Report No. 44072-NE Niger Food Security and Safety Nets February 18, 2009 Human Development, AFTH2 Country Department AFCF2 Africa Region Document of the World Bank QUIBB Questionnaire sur les Indicateurs de Base et de Bien-&tre UEMOA UnionEconomique et Mone'taire OuestAfricaine UNICEF UnitedNations Children's Fund UNDP UnitedNations Development Program USAID UnitedStates Agency for International Development FEWS NET USAID-Famine Early Warning SystemNetwork WAEMU West African Economic and Monetary Union WFD World Food Program WHO World Health Organization Vice President: Obiageli KatrynEzekwesili Country Director Madani M.Tall Sector Director Yaw Ansu Sector Manager Eva Jarawan Task Team Leader Setareh Razmard Carlo Del Ninno ii ACKNOWLEDGMENTS This report was preparedby a team consisting of Jenny C. Aker (Consultant), Paul Dorosh (FEU), Menno Mulder-Sibanda (AFTH2), Carlo Del Ninno (HDNSP, co-task team leader), and Setareh Razmara (AFTH2, co-task team leader). Valuable inputs were provided by Phillippe George Leite (HDNSP), Koffi Akakpo (VAM Officer WFP - Niger), and Khamada Baye (Local Consultant). The report draws from two background papers that provided substantial inputs:(i) Food Security, Social Safety Nets and Food Crises: Country Experience from Niger (Jenny C. Aker, Consultant); and (ii) Production Shocks, Regional Trade and Cereal Prices in Niger: A (Multi-Market) Simulation Analysis (Paul Dorosh). Valuable guidance and suggestions were provided by Ousman Diagana (Country Manager for Niger); Patrick Verissimo, El Hadj Adama Toure, and Amadou Alassane (AFTAR); Xiao Ye (AFRCE); Richard Seifman and Leandre Bassole (AFTHV); and Amadou Ibrahim (AFTP4). Peer reviewers were Dominique van de Walle (PRMGE) and Harold Alderman (AFTHD). Kathryn Bach provided valuable editorial inputs. Administrative support was provided by Aissatou Chipkaou (AFTH2) and Mohamed Diaw (AFTH2). The preparation of this study also benefitedfrom Trust Fundsmanaged by HDNSP. Valuable support from the Nigerien counterpart team in the Prime Ministry Office (Primature), Executive Secretariat of Rural Development (SE/RDS), Ministryof Economy and Finance (MEF), Office des Produits Vivriers du Niger (OPVN), Cellule de Crise Alimentaire (CCA), Systeme d'Alerte Precoce (SAP), and Institut national de la statistique (INS), are gratefully acknowledged. Particularly the team would like to thank M. Malla Ari, Ex-Director of Cabinet of Prime Minister and Director of the SRPII, M.Ousmane Mahaman, Director of Cabinet of Prime Minister, Mme Zeinabou Maikorema, Executive Secretary of Rural Development Strategy, M. Azara Sully, Ex-Senior Advisor to Prime MinisterCabinet, and M.DiamoitouBoukari, Rural Development Senior Advisor to Prime Minister Cabinet for supporting this study at early stage and facilitating its preparation. Preliminary results were presented through a workshop in Niger (December 2007) to the Government officials as well Development Partners and researchers, and valuable comments were received from participants. In October 2008, the draft report was discussed with the authorities and Technical and Financial Partners (TFPs). There was consensus on the recommendations and strategic messages of the report, and with close collaboration of the GoN and the TFPs a concrete action planfor an effective food security and social protection strategy (with short and medium term measures) was finalized. The team also received valuable advice and supports from, the TFPs (particularly EU, GTZ, WFP, UNICEF, FAO, FMI, etc.). Niger Food Security and Safety Nets Table of Contents EXECUTIVE SUMMARY ........................................................................................................... 4 CHAPTER 1:INTRODUCTION ............................................................................................... 18 BACKGROUND AND MOTIVATION .................................................................................... 18 STRUCTUREOF THE REPORT .............................................................................................. 19 CHAPTER 2: POVERTY AND FOODINSECURITY ........................................................... CHRONIC AND SEASONAL FOODINSECURITY.............................................................. 21 RISK, VULNERABILITY, AND HOUSEHOLDFOOD SECURITY..................................... 22 29 CHAPTER3: FOODAVAILABILITY AND ACCESS .......................................................... .............................. 39 AGRO-FOODMARKETS AND PRICES ................................................................................ FOODAVAILABILITY: PRODUCTION, IMPORTSAND FOODAID 39 47 CHAPTER 4: GOVERNMENT INSTITUTIONS AND PROGRAMSAND RESPONSE TO THE 2005 FOOD CRISIS ........................................................................................................... INSTITUTIONAL MECHANISMS FOR RESPONSE TO FOOD CRISIS ............................. 53 CAUSES OF THE 2005 FOODCRISIS.................................................................................... 53 57 GOVERNMENT SAFETY NETAND SOCIAL PROTECTIONPROGRAMS...................... RESPONSES DURING THE FOODCRISIS ........................................................................... 64 66 NIGER 2008: PREPARINGFOR AND RESPONDINGTO A POTENTIAL FOODCRISIS69 CHAPTER 5: SUMMARY AND POLICY OPTIONS ............................................................ 73 OUTLINE OF A FOOD SECURITYAND SOCIAL PROTECTIONSTRATEGY................73 SUMMARY OF FINDINGS...................................................................................................... 75 REFERENCES ............................................................................................................................. 81 ANNEX I:DESIGNOF EARLY WARNING SYSTEMS ....................................................... 86 TABLES Table 1: Socio-economic Indicators 2005 Table 2: Poverty Status inNiger by Geographic Area Table 3: Sources and level of household food consumption in2005 and 2006 Table 4: Prevalenceof Chronic Food Insecurity inNiger Table 5: Prevalenceof Undernutrition inChildren under 5 inNiger. 2006 Table 6: Percentageof Households' Exposedto Shocks by Poverty level and Quintile Table 7: Distributionof householdsExposedto Drought across Regions of Residence Table 8: Households Coping Strategies after Exposure to Drought Table 9: Percentage of drought affected households receiving food aid in 2005 by poverty category and location . ii Table 10: Coping mechanism after drought by ownership of livestock in rural and urban areas Table 11: Household food insecurity levels Table 12: Chronic, transient and vulnerable levels of food insecurity Table 13: Dietary diversification index Table 14: Milletand Sorghum Production inNiger, 1985-2004 Table 15: Official Cereal Imports for Niger, 1998-2004 Table 16: Niger: Cereal Production and Availability, 1990-2005 Table 17: Cereal Production inNiger andNigeria, Average 2003-2004 Table 18: Average GrainPrices inNiger, Nigeria and Chad, 1996-2006 Table 19: Simulated Prices Effects of Productions Shock and Net Imports, 2004/005 Table 20: Grain Prices by Food Crisis Region iNiger, 2004/2005 Table 21:Distributionof Safety net programs between2001and 2006 Table 22: Source of financing of Safety nets 2001-2006 Table 23: Governmental (Emergency) Safety Net Programs, by Year and by Program BOXES Box 1: Howto measure Food Security? Box 2: Steps for Identifyingthe Most Vulnerable inNiger Box 3: Methodologies to estimate household food insecurity Box 4: National Mechanism for the Prevention andmanagementof Food Crises inNiger Box 5: History of Food Crises inNiger Box 6: Ethiopia's Productive Safety Net Program (PSNP) Box 7: Enhancing Food Security inBangladesh FIGURES Figure 1: Poverty Rates inNiger Department, 2005 Figure 2: Share of Per Capita Caloric Consumptionby Commodity Group (2005-2006) Figure 3: Stunting in Sub-SaharanAfrica Figure 4: Number of Cases of ReportedWasting (by Week), 2006 and 2007 Figure 5: Percentage of DepartmentsAffected by ProductionShocks by Year 2000-2006 Figure 6: Percentage ofthe PopulationVulnerable to FoodInsecurity, 2007 Figure 7: Predictions of Areas Vulnerable to Food Insecurity inNiger, 2006 Figure 8: Administrative regionsand agro-ecological Figure 9: Key Grain Markets inNiger Figure 10: Grain Prices inNiger and key Regional Markets, 1996-2006 Figure 11: Key ForecastingMarkets for Niger Figure 12: ComparisonofMillet Prices inNiger and Nigeria Figure 13: Quantities of FoodAid Obtained and Distributedby OPVN and CCA during 2001-2007 Figure 14: Grain Price Levels of Key ForecastingMarkets during the Harvest Period. Figure 15: Intra-SeasonalChange of Millet Prices by Year Figure 16: Distribution of Children with an Infectious Illness and Acute Under-nutrition by Age Group Figure 17: InternationalPrices of Grains (Wheat, Rice and Maize), 2000-2007 Figure 18: MilletPriceson Key ForecastingMarkets during the Harvest Period in200,2004,2007 Figure 19: Changes in Millet Prices betweenthe Harvest and Hungry Periods iii EXECUTIVESUMMARY Niger is a very poor country thatfaces serious problems of poverty and householdfood insecurity. With a per capita Gross National Income (GNI) of US$240 and an estimated 62 percent of the population living below the poverty line, Niger is one of the lowest-ranked countries on the United Nations' Human Development index. The incidence of poverty has remained constant since the early 1990s (63% in 1993), mainly due to lack of resources, low agricultural productivity and high population growth. Currently approximately 85 percent of the population lives on less than US$2 per day, with 65 percent living on less than US$1 per day. Although the magnitude of poverty and food insecurity is most evident during years of drought, the problem is in fact a chronic one. Low levels of food availability, high population growth, limited economic access to food because of low incomes relative to market prices, and inadequatehealth care have resultedinpoor nutritional status for much of the population, even in years of normal harvest. Reducing vulnerability and ensuring food and nutrition security is an overarching priority for the Government. Maintaining food security at the national and household level is an important priority for developing countries in general, both for the welfare of the poor and for political stability. Inorder to ensure food security, governments have adoptedvarious strategies, including efforts to increase staple food crop production (often with the explicit goal of food self-sufficiency), market interventions, and a variety of safety net programs, especially during emergencies. InNiger, where profound vulnerabilities combined with a high level of population growth (3.3 percent per year) have resulted in endemic food insecurity, the Government is faced with a serious challenge. Food aid has served as an important resource for the Government and i s considered to be integral to the provision of safety net interventions, especially emergency response. In this framework, since 1998, Niger's government and major food aid donors have managed a DispositifNational de Pre'vention et de Gestion des Crises Alimentaires (DNPGCA), a National Mechanismfor the Prevention and Management of Food Crises inNiger mandatedto (i)helpthegovernmenttobuildcerealandfinancialreservesforfoodaidduringcrises;and(ii) implementsupport actions for populations during periods of food crisis. Moreover, in its Rural Development Strategy (RDS) as well as the second Poverty Reduction Strategy (PRSP2), the Government of Niger (GoN) has established the strategic vision, policies and institutional framework to address food security issues. Particularly the RDS, which is the national agriculture policy, has for objectives to: (i) secure the living conditions of the population through prevention of food crisis; (ii)improve the revenues of rural population through access to economic opportunities; (iii) strengthencapacitiespublic institutions and agriculture professional organizations; and (iv) provide support to increase food production. I n this context, thepurpose of this study is to contribute to the existing strategy and assist the Government in developing a holistic, multi-sectoral, and institutional approach to reducing the population's vulnerability to food insecurity. A comprehensive food security strategy needs to address ways to enhance food security through increases in agricultural productivity, as well as through policies related to education, health & nutrition and population growth that are crucial for diversifying and increasing household incomes (and thereby access to food). Since safety nets are also essential to respond to chronic food insecurity, the report discusses the food security strategy in the context of the social protection interventions, while complementing other existing and ongoing analytical works (Le., on irrigation, agriculture, 4 population, health). Fromthe perspective o f social protection, this study is designedto synthesize considerable existing analysis, review food and nutrition security policies and programs inNiger, and provide an action plan for strengthening the existing system and developing an effective food security and safety net strategy, in the context o f the Government's poverty reduction strategy. This report adds value to the ongoing policy discussions intwo ways: First, it presents new empirical analysis o f i)food insecurity and vulnerability o f households during the period o f food crises as well as during normal period, ii)the structure and integration o f cereal markets within Niger and with markets in neighboring countries, and iii)causes o f the 2005 food crisis, and lessons learned on implications o f various levels o f cross-border flows betweenNiger and Nigeria. Second, it provides concrete short- and medium-term recommendations for helping government to improve the performance o f existing programs to increase food security, particularly related to preparedness for and responses to food crises, and to design efficient safety nets mechanisms for vulnerable population. The policy options provided in this report are aligned with the strategies proposed in the SDRP2 and RDS and complement the Government national contingence plan(le Plan National de Contingence sur la Securite Alimentaire) (November 2007). They are also aimed to help Government respond to the recent and future high food world prices, which may contribute to higher domestic prices and reduce availability o f food aid to supplement Niger's domestic food supply. What are mainfindings of the study?First, the data suggest that more than 50 percent o fthe population suffers from some form o f food insecurity, with 22 percent o f the population chronically extremely food insecure. Second, poor households are more exposed to shocks, the poorest regions are the most vulnerable to food insecurity, and coping mechanisms often increase the vulnerability o f households to future food insecurity. Third, since two thirds o f daily caloric consumption comes from cereals, cereal production, availability, and market performance are crucial determinants o f food security. Fourth, evidence suggests that the targeting o f food aid has been either non-existent or has focused on helping all those affected by drought rather than only poor households affected by drought. Fifth, although the need to support poor and food insecure households i s substantial, safety net programs are small, receive limited government funding, and are designed for emergency food crises. Finally, faced with limited resources, a high prevalence o f poverty and periodic severe droughts and other shocks, as mentioned in the RDS (program 9), an effective food security and social protection strategy is essential for Niger. This strategy needs to focus on: (i)improving the efficiency and scope o f safety net programs; (ii) promoting effective medium-term strategies and investments to improve food availability, access and utilization; and (iii)improving emergency responses and the information system. POVERTY FOOD AND INSECURITY Niger is classifiedas one of thepoorest countries in the world with substantialpoverty variation at the regional level. According to 2005 household survey on basic indicators (QUIBB),about 62 percent o f the population has an income that places them below the poverty line. Since the economy i s mainly centered around subsistence crops and livestock, and given the highlevels o fpopulation growth (3.3 percent pea.)and modest growth inreal GDP, the incidence o f poverty has remained constant since 1993. As inmany other countries in sub-Saharan Africa, poverty is more prevalent in rural areas (65.7 percent) than in urban areas (55.5 percent). 5 Regional variations of poverty follow this pattern as well: predominantly rural regions such as Maradi (79.7 percent), Tillaberi (68.9 percent) and Dosso (67.3 percent) have the highest incidences of poverty, while in the predominantly urban region of Niamey, the incidence of poverty is only 27.1 percent. Since grains accountfor a high share of total household expenditures and almost all households are net purchasers of grains,fluctuations in grain production and prices have a major impact on household food security. Between 61-64 percent of total household expenditures are for basic foodstuffs, and grains represent about 50 percent of total household spending. Grains (millet and sorghum) are also the dominant sources of caloric consumption, accounting for over 75 percent of food consumption. Moreover, although the majority of households are engaged in agricultural activities, almost all are net purchasers of food and are thus negatively affected by increases infood prices. More than 50 percent of Niger's population is estimated to be chronically food insecure, with 22 percent of the total population suffering from extreme chronic food insecurity (per capita caloric consumption of less than 1800 kcal/person/day). A variety of indicators demonstrate Niger's food insecurity, including low per capita food consumption, the limited dietary diversity of the population, the high prevalence of stunting in children under 5, and high levels of micronutrient deficiencies (primarily iodine, iron and Vitamin A). The causes of food security are complex, however, and result from the interaction between low food availability, poverty and limited economic access, the poor health `and nutritional status of the population, and the country's high vulnerability to shocks. Although poverty rates lower in urban areas, available household data for 2006 suggest that the percentage of the population with inadequatecaloric consumption (less than2100calories per day) is actually higher inurbanareas (58 percent) in all regions, compared to the rural areas (51 percent), with the exception of Dosso. .The 2007 survey also confirms these findings, showing that the 50 per cent of the population which are chronically food insecure are composed of 30 per cent suffering from extreme chronic food insecurity and 20 per cent at risk. There were substantial variations in food insecurity across regions between 2005 and 2006. The high level of extreme chronic food insecurity has affected regions differently inNiger. In2005, Maradi, Dosso andNiamey experiencedthe highest levels of extreme chronic and total food insecurity, with over 48 percent of the population in a state of extreme chronic food insecurity. In 2006, although extreme chronic food insecurity had been reduced, Agadez (with about 44 percent of the population vulnerable to chronic food insecurity) had taken Maradi's place as the region with the highest level of food insecurity in the country because the level of food insecurity fell inMaradi. I n addition to chronic food insecurity, much of Niger's population suffers from seasonal and transitoryfood insecurity. Nutritional data show that much of the rural population suffers from seasonal food insecurity, evidenced by the annual hungry season (soudure) between June and September. Both rural and urban populations also are vulnerable to transitory food insecurity, definedas temporarily reducedfood consumption after a shock. Households havefairly limited coping strategies to deal with shocks to their incomes andfood access, and some coping mechanisms make affected households more vulnerable to future food insecurity. Qualitative surveys report that the most common strategies used by 6 Nigerien households to respond to natural shocks include migration, asset and livestock sales, and borrowing money. In a 2005 household survey, food aid was rated as the most common strategy used by households to mitigate the effects of shocks, including droughts and economic and health shocks. In the case of droughts, almost 1 in 4 households used food aid as the main coping strategy. Severe shocks leave households more vulnerable to future food insecurity: 60 percent of households were still in debt after 2005 food crises (Marinho and Gerard, 2008). In some cases, coping strategies exacerbatedvulnerability. The probability of receiving food aid is not correlated with poverty in rural areas. Among those affected by the drought inrural areas, only 72 percent of the poor received food aid as compared to almost 80 percent of the non poor. In urban areas, however, poor households affected by drought were more likely to receive food aid than non-poor households. These results suggest that the targeting of food aid has been either non existent or it has focused on helping all those affected by drought rather thanonly poor householdsaffected by drought. FOOD PRODUCTION MARKETS AND In light of the importance of staple cereals (millet and sorghum) for producers' and consumers' welfare (over two thirds of daily caloric consumption comes from cereals), the structure, conduct andperformance of food production andmarketing has important implications for food security inthe country. Despite unfavorable agro-climatic conditions, agriculture and livestock-raisingplay a key role in the economy of Niger. Therefore, the prevalence and severity o f the food security problem in Niger are directly related to the structure, conduct and performance of food production and marketing. Overall, agriculture employs more than 80 percent of the total population and contributes to approximately 40 percent of Gross Domestic Product (GDP). Agriculture accounts for about 52 percent of rural output, with livestock contributing another 30 percent. Niger relies mainly on its own domesticproduction of grains (dominated by millet and sorghum) for its cereal supply. Agriculture is primarily characterized by subsistence farming and production of food crop is primarily basedon rainfed agriculture. Over 80 percent of Niger's total cereal availability comes from domestic production, primarily millet and sorghum. Millet is by far the dominant staple food crop (accounting for 76 percent of cereal production), followed by sorghum (accounting for 22 percent of cereal production). These crops are major determinants of bothnational food availability and rural household incomes. Given irregular rainfall, cereal production varies substantially and rural households are exposed to production shocks. Over the past two decades, the lowest cereal production years (1993, 1997, 2000 and 2004) coincided with natural shocks (drought). Although national cereal production increased by 48 percent between 1985 and 2004 (from 1.8 million tons in 1985 to 2.7 million tons), the increasewas primarily due to the expansion of cultivated areas (from 4.3 million hectares to over 7 million hectares). Becauserainfall is highly variable both across space and from year to year, there are substantial spatial and inter-annual variations in cereal production: the average coefficients of variation for national millet and sorghum production are .20 and -35, respectively. There is significant spatial variation in millet production as well: the regions of Maradi and Zinder (which account for approximately 40 percent of national millet 7 production) have the lowest coefficients of variation, while the regions of Tillaberi and Tahoua have the highest, suggesting that rural households in the latter regions are exposed to relatively higher production risk than households in other areas of the country. Inaddition, the magnitude of production shocks varies quite substantially across regions. Given the strong intra-annual variation in staple food crop production, total food availability in Niger depends strongly on commercial imports (specifically supply and demand conditions in Nigeria), whilefood aid does notplay an important role. Data on cereal imports and exports inNiger are highly unreliable, due in part to the large volume of informal trade that occurs between Niger and its neighbors (Benin, Burkina Faso, Chad, Mali and Nigeria), and unofficial net imports of sorghum and millet from Nigeria are likely much larger in most years than recorded official imports. Estimates of these imports vary widely by year, but analysis of annual supply, demand and prices suggests that millet imports may have been as high as 200 thousand tons in2004, equivalent to about 10percent of total net millet supply. Moreover, since production of millet, sorghum and maize are much larger inNigeria than inNiger, the supply and demand conditions in Nigeria, especially in the northern regions, have a major influence on prices and availability in Niger. In comparison with unofficial imports, food aid inflows are small, accounting for less than20 percent of total cereal imports and only about 1percent of total net cerealavailability. The role of the Government of Niger in cereal production and marketing has been significantly reduced since the 1990s andfood prices are now determined by market forces. These reforms have had important implications for the functioning of Niger's agro-food sector, particularly in terms of pricing, marketing, and agricultural development and investment. With the cereal market becoming liberalized, the role of the OfJice des Produits Vivriers uu Niger (OPVN), established in 1984 to manage the purchase and sales of staple cereals (millet and sorghum), is now reduced to monitoring the food security situation within the country and to managing the country's strategic grain reserves. As a result, during 2000 and 2004, OPVN was responsible for distributing subsidized food to the population, including to vulnerable groups, and domestic cereal prices are influenced by a host of factors, such as domestic supply and demand, commercial imports, import taxes and tariffs, regional trade patterns and market structure. However, depending on domestic cereal production ina particular year, the GoN may regulate trade flows by limiting either exports or imports of specific commodities. Although grain markets in Niger are well integrated in general, this varies both temporally and spatially. Staple food crop markets in Niger are somewhat integrated, with an average correlation coefficient of .55 for all markets between 1996 and 2006. Nevertheless, the degree of domestic market integration varies both over time and across space, with higher levels of integration during low-production years. The degree of integration betweenmarkets inNiger and those inborder countries (Benin, Burkina Faso, Chad and Nigeria) follow the same pattern as the domestic markets. Market price integration between Chad and Niger i s minor, however, as i s the degree of integration between Burkina Faso and Niger. The highest degree of integration occurs betweenNiger, Beninand Nigeria, with correlation coefficient o f grain prices averaging 0.65. There is also evidence of strong market integration betweenthree-quarters of the major markets in southern Niger and markets just across the border in Malanville (Benin) and Jibia, Illela and Mai-Adua (Nigeria). 8 THE2005FOOD CRISIS Thefood crisis in Niger in 2005provided important lessons to national policy-makers and the international community. Estimates o f national food production in late 2004 suggested only a modest decline relative to medium-term trends gave little cause for concern. Inone sense, the major cause o f the crisis was the extremely low level of household income, which leaves many households on the brink o f serious malnutrition even in years o f normal harvests and prices. The additional shocks in 2005 that tipped the balance from risk to reality o f a food crisis were: 9 losses infood production and incomes for those farmers whose late 2004 harvests were affected by drought and locusts; and 9 a sharp rise in food prices (related to even more dramatic price increases in some of Nigeria's markets beginning in mid-2004) that reduced household access to food for all net purchasers. Market analysis suggests that Nigeria's production shortfall and high prices likely led to a decline in net imports o f millet that significantly worsened the effects o f Niger's production shortfall. The Government response to the 2005 food crisis included various mechanisms to support households affected by income losses and production shocks. The government response, with support from non-governmental agencies and international organizations, consisted mainly o f a range o f emergency schemes: (i)food aid for direct emergency distribution and "Food For Work" (FFW) programs; (ii) sale o f approximately 40 thousand tons o f cereals at subsidized prices; and (iii)provision o f "Cash-for-Work" (CFW) programs to raise the purchasing power o f affected households. These programs relied upon both imported and local food aid, with some local purchases occurring within Niger (Niamey and Maradi) and in northern Nigeria. In spite o f these efforts, the gross mortality rate reached 1.5 deaths per 10,000 per day in some o f the worst-affected regions, with child mortality rates o f 4.1 deaths per 10,000 per day, levels well above international thresholds for humanitariancrises. Since criteria for targeting vulnerable regions for the implementation of emergency programs were unclear, a list identifying the most severely affected regions during the 2005 food crisis is still not available. Due to the lack of explicit and transparent criteria (or thresholds) for determining a food crisis and identifying vulnerable regions, the areas most affected by the crisis may not have been reached in2005. A list o f vulnerable villages was first produced in May 2005, however the list was not aligned with qualitative and quantitative evidence observed by international and non-governmental organizations. To improve the functioning of its emergency response interventions, the GoN has developed a national contingencyplanfor food security and nutrition with theparticipation of a variety of governmental and non-governmental actors. Developed in 2007, the document outlines the strategies priorities (including early warning indicators, i.e., cereal deficit, food prices and severe malnutrition) for the Government and development partners in preventing and managing food crises in the country. The overall objective o f the plan i s to minimize the impact o f food crises by ensuring households' access to staple foods and protecting their assets, mainly via the national security stock and emergency cash resources (Le., general food distribution, Food for Work (FFW), Cash for Work (CFW), subsidized sales, seeds distributions). However, there is still lack o f agreement regarding the relative effectiveness o f safety net interventions in 9 Niger. The list of response interventions outlined by the national contingency plan includes most of the same instrumentsthat have beenused inthe past few years, with an increased emphasis on CFW rather than FFW and cereal banks. However there i s no good evidence regarding which interventions have been most effective. The data on cereal banks from the 1990s shows that these have not been sustainable. CFW programs, which are usually preferred to FFW, are fairly new interventions that have not beenfully evaluatedinNiger. POLICY OPTIONS TO RESPONDTO FOOD INSECURITY International experience has shown that social safety net programs play a very important role in reducing and attenuating chronic poverty and in helping the poorest people after natural catastrophes. Inthis study, social safety nets (safety nets for short) are non-contributory transfer programs targeted to the poor and to individuals vulnerable to poverty and to shocks. These programs have the simultaneous goals of (i) providing assistance to households in poverty, thus contributing to a reduction in the impact of poverty and to helping them get out of poverty, and also to help households face the impact of shocks; and (ii) reducing the effects of globalization, macroeconomic shocks, and structural changes, and thus contributing to more sustainable growth'. These are programs such as: Cash transfers or food coupons, by category or subject to resource conditions, such as family allowances or social pensions. In-kind transfers, school meal programs or supplements designed for mothers and children are the most widespread, but they also include distribution of meals ready to eat, school supplies, uniforms, etc. Generalprice subsidies, often for food or energy, targetedto households. Employment through labor-intensive public works programs, sometimes called "workfare" (conditional transfers). Cash or in-kind transfers to poor households, subject to meeting specific educational or health conditions. Exemption from duties for basic services, health services, education, public services, or transportation. Given its limited resources, high poverty incidence and periodic climatic shocks, an effective food security and social protection strategy i s essential for Niger. Such a strategy should include three key components: (i)improving the efficiency and scope of safety net programs; (ii) promoting effective medium-termstrategies and investments to improve food availability, access and utilization; and (iii) strengthening emergency responseandthe information system. Avoiding future food crises will also require greater recognition of the major importance of regional markets (particularly with Nigeria) in influencing market prices, food imports (and exports), and ultimately household food consumption in Niger. Finally, any social protection strategy, including the safety nets, will needto be consistent with the country's overall strategy for growth and poverty reduction, which represent the ultimate solutions for food insecurity and the poor See Grosh et al. (2008) for a complete description of social safety nets. 10 health and nutritional outcomes associated with both poverty and household food insecurity. An action plan has been developed with close collaboration o f the GoN and development partners proposing policy measures that need to be implemented in short and medium terms (see the policy matrix). Safety nets toprotect poor and vulnerable people: Currently, social safety net programs receive only a small portion o f the government's total budget, representing between 1 and 5 percent o f total expenditures between 2001 and 2006. During this period, almost 70 percent o f the safety net expenditures focused on food crises, with the Food Crisis Cell (CCA) largely responsible for coordination o f these programs and Niger's Food Production Office (OPVN) and a variety o f NGOs serving as the main implementing agencies. Safety nets are needed even in years of normal harvests, however, toprevent vulnerable and poor peoplefrom falling deeper into poverty and debt. The main challenge i s to design an effective safety net system and provide adequate funding to ensure its sustainability. The development o f an effective safety net program could include the following steps: > Development of a comprehensive safety nets syste. The mechanisms for prevention and response to food crises as currently implemented by the National Mechanism for the Prevention and Management o f Food Crises in Niger (DNPGCA) constitute only a part and one stage inthe implementation o f a safety net system. Consequently, to develop a general safety net system for households suffering from chronic and seasonal and transitory food insecurity, it i s necessary to (i) formulate a safety net program aligned with the PRSP2 and RDS; and (ii)ensure coordination between the State and the technical and financial partners (TFPs). This in turn will require the inclusion o f safety net programs in the government budget and a long-term financial commitment from the TFPs. P Improvement of the effectiveness of current safety nets systems. This would require (i) an assessment o f the effectiveness o f existing programs (HIM0 [labor-intensive public works programs] such as "Food for Work" (FFW) and "Cash for Work" (CFW); free distribution o f foodstuffs; subsidized food sales; cereal banks; school feeding programs; etc); (ii)a review o f the existing targeting system to determine the challenges o f addressing the needs o f chronically poor people; and (iii) an improvement inthe program monitoring system. Within this framework, to ensure that the safety net system is effective (in crisis and normal periods) it must be possible (i) to guarantee that the "Food for Work" (FFW) and "Cash for Work" (CFW) programs are well organized and give productive results; (ii)to limit the amount o f free distribution o f food as well as subsidized sales, and pay specific attention to synchronization and localization o f sales; (iii) ensurethatthepurchaseoflocalfoodstuffsiswellplanned; and(iv)thattargeting to o f programs follows transparent and efficient procedures for the selection o f the regions and households to benefitfrom the program. > Introduction of new safety nets programs. Once the vulnerable groups are identified and evaluation o f the existing programs i s completed, it will be necessary to identify pilot programs that could be tested based on transparent eligibility criteria, coverage and cost. Among these new programs to pilot are cash payments and conditional cash transfers (i.e., linked to school attendance and health service utilization for children). 11 Medium term policies and investments to improvefood availability, access and utilization: Medium and long term policies are alsb needed to reduce vulnerability to production shocks, raise incomes of the poor, and enhance market efficiency. These targeted actions are already integrated in the Government programs such as under axis no.1 of RDS as well as in the in the Accelerated Development Strategy and the Poverty Reduction Strategy (PRSP2). In this context, policies could focus on the following: 9 Increase availability of staplefood products. This could be achievedthrough improving agricultural productivity and efficiency of agriculture markets, as well as reducing production risks for farmers. Promoting eflcient domestic agriculture production could be achieved through investments in agricultural research and extension, investments in road construction and maintenance, increasing market infrastructure (e.g. storage facilities), and improving access to water for agriculture through investments in irrigation and water retention. 9 Ensure that commercial policies promote the development of cereals markets. In this context it is necessary to (i)reduceprice riskfor farm products to increase incentivesfor production; and (ii) reduce variability of rural incomes through promotion of farmers' groups. In addition it is crucial to improve market information systems on prices and volumes of production and market supplies. To ensure that commercial policies do not impede the development of cereals markets, it will be necessary to promote market efficiency and transparent policies on food aid flows, imports and releases of food stocks. 9 Expand access to food for the poor (in rural and urban areas). In this context it is important to support various household income-generating activities (such as micro- credit through local NGOs, investment in education and skills development to raise labor productivity). 9 Improve utilization of food. This will require focus on the overall health and sanitation environment, particularly among the poorest households. Strengthen emergency response and information system. Despite recent progress, further steps are needed to improve the functioning of the early warning systems and emergency response as well the existing informationsystem. This could include: 9 Reviseand implement the GovernmentNational Contingency Planfor FoodSecurity and set up a comprehensivemonitoring systemfor food procurement, distribution and stocks. 9 Improve the information system and the emergency program by (i) monitoring international price movements, particularly in Nigeria. Particularly it is critical to recognize that, depending on the price incentives for trade between major markets in Niger and northern Nigeria, net imports of millet could vary by about 200 thousand tons (about 10 percent of millet availability in a normal year); and (ii)improve information dissemination of the early warning system to decision makers and other actors; and (iii) strengthen the analytical capacities of the key government institutions (CCA, SAP, SIMA, CIC). 12 P Use current and new safety nets programs for emergency response. Particularly it is necessary to strengthen efficient safety nets programs and to define specific targeting mechanisms for emergencies. 13 . . . . e . . . I . . . . 0 . . . . 0 . 0 . . . . . . 0 0 0 0 0 0 0 0 . 0 0 P 0 0 0 a 0 e - e o e e e E r cU c c c Bct k G B- e 0 U m e .-e v) 0 U Y Uc c c o( c c c.1 e 0 0 0 0 0 e -L CHAPTER 1:INTRODUCTION BACKGROUNDAND MOTIVATION countries` Poverty remains widespread Countries Niger and social indicators compare poorly to G N I per mplta, Atlas method (current us$) 146 240 population growth (annual 2 3 3 3 Sub-Saharan African averages as Life expectancy at birth. total (years) 41 45 Feniliry rate (briths per woman) 5 3 I 1 presentedintable 1 below. The share of Infant mortality rate (per i 000 live births) 96 I50 Access to improved water source (% of total pop ) 56 46 the population living in poverty, Access to improved sanitation (Oh of total pop ) 53 43 Gross primary enrollment. total (ohofage group) 92 41 estimated at 62 percentin 2005, has been Adult literac ,total (% of a es IS and older) 61 29 I ~ r o TheI~li~dauBmk.1007k s WorldBenk c 2. Niger's vulnerability to several exogenousfactors, as well as its low level of socio-economic development, result in chronicfood insecurity. Past experience shows that Niger's economy has been affected by high variability of rainfall, terms of trade shocks, and volatility of aid flows. Evidence shows that these shocks result in lower growth performance, severely affect human development, and cause chronic food security in Niger. Food insecurity is aggravated by high incidences of rural poverty, which translate at the household level into low purchasing power, lack of access to food by vulnerable groups, rampant malnutrition, and poor access to health facilities. Droughts often result in increases in the prices of millet, sorghum and other staple foods of the population, and dramatic decreases in prices of livestock ("savings account" of agro- pastoralists and pastoralists).2 Since there is no safety net scheme in place, the rural population often migrates out of the most vulnerable zones during these difficult times. 3. Reducing vulnerability and ensuring food and nutrition security is an overarchingpriority for the Government. Food aid has served as an important resource for the Government and i s considered to be integral to the provision of safety net interventions, especially emergency response. In this framework, since 1998, Niger's government and major food aid donors have managedaDispositifNational de Pre`vention et de Gestion des Crises Alimentaires (DNPGCA), a National Mechanism for the Prevention and Management of Food Crises in Niger mandated to (i)help the government to build cereal and financial reserves for food aid during crises; and (ii) implement support actions for populations during periods of food crisis. Moreover, the * Themorerecent drought in2005 ledto food shortages for millions ofNigeriens. 18 government's Rural Development Strategy (RDS) 3action plan has set the strategic vision, policies, and institutional framework to address food security issues. More specifically, this strategy suggests that the Government (i)pursue the ongoing efforts to increase the performance of the existing national system for crisis prevention and mitigation; and (ii)adopt a multisectoral approach to focus on both increasing agricultural production and reducing the country's vulnerability to food crises. In this context, the existing national system for crisis prevention and mitigation, createdin 1989, has been improved over time to better target vulnerable zones and households. Moreover, in light of lessons learned from the 2005 crisis, the Government is pursuing efforts to increase the performance of the system by: (i) focusing the alert approach on accessibility (monitoring of purchasing power and anticipation of market behavior and household strategies); (ii)integrating the nutrition and health dimensions of food insecurityinto the vulnerability analysis; (iii) improvingthe targeting of beneficiaries and the estimation of their needs; (iv) strengthening the role o f the communes in food crisis management; and (v) developing a communication strategy with the media. Moreover, with donors' collaboration, an evaluation study on the national system for food crisis management and prevention, along with an action plan have been prepared and validated by the Government. This action plan aims to establish a better emergency response mechanismto food crisis inNiger, but it is not expectedto address the structural issues to prevent food crisis. STRUCTUREOF THE REPORT 4. Aligned with Government thinking, and taking stock of the existing information as well as new analysis on vulnerability, the purpose of this report is to provide an analysis of the food security situation inthe country, draw lessons from the experience of the 2005 food crisis, and discuss policy options for improving food security. More specifically, this report: > Presents a detailed analysis of households food security in Niger, including main sources of risks, household vulnerability to food insecurity and coping > mechanisms; Provides an overview of agriculture production, food availability and access to grain markets, highlighting the role of private unofficial trade with > neighboring countries; Reviews the experience of the food and nutrition crisis in 2005 and an overview of the existing national food security system and safety nets for > crisis prevention and mitigation; Suggests policy options, aligned with the Government Development Strategy and the Poverty Reduction Strategy (PRSP2) and the Rural Development Strategy (RDS), for: (i)introducing a safety net to protect the poor and vulnerable groups; (ii)improving the efficiency of mediumterm policies and investments to improve food availability and accessibility; and (iii)strengthening the existing emergency response. R D S Program9 is focused on "Reduction de la vulnerabilite des menages" - 19 5 . Given the multi sectoral aspects of thefood security, the report addresses both enhancement of food security and prevention of reductions in food security. A comprehensive food security strategy needs to address ways to enhance food security through increases in agricultural productivity, as well as through policies related to education, health & nutrition and population growth that are crucial for diversifying and increasing household incomes (and thereby their access to food). Since safety nets are essential to respond to chronic food insecurity, the report is discussing the food security strategy in the context of the social protection interventions, while complementing other existing and ongoing analytical works (i.e., on irrigation, agriculture, population, health). 6. This report relies on several primary and secondary sources of information, including studies on risks, shocks and vulnerability to food insecurity in Niger, conducted between 2004 and 2008. Primary sources include surveys of governmental and non-governmental interventions during and after the food crisis, prepared by the National Statistical Office (INS), as well as a cereal market survey conducted by four non-governmental organizations in Niger (CRS, CARE, HKI and World Vision International). The INS has completed several consumption and expenditures surveys: (i) in May 2005 an extended national survey (Questionnairesur les Indicateurs de Base et de Bien-etre, or QUIBB) and (ii)in November 2006 a survey on food security of the Households (Enquztesur la Conjoncture et la Vulne'rabilite'Alimentaire des Me'nages,or ECVAM). These two surveys have beencollected at different point in time: one during the crises of 2004-5 and the other after a good harvest inNovember 2006. They provided the basis for the analysis of poverty, conducted by the INS in collaboration with the World Bank, and the household food insecurity. Secondary sources of information include the series of poverty and vulnerability reports preparedby various international partners (including WFP, the World Bank, etc.) and various consultants between 2006 and 2008. These studies provide detailed information on the food security situation in Niger, poverty and consumption, cereal production and marketing, health and nutrition and the types of shocks that affect the country. 7. The rest of this study is organized as follows: Chapter 2 presents a profile of the poverty and food security situation in Niger, including the sources of risk and vulnerability to food crises. Chapter 3 provides an overview of the agricultural production and marketing system in the country, including information on trade flows and food aid. Chapter 4 summarizes the 2005 food crisis, its causes and the responsesof the Government of Niger, non-governmental and international organizations prior to, during and after the food crisis. Chapter 5 outlines policy recommendations to provide protection to householdswho are food insecure and respondto future food crises. 4Notethat the comparisonbetweenthe two surveys is firther complicated by the differenceinthe structure of the consumption modulesbetweenthe two surveys. Howeverthey do represent an illustration ofthe consumption patternbetweenthe periodof crises in 2005 and after a good harvest in fall 2006. 20 CHAPTER2: POVERTYAND FOODINSECURITY According to the available information, more than 50 percent of Niger's population suffers from someform of food insecurity, including chronic, seasonal and transitory food insecurity (defined as reduced access tofood after a shock). This is due to the complex interaction between low food availability, limited economic access, the poor health and nutritional status of the population, and the country's high susceptibility to shocks. According to available household surveys, over two thirds of people's daily caloric consumption comes from cereals, with regional variations in consumption levels mirroring the spatial distribution of poverty. Nutritional outcomes are consistent with the prevalence of seasonal and transitory food security. Coping strategies to the various risks and shocks affecting urban and rural households include the use of food aid, reduction in number of meals per day, migration, and asset and livestock sales. Overall, poor households are more exposed to shocks, thepoorest regions are the most vulnerable to food insecurity and coping mechanisms make households more vulnerable tofuture food insecurity. 8. More than 50 percent of Niger's population is estimated to suffer from some form of food insecurity. Regardless o f the indicators used, survey-based evidence consistently shows that much o f Niger's population suffers from widespread chronic, seasonal and transitory food insecurity. This i s due to the complex interaction between low food availability, poverty and limited economic access, the poor health and nutritional status o f the population, and the country's high vulnerability to shocks due to its economic dependency on subsistence crops and livestock. Per capita food consumption levels are low for about half o f the households in Niger; there i s limited dietary diversity for much o f the population; prevalence o f stunting inchildren under 5 i s high; and micronutrient deficiencies are widespread (see Box 1). Box 1: How to Measure Food Security? I International Organizations (UN,FAO, WFP and the World Bank 1986)define food security as "access - by all people at all times to sufficient food for an active, healthy life". Food insecurity is therefore defined as the absence o f one or all o f the conditions required, and can be classified into one o f three categories: chronic food insecurity, defined as insufficient access to sufficient food on a continuous basis; seasonal food insecurity, or insufficient access to food on a cyclical basis; and transitory food insecurity, whereby households do not have access to sufficient food following a shock, such as a man-made or natural disaster. Although most definitions o f food security used by governmental, non-governmental and international organizations are closely related to the World Bank's definition, measurability remains problematic. Maxwell and Frankenberger (1992) highlightedthe diversity o f the points o f view regarding food security, and concluded that there is not a "gold standard" for measuring food security (Maxwell 1999). The most common approaches used include analysis o f poverty, consumption patterns, such as daily caloric intake, and nutritional status o f children under 5 (stunting, wasting, under-nutrition) and micronutrient deficiencies. Another approach, presented later in this chapter, uses a risk and vulnerability analysis to relate the main sources o f risk and vulnerability to household level food security indicators, distinguishing between chronic food insecurity and vulnerability to food insecurity. 21 CHRONIC AND SEASONALFOODINSECURITY Linksbetweenfood InsecurityandPoverty5 Table 2: Poverty Status in Niger by Geographic 9. All methods used for measuring poverty - includingliving Area the Region Incidence Depth Severity monetary approach, the Agadez 45.9 16.1 8.8 conditions approach and the Dosso 67.3 28.8 15.3 subjective approach - show that the Maradi 79.7 35.1 19.0 incidence of poverty remains high in Tahoua 45.9 14.5 6.2 Niger.6 With an average GDP per Tillaberi 68.9 26.8 13.9 capita of US$ 240 (2005), Niger is Zinder-Diffa 63.1 23.0 23.0 classified as the forth poorest country Niamey 27.1 7.2 7.2 in the world, and is the lowest-ranked National Source: UIBB200562.1 24.1 12.3 on the UN's Human Development Index (HDI).7 According to the Survey on Satisfaction of Basic Needs (DSBE) in 2004, 70 percent of households inNiger do not meet their basic needs, including food, income, employment, farm equipment, land, education, health, and drinking water. These results are further confirmed by the monetary evaluation of poverty (QUIBB 2005), which shows that the prevalence of the population living below the poverty line was 62.1 percent in 20058(see Table 2). Infact, the prevalence of poverty inNiger has remained essentially constant since 1993 (in 1993 the incidence of poverty was 64 percent as compared to 62 percent in 2005). Furthermore, according to a participatory poverty survey (2006/07), approximately 66 percent o f households inNiger consider themselves to be poor, and 20 percent consider themselves to be extremely poor. Updated estimates and further Three poverty surveys have been conducted in Niger since 2000: The Survey on Combined Questionnaire of Basic Welfare Indicators (QUIBB) in 2005; the Survey on Satisfaction of Basic Needs (DSBE) in2004; and the second participatory survey on poverty (EPP2) in2005. 6 The incidence of poverty measures the percentage of individuals or households whose consumer spending is below the monetary poverty line.The poverty line correspondsto a minimum annual consumer spending for an individual or household; the depth of poverty measures the average gap in percentage between the level of well-being of the poor households and the poverty line. It is used in estimating the minimum amount of additional resources to be transferred to poor households to place them on the monetary poverty line; and the severity of poverty measuresthe average gap between consumption by the poor and the poverty line.It gives relative proportion ofpoorer people in a given group. While there i s no single and universal definition of poverty, poverty i s often defined as "a state of individual or collective destitution whichplaces man in a situation of shortage or lack of essential needs" This definition (WB 2000) reflects the lack of adequate income to satisfy basic needs in feeding, health, education, drinking water, decent housing and results in lack of opportunities to participate in social and *economic life,Millenium an in greater vulnerability of the populations concernedto shocks of various types. See "Niger Development Goals: Diagnostic and Tools"; WB, June 2005; Q. Wodon. Based on this study, the poverty thresholds for Niger in 2005 were 144,750 FCFA (equivalent to about $USD 175) per capita and per year for urban areas and 105,827 FCFA (equivalent to about 110 USD)for rural areas. 22 analysis o f poverty trends are expected after the completion o f the ongoing Household Budget Survey (2007/08). Moreover, besides its lower level o f consumption, food expenditure accounts for a significant amount o f total household expenditures leading to highlevels o f food insecurity. 10. Poverty incidence varies by region, though is on average higher in rural areas than in urban areas. The prevalence o f poverty and extreme poverty are relatively higher inrural areas, with 65.7 percent of the population inrural areas living below the poverty line, compared to 55.5 percent in urban areas. Regional variations in poverty are also significant as presented intable 2. The incidence o f poverty i s highest inthe Maradi (79.7 percent), Tillaberi (68.9 percent) and Dosso (67.3 percent) regions; but it reaches 27.1% inthe capital Niamey. Likewise, the depth ofpoverty that indicates whether families are living in abject poverty or just below the poverty line, also confirm higher levels o f poverty inthe same three regions: Maradi, Tillaberi and Dosso. 11. Variations in poverty rates are even greater at the deparment level (see Figure 1J9 Poverty is most widespread in Mayahi (Maradi) department, where almost 90 percent o f the population i s living below the poverty line; departments located in the regions o f Maradi, Tillaberi, Dosso and Zinder also have very highincidences o f poverty, with more than 50 percent o f the population living below the poverty line. Other regions (Tahoua and Agadez) have localized pockets o f poverty, but on average poverty rates range from 35 to 50 percent. Figure 1: Poverty Rates in Niger by Department, 2005 12. As in other countries, poverty in iwger is strongly correlated with a variety of household characteristics, such as household size, the gender of the household head, educational level and the economic status of the head of the households. According to the 2005 QUIBBsurvey, larger households are relatively poorer, as are households with more children. All else being equal, female-headed households are relatively poorer, as are households in which the household head did not complete primary education. The See WB 2006; INS and Q. Wodon. 23 household's source of income also has a major impact on per capita consumption and the household's probability of being poor. In general, households inNiger rely on a limited number of income-generating activities: according to the 2006 ECVAM survey, 20 percent o f households generate income from a single activity (agriculture, trade or livestock)." Moreover, approximately 65 percent of households whose head is self- employed are classified as poor, followed by those headed by unpaid apprentices (61.3 percent) and domestic servants (54.3%) (INS QUIBB 2005). By contrast, household heads whose main source of income is a formal public sector salary job have a higher standardof living. Consumptionpatterns 13. Foodaccounts for over 60 percentof total householdexpenditures. According to the 2005 and 2006 household budget surveys, between 61-64 percent o f total household expenditures were for basic foodstuffs. Among these expenditures, cereals (including millet, sorghum and other cereals) representedapproximately 50 percent of all spending 14. Though the majority of households engage in some farming, almost all households are net purchasers of food. Over 60 percent of households rely in part on their own production to meet their consumption needs. Nevertheless, over 60 percent of households were net purchasers of millet in 2005. This implies that households do not produce sufficient quantities to meet their consumption needs (see Table 3). loMore recent information confirms that 40 percent o f the population generates 80 percent o f their income from one activity (see Marinho and Gerard, 2008). 24 Table 3: Sources and level of household food consumption in 2005 and 2006 Source O f COnSUmptiOn consumed Quantity Expenditure Calorie Consuming Own Netmarket per capita bugdet shares Commodity groups households Production Purchase per day shares Grams 2005 Millet 86.7 60.4 62.5 385.4 48.2 56.5 Sorghum 42.5 31.7 21.5 75.9 9.5 12.6 Other Cereals 79.2 10.2 77.0 209.3 21.3 16.7 Rootsand Tubers 36.2 4.0 33.3 13.6 1.2 0.6 Nuts and pulses 56.3 23.2 40.0 21.5 2.7 2.7 Fruit Legume 83.7 18.9 79.9 101.2 1.4 Meat and fish 54.1 22.2 41.1 11.8 5.2 2.7 Milk,Eggs, cheese 65.1 19.0 53.2 51.0 6.2 2.0 Oil 36.0 0.0 36.0 3.8 2.4 3.3 Sugar 65.2 0.0 65.2 9.6 3.3 1.3 2006 Millet 81.8 57.8 26.3 346.7 46.1 42.0 Sorghum 39.2 27.1 13.3 107.5 8.0 13.0 Other Cereals 76.7 7.8 74.7 130.2 26.8 14.3 Roots and Tubers 14.3 0.1 14.3 12.2 1.9 1.4 Nuts and pulses 65.2 34.8 33.3 67.9 6.4 9.0 Fruit Legume Meat and fish 52.1 4.5 50.3 53.8 4.9 2.7 Milk,Eggs, cheese 21.7 7.9 15.0 58.0 3.8 2.6 Oil 7.8 0.7 7.2 48.3 2.1 14.9 Sugar 32.5 0.3 32.2 1.o 0.1 0.1 Souurce: INS QUIBB 2005 and ECVAM2006 15. Cereals are the dominant source of calories and dietary diversity is limited especially after a crisis. According to the QUIBBand E C V A M surveys, over 80 percent o f daily caloric consumption came from cereals (millet, sorghum and other cereals) in 2005 and 70 percent in2006. Less than 5 percent o f daily caloric consumption is derived from meat and fish, dairy products, and fmits and vegetables, which are important sources o f micronutrients, such as iron, iodine, vitamin A and calcium (Figure 2). l1 'ICalculations of per capita daily caloric consumption are based on different number of commodities and methodology in 2005 and 2006 expenditure surveys. However, the ordering of the commodities remains roughly the same. 25 Figure 2: Share of Per Capita Caloric Consumption by Commodity Group (2005-2006) Shm of Psroapita Cslons Conrumplionin ZWS, hy commndily gmup Share of Percapila Cslonc Consumptiontn 2lUA'i. by tommodit).group Source: I N S QUIBB 2005; ECVAM2006. 16. The majoriw of Niger's population consumesfewer than 2100 caloriesper day. In 2005, 59 percent of the population consumed less than the minimum caloric requirement for adults population - 2100 calories per day - as defined by the World Health Organization'2, while in 2006, about 52 percent o f the population consumed less that this minimum requirement. Table 4 also shows important geographical variation among regions. Although poverty rates are lower in urban areas, the percentage o f the population with inadequate caloric consumption i s actually higher in urban areas in all regions with the exception o f Dosso. This difference between patterns o f poverty as measured by the value o f expenditures and estimated calorie consumption reflects differences in food and non-food consumption patterns, as well as variations in prices between rural and urban areas, confirming the multi-dimensions o f food security. For example, urban households tend to eat more rice, a higher per calorie cost staple; rural households tend to consume more millet, a staple with a lower per calorie cost. More over, a shortage o f food affect urban areas through prices o f products while rural areas by the unavailability to cover own needs from own production. Hence, better indicators o f l2Daily caloric requirements for specific populations should be calculated based upon age, height, gender and the individual's activity level (mild, moderate and intense). In the absence o f such information, the standardthreshold of 2100kocalories per adult per day is used. 26 food insecurity in rural areas can be correlated to the fact they are more likely to receive exterior assistance in case of urgency or of threat of famine given its dependence on subsistence crops. (YO rthe Population Consuming less Table 4. Prevalence of Chronic Food Insecurity in Niger Ian 2100 calories per day) 2005 (Yo) 2006 (Yo) Region Rural Urban Total Rural Urban Total Agadez 41.4 64.4 53.3 66.0 64.6 65.2 Diffa 35.2 43.4 36.4 44.5 54.1 45.7 Dosso 68.7 59.1 67.9 58.5 51.6 57.7 Maradi 74.8 72.9 74.6 44.5 52.8 45.2 Tahoua 37.5 39.8 37.7 52.8 55.7 53.0 Tillaberi 55.7 55.5 55.7 55.4 58.9 55.7 Zinder 58.5 73.4 60.3 46.5 68.0 49.6 Niamey 67.3 67.3 50.7 50.7 Total 58.2 64.1 59.2 51.0 58.1 51.9 Source: INS OUIBB 2005 and 2006 ConsumDtionSurvev The nutritionaldimension Table 5. Prevalence of Undernutrition in 17, Nutrition-based indicators of Children under 5 in Niger, 2006 food insecurity confirm the high level Region Children 0-59 months of chronic food insecurity in Niger. WastinP Stunting According to the 2006 Demographic Zinder 10.6 58.7 and Health Survey (DHS), the Maradi 11.6 62.2 prevalence of stunting (low height-for- Diffa 12.7 46.6 age) inNiger is estimated at 50 percent. Dosso 10.3 46.6 The situation varies across regions, with Tahoua 10.3 46.7 Maradi and Zinder having the highest TillabCri 8.7 46.7 prevalence of stunting (Table 5). Agadez 10.5 38.8 Niger's stunting rate makes it the Niamey 6.7 21.3 Source : DHS/MICS(2006) second worst-affected country in Sub- Saharan Africa (Figure 3). The l3 Multiple rounds of DHS surveys conducted in Niger since 1992 suggest that chronic food insecurity, as measuredby stunting, has increased from 40 percent (1992-2000) to 50 percent (2006). 27 Figure 3: Stunting in Sub-Saharan Africa Prevalence Source : State ofthe World's children, UNICEF, 2007 18. Micronutrient deficiencies in the country provide further evidence of chronic food insecurity. The main micronutrient deficiencies are iodine, iron and vitamin A, which are crucial for vital bodily functions. It is estimated that 20 percent of the population suffers from goiter (iodine deficiency), with higher rates in the Dosso and Tahoua regions. Furthermore, an estimated 42 percent of children under six suffer from sub-clinical vitamin A deficiency, with 7 percent of their mothers reported to have suffered from night blindness during pregnancy. Interms of iron deficiencies, 84 percent of all children, and 46 percent of women aged 15-49, are anemic (DHS/MICS 2006). 19. Nutrition outcomes also provide evidence that Niger's rural population suffers from seasonal food insecurity. There i s Figure4: Numberof Cases of ReportedWasting(by Week), 2006 and qualitative evidence that 2007 during the hungry season (soudure), between June and September, household food consumption is lower (both in terms of quality and quantity). This i s supported by the prevalence of wasting in Niger (10 percent), classified as "high" according Semmes to the World Health thresholds for public(WHO) Organization's health **"" s'nNspMnr problems. Figure 4 shows the number of reported cases of moderate and acute malnutritionby week for 2006 and 2007. While the data are subject to potential selection 28 bias (as a higher number of nutritional rehabilitation centers opened during this period), a general pattern of seasonal food insecurity emerges. In 2006, there was a marked increase in reported cases of moderate malnutrition between April (week 16) to October (week 39), coinciding with the pre-harvest period and hungry season. The same pattern, albeit not so strong, was also observable in2007. RISK,VULNERABILITY, AND HOUSEHOLDFOODSECURITY 20. Niger's populations are also vulnerable to transitoryfood insecurity, defined as reduced access to food after a shock. Shocks, such as droughts, floods, pest attacks, health epidemics and economic not only have an immediate impact on households' food security, but can also have longer-term consequences and leave households in debt or without the necessary assets to guarantee a minimum level of production.l5 As vulnerability is an ex-ante measure of well-being, an ideal vulnerability analysis would follow the approach outlined in Box 2. While the data available will not permit covering all the steps outlined in the box, the section first summarizes the results of other reports and attempts to using some of the available data surveys (Le., QUIBB 2005 and ECVAM 2006). Box 2 :Steps for Identifying the Most Vulnerable in Niger o Identify potentialshocksor hazards that could affect communitiesor individuals o Determinethe levelof risk(probability ofbeingaffectedby this shock) ofthe affected populations o Notethe negativeimpactsof theseshocks on individual,householdand community- levelfood security and how householdscopewith these shocks. This can or should includea percentage reductioninthe productionof staple food crops, or an increase in staple food crop prices (or a reductionin livestockprices). o Determinethe levelof householdsfood insecurityand the characteristics of individuals,households,communitieswho are the mostvulnerableto shocks (ie, they are the most at risk andthe most affectedby the negative impacts) o Determinewhethervulnerablehouseholdswill have"unacceptable" levelsof food insecurityas a resultof the shocks.For example, do householdreducethe number of meals. dietarv diversitv? 1s there an increase in malnutrition? Source: See Tesliuc, E. and Kathy Lindert (2004); Hoddinott, J. and Agnes Quisumbing (200. Hoogeveen, J., E. Tesliuc, R. Vakis, with S. Dercon (2004); del Ninno, C. and A. Marini(2005). Risks and Shocks 21. Although rural and urban households in Niger are susceptible to a variety of risks and shocks, not all of these shocks result in food insecurity or food crises. Whether or not food insecurity occurs depends on households' assets, risks and risk I4UNDP (2007) notes two main types of risk factors inNiger: conflicts (farmer-herder and farmer-farmer) and naturalshocks, such as droughts and production shocks. IsSee "Carter and Barrett paper" 29 management strategies. This section focuses on national and regional covariant shocks16 that are highly correlated with the production of and access to food; namely, droughts, harvest failures, and rising grain prices (Holzmann and Jorgensen, 2000). l7 22. Over the past 12 years, Niger has experienced three severe droughts, two of which were associated with food crises. During the last 12 years, the country experienced droughts in 1996, 2000 and 2004.18 These weather related shocks coincided with lower cereal production, higher cereal prices, and lower incomes for the rural poor. Among these drought years, 1996/1997 and 2004/2005 were identified as severe food crisis years.lg In 2005, an estimated 2.4 million Nigeriens were affected by severe food shortages, with more than 800,000 of these classified as critically food insecure (USAID FEWSNET 2005).20 Although the 2004/05 food crisis was not of the scale of the 1968- 74 or 1983-84 famines, the gross mortality rate reached 1.5 deaths per 10,000 per day in certain regions, and the child mortality rate reached 4.1 deaths per 10,000 per day (Medecins sans Frontieres 2005). 23. While droughts consistently cause production to decline, the severity of the impact has varied sharply across sub-regions within Niger and also over time. In spite of the fact that 2000/01 drought had affected per capita production decrease - of more than 50 percent - in about 15% of departments, the 2004/05 drought affected over 25 percent of departments. Comparatively, during non-drought years (e.g., 2001/2002 and 2005/2006), less than 3 percent o f all departments experienced a per capita drop at the same level of more than 50 percent. l6 Shocks that affect all the households in an area or group are called covariant shocks, as opposed to idiosyncratic shocks that affect only some households. 17 This paper adopts the definitions proposed by the World Bank (2001) for the concepts of risks, shocks, poverty and vulnerability. Poverty is defined as an ex-post measure of well-being, whereas vulnerability is an ex-ante measure of well-being (or expected outcome). Central to the concept of vulnerability is risk, which is aprobability distribution of events. Examplesof risks include natural risks, health risks, economic risks, life-cycle risks, social risks, political risks or environmental risks. Whether a risk materializes, it is known as a shock. l8 Annual rainfall was lower than one standarddeviation from the mean. See Nicholson et a12000. I 9 Since the 1960s, Niger has experienced drought-related transitory food insecurity in 1966/67 (Bandabari), 1973/74 (Dakoussou), 1984/85 (El Bohari), 1990/91, 1993/94, 1996/97 (Matche mai), 2000/01 and 2004105 (Wazaka gaya ma). O f these years, only 1973/94, 1984/85, 1990/91, 1996/97 and 2004/2005 were classified as severe food crises. (Government ofNigerDNPGCA 2007). 2o As the definitions of a food crisis vary, identifying clear thresholds for food crises is problematic. While a number of householdswere affected by food insecurity in 1997/1998and 2000/2001, data on the number of households affected are not readily available. Nevertheless, it is commonly agreed that the magnitude of the 2004/2005 food crisis is the most severe, both interms of number of persons affected and the severity of the problem. 30 I Figure 5: Percentage of DeDartments Affected bv Production Shocks bv Year. 2000-2006 1 e 4"W j 1930% 20% IO% 0% 2a00~2001 Source: Aker 2007 24. These marked differences in regional climatefluctuations are also reflected in higher dispersion of grain production in drought-prone departments around the observed production mean in Niger. These fluctuations measured by the coefficient o f variation, CVs2', ranges from -11 in the Magaria department (Zinder) to .75 in the Tillaberi department. Overall, the regions of Maradi and Zinder (which account approximately 40 percent of national millet production) have the lowest coefficients of variation (.22 and -27, respectively), whereas the deficit regions of Tillaberi and Tahoua have the highest coefficients of variation. This suggests that rural households living in drought-prone areas (Tahoua and Tillaberi) are more likely to be affected a higher degree of output risk. Table 6. Percentage of Households' Exposed to Shocks by Poverty Level and Quintile Illness Production (severe or Land Drought Shock Death chronic) Flood Job Loss loss All HH 47.62 16.59 11.22 10.55 1.99 1.48 0.90 Poor 49.03 19.16 10.91 10.84 2.09 0.63 1.59 Non- poor 46.04 13.68 11.57 10.22 1.88 1.21 1.36 Quintilel 47.57 21.87 10.28 10.9 2.28 0.56 1.94 Quintile2 46.76 14.73 10.82 10.5 1.97 1.15 0.8 Quintile3 45.60 16.34 10.52 10.56 2.01 0.86 1.39 Quintile4 47.85 13.98 10.99 10.20 1.65 0.61 1.10 QuintileS 50.50 15.59 13.99 10.51 2.01 1.37 2.15 Source: INS QUIBB 2005 2 'The coefficient of variation is the standard deviation of per capita grain production in a particular departmentover the entire period, divided by meanper capita grain production. 31 25. Vulnerability to drought is associated with low level of household's assets and human capital. Poor households are not more exposed to shocks. Table 6 shows the relative importance of different types of shocks in Niger by poverty levels and quintiles. According to available information, drought is the most prevalent type of shock experienced by households in Niger, followed by production shocks. On average, poor householdsappear to be more likely to be exposed to natural and health shocks (drought, flood, production losses and illnesses). When stratifying the results by income quintile, however, the correlation betweenpoverty and risk is less clear. Econometric analysis by Ouedraogo (2008) shows that households whose head is unemployed, or work in informal sector; or is female; or does have low levels of education; or does not have livestock; or does have high dependency ratio are more likely to suffer from drought. In terms o f location, the households in Tahoua have the highest combination of people residing and reported having been exposed to drought and production shocks the most frequently, followed by those inMaradi,Tillaberyand Diffa (see Table 7). Table7. Distributionof householdsExposedto DroughtacrossRegionsof Residence (%) Region Residing Affected Incidence Agadez 3.2 2.4 0.74 Diffa 3.7 5.9 1.59 Dosso 9.7 5.1 0.53 Maradi 19.7 20.8 1.06 Tahoua 20.8 32.5 1.56 Tillabery 15.9 16.5 1.04 Zinder 20.2 16.7 0.83 Niamey 6.6 0.2 0.03 Total 100 100 1 Source: QUIBB2005 26. According to the information from 2005 QUIBB (INS), households havefairly limited coping strategies to deal with covariate and individual shocks. Based on this survey, food aid was rated as the most common strategy used by households to mitigate the effects of shocks, including natural, economic and health shocks (see Table 8). For drought alone, almost 1 in 4 households stated that they had used food aid as a coping strategy. Following food aid, households cited requesting assistance and the sale of agricultural products as the most common strategies, followed by the sale of household assets, borrowing, and assistancefrom NGOs. 27. The probability of receiving food aid is not correlated with poverty in rural areas. Among those affected by the drought in rural areas, only 72 percent of the poor received food aid as compared to almost 80 percent of the non poor. In urban areas, instead, poor households affected by drought were more likely to receive food aid than non-poor households. These results suggest that the targeting of food aid has been either non existent or it has focused on helping all those affected by drought rather than only poor households affected by drought (see Tables 8 and 9). 32 Table 8. HouseholdsCoping Strategiesafter Exposure to Drought (YO) Strategies Drought Food aid 74.8 Sell agricultural products 15.3 sell agricultural materials 3.O Sell HH assets 8.0 Get help from Other households 40.0 Assistance from NGOs 8.1 International assistance 6.2 Borrowing 7.3 Others 1.2 Source: INS QUIBB 2005 Table 9: Percentageof drought affected householdsreceiving food aid in2005 by poverty category and location Urban Rural All IPoor 73.0 72.2 72.3 Non Poor 52.4 79.3 78.1 ] Source: I N S QUIBB 2005 28. Qualitative studies conducted by numerous NGOs and WFPpaint a notably different picture of household coping strategies. WFP's Emergency Food Security Assessment (2005) noted that the most common strategies used by householdsto respond to natural shocks include a reduction in the number of meals per day, collecting wild foods, borrowing money or food from other households, migration, asset sales, land sales and livestock sales. While the study did not focus on the prevalence of these strategies (overall or by region) or their relative importance, migration, the sale of household assets, livestock stales and borrowing were highlighted as the most common strategies used by Nigerien households. 29. Some coping mechanisms make households more vulnerable to future food insecurity. Existing information confirms that although borrowing and selling strategies may enable households to cope with shocks in the short-term and help them to smooth income, they effectively decapitalize the household of its productive assets (natural capital, humancapital and financial capital), thereby making it more vulnerable to future food insecurity. For example, Marinho and Gerard (2008) estimate that 47 percent of households had sold livestock, and as a consequence, 60 percent of households in the national poverty survey were still indebt after the 2005 food crisis, with the average debt equivalent to 427 kg of cereals. 30. Livestock ownership affects coping mechanisms. In rural areas food aid is the most common form of coping mechanisms, especially for those that do not have any livestock. In Urban areas, instead, households without any livestock are more likely to receive help from other households (70 percent). 33 Table 10. Copingmechanismsafter drought by ownershipof livestockin ruraland urban areas (YO) RURAL URBAN Livestock Livestock Livestock Livestock value value value value No below above No below above Copingmechanisms livestock median median livestock median median Foodaid 80.7 74.7 72.7 55.8 72.7 76.1 Sell agriculturalproducts 10.7 14.7 18.6 18.6 6.3 33.0 Sell agriculturalequipments 3.6 2.5 2.8 1.5 7.4 5.9 Sales of HHassets 8.0 8.3 8.2 1.3 11.3 3.5 Get help from other HH 42.8 33.8 40.5 70.0 48.4 61.3 Assistance fromNGO's 7.6 7.5 10.1 4.0 1.o 2.0 Internationalassistance 6.0 4.4 8.3 11.4 1.o 0.0 Borrowing 9.4 5.6 7.0 20.5 8.0 5.4 Others 2.2 1.1 1.5 0.0 0.0 2.0 Source: QUIBB 2005. Vulnerability to Food Insecurity 31. Despite the wealth of data collected on householdfood security subject, it is not possible to use a common classification of the most vulnerable households within the country and over time. There have been numerous expenditure, consumption and vulnerability surveys conducted by the Government of Niger (GoN), WFP and various NGOs between 2004 and 200722. Each study relies upon a different set of variable for the definition of food security and vulnerability, in addition to different approaches and methodologies to data collection. For example only the QUIBB 2005 and ECVAM 2006 contain information on expenditure and consumption. Therefore it i s difficult to have a consistent definition and sources of information for estimating household food insecurity (see Box 3). 22Vulnerability surveys include: Enquete Conjointe sur la Vulnerabilite' a 1'Insecurite' Alimentaire des Mknages au Niger (INS and SAP, November 2006); Enquete sur la Vulnerabilite' a l'lnsecurite' Alimentaire des Me'nages au Niger (INS, WFP, FAO, UNICEF, FEWS NET, CILSS, December 2007); Enquetesur la Securtie Alimentaire en Situationd'Urgence au Niger (WFP, October 2005). 34 Box 3 :Methodologies to estimate householdfood insecurity Besides the simple caloric consumption analysis presented in the first sectionof this chapter, there are several methodologies that can be usedto calculate householdfood insecuritv as: The fikt measureestimates the expected level of caloric consumption, based on householdhuman and physical assets and capabilities, and compares it with the observed level of caloric consumption. below 2100 calories per capita per day into 3 categories of risk: o Extreme chronic level (A) of food insecurity level reflects both observed and expected levels of consumption below the minimum level of caloric consumption; o Vulnerabilty to chronic level (B) food insecurity summarizes the share of householdswith observed consumption levels below minimum level ofcaloric consumption, but have the human and physical assets that would allow them to consume adequate level of calories, but they do not becauseof particular circumstances like drought. o Vulnerability to food insecurity level (C) that summarizes the share of householdsexposed to risk and uncertainty, which had affected their levels of consumption. They are those who are expected to consume less than 2100 calories per capita a day inresponseto a shock, but manageto consume more. o The overall level of food insecurity is measured by the sum of chronic (A) and transient food insecurity. The second measure calculates the food variety index named Dietary diversity. This index is a simple, or weighted, count of foods or food groups over a given reference period. It emphasizes the importance of consuming a wide variety of food so as to enhance dietary quality. The main disadvantage of this method is that it does not take into account quantity nor control diets regarding caloric composition. However, developing countries review confirms positive correlationof dietary diversity and nutrient adequacy. The third measure uses the principal component analysis to reduce multidimensional data sets to lower dimensions for analysis of different outcomes. The method provides a way of identifying patterns in multidimensional data, expressing it in way to highlight similarities and differences by reducingthe number of dimensions, without much loss of information. The last one uses a spatial econometric model for determination of local household food insecurity on the basis of the ECVAM 2006 and GIS data. This model follows a theoretical methodology to spatially predict food vulnerability using all available source of information at a given geographical level controlling for their geographical correlations. Source: See Coates, Webb and Houser (2003); Hoddinott, J., and Y. Yohannes (2002); Migotto et al. (2006); del Ninno, Vecchi and Hussain(2006) 32. I n Niger at least 22percent of the households are classified as extreme chronic food insecure, despite a small reduction of its level of food insecurity in 2006 compared to 2005. According to the first measure o f food insecurity presented above in Box 3, the overall food insecurity (extreme chronic and vulnerable to chronic) measured here i s by definition similar o f those presented in table 4. The reduction o f 7.3 percentage point in 35 food insecurity between 2005 and 2006 was the result of the improvement in the consumption levels of Niger population after the food crisis that has affected calorie consumption of population. In particular the reduction in the level of extreme chronic food insecurity was very large, going from 37 to 22 percent. However, the reduction in chronic food insecurity was accompanied by smaller increase of vulnerable to chronic levels of food insecurity of about 8 percent (see Table 11). Table 11:Household food insecurity levels 2005 2006 Food insecurity Extreme Chronic' 37.4 22.0 Vulnerable to Chronic' 21.9 29.9 59.3 51.9 -Vulnerable to food insecurity3 9.7 5.3 33. There are substantial variations infood insecurity across regions between 2005 and 2006. The high level of extreme chronic food insecurity has affected regions differently inNiger. In2005, Maradi, Dosso andNiamey were the region with the highest levels of extreme chronic and total food insecurity. In 2006, Agadez had taken Maradi's place as the region with the highest level of food insecurity in the country because the level of food insecurity fell inMaradi (see Table 12). 34. Given the incomparability of products list consumed in 2005 and 2006 surveys, the proposed dietary diversity index should not be directly compared over time. Comparing ranks of dietary diversity with previous results, Maradi and Dosso are the only two regions misclassified compared to table 4. In 2005, Dosso had the lowest level of insecurity among regions while Maradi has medium level of insecurity, compared to their higher levels of chronic food insecurity presentedintable 2. However, the dieatary diversity index of Dosso and Niamey can be explained by the fact that while rural/poor regions as Dosso had received a lot of assistance to deal with the famine23while urban households as those from Niamey reduce their diet as result of price increase. Measured by this index, consistently with food insecurity from above, Niamey and Zinder had higher levels of food insecurity in 2005; Tahoua and Diffa had increased levels of food insecurity between 2005 and 2006 while Tillaberi improved a little its levels of insecurity; and Dosso and Agadez had higher levels of food insecurity in 2006 (see Table 13). 23For example, as immediate response to the famine, the CADAJEEPN (Union of Evangelical Protestant ChurchesofNiger) had provided food supplies to 600 families inthree rural communes inthe Niger, Dosso Province 36 Table 12: Chronic. transient and vulnerable levels of food insecuritv 2005 2006 Extreme Extreme Chronic Vulnerable to Chronic Vulnerable to I Chronic2 Vulnerable3 1 Chronic2 Vulnerable3 Agadez 33.2 20.1 9.5 21.5 43.9 2.9 Diffa 23.6 12.8 17.3 14.4 31.3 4.5 Dosso 49.4 18.5 9.7 31.5 26.0 5.4 Maradi 48.1 26.5 9.4 18.2 26.6 5.8 Niamey 50.9 16.4 15.3 25.3 25.4 5.9 Tahoua 19.1 18.6 8.0 21.7 31.3 5.8 Tillaberi 31.6 24.1 7.6 25.2 30.4 5.1 Zinder 36.2 24.2 10.0 16.5 32.9 4.9 Urban 42.9 21.1 14.0 25.1 33.0 5.7 Rural 36.2 22.0 8.8 21.3 29.4 5.2 Niger 37.4 21.9 9.7 22.0 29.9 5.3 Source: 2005 QUIBB 2005 and ECVAM 2006 Table 13: Dietary diversification index 2005 2006 Agadez 14.0 15.9 Diffa 18.1 18.6 Dosso 20.5 18.3 Maradi 18.0 21.5 Niamey 13.6 23.2 Tahoua 18.2 19.0 Tillaberi 16.6 20.8 Zinder 15.5 19.5 Urban 12.7 19.6 Rural 18.2 19.9 Niger 17.2 19.8 Source: 2005 QUIBB 2005 and ECVAM 2006 35. According to the principal component index the poorest regions measured by household per capita expenditure are also the most vulnerable to food insecurity. Based on this 2007 survey, the regions of Tillaberi, Dosso and Maradi were the most vulnerable to food insecurity, with approximately 17, 13 and 11percent of the population affected, respectively. This is consistent with the results of 2005 poverty survey, which identified Maradi, DOSSO, Tillaberi and Zinder as the poorest regions. Overall, the survey data confirms that vulnerability is more heavily concentrated in the rural regions of 37 Niger, with 12 percent o f the population vulnerable to food insecurity, as compared with 6 percent inurbanregions. Figure 6 shows the percentage o f the populationvulnerable to food insecurity, based upon geographic region (INS et al, 2007). Although interesting results, they are not comparable with previous two index o f food insecurity. Figure 6: Percentage of the Population Vulnerable to Food Insecurity, 2007 36. Based on 2006 data and using a spatialframework for vulnerability, it can be shown that proximity to paved roads and access to water are key determinants of vulnerability to food insecurity. Figure 7 present results from Marinho and Gerard (2008) which had estimated that most vulnerable households in Niger are located in the south but not the extreme south o f Tahoua, the middle and the south o f Dosso and the west o f Tillaberi. According to this work, the strongest determinants o f food insecurity in Niger are proximity to a paved road, access to water sources, agro-climatic zone, rainfall or normalized difference vegetation index, NDVI, and recent shocks or food crises. While the identification o f Dosso and Tillaberi as vulnerable regions is consistent with the I N S 2007 survey and previous findings as from table 2 or table 9, Maradi, as for dietary diversification, i s not classified as vulnerable to food insecurity. Figure 7 :Predictions of Areas Vulnerable to Food Insecurity in Niger, 2006 CHAPTER 3: FOOD AVAILABILITY AND ACCESS Niger relies mainly on domestic production of grains (primarily millet and sorghum) for its cerealsupply. As production variessubstantiallyon an inter-annual basis due to extreme rainfall, imports (consisting mainlj of wheat, maize and rice) are required to meet local consumption needs. However,food aid and commercial imports typically accountfor a small percentage of total cereal supply on an annual basis, averaging 7 percent of total cereal availability between 2000 and 2005. Trade data in Niger are unreliable and underreported: unofficial imports of grainsfrom Nigeria may account for as much as 200 thousand tons of total supply. Survey evidence and analysis of prices indicate that Niger's wholesale cereal markets are generally well integrated; however, lack of production data in Nigeria prevents detailed analysis of market links between the two countries. Nonetheless, simulation analysis suggests that a reduction in net importsfrom Nigeria in 2005 was a major factor in the sharp rise in market prices during the 2005food security crisis. FOODAVAILABILITY: PRODUCTION,IMPORTSAND FOODAID AgriculturalProduction 37. Despite unfavorable agro-climatic conditions and recurrent natural disasters, agriculture and livestock-raising play a key role in Niger's economy. The agro-food sector accounts for approximately 40 percent of Niger's Gross Domestic Product (GDP), and is the major source of employment for more than 80 percent of the total population, most of who live below the poverty line. Inaddition to agriculture, which accounts for about 52 percent of rural output, livestock-raising is an important source of income for pastoral and agro-pastoral households, with an estimated 7 million animal head (WFP 2005). Agro-pastoral products account for about 85 percent o f export revenues for Niger (excluding uranium), with the livestock sector alone accounting for 56 percent (mostly live exports). 38. While there are a variety of livelihood zones within the country, agriculture is primarily characterized by subsistencefarming, and production of staplefood crops is based upon rain-fed agriculture. Niger is a large country located within the Saharan, Sahelo-Saharan and Sahelian agro-ecological bands of Africa, with rainfall ranging from 250-800 mm per year. Niger is characterized by three distinct agro-ecological areas: the pastoral zone, the agro-pastoral zone, and the agricultural zone (Figure 8)24,25(USAID 24 In its 2005 Livelihood Profile for Niger, USAIDiFEWS NET identified eight livelihood zones for the country: the desert, the Bilma oasis sub-zone, Air mountain cultivation, pastoral, agro-pastoral, rainfed agriculture, sub agriculture zones with high out-migration, irrigated cash cropping, Lake Chad region and *'Nigeri irrigatedrice. USAID/FEWSNET,2005 s divided into eight (8) administrativeregions: Agadez, Diffa, Dosso, Maradi, Niamey, Tahoua, Tillaberi and Zinder. Eachregionhas several dkpartements. 39 FEWSNET 2005). In the pastoral zone, which is located in the northern part of the country, immense spaces are exploited using a pastoral and trans-human system. Though natural conditions make this very good pasture land, demographics have brought about a development of agriculture along the southernbelt of this area, and a reduction of pasture space. The agro-pastoral and agricultural zones are located in the southern regions, which involves various sub-systems: a dune system, the Eastern plains and the Western plateaux. The dune system primarily produces millet, characterized by low yields and deficit household production. The Easternplains system includes the Maradi and Zinder regions, where staple food crops (millet and cowpea) are produced using semi-intensive agricultural techniques, intercropping and the use of small livestock. The Western Plateaux system (Tillaberi and Dosso) also cultivates staple food crops, with localized wetlands usedfor market garden production. Production of staple food crops inthe agro- pastoral and agricultural zones are primarily based upon rain-fed agriculture, with irrigated agriculture practiced in localized areas (Niger River basin, the Goulbis in Maradi and Zinder, the Maggia and Tarka, Lake Chad and Dallols (Dosso region). Agriculture is characterized by a family approach, on plots averaging five hectares (WFP 2005). Overall subsistence farming and animal husbandry are cultivated using traditional techniques. With increasing population density in the most productive zones - the population growth rate in Niger averages 2.9 percent per annum -- the sustainable management of productive natural resources, particularly land and water, has become essential to Niger's agricultural growth andrural poverty reduction. 39. Millet and sorghum are the main staple food crops in Niger. Figure 8: Administrative regions and agro-ecological The main grains cultivated are millet, zones sorghum, rice, fonio and maize, with moor cash crops including cowpea, peanuts, cotton, sesame, vouandzou, and souchet.26 Millet is by far the dominant staple food crop in Niger and is grown in almost every agro- ecological zone, accounting for 76 percent of total cereal production and 72 percent of net availability between 2000-2005. Sorghum, which is less drought-resistant than millet, is grown primarily in the southern regions of Niger; it, accounted for 22 percent of cereal production and 20 percent of net availability during the same period. Rice production accounted for only 2 percent of total net availability of cereals, and is grown primarily in the Niger River basin. The regions of Maradi and Zinder are considered to be the "bread baskets" of Niger, accounting for approximately 40 percent of national millet production. 26Vouandzou is a legume (also known as Bambara groundnut) that requires 500 mm o f rainfall, and is therefore well-suited to the southern Sahelian and northern Sudanian zones o f West and Central Africa. Souchet is a type o f tuber known as Cyperus esculentus, or Chufa Sedge, 40 40. Per capita increases in production are the result of an expansion in cultivated land, rather than higher yields.27 Although national cereal production increased by 48 percent between 1985 and 2004 (from 1.8 million MT in 1985 to 2.7 million MT), the increase was primarily extensive in nature; cultivated areas increased by more than 84 percent over the same period, from 4.3 million hectares to over 7 million hectares (Table 12). Cereal yields remained fairly stable over this period, averaging 372 kgha. 2.5 times less as compared to other West African with comparable conditions. These relatively low cereal yields are related to low rates of technological adoption, the use of rainfed agriculture and the frequency of natural shocks. These, in turn, are affected by limited agricultural researchand extension projects. 41. The onion and cowpea sub-sectors offer the best growth potential for exports. The World Bank's Agro-Pastoral Export Promotion Project identified five priority sub-sectors in agriculture: onions, cowpeas, sesame, souchet, and gum arabic. Onions are produced around the vast periodic watercourses (broad dry river beds) that cross much of south-central Niger, and account for 50 percent of the value of the exports o f these five sub-sectors. Niger i s also a major producer of cowpeas, which has replaced groundnuts as a cash crop for rural households. As is the case with staple food crops, the increaseincowpea production over the past two decades has occurred largely through an expansion o f the cultivated area. As per capita cowpea consumption is fairly modest (6 kglcapita), 50-75 percent of Niger's cowpea production is exported to Nigeria, mostly on an informal basis. The potential for higher yields and exports for these crops is constrained by storage facilities and practices, marketing channels (especially access to market information) and improved agricultural techniques. 42. Niger's livestock herd and exports are some of the largest in West Africa Despite an austere natural environment characterized by low rainfall; Niger's agro-pastoral space is vast and diversified. Niger's livestock herd, estimated at 28 million head in 2006 (7.3 million heads of cattle, 9.2 million sheep, and 11.2 million goats), is the largest in West Africa. Livestock exports to Nigeria account for more than 90 percent of overall livestock exports, although little meat is exported from Niger. Major constraints to increasing herd sizes and improving productivity include limited veterinary and other inputs, a feed balance deficit and low availability of agricultural and industrial by-products. However, there are substantial opportunities for increasing the value of livestock exports and meat exports to coastal countries, primarily Nigeria, Ghana and Ivory Coast. 27Productionshocks are defined as years when per capita cereal production in a department declined by morethan25 percentas comparedto the six-year average. 41 Table 14: Millet and Sorghum Production in Niger, 1985-2004 Cultivated Production Area per Production Year Population Areas (ha) (MT) Capita per capita 1985 6565000 4310931 1774113 097 270 1986 6783000 4348597 1743 559 0,6 257 1987 7008000 4359029 1362777 0,6 194 1988 7240000 4995768 2326505 037 321 1989 7480000 5094042 1754605 0,7 235 Average (85-89) 4 621 673 1792 312 0,7 256 1990 7728000 6942899 2045960 0,9 265 1991 7967568 6456771 2314991 038 291 1992 8 214 563 7 519314 2 171 693 0,9 264 1993 8469214 6099 128 1714310 0,7 202 1994 8731760 6950251 2368538 0,8 271 Average (90-94) 6 793 673 2 123 098 0 3 259 1995 9002444 7 164356 2034983 098 226 1996 9286395 7138358 2 172213 098 234 1997 9574274 6386922 1641530 097 171 1998 9871071 7607398 2894013 0,8 293 1999 10177080 7449871 2772346 0.7 272 Average (95-99) 7 149 381 2 303 017 097 239 2000 10492569 7306951 2049890 097 195 2001 11060291 7835456 3022350 0,7 273 2002 11403 160 7 816590 3 236927 0,7 284 2003 11 756 658 8 041 222 3 502 464 0,7 298 2004 12 121 114 7823260 2637242 036 218 Average (00-04) 7 764 696 2 889 775 0,7 254 Average (80-04) 6 085 875 2 128 794 0,72 253 Source: Productionstatistics from the Ministry of Agriculture, various years; WFP 2005. 43. Over thepast 22 years, climatic shocks -primarily drought -- have resulted in high inter- annual and spatialfluctuations in staplefood crop production. Between 1990-2004, the years of lowest cereal production were 1993, 1997, 2000 and 2004, coinciding with natural shocks (drought). The high degree of inter-annual variation in cereal production is evident when looking at the coefficient of variation (CV).28 The average coefficient of variation for national millet and sorghum production i s -20and .35, respectively, with an inter-crop correlation coefficient of .91.29 In general, sorghum production i s more variable than millet production, as the most common varieties of sorghum are of longer duration and less drought tolerant than millet. Beyond the national-level variation, there i s also significant spatial variation inmillet production, as 28The CV is calculated as the ratio ofthe standard deviationof cereal productionto the mean of cereal productionover the time period. 29The CVs for millet and sorghum inNigeria, which has a higher annual average rainfall than Niger, are only 0.16 and 0.10, respectively. ProductionbetweenNiger and Nigeria are correlated to a limited extent, with correlationcoefficientsof .43 for millet and .51for sorghum, respectively (Dorosh2008). 42 i s evident when assessing the CVs by department. CVs range from .l in the Magaria 1 department (Zinder) to .75 in the Tillaberi department. Overall, the regions of Maradi and Zinder have the lowest CVs, whereas the regions of Tillaberi and Tahoua have the highest CVs. This suggests that rural households in Tillaberi and Tahoua are exposed to relatively higher production risk. 44. Departmental variation in staplefood cropproduction - as opposed to national -level production variations -plays an important role in grain marketperformance and hencefood prices, which affects households' access tofood. The spatial distribution of drought and production shocks varies substantially across drought and non-drought years. In2000, a severe drought resultedinlower cerealproduction, with national production 21 percent lower than the ten-year average. During this year, only 15 percent of the departments experienced a per capita decrease of more than 50 percent. In 2004, a drought also contributed to a serious reduction in its staple food crop production; with cereal production 12 percent lower than the ten-year average. Although national cereal production was higher in 2004, 25 percent of departments experienced a per capita decrease of more than 50 percent. During years of adequaterainfall, however, less than 3 percent of all departments experience a per capita drop in production by more than 50 percent. This suggests that the spatial distribution of drought is important for grain market performance and prices. Table 15. Official Cereal Imports for Niper, 1998-2004 Commodity 1998 1999 2000 2001 2002 2003 2004 Average Percentage Millet 27,731 4,056 84,004 81,657 4,530 6,699 29,811 34,070 0.20 Sorghum 13,363 1,469 15,872 19,923 950 715 7,470 8,537 0.05 Maize 88,262 56,103 85,116 80,727 12,184 19,232 48,803 55,775 0.34 Rice 91,497 79,035 62,904 70,041 74,884 37,430 59,399 67,884 0.41 Total 220,853 140,663 247,897 ' 252,349 92,548 64,076 145,484 166,267 Source: DPP and Ministry of Agriculture, various years; WFP (2005). 45. Levels of governmental support for the agriculture sector are some of the lowest in WestAfrica. In 2004, the Government of Niger's total expenditures on agriculture were USD 16 million, which represented 1 percent of total governmental spending for that year and .5 percent of agricultural GDP. Only Guinea Bissau had a lower percentage of spending on agriculture (,5 percent) for that year. Importsand Food Aid 46. I n light of the strong inter-annual variation in staplefood crop production in Niger, total food availability depends strongly on commercial imports, imported food aid andpublic stocks.3oOverall, the country's agricultural trade balance remains heavily 30National-level data on private stocks are unavailable. However, according to the cereal marketing survey conducted by UC-Berkeley and four NGOs (CARE, CRS, HKI and World Vision), the duration and magnitude of private stocks of millet and sorghum are fairly limited for both farmers and traders. Traders store an average of 30 days, ranging from 7 days until 60 days. In addition, a small percentage of traders store inter-annually. The average duration of storage for farmers is longer, with an average of 6 months. However, a small percentage of farmers store cereal crops inter-annually. 43 in deficit, and imports of foodstuffs represent a third of the country's import bill. However, data on cereal imports and exports inNiger are highly unreliable, due inpart to the informal trade that occurs between Niger and its neighbors (Benin, Burkina Faso, Chad, Mali and Nigeria (Table 13). Nevertheless, existing data indicates strong inter- annual fluctuation in cereal import^.^' Official cereal imports averaged 166,000 M T between 1998 and 2004, with rice and maize representing the highest percentage of total cereal imports. Official millet imports averaged 34,000 MT between 2000 and 2005, representing approximately 1 percent of total millet availability (Table 14). Millet and sorghum imports through official channels increased dramatically in 1997/1998, 2000/2001 and 2001/2002, coinciding with drought years. Nevertheless, official import data suggest that cereal imports were relatively lower during the first half of the 2004/2005 marketing season. 3'These figures are provided by calendar year, rather than marketing year, which are the most relevant for the analysis of cereal availability. 44 Table 16: Niger: Cereal Productionand Availability, 1990-2005 Average Quantity (thousand tons) Coefficient of variation 1990-1994 1995-1999 2000-2005 1990-2005 1990-2005 Production 2,238 2,363 3,075 2,59 1 0.23 Millet 1,803 1,914 2,339 2,038 0.20 Sorghum 363 388 661 482 0.40 Maize 1 3 4 3 0.64 Rice 72 57 71 67 0.14 Imports 247 173 181 199 0.21 Millet 47 31 15 30 0.56 Sorghum 15 5 2 7 1.17 Maize 76 56 51 60 0.3 1 Rice 41 37 42 40 0.26 Wheat 69 44 71 62 0.25 Net Supply 2,149 2,181 2,794 2,40 1 0.20 Millet 1,579 1,657 2,003 1,762 0.19 Sorghum 323 335 563 417 0.39 Maize 77 59 55 63 0.30 Rice 101 86 102 97 0.15 Wheat 69 44 71 62 0.25 Net Supply/capita (kgs) 263 224 237 241 0.14 Millet 193 170 170 177 0.14 Sorghum 40 34 47 41 0.26 Maize 9 6 5 7 0.40 Rice 12 9 9 10 0.23 Wheat 8 4 6 6 0.30 Net ImportdSupply 11.5% 8.1% 6.7% 8.7% 0.30 Millet 3.0% 1.9% 0.8% 1.8% 0.60 Sorghum 4.5% 1.6% 0.3% 2.0% 1.09 Maize 98.5% 94.6% 93.7% 95.5% 0.03 Rice 40.1yo 42.5% 40.0% 40.8% 0.17 Wheat 100.0% 100.0% 100.0% 100.0% 0.00 Notes:Net supply calculated as production less assumed 15% losses plusnet imports. Source: Calculated from FA0 data. 47. Supply and demand conditions in Nigeria, especially the northern regions, have a major influence onprices and availability of staplefood crops in Niger. Covering an area o f 923,000 km2andwith apopulation of over 100million, Nigeriaproduces significantly larger quantities of millet, sorghum and maize as comparedwith Niger. This i s primarily due to the diversity of Nigeria's agro-ecological zones: annual rainfall averages from 500 mm inthe northern regions to 4000 minthe southeast. As is shown inTable 15, Niger's averagemillet production in2003 and2004 accountedfor 30 percent of total millet production inthe two countries. Similarly, Niger's sorghum production 45 representedonly 9 percent of total sorghum production inthe countries. Almost all of the maize produced is cultivated inNigeria. Thus, Nigeria dominates the sub-regional supply of these cereals, and fluctuations incereal supply and demand inNigeria have important impacts on cereal supply and prices inNiger. While unofficial imports of sorghum and millet from Nigeriaare relatively low, analyses of millet supply, demand and prices suggestthat millet imports from Nigeriamay have reached200,000 MT in2004, equivalent to about 10percent of total net millet supply inNiger (Dorosh, 2007). Table 17: Cereal Production in Niger and Nigeria, Average 2003-2004 Millet Sorghum Other Total Niger 2.7 0.8 0.1 3.6 Nigeria 6.3 8.0 1.2 21.5 Total 9.0 8.8 1.3 25.1 Niger share 30.4% 8.6% 1.1% 14.3% Source: Calculated from FA0 data. 48. Unlike other countries in Sub-Saharan Africa, importedfood aid has notplayed an important role in Niger's total food availability since the mid-1990s. Between 1997and 2007, Niger programmed both imported food aid and food aid from local or triangular purchases. According to WFP (2005), approximately 29,000 MT of food aid are imported annually, comprised of rice (11,000 MT), wheat (4,000 MT), cereals (14,000 MT) and other commodities, such as vegetable oil and soy-fortified bulgur wheat (SFSB). Using official import data, this suggests that food aid represents less than 20 percent of total cereal imports and 1 percent of total cereal net availability. Overall, emergency and development food aid programs are managedand executedby a variety of governmental, non-governmental and international organizations. Food aid programs between 2000-2007 were implemented by NGOs, including Action Contre la Faim (ACF), Africare, CARE, Catholic Relief Services (CRS), Helen Keller International (HKI), Concern Worldwide, Medecins sans Fronti&res, and Plan-Niger; international organizations, including UNICEF and the World Food Program (WFP); and the GoN, including the Cellule de Crise Alimentaire (CCA), the DispositifNational de Prevention et de Gestion des Crises Alimentaires (DNPGCA) and the OfJicedes Produits Vivriers au Niger (OPVN). 49. Food aid in Niger is usedfor both direct distributionprograms andfor moneti~ation,.~~ Since the mid-1990s, food aid for distribution has beenusedinthree types of programs inNiger: emergency response, safety net and Food for Work (FFW).33 32 Monetizationis defined as "imported food aid sold commercially to local buyers in order to generate proceeds for longer-term development programs.. )) Between 2000-2007, the primary monetization programs were financed by USAID Title I1and the US. Department of Agriculture (USDA). During this period, four NGOs (including Africare, CARE, CRS and HKI) imported rice and wheat for sale to local buyers, using the proceeds to implement multi-year agriculture and health programs in Agadez, Dosso, Tahoua, Tillaberi and Zinder. These imports have averaged3,000 MT (rice) and 5,000 MT of wheat on an annual basis. 33There has also beenarelatively small school feeding program inNiger for nomadic populationssince 2000/2001. However, the geographic coverage has been limited to 250 schools (2001) to 500 schools (2005). 46 Emergency responseprograms have typically beenof short-duration (ie, during the year of shock or localized food crisis) and include general distribution, nutritional rehabilitation of childrenunder 5 and the subsidized sales of cereals. Safety net programs have typically provided food aid to extremely vulnerable persons, such as the elderly, persons living with HIV andAIDS (PLHAs), andorphans. Finally, FFW programs have provided food aid to communities during the hungry season inreturnfor work on a community-based infrastructure project, such as health centers, demi-lunes incommunity pastures or farm-to-market roads. FFW projects have typically beenmulti-year projects as part of larger sectoralprograms, althoughthey have also beenusedas part of emergency response efforts during drought years. 34 Inlimitedcases, safety net programs and nutritionrecuperation programs also exist on a multi-year basis. The sale of staple cereals at subsidized prices has beenprimarily executed by the Government ofNiger (via the OfJice des Produits Vivriers au Niger (OPVN) and the Cellule de Crise Alimentaire. (CCA)) using locally purchased commodities (see Chapter 4). AGRO-FOOD MARKETS AND PRICES Market Structure I Figure 9: Key Grain Markets inNiger I 50. Performance of agro-food markets affect households' economic access tofood products. Inlight of the importance of staple cereals (millet and sorghum) for producers' and consumers' welfare, the structure, conduct and performance of food production and marketing has important implications for food security inthe country. Staple food crops (millet, sorghum), cowpea, and rice are traded through an extensive system of traditional markets, which run the length o f the country.35 Figure 9 shows the location . I n e of key grain markets in the country. Cereal markets in Niger are classified into four categories, depending upon their primary role and geographic location: collection; consumer; wholesale; and cross-border. The average distance betweenthese markets is 350 km, but distances between the markets inwhich trade takes place ranges from 10-900 km. While the density of grain markets varies considerably by department, market density appears to be evenly distributed throughout the country, although there are very few markets inthe northernpastoral zones (Agadez, northern Tahoua and Diffa). 34 During previous drought years, some NGOs also implemented emergency FFW projects intargeted villages. Inthese cases, food aid was provided during the hungry season for one year only to households who constructeda community-based infrastructure, such as aroad or awell. 35According to a census of grain markets conductedin2006 and 2007, there are over 2,000 grain markets inthe country. The number oftraders per market ranges from 38 tradersto over 350 traders (Aker2007). 47 51. A variety of market actors are involved in moving cerealsfrom producers to consumers. Market actors involved in the exchange of cereals include: (i) farmers, who produce, sell and buy millet, sorghum and cowpea; (ii)traders, including retailers, intermediaries, semi-wholesalers and wholesalers; (iii) transporters, who are responsible for moving goods via truck, car or boat; (iv) rural and urban consumers,who purchase the final good inrural or urban markets; and (v) state structures, such as OPVN and the CCA. Cereals are first produced by farmers, who sell their production directly to intermediaries located in the village or in the market. Intermediaries in turn sell directly to semi-wholesalers and wholesalers in local retail or wholesale markets. These wholesalers sell to other traders, such as wholesalers located in other markets or retailers located in the same market. Wholesalers can also sell to consumers directly, although usually do not sell in small quantities (i-e., less than 100-kg bags). Retailers in turn sell directly to both urban and rural consumers. Retailers represent the highest percentage of all traders (57 percent) in the market, followed by wholesalers (16 percent), semi- wholesalers (15 percent) and intermediaries (10 percent) (Aker 2007). GovernmentAgro-Food Policies 52. The role of the Government of Niger in cereal production and marketing has been significantly reduced since the 1990s. Inthe two decades following independence (1960), the Government of Niger (GoN) was involved in all aspects of cereal production and marketing. During this time, OPVN was established inorder to manage the purchase and sales of staple cereals (millet and sorghum), functioning as a state-owned monopoly until 1984. With the implementation ofthe government's Structural Adjustment Program (SAP) inthe 1 9 9 0 however, the Nigerien cereal market became increasingly liberalized. ~ ~ By the mid-1990s, OPVN's role was reduced to monitoring the food security situation within the country and managing the country's strategic grain reserves. These reforms had important implications for the functioning of Niger's agro-food sector, particularly in terms of pricing, marketing, and agricultural development and investment (Terpend 2005). During 2000 and 2004, OPVN was responsible for distributing subsidized food to the population, includingvulnerable groups. 53. I n contrast to thepricing supportpolicies of the 1960s and 1970s)food prices in Niger are now determined by marketforces. Domestic cereal prices are influenced by a host of factors, such as domestic supply and demand, commercial imports, import taxes and tariffs, regional trade patterns and market structure. Traders are free to import cereals and export cowpeas and livestock, provided that they are registered and respect all trading and tax procedures. However, depending upon domestic cereal production in a particular year, the GoN may regulate trade flows by limiting either exports or imports of specific cornm~dities.~~ 36While the free exchange of goods and services is permitted within and among West African countries (including between Niger and Nigeria), the GoN may attempt to control cereal exports or imports in particular markets during low production years. Most traders interviewed during the 2005-2007 market survey believethat the exportationof millet and sorghum is prohibited. 48 54. Trade tax regulations in Niger are officially aligned with UEMOA rules. The free exchange of goods and services i s permitted within and among West African countries. Within the Economic Community of West African States (ECOWAS), there are no official export taxes for local products traded among the West African Economic and Monetary Union (WAEMU) countries. Nevertheless, import and export tax regulations inNiger are integrated into the rules of UEMOA. Customs duties still exist for imports of cereals originating from outside of the WAEMU area, namely Nigeria and Ghana.37 Table 18. Average Grain Prices in Niger, Nigeria and Chad, 1996-2006 (CFA/kg) 10-year Variable Obs mean Std. Dev. Average Prices (CFAlkg) 120 145.98 40 Niger (Region) Diffa 120 167.51 47.88 Dosso 120 151.58 40.33 Maradi 120 124.20 38.50 Niamey 120 143.36 30.90 Tahoua 120 155.99 42.39 Tillaberi 120 158.10 41.62 Zinder 120 129.02 41.82 199711998 Prices 12 173.25 33.35 2000/2001Prices 12 160.76 36.89 200412005 Prices 12 194.10 53.40 Benin (Malanville) 81 127.17 39.07 Burkina Faso (Kantchari) 47 125.95 32.17 Chad 120 156.18 44.06 Nigeria 81 137.03 40.49 Notes: Prices are deflatedby the consumer price index. Source: SIMA/Niger Grain Pricesand Market Integration 55. Millet and sorghum prices in Niger are subject to a high degree of inter- and intra-annual variation. Table 16 shows average grain prices in Niger by region, and Figure 10 shows average monthly grain prices in Niger and in the sub-region (Benin, Chad andNigeria) between 1996 and 2006. High-production years inNiger are followed by relatively lower prices (1998/1999, 1999/2000, 2003/2004), and low-production years are followed by relatively higher prices (1997/1998, 2000/2001, 2004/2005). Grain price 37Duringthe trader surveys in cross-border markets, specifically those inNigeria, traders cited excessive official and unofficial import taxes on millet and sorghum as a constraint to marketing among the two countries. 49 levels inNiger closely follow those in Benin and northern Nigeria (Illela, Jibia and Mai Adua), but not those in Burkina Faso and Chad. In addition, the seasonal variation of millet prices is also important. Millet prices range from 20-89 percent higher in August (the hungry season) as compared to October (the harvest period), with an average intra- seasonal price difference of 44 percent. Millet prices increased by 89 percent between October 2004 and August 2005, and by 75 percent between October 2000 and August 2001.38 Markets are therefore subject to relatively high price instability, which suggests that householdsand traders are subject to highto inter-and intra-annual annual price risk. 56. Although grain markets in Niger are relatively integrated, this varies by geographic region and by year. Staple food crop markets in Niger are somewhat integrated, with an average correlation coefficient of .55 for all markets between 1996 and 2006. This is well below the price correlation coefficients computed for other agricultural products in the developing world (Timmer 1974). Nevertheless, the degree of market integration varies on an inter-annual basis, with higher levels of integration during low-production years. The degree of integration between markets in Niger and those in border countries (Benin, Burkina Faso, Chad and Nigeria) follow the same pattern. On average, the degree of market integration between Chad and Niger i s small, averaging .25. Similarly, the degree o f integration between Burkina Faso and Niger i s also quite limited, averaging .47 between 1999 and 2006. The highest degree of integration occurs between Niger, Benin and Nigeria, with correlation coefficients averaging -65. On average, grain markets in Niger are more integrated with Benin and Nigeria during drought years, with the highest degree of integration occurring during the year of the 2005 food crisis. I Figure 10: Grain Prices in Niger and Key RegionalMarkets, 1996-2006 (CFA/kg) I 38During non-droughtyears, prices increase by 10-20percentbetweenthe harvestand hungryperiods. 50 57. Grainprices in Niger respond to supply shocks, andprice movementsin surplus regions within Niger and regional markets are useful' in predicting grain price changes in Niger (Granger causality forecasting tests). Grain prices in Niger follow well-defined paths: they start inproduction centers (ie, Maradi and Zinder) and then spread across the country. This means that grain prices in Niger respond to supply shocks (such as droughts or locusts), rather than demand shocks Figure 11 shows the percentage o f times that a market i s useful for predicting price changes in other markets in Niger. The map shows that markets located in surplus regions (the southern and eastern parts o f the country, such as Maradi, Zinder and southern Dosso) are useful for predicting price changes in other markets. In addition, the cross-border markets o f Malanville (Benin) and Jibia, Illela and Mai-Adua (Nigeria) predict price changes in over 75 percent o f the markets inNiger.39 Notably, price changes in the capital city, Niamey, are not useful for predicting price changes inNiger. Overall, this means that prices in production regions can be useful inpredicting price changes inNiger. This also suggests that markets located in deficit regions - primarily Tillaberi, Tahoua and Diffa - follow price movements in other markets more frequently. 58. I n addition to supply shocks, grain market performance in Niger is partially determined by access to information. Market actors' (farmers and traders) access to price information has an important impact on grain market performance in Niger. Between 2001-2007, the introduction o f cell phone towers throughout Niger enabled traders to search for price information over a larger number o f markets, allowing them to move commodities to excess demand markets. This, in turn, reduced grain price dispersion betweenmarkets and lowered average grain prices (Aker 2008). . 59. Niger needs to import, but arbitrage opportunitiesfor imports vary on a intra- and inter-annual basis. . Duringmost years, import parity prices for cereals (millet and sorghum) from northern Nigeria (Illela, Jibia and M a i Adua) are typically higher than domestic millet prices from October until May, meaning that there are no incentives to import (Figure 12).. This pattern changes between June until August, when domestic prices in Niger are higher than import parity prices, implying that there are positive spatial arbitrage opportunities. During this period - which coincides with the hungry season in Niger -- traders typically import grains from Nigeria.40 This pattern was markedlydifferent duringthe 2004/2005 marketing season. 39Kano, an important grain market located innorthern Nigeria, does not appear useful in forecasting millet prices in a significant number of markets inNiger; it only "predicts" price changes in29 percent of markets. However, price data for Kano are only available for 2003-2007, arelatively short time series for forecastingtests. 40When lookingat trade with Burkina Faso and Benin, net marketing margins are consistently positive, and there are fewer intra- and inter-annual fluctuations. . 51 Figure 11: Key Forecasting Markets for Niger Figure 12: Comparison of Millet Prices in Niger and Nigeria Om-Jan Feb-May Juiie Aug 52 CHAPTER4: GOVERNMENT INSTITUTIONSAND PROGRAMSAND RESPONSETO THE 2005 FOODCRISIS Although the need to support poor andfood insecure households is substantial, safety net programs in Niger are small, designed for emergency food crises, and receive limited government funding. Thefood crisis in Niger in 2005 caught national policy- makers and the international community by surprise. Estimates of national food production during the 2004 harvest suggested only a 12percent decline relative to the ten-year average and gave little cause for concern, However, predictions of a large- scalefood crisis did not occur until April 2005. I n one sense, the major cause of the crisis was the extremely low level of household income, which makes households vulnerable to serious malnutrition even during normal production years. During 2004/2005, two additional shocks precipitated the food crisis: 1) drought and corresponding production shocks in 2004, which lowered household-level production and incomesfor the 2004/2005 marketing season; 2) a sharp rise infood prices in May 2005, which reduced access tofood for all net buying households. The response of the GoN and the international community (donors, NGOs and WFP) consisted primarily of the use of food aid for emergency distribution, FFW, the sale of approximately 40,000 MT of cereals at subsidizedprices and "Cash-for- Work" (CFW programs. I n spite of these efforts, in some of the worst-affected regions, the gross and child mortality rates reached well above international thresholdsfor humanitarian crises. Understanding the factors that contributed to the 2005 food crisis is important for preparingfor and responding tofuture food crises. INSTITUTIONAL MECHANISMSFORRESPONSETO FOOD CRISIS 60. Since the Sahelian famines of the 1970s, a variety of governmental and non- governmental institutions have been established to prepare for and respond to food crises in Niger. he Niger's Food Production Office (Ofjce des Produits Vivriers du Niger OPVN) was established inorder to managethe purchaseand sale of staple cereals - (millet and sorghum) in the country. Between 1990-1998, OPVN focused on monitoring the food security situation andmanaging the country's strategic grain reserves (definedas the stocks held by the end of the calendar year). During this time, Niger also drastically reduced its strategic grain reserves, previously an important part of the country's emergency response strategy. In 1992, for example, Niger's strategic grain reserves were 85,000 MT, and fell to 12,000 MT in 1998. Since 2000, the OPVN's strategic grain reserves have averaged27,000 MT. 53 Figure 13: Quantities of Food Aid Obtained and Distributed by OPVN & CCA during 2001-2007 140,000 I20,OOO I20,OOO l00.000 100,000 80,000 80,000 60,000 60,000 40,000 40,000 20,000 20,000 0 t 0 2001 2002 2003 2004 2005 2006 2007 2008 1-Sources utilization * End Stocks OPVN -*-End Stocks (OPVN+CCA)1 Notes: "Sources" refer to the total quantity o f food aid obtained (as foreign aid or other sources) by OPVN and CCA during a particular year. "Utilization" refers to the total quantity o f food aid distributed by both agencies. End stocks are the quantities o f food aid remaining at the end of the fiscal year. 61. In 1998, the GoN created a new structure for managing food crises, the National Body for the Prevention and Management of Food Crises (Dispositif National de la Prevention et la Gestion des Crises Alimentaires - DNPGCA). The objective of this new structure was to reduce food insecurity by ensuring better coordination among the various organizations. The DNPGCA, overseen by a National Consultation Committee (Comission Mixte de Concertation, or CMC) is responsible for overseeing and coordinating (i) the interventions related to the country's strategic grain reserves, and (ii) funds provided to governmental structures. Under the broader donor umbrella of the CMC, the four local Government structures involvedinthe preparation of and responseto food crises include: the Cellule de Crises Alimentaires (CCA), primarily responsible for decision-making with respect to food crises; the Systeme d'Alerte Pre`coce (SAP) or local early warning systems, responsible for collecting and analyzing data on food insecurity; OPVN, responsible for managing the country's strategic grain reserves; and local market information systems for agriculture and livestock (SIMA)(see Box 4). 62. I n addition to the DNPGCA, there are a number of non-governmental and international actors involved in the prevention of and response tofood crises. These include non-governmental organization, such as Africare, Catholic Relief Services, CARE, and WVI; international bodies (WFP, FAO, UNICEF), regional structures (Le., CILSS) and donors (ie, the European Union, USAID's Famine Early Warning Systems 54 Network (FEWSNET). These institutions are all involved in the prevention, preparation and management o f response to food crises. In this context, these organizations collaborate with the GoN in establishing famine early warning systems, executing activities to mitigate the effect o f food crises; and responding to food crises once they occur. Box 4: National Mechanism for the Prevention and Management of Food Crises in Niger: the Dispositif National de Prbvention et de Gestiondes CrisesAlimentaires (DNPGCA) Since 1998, Niger's government and major food aid donors have managed a National Mechanism for the Prevention and Management of Food Crises in Niger, the DispositifNational de Pre'vention et de Gestion des Crises Alimentaires (DNPGCA). Its primary mission is (i) to help the Government build cereal and financial reserves designed for food aid during crises; and (ii) to implement actions supporting the populationduring food crises. Management: The DNPGCA is managed by the National Consultation Committee (CMC) for State donors, which is responsible for (i) defining the operational goals, and (ii)handling programming, coordination, and management of resources made available to prevent and moderate food insecurity. The Prime Minister and his office representthe Government. The Technicaland FinancialPartners consist of 12 members including 5 from the EU: Germany, Belgium, France, Italy, and the European Commission; 1 from Europe outside the EU: Swiss Confederation; 2 from outside Europe: Canada and the United States; and 4 United Nations bodies: UNDP, WFP, FAO, and UNICEF. A framework agreement, signed by the State and all 12 partners in 2005, defines the methods of operation and for the utilization of DPNGCA resources. Within this framework, two organizationswere created: 9 The National ConsultationCommittee (CMC): This is a strategic organizationat the level of the Prime Minister, Ambassadors and Representativeswho meet once per year to assess the status of the previouscampaignand outlinethe major lines for interventionfor the comingyear. 9 The CRC, ComitC Restreint de Concertation (Special Advisory Committee): This is a technical organizationat the level of the Prime Minister's cabinet and of the delegates from the Financial and Technical Partners who negotiate and approve programs (unanimously), authorize expenditures and manage the available funds, assess the operations implemented, and monitor the CMC's actions. DNPGCA Tools: The DNPGCA has mobilizedtwo tools: 1, The Stock National de Rbserve, SNR (National ReserveStock): The SNR is used only in years o f national or regional crisis to implement a rapid response while waiting the mobilization of internationalhumanitarian and nationalsolidarity aid. The SNR's optimal volume is 110,000 tons of cereals. The SNR is composed of: 9 The Stock National de SCcuritC. SNS (National food security stock): A physical stock of 80,000 tons of cereals (millet, sorghum, maize, and rice) stored in the warehouses of the Office des Produits Vivriers du Niger (OPVN). This stock is usedto avoid the developmentof speculationby implementingsubsidizedsales. 9 The Fonds de SCcuritC Alimentaire. FSA (Food Security Fund): A financial reserve to purchase 30,000 tons of cereals in case of major crisis and low national stocks. Purchase decisions will be made by common agreement with the partners and entrusted to the OPVN, which will receive a lump sum remuneration of 10,000 FCFA per ton. 2. The Fonds Commun des Donateurs, FCD (Common Donor Common Fund): This fund is the frst level of resourcesthat can be mobilized for financing actions to respond to small food crises. This fund will primarily finance actions for the preventionand moderationof crises (Cereal banks, cash for work for soil reconstitution, subsidizedcereal sales, targeted free distributionof cereals). Primary DNPGCA Bodies: Aside from the CRC, the other bodies are: 9 The Cellule de Crises Alimentaires, CCA (Food Crisis Cell): This is the executive body of the DNPGCA and is responsible for the implementationand monitoring of the programs retained by the CMC. The CCA (created in 1998) is primarily responsible for (i)ensuring the proper coordination of assistance programs; (ii)ensuring the proper operation of the DNPGCA; (iii) 55 assessing food aid needs; and (iv) keepingrecords of food aid received and actions implemented to moderatecrises. l+ The Cellule de Coordination du Systkme d'Alerte Prdcoce, SAP (Early Warning System Coordination Cell): responsible for collecting and processing information from various information systems (SIMAgricole, SIMBetail, Enqudte permanente d'estimation des rdcoltes, (EPER (agriculture production survey), etc.). Additionally, SAP monitors vulnerable areas on a monthlybasis. l+ The Office des Produits Vivriers du Niger, (OPVN (Food Crop Authority): it is responsible for maintainingthe Stock National de Reserve (storage and renewal of the stock) and the logistical managementof food aid (distributionof services, commercializationof food aid destined for sale, > and processingof stocks). The InformationCenter (IC): Attached to the Prime Minister`s Cabinet, this is the distribution and communicationbody for the DNPGCA. Source: GoN and FTPs. Early Warning Systems 63. A variety of local, regional and international famine early warning systems exist in Niger to collect, analyze and disseminateinformation concerningpossiblefood crises. These include the governmental early warning system (EWS, or SAP) and Agricultural Market Information System (SystBme d'Informations sur le Marche`Agricole - SIMA); the USAID-funded famine early warning system network (FEWS NET); international organizations, such as World Food Program (WFP), UNICEF and various community-based early warning systems establishedby international NGOs; and CILSS, a regional body. In general, these systems rely on rainfall data, agricultural production, remote sensing and market information (i.e,, prices) to generate predictions regarding the food security situation inthe country. 64. Despite important efforts and available climatic, production, price and food security data, the national, regional and international early warning systems were not able to complete their mission: predicting in an accurate and timely manner the magnitude, scope and location of the food crisis in 2005. While some institutions warned of a food crisis in drought-affected regions as early as December 2004, these predictions were based primarily on data concerning drought and production shocks. In addition, once the most vulnerable regions were identified in 2005, this process was not without problems.41 This was related to a variety of factors, includingthe following: > each organization -even among the early warning systems -- uses its own Different conceptions of food insecurity and vulnerability. Ingeneral, framework for food security and vulnerability. While there is considerable overlap among these frameworks, there i s not a common understanding of these concepts among different actors. Consequently, developing > common andtransparent early warning indicators is difficult. Early warning system indicators exist, but there are no thresholds to trigger action. Unlikemalnutritiondata -- where there are clear thresholds for a crisis -thresholds for the drought, production or price 41Government of Niger, Prime Minister's Office, EC, SAPICG. Situation Alimentaire dans les Zones Vulnkrables (May 2005). 56 data do not exist. Although there i s substantial data available inNiger - and the early warning systemorganizations (CCA, SAP, SIMA and FEWSNET) do an excellent job of analyzing these data - -the analyses focus on production and price trends, rather thanthreshold levels to trigger a crisis. Such thresholds are requiredinorder to determine whether a crisis will occur. CAUSES OF THE 2005 FOODCRISIS 65. I n 2004, Niger experienced a severe drought that contributed to a 12 percent reduction in cereal production relative to averageproduction of theprevious ten years. Between January and June 2005, millet prices rose to a level that was 25 percent higher than the ten-year average, with the price of a 100-kg sack of grain representing approximately 37 percent of national per capita income during the height of the food crisis. By June 2005, an estimated 2.4 million Nigeriens were affected by severe food shortages, with more than 800,000 of these classified as critically food insecure (USAID FEWSNET 2005). Although the 2005 food crisis was not of the scale o f the 1968-74 or 1983-84 famines, the gross mortality rate reached 1.5 deaths per 10,000 per day in targeted regions, with a child mortality rate of 4.1 deaths per 10,000 per day (MCdecins sans Frontikres 2005) (Box 5). Both of these are above the threshold mortality rates used to define a humanitarian Box 5: History of Food Crises inNiger During the first post-colonial decade, the country succeeded in reaching food self sufficiency due to low demography, good rainfall, and acceptable crop yields from rudimentary agricultural systems. Until mid 1960s, the cereal balance sheets were in excess: a cereal surplusof 49 kg per personper year was recorded between 1960 and 1965.However, by late 1960s, the country registeredan important food shortage for the agricultural campaign of 1968-1969. In 1972 and 1973, food production dropped dramatically: food shortage was estimated at about 400, 000 tons in 1973. The main causes of these food crises were: the lack of valorization of marketable agro-pastoral products; fiscal pressure; and the decline in cash crops. All these factors combined with the effects of harsh climatic conditions resulted in the most serious food crisis that the country had faced since its independence. During the 1976-1982 period, conditions were favorable (substantial uranium revenues, fiscal deficit reduction with the suppression of capitation taxes, increase in support to farmers, and favorable climatic conditions). Between 1975 and 1982, uranium replaced peanuts in terms of foreign currency injection into the development of the public sector. From 1982, the State disengaged from the market and production sectors: producers were made responsible through sale and production co-operatives,and seeds and pesticidessubsidies were eliminated. In 1983-1984, the country faced a new food crisis, due to the cumulated effects of the deterioration of climatic conditions, the structural adjustment policy and the closing of the Niger-Nigeria border (1983). Another food shortage comparable of that of 1973-83 took place due to aheavy drought. In 1984- 1985, the Nigerien farmers, who never knew a stable, sustainable and good food situation, again faced a big crisis. The needs for food aid were estimated at 410, 000 tons (in addition to imports of 125, 000 tons). 1990-1991 and 1993-1994: during these years, people faced food shortages. The 1996-1997 crisis: This crisis is still remembered by the Zarma (western part of Niger) and the Haoussa (central part of Niger) regions. The important campaign shortage resulted from the severe drought period of August 1996. 42The crude mortality rate used to define a humanitariancrisis is one deathper 10,000 people per day. 57 2001, Niger was part of atrans- regional food crisis, which was particularly severe in many regions due to depletionof both farmers' food stocks and national security food stocks (the national security stock dropped from 200.000 to 100.000 tons). Food scarcity was observed in almost all the regions.The price of a 100 kg bag of pearl millet reached CFA 25, 000 in many regions of the country. Even other neighboringcountries suchas Mali, Burkina Faso and Nigeria were affected. The 2004-2005 crisis: even though it was not as severe as that of 1973-1984,it exceeded that of 1996-1997.One of the consequences of food crises is the increasingevolution of cereal imports. Since 1973, staple food importing (millet, sorghum) has considerably increased. But it is mainly rice that constitutes the major part of the introduced products (more than 50% ofthe imports of cereals). 66. As is the case with chronic and seasonalfood insecurity in Niger, the causes of the 2004/2005food crisis are multiple and complex. Although rural households inNiger are susceptible to a variety of risks, including droughts, floods, pest attacks, health epidemics and economic shocks, not all of these shocks result in food insecurity or food crises. Whether or not the shock results in food insecurity depends on households' assets, risks and risk management strategies. The inter-related factors that contributed to the food crisis in2004/2005 include production and marketing factors that have an impact on food availability and market prices; the extended level of poverty and low level of assets the limited household's income and access; the poor level of heath and sanitation and ultimately the delays inthe relief response. Domestic and Regional Food Availability in 2004/2005: Drought, Production Shocks and CommercialImports 67. The lower domesticfood availabiliy was a result of widespresd regional droght and market prices. The food production decline was certainly an important factor, but staple food crop production was even lower in the 2000 drought year. Yet, according to the local early warning systems, a severe food crisis did not occur in2000/2001.43Thus, it is important to consider other factors that contributed to the 2005 food crisis. These include the large number of departments affected by production shocks in 2004 (that dramatically reduced farm household incomes in these areas); production shocks on key forecasting markets in Niger and Nigeria; changes in the quantities imported from neighboring countries (especially Nigeria) due to unfavorable marketing conditions. More specifically, the inter-related factors that contributed to the food crisis in2004/2005 includethe following: > Factor #I: Drought contributed to up to a certain extent to the loss of production. Between 1996-2006, Niger experienced droughts in 1996, 2000 and 2004, with annual rainfall lower than one standard deviation from the mean (Nicholson et a12000). 32 percent of departments experienced drought in 2000, compared with 50 percent of departments in 2004. Nevertheless, staple food crop production was 10 percent higher in 2004 as compared to 2000; total production of cereals (millet, sorghum, fonio and rice) was estimated at 2.05 million MT in 2000 (195 kgkapital), as compared with 2.63 43Although a severe food crisis did not occur in2001, householdsdid suffer from seasonal food insecurity during the hungry season (soudure) betweenMay-August. . 58 million M T in 2004 (218 kg/capita). This suggests that drought is not > perfectly correlated with levels ofper capita grain prod~ction.~~ Factor #2: A Higher Percentage of Departments wereAffected by Drought in 2004. In2000, only 15 percent of the departments experienceda per capita decrease of more than 50 percent. By contrast, in 2004, over 25 percent of departmentssuffered aper capita decrease ingrain production ofmore than 50 percent.45 This means that the percentage of departments affected by production shocks -as opposedto national-level production - is important for market performance, grain prices and food crises inNiger. 9 Factor #3. Key Markets in Niger and Nigeria wereAffected by Drought in 2004. In addition to the percentage of markets affected by drought, which markets are affected is also important. Since prices in Niger respond to supply shocks, drought in surplus-producing regions (such as Maradi and Zinder) will have a larger impact on price levels than drought in deficit regions (such as Tahoua and Tillaberi). In 2004, the markets affected by drought were key surplus-producing markets. Average (deflated) milletprices in these markets were 15 CFNkg higher in October 2004 as compared to October 2000.46 Therefore, while total and per capita grain production was higher in 2004, price levels inkey forecasting markets were 17percent higher at the start of the 2004/2005 marketing season. This suggests that monitoring prices on these markets during the 2004 harvest could have served as an > indication of a potential food crisis. Factor #4. Prices were Higher in northern Nigeria, Making it Unprofitable to Import. Following the 2004 harvest, the FA0 and WFP estimated that there would be a cereal "deficit" of 278,350 MT in 2004/2005, taking into account commercial imports (FA0 and WFP, 2004).47 This prediction assumed that Niger would be able to import from neighboring countries (Benin, Burkina Faso, Mali and Nigeria) in order to reduce its cereal deficit. During most years, domestic millet prices in Niger are lower than prices in Nigeria from October until May. This pattern changes from June until August, when prices in Niger are higher than those in northern Nigeria. During this period, traders typically import grains from Nigeria.48 The situation during the 2004/2005 marketing season was markedly different. Figure 12 shows the millet prices between Jibia (Nigeria) and Maradi (Niger) 44There was not a strong correlation betweendrought-affected and food crisis zones during the 2004/2005 '` marketing season The correlation between food insecure and drought-affected regions in 2004/2005 was .02. 45 During non-drought years, less than 3 percent of all departmentsexperience a percapita decrease ingrain roduction greater than 50 percent. The forecastingmarkets included those markets that Granger-causemore than 75 percent of markets in Niger: Gaya, Maradi, Guidan Roumdji, Tessaoua, Tounfafi and Zinder (Niger); Jibia, Illela and Mai Adua (Nigeria); and Malanville (Benin). 47Thenorm used for cereal consumption by the GoN is 240 kglcapita. This is based upon a population- based weighting of the 200 kglperson norm for nomadic populations and the 250kglperson norm for agricultural populations. In 2004, the FAO/CILSS/PAM 2004 mission adopted a norm o f 239 kg of dry cereals (millet, sorghum) and 20,5 kg of wheat and rice. 48When lookingat trade with BurkinaFasoandBenin,net marketingmarginsare consistently positive,and there are fewer intra-and inter-annualfluctuations.. 59 for several years. Between October 2004 and May 2005, millet prices in Niger were lower than those in northern Nigeria, as is the case during most years. However, millet prices inNigeria remained higher thandomestic millet prices for the entire 2004/2005 marketing season. Regardless of the reasons for higher millet prices in Nigeria during this period, the impact i s clear: It was not profitable for traders to import grains from Nigeria during the year of the food crisis (although it was profitable in2001). While prices inBeninand Burkina Faso were lower than those in Niger during 2005, it i s unclear whether sufficient quantities of grains could have been imported from these countries. 49 68. Simulations of the effects of the shorifall in Niger'sfood production in 2004-05 on market prices suggest that a decline in net imports was partly responsiblefor the huge spike in millet prices. If there had been no change in net trade in cereals in unofficial markets in 2004-05, the 26 percent decline in millet production in that year would have resultedin an estimated 39 percent increase in the average real market price of millet (Table 17).50The actual real price increase in Maradi in the October 2004 - September 2005 crop marketing year was even larger -- 58 percent. This suggeststhat, in addition to the production shortfall, net availability was further reducedby a reduction in net imports, a conclusion which i s consistent with interviews with traders (Care, CRS, HKIandWVI, 2007). The normal flow of informal sector millet imports from Nigeria is likely on the order of 100-200,000 MT per year. The model simulations suggest that net imports may have fallen by 200 thousand tons (equivalent to 10 percent of net millet supply (1.98 million tons) in 2003-04 and 15 percent estimated net millet availability from production only in 2004-05). In combination with the 26 percent drop in millet production, a decline in net imports of this magnitude would have been enough to result inan increaseinreal millet prices of 56 percent, very close to the actual price increase in Maradi inthis period.'l 49Although official import data is of dubious quality, there are severalsources that suggest that the quantity of cereal imports fell dramatically during the first half of 2004/2005, particularly from Burkina Faso and Nigeria. WFP (2005) notedthat the level of imports between October 2004 and June 2005 was 35 percent of the five-year average, Le., 25,000 MT as compared to 71,000 MT. This was not the case during the 2000/2001 marketing year. 50Note that the model simulations take into account the effects of lower householdincomes inthat year that reducedmarket demand for millet. " Becauselevelsofnetimports, demandparametersand the extent of market integration are uncertain, these estimates can only provide a rough approximation of the effects of changes in production and net imports on market prices. See Dorosh (2008) for sensitivity analysisregarding these assumptions. 60 Table 19: SimulatedPriceEffectsof ProductionShock and Net Imports, 2004/05 Sim 2: Sim 3: Sim 1: Production SimCross- 2 with Production Shock and Shock Lower Price Only Imports Effects 2003104 2004105 2004105 2004105 Production (mn tons) 2.745 2.038 2.038 2.038 Production Shock __- -26% -26% -25.8% Net Production Change (mntons) _-_ -0.460 -0.460 -0.460 Net Trade (mn tons) 0.200 0.212 0.012 0.012 Total Supply (mn tons) 1.984 1.537 1.337 1.337 Change in Supply (percent) --- -22.6% -32.6% -32.6% Per capita consumption (kglperson) 147.0 113.8 99.0 99.0 Simulated Real Price Change (percent) _-- 39% 56% 0.618 Actual Real Price (Maradi, October-Sept) --- 58% 58% 0.578 Source: Model Simulations, (Dorosh, 2008). Access in2004/2005: Incomes,GrainPricesand Cereal-LivestockTerms of Trade 69. The drought and subsequent production shock in 2004 not only reduced domestic supply, but also affected households'purchasing power. The drought reduced households' production of staple food and cash crops (such as cowpea), thereby resulting inlower householdincomes for the 2004/2005 marketing year. Infact for the country as a whole, a 0.6 percent drop in real GDP (a proxy for the change in total household incomes) and a 3.3 percent increase in the population implied a 3.9 percent drop in per capita incomes in2004/2005. 70. Grain price increases in 2004/05 weregreater than during the 2000/01 drought, suggesting a sharper decline in household access to food for net food purchasers. Average grain prices in 2004/2005 were 25 percent higher than the 10-year average. By contrast, prices during the 2000/2001 marketing season were only 12 percent higher than the IO-year average. Inaddition, grain prices infood crisis regions were`8 percent higher as compared to non-crisis regions (Table 18). Although a higher percentage o f markets infood crisis regions were affected by drought, the regions were similar with respect to most other characteristics (Aker 2007). 61 Food Crisis Non Food Crisis Difference in Difference in Regions Regions Means Distributions D- P- Mean (sed.) Obs Mean (s.d.) Obs s.e. statistic value Price dispersion between markets (CFAkg) 29.6(22.7) 1392 24.99(19.7) 3828 4.602***( 1.6) 0.0872 0 Grain Price (CFAkg) 164.96(43.4) 183 151.8(39.0) 155 13.10***(6.63) 0.148 0.051 Notes: Datafrom SIMA/Niger. *i s significant at the 10% level, ** significantat the 5% level, ***is significant at the 1% level. All prices are in 2001 CFA. 71. Although grain prices 130 during the 2004 harvest were ,20 relatively higher as compared A to 2000, they initially followed 1 1 0 - A a similar pattern to that of A other drought years. i w . A A However, prices increased A I t iw Pf significantly during the last ' t t four months of the 2004/2005 A t marketing season. Figure 15 BO . t Uarndi shows the seasonal price t TOuni.8 ,o. Ovidvl Rourndji Lindcr change for cereals in domestic Mu Adua and cross-border markets for 6 0 1 Niger between 1999 and 2006. The greatest intra-seasonal price Figure 14: Grain Price Levels of Key Forecasting Markets dllrino thPHarvest Pprind Figure 15: Intra-Seasonal Change of Millet Prices by Year __________ .*............ ." *........, a +__--. I"" _ _ _ _ _ _ _ -------- __---__---------- ___+__- 62 Utilization: Diseases, Health Services and Behavior 72. Although the immediate causes of the 2004/2005 food crisis were related to drought, production shocks and price increases, the situation was exacerbated by structural factors, especially those related to the poor health and nutritional status of the Nigerien population. Poor nutritional outcomes are determined not only by food intake, but also by a variety o f other factors, including infectious diseases and the care and management o f these diseases. Based upon the data from targeted health and nutrition surveys in 2005 (i.e., HKI, MSF and UNICEF, among others), there i s little to suggest that the 2005 food crisis was due to a health shock, such as an outbreak o f meningitis, malaria, typhoid or other diarrheal diseases. The DHS/MICS survey o f 2006 provides important information on the health and nutritional status o f the population. Figure 16: Distribution of Children with an Infectious Illness and Ariitr Ilndrr-niitritinn. hv AUP Crniin PercentaQe --CWasting -C-Diarrhea .. ..ARI --X-Fever 0-5 6-11 12-23 24-35 36-47 48-59 Source EDS el MlCS (2006) ' 73. Infectious diseases,particularly those associated with limited access to potable water and health-seeking behaviors, are of particular concern in Niger. Infectious diseases affect the growth and nutritional well-being o f young children. Figure 16 shows the relationship between infectious diseases and under-nutrition (wasting) in Niger, according to age group. The relationship between the two is clear: (i) a sick child loses weight as a result o f appetite loss, reduced absorption and higher energy requirements to fight off the disease; and (ii)a malnourished child has lower immunity, therefore is at higher risk o f falling illand getting more severe infections. Reduced consumption and absorption result inunder-nutrition. 74. While there is no evidence that child care practices significantly changed during the 2004/2005 food crisis, limited access to sufJicient quality and quantity of food exacerbated thepoor health and nutritional status of thepopulation. The duration and severity of a child's illness is partly determined by preventive practices and child care behaviors once the child becomes ill.According to DHS/MICS survey (2006), child care practices during illness were poor: only 36 percent o f children under five suffering from diarrhea were given more fluids than usual, 29 percent o f children were treated with 63 an oral rehydration therapy, and 6 percent were offered more food than usual. In addition, only 13.5 percent of infants aged 0-6 months were exclusively breastfed, an important protective measureto avoid exposure to pathogens (DHS/MICS, 2006).s2 RESPONSESDURINGTHE FOODCRISIS 75. I n 2004/2005, a varie o governmental, local and international organizations respondedlo thefood ~risis.~ PThese f interventions involved (i)(i) based programs, Food such as free distribution of food, therapeutic feeding programs (ii)Food for work programs (FFW); (iii)the sale of cereals at subsidized prices; (iv) Other interventions, such as Cash-for-Work (CFW) programs and Seed Vouchers and Fair (SVF) programs. In spite of these efforts, in some of the worst-affected regions, the gross mortality rate and childmortality rates rose well above international thresholds for humanitarian crises. > Food based programs: During the 2005 food crisis, food aid resources were used in two types of programs: vulnerable feeding programs and cereal banks. Vulnerable feeding programs during the crisis included (i) general distribution programs, which provided general rations to entire communities, regardless of gender, age or nutritional status; and (ii)therapeutic feeding programs, which involved providing intensive curative care for severely malnourished recipients, primarily children under 5,54 Finally, insome communities, food aid was provided to replenishthe stocks of existing cereals banks, or new cereal banks were created (inaddition to aninitial stock). > FFW proprams during the crisis provided food resources (usually to all households in the community) in return for work on community-based infrastructure projects, such as demi-lunes in community pastures or farm-to- marketroads. > Cereal sales at subsidized prices. Along with community-based interventions targeted at individual households, the GoN also organized the sale of cereals at subsidized prices during the 2005 crisis. The GoN sold approximately 40,000 MT of subsidized cereals in Niamey and other regions between February and June 2005 (WFP 2005b), with 12,750 MT sold inFebruary-April 2005, and 30,000 MT sold betweenApril-May 2005. Both of these sales used food aid resources.. '*Infant and child care and feeding practices are often closely associatedwith aspects of motherhood, such as teenage motherhood, short birth intervals, and maternal mortality. While data on health behaviors and motherhood status are rare, maternal mortality in Niger is estimated at 1,600 maternal deaths per 100,000 live births, one of the highest maternal mortality rates inthe world. (DHSMICS 2006). 53NGOs that implemented food aid programs between 2000-2007 includeAction contre la Faim, Afiicare, CARE, Catholic Relief Services (CRS), Helen Keller International (HKI), Concern Worldwide, Medecins sans FrontiBres, Plan-Niger;internationalorganizations includeUNICEF, the World Food Program (WFP); and the GoN, includingthe Cellule de Crise Alimentaire (CCA), the DispositifNationalde Preventionet de Gestion des Crises Alimentaires (DNPGCA) and OPVN. 54The establishment of centers for therapeutic feeding should be considered when the rate of malnutrition amongunder-fivesexceeds 10%and the capacity of existing facilities is exceeded. 64 9 Other emeraency responseprograms: Inadditionto food-basedinterventions, two other types of emergency interventions were implemented during the 2005 crisis: Cash for Work (CFW) and seed vouchers and fairs (SVFs). The GoN and the IRC (International Rescue Committee) both implemented CFW programs in targeted areas; based upon the FFW approach, community members were provided with cash payments in return for work on a community-based infrastructure projects. Additionally, CRS implementedSeed Vouchers and Fairs (SVFs), seed fairs in particularly food insecure areas. Participants were provided with vouchers (worth a specific amount) to purchase the seeds; seed sellers included farmers and/or traders in the area. The primary idea behind the SVFs was to mitigate the medium-term impact of the food crisis by providing households with seeds for the next harvest. However, rather than importing seeds, the seed fair sought to facilitate exchangesbetween buyers and sellers. 76. Food distributed and sold during the food crisis came from two primary sources: importedfood aid, mainlyfrom the USand the EU; and local and triangular purchases, the latter of which was provisionedfrom Nigeria. Although quantitative data on the location, magnitude, timing and price of these purchasesare not available, the largest local purchase appearsto be the GoN's purchaseof 40,000 MT betweenFebruary- May 2005. WFP purchased 5,800 M T of cereals locally in 2004/2005, including 3,800 MT of millet inOctober 2004 and 2,000 MT of imported rice inJuly 2005. Inaddition to these local purchases, WFP also bought approximately 10,500 MT o f sorghum in Nigeria. Several other NGOs also acquired food aid via local and triangular purchases in 2004/2005, although the quantities and locations of these purchasesare difficult to verify. Nevertheless, these quantities appear to be relatively small in magnitude (Le., less than 1,000 MT) 77. Understanding the impact of local purchases, and developing guidelines and criteriafor using local or triangular purchases as comparedto importedfood aid, is an importantpolicy concernfor Niger. The impact of these local and triangular purchases on domestic supply and food prices in Niger is difficult to measure without detailed information on their location, magnitude and timing. Iflocal purchasestook place inkey "forecasting" markets in Niger or Nigeria - such as Maradi, Jibia or Mai Adua - and significant quantities were involved, then the impact on supply and hence local prices could have beensubstantial. 78. Criteriafor targeting vulnerable regionsfor the implementation of emergency programs were unclear. Due to the lack of explicit and transparent criteria (or thresholds) for determining a food crisis and identifying vulnerable regions, the areas most affected by the food might have not been reached in 2005. Although a list of vulnerable villages was first produced in May 2005, this initial list did not coincide with qualitative and quantitative evidence observed by international and non-governmental organizations. This prompted considerable friction between the GoN and the international community. Although numerous institutions attempted to update this information by conducting additional food security assessments, to-date, a list identifying the most severely affected regions during the 2005 food crisis is still not available. 65 GOVERNMENT SAFETY NET AND SOCIAL PROTECTIONPROGRAMS The mechanisms for preventing and responding to food crises discussed above represent only one o f a range o f interventions that comprise a social protection system. Social safety nets (safety nets for short) are non-contributory transfer programs targeted to the poor and to individuals vulnerable to poverty and to shocks. These programs have the simultaneous goals o f (i) providing assistance to households in poverty, thus contributing to a reduction inthe impact o f poverty and to helping them get out o f poverty, and also to help households face the impact o f shocks; and (ii) reducing the effects o f globalization, macroeconomic shocks, and structural changes, and thus contributing to more sustainable These are programs such as: Cash transfers or food coupons, by category or subject to resource conditions, such as family allowances or social pensions. In-kindtransfers, school meal programs or supplements designed for mothers and children are the most widespread, but they also include distribution o f meals ready to eat, school supplies, uniforms, etc. General Price subsidies, often for food or energy, targeted to households. Employment through labor-intensive public works programs, sometimes called "workfarel' (conditional transfers). Cash or in-kindtransfers to poor households, subject to meeting specific educational or health conditions. 79. Exemption from duties for basic services, health services, education, public services, or transportation. 80. Social safety net programs in Niger have primarily been implemented by governmental, non-governmental and international actors on an annual basis. These programs inNiger include the subsidized sale o f cereals, free food distribution, and, in a few cases, food or cash for work. In addition, there are a few safety net programs that support the access o f the poor to health and education programs, such as school feeding programs and nutritional support. 81. Although the need to supportpoor andfood insecure households is critical in Niger, government spending on food-crisis safety nets represents only a small percentage of total government expenditure on social programs. Between 2000 and 2006, education represented the largest percentage (54 percent) o f total governmental expenditure on social programs, followed by health (32 percent) and social safety nets (14 percent). These percentages remained relatively constant over the time period, with the exception o f 2001 and 2005. In general, social safety net programs remain a " See Grosh et al. (2008) for a complete description of social safety nets. 66 and Public Works (HIMO) 0.1 3.0 7.3 8.0 5.3 12.9 Preventio n and managem ent of food crises 95.8 53.6 60.3 19.5 79.6 50.5 Others 0.5 9.2 24.8 19.4 0.6 2.5 82. Since the late 1990s, governmental safety net programs have been primarily used to help poor households respond to food crises, and most of the programs are financed by external resources. Between 2001 and 2006, more than 80 percent o f the safety net expenditures of the Government were used for preventing and managing food crises. 67 percent of funding for these programs came from external resources. Thus, in that period, 70 percent of total spending on safety nets - Government and external - focused on food crises, with external financing accounting for the majority of spending (Table 20). While these data may exclude emergency and non-emergency food aid distributed by other non-governmental and international actors during this time period, two things are clear: (i)the government has only limited resources available for safety nets; and (ii)spending on safety nets strongly favors emergency response, particularly food crises. ~ 56 Data providedon governmentalexpenditures on safety net programs over time, as well as the percentage ofthe total budget expenditures needto be reviewedand double checked with the MEF. 67 Table 22 -Sourceof financing of Safety nets 2001-2006 , Total 4.0 I 6.4 I 80.0 I 9.5 I 95.9 I 100.0 83. Since most of the safety net programs are implemented duringfood crises, they are managed by institutions responsible for food emergency interventions. A large percentageof safety net program have beencoordinated by CCA and implementedby the CCA, OPVN and a variety of NGOs (Section 1). Table 21 shows the quantities of in- kind programs managed by the GoN, by year and by type of program, between2001 and 2007. The CCA is primarily responsible for FFW and cereal banks, whereas both agencies are involved inthe sales of cereals at subsidized prices and general and targeted distributions. Over the past seven years, the greatest quantities were distributed in 2000/2001 and 2004/2005, the years of drought. Therefore, during non-drought years, there were relatively limited safety net interventions available in the country, despite the fact that over 50 percent of the population i s food insecure each year. CCA and OPVN rely on strategic grain reserves, imported food aid and local purchases in order to obtain food aid. Between 2000 and 2007, approximately 30 percent o f all food aid used for governmental safety net programs was obtained via local purchase^.'^ Table 23. Governmental (Emergency) Safety Net Programs, by Year and by Program Year Subsidized General Cereal Livestock CFW (CFA) FFW (MT) Seeds Sales (MT) distribution Banks Banks (MT) (MT) (MT) 2001 14,500 0 1596 11 193,294,625 342 536 2002 3,140 0 0 0 0 0 305 2003 5,000 0 605 25 17,606,250 137 0 2004 0 0 0 0 0 0 0 2005 38,209 65,029 7,235 107 576,663,000 1,160 500 2006 10,130 10,606 4,080 0 1,891,8 14,958 0 0 2007 0 0 0 0 202,000,000 0 450 2008 (Planned) 7,000 11,561 0 0 1,530,047,648 0 1.100 Source: CCA 2007 Notes: Cereal banks and seed distributions are planned for May-August 2008, although the tonnage is not available. 57Although this is provided in the official figures, the percentage derived from local purchases appears to be greatly underestimated. This could be due to a different classificationof local and triangular purchases. For example, in2005, it is estimatedthat OPVN purchased2,300 MT of grains for sales and distribution; at the same time, other documents report that approximately40,000 MT were purchased in 2005 by the GoN. However, the CCA reportsthat it purchased50,000 MT of food from neighboringcountries. 68 84. To improve the functioning of its emergency response interventions, the DNPGCA developed a national contingencyplan for food security and nutrition with the participation of a variety of governmental and non-governmental actors. Developed in 2007, the document outlines the strategies priorities for the DNPCGA in preventing and managing food crises in the country. The overall objective of the plan is to minimize the impact of food crises by ensuring households' access to staple foods and protecting their assets, mainly via the national security stock and emergency cash resources. The document also outlines a variety of early warning indicators (Le. cereal deficit, food prices and severe malnutrition) in order to identify different food crisis scenarios, and to implement appropriate interventions (Le., general food distribution, FFW, CFW, subsidized sales, seed distributions). 85. There is a lack of agreement on what are the most effective safety nets interventions in Niger. The list of response interventions outlined by the national contingency plan includes most of the same instruments that have been used in the past few years and puts more emphasis on CFW versus FFW and cereal banks. However there i s no good indication of what has been more effective. The data on cereal banks from the 1990s shows that these have not been sustainable. CFW programs, which are usually preferred to FFW, are fairly new interventions that have not been fully evaluated in Niger. Inaddition, CFW programs are only appropriate ifprices are sufficiently low. NIGER2008: PREPARINGFORAND RESPONDINGTO A POTENTIALFOOD CRISIS 86. International prices of major cereals (rice, wheat and maize) have been on an upward trend since 2006. International prices of rice, wheat and maize have risen sharply since early 2006 (Figure 17). Maize prices have increased by 45 percent in Euro (and CFA) terms between March 2006 and March2007, due in large part to increased demand for maize for ethanol production in the US.. Wheat prices rose sharply in late 2007 following a poor harvest in Australia (a major exporter of wheat), which aggravated a situation of low international wheat stocks. By March 2008, international wheat prices (in Euro and CFA terms) were 96 percent higher as compared to those of March 2007. The increase in international rice prices has been the most dramatic. Following India's banon private sector exports of non-Basmati rice inlate 2007, as well as similar bans by Vietnam and other exporters, international rice prices rose to 449 Euro/MT inApril 2008 (nearly US$700/MT), 126 percent higher than in March 2007. Overall, prices of maize, wheat and rice were 106, 71 and 90 percent higher, respectively, inEuro terms, inMarch 2008 than inMarch2006.58 58InUSDterms, international maize, wheat and rice prices were 166, 120 and 145 percent higher inMarch 2008 as compared with the previous year. 69 Figure 17: International Prices of Grains (Wheat, Rice and Maize), 2000-2007 InternationalPrices of Cereals, 2000-07 -Wheat (US HRW#2, FOB Gulf) -- -- Maize (US #2 Yellow, US Gulf) Rice (FOB Bangkok,AI Super) 87. The increase in internationalfood prices should have a limited impact onfood security in Niger, which primarily imports millet, sorghum and maize. In this environment of high international grain and petroleum product prices, poor consumers, especially those in net food importing countries, face serious threats to their food security. To some extent, Niger i s insulated from the direct effects of these international price increases, as neither sorghum nor millet i s internationally traded in large volumes. Niger's prices for these commodities are affected by supply and demand conditions in Nigeria. Similarly, Niger imports maize from Benin and Nigeria, which are important local producers of this commodity. Consequently, changes in international maize prices are filtered through Nigeria. And finally, while wheat and rice prices in Niger are more closely linked with those in international markets, these commodities are not major staples of the poor inNiger, especially rural areas. 88. Price increases of staple grains (millet, sorghum and maize) in 2007/2008 appear to be driven by irregular rainfall and poor harvests in late 2007, not by international price movements. While departmental-level production data are not available, Figure 18 shows millet prices in key forecasting markets inNiger and Nigeria for October 2007. The prices on several of these markets - namely, Gaya, Guidan Roumdji, Tessaoua, Illela, Tounfafi, Mai Adua and Dogondoutchi - are higher or close to the price levels of the 2004 harvest. The prices for Benin(Malanville), Maradi and Jibia, however, are below the October 2004 levels. This suggests that there were production shocks on these markets during the 2007 harvest, similar to those in 2004, but not of the same magnitude. Therefore, while 2007 was considered to be a normal harvest, the price data suggest that perhaps key markets were affected during the harvest period. 70 Figure 18: Millet Prices on Key Forecasting Markets during the Harvest Period in 2000,2004,2007 IS0 140 Maradi 130 Gaya Jibia Zinder 120 i 01 e,8 Tessaoua Mal Adua 110 -g- 0- - IOU 2 90 80 70 60 10Oa-00 8Ocf.03 PIOct-04 Oct-07 89. The GoN, early warning systems and NGOs should monitor price differences between Nigeria and Niger to determine whether grain imports will be profitable in 2007/2008. Duringnormal years, prices innorthern Nigeriaare lower thanthose inNiger as of May/June, thereby making it profitable to import grains from Nigeria. In 2004/2005, prices in northern Nigeria were above those inNiger for the entire marketing season, suggesting that imports were not profitable during this year. In order to determine whether imports from Nigeria will be profitable in 2007/2008, the early warning systems and NGOs should monitor grain prices on cross-border markets and compare them with prices in Niger. If prices in northern Nigeria remain higher than those in Niger, this suggests that Niger would need to import from other countries. The following price pairs should be compared: Malanville (Benin)-Gaya, Illela (Nigeria)- Konni, Jibia (Nigeria)-Maradi/Dan Issa, Mai Adua (Nigeria)-ZindedMatameye, and Kantchari (Burkina Faso)-Torodi. Figure 19 shows how millet prices have been changing in 2007/2008, as compared with other drought and non-drought years (2000/2001, 2003/2004 and 2004/2005). Between October and March, grain prices in Niger and northern Nigeria have been following a pattern similar to a drought year. However, as of April 2008, prices have fallen slightly on these markets. This suggests that price movements in May/June should be monitored carefully, and any emergency preparedness and response interventions should not put pressure on local and regional supply* 71 Figure 19: Changes in Millet Prices between the Harvest and Hungry Periods 90. In the meantime Government and International Organizations are preparing a response program. To cope with the price increases, as part of the emergency measures, the Government has already taken temporary fiscal measures to mitigate the impact while also replenishingcereal stocks. More specifically, the VAT and customs duties on rice imports have been suspendedfor three months in an effort to alleviatethe impact of the increase in food prices on consumers. In addition, with international supports, existing governmental and non- governmentalinstitutionsare in the process of preparinga response basedon existing emergency mechanisms, which consist of food based programs for emergency distribution and "Food for Work"; subsidized cereals; and "Cash-for-Work" programs to raisepurchasingpower of affected households. 72 CHAPTER5: SUMMARY AND POLICY OPTIONS Niger faces serious problems of poverty and householdfood insecurity. Although the magnitude of the problem is most evident during years of climatic shocks, especially drought, the problem is in fact a chronic one. Low levels of food availability, limited economic access tofood and poor health and nutritional status have resulted in food insecurity and malnutrition for much of the population, even in years of normal harvests. Faced with limited resources, a high prevalence of poverty and periodic droughts and other shocks, an effective food security and social protection strategy is essentialfor Niger. Such a strategy should include three key components: (i) improving the efficiency and scope of safety net programs; (ii)promoting effective medium-term strategies and investments to improve food availability and utilization; and (iii) strengthening emergency response. Avoiding future food crises will also require greater recognition of the major importance of regional markets (jarticularly Nigeria) in influencing market prices, food imports (and exports), and ultimately household food consumption in Niger. Finally, any social protection strategy will need to be consistent with the county's overall strategy for growth and poverty reduction, the ultimate solutionsfor food insecurity and the poor health and nutritional outcomes associated with both poverty and householdfood insecurity. SUMMARYOFFINDINGS Poverty and Food Security 91. More than 50 percent of Niger's population suffers from some form of food insecurity and most of the population does not consume the minimum daily caloric requirements for adequate nutrition. Nutritional data show that much of the rural population suffers from seasonal food insecurity and both rural and urbanpopulations are vulnerable to transitory food insecurity, defined as reduced access to food after a shock. Analysis of household survey data shows that although there was some improvement between 2005 and 2006, the food security situation of the poor remains very fragile, with many households consuming only slightly above the minimumdaily caloric requirement. Higher food prices would most likely prompt consumption in these households to drop below that level, unless there were compensating gains in household incomes (including transfers). Among coping strategies used by households to mitigate shocks are: use of food aid; reduction innumber of meals per day; migration; and asset and livestock sales. Overall, poor households are more exposed to shocks, the poorest regions are the most vulnerable to food insecurity and coping mechanisms make households more vulnerable to future food insecurity. Food Availability and Markets 92. Because over two thirds of daily caloric consumption comes from cereals, production, total availability, market prices andperformance of marketsfor cereals are 73 crucial determinants of food security. Typically, over 80 percent of Niger's total availability of cereals comes from domestic production, mainly millet and sorghum. Droughts and pest attacks cause large fluctuations indomestic production, and production losses vary substantially across regions within Niger. Formal sector imports (food aid and commercial imports), consist mainly of wheat, maize and rice, and are on average only a small share of supply (7 percent of the average 2.8 million tons of annual total net cereal availability from 2000 to 2005). Unofficial imports of millet from Nigeria may have added as muchas 200 thousandtons to total net supply inrecent years, though there i s considerable uncertainty over these magnitudes. Survey evidence and analysis of prices indicate that Niger's wholesale cereal markets are generally well integrated. Markets in southern Niger are also well integrated with markets in Nigeria close to the border. However, lack of data prevents detailed analysis of price links between major markets in northernNigeria (e.g. Kano) andNiger. The 2005 Food Crisis,current surge in international prices and Government response 93. I n early 2005, national policy-makers and the international community were caught by surprise by reports of sharp price increases and serious malnutrition in southern Niger. Estimates of national food production in late 2004 had suggested only a modest decline inproduction relative to medium-termtrends, though it was clear that the October 2004 harvest was likely to be smaller than that of 2003. One major cause of the crisis was the extremely low income level of households, which contributes to their low nutritional status innormal years and makes them highly susceptible to major income and price shocks. In 2005, two major shocks occurred: a sharp rise in food prices (related to even more dramatic price increases insome of Nigeria's markets beginning inmid-2004), and losses in food production and incomes for those farmers whose late 2004 harvests were affected by drought and locusts. 94. The Government's response was not adequately sufficient to mitigate the impact of the 2004.4 crises on the vulnerable groups. The government's response, with international support, consisted mainly of food aid for emergency distribution and "Food for Work"; the sale of approximately 40 thousand tons of cereals between February and June 2005; and "Cash-for-Work" programs to raise purchasing power of affected households. In spite of these efforts, in some of the worst-affected regions, the gross mortality rate and child mortality rates rose well above international thresholds for humanitarian crises. Avoiding such crises in the future will require better monitoring of production and prices at local levels to capture information regarding households experiencing large shocks to incomes and purchasing power, monitoring and ongoing analysis of cereal markets and likely trade flows from neighboring countries (especially Nigeria), and improved response capacity of the government. 95. I s Niger affected by the International Increase in Food Prices and How should the Government respond in 2007/8? High international world prices for grains are cause for concern because o f their potential effects on domestic markets, as well as the potential for reduced availability o f food aid ifNiger (or even Nigeria) should suffer a production shortfall, (as appears to have been the case in late 2007). To some extent, Niger is 74 insulated from the direct effects of these international price increases because neither sorghum nor millet is internationally traded in large volumes. Niger's prices for these commodities are affected by supply and demand conditions in Nigeria. Wheat and rice prices in Niger are more closely linked with those in international markets, these commodities are not major staples of the poor, especially rural areas. However the price increases of maize, sorghum and millet inNigeria in 2007 and early 2008, driven mainly by irregular rainfall and poor harvests in late 2007 and not by international price movements, might raise some concerns. In the short run, it is important that the government monitor international and domestic market prices and recognize that increases in market prices in Niger that coincide with price movements in Nigeria (for maize, sorghum and millet) and international markets (for rice and wheat) are likely not due to non-competitive behavior on the part of domestic traders, but on increases in the cost of additional supply. As such, the government should avoid placing restrictions on storage and trade of cereals, but instead make efforts to promote market efficiency through transparency of government policies on food aid flows, other official imports, and releases of stocks. Open consultations with private sector traders and importers to share information i s one way of building confidence (Minten and Dorosh, 2006; Dorosh, 2008). In addition, the GoN should consider not engaging in local or triangular purchases of grains in Niger or in Nigeria, to avoid putting pressure on local supply and hence prices. OUTLINE OFA FOODSECURITY AND SOCIAL PROTECTIONSTRATEGY 96. To better address the chronic and severe transitory food insecurity problems of Niger, broad response is needed, including: (i) nets to protect poor and vulnerable safety households; (ii) mediumterms policies and investments to improve food availability and utilization; and (iii) measures to strengthenthe emergency response. Safetv nets to orotect Door and vulnerable people 97. Although the food insecurity problems that occur in drought years are most visible, safety nets are needed even in years of normal harvests to address the need of chronic food insecure households. Every year at least 20 percent of the population is severely chronic food insecure. In addition there are many poor and vulnerable households, which incase of face individual level shocks (e.g. injury, sickness or death in the family, loss ofjobs) they will be forced to greatly reduce household purchasingpower and access to food and fall deeper into poverty. Farmers also face price risks and suffer income losses if post-harvest prices are low; making it difficult to meet current consumption needs and repay debts. 75 Box 6: Ethiopia's Productive Safety Net Program (PSNP) In Ethiopia, over 40 percent of the population lives below the national poverty line and over 20 percent of the population is extremely poor (below 1,650 kilocalories per person per day). Since the variability in rainfall is among the highest in the world, and fluctuations in rainfall are inversely related to mean incomes, every year for more than two decades the government of Ethiopia has launched an international emergency appeal for food aid. This annual emergency assistance was designed to meet the consumption needs of both chronically and transitorily food-insecure households. Despite substantial amount of humanitarianassistance, evaluations have shown that emergency assistance was unpredictablefor both planners and households, often arriving late relative to need. As a result of the delays and uncertainties, the emergency aid could not be used effectively and did little to protect livelihoods, prevent environmentaldegradation,generate community assets, or preservehouseholdassets (physicalor humancapital). Characteristics of the Program: Given these shortcomings of the emergency aid regime, in 2005 the Ethiopian governmentstarted implementationof anew program, the ProductiveSafety Net Program(PSNP). The PSNP replaced the emergency humanitarianappeal systemas the chief instrumentinthe country's safety net. The programis currently operationalin 234 chronically food-insecure districts(of atotal of 692 districts), and targeted about 7 million people in 2006. The PSNP providesresourcesto chronically food-insecure households in two ways: (i)through payments to the able-bodied for participation in labor-intensive public works activities; and (ii) through direct grants to households composedof the elderly or those who cannot work for other reasons. Impact of the PSNP: A 2005 beneficiary survey found that the PSNP had had a significant positive effect on beneficiaries'well-being as calculatedby both subjective and objective indicatorsl. The survey found that three in five beneficiariesavoided havingto sell assets to buy food in 2005, and accordingto 90 percent of the households, this was a result of their participation in the PSNP. Moreover, almost half the beneficiariessurveyed stated that they had used healthcare facilities more and 76 percent of these households creditedthe PSNP with this enhanced access. More than one-third of surveyed households enrolled more of their children in school and 80 percent of them attributed this to participationin the PSNP. Ongoing reforms: Significantwork is plannedto further improve implementationcapacity and bring systems to a level of functioning not previously possible with fragmented and temporary programs. Work is also beginning on a contingent grant mechanism (conditional cash transfer) to provideresourcesin the same districtsto help transient food- insecure households during periods of drought. The mechanism will use a rainfall-based index that uses 30 years of rainfall data to trigger funding. Moreover, the PSNP is complemented by a larger food security program that tries to help householdsraisetheir incomesby meansof resettlementgrants, household income-generatingpackages, and water harvesting. Householdsthat benefit from the PSNP are also entitledto assistance under other parts of the food security program. Food security interventions financed by donors that fall outside the PSNP are, however, rarely coordinated at local levels, and their links to basic rural servicesare also weak. Lessons Learned: The PSNP illustrates many of the issues that surround safety nets in very low-income countries, namely: > The programis moving in a clearly beneficialdirectionby means of a basic design that not only seeks to use resources in ways that save lives, but also assist in livelihoods. The progress in implementation to date > suggests that this is possibleeven in avery low-incomesetting. The design process and implementation planning have undergone a fairly harsh triage. Even when fully realized, the programwill only providea safety net in about a third of the country. The districts selected are appropriately the poorest, but many poor people also live in the unserved districts. Moreover, the program has phased its implementation. It is focusing first on consolidating the basic PSNP. It hopes to enrich it eventually in a number of dimensions, but programmanagers and donors have realizedthat everythingcould not be accomplished right away. Thus, for example, the contingent fund for droughts was not implemented > until the third year of the PSNP. Good implementation requires diligent and sustained effort. By 2007, the program had many positive outcomes, and early qualitative assessmentso f its targetingand impacts are positive, but more remains to be done to consolidate implementation. Good implementation also requires flexibility and innovation. For example, the government was initially having problems with the program's monitoring system, but in the interim, it deployed so-called rapid response teams to visit districts to identify and solve implementation problems. This gave managers a sense of what was going well and what was not and whether adjustments were needed in individual districts or at a more systemic level. Meanwhile, the design of the monitoring > systemwas simplified and apilot to computerize it is under way. An important part of the reform is the shift to a multidonor, multiyear framework rather than an annual emergency appealsystem with each donor runninga separate initiative. This is complementedby the decision to deliver the programthrough regular government systems rather than special implementationunits common in donor funded programs. The multiyear framework and the reduction in fragmentation should permit the development of much more effective administrative systems. The multidonor framework should also aid in resilience. in that withdrawal or a reducedcommitment bv a sinele donor will have a less deleterious effect. Source: W 2008 Safkty Net Book Y 76 98. The challenge lies in developing a general safety net system and of designing an effective system that can be adequately financed. The mechanisms for prevention and response to food crises as currently implemented by the National Mechanism for the Prevention and Management o f Food Crises in Niger (DNPGCA) constitute only a part and one stage in the implementation o f a safety net system. Consequently, to develop a general safety net system for households suffering from chronic and seasonal and transitory food insecurity, it i s necessary to (i) formulate a safety net program suitable for the PRSP2 andRDS; and (ii) coordinate between the State and the technical and financial partners. This in turn will require the inclusion o f safety net programs inthe government budget and the confirmation o f a long-term financial commitment on the part o f the technical and financial partners (TFPs). Decisions on the magnitude o f safety nets must take into account the costs o f the program and possible tradeoffs with investments related to the overall poverty reduction strategy. The decision-making process i s critical and should include consultation with civil society, concerned government agencies and donors. Once a sustainable safety net program is inplace, it could be scale up both inthe number o f beneficiaries and the amounts transferred in the event o f a major production shortfall. In order to ensure support for safety net programs, it i s crucial that their financing be included in the government budget, perhaps with a commitment o f long- term financial support from donors. For example, in Ethiopia, a country also confronted with recurring food crises, the government and donors have agreed upon multiyear financing programs to replace the annual calls that often led to delays inrouting food aid or to inappropriate utilization o f funds. (See Box 6.) This medium term finance mechanism has increased effectiveness o f the design and implementation o f programs and has reduced uncertainty among the participants. 99. Existing government programs involving distribution of food could be included in the overall safety net system, though steps should be taken to improve the efficiency of their operation. This would require (i)an assessment o f the effectiveness o f existing programs (HIM0 [labor-intensive public works programs] such as "Food for Work" (FFW) and "Cash for Work" (CFW); free distribution o f foodstuffs; subsidized food sales; cereal banks; school cafeterias; etc); (ii) a review o f the existing targeting system to determine the challenges o f addressing the needs o f chronically poor people; and (iii) an improvement in the program monitoring system. Within this framework, to ensure that the safety net system is effective (in crisis and normal periods) it must be possible (i) to guarantee that the "Food for Work" (FFW) and "Cash for Work" (CFW) programs are well organized and give productive results; (ii) limit the amount o f free distribution o f to food as well as subsidized sales, and pay specific attention to synchronization and localization o f sales; (iii) ensure that the purchase o f local foodstuffs i s well planned; to and (iv) that targeting o f programs follows transparent and efficient procedures for the selection o f the regions and households to benefit from the program. Free distribution o f food and subsidized sales o f food should be done only on a limited scale, with careful attention to timing and location o f sales, avoiding sales at the time o f harvests that could reduce farmers' revenue. New safety nets programs could be introduced. Once the vulnerable groups are identified and evaluation o f the existing programs is completed, it will be necessary to 77 identify pilot programs that could be tested based on transparent eligibility criteria, coverage and cost. Among these new programs to pilot are cash payments and conditional cash transfers. Cash programs have an advantage over in-kind food distribution because they avoid operational costs involved in storage and movement o f food, as well as possible disincentive effects on production during periods when food availability i s not a constraint (e.g. harvest periods or, more generally, good harvest years). Conditional cash transfers, such as those linked to school attendance, offer benefits both in terms o f increasing incomes o f targeted households and increased participation in socially desirable programs (i.e., linked to school attendance and health service utilization for children). Medium term policies and investments to improvefood availabilitv and utilization 100. Medium term policies and investments are also needed to reduce vulnerability to production shocks, raise incomes of the poor, and enhance market efficiency. For example, inBangladesh, longer term investments inrural roads (often built through food for work or cash for work programs) and agricultural research and extension were combined with effective government policies that promoted expansion o f small-scale irrigation through liberalization o f the import o f small pumps and food policy management to ensure that food aid flows did not have market price disincentives for domestic food production (see Box 7). As a result, over a period o f about two decades, the country was able to double its food production, raise rural incomes and reduce dependence on food aid imports. India, likewise, used food for work and public works programs to develop irrigation that increased agricultural production and rural incomes. Box 7: Enhancing Food Security in Bangladesh Bangladesh has made considerable progress in increasing food production and enhancing food security through support of agriculturalresearchand extension, investments in roads, maintainingprice incentivesfor domestic production, liberalizationof domestic and internationaltrade in cereals, andtargeted public food distribution. Bangladesh suffered a major famine in 1975 following flood-induced production shortfalls in late 1974, as the country lacked foreign exchange for imports and government stocks for targeted distribution. Thereafter, the country invested heavily in small-scale irrigation (including major private sector investments in tube-wells), agricultural technology and roads. As a result, Bangladesh was able to double its rice production between the mid- 1970s andthe late 1990s. Particularlyimportantfor food security was the increase in winter season crops (boro rice and wheat) that enabled amajor harvestfollowing the traditional monsoonseasonharvest. Following major production shortfalls due to massive floods in 1988 and 1998, Bangladesh was able to prevent recurrence of famine. In 1988, the government relied on large public stocks of rice and wheat, government commercial imports and food aid inflows to help stabilize market prices and permit large-scale public distribution of grains. A decade later, following the 1998 floods, trade liberalization made private sector imports of rice and wheat possible, and the private sector respondedwith over 2 million tons of rice imports from neighboringIndia. As a result, rice prices were stabilizedwithout large-scalepublic distribution of grain or food aid flows, and at minimal cost to the government. Food aid-supported and other public distribution programs were used, however, to target flood-affected and other poor households. As shown by the Bangladesh experience, an appropriate mix of public and private investments, combined with adequate price incentives, can lead to increased productionand reduced variability of major cereals. Moreover, it i s possible to promote private trade to increase availability of food and stabilizeprices while operatingmajor targeted food transfer schemes to directly increase access to food of poor households. Such a combination of policies can substantially lower fiscal costs, but requires transparency on the part of the government in order to reduce risks for privatesector trade. Source: del Ninno et al., (2001) ;del Ninno, Dorosh and Subbarao (2004): 78 101. There is need to promote efficient domestic agriculture production. For Niger, increasing agricultural productivity and the level and stability o f domestic production would help increase food availability, raise rural incomes and reduce output and rural income fluctuations. To achieve this, designing and implementing policies to promote efficient domestic agricultural production are crucial. This will include investments in agricultural research and extension, investments in road construction and maintenance, and increasing market infrastructure (e.g. provision o f basic storage facilities in major markets). Also essential in Niger's semi-aridenvironment i s to improve access to water for agriculture through investments in irrigation and water retention (e.g. check dams, where appropriate). 102. Ensure trade policies do not impede cereal market development. Internal and external trade policies should consistently support short-term market efficiency, especially in emergencies, and avoid measures that impede medium-term market development. In particular, the government should avoid placing restrictions on storage and trade o f cereals, but instead make efforts to promote market efficiency through transparency o f government policies on food aid flows, other official imports and releases o f stocks. Open consultations with private sector traders and importers to share information are one way o f building confidence. 103. Reducing price risk for farm products would also increase incentives for production and reduce variability of rural incomes. This need not involve direct government intervention in markets, which in many countries has proven to be ineffective and expensive, often reducing the efficiency o f markets by reducing incentives for private sector processing, trade and storage. Instead, promotion o f farmers' groups can help small producers to pool their products and make use o f group storage or transport facilities. Improvements in market information systems on prices and volumes o f production and market supplies can also help farmers to increase their bargaining power and make more informed decisions on their sales. Already, the use o f cell phones has proven to be effective inimproving transmission of price information across markets. 104. Increasing access to food for the rural poor could also include measures to support diverse household income-generating activities. Micro-credit through local NGOs has been successful in supporting these activities, particular in situations where there is sufficient effective market demand for the outputs. In the medium term, investments in education and vocational training are also needed to increase human capital and raise labor productivity. 105. . Finally, improvements in heath and sanitation are needed to promote better nutritional outcomes and overall health. Increases in food consumption alone may not be lead to improved nutrition if disease or otherwise poor health limits the effective utilization o f food by individuals, especially pregnant and lactating women, small children and infants. Programs that address the overall health and sanitation environment, as well as care and feeding practices o f infants, need to be supported, particularly among the poorest households. 79 Strengthening emergencv response and improving the information svstem 106. Significant progress has been made in monitoring and responding to acute disasters. Nonetheless, further steps are needed to strengthen early warning systems and emergency response (see Annex 1). 107. Revise the Government National Contingency Plan. As discussedearlier, several institutions are involved inearly warning and emergency response. Implementation of the government's National Contingency Plan for Food Security is needed to strengthen the key institutions involved, particularly with regard to timely analysis and policy design. In this context, the role of the OPVN should be more clearly defined, along with the operational rules regarding use of the emergency food security stock. Given the multiplicity of actors (various government agencies, NGOs, donors), a more comprehensive system o f monitoring food procurement, distribution and stocks through all the various programs should be set up. 108. Improve the information system of the early warning system. Current early warning systems provide valuable informationon markets inNiger. However, monitoring and analysis of price movements, particularly inNigeria (not just inborder markets), and the broader international markets are also crucial in order to allow as much time as possible to arrange for additional food aid or government commercial imports in the event of a poor domestic harvest. This will require significant strengthening of analytical capacity within the key government institutions (CCA, SAP, SIMA). Likewise, encouraging the free flow of information innewspapers and other media can help inform of the government of situations where urgent action is needed. 109. Use safe@ net program to complement the emergency response. Once sustainable safety net programs are developed, the government could substantially scale them up in the event of a major production shortfall. This will require effective contingency planning, including arrangements for financing and effective program implementation, with areas of responsibility clearly delineated. 80 REFERENCES Africare, Catholic Relief Services, CARE and HelenKeller International (2004). Initiatives de Se`curite` Alimentaire au Niger (ISAN) - Etude quantitative Jinale, 2004. Niamey, Niger. Aker, Jenny C. (2008). "Does Digital Divide or Provide? The Impact of Cell Phones on Grain Markets inNiger." BREAD WorkingPaper. Aker, Jenny C. (2007). "The Structure, Conduct and Performanceof the Cereals Market in Niger: Implications for Preparing for and Responding to Food Crises." A Report prepared for the World Bank, Washington, D.C. Alderman, H., S. Appleton, L.Haddad, L. Song, and Y. Yohannes (2005). Reducing Child Malnutrition: How Far Does Income Growth Take Us? Centre for Research inEconomic Development and International Trade, University ofNottingham. Barrett, Chris and DanielMaxwell. (2004). FoodAid After Fifty Years; Recasting Its Role. Rutledge. Beekhuis, Geert (2005). Niger: Profile of Cereal Markets". World FoodProgram, Rome, Italy. Tesliuc, E. and Kathy Lindert. 2004 Risk and Vulnerabilitv in Guatemala: A Quantitative and Qualitative Assessment Social Protection Discussion Paper No. 0404; Publication Date: 03/04. World Bank Coates, J., P. Webb, and R. Houser (2003). `Measuring Food Insecurity: Going Beyond Indicators of Income and Anthropometry'. Food and Nutrition Technical Assistance Project. Washington, DC: Academy for Educational Development. del Ninno, C. Giovanni Vecchi and Noshin Hussain, 2006. Poverty, Risk and Vulnerability in Pakistan.CRPRID, Islamabad, Pakistan. http://undp.orrr.t-lk/documents/PAKVulnerabilitvtoPovertv.pdf del Ninno, Carlo, Paul A. Dorosh and Kalanidhi Subbarao. Food aid, domestic policy and food security: Contrasting experiences from South Asia and sub-Saharan Africa. FoodPolicy, Volume 32, Issue4, August 2007, Pages 413-435 delNinno, Carlo Paul A. Dorosh, Kalanidhi Subbarao (2004), "FoodAid and Food Security in the Short and long Run: Country Experience from Asia and Sub- Saharan Africa ";SP DiscussionPaper #0538, World Bank Washington DC. delNinno, C. andA. Marini, 2005; "Household's Vulnerability to Shocks inZambia"; SP Discussion Paper #0536, World Bank Washington DC. 81 CARE International, Catholic Relief Services, Helen Keller International and World Vision International (2007). The Cereal Market in Niger: Findings from the 2005-2007Market Surveys. Baltimore, MD. Cooperation Francaise, European Union, FAO, FEWSNET, INS, SAP, SIMA and WFP. Enqubte sur la conjoncture et la vulne`rabilite` alimentaire des me`nages (May 2006). Niamey, Niger. Dorosh, Paul (2008). "Production Shocks, Regional Trade and Cereal Prices inNiger" A Report Preparedfor the World Bank,Washington, D.C. Dorosh, Paul (2008). "Food Price Stabilization and Food Security: The International Experience", Bulletin of Indonesian Economic Studies 44( 1): 93-114. C;rosh, Margaret. Carlo del Ninno iiiid t;inil l'esliuc, A/cdine Oucrghi. 2008. For Protection and Promotion: The Design and Implementation of Effective Safety Nets. World Bank Publications, Washington DC. Hoddinott, J., and Y. Yohannes (2002). `Dietary Diversity as a Food Security Indicator'. FCND Discussion Paper #136; Washington, DC: International Food Policy ResearchInstitute. Hoddinott, J. and Agnes Quisumbing (2003); "Methods for Microeconometric Risk and Vulnerability Assessments"; SP Discussion Paper #0324; World Bank Publication. Hoogeveen, J., E. Tesliuc, R. Vakis, with S. Dercon (2004); "A Guide to the Analysis of Risk, Vulnerability and Vulnerable Groups"; The World Bank. Macro International (1998). Enqubte de`mographiqueet de la sante`, 1998. Maryland. Macro International (2002). Enqubte de`mographiqueet de la sante`, 2002. Maryland. Migotto, Mauro, BenjaminDavis, Gero Carletto, and Kathleen Beegle. 2006. Measuring Food Security Using Respondents' Perception of Food Consumption Adequacy. ResearchPaper No. 2006/88. Wider. Minten,Bartand PaulDorosh(eds.). 2006. Rice Markets in Madagascar in Disarray: Policy Optionsfor Increased Eflciency and Price Stabilization. Africa Region Working Paper Series, Number 101, September.Washington, D.C.: World Bank. Famine Early Warning SystemsNetwork (FEWSNET). (2005). "Niger: An Evidence Base for Understanding the Current Crisis", USAID/FEWS NET, Washington, D.C. FEWSNET, FAO, INS, RGAC, SAP, UNICEF and WFP (November 2006). Enqubte 82 conjointe sur la vulne'rabilite' a I 'insecurite' alimentaire des me'nages au Niger. Niamey, Niger. Fongillo EA, Nanama S (2006). Development and validation of an experience-based measure of household food insecurity within and across seasons in Northern Burkina Faso.Journal of Nutrition 136: 1409s-19s. Government ofNiger/Prime Minister's Office (August 2007). Accelerated Development and Poverty Reduction Strategy: 2008-2012. Niamey, Niger. Government o fNigedDNPCGA (November 2007). Plan National de Contingence: Volet Securite Alimentaire et Nutritionnelle Niamey,Niger. Guerrant R. L.(1992). Diarrhea as a cause and an effect o f malnutrition: diarrhea prevents catch-up growth and malnutrition increases diarrhea frequency and duration. American Journal of Tropical Medical Hygiene 47: 28-35. Haddad, L.and Frankenburger, T. (2003). "Integrating Reliefand Development to Accelerate Reductions in Food Insecurity in Shcok-Prone Areas." USAID Office of Foodfor Peace, Occasional Paper No. 2, Washington, D.C. Hamer, D. H.(2006). Malnutrition, Micronutrients, Measles andAcute Respiratory Infections. Helen Keller International, InstitutNationalde la Statistique, Ministere de SantC Publique, UNICEF, World Food Program (2007). Enqubte nutrition et survie des enfants de 6-59 mois, 2007. Niamey, Niger Horton, S. and J. Ross (2003). The economics of iron deficiency. FoodPoIicy 28 (1): 51-75. InstitutNationalde la Statistique (INS), Macro International andUNICEF. (2006). Enqubte de'mographique, de la sante' et de la nutrition / Multiple indicator cluster survey. Niamey, Niger and Maryland. InstitutNational de la Statistique, UNICEF and WFP (2006). Rapport d'enqugte nationale nutrition et survie de I 'enfant. Niamey, Niger. InstitutNational de la Statistique (2004). ProJil de lapauvrete' au Niger, 2004 : Application de la me'thode de degre' de satisfaction des besoins essentiels (DSBE). Niamey, Niger. 83 InstitutNationalde la Statistique (2005). Questionnaire des Indicateurs de Base du Bien- &re (QUIBB), ProJl de Pauvretk. Niamey, Niger. Marinho, Eduardo and Bruno Gerard (2008). "Tracking Vulnerability." Unpublishedmimeo. Mason, J. B., P. Musgrove, and J.-P. Habicht (2003). At Least One-Third of Poor Countries' Disease Burden Is Due to Malnutrition. Disease Control Priorities Project Maxwell, S., Frankenberger, T., 1992, "Household Food Security: Concepts, Indicators, Measurements. A Technical Review", UNICEF, New York and IFAD, Rome. Maxwell, D., Ahiadeke, C., Levin, C., Armar-Klemesu, M.,Zakariah, S., Lamptey, G.M., 1999, (( Alternative Food-Security Indicators : Revising the Frequency and Severity of `Coping Strategies"', FoodPolicy, Vol. 24, pp.411-429. Nicholson, S.E. ,B. Some and B. Kone (2000). "AnAnalysis of Recent Rainfall Conditions in West Africa, Including the Rainy Seasons o f the 1997 El Nino and the 1998 LaNinaYears." Journal of Climate. Vol. 13: Pelletier, D. L.,and E. A. Frongillo (2003). Changes inChild Survival Are Strongly Associated with Changes in Malnutrition in Developing Countries. Journal of Nutrition 133 (107): 119- Rawlings, L.(2005). Evaluating the Impact of Conditional CashTransfer Programs. WorldBank Res.Obs. 20 (1): 29- Ricci et al. (2006). Reducing stunting among children: thepotential contribution of diagnostics. Nature Publishing Group. Subbarao, K.and W. James Smith(2004). "Safety Nets versus Relief Nets: Towards a Medium Term Safety Nets Strategy in Ethiopia." Unpublished mimeo. World Bank, Washington, D.C. UNDP and the Government ofNiger (December 2007). Analyse Integree des Facteurs de Risques au Niger. Niamey, Niger. UNICEF (2000). Enqugte indicateurs multiple de lafin de la dkcennie -MICS2, 2000. Niamey, Niger. UNICEF (2005). Report of an emergency nutrition survey in Niger. Niamey, Niger. 84 UNICEF (2007). TheState of the World's Children. New York, UNICEF. Victora, C. G., B. R. Kirkwood, A. Ashworth, and et al. (1999). Potential interventions for the prevention of childhood pneumonia in developing countries: improving nutrition. American Journal of Clinical Nutrition 70: 309-320. World Bank (1986). "Poverty and Hunger: Issues and Optionsfor Food Security in Developing Countries", World Bank, Washington, DC. World Bank (2005).Niger MilleniumDevelopment Goals: Diagnostic and Tools; June 2005, GreenCover report; Q. Wodon. World Bank (2006). Carte de Pauvre'te'pour le Niger. Unpublishedmimeo. Q. Wodon. World Bank (2006). Repositioning Nutrition as Central to Development: A Strategyfor Large-Scale Action. Washington, DC: World Bank. World FoodProgram(WFP) (2005). Enqugte sur la se'curite'alimentaire en situation d 'urgence (EFSA) au Niger. Niamey, Niger. World FoodProgram(WFP) (2005) 85 ANNEX I: DESIGNOF EARLY WARNING SYSTEMS There is considerabledebateas to the appropriate design of early warning systems for the prediction of drought and famine in Africa. Much of the debate centers on the ability of current or prospective indicators to provide location-specific, timely, and cost- effective information. Macro-level environmental data are generally agreed to be useful, but they are often unable to identify the individual communities most in need of relief. Furthermore, while grain prices and dispersion in Niger are strongly affected by climatic shocks, there was not a strong correlation between drought and food crisis regions in 2005. Consequently, an early warning system based solely upon climatic indicators would have over (or under) estimated the severity of the crisis. Similarly, although per capita grain production was higher in 2004 than 2001, the severity of the food crisis was much greater in 2005 compared to 2001; consequently, an early warning system that depends primarily upon production levels would have underestimated the severity of the food crisis. In an effort to minimize concerns related to using macro-level indicators, the use of local-level indicators has been recommended to generate predictions on the food supply at the village or sub-district level. However, the choice of indicators remains controversial. Drought and Agricultural Production Indicators of drought and agricultural production are timely and easy-to-use, and can provide early information on a possible food crisis in West Africa. Nevertheless, as experience has shown in Niger, droughts do not necessarily lead to food crises, or vice versa. De Waal (1998) notes that effective early warning indicators should be reliably associatedwith the development of food crisis conditions; more specifically, they should generate few responses to situations where food crises do not actually develop ("false positive" responses) and should not fail to respond when an actual famine is approaching ("false negative" responses). Relying solely upon drought and agricultural production as early warning indicators for Niger would have generated a "false positive" in 2001, but yielded a "false negative" in2004. While drought and agricultural production-based predictions are prone to error, these indicators can be strengthened by changing the level of analysis. Rather than focusing on the presence of drought or national per capita production, a more appropriate indicator would be the percentage o f departments affected by drought or by production shocks. For example, while overall per capital production was higher in 2004, the percentage of departments affected by production shocks was much higher. Similarly, drought andproduction levels inthe northern regions of Beninand Nigeria should also be monitored. 86 Monitoring Domestic and Regional Market Prices Market activity at the national, regional, and local levels has been used to signal an impending food shortage in a variety of early warning systems (Huss-Ashmore 2007). In general, two aspects have been monitored: the amount of grain or livestock being traded, and changes in the prices of these commodities. These indicators are interconnected, as changes in the quantity of commodities traded affect their prices. However, monitoringthe amount of grain traded - either domestically or via imports - is difficult, especially inthe West African context. One indication of impending food crisis i s an unusually high or rapid increase in the price of staple food grains (McCorkle, 1987; Walsh, 1986). As with all market data, the interpretation of price changes inagricultural markets depends on local circumstances (Huss-Ashmore, 2007). Normal seasonal fluctuations in prices need to be considered when assessing the meaning of sudden market shifts. Only increases that significantly exceed those expected for a particular period can be consideredas signs of unusual ~hortfall.'~ The relationship between market performance and food crises in Niger provides some indications of the appropriate focus, timing and interpretation of changes inmarket prices. Although the current early warning systems in Niger monitor a variety of markets, not all markets affect grain prices or food security to the same degree. First, early warning systems should carefully monitor key "forecasting" markets within Niger and in the sub-region (Benin and Nigeria). To be precise, strategic forecasting markets could be defined as those markets that "Granger-cause" a significant number of markets inNiger (e.g., over 75 percent), includingthose located inBeninandNigeria. A secondary consideration is the timing of the monitoring of market prices. While current systems monitor prices on a monthly basis, the harvest (October- November) and hungry (June-August) periods are of most concern. The relatively higher prices in Granger-causing markets during October 2004 could have served as a key indicator of the potential crisis during the 2004/2005 marketing season. This is especially important in light of the slow adjustment speed of market prices. Therefore, early warning systems should also monitor prices on key forecasting markets in Niger, Benin and Nigeria during the harvest period, in order to determine whether price levels are higher at the outset of the marketing season. 59 Variations in local circumstances mean that economic activities, as famine indicators, do not function equally well for all populations. Cutler (1985) shows that boththe retail price o f coarse rice and the demand for fertilizer served as accurate warning signals for famine conditions in Bangladesh in 1979. Similarly, McCorkle (1987) argues that flux in market prices o f cereals and livestock is one o f the most promising quantitative markers o f famine onset in West Africa. Using data from Burkina Faso, she shows that the food crisis o f 1983-84 was accurately signaled by a tripling o f the price o f sorghum, a sharp fall in the selling price of cattle, and an increase in the number o f farmers defaulting on their advance sales agreements with grain merchants. B y contrast, de Waal (1988) argues that economic indicators were poor predictors o f famine for D a r k , Sudan, in 1984-85. Despite two complete harvest failures in northern Darfur and one in southern Darfur, economic data proved to be neither sensitive to nor specific in the prediction o f increased mortality. 87 While monitoring prices and price changes in strategic markets is important, knowing that prices are "higher" or have "increased" does not provide an effective indicator for determining whether there will be a food crisis. Based on the livelihood security studies and poverty studies conducted by international organizations, a threshold grain price in key national andregional marketsshould be identified. FuelPrices Traditionally, fuel prices are not included as early warning indicators in the West Africa region. While grain prices are fairly inelastic to changes in gas prices or transport costs, such changes do affect spatial arbitrage opportunities. Consequently, changes in fuel prices can impact the spatial allocation of goods, thereby affecting food supply and prices. Monthly fuel prices and transport costs could be a timely and cost effective indicator of food crisis, especially for remote or distant markets. Market Information Most early warning systems focus on providing relevant and timely information to policymakers, rather than the market actors themselves. Market performance, however, depends on symmetric access to appropriate information. Asymmetric information prevents market actors from engaging in optimal arbitrage opportunities. Although market information systems collect and disseminate price information in several West African countries, survey evidence suggests that these are not always used by traders and farmers. A recent study in Niger found that cell phone coverage is associated with a statistically significant reduction in grain price dispersion, as well as lower grain prices. The main mechanism for this result is a change in traders' behavior, as traders use cell phones to gather market information and to make decisions about the optimal allocation of goods. This suggests that the data and analysis of early warning systems should not only be available to policymakers, but also to market actors 88