paper relegated population issues to seventh place out of seven development priorities. 5.50 Nepal does not face the same cultural and religious barriers to fertility control that hinder many of its neighbors, and if the Government could put into place an effective program to promote and support temporary methods - then there is the possibility that population growth could be effectively curbed. There is some evidence that desired family size is 11/ Source: Social Sector Strategy Review; Total Fertility Rate (TFR) - is the number of children ever born per woman of reproductive age; the Contraceptive Prevalence Rate (CPR) is the percentage of married women of reproductive age practicing some form of contraception.  - 84 - lower than actual family size (although perhaps not among the poor), suggesting that there is unmet demand for family planning services. Forces are thus in place for a decline in population growth rates - but only if a major national effort is made. 5.51 What is required is: (i) a major shift to temporary methods of contraception, and (ii) a sustained sterilization effort, but focussed on those in lower age groups who have reached their desired family size. The success of temporary methods depends on a program of sustained support at the individual level, such as can only be delivered by a cadre of well- trained (mostly female) outreach workers. Relying on incremental improvements to the current service delivery system will not yield results quickly enough to have any impact. Where such programs have worked in other countries, they have been central national priorities: well-funded, well-pub,licized programs that receive emphatic public support from the head of state and all levels of Government. Nepal needs to allocate major funding for such a program. Given the weakness of the service delivery system, this is admittedly a gamble, but it is one the country cannot afford not to take. 5.52 The elements of such a program might include, for example: (i) recruiting and training 25,000 highly paid female family planning workers, with adequate support and incentives to operate at the periphery; (ii) population-linked incentives in other sectors; (iii) possibly establishment of a separate delivery framework - outside of the health services; and, (iv) development of a cadre of population specialists and program managers. This would cost about Rs. 500 million per annum ($18 million - or $0.95 per capita, similar to the costs of a successful program in Indonesia). It is recommended that HMG develop such a program as a matter of the utmost urgency. 5.53 Provision of family planning services alone will not, of course, necessarily reduce fertility among the poor. Children have a high perceived economic value in the subsistence economy - both because they expand the family's production possibility frontier, and because they are seen as a source of security in old age. Surveys indicate that about two- thirds of families do not use contraception because they are unwilling to limit family size, largely for economic reasons.12/ This is reflected in the preference for sterilization after achieving a desired family size, and reluctance to use temporary methods. Similarly women are much more willing to consider contraception after giving birth to two or three sons - while the birth of daughters (whose economic value is lost to the family at marriage) has little impact on family planning decisions. 12/ UNICEF; Women and Children of Nepal, 1987.  - 85 - 5.54 The private economics of family-size decisions in Nepal are not well understood. The one available piece of analysis (of 300 terai families)13/ found only child labour and duration of marriage to be statistically significant in explaining fertility. More work needs to be done on this as part of designing major population program interventions. (It may be cost-effective, for example, to allocate resources to female employment; or irrelevant to provide family planning services if there are over-riding incentives againEt limiting family size.). An understanding of the household income effects of having fewer children among the poor is also important to ensure that population control initiatives do not have an adverse effect on household incomes - or if they do, to incorporate compensating measures. 5.55 Raising average incomes and increasing financial security are clearly conditions which facilitate rapid declines in fertility. However experience in other countries shows that significant fertility declines are possible in the absence of rapid economic growth. The principal determinants are average levels of education (and especially female education), the incidence of female labour force participation; and to a lesser extent levels of infant and child mortality (since total fertility may be explained in part by expectations regarding the probability of child survival.) Measures in all of these areas, in which Nepal ranks among the lowest in the world, will also pay dividends in curbing population growth. Health 5.56 Health status is uniformly poor in Nepal - with infant mortality rates (113 per thousand) among the highest in the world, and life expectancy (52 years) among the lowest. There are remarkably high incidences of parasite infestations, diarrhoeal diseases, goitre, acute respiratory infections, and morbidity associated with pregnancy and childbirth. Cretinism, blindness, and other permanent disabilities afflict a large proportion of the population. 5.57 There is no data on health status by income level - but it is almost certainly worse among the poor than on average - because their lower nutritional status leaves them more susceptible to disease, and because they are least likely to have access to either curative or preventative health services. The health of many is sufficiently bad to impair their ability to work - draining total labour productivity, as well as reducing personal incomes and increasing the dependency burden within poor families. 5.58 The health service is at a very rudimentary level - and reaches very few - let alone the poor. There were estimated to be only 863 physicians in 1987. Health infrastructure is severely underdeveloped - in most parts it is several days' or hours' walk to the nearest health post, which is often little more than a shack lacking staff or supplies. The system is plagued by a range of institutional and service delivery weaknesses. As in most countries the composition of expenditure within health is biased (both geographically and by type of care) in favor of the 13/ K. Rauniyar: Demand for Children in the Nepal Terai; Winrock, 1985.  - 86 - non-poor. Although there is some scope for re-allocation towards rural, primary health care, the institutional barriers to developing an effective health service in the near term are formidable (see para. 5.62). 5.59 There are, however, a number of interventions which could have a dramatic impact on the health status of the poor, at relatively little cost. These include improved nutrient retention via improved hygiene (since many illnesses are the result of poor sanitation); improved nutrition - through increased food security, and better feeding practices; and measures to reduce fertility. The most useful interventions would be in the areas of female education, rural water supplies, and to some extent targeted nutrition (see nutrition section). One option may be a national hygiene campaign - which would incorporate hygiene awareness with improved water supplies and sanitation. 5.60 There are also selected health care interventions which are particularly cost-effective - including immunization (only 112 of children are currently fully immunized), the use of oral rehydration salts; micro- nutrient supplementation (iron, iodine, and vitamin A). These have a disportionate impact on health status, and for each there is an existing program which is amenable to expansion, without depending on major institutional improvements. Expansion of these should be the first line of attack. Reductions in fertility will also go a long way towards improving the health status of the poor - both because health status of children increases dramatically with spacing between births, and because complications of pregnancy and childbirth account for a large share of female morbidity and mortality. 5.61 Access by the poor to health services is hindered both by cost and by the weakness of the health care system. The direct costs of medical care represent a relative small a,erage proportion of expenditure by the poor (2-32 of total income, 7? of cash incomes). However numerous village- level reports suggest that it is a major cause of catastrophic expenditure, and is often cited as the reason for going into unsustainable debt. Unfortunately many of these payments are made for traditional cures of dubious value. 5.62 The problems of the modern health sector have been well-documented elsewhere,14/ and include: - inadequate resources ($1.85 per capita, compared to an average of about $4.00 among low income countries; - the absence of a sectoral strategy or expenditure plan which allocates resources where they would have the greatest health impact; - excessive centralization of staff and resources, and unwieldy funds release mechanisms; 14/ See for example, the Social Sector Strategy Review.  - 87 - - weak supervision and management, and a confused institutional structure. Steps are underway to address most of these - however overcoming them will be a long, slow process involving subtle service-delivery improvements. The point is that one can improve the welfare of the poor at existing income levels by improving their health status - but it should be done through measures which are not reliant on institutional strengthening of the health service. 5.63 HMG should by all means continue with upgrading of the health service - but given the institutional barriers involved (many of which are civil service-wide), and the lack of effective access by the poor, the focus should be mostly on non-medical interventions.  - 88 - VI. NEPAL - POVERTY RELATED PROGRAMS AND POLICTES A. HMG's Policies and Programs General Policies 6.1 The first few development plans during the 1960's quite correctly emphasized infrastructure investments. Recognizing that these would not create growth quickly enough the emphasis shifted (around the Fifth Plan) to raising aggregate income through production incentives and investments. Most recently the government has recognized that the benefits of this growth will reach only some segments of the population. Consequently HMG has started to concern itself with distributional aspects - particularly with assuring the provision of basic minimum needs, and with employment generation. 6.2 Historically, HMG has not in general followed redistributive policies. It uses foreign aid to finance services which benefit the poor (eg. IRDP's, health services, food aid) rather than redistribute through direct taxation. This is not an irrational strategy - as long as aid flows keep up - given the lack of resources to redistribute; although: (a) the level of taxation is remarkably low, and (b) it suggests problems of long- run sustainability of, and commitment to, these programs. 6.3 There are some notable exceptions to this trend - for example, the 1964 land reform effort, and the stated aims of the Basic Needs Program. HMG's ambivalent approach to redistributive issues has reflected the competing pressures of, on the one hand, the efficiency gains and popular support to be had from dismantling the feudal system, and on the other hand the political and economic interests of traditional elites. 6.4 Tax revenue has averaged only 6-7Z of GDP, although it has risen recently (to 8.4%) as a result of measures under the structural adjustment program to improve fiscal discipline. This puts Nepal among the lowest taxing countries in the world. The ratio of revenue-to-GNP reported in the 1989 WDR is about one-third that of the average for low income countries as a whole, with lower rates reported only for Chad and Sierra Leone. 6.5 Indirect taxes account for almost 802 of tax revenue, of which import-related taxes account for half. These are progressive in that they are directly related to consumption of imported goods, which the poor do not consume. More striking is the virtual absence of income or wealth related taxes. It is admittedly difficult zo administer an income tax in a subsistence economy, and a more useful proxy would be a land tax - although with very small average holding sizes, care needs to be taken to differentiate between true smallholders and surplus producers with fragmented holdings. The Composition of Public Expenditure 6.6 The overall composition of public expenditure (Table 6.1) is relatively consistent with a poverty alleviation str&tegy - i.e. with the emphasis on agriculture and infrastructure (17% and 23? of expenditure  - 89 - respectively), and a reasonable share of public resources going to social services (162). The problems lie more in (a) the low absolute level of financing (as a consequence largely of the low level of GDP), which results in per capita expenditure among the lowest in the world; and (b) the institutional and service delivery problems which hinder the effective use of these resources (see section on Institutional Issues below). Table 6.1: Composition of Public Expenditure - 1988189 Share of Public Per Capita US$ Share of Expenditure Expenditure Equivalent GDP (Re.) Administration 7.7Z 64 S2.57 1.6t Defense and Order 10.8t 90 $3.61 2.22 Education 9.1S 76 $3.04 1.92 Health 5.52 46 $1.84 1.12 Other Social Services 1.7Z 14 $0.57 0.32 Agriculture 17.42 146 $5.85 3.62 industry and Mining 5.02 42 $1.67 1.02 Utilities 14.12 118 $4.72 2.92 Transport 9.52 80 $3.19 1.92 Interest, Transfer and Others 19.32 162 $6.47 3.92 TOTAL 100.0S 838 $33.53 20.52 6.7 There is some room for improvement of the allocation within sectors - for example from tertiary to secondary and primary education; from curative to preventive health care; and from bulky irrigation investments to smaller schemes.l/ However in aggregate these will not make a substantial enough difference to the level of service delivered. That will require greater resource mobilization in total, as well as improvements in operational effectiveness. 6.8 In addition, many programs with a potential aileviation impact need a change in focus to more accurately address the needs of the poor. Many of the expenditures which are nominally justified on poverty-alleviation grounds are in fact only weakly poverty related. (See for example the discussion of subsidies and transfers, and targetted credit later in this chapter). 1/ See discussion in the Social Sector Review, and in relevant sectoral chapters for some specific recommendations.  - 9o - The Basic Needs Program and the Eighth Plan 6.9 Recognizing that growth alone would not lift many Nepalese out of poverty in the near future, and that general development programs were not reaching the poor, HMG announced the Basic Needs Program (BNP) in 1987. The BNP was intended to achieve minimum Asian standards of living for all by the year 2000. It included the following specific elements: Agriculture - expansion of irrigated areas; - intensive cultivation to reduce food deficits in hill areas (including rainfed research); - increased use of modern inputs - through both public and private sector distribution; - expansion of the SFDP, PCRW, and land reform, and other programs with a poverty focus. Primary Education: Expanded teacher training; distribution of free textbooks; curriculum revision; introduction of measures to encourage girls to attend school; and expanded non-formal education. Primary Health Care: Expansion of family planning, maternal and child health, immunization, and respiratory disease programs. It also included measures to increase the supply of clothing and housing by the private sector. 6.10 The BNP contains most of the elements of a successful poverty alleviation program in Nepal although it suffers from some serious flaws in approach. In some sectors it is excellent (eg. education), in others it gives the correct emphasis, but does not go far enough (eg. population). The main shortcoming is that it is centered on intensification of existing programs - often without addressing the reasons why they have failed in the past. The target-driven nature of the program could distort efforts, and in some cases force the allocation of resources to admittedly unproductive programs in an effort to achieve unrealistic targets; there is also a danger that the input targets are not interpretted in a sufficiently flexible manner, and could stifle more imaginative interventions. 6.11 The Government realized that the BNP did not adequately address employment and income generation issues. The Approach Paper for the 8th Plan (1990-1995) emphasizes the need to generate employment, and the fact that some targetting is necessary, because general programs tend not to reach the poor. The paper proposes an emphasis on agriculture, and especially irrigation, as well as suggesting that most productive sector activities will be expanded. It does not spell out the specifics of how this is to be achieved - although this may be because it is in the nature of a broad strategy document. One area of concern is the relatively low priority accorded the population growth issue in the paper. While it is acknowledged that this is important, the document does not give the  - 91 - impression that policy-makers have accepted the fact that curbing population growth is the central element of any effort to raise incomes. B. Poverty Alleviation Programs 6.12 The table below shows the salient features of the major poverty- related programs in Nepal. The subsequent sections summarize our assessment of their role 'n contributing to poverty alleviation, based on the more detailed program reviews contained in Annex II. Table 6.2: Nepal - Poverty Alleviation Programs at a Glance Approximate Annual Cost (1988/89) (Rs. millions) (US$ million) Subsidy and Transfer Programs Nepal Food Corporation 209 $8.4 Fertilizer Distribution 200 $8.0 Agricultural Credit Subsidies 48 $1.9 Irrigation Operations Subsidy 38 $1.5 Integrated Rural Development Projects 640 $25.0 Food and Feeding Programs Nutritious Foods Program 26 $1.0 Joint Nutrition Support Program 33 $1.3 Targetted Credit Programs Small Farmers Development Program 105 $4.1 Production Credit for Rural Women 12 $0.5 Intensive Banking Program 60 $2.4 Employment Programs Food For Work 38 $1.5 Special Public Works Program 13 $0.5 Income-Generating Programs Training for Rural Gainful Activities n/a n/a Labour Department Centers 10 $0.4 NGO Income-Generating Programs 25 /a $1.1 /a a/ Rough estimates only. * Full time equivalent employment. Subsidies and Transfers 6.13 Nepal does not operate the kind of large-scale price subsidy or income-transfer programs which are common in many other countries. Those which do exist are aimed primarily at providing incentives for agricultural investments, or transport subsidies to compensate for the high cost of inputs in the hills. Expenditure on the four main subsidies programs is estimated at about Rs. 500 million p.a. (US$20 million equivalent) or only 3.32 of public expenditure.  - 92 - 6.14 Food Subsidies. The National Food Corporation (NFC) provides food at below market prices to civil servants, and to the population as a whole in remote areas. The principal benefit is less in the price subsidy element than in the fact that the program makes food available in areas where there is otherwise no food for sale at all for large parts of the year. However, the majority of its food goes to Kathmandu, and to civil servants to compensate for remoteness. Probably *ess than 25Z remains as a subsidy to those living in remote areas, and vh;-. there may be some benefit to the poor, the spread of benefits in remote areas is generally at best independent of income level. (The impact of the NFC program on food security is discussed in the subsequent section on food programs). 6.15 In addition to the food distributior. program, the NFC is responsible for encouraging agricultural production with a system of grain price supports. The price supports are meant to increase production in the terai and thereby make more food available for transport to the hills. This program has had no impact, because the amount of grain the NFC buys is too small to influence the market, and the support price has almost always been lower than the market price. Moreover, prices on the Indian market dominate the terai markets and limit the potential impact of government price interventions. 6.16 Fertilizer Subsidies. The fertilizer subsidy consists of a transport element and a price element - which is to some extent a function of the need to maintain parity with .;dian prices. The net effect is that farmers in the terai and Kathmandu probably pay something close to the real cost of fertilizer, while hill and mountain farmers are paying below cost. As mentioned in Chapter 4, the subsidy has very little impact on the poor, because: - they do not use fertilizer (because they are not involved in the cash economy, and because they practice rainfed agriculture with limited scope for fertilizer use); - supplies through AIC are unreliable, and unavailable at key times; and, - subsidized fertilizer is deflected to India, to Kathmandu, or goes to the better off. 6.17 Under the second Structural Adjustment Credit HMG is starting to deregulate fertilizer distribution, and reducing subsidies to discourage deflection to India, although the total cost of the subsidy has not yet come down. At any rate it is unlikely that any fertilizer subsidy would have a substantial direct impact on the poor - given the constraints on their effective use of fertilizer. 6.18 Credit Subsidies. Through the Agricultural Development Bank of Nepal (ADBN) HMG provides interest and capital subsidies to encourage productive investments in agriculture (mostly for irrigation, cash crops, and livestock). These subsidies have little direct impact on the poor; less than 10X of such credit is estimated to go to small farmers; they generally have neither the access nor capacity to use institutional credit; and these are not the type or scale of investments which the poor undertake.  - 93 - 6.19 Irrigation Operations. The Government provides an indirect subsidy by bearing the operating costs of public irrigation schemes. This implicit subsidy probably benefits the poor to some extent, although the poor tend not to be on irrigated land, so the incidence of the subsidy is mostly on the non-poor, and probably on the wealthy (since it varies in direct proportion to area of landholdings). Our calculations show that if they are on irrigated land, the poor could afford to pay irrigation charges, and in the long run the failure to recover costs will hurt everyone, including the poor, as operations and maintenance deteriorate. HMG's policy is to move towards full cost recovery, either through higher user charges or transfer of responsibility for O&M to beneficiaries. 6.20 Grants are also provided for the capital costs of small irrigation schemes. It has not been possible to estimate the total costs of this subsidy, as it is distributed among a number of projects in the budget, but we suspect it is large. While promoting small scale irrigation is a legitimate poverty-alleviation strategy in Nepal (in fact probably one of the most effective), given the very high private returns it is not clear that a capital subsidy should be necessary except for very small farmers. It is recommended that consideration be given to replacing the capital subsidy over time with a credit program, with an exemption for very small landowners, while retaining the technical assistance elements of such programs. 6.21 Conclusions. Existing subsidies and transfers do not result in any substantial transfer of benefits to the poor. There may be other reasons for continuing these programs (eg. to accelerate sectoral growth) but they do not appear to be justified on poverty alleviation grounds. If the objective, in the minds of policy-makers, is to alleviate poverty, then these funds would be better spent on other programs. Even the non-poverty objectives may be better achieved otherwise - eg. by increasing civil servants salaries rather than providing subsidized grain, or by credit rather than capital subsidies for irrigation works. 6.22 Some elements of the transfers and subsidies are poverty-related (eg., NFC's general food distribution for remote areas, SFDP subsidized credit) and these could benefit from being strengthened. Wholesale transfer programs are probably not viable in Nepal in the foreseeable future, given the magnitude of the problem and the limited financial resources available to HMG. (For example, with 7 to 10 million poor, and no easily identifiable sub-group of ultra-poor, the cost of a transfer program which supplemented incomes of the poor by only 202 would be about Rs. 3.9 billion annually, or 251 of public expenditure.) 6.23 The transfer interventions which do make sense in Nepal are those which improve food security for the poor (especially in remote areas), and selected transfers to poor groups or individuals for income-generating activities, where these can be effectively identified and targetted - eg., under some NGO and targetted credit programs. RecommendatioaLs to strengthen existing programs (NFC, SFDP, PCRW) are made elsewhere. In addition, HKG may wish to consider additional programs for targetted feeding, nutrition supplementation, and public works employment. 6.24 Subsidy programs (especially for foodstuffs) have had a substantial impact on the welfare of the poor in many countries (eg., Sri Lanka, Mexico,  - 94 - India), although they have operated largely to the benefit of the urban poor. With the emergence of a large number of urban poor in Nepal over the next 20 years the pressure (and need) for such a program will mount. At that point it will be necessary to re-examine the need for price subsidies, but for the moment there is no justification for subsidizing foodstuffs to urban consumers. The remote areas food distribution element of NFC's subsidy is legitimate, and should be refined (see para. 0.45) - possibly utilizing some of the resources which currently benefit Kathmandu. 6.25 Finally, a whole constellation of subsidies, incentives, and payments in Nepal are designed to compensate for the high costs of access in the hills. In our view it would, in general, be more cost-effective to spend these resources on improving physical access (at least in those areas where it is economically viable) rather than on an inefficient conglomeration of subsidies which are difficult to target, costly to administer, and prone to mis-use. Integrated Rural Development Proiects 6.26 There are at least nine on-going IRDPs, at a total cost of over $250 million. Like IRDPs elsewhere, they take an area focus, include a wide range of sub-projects (agriculture, roads, health posts, schools, water supplies, income-generating projects, etc.), and attempt - at least in theory - to integrate thet in a regional development strategy. In fact the sub-components are often indistinguishable from general sectoral programs, except sometimes for implementation arrangements. 6.27 IRDPs are not poverty alleviation projects. However to the extent that they are in relatively poor areas of an absolutely poor country, many of the intended beneficiaries are poor. They tend, by design, to focus on smallholders and those with some productive assets - they do not address the policy issues (eg. land reform, employment, tenant's rights, migratipn) which most effect the landless and rural labourers. In Nepal, instead, they have increasingly focused on "process' issues of beneficiary participation and organization. 6.28 Internationally, the experience with IRDPs has been less than satisfactory - so much so as to call the validity of the approach into question.2/ The main problems have been: the lack of sustainability of project benefits when activities are financed at a higher level than in the economy as a whole; (ii) unmanageable complexity; (iii) excessive reliance on special project implementation groups: and, (iv) failure of ihe expected benefits of integration to materialize. The IRDP's in Nepal manifest some or all of these of these problems. 6.29 None of the IRDPs reviewed under this study can claim to have been very successful either in promoting rural development in general or in meeting the needs of the rural poor in particular (see Annex II.3). In all six cases there is an absence of documentary evidence of project effects that goes much beyond the achievement of targets or accounting for inputs. In most cases the projects are able to claim that they achieved their 2/ See for example, World Bank Experience with Rural Development Projects 1965-1985. IBRD, 1937.  - 95 . targets within budget, but unable to claim improvements in household income or nutritional status. 6.30 Comparing the performance of agriculture in six IRDP areas with the regions in which they are located reveals no clear trends. The best one can conclude is that the performance of IRDP's may have been no worse than that of other development initiatives in agriculture or forestry, but the performance in these sectors during the last two decades has been marked by almost systematic deterioration, which intensive inputs under IRDP's has not arrested. To the extent that benefits were generally postulated in terms of increased agricultural productivity, one might have hoped for more from expenditures of $25 million p.a.. 6.31 There have been some particularly successful sub-components (eg. the Pakribas Agricultural Centre under the Kosi Hills project, financed by ODA); and the design of more recent IRDPs reflects some learning from experience - both with more appropriate implementation arrangements, and many fever components.3/ However, in the absence of any clear indi.ation of success, and given the verv high levels of expenditure, one needs to question whether or not it is more effective to focus instead on; (i) strengthening overall sectoral programs (eg. agriculture, forestry, health); (ii) financing essential infrastructure; (iii) building up the capacity of line Ministries to deliver the services for which they are responsible (including building in the useful 'process, reforms being sponsored under some IRDPs); and, (iv) undertaking the necessary countrywide programs and policy reforms which will affect the poor (eg. population measures, developing an appropriate technical package for hills agriculture, investing in road access and education). 6.32 It is recommended that HMG consider adopting this approach rather than pursuing new IRDPs, or, if it is felt to be desirable to focus efforts on some less-developed areas because the strengthening of general programs will take too long, then HMG should consider limiting IRDPs to a restricted number of activities (probably irrigation, and road and trail building - factors which can make a quantum difference to productive potential). Other Intensive-Agriculture Programs 6.33 The Pocket and Block Production programs of the Ministry of Agriculture provide intensive agricultural support services to selected rural areas which have irrigation and which are accessible. These programs are intended to accelerate general agricultural growth, rather than to help the poor specifically. In the areas covered, Ministry staff are supposed to construct cropping plans in collaboration with local farmers and ensure the timely availability of inputs. Local studies commissioned for this report show that in areas where adequate staff is available, substantial increases in production can occur. However, staffing levels in many areas have not been adequate and the Ministry has not been able to insure availability of inputs. It is not clear that these programs are more effective than a well functioning general extension program, and they have concentrated on areas with superior resource endowments, drawing resources away from poorer areas. Given the uneven experience in implementation it ma-- be better if these 3/ Dhading (1985), for example, relies on local implementation capacity, and Seti and hechi (1987) both focus only on irrigation and roads and trails.  - 96 - programs were discontinued and funding and staff shifted to improving the general extension program. 6.34 The Lumle and Pakhribas Agricultural Centers, both funded by ODA, have developed innovative methods for integrating research and extension services so that the focus of research is relevant to local farmers. Both Centers stress on-farm varietal testing and frequent contact between researchers and farmers. The high levels of staffing needed may limit the ability of these Centers to expand their areas of coverage. In addition. the areas currently covered are somewhat better endowed in terms of agronomic characteristics than many hill areas, and this may have contributed to the success of the Centers. However, the emphasis on frequent contact between farmers and researchers is the most effective method of developing agricultural technologies relevant to local conditions. Food and Feeding Programs 6.35 The Nepal Food Corporation attempts to reduce inter-regional food imbalances by buying food in surplus areas and selling it at subsidized prices to the poor and government officials in food deficit areas. Political considerations channel most NFC-distributed food to the Kathmandu Valley, despite the relative lack of poverty in the area and the well established market system. In remote areas, NFC food is flown in at great expense, and even in less remote areas the transport costs are heavily subsidized. 6.36 The food distributior program has not had a significant impact on the food security of the poor. The quantities distributed have been too small, usually less than 152 of the deficit in areas outside the Kathmandu Valley, many of the beneficiaries are government officials, and large amounts of the grain goes to people too wealthy to meet eligibility requirements. Despite the problems with this program, the need for it is increasing as population growth continues to outstrip agricultural growth. The private market will not provide sufficient food in remote areas in the near future because effective demand is too low at the prices private traders must charge to cover transportation costs. 6.37 The NFC is starting a local grain storage program to try to reduce seasonal price fluctuations by giving poor farmers the opportunity to hold grain after harvest until prices rise. The program centers on organizing poor farmers into savings groups with rotating loan funds, eo that they can escape the necessity of selling their crops immediately after the harvest in order to repay debts. 6.38 Food-for-work projects supported by the World Food Program appear to be more successful in reaching the poor than the NFC food distribution program. The potential impact of this program on food security is quite large, although at the moment the total food distributed as wages is less than 2,000 mt p.a., compared to the NFC which distributes 34,000 mt/year. and relative to a total average annual deficit of 300,000 tons. 6.39 The Nutritious Food Programme, also supported by VFP, reaches a larger number of people (190,000 in 1987), and distributes a larger amount of food (5,500 mt in 1987), but some aspects of the program have targetting problems and the full development potential of the food distribution program  - 97 - is not utilized. It includes distribution of food to malnourished children and pregnant and lactating women, and institutional feeding programs at primary schools, orphanages, and child care centers. 6.40 The distribution of food to women and malnourished children is the largest segment of the program. Children are screened once a year, and if they are malnourished, they become eligible for supplementary food. Pregnant and lactating women are automatically eligible for food supplements. Targetting is a problem, with the food often distributed irrespective of income, gender, or age, and case studies have shown massive amounts of leakage. If the targetting problems could be overcome, and if the food distribution was accompanied by health and nutrition education, the impact of the program could potentially be substantial. 6.41 The Joint Nutrition Support Programme, under the guidance of UNICEF, has attempted to coordinate the activities of four government ministries on a wide range of activities (nutrition, agriculture, food storage, sanitation, education, women's income generation, etc.) on the grounds that the multiple causes of malnutrition need to be addressed as a group. The program has largely failed because it is excessively complex, because it is not the sole responsibility of any single ministry, and is not a high priority in any ministry. 6.42 The combined impact of all the programs discussed in this section is relat:vely small compared to the magnitude of the food security picblem in Nepa. All the food handled by the NFC and the World Food Program represents less than 102 of the national food deficit, and much of it goes to households that are not food insecure. In addition, most of the programs have not changed long-term food security at the household level. However, if expanded and improved, they could be important parts of household coping mechanisms, either by supplying subsidized food when supplies are low or by providing seasonal employment. 6.43 The main elements of a strategy to improve food security in Nepal need to revolve around improving agricultural productivity and physical accessibility and raising incomes. However, given the length of time this will take and the remoteness of many communities, some directly food-related measures also form a legitimate part of the solution. 6.44 There is some role for distribution of food to deficit areas; however, one has to question both the sustainability of such transfers, and the capacity of the existing NFC framework to deliver food efficiently without leakage. The preferred solution would be to develop the capacity of private traders to supply food as purchasing power increases in hill areas. However this will not happen quickly enough to relieve food shortages for many, and therefore it is legitimate to distribute food, the question is how best to do it. 6.45 The NFC would have to distribute more food than it currently does to have a significant impact. A restructured program would be beneficial if: (i) the supply of food to public servants were divorced from other feeding functions (as long as it remains it will divert effort, and scarce supplies will always tend to be allocated first to public servants); (ii) better geographical targetting is necessary, as a first step the Kathmandu  - 98 - valley should be eliminated entirely from the program; and, (iii) the scope for more effective beneficiary targetting should be investigated - although we remain skeptical of the capacity to effectively target general food distribution in Nepal. (Food-for-work, for example, may be a more effective mechanism). 6.46 Storage schemes to reduce seasonal variations may have some impact. They would have to be accompanied by a credit component since most poor households must sell crops immediately in order to pay debts. However, the experience worldwide with such programs has been discouraging, and the relevance of storage programs is questionable in the absence of increases in incomes. While tnere may be an element of oligopolistic pricing which could be reduced by more cooperative storage, one must assume that those currently purchasing and storing grain would be willing to sell it within the area, if those living there had sufficient purchasing power. 6.47 Targetted group feeding programs have proved very difficult in Nepal. Most have failed, and Freedom from Hunger, one of the largest NGO's, is shutting down its feeding operations because of the difficulty in targetting. The Vulnerable Group Feeding program has recently been restructured to aim more specifically at the malnourished - it is recommended that the experience with this program be reviewed in about a year's time, along with that of similar programs in other countries before proceeding with further targetted feeding initiatives. It should be noted, however, that all of the staff with field experience in Nepal interviewed by this mission questioned the capacity to delives. ny targetted program. 6.48 There is a lot of mileage to be gained by changing feeding practices. A hygiene/nutrition education program, if it could be well designed and adequately staffed, could thus potentially play a useful role. While changing behavior is always a problematic proposition, experience to date suggests people are receptive, and that the messages to be transmitted are not that complex. (They involve not withholding food from diseased persons and women during pregnancy, better use of weaning foods, boiling water, etc.). Targetted Credit 6.49 Targetted credit (perhaps due partly to the the relatively poor performance of most IRDPs and sectoral programs) has become an important element in Nepal's poverty alleviation strategy. Currently, considerable attention is being focussed on three programs: the Small Farmers Development Program (SFDP), Production Credit for Rural Women (PCRW), and the Intensive Banking Program (IBP). All three combine elements of community development, income-generation activities, and the provision of credit.  - 99 - Table 6.3: Targetted Credit Programs INTENSIVE SFDP PCRW BANKING Interest Rate 151 161 161 No. of Beneficiaries 15,000 p.a. 2,700 p.a. n.a. (Cumulative) (78,500) (6,650) (62,150) Disbursements Rs. 150 m. p.a. Rs. 8 m. p.a. Rs. 140 m. p.a. (Cumulative) (Rs. 644 m) (Rs. 14 m) (Rs. 814 m) Recovery Rate 48? 911 57? Average Loan Size Rs. 1,900 lb Rs. 2,105 Rs. 13,100 Estimated Delivery Cost Rs. 2,050 Rs. 4,075 Rs. 1,010 per Beneficiary \a a/ Administrative costs only, much higher allowing for provision for bad debts. b/ May be substantially higher, average lending per beneficiary is Re. 8,200. 6.50 SFDP, under the Agricultural Development Bank of Nepal (ADBN), relies on a network of facilitators who organize small farmers into groups for credit and developmental activities. It appears that most credit is used for working capital and livestock, and that about 20-30? finances longer-term improvements (eg. irrigation and horticulture). Program design incorporates most of the successful elements of similar programs elsewhere (group liability and savings, intensive technical support), however these appear to have lost impetus in the process of implementation - especially following recent expansion. 6.51 PCRW is a government program under which Women's Development Officers form groups of village women for developmental activities and small-scale business undertakings, for which they then receive slightly subsidized commercial bank financing under the IBP (below). PCRW is not primarily a credit program - it has lent to only 6,500 persons (the vast majority for livestock), and 60S of its costs are non-credit related. It is a general development program for women, which uses credit as an entry point to organize groups as vehicles for such activities as literacy, health and family planning, and small public works. Unit costs are high, although this may reflect naturally expensive mobilization efforts. 6.52 In neither program are interest rates much below average commercial lending levels. Delivery costs are very high, and (with the exception of PCRW) recovery rates, while no worse than for many other types of lending in Nepal, are not high enough for the programs to be financially self- sustaining. HMG currently supports both programs by utilizing concessional financing from donors (mostly the ADB, IFAD and UNICEF).  - 100 - 6.53 Despite the weakness of evaluation reports in distinguishing between credit-related benefits and general income increases, it appears both programs are having some impact on incomes among the moderately poor. Case studies of borrowers however reveal many instances of business ventures which fail, or use of fands for non-productive purposes. Moreover, experience from other countries (as well as information on landholding sizes among SFDP clients) indicates that credit-based approaches work best for the 'middle poor' - those within say, two deciles of the poverty line, who have some assets, skills, and their socio-economic connections sufficiently intact to take advantage of income-earning opportunities. 6.54 The two programs currently reach an estimated 52 of the population below the poverty line. There are serious questions about the extent to which they can expand effectively to reach a more significant proportion, because their success is dependent on the intensive use of dedicated field staff, who are in short supply. Already there are signs that both programs are suffering from too-rapid expansion, and inadequate training of new organizers. It is recommended that: (i) both programs should consolidate before further expansion; and (ii) training of group organizers should be strengthened (see Annex II). 6.55 The Intensive Banking Program is different from the other two in that it involves HMG directing the commercial banks to lend to particular groups and purposes. Banks are required to lend 252 of total advances4/ to the productive sectors (broadly defined), and 82 to "Priority Sectors, (eg. small agriculture and cottage industry), of which 602 is to be directed to families below the poverty line. Banks are also expected to provide some of the development functions provided under the other two programs. While there has been some success in raising incomes of the moderately poor, most credit has gone to larger cottage industry projects. The program is characterized by non-compliance by the banks, because the unit costs of small loans are sufficiently high that the banks are losing money on them, banks are not reimbursed for the costs of non-banking services, and the penalty for non-compliance (holding of deposits in non-interest bearing accounts with the central bank) is not a sufficient deterrent to ensure compliance. 6.56 Banks have traditionally been extremely conservative in Nepal - lending only against very safe investments (goods in warehouses, land, jewelry). It is legitimate for HMG to try to encourage the financial sector to undertake more productive lending, and to hasten its spread into rural areas. However, experience worldwide raises doubts as to the sustainability of directed credit; because: (i) it is inefficient, if one is trying to develop a commercial banking system, to then interfere with commercial decision-making; (ii) if banks are not interested they will ultimately not do it; and (iii) if credit is subsidized, it tends to be captured by the non-poor. In Nepal these problems are compounded by the extreme weakness of the banking system, which makes it more difficult for banks to bear the costs of delivering non-credit services required under IBP (and administratively less able to do so). 4/ Defined as investments plus loans.  - 101 - 6.57 There has been a tendency ttlroughout to combine legitimate (but largely non-credit) objectives of group formation, organization of the poor, and assistance in project preparation, with the provision of credit. The use of small-scale credit-led programs as an entry point to reach the moderately poor is legitimate, but for larger scale operations, such as those under IBP, a more useful approach may be to provide these services under the auspices of some developmental program, which would, among other things, help its clients secure credit (either from SFDP, commercial banks, or non-formal sources). It is recommended that for commercial scale investments, HMG divorce the purely commercial banking functions from these developmental functions, and that commercial banks not be expected to deliver developmental services. (At a minimum, if IBP is to continue, HMG must explicitly subsidize the non-credit costs of the program). For micro- level investments which affect the poor, it is recommended that expansion be concentrated on SFDP and PCRW, and not involve the commercial banks. One option might be establishment of a 'Poverty Bank", or development fund which could incorporate financing for both programs outside of the commercial banking system. 6.58 Under a recent variant of IBP (the Lead Bank Scheme), one bank in each area is also to help coordinate the delivery of other developmental services. A major expansion of the scheme is proposed as part of the Basic Needs initiative - which would involve the banks increasing by ten-fold their lending to the priority sectors, and directing a quarter of their lending to the poor. In early 1990, HMG announced a further new initiative, whereby injections of credit to poor families (mostly through the commercial banking system) would be combined with local and national-level coordination of supporting development services in a major poverty alleviation effort. This proposal appears to be a combination of the SFDP and Lead Bank Scheme approaches. 6.59 It is questionable whether the strong emphasis on credit as an income-raising strategy is warranted. It is predicated on the assumption that credit (rather than the lack of investment opportunities) is the binding constraint; and that to the extent that such opportunities exist, the poor will be in a position to take advantage of them (which they seldom are). There are severe limitations on productive investment opportunities (lack of transport, markets, and of a monetized rural economy) - any one of which would be more binding than the investment-financing constraint. Certainly credit is one of a number of important inputs, and HMG should continue to ensure it is available as part of its small business development strategy, however to the extent that such limited opportunities exist, it is unlikely to be the poor who will be in a position to take advantage of them. 6.60 Furthermore, we have seen (in Chapter 3) that the problem of indebtedness among the poor in Nepal is one of insufficient incomes to meet consumption expenditure - in these circumstances the insufficiency of credit is a symptom of poverty, not its cause. This is confirmed by disturbing stories suggesting that many SFDP and other loans intended to raise incomes are used to finance consumption or repay previous debts, rather than to finance productive investments. There are three ways in which credit may be of use to the poors some micro-investments (eg. purchase of a goat or buffalo) can improve their incomes marginally; providing formal credit may  - 102 - reduce the consumption debt-service burden, and reduce dependence on landlords or employers; and small amounts of credit for working capital may be useful to small farmers (who could not otherwise get access to it on affordable terms). These are arguments for providing small amounts of highly directed credit which will have a useful but limited welfare effect. Such micro-credit should not be confused with more substantial poverty alleviation through the massive provision of credit, such as is currently being proposed. Employment Creation Projects 6.61 There are no large-scale employment programs as such in Nepal; the Special Public Works Program (SPWP, sponsored by ILO) is focussed on small irrigation projects with a labour-intensive emphasis, and the Food-for-Work program is part of the WFP's food distribution efforts. SPWP has employed perhaps 10-20,000 rural people in construction, many of them poor - but the main benefit has been from irrigation improvements, which tend to benefit households in direct proportion to their landholdings. It has had success in raising the labour intensity of rural works, achieving rates of 56Z of expenditure on wages, as opposed to 30-40Z otherwise - but at the cost of intensive supervision and support for very small local petty contractors. The project has also experienced equity problems in the use of *voluntary' labour, because despite being designed to maximize beneficiary participation, in practice the workers involved in construction are often not from the same groups as those who benefit from the irrigation works. 6.62 SPWP is useful as a tool for increased irrigation and agricultural intensity - but in this respect not differentiable from intensifying -rrigation in general. As a direct employment program it is unlikely to have a major impact on poverty at its current level, as it creates only an average of 1,350 full-time construction jobs. Wider use of the same approach - through an expanded SPWP or otherwise - could however have a significant impact on construction employment. 6.63 Irrigation (and other rural works) are probably more likely to benefit the not-so-poor - careful attention needs to be paid to selection of sub-projects, and consideration given to compensating measures which spread the benefits more equitably (the example of a water tax, combined with paid maintenance workers, is a good one). 6.64 Subject to a review of cost-effectiveness, the program probably should be expanded as rapidly as its absorptive capacity allows. The approach should be replicated in other areas - most notably on rural roads, with attempts to refine the system of small-contractor use. Water taxes or user charges should be introduced once schemes are completed, and the proceeds used to pay wages to poorer villagers for operations and maintenance. 6.65 The Food-for-Work program (see para. 6.38) provides about 1 million days of employment per year (6,600 full-time equivalent jobs). It pays an in-kind wage, the value of which is primarily attractive to the poor. In addition the work, building roads and trails, is offered during the dry season, when other employment opportunities are scarce.  - 103 - 6.66 Some useful lessons emerge from rural employment schemes elsewhere. In particular, the experience with programs in India suggest that they must provide employment on a significant scale, and on a reliable basis, to have a meaningful impact on poverty status. The Employment Guarantee Scheme in Maharastra has been particularly successful by providing an employment guarantee which has broadened the material options for the poor, and reduced the extreme vulnerability which is a hallmark of poverty. 6.67 It incorporates a number of instructive features. First, the wage has been intentionally kept below prevailing market rates - dissuading healthy adult males from participating except during seasonal lulls or when disasters have hit. Labourers, who are invariably drawn from landless and marginal cultivator households, are primarily (at least 60Z) women who are attracted by the proximity of work sites, by the self-pacing nature of the work (piece-rates apply), by the pattern of working in groups, and by the practice of paying wages directly to laborers without intermediation by contractors or the panchayat. It also involves extensive monitoring and supervision of work sites. 6.68 However, what distinguishes EGS from other rural employment programs is the statutory work guarantee which represents precisely the kind of insurance which is needed by the poorest. The susceptibility of very poor households to sharp reductions in income and loss of assets can only be counterbalanced by a non-alterable fallback arrangement of some type. 6.69 The feasibility of such a broad-based program in Nepal is questionable, both because of the high costs and because of limited administrative capacity. However, some version of the EGS may be useful in particularly poor areas, and elements of the approach could certainly guide design of a labour-intensive rural works program. It is recommended, that following preparatory analytical work (see para. 4.44) HKG undertake a program of public works employment aimed particularly at the poor which incorporates features of the SPWP and food-for-work, and EGS programs, as well as drawing on experience of similar programs elsewhere. Income Generating Projects 6.70 There are many small income-generating projects focussed on activities such as tailoring, bee-keeping, sericulture, and handicrafts. They are usually wrapped up with more general community development efforts of NGOs which combine agriculture with some health care, education, and small business projects within a few villages or districts. 6.71 In many cases they involve very intensive support (eg. helping to supply inputs, providing the technology, and marketing of outputs). Reliance on such extreme levels of external support raises questions about the sustainability of these projects. While such intensive interventions may dramatically affect the incomes of individual poor households, they do not address the more general issue of constraints on the overall expansion of non-formal sector activity (Chapter 4).  - 104 - 6.72 The aggregate numbers being reached are small (although not insignificant). There are almost no quantitative evaluations of these programs. There are some reports of localized successes - often where access has recently been improved - and many accounts of financial failures. It may be indicative that the incoD,e-generating components of several IRDPs have been dropped as non-performing. The areas of success appear to be those associated with forming the poor into groups, promoting self-reliance, and the like. 6.73 The intensive use of resources to reach small numbers of families suggests that this is not a major element of poverty alleviation strategy. However it is possible that the high costs of supporting these enterprises represent transition costs associated with the change to a cash economy - which will decline as people become more accustomed to starting small businesses, and as input and marketing systems become more developed. These are the sort of interventions in which NGOs are particularly useful at targetting, although their capacity to pick financially sound projects appears weak. It is recommended that consideration be given to establishing a fund for financing small income generating projects for NGO implementation with strict beneficiary criteria, and a small technical group to help appraise the financial feasibility of sub-projects. 6.74 A wide range of programs also provide training for self-employment or to provide marketable skills. The largest are Labour Training Centers run by the Labour Department and the Training for Rural Gainful Activities program (TRUGA). It is not obvious that they are particularly selective of the poor, rather than of the population in general. Most evaluations report only limited success - mostly because of the lack of employment opportunities for those trained. TRUGA, for example, while having success in integrating into local villages, has trained only 1,000 people over four years. A multitude of small programs provide training specifically for women - in the usual areas of sewing and handicrafts - with little analysis of whether there are opportunities for them to profitably use these skills, and equally little employment of those trained. While skills training potentially has a contribution to make to long-term growth in Nepal, the projects tried to date would need better design and implementation to have an impact on the incomes of the poor. C. Institutional Issues Civil Service and Service Delivery Issues 6.75 The civil service is still in the early stages of development - prior to the 1960's its functions were limited to tax collection and maintaining public order. A wide range of problems plague the delivery of public services. They include: - Inadequate Renumeration and Incentives and Poor Supervision - which result in high absenteeism, low morale, and under- staffing in rural areas; - Promotions and Postings Practices - based on narrowly - defined performance criteria, and influenced by patronage connections;  - 105 - - Inadequate Operating Funds - coupled with unwieldy funds release mechanisms - which result in inadequate supplies and travel - especially to outstations; - Excessive Centralization of resources, staff, and decision- making; _ Attitudinal and Communications Problems - which prevent effective contact between public servants and illiterate peasants. 6.76 These affect all HMG activities, they are not specific to poverty- related programs, However, they have a disproportionate impact on: (a) interventions which require complex service delivery, and (b) the quality of services in outlying areas; both of which are particularly important in reaching the poor (services such as health, family planning and agricultural extension are especially hard hit). 6.77 The problems of public service appointments, posting, supervision, etc. - while deeply entrenched - can probably be ameliorated by development of more systematic personnel and administrative systems. It is recommended that HMG undertake a systematic review and reform rather than addressing them sector-by-sector. In the meantime, there is an acute shortage of local service delivery capacity, and this needs to be borne in mind in designing poverty-related interventions. Decentralization and Popular Participation 6.78 There are two elements of "process reform which have a potential bearing on poverty alleviation efforts in Nepal. The first is greater self- reliance, participation in development initiatives, and organization among rural people. The second is decentralization of administrative responsibilities to District and local level, along with the establishment of local representative councils (pursuant to the Decentralization Act of 1982). 6.79 With respect to the first, many engaged in development in Nepal believe that the predominantly top-down approach followed to date has resulted in few benefits, selection of inappropriate projects, and an erosion of self-reliance; and that community mobilization can lead to more effective interventions. Some success has been had in forming groups (eg. of forestry and irrigation users, and under some NGO operations and IRDPs) - which allow the poor to more effectively demand and utilize services. While such initiatives can be very useful, their success: (a) is often dependent on the presence of a skilled organizer, and (b) has been limited to the extent that they do not challenge the existing political (and economic) order. They yield useful lessons for the design of other interventions, but as a broad poverty alleviation strategy their effectiveness may be limited to the extent that the binding constraints (i.e. the distribution of real political power) is not being lifted. 6.80 Administrative decentralization can help the poor to the extent that resources are moved closer to poor beneficiaries. So far, however, there has been a natural reluctance to transfer real control over staffing and budgets. In addition, if revenue generation for decentralized functions  - 106 - is transferred at the same time, the poor may lose - as poorer Districts, which have the weakest resource base, will be less able to support services or finance local projects. In addition, placing responsibility for delivery of complex services at the level of government with the fewest resources and the least experience of planning and implementation seems unreasonable. 6.81 The transfer of political power to local levels can potentially increase responsiveness. However, experience with local government elsewhere in the world does not suggest that it is particularly more representative of, or responsive to, the needs of the poor. (Nor does experience in Nepal suggest that this is the case here (see Chapter 3). Decentralization is a necessary stage in the evolution of administration in Nepal, but it should not be counted on to have major poverty-alleviation effects. Non-Governmental Organizations 6.82 There are a large number of international NGO's active in Nepal (although mostly not engaged directly in income-generating activities), while the national NGO sector is still in a fledgling state. There have traditionally been a wide range of indigenous self-help groups and community service organizations - but Nepal does not have the range of local developmental NGO's found in neighboring countries. 6.83 The activities of some 150 NGO's were surveyed for this study; most of them are engaged in community services or the provision of health and education; although some (about 15) are increasingly engaged in income- generating activities. Among the most active are Action Aid, Save the Children, CARE, Lutheran World Services, World Neighbors, Redd Barna, the Association of Craft Producers, and Integrated Development Systems, a consultancy-cum-development agency. 6.84 The benefits of NGO's are that they can try more innovative solutions, are not constrained by government staffing and financing procedures, and can work intensively at the local level. This intensity is also a weakness, in that their capacity to operate on a large scale is limited by the availability of dedicated staff. Major donors, on the other hand, have the resources to finance programs large enough to have an impact on poverty, but not the capacity to design or implement the intricate service delivery and social mobilization mechanisms needed for theta to succeed. There are thus opportunities for the greater involvement of NGOs in executing large-scale projects, as well as in enhancing the *process* aspects of implementing government's general development programs. 6.85 The principal benefit of NGOs may be that they can operate free of the political and social constraints which limit the capacity of government programs to help the poor. With respect to poverty alleviation efforts, caution should be exercised, because to the extent th&t existing economic interests are threatened by organizing the poor, it seems unlikely that NGOs will be allowed to operate freely. 6.86 NGO activities are governed by the Social Services National Coordinating Council (SSNCC), through which both funds and approvals for NGO activities are channeled. HMG's attitude towards NGOs has been ambivalent;  - 107 - the SSNCC provides a central monitoring and coordination function, but is also an agent of control. This reflects a legitimate concern to prevent a plethora of agencies (many of them foreign) operating in the countryside un- policed, but it may also limit the effectiveness and flexibility of NGOs. There is room for revamping the framework for NGO coordination, especially if NGOs are to play the active role prescribed for them under the Basic Needs Program. It is recomiended that a review be undertaken of (a) the regulatory framework for NGOs, taking account of arrangements in other countries; and (b) the scope for establishing a fund through which modest levels of donor and NGO financing can be cbanneled to local NGOs; and assistance can be provided in strengthening their administrative and project selection capacity.  - 108 - VII. COUNTRY STRATEGY IMPLICA'IONS 7.1 This chapter consolidates the findings on the potential contributions of the various sectors to raising incomes and absorbing labour. The objective is to point to areas in which to concentrate efforts, and to frame the bounds within which HMG needs to design a poverty alleviation strategy. Specific policy and program recommendations are discussed in the following chapter. 7.2 Nepal is a very poor country, with a relatively even distribution of income and land, therefore there is limited scope for poverty alleviation through redistribution or welfare measures, the need is for targetted growth. However at current rates of population growth (2.72 p.a.), and any expected levels of economic growth, there is no hope of seriously reducing the number of poor. Curbing population growth is thus one of the central elements of poverty alleviation strategy for Nepal. To reduce the incidence of poverty to say 202 by 2010 (even with optimistic assumptions regarding GDP growth) would require reducing the population growth rate to around 1.72 p.a.. In the absence of this sort of achievement, any other income increasing measures will be meaningless. 7.3 While effectively curbing population expansion is central to poverty alleviation efforts, there can be little hope of major advances as long as economic growth averages below 3Z per annum. Nepal faces formidable natural disadvantages, and in this restricted environment, government policy must be particularly effective to have a meaningful impact. The Government will need to pursue a more active program of reforms than it has generally over the last three decades, and in particular it needs to follow through firmly on the implementation of reforms and on the delivery of programs. 7.4 We have seen that about two-thirds of rural hill households and just under half of terai households could never be expected to rise out of poverty on the basis of agriculture alone - their holdings are too small to do so. it is incorrect to keep thinking of these people as poor farmers, rather they are poor households, who happen to own some land. Their coping strategies thus have to depend on a range of interventions. This is not to say that agriculture is not important, especially in the terai. The poor receive half or more of their current income from agriculture, so even the increases of 20-302 in output which are achievable would have a substantial impact. In addition, they are disproportionately affected by the demand for agricultural labour. Equally importantly, experience in other countries shows that while agriculture itself may not be the solution to rural poverty, reasonably equitable growth in agricultural incomes leads to very significant expansion of non-farm rural employment. 7.5 Furthermore, it should be remembered that in a heavily agricultural economy which is not heavily taxed, and where the agricultural transition has scarcely bagun, the social safety net needs to be built of resources found in farming itself; resources are not yet being transferred out of the sector for welfare purposes, that is, to take care of people crowded out of the economy for lack of adequate resources, or because of age or infirmity of one kind or another. So while these small plots may not produce much, their major use is for the succor of the families that live on them in a job-scarce economy. In economies such as that of Nepal  - 109 - agriculture must serve not only the functions we usually attribute to farming, but also as a social safety net. 7.6 About a third of the agricultural poor in the hills have enough land to potentially raise themselves above the poverty line through agricultural improvements (but only if their land is irrigable). In the terai, perhaps 402 of the poor have large enough holdings (say abo_. 0.6 ha.) to produce above-poverty household incomes, if fully irrigated. Perhaps another third of the agricultural poor would, with foreseeable technical improvements, produce a significant share of poverty-line incomes from agriculture alone. Labour absorption as the result of transformations in agricultural technology could potentially result in employment for about an additional 250,000 persons in the hills and one million in the terai.l/ It can be assumed that most of these jobs will absorb the poor (or family members of non-poor households who would otherwise become poor as a result of land sub-division). In sum, the best that could be hoped for from agriculture in the foreseeable future might be to support another four or five million of the poor (about 700,000 households). Scope for land redistribution is limited, especially in the hills; however a redistributive reform in the terai - if carefully targetted - might raise 400,000 households out of poverty. 7.7 We have seen (Chapter 4) that the formal sector might be expected to generate an additional 1.2 million jobs by 2010, although outside construction, few of them absorb the poor. (Bearing in mind that at current wage levels and family sizes, off-farm jobs do not generate sufficient income to raise the poor out of poverty unless several family members are employed.) Hydro electricity exports have the potential to increase GDP by about 10X. However beneficial income effects depend on following consciously redistributive policies, and taking very specific steps to counter terms-of-trade effects which would be biased against the poor. The scope for macro-economic adjustments which would significantly impact on incomes of the poor appears to be extremely limited. The only caveat is that if India were to undergo a major liberalization of its trade regime, then Nepal, by following suit, could potentially open up much faster formal sector growth. 7.8 The informal sector can at best follow growth led by the other sectors - with allowance for a catch-up effect due to the current low level of monetization. Measures tried in other countries to 'unleash' self- employment in the non-formal sector (eg. credit, skills training) may not be very effective due to difficulties of access and the lack of an effective cash economy - although the scope in urban areas and the terai is growing. Analysis in Annex I.6 suggests that about an additional 1.8 million persons might be absorbed in informal activities within the next twenty years. 7.9 Table 7.1 shows that if our (relatively optimistic) expectations are met regarding growth of GDP and formal sector employment, and intensification of agriculture, then labour absorption can just about keep pace with labour force growth. The absolute number of under/unemployed (as 1/ See Annex 1.4.  - 110 - a proxy for the number of poor) would increase by about one million, although the proportion of the LF underemployed is substantially lower than today. If, on the other hand, GDP growth is slower, and agriculture performs only slightly better than it has done to date, then the number of unemployed will increase dramatically - perhaps more than doubling. Finally, the last line in Table 7.1 corresponds to an absolute transformation of the Nepalese economy, such as has been experienced in Malaysia or the Republic of Korea. Such a transformation, however, appears unlikely at this stage. Table 7.1: Possible Labour Absorption - 2010 (millions of full-time equivalent jobs) Non-Formal Formal (Under/ Agriculture/a Sector Sector Total LF Unemployed) 1990 2.5 1.2 0.8 4.5 7.9 (3.4) '.ower Bound 3.1 2.1 1.4 6.6 13.6 (7.0) |2010 Medium 4.2 3.0 2.1 9.3 13.6 (4.3) l Upper Bound 4.9 4.3 2.9 12.1 13.6 (1.5) a/ Full-time equivalent jobs in cropping, at 180 days/year - actual number of persons engaged is much higher due to shared underemployment, and effects of non-crop labour requirements. 7.10 Table 7.2 illustrates, in very rough terms, the impact various developments could have on reducing the number of poor. What is notable is that: (i) no one set of measures stand out; (ii) even achieving these results is dependent on successful new initiatives (eg. in expanding irrigation, intensifying cropping, etc.); (iii) most of these measures will only have the indicated effect if intensive efforts are made to ensure that the poor (as opposed to the non-poor) benefit; and (iv) the potential impact of redistributive measures is not insignificant - however the capacity to effectively target them is questionable.  - ill - Table 7.2: Potential Impact of Various Developments on the Incidence of Poverty (Number of poor households potentially raised above poverty Ino by various events) Increased Toetal Agricultural Increased Opening FormaI Informal Land Income Production Agricultural up Toral Sector Sector Redistri- Redistri- On Own Land / Emaployment /c Lands /_ Employment JI Employment /f bution /g butlon /h Terat Hills Toras Hills 0.15 a 0.21 s. 0. 2S a. 0.18 a. 0.4 0.25 B. 0.6 a. 0.4 a. 0.8 a. (0.23)/k (0.1)/b a/ Currently poor households with enough land (0.5-1.0 ha.) to produce above poverty income from cropping alone. k/ Poor households with enough land to produce a substantial share of poverty level income (say SOX +) from cropping alone, with Improved technologies (15% of hills poor, 20% of rural toral households). Si On holdings larger than 1 ha.; as a result of Irrigation and Intensified cropping; 2 jobs per household. -/ Conversion of 400,000 ha. of remaining forests to farm land - if distributed equally In 1 ha. lots among poor households. 2f Incremental growth to 2010 - assuming 60 of construction jobs and 20S of others absorb the poor; 2 jobs per household. y As with (e); assuming 70% absorb the poor. / Redistribution of holdings abovo 4 ha. (20% of toral land) - if redistributed In equal plots excluoivoly among poor households. ki 16% tax on Incomoe of the top 10 (20X in urban areas) if distributed among those in the bottom 6 deciles. 7.11 There is then no easy poverty alleviation strategy for Nepal. The solution must lie in growth. However growth fueled by the modern sector (eg. cash-cropping, urban-based, industrial) has tended not to reach the poor, except in selected cases. Such a growth strategy thus needs to be focussed: on raising lower-level (eg. subsistence) agricultural productivity, on labour-absorbing measures, and on "enabling' measures - setting the preconditions which would allow the poor to help themselves - by equipping them with education and skills and by providing infrastructure. Particular proposals are discussed in the next chapter. 7.12 It must be emphasized that if the Government acts decisively there is scope for a major impact on the incidence of poverty through public policy. For example. an effective program to curb population growth would reduce the number of poor by at least 5 million by the year 2010. Similarly a combination of measures to raise the agricultural productivity of smallholders, and to increase agricultural employment on larger holdings (mostly through irrigation), coupled with selected rural works employment and redistributive measures (see Table 7.2) could in aggregate produce above-poverty line incomes for over 2.5 million poor households, effectively eliminating most absolute poverty by the year 2010.  - 112 - 7.13 Conversely, failure to act on some of these key areas would likely lead to a catastrophic increase in the incidence of poverty. In particular, if population growth is allowed to continue unabated at its current level, this will likely add another 4-5 million persons living at close to starvation levels by 2010, and a further 10-15 million iii the following twenty years. 7.14 The numbers of the poer could easily increase over the next two decades, and the conditions under which they live could simultaneously worsen. Firstly, the rural land base is saturated in the hills - average plot sizes amonig the poor are alreadv below minimum economic size, and with further intra-family subdivision, many non-poor families will join the ranks of the poor witb.in one generation. In the terai, average holding size is projected to fall by 502 as the result of population growth - perhaps marginalizing 25Z of currently non-poor landholders. Secondly, forces are in place for a deterioration in the distribution of income which usually accompanies monetization and urbanization. We have seen (Chapter 3) that if the distribution of income were to shift to one more typical of developing countries, then at expected levels of GDP and population growth, the bottom 60Z of the population would have incomes substantially below poverty levels. Finally, Nepal will soon witness the emergence of an urban under-class who lack even the weak asset base and social networks which currently support many of the poor. Only imaginative counter-measures can prevent this decline in conditions of the poor, emphasizing the need for the type of 'balanced', labour-absorbing growth referred to above. 7.15 It will take 15 or 20 years for family planning measures to have a major impact on population pressure. In the meantime a large proportion of the poor will continue to cope through migration - to the terai, and increasingly to urban areas. Managing this transition needs to be a key element in country strategy. The rural terai has the potential to support a substantially larger population (about double its current level), but only if: (a) further forest lands are opened up, and (b) steps are taken to raise agricultural productivity, mostly through irrigation. Even tnen saturation will be reached within 20 years, and thereafter all incremental growth (about 650,000 persons per year) will have to migrate to urban areas or to India. The challenge will be to provide basic urban services, especially those used by the poor, while resisting the pressure to provide benefits for the urban non-poor, and avoiding implicit subsidies which will increase the incentive to migrate. One option would be to focus on middle-sized settlements (eg. roadhead bazaars) which can provide growth loci for rural regions. 7.16 Migration has been an adjustment mechanism in the hills for some time. Incomes there have stabilized at just about survival level. Unfortunately, measures to improve incomes in the hills may do little more than slow the pace of out-migration. More broadly, with the open border, it is likely that real wages will equilibriate at about the level of northern India. Income growth for the Indian poor is thus a major factor in raising incomes of Nepal's poor.  - 113 - 7.17 Twenty one of Nepal's 75 districts have no road, and outside Kathmandu only two districts have minimal road densities of 20 km. per 100 sq. km. In many areas providing improved access is the single most important intervention HMG can make, following which market forces alone can provide tremendous increases in both agricultural opportunities and off-farm employment. However, given the very high costs of road building and maintenance, care has to be taken in defining a subset of areas in which improved access is economically justified; those not accessible at reasonable cost in which it may be efficient to provide a minimal package for increasing incomes in-situ (for example, through improved food production); and those from which out-migration should be encouraged (see para. 8.09). 7.18 Developing effective education strategies is doubly important for poverty alleviation in Nepal. Given Nepal's limited resource base, developing a skilled and trained workforce will be central to any long-term development strategy. This is admittedly a long-term proposition, which will take several generations, however a start has to be made now. Fortunately, education also makes an important near-term contribution to the earning prospects of the poor. 7.19 It will take time for population programs and productivity improvements to affect incomes on a significant scale. In the meantime there will remain a very large number of absolute poor (perhaps 5-10 million over the next twenty years), it is legitimate to consider a sustained program of support for them. It is not very meaningful to make the distinction (useful in many countries) between the poor and a sub-class of the *ultra-poor'. It is therefore not particularly useful to target the poor for relief measures - there are too many of them. The emphasis instead has to be on balanced growth through effective general programs - with efforts to ensure that they reach the poor. There are some exceptions - particularly related to food security, and these are discussed in the following chapter. 7.20 Projections show that food balances are likely to get worse, at least over the next ten years, with little chance of there being sufficient purchasing power to import the shortfall. Individual food insecurity is ultimately a function of insufficient incomes. In Nepal, however, due to high transport costs and the absence of marketing networks many people in many areas would not be able to purchase food at any expected income levels in the near future. There are sufficiently large numbers of people in these remote, food deficit areas that efforts designed to address food security per se are a legitimate part of any poverty alleviation program. 7.21 There is also scope for improving the welfare of the poor at existing income levels, by: (i) improving the retention of calories consumed but not utilized because of parasites and other infestations; (ii) reducing the incidence of common, preventable diseases; (iii) changing feeding and hygiene practices; and (iv) reducing time spent in gathering water, fuel and fodder. 7.22 Measures which focus on women in particular will probably improve the equity aspects of growth, because women predominate in poor households, because they are disproportionately not participating in the labour force,  - 114 - and because they tend to be most concerned with the family welfare aspects of income. Given the limited labour market opportunities, measures which concentrate on on-farm income earning opportunities are probably the most fruitfu.l. Women are also the key to successful population control initiatives, and to the behavioral changes (eg. with respect to hygiene and feeding practices) which can improve living conditions at existing income levels. Female education is important for the contribution it can make in all of these areas. 7.23 Finally, we have seen that the poor could fare better under existing economic conditions by freeing up the political and social environment at the village level to allow more self-reliant activity; and by eliminating some of the more exploitative aspects of labour and construction contracting arrangements. Also the current tenancy system provides disincentives both to maximizing output, to equitably sharing the costs of inputs, and to allowing security of tenure.  - 115 - Vill. POLICY CONCLUSIONS 8.1 With its substantial resource limitations and a disadvantageous location, Nepal faces a difficult task in raising Incomes. That is not to say that the situation is hopeless - the foregoing analysis demonstrates that growth balanced across the key sectors could be sufficient to raise many of those who are currently poor out of poverty. It is important to recognize, however, that there is no easy solution to poverty in Nepal. In the previous chapter we have seen that there are a constellation of measures which in combination with one another, and if effectively bnplemented, could in aggregate have a substantial impact in reducing the incidence of poverty. 8.2 The main thrust of a poverty alleviation strategy in Nepal needs to revolve around: (i) a national population campaignl (ii) agricultural intensification - particularly through improved irrigation; (iii) increased rural access in the hills and the terai; and, (iv) education - focussing in particular on the lower levels. Important incremental gains will also come from a supporting program of: - Selected off-farm income-generating activities - particularly a program of rural public works, and small amounts of focused credit; - Some relief measures - especially to improve access to food; and, - Measures which improve the quality of life in the absence of income increases, particularly hygiene and health related ones. In a number of areas which hold promise, additional analysis is needed before proceeding, particularly on strategies for managing population movements and rural access, for improving food supplies in remote areas, and for agrarian reform. The analysis has also shown that in some areas government interventions will not particularly help the poor - including large-scale use of targetted credit, price controls or subsidies, or industrial sector interventions. Income-generating and skills training projects could have some Impact, but would need better design and implementation. The following sections briefly summarize the main policy conclusions by sector; specific recommendations are listed in the Appendix table which follows. S. Sectoral Policies Population 8.3 The need is to implement an effective program of support for temporary methods of contraception - this should be a central national  - 116 - priority. Such a program needs to rely on a large cadre of well-trained and well paid female family planning workers at the village level, to provide continuous support to contraceptive users. Every effort needs to be put into strengthening the institutional and managerial capacity to run such a program; in doing so all necessary financial and managerial resources should be made available without hesitation. At the same time the Government should continue to revitalize and strengthen the established sterilization program. Agriculture 8.4 HKG needs to pursue two parallel tracks - one, for growth, consisting of traditional agricultural programs focusing on those farmers with enough resources to make use of modern inputs - these will yield the greatest returns, but will bring direct benefits to only about 252 of farmers. Therefore the other track needs to focus, in the hills on marginal improvements in rainfed agriculture and systems largely free of purchased inputs; and in the terai on delivering effective irrigation to small farmers (including a package of tenancy reform, credit, and support for shallow tube-wells). Recent initiatives in community forestry and small farmer irrigation are steps in the right direction, as are changes in the research and extension system which should make it more appropriate to small farmers. These initiatives should be closely monitored - and adjusted if necessary - to ensure that they do in fact reach the poor. There is also a need to deregulate input supply, and encourage development of a network of small agricultural traders. Employment and Wages 8.5 Absorbing the rapidly growing numbers of landless in employment represents a major part of the poverty challenge in Nepal. Unfortunately there are few obvious measures open to HMG to achieve this. The greatest impact will probably come from measures to increase labour absorption in agriculture - particularly getting irrigation working in the terai. However, the scope for policy measures to increase off-farm employment is limited by the natural constraints to formal sector growth and the open border with India. HMG's main role should be in providing enabling mechanisms: transport and communications infrastructure, general education, and to a lesser extent skills training. A rural public works program will also have some effect. In addition, gi-an the magnitude of the emerging employment problem, at the margin policies should be followed which maximize labour absorption - through following labour-intensive technologies in public works, providing reasonable incentives to labour- employing enterprises, and avoiding incentives whicL encourage mechanization or capital intensive production. 8.6 The dynamics of the informal sector are not well enough understood to know what measures would help the poor, especially since interventions tried to date seem to have had little effect. Again HEG can probably achieve the greatest impact by providing infrastructure and education. Analytical work should be undertaken to identify whether or not there is a set of more specific measures which could raise informal sector incomes among the poor. On the basis of the preliminary review carried out under this study no obvious candidates emerge.  - 117 - 8.7 There is also little scope for influencing wages, because at the margin they are driven by agricultural productivity, and by the free flow of labour to and from India. The focus instead should be on helping the poor overcome the most exploitative aspects of current labour relations bys selective reform of labour and construction contracting practices, and empowering the poor through group formation, education, and selected credit programs. In the long run the most substantial impact may come from a tightening of the labour market following effective population efforts. Subsidies, Transfers, and Price Policies 8.8 Existing subsidies and transfers do not reach the poor - they should be re-allocated to programs that do (eg. food-for-work), or vhere possible re-targetted. Most of the collection of subsidies intended to compensate for remote access in the hills should be abolished and the funds put into improving access instead. There is little justification for price interventions of any kind, most do not benefit the poor, and anyway they are largely irrelevant given the open border - except in the case of foodstuffs for remote areas, and even those are questionable as they often cannot be effectively targetted. Infrastructure and Spatial Development Issues 8.9 Providing transport infrastructure is one of the most effective interventions HMG can make - however only in some instances. A poverty alleviation strategy for the hills needs to distinguish between areas for which varying degrees of support are justified. The hills should be partitioned conceptually into three categories: (a) areas where road access is economically justified - for which HMG should finance an accelerated program of access; (b) those areas with sufficient potential to justify limited access improvements (eg. suspension bridges and trails, minor roads), and commitment to a continuing program of low level support - perhaps in the form of food distribution, public employment, etc.; in these areas investments in infrastructure should be limited to those with a welfare effect - eg. low-cost water supplies, schools, and health centers; and (c) finally, those areas with no hope of access and where continued expenditure is not justified in the long run - ultimately these areas should depopulate. HMG should undertake only minimal investments in these areas, along with a time-bound program of essential support justified on welfare grounds - but with a timetable for phasing them out.l/ Access to Food 8.10 Improved access to food in remote areas depends mostly on raising incomes and improving physical access, and in some cases outmigration. In many areas this will not happen fast enough to relieve chronic food deficits. In these areas the most cost-effective solution is to produce food in-situ; even this will be a slow process, in the interim limited food distribution is justified. Measures include: (i) encouraging the growth of a private food sales and distribution network; (ii) selected interventions 1/ There will be a fourth category, of isolated pockets with high potential which can support a significant population in comfort; solutions need to be tailored for them on a case-by-case basis.  - .18 - to compensate for remoteness and seasonal variation in food supply: (iii) some feeding of vulnerable groups, if and where they can be cost- effectively targetted; and, (iv) measures to improve nutritional retention at existing consumption levels (eg. through nutritional education and changed feeding practices). 8.11 As a first step HMG and the relevant donors should undertake a food security programming exercise, to include: - agreement on the appropriate level and form of food aid, its financing, and how best to distribute it. - reform of the National Food Corporation's program - eliminating mis-targetted subsidies, identifying which distribution measures reach those suffering food deficits, and how they should be expanded and financed. - identification of a program of effective interventions - including evaluating the relative roles of food distribution, agricultural productivity measures, food storage, vulnerable group feeding, and promotion of effective practices (eg. weaning foods, feeding during pregnancy, etc.); followed by preparation of specific projects and agreement on financing for them. 8.12 Experience elsewhere shows that very targetted nutrition and feeding programs (eg. for malnourished children below the age of 36 months) can have a dramatic effect, although the capacity to administer them in Nepal is weak. No new feeding initiatives should be undertaken until the restructured Vulnerable Group Feeding program has operated for some time and is reviewed. Poverty Alleviation Programs 8.13 The experience with programs to date has been discouraging. Transfer programs do not reach the poor, except for parts of NFC's remote area food distribution. Feeding programs have had only limited success, delivery is sporadic, and they are not well targetted. The food-for-work program is self-targetting, and while it has problems, is an area to consider expansion. The IRDP's show little sign of success in raising incomes of the poor; some intensive agricultural programs do, but their replicability is questionable (i.e. they may be successful because they are in the right places), and unit costs are high. Labour intensive construction and public works employment shows some promise, there are problems to be ironed out, but this is an area for expansion. The targetted credit programs, while partially successful, reach only a small proportion of the poor - their benefits appear to be more in group formation and other externalities than the provision of credit as such. 8.14 HMG should focus its efforts on food-for-work and rural public employment schemes - not because they are necessarily the best solutions (ultimately their sustainability is questionable), but because they are the ones which can reach large enough numbers of the poor to make a difference, and which the administrative capacity exists to deliver. Many small income-generating programs (eg. under NGO's, IRDP's, and targetted credit  - 119 - projects) have had only very limited success. They require labour- intensive interventions, they have proven difficult to deliver, the capacity to expand them is limited by the shortage of skilled staff and by their high unit costs. EMG is probably better off: (i) concentrating its efforts on improving overall administrative and service delivery capacity; (ii) providing a fund for NGO's to continue to undertake small-scale income-generating projects; and (iii) pursuing a slow expansion of SFDP and PCRW - or a revised variant of them. Credit Issues 8.15 HMG should exercise caution in pursuing major infusions of credit to fuel income growth among the poor - our analysis suggests that it will not lift binding constraints on the poor, and that the banking system can probably not efficiently deliver it. Certainly credit should form part of the Government's small business development strategy, but it should not be a centerpiece of poverty alleviation efforts. Providing small amounts of credit to the poor for selected investments is legitimate - for this HMG should focus on expansion of the SFDP and PCRW programs, but only after they have been consolidated and strengthened. Education 8.16 HMG is following largely the right policies in education development - the capacity of the system needs to be expanded if it is to absorb the poor, and measures have to be found which will atcract them into schools. HMG and the donor community need to be ready to allocate substantially more funds to general education (about double in real terms within the next 20 years). Analytical work should be undertaken on the determinants of school attendance, and to frame measures to increase enrollments. The following particular measures are recommended: (i) a national literacy campaign; (ii) revamping the curriculum and examination systems to improve relevance and quality; (iii) measures to increase female participation, including expansion of programs that allow alternative scheduling of classes, more female teachers, and possibly a scholarship program for girls. Health 8.17 The greatest impact will come from non-health measures: water supply and hygiene, improved access to food, and reduced fertility. For the poor HMG should focus on a subset of simple health interventions - immunization, oral rehydration, diarrheal disease management, and micro- nutrient supplementation - which do not rely on improvements in the health service for delivery. 8.18 A national mass hygiene campaign should be undertaken, which would combine nutrition education and awareness with a program of rural water supplies, and provision of sanitation facilities in both urban and rural areas. General 8.19 There has been a startling failure to measure the income effects of massive expenditures on programs and projects designed with the intention of raising incomes (including those financed by the World Bank).  - 120 - In addition, there is no source which allows HMG to track overall changes in incomes or living standards. It is recommended that HMG establish the capacity to monitor household incomes on a periodic basis (the MPHBS is a good model), as well as requiring major projects to include specific income monitoring measures. 8.20 Given the critical importance of employment generatiorn in off-farm activities surprisingly little is known about labour force and wage issues in Nepal. The Department of Labour focuses exclusively on the manufacturing sector (22 of employment). HMG should establish a data base and analytical capability covering employment statistics, wages and employment conditions. Complementary to tbis, a group of studies should be undertaken to improve the (almost non-existent) understanding of employment and incomes in the informal sector and in construction. C. Programming Implications 8.21 To promnte equitable growth the Government needs to strengthen and expand key sectoral programs - including the national population campaign, modified approaches in agriculture, an expanded program of access, and the spread of basic education. In most of these areas steps are already underway - they need to be intensified, and implemented effectively. Major programming recommendations in these sectors are beyond the scope of this report, although additional suggestions are made in the text as to how they could have a greater impact on the reduction of poverty. In addition, to directly improve the conditions of the poor, the following six point action program is recommended as a framework around which HMG and donors could mobilize support. (1) Expansion of Selected Transfer Programs - probably consisting of a rural works program and Food-for-Work. (2) Incremental Improvements in Existing Income-Generating Projects - including the SFDP and PCRW, and a fund for NGO income-generating activities. (3) Food and Food Aid Programs - including an agreed program of support, reform of the NFC, and identification of the most effective interventions. (4) Strengthening Non-Income Alleviation Measures - which improve the welfare of the poor in the absence of income increases: including a mass hygiene campaign, simple preventive health interventions, and possibly targetted nutrition/feeding interventions. (5) A Package of Policy Reforms - involving tenancy, labour contracting arrangements, service delivery and decentralization, and the regulatory framework for NGOs.  - 121 - (6) Analytical Work in Priority Areas - including: (i) labour force issues, (ii) measuring income and income changes; (iii) development of migration and re-settlement strategies; (iv) tenancy and agrarian reform; (v) the construction industry; (vi) the informal sector; (vii) determinants of education participation; (viii) the private economics of family size; and (ix) the cost-effectiveness of simple health interventions. 8.22 In some of these areas further study is needed before program reforms can be finalized. The first steps in framing this program would be: (a) preparation of sub-projects - including agreeing on modifications required to existing programs; appraisal of the new projects (eg. for food security and rural works); and agreement on the shape of the sectoral programs (population, agriculture, and rural access) - to the extent that they differ from, or represent an intensification of, on-going initiatives; (b) in parallel, undertaking necessary analytical work (eg. for agrarian reform, food security measures); and, (c) agreeing on a financing package for the main elements of the program.  - 122 - nAUL OF R 0 RNC SIOC  AGUtCULTURE - Continue to pursue small fazmer Irrigation programs in the hills and Paras. 4.29, 6.63 the terai; with careful design to Improve benefits to the poor; consider using criteria under Irrigation Sector Program to select poor beneficiaries, where feasible. For groundwater Irrigation developrment In the terais form m1ll Pare. 4.29 farmers into groups; regularize land tenure and registration; provide a package of credit and technical support. Encourage development of a small traders to supply agricultural Pare. 4.7, 4.30 inputs - possibly by providing training and credit. Continue to pursue research and extention (RU) based on a Paras. 4.11, 4.31 farming systems approach, Including: (i) expansion of RIB based on systems free of purchased inputs; (ii) conduct field tests using poor farmers as models; (III) more research on subsistence livestock: and, (iv) consider a small-scale horticulture program for farmers In Para. 4.19 accessible areas. Train agricultural extension workers to deal with female farmers Para. 4.19 DE?WIYaNIT - Consider a major rural employment program, based around expansion Para. 6.69 of Food-for-Work, and SPm. In combination with above, consider a program of more labour-intensive Para. 4.44 and Annex 1.5 construction works - as a first step do an analysis of costs and efficiency trade-offs. Increase monitoring and supervision of construction contract-rs. Para. 3.70 and Annesx 1.5 Enforce District wages. Para. 3.69  SECTOR RECOIHENDATIONS RRnuRu - Review construction contracting and employment practices, including Pars. 4.42. Annex 1.5 scope for use of smaller contractors. - Increase labour-intensive maintenance, especially on rural roads. Annex 1.5. Para. 4.43 - Rather than beneficiary operations and maintenance (eg. of irrigation Para. 6.63 schemes), consider a user charge combined with wage employment of poorer villagers. LAND TENURB - De-link tenancy reform from land redistribution issues. Para. 3.42 - Design and implement a tenancy and agrarian reform package. Para. 3.43 As a first step, analyze the extent and type of tenancies, and their Psra. 3.43 economic effects. - Streamline procedures for land registration and transfer. Annex 1.1 and Para. 3.42 - Divorce issues of urban land management from those of agrarian reform. Para. 3.41 and Annex 1.1 FOOD SECURITY - Separate food distribution for civil servants from NFC's other Para. 6.45 activities. Preferably discontinue distribution to civil servants and replace Paras. 6.21, 6.45 with salary increases. Drop Kathmandu entirely from NFC's operations. Para. 6.45 Review and determine scope for benefit targetting in remote areas Para. 6.45 under NFC. Undertake no new feeding programs at this stage. Pars. 6.47 Review experience with the Vulnerable Group Feeding program next Para. 6.47 year, along with experience elsewhere. NMG and donors should undertake a general assessment of options Para. 8.11 for food security interventions, ar;d agree on a program.  SECTOR ucoiomuMTzous izinzncR POPULATION - Mount a major national population campaign, it should includet Para. 5.52 Recruitment of a large number of female family planning workers (maybe 25,000). highly paid and well-trained. Population linked incentives in other sectors. Possibly a separate service delivery system, outside the health service. Develop a cadre of population specialists and managers. Par. 5.52 - Undertake analytical work on the income effects and determinants of Para. 5.54 of family size. EDUCATION - Concentrate on basic education and literacy Para. 5.43 - Be prepared to increase financing for primary education from Para. 5.44 US$24 million to US$50 million p.a. - Undertake a national literacy campaign. Para. 5.45 - Improve curriculum relevance, reduce rote-learning and dependence Para. 5.46 on examinations. - Consider automatic promotion in the first three years, and possibly Para. 5.46 instruction in non-Nepali languages. - Consider ti) siting schools closer to families; (ii) more flexible Para. 5.46 hours, (iii) recruiting teachers locally. - Increase the use of female teachers (postpone SLC requirement), Para. 5.47 consider providing female scholarships and uniforms. BHALTH - Focus on immunization, oral rehydration. micro-nutrient programs. Para. 5.60  SECTOR IUCCIUKRDA?IONS _E__R__ - Consider a mass national hygiene campaign - combining hygiene Paras. 6.48, 5.59 education, and changing hygiene behavior, with rural water supplies and sanitation facilities. - In conjunction with the above - a nutrition behavior program to Para. 6.48 change weaning and feeding practices. CREDIT - Consolidate SFPD and PCRV before further expansion. Para. 6.54 - Strengthen the training of group organizers. Para. 6.54 and Annex 11.4 - Divorce development functions from comercial banking activities. Para. 6.57 - Consider establishment of a *Poverty ank* to finance small maounts Para. 6.57 of directed credit for income increasing activities among the poor. - Reconsider massive expansions of commercial credit to the poor. Paras. 6.60. 8.15 SPATIAL ISUPS - Undertake an analysis of measures needed to mange population Para. 3.22 transition out of the hills - including scope for conversion of teral forests. - Consider concentrating on growth of small-to-medium sized Para. 7.15 settlements. - Adopt a hierarchy of support for hills areas, depending on degree Pares. 8.09, 7.15 of accessibility and sustainability of welfare support and economic activity. QCJEAL - Consider a land tax - except on small landowners. Para. 6.05 - Strengthen general programs, and provide Infrastructure, rather Para. 6.32 than finance additional IRDP's. - If there are to be further IlDP's, they should focus on very few Para. 6.32 activities - probably roads and Irrigation.  8DCn cii umc_i 8nforce collection of Lrrigation charges. Par. 6.19 - Replace irrigation capital subsidies wLth a credilt program, with an Para. 6.20 exemption for very mall landholders. Undertake analytical work to identify poseLble micro interventions Para. 8.06 to accelerate informal sector income growth. _ Undertake a systematLe review of adminLstration and personnel factors Para. 6 77 effectLng service-dellvery, rather than tackling then sector-by- sector. Pursue programs whh put more ncome In the hands of women (eg. PCKV). Pa. 3.91 - Brpmad program wbLch reduce the tim spent (wostly by waeen) in gatbering fuel fodder, and water - eg. rural water upply, coinmnity forestry, trail improvemnt. * istablish a data-base on employment d wagps Issues. fPr. r .20, Annex 1.5 - Improve monitoring of incomes, income cbas. and the lnce_ Pars. 8.19 effects of projects and programs. * Consider partnerships of Gos executing soe projects financed Pars. .84 by large donor. - Considor a fund for small income-generating projects to be Pars. 6 n7 imple_nted by NOO's, along with tecbhncal support. - evLew the regulatory framework for 0GO's, with a view to Pers. 6.86 relaxing lt. - BuLld up the capacity of local 0GO'9, consider a mall fund through Para. 6.86 which SRGO's and others could support them.  I L 'k 8 r z IBRD 10154R2 31'. rr * .N E P A L SAk S'A!,/ > AAP / < -"'m>% t w 2 t ;