Mozambique - Food security study

Report No. 7963-MOZ Mozambique Food Security Study October 12, 1989 Agriculture Division Southern Africa Department FOR OFFICIAL USE ONLY Document of the Worl Ban& 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 Worhl Bank authorization.  CURRENCY EQUIVALENTS Meticais/US$ 1975 22.55 1980 32.40 1981 35.35 1982 37.77 1983 40.18 1984 42.44 1985 43.18 1986 40.43 1987 (average) 286.00** 1988 (average) 526.00** 1989 (average Jan-June) 710.00** SOURCE: People's Republic of Mozambique - Recent Economic Developments, IMF, November 1987 **Government of Mozambique  - ,, sAL t - PREFACE This report is part of a special initiative by the World Bank Africa Region, in conjunction with the Food and Agriculture Organization and the World Food Program, to develop food security strategies for a small number of countries during 1989. It presents the findings of a mission which visited Mozambique between February 8-28, 1989. The mission was comprised of Jane Armitage (mission leader and principal author, World Bank), Neeta Sirur (economist and co-author, World Bank), Francois Sizaret (nutritionist, FAO), Tadj Zekrya (Desk Officer for Mozambique, WFP), Barry Riley (economist, USAID), Sandra Laumark (CARE), Roger Hay (food policy consultant) and James Levinson (nutrition consultant). The Government nominated a ccanterpart team with participants from the Ministries of Commerce, Agriculture, Health and Finance to work with the mission. The working team was comprised of Maria Nita Dengo, Iolanda Fortes, Agy Mamunune (FSD), Admir Bay (MOA), Eduardo Franca, Leonardo Lopes (MOF), Kerry Sylvester (KOH), and Juan Del Canto (FAO Resident Food Security Advisor). Food security is a complex subject, cutting across many different sectors, and at times difficult to distinguish from poverty alleviation or even economic development itself. In Mozambique, the complexity of the issues is particularly great, because, in addition Lo the problems arising from the acute scarcity of both financial and human resources, the implementation of effective strategies with long term objectives is severely limited by the poor security situation. As a result, the present report is lengthy and broad in scope, and at this stage does not attempt to develop in detail the costs and benefits and implementation plans of the various recommended policy and programmatic interventions. As such, the report represents the first step in an ongoing process of developing with the Government a comprehensive but realistic food security strategy for Mozambique. Discussions of the report with the Government during September 1989 focused on identifying which interventions are considered to be the highest priority and feasible given the war and limited institutional capacity. Based on these discussions, a short paper laying out a focused strategy and action plan to address food security issues is being prepared. It is anticipated that this strategy and an accompanying matrix of policies and actions for the next year to eighteen months will be presented to the Consultative Group in November 1989. It is expected that the strategy paper and action plan would be updated annually. The intention is that the strategy paper and action plan will assist in identifying components which could be included in future Bank lending to improve food security and, in addition, provide a framework for bilateral and multilateral donors and NGOs within which current and future external financial and technical assistance can be evaluated and where necessary strengthened or redesigned to enhance the impact on food security. 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.  ABSTRACT Food security problems are enormous in Mozambique today - affecting about half of all households in some degree and severely affecting at least 602 of these in both rural and urban areas. Levels of acute and chronic malnutrition are exceedingly high (at the upper end of the range seen in Sub-Saharan Africa) and demonstrate the extremely precarious nutritional situation of Hozambican families, suggesting that even small further shocks could precipitate mass starvation. The factors contributing to food insecurity and nutritional stress are various, including the continuing warfare which has meant widespread displacement of rural households and disruption of trading and communications networks, climatic disasters, past failures in economic policy and the inevitable shocks from needed adjustments, and poor health and sanitary conditions. This report examines the effects of the above factors on national food supply and distribution, on the ability of poor households to produce or purchase food, and on the physiological utilization by household members of nutrients contained in the food. In addition, bearing in mind the constraints imposed by poor military security and weak institutional capacity, the report analyzes a number of options for improving food security through greater efficiency in utilization of available food and other resources and through better targeting to vulnerable groups. The basic strategy proposed is to concentrate resources on increasing productivity of smallholders and on creating sustainable employment and income-generating opportun1t-ies for the most vulnerable houssholds rather than on distribution of free or subsidized food on a large scale. However, these efforts need to be complemented by improvements in access to basic public services such as health, education, clean water and sanitation, and by a limited 'safety net" of direct and indirect income transfers as necessary to prevent serious malnutrition or starvation. The main issues examined and principal recommendations are outlined below. One major topic analyzed in the report is the resource transfer efficiency of food aid, which, by augmenting national food supplies, is an essential component of food security in the current Mozambican situation of disrupted domestic production and foreign exchange scarcity. The report argues that food aid constitutes a resource of considerable magnitude in Mozambique, so that its use should be subject to the same efficiency standards that are applied to the use of financial resources. Several recommendations are made to increase the efficiency with which food aid (and revenue from food aid sales) is provided and used, inter alia: (a) developing a single framework for food aid valuation, management of revenue from food sales and setting of priorities for use of revenue; (b) reviewing the financial and management position of public food trading enterprises to reduce discrepancies in the value of food aid sales and actual rece-4ts by the Ministry of Finance; (c) better coordinating food aid arrivals; and (d) where feasible, increasing monetization of food aid once inside the country, as money is cheaper to move and can be used more flexibly. Another emphasis rZ the report is analysis of options for increasing productivity in agriculture, where, despite the uncertain security situation, performance could be strengthened. Taking into account recent Government action to restore price incentives for agricultural production, the report's focus is on next steps required to encourage recovery in the sector. The chief argument is that public investment  should concentrate on the family (rather than the state) aector, from the standpoints of both enh&ncing sectoral productivity and increasing housaiold food security. This will involve improving provision of basic agricultural servicea (research, credit, training and extension), improvements to the marketing system and development of specific interventions for areas with high drought risk or land scarcity. Initial attention should be concentrated in the 40 districts selected by Government as having high agricultural potential and/or strategic safe location. Although increasing productivity of family farmers should help improve food security for many households, the potential for widespread rural development remains limited until the security situation is resolved -- thus necessitating complementary measures to increase non-agricultural employment and income-generation in rural as well as urban areas. Therefore, the report explores ways to increase employment through a combination of: (a) assistance to labor-intensive small businesses in the private sector; (b) training programs targeted to unemployed members of the poorest households; and (c) public employment programs focused on developing/maintaining basic infrastructure. It is recommended that jobs so-created be targeted explicitly to members of vulnerable households, using self-targeting devices such as paying less than minimum wage or promoting micro-enterprises and informal sector activities which tend to have a greater proportion of low-wage workers. Taking into account the likely difficulties and time involved in putting in place large-scale agricultural development/employment programs and the fact that there exist labor-scarce households (e.g., with disabled household heads) unable to benefit from emJloyment opportunities, the report also explores options for establishing a temporary safety net for households currently at risk of food insecurity. The recommended safety net comprises three complementary parts, each designed to address the needs of particilar target groups: the first involves provision of indirect income trensfers through subsidies and reduced turnover taxes on so-called 'inferior' commodities (e.g., sorghum or yellow maize) consumed mainly by the poor; the second involves adjustments to the urban rationing system to improve coverage of the urban population; and the third consists of direct income transfers to the roughly 10-15X of households identified by local leaders to be most vulnerable on the basis of pre-established criteria. By helping to maintain human capital in a difficult petiod, the safety net would not only help families in need but also be a vital contributor to long'-r-term socio-economic development and growth. Existing widespread problems of food security/malnutrition in Mozambique are primarily related to absolute food shortages and inadequate purchasing power on the part of consumers. However, problems of inadequate food availability are severely exacerbated by the frequent occurrence of infectious and parasitic diseases which contr:oute to physical debilitation and malabsorption of nutrients in food taken. The report analyzes public health conditions and recommends increased investments in rehabilitation of water and sanitation systems and healt:. infrastructure and services. In addition, it suggests measures to improve control of specific diseases, in particular diarrhea, measles, malaria and intestinal parasites, which have severe nutritional consequences. A common thread in the proposed measures is the need to increase community participation or utilize alternative channels/organizations outside the formal government health system.  TABLE OF CONTENTS PaRe No. Executive Summary ................ i I. OVERVIEW OF FOOD SECURITY IN MOZAMB E .. 1 A. Main Themes of the Report. 1 B. Economic Context. 4 C. National Food Balance. 7 D. Food Distribution System .14 E. A Strategy for Household Food Security .25 II. HOUSEHOLD FOOD SECURITY AND NUTRITION . .................... 29 A. Nutritional Situation ................................. 29 B. Food Availability, Incomes and Household Food Security . C. Morbidity and Malnutrition/Food Security. . 48 III. FOOD AID ............................................. 53 A. Introduction .......................................... 53 B. Triangular Food Aid ................................... 56 C. Food Aid and 'Counterpart Funds" . . 63 IV. THE ROLE OF THE AGRICULTURAL SECTOR . .73 A. Background ..................................... ...... 73 B. Food Pricing Policy .. 75 C. Agricultural Development Strategy .. 83 D. Assistance to Populations Affected by Drought and War .. 87 V. EMPLOYMENT AND INCOME GENERATION .. 93 A. Urban Areas .. 93 B. Peri-Urban Areas ..101 C. Rural Areas ....................... 106 VI. THE "SAPETY NET" ...... 1lll A. Introduction .111 3. Commodity-Specific Subsidies and Taxes ................ 114 C. Improvements to the Urban Rationing System (NSA) ...... 123 D. Direct Income Transfers ............................... 129 E. Budgetary Impact of Proposals ........................ 137 VII. HEALTH AND MALNUTRITION/FOOD INSECURITY ...... . 143 A. Health and Sanitation Services ............ 143 B. Proposed Health and Nutrition Interventions ........... 146  VIII SUMMARY OF RECOMMENDATIONS ......................... ....... 153 A. Overview of Strategy .3................................. 13 B. Specific Actions to Improve Food Security ............. 154 ANNEXES 2A Food Price and Minimum Wage Data 2B Official and Parallel Market Prices for Food Commodities 6A Pilot Testing of the Targeting System for Vulnerable Households in Urban Areas 7A Health System Utilization 1980-1987 7B Prospects for Developing Health Care MAP 1 Adjusted Agroclimatic Zones MAP 2 IBRD #18436 Tables in Text 1.1 Marketed Production of Food Crops, 1980-1988 ......... 7 1.2 Per Capita Production in Family Sector ............... 8 1.3 USDA Estimates of Production of Major Cereals in Mozambique .. 8 1.4 Population Dependent on Food Distribution (Commercial, Food Bank and DPCCN) .................. 11 1.5 Focd Aid Requirements Included in 1989/90 Appeal ..... 12 1.6 Response to the April 1988 Appeal for Food Aid for May 1988-April 1989 ............................ 13 1.7 DPCCN Commodities Managed by Type: 1984-1988 ......... 17 1.8 DPCCN Fleet Composition: 1984-1988 ................... 18 1.9 DPCCN Storage Capacity: 1984-1988 .................... 18 1.10 DPCCN Staff: 1984-1988 .. 18 1.11 Assessment of Truck Fleet Capacity, 1988 ............. 19 2.1 Incidence of Chronic and Acute Malnutrition: Data from Spot Surveys 1985-87 ..................... 30 2.2 Prevalence of Grovwh Falterings 1987-88 Data by Province ... ..................................... 31 2.3 Prevalence of Growth Faltering: 1988 National Data by Month .. o ..*... 32 2.4 Estimated Food Security Status of the Rural Population .......................................... 36 2.5 Household Food Requirements and Estimated Costs ...... 42 2.6 Household Expenditures: Maputo City 1988 ............. 46 2.7 Characterization of Vulnerable Households in Maputo City ....................... 47 2.8 Nutritional Status of Children < 3 Yrs by Living Environment.....* ................................ 50 3.1 Food Aid 1988-89: Pledges by Donor and Commodity ..... 54 3.2 Commodity Allocations by Use 1988-89 ................. 55 3.3 Triangular Food Aid Imports 1988-89 by Commodity ..... 56 3.4 Timeliness of Food Aid: Bilateral vs. Triangular ..... 58 3.5 Triangular Food Aid Transactions: Value of White Maize Compared with Ye:ow in 1988-89 .............. 60  3.6 Triangular Food Aid Transactions: Mozambique 1988-89 Resource Transfer Efficiency Measures ...... 61 3.7 Regulations Governing Counterpart Funds .. . 65 3.8 Value to Mozambique of Food Aid for Market Support 1988 ....................................... 66 3.9 Counterpart Fund Valuation and Receipts by Ministry of Finance 1988 ........................... 67 3.10 Planned Utilization of Food Aid Sales Revenue in 1988 ............................................... 70 4.1 World Commodity Prices and Price Projections ......... 78 4.2 Illustrative Model to Estimate Producer Prices Based on International Prices ...................... 79 4.3 Producer Prices for Basic Foods at Selected Exchange Rates .. 80 4.4 Illustrative Model to Estimate Consumer Food Prices Based on International Price ................ 81 4.5 Consumer Prices for Basic Foods at Selected Exchange Rates ..................................... 82 5.1 Structure of Production in Maputo Green Zones, 1989 ........ 103 6.1 Safety-Net: objectives and Structure ................ 112 6.2 Supply Implications of Proposals for Maputo NSA ...... 125 6.3 Private Versus Public Fleet Indicators ............... 136 6.4 Government Revenue ................................... 137 6.5 Turnover Tax on Some Basic Foods, 1988 ............... 133 6.6 Current Expenditure .............. 139 6.7 Consumer Subsidies by Commodity, 1988 .... 139 6.8 Estimated Value of Wheat Subsidies, 1988 ............. 140 Figures in Text 1.1 The Food Security/Nutrition Cycle .................... 1 1.2 Cereal Supply, 1955-1988 ............................. 10 2.1 Conditions of Accessibility February 1989 ............ 37 2.2 Manor Food Staples: Price-Wage Indices ............... 41 2.3 Reported Food Consumption by Per Ca:ita Income Group: Maputo 1988 ................................. 44 6.1 Food Consumed by Income Group ........................ 116 6.2 Engel's Curves Relating Per Capita Monthly Consumption of a Commodity to Per Capita Monthly Income ............................................. 117 6.3 Logarithmic Figures of Per Capita Monthly Consumption of a Commodity to Per Capita Monthly Income .....18...... l  GLOSSARY OF ABBREVIATIONS AGRICOM - State Enterprise for Agricultural Marketing BPD - Banco Popular de Desenvolvimento (People's Development Bank) CCA - Caixa de Credito Agrario (Fund for Agricultural Credit) CFW - Cash for Work CIF = Cost, Insurance and Freight COE - Comite Operacional para a Emergencia (Operational Committee for the Emergencies) COGROPA = State Enterprise for Provincial Food Marketing DNDR - Direccao Nacional de Desenvolvimento Rural (National Directorate for Rural Development) DPCCN = Departamento de Prevencao e Combate das Calamidades Naturais (Department for the Prevention and Control of Natural Calamities) DPCS = Direccao Provincial do Comercio de Sofala (Provincial Directorate of Commerce of Sofala) EACB = Empresa de Abastecimento a Cidade de Beira (Supply Company for the City of Beira) EACM - Empresa de Abastecimento a Cidade de Maputo (Supply Company for the City of Maputo) EEC - European Economic Community ERP = Economic Rehabilitation Program FAO = Food and Agriculture Organization FAS = Free AlongSide FFL = Fundo Fomento Agrario (Fund to Promote Agriculture) FEW = Food For Work FRELIHO = Front for Liberation of Mozambique FSD - Food Security Department GDP = Gross Domestic Product GOAM = Gabinete de Organizacao de Atastecimento a Cidade de Maputo (Office Managing Food ;.'upplies to Maputo City) GOM = Government of Mozembique GZV - Gabinete das Zonas Verdes (Greer. Zones Office) IDA - International Development Association IDIL = Instituto de Desevolvimento da Industria Local (Institute for Development of Local Industry) IEC = Information, Education, and Communication IMBEC = State Enterprise for Import of Consumer Goods INIA = Instituto Nacional de Investigacao Agraria (National Agricultural Research Institute) INPF = Instituto Nacional de Planeamento Fisico (National Institute for Physical Planning) MIC = Ministry of International Cooperation MIS = Management Information System MNR - Mozambique National Resistance MOA = Ministry of Agriculture MOC - Ministry of Commerce MOH = Ministry of Health MOL - Ministry of Labor NGO - Non-governmental Organization NSA - Novo Sistema de Abast.cimento (New Supply System) ODA - Official Development Assistance  SADCC - Southern Africa Development Coordination Conference SDA . Social Dimensions of Adjustment SIDA - Swedish International Developmen. Agency TFAT - Triangular Food Aid Transactions UNDP - Unit( Nations Development Programme UNICEF - United Nations Children's Fund USAID - United States Agency for International Development US . United States USDA . United States Department of Agriculture WFP - World Food Programme  Executive Summary The Food Security Problem in Mozambique i Food security may be defined as access by all people at all times to svwficient food for a healthy and productive life. In Mozambique, problems of food security exist both because of national food supply shortages and distribution problems and because many poor households in rural and urban areas do not have the ability to produce or purchase enough food. In addition, nutritional deficiencies resulting from inadequate food intake at the household level are exacerbated by the common occurrence of many infectious and parasitic diseases which are associated with malnutrition. Overall, food security problems are enormous in Mozambique today - affecting about half of all households in some degree and severely affecting at least 60Z of these in both rural and urban areas. Levels of acute malnutrition are at the upper end of the range seen in sub-Saharan Africa, and, when placed in the context of extremely high levels of infant and child mortality - about one-third of all children die before the age of five - begin to assume alarming proportions. Moreover the exceedingly high levels of chronic malnutrition (between 40-60 Z according to many spot surveys) demonstrate the extremely precarious nutritional situation of Mozambican families, suggesting that even small further shocks could precipitate mass starvation. The factors contributing to food security problems and nutritional stress are various, including the continuing war- fare and banditry which has led to widespread displacement of rural households and to the disruption of trading and communications networks, climatic disasters such as droughts and floods, past failures in economic policy and the inevitable shocks associated with necessary adjustments, and generally poor health conditions throughout the population. Overview of Strategy to Enhance Food SecuritX ii There are a number of constraints which mvt-t be taken into account in the development of a realistic strategy for food security in Mozambique Overall, althougL the total resources available are scarce in comparison to needs, the implementation of effective strategies with long term objectives is probably limited most by poor security and weak institutional capacity. First and foremost, Mozambique is a country at war. The difficulty of access to many areas creat a a fundamental dilemma between allocating resources according to needs or according to feasibility. A second set of constraints is imposed by the weak technical and administrative capacity of Mozambican institutions. A third set of constraints relate to the need to ensure that Government expenditure moves in line with its revenue. An implication of the strategy proposed in the report is increased public spending to address food security problems - financed largely with external resources including food aid. It is therefore essential that public expenditure on food security be directed most towards investment in future economic growth an" least towards transfers which create long-term dependence on unfinanceable social programs. The basic approach proposed in the report is to concentrate resources on increasing productivity of smallholders and on creating  - ii - sustainable employment and income generating opportunities for the most vulnerable households, rather than on "handouts, of free or subsidized food on a large scale. However, these efforts need to be complemented by improvements in access to basic public services such as health, nutrition, education, clean water and sanitation, and by a limited safety net of direct and indirect income transfers where necessary tc prevent serious malnutrition or starvation. There is no question that the root causes of food insecurity in Mozambique cannot be fully addressed until peace is restored. However, a major theme throughout the report is that a number of improvements in food security can be made in the interim through greater efficiency in utilization of available food and other resources and through better targeting to vulnerable groups. Food Supply and Distribution iii As a result of the sharp falls in domestic food production in recent years, together with acute shortages of foreign exchange, Mozambique has become increasingly dependent on external food aid to feed much of its population. In 1988, food aid cereal imports were in excess of half a million tons. This represented almost 90! of the total supply of cereals marketed commercially in urban/peri-urban areas or distributed free throughout the countryside. Guvernment/UNDP estimates of food aid requirements for 1989/89 - based on an analysis of national food supplies relative to aggregate needs - exceed 900,000 tons. Although this seems an extraordinarily large request, Mozambique's recent appeals to the donor comaunity under the so-called 'emergency program" appear to have been so successful that the availability of food in the country as a whole is now limited more by internal logistical constraints than by overall supply. Food distribution in Mozambique is particularly difficult due to a large number of transportation and logistical problems, including, inter alia, the poor state of the roads, shortages of trucks and small vessels for coastal shipping, inadequate cargo handling equipment at many small ports, delays in access to fuel and spares, and inadequate repair and maintenance facilities and technicians for both trucks and shipping, compounded by the lack of security and insufficient availability of military vehicles to escort convoys. Ccnsequently, the role of the Government in distributing food staples to consumers has grown significantly in the latter part of the decade as normal trading networks have been seriously disrupted by the war and food shortages have become critical. The report describes in some detail the various food distribution systems operating in Mozambique currently, specifically: (a) the "normal" commercial trading system; (b) the system of free food distribution under the emergency program and (c) the urban rationing system operating in Maputo and Beira, and offers a number of suggestions to increase their effectiveness and efficiency. A potential solution to problems of transportation may lie in greater privatisation of trucking operations particularly in relatively secure areas. This could help reduce costs (evidence suggest that the operating costs of private sector vehicles are around one-half of the operat$ng costs of public sector vehicles) and increase accountability and help build a private transport sector in Mozambique.  - iii - Identification of the Food Insecure iv The report draws on a variety of sources to build up a picture of the incidence and severity of malnutrition in Mozambique today, and concludes that malnutrition is a serious and widespread problem, particularly among the high risk groups (infants, young children and pregnant women). It also attempts to identify the causes of household food insecurity and the characteristics of the families most vulnerable to malnutrition. Overall, it appears that malnutrition is primarily related to absolute food shortages in some areas and to inadequate purchasing power on the part of consumers, rather than to unsatisfactory dietary patterns or inadequate feeding practices for infants and weaning-age children. In rural areas, food security problems stem principally from the war and climatic disasters which have displaced millions of rural households or otherwise reduced their capacity for self-provisioning. Many rural households have also suffered declines in their incomes due to the fall in employment opportunities in South Africa and related remittances, the impact of drought and banditry on cattle holdings, and the decline in sales of cash crops such as cashew, copra and cotton due to the virtual collapse of the marketing system. The food security status of these households depends on the ability of DPCCN (the Government department responsible for distribution of relief supplies under the emergency program) to reach them with supplies of food on a regular basis - a task immensely complicated by the extraordinarily difficult logistics conditions prevailing in the country. A few areas of Mozambique are totally inaccessible, while other extensive areas are accessible only with heavily armed escort or by extremely expensive air-lift. In 1988, total DPCCN food distribution represented only about 50S of estimated needs, implying that food shortages and food insecurity has been widespread among displaced and affected groups in rural areas. v In urban areas, problems of absolute food shortages have been largely ameliorated in the two main cities of Maputo and Beira, although temporary scarcity of marketed food supplies is still common in many smaller cities/towns. In general, issues of household financial access seem to play the major role in urban food insecurity, given recent changes in consumer prices of food vis a vis wage policies and employment conditions. Analyc:s presented in the report suggests that poor urban households have experienced sharp declines in their real incomes as a result of the substantial increases in food prices over the last two years. The report also analyzes the food consumption of different household income groups, utilizing the household expenditure survey carried out in Maputo in 1988. The findings suggest that up to one third of urban households are currently facing serious food deficits relative to needs. Therefore an attempt is made to identify these vulnerable households as precisely as possible with a view to targeting special assistance to them. A number ef easily measurable household characteristics are identified which are highly correlated with household per capita income/nutritional status of children, including total household size, number of displaced or unemployed persons in the household, education of the household head, type of dwelling, water and sanitation facilities and so on.  - iv - Food Aid vi Food aid is fulfilling two important food security functions in Mozambique. It is substituting for market flows of domestically produced food and commercial imports in the face of foreign exchange scarcities and disrupted domestic food markets. It is therefore preventing price rises which would otherwise imperil the food security of poor households who depend on markets for food. It also supplements Government transfers to households - by counterpart fund contributions to Government expenditure from the sale of food aid and in the form of free food distributed through the emergency program. Food aid is an additional resource of considerable magnitude in Mozambique and therefore its use must figure prominently in overall food security strategies. However, the report argues that, ill many cases, the same objectives would be achieved more efficiently by using money instead of commodities. ALthough food commodity aid is necessary to food security at the national level (because it augments market supply), once inside the country it has few advantages over money. Money can be moved about more cheaply (a major advantage given transport difficulties in Hozambique) and can be used by recipients more flexibly. This does not discount the value of food aid in Mozambique. On the contrary, strategies which raise the level of food demand effectively will make additional demands on market supply which at the moment can only be met with food aid. vii Triantular Food Aid Transactions. The report looks at ways to improve the resource transfer efficiency of food aid, with particular focus on an evaluation of the effectiveness of triangular food aid transactions MTFATs) which currently represent almost 30Z of Mozambique's food aid. Possible advantages of TFATs include increased efficiency in resource transfers as a result of cheaper commodities and reduced transport costs, improved timeliness in food aid deliveries, increased access to commodities which are preferred by consumers, and strengthe.ting of regional markets. The counter view is that TFAT commodities and transport are no cheaper (and sometimes considerably more expensive) than they would be if supplied from donors' stocks or procured in international markets, that logisticai problems impose undue administrative costs, that sources and quality are unreliable a.nd deliveries are prone to delays, and that donor trading in oligopolistic regional markets allow parastatal traders to bid up their export prices. The report attempts to analyze some of these issues with respect to the triangular procurement of white maize for Mozambique. Unfortunately, due to problems with the data, the conclusions from the analysis shoald be regarded as indicative rather than definitive; however, the analysis serves to illustrate methods of assessing the relative value of food aid which might be useful for food policy makers in Mozambique. In general, it appoars that both yellow maize from developed countries and white maize procured from Argentina were delivered more promptly than white maize procured from other developing countrie' in the African region. African white maize also appears to be somewhat more expensive than maize from other sources, although there are very significant differences in the white maize prices negotiated by different donors. However, evidence from informal market surveys suggests that white maize is strongly preferred by Mozambican consumers and carries a premium of between 33-50Z over yellow maize in the parallel market. The report concludes that white maize has a resource transfer advantage over yellow maize as long as its price  - v - differential in local markets exceeds any premium donors need to pay in external markets to procure it. viii Management of Counterpart Funds. While part of food aid is distributed free under the emergency program by DPCCN, almost 602 is sold throughout the country, with the bulk sold through the rationing system operating in Maputo and Beira. The revenue from the sale of food aid is potentially an important resource for financing government programs which are consistent with both growth and a sustained reduction in food insecurity at the household level - including support for government recurrent expenditures in the social sectors, programs of employment creation, direct income transfers targeted to vulnerable households, food subsidies and so on. However, there have been a number of difficulties concerning counterpart funds which have reduced their contribution to economic development and social welfare. One problem is that donor regulations - concerning the method of valuing the food aid, the payment and accounting procedures to be followed, and the uses to which the funds can be put - have been developed piece-meal as a result of annual bilateral donor-Government negotiations. The multiplicity of donor rules and expectations imposes a considerable administrative cost on the government and complicates accountability. In addition, particularly before 1988, much of the funds were earmarked by donors for special purposes that were not necessarily in accordance with the priorities of the government and did not always represent the most productive use of scarce resources. A final problem concerns the major discrepancy between the value of food aid sales and actual receipts by the Ministry of Finance. In 1988, actual transfers to the Ministry represented only 43Z of the revenue available from sale of food aid. It appears that late payments by state wholesalers and millers are the major source of the problem, suggesting that food aid imports are generating a significant, albeit hidden, subsidy for millers and traders. ix The report recommends that donors and Government agree on a single overall framework for the management of counterpart funds, including standard food aid valuation methods and priorities for counterpart fund utilization. This would have a number of advantages. First, it would provide the basis for a rapid conclusion to existing Government-donor negotiations and for releasing accumulated counterpart funds against a set of expenditure criteria, and would obviate the need to repeat negotiations every year. Second, it would simplify the current administration of food aid counterpart funds, with its multiplicity of regulations and accounts, thereby liberating scarce skilled manpower in the Ministry of Finance. Third, it could provide the basis for a joint Government-donor procedure for monitoring payment of counterpart revenues, based on transparent accounting principles which would discriminate clearly between stock financing requirements, losses, cost of specific transport or commodity subsidies and trading or processing inefficiencies. Finally, it would permit the allocation of food aid counterpart funds to finance food security strategies to be made in accordance with overall priorities for public expenditure, as determined jointly by the Government and donors, rather than being earmarked by individual donors for special purposes.  - vi - Agricultural Sector Strategy x Agricultural Policy Reform. In the short run, large volumes of food aid will continue to be necessary to feed much of the population. However, in the medium term, given peace and a supportive policy and institutional environment, in most years Mozambique's agricultural sector could produce sufficient food for rural consumption and substantial marketed surplus to feed the urban population, supply raw materials to local processing industries and make a major contribution to export earnings. Under the ERP, the Government has embarked on a series of policy and institutional reforms at the macro and sectoral level to bring about the :ecovery of economic production and increase the efficiency with which resou).ces are allocated in the economy. The initial focus of agricultural reforms under the ERP has been on pricing policy, with a view to restoring the incentives for agricultural production and marketing. At present, given continued civil war in the countryside, the supply response from the agricultural sector is likely to be limited. However, the intention of the pricing and other reforms is to lay the foundation for rapid economic recovery once the security situation is resolved. Nevertheless, in the short run, it is likely that real income and access to food of many urban households will be adversely affected by the increases in food prices. Therefore, the report suggests some possible modifications to the food price system which could cushion the impact of further -rice increases on the nutritional status of the poorest while respecting the basic principles of pricing policy and the limitation on Government expenditure and subsidies under the ERP. xi An additional area for reform concerns the allocation of public expenditure within the agricultural sector. In the past, public expenditure in agriculture has been concentrated on large scale, capital- intensive investments in the state sector, while the family sector has been neglected. However, estimates of total agricultural production show that, compared to state or private producers, family producers have an overwhelming weight in production and generally have much lower unit costs of production than state farms. In the future, therefore, it is recommended that the main focus of Government strategy and resource allocation be on increasing production and yields in the family sector. Promotion of the family sector makes sense not only on efficiency grounds but also on equity grounds. As the bulk of the population derives their livelihood from small family farms, increasing their productivity should contribute directly to improvements in food security at the household level as well as increasing overall domestic food supply. Furthermore, over 60Z of those working in the smallholder sector are women, reflecting the legacy of the colonial period where the men were forced to work on settler plantations and estates or left to work in neighboring countries. Supporting the growth and development of smallholder agriculture wi_l therefore directly benefit rural women and thence should lead to improvements in the nutritional and health status of their children. xii Agricultural Services. While restoring the incentives for agricultural production and marketing through pricing reform is a necessary condition for recovery of the sector it is by no means sufficient. Adequate provision of inputs to producers is essential to enable them to  - vii - respond to the increased incentives. More generally, a well functioning marketing system needs, inter alia: (a) an adequate transportation system, (including roads, vehicles, repair and maintenance facilities, spare parts and gasoline); (b) storage facilities; (c) credit facilities (both for farmers to purchLae inputs and for traders to finance vehicles and the purchase and storage of crops for sale later in the year); and (d) a supply of consumer goods for farmers to buy. In addition, in order to increase yields, it will be necessary for the Government to improve the provision of basic agricultural services such as extension, credit and research. With respect to extension, given the critical financial and institutional constraints facing Mozambique, which are exacerbated by the insecurity, the report recommends that efforts be focused on developing a cost-effective and appropriate methodology for delivering extension services through a number of pilot operat:.ons in strategic safe areas with high agricultural potential. It also recommends that extensionists focus on delivering a simple message to farmers that requires minimal transportation of inputs, foreign exchange or credit. For example, researchers in the Ministry of Agriculture estimate that yields in the family sector could be doubled just through utilization of improved ciltivation practices, such as more thorough weeding and land preparation, planting at the right time and with the appropriate densities, intercropping of maize with cassava or pulses, incorporation of organic materials and so on. xiii In the medium term, increasing yields substantially in the family sector will require the development of a cost-effective credit system to assist smallholders to adopt improved technical packages that rely more heavily on purchased inputs. Furthermore, it is essential that, in the future, agricultural research in Mozambique devote much more attention to the needs of the family sector. The report recommends that activities focus on highly applied research oriented to the development of high yielding seed varieties which are disease-resistant and adapted to low input technologies, and linked to a program of seed multiplication. Specific attention should be given to the development of packages suitable for areas with high drought risk and for areas with acute land scarcity, for example around emergency camps or in the "Green Zones' around urban centers. In general, greater emphasis should be given to programs of adaptive trials and on-farm tests and demonstrations based on existing findings particularly from neighboring African countries. xiv Assistance to Drought Affected Households. In drought-prone areas (principally the areas south of the Save River, i.e., Inhambane, Gaza, and Maputo Provinces) food insecurity will continue to be a problem even under peace and a perfect policy regime. Therefore, the report outlines a number of specific interventions to assist the population practicing rain-fed agriculture in these areas including: (a) replanting tree cash crops such as cashew, copra and mafurra; (b) promotion of greater crop diversification and inter-cropping, particularly by growing more drought resistant crops such as cassava, sweet potato, sorghum, millet, cowpeas and lab-lab beans; (c) incorporation of small species such as goats, rabbits, fish and bees into the farming system; (d) dissemination of low cost measures to provide dry-season cattle feed; (e) reduction of pre- and post-harvest losses and improved food conservation techniques in the home; and (f) development of small-scale low cost irrigation facilities.  - viii - Employment and Income Generation xv Employment Creation. In the medium term, the strategy of raising agricultural producer prices and increasing the productivity of smallholder agriculture should contribute directly to improvements in food security for the majority of Mozambican households. Until the security situation is resolved, however, the possibilities for widespread rural development are likely to be rather limited, and many rural households are unable to produce more than a fraction of their food requirements as a result of the insecurity in the countryside. In addition, poor urban families suffer declines in real income when food prices rise, with potential serious implications for food consumption and nutritional status. Therefore the report explores ways to increase employment of poor families in urban and rural areas, through a combination of: (a) assistance to labor-intensive micro-enterprises in the private sector; (b) training and "skills liberation " programs targeted to unemployed members of the poorest households; and (c) public employment programs focused on development and maintenance of basic economic and social infrastructure. While any programs in Mozambique which result in increasing the number of jobs are likely to have some beneficial effects on the poorest segments of the population, where possible the objective is to create jobs explicitly for those fron. the most vulnerable households. However, the programs should be designed to be self-targeting, as otherwise the need to identify those persons who would qualify and exclude those who do not is likely to be administratively and financially burdensome. Paying less than the minimum wage is one unfortunate but probably necessary self-targeting device to help ensure that only workers from the very poorest households would be candidates for the jobs created. Another way is to direct resources, Information and training to those enterprises most likely to increase employment of the most impoverished. xvi Given the objectives of restraining public expenditure under the ERP in order to restore fiscal balance, together with the extreme scarcity of public managerial, technical and administrative resources, the report recommends that the emphasis in employment creation should, to the greatest possible extent, be on the private sector. Promotion of micro-enterprises and informal sector activities appear to offer a particularly promising avenue. Although total employment in the industrial sector fell by 58? between 1973 and 1987, employment in the micro-enterprise sector tripled. Since micro-enterprises tend to have a greater proportion of low-wage workers than larger, more formal enterprises, there is a double justification in attempting to promote further growth in the sector. First, it is likely to benefit the most vulnerable groups through self selection via low wages, and second, it has shown economic dynamism in a time of general industrial decline. In addition, micro-enterprise expansion tends to create more jobs at less cost than the growth of large scale industries which are generally more capital intensive. The report proposes a number of ways to encourage growth of employment in labor- intensive private enterprises, including provision of assistance and training in business planning, management, budgeting and accounting, in conjunction with small lines of credit and start up grants.  - ix - xvii However, the report notes that generating viable jobs in the private sector is likely to take time. Employment-generating public works programs will therefore be required to undergird more general income and employment strategies. It is often suggested that such schemes tend to have a high opportunity cost resulting from the diversion of labor from other productive activities and high administrative costs, while the output has little economic value. However, at present, in Mozambique, the opportunity cost of labor is likely to be low due to the high levels of unemployment resulting from massive population displacement caused by the war. In addition, given the widespread destruction and deterioration of basic economic and social infrastructure in rural and urban areas, it should be possible to identify labor intensive projects, such as rehabilitation and construction of roads, schools, health facilities, marketing infrastructure, small scale irrigation facilities, land improvement schemes and so on, that have high economic returns and are essential for long term economic recovery. The critical bottleneck to developing economically sound public works programs is the shortage of skilled manpower to plan, implement and monitor such programs. Therefore, the report recommends that, where possible, existing administrative structures are utilized, and that mechanisms are developed to reduce the involvement of the Central Government by relying on local community resources, NGOs and the private sector. It is further suggested that all donors engaged in rehabilitation, reconstruction, social services and related programs review the designs rf their interventions to determine whether minor modifications might allow for additional employment creation and whether there is scope to target employment opportunities to the most vulnerable groups. xviii Food for Work (FFW) versus Cash for Work (CFW). There is a strong belief in the Government that food aid should be used increasingly in support of productive activities. 'Food for work" public works programs have been used in many countries to try to increase the developmental impact of food aid. Their experience suggest that middle level management requirements are generally high, as are supervisor-to-worker ratios. These skills are in extremely short supply in Mozambique. Furthermore, the food commodities must be stored, transported, weighed, packed, monitored and distributed equitably against work completed. In contrast, money can be moved around, stored and distributed much more cheaply than food. In many situations, therefore, it will make more sense to sell food aid through existing market channels and use the proceeds to finance measures to improve food security, including employment-generating programs. In particular, in urban areas where poor households lack access to food primarily because of insufficient income rather than the physical unavailability of food in the marketplace, CFW would appear to be more appropriate since wage payments are considerably cheaper and easier to administer than payments in food. Of course, CFW employment programs in urban areas will lead to increased demand for food, which, in the short run, will have to be met largely through increased food aid if the increased demand is not merely to drive up food prices. xix In remote rural areas where war or drought has led to severe food shortages and normal food marketing channels have been disrupted, there is a danger that CFW employment programs could have an inflationary  impact on food prices. In this situation, FFW may be appropriate. Again, however, the important point is that the increased demand for food resulting from the increase in employment be backed by an increase in the supply of food. In general, where a trading system is functioning, the preferred approach would be for the Government to pay workers in cash and to allocate more food aid supplies to the area, to be sold through normal market channels. This avoids the need to set up an expensive parallel food distribution system which undermines existing market channels. The increased cash incomes may also scimulate domestic food production where the security situation permits a supply response. Food aid that is "monetized" in this way helps strengthen domestic marketing and production of food rather than substituting for it. However, where the trading system has been totally destroyed by the war, direct food payments to workers may be necessary. In these cases, the report notes that experience from other countries suggests that close coordination between Government, NGOs, and local communities/participants is absolutely essential to success. xx Green Zones. In peri-urban areas, many families can produce at least part of the food they consume on small plots of land in the so-called "Green Zones" - the belt of unutilized land around the cities. However, in Maputo, the contribution of the Green Zones to urban food sucurity is limited by the shortage of land and the poor quality of the soils. In addition, there is a high risk to rainfed agriculture due to the irregular rainfall, while only a small part of the available land is suitable for irrigation. While heavy doses of fertilizer could provide a way to increase the productivity of the land, widespread use of fertilizer by family farmers is impossible in the short run, because families are totally unfamiliar with its use and the facilities for extension and credit to promote it do not exist. Furthermore, many peri-urban families could not afford to buy fertilizer for subsistence food production even if credit were available as they would have insufficient income to pay off the loan. There is a real danger therefore, that intensifying cultivation in the Maputo Green Zones, without adequate replenishment of soil nutrients, vill lead to exhaustion of the land. Therefore, the report argues that resettlement of parts of the peri-urban population in more fertile areas outside Maputo (in close coordination with the military to ensure safety), must also be part of the strategy to improve peri-urban food security in Mozambique. xxi In other cities, land availability may be somewhat more favorable than in Maputo, although even in relatively small cities the increasing insecurity in the countryside has led many rural families to cluster atvund urban centers, putting pressure on the land available for food production. Despite the overall constraints on land availability, however, there is significant potential to increase productivity in the Green Zones, through: (a) improved cultivation techniques; (b) greater use of purchased inputs such as better quality seed, improved high yielding varieties and pesticides; (c) development of small-scale, low-cost manual irrigation systems; and (d) concentration on crops with high nutritional value such as beans, cow peas and lentils, or high financial value such as fruits and vegetables. Given the existing low level of agricultural practices in the Green Zones, the report recommends that donor assistance be sought to establish more Casas Agrarias to enable more intensive  - xi - interaction of extension agents with farmers. It is further recommended that, in the short run, the extensionists focus on teaching low-cost, basic cultivation techniques such as planting seeds at the right time, greater attention to weeding and adequate land preparation, inter-cropping cereals and legumes, sowing in straight lines and in appropriate densities, and incorporation of organic materials to enhance soil fertility. A combination of mixed cropping and better cultivation practices could double yields at virtually no additional cost, and without the need for credit. Over time, additional increases in productivity can be achieved by use of improved seeds and low doses of pesticides and fertilizers, and low cost irrigation where feasible. The Safety Net. xsii Justification. The main focus of the food security strategy outlined in the report is to develop viable approaches to creating a sustained increase in demand for food at the household level - principally by increasing the productivity of smallholders and encouraging growth in both private and public sector employment. However, in all likelihood, many households will, in the short- to medium-term, remain beyond the scope of strategies designed to stimulate growth due to a variety of interacting factors. These include: (a) th3 continuing security crisis, its accompanying large-scale displacement of rural households and disruption of the trading and communications networks essential for the development of durable productive activities and trade; (b) the inevitable adjustment period required for large numbers of the work force to learn the new skills required to take advantage of employment opportunities/incentives created within the ERP framework; and (c) the limited institutional capacities in Mozambique for rapid adoption of new measures or for development of substantial new employment- or income-generating enterprises. Thus it is likely that, in the current period of economic transition, some households with low or uncertain incomes would require special assistance on a temporary basis to allow for adequate access for foo4. In addition, there is a small proportion of labor-scarce households whose effective demand for food can not be substantially increased through improved opportunities for employment or through enhanced labor productivity. For these households, there will always be a need for some type of social security system to ensure their food security. Therefore, a number of "safety net" measures are proposed in the report to ensure that vulnerable families have adequate access to food. This safety net is made up of three complementary parts, each designed to address the needs of particular population groups: the first consists of the provision of indirect income transfers through subsidies on so-called "inferior" commodities, such as sorghum or yellow maize, which are consumed predominantly by poorer households; the second involves adjustments and/or extensions to the existing rationing system to improve coverage of the urban population and the third consists of direct income transfers targeted to those most in need in rural and urban areas. By helping to maintain human capital during a difficult economic period, the safety net would not only assist families in need but also be a vital contributor to longer-term socio-economic development and growth.  - xii - xxiii Commodity Specific Subsidies. General food subsidies were part of Government policy to protect the market-deperdent poor for many years. However, they impose a heavy fiscal burden which is likely to be incompatible with the objectives for the current account set out in the PFP. Furthermore, general food subsidies are inefficient because the benefits are distributed across the entire population, including well off households, rather than directed to the target group of the poorest households. Although most food subsidies were eliminated when the prices of major food commodities were raised several fold during 1988, wheat flour and bread continue to be highly subsidized. The report recommends that these subsidies be removed and that a subsidy be placed on a food commodity which is less preferred by the rich but is an imDortant staple for the poor. Utilizing information from the 1988 Maputo household survey, the report analyzed the food consumption behavior of households from different income groups. Sorghum and whole yellow maize appear to be the most promising candidates for a self-targeting food subsidy which would indirectly transfer income to the poor. Therefore the report suggests that market trials with subsidized sorghum and yellow maize be conducted on a pilot basis to test consumer response. xxiv Urban Ration System. One of the Government's main instruments for promoting equitable access by urban consumers to scarce food supplies is the urban rationing system or NSA. By making about half of food requirements available to consumers at official prices, which though now in line with border parity prices are still well below parallel market prices which incorporate a significant scarcity premium, the NSA remains an important element of a food security safety net in Maputo and Beira. The report makes a number of suggestions to improve the ration system including: (a) increasing the ration allowance for basic cereals, ensuring the availability of the full ration of oil and reintroducing beans into the NSA; (b) promulgating widely new regulations concerning greater freedom of choice for consumers under the ration system; (c) issuing new ration cards which would permit the poorest households to purchase their monthly ration in smaller quantities; (d) streamlining the procedures for registration of newly arrived households; and (e) assessing the costs of extending the system to other cities such as Nampula and Inhambane. xxv Direct Transfers to Vulnerable Households. Finally, the safety net proposed in the report includes schemes in both urban and rural areas designed to target income transfers directly to the poorest households. A major difficulty in achieving direct income transfers to vulnerable groups lies in developing a selection process that is both relatively cheap to administer and reasonably accurate in identifying the intended beneficiaries. A number of possible selection mechanisms were considered and rejected including: (a) targeting by geographic location in urban areas; (b) use of employment registration information to identify households with low incomes; and (c) use of growth faltering data from health posts to identify food insecure families. The mechanism that appears most likely to be abie to identify the most vulnerable families in urban areas is the neighborhood-based system of community organization which presently certifies households for inclusion into the NSA system in urban areas and identifies households for relief food supplies in rural areas. Selection of families for the dixect income transfer would be based  - xiii - on a series of objectively and quickly assessable criteria such as female or disabled head of household, absence of formal employment, high dependency ratio, low education of housebold head, type of dwelling, water source, etc.. The report recommends that a pilot operation be undertaken to pretest the selection criteria and the efficacy of the community organization in applying them. xxvi In rural areas, the distribution of free food by DPCCN is the principal means by which a safe.ty net of direct transfers is extended to the most vulnerable families. Although it is recognized that the war places absolute constraints on the scope of DPCCN's operations, the report explores a numbers of wdys to improve the cost-effectiveness of relief food distribution and to utilize food aid in ways that strengthen and stimulate rural markets rather than substitute for them. It is recommended that DPCCN and AGRICOM (the agricultural marketing parastatal) explore possibilities for sub-contracting transportation of emergency food to private haulers. The report also proposes that substantially greater attention be paid to the possibilities for food 'swap" arrangements between DPCCN and AGRICOM, particularly in areas where logistics are difficult and food is availeble in local markets. By 'swapping' part of its food imports with AGRICOM for locally procured food, DPCCN could distribute the food thus obtained to needy families near the point of procurement, while DPCCN's imported food could be sold by AGRICOM in urban markets according to its usual practice. This could reduce transportation costs considerably, as well as improving the timeliness of relief food distribution, and facilitating consumption of locally produced food by poor households who lack the means to purchase it. In addition, in areas where food surpluaes are available, it is recommended that DPCCN consider selling some of its food to AGRICOM or private traders for money which could be given to households normally eligible for relief food so that they could buy food for themselves. The advantages of a scheme such as this is that it would help stimulate local food production and marketing rather than substituting for it. It would therefore link the provision of a safety-net with growth objectives and also allow for a smoother transition from emergency to normal sources of food supply as conditions improve. Health snd Nutrition Services xxvii Relationship of Health and Nutritional Status. To the extent that household food security is not an end in itself but rather a means to achieving an active and healthy life, it also requires satisfactory social practices and health conditions - to ensure efficient and equitable allocation of available food among household members, and optimal physiological utilization of the nutrients contained in the food by individual household members. The "normal" Mozambican diet is well- balanced from the standpoint of nutritional requirements and, in the absence of real resource constraints, families make generally rational decisions regarding intra-household food allocation and the feeding and care of young children. Existing widespread problems of malnutrition in Mozambique are primarily related to absolute food shortages and inadequate purchasing power on the part of consumers. However, nutritional problems resulting from inadequate intake of food at the household level are  - xiv - exacerbated by the common occurrence of many infectious and parasitic diseases. Infection or disease appears to affect nutritional status in three ways, depending on the illness: (a) by inducing anorexia in an infected person, although food requirements may actually increase during illness; (b) by contributing to malabsorption of the nutrients i4n food taken; and (c) by causing physical debilitation over a relatively long period and thus preventing adult sufferers from working to provide food. In addition, malnutrition is itself associated with susceptibility to illness, implying an interactive relationship between diriase and nutritional status. xxviii Health Interventions. Given this strong association between malnutrition and ill-health, a critical element in the strategy to achieve food security in the long term is an improvement in household living conditions - housing, sanitary facilities and clean water, and in access by the population to medical services and preventative health interventions. In general, however, sustained improvement in these areas can only be achieved in the medium to longer term, given both the problems created by the continuing security, and constraints in the availability of the physical, financial and skilled human resources required to implement the large-scale reconstruction efforts necessary. Therefore, from the standpoint of food security, there is an immediate need to identify or strengthen ways to control specific diseases, in particular, diarrhea, measles, malaria and intestinal parasites, which have severe nutritional consequ)nces. The report proposes a number of measures to strengthen or supplement immunization programs, and programs for control/treatment of diarrh al disease, malaria, and intestinal parasites. In addition, it recowmends an increase in resources devoted to information, education and communication (IEC) activities, with particular focus on promoting child spacing and maintaining satisfactory young child feeding practices. Given the acute shortage of skilled personnel within the Ministry of Health, and the destruction of health facilities in many rural areas, a common thread throughout the recommendations is the need to utilize alternative organizations and channels outside the IftLAstry (such as local community leaders, schools, FRELIMO mass organizatiot'iz such as the women's organization, DPCCA depots, AGRICOM outlets and private retail establishments) in implementing the proposed measures.  "ZAMBIQUE FOOD SECURITY STUDY I. OVERVIEW OF FOOD SECURITY IN MOZAMBIQUE A. Main Themes of the Report 1.01 Definition of the Problem. As discussed in the recent Report of the Task Force on Food Security in Africa (World Bank, 1988), food security is best defined as access by all people at all times to sufficiert food for a healthy and productive life. Within the framework of this definition, attainment of food security requires fulfilling certain conditions for the supply, demand and household-level utilization of food. At the national level, an adequate aggregate supply of food (either from domestic production or imports) is a necessary condition for food security - however, it is not sufficient. The "access' part of the definition of food security is also critical. Food security requires that all households have the ability to acquire enough food, either by producing it themselves or by generating sufficient income to purchase needed food supplies. 1/ Moreover, to the extent that household food security is not an end in itself but rather a means to achieving an active and healthy life, it also requires satisfactory health conditions and social practices - to ensure efficient allocation of available food among household members and optimal physiological utilization by individual ho'lsehold members of nutrients contained in the food. These factors, in turn, have an impact on nutritional status and, hence, on the capacity of household members to work and so provide food. Figure 1.1: The Food Security/Nutrition Cycle and Fioa Watso, Marc 1988.AV |PRODUCTION ||AT THE HOUSEHOLD| | WORlt g i FOOD SNT~~~~~~~AKEl | APACITY ||BY IN;DIVIDUAL| V PfYSSOLOGICAL ||NUTRS | Uh'CTION UT2LIZATION| SOURCE: Interaction of Nutrition and Infectiont A Review, Andrew Tomkins and Fiona Watson, March 1988. 1/ There are examples where countries heve had large stocks of food at the national level, or have even exported food surpluses, while at the same time many sections of the population suffered from malnutrition because they were not able to produce elough food to cover their needs and were too poor to buy it.  1.02 The simplified model of the food security/nutrition cycle, illustrated in diagrammatic form in Figure 1.1, encompasses a number of factors which affect household food security a- nutritional status. Inter alia, these include: the economic environment and employment opportunities; food imports and the ra-.ge of variables (e.g., land, climate, prices, input availability etc.) which influence domestic agricultural production; the physical fitness and health of the population; household characteristics including age distribution and dependency ratios; and socio-cultural beliefs surrounding diets and feeding practices. In Mozambique, as elsewhere, an attempt must be made to examine the effects of these factors on households in order to assess the magnitude of the food insecurity/malnutrition problem and to develop effective strategies to address it. This is done in the following Chapters. 1.03 Food insecurity is a problem of gargantuan proportions in Mozambique today, as evidenced by prevailing high levels of malnutrition (documentei in Chapter II). In large part, this is due both to overall supply shortages and distribution problems at the national level as well as the fact that many households are unable to produce or purchase sufficient food to meet their nutritional requirements. Domestic food production declined sharply during the 1980s, due mainly to a series of climatic disasters and the worsening security situation which has led to the displacement of large sections of the rural population. Marketed production of basic food crops also fell substantially, from about 135,000 tons in 1980 to around 50,000 tons in 1986. At the same time, there has been a massive increase in the numbers of people who are no longer self- sufficient in food because of the disruptions caused by the war. In 1989, approximately 8.4 million people (around 55Z of the population) were dependent on purchases through commercial channels or on free food distribution for part or all of their nutritional requirements. Given the acute scarcity of foreign exchange which precludes any commercial imports of food, the country has become increasingly dependent on food aid. In 1988/89 over 600,000 tons of food aid were imported, about 40Z of which was distributed free to emergency camps, hospitals, health and social centers and education institutions, and the remainder sold. Although the volume of food aid supplied by the donor community in response to Mozambique's emergency appeal has been reasonably satisfactory, the country currently faces severe internal distribution problems, due largely to difficulties of access caused by the war, compounded by the inadequate supply of trucks, spare parts and fuel. As a result, in 1988/89 only about 65Z of initial food stocks plus food aid receipts - representing about 50Z of needs - were actually distributed, leading to critical shortages of food vnd even starvation in the more remote and inaccessible parts of the country. 1.04 In addition to the problems caused by insufficient supply of food in some areas, it appears that many households have suffered from food insecurity due to inadequate purchasing power to acquire sufficient food. This is a problem particularly for low income urban households who depend on the market for their nutritional requirements and who have experienced sharp declines in their real incomes as a result of the substantial increases in food prices over the last two years. In addition, many rural households have been affected by the war and drought and are consequently no longer fully self-sufficient in fooa production. Other rural households  - 3 - have also suffered declines in their incomes due to the fall in employment opportunities in South Africa and related remittances, the impact of drought and banditry on cattle holdings, and the decline in sales of cash crops such as cashew, cotton and copra due to the virtual collapse of the marketing system. Finally, nutritional problems resulting from inadequate intake of food at the household level are exacerbated by the common occurrence of many infectious and parasitic diseases, in particular, diarrhea, measles, malaria, respiratory diseases and intestinal parasites, which are associated with malnutrition. The net result of all these factors is that malnutrition is a serious and widespread problem in Mozambique. 1.05 Outline of the Revort. One of the main objectives of the report is to identify the food insecure in Mozambique with sufficient detail and accuracy with respect to: (i) the numbers, location and characteristics of vulnerable groups; (ii) the specific nature of nutritional problems (both protein-calorie-malnutrition and micronutrient deficiencies); and (iii) the principal causes and severity of these problems. The report also aims to develop a country strategy for food security in Mozambique, including an outline of broad policies and programs to assist the food insecure. In the short run, given the constraints imposed by the war, much of the focus will be on improving the management and utilization of food aid. Food aid is increasing the supply of food in Mozambique, thereby reducing price rises which would imperil the food security of those households who are not self- sufficient in food. In addition, however, household food security requires that the supply of food be matched by sustained demand for food by all Zamilies. Free distribution of food aid to poor families achieves this directly, but does not contribute to the long-range sustainab.e development of Mozambique (except by helping to prevent destruction of human capital). The challenge facing the Government is to expand the role of food and other aid beyond the temporary relief of hunger to become a source of economic growth which raises the incomes of the most vulnerable in a self-sustaining way. The basic approach proposed is to concentrate efforts as much as possible on restoring domestic agricultural production particularly within the family sector and on creating employment and income generating opportunities in both urban and rural areas, rather than on 'handouts' of free or subsidized food on a large scale. These efforts need to be complemented by improvements in access to basic public services such as health, nutrition, education, clean water and sanitation, and by a targeted safety net of direct and indirect income transfers where necessary to prevent serious malnutrition or even starvation. 1.06 Chapter I provides some brief background information on the overall economy and the Economic Rehabilitation Program (ERP), followed by a detailed analysis of the national food balance and the food distribution systems operating in Mozambique at present. The final section proposes some broad strategies for improving the use of food and other aid to address the underlying causes rather than the symptoms of household food insecurity. Chapter II draws on a variety of sources to build up a picture of the incidence and severity of malnutrition in Mozambique today. It also attempts to identify the causes of household food insecurity and the characteristics of the families most vulnerable to malnutrition. Chapter III looks at ways to improve the management and utilization of food aid,  - 4 - including an analysis of the pros and cons of triangular food aid transactions (which currently represent almost 302 of Mozambique's food aid) and an evaluation of the system for accounting and management of the counterpart funds generated from sales of food aid. Chapter IV discusses the objectives, constraints and strategy for development of the agricultural sector. Around 80S of the total work .ng population in Mozambique is in the smallholder or family sector, and over 60? of those working in the sector are women. Unit costs of production in the family sector are generally well below those on state farms. The Government's ctrategy to focus resources on increasing productivity of the family sector therefore makes sense on both efficiency and equity grounds and should enable Mozambique to meet both the overall supply and the household access conditions for food security in the medium term. 1.07 Until the security situation is resolved, however, the possibilities for widespread rural development are likely to be rather limited. Therefore, Chapter V explores ways to increase employment and income of poor families in urban and rural areas, through a combination of assistance to labor intensive micro-enterprises in the private sector (e.g., credit and small business management training); training and "skills liberation" programs targeted at able-bodied unemployed members of poor households; as well as public employment programs focused on development and maintenance of basic economic and social infrastructure. However, in order to prevent serious malnutrition and starvation, a limited asafety- net, may also be required for those who are beyond the reach of employment strategies. In Chapter VI, various Isafety-net' measures to assist poor households are discussed, including ways to improve the existing urban food ration system, possibilities for subsidizing a less preferred food commodity consumed primarily by the poor, increases in the minimum wage, and direct income ti.snsfers to targeted sections of the population identified as especially vulnerable. Chapter VII analyses the rale of basic public services such as health, nutrition, clean water and sanitation in bringing about improvements in food security. Finally, Chapter VIII summarizes briefly the main recommendations made in the preceding chapters. A major theme throughout the report is first, the importance of targeting both growth ard welfare iuterventions to the most vulnerable households, and second. in a situation of acute scarcity of financial and human resources, the need to find cost-effective and administratively simple measures to assist the poor. B. Economic Context 1.08 Poit Independence Economic Performance and Constraints. Despite Mozambique's well-endowed and diversified resource base, since Independence there has been a steady decline in real Gross Domestic Product (GDP) per capita. In 1988, GDP per capita was estimated at around US$100-120, making Mozambique one of the poorest countries in the world. Between 1980 and 1986, overall production fell by about 301 and exports by nearly 75?, while imports were compressed by more than one third from their 1980 level by 1985, before recovering somewhat in 1986. External imbalances, aggravated by an increasingly overvalued exchange rate, led to the rapid accumulation of external debt service arrears. Internally, due to large fiscal deficits  - 5 - and unconstrained bank financing of enterprise losses, the money stock tripled despite shrinking output. Administrative controls on allocation and prices, although intended to promote efficiency and social welfare, introduced rigidities and disincentives that proved inimical to the efficient allocation of resources. As shortages increased, barter, inflation and parallel markets for goods and foreign exchange became widespread. 1.09 The difficulties faced overall by the Mozambique economy have been felt with particular force in the agricultural sector. Agriculture is the most important sector in the Mozambican economy. accounting for about 40-50_ of GDP, employing over 80? of the labor force and generating around 80O of the country's export earnings. Land is plentiful relative to population - only 20-30? of the cultivable area is currently farmed - and soil is reasonably fertile and rainfall generally adequate, although there is a drought risk in the South. Despite the considerable potential of the sector, however, in recent years its performance has deteriorated alarmingly. Between 1980 and 1986, the current value of agricultural exports declined by around 60?, and marketed production of basic food crops fell by about 65?. Subsistence production in the family sector is also believed to have declined significantly. Rural insecurity has led to over- concentration of population and cattle in safe areas, resulting in environmental degradation due to deforestation, soil erosion from overgrazing and excessive land clearing, and declining soil fertility as the traditional practice of shifting cultivation is curtailed. 1.10 Many of the difficulties in the agricultural sector are the result of external factors, including a series of droughts and floods, and the worsening security situation which has caused severe disruption in the rural economy and widespread population displacement. In addition, years of colonial rule left Mozambique with an almost total lack of basic physical economic and social infrastructure, one of the lowest literacy rates in the world, and an acute shortage of skilled managerial and technical manpower which severely constrains the ability of the Government to undertake even the most basic policy analyses or planning activities. Inappropriate economic policies were also factors behind the poor performance of the sector, in particular, the rigid pricing policy and the focus of Government expenditure on large capital-intensive investments in the centrally-planned state sector. Little attention has been paid to the development of institutions and infrastructure to deliver essential agricultural services such as research, extension and credit to family and private producers. In addition, basic marketing activities such as crop purchase, and input and consumer goods supply have been adversely affected by a number of factors including attacks on vehicles by the bandits, the poor state of the roads, the scarcity of vehicles, spares and fuel, and inadequate vehicle repair and maintenance facilities. 1.11 The Economic Rehabilitation Program. In response to the deteriorating economic situation, the Government of Mozambique embarked on a comprehensive ERP in January 1987. Economic reforms are focused on several key macro-economic policy issues, including the exchange rate, foreign exchange allocation system, trade policies, pricing policy, the budget and credit. In addition, the reforms address policy and  - 6 - institutional issues in the key sectors. In agriculture, the main focus of reforms under the ERP so far has been on pricing and marketing policy, with a view to restoring the incentives for agricultural production and marketing. Because food prices play such an important role in food security, the rationale for the pricing policy reforms under the ERP and their implications for prices of the main food staples are analyzed in detail later in the report. Since 1987, prices of all agricultural commodities, including food crops, have been increased several-fold. Around 0B of households in Mozambique consist of small farmers who, given peace and a supportive policy and institutional environment, could in most years produce enough food for their own requirements and some surplus for sale. increases in food prices should therefore lead to direct improvement in incomes and food security for most rural households. 2/ In the medium term, urban employment and incomes should also gradually improve as increased agricultural growth and incomes stimulate agro-industries and industrial production of investment and consumer goods. In this context, an underlying assumption throughout this report is that the reform measures taken by the Government under the ERP, including exchange rate and domestic price adjustments, and cuts in subsidies to restore fiscal balance, are fundamentally sound and, indeed, essential for the resumption of long-range economic growth and development. 1.12 On the other hand, the substantial increases in official food prices that have occurred and will continue to occur under the ERP lead to serious concerns about the short-term food security and nutritional status of low-income consumers who need to purchase their food. This latter group includes the bulk of urban households but will also include agricultural households who are not fully self-sufficient in food because of drought, shortages of labor, inadequate access to basic inputs or other difficulties caused by the security situation. Indeed, available evidence suggests that the real income and access to food of many urban households have already been adversely affected by the sharp increase in food prices. Therefore, this report places considerable emphasis on: (a) assessing the impact of recent food price increases on poor urban households; and (b) identifying some possible modifications to the food price system which might cushion the impact of further food price increases on the poor while respecting the basic principles of pricing policy and the limitations on Government expenditure and subsidies tlnder the ERP. 2/ However, those sections of the rural population who have been displaced or othervise affected by the war and are unable to produce enough food for their own subsistence needs (much less to respond to the incentives for marketing created by the ERP) will continue to depend on public assistance, including free food, to meet their nutritional requirements.  - 7 - C. National Food Balance 1.13 Food Supply. As previously emphasized (para 1.01) an adequate supply of food at the national level - from both domestic and external sources - is a key precondition to the achievement of household food security. This section, therefore. attempts to review overall food availability in Mozambique in relation to aggregate need. A major difficulty with evaluating the national-level food supply situation in Mozambique is the lack of data on domestic food production in the family sector. The only data published are for officially marketed production, which for food crops is likely to be a very small and variable proportion of total production in the family sector. These data show that after 1982, there was a marked decline in marketed production for most food crops (see Table 1.1). Prolonged drought during the early eighties and severe flooding in 1984, together with the increasingly disruptive activities of the armed bandits in rural areas were impcrtant contributory factors. Rigid pricing policy also played a role. In real terms most producer prices declined, reducing the incentives to market produce. This effect was compounded by the near total absence of consumer goods for peasants to buy in the rvral areas. In addition, traders faced severe difficulties in access to vehicles, spare parts and gasoline, due to shortage of foreign exchange, compounded by the security problems caused by the bandits who attacked and destroyed large numbers of vehicles and shops. Production from state farms also fell during this period, due largely to the lack of capacity to maintain and repair machinery, much of which ceased to be operational, and the inadequate supply of necessary inputs in sufficient quantities and at the proper time. 1.14 One encouraging sign from Table 1.1, however, is that at least for the two vuain cereals produced - maize and rice - marketed production in 1987 and 1988 appears to have increased in response to the reform measures under the ERP. However, levels still remain far below the levels of marketed production of these cereals achieved in the past. Between the mid-SOs and late 60s, domestic marketed production of maize and rice ranged between 70,000-125,000 tons a year. From 1970 until just after Independence in 1975, marketed production of maize and rice ranged between 150,000-230,000 tons a year. From 1977 to 1984, volumes averaged about 100,000 tons a year, falling to between 40-75,000 tons from 1985 to 1988. Table 1.1: Marketed Production of Food Crops. 1980-1988 1980 1981 1982 1988 1984 1985 1986 198? 1988 (000 tons) Mal" 65.00 78.30 86.20 56.80 82.66 58.66 21.60 27.99 48.20 RIc 43.60 28.90 41.60 17.80 19.16 17.90 19.00 81.69 80.70 Sorghum 6.60 1.60 1.60 1.86 2.10 1.80 6.66 0.61 0.76 Bean. 9.60 14.90 6.96 4.80 3.60 8.60 2.80 9.20 8.39 Cassava 8.80 10.90 9.60 8.60 6.90 6.40 6.o0 7.27 6.62 Groundnut, 6.80 5.00 1.60 0.70 2.00 2.00 0.90 2.18 1.82 SOURCE: Ministry of Commerce  - 8 - 1.15 Although no official data are available for total food production, there are a number of reasons to believe that it has also fallen in recent year. Because of the worsening security situation, as well as the effects of natural calamities, a large number of peasant families have been displaced. These persons, numbering about 1.7 million according to the latest estimates of the Emergency Appeal, have left their homes and land, lost most of their possessions, and are living in accommodation centers or on the edge of urban areas where there is little surplus agricultural land, and therefore are unable to produce more than a small fraction of their food requirements. A further 2.9 million persons are considered 'affected' by the war and are only able to produce around 40X of their food requirements. In addition, it appears that the already low yields in the family sector have declined by up to 20? in recent -ears because of shortages of inputs such as seeds and hand tools due to the virtual collapse of the rural marketing system. A rough attempt was made in the Mozambique Agricultural Sector Survey (World Bank, 1988) to estimate agricultural production in the family sector, taking into account the above factors. The results of the model, using 1966 as the base year, suggest that per capita food production in the family sector is likely to have fallen significantly during the eighties (see Table 1.2). The model is based on data collected during the last Agricultural Census (carried out in the early sixties) plus a large number of assumptions, and therefore the figures shown in the table should be taken as illustrative rather than definitive. However, the general trend shown is plausible and consistent with other estimates of food production trends, e.g., those made by the USDA which are shown in Table 1.3. Table 1.2: Per Capita Production in Family Sector (kg per person per year) Sweet Cassava Sorghum Maizo Rico Groundnuts Beans Millet Potato Sesame 1966 652.1 86.6 68.0 6.9 1a.1 18.0 6.2 8.2 6.8 1980 712.6 88.8 70.4 6.5 18.8 18.7 6.8 8.4 0.8 1986 627.6 23.8 89.0 4.9 9.0 8.6 8.8 1.9 0.2 SOURCE: Chapter I1, Agricultural Sector Survey, World Bank (1988), Report No. 7094-MOZ Table 1.8: USDA Estimates of Production of MaJor Cereals In Mozambique 80/81 81/82 82/88 88/84 84/86 86/86 86/87 87/88 Production ('000 tons) 5a8 "65 570 872 429 518 486 483 Index (80/81=100) 100 112 106 69 80 95 90 80 1.16 An attempt is being made, witn support from FAO, to set up a Crop Monitoring/Early Warning System (EWS) in the Ministry of Agriculture to forecast total food production, including the output in the family sector. The effectiveness of such a system would depend on early information on area planted and on growing conditions, including a network of rainfall statious which would provide information on rainfall, soil moisture and other agrometeorological indicators. Historical information on area under different crops, yields and the relationship yields and agrometeorological conditions are required in order to obtain forecasts with a satisfactory  degree of reliability. Also required for the forecasts is information on pest outbreaks, input supplies and market prices. An important constraint to the establishment of the system is the lack of information on the area under food crops and on the level of production in recent years: the only information that does exist dates from the mid-sixties and hence is of very limited usefulness. Recently, however, with assistance from FAO, a methodology to collect the necessary information has been designed and field tested in Manica province. If the results are good, the methodology could be applied to the other provinces (security conditions permitting) over the next three years. Staff within the Statistics Office of the MOA will need to be trair.ed so that in the future they will be able to manage the necessary data collection. In addition, a serious effort is also necessary to strengthen the collection of meteorological information on a regular and nationwide basis. At present only 10-151 of the 134 different stations are sending information to Maputo on a regular basis. However, it has to be recognized that the problems of security in rural areas and the limited availability of trained local officials represent severe constraints to the rapid implementation of an effective EWS in Mozambique. 1.17 Notwithstanding these limitations of the data on agricultural output, it is clear that, on the one hand, domestic marketed production of food has declined sharply over the last few years, while, on the other hand, there has been a massive increase in the numbers of people who are no longer self-sufficient in food. In total, over 4.5 million rural persons (30? of the population), who in Onormal" conditions would be expected to be fully self-sufficient in most years, are now dependent on free food distribution under the emergency program or on commercial channels for between 60 and 70X of their needs (see Table 1.4). Given the acute scarcity of foreign exchange which severely limits the possibilities for commercial food imports, the result is that Mozambique has become increasingly dependent on food aid to feed much of its population. Figure 1.2 shows how the declines in both domestic marketed cereal output and commercial imports have been mirrored by a rise in food aid cereal imports (maize, rice and wheat). In 1988, cereal food aid reached over 500,000 tons, which represented almost 902 of the total supply of cereals marketed commercially or distributed free under the emergency program.  - 10 - FOOD GRAIN SUPPLY Intemol m4rketing, imports, food aid 600 - 500 - 2009 2 300 - 0 20 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 19E8 EZI mcrketed E i imports aid Figure 1.2: Cereal Supply. 1955-1988 SOURCE: Food Security Department. Ministry of Commerce 1.18 Food Needs Assessment. The Food Security Department (FSD) in the Ministry of Commerce (MOC) makes an annual assessment of the overall f.od balance in the country. The assessment of food distribution needs is based on estimates of the population normally supplied by the market network of the country (essentially the urban and peri-urban population living in provincial and district capitals plus workers on projects and in large enterprises located in rural areas) and of the rural population affected or displaced as a result of war and natural calamities, who are in need of emergency food relief or food purchased from commercial channels to cover part or all of their nutritional requirements. Specific assumptions are made about the degree of self-sufficiency of the different population groups, and whether the balance of their food needs is to be provided through commercial channels, Food Bank or distributed free under the emergency program by the Departamento de Prevencao e Combate das Calamidades Naturais (DPCCN) (see Table 1.4).  - 11 - Table 1.4: Population Dependent on Food Distribution (Commercial, Food Bank and DPCCN) (1) (2) (3) (4) Province Total Urban + Rural Displaced Affected Pop Peri-urban Employees (thousands) Maputo 1850.00 1246.00 198.00 173.00 225.00 Gaza 1200.00 91.50 87.50 38.00 658.00 Inhabane 1310.00 197.50 152.00 188.00 425.00 Sofala 1242.00 255.30 142.50 125.00 195.00 Manica 754.00 53.00 38.00 93.00 147.00 Tete 1010.00 49.30 138.20 97.00 125.00 Zambezia 3078.00 227.80 191.20 539.00 365.00 Nampula 2943.00 330.00 209.20 200.00 536.00 Niassa 635.00 80.00 30.00 199.00 119.00 Cabo Delgado 1134.00 69.00 6.00 36.00 79.00 TOTAL 15166.00 2599.40 1192.60 1688.00 2874.00 (1) Assumed 100Z dependent on comriiercial ;-.arket purchases. (2) Assumed 60? self-sufficient. Other 4(4 to be provided by Food Bank. (3) Assumed 30? self-sufficient. Other 702 to be provided free by DPCCN. (4) Assumed 40Z self-sufficient. Of remaining 6CZ, 80Z assumed purchased in market and 20? to be provided free by DPCCN. -------------------------------------------------------------------__----- 1.19 These population data and the assumptions about the degree of self- sufficiency of each group are combined with estimates of per capita daily food needs and assumptions about milling losses to arrive at estimates of tons of food requirements for market sales, the Food Bank and the emergency program. 3/ In turn, these data, together with forecasts for domestic marketed food production, are used to estimate the need for food aid, which is the basis for the emergency appeal for food aid from the international community. This method has the virtue of simplicity and refinements might not improve overall precision, given the difficulties in obtaining reliable data for the numbers of people who require food assistance and their degree of sel} sufficiency. However, it is worth noting that, in the case of market support, the appropriate indicator is the difference between market demand and supply from domestic sources. However, estimating the demand for food is difficult and typically unsupported by the required data and therefore the report does not recommend an immediate change in the methodology being used. Nevertheless, as policy reforms affect the demand 3/ Daily requirements (grams per capita) Maize Wheat Rice Oil Pulses Sugar Market sales 157 (flr) 123 70 10 40 25 Food Bank 260 (flr) 0 90 10 40 25 Emergency 350 (grn) 0 0 10 40 25 Milling losses of 152 for maize and 25Z for wheat are added to obtain the gross requirements for grain.  - 12 - for food at the household level and its distribution among households, the difference between aggregate demand and supply from domestic sources will change and thus have an impact on food aid requirements both in terms of commodity mix and the aggregate quantitles required. As reforms in Mozambique proceed, it will become increasingly important to attempt to predict the likely direction of changes in the demand for food and their implications for food aid requirements. Table 1.5: Food Aid Requirements Included in 1989190 Appeal Commodity Market Emergency Food Bank Total (tons) Wheat 179000 0 0 179000 Heize 228000 195000 61000 484000 Rice 102000 0 0 102000 Pulses 58000 22000 7000 87000 Edible 15000 6000 1700 22700 Sugar 36000 1000 4300 41300 TOTAL 618000 224000 74000 916000 SOURCE: "The Emergency Situation in Mozambique" 1989, GOM/UNDP 1.20 Based on the previously-described data and methodology, the Government has prepared a request for food aid which was presented to the donor community during the 1989/90 Emergency Appeal (see Table 1.5). Requirements for food aid are estimated to exceed 900,000 tons for 1989/90. While this seems an extraordinarily large request, Mozambique's recent appeals to the donor community for food aid appear to have been so successful that the availability of food in the country as a whole is limited more by internal logistical constraints than by overall supply. The situation for the previous year is illustrated in Table 1.6. In 1988/89, donors p edged 665,000 tons of food aid to Mozambique, which represented 74% of requirements. It is estimated that 90Z of these pledges will have arrived by end AprJl lS89. However, it is projected that only 65% of receipts plus opening stocks will have been distributed by end April, due mainly to transportation problems created by the war and scarcities of trucks, parts and fuel. This suggests that a build-up in stocks can be expected in the current year. During 1988-89, DPCCN distributed 115,000 tons of emergency food. Albeit an increase of 17Z over deliveries in the previous year on the part of DPCCN, these deliveries represented only 60% of receipts plus opening stocks for emergency distribution and only about 50% of needs. Parastatals under the HOC did somewhat better, marketing 330,000 tonnes of food from food aid sources, which represented 71% of receipts plus opening stocks and 66% of needs. However, HOC parastatals have a clear advantage as they supply more urban areas close to ports of entry. As a result of the 1988/89 experience, World Food Program (WFP) has formed the view that the Government's requests for 1989-90 (916,000 tons) cannot be distributed. The Government argues, on the other hand, that its distribution capacity is increasing and would be enhanced still further if its requests for transport capacity were met.  - 13 - 1.21 In general, the information on numbers of the population who are 4isplaced or affected and their degree of food availability is very poor. High priority is now being given to improving both the data base on food consumption and production in rural and urban households and the capacity of the Government to analyze information and develop policies to improve food security. During 1988, two urban household surveys in Maputo and Tete were carried out by the FSD and the MOH Nutrition Section (with Swiss Government and UNICEF assistance). As noted in para 1.16, an Early Warning S7stem is being developed in the Ministry of Agriculture with support from FAO which will provide information on domestic food production. The Government has also requested support from FAQ, under its Technical Cooperation Project, to improve the capacity of the FSD to enalyze survey data. In addition, technical assistance is being providod under the IDA- supported Health and Nutrition Project to enhance the capability of the FSD to carry out policy analysis. Joint preparations by the FSD, the Directorate for Agricultural Economics and the National Directorate of Statistics are currently underway, with assistance from the EEC, FAO and UNICEF, to carry out surveys of rural households in five provinces. A complementary survey for urban areas of the same five provinces is planned under the Bank-supported Social Dimensions of Adjustment Project. While all these initiatives address important gaps in both the data base and the institutional capacity in the area of food security, care must be taken to avoid duplicating efforts or overwhelming the absorptive capacity of the small number of Government counterparts. Consideration should also be given to providing resources to enable staff working on food security issues to participate in short training courses in data collectionlanalysis and policy formulation. Finally, special attention should be paid to developing a single, centralized data base of the food and nutrition surveillance information now being collected and analyzed by a number of ministries, including, inter alia, Agriculture, Commerce, Finance and Health. Table 1.6: Rospons bo the April 1988 Agpeal for Food Aid for Maw 1985-April 1989 Tons Free Market FB TOTAL RequI roents as per AprlI 1988 Appeal 229000 6sees" 169000 598000 Pl dges 213690 419647 81541 664770 Pledges as X of Requirements 93% 84% 19% 74% Rec ipta 151721 419647 27941 699869 Receipts as X of Pledges 71% 100% 89% 90% Opening Stocko plus recelpts 196836 468118 27941 691891 Distribution 115412 830198 8e65 448616 Distributlon *a X of stocks + rceipts eox 71X llX esx Distribution as X of requirements seX 66X 2X 6s% SOURCE: 'Tho Emorgency Situation In Mozambique* 1989, GOM/UNDP  - 14 - D. Food Distribution System 1.22 The preceding section analyzed the aggregate availability of food supplies - from food aid, domestic production and commercial imports - in relation to the estimated food needs of Mozambique's population. This section discusses the ways by which available food is distributed to consumers, many of whom are largely or sol.ely dependent on access to marketed or free food to meet nutritional requirements. Food distrabution in Mozambique is particularly difficult in view of security problems, transport difficulties, limited warehousing capacity, and the scattered points of entry for food donations which make up around 901 of the food distributed in the country. At present, there are 18 main points of entry, some by sea and others by land. Clearing times at many of these are inordinately lengthy, with a concomitantly greater risk of leakages at entry points. Road, rail, water and air transport are all used to move food supplies from entry points to appropriately-located warehouses. While warehousing is estimated to be satisfactory if judiciously managed, major bottlenecks to food distribution result from a large number of transportation rnd logistical problems including, inter alia, the poor state of the roads, shortages of trucks and small vessels for coastal shipping, inadequate cargo handling equipment at many small ports, delays in access to fuel and spares, and inadequate repair and maintenance facilities and technicians for both trucks and shipping, compounded by the lack of security and insufficient availability of military vehicles to escort convoys. A thorough analysis of the logistical constraints to food distribution in Mozambique was prepared by WFP for the 1989 Emergency Appeal. 4/ In addition, a detailed description and analysis of Mozambique's transportation system can be found in .a number of reports prepared during 1988 for the United States Agency for International Development (USAID) by Louis Berger International, Inc. 5/ Finally, the World Bank is currently carrying out a general review of the transport sector in Mozambique which aims to identify the main issues and constraints in the sector and develop a transportation strategy, including recommendations for policy reforms and investment priorities. Therefore this report does not cover in detail the existing capacity and investment needs of the various transportation modes (sea, rail, air, road), but merely highlights some broad themes of relevance for the efficiency with which food is distributed within the country. 1.23 The role of the Government in distributing food staples to consumers has grown significantly in the latter part of this decade as normal trading networks have been seriously disrupted by the war and food shortages have become critical. Within the Government. chief responsibility for domestic food distribution is borne by the MOC and the Ministry of International Cooperation (MIC). MOC is responsible for 4/ See "The Emergency Situation in Mozambique a GOM/UNDP, April 1989. 5/ Louis Berger International. Inc., "Southern Africa Regional Transportation Strategy Paper", March 1988, 'Southern Corridor Transportation Study', ,March 1988, 'Evaluation of OFDA and FFP Grants to CARE -International in Mozambique", June 1988.  - 15 - monitoring the overall availability of food, and for planning and coordinating the distribution of food for sale. Attached to the MIC is the DPCCN which is responsible, under the emergency program, for the distribution of free food to rural families displaced or otherwise seriously affected by bandit activities or climatic factors such as drought and floods. Given the prominent public sector role in food distribution, and taking into account the continuing food supply crisis, the Government has established special institutional arrangements to track food availability and help ensure regular consumer access. At the central level, the principal instrumeat of coordination is an inter-sectoral committee, the Comite Operacional para a Emergencia (COE), chaired by the Vice-Minister of Commerce and including representatives from MOC, DPCCN, the Ministry of Transport, the Nutrition Section of the Ministry of Health (MOB), and the Ministry of Agriculture (MOA). In addition, officials of relevant United Nations agencies, bilateral donors and Non-Governmental Organizations (NGOs) are members of the CCE. Escaping the fate of many such committees, the COE meets regularly and appears to be quite effective. It is advised by a small technical group consisting of the MOC Department of Food Security and DPCCN. The WFP assists the COE by monitoring and coordinating food aid receipts. The COE also has a Food and Logistics subcommittee to help minimize distribution bottlenecks. 1.24 Affected provinces each have a provincial emergency committee chaired by the Governor and with representation from the provincial departments of Commerce, DPCCN, Health, Agriculture, Education, and Construction and Water. For matters relating to food availability and distribution, its executive agents are the provincial directors of Commerce and of DPCCN. In accordance with the previously mentioned division of responsibilities among their two agencies, the provincial directors of Commerce are responsible for distribution of food for sale to the public, while the provincial DPCCN directors are primarily concerned with emergency relief distribution. There are analogous committees in affected districts. The provincial committees, supported by technical advice from sectoral departments (particularly DPCCN), maintain a current assessment of needs through radio contact with the districts, and available food supplies are distributed accordingly, often by army convoy. Information from a MOH nutrition surveillance system is also supposed to be fed into the needs assessments, but tnis does not always happen in practice. The administration of the rationing system in Maputo and Beira is separate, being the responsibility of the Direccao de Comercio, working in conjunction with municipal authorities. Further description of the commercial trading system, DPCCN operations and the urban ration system is given in paras 1.25 - 1.39 below. 1.25 Commercial Trading System. The MOC has the responsibility for planning the allocation of food for the whole country under the so-called 'Plano de Abasticimento' or Internal Allocation Plan. This plan covers basic food staples, as well as a number of important consumer goods and utensils. The MOC prepares a plan for the quota of these products that each province will receive which is sent to the National Planning Commission for approval. The quota for each province is based on various criteria including population; economic importance, especially in agricultural marketing; consumption patterns and local transport  - 16 - conditions. At the provincial level, a distribution plan determining the quota of each product for each district is prepared based on criteria defined by the MOC. The instrument to implement the allocation plan is the 'Orden de Distribucao". This form instructs each importing enterprise and factory to which warehouse to send its output, and each wholesaling enterprise where to distribute its products. Food imports (virtually 1002 food aid) are handled by the importing parastatal IMBEC. Based on the plan, IMBEC allocates part of the food to the state wholesaling enterprises respons4ble for supply of food under the ration system in Haputo and Beira, and part to COGROPA, the state enterprise resronsible for food supplies to the provinces. COGROPA centralizes national food supplies (mostly food aid, but some domestic production) in three national warehouses in Maput3, Beira and Nacala, and then sends appropriate quantities to the provinces in accordance with the allocation plan. At the provincial level, wholesalers are predominantly state enterprises although some private enterprises exist, while at the district level, 902 of wholesalers, and virtually all retail traders, are private. 1.26 Although the final sale of food to consumers and the initial purchase of agricultural commodities from farmers are generally carried out by private traders, their activities are severely circumscribed by the Government. First, as described above, the quantities of basic foods available for sale in any area are centrally determined. Second, the prices and marketing margins for basic foods are also established by the state. Finally, only certain enterprises and shops are permitted to sell products under the plan. These trading enterprises are chosen by the HOC on the basis of criteria such as experience, storage capacity and transport facilities. Since the early eighties, the network of private traders has diminished, particularly in -ural areas, due to attacks on shops and vehicles by bandits, while ;he activities of DPCCN in supplying free food to rural areas under the emergency program have increased dramatically. Although there is some indication of an increase in domestic cereal marketing during 1983 and 1989, it appears that in some areas the official marketing margins are insufficient to cover the transport costs of the traders. Anecdotal evidence inedicates that some traders are therefore paying farmers less than official prices for their produce. A further problem relates to the inability of AGRICOM (the parastatal for agricultural marketing) to move food commodities from surplus to deficit provinces because of lack of funds to cover the war-inflated transportation costs. This has led to a build up of food stocks in AGRICOM warehouses in some provinces and, in turn, to a deterioration in AGRICOM's financial status and its ability to continue to purchase produce from traders and farmers. Some possible solutions to this problem are suggested in Chapters IV and VI. 1.27 DPCCN. Originally established in 1980 to respond to drought, cyclones and other natural disasters affecting the country, the DPCCN has since 1985 also been obliged to respond to the widespread and long-term effects of Mozambique's ongoing security crisis. Today, its primary role is to provide food supplies and other basic necessities (e.g., soap, clothes, seeds and hand-tools) to rural households displaced or otherwise seriously affected by bandit activities. Recipient families are identified by local community leaders, responsible both to the communities in which  - 17 - they live and to the Front for Liberation of Mozambique (FRELIMO) party organization. The actual distribution of food to households is also handled by the community leaders (under the supervision of district administrators and DPCCN), once DPCCN provides relief food supplies at the local level. As this suggests the main role of DPCCN is in handling the logistical aspects of food supply and in working closely with the COE and provincial and district emergency committees to assess food needs and provide supplies as required. Due to the deteriorating sec'rity situation, the numbers of households and communities in need of emergency assistance has grown dramatically over the past few years, with the result that DPCCN has had to substantially expand its activities in the distribution of emergency commodities. For example, food quantities delivered to the district level by DPCCN increased from about 60 thousand metric tons in 5 provinces in 1985 to over 110 thousand metric tons in 10 provinces by 1988. This, in turn, has implied significant expansion of DPCCN's own operational capacity in terms of needs assessments, in-country transport, warehousing, distribution, accountability and reporting. Today, DPCCN, whose 1988 budget was Mt 2.9 billion (US$5.5 million equivalent), has some 1,700 staff and about 340 trucks at its disposal. Between 50-200 staff are deployed at the headquarters of each province, with one or two assigned to each of the 115 districts in which DPCCN is active. Other staff are employed at the DPCCN headquarters in Maputo. The expansion of DPCCN's operational capacity is illustrated by the figures in Tables 1.7 - 1.10 below, which show increases in food quantities managed, transport capacity, storage capacity and DPCCN staff since 1984. rable 1.7: DPCCN Commodities Managed by Type: 1984-1988 (metric tons) Commodities 1984 1985 1986 1987 1988 Total Cereals N/A 52,871 91,081 98,000 151,862 393,814 Beans 1,499 7,933 23,000 23,034 55,466 Edible Oils 701 3,390 10,703 10,564 25,358 Sugar - 222 1,068 3,452 4,742 Milk Powder 1,476 513 3,307 864 6,160 Other Food Items 1,203 1,009 3,150 859 6,221 Sub-Total Food N/A 57,750 104,148 139,228 190,635 491,761 Seeds N/A 1,778 688 8,233 2,549 13,248 Tools 261 46 1,103 326 1,736 Clothes/Blankets 1,311 605 1,655 1,942 5,513 Medicines 18 54 2 83 157 Soap 269 675 2,053 1,859 4,856 Other Non-Food 163 325 979 469 1,936 Sub-Total/Non-Food N/A 3,800 2,393 14,025 7,228 27,446 Total Commodities 61,550 106,541 153,253 197,863 519,207 SOURCE: CARE, February 1989.  - 18 - Table 1.8: DPCCN Fleet Composition: 1984-1988 VEHICLE TYPE -----YEAR DLPLOYED---- BY CAPACITY 84 85 86 87 88 SUBTOTAL OFFROAD OPERATIONAL 1OT and over 45 19 28 88 23 180 38 142 <10T 14 14 7 137 -- 172 49 123 Tractors -- 7 25 47 -- 79 2 77 Light Vehicles 2 17 12 27 12 80 0 80 TOTALS 61 57 72 299 36 511 89 422 SOURCE: CARE, February 1989 ---------------------------------------------------------------__--------- Table 1.9: DPCCN StoraRe Capacity: 1984-1988 (Metric Tons) 1984 1985 1986 1987 1988 Maputo Cntrl Ops. Own 1,500 2,500 2.500 2,500 5,000 Rental 2,500 25,000 ProvincialHeadquarters 4,250 25,000 34,100 50,300 77,500 Districts Headquarters 1,000 3,500 9,100 10,415 17,745 GRAND TOTALS 6,250 31,000 45,700 65,715 125,345 SOURCE: DPCCN/CARE, February 1989 ---------------------------------------------------------------__--------- Table 1.10: DPCCN Staff: 1984-1988 Nos. of Staff 1984 1985 1986 1987 1988 Management 7 9 11 15 15 Mid-level 20 35 66 145 184 Support 350 632 987 1305 1535 Totals 377 676 1064 1465 1734 SOURCE: DPCCN, February 1989  - 19 - Table 1.11: Assessment of Truck Fleet Capacity. 1988 (*) Fleet Size Capacity A. UN-GOM Appeal estimate of existing fleet for 1988 - 279 2908 T B. UN-GOM Appeal solicited additional fleet for 1988 - 203 2100 T C. Estimate by UN-GOM on total fleet capacity required to deliver 260,000 tons to affected areas - 482 5008 T D. Actual operational fleet available 1988- 244 2406 T E. Vehicles received 1988 - 21 260 T F. Total fleet availability (D+E-F) - 265 2666 T G. Difference between fleet requiired and availability (C-F-G) 217 2342 T SOURCEt CARE, December 1988 (*) Does not include tractors and light vehicles. -------------------------------------------------------------__----------- 1.28 Given DPCCN's critical role in moving food supplies to rural areas where they are needed, its logistics and distribution capability are of special concern. Table 1.11 above shows an assessment of DPCCN's current trucking capacity. Despite the expansion of DPCCN's trucking fleet, it appears that fleet capacity in 1988 was just over half of that requiLed to transport the 260 thousand metric tons of food estimated to be required by displaced and affected rural households. This is certainly a large part of the explanation why actual DPCCN food distribution last year (115 thousand metric tons) was under half of requirements, and, without doubt, donors should give the highest priority to meeting the GoverDment's request for additional trucks and other logistical support. Nevertheless, as pointed out in the Government of Mozambique (GOM)/United Nations Development Program (UNDP) document prepared for the 1989/90 Emergency Appeal, shortfalls in distribution cannot be wholly attributed to the insufficient and aging fleet of trucks, inadequate donor response to appeals for increased logistical resources, and attacks by the bandits. In addition, there is increasing concern, within both the Government and the donor community, about the poor management and apparent lack of accountability and cost-consciousness within DPCCN, particularly at the provincial and district levels, which has resulted in stockpiling of food =n urban areas, significant inefficiencies in the utilization of transport facilities, and in considerable 'losses" at various points in the distribution chain. 61 6/ For example, it is widely acknowledged that port losses alone average about lOX in Maputo and are even higher in Beira. To date, there has been little investigation into clairnc concerning corruption and losses throughout the distribution system.  - 20 - 1.29 There is no question that DPCCN's role in providing relief food to needy households has been significant, and that the Department has been a key actor in averting large-scale famine in Mozambique's rural areas. However, the extremely rapid growth in responsibilities, staff and physical resources, illustrated in the tables above, has placed considerable strain on the Department's limited management capabilities to plan, implement and monitor the distribution of food. A more general concern is that, in response to the emergency situation over the last few years, a massive food distribution system, parallel to the existing commercial system, has been built up, which is not run according to principles of efficiency and cost minimization. While this is understandable in the light of the urgent need to move food rapidly to prevent widespread starvation, the likely continuation of the war and consequent emergency situation for the next few years make it imperative to devise mechanisms to ensure that food distribution becomes more cost effective and efficient. In the very short run, this is likely to necessitate an improvement in the managerial and operational capabilities of DPCCN and in its accountability. In the medium-term, it is likely to imply a shift away from distribution of food by Government Departments or parastatals towards increased participation by the private sector. 1.30 Recently, a number of NGOs and donors have devoted significant resources to the provision of training and technical assistance explicitly designed to increase DPCCN's managerial capacity. For example, the recent development of a computerized management information system (MIS) (with assistance from CARE, other NGOs and multi- and bi-lateral donors), has been an important element in improving the planning and tracking of commodities and timely identification of impending gaps in commodity availability on a province-by-province basis. The MIS provides the following information: (a) national/provincial food availability status; (b) status of food pledges; (c) requirements versus pledges; (d) availability projections for allocation planning; (e) commodity flow status; (f) offtake information; and (g) districts reached, with food types and quantities. The reports are shared with national/provincial emergency committees and are the basis for weekly deliberations by the Food Aid Sub- committee. While the MIS in its cu1rrent form has already had some impact in increasing the efficiency of DPCCN's operations, more remains to be done in improving the comprehensiveness of data on cross-border food arrivals and oh arrivals and distribution of food at the district level. A major problem is that, as a result of the rapid expansion, many DPCCN staff, especially those in remote locations, generally have very low levels of literacy and numeracy, leading to significant gaps and errors in data reporting. Improving staff training at the local level is critical, therefore, if the MIS is to provide accurate information for monitoring and planning purposes. In addition, in order to increase accountability, DPCCN needs to introduce some "private sector' pressures on its staff, with penalties such as fines or firing for corruption and carelessness, and productivity-related bonuses. 1.31 In the medium-term, a potential solution to problems of transportation may lie in greater privatization of trucking operations particularly in relatively secure areas. This would help reduce costs and increase accountability and help build a private transport sector in  - 21 - Mozambique. Some steps are currently being taken towards this goal by establishing line items in donor grant agreements specifically for contracting private transporters to distribute emergency food aid. In addition, DPCCN is exploring what incentives, such as hard currency payments or preferential access to spare parts, are needed to attract transporters to emergency transport activities. However, there are a number of difficulties associated with the war which need to be addressed in order to encourage wider participation of the private sector. Military conveys with armed protection are required for about 85Z of operations, which frequently results in significant delays as convoys may take up to two weeks or more to form. This leads to vehicles being tied up inactive for long periods, considerably increasing their operating costs. Explicit recognition of these additional costs in contracts with private truckers would be necessary to make private participation attractive. In addition, despite protection from military convoys, DPCCN has lost about 7Z of its fleet annually in the list 3-4 years due to attacks by the bandits. (In 1988 alone, 37 DPCCN trucks were destroyed). Private truckers would need compensation for the risks of operating under war conditions, either through the inclusion of specific risk premia in the contract, or perhaps through the introduction of some type of donor-sponsored insurance scheme. 1.32 The role of NGOs in supporting the emergency food distribution effort is another area that needs urgent attention from the Government. Recently, frustrations with losses and inefficiencies within DPCCN, and concerns that not enough food was getting to the truly needy in very remote areas, has led some NGOs to set up their own distribution systems, with separate warehouses, trucks and staff. While from an individual NGO perspective this may be the most efficient way to reach its specific target of food distribution, in aggregate, there is a real danger of duplication of resources and of further undermining the weak commercial trading system, as well as possible inconsistencies with the Government's objective of diminishing free food handouts. Nevertheless, NGOs can clearly assist the Government in food distribution in many ways, for example, in training DPCCN staff, as a *watchdog for Government and donors to verify and monitor what is actually happening at the local level, in facilitating participation by the private sector and in directly managing local food distribution operations where appropriate. 1.33 Urban Food Rationing Schemes. While DPCCN's focus has been to help ensure food security for rural-dwellers, the Government's main instrument for promoting equitable distribution of food supplies to urban households has been the Novo Sistema de Abastecimento (NSA) or the food rationing system operating in Maputo (since 1981) and Beira (since 1986). Taking the estimated populations of the two cities together, the NSA now covers about 1.3 million urban-dwellers or over half the estimated total urban population of 2.3 million. Plans do exist to extend the NSA to cover other cities/towns, beginning with Nampula and Inhambane (combined population of about 0.5 million), but progress has been slow given limited administrative capacity. When originally established, the NSA's main role was to provide to registered households the opportunity to purchase at subsidized prices a fixed ration of food staples and some other necessities such as cleaning supplies. However, in line with the priorities of the ERP, subsidies on all NSA commodities have been reduced since 1987, until,  - 22 - today, pricea in the NSA are approaching border parity 7/. Nonetheless, prices within the NSA are still considerably lower than in unofficial markets where prices reflect a scarcity premium, suggesting that the NSA continues to play a key role in helping to allow more equitable access to basic food staples than would be possible under free market conditions. 1.34 As presently set up, the aim of the NSA is to provide each member of a registered household the opportunity to purchase at official prices a ration equal to roughly half of daily caloric requirements or about 1,100 calories per day. In general, this ration is composed of a monthly per capita allocation oft 4 to 5 kg of cereals (rice, maize and maize meal); 1 kg sugar; 1 liter oil; and salt. Until recently, 1 kg of beans was also included in the allowable ration but sale of beans has now been decontrolled. It should be noted that the actual amount of each food staple provided through the NSA varies monthly according to overall availability. For example, given a severe shortage of cooking oil at the time of the mission's visit, the ration for oil was 1 liter per three household members rather than 1 liter per capita as is usually the case. Until recently, the NSA was administered in a notably rigid manner, with consumers required to purchase a fixed basket of commodities on a bi-weekly or monthly basis - a burden sometimes insupportable for the poorest consumers. However, since April 1988 when food prices were raised substantially in relation to minimum wages, there has been an attempt to introduce more flexibility by amending the NSA to enable customers to buy their rations in much smaller quantities and to substitute among the various cereals to achieve a more affordable mix. The main caveat to this latter innovation is that rice purchases through the NSA are still limited to a maximum of 1.5 kg per capita per month reflecting the scarcity of this commodity in relation to demand. In addition, among cereals, only rice and maize meal are available through the NSA in Maputo's 'cement city' since home milling of whole maize is believed to be infeasible in high- or medium-rise apartment buildings. 1.35 Responsibility for organization and overall management of the NSA (for both Maputo and Beira) lies with the MOC, which advises the Ministry of Finance on the appropriate levels for official food prices, monitors overall food availability for the NSA and determines the monthly per capita ration allowable. Day-to-day administration of the NSA is, however, principally the responsibility of the relevant local authorities: specifically, the Gabinete de Organizacao do Abastecimento a Cidade de Maputo (GOAM) and in Beira, the Direccao Provincial do Comercio de Sofala (DPCS). Both agencies report to the MOC as well as local authorities. Actual warehousing and monthly distribution to NSA retailers of ration food supplies is done by parastatals created for the purpose - the Empresa de Abastecimento a Cidade de Maputo (EACM) and the Empresa de Abastecimento a Cidade de Beira (EACB). In turn, consumers purchase ration commodities from NSA retail-shops (usually two or three per bairro), nearly all of which are privately owned and participate in the NSA scheme for a fixed commission (52 of the value of goods sold). 7/ See Chapter IV for discussion of food pricing policy.  - 23 - 1.36 In order to become eligible for the ration scheme, household heads must register their households with GOAM/DPCS. The process of registration requires verification by the community leadership at three levels - agregado familiar (10-households), quarteirao (50-80 households) and bairro (250-350 households) - of the authenticity of information given regarding residency, number of household members and other pertinent information. Provided the information is certified as accurate, the household is registered with the NSA and a ration card is issued to the household. At the same time, the household is assigned to the NSA retail shop of its choice (usually the one closest to its residence), from where it is eligible to purchase its monthly rations on presenting the ration card. For planning and control purposes NSA retail shops are also provided with information regarding the households eligible to purchase rations from them, and are expected to keep detailed records of purchases by each household. Once issued, ration cards are usually valid for twelve months at which time a new card is provided to all registered households. 1.37 The coverage of the NSA in both Maputo and Beira is extremely good, with nearly all long-term residents (over 90Z of the population) registered and eligible for food rations. Nost of these were registered immediately upon initiatior. of the ration scheme, based on neighborhood/household surveys conducted by the community leadership under GOAKIDPCS supervision. For more recent residents - e.g., newly-arrived displaced families - the existence of an extremely lengthy registration process probably inhibits them from participating in the NSA, with detrimental consequences for their food security/nutrition status in the interim. In fact, the system of certification for these families appears to be intentionally slow and carried out on a case-by-case basis, probably In an effort to avoid giving displaced families any incentive to remain in Maputo. In order to gain a ration card, the head of the family must go through the following process: (a) Procure a letter from the Chefe do Agregado Familiar (10-cell leader) in the areas where the family resides ascertaining the family's residence and composition; (b) Have the letter signed by the Chefe do Quarteirao; (c) Take this letter to the Grupo Dinamizador, the head at the bairro level, where a new letter is prepared; (d) Take this new letter to one of eight GOAM offices in Maputo, and fill out a form provided by that office; (e) Have the form signed by the community leaders at each of the three (10-cell, quarteirao and bairro) levels; (f) Leave the form with the Grupo Dinamizador in order that he might collect several similar applications and make a special request to the District Administrator for inclusion of the families in the NSA;  - 24 - (g) If approved by the District Administrator, take the form back to the GOAM office to be issued a provisional ration card valid for the period until regular (computerized) ration cards are issued to all families. Similarly lengthy procedures are required when adjusting the number of household members listed on an existing ration card - e.g, when a baby is born or a household provides shelter to relatives from rural areas. While unlikely to discourage rapid rural-urban migration, which will probably continue as long as the security crisis persists in the country-side, the slow NSA registration process probably increases the food insecurity and nutritional stress of the new arrivals and their host families and should therefore be streamlined as soon as possible. Indeed, recent data from the NSA indicate that there Las been some decline since mid-1Q88 in the total numbers of clients registered with the system. While the reasons for this decline are not clear and require study, they probably do relate in some measure to the difficulties associated with registering. 1.38 Other issues affecting the effectiveness of the NSA relate to weak management, in turn affecting the regular supply of ration commodities to consumers. At present, management of the NSA suffers from inadequate skills of staff, absence of basic office equipment and supplies and inadequate procedures for maintaining and updating records of NSA clients. Moreover, information on food stocks, expected deliveries and offtake is maintained only in an ad hoc manner and long delays occur in invoicing of retailers and in collection of monies owed. Improvements in these areas could substantially reduce the administrative costs of the NSA and improve distribution, thereby helping to improve the access of consumers to food supplies. Indeed, efforts to develop a better, computerized MIS are now underway in both GOAM and DPCS, with assistance from a recently-approved International Development Association (IDA) credit. In addition, regular food availability at ration shops is often hindered by the shortages of transport at EACH and EACB to deliver food supplies to retailers on a regular basis. The MOC, with assistance from IDA, is taking steps to alleviate this problem by encouraging the retailers themselves to take greater responsibility for food transport through providing them with the foreign exchange necessary to purchase vehicles. 1.39 In the past, the ration system guaranteed the urban population in Maputo and Beira access to at least part of their nutritional requirements at official prices substantially below international prices. However, as part of the reforms under the ERP, official food prices in the ration system have increased sharply in the last couple of years (although they are still between one third to one half of parallel market prices), which has led to a significant decline in the real purchasing power of urban consumers. The impact of food price increases on urban household food consumption is analyzed in more detail in the following Chapter. A number of recommendations for improvements in the NSA system, together with other measures to increase the access to food of the poorest households, are made in Chapter VI.  - 25 - E. A Strategy for Household Food Security 1.40 There are a number of constraints that must be considered in the development of a realistic strategy for food security in Mozambique. Overall, although the total resources available to Mozambique are scarce in comparison to needs, the implementation of effective strategies with long- term objectives are limited most by poor security and institutional capacity. First, and foremost, Mozambique is a country at war. The difficulty of access to many areas creates a fundamental dilemma between allocating resources according to 'needs' or according to feasibility. A second set of constraints is imposed by the weak technical and administrative capacity of Mozambican institutions. 8/ These constraints imply a priority to utilize existing structures wherever possible and to use resources in ways which create and strengthen local institutions -- in the countryside as well as the cities and in the private as well as the public sectors. A third set of constraints is imposed by the need to ensure that Government expenditure moves in line with its revenue. An implication of the strategy proposed here is increased public spending for various food security related projects and programs - financed larg-aly with external resources including food aid. It is therefore important that public expenditure on food security be directed most towards investment in future economic growth and least towards transfers which create long-term dependence on unfinanceable social programs. Furthermore, given the tight limits on the government budget under the ERP, increased spending OD food security interventions will have to be financed largely through a reallocation of public expenditure, rather than increases in aggregate expenditure. Finally, budgetary constraints make it especially critical that food security interventions are "targeted', i.e., designed so that their benefits go principally towards the poorest and most vulnerable sections of the population. 1.41 Bearing in mind these observations, this section attempts to provide an overview of ideas elaborated in this report to improve food security at the household level within a consistent policy framework. As discussed in para. 1.01, household level food security is an access issue. Food aid imports can increase the supply of food, but household food security requires that this be matched by a sustained increase in the demand for food by the poorest families. There are three main sources of food for Mozambican families: the farm, the market and the Government. Most effort in Mozambique has gone into creating an effective public food distribution system. The theme here is that the contribution of food aid and other aid should now be directed primarily towards improving access to farm and market food sources by raising labor productivity, improving the value of assets and increasing employment, thus reducing the need for 8/ The first can be relieved by the provision of technical support from abroad. The second is often managed by contracting out the design and management of projects, for example to NGOs. However, the rapid growth of NGOS in Mozambique, from 2 in 1983 to more than 80 in 1989, while augmenting scarce Government technical and managerial skills, impose their own management costs on central and local national authorities.  - 26 - entitlements from the state - albeit maintaining this 'safety net' for those who remain beyond the reach of inccue and employment strategies. It should be recognized that much of this discussion is not specific to food aid - it could apply equally well to the use of financial resources. Food aid is an additional resource of considerable value in Mozambique and therefore must figure prominently in overall food security strategies. But, in many cases, the same objectives would be achieved more efficiently by using money instead of commodities. Although food commodity aid is necessary to food security at the national level (because it augments market supply), once inside the country it has few advantages over money. Money can be moved about more cheaply (a major advantage given transport difficulties in Mozambique) and can be used by recipients more flexibly. This does nct discount the value of food aid in Mozambique. On the contrary, in Mozambique, strategies which raise the level of food demand effectively will make additional demands on market supply which at the moment can only be met with food aid. 1.42 At the household level, where food is available in local markets, food security is imperilled by two main factors: low labor productivity and unemployment. Low labor productivity is an underlying and long-term characteristic of much of Mozambique's rural economy. Displacement by war has also created a massive increase in rural unemployment. In the cities, real incones of many urban households have been depressed by the sharp increases in food prices under the ERP, while unemployment has increased as a result of a reduction in state sector employment. A food security strategy which addresses causes should be directed at increasing household incomes by increasing labor productivity and creating new jobs, thus creating a sustained increase in the demand for food at the household level. The strategy proposed in this report therefore explores ways to increase the productivity of family farmers through more relevant research and improved provision of agricultural services such as input supply, extension, and credit, channelled for example through farmers' credit clubs, and associations or other community based groups. In addition, it proposes ways to encourage the growth of employment in labor-intensive private enterprises, including targeted credit schemes to small-scale ventures to help them expand, training programs and start-up grants to new small-scale businesses. While measures which create more jobs of a permanent kind should receive most attention, they require time and organization to develop. Public employment schemes will therefore be required to undergird more general income and employment strategies. These should not be 'make-work' schemes nor displace labor from other jobs but, instead, should contribute actively to infrastructure which will benefit the poor most. For example, one way of improving rural incomes and increasing labor productivity is to improve access by the poorest to markets and hence the repair and construction of rural roads by labor- intensive means must be a priority. Land improvement schemes also offer considerable potential for temporary employment of under-utilized seasonal agricultural labor or the unemployed, while enhancing long-run agricultural productivity. 1.43 Another major priority is to improve the value of assets owned by the poor - principally their human capital. Basic health services and education are vital to the development of human capital and it is important  - 27 - to ensure that cost recovery measures do not have a negative impact on access by the poor. Given the serious implications of infections and disease for nutritional status and food security, priority needs to be accorded to improving the quality and coverage of both preventive/curative health services per se as well as to ameliorating the environmental conditions (mainly inadequate vater supply and sanitation) which contribute to disease risk. With regard to education, all training need not be fort l or expensive. Community literacy and numeracy programs for adults which are already given high priority in Hozambique should be maintained and extended. In addition, artisans might be encouraged by supplementary grants to take on apprentices for on-the-job skills enhancement. Now that agricultural prices have been increased, strengthening the technical extension services is one of the most important priorities in creating conditions for sustained agricultural growth. 1.44 It would be unrealistic to expect that measures such as these could be put in place everywhere or quickly enough to ensure adequate food security for all. For families without employed members or with low or uncertain incomes, 'safety net" measures are called for. In rural areas, these families will typically be the displaced and those affected by conflict or drought. In urban areas, they will typically be recent migrants and families with only one or no worker and large numbers of dependents. The safety net proposed in this report consists of direct and indirect income transfers from public sources which augment but do not displace income earned from other sources. While, in general, these transfers should be regarded as temporary expedients to be reduced as soon as possible, it is likely that some type of permanent social security system will always be required for labor-scarce households. However, it is essential that transfer income from public resources is affordable and kept within limits which do not create permanent budgetary imbalances for the Government. 1.45 The main safety net currently operating in the cities involves subsidies on food items and the food ration system3 in Maputo (and, more recently, in Beira). In rural areas, free food distribution by DPCCN constitutes transfer income of considerable proportions to needy families. In addition, in both rural and urban areas, Government services - free education, health and sanitation - have been and are an important source of transfer income to poor households. A number of recommendations are made in the report with a view to making this safety-net more efficient. They are based on two main lines of argument. The first is that transfer income in rural areas, including free food aid, should be switched where possible to finance rural investment - in jobs, land Improvement and infrastructure - and that wherever feasible it should strengthen and stimulate rural markets rather than bypassing them. The second is that transfer income should be better targeted to the most vulnerable. However, this involves administrative costs which may be considerable unless measures which are self-targeting can be designed. 1.46 In reality, given the difficulties associated with the security situation and the weak institutional capacity, the best may sometimes have to be foregone for the good when designing and implementing a strategy for food security in Mozambique. For example, although the creation of  - 28 - permanent jobs might otherwise be given highest priority, in rur&l areas where access is difficult and people are urgently in need of immediate nutritional assistance, preference may need to be given to simpler safety net measures. When targeted income transfers would be more efficient but impose undue administrative costs, they may need to be set aside for less efficient ways of ensuring food security. In inaccessible areas, giving resources to local communities to allocate and manage themselves may be a better strategy than doing nothing, even if it means forgoing the imposition of technical and administrative criteria and accepting some inefficiencies and loss of control.  - 29 - II. HOUSEHOLD FOOD SECURITY AND NUTRITION A. Nutritional Situation 2.01 Data Sources. Sources of data on the nutritional situation of the Mozambican population are limited to: (a) a number of small surveys conducted in selected rural and urban areas over the past 5-6 years; and (b) the "Vigilancia Nutricionall (Nutrition Surveillance) system developed by the MOH to monitor the nutritional status of children through weight data collected at health facilities. The latter system uses an anthropometric indicator - growth faltering or failure to gain weight between monthly visits to Maternal and Child Health (MCH) clinics - together with relevant observations from health staff concerning epidemics, population movements, etc. to monitor and analyze the nutritional situation of children under three years. Thus, the ability of the system to provide an accurate picture of the nutritional status of young children nationwide is necessarily limited by the effectiveness of the coverage achieved by the MCH system itself. It is further constrained by irregularity in the flow of information from health facilities to the central MOH Nutrition Section. In the past two years, however. progress has been made in improving information flows, with surveillance results now available at least intermittently from most districts, though scmetimes only for district capitals because of security problems. Information from health facilities on the numbers of severely malnourished children admitted to pediatric wards and the incidence of low birth-weight births is also useful in determining the seriousness of Mozambique's nutritional problems. 2.02 Nutrition surveys are conducted mostly by the MOH Nutrition Section, frequently in places with suspected particular problems. These surveys focus on young children and assess nutritional status through anthropometric measures, including: height for age, generally regarded as an indicator of chronic malnutrition when below accepted standards (stunting) 1/; weight for height, which when low indicates presence of acute malnutrition (wasting); and measurement of arm circumference which when less than 80? of standard measurements also indicates acute malnutrition. In addition, recent surveys have placed some emphasis on surveying family food stocks, for early warning purposes. Over 75 spot nutrition surveys have been conducted over the past five years (by Government and by NGOs) in selected rural and urban areas and, together with surveillance information, help give a reasonable picture of Mozambique's nutritional situation. Their usefulness in determining nutritional trends is, however, limited because the surveys are rarely repeated in the same geographic locations or with comparable sampling techniques. Moreover, to the extent that the surveys are necessarily conducted in accessible, relatively secure areas, they probably underestimate the national prevalence of acute malnutrition and, to some extent, of chronic malnutrition as well. 1/ Some nutritionists argue that low height for age can also be an indicator of acute malnutrition if occurr-ing in children aged 0-3 which are considered to be 'growth years".  - 30 - Table 2.1 Incidence of Chroilc and Acute Malnutrition: Date from Spot Surveys 1985-U7 ~~~~~~- - - - .- - -_ -- -- -- -- -- -- -- - -- - -- -- -- - -- - ---- X Malnourished No. of Province/diat Hnth/Yr ---- ---------------- Children low hoeght low weight low arm Surveyed for age for height circumforence MAPUTO CITY 1. 7 bairro. 13/65 23X 3X -- 51 2. B. Malanga 1/6 SoX aX 403 INHAMBANE 1. Massing& 1/85 3SX 2 -- U83 2. Morrumkene 12/85 47X 2X 497 GAZA 1. 8 VIllag es Chicualcuala 1/U8 - C X 457 2. Chicuaia- cuala 1/86 12% 5X -- 253 8. Dlndiza 11/8S 8 4X -- HSC. 4. Chibuto 5/87 6sx 6x as S. Hokwo 7/87 aox -- -- NA 6. GuiJa 8/87 - x -- HSC 7. Guij (dilpl 8/87 - 5% HSC* 8. Chibuto 9/87 - 6% "SC* 9. Manjaecze 9/87 - 12X - HSC MANICA 1. Mossurize 5/86 sox 4X 214 2. Espungs 7/86 6OX -- -- NA S. Espungabeira 11/86 35x ex 288 4. Machaze 2/88 a82 7x -- 141 5. Mossurize 8/87 47 -- NA 6. Machaze 9/87 as 1X HSC* 7. Vanduzi 9/87 -- ax HSC* SOFALA 1. Nhamatanda 2/88 29X 7x 281 2. Nhamatandn 2/87 4CX 6X -- 248 8. Machangs 4/87 26X lo - x TETE 1. Changara 3/85 8SX SX -- 28 2. Estim 12/85 6% e8x - 227 S. Changars 1/U6 42X SX - 818 4. Dengue 8/36 8sx - NA S. Canongoln 9/86 SoX -- NA S. Moatlt 2/87 5X 6X -- 1ol 7. Senga 2/87 Sax ex Sl1 8. Macanga 3/87 s9o 18X 102 ZAMBEZIA 1. quelimane /85 O2S S -- eS0 2. Mocuba 6/87 - 9X 828 8. Niced'ne 8/87 4X NSC. 4. Nante 8/87 -- 4X "SC S. Chinde 8/87 24X "SC* 6. Maln 9/87 -- 19X HSC* NAMPULA 1. Murrupula 8/87 - ax HSC. 2. Mmba 6/87 -- -- 1X HSC. NIASSA 1. Muses 9/87 3x H SC* * H"lth Systm Clients Source: Ministry of Health, Nutrition SectIon  - 31 - 2.03 Nutritional Status. Based on available data, there can be little doubt that malnutrition, particularly among the highest at-risk groups (infants, young children and pregnant women) is a serious and widespread problem in Mozambique. As the figures from 1985-87 spot surveys show (Table 2.1), the prevalence among children of "long-term' or chronic malnutrition, measured by height-for-age, is high: stunting was below 30? in only 4 out of 25 cases; between 30-402 in 10 cases; and over 40? in the remaining 11 cases, thrice rising to 60Z or over. Incidence of acute malnutrition, when measured by weight for height, averages slightly over 6Z and ranges from 1-13Z. If data on low arm circumference are included, the average rises slightly, but is still within the 6-7Z range. The results of the MCH-based system of child growth monitoring (Tables 2.2 and 2.3 below) also point to the poor nutritional status of Mozambican children. At the national level, there seems to be no seasonality to the incidence of growth faltering (Table 2.2), but this is probably due to the urban bias of national growth monitoring data. Indeed, studies in the green zones surrounding Maputo (and some other rural areas) indicate that there is substantial seasonality in incidence of malnutrition. Overall, there is little doubt that the nutrition situation in most parts of Mozambique is chronically precarious, with a high risk of accelerating to crisis proportions in the face of any further adversity. Table 2.2: Prevalence of Growth Faltering: 1987-88 Data by Province 1987* 1988 Total Total Z Total Total Z Province Children Growth Children Growth Weighed Faltering Weighed Faltering -----------------------------------------------------------__--------- Maputo City 96473 7332 7.3 48389 7417 15.3 Maputo Prov. 65373 15282 23.4 Gaza 66009 4732 7.1 108843 7687 7.1 Inhambane 73982 24729 33.4 158881 32089 20.2 Sofala 60807 10069 16.5 102938 12770 12.4 Manica 5283 1008 19.0 66743 15629 23.4 Tete 10843 1496 13.7 Zambezia 44822 6661 14.8 53899 7303 13.5 Nampula 40676 9406 23.1 37546 5990 16.0 Cabo Delgado Niassa --------------------------------------------------------------__------ Total 398895 65433 16.4 642612 104167 16.2 -----------------------------------------------------------------__--- * 1987 data are for 10 months only (January-October) SOURCE: Ministry of Health, Nutrition Section  - 32 - Table 2.3: Prevalence of Growth Faltering: 1988 National Data by Month --------------------------------------------------------------------__--- Total Total Total Total Month Children Growth 2 Month Children Growth S Weighed Faltering Weighed Faltering -----------------------------------------------------------__------------ January 62095 8755 14.1 July 47857 7575 15.8 February 59500 9905 16.6 August 56687 9209 16.2 March 60676 11335 18.7 September 57663 9390 16.3 April 54138 8952 16.5 October 56590 7883 13.9 May 62181 11299 18.2 November 58326 9875 16.9 June 52384 8575 16.4 December 14545 1414 9.7 --------------------------------------------------------------------__---- SOURCE: Ministry of Health, Nutrition Section -----------------------------------------------------------__------------ 2.04 Another indicator of the poor nutritional status of Mozambicans is the high incidence of low birth weights, usually the result of deficiencies in maternal nutrition. The last nation-wide study (1982) estimated that about 16Z of births fell below the accepted minimum of 2.5 kg. Recent hospital data show a similar incidence, but suggest that, if anything, incidence of low birth weight has risen slightly. For example, data from the Maputo Central Hospital for the first three months of 1988 show that an average of 17.1S of births were low birth weight, with the situation presumably worse in areas of the country less well supplied with social and economic services. Indeed, United Nations Children's Fur. (UNICEF) reports that up to 27S of births are low weight in some Mozambican provinces 2/. Data from Maputo Central Hospital also show a dramatic increase over the past five years in the numbers of patients admitted to pediatric wards for treatment of acute malnutrition (both kwashiorkor and marasmus). Between 1982 and 1986, the total number of cases of acute malnutrition admitted annually were in the 5-50 range, but a sharp increase occurred in 1987 and 1988, with nearly 250 cases admitted in 1987 and 125 admitted in the first four months of 1988. 2.05 While the levels of malnutrition described above are not unprecedented in Sub-Saharan Africa, they are, nonetheless, at the highest end of the range. Moreover, the situation is likely to be much worse in areas overrun by bandit activity and/or severely affected by adverse climatic conditions for agriculture. Anecdotal evidence from relief workers indicates that it is not unusual to find levels of acute malnutrition approaching 20-252 among displaced children newly-arrived at resettlement camps, many requiring nutritional rehabilitation at paediatric wards. Provided adequate services are available to treat them, however, many of these children soon recover to normal levels. In this context, of the approximately 35 spot surveys conducted among insecurity- and drought- affected populations in the first eight months of 1988, only 7 found incidence of acute malnutrition (measured by weight for height) to be higher than 202, with 17 finding incidences of over 10S. 2/ Children on the Frontline: A Report for UNICEF, UNICEF 1989  - 33 - 2.06 Micronutrient deficiencies in Mozambique appear to be somewhat less critical than is the case in many other low income countries, although anemia, Vitamin A deficiency (and consequent xerophthalmia), goiter and pellagra have been found in at least pockets of the country. Of these, anemia, widespread among women, is the most serious. Vitamin A deficiency has been identified as a problem in Zambezia and Manica Provinces, although the data are somewhat ambiguous. Accommodation centers for displaced persons aiso report a high level of Vitamin A deficiency. Problems of goiter appear to be concentrated in the northern areas of Tete and Niassa provinces, while pellagra has been identified among displaced populations in Tete who have been forced to subsist solely on maize. Although micronutrient deficiencies are sometimes related to diet, they are more often the result of food shortages or low incomes leading to an inadequate or uncharacteristically monotonous diet - as in the case of pellagra. The common occurrence of anaemia is, in turn, more the result of parasitic illnesses, particularly malaria and intestinal worm infestations, than due to unduly low intake of iron-rich foods. The 'normal" dietary patterns of Mozambicans (i.e., in conditions of normal food availability and buying power) are discussed in paras. 2.07-2.09 below. 2.07 Dietary Patterns. Although information on historical household food consumption patterns is incomplete, it appears that in pre-colonial times, the mainstays of the Mozambican diet were millet and sorghum porridges. During the colonial period, a number of new crops were introduced by Portuguese settlers, including maize, rice and cassava, all of which became incorporated in the Mozambican diet. In urban areas, bread and pasta (made from imported wheat), were introduced, but consumed mainly by the colonial elites until Independence. In the period since Independence, the normal Mozambican diet includes nearly all the staples mentioned above, but dietary patterns vary by geographic region, and reflect previously discussed changes (Chapter I) in agricultural production, food imports and life-styles. In the south of the country, the main household food staples are maize (preferably white maize) and rice, with northern households relying mainly on maize, and in remote drought- prone rural areas, on sorghum, for the bulk of caloric requirements. Dried maize and sorghum are usually ground into meal and eaten in the form of stiff porridges. A common supplement to these staples in rural areas is cassava, eaten either boiled or made into dried flour for porridge; and, in urban areas, wheat products are increasingly popular, due to their convenience for the urban life-style and wider availability, given food aid in the form of wheat. These staples are complemented by sauces based on varied ingredients such as green leaves, beans, pumpkin leaves, groundnuts and dried fish, depending on local availability. Consumption of fruit Is not very common, except in urban areas and as income increases. Use of oil, sugar and animal products is dependent on their availability, with use of the former two items (in sauces and in tea resF etively) more widespread, provided income is adequate. 2.08 The number of meals consumed daily varies according to geographic area (urban/rural) and, of course, income in cash and kind. In rural areas, people traditionally take a light breakfast accompanied by sweetened tea before leaving for their farms at dawn and a main meal at around 10-11  - 34 - a.m. after which they may return to agricultural work. Dinner, if eaten at all, is a light meal, taken around sun-down. In urban areas, meals are generally eaten three times a day - usually a light breakfast with tea or coffee, lunch at midday, and a late dinner. However, due to transport problems and the absence of cooked food vendors common in other African cities, many workers skip lunch or carry a light snack. 2.09 With regard to feeding of infants and weaning-age children, frequently the most vulnerable age-groups from a nutritional standpoint, existing practices appear to be satisfactory - i.e., deleterious belief patterns and practices concerning child feeding do not appear to be a significant factor affecting the growth of young children anywhere in the country. Breast feeding is near-universal and continues on average for 16 months. Solid food supplementation normally commences at 6 months, and, when resources are adequate, mothers have no problem preparing adequate maize-based porridges for the children. In many parts of the country, particularly in the south, weaning-age children are fed a fermented porridge made of sprouted maize. This porridge does not harden to the stiffness of ordinary maize porridge during cooling, meaning that it need not be diluted, with the resulting decrease in caloric density. when fed to children. The porridge, which is believed to have infection-fighting properties, is also fed to older children and adults, when sick. Finally, such information as does exist (mostly anecdotal) on intra-family food distribution indicates that, if anything, children are favored when problems of absolute food shortage occur. For example, relief workers frequently report that children of families arriving at accommodation centers are better nourished than tLe adults. 2.10 As the above suggests, the "normal" Mozambican diet is well- balanced from the standpoint of nutritional requirements and in the absence of real resource constraints, families make generally rational decisions regarding food consumption and feeding and care of young children. Existing widespread problems of malnutrition in Mozambique are primarily raiated to problems of absolute food shortages and to inadequate purchasing power on the part of consumers. Persistently poor environmental health conditions, including unsafe drinking water supplies, inadequate waste and garbage disposal systems and other hygienic problems, have also had negative consequences for nutritional status, given the close interrelationships between nutrition and morbidity. In addition, in the Mozambican context, the seriousness of these factors has frequently been dramatically increased (particularly in rural areas) by recent adverse climatic conditions (mainly droughts) and by the continued insurgency of Mozambique National Resistance (MNR) bandits, with attendant disruption of communication channels, administrative and economic systems, and social services. The remainder of this chapter examines in some detail the relative impact of these factors on nutritional and household food security status and attempts to characterize the households most at risk.  - 35 - B. Food Availability. Incomes and Household Food Security Rural Food Security 2.11 As the analysis in Chapter I of declining trends in family sector agricultural production indicates, there has, in recent years, been a significant deficit in rural food supplies. As discussed, the decline in production has been partially the result of economic and policy factors - e.g., distortions in domestic agricultural pricing policies, neglect of the family sector in favor of (relatively inefficient) mechanized state farms, shortages of key inputs and reduced efficiency/incentives of marketing systems - many of which the Government has begun to address. It is also attributable to "emergency' factors including a series of climatic disasters (droughts and floods) during the 19809 and lack of security in much of the countryside caused by armed bands of anti-Government insurgents. These emergencies have displaced millions of rural households or otherwise reduced their capacity for food self-provisioning, leaving them primarily dejendent on relief supplies of food or on marketed food to meet nutritional requirements. 2.12 Table 2.4 below shows figures prepared for the 1989 Emergency Appeal of the rural population estimated to be dependent on external food supplies. These include: (a) agricultural employees (i.e., 60Z self- provisioning with the remainder supplied through the WFP-supported Food Bank); (b) displaced (i.e., resettled within the past 12 months and assumed to be 30Z self-provisioning); and (c) "affected" (i.e.. displaced population, resettled in a new, secure location for one year or more and assumed to be 402 self-provisioning) 3/. Figures in the ^other' column indicate the estimated numbers of rural dwellers whose food security status is not known. but who are (optimistically) assumed to be self-provisioning. In practice, a significant proportion of households in this group might be expected to be food insecure but located in areas which are currently inaccessible to systematic relief operations, either because of widespread bandit activities or other logistical difficulties. 3/ There is some debate within Government and among informed donor agencies about the accuracy of these estimates -- not surprisingly given the continued disruption caused by bandit activity and the difficulties of verifying the self-provisioning capacity of displaced and affected households. It is extremely important to improve the data on the degree of self sufficiency of rural households. Information collected under the EWS field-test in Manica and during the planned rural survey in five provinces should provide more up-to-date estimates on rural households, food production and consumption levels. There is, however, general agreement that the above estimates do reflect the order of magnitude of the problem. It should also be noted that as of 1989, the gap between food requirements and self-provisioning for affected households would not be met entirely through relief supplies. Rather free food distribution would amount to only 202 of this gap with the remainder sold to them from MOC supplies at official prices.  - 36 - Table 2.4: Estimated Food Security Status of the Rural Population Province Total Rural % Displaced X Affocted X Other X Rural Pop Employ..a Pop. Pop. Maputo City a 0 0.00 0 0.oo 0 o.o0 0 o.0 Maputo Prov 8soooo 19806 30.46 173300 26.66 224900 34.80 56S10 6.28 Gaza 1150000 87600 7.61 38500 8.86 667800 57.20 28682 81.84 Inhambane 1190000 162000 12.77 187900 15.79 42MM 35.71 42511 35.72 Sofala 103010 142638 13.84 126000 12.14 195010 18.98 567402 55.09 Manica 727000 38000 6.28 93100 12.81 147500 20.29 446410 61.6 Tote 986000 138234 14.03 970G0 9.86 125100 12.69 624766 68.43 Zambezia 2953000 191227 6.48 538900 18.25 864900 12.86 1657978 62.92 Nampula 2656486 209185 7.88 200560 7.65 635590 20.18 1709910 64.89 Cabo Delgado 1094000 80000 2.74 36410 8.83 79400 7.26 946210 6.67 Nass 590000 8000 1.02 199800 88.78 119800 20.22 266410 44.98 Total 13024485 1192684 9.16 1689900 12.97 2874710 22.07 7207201 65.8J SOURCE: Ministey of Commerce, Department of Food Security 2.13 As the table above indicates, under 102 of rural dwellers are estimated to be assured of food supplies through self-provisioning and agricultural employment, with another 35Z or so (in the displaced and affected groups) heavily dependent on relief or marketed food distribution to meet nutritional requirements. Moreover, to the extent that relief supplies are intended to be supplemented by self-provisioning to meet food needs, households in this latter group are also often dependent on emergency supplies of seeds and tools. The food security status of the displaced and affected groups is, therefore, necessarily related to DPCCN's ability to reach them on a regular basis - a task immensely complicated by the extraordinarily difficult logistics conditions prevailing in the country. As the map in Figure 2.1 shows, a few areas of Mozambique are totally inaccessible. Other extensive areas are accessible only with heavily armed escort (not adequately available in Mozambique) or by extremely expensive airlift. DPCCN records show that in 1988 an average of only about two-thirds of the courntry's 125 districts were reached in any given month, suggesting that many displaced and affected households suffered at least temporary food insecurity over the year. In addition, the fact that total DPCCN food distribution in 1988 amounted to about 115,000 metric tons (not including cross-border or direct distribution in areas around ports) - versus an estimated requirement of over twice that amount - indicates that food insecurity has been widespread among the displaced and affected groups. 2.14 As discussed earlier, few firm data are available about the food security situation of the population included in the 'other' category. However, based on information about agro-climatic conditions (see Map 1 in annex) and anecdotal accounts regarding agricultural production, it appears that in 1988 and 1989, food supplies have probably been satisfactory in the northern and central parts of the country (i.e., in Niassa, Cabo Delgado, Nampula, Zambezia and some parts of Tete, Manica and Sofala). In the relatively arid south, however, subsistence food crop production has probably been far more difficult. In addition, given the uncertain security situation, it seems likely that food insecurity prevails even  - 37 - MOZAMBIQUE CONDITIONS OF ACCESSI8ILITY FEBRUARY 1989 / ) 5 ~UNITED REPUBLIC OF NI ELGADO Pemba - Free overland accesuelimane ZIMBABWE }a - 'V',dnculos rr lake jJJJMilitary escort access* GAZ Access only by airei Non accessible SOUTH .External access * Prolecion to guard aainst random aftacka by 'Acess by air due to mnaoMaseblt caused by .: - -;*tor ro dtian or curiy tat 1989190 Emr gency App ncluds ,O, fAAPUTO ~~~~~~~~~~~~~~~~~road/brndge repai to educe arasf airil :>.i .--puto 0 200 400 m Wss[1 '0. I t*. 1 wow rEtRV .A1I%~D 11111 0 10O 200mi EJ~ 11 bCiMiUi~ z lnupo gi,l the UndW l ti MAP NO 34771 Rsw UI TSO kATON FEINUARV lug  - 38 - where agro-climatic conditions are favorable. For example, despite good yields last year in bandit-overrun areas of Zambezia, spot nutrition surveys conducted in areas resecured by Government forces indicated that levels of acute malnutrition (measured by weight for height) approached 502 - largely the result of household food supplies being stolen or destroyed by bandits. Bandit activity .s currently most concentrated in Nampula, Sofala and the south. The pr;.:arious food security status of rural households was also highlighted by surveys conducted by the HOH Nutrition Section in selected areas of Tete, Manica and Sofala in 1986-87 which showed that in most cases over half the households surveyed had only 0-4 months of food stocks available versus 10-11 months stocks needed until the next harvest. Although current information does not permit a quantification of the numbers of rural households subject to food insecurity, a better picture should be possible once data from the Early Warning System field-test in Manica and the planned rural household survey in five provinces become available. 2.15 Apart from subsistence food crop production, the food security of many rural households is also dependent on income from remittances by migrant labor, exchanges and sales of cash crops and non-food items produced by the family, livestock holdings, and employment on plantations and large farms. In the past few years, mainly as a result of bandit- related insecurity and unfavorable employment prospects in neighboring South Africa, serious declines have been apparent in income earned from these activities: 41 (a) Firstly, revision of hiring policies by the South African Government and the Chamber of Hines since 1985 has significantly reduced employment of Mozambicans in that country, causing substantial hardship to rural households in the arid southern provinces of Mozambique from which many miners had been recruited. Available data indicate that in a single year mining employment fell by about 15Z, from nearly 56,000 jobs in 1985 to just under 48,000 in 1986. Moreover, employment opportunities for Mozambicans in South African agriculture have also been affected by increased expulsions of Mozambican agricultural workers by South African authorities since 1986. (b) Secondly, the combination of drought and insurgency has had serious repercussions for livestock holdings in the southern provinces and parts of Tete, where the rural population has traditionally depended on income from cattle-raising, given the absence of the tse-tse fly, and poor agricultural conditions. The number of cattle is estimated to have fallen from 1,400,000 in 1980 to 750,000 by 1986. More recent data are not available but informed opinion suggests that the declining trend has continued. 4/ Much of the discussion in this paragraph is based on the excellent analysis of rural food insecurity contained in the 1989 GOM/UNICEF Situational Analysis.  - 39 - (c) Thirdly, in the northern provinces (and to some extent Inhambane and Gaza in the south), where cashew and cotton production is an important contributor to household income, marketing declined by over 50? between 1980 and 1986 - with obvious consequences for household food security. Increased producer prices under the ERP have led to some increases in marketing in the past two years, although sales are still well below their 1980-81 levels. (d) Finally, potential for income-generation from production and sale (or barter) of food and non-food items has been sharply reduced because of the severe reduction in coverage of rural shops and other trading enterprises, many nf which have been destroyed by MNR insurgents. This reduction in rural trade has also meant that many rural communities are now cut off from access to key agricultural inputs and basic consumer goods. Urban Food Security 2.16 In urban areas where only a small proportion of food consumed can be derived from household production, food security is largely dependent on physical availability of marketed food from domestic and external sources and the ability of households to purchase adequate amounts of food at prevailing official and parallel market prices. Since 1987, given increases in food donations for market support, increased marketed production from green zones surrounding cities, improved logistics and establishment of the ration system (NSA) 5/, problems of absolute food shortage have been largely ameliorated in the two main cities of Maputo and Beira (which together contain about two-thirds of the urban population). Temporary scarcity of marketed food supplies (due mainly to logistical difficulties) is still common in many smaller cities/towns, where, however, there is often greater scope for partial self-provisioning through cultivation of garden plots. Thus, although shortages of marketed food supplies are not inconsequential contributors to problems of food insecurity/malnutrition in urban areas, the role of household financial access issues appears to be far greater given recent changes in consumer prices for food vis a vis wage policies and employment conditions. 2.17 A time series of monthly official food prices and minimum wages from January 1987 to February 1989 is presented in Annex ZA. Some important observations can be made from the data contained there.n. With regard to official food prices, it appears that over the 26 month period, average nominal price increases were roughly equal to the aggregate devaluation of the metical 6/, so that the prices of the major food staples (rice, maize meal, maize and sugar) rose between 1,500 and 2,000X. Nominal 5/ Chapter I provides more detail regarding food availability and distribution systems. 6/ The USc/kg prices of rice, maize meal, maize and sugar in February 1989 were virtually identical to the corresponding prices in January 1987. The prices of oil and beans, lhowever, were about 30Z lower in US dollar terms in February 1989 as compared with January 1987.  - 40 - meticais prices of oil and beans rose at a slightly slower rate - about 1,100 and 1,0002 respectively. Moreover, it should be noted that these price increases did not occur in a gradual fashion: the prices of the major food staples doutled between January and February 1987, then held steady until August 198i when they went up by another 50Z or so; from then on, there were no further price increases until April 1988, when sharp increases of 400-600Z (depending on commodity) were put into effect. While these data on nominal food price increases provide some indication as to the difficulties urban dwellers have been facing in meeting food requirements, they need to be related to changes in income (i.e., employment and earnings) if their impact is to be better understood. The analysis in para 2.20 below therefore examines the ratio e. some food prices to urban wages in an effort to further elucidate the likely food security status of urban households. 2.18 EAamination of the relationship between food prices and wage levels reveals the following: (a) that during 1987 and the first quarter of 1988, the prices of the major food staples important for low income consumption relative to the urban minimum wage did not change drastically 7/; (b) that the price-wage index for almost every staple commodity (except oil and beans) 8/ shot up dramatically in April 1988, although somewhat less steeply for maize than for rice, sugar and maize meal; and (c) that some decline in prices relative to wages occurred in October 1988, when new wage guidelines were promulgated by the Government. These trends are illustrated graphically in Figure 2.2. The price-wage index presented in the figure has been calculated by comparing the price of each food commodity deflated by the prevailing minimum wage with January 1987 prices and wage rates. The results are expressed as an index where the January 1987 price-wage ratio is 100 9/. 7/ Relative to minimum wages prices of the main cereals (rice, maize meal and maize) and sugar rose in August 1987, but not much ab.-.,e their January 1987 levels (except for maize meal). Prices of oil and beans rose more sharply, but given the relatively lower weight of these commodities in the household food basket, these increases probably had a limited impact on household food consumption. 8/ The official price of beans rose sharply in early 1987 when beans were removed from the ration system. Since then, prices have remained roughly level, implying a declining trend in the price-wage index, given increases in the minimum wage. Similarly, official oil prices have remained steady since about July 1987 when the price was increased about five-fold. 9/ The price-wage ratio approximates a terms of trade index for urban consumers (which is not presently available in Mozambique) but is inferior to it because it does not take into account the cost of other consumer goods, accommodation and other services such as transport, education and health. It also does not take into account the amount of each food item consumed by low and high income groups. AGRICOM and the MOA are now attempting to calculate an internal terms of trade index -- a task which should be given high priority because of its relevance for policy analysis.  - 41 - Figure 2.2 Major Food Staples: Price-Wage Indices UO -. .. 350 * 300 250 ISO Jon 87 Apr807 Jul 87 Oct387 Jan 83Apr 83 Jul88 Oct808 Jim 89 Munt,/rw a N 4 mm M Ad S x 0 S Note: R. rice; MM - maize meal; M - maize; S= sugar; 0= oil;& B beans 2.19 The picture presented in the preceding graph indicates that there has been a significant increase in the official prices relative to wages of all the basic food commodities over the past 26 months. This implies that significant real declines have taken place in the purchasing power of even wage-earning households - particularly since food purchases account for over half of household expenditures except perhaps in the highest income brackets (see para. 2.25). Of course, it should be noted that urban families can generally purchase only up to half of food requiremefts at official prices 10/, with the remainder necessarily purchased on parallel markets, where prices are much higher (see Annex 2B for a list of parallel market prices for some common products). As such, data on trends in the parallel market prices of food would be necessary to examine the full impact of food price changes on household buying power. In the absence of this latter data, however, an attempt has been made to estimate the necessary food expenditures (including both official and parallel market purchases) for the average urban household dependent on purchased food to meet nutritional requirementsI The estimate, presented in table 2.5, is 1Oj I.e., in Maputo and Beira where the NSA helps ensure that at least hF.f of individual/household food requirements are available at official Frices. In. other urban areas the amount of food available at official ,prices is probably variable, dependent on the availability of supplies at itfair-price" shops.  - 42 - based on an MOH assessment of the daily food requirements of Mozambican adults. Tablo 2.5: Household Food Requirements and Estimated Costs Food Items Daily ReqMthly ReqMthly Reqqty Aval lEst Req Pur Est Total Per Cap Per Cap Per NHi In NSA Cost on Mkt Coot Cost gu/mIs kg/l kg/I kg/I (mt)** kg/I (at)** (mt) Ceresls 8S0 10.5 70.4 32.6 8e"6 86.4 6401 166S1 Pulses 6 1.6 12.1 6.0 6 12.1 4646 464 SuIar s6 1.5 10.1 8.6 2112 2.1 945 8657 Oil 40 1.2 8.0 4.6 2886 8.5 3266 8641 Totels 8547 16192 28989 * The 1988 Maputo Household Incom, and Nutrition Survey found average urban housohold sizo to bo 8. Ago distribution was not available but for purpose of this analysis is assumed to be 1 member >66 yrs., 8 *ged between 15-45, 1 each gagd 10-15 yr. and 6-10 yrs rospectively and 2 aged below 6 yrs. As such, housohold size In adult equivalents for estimting food requirement. to assumed to be 6.7, based on the following woights: >S6-.8, 16-46=1, 16-153.9, 5-106.8 and 0-6=.6. - Where a choice of nutritionally comparable commodities Is avallable It Is assumed that the choapest kind Is bought -- e.g., wholo maize among cereals or brown rather than white sugar. as These per capita daily requiremnts ropresent the barest minimum ne ded as reflected In the official emergoncy ration which allows for 2106 Koal. A more sustainable diet should Include between 2866 and 2660 gcal. . Unit costs per kg of listed food staples In the NSA shops and In the perallel market are found In Annexes 2A and 28 respectively. SOURCE: Mission estimates 2.20 Based on the analysis in Table 2.5, it appears that a typical urban household without access to a garden plot for self-provisioning (about 702 of households) would have to spend about Mt. 24,000 per month on food to meet nutritional requirements (even excluding the vegetable purchases needed to avoid micronutrient deficiencies 11/). When compared with the prevailing urban minimum wage of Mt 17,000 per month, this suggests that households require at least two formal wage earners to ensure adequate household food supplies and meet costs of cooking fuel (about Mt 11/ Green leafy vegetables and other items (such as fruit) are also necessary for an adequate diet. The cost of these items is not included in the table due to difficulties of estimating prices accurately for the range of fruits and vegetables that might be consumed. In addition, many households in the periphery of urban areas are able to grow leafy vegetables in the area surrounding their homes and may receive them as gifts from rural relatives. If all purchased at market prices, inclusion of vegetables and fruit would probably imply expenditure of an additional Mt 3,000-4,000 per family per month.  - 43 - 4,000-6,000 per month), transport and housing costs 12/ and any fees for schooling and health services. Alternatively, one formal wage earner, some access to informal income or remittances and possession of a garden plot (machamba) 13/ could allow for a precarious food security. Indeed, in view of recent cutbacks in public employment, a sizable proportion of urban households probably have only one formal wage earner and, in some instances, none at all. 2.21 At present few data are available on aggregate urban employment. According to a recent Government study (Green 1989), it is estimated that nationally there were some 580.000 formal sector jobs as of mid-1988, of which some 400,000 were urban. These estimates are broadly confirmed by a 1988 Ministry of Labor (MOL) survey of employment conditions in Maputo which indicates that there are about 200,000 regular jobs available in the city, which contains about half the country's urban population. Given a total urban population of about 2.3 million, the estimate of 400,000 urban jobs would mean that approximately one in 5.8 urban dwellers has regular wage employment - an insufficient number to allow for two wage earners per household. In fact, the MOL Maputo survey found that of a total of 567,000 persons of working-age, only about 260,700 or 45.92 claimed to have any kind of employment (formal or informal), including about 61,000 engaged in agriculture, fishing or petty trade. The remainder were seeking but unable to find gainful work, or were retired (8,700) or physically incapacitated (5,500). A particularly striking finding was that among female-headed households (21.72 of total households), 51.22 had no regularly employed member. This suggests that female-headed households as well as those with disabled or aged adults constitute a particularly vulnerable group from the standpoint of food security. 2.22 Analysis of the 1988 MOC/MOH Maputo household survey indicates that while food insecurity is probably widesp-ead in urban areas at present, some households are likely to be much worse off in both relative and absolute terms 14/. While there was probably some undercounting both of household income and of food purchases in conducting the Maputo survey, the data nonetheless indicate a clear pattern of steadily declining food purchases with decreasing household per capita income. Moreover, those at the lowest per capita income levels appear to be facing serious food 12/ For families living in self-constructed "canicol houses rents are usually not applicable, but transport costs are probably high given that canico areas are located at a distance from most workplaces. By contrast, families living in close-in locations probably can walk to work but generally owe rents (Green, Feb 1989). 13/ The 1988 MOC/MOH Hcusehold Nutrition and Expenditure Survey found a high correlation between access to a machamba and adequate nutritional status. 14/ It should be noted that given some methodological problems in the conduct of the 1988 Maputo Household Survey, its results should be (and have been in this analysis) interpreted with caution.  - 44 - deficits relative to estimated requirements, even after allowing for undercounting of food purchases. 2.23 The graph in figure 2.3 below shows survey findings with regard to reported household food consumption (i.e., quantities acquired through purchase, transfer, donations and self-provisioning). For purposes of the analysis, surveyed households were categorized into six groups by monthly per capita income as follows: Classes A B C D E F Income Per/Capit(Mt)<2u00 2000-2999 3000-3999 4000-6999 7000-9999 10000< Class average (Mt) 1628 2529 3079 4076 6390 17818 Class midpoint (Mt) 1000 2500 3500 5500 8500 20000 No. in class 118 236 259 365 97 40 Class share 0.11 0.21 0.23 0.33 0.09 0.04 Figure 2.3: Reported Food Consumption by Per Capita Income Group: Maputo 1988 2t- 20 '9 - 26 E IS IB \ 3 22 31 , toS A H00 Cf IHAaehdds (tmtupd by Ericme/pe/mth _mdni monthly cereal requiremDent per capita Note: Per capita food consumption fs calculated on the basis of adult equivalents for an average household using the assumptions given in Table 2.5 above. Actually, average household size differs by income group (see Table 2.7 below) suggesting that the food security situation of the lowest income groups may be worse than that depicted in the graph. SOURCEt MOC/MOH Maputo Household Survey, 1988 ----------------------------------------------------------------__-------  - 45 - 2.24 Overall, as noted earlier, the above figure (based on data from the MOC/MOH 1988 Maputo household survey) probably underestimates household food conrumption. Primarily, this is because bread, a relatively important commodity in urban diets, and meals outside the home were omitted from the questionnaire. In addition, given imperfect recall on the part of those surveyed and a relatively long recall period (one month), there is likely to have been some underreporting of food items acquired and Mt amounts spent 15/. Taking these factors into account, it nonetheless appears that households in the lowest per capita income group and some proportion of those in the second lowest are currently facing serious food deficits in relation to needs. For example, the reported monthly food consumption for the lowest income households equals less than two-thirds of basic cereal requirements, a deficit unlikely to be entirely accounted for by bread purchases or better recall. The reported food consumption of the second lowest income group, amounting to about 80% of requirements, suggest that some households in this latter group are also likely to be facing at least intermittent food insecurity. 2.25 Another indicator of the precarious food security status of many urban households is the high share of food expenditures in their total expenditures. Data from the Maputo survey indicate that average household food expenditures amount to over 60X (61.2X) of total expenditures - an indicator of absolute poverty - and may well reach 70Z (a level associated with "desperate" poverty 16/) or more in the lowest income households. Moreover, survey data indicate that cooking fuel and water expenditures respectively amount to an additional 17Z and 2% of total expenditures, indicating that between 80Z and 901 of the total household expenditures of the lowest income group are devoted to meeting nutritional requirements. Table 2.6 below shows the average shares of various expenditure items for households included in the survey. The figures in the *numbers" column indicate the number of households of a total 1,115 households surveyed reporting any expenditures on the specified item. 15/ Also should be noted that findings on purchase/expenditures on potatoes, sweet potatoes, pasta and meat which were included in the questionnaire are not included in the data aggregated for the graphs -- need to get figures while in Maputo. 16/ Data from the Maputo survey allows for the construction of household budgets by income group which would provide a better understanding of the constraints faced by the lowest income households. The Food Security Department intends to carry out this analysis.  - 46 - Table 2.6: Household Expenditures: Haputo City 19e8 Items 2 No. Item Z No. Food 61.2 1115 Transport 2.9 NA Rent 2.1 367 Water 1.9 598 Energy; (17.3) Electricity 3.1 332 Coal 3.3 278 School 0.8 838 Wood 11.6 665 Health 1.3 531 Gas 0.6 207 Other 9.4 NA Petrol 1.8 679 SOURCE: MOC/MOH Maputo Household Survey, August 1988 -----------------------------------------------------------------__-------- 2.26 As the preceding discussion indicates, findings of the Maputo survey do support the view that households in the lowest per capita income group (and probably a limited nurmber of those in the second lowest) are unacceptably vulnerable to the risk of food insecurity and consequent malnutrition. There is, therefore, a need to try and identify these households as precisely as possible with a view to targeting special assistance to them. In this context, and given the difficulties inherent in attempting to precisely determine household income without complex survey efforts, an attempt was made to identify more easily measurable household characteristics to serve as proxies for income. The cross- tabulation in Table 2.7, based on the Maputo survey data, presents the relationship between the previously specified per capita income groups and some measurable household characteristics.  - 47 - Table 2.7: Characterization of Vulnerable Households in Maputo City Hhold per Total Total Total Education Type capita Hhold Displ. Adults of Hhold of Income Size in Hhold Unempl. Head (yrs House (Mt) (No) (No) (No) of form sch) --------------------------------------------------------------------__- < 2000 Mean 9.58 0.74 0.56 Mode 10 0 0 0 canico 2000-2999 Mean 8.94 0.38 0.51 Mode 9 0 0 4 brick 3000-3999 Mean 8.05 0.39 0.52 Mode 7 0 0 4 brick 4000-6999 Mean 7.41 0.19 0.33 Mode 6 0 0 4 brick 7000-10000 Mean 6.37 0.34 0.25 Mode 7 0 0 4 brick 10000< Mean 5.45 0.10 0.18 Mode 4 0 0 4 brick SOURCE: MOC/MOH 1988 Maputo Household Survey 2.27 The data in Table 2.7 above suggest that it is likely to be possible to identify households in the lowest per capita income group by means of some of the variables examined. A number of the variables examined show a clear correlation with per capita income level, suggesting that they could be used to identify the most vulnerable households. It should be noted, however, that, singly, none of these household characteristics is likely to prove sufficient to identify households most in need of assistance. Presence of three or four characteristics could, however, identify them with reasonable accuracy. Other measures of household food insecurity and malnutrition include characteristics related to the household's living environment such as density per room, access to safe drinking water, presence/absence of latrines etc. Given the close relationship of these factors to morbidity, they are discussed in the section on health and nutrition below.  - 48 - C. Morbidity and Malnutrition/Food Security 2.28 As discussed in Chapter I, physical fitness and health have a key role to play in assuring adequate utilization of household food supplies by individual members, achieving satisfactory nutritional status and so producing or earning income for food to maintain/improve household food security. Investments in human capital by improving the nutritional status and health of the population have a large -- though difficult to quantify - - impact on labor productivity and economic growth. In the Mozambican context, given extremely poor health conditions, it would appear that the impact of infection/illness on nutritional status and, hence, on household food security deserves special attention. Today, life expectancy at birth - a key indicator of health status - is only about 48 years in Mozambique, well below the average for low income countries (62 years). Estimates of deaths of children under 5 years range from about 325 to 375 per thousand live births 17/, based on fragmentary evidence taking into account the impact of the war (UNICEF 1989). Although data on causes of illness/death are limited due to the difficulties of establishing regular flows of information in the current security crisis, available information clearly indicates that the principal causes of morbidity and mortality are infectious and parasitic diseases. For all age groups combined, malaria, gastrointestinal illnesses, pneumonia, measles, neo-natal tetanus and tuberculosis are the most important diseases and account for over 75Z of non-MCH visits to health facilities. In addition, a review of available hospital admissions records indicates that trauma is frequently the leading cause of hospitalization - not surprisingly in view of the ongoing security crisis. In general, however, the gamut of medical problems arising from the prolonged bandit activities in Mozambique, ranging from civilian injuries during bandit attacks to the mental trauma suffered by children witness to or involved in atrocities have probably been underreported and catered for by a health system designed for times of peace. 2.29 While the impact on household food security of disability resulting from war-related or other trauma is relatively clear - the work capacity of disabled persons is necessarily limited - the relationship between infection and nutritional status is more complex and requires some elaboration. In general, it is believed that infection or disease affects nutritional status in three ways, depending on illness: (a) by inducing anorexia (or a reduced desire for food) in an infected person - although fc i requirements may actually increase during illness; (b) by contributing to malabsorption of nutrients in food taken; and (c) by causing physical debilitation over a relatively long period and thus preventing adult sufferers from working to provide food. In addition. malnutrition is itself associated with susceptibility to illness, implying an interactive relationship between disease and nutritional status. A recent review 17/ These high under-five mortality figures suggest that the overall nutritional situation in Mozambique is more serious than that indicated by surveys and growth faltering data on survivors -- particularly as malnutrition is a key factor in many child deaths.  - 49 - article on the interaction of nutrition and infection 18, concludes that prospective studies of growth and morbidity in children have identified certain infections as particularly important as causes of poor growth. In particular, diarrhoea, measles, some respiratory illnesses, malaria and intestinal parasites (all very widespread in Mozambique) seem to be closely associated with malnutrition - either failure to gain weight/height or development of micronutrient deficiencies (usually iron or vitamin A). In adults, debilitating infections such as schistosomiasis and malaria, both endemic in parts of Mozambique and common elsewhere, have been shown to cause entire households to become food insecure and hence, nutritionally vulnerable as economically active members are prevented by disease from working. 2.30 The high levels of mortality and morbidity in Mozambique, and, in particular, the common occurrence of many infectious and parasitic diseases associated with malnutrition, suggest that households inadequately covered by basic health services or those living in unhygienic environments are likely to be extraordinarily vulnerable to nutritional problems and food insecurity. This is supported by findings of the 1988 MOC/MOH Maputo household survey which show strong correlations between poor sanitary conditions and the incidence of malnutrition or insufficient growth. Table 2.8 below shows the relationship between some measures of household living conditions and the incidence of malnutrition in children under 3 years. 18/ Andrew Tomkins and Fiona Watson, Interaction of Nutrition and Infection: A Review, March 1988  - 50 - Table 2.8: Nutritional Status of Children c 3 Yrs by Livina Environment Low Height Low Weight Low Weight Growth for Age for Age for Height Faltering (-) (z) (X) (X) 1. Tvpe of Dwelling Canico 34.9 11.9 2.7 13.4 Wood/Zinc 35.3 11.9 2.1 10.2 Brick 31.2 9.9 2.9 15.2 2. Sanitation with Latrine 23.0 22.0 12.0 18.0 without Latrine 77.0 78.0 88.0 82.0 3. Density/Room <2 27.2 9.4 2.0 14.7 2-3 32.9 10.0 3.3 12.4 3-4 33.3 11.6 3.4 16.0 4-5 36.7 11.7 1.2 15.3 >5 38.6 12.8 4.0 11.8 4. Water Facilities Fountain inside 30.0 7.9 2.0 13.3 Fountain outside 37.6 12.3 2.7 12.9 No Fountain 35.9 13.2 2.9 20.0 SOURCE: MOC/MOH Maputo Survey 1988 -----------------------------------------------------------------__------- 2.31 The data above show the strong correlation between non-food factors and nutritional status. In general, however, the relationship between the availability of sanitary facilities/safe water sources and nutritional status is strongest, consistent with the hypothesis that water- borne infections ard parasitic diseases have an important negative impact on nutritional status in Mozambique. Indeed, even taking into account likely income effects, the finding that children living in homes without latrines run a seven times greater risk of acute malnutrition (low weight for height) indicates an unusually close relationship between hygienic conditions precipitating infection and disease, and malnutrition. While less dramatic, relationships between crowded living conditions (density per room) and malnutrition also suggest that the respiratory diseases that they are associated with may be important determinants of nutritional status. 2.32 In summary, the analysis in this chapter demonstrates clearly that household level malnutrition/food insecurity is a problem of enormous proportions in Mozambique today - affecting about half of all households in some degree and severely affecting at least 602 of these in both rural and urban areas. The factors contributing to it are various, including exogenous factors such as climatic disasters, political problems manifested  - 51 - in the continuing security crisis, past failures in economic policy and the inevitable shocks associated with necessary adjustments, difficult conditions for agriculture and logistics, and generally poor health conditions. To combat malnutrition/food insecurity in these circumstances will require a multi-pronged approach aimed both at alleviating identified constraints where this is possible, and, where it is not, developing or strengthening strategies for coping better within them. The remaining chapters of this report propose a number of options for reducing household food insecurity in Mozambique ranging from short and medium term measures to improve efficiency and equity in food distribution, allocation and access, to putting in place the appropriate agricultural growth and employment strategies for sustained food security in the longer term.  - 53 - III. FOOD AID A. Introduction 3.01 Food aid is fulfilling two important food security functions in Mozambique. It is substituting for market flows of domestically produced food and commercial imports in the face of foreign exchange scarcities and disrupted domestic food markets. It is therefore preventing price rises which would otherwise imperil the food security of poor (and probably most) households who depend on markets for food. It also supplements Government transfers to households - by counterpart fund contributions to Government expenditure from the sale of marketed food aid and in the form of free food distributed through the emergency program. Food aid is a fungible resource - interchangeable by trade for money. Its significance, therefore, does not lie so much in the physical nature of the commodities supplied by food aid donors, but in their value to the economy as a whole. The challenge presented by food aid in Mozambique is to expand its role from the temporary relief of hunger to become a source of economic growth which raises the incomes of the most vulnerable in a self-sustaining way. The basic strategy is to concentrate resources as much as possible on creating employment and income generating activities in both urban and rural areas, and on restoring domestic agricultural production particularly in the family sector. In addition, these efforts need to be complemented by improvements in access to basic public services such as health, nutrition, clean water and sanitation, and by a "safety-net" of direct and indirect income transfers, where necessary, to prevent serious malnutrition or even starvation. This strategy is explored in more detail in Chapters IV to VII. In this Chapter, the main thrust is to explore ways to improve the resource transfer efficiency of food aid, with particular focus on an evaluation of triangular food aid transactions, and to improve the management of the revenue from food aid sales. 3.02 Mozambique's dependence on food aid has increased dramatically over the last few years as shown in Figure 1.2. It is the task of the FSD in the MOC to match Mozambique's estimated requirements in terms of uses (emergency, market support, Food Bank, projects) and commodities with food aid pledges which reflect donor capacities and interests. This is a formidable task. It involves scheduling imports of 9 or so different commodities for various uses from 16 main donor sources through 18 points of entry to multiple destinations in three regions. Table 3.1 shows how total food aid pledges were made up by contributions from donors in 1988- 89. 1 The United States (US) and the European Economic Community (EEC), followed by Canada, Australia and WFP, together contributed more than 70Z 1/ The data in the table come from the data base on food aid comp4.led by the Food Security Department. The overall volume pledged (710,000 tons) exceeds the volume showr. in Table 1.6 (665,000 tons) because the former includes unmet pledges from the previous year which have been carried over into 1988/89 while the latter includes only new pledges for 1988/89 made during the April 1988 appeal.  - 54 - of the total. The table also shows the kind of bias that might be expected between donor and commodity offered. Bilateral donors generally supply commodities which are surplus to domestic demand. Thus, the US was the major supplier of yellow maize, followed by the EEC; Italy, Australia and the EEC of rice; and Canada, the US, the EEC and Italy of wheat. Clearly, WFP's contributions do not reflect such a domestic production pattern. More recently, however, most bilateral donors have also financed 'triangular food aid - i.e., food aid purchased in a developing country. The most important *triangular' commodity in the case of food aid for Mozambique is white maize. All of the major food aid donors with the exception of Japan bought white maize for Mozambique in 1988-89. The characteristics of these triangular contributions are analyzed in more detail in the following section. Table 3.1s Food Aid 1988-89: Pledges by Donor and Commodity M A I Z E rin m8BEAM FM DSI YELLOW WH1TE RICE WHAT SUCA OIL TOTAL 51A03 LANED S (to Fob )PL BTMALJA Qt 720 21600 10861 13040 84041 OS 80407 0.56 CANADA Q* O0s 17740 89000 62815 95 85443 0.56 Mc Q2t m0 40000 32020 16667 40000 1000 2000 187187 191 117597 0.66 PM0 9% 124 1X 1 27a 17820 2000 60 410 21698 3 12as 0 s5 ITALY 9t, 823 14787 12000 10000 62s8 43407 6e 26663 0.62 JAPAN ty 012 l6o00 16812 26 240 0.01 NEMI.ANDS Q9 950 165 26700 1400 98 29318 4S 13114 0.52 M 9, 21844 21344 S 9106 0.43 us Q% 268 1s30 1100 25000 500 9425 216511 30S 16580 0.78 WFP ft 70 11Q 0 1097 126I0 17816 4698 5241 46317 71 54248 0.74 OTHERS Qty a 16 5100 827 2915 7X00 1050 221? 61143 U 4113 0.67 TOTALS 11y 49707 2179 140 283130 200413 60642 121000 24483 17391 710505 1001 490624 0.69 Np+tmr 9%W * buautitW (m't'ic tons). Food aid year is -from Ma to April WwRCE. PSO0NC 3.03 Table 3.2 shows the way food aid commodities were allocated in 1988-89 between two main and two subsidiary purposes: market supply support, emergency distribution to the people affected by war and drought, specific projects and, more recently, contributions to a 'food bank" which will be described in more detail in Chapter V. In 1988-89, 362 of food aid by volume was distributed under the emergency program and 582 was used to augment market supply. The former was distributed free, and the latter was sold. Revenues from these sales were intended to generate "counterpart fundsO (see Section C). Only minor quantities were used for specific developient projects 2/ and as contributions to the newly established food bank. Although the overall amount pledged compares fairly well with the Government's requests, the amounts of oil and sugar, in particular, fell short of requirements. While cereals tend to dominate all international 21 It is possible that the allocations to project support are under- estimated. More mAy be included within 'emergency' allocations.  - 55 - food aid donations, beans and oil transfer value as well as nutrients more efficiently than cereals, because transport costs comprise a lower share of total costs. There would therefore be advantages to increasing these hIgh value items in the food aid pledges and deliveries. Table 3.2: Commodity Allocations by Use 1988-89 Beans Fish DSM Maiz RicO Wheat Suar Oil Totals Emergency aid Qty 88825 1262 778 196496 6516 12941 253JJ9 Market support Qty 14767 412 e 203966 8942 121069 12439 1699 414816 Projects Qty 1216 798 1097 12499 185 676 16829 Food bank Qty 2860 16884 420o 2907 27971 Totals Qty 49707 2279 1640 488648 6042 121999 24488 17391 710635 Share X 7 0.8 0.2 el 9 17 a 2 1i9 Value I "NO 28866 5769 2087 60184 10494 17489 6186 14297 138208 Unit value 3 $/ton 479 2631 1366 116 396 144 253 622 Note: Qty * pledged quantities (metric tonnes) SOURCE: FSD, MOC 3.04 The comML'ity pattern shown in Table 3.2 indicates that the emergency program received proportionately more beans, fish, dried skimmed milk and oil. Among cereals p-oportionately more wheat and rice - the preferred cereals in Mozambique - were allocated to market support, with maize roughly evenly divided between the emergency program and markets. This distribution seems to be sensible, but it is not clear whether the overall commodity allocation pattern is driven more by needs or the availability of exportable surpluses within donor countries and donor preferences for different parts of the food aid program. For example, because WFP is not permitted to sell food, it did not contribute to market support at all. In contrast, the EEC, Italy, and the US have a preference for market support while the smaller donors, aggregated in the tables as "other donors', have a preference for emergency contributions. 3.05 Taking the donor community as a whole, preferences and regulations seem to have been complementary enough for Mozambique to fulfill its needs for food aid fairly satisfactorily. In the past, Mozambique's dependence on multiple donor sources for food aid typically resulted in cycles of gluts and scarcities. Although donor coordination has improved considerably, 4/ Table 3.1 shows that only 69Z of the food aid pledged for 1988-89 had arrived by the middle of the 10th month of the food aid year (April - March). Six donors had failed to deliver 60Z of their pledge. It mignt be argued that additional imports could not have been stored or transported. This may be true, but limited storage and transport capacity can only be used efficiently if import scheduling is precise, 3/ The unit values for each commodity were derived by dividing the value of shipments provided by donors by their weight. 4/ There is a weekly meeting to discuss coordination of food aid schedules attended by the principal Ministries, DPCCN, UNDP and the major food aid donors and NGOs.  - 56 - well-coordinated between donors and known to the Government in advance. Overall, however, in contrast to previous years, Mozambique's recent appeals to the donor community for food aid seem to have been so successful that the constraints limiting the availability of food throughout the country are logistical rather than supply-related. Transportation problems created ',y the war and scarcities of trucks, parts and fuel are severe and form ar underlying theme limiting the implementation of a satisfactory food securiliy strategy. The management of internal food aid distribution and recomm,'ndations for improvement are discussed more fully in Chapters I and VI. B. Triangular Food Aid 3.06 The relative merits of triangular food aid are a matter of some debate both within Mozambique and outside. The proponents of triangular food aid transactions (TFATs) allege, on the side of the donor, increased efficiency in resource transfers as a result of cheaper commodities and reduced transport costs and, for the recipient, improved timeliness in food aid deliveries and access to commodities which are preferred by consumers. They also point to potential advantages to TFAT suppliers in terms of additional access to regional markets for agricultural exports; some donors have expectations of a supply response as a result of the increased demand for food produced in developing countries. The counter view is that TFATs are an inefficient use of donor funds because commodities and transport are no cheaper than they would be if supplied from donors' stocks or procured in international markets; that quality is uncertain and logistical problems impose undue administrative costs; that sources are unreliable and deliveries are prone to delays; that a supply response cannot be expected in supplying countries and that donor trading in oligopolistic regional markets allows parastatal trader. to bid up their export prices which in turn can lead to misallocations within the supplying economies. These issues are important to Mozambique as a recipient of triangular food aid and to Southern Africa as a whole, as food aid donors are now financing more than 40Z of the cereal trade amongst Southern Africa Development Coordinating Conference (SADCC) countries. Table 3.8: Triangular Food Aid Import. 1988-89 by Commodity P L E 0 0 E D L A N D E D quant;ty Share of Value Valuo Unit Landed Landed/ (tonnes) total (8'000) *hare value (Feb 1989) pledged) food aid (1/tonne) Beans s00 0.001 270 0.002 450 s0e 1.00 Sugar 1200 0.002 300 0.002 260 1044 8.87 oil 515 0.001 412 0.003 800 465 0.88 White 200413 0.282 23619 0.170 117 183294 0.87 maize of which from EASA 146497 0.206 17405 0.128 120 89880 8.82 SOURCE: FSD, MOC