96289 AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER MOZAMBIQUE: AGRICULTURAL SECTOR RISK ASSESSMENT RISK PRIORITIZATION Kilara C. Suit and Vikas Choudhary WORLD BANK GROUP REPORT NUMBER 96289-MZ AUGUST 2015 AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER MOZAMBIQUE: AGRICULTURAL SECTOR RISK ASSESSMENT Risk Prioritization Kilara C. Suit and Vikas Choudhary © 2015 World Bank Group 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved This volume is a product of the staff of the World Bank Group. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank Group or the governments they represent. The World Bank Group does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. World Bank Group encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clear- ance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone: 978-750-8400, fax: 978-750-4470, http://www.copyright .com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank Group, 1818 H Street NW, Washington, DC 20433, USA, fax: 202-522-2422, e-mail: pubrights@worldbank.org. Cover photo credits from left to right: Peter Casier/UN World Food Programme, Kate Raisz/Bread for the World, and Marcos Villalta/ Save the Children. CONTENTS Acronyms and Abbreviations ix Acknowledgments xi Executive Summary xiii Chapter One: Introduction and Context 1 Chapter Two: Mozambique’s Agricultural System 5 Agro-Ecological Conditions 5 Land and Water Resources 6 Crop Production Systems 7 Livestock Production System 9 Principal Constraints 10 Chapter Three: Agricultural Sector Risks 11 Production Risks 11 Market Risks 18 Enabling Environment Risk 21 Independent Risks Occurring Simultaneously 21 Dependent Risks 21 Regional Shocks 22 Differential Impact of Risk 22 Climate Change 22 Chapter Four: Quantification of Losses and Impact of Agricultural Risks 25 Conceptual and Methodological Basis for Analysis 25 Indicative Quantification of Losses 25 Expected Losses and Risk Priorities for Production 26 Impact of Agricultural Risks on National GDP 29 Chapter Five: Stakeholder Risk Assessments 31 Government 31 Producers 32 Vulnerable Groups 33 Chapter Six: Risk Prioritization and Management 35 Risk Prioritization 36 Risk Management 36 References 47 Appendix A: Methodology/Approach to Production Risk Losses Calculation 49 Appendix B: Weather Yield Report 51 Background 51 Rainfall Patterns in Mozambique 52 Risk Prioritization iii Drought and Excess Rainfall Analysis 52 Summary52 Rainfall—Yield Regressions 54 Maize (Milho) 54 Rice (Arroz) 55 Sorghum (Mapira) 56 Groundnuts (Amendoim) 58 Appendix C: Climate Change Impact Assessment on Agriculture in Mozambique: Literature Review 61 Introduction61 Principal Findings 61 Brief History of Climate Change Impact Assessments 62 Methodologies and Temperature/Precipitation Projections 62 General Findings 65 Maize  66 Cassava66 Soy66 Sorghum68 Other Crops 68 Conclusion69 Limitations73 Appendix D: Vulnerability Analysis 75 Definitions75 Recent General Trends in Vulnerability 75 Vulnerable Groups and Characteristics 76 Appendix E: Commodity Risk Profiles 81 Poultry82 Sorghum84 Tobacco85 Vegetables87 Groundnuts89 Cashew Nuts 91 Cotton93 Sugarcane95 Cassava97 Maize99 Appendix F: Risk Transfer and Financing Options in Mozambique 101 Microlevel Options 101 Mesolevel Options 102 Macrolevel Options 102 BOXES Box 4.1: Crop Losses According to the Agricultural Census 28 iv Mozambique: Agricultural Sector Risk Assessment FIGURES Figure ES.1:  Major Shocks to Crop and Livestock Production: Annual Percentage Growth in Agriculture Value Added xiv Figure ES.2: Acute Food Insecure Population in Mozambique xiv Figure ES.3: Estimated Aggregate Total Losses by Risk Event xvi Figure 2.1: Agro-Ecological Conditions—Mozambique 6 Figure 2.2:  Maize Production, 1992–2011 7 Figure 2.3: R  ice Production, 1992–2011 8 Figure 2.4: Cassava Production, 1992–2011 8 Figure 2.5:  Mozambique Agricultural Export Shares in US$, 2010 9 Figure 2.6:  Consumption and Production of Poultry in Mozambique, 2005–09 9 Figure 3.1:  Major Shocks to Crop and Livestock Production: Annual Percentage Growth in Agriculture Value Added 12 Figure 3.2: D  eficit Rainfall Events Based on Weather Station Data Analysis 13 Figure 3.3:  Excess Rainfall Events Based on Weather Station Data Analysis 14 Figure 3.4:  Monthly Retail Prices of Maize (White) in Key Markets in US$/Ton 18 Figure 3.5:  Monthly Retail Prices of Rice in Key Markets in US$/Ton 18 Figure 3.6:  International Monthly Prices for Sugar 19 Figure 3.7:  International Monthly Prices for Cotton “A” Index 19 Figure 3.8:  Exchange Rate Risks—Metical/Rand 20 Figure 3.9: E  xchange Rate Risks—Metical/Euro and Metical/US$ 20 Figure 4.1:  Estimated Aggregate Total Losses by Risk Event 26 Figure B4.1.1:  Principal Cause of Crop Losses 28 Figure 4.2: Annual GDP Growth and GDP per Capita 28 Figure 5.1:  Adverse Impact of Agricultural Risks on Inflation 32 Figure 5.2: Acute Food Insecurity in Mozambique 34 Figure B.1: Provinces of Mozambique 51 Figure B.2:  Location of Meteorological Stations 52 Figure B.3: Mean Cumulative Rainfall per Month 53 Figure B.4: Drought Events, 1979–2009 53 Figure B.5: Excess Rainfall Events, 1979–2009 54 Figure B.6:  Midseason Regression Models for Niassa, Inhambane, and Maputo Provinces 55 Figure B.7: 2008 Rice Surface Sown by Region (hectares) 56 Figure B.8: 2008 Sorghum Surface Sown by Region (hectares) 57 Figure B.9: Sorghum Yield Histogram for All Regions 57 Figure B.10: 2008 Groundnuts Surface Sown by Region (hectares) 58 Figure B.11: Groundnut Yield for All Regions 59 Figure C.1:  Annual Cycle of Rainfall, Maximum Temperature, Potential Evapotranspiration, and Potential Moisture Index 63 Risk Prioritization v Figure C.2:  Climate Change Effects on Yield for All Major Crops 65 Figure C.3:  Change In Yield with Climate Change: Rain-Fed Maize  68 Figure C.4: M  aps of Land Suitability and Hotspots Resulting from Climate Change, for Maize 68 Figure C.5:  Expected Changes in the Future (2046–65) for Maize (Expressed in kg/ha) under Rain-Fed Agriculture Based on the Median of All Seven GCMs 69 Figure C.6:  Change in Land Suitability per Crop Resulting from Climate Change 69 Figure C.7:  A. Projected Changes in the Future (2046–65) for Cassava in Percentage of Present Yields. B. Projected Changes in the Future (2046–65) for Cassava in kg/ha 71 Figure C.8:  Maps of Land Suitability and Hotspots Resulting from Climate Change, for Soy 71 Figure C.9:  A. Projected Changes in the Future (2046–65) of Soybean Yields in Percentage of Present Yields. B. Projected Changes in the Future (2046–65) for Soybeans in kg/ha 72 Figure C.10: M  aps of Land Suitability and Hotspots Resulting from Climate Change, for Sorghum 72 Figure C.11:  A. Projected Changes in Future (2046–65) for Sorghum (Median of All Seven GCMs), Expressed in Percentage of Present Yields. B. Projected Changes in the Future (2046–65) for Sorghum (Median of All Seven GCMs), Expressed in kg/ha 73 Figure C.12:  A. Projected Changes in the Future (2046–65) for Cotton (Median of All Seven GCMs), Expressed in Percentage of Present Yields. B. Projected Changes in the Future (2046–65) for Cotton (Median of All Seven GCMs), Expressed in kg/ha 74 Figure C.13:  A. Projected Changes in the Future (2046–65) for Groundnuts in Percentage of Present Yields. B. Projected Changes in the Future (2046–65) for Groundnuts, Expressed in kg/ha 74 Figure D.1:  Vulnerable Groups within Mozambique 76 Figure D.2: Vulnerability Characteristics 76 Figure D.3:  Provinces Where Vulnerability Characteristic Is More Common 79 Figure E.1: Domestic Price Volatility, Poultry 82 Figure E.2:  Domestic Price Volatility, Sorghum 84 Figure E.3:  Domestic Price Volatility, Goundnuts 89 Figure E.4:  Domestic Price Volatility, Cashew Nuts (with Shell) 92 Figure E.5: Cotton Yields, 1965–2007 (kg/ha) 94 Figure E.6:  International Cotton Prices, 2005–10 94 Figure E.7:  Domestic Price Volatility, Sugarcane 96 Figure E.8: I nternational Price Volatility, Cassava 98 Figure E.9:  Domestic and International Price Volatility, Maize 100 TABLES Table ES.1:  Filtering of Risk Management Interventions and Potential for Multiple “Wins” xvii Table 2.1:  Number of Agricultural and Livestock Production Units in Mozambique 6 Table 2.2:  Agricultural GDP Share in Mozambique, 2009 7 Table 3.1: Summary of Drought Events 13 Table 3.2: Summary of Flood Events 15 vi Mozambique: Agricultural Sector Risk Assessment Table 3.3: Summary of Pest and Disease Events 16 Table 3.4: Summary of Wildfire Events 16 Table 3.5: Summary of Cyclone Events 16 Table 3.6:  Projected Changes for 2046–65 in Average Temperatures during the Growing Season, Crop Yields under Rain-Fed Conditions, and Rainfall during the Crop Growing Season 17 Table 3.7:  Major Crop Systems and Their Principal Hazards in Mozambique 23 Table 3.8: Provinces and Principal Hazards 23 Table 3.9: Principal Risks by Commodity 24 Table 4.1:  Estimated Aggregate Total Losses by Risk Event, 1996–2015 27 Table 4.2:  Estimated Aggregate Total Losses by Year, 1996–2015 27 Table 6.1: Risk Prioritization 36 Table 6.2: Potential Interventions for Risk Management 38 Table 6.3: Decision Filters for Prioritizing Interventions 42 Table 6.4: Multiple “Wins” 44 Table 6.5: Risk Management Intervention Integration with PNISA 45 Table B.1:  Determination Coefficient (Linear Regression Model, Rainfall, and Rice Yield) 56 Table B.2:  Determination Coefficient (Linear Regression Model, Rainfall, and Sorghum Yield) 58 Table B.3:  Determination Coefficient (Linear Regression Model, Rainfall, and Groundnut Yield) 59 Table C.1:  Projected Changes for 2046–65 in Average Temperatures during the Growing Season, Crop Yields under Rain-Fed Conditions, and Rainfall during the Crop Growing Season 64 Table C.2:  Climate Change Effects (Rain, Temperature, CO2, and O2 Changes) on Crop Yielding in the 2046–65 Period 66 Table C.3:  Impact of Midcentury Climate Change on Crop Yield in Mozambique: Effect of Rise in Temperature, Background Ozone, and Atmospheric CO2, a Layered Approach 67 Table C.4:  Average of the Percentage Change in Yield for Mozambique 70 Table D.1: Underlying Factors of Food Security in Mozambique 77 Table D.2: Major Shocks to Food Security in Mozambique79 Risk Prioritization vii ACRONYMS AND ABBREVIATIONS AES Agriculture and Environment Services IIAM Agricultural Research Institute of AfDB African Development Bank Mozambique ALES Automated Land Evaluations System INAS National Institute for Social Action ARMT Agricultural Risk Management Team, World INGC National Institute for Disaster Management Bank IPM Integrated Pest Management CBSD Cassava Brown Streak Disease MASA Ministry of Agriculture and Food Security CCGC Coordinating Council for Disaster Risk MIC Mozambique Ministry of Industry and Management Commerce CIMMYT International Centre for Maize and Wheat MICOA Ministry for Coordination of Environmental Improvement Affairs CMD Cassava Mosaic Disease MMAS Ministry of Women and Social Action DANIDA Ministry of Foreign Affairs of Denmark MPF Ministry of Planning and Finance EACC Economics of Adaptation to Climate Change: MT Metric Ton Mozambique ND Newcastle Disease FAO Food and Agriculture Organization NGO Nongovernmental Organization (of the UN) OIE World Organization for Animal Health FAOSTAT Food and Agriculture Statistical Databases (UN) PEDSA Strategic Plan for Agricultural Development FCMNB Financial and Private Sector Development, Non-Bank Financing Institutions Unit PES Economic and Social Plan FEWSNET Famine Early Warnings System Network PNISA National Investment Plan for the Agrarian Sector in Mozambique FRELIMO Mozambique Liberation Front PPP Public-Private Partnership G-8 Group of Eight RENAMO Mozambican National Resistance GCM Global Climate Models GDP Gross Domestic Product SADC Southern African Development Community GFDL Geophysical Fluid Dynamics Laboratory SETSAN Technical Secretariat for Food Security and Nutrition GIEWS Global Information and Early Warning System SIMA Agricultural Market Information System ha Hectare TIA Mozambique Agricultural Census ICT Information and Communication Technology UNDP United Nations Development Programme IFPRI International Food Policy and Research USAID U.S. Agency for International Development Institute WFP World Food Programme Risk Prioritization ix ACKNOWLEDGMENTS The Mozambique Agricultural Sector Risk Assess- the findings. Their insights obliged the team to be realistic ment was conducted by the Agriculture Global Prac- and practical. tice (GFADR). This report was prepared by a team led by Vikas Choudhary (GFADR, Senior Economist) and The team would like to thank Mark Lundell (AFCS2, Kilara Suit (GFADR, Agricultural Specialist) and com- Country Director, Mozambique), Mark Austin (AFCS2, prising Jan Nijhoff (GFADR, Senior Economist), Pedro Program Leader), Laurence Clarke (then-Country Direc- Arlindo (GFADR, Agricultural Economist), Daniel tor, Mozambique), Severin Kodderitzsch (GFADR, Prac- Clarke (GFMDR, Senior Insurance Specialist), Barry tice Manager), and Marc Sadler (GFADR) for their valuable Maher (GFMDR, Financial Sector Specialist), Elon Gil- guidance and support. Ross Hughes (GENDR, Senior bert (Consultant), and Carlos Costa (Consultant). Mary Natural Resources Management Specialist), ­ Aniceto Bila Rose Parish (STC) and Luisarturo Castellanos conducted (GFADR, Senior Rural Development Specialist), and the Vulnerability Assessment and weather analysis for the Maria de Sousa Amazonas (GFADR, Senior Rural Devel- report, while Hermes Sueia (Consultant) provided local opment Specialist) were peer reviewers of the assessment. coordination and data collection support. James Cantrell and Damian Milverton provided valuable editorial and desktop publishing input. Administrative The team is grateful to the Government of Mozambique, support was provided by Celia Dos Santos (World Bank, in particular the Cotton Institute and their staff for their Maputo Office) and Ramon Yndriago (GFADR). helpful collaboration and contributions to the work dur- ing and beyond the field mission. The team would also The authors gratefully acknowledge the generous contri- like to express its gratitude to all the stakeholders who par- butions from USAID; Ministry of Foreign Affairs of the ticipated during the field work and workshop exercises for Govern­ment of the Netherlands and State Secretariat sharing their valuable time and perspective and discussing for Economic Affairs (SECO) of the Government of Switzerland. Risk Prioritization xi EXECUTIVE SUMMARY Agriculture plays a key role in the economy of Mozambique. It accounts for 31.8 percent of gross domestic product (GDP), providing a livelihood to almost 81 percent of the labor force. The majority of production undertaken by smallholder families is for con- sumption, and the main staple crops produced in the family agricultural sector are maize, sorghum, rice, millet, potatoes, sweet potatoes, cassava, and beans. Cash crops frequently grown by households include cotton, tobacco, copra, cashew nuts, sesame, sugar beans, sunflower, bananas, and sugarcane. Many family farms combine food crops with a single cash crop. The majority of Mozambican agriculture is rain fed and therefore very sensitive to climatic conditions. Climate change models indicate an increased likelihood of extreme weather events such as flood, drought, and cyclones leading to severe negative impacts on the agricultural sector in Mozambique. Improved agricultural risk management is one of the core enabling actions of the Group of Eight’s (G-8) New Alliance for Food Security and Nutrition. The Agricultural Risk Management Team (ARMT) of the Agriculture and Environment Services (AES) Department of the World Bank, at the request of the G-8, conducted an agricultural sector risk assessment in Mozambique with several goals: to provide a robust analytical underpinning to the Strategic Plan for the Development of the Agrarian Sector (PEDSA) and the National Investment Plan for the Agrarian Sector in Mozambique (PNISA), to incorporate agricultural risk perspective into deci- sion making, and to build the capacity of local stakeholders in risk assessment and management. Figure ES.1 shows the annual percentage growth in agriculture value added and the impacts of major shocks to the sector. The periods of intense civil war (1985–86 and 1990–92), drought (1994), and flood (2000) resulted in negative growth rates in the agri- cultural sector. Since 2000, the growth rate has not dipped into negative territory, but the occurrence of agricultural risk has had an adverse impact on agricultural growth. The frequent occurrence of agricultural risks creates food affordability and availability problems for vulnerable rural populations and urban consumers, and results in sudden spikes in the food insecure population (figure ES.2). Risk Prioritization xiii FIGURE ES.1. MAJOR SHOCKS TO CROP AND LIVESTOCK PRODUCTION: ANNUAL PERCENTAGE GROWTH IN AGRICULTURE VALUE ADDED 30 20 Drought (2003 and 2005) Drought, Flood and Percentage 10 Cyclone (2007) Drought (2010) 0 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Civil War/Conflict Drought Flood (2012) –10 (1985-1986) (1994) Flood (2000) Civil War/Conflict –20 (1990-1992) Source: World Development Indicators. Update: Annual percentage growth in agriculture value added has dropped from 2011 at 5.8 percent to 2 percent in 2012 and 3.5 percent in 2013. A drop would be expected to be seen after the large flood of 2015.The figures that would reflect the effect of the flood in 2015 are currently not available. FIGURE ES.2. ACUTE FOOD INSECURE POPULATION IN MOZAMBIQUE 900,000 Severe Severe Severe drought Multiple risk events Flooding Drought 55% Drought 800,000 drought drought Flood 25% Cyclone 19% 700,000 600,000 Population 500,000 400,000 300,000 200,000 100,000 0 0 1 2 3 4 5 6 7 8 9 0 1 2 00 /0 /0 /0 /0 /0 /0 /0 /0 /0 /1 /1 /1 00 01 02 03 04 05 06 07 08 09 10 11 /2 99 20 20 20 20 20 20 20 20 20 20 20 20 19 Source: SETSAN reports. Production risks: The main sources of production tion between production risk and production declines risk observed through analysis of data and interviews could not be performed. The bulk of the analysis relied with a range of farmers and other stakeholders are on data gleaned from multiple sources. drought, flood, cyclone, and pests and diseases. The absence of consistent, time-series data on disaggregated Drought was observed to be the most important agri- production output posed a challenge, and thus correla- cultural risk, considered probable with catastrophic xiv Mozambique: Agricultural Sector Risk Assessment consequences. An agricultural drought occurs when soil caused by production risks triggers price spikes in local moisture is significantly deficient, resulting in reduced markets. crop yields and output. This can be the result of low over- all annual rainfall or variations in the timing and duration Abrupt and steep price spikes and falls, often driven by of seasonal rainfall, that is, late onset of rain, early cessa- underlying production deficits, market factors, or other tion of rain, long rain-free periods, and so on. Weather exogenous factors, are a cause of serious concern with analysis shows significant dry years in 1979 and 1992, with severe implications for consumers and producers alike. a lack of rain in specific regions indicated in 1983, 1988, Although the northern region is more self-sufficient in food, 1994, 2005, 2008, and 2009. The analysis also depicts a local events and regional conditions (in Malawi, Tanza- declining trend of deficit rainfall events, highlighting that nia, and so on) contribute to price volatility. The southern cumulative precipitation in Mozambique is increasing, on regions are more likely to experience food deficits and often average. rely on food imports from South Africa. These regions are more exposed to international price volatility, passed down Flooding is the second-most important risk in Mozam- to the domestic markets, as well as local production failures. bique, with a high probability of occurrence and critical Commodity price analysis demonstrated relatively fewer consequences. Incidence of this risk generally occurs episodes of sudden spikes and a general trend of increasing between December and February during the wet season prices, especially over the past five years. The price spike and is often the result of heavy rainfall in a short period of white maize in 2006 was driven by local events (largely of time. This heavy rainfall not only causes flash flood- drought in 2005), whereas the 2008–09 jump was a pass- ing, but also causes rivers to burst their banks and, par- through effect of the global food price crisis. Because of ticularly for farmers around irrigation infrastructure, heavy reliance on imports for rice, there is a resulting direct for dikes to burst and dams to fail. There is often little transmission of international price volatility to domestic time to prepare and effects tend to be localized, com- markets. Among export commodities, tobacco and cash- pared with the often wide-reaching nature of drought. ews do not demonstrate high volatility. Cotton, conversely, Weather analysis highlights significant events in 1981 in is exposed to severe price volatility. the south, 1989 in the north, and in 1999–2001 across the country (including devastating floods seen in 2000). Enabling environment risks: Enabling environment The year 2013 witnessed catastrophic flooding, lead- risk covers many different aspects of legal, institutional, ing to severe losses across the country. The analysis also fiscal, and policy volatility and/or uncertainty that affect indicates a trend of rising excess rainfall events, sug- stakeholders’ ability to undertake their business within a gesting that flooding might become more frequent in sector. In Mozambique, the enabling environment is rela- future. tively weak, but stable, and thereby is not much of a risk for the agricultural sector. Conflict and insecurity, political Cyclones are frequent along the Mozambican coastline instability, and regulatory risk can have an adverse impact during the wet season, but are usually not considered on the agricultural sector. particularly destructive if they are below Category 4. The main damage caused by cyclone events is inflicted Available data on actual losses resulting from adverse on farm infrastructure: houses, storage, and so on, but in events in Mozambique are not particularly accurate or terms of agriculture, most damage is borne by tree crops: consistent within individual data sources. In an attempt to cashews, coconuts, and fruit trees. facilitate comparison and ranking of the costs and losses resulting from various events, different data sources were Market risks: Price volatility (domestic as well as inter- combined to generate a more or less consistent time series. national), exchange rate volatility, input volatility, and Figure ES.3 uses 2000–2015Q1 data from the Technical counterparty risk are some of the major market risks in Secretariat for Food Security and Nutrition (SETSAN) Mozambique. They, however, are less significant than and other sources to quantify the frequency and severity production risks, and in many instances, crop failure of the impacts of major production risks. Risk Prioritization xv FIGURE ES.3. ESTIMATED AGGREGATE TOTAL LOSSES BY RISK EVENT Bubble volume = estimated maximum loss from one event Estimated total losses per event $600,000,000 (constant 2006–2007 I$) $500,000,000 Drought/Intense heat $400,000,000 Flood/Irregular rainfall $300,000,000 $200,000,000 Pest and disease $100,000,000 Wildfire Cyclone $- 0.0 0.1 0.2 0.3 0.4 0.5 Frequency of event 1996–2015Q1 Based on the assessment team’s evaluation, which com- ex post actions, coping solutions are generally expensive bined qualitative and quantitative analysis, drought and and do not transfer or mitigate the risk or help with making flood emerge as the two biggest agricultural risks in the overall sector and those within the sector more resil- Mozambique. These are followed by pest and disease out- ient in the long run. Risk mitigation is perhaps the most breaks, international price volatility, and domestic price needed, and the most ignored, with the highest returns volatility, which are often difficult to qualify. in addressing short- and long-term risk issues in Mozam- bique’s agricultural sector. It is important to highlight that To address the priority risks, the assessment deployed a most of these potential interventions are complementary holistic agricultural risk management framework, com- in nature, and most of them are required to effectively posed of mitigation (action taken to reduce the likelihood address agricultural risks. Nonetheless, considering the of events, exposure, and/or potential losses), transfer resource limitations, decision filters (see table ES.1) were (risk transfer to a willing party, at a fee or premium), and used to help evaluate and prioritize interventions with the coping solutions (activities geared toward helping cope greatest potential to generate sizable risk management with losses) to identify a list of potential interventions. benefits. Risk transfer solutions (insurance and hedging), owing to Mozambique’s specific context, have limited applicability Based on the prioritization of risk and intervention meas- and will be quite challenging to implement. Microlevel ures, the following three intervention categories might options such as public-private partnership would require yield greatest risk management benefits: significant public investments in data and substantial 1. Water management: Flood and drought are growth in credit or input utilization. At the mesolevel, the two biggest risks for the agricultural sector low levels of rural lending and lack of direct regulatory in Mozambique and effective and efficient water incentives for lenders to transfer portfolio-level agricul- management is a necessary precondition for man- tural risk means any form of unsubsidized mesolevel agri- aging both. This might entail (a) improved irriga- cultural insurance program would have limited uptake tion for drought management, (b) flood control and be unsustainable. Macrolevel sovereign agricultural infrastructure, and (c) improved water manage- risk financing and insurance would need to be looked at ment practices. carefully through the lens of what the government thinks a. Irrigation: Irrigation has the potential to gen- their contingent liability to the agricultural sector is. Cop- erate a sizable gain in household welfare, boost ing solutions (social safety net programs) are required and agricultural growth, improve food security, quite important in Mozambique; however, they do not mitigate the impact of drought, and promote address fundamental risk issues in the agricultural sector overall economic growth in Mozambique. and have limited applicability as a long-term solution. As However, the performance record of irrigation xvi Mozambique: Agricultural Sector Risk Assessment TABLE ES.1. FILTERING OF RISK MANAGEMENT INTERVENTIONS AND POTENTIAL FOR MULTIPLE “WINS” Reduces Reduces Compensates Improves Climate Change Climate Change the Risk the Losses after the Loss the Yield Mitigation Adaptation Soil and water conservation Y Y N Y Y Y measures (for example, conservation agriculture) Improved access to Y Y N Y Y Y extension services Improved water Y Y N Y N Y management practices Altered cropping patterns Y Y N N Y Y Flood control Y Y N Y N Y infrastructure investment (dikes, drainage, and so on) Small-scale irrigation Y Y N Y N Y Large-scale Irrigation Y Y N Y N Y Improved market Y Y N N N N information system Building tolerant varieties Y Y N N N Y system (flood, drought and disease-tolerant varieties) Timely and reliable N Y N Y N Y availability of weather information to farmers and other stakeholders Regional coordination Y Y N N N Y Promotion of integrated Y Y N Y N N pest management (IPM) Subsidized crop insurance N N Y Y N N (for example, bundled with credit or input) Commercial catastrophic N N Y N N N weather insurance Sovereign risk financing N N Y N N N Saving/credit N N Y N N N On-farm storage N N Y N N N Social safety net programs N N Y N N Y (for example, food/cash/ vouchers for work, food aid) Facilitate temporary N N N N N N migration Risk Prioritization xvii schemes is very mixed, which at times could dams, Ngare or Mhindu ridging, afforestation/ result in increasing exposure to risk (for exam- reforestation, conservation agriculture prac- ple, regular flooding). To reverse the declin- tices) are effective and efficient mechanisms for ing investment in irrigation, the government mitigating the risk of droughts and/or floods. of Mozambique approved a National Irriga- In addition, they yield significant productivity tion Strategy for 2011–19. If designed appro- gains and help in mitigating the effects of cli- priately, irrigation systems could help reduce mate change. drought risk and manage flood risk. b. Improved access to extension services: b. Flood control infrastructure: Thirteen big Improved access to extension services would rivers flow across from southern Africa and allow producers to be better informed, and to through Mozambique, exposing the agricul- access advice, technology, and inputs to alter tural sector relying on these rivers to frequent their agronomic practices in view of the cur- flooding. Dams, dikes, and drainage systems rent and emerging risk profile of agricultural are some of the flood control infrastructure sector. Although the coverage of extension ser- that can effectively mitigate the impact of vices is increasing, it only reaches 12 percent flood. Unfortunately, the existing condition of of farming households in Mozambique. The flood infrastructure in Mozambique is poor bulk of the farming households do not have and in dire need of repair. Furthermore, the access to any extension services and further current flood control infrastructure must be investment and expansion of extension services significantly expanded to adequately address and development of new delivery channels will frequent flooding as well as extreme flood assist in improved management of agricultural events. risks. c. Improved water management practices: c. Altered cropping patterns: Over the past Water deficit (drought) and water excess (flood) few decades, Mozambique has witnessed are likely to pose greater risks as a result of changes in traditional cropping patterns. For climate change; hence, effective water man- instance, millet and sorghum (more drought agement has to be the cornerstone of any risk tolerant) have replaced maize (a more sensitive management strategy. This involves the plan- crop), especially in low rainfall areas, resulting ning, developing, distributing, and managing in increased risk exposure. Losses from drought the optimal use of water resources. Activi- could be significantly reduced by replacing ties such as improving data and reforming maize with sorghum, millet, or root crops in water governance—along with education and areas where drought is particularly common. training on water management—would aid Root crops are generally more drought tolerant in water availability, particularly in drought- than are cereals. The trend of climate change prone areas. will likely alter cropping calendars and seasonal 2. On-farm practices for improved risk agro-climatic conditions, which will necessitate management: Management of drought, flood, changing cropping pattern and practices to and pest and diseases requires effective on-farm better adapt to climate change. practices on individual agricultural plot and com- 3. Farmer access to information (for exam- munity structures. This includes (a) soil and water ple, weather, price, diseases, early warn- conservation measures, (b) improved access to ing): Ready access to timely, accurate, and extension services, and (c) altered cropping pat- localized information about impending events terns. that could have severe impacts on crops is a pre- a. Soil and water conservation measures, requisite for enabling any preemptive actions by including conservation agriculture: Soil farmers to reduce exposure or losses. This may and water conservation measures (such as sand mean relocating and minimizing losses before a xviii Mozambique: Agricultural Sector Risk Assessment flood or postponing planting or early harvesting, meaningful impact on the agricultural sector in Mozam- as well as altering agronomic practices. Besides bique. This would require understanding the landscape helping managing risks, this information could of these interventions, assessing their relative efficacy, also help farmers manage inputs better and understanding principal challenges to success and scale, improve yields. and identifying leverage points and necessary interven- tions to increase access to a wide majority of agricultural These three intervention categories align with PNISA sector stakeholders. Assessing solutions to help prioritize and PEDSA, and many of them are being implemented specific interventions, scaling up priority programs, and and are having positive impacts, albeit at a smaller, local- putting in place a risk management implementation plan ized level. Greater emphasis should be placed on scal- will be the next steps in the process of building resilience ing up these interventions to the national level to make a in Mozambique’s agricultural sector. Risk Prioritization xix CHAPTER ONE INTRODUCTION AND CONTEXT Agricultural risk management is a central issue that Mozambique faces in develop- ment, and multiple stakeholders have analyzed this challenge, sometimes with differ- ent terminology and focusing on varying aspects.1 Risk is considered the probability of harmful consequences or expected losses resulting from interactions between nat- ural or human-induced hazards and vulnerable conditions.2 This is in comparison with trends that are longer term or “chronic” patterns (reversible or irreversible) that provide context for impact on agriculture or constraints that are existing conditions/ bottlenecks that hamper the smooth functioning of the sector and lead to suboptimal performance. The government of Mozambique has adopted the Strategic Plan for Agricultural Development ([PEDSA]3 2010–19) that focuses on (i) increasing the avail- ability of food in order to reduce hunger through growth in small producer produc- tivity and emergency response capacity; (ii) enlarging the land area under sustainable management and the number of reliable water management systems; (iii) increasing access to the market through improved infrastructures and interventions in marketing; and (iv) improving research and extension for increased adoption of appropriate tech- nologies by producers and agro-processors. To help implement PEDSA, in May 2013 the government adopted the PNISA, which has also been accepted by all the major stakeholders as the roadmap for national agricultural development. Components of the 2010–19 PEDSA seek to minimize and mitigate agricultural risks as well as to prepare the sector for dealing with climate change. There are also World Bank operations of note involved in making the sector more resilient and improving access to predictive information for farmers and early warning systems including the ongoing climate change development policy operation 1 Related reports that address aspects of agricultural risks include, among others: (FAO/WFP, 2010); (SETSAN, 2010); (IFPRI, 2011); the World Bank’s Agribusiness Indicators: Mozambique (2012); (System), 2012); (FEWSNET, 2006); and (INGC, 2009b). 2 Definition as found on World Bank—Forum for Agricultural Risk Management for Development (FARMD) website. 3 PEDSA is integrated into the instruments established by the National Planning System such as, the Green Revolu- tion Strategy, the Priorities of the Agriculture Sector, the Research Strategy, the National Extension Programme, the Re-forestation Strategy, the National Forestry Plan, the Irrigation Strategy, the Food Production Action Plan, and the Strategic Plan for Livestock. Risk Prioritization 1 series that has so far supported thirteen cross-cutting and 3. Analyze existing and potential risk management sector-level reforms.4 However, the proposed actions are strategies (mitigation, transfer, and coping); somewhat piecemeal and do not constitute a coherent 4. Identify, analyze, and prioritize principal risk man- approach to addressing the risks to the agricultural sector. agement interventions; and 5. Build capacity of local stakeholders in agricultural Improved agricultural risk management is one of the risk management. core enabling actions of the G-8’s New Alliance for Food Security and Nutrition. Accordingly, the Agricultural Risk The World Bank’s Agriculture Sector Risk Assessment Management Team of the Agriculture and Environment takes a holistic approach and relies on long time-series Services Department of the World Bank, at the request of historical data to arrive at an empirical and objective the G-8, conducted an agricultural sector risk assessment assessment of agricultural risks and their impacts on in Mozambique. This exercise aimed to provide a robust Mozambique. A lengthy desk study of existing data and analytical underpinning to PEDSA and PNISA, incor- analyses was completed followed by an in-country field porate agricultural risk perspective into decision making, study to corroborate findings and present the risk assess- and build capacity of local stakeholders in risk assessment ment to stakeholders. A joint World Bank/USAID team and management. undertook a technical mission to Mozambique April 2–19, 2013, for this purpose, including focus group dis- This activity was carried out in collaboration with the cussions with farmer groups and a multitude of field Ministry of Agriculture and the financing was provided interviews with various stakeholders (farmers, input sup- by the Feed the Future program in Mozambique, which is pliers, service providers, traders, processors, exporters, an initiative of the U.S. Agency for International Devel- financial intermediaries, and government agencies). The opment (USAID). In addition, a multidonor trust fund team presented their preliminary findings to stakeholders supported by the Dutch Ministry of Foreign Affairs and in Maputo on April 18, 2013. This assessment will form the Swiss Secretariat of Economic Affairs (SECO) also the basis of the second step, solution assessment, whose made a financial contribution toward this work. final findings will inform PNISA. The solution assessment conducted in November 2014 looks into the three prior- The ARMT of the Agriculture and Environmental Ser- itized interventions that were discussed with the govern- vices Department and representatives from Financial ment after delivery of this report and include an extension and Private Sector Development, Non-Bank Financing risk landscape mapping for each intervention considering Institutions Unit FCMNB (supported by the Agricul- existing tools and interventions. More information can be tural Insurance Development Program) conducted this found in the Mozambique Agricultural Sector Risk Solu- Mozambique Agriculture Sector Risk Assessment to: tions report, but they include mitigation interventions in 1. Identify, analyze, quantify, and prioritize princi- relation to water management, conservation agriculture, pal agricultural risks (production, market, and and access to information systems and extension services. enabling environment risks) facing the agricultural sector; This document considers the many aspects of assessing 2. Analyze the impact of agricultural risks on risk in the Mozambican agriculture sector. Chapter 2 major agricultural sector stakeholders (farmers, introduces the major characteristics of the agricultural vulnerable populations, commercial sector, and system leading into chapter 3, which presents a compre- government); hensive picture of the risks that exist in the sector. Chap- ter 4, in quantifying the risks that have been observed, comments on the losses that have been incurred by the 4 These include introduction into MASA of a national action plan on climate sector because of production risks, whereas chapter resilient agriculture to support PEDSA implementation, ongoing WB support 5 provides a qualitative discussion of how risk has an for the Transforming Hydro-Meteorological Services Project and development of digital elevation models to improve flood modeling in the Zambezi and Lim- effect on the different stakeholders present in the sector. popo Valleys. Chapter 6 delves into the risk prioritization carried out 2 Mozambique: Agricultural Sector Risk Assessment by the team and then comments on various manage- calculating the production risk losses; a report produced ment measures. The report concludes with chapter 6, in as background for the report on available weather yield which recommendations are provided for improving risk data; elements on vulnerability analysis; and a litera- management in Mozambique. There are several appen- ture review to provide an impact assessment of climate dixes that include information on the methodology for change on agriculture. Risk Prioritization 3 CHAPTER TWO MOZAMBIQUE’S AGRICULTURAL SYSTEM Agriculture plays a key role in Mozambique’s economy, making up 31.8 percent of GDP and providing a livelihood to almost 81 percent of the labor force (CIA 2013). The majority of Mozambican agriculture is rain fed and therefore very sensitive to climatic conditions. The agricultural sector in Mozambique is dominated (99 percent) by smallholder farmers using family labor, most of whom cultivate small plots of land (0.5 to 1.5 hectares [ha]) with limited inputs and mechanization.5 According to data from the Mozambique Agricultural Census (TIA) surveys, the total farmer population is estimated at 3.8 million. Producers were organized into small organizations and forums, although these only accounted for 7.2 percent of farmers in 2008 (Government of Mozambique 2008). Following the Mozambique civil war, which ended in October 1992, increases in agri- cultural production have been credited for a reduction in poverty, as people were able to return to previously abandoned land. Despite the increased use of land for agricul- tural activity, low productivity is still considered a major constraint on development (Cungara and Garret 2011). Mozambique is facing a natural resources boom, and agricultural exports have declined from 31.4 percent of total export revenue in 2002 to 14.4 percent in 2008 because of large-scale expansion of aluminum and electricity exports (World Bank 2012a). Even so, the agricultural sector remains largely responsible for job creation and the government and donors alike consider agricultural development as a driving force for poverty reduction. AGRO-ECOLOGICAL CONDITIONS The agro-ecological zones in Mozambique include three macro-agro-ecological zones: northern, central, and southern. These macro-agro-ecological zones are based on climate, vegetation, altitude, soils, and farming systems. 5 A national survey in 2007 found that only 4 percent of farmers use fertilizer (FAO/WFP 2010). Risk Prioritization 5 TABLE 2.1. NUMBER OF AGRICULTURAL AND LIVESTOCK PRODUCTION UNITS IN MOZAMBIQUE Type of Production Unit Small Medium Large Total Countrywide 3,801,259 25,654 841 3,827,754 Percentage of distribution per type on total # of units 99.31% 0.67% 0.02% Cultivated area Countrywide 5,428,571 130,651 73,565 5,632,787 Percentage of cultivated land per type of units 96.37% 2.32% 1.31% Average land per type of unit (ha) 1.43 5.09 87.47 Source: Agricultural and Livestock Census 2009/10. In the north, between the Zambezi and Rovuma rivers, FIGURE 2.1. AGRO-ECOLOGICAL the major crops are maize, cotton, coconuts, cashews, cas- CONDITIONS— sava, sorghum, millet, and groundnuts. The central region MOZAMBIQUE between the Save and Zambezi rivers produces maize, cotton, cassava, bananas, citrus, sugarcane, vegetables, sorghum, cashews, and rice. In the southern region, south of the Save River, crops include maize, rice, groundnuts, cowpeas, cassava, citrus, sugarcane, vegetables, and cash- ews (Mucavele 2000). These three macrozones, are composed of 10 diverse agro- ecological zones. They are pictured here in figure 2.1. Arid and semiarid zones dominate the south and south- west, whereas subhumid zones make up the majority of the center and north, and humid highlands are found pri- marily in the central provinces (United Nations Food and Agriculture Organization [FAO]/World Food Programme [WFP] 2010). Agriculture is practiced in all zones, with the exception of highly arid regions in the south and south- west part of Gaza province, which are only suitable for livestock. As noted in appendix B, this is due in particular to the increased incidence of extreme climatic events. The impact of climate change, if not addressed, may reverse progress made in agricultural development. area (FAO/WFP 2010). Access to basic infrastructure in the rural areas is limited, as are support services including LAND AND WATER information and institutions. Irrigation potential is about RESOURCES 3.0 million ha, but the Ministry of Agriculture reports The country’s climate and land are suitable for a wide vari- only 120,000 ha having irrigation infrastructure, and only ety of annual and perennial crops, along with livestock. 60,000 ha as operational under public irrigation schemes in The total cultivated area grew from 3,867,000 ha in 2002 2010 (Chiloda et al. 2012). According to the joint Food and to 5,632,787 ha in 2009/10 (Government of Mozambique Agriculture Organization /World Food Programme Crop 2009/2010), but according to the World Bank in 2011, and Food Security Assessment Mission to Mozambique in permanent cropland made up only 0.3 percent of the land 2010, only 55,000 ha of land are irrigated, of which 35,000 6 Mozambique: Agricultural Sector Risk Assessment ha are under sugarcane with the remainder mostly under TABLE 2.2. AGRICULTURAL GDP SHARE rice and vegetables (FAO/WFP 2010). As the center and IN MOZAMBIQUE, 2009 south are characterized by irregular rainfall, irrigation Agricultural infrastructure is concentrated in Gaza, Zambezia, Tete, Commodity GDP Share (%) Category and Manica (Government of Mozambique Ministry of Maize 21.23 Cereals (food crop) Agriculture 2010). The World Bank in 2011 estimated that Cassava 18.34 Root crop (food 73.9 percent of annual total freshwater withdrawals were crop) for agriculture. Lack of irrigation is one of the main con- Vegetable 8.21 Horticulture straints to increasing production and mitigating the effects Poultry 4.70 Livestock of droughts in many regions of Mozambique. Rice 4.41 Cereals Groundnuts 3.06 Pulses, nuts, and CROP PRODUCTION oilseeds SYSTEMS Cashew nuts 2.51 Export-oriented crop Cotton 0.75 Export-oriented crop Main staple crops produced in the family agriculture sec- Tobacco 1.17 Export-oriented crop tor are maize, sorghum, rice, millet, potatoes, sweet pota- Sugarcane 0.98 Export-oriented crop toes, cassava, and beans. Cash crops frequently grown by Sorghum 2.56 Cereal food crop households include cotton, tobacco, copra, cashew nuts, Total 67.92% sesame, sugar beans, sunflower, bananas, and sugarcane. Source: Pauw et al. 2012. Many family farms combine food crops with a single cash crop. The bulk of the agricultural sector is represented by such as maize, rice, and cassava, and concession agricul- the 11 commodities listed in table 2.2, which accounted ture such as sugarcane and bananas) (Cungara and Gar- for 67 percent of agricultural GDP in 2009 (Cungara and ret 2011). The large increase in area harvested since 1992 Garret 2011). These commodities and risks affecting their can be seen in figures 2.2 and 2.3. The lack of increase production are examined in chapters 3–6. in area harvested for cassava is notable in figure 2.4, although this confirms its place as a crop primarily for PRODUCTION TRENDS self-consumption and food security, with little commer- The growth of the agricultural sector between 2003 cialization. Figures 2.2 and 2.3 further confirm that there and 2009 averaged between 7.4 and 7.5 percent a year. has not been a huge improvement in yield for the major Much of this growth is attributed to expansion of the food crops over the same time period, apart from a slight cultivated area (particularly increased area for food crops improvement in maize. FIGURE 2.2. MAIZE PRODUCTION, 1992–2011 Area harvested (ha) Production (tons) Yield (tons/ha) 2,500,000 1.40 1.20 2,000,000 1.00 Ha and tons 1,500,000 Tons/ha 0.80 1,000,000 0.60 0.40 500,000 0.20 0 0.00 11 92 93 94 95 96 98 99 00 01 02 03 04 05 06 08 09 10 97 07 20 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 19 20 Year Source: FAOSTAT, Author’s Calculations. Risk Prioritization 7 FIGURE 2.3. RICE PRODUCTION, 1992–2011 Area harvested (ha) Production (tons) Yield (tons/ha) 300,000 1.40 250,000 1.20 1.00 Ha and tons 200,000 Tons/ha 0.80 150,000 0.60 100,000 0.40 50,000 0.20 0 0.00 11 92 93 94 95 96 98 99 00 01 02 03 04 05 06 08 09 10 97 07 20 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 19 20 Year Source: FAOSTAT, Author’s Calculations. FIGURE 2.4. CASSAVA PRODUCTION, 1992–2011 Area harvested (ha) Production (tons) Yield (tons/ha) 7,000,000 9 6,000,000 8 7 5,000,000 6 Ha and tons Tons/ha 4,000,000 5 3,000,000 4 3 2,000,000 2 1,000,000 1 0 0 11 92 93 94 95 96 98 99 00 01 02 03 04 05 06 08 09 10 97 07 20 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 19 20 Year Source: FAOSTAT, Author’s Calculations. In general, however, production growth and productivity in food crops such as maize, cassava, sweet potatoes, sor- are significantly below the averages for Southern Africa as a ghum, and groundnuts (World Bank 2012). region. More than half of the rural population remains poor and food insecure. According to the Ministry of Agriculture Nationally, however, there is a large deficit in rice pro- and Food Security (MASA)/Mozambique Ministry of Indus- duction and wheat is a major import for domestic con- try and Commerce (MIC) Food Balance data, Mozambique sumption, with imports of cereals almost equivalent to all faces a yearly food deficit of about 500,000 metric tons (MT) agricultural exports combined. in cereals, even after commercial imports and food assistance. CASH CROPS FOOD CROPS The major cash crops in Mozambique are cotton, tobacco, The majority of production undertaken by smallholder and sugarcane. Other cash crops include copra, cashews, families is for consumption. This production has espe- sesame, sunflower, coconuts, bananas, and other horticul- cially low yields and productivity. The main food crops in ture. Cotton has seen a steady increase in production since Mozambique include cassava and sweet potatoes, maize, the late 1990s. Tobacco and sugarcane have both seen rice, sorghum, millet, and pulses. Although characterized huge spikes in production beginning in the early 2000s by “generally weak performance,” staple food crop pro- following the installation of processing facilities to cap- duction has increased and Mozambique is self-sufficient ture value and reduce postharvest losses. Figure 2.5 shows 8 Mozambique: Agricultural Sector Risk Assessment FIGURE 2.5. MOZAMBIQUE AGRICULTURAL EXPORT SHARES IN US$, 2010 2010 Agricultural Export Shares Mozambique in US$ Value (FAOSTAT) Maize Groundnuts shelled 1% + groundnuts, with shell 2% Molasses Other 2% 11% Cashew nuts shelled + cashew nuts, with shell 4% Tobacco, unmanufactured 36% Cotton lint + cottonseed 7% Sugar raw centrifugal and sugar Sesame seed refined 6% 25% Pulses, nes 6% Source: FAOSTAT. the quantity and value of exports in 2010 and the large FIGURE 2.6. CONSUMPTION AND percentage of exports that came from tobacco and raw PRODUCTION OF POULTRY sugar. Starting in 2011, cassava is becoming a cash crop as IN MOZAMBIQUE, 2005–09 the result of SABMiller’s cassava beer project in Nampula. 30,000,000 Consumption 25,000,000 Production LIVESTOCK PRODUCTION 20,000,000 Unit/Kg SYSTEM 15,000,000 Livestock has an important place in the livelihoods of 10,000,000 small and medium producers in Mozambique. In 2008, 5,000,000 58.8 percent of small and medium producers had chick- 0 2005 2006 2007 2008 2009 ens, 26.2 percent had goats, 12.1 percent had pigs, Source: Associação Moçambicana de Avicultura. 11.4 percent had ducks, and 6.7 percent had cattle. How- ever, only 10.9 percent used any sort of animal traction for production on their land (Government of Mozambique ing countries (Government of Mozambique, Ministry of 2008). Livestock made up 10 percent of total agricultural Agriculture 2010). Almost all of the inputs (rations, con- production but contributed only 1.7 percent of GDP in centrates, medicines, vaccines, veterinary instruments, 2008 (OIE 2008). and equipment) for the sector are also imported. Mozambique relies heavily on imports for every com- There are some constraints that are specific to livestock modity in the livestock sector and figure 2.6 clearly shows production and have an impact on the productivity of the consumption outstripping production for poultry. It was sector: estimated by the government that over 40 percent of beef » The low genetic quality of the breeding animals, as consumed in Mozambique is imported from neighbor- well as unsuitable management practices. Risk Prioritization 9 » The weak network of veterinary assistance to The principal constraints to the agricultural sector in smallholders. Mozambique include: » The lack of infrastructure for watering and man- » Limited input usage, availability, and market aging cattle. dynamics. » The vast majority of slaughterhouses that are » Limited access to agricultural extension services: without running water, refrigeration, separation approximately 8.3 percent of farmers in 2008 between clean and dirty areas, and pens in which had access to agricultural extension services the animals can be held before slaughter. despite efforts by the government and partner- ing nongovernmental organizations (NGOs) The poultry subsector has experienced significant (Government of Mozambique, Ministry of Agri- improvements in production (figure 2.6) and productiv- culture 2010). ity in recent years. The growth in poultry is traceable to » Limited access to credit: in Mozambique, less than a small but dynamic commercial subsector that has been 6 percent of total lending in 2010 was dedicated successfully replacing imports in meeting the growing to agriculture, down from about 10.5 percent in demand in the urban centers for poultry. 2004. A limited group of so-called traditional prod- ucts (tea, sugar, cashews, sisal, coconuts, and cot- ton) are the main recipients of agriculture credit, PRINCIPAL CONSTRAINTS account for 67.7 percent. Since 2004, only sugar This document primarily analyzes agricultural risks, but it is and cashews show consistent growth in financing. also important to understand some underlying constraints, In contrast, tea, coconuts, sisal, and, most recently, which can, in turn, aggravate the risk profile of a number of cotton have decreased. commodities and play a role in risk management. Whereas » Lack of infrastructural facilities such as transporta- risks are defined by attributes of uncertainty, events, and tion, storage, and irrigation. losses, constraints are classified as conditions that lead to » Limited investment in the agricultural sector by the suboptimal performances by supply chain actors. government and private sector. 10 Mozambique: Agricultural Sector Risk Assessment CHAPTER THREE AGRICULTURAL SECTOR RISKS This chapter considers the types of risks that are observed in the agricultural sector and groups them into three main categories: production, market, and enabling envi- ronment. This chapter is a synthesis of analysis conducted by the assessment team as well as detailed review of secondary data. A detailed weather analysis (appendix B) was conducted to understand weather patterns (deficit and excessive rainfall) and their impact on crop yields. Appendix C reviews the existing literature on climate change and appendix D summarizes the vulnerability situation in Mozambique. Detailed com- modity risk profiles highlight principal risks to nine major commodities in appendix E. PRODUCTION RISKS The main sources of production risk observed through analysis of data and inter- views with numerous farmers are drought, flood, cyclone, and pests and diseases. Figure 3.1 shows the annual percentage of growth in agriculture value added and the declines that occur around major shocks to the sector. The periods of intense civil war (1985–96 and 1990–92), drought (1994), and flood (2000) resulted in negative growth rates in agriculture. Since 2000, the growth rate has not dipped into negative territory, although it has reflected the impact of several critical events (figure 3.1). The weather analysis in appendix B provides further information on provincial trends in rainfall using data from 31 weather stations from January 1979 to December 2009. The climate is considered humid tropical in the north and on the coast, and dry tropi- cal in the south and interior (Bay 1997). Soils with good agricultural potential exist in the north, central, and western parts of the country and those with limited potential occur in the southern part, in the low plateau (Bay 1997). The rainy season occurs in November/December and lasts until March/April. Annual rainfall is generally up to 400 mm in the south and central western areas and more than 1,200 mm in the north. DROUGHT Drought was observed to be the most important agricultural risk, considered probable and with catastrophic consequences. An agricultural drought occurs when soil moisture is significantly deficient, resulting in reduced crop yields and output. This can be the Risk Prioritization 11 FIGURE 3.1. MAJOR SHOCKS TO CROP AND LIVESTOCK PRODUCTION: ANNUAL PERCENTAGE GROWTH IN AGRICULTURE VALUE ADDED 30 20 Drought (2003 and 2005) Drought, Flood and Percentage 10 Cyclone (2007) Drought (2010) 0 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Civil War/Conflict Drought Flood (2012) –10 (1985-1986) (1994) Flood (2000) Civil War/Conflict –20 (1990-1992) Source: World Development Indicators. Update: Annual percentage growth in Agriculture Value Added has dropped since 2011 at 5.8 percent to 2 percent in 2012 and 3.5 percent in 2013. A drop would be expected to be seen after the large flood of 2015. The figures that would reflect the effect of the flood in 2015 are not currently available. result of low overall annual rainfall amounts or variations the people reported to have been affected, deaths that in the timing and duration of seasonal rainfall, that is, late occurred, and any indication found in regard to economic onset of rain, early cessation of rain, and so on. Adverse damages and hectares of planted crops or production lost timing and duration are particularly critical if they occur because of negative abnormalities in precipitation. during the crop production cycle (seeding, flowering, and grain filling) with even a short absence of rain potentially HEAT AND EXCESSIVE TEMPERATURE having a detrimental effect on production. Long dry spells Heat and excessive temperature can be serious risks, espe- (consecutive days without rain); limited, sporadic rainfall; cially for poultry and horticultural commodities. This was as well as excessive participation are collectively respon- indicated by poultry farmers as a source of mortality for sible for significant crop losses in Mozambique. chickens and by horticultural producers as a source of losses. At times, heat waves and extreme temperatures are The weather analysis graph (figure 3.2) shows that the experienced alongside drought and are difficult to identify weather stations indicated a negative anomaly of precipi- separately. tation, in other words, a possible drought. By counting all the stations that had a negative anomaly of precipita- tion (considered as a drought event), the following chart FLOOD (figure 3.2) summarizes the number of total stations that Flooding is also a significant risk in Mozambique, with a had drought events per year with colors representing the high probability of occurrence and critical consequences. different provinces (red = North, green = Central, blue Incidence of this risk generally occurs between December = South). The weather station analysis shows significant and February during the wet season and is often the result dry years in 1979 and 1992 with a lack of rain in specific of heavy rainfall in a short period of time. This heavy regions indicated in 1983, 1988, 1994, 2005, 2008, and rainfall not only causes flash flooding, but also leads rivers 2009. Figure 3.2 also shows a declining trend of deficit to burst their banks and, particularly for farmers around rainfall events, highlighting that cumulative precipita- irrigation infrastructure, for dikes to burst and dams to tion in Mozambique is increasing, on average. Table 3.1 fail. There is often little time to prepare and effects tend provides further details on drought events, including to be localized compared with the often wide-reaching 12 Mozambique: Agricultural Sector Risk Assessment FIGURE 3.2. DEFICIT RAINFALL EVENTS BASED ON WEATHER STATION DATA ANALYSIS Map Gaza Inh Sof Man Zam Tete Nam 18 Nia C.Del 16 14 12 Number of stations 10 8 6 4 2 0 19 9 19 0 19 1 19 2 83 19 4 85 19 6 19 7 88 19 9 90 19 1 19 2 19 3 19 4 95 19 6 97 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 08 09 7 8 8 8 8 8 8 8 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 19 19 19 19 19 19 20 Year TABLE 3.1. SUMMARY OF DROUGHT EVENTS Estimation of People Total Economic Year Province Affected Damages (US$) Hectares/Production Lost 1983 Cabo Delgado, Nampula, Tete, 4,750,000 419,200,000 Unknown Manica 1991/92 Limpopo River Basin 3,300,000 50,000,000 Unknown 1999 Nampula Unknown Unknown 10,000 ha of maize, sorghum, peanuts, rice, beans, cotton, and horticulture 2002 Maputo, Gaza, Inhambane, Sofala, 600,000 Unknown 160,629 ha Tete, Zambezia, Manica 2003 Central and Southern Unknown Unknown Almost total failure of maize crop 2005 Maputo, Gaza, Inhambane, 1,400,000 Unknown 317,000 ha Manica, Sofala, Nampula 2007/08 Inhambane 520,000 Unknown Unknown 2010 (March) Central and South 465,000 3,000,000 605,000 ha Source: The annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), Famine Early Warnings System Network (FEWSNET)/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from the National Institute for Disaster Management (INGC)—Government of Mozambique data. nature of drought. However, the number of physically Figure 3.3 shows the weather station data analysis for excess displaced people is higher with floods and occurs immedi- rainfall events, indicating positive anomaly records for rain- ately, which brings additional problems. fall, with significant events in 1981 in the south, in 1989 Risk Prioritization 13 FIGURE 3.3. EXCESS RAINFALL EVENTS BASED ON WEATHER STATION DATA ANALYSIS 16 Map Gaza Inh Sof Man Zam Tete Nam Nia C.Del 14 12 Number of stations 10 8 6 4 2 0 19 9 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 19 9 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 20 8 09 7 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 19 Year in the north, and in 1999–2001 across the country (includ- » Pests: Wild animals (boars, monkeys, elephants, ing devastating floods in 2000). Figure 3.3 helps to visualize birds), insects that there were four dry years: 1979 stands out as a year in which almost all provinces experienced insufficient rainfall There is a lack of detailed information on losses caused by whereas 1992 follows closely as another dry year. Both 1983 pests and diseases, but table 3.3 gives an indication of the and 1988 were also dry years with the lack of rain expe- occurrences. Farmers stated that although pest damage rience in some provinces only. The graph also indicates a is highly probable and causes considerable damage, pests trend of rising excess rainfall events, highlighting that flood- are controllable if inputs are available. Subsequently, the ing might become more frequent in future. As the number government, along with commercial organizations and of events listed in table 3.2 indicates, for some provinces NGOs,6 has undertaken aerial spraying for locusts and localized flooding is a regular occurrence, although farmers grain-eating birds, as well as insecticides, pesticides, fungi- will remember 2000 and 2013 as being the most significant cides, traps for fruit fly management, and so on. events during which they lost the most in recent years. WILDFIRES PESTS AND DISEASES As table 3.4 indicates, wildfires can lead to a large number The main crop pests and diseases are as follows: of hectares being lost to agriculture. In Mozambique, wild- » Insects: Red and elegant locusts, lizards, wheat fires can be caused by strong winds and high temperatures. worm, large grain borer, stinkbugs, armyworms, Losses are also often recorded in various data sources bedbugs, fruit fly that have been caused by fires started by arson/human » Diseases: Cassava fungus, peanut leaf curl, root error/poaching, which can also lead to significant local- rot, nematodes (tobacco, tomatoes, beans, fruit), brown streak disease, avian flu, Newcastle disease, 6 ACDI/VOCA in coconuts, IRRI in rice, FAO for locust spraying, and FAO Trypanosoma, powdery mildew, tomato virus and World Bank for fruit fly pest management. 14 Mozambique: Agricultural Sector Risk Assessment TABLE 3.2. SUMMARY OF FLOOD EVENTS Estimation of Total Hectares/Production Year Province People Affected Economic Damages (US$) Lost 1985 Maputo 500,000 Unknown Unknown 1988 Maputo 90,000 Unknown Unknown 1990 Central 12,000 Unknown Unknown 1996 Central and South 200,000 Unknown 170,000 ha of food crops including 45,000 ha of maize 1997 Tete, Sofala, 43,000 Unknown 103,000 ha Zambezia, Manica 1999 Inhambane Unknown 12,400,000 63,000 ha 2000 South 4,500,000 100,000,000 12% of cultivated land, about 198,000 ha 2001 Zambezia, Northern 549,326 36,000,000 71,000 ha Sofala, Tete, Manica 2003 Nampula, Zambezia 100,000 Unknown 237,000 ha of crop fields of beans, cassava, and thousands of cashew trees 2006 Zambezia Unknown Unknown 6,754 ha 2007 (Jan) Inhambane, Sofala, Unknown 100,000,000 288,000 ha in crops Tete, Manica, Zambezia 2007 (Dec) Sofala, Manica Unknown 71,000,000 Failure of staple maize crop and reduced yields in others. 2008 Tete, Zambezia, Tens of thousands Unknown More than 160,000 ha Manica, Sofala 2011 Zambezi, Lucheringo, 180,000 Unknown 10,000 ha Pungue rivers 2011 Maputo 123 families Unknown 12,974 2012 Gaza Unknown Unknown 4,898 ha 2013 Gaza, Inhambane, 240,000 Unknown 153,000 ha approximated Zambezia 2015 Zambezia, Manica 157,172 0.5% drop in predicted 87,000 ha economic growth rate Source: The annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data, and FAO. ized losses. However, wildfires are not cited by farmers and The main damage caused by cyclone events is inflicted stakeholders as a major risk and are classed as having a on farm infrastructure: houses, storage, and so on. In remote probability of occurring with negligible loss. terms of agriculture, most damage is done to tree crops (cashews, coconuts, and fruit trees) where trees are CYCLONES/TROPICAL STORMS blown down and destroyed. Table 3.5 gives an indica- Cyclones occur often along Mozambique’s coastline tion of the Category 4 and above cyclones that have hit during the wet season, but are usually not considered the Mozambican coastline since 1985 and the damages particularly destructive if they are below Category 4. incurred. Risk Prioritization 15 TABLE 3.3. SUMMARY OF PEST AND DISEASE EVENTS Type of Pest/ Year Province Disease People Affected Hectares Lost 1998/99 Cabo Delgado Red locust Unknown Rice and sorghum (four districts) destroyed 1998/99 Nampula Coast Locusts and fungi Unknown Sorghum, millet, and rice and Cabo crops severely affected Delgado 2000 Inhambane Locusts Unknown 160,000 ha of mainly cassava 2001/02 Unknown Red-beaked sparrows Unknown 30% of planted area of rice 2001/02 Cabo Delgado Wild animals, others Unknown 66,490 ha 2004 Nampula Brown streak in cassava 14,500 Unknown 2004 South Powdery mildew in Unknown Unknown cashews 2006 Gaza Locusts, lizards, wild Unknown 64,000 ha animal 2008 Zambezia Lizards Unknown 2,786 ha Source: The annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data. TABLE 3.4. SUMMARY OF WILDFIRE EVENTS Year Province Origination People Affected Hectares Lost 2004 Maputo Unknown Unknown 6,860 ha 2008 Manica Strong winds and high temperatures 89 3 million ha of farmland and forest Source: The annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data. TABLE 3.5. SUMMARY OF CYCLONE EVENTS Estimation of Type of Event/ People Total Economic Year Province Name Affected Damages (US$) Hectares Lost 1985 Maputo Cyclone Domoina Unknown 75,000,000 Serious damage to root crops 1994 Nampula, Ilha de Cyclone Nadya 2,502,000 700,000,000 600 tons of rice Moçambique 2000 North/Coastal areas Tropical Storm Delfina 1,000,000 Unknown 5,600 ha including 2,000 ha of beans 2003 Central Cyclone Japhet Unknown Unknown Unknown, included in flood losses 2006 Nampula Tropical Storm Unknown Unknown 5,000 ha 2008 Nampula, Zambezia Cyclone Jokwe +200,000 20,000,000 1,500,000 cashew trees. 68,522 ha 2012 Mozambique Cyclone Funso/ 5,000 Unknown 53,130 cashew trees, 157 mango trees, Channel Tropical Storm 252 coconut palms. 41,979 ha Source: The annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data. 16 Mozambique: Agricultural Sector Risk Assessment TABLE 3.6. PROJECTED CHANGES FOR 2046–65 IN AVERAGE TEMPERATURES DURING THE GROWING SEASON, CROP YIELDS UNDER RAIN-FED CONDITIONS, AND RAINFALL DURING THE CROP GROWING SEASON Changes in Temperature Changes in Yield Changes in Rainfall Median Change in Median Change in Median Change in Crop (past) Future (past) Future (past) Future °C °C % mm mm % mm mm % Cassava 23.8 2.0 8.5 0.397 −0.02 −4.2 633.7 −17.3 −2.7 Cotton 24.1 2.1 8.5 0.517 −0.02 −2.9 610.0 −20.0 −3.3 Groundnut 24.5 2.1 8.5 0.599 −0.03 −4.6 487.9 −5.1 −1.1 Maize 24.5 2.1 8.5 0.373 −0.04 −11.1 454.2 −5.8 −1.3 Sorghum 24.6 2.1 8.5 0.572 −0.02 −3.5 438.9 −3.9 −0.9 Soybeans 24.6 2.1 8.4 0.217 −0.03 −6.4 377.4 −4.5 −1.2 Source: Brito and Homan 2012. CLIMATE CHANGE average temperature or precipitation, may pose the Agriculture is highly vulnerable to climate change in greatest threat to agriculture in Mozambique. This Mozambique, and the effects are heterogeneous based includes flooding, drought, and tropical cyclones. on model assumptions and across regions, socioeconomic Risk management therefore becomes increasingly groups, and crops and livestock. There are direct impacts, important. such as changes in crop yields caused by precipitation » Mozambique is highly vulnerable to climate changes, and indirect impacts, such as rising food prices change because of its geography, in particular, its caused by production changes and land tenure conflict long coastline. stemming from shifting agro-climatic zones. If climate » Temperature projections vary in various models change is left unaddressed, then progress in agricultural and scenarios, but generally Mozambique expects development, food security, and poverty alleviation in to see a rise of 1°C–2.5°C by midcentury and an general may be reversed. increase of 1.4°C–4.6°C by late century.7 » Projections on precipitation vary from both posi- In the 2012 Responding to Climate Change in Mozambique: tive to negative changes, but increases in the pro- Theme 6: Agriculture, the INGC sought to quantify the portion of rain that falls during the rainy period effects of increased temperatures, changes in rain, and may occur. increased concentrations of carbon dioxide and ozone » Crop yields and land suitability: on six main crops (cotton, groundnuts, cassava, sorghum, – With some variations, generally there will be maize, soy) (table 3.6). The report ran seven general cir- no significant change in areas suitable for crops culation models to project temperature and rainfall data, (cassava, maize, soybeans, sorghum, groundnuts, and then used CliCrop to estimate yields based on soil and cotton). humidity daily diary. Data came from 47 meteorological – Likewise, the average change in yields in crops stations (Brito and Homan 2012). is projected to change in small increments, but generally will decrease slightly (cassava, sor- The principal findings from research carried out for this ghum, soybeans, sweet potatoes and yams, report, further details of which can be found in appendix maize, groundnuts, millet, potatoes). C, are the following: » The increase in the likelihood of extreme events caused by climate change, as opposed to changes in 7 IFPRI, UNDP. Risk Prioritization 17 FIGURE 3.4. MONTHLY RETAIL PRICES OF MAIZE (WHITE) IN KEY MARKETS IN US$/TON 600 Maxixe, Inhambane Nampula, Nampula 500 400 US$/ton 300 200 100 0 Jan-98 Aug-98 Mar-99 Oct-99 May-00 Dec-00 Jul-01 Feb-02 Sep-02 Apr-03 Nov-03 Jun-04 Jan-05 Aug-05 Mar-06 Oct-06 May-07 Dec-07 Jul-08 Feb-09 Sep-09 Apr-10 Nov-10 Jun-11 Jan-12 Aug-12 Source: GIEWS Food Price Data and Analysis Tool. FIGURE 3.5. MONTHLY RETAIL PRICES OF RICE IN KEY MARKETS IN US$/TON Maputo, Maputo City Maxixe, Inhambane 1400 Nampula, Nampula Ribaue, Nampula 1200 1000 US$/ton 800 600 400 200 0 Jan-00 Jul-00 Jan-01 Jul-01 Jan-02 Jul-02 Jan-03 Jul-03 Jan-04 Jul-04 Jan-05 Jul-05 Jan-06 Jul-06 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 Jan-10 Jul-10 Jan-11 Jul-11 Jan-12 Jul-12 Source: GIEWS Food Price Data and Analysis Tool. concern with severe implications for consumers and MARKET RISKS producers alike. The geography of Mozambique, cou- Price volatility (domestic as well as international), pled with limited transportation infrastructure, results exchange rate volatility, input volatility, and counterparty in three regions of Mozambique (North, South, and risk are some of the major market risks in Mozambique. Central) often behaving economically independently. They are, however, less significant than production risks, Although the northern region is more self-sufficient in and in many instances, crop failure because of production food, local events and regional conditions (in Malawi, risks triggers price spikes in local markets. The following Tanzania, and so on) contribute to price volatility. The section describes some of the major market risks. southern regions are more likely to experience food defi- cits and often rely on food imports from South Africa. These regions are more exposed to international price DOMESTIC PRICE VOLATILITY volatility passed down to the domestic markets, as well Seasonal price volatility is a feature of many agricultural as local production failures. Figures 3.4 and 3.5 show commodities in Mozambique and is driven by seasonal- the volatility of maize and rice in the main markets of ity of production, oversupply and deficit during particu- Mozambique. The graph in figure 3.4 depicts a few epi- lar months, and fluctuating demand for commodities. sodes of sudden spikes and a general trend of increas- This, however, is often considered a “normal” cost of ing prices, especially over the past five years. The price doing business in Mozambique, similar to many other spike of white maize in 2006 was driven by local events countries. Abrupt and steep price spikes and falls, often (largely drought in 2005), whereas the 2008–09 price driven by underlying production deficits, market fac- spike was a pass-through effect of the global food price tors, or other exogenous factors, are a cause of serious crisis. Because of heavy reliance on imports for rice, this 18 Mozambique: Agricultural Sector Risk Assessment FIGURE 3.6. INTERNATIONAL MONTHLY PRICES FOR SUGAR Sugar, European import price, CIF Europe Sugar, free market, coffee sugar and cocoa exchange (CSCE) contract no.11 nearest future position 45 Sugar, U.S. import price, contract no.14 (revised to no. 16) nearest futures position 40 35 30 US ¢/lb 25 20 15 10 5 0 1991M01 1991M10 1992M07 1993M04 1994M01 1994M10 1995M07 1996M04 1997M01 1997M10 1998M07 1999M04 2000M01 2000M10 2001M07 2002M04 2003M01 2003M10 2004M07 2005M04 2006M01 2006M10 2007M07 2008M04 2009M01 2009M10 2010M07 2011M04 2012M01 2012M10 Source: IMF Primary Commodity Prices. FIGURE 3.7. INTERNATIONAL MONTHLY PRICES FOR COTTON “A” INDEX 100 90 80 70 Cents/Pound 60 50 40 30 20 10 0 Aug-95 Mar-96 Oct-96 May-97 Dec-97 Jul-98 Feb-99 Sep-99 Apr-00 Nov-00 Jun-01 Jan-02 Aug-02 Mar-03 Oct-03 May-04 Dec-04 Jul-05 Feb-06 Sep-06 Apr-07 Nov-07 Jun-08 Jan-09 Aug-09 Mar-10 Source: National Cotton Council of America. commodity leads to direct transmission of international on food crops means that the risks extend far beyond those price volatility to domestic markets. that just have an effect on farmers and producers. INTERNATIONAL PRICE VOLATILITY EXCHANGE RATE RISK International markets can also be volatile and will often Given Mozambique’s geographic location close to Sub- have a knock-on effect for domestic prices, as figures 3.6 Saharan Africa’s biggest market (South Africa), along and 3.7 demonstrate. Mozambique relies on imports to with its heavy reliance on trade with this neighbor, supplement much of its domestic consumption needs movement in the rand/metical exchange rate can be a each year, meaning the country inherits some interna- potential source of risk. Figure 3.8 shows the historical tional market volatility. This was in evidence in 2008 trends of the metical against the rand, indicating signifi- when international food price crises caused spikes in inter- cant troughs and peaks. As figure 3.9 indicates, although national prices that were also seen in the domestic market, the metical has been steadily appreciating against the leading to political unrest in Mozambique. Transmission rand, there has not been marked movement in the past of this volatility to domestic producers is significant; this 10 years between the metical and the euro/U.S. dollar. can be seen by the mirroring of peaks in domestic versus These peaks and troughs, driven by macroeconomic fac- international prices and is also reflected in export prices. tors, monetary policies, and currency market events, can Apart from sugar, the impact that this price volatility has have a marked effect on the prices that producers receive Risk Prioritization 19 FIGURE 3.8. EXCHANGE RATE RISKS—METICAL/RAND 0.6 0.5 MZN/ZAR 0.4 0.3 0.2 0.1 0 3/1/2000 9/1/2000 3/1/2001 9/1/2001 3/1/2002 9/1/2002 3/1/2003 9/1/2003 3/1/2004 9/1/2004 3/1/2005 9/1/2005 3/1/2006 9/1/2006 3/1/2007 9/1/2007 3/1/2008 9/1/2008 3/1/2009 9/1/2009 3/1/2010 9/1/2010 3/1/2011 9/1/2011 3/1/2012 9/1/2012 Source: Oanda. FIGURE 3.9. EXCHANGE RATE RISKS—METICAL/EURO AND METICAL/US$ MZN/EUR MZN/USD 0.08 MZN/Currency 0.06 0.04 0.02 0 3/1/2000 9/1/2000 3/1/2001 9/1/2001 3/1/2002 9/1/2002 3/1/2003 9/1/2003 3/1/2004 9/1/2004 3/1/2005 9/1/2005 3/1/2006 9/1/2006 3/1/2007 9/1/2007 3/1/2008 9/1/2008 3/1/2009 9/1/2009 3/1/2010 9/1/2010 3/1/2011 9/1/2011 3/1/2012 9/1/2012 Source: Oanda. if they are engaged in regional trade (South Africa is the Considering that a very small percentage of the Mozam- biggest regional market partner for Mozambique) and the bique agricultural sector is reliant on input markets (seed, prices that consumers have to pay at market for imported fertilizer, diesel, and so on), the volatility in input prices commodities. The peaks and troughs shown in figure 3.8 does not have significant impact on the broader agricul- will have a similar effect to those against the rand because tural sector. Europe is a substantial export market for Mozambique and various commodities are imported in U.S. dollars. COUNTERPARTY RISK Going forward, Dutch disease, driven by natural gas and Counterparty risk/default risk is particularly important the energy economy (and the corresponding threat of for stakeholders further up the agricultural value chain underinvestment in agriculture and significant overvalu- involved, for instance, in off-taker agreements or with ation of the metical), could have disastrous impacts on the out-grower arrangements, particularly in the cotton, sug- agricultural sector. arcane, and tobacco supply chains. Various contractual arrangements exist, including provision of inputs in the INPUT PRICE RISK form of fertilizer or seed to be paid back based on an In 2008, only 10 percent of maize farmers and 4 percent agreed price for a commodity or signed contract based of rice farmers used improved seeds (Government of on quantity and a fixed price. Although Mozambique has Mozambique 2008), and most of these were engaged in relatively sound legislation for the enforcement of con- commercial production. The cases of fertilizer and die- tracts, various economic and sociocultural issues sharply sel are similar to that for seeds, and input application is limit the use of formal, written contracts in the agriculture currently the domain of commercial farmers. Although sector. Although this is a perennial problem in some com- the input prices are not enormously volatile, they are con- modities, the impact on the broader agricultural sector is sistently high, which reduces their widespread adoption. not particularly significant. 20 Mozambique: Agricultural Sector Risk Assessment stable in Mozambique, for better or worse, and have not ENABLING been registered by stakeholders as a major risk. ENVIRONMENT RISK Enabling environment risk covers many different aspects LINKAGES BETWEEN RISKS of legal, institutional, fiscal, and policy issues that influ- Although individual risks and their occurrence as ence the ability of stakeholders to undertake their busi- described above are significant, it is important to under- ness within a sector. Domestically, this includes factors stand the dynamics of the linkages among and between such as competitiveness (exit/entry conditions, incen- these risks. This is particularly pertinent in Mozambique tives/subsides for production units by size, tax consid- because of the differing agro-climatic zones9 and the con- erations); regulation and enforcement (for example, tax current emergence of several different types of risk.10 code, property rights, resource management); and trade policy barriers. On an international level, several factors should be taken into consideration in assessing risks to INDEPENDENT the Mozambican agricultural sector: international trade RISKS OCCURRING regulations and treaties; other international protocols, policies, regulations of nationals, and trading blocs; and SIMULTANEOUSLY the extent to which international trade agreements and Although not presented in detailed timeline form in this conventions affect commodity performance (Jaffee, Siegel, chapter, it is clear that Mozambique appears to suffer from and Andrews 2008). some type of risk every year and sometimes several in the same year, leading at times to significant losses. Flood and drought are often considered to be risks independent of CONFLICTS AND INSECURITY each other however there have been years where both Rising pressure on land because of insecure land tenure risks were experienced in the country, such as 1999, 2007, rights,8 land grabs, and other issues are not yet major and 2010. sources of risk in Mozambique. As all land is state owned this can create a certain amount of protection for commu- nities, although frequent disputes between communities, DEPENDENT RISKS government, and investors have occurred. Drought, one of the most significant risks in Mozam- bique, is a clear example of one risk that can trigger oth- POLITICAL INSTABILITY ers. These could include price volatility caused by supply AND REGULATORY ISSUES deficits and import or export restrictions, which can in After the peace accords were signed in October 1992 to turn also bring about domestic price volatility and other end civil war between The Front for the Liberation of risks such as exchange rate risk, and so on. Similarly, Mozambique (FRELIMO) and Mozambican National pest and disease outbreaks, such as incidents of fruit flies Resistance (RENAMO), conflict has not been a serious in 2008 and 2010,11 could lead to export restrictions in threat to the country (the recent disturbances in Sofala regional and international markets. Flood risk is often District notwithstanding). FRELIMO has remained in accompanied by alluvion,12 in which sediment is depos- power since 1992, with elections held in 1994, 1999, ited when rivers break their banks, and also the tempo- 2004, and 2009. Although the elections have been more contested in recent years, they did not generally lead to frequent changes in the policy or institutional environ- 9 Mozambique lies between 10˚S and 27˚S. Chapter 2 gives further information on the agricultural system of Mozambique. ment as FRELIMO remained in control of the govern- 10 In 2007, for example, Mozambique suffered from drought, flood, and cyclone, ment. Policies and regulations have also been relatively resulting in an estimated loss of 563,885 ha combined. 11 Fruit imports from Mozambique to South Africa and Zimbabwe were banned in September 2008 and February 2010, respectively. 8 Only about 10 percent of communities have their rights registered according 12 In 1999, Inhambane registered losses of 22,018 ha directly linked to alluvion to the IS Academy on Land Governance—(LANDac 2012). according to the INGC. Risk Prioritization 21 rary movement and relocation of communities when the south. There is also a strong coastal-to-inland oro- flooding is especially bad. A specific example of risk graphic, or elevation gradient, effect on weather pat- dependence occurred in 2008 during the international terns in Mozambique. Weather patterns change as they food price crises, with international price volatility, inse- move west from the southeastern, low-elevation, coastal curity, and changes in government policy seen in Mozam- belt into the central and north-central plateau regions of bique. Riots took place in urban centers as food price the country. Some areas of Mozambique such as coastal increases coincided with a reduction in fuel and food sub- zones do not conform to the “administrative division of sidies. These interdependencies, as with the independent Mozambique” but characterize specific geographic areas risks that occur simultaneously, demand special attention vulnerable to droughts, floods, and storm surges directly when addressing the combined issue rather than the sep- and indirectly related to sea level rise. Other parts of dif- arate symptoms. ferent provinces are influenced by basins of major riv- ers flowing to the Indian Ocean. Table 3.7 highlights the major crop production systems in the country and princi- REGIONAL SHOCKS pal hazards they face. It is important to consider regional systemic shocks, par- ticularly as Mozambique shares borders with six coun- However, despite this type of classification, a picture of tries. The Zambezi River System, for example, crosses all the vulnerability of the 10 different provinces can be the way from Angola through Zambia, Malawi, Tanzania, inferred from the number of occurrences registered in and Zimbabwe before ending in Mozambique and empt- the last two decades as highlighted in table 3.8. Prov- ing into the Indian Ocean. Regional events can affect the inces such as Gaza and Sofala are equally exposed to ability of people to cope with risks as they occur. Refugees drought and flood risks. Flood is a more serious con- coming across the borders have been known to induce the cern in Zambezia, whereas drought is a bigger issue in risks associated with conflicts and insecurity and present a Inhambane. risk that is hard to mitigate or cope with. COMMODITY RISK PROFILES DIFFERENTIAL IMPACT The adverse impact of risk varies greatly by agricultural OF RISK commodity, as well as by province/region. Risk profiles of cash crops, food crops, and the livestock sector can Location and timing of risks can also have an important be quite different. Table 3.9 summarizes principal risks impact on the sector. For instance, where flooding or for major commodities analyzed in Mozambique, high- drought or wildfires take place, the type of land that is lighting the risk of droughts, floods, pest and disease out- affected will determine the extent of the effects. The 2008 breaks as well as price volatility. Detailed commodity risk wildfire that hit Manica, and is reported to have led to profiles, providing more information about individual losses of 3 million hectares, is generally agreed to have commodities, their risks, and impacts can be found in affected mainly uncultivated forest land. More precise dis- appendix E. tinctions would help to identify who was affected as well as the appropriate measures based on these stakeholders for management of the risk. CLIMATE CHANGE Considering its importance in inducing structural shifts PROVINCIAL RISK PROFILE and altering the risk profile of the agricultural sector, it The impact of agricultural risk varies greatly by region is important to examine climate change separately, even depending on agro-climatic conditions, agricultural though it is a trend, and not a risk per se. Agriculture in system and composition, and regional institutional Mozambique is highly exposed and vulnerable to climate arrangement. In the northern and central regions, the change, and the effects are heterogeneous based on model climate can be classified as tropical and subtropical; in assumptions and across regions, socioeconomic groups, contrast, steppe and dry arid desert conditions exist in and crops/livestock. There is, however, consensus across 22 Mozambique: Agricultural Sector Risk Assessment TABLE 3.7. MAJOR CROP SYSTEMS AND THEIR PRINCIPAL HAZARDS IN MOZAMBIQUE Location Principal Hazards Coastal urban areas (most This zone is marked by highly differential vulnerability across income groups, with large peri-urban important, Maputo and Beira) areas vulnerable to flooding from both rivers and the ocean. Nonurban coastal strip This zone is marked by high vulnerability to coastal flooding and storm surges from tropical cyclones as well as threats of erosion. It is relatively food secure, with low rates of poverty; it encompasses the coastal provinces of Maputo, Gaza, Inhambane, Sofala, Zambezia, Nampula, and Cabo Delgado but the Central and Northern Provinces are more affected. Limpopo river valley districts This zone is unique in being highly exposed to two very different threats: river flooding and upstream of Xai-Xai drought. It has relatively high population density, and thus high numbers of poor people. Gaza and Southern Inhambane are under its influence. Other flood-prone river valleys These zones, in particular in the Buzi and Zambezi river valleys, are highly susceptible to floods (especially those caused by tropical cyclones), but less so to droughts. Sofala and Low Zambezia are within this zone. Drought-prone inland areas These areas are highly susceptible to drought: adequate rainfall to support agriculture is an (especially in the South exception rather than the rule. Inhabitants of this region are often dependent on remittances Inhambane) for survival. Population densities are low. Inland areas of higher These areas are perhaps the least vulnerable in Mozambique, facing adequate rainfall agricultural productivity most years, and no extreme risks from flooding or tropical cyclones. They are somewhat (including the highly productive heterogeneous in terms of poverty rates and food security. The highly productive regions stand and populated areas in Zambezia) out for their high population density and relatively low vulnerability. Source: EACC publications; World Bank Group. TABLE 3.8. PROVINCES AND PRINCIPAL HAZARDS Events Maputo Gaza Inhambane Sofala Manica Tete Zambezia Nampula C. Delgado Niassa Drought XX XXX XXX XXX XX XXX X XX X X Floods X XXX XX XXX XX XX XXX XX X X Storms XXX XXX XXX XXX Flash X XX XX XX X XXX XXX X XX X floods Cyclones X X XX XXX X Rains XXX XXX XXX XXX Alluvions XXX XX X X X Conflict XX XXX XX X Epidemics X X X XXX XX XX XX X X Fire XX XXX XX XX Source: Compiled based on data from Desinventar-Disaster Information System—UNISDR. Note: Level of impact: XXX = severe, XX = considerable, X = moderate. different models that the increase in the likelihood of tions, models generally indicate there will be no significant extreme events (flooding, drought, and tropical cyclone) change in areas suitable for crops (cassava, maize, soy- caused by climate change, as opposed to changes in aver- beans, sorghum, groundnuts, and cotton). Likewise, the age temperature or precipitation, may pose the greatest average change in yields in crops is projected to change threat to agriculture in Mozambique. With some varia- in small increments, but generally will decrease slightly Risk Prioritization 23 TABLE 3.9. PRINCIPAL RISKS BY COMMODITY Commodity Principal Risk 1 Principal Risk 2 Principal Risk 3 Maize Drought Domestic price volatility Pests and diseases Rice Drought Flood International price volatility Cotton Drought International price volatility Pest and disease outbreak Groundnuts Drought Pests and diseases Price volatility Sugarcane Flood Drought Pests and diseases Poultry Pests and diseases Input price volatility Poultry price volatility Sorghum Pests and diseases Drought Domestic price volatility Cassava Pests and diseases Flood Price volatility Cashews Pests and diseases Cyclone Drought Vegetables Pests and diseases Flood Price volatility Tobacco Drought Flood Pests and diseases (cassava, sorghum, soybeans, sweet potatoes and yams, culture sector in Mozambique. In Mozambique, negative maize, groundnuts, millet, potatoes). Projections on pre- impacts on agriculture from climate change will primarily cipitation vary from both positive to negative changes, but be from the increased likelihood of extreme events such as increases in the proportion of rain that falls during the flooding and droughts. However, outside of the expected rainy period may occur. These models, showing increased increase in extreme events, agriculture in Mozambique variability, would indicate that it will be ever more difficult will see little change in land suitability and yield. Thus risk for farmers to predict appropriate times for planting and management becomes increasingly important as uncer- harvesting. Appendix C provides an exhaustive literature tainty, frequency, and severity of risk events will increase review of potential climate change impacts on the agri- as a result of climate change. 24 Mozambique: Agricultural Sector Risk Assessment CHAPTER FOUR QUANTIFICATION OF LOSSES AND IMPACT OF AGRICULTURAL RISKS This chapter outlines the conceptual and methodological basis used for analysis and seeks to quantify the impacts of production risk events. The various sources of risk are then prioritized on the basis of indicative losses and consideration given as to how to manage these risks in chapter 5. CONCEPTUAL AND METHODOLOGICAL BASIS FOR ANALYSIS For the purposes of this study, risk is defined as an exposure to a significant financial loss or other adverse outcome whose occurrence and severity is unpre- dictable but for which some probability of occurrence can be estimated on the basis of historical experience. Accordingly, risk implies exposure to substantial losses, over and above the normal costs of doing business. In agriculture, farmers incur small losses each year caused by moderately adverse climatic conditions and fluctuations in output or input prices. Risks discussed here refer to more severe and unpredictable events. INDICATIVE QUANTIFICATION OF LOSSES Available data on actual losses caused by adverse events in Mozambique are not par- ticularly accurate or consistent within individual data sources. In an attempt to facilitate comparison and ranking of the costs and losses caused by various events, different data sources were combined to generate a more or less consistent time series. Appendix A describes the methodology for quantifying risk. The figures used (refer to tables 3.1–3.6) Risk Prioritization 25 are from actual data sources13 and attempts have been made FIGURE 4.1. ESTIMATED AGGREGATE TOTAL to prevent double-counting.14 LOSSES BY RISK EVENT Bubble volume = estimated maximum loss from one event Estimated total losses per event EXPECTED LOSSES AND $600,000,000 (constant 2006–2007 I$) $500,000,000 Drought/Intense heat RISK PRIORITIES FOR $400,000,000 Flood/Irregular rainfall $300,000,000 PRODUCTION $200,000,000 Pest and disease $100,000,000 Figure 4.1 shows estimated aggregate total losses by risk Wildfire Cyclone $- event for 1996–2015Q1 clearly indicating that drought 0.0 0.1 0.2 0.3 0.4 0.5 Frequency of event 1996–2015Q1 generates the most total losses but occurs slightly less than Source: FAOSTAT (April 2015), the annual Plano Economico e Social (Eco- floods. It is also indicated that it was a drought that is esti- nomic and Social Plan), the annual Balanço de Plano Economico e Social (Bal- mated to have led to the single largest loss from one event ance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, (US$182 million—drought 2009/10). SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozam- bique data, and Author’s calculations. Table 4.1 presents the estimated losses in numbers per event over the time period 1996–2015Q1 that total two events (1999, 2002, 2003, 2006, 2009, 2010, and US$900 million from 3 million hectares lost. According 2012). In 2000, widely accepted as the most memora- to our calculations, 2000, 2007, and 2010 (see table 4.2) ble year as far as losses are concerned among farmers, all had losses of over US$85 million in constant 2004– it is estimated that 363,600 hectares were lost at an esti- 06 U.S. dollars. Table 4.1 also clearly indicates that in mated cost of US$85.9 million. Of this, calculations regard to frequency, floods and droughts happen regu- suggest that 198,000 hectares or US$46.8 million was larly whereas catastrophic pests and disease outbreaks caused by flood, 160,000 hectares or US$37.8 million and cyclones (that incur severe damage) happen far less caused by locusts, and 5,600 hectares or US$1.3 million frequently. The incidence of wildfire was extremely low. caused by cyclone. The extensive losses estimated in 2007 However, the frequency of pest and disease outbreaks, in of 563,885 hectares demonstrate the impact of having general, is much higher but their aggregate impacts are flood and drought events in the same year. It was esti- either localized or much smaller, thus not leading to severe mated that 203,263 hectares or US$50.5 million was lost losses at the national level. caused by drought, 288,000 hectares or US$71.5 million because of flood and 72,622 hectares or US$18 million Table 4.2 presents the same data broken down by year because of cyclone. The same is seen in 2010 with a total rather than risk event and shows that between 1996 and of 607,950 hectares or US$183.3 million (605,000 hec- 2015Q1 there was an event that led to recorded losses tares of US$182.4 million caused by drought and 2,950 every year except 2014, with some years experiencing as hectares or US$0.9 million caused by cyclone). many as three events (2000 and 2007) and several with Unfortunately, sufficient data did not exist to be able to 13 Data sources used for the calculations include: FAOSTAT, the annual Plano undertake sensible quantification on livestock losses, par- Economico e Social (Economic and Social Plan), the annual Balanço de Plano Eco- ticularly for poultry. Given the importance of livestock nomico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID in Mozambique’s GDP,15 as well as to food security more monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC— generally, it would have been useful to attempt to quantify Government of Mozambique data. the losses. Responses from farmer groups as well as a desk- 14 Estimated losses were calculated as follows: the share of production from based study strongly suggested that livestock diseases are a crops/agriculture was assumed to the 60 percent of net production value in source of significant losses. Vaccines exist and producers constant 2004–06 US$, which was taken from FAOSTAT. This was divided by the amount of land in hectares planted for arable and permanent crops, also from FAOSTAT, to formulate an estimated cost per hectare. This estimated cost per hectare was then applied to the number of hectares lost per event per year that was derived from the various sources previously mentioned. Additional 15 According to IFPRI 2007 calculations, poultry was a 4.7 percent share of details on methodology are found in appendix A. agricultural GDP. 26 Mozambique: Agricultural Sector Risk Assessment TABLE 4.1. ESTIMATED AGGREGATE TOTAL LOSSES BY RISK EVENT, 1996–2015 Estimated Costs of Estimated Maximum Losses (constant Number of Events, Loss per Event Risk Event Hectares Lost 2004–06 US$) 1996–2015Q1 (constant 2004–06 US$) Drought/ 2 million 475.5 million 8 182 million intense heat Flood 900,000 300 million 13 72 million Pest and disease 300,000 70 million 4 38 million Cyclone 100,000 39 million 4 18 million Wildfire 7,000 2 million 1 2 million TOTAL 3 million 900 million 30 Source: FAOSTAT, the annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data, and Authors’ calculations. TABLE 4.2. ESTIMATED AGGREGATE TOTAL LOSSES BY YEAR, 1996–2015 Estimated Cost of Hectares Losses (constant Number of Year Lost 2004–06 US$) Events 1996 170,000 $39.3 million 1 1997 103,000 $24.5 million 1 1998 60,600 $15.0 million 1 1999 73,000 $18.4 million 2 2000 363,600 $85.9 million 3 2001 116,208 $28.8 million 1 2002 227,119 $53.6 million 2 2003 258,760 $63.4 million 2 2004 6,860 $1.7 million 1 2005 317,200 $70.1 million 1 2006 70,754 $17.3 million 2 2007 563,885 $140.0 million 3 2008 2,786 $0.8 million 1 2009 185,000 $47.7 million 2 2010 607,950 $183.3 million 2 2011 13,974 $4.5 million 1 2012 46,877 $15.2 million 2 2013 153,000 $50.0 million 1 2014 − − − 2015Q1 87,000 $28.2 million 1 TOTAL 3,427,573 $789.8 million 30 Source: FAOSTAT, the annual Plano Economico e Social (Economic and Social Plan), the annual Balanço de Plano Economico e Social (Balance of the Economic and Social Plan), FEWSNET/USAID monthly bulletins, SETSAN (Food Security) Reports for the Government of Mozambique and the Disaster Information System, UISDR from INGC—Government of Mozambique data, and Authors’ calculations. Risk Prioritization 27 BOX 4.1. CROP LOSSES ACCORDING TO THE AGRICULTURAL CENSUS The Mozambique Agricultural Census 2002–07 asked small and medium producers (95 percent of all production in Mozam- bique) about the principal cause of crop losses. Figure B4.1.1 takes data from the 2002, 2003, 2005, 2006, and 2007 surveys. It can be clearly observed that lack of rain was the largest cause of loss: over 50 percent for maize, rice, and others. Losses attributable to pests and wild animals were lower, although still significant at over 10 percent, whereas floods, excess rain, and disease/rot (in other crops) came in at about 5 percent. FIGURE B4.1.1. PRINCIPAL CAUSE OF CROP LOSSES 70 60 50 Crop loss (%) 40 30 20 10 0 Diseases, Domestic Excess Lack of Wild Fires Floods Others Pests rot animals rain rain animals Maize 2.6 2.2 6.8 0.3 3.6 59.5 3.7 12.7 12.4 Rice 1.6 0.8 6.9 0.2 7.1 57.0 3.1 16.1 11.0 Others 6.1 2.2 5.3 0.4 2.1 55.5 2.8 18.6 12.2 FIGURE 4.2. ANNUAL GDP GROWTH AND GDP PER CAPITA 20 GDP growth (annual %) GDP per capita growth (annual %) 15 10 5 Percent 0 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 –5 –10 –15 –20 Source: World Development Indicators Database 2012. are generally aware of their benefits, but they are not widely Mozambique Agricultural Censuses. Although exact used because of their costs and limited availability. figures do not exist, this information gives a good indi- cation of the major risks as far as they are assessed in Box 4.1 illustrates the percentage of crop losses that can the minds of small and medium producers (95 percent of be attributed to different types of risk according to the those involved in agricultural production in Mozambique). 28 Mozambique: Agricultural Sector Risk Assessment substantial resources. Figure 4.2 indicates the volatility of IMPACT OF AGRICULTURAL GDP growth and GDP per capita growth. The results RISKS ON NATIONAL GDP suggest a close relationship between drops in GDP In addition to having an effect on agricultural production, growth rates and risk events. Prior to 1992, the volatility agricultural risks affect foreign exchange earnings, is probably primarily the result of the civil war and politi- the GDP growth rate, per capita income, and govern- cal unrest, but these have not been important factors in ment revenues. Responding to these impacts requires the past two decades. Risk Prioritization 29 CHAPTER FIVE STAKEHOLDER RISK ASSESSMENTS How different stakeholder groups perceive and manage production risks varies sig- nificantly depending on the commodity as well as on the priorities and capacities of individual stakeholders. The importance and character of market risks for individual stakeholder groups is very much a function of the nature and extent of interaction with markets. A minority of producers use the market as a source of inputs and an out- let for their production to any significant extent. Almost the entire population depends on markets to some degree as a source of food. Enabling environment risks also affects everyone, but again by varying degrees at different points in time. This chapter discusses the impacts of the range of risks described in chapter 3 on the major stakeholder groups, namely, government, producers, consumers, and the most vulnerable. Producers are a heterogeneous group that can be differentiated in several ways, including size of operations and relationships to markets. The vulnerable can be considered as components of producer and consumer groups, but are treated sepa- rately here in recognition of the severity of the impacts of risk-related events on this segment of the population. This discussion disaggregates the impacts of the risks presented in chapters 3 and 4 by stakeholder group indicating which types of risks are most significant to each group as well as how the group is affected. Although there are insufficient data for formal quantification, the results provide additional guidance for the prioritization of risks and their management, which is the focus of chapter 6. In addition to severity and frequency of risk-related events, their differential impacts on stakeholder groups are considered in the prioritization and management response planning processes. GOVERNMENT The impact of agricultural risks is far reaching for governments, particularly when the econ- omy and livelihoods are agriculturally dependent, and the government of Mozambique is no exception. Not only does it have to provide financing for immediate relief and coping, it also incurs loss of revenue from the agricultural sector, reduced GDP, and financial losses across the economy. The government has to confront the devastating effects on people’s Risk Prioritization 31 FIGURE 5.1. ADVERSE IMPACT OF AGRICULTURAL RISKS ON INFLATION Food and non-alcoholic beverage inflation Total inflation 35 Flood in north and 2010 droughts in south (2000/01) 20% depreciation of 30 2005/06 drought 2008 global currency, spike in and floods food and food (imported) 25 financial crisis price 20 Percent 15 10 5 0 –5 Jan-00 Apr-00 Aug-00 Dec-00 Jan-01 Apr-01 Aug-01 Dec-01 Jan-02 Apr-02 Aug-02 Dec-02 Jan-03 Apr-03 Aug-03 Dec-03 Jan-04 Apr-04 Aug-04 Dec-04 Jan-05 Apr-05 Aug-05 Dec-05 Jan-06 Apr-06 Aug-06 Dec-06 Jan-07 Apr-07 Aug-07 Dec-07 Jan-08 Apr-08 Aug-08 Dec-08 Jan-09 Apr-09 Aug-09 Dec-09 Jan-10 Apr-10 Aug-10 Dec-10 Jan-11 Apr-11 Aug-11 Dec-11 livelihoods, in particular income loss (crop failure) and asset by international donors. Emergency financing put in place loss (livestock and infrastructure damages, for example, dur- because of drought, flood, and cyclone leads to diversion ing flooding). Because producers can lose a substantial part of funding previously allocated for development activities. of their crop production and much of the livestock in the The ad hoc crisis response mode, rather than risk manage- areas where the disaster is most severe, the government must ment, diverts attention from development priorities of the cope with the consequences, launching programs to relieve country, with significant opportunity costs. The frequent instances of food insecurity and malnutrition. occurrence of agricultural risks leads to sudden spikes in inflation (figure 5.1), contributing to macroeconomic In the cases of floods and cyclones, the government has instability with adverse consequences for consumers. to bear the costs of resettling families in less flood-prone areas. There are also social consequences of moving pop- ulations a considerable distance from their farms, as natu- PRODUCERS rally some are reluctant to leave their traditional zones Producers are directly and immediately affected by occur- of living and can be traumatized by the need, in certain rence of market, production, and enabling environment cases, to forge new livelihoods. risks. However, their exposure to risk and their capacity to manage are dependent on their production systems, size One of the immediate effects of crop failure or price vola- of operations, commodity spread, and so on, which even- tility is the drastic fall in volume and/or value of com- tually determine their losses. modities trading, which reduces the government’s capacity to collect taxes. It also reduces foreign exchange earnings because of the fall in production or value of export crops SMALLHOLDERS such as cotton, cashews, tobacco, tea, sugar, and other tra- The agricultural economy is still heavily subsistence ori- ditionally exported products. ented; less than 10 percent of households sell their sur- pluses of maize, cassava, and so on. By being largely The government annually provides US$3.5 million– subsistence oriented, the majority of smallholders reduce US$5 million to INGC for disaster risk management and their exposure to market-induced risks (World Bank response, which may increase depending on the magni- 2006). Smaller farms commonly have diversified sets of tude of the disaster. The contingency plan is also funded farm and nonfarm enterprises, partially in an effort to 32 Mozambique: Agricultural Sector Risk Assessment ensure they have some production and income for their some food crops such as beans, maize, rice, and cassava. In survival when drought or other risks strike and result in the lowlands of Sofala, Zambezia, and Gaza, huge losses crop or income loss. Nationally, 42.3 percent of family have been reported in rice production relating to floods. sector farmers experienced shortages of food at some time during the year (CAP 2009/10, INE). According to CAP Despite the difficulties in overcoming the effects of the 2009/10, irregularity of rainfall (51.3 percent) affecting major hazards such as floods and droughts, their access 822.713 small farm units is the major reason for food defi- to production means and financing can be substantially cit during the season 2009/10. easier than the smaller farms. The difference is that they can “bet” a little more on pure cash crops and have the Although the small farms have diversified their crop means to conduct their business more professionally, using composition and livelihood profile, they have only lim- the right technology with access to funds, and so on. This ited access to resources (information, technology, finance, also increases their potential losses because their invest- infrastructure, and so on) and low capacity to manage. ment is higher than that of the smallholder farmers. Their lack of use of inputs will also exacerbate losses experienced from pests and disease. Crop failures force LARGE FARMS small farms to become even more dependent on the mar- Large farmers tend to specialize, which gives them greater ket for their basic needs at times when prices are likely to exposure to risks that may be specific to a particular commod- be high and availability low. ity. They are exposed to drought, flood, and pest and disease outbreaks as well as international price volatility, given many COMMERCIAL FARMERS are involved in production of export commodities. Some are Commercial farmers are those who regularly produce a also engaged in contract farming operation with small and surplus of one or more commodities that they sell locally medium farms, increasing their exposure to counterparty in markets or to companies under contract. They are not risk. Considering large farms have significant infrastructure necessarily self-sufficient in basic staple foods, but rather (irrigation, machinery, and so on), they are exposed to severe may opt to produce cash crops and buy at least a portion of and long-term asset loss, especially from flooding. their staple food needs from the market. They have a plot of five or more hectares with limited access to credit and Although exposure is high, large farms have relatively bet- can employ people for planting, weeding, and harvesting. ter capacity to manage because of their access to informa- tion (weather, prices, and so on), finance, infrastructure, These farmers do not differ very much from the vast and technology, which help them in better decision mak- majority of smallholders whose land on average is 1.5 ha. ing to reduce losses. Although the local availability of As such, their vulnerability to the major natural hazards information, technology, and other instruments for man- is similar to the vast majority of small farms. In many aging risks is relatively limited, some of the larger firms cases, these natural hazards may have more dramatic are able to access these from South Africa. consequences on the welfare of the family because they have considerable costs of production (mechanized land preparation, significant lab costs, planting and harvesting, VULNERABLE GROUPS and so on) and potentially have loans that must be repaid. Agricultural risk is one of the biggest poverty traps, other Their size is reported to be less than 1 percent of the total than health risk, in Mozambique, leading to adverse farmers involved in agriculture. impact on two of the biggest vulnerable groups: the rural poor and the urban poor. The bulk of the vulnerable pop- MEDIUM-SIZE FARMS ulation in rural areas is composed of small-scale farmers Medium-size farms follow more or less the same produc- and female-headed households, with both groups signifi- tion pattern as do small farms. They cultivate larger plots, cantly affected by occurrence of agricultural risks. Fre- with their chosen crops based on cash crops and include quent crop failures and price spikes create food availability Risk Prioritization 33 FIGURE 5.2. ACUTE FOOD INSECURITY IN MOZAMBIQUE 900,000 Severe Severe Severe drought Multiple risk events Flooding Drought 55% Drought 800,000 drought drought Flood 25% Cyclone 19% 700,000 600,000 Population 500,000 400,000 300,000 200,000 100,000 0 0 1 2 3 4 5 6 7 8 9 0 1 2 00 /0 /0 /0 /0 /0 /0 /0 /0 /0 /1 /1 /1 00 01 02 03 04 05 06 07 08 09 10 11 /2 99 20 20 20 20 20 20 20 20 20 20 20 20 19 Source: SETSAN reports. and affordability problems for poor rural producers and major occurrence of agricultural risk has resulted in a sud- poor urban consumers. A social safety net assessment den increase in food insecurity in Mozambique. Not only of Mozambique produced by the World Bank (2012b) does agricultural risk aggravate the situation for acute showed that, in the previous five years, households were food insecure populations, it also is the principal cause of exposed to a series of covariate and idiosyncratic shocks transient food insecurity. such as increases in food prices (reported by 36.5 percent of households) followed by drought (19.8 percent), the The capacity to manage risk is weakest among the vulner- death of a household member (10.8 percent), or an illness able population in Mozambiaque and this increases their in the household (8.1 percent). vulnerability to a range of agricultural risks. Appendix D provides a detailed vulnerability analysis in Mozambique, Severe drought in 2004/05 resulted in a fourfold increase with desciptions of vulnerable groups, recent trends in in acute food insecurity, with the population at risk growing vulnerability, underlying factors to food security, and from 200,000 to 800,000. Figure 5.2 highlights that every major shocks to food security in Mozambique. 34 Mozambique: Agricultural Sector Risk Assessment CHAPTER SIX RISK PRIORITIZATION AND MANAGEMENT Risks for the agricultural sector differ significantly from one another not only in their character, but also in the severity, frequency, and distribution of their impacts on dif- ferent areas, commodities, and sections of the Mozambican population. Chapter 3 describes the major risks affecting the agricultural sector. Production-related risks, notably drought and flood, are prominent, but there are also a range of risks associ- ated with markets and the enabling environment. Chapters 4 and 5 attempt to quan- tify the losses associated with these major risks and assess their impacts on government and consumers as well as different categories of agricultural producers and service providers. Special attention has been given to the impacts on the most vulnerable groups in this process. The information and analysis in the preceding chapters enable a prioritization of risks according to the magnitude, frequency, and distribution of the impacts of risk-related events. However, those results must be matched with measures to more effectively manage risks to be of use to those affected: individuals, communities, private enter- prises, and public agencies. There is already an array of programs and activities that government, businesses, and individual farmers use to manage risks. Consumers man- age higher prices and limited availability for some commodities by substituting with others. How effective are these activities and how adequate is their coverage? What are the options for better managing these risks and which are most feasible to implement in Mozambique? This chapter summarizes the approaches and results of the risk prioritization (see “Risk Prioritization”) and risk management (see “Risk Management”) that are the focus of this initial phase of the agricultural sector risk assessment study. These find- ings in turn provide the background information for an examination of ways in which Mozambique can more effectively manage risks. It is important to note that difficulties in accessing good data and information to assess risks were encountered, which may have affected the prioritization undertaken. Interventions to better manage the risks prioritized, including access to information, are the focus of a proposed second phase for this exercise that took place in November 2014. Risk Prioritization 35 TABLE 6.1. RISK PRIORITIZATION Degree of Impact Probability of Event Negligible Moderate Considerable Critical Catastrophic Highly Pests and Flood probable diseases (for example, locusts, wild animals, cassava mosaic virus) Irregular rainfall Probable Alluvian/soil Domestic price International Drought (late erosion volatility price volatility onset of rain, early Input price Exchange rate cessation of rain, volatility (for risk low cumulative example, rainfall) Execution risk fertilizer, diesel) Cyclone/storm Counterparty risk Infrastructure disruption Heat/excessive temperature Occasional Remote Wildfires Violent conflict pests and diseases (locusts, other insect pests, wild animals, RISK PRIORITIZATION plant and animal diseases); (3) price volatility (domes- The identification and prioritization of risks are critical tic and international markets for products and inputs); initial steps in designing a more comprehensive and effec- (4) infrastructure disruption; (5) political instability; (6) tive set of measures to manage those risks. To better uti- soil erosion; (7) counterparty risks; (8) wildfires; and (9) lize scarce resources, it is important to understand which execution risks. risks, or subset of risks, are causing maximum losses, and at what frequency. Figure 4.1 in chapter 4 highlights the priority production risks, using quantitative measures, for the crop subsector. Because of the paucity of data, some RISK MANAGEMENT of the risks could not be quantified. The ranking of risks Some risks are much more readily managed than are oth- combines qualitative and quantitative measures, based on ers, though there is no silver bullet to manage any given the assessment team’s evaluation. The relative significance risk. Effective risk management typically requires a combi- of these risks for different segments of the population var- nation of measures; some are designed to remove under- ies, as discussed in chapter 5. lying constraints and others to directly address the risk. One cannot control the weather, and significant resources A more in-depth discussion of agricultural risks is under- are required to effectively offset the effects of drought taken in chapter 3, whereas the analysis in this section and floods (for example, irrigation and flood control mea- highlights principal risks according to frequency and sures). Resource availability will often determine what is severity (see table 6.1). These include (1) weather-related possible, and integrated risk management programs are risks (drought, flood, cyclones, high temperatures); (2) often more effective than stand-alone programs. 36 Mozambique: Agricultural Sector Risk Assessment CATEGORIES OF RISK MANAGEMENT addition, they yield significant productivity gains and help MEASURES in mitigating the effects of climate change adaptation. The Risk management measures can be classified into the bulk of these measures are undertaken on individual farm- following categories: land or at the community level, whereas those involving a » Risk mitigation (ex ante). Actions designed broader watershed or landscape approach require coordi- to reduce the likelihood of risk or to reduce the nated measures across a number of communities. Under severity of losses (for example, soil and water con- the umbrella of conservation agriculture, a number of soil servation measures, changes in cropping patterns, and water conservation practices are being promoted in adoption of improved practices that improve per- Mozambique by projects funded by International Fund for formance and reduce risks such as conservation Agricultural Development (IFAD), Millennium Challenge farming, using short duration and tolerant vari- Corporation (MCC), USAID, and the African Develop- eties; irrigation and flood control infrastructure). ment Bank (AfDB) and implemented by the Ministry of » Risk transfer (ex ante). Actions that will trans- Agriculture and NGOs. Many of the practices of conser- fer the risk to a willing third party. These mecha- vation agriculture contribute to soil and water conservation nisms usually will trigger compensation in the case and help in mitigating drought and flood risk. Recent years of a risk-generated loss (for example, purchasing have witnessed growth of projects, with successful results insurance, reinsurance, financial hedging tools). at the local level. However, on a broader level, the inter- » Risk coping (ex post). Actions that will help the ventions are not yet on a scale to make any sizable impact affected population and the government cope with on mitigating effects of drought or flood at a regional or the loss. They usually take the form of compensa- national level. More coordinated landscape and/or water- tion (cash or in-kind), social protection programs, shed approaches and further scaling up of investments, and livelihood recovery programs (for example, at national and regional levels, with appropriate soil and government assistance to farmers, debt restructur- water conservations measures will yield multiple dividends ing, contingent risk financing). and should be the cornerstone of a stronger risk manage- ment plan in Mozambique. Table 6.2 uses these classifications to highlight some of the indicative interventions that could be undertaken to man- Tolerant seed varieties: Widespread availability of tol- age selected risks in Mozambique, grouped by manage- erant seed varieties and short-maturing varieties will help ment strategy. This is followed by a brief description of 11 in ensuring crop production during drought and flood risk management interventions. Although agricultural risk years in addition to reducing losses from pest and disease management measures are discussed individually and/or outbreaks. In essence, short-maturing varieties escape the sequentially, many of these interventions, if implemented effects of drought at either end of the growing season jointly, can have positive and complementary impacts when the frequency of rains tends to be most unpredict- while addressing multiple risks and contribute to improved able. Tolerant varieties, conversely, are better able to sur- risk management in the short, medium, and long term. vive periods of moisture stress or excess water and build resistance against specific pests and diseases. This needs to Interventions are described in more detail below, where be coupled with early warning about impending weather appropriate flood, drought, and pest and disease out- and disease outbreaks to help inform farmer decisions on breaks have been grouped together if the intervention using appropriate varieties and thus mitigate the risk of relates to various risks. total crop failure. The national system is currently unable to ensure sustainable delivery of tolerant seed, at afford- Soil and water conservation measures: Soil and able or subsidized prices, to all the farmers who need water conservation measures (such as sand dams, Ngare or them. As a result, in spite of all the work being done by Mhindu ridging, afforestation/reforestation, conservation the government, its partners, and the private sector, only agriculture practices) are effective and efficient mechanisms 10 percent of maize farmers and 4 percent of rice farmers for mitigating the risk of droughts and/or floods. In (Governement of Mozambique 2008) (mostly large- and Risk Prioritization 37 TABLE 6.2. POTENTIAL INTERVENTIONS FOR RISK MANAGEMENT Mitigation Transfer Coping Drought • Soil and water conservation • Crop insurance (farmer level) • Social safety net • Drought-tolerant varieties • Macro (government) level programs • Changing crop pattern crop insurance • On-farm storage • Irrigation (small and large scale) • Savings/credit • Improved extension services • Sovereign risk financing • Improved weather information and early warning • Improved water management Flood • Soil and water conservation • Social safety net • Flood control infrastructure programs investments (for example, dikes, • On-farm storage drainage) • Savings/credit • Regional coordination • Sovereign risk financing • Flood-tolerant varieties • Altering crop pattern • Improved extension services • Improved weather information and early warning • Improved water management Pest and disease • Promotion of integrated pest • Facilitate temporary outbreak management migration • Pest- and disease-tolerant varieties • Improved extension services • Improved information and early warning International • Improved market information • Price hedging • Social safety net price volatility systems (production, stocks, programs prices, and so on) • On-farm storage • Long-term forward contracts • Savings/credit • Reducing postharvest losses • Sovereign risk financing • Improving storage Domestic price • Improved market information • Price hedging • Social safety net volatility systems (stocks, production, prices programs and so on) • On-farm storage • Reducing postharvest losses • Savings/credit • Improving storage • Sovereign risk financing • Investment in transportation and storage infrastructure • Facilitate regional trade medium-scale commercial farmers) have access to these expand a number of tolerant varieties (HAR05 for resist- improved seeds that are largely aimed for yield improve- ance to downy mildew in maize, testing a flood-tolerant ment and not tolerance. A number of interventions are rice variety from East Asia, and developing drought-toler- currently being undertaken by the Instituto de Investigação ant varieties of maize, and so on), which is a positive step. Agrária de Moçambique (IIAM), in collaboration with Addressing this issue will require a national- or regional- the International Centre for Maize and Wheat Improve- level approach to develop a seed system that ensures the ment (CIMMYT) and other institutions, to develop and sustainable delivery of tolerant varieties. This might entail 38 Mozambique: Agricultural Sector Risk Assessment further support and expansion of seed multiplication by als. The trend of climate change will likely alter cropping seed producers and cooperatives, expansion of seed mul- calendars and seasonal agro-climatic conditions, which will tiplication by producers’ organizations and private com- necessitate altering cropping patterns and practices. This mercial firms, further support to agricultural input shops, is a complicated endeavor as altering farmers’ behavioral and continuation of social protection activities whereby patterns, which are rooted in psycho-socioeconomic belief NGOs provide improved seeds to poor smallholder house- systems, can be very challenging. Nonetheless, cropping holds. Furthermore, the possibility of research and devel- patterns that respond to appropriate weather and crop risk opment into newer varieties that are tolerant as well as profiles are fundamental to managing agricultural risks high yielding (during normal years) could be explored. while building resilience to climate change. Risk Transfer solutions: Agricultural insurance and Irrigation: Irrigation has the potential to generate siz- commodity price hedging (using forward contracts and able gains in household welfare, boost agricultural growth, futures) could be useful risk management instruments. Suc- improve food security, mitigate the impact of drought, and cessful functioning of farmer-level agricultural insurance promote overall economic growth in Mozambique. However, requires the presence of a number of necessary precondi- the performance record of irrigation schemes is very mixed, tions: affordability (ability and willingness to pay premiums); at times potentially increasing risk exposure (regular flooding relatively low frequency of events; robust crop and weather and so on). Nonetheless, there is a strong case for investment data infrastructure; farmers’ access to financial products and in irrigation. Although it might not be able to address severe services, and so on. Most of the necessary preconditions for systemic droughts, it could help ensure food availability in farm-level agricultural insurance, unfortunately, do not exist food-deficit areas in the case of localized drought or poor currently in Mozambique, which places serious limitations rainfall distribution. Irrigation infrastructure in Mozambique on rolling out a large-scale agricultural insurance program in is less developed than in the average Sub-Saharan African Mozambique. Appendix F provides further details of poten- country. As of 2007, 2.7 percent of the country’s cultivated tial risk transfer and risk financing products in Mozambique. area was equipped for irrigation, below the Sub-Saharan Price hedging using futures and forward contracts to man- average of 3.5 percent. The equipped irrigation area con- age price volatility in cotton and sugarcane is currently being tributes just 4.8 percent to the total agricultural output. In undertaken by some of the larger exporters in Mozam- order to reverse this trend, the government of Mozambique bique. Such instruments could potentially be used in other approved a National Irrigation Strategy for 2011–19. The commodities and by small exporters, traders, and farmers’ strategy is estimated to cost US$645 million and aims to dou- groups. However, they too require a large number of prereq- ble total irrigated land in the provinces of Sofala, Manica, uisites: homogenization/standardization of commodities, and Zambezia from 66,000 hectares to 113,000 hectares by transparency in commodity markets, limited interventions by 2019. There is an increased push for small-scale irrigation, the government, the presence of infrastructure (storage and owing to its low cost and higher sustainability. For large-scale warehousing), and so on. In Mozambique, many of these infrastructure projects, there is an emphasis on rehabilitat- preconditions do not exist, which limits the potential use of ing existing irrigation infrastructure and improving manage- hedging as a tool for managing price risk at a broader level. ment. If designed appropriately, irrigation systems could help reduce drought risk and manage flood risk. Altered cropping patterns: Over the past few decades, Mozambique has witnessed changes in traditional cropping Weather, pests and disease, and early warning patterns. For instance, millet and sorghum (more drought information systems for farmers: Ready access to tolerant crops) have replaced maize (a more sensitive timely, accurate, and localized information about impending crop), especially in low rainfall areas, resulting in increased events that could have a severe impact on crops is a prerequi- risk exposure. Losses from drought could be significantly site to enable preemptive actions by farmers to reduce expo- reduced by replacing maize with sorghum, millet, or root sure or losses. This may mean relocating and minimizing crops in areas where drought is particularly common. Root losses before a flood or postponing planting or early harvest- crops are generally more drought tolerant than are cere- ing, as well as altering agronomic practices. Besides helping Risk Prioritization 39 managing risks, this information could also help farmers principal risks for Mozambique’s agricultural sector that to better manage inputs and improve yields. The needs for will be further aggravated attributable to climate change, such information will further increase because of weather hence effective water management has to be the corner- uncertainties induced by climate change. Despite a num- stone of any risk management strategy. This involves plan- ber of small initiatives and rapid growth of information and ning, developing, distributing, and managing the optimal communication technology (ICT), the majority of the farm- use of water resources. Activities such as improving data ers still do not have access to such information. Although and reforming water governance along with education there is an effective early warning system in Mozambique and training on water management would aid in water for flood, cyclone, and food security, its emphasis is on dis- availability, particularly in drought-prone areas. Improved aster response and saving human lives. Using information water management both at the public sector level by systems to guide and influence decision making by farmers government and private sector level (individual farming and communities to reduce hazard, exposure, and losses at household and community) will be required for tackling the farm level is still not high on the development agenda, the issue. The Ministry of Foreign Affairs of Denmark despite the multiple benefits such systems could generate. (DANIDA) supports activities to encourage those in arid and semiarid zones with water management. Flood control infrastructure investments: Thirteen big rivers flow across from southern Africa through Mozam- Improved access to extension services: Improved bique, exposing the agricultural sector relying on these rivers access to extension services would allow producers to be bet- to frequent flooding. Dams, dikes, and drainage systems are ter informed and to access advice, technology, and inputs to some of the flood control infrastructure that can effectively alter their agronomic practices in view of the current and mitigate the impacts of such events. The rehabilitated Mass- emerging risk profile of the agricultural sector. The public ingir dam on the Limpopo River prevented floods in 2008 extension service in Mozambique has been characterized that could have affected the cities of Chókwe and Xai-Xai by great variability in terms of availability of extension and protected the largest irrigation scheme in the country in staff. Efforts of public extension are being complemented Chókwe. Protection dikes have also been popular measures by NGOs and the private sector. In 2009, it was reported to protect settlements from floods: during 2007 and 2008, that 378,043 households were covered by public extension for example, dikes along the Zambezi were seen to have pro- services; 203,683 households by NGOs; and 375,351 house- tected the towns of Luabo and Marromeu from inundation. holds by private sector extension services (SADC 2010). Unfortunately, the existing condition of flood infrastructure Though the coverage of extension services is increasing, it is poor and in dire need of repair. Furthermore, the current reaches only 12 percent of farmer households in Mozam- flood control infrastructure must be significantly expanded bique. The bulk of farming households do not have access to to adequately address frequent flooding, as well as extreme any extension services and further investment and expansion flood events. Although the principal objective of such infra- of extension services and development of new delivery chan- structure has to be saving human settlements and infrastruc- nels will assist in improved management of agricultural risks. ture, preventing crop losses, and diverting and storing excess water for irrigation could be a useful secondary aim. Although Social safety net programs: Despite the vulnerabil- national- and regional-level planning and coordination are ity of the population to agricultural risk, other perils, and necessary, the nature of the problem requires regional coor- acute poverty, the social safety net programs are relatively dination among neighboring countries. There are already weak in Mozambique. The World Bank’s recent safety net regional bodies, such as the Zambezi River Basin Initiative, assessment for Mozambique identified about 40 differ- that address issues related to shared water resources. Regional ent social protection-related programs, but most of them coordination also would allow countries to come together and were fragmented, had low coverage, and were not well provide critical mass to address issues before and as they arise. targeted. There are limited mechanisms to help affected populations cope with high-frequency and high-impact Improved water management practices: Water covariate shocks. Although emergency food aid and disas- deficit (drought) and water excess (flood) are two of the ter relief from donors and government partially help the 40 Mozambique: Agricultural Sector Risk Assessment affected population to survive, aid alone is not sufficient to group of stakeholders, other interventions have an help the population recover from income and asset losses. opportunity to reach many more stakeholders. There is an urgent need to consolidate, scale up, refocus, » Replicability. Some interventions with the and improve targeting of social safety net programs. A potential to reach larger numbers of stakeholders recently approved US$50 million social protection pro- can be more easily replicated than others. ject by the World Bank aims to provide temporary income » Cost. Without detailed assessments, it is difficult support to extremely poor households and to put in place to estimate the cost of some interventions. How- the building blocks of a social safety net system. ever, based on the experience of the assessment team, the relative cost of interventions can be Improved market information system: Accurate, assessed. The cost involved in a large-scale irriga- timely, and transparent availability of information about tion project, for example, is generally much higher the production and stocks, trade flows, and prices in dif- than the cost involved in setting up a system for ferent markets can help manage price volatility in domes- seed distribution. tic markets. Although Mozambique’s Agricultural Market » Difficulty of implementation. The technical Information System (SIMA) collects and provides price complexity of interventions and the capacity of information concerning major agricultural commodities, local stakeholders to implement them are filters there is limited reliable information about stocks and pro- that could be used to prioritize decisions. Simpler duction, which hampers SIMA’s efficacy. Strengthening interventions might find greater acceptability and agricultural statistics and collection and compilation of will be easier to implement. production data, along with collecting reliable informa- » Return time. Some interventions have a long tion on stocks and trade flows, will improve transparency gestation period, whereas others could yield quick in agricultural markets and help in price risk management. results. Although risk management will require short-, medium-, and long-term perspectives, quick DECISION FILTERS wins are often a high priority for decision makers. In a resource-constrained environment such as Mozambique, » Sustainability of benefits/interventions. decision makers are compelled to find the quickest, cheapest, Given the scarcity of public resources, it is critical and most effective measures among myriad possibilities. A to ensure that money spent on these interventions detailed and objective cost-benefit analysis can help in select- will have a lasting impact. ing the most appropriate intervention options. But conduct- » Environmental impact. Some of the risk man- ing a cost-benefit analysis of many different options in itself agement interventions, especially large-scale spray- can be costly and time consuming. Further, there is a range ing of chemicals for locust destruction, could have of not easily quantifiable considerations/criteria that are not long-term catastrophic consequences for the envi- easily factored into such types of analysis. Using decision fil- ronment. Hence, it is important to scrutinize the ters to evaluate and prioritize among a list of potential inter- potential adverse environmental impacts of a given ventions could help in making rational resource allocation intervention. decisions in lieu of a detailed cost-benefit analysis. » Potential impact on poverty alleviation. Whereas some interventions would directly con- The following criteria were used by the World Bank team. tribute to improved income and poverty allevia- There are a number of complex analytic screening tools tion, others might indirectly contribute to the goal. to assess all of these decision filters and this study does not Using this filter helps to identify risk management claim methodological rigor while assessing these filters. interventions that might yield large poverty allevia- Instead, the study team applied these filters as a sort of tion dividends. rapid assessment to obtain first order of approximation, based on their assessment of the situation on the ground: The results after application of the decision filters (table 6.3) » Outreach. Although some interventions, because are indicative and imperfect; nonetheless, they present a of prerequisites, might be able to benefit a small first step toward the development of a more comprehensive Risk Prioritization 41 42 TABLE 6.3. DECISION FILTERS FOR PRIORITIZING INTERVENTIONS Potential Return Sustainability Impact on Time for of Benefits/ Environmental Poverty Outreach Replicability Cost Difficulty of Impact Interventions Impact Alleviation (low, (low, (low, Implementation (low, (low, (positive, (high, medium, medium, medium, (low, medium, medium, medium, neutral, medium, high) high) high) high) high) high) negative) low) Building tolerant H H L M S M Neutral H systems (flood-, drought-, and disease-tolerant varieties) Small-scale M H L M M H Neutral H irrigation Soil and water M H M M M H Positive H conservation measures (for example, conservation agriculture) Improved market H M M M L H Neutral L information system Altered cropping M M L H S H Neutral M patterns Improved access to M M H M S H Positive H extension services Saving/credit L H M M S H Neutral H/M On-farm storage M H M M S H Neutral H Social safety net M H H M S M Neutral H programs (for example, food/ cash/vouchers for work, food aid) Mozambique: Agricultural Sector Risk Assessment Improving water H H L M S H/M Positive M management practices Large-scale L L H H H M Positive or L irrigation Negative Risk Prioritization Flood control M M H H M/L H Negative or M infrastructure positive investment (dikes, drainage, and so on) Timely and H H H H S H Neutral M reliable availability of weather information to farmers and other stakeholders Regional H L H M H Neutral M coordination Promotion of H H M M S M Positive M integrated pest management Subsidized crop L L H L M M Neutral M insurance (for example, bundled with credit or input) Commercial L L H L M L Neutral L catastrophic weather insurance Note: H = high, M = medium, L = low. 43 TABLE 6.4. MULTIPLE “WINS” Climate Climate Reduces Reduces Compensates Improves Change Change the Risk the Losses after the Loss the Yield Mitigation Adaptation Soil and water conservation Y Y N Y Y Y measures (for example, conservation agriculture) Improved access to extension Y Y N Y Y Y services Improved water management Y Y N Y N Y practices Altered cropping patterns Y Y N N Y Y Flood control infrastructure Y Y N Y N Y investment (dikes, drainage, and so on) Small-scale irrigation Y Y N Y N Y Large-scale Irrigation Y Y N Y N Y Improved market information Y Y N N N N system Building tolerant systems Y Y N N N Y (flood-, drought-, and disease- tolerant varieties) Timely and reliable N Y N Y N Y availability of weather information to farmers and other stakeholders Regional coordination Y Y N N N Y Promotion of integrated pest Y Y N Y N N management Subsidized crop insurance (for N N Y Y N N example, bundled with credit or input) Commercial catastrophic N N Y N N N weather insurance Sovereign risk financing N N Y N N N Saving/credit N N Y N N N On-farm storage N N Y N N N Social safety net programs (for N N Y N N Y example, food/cash/vouchers for work, food aid) Facilitate temporary migration N N N N N N strategy for managing risks to the agricultural sector. Table The process put forth here and carried out by the team 6.4 goes a step further and considers the various outcomes allowed for an initial consideration of all options. Follow- and whether multiple “wins” can be incurred. In both ing this, further discussions took place with the government tables, potential outcomes are grouped by highest prioriti- and ultimately this informed the decision of MASA to con- zation (table 6.3) and most “wins” (table 6.4). tinue with the three areas detailed in the ASRSA report. 44 Mozambique: Agricultural Sector Risk Assessment TABLE 6.5. RISK MANAGEMENT INTERVENTION INTEGRATION WITH PNISA Intervention PNISA Components PNISA Components Soil and water conservation Component 1: Production Program 6: Agrarian Extension measures (for example, and Productivity conservation agriculture) Improved access to extension Program 6: Agrarian services Extension Improving water management Component 1: Production Program 6: Agrarian Extension practices and Productivity Access to information (weather, Program 6: Agrarian Program 19: Cartography and price, diseases, early warning, Extension Remote sensing and so on) to farmers Altered cropping patterns Program 6: Agrarian Extension Flood control infrastructure Program 5: Agrarian Program 7: Irrigation investment (dikes, drainage, and Investigation so on) Small-scale irrigation Program 7: Irrigation Large-scale irrigation Program 7: Irrigation The multiple “wins” achieved by the different interven- All seven priority interventions align with PNISA (see tions illustrated in table 6.4 include reducing the risk, table 6.5) and many of these interventions are already reducing the losses incurred because of the risk, whether in place with successful positive outcomes. Many of they lead to a compensation for those affected after the these interventions are being implemented on a much loss, whether they improve the yield of the crop affected, smaller scale and are having positive impacts but at a and whether there is any aspect of climate change mitiga- localized level. Greater emphasis should be placed on tion of adaptation. scaling up these interventions to the national level to make a meaningful impact on the agricultural sector of Based on prioritization of risk and intervention measures, Mozambique. the following seven intervention categories might yield the greatest risk management benefits: Expansion of the scale of these interventions would require 1. Soil and water conservation (including conserva- understanding the landscape of interventions, assessing tion agriculture) their relative efficacy, understanding principal barriers/ 2. Improved access to extension services challenges to success and scale, and identifying leverage 3. Improved water management practices points and necessary interventions to increase their access 4. Access to information (weather, price, diseases, to a wide majority of agricultural sector stakeholders. early warning, and so on) to farmers Assessing solutions to help prioritize specific interventions 5. Changes in cropping practices, including cropping to scale up priority programs and putting in place a risk patterns and use of improved varieties management implementation plan will be the next steps 6. Flood control infrastructure in the process of building resilience in Mozambique’s agri- 7. Irrigation (small and large scale) cultural sector. Risk Prioritization 45 REFERENCES Bay, A. 1997. “The Seed Sector in Mozambique.” Seed Science and Technology 25 (3): 427–42. Brito, R., ed., and E. H. A. Homan. 2012. Responding to climate change in Mozambique: Theme 6: Agriculture. October. Maputo: INGC. Chilonda, P., V. Xavier, L. Luciano, H. Gemo, A. Chamusso, Precious Zikhali, A. Faria, J. Govereh, E. Musaba, S. Manussa, B. Acubar, L. Osvaldo, N. Alage, E. Macome, and A. Manganhela. 2012. Monitoring and Evaluating Agricultural Growth, Trade and Poverty in Mozambique. CIA. 2013. “Mozambique.” Factbook. Cungara, B., and J. Garret. 2011. Agricultural Sector in Mozambique: Situation Analysis, Constraints and Opportunities for Agricultural Growth. IFPRI. Donovan, Cynthia, and Emilio Tostao. 2010. “Staple Food Prices in Mozambique.” Prepared for the Comesa policy seminar, “Variation in Staple Food Prices: Causes, Consequence, and Policy Options,” Maputo, Mozambique, January 25–26, under the Comesa-MSU-IFPRI African Agricultural Marketing Project (AAMP). FAO (Food and Agriculture Organization [of the UN])/WFP (World Food Programme). 2010. Crop and Food Security Assessment Mission to Mozambque. Rome: FAPWFP. FEWSNET. (2006). Baseline Survey of Food Security and Nutrition in Mozambique. Government of Mozambique, M. o. 2008. TIA (Trabalho de Inquerito Agricola). Government of Mozambique, M. o. 2009/2010. TIA (Trabalho de Inquerito Agricola). Government of Mozambique, Ministry of Agriculture. 2010. Strategic Plan for Agricul- tural Development (PEDSA), 2010–2019. IFPRI (International Food Policy and Research Institute). 2011. Agriculture Sector in Mozambique: Situation Analysis, Constraints, and Opportunities for Agricultural Growth. ———. 2012. Southern African Agriculture and Climate Change: A Comprehensive Analysis—Mozam- bique. CGIAR. December. (Short summary only—main document to be published shortly) INGC (National Institute for Disaster Management). 2006. Indicative Plan for Prevention and Mitigation of Natural Disasters. PDPMCN. ———. 2009a. Study on the Impact of Climate Change on Disaster Risk in Mozambique—Synthesis Report. ———. 2009b. Perfis de Calamidades—Cultivando a Cultura de Prevenção: Perfil de Acidentes, Seca, Pragas e Ciclones. Jaffee, S., P. Siegel, and C. Andrews. 2008. Rapid Agricultural Supply Chain Risk Assessment: Conceptual Framework and Guidelines for Application. World Bank, Commodity Risk Management Group, Agriculture and Rural Development Department. LANDac (IS Academy on Land Governance). 2012. Mozambique Food Security and Land Gov- ernance factsheet. KIT on behalf of LANDac—The IS Academy on Land Governance. McSween, S., T. Walker, V. Salegua, and R. Pitoro. 2006. “Economic Impact on Food Security of Varietal Tolerance to Cassava Brown Streak Disease in Coastal Mozambique.” Research Report No. 1E, August, Maputo, Institute of Agricul- tural Research of Mozambique. Risk Prioritization 47 McSweeny, C., M. New, and G. Lizcano. 2008. Mozambique. UNDP Climate Change Country Profiles. MICOA (Ministry for Coordination of Environmental Affairs). 2003. Mozambican Ini- tial Communication to the United Nations Framework Convention on Climate Change. Mucavele, F. 2000. Analysis of Comparative Advantage and Agricultural Trade in Mozambique. Technical Paper No. 107, USAID, SD Publication Series, Office of Sustainable Development, Bureau for Africa. OIE (World Organization for Animal Health). 2008. Pauw, Karl, James Thurlow, Rafael Uaiene, and John Mazunda. 2012. “Agricultural Growth and Poverty in Mozambique: Technical Analysis in Support of the Com- prehensive Africa Development Program (CAADP).” Working Paper 2, Novem- ber, Mozambique Support Program, International Food Policy Research Insitute, Washington, DC. SADC (Southern African Development Community). 2010. “Proceedings of the Work- shop on Information Sharing among Extension Players in the SADC Region.” http:// www.sadc.int/fanr/agricresearch/icart/inforesources/ImprovingExtension.pdf. Sarris, Alexander, and Panayiotis Karfakis. 2006. “Household Vulnerability in Rural Tanzania.” Paper presented at the CSAE conference, “Reducing Poverty and Inequality: How Can Africa Be Included?” March. SETSAN (Technical Secretariat for Food Security and Nutrition). 2010. Comprehen- sive Food Security and Vulnerability Analysis. Republic of Mozambique. Vulnerability Assessment Group. World Food Programme. September. Tschirley, David L., Danilo Carimo Abdula, and Michael T. Weber. 2006. “Toward Improved Maize Marketing and Trade Policies to Promote Household Food Secu- rity in Central and Southern Mozambique.” Ministry of Agriculture and Rural Development, Directorate of Economics, Republic of Mozambique. van Logchem, B., and R. Britos, eds. 2009. Study on the Impact of Climate Change on Disas- ter Risk in Mozambique: Synthesis Report. Mozambique: INGC. Waterhouse, Rachel. 2009. “Vulnerability in Mozambique: Patterns, Trends, and Responses.” Paper presented at the IESE conference, “Poverty Dynamics and Pat- terns of Accumulation in Mozambique,” Maputo, April 22–23. Wheeler, David. 2011. “Quantifying Vulnerability to Climate Change: Implications for Adaptations Assistance.” CGD Working Paper 240. Washington, DC: Center for Global Development. (and accompanying data set and maps) World Bank. 2010a. The Economics of Adaptation to Climate Change: Mozambique. ———. 2010b. “Cotton Supply Chain Risk Assessment.” ———. 2012a. Agribusiness Indicators: Mozambique. Geneva: World Bank. ———. 2012b. “Social Protection Assessment. Review of Social Assistance Programs and Social Protection Expenditures.” February, Africa Region, Social Protection Unit. Report No. 68239-MZ, World Bank, Geneva. Zavale, Helen, Edward Mabaya, and Ralph Christy. 2006. “Smallholders’ Cost Effi- ciency in Mozambique: Implications for Improved Maize Seed Adoption.” Paper Prepared presentated at the International Association of Agricultural Economists Conference, Gold Coast, Australia, August 12–18. 48 Mozambique: Agricultural Sector Risk Assessment APPENDIX A METHODOLOGY/APPROACH TO PRODUCTION RISK LOSSES CALCULATION » Gross production value [A] is taken from FAOSTAT. » Share of production from crops/agriculture [b] was assumed to be about 60 percent of gross production value. – b = 0.6 × A » The amount of land planted in hectares for arable and permanent crops [C] per year was taken from FAOSTAT. » The estimated value per hectare of land planted with arable and permanent crops [d] was derived by dividing the figure for the share of production from crops/agriculture using constant 2004–06 U.S. dollars by the estimated amount of land planted for arable and permanent crops. – b/C = d » The hectare risk event number of hectares lost (e) was taken from analysis of annual reports from SETSAN (food security situation reports), the PES (The Annual Balance of Economic and Social Plan and Annual Economic and Social Plans), FEWSNET reports, Global Information and Early Warning System (GIEWS)/FAO reports and data downloaded from the Early Warning System. » The estimated loss (f) was calculated by multiplying the estimated value per hectare by the hectares lost in a given year because of risk events. – d×e=f Risk Prioritization 49 APPENDIX B WEATHER YIELD REPORT BACKGROUND The World Bank is conducting a study of the relationship that several climatic events have on different crop yields for Mozambique. The purpose of the study is to deter- mine whether yield is affected by climatic events and by how much these events affect it. Figure B.1 shows the political division of Mozambique, which comprises 11 provinces, sorted here alphabetically. FIGURE B.1. PROVINCES OF MOZAMBIQUE 1 1. Cabo Delgado 8 2. Gaza 10 7 3. Inhambane 11 4. Manica 5. Maputo (city) 9 6. Maputo (province) 4 7. Nampula 8. Niassa 9. Sofala 3 2 10. Tete 11. Zambezia 6 –5 Source: https://en.wikipedia.org/wiki/Provinces_of_Mozambique. Agricultural information is provided on a regional basis but only for 10 provinces because the city of Maputo has no available data. The database comprises two vari- ables: sowed area in thousand hectares, and production in thousand tons. Yield is not provided, but can be estimated as follows: Production Yield = Area Risk Prioritization 51 RAINFALL PATTERNS IN FIGURE B.2. LOCATION OF MOZAMBIQUE METEOROLOGICAL STATIONS A weather database was provided that consists of 31 weather stations spread across the country with data on a daily basis from January 1, 1979, to December 31, 2009. The available variables are precipitation, maximum tem- perature, and minimum temperature. Figure B.2 shows the geographic distribution of the weather stations. On a general basis, rainfall follows a clear pattern through- out the whole country. From November to April rain is plentiful but is scarce from May to October it. Figure B.3 shows the mean cumulative rainfall per month for all the stations in each region. DROUGHT AND EXCESS RAINFALL ANALYSIS Because of the rainfall pattern described above, the year is not considered to be the calendar year, but rather the period from October of the previous year to September of the next year, in order to consider that rain falls mainly in the November–March period. Instead of showing the cumulative rainfall in each year, standardization was Source: Uribe. applied according to the following formula: Note: Labels indicate station ID. StdRain = (Σ sep e i = nov Preci − mi ) representing the different provinces (red = North, green i si = Central, blue = South). Where Figure B.4 helps to visualize that there were four dry years: StdRain, standardized cumulative rainfall 1979 stands out as a year in which almost all provinces Prec, daily rainfall experienced insufficient rainfall, whereas 1992 follows μ, mean yearly rainfall closely as another dry year. Both 1983 and 1988 were also ∑, standard deviation of yearly rainfall dry years with the lack of rain experienced only in some i, year provinces. This way, it is easier to highlight drought and excess rain- At the other end of the rain spectrum, figure B.5 shows fall events. that there were three consecutive years in which rain was abundant in Mozambique. From 1999 to 2001, stations SUMMARY in many provinces experienced plentiful precipitation, By counting all the stations that had a negative anomaly particularly in the southern provinces of Inhambane, of precipitation (considered as a drought event), the fol- Gaza, and Maputo. Additionally, 1981 was humid, lowing chart (figure B.4) summarizes the number of total mainly in the south, whereas 1989 was humid, but stations that had drought events per year with colors mainly in the north. 52 Mozambique: Agricultural Sector Risk Assessment Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) 0 50 100 150 200 250 300 0 20 40 60 80 100 120 140 160 180 200 0 50 100 150 200 250 Ja Ja Ja n n n Risk Prioritization Fe Fe Fe 13 b b b M M M 7003 3002 ar ar ar Ap Ap Ap M r M r M r Number of stations 3053 ay ay ay Ju Ju Ju 0 2 4 6 8 10 12 14 16 18 19 n n n 14 22 34 7 Ju Ju Ju 19 9 l l l 8 Au Au Au Tete Region 19 0 g g g 8 Se Se Se p p p 7004 7007 7010 19 1 Inhambane Region O O O Cabo Delgado Region 8 Nia c c c Man Map 19 2 83 N t N t N t 19 ov ov ov 8 D D D 19 4 ec ec ec Zam 85 Gaza C.Del 19 8 Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) 19 6 8 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 0 20 40 60 80 100 120 140 Inh 19 7 8 Tete Ja Ja Ja 19 8 n n n 8 Fe Fe Fe 19 9 b b b 9 M M M Sof 1002 4001 19 0 ar ar ar Nam 91 Ap Ap Ap 19 M r M r r 9 M 4029 1003 19 2 ay ay ay 9 Ju Ju Ju 19 3 n n n 9 Ju Ju Ju 19 4 l l l Year 9 Au Au Au 19 5 9 g g g Gaza Region Niassa Region 19 6 Se Se Se p p p Zambezia Region 9 O O O 19 7 ct ct ct 9 N N N 19 8 ov ov ov 9 8007 8010 8032 8035 8050 20 9 D D D ec ec ec 0 20 0 0 FIGURE B.4. DROUGHT EVENTS, 1979–2009 20 1 Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) Mean cumulative rainfall (mm.) 0 FIGURE B.3. MEAN CUMULATIVE RAINFALL PER MONTH 20 2 0 20 40 60 80 100 120 140 160 0 50 100 150 200 250 300 0 50 100 150 200 250 300 350 0 20 3 Ja Ja Ja 04 n n n 20 Fe Fe Fe 0 b b b M M M 5015 2006 20 5 06 ar ar ar 20 Ap Ap Ap 0 20 7 M r M r M r 9005 9044 0 ay ay ay 20 8 Ju Ju Ju 09 n n n Ju Ju Ju 5032 5045 l l l 9052 Au Au Au g g g Manica Region Nampula Region Se Se Se Maputo Region p p p 9063 2008 2049 2051 O O O ct ct ct N N N ov ov ov D D D ec ec ec 53 FIGURE B.5. EXCESS RAINFALL EVENTS, 1979–2009 16 Map Gaza Inh Sof Man Zam Tete Nam Nia C.Del 14 12 Number of stations 10 8 6 4 2 0 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 2099 2000 2001 2002 2003 2004 2005 2006 2007 2008 09 19 Year season. The determination coefficient is a measure of RAINFALL—YIELD the proportion of the variability in yield that is being REGRESSIONS explained by cumulative rainfall. Therefore, whenever it It is significant that geographic resolution of data is not the is high (<40 percent), it is a good indication that rain and same. Rainfall data are available on point estimates; at the yield are related. same time, yield data are available regionwide, making up the whole political division as described above. Therefore, it is needed to make equivalent the geographic resolution MAIZE (MILHO) of both data sets. As there is no information regarding the Maize is grown in most of Mozambique with an average sowing zones within each region; all available stations within of approximately 1.2 million hectares sown countrywide. the region were considered to match the yield information But the area has been steadily increasing each year, with of each region. Thus, the average of the available stations 1.6 million hectares sown in 2011. Production, conversely, within a region was used as a proxy of each region’s rainfall. has been increasing at a higher rate, with more than 2 mil- lion tons produced during 2011. Linear regression models use each crop’s growing phase cumulative rainfall as the explanatory variable for yield. Thus, yield on a national level follows a similar pattern, with a steady increase, particularly since 2006, up to 1,300 l Yield b0 b1CumRain1 kg/ha, whereas the average yield is 894 kg/ha. From 1992 Yield l b0 b2CumRain2 to 1995, yield was low, but it has been steady since then Yield l b0 b3CumRain3 (except in 2005, when yield was 769 kg/ha, lower than average). Where CumRain1 is the cumulative rainfall of the sowing season The distribution of surface sown differs by region; in the CumRain2 is the cumulative rainfall of the midseason central regions of Zambezia, Tete, and Manica, the most CumRain3 is the cumulative rainfall of the harvest season surface is sown, whereas Cabo Delgado, Inhambane, and Maputo are the regions where the least surface is sown. The main objective of the regression analysis is to Figure B.6 shows the distribution of surface by region calculate the determination coefficient (R2) for each during 2008. 54 Mozambique: Agricultural Sector Risk Assessment FIGURE B.6. MIDSEASON REGRESSION MODELS FOR NIASSA, INHAMBANE, AND MAPUTO PROVINCES All studies of charts show a positive slope, indicating Production follows the same pattern as surface, almost that the more rain, the more yield; as such, this signals a emulating the behavior of the surface sown. Yield, con- drought risk for the midseason. It is worth noting that in versely, has been quite steady. Since 1994, yield has oscil- in the Niassa province, the two lowest yield years (2005 lated between 800 and 1,200 kg/ha, with little variation and 2007), in which yield was approximately 600 kg per other than a peak in 2008 of 1.19 tons per hectare. The hectare, were also the years with the lowest rainfall (227 average yield is 930 kg/ha. Zambezia province provides mm in both instances). almost half of the national surface, whereas Sofala and the northern provinces of Nampula, Cabo Delgado, and In the province of Inhambane, there are also two observa- Niassa provide the other half. Figure B.7 illustrates the tions with low rainfall as well as low yield. During 2002 distribution of surface sown by region. and 2005, less than 60 mm fell and yield was low (less than 180 kg per hectare), but the lowest yield year (2003, with As with maize, regional data are only available from 156 kg/ha) had 143 mm of rainfall. 2002 to 2008 (with the exception of 2004). Regional mean yield is 315 kg/ha, which is significantly lower In Maputo, yield shows more volatility with a minimum of than the national mean yield (930 kg/ha). The standard 194 kg/ha, matching a dry year (2003 with 110 mm), and deviation is 204 kg/ha, which is an indication of low a maximum of 1,028 kg/ha, matching a rainy year (2006, volatility in yield. with 273 mm), so the drought signal is clear for this stage. Given the absence of a specific sowing calendar for each crop, the same general sowing calendar used for RICE (ARROZ) maize was used for rice. Cumulative rainfall was calcu- Rice is sown mainly in the coastal central region of lated for each weather station for each stage in the cal- Mozambique, but the surface dedicated to its production endar in order to determine the relationship between has varied significantly during the past 20 years. In the rainfall and yield. Simple linear regressions were run late 1990s, almost 200,000 ha of rice were sown, but then per region in order to determine the relationship of at the turn of the century it was halved to a minimum of yield and rainfall. Table B.1 shows the determination 67,000 ha in 2005. Subsequently, the area sown grew to coefficient (R2) of the linear regression models applied reach approximately 200,000 ha again. by region. Risk Prioritization 55 FIGURE B.7. 2008 RICE SURFACE SOWN BY REGION (HECTARES) Source: Instituto Nacional de Estatística. TABLE B.1. DETERMINATION COEFFICIENT For Manica province, the outcome is negative, indicating (LINEAR REGRESSION MODEL, that the more rain, the lower the yield. This is particularly evident because during 2007, yield was high (714 kg/ha) RAINFALL, AND RICE YIELD) with only 73 mm of rain during the harvest season. Con- Determination Coefficient versely, during 2002, yield was quite low (61 kg/ha) but Sowing Midseason Harvest 276 mm of rain fell, indicating an excess rainfall event Province (%) (%) (%) affected yield during that season. Cabo Delgado 40 5 10 Niassa 0 95 14 However, for Sofala the outcome is positive given that Nampula 0 16 0 during very humid years (2003 and 2006) when approx- Tete 34 20 1 imately 800 mm of rain fell during the harvest season; Zambezia 39 11 1 the higher rainfall matched higher yields (300 kg/ha). It Manica 10 27 36 is worth noting that in Sofala province there is only one Sofala 6 1 37 meteorological station, so it is possible that it is not located Inhambane 0 2 20 near the areas where rice is sown. Gaza 8 9 27 Maputo 31 60 26 SORGHUM (MAPIRA) The surface dedicated to sorghum production is spread For Zambezia, the most important region in terms of sur- more widely throughout the country. During the 1990s, face, and Cabo Delgado, cumulative rainfall during the an average of 400,000 ha were sown with sorghum, fall- sowing season shows a relatively important correlation ing to less than 300,000 ha in the new century. In recent with yield. In these regions, rain explains roughly 40 per- years, this has doubled with approximately 600,000 ha cent of the variability in yield. sown nationwide. 56 Mozambique: Agricultural Sector Risk Assessment FIGURE B.8. 2008 SORGHUM SURFACE SOWN BY REGION (HECTARES) Source: Instituto Nacional de Estatística. FIGURE B.9. SORGHUM YIELD HISTOGRAM FOR ALL REGIONS 4.5 4.0 3.5 Frequency density 3.0 Input 2.5 Minimum 0.06135 Maximum 0.77165 2.0 Mean 0.30132 Std dev 0.14898 Values 54 1.5 Median = 0.27329 1.0 0.5 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Yield (tons per hectare) Production follows the same pattern as surface, almost As with the other crops, regional data are only available emulating the behavior of the surface sown, but always from 2002 to 2008 (again, with the exception of 2004). with less production than surface sown. Yield has been Regional mean yield is 301 kg/ha, which is significantly quite steady. From 1995 onward, yield has had little vari- lower than the national mean yield (556 kg/ha). The ation above or below the average of 550 kg/ha. There standard deviation is 148 kg/ha, which is an indication of seems to have been an increase recently, as 2011 recorded low volatility in yield. Figure B.9 shows the distribution of a maximum of 787 kg/ha. yield in all regions. Sofala is the main province where sorghum is grown, with Because specific sowing calendars do not exist for each almost 90,000 ha sown during 2008. Nampula, Cabo Del- crop, the same general sowing calendar used for maize gado, Zambezia, and Manica follow with approximately was used for sorghum. Cumulative rainfall was calculated 50,000 ha in each province. Figure B.8 illustrates the dis- for each weather station for each stage in the calendar in tribution of surface sown by province. order to determine the relationship between rainfall and Risk Prioritization 57 TABLE B.2. DETERMINATION COEFFICIENT GROUNDNUTS (AMENDOIM) (LINEAR REGRESSION MODEL, The surface dedicated to groundnut production has been RAINFALL, AND SORGHUM steady during the past 20 years. Approximately 300,000 YIELD) ha have been sown annually, with a maximum of 357,000 Determination Coefficient ha sown during 2009. Sowing Midseason Harvest Province (%) (%) (%) Production has also been steady at about 100,000 tons per year, but recently it has decreased to a level of 70,000 Cabo Delgado 44 10 21 tons per year. Thus, yield shows a peak and a valley. In the Niassa 5 20 19 late 1990s, yield reached a peak of more than 500 kg/ha, Nampula 5 3 8 but recently, yield has been below the 20-year average of Tete 58 16 77 353 kg/ha, with a minimum of 190 kg/ha during 2009. Zambezia 59 6 2 Manica 1 8 4 Nampula is the most important province in terms of sur- Sofala 6 2 48 face, with more than 160,000 ha sown during 2008. Zam- Inhambane 4 0 0 bezia, Tete, and Inhambane follow, with approximately Gaza 0 4 3 60,000 ha in each province. Figure B.10 illustrates the dis- tribution of surface sown by province. yield. Simple linear regressions were run per region in order to determine the relationship of yield and rainfall. Once again, regional data are available only from 2002 to Table B.2 shows the determination coefficient (R2) of the 2008 with the exception of 2004. Regional mean yield is linear regression models applied by region. 211 kg/ha, which is slightly lower than the national mean yield (353 kg/ha). The standard deviation is 117 kg/ha, There are three provinces for which rain during the sow- which is an indication of low volatility in yield. The histo- ing season is significant to explain variation in yield: Cabo gram in figure B.11 shows the distribution of yield in all Delgado, Tete, and Zambezia. regions. FIGURE B.10. 2008 GROUNDNUTS SURFACE SOWN BY REGION (HECTARES) Source: Instituto Nacional de Estatística. 58 Mozambique: Agricultural Sector Risk Assessment FIGURE B.11. GROUNDNUT YIELD FOR ALL REGIONS Groundnuts yield 6 5 Frequency density 4 Input Minimum 0.03106 Maximum 0.69388 3 Mean 0.21105 Std dev 0.11769 Values 60 2 Median = 0.19087 1 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Yield (tons per hectare) In the absence of a specific sowing calendar for each crop, TABLE B.3. DETERMINATION COEFFICIENT the same general sowing calendar used for maize was (LINEAR REGRESSION MODEL, used for groundnuts. Cumulative rainfall was calculated RAINFALL, AND GROUNDNUT for each weather station for each stage in the calendar in YIELD) order to determine the relationship between rainfall and Determination Coefficient yield. Simple linear regressions were run per region in Sowing Midseason Harvest order to determine the relationship of yield and rainfall. Province (%) (%) (%) Table B.3 shows the determination coefficient (R2) of the linear regression models applied by region. Cabo Delgado 39 0 10 Niassa 40 29 49 As table B.3 indicates, there are four provinces for which Nampula 12 6 19 rain during the sowing season is significant to explain Tete 15 43 17 variation in yield: Cabo Delgado, Niassa, Manica, and Zambezia 9 13 6 Inhambane all have a significant correlation. Manica 50 35 1 Sofala 6 6 7 Inhambane 45 70 26 Gaza 34 2 14 Maputo 2 25 66 Risk Prioritization 59 APPENDIX C CLIMATE CHANGE IMPACT ASSESSMENT ON AGRICULTURE IN MOZAMBIQUE: LITERATURE REVIEW INTRODUCTION Agriculture is highly vulnerable to climate change in Mozambique, and the effects are heterogeneous based on model assumptions and across regions, socioeconomic groups, and crops and livestock. There are direct impacts, such as changes in crop yields attributable to precipitation changes, and indirect impacts, such as rising food prices because of production changes and land tenure conflict stemming from shifting agro-climatic zones. If climate change is left unaddressed, then progress in agricultural development, food security, and poverty alleviation in general may be reversed. In the Mapping the Impacts of Climate Change index under “Agricultural Productivity Loss,” the Center for Global Development ranks Mozambique 85th out of 233 coun- tries globally for “direct risks” attributable to “physical climate impacts” and 28 out of 233 for “overall vulnerability” attributable to “physical impacts adjusted for coping ability” (Wheeler 2011). As agriculture accounts for 31.8 percent of the GDP and 81 percent of the labor force are involved in agriculture, there is great potential for wide- spread impact (CIA 2013). Varying impacts, in combination with numerous approaches to impact studies, make it difficult to generalize the potential effects of climate change on agriculture in Mozambique. This appendix discusses possible outcomes in the context of agricultural productivity and yield. PRINCIPAL FINDINGS » The increase in the likelihood of extreme events caused by climate change, as opposed to changes in average temperature or precipitation, may pose the great- est threat to agriculture in Mozambique. This includes flooding, drought, and tropical cyclones. Risk management therefore becomes increasingly important. Risk Prioritization 61 » Mozambique is highly vulnerable to climate related to climate change—the Climate Resilience Pilot Pro- change because of its geography, in particular, its gramme and the Global Facility for Disaster Risk Reduction. long coastline. » Temperature projections vary in various models METHODOLOGIES and scenarios, but generally Mozambique expects to see a rise of 1°C–2.5°C by midcentury and an AND TEMPERATURE/ increase of 1.4°C–4.6°C by late century.16 PRECIPITATION PROJECTIONS » Projections on precipitation vary from both positive In 2007, the Mozambican Initial Communication to to negative changes, but increases in the proportion the United Nations Framework Convention on Climate of rain that falls during the rainy period may occur. Change (UNFCCC) found that with a doubling of CO2 » Crop yields and land suitability: the mean air temperature would increase 1.8°C–3.2°C, – With some variations, generally there will be rainfall would be reduced by 2–9 percent, solar radiation no significant change in areas suitable for crops would increase by 2–3 percent, and evapotranspiration (cassava, maize, soybeans, sorghum, groundnuts, would increase by 9–13 percent (MICOA 2003). and cotton). – Likewise, the average change in yields in crops The UNDP Climate Change Country Profile for Mozam- is projected to change in small increments, but bique from 2008 notes projected increases in temperature generally will decrease slightly (cassava, sor- of 1.0°C to 2.8°C by the 2060s, and 1.4°C to 4.6°C by the ghum, soybeans, sweet potatoes and yams, 2090s (McSweeny, New, and Lizcano 2008). Two notable maize, groundnuts, millet, and potatoes). characteristics will be warming occurring more rapidly in the interior and high numbers of “hot” days and nights.17 BRIEF HISTORY OF Using data from the National Institute of Meteorology of CLIMATE CHANGE IMPACT Mozambique (INAM), the 2009 INGC Climate Change ASSESSMENTS Report ran seven general circulation models18 forced with The government of Mozambique has been very involved the SRES A2 emissions scenario, focusing on 2046–65 (mid- in climate change adaptation initiatives and disaster risk century) and 2080–2100 (late century) to project downscaled reduction strategies, including the 2007 National Adapta- future climate scenarios. The seven Global Climate Models tion Programme of Action (written following the United (GCMs) were also used in the IPCC 4th Assessment in 2007. Nations Framework Convention on Climate Change) and The models incorporated the Geospatial Stream Flow Model the Mozambican Institute for Disaster Management’s (GeoSFM) to determine water resource conditions, and a Climate Change program. dynamic and automated land evaluations system (ALES) to determine land utilization. ALES allows for the simulation DANIDA and the United Nations Development Program of crop yields/performance under different management (UNDP) funded the 2009 INGC Climate Change Report, levels, and for the purposes of Mozambique associated yields which was coordinated by the Mozambican Institute for with “prevailing smallholder traditional low-input farming Disaster Management and focused on cropland “suitabil- systems and potential yields corresponding to limitation free ity.” In October 2012, the INGC published another study highly managed commercial crop production systems.” To (Phase II) on climate change digging further into impacts on look specifically at land suitability for six major crops, the agriculture by looking specifically at production and yields. report used three GCMs (Geophysical Fluid Dynamics Lab- oratory [GFDL] as the dry scenario, Pierre Simon Laplace In 2010, the World Bank published a report, Economics of Global Institute of Science of the Environment [IPSL] as Adaptation to Climate Change: Mozambique (EACC). Currently, the World Bank has two main activities in Mozambique 17 “Hot” days projected to increase by 17–35 percent by the 2060s and 20–53 percent by the 2090s. 18 The seven different GCMs: CGCM3.1 (T63), CNRM-CM3, CSIRO-Mk3.0, 16 IFPRI, UNDP. ECHAM5/MPI-OM, GFDLCM2.1, GISS-ER, and IPSL-CM4 62 Mozambique: Agricultural Sector Risk Assessment FIGURE C.1. ANNUAL CYCLE OF RAINFALL, MAXIMUM TEMPERATURE, POTENTIAL EVAPOTRANSPIRATION, AND POTENTIAL MOISTURE INDEX (a) Rainfall (b) Maximum temperature North South North South 7 7 2 2 6 6 1 1 5 5 4 4 0 0 3 3 2 2 –1 –1 Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Month Month Month Month Central Coastal Central Coastal 7 7 2 2 6 6 1 1 5 5 4 4 0 0 3 3 2 2 –1 –1 Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Month Month Month Month (c) PET (d) PMI North South North South 2 2 1.5 1.5 1 1 1 1 0 0 0.5 0.5 –1 –1 Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Month Month Month Month Central Coastal Central Coastal 1.5 1.5 2 2 1 1 1 1 0 0 0.5 0.5 –1 –1 Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Month Month Month Month Source: INGC, 2009b. Figure 5: Changes in the annual cycle of a) Rainfall (mm day-1); b) Maximum temperature (°C); c) potential evapotranspira- tion (PET) (mm day-1) and d) Potential Moisture Index (PMI) (Rainfall – 0.5*PET) (mm day-1) simulated by seven GCMs for the North, Central, Southern and Coastal zones. Green shading indicated the range (olive line the median) for the 2046–2065 period, blue shading the range (blue line the median) change for the 2080–2100 period (Todross 2009). wet, and ECHAM as intermediate) (vanLogchem and Britos increase in maximum temperatures are 2.5°C–3°C. 2009). FAO’s CropWat model was used. Other observations from the INGC include that during the September, October, and November period, similar By midcentury, the models project increases in rainfall minimum temperature increases are projected for the during two periods: (1) December, January, and February; Limpopo and Zambezi valleys and seasonal variability and (2) March, April, and May. Higher increases in rainfall decreases for the maximum temperature in the north. and seasonal variability are expected along the coast and During the March, April, and May period and the June, in the south. Evotranspiration increases are greater than July, and August period, seasonal variability in maximum rainfall increases during two periods as well: (1) June, July, and minimum temperature increases in the north. and August; and (2) September, October, and November. Averaged across the country, a 10–25 percent increase in As the latter part of the 21st century approaches, the rainfall is expected (van Logchem and Britos 2009). projections predict increases in temperature in the center of Mozambique of up to 5°C–6°C during the September, Also by midcentury, higher temperatures are projected, October, and November period (van Logchem and Britos particularly in the inland areas during the September, 2009). The annual cycles are represented graphically in October, November period. The median estimates of figure C.1. Risk Prioritization 63 TABLE C.1. PROJECTED CHANGES FOR 2046–65 IN AVERAGE TEMPERATURES DURING THE GROWING SEASON, CROP YIELDS UNDER RAIN-FED CONDITIONS, AND RAINFALL DURING THE CROP GROWING SEASON Changes in Temperature Changes in Yield Changes in Rainfall Median Change in Median Change in Median Change in (past) Future (past) Future (past) Future Crop °C DC % mm mm % mm mm % Cassava 23.8 2.0 8.5 0.397 –0.02 –4.2 633.7 –17.3 –2.7 Cotton 24.1 2.1 8.5 0.517 –0.02 –2.9 610.0 –20.0 –3.3 Groundnut 24.5 2.1 8.5 0.599 –0.03 –4.6 487.9 –5.1 –1.1 Maize 24.5 2.1 8.5 0.373 –0.04 –11.1 454.2 –5.8 –1.3 Sorghum 24.6 2.1 8.5 0.572 –0.02 –3.5 438.9 –3.9 –0.9 Soybeans 24.6 2.1 8.4 0.217 –0.03 –6.4 377.4 –4.5 –1.2 Source: Brito and Homan 2012. Finally, the INGC Climate Change Report (INGC 2009b) of increased temperatures, changes in rain, and increased suggests that there may be an increase in frequency and concentrations of carbon dioxide and ozone on six main intensity of cyclones. Under the INGC report, there are crops (cotton, groundnuts, cassava, sorghum, maize, soy) two general scenarios for sea level by 2100: (1) low sea level (see table C.1). The report ran seven general circulation rise of 30 cm, and (2) high sea level rise of 500 cm. Under models to project temperature and rainfall data, and then the low scenario, tropical cyclones are the principal threat, used CliCrop to estimate yields based on soil humidity along with coastal erosion and setback (approximately 30 using a daily diary. Data came from 47 meteorological m). In the high scenario, the coast and low-lying areas stations (Brito and Homan 2012). may be permanently flooded, with setback at about 500 m (devastating impact) (van Logchem and Britos 2009). International Food Policy and Research Institute’s (IFPRI) study is based on the four downscaled global cli- The World Bank Mozambique EACC derived climate mate models from the IPCC AR4 (CNRM, ECHAM, outcomes from four general circulation models: global CSIRO, and MIROC). Based on these models, the wet, global dry, Mozambique wet, and Mozambique dry. IFPRI study uses the Decision Support System for Agro- These scenarios were used to estimate changes in yield for technology Transfer (DSSAT) crop modeling software rain-fed and irrigated crops, alongside demand for irriga- projections for crop yields, comparing yield projections tion in six cash and eight food crops. This study was meant for 2050 against real 2000 yields. The CSIRO model to supplement the INGC, and all four scenarios were also projects little change in precipitation (although a slight used by the INGC (World Bank 2010). The World Bank decrease in the eastern part of Inhambane province, used CliCrop, a generic crop model,19 to calculate the and an increase in part of Tete province), and projects impact of changes in precipitation and increased CO2 on the smallest increase in temperature, 1°C–1.5°C gener- crop yields and irrigation demand. ally, but almost 2°C increases in parts of the south. The MIROC model projected little change in precipitation In the 2012 Responding to Climate Change in Mozambique: in the southern regions and most of the coast, but away Theme 6: Agriculture, the INGC sought to quantify the effects from the coast in the north and northwestern regions. it projected increases in rainfall (occasionally exceeding 200 mm). MIROC projected warmer temperatures than 19 CliCrop was developed jointly by the World Bank and the INGC to improve yield predictions and fix problems with monthly water estimations in previous CSIRO, with increases of 2°C–2.5°C (particularly in the models. northeast) (IFPRI 2012). 64 Mozambique: Agricultural Sector Risk Assessment GENERAL FINDINGS FIGURE C.2. CLIMATE CHANGE EFFECTS ON YIELD FOR According to the 2009 INGC Climate Change Report, Mozambique is considered particularly vulnerable to ALL MAJOR CROPS South climate change because of geographic factors. These Central Moz. wet geographic factors include a 2,700 km coastline, the con- North vergence of multiple international rivers destined for the Indian Ocean, and large tracts of land below sea level. Moz. dry Other factors that increase Mozambique’s vulnerability to climate change include already high temperatures, aridity, infertile soil, endemic disease, lack of a communications Global wet infrastructure, high illiteracy levels, high population growth rate, high absolute poverty rates, and dependence on natu- ral resources requiring “predictable rain” (INGC n.d.). Global dry The central zone will be hit hardest, with increased drought –6% –4% –2% 0% 2% 4% risk during the October, November, and December period, Source: World Bank 2010. resulting in the highest likelihood of crop failure across the Note: The crops modeled are cassava, sorghum, soybeans, sweet pota- toes and yams, wheat, groundnuts, maize, millet, and potatoes. country according to the INGC assessment, particularly low-lying regions, such as the Zambezi valley, that already have high temperatures (van Logchem and Britos 2009). soybeans, sorghum, groundnuts, and cotton). Area may Generally, the increased risk of drought centers on Cahora increase in the center and the north, and decreases will Bassa in Mozambique, but also covers most of Zimbabwe occur primarily where there are already issues of irreg- and Zambia. It is important to consider neighboring coun- ular and extreme climatic events (including the mixed tries as well, because increased drought risk in Zimbabwe arid-semiarid systems in the Gaza, semiarid systems in during January, February, and March will have significant northern Inhambane and south of Tete, coastal zones implications for trans-boundary water usage and agricul- in the south, southern central zones, and drier areas of tural trade (van Logchem and Britos 2009). major river systems such as the Save, Limpopo, and Zam- beze) (van Logchem and Britos 2009). In the southern zone, increased temperature leads to a 10 percent increase in evapotranspiration and higher crop River flow is not projected to decrease significantly in the water requirements. But five of the seven models show north, and there is irrigation potential. The INGC pro- that the risk of drought, damage to crops, and crop failure jects that “increases in yields attainable with the intensifi- in southern Mozambique will be unchanged. In the north, cation of agriculture and technological development are all seven models project no change in drought risk or crop higher than the expected decreases in yields caused by cli- failure during the January, February, and March period. mate change” (van Logchem and Britos 2009). In the October, November, and December period, projec- tions suggest low risk for drought, but possible increased or The World Bank EACC Mozambique projected a 2–4 decreased crop failures (although along the coast it is more percent decrease in major crops by 2050, particularly certain that there may be mild reductions in the frequency in the central region. This loss will be exacerbated by of crop failure) (van Logchem and Britos 2009). The pro- increased frequency of flooding (particularly along rural jections make it difficult to generalize about changes in roads) and agricultural practices such as “slash and burn.” flooding, but generally, increased flood peaks will occur in The EACC predicts a 4.5–9.8 percent loss of agricultural small watersheds wherever storms make landfall. GDP (World Bank 2010a). With some variations, generally there will be no signifi- In the 2012 INGC report, yields are examined across the cant change in areas suitable for crops (cassava, maize, six main crops (see table C.1). According to their models, Risk Prioritization 65 TABLE C.2. CLIMATE CHANGE EFFECTS (RAIN, TEMPERATURE, CO2, AND O2 CHANGES) ON CROP YIELDING IN THE 2046–65 PERIOD Rain and Crop Temperature Temperatura Co2 O3 Total Cotton –2.9% –11.0% +27.0% –37.0% –23.9% Groundnut –4.6% –11.0% +10.0% –14.0% –19.6% Cassava –4.2% +6.0% +10.0% –14.0% –2.2% Sorghum –3.5% –11.0% +7.0% –9.0% –16.5% Maize –11.1% –11.0% +7.0% –9.0% –24.1% Soy –6.4% –11.0% +20.0% –28.0% –25.4% Average –5.5% –8.2% +13.5% –18.5% –18.6% Source: Brito and Homan 2012. the most affected crop appears to be maize (with an overall similar charts in subsequent sections (van Logchem and 24.1 percent decrease in yield in the 2046–65 period), and Britos 2009). the least affected crop as cassava (with an overall decrease of 2.2 percent) (Brito and Homan 2012). The EACC Mozambique showed regional variability in change to maize yields, but when averaged across sce- The 2012 INGC report went a step further and used a narios and regions saw a very small projected change layered approach to assess the effect of a rise in tempera- of –0.66 percent. There is a stark contrast in suitability ture, background ozone, and atmospheric carbon diox- between the north and the south. ide. Even using a layered approach, ground level ozone and atmospheric carbon dioxide have an interaction (see table C.3). The model used by INGC shows that a pro- CASSAVA jected 15 parts per billion (ppb) increase in background The EACC Mozambique projects an ambiguous, but ozone by midcentury virtually cancels out the increase primarily a decrease, change in yields for cassava when in yield seen from a 178 parts per million (ppm) atmos- averaged across scenarios and regions (–3.65 percent). pheric carbon dioxide. Yields performed similarly under the 2012 INGC, showing decreases in a band across the center in Tete, MAIZE and Sofala. The IFPRI impact assessment results varied regionally, The 2009 INGC assessment found “no significant change” but CSIRO and MIROC generally showed 5–25 percent in land suitability for cassava. This is primarily because of yield gain for rain-fed maize in the north, and differed cassava’s relative drought and poor soil fertility tolerance. in the south by 2050. The CSIRO model predicted both Interestingly, the greatest area of change away from land yield gains and losses of over 25 percent. MIROC, con- suitability is in the north. There is also an increase across versely, showed significant yield increases, most in excess parts of the center, principally Sofala. of 25 percent. Both the CSIRO and MIROC show sub- stantial yield gain across southern Gaza, along the conflu- ence of the Limpoto and Olifants rivers (see figure C.3). SOY The EACC Mozambique showed little change in soybean The INGC assessment, conversely, found a relatively con- yields, –1.28 percent, when averaged across scenarios and stant, “no significant change” in suitable land for maize regions. According to the INGC assessment, the primary in 70 percent of the center zone, 69 percent in the south, finding was “no significant change” for soy land suitabil- and 76 percent of the coastal area. Figure C.4 is taken ity except in the south (inland) (van Logchem and Britos from the 2009 INGC Climate Change Report, as are 2009). Yields similarly remained stable in the 2012 INGC 66 Mozambique: Agricultural Sector Risk Assessment Risk Prioritization TABLE C.3. IMPACT OF MIDCENTURY CLIMATE CHANGE ON CROP YIELD IN MOZAMBIQUE: EFFECT OF RISE IN TEMPERATURE, BACKGROUND OZONE, AND ATMOSPHERIC CO2, A LAYERED APPROACH Surface Air Back Ground Ozone Atmospheric CO2 +178 Temperature +15 a 30 ppb ppm (550 ppm) +1.8 a 2.4 °C Sum Sensitivity +15 +23 +30 +15 +23 +30 Factor for ppb ppb ppb ppb ppb ppb +1.8 +2.1 +2.4 Min Med Max Crop Ozone O3 O3 O3 O3 O3 O3 °C °C °C Total Total Total Cotton 1.6 –24% –37% –48% +21% +27% +33 –9% –11% –13% –12% –21% –28% Soybeans 1.2 –18% –28% –36% +16% +20% +25 –9% –11% –13% –11% –19% –24% Groundnuts 0.6 –9% –14% –18% +8% +10% +12 –9% –10% –12% –10% –14% –18% Cassava 0.6 –9% –14% –18% +8% +10% +12 +5% +6% +7% +4% +2% +1% Maize 0.4 –6% –9% –12% +5% +7% +8 +10% –13% –15% –11% –15% –19% Sorghum 0.4 –6% –9% –12% +5% +7% +8% –8% –10% –11% –9% –12% –15% 67 FIGURE C.3. CHANGE IN YIELD WITH assessment, with the exception of decreases across the CLIMATE CHANGE: center, particularly in Tete. RAIN-FED MAIZE Land suitability appears to increase slightly in the south with areas of “significant increase risk” in Zambezia (INGC 2009). SORGHUM The 2009 INGC assessment also found little change in land suitability for sorghum, although in a small section in the northern inland near Lake Nyasa, a sharp increase in risk is expected. Similarly, the EACC Mozambique found little change in yield across scenarios and regions. Averaged across scenarios and regions, the EACC Mozambique predicted a 0.51 percent decrease in yield. Very small changes in yield were noted by the 2012 INGC, concentrated in Tete. These changes in yields are among the lowest predicted for various crops. OTHER CROPS The EEAC Mozambique also projected changes in yield for sweet potatoes and yams, wheat, groundnuts, millet, FIGURE C.4. MAPS OF LAND SUITABILITY AND HOTSPOTS RESULTING FROM CLIMATE CHANGE, FOR MAIZE Source: IIAM 2008. 68 Mozambique: Agricultural Sector Risk Assessment and potatoes. The results are displayed in table C.4 from FIGURE C.5. EXPECTED CHANGES IN THE the report. FUTURE (2046–65) FOR MAIZE (EXPRESSED IN KG/HA) UNDER CONCLUSION RAIN-FED AGRICULTURE In Mozambique, negative impacts on agriculture from BASED ON THE MEDIAN OF climate change will primarily be caused by the increased ALL SEVEN GCMS likelihood of extreme events such as cyclones and flood- ing. However, outside of the expected increase in extreme events, agriculture in Mozambique will see little change in land suitability and yield (see figures C.5 through C.13). Risk management then becomes increasingly important as risk incidents increase across the country in a vari- ety of manifestations. Among other tools, information systems (such as early warning systems, the disaster risk database) are paramount to mitigating these risks. Other risk management instruments noted throughout this agriculture risk assessment will also be crucial to protect producers and consumers alike as the uncertainty of cli- mate change unfolds. Source: Brito and Homan, 2012. FIGURE C.6. CHANGE IN LAND SUITABILITY PER CROP RESULTING FROM CLIMATE CHANGE Risk Prioritization 69 70 TABLE C.4. AVERAGE OF THE PERCENTAGE CHANGE IN YIELD FOR MOZAMBIQUE North Central South csiro30_ ncarc_ ukmo1_ Ipsl_ csiro30_ ncarc_ ukmo1_ Ipsl_ csiro30_ ncarc_ ukmo1_ Ipsl_ a2 a2 a1b a2 a2 a2 a1b a2 a2 a2 a1b a2 Crop Global Global Moz. Moz. Global Global Moz. Moz. Global Global Moz. Moz. Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Average Cassava –3.44% 2.01% –6.51% –0.09% –6.24% –4.75% –6.21% –3.10% –3.27% –9.36% –3.20% 0.36% –3,65% Sorghum –0.99% 0.66% –6.08% –1.59% 0.25% –0.74% –0.66% –1.97% 0.55% –1.57% 1.33% – 0.68% –0.51% Soybeans –0.40% 0.06% –2.58% –1.00% –0.52% –3.63% –5.81% –1.46% –1.32% –6.06% 5.91% 1.47% –1,28% Sweet Potatoes 0.29% 0.58% –5.70% –1.39% –1.45% –4.05% –6.70% –5.70% –0.32% –3.69% –4.45% –0.63% –2.77% and Yams Wheat –2.18% –2.31% –5.11% –3.20% –0.93% –4.33% –3.03% –2.93% –1.64% 5.11% 2.48% 0.20% –2.10% Groundnuts 0.71% 1.65% –3.23% –1.84% 1.17% –0.08% –4.73% –3.66% –1.66% –2.90% –3.72% 0.58% –1.48% Maize –1.32% 1.27% –1.87% –2.92% 0.64% 0.34% –2.59% –3.04% 6.37% 3.49% –3.95% –4.36% –0.66% Millet –6.82% 10.03% –17.38% –8.40% –1.35% –3.45% –1.78% –6.29% –2.78% –10.07% 7.85% 0.29% –3,34% Potatoes –0.36% 4.15% –5.87% –1.10% –3.20% –1.15% –8.05% –1.46% –4.09% –6.78% –3.10% 1.29% –2.69% Average –1.61% 2.01% –5.44% –2.07% –1.29% –2.43% –4.40% –3.29% –0.91% –4.67% –0.09% –0.45% –2,05% Source: World Bank 2010. Mozambique: Agricultural Sector Risk Assessment FIGURE C.7. A. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR CASSAVA IN PERCENTAGE OF PRESENT YIELDS. B. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR CASSAVA IN KG/HA Source: Brito and Homan, 2012. FIGURE C.8. MAPS OF LAND SUITABILITY AND HOTSPOTS RESULTING FROM CLIMATE CHANGE, FOR SOY Source: IIAM 2008. Risk Prioritization 71 FIGURE C.9. A. PROJECTED CHANGES IN THE FUTURE (2046–65) OF SOYBEAN YIELDS IN PERCENTAGE OF PRESENT YIELDS. B. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR SOYBEANS IN KG/HA Source: Brito and Homan, 2012. FIGURE C.10. MAPS OF LAND SUITABILITY AND HOTSPOTS RESULTING FROM CLIMATE CHANGE, FOR SORGHUM Source: IIAM 2008. 72 Mozambique: Agricultural Sector Risk Assessment FIGURE C.11. A. PROJECTED CHANGES IN FUTURE (2046–65) FOR SORGHUM (MEDIAN OF ALL SEVEN GCMS), EXPRESSED IN PERCENTAGE OF PRESENT YIELDS. B. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR SORGHUM (MEDIAN OF ALL SEVEN GCMS), EXPRESSED IN KG/HA Source: Brito and Homan, 2012. have suggested increased foliage growth, potentially a pos- LIMITATIONS itive impact on livestock, and there are also implications in The glaring limitation of a climate change impact assess- heat fluctuations for male livestock fertility. ment in Mozambique is the uncertainty of extreme events (such as drought, flood, and cyclones) and their depth of Other gaps include information on the impact of climate impact on the agricultural sector. change for some of the primary cash crops such as cot- ton and sugar cane. The 2009 INGC assessment briefly Additionally, climate change impact on livestock is one mentions cotton in the methodology section, but does not avenue for further research. There are several studies that elaborate on its findings. Risk Prioritization 73 FIGURE C.12. A. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR COTTON (MEDIAN OF ALL SEVEN GCMS), EXPRESSED IN PERCENTAGE OF PRESENT YIELDS. B. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR COTTON (MEDIAN OF ALL SEVEN GCMS), EXPRESSED IN KG/HA Source: Brito and Homan 2012. FIGURE C.13. A. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR GROUNDNUTS IN PERCENTAGE OF PRESENT YIELDS. B. PROJECTED CHANGES IN THE FUTURE (2046–65) FOR GROUNDNUTS, EXPRESSED IN KG/HA Source: Brito and Homan 2012. 74 Mozambique: Agricultural Sector Risk Assessment APPENDIX D VULNERABILITY ANALYSIS The World Bank defines vulnerability as exposure to uninsured risk, leading to a socially unacceptable level of well-being. An individual or household is vulnerable if it lacks the capacity and/or resources to deal with a realized risk. It is generally accepted that in low-income countries, rural populations are both poor and vulnerable, and that primary risks to these populations may include climate and market shocks (Sarris and Karfakis 2006). Vulnerability is a useful lens through which to view shocks, because it allows for determination of impacts on populations and who will be most affected. Vulnerability is discussed here particularly in the context of food security. SETSAN performs identification of “vulnerability to external shocks,” measured through food security (see tables D.1 and D.2). As of 2009, other leading government bodies mandated with addressing the needs of vulnerable groups were the Ministry of Women and Social Action (MMAS) and a subordinate institution, the National insti- tute for Social Action (INAS) (Waterhouse 2009). DEFINITIONS SETSAN views vulnerability as “associated with exposure to risks and [it] determines the susceptibility of people, places, or infrastructures to a particular natural disaster.” The Ministry of Planning and Finance (MPF) defined vulnerability as, a “lack of defense against adversity (including) . . . exposure to external shocks, tension and risks, and lack of internal defense, of means to compete without suffering serious losses” (MPF 2000; Waterhouse 2009). RECENT GENERAL TRENDS IN VULNERABILITY » Sustained high levels of food insecurity » “Deepening” of the HIV/AIDS crisis » More frequent and severe climate-related events such as drought, floods, and cyclones » “Increasing feminization of chronic rural poverty” (Waterhouse 2009) Risk Prioritization 75 FIGURE D.1. VULNERABLE GROUPS WITHIN MOZAMBIQUE • Households perpetually live on the edge Marginal households (very • Low access to all resources poor) • Concentrations: Cabo Delgado, Nampula, Inhambane Low income laborers (poor • Concentrations in northern provinces: Nampula, Zambezia, households) Tete, and Inhambane • Lower levels concentrated in Tete, Cabo Delgado, Niassa, and Nampula Households with low levels of • (Using a composite index of well-being based in the analysis of the "well-being" five capitals by FEWSNET) Households with lower dietary diversity ("very inadequate • Concentrations in: Tete, Manica, Inhambane diet") Source: FEWSNET. FIGURE D.2. VULNERABILITY CHARACTERISTICS Provinces where the Vulnerability Characteristic Vulnerability Characteristics Is More Frequent Highest % female-headed households Gaza (47%) Cabo Delgado (36%) Inhambane (37%) Highest % elderly-headed households Cabo Delgado (3l%) Maputo (30%) Highest % of never enrolled children Girls in Niassa (l8%) Boys and girls in Nampula (15%) Lowest enrolment level for orphans Boys in Niassa (61%) Boys in Maputo (65%) Boys in Nampula (63%) Highest level of never enrolled orphans Boys in Niassa (23%) Nampula (23%) Highest & of HHs with Chronically ill members Sofala (12%) Highest % of HHs with disabled Inhambane (29%) Highest % of HHs hosting orphans Gaza (26%) Zambezia (22%) Highest % of HHs with recent death of family Cabo Delgado (11%) Sofala (11%) member Note: Niassa and Nampula seem to have the major shortcomings in terms on education as they particularly low enrolment levels for both children and orphans. In terms of demographic characteristics of the households, Cabo Delgado shows serious results on the percentage of female and elderly headed households and households with a recent death. households hosting orphans are considered particularly VULNERABLE GROUPS vulnerable to shocks. Figure D.2 details provinces with high AND CHARACTERISTICS levels of vulnerable populations as assessed by the SET- Certain types of households, livelihood activities, and SAN Vulnerability Assessment Group (SETSAN 2010). populations are more vulnerable to shocks than others. Figure D.1 identifies four types of vulnerable groups and Figure D.3 highlights provinces with poor infrastructure, lists regions in Mozambique with particularly high con- characteristics that will in effect influence levels of vulner- centrations of vulnerable households. ability. Sanitation, particularly access to water, is an issue, deepening vulnerability in crises (SETSAN 2010). This At the household level, certain characteristics, apart from issue is particularly salient for food crop farmers and fish- income and well-being, are associated with high levels of ermen, who are less likely than other types of workers to vulnerability. For instance, female-headed households and have access to safe water and sanitation. 76 Mozambique: Agricultural Sector Risk Assessment Risk Prioritization TABLE D.1. UNDERLYING FACTORS OF FOOD SECURITY IN MOZAMBIQUE Description Vulnerable Groups Vulnerable Period Vulnerable Areas Climate Recurrent droughts (increasing Drought: Rain-fed agriculture Drought: Southern region, frequency), major devastation. dependent households semiarid central regions. Results in almost complete failure of Cyclones: Coastal districts of nontolerant crops (maize). Inhambane, Nampula, Sofala, Increased frequency and magnitude of and Zambezia. cyclones. Floods: Central and southern Floods result in temporary deficient food rivers. access. Poverty and Poverty weakens the ability of households 50% of households engage in food Rural and peri-urban residents. fragile livelihoods to mitigate recurrent shocks. production as their main economic activity. Vulnerable to production shocks, depending on diversification of other economic activities. Food reserves Normal year → households produce Households with no other sources of Years with low The southern regions. enough food for 2–5 months postharvest food/income agricultural production to be self-sufficient, and then begin and the hungry period: purchasing. October–January. Food deficit years → intensive, unsustainable coping strategies. Markets Poor functioning markets, lack basic Populations in remote areas with a products. Major constraint for markets is lack of income options. the “lack of effective demand (poverty).” (continued) 77 78 TABLE D.1. UNDERLYING FACTORS OF FOOD SECURITY IN MOZAMBIQUE (continued) Description Vulnerable Groups Vulnerable Period Vulnerable Areas Limited options Remote markets have high food prices Poor households with limited income South and central because of transportation costs, limiting options, reliance on purchased food food-deficit areas. access. for consumption. Chronic Stunting rates are high. Nampula and Tete have the malnutrition highest proportions of stunted children (63 and 51 percent). Gaza and Inhambane have the lowest (31 and 32 percent). Limited public Low access to basic services. Health Northern provinces. service delivery problems may compound food insecurity (health, water, and conditions even in high-surplus areas. sanitation) Increase in Many households have lost an HIV/AIDS–affected groups. the effective economically active member. dependence ratio Government A decrease in government investment in Households dependent on the policies the agricultural sector. agricultural sector. Source: Mozambique, FEWSNET. Mozambique: Agricultural Sector Risk Assessment FIGURE D.3. PROVINCES WHERE VULNERABILITY CHARACTERISTIC IS MORE COMMON Provinces where the Vulnerability Characteristic Vulnerability Characteristics Is More Common Sofala (5l%) Highest % HHs with poor quality housing Inhambane (44%) Gaza (31%) Niassa (95%) Highest % of HHs with water from unimproved sources Nampula (94%) Cabo Delgado (93%) Niassa (6%) Lowest % of HHs treating water before drinking Tete (6%) Highest % of HHs with unsafe sanitation Nampula, in particular rural Nampula (100%) Tete (37%) Highest % of asset poor households Niassa (31%) Zambezia (32%) Source: SETSAN. TABLE D.2. MAJOR SHOCKS TO FOOD SECURITY IN MOZAMBIQUE Vulnerable Time Hazard/Shock Description/Impact Period Vulnerable Areas Drought High frequency, major devastation. Southern and central regions. Recurrent and severe droughts recently. Floods Impact is difficult to assess—potentially More than 50% of Mozambican destructive, positive externalities for flood- land is classified as part of recession agriculture. international river basins. Cyclones Households may lose houses, food reserves, November–April The majority of the coastal area, crops, and fruit trees, and often face acute (Coincides with the the most at risk between Pemba food shortages. (Despite the damage, may main agricultural and Angoche, and near Beira. bring needed rainfall.) season). Cassava brown streak Plant disease, affects cassava (food crop) The northern regions. Lumpy skin disease Bovine disease, limits the movement and sale The southern provinces. of animals, which reduces income and assets. Human diseases Cholera, other diarrheal diseases, malaria, Illnesses such as HIV/AIDs. Because of HIV/AIDS, malaria increase child-headed households, elderly-headed during rainy seasons. households, and orphans are increasing. Two impacts on food security: (1) loss of an economically active adult and additional children to be fed without a caretaker and (2) the malnourishment trap (sick → less work → less money/food → sickness continues). Earthquake Relatively new potential shock. The central and northern provinces, with close proximity to the southern end of the East Africa Rift System. Source: Mozambique, FEWSNET. Risk Prioritization 79 APPENDIX E COMMODITY RISK PROFILES Risk Prioritization 81 POULTRY The poultry sector in Mozambique represents an estimated interviews, in Chokwe, Manica province, during the 4.7 percent of agricultural GDP20 and is the primary source recent floods of 2013. of meat and the most common form of livestock for rural households. In 2008, 58.8 percent of small and medium PESTS AND DISEASES producers surveyed had chickens, but only 4.4 percent of Infectious bursal disease, avian salmonellosis and pas- them were vaccinated.21 teurelosis, parasitosis, and Newcastle disease (ND) all jeopardize poultry health. Newcastle disease ranks first PRODUCTION RISKS as the most devastating disease for village poultry, often killing from 50 to 100 percent of chickens in a flock.22 DROUGHT Major losses were recorded in 2005/06 as a result of Drought not only affects poultry directly but also indi- Newcastle disease (ND),23 which is transmitted and culti- rectly through the availability of grain feed. In particular, vated by chickens ingesting contaminated feed, water, and the lack of rainfall limits harvests and redirects potential feces. Although poultry may be vaccinated against ND, feed to addressable markets and not poultry. Whereas the presence of any of the above diseases or predatory drought per se does not lead to fatalities in poultry, high threats—birds, reptiles, mammals—results in the total loss temperature does. Interviews with farmer groups also of the animals and any future income. According to the confirmed that high temperatures have at times led to 2008 TIA, of 17,220,796 chickens that small and medium a total loss of chickens. A secondary effect of drought is producers surveyed said they had, 12,257,403 were lost to selling to neighboring provinces, which not only affects disease—71.2 percent. prices in the province from which they are being sold, but also increases the likelihood of the transmission of As avian flu affected the global sector, it was also devastat- diseases and pests. ingly felt in Mozambique. Farmer interviews confirmed that along with the birds that contracted the disease and FLOODING died, many potentially healthy birds had to be culled, Although an indirect effect, the type of storage of chickens under advisement from the National Department of Vet- in rural areas means that when flooding is high enough, erinary Services of the Ministry of Agriculture (MASA/ infrastructure, along with the chickens themselves, will DNSV) to prevent further transmission. In 2004, there be washed away. This occurred, as related during farmer was a nationally recorded outbreak of avian flu.24 FIGURE E.1. DOMESTIC PRICE VOLATILITY, POULTRY Producer price (USD/ton), chicken live weight 3,500 Producer price (USD/ton), chicken meat 3,000 USD/ton 2,500 2,000 1,500 1,000 500 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Source: FAO PriceSTAT. 22 Control of ND on Dropbox. 20 IFPRI document. 23 Report on Current Vulnerability in Mozambique, 2006. 21 TIA 2008. 24 Direccao Nacional de Pecuaria—Relatorio Annual 2004. 82 Mozambique: Agricultural Sector Risk Assessment and equipment) are imported to Mozambique and the MARKET RISKS traders are generally located only in the capital city, with INPUT VOLATILITY links to the provincial capitals. Traditionally, the raising of chickens is undertaken either with free-range scavenging or on a commercial scale. Although the more common free-range scaveng- INTERNATIONAL PRICE VOLATILITY ing requires minimal inputs, commercial production The global poultry market forecast and price index is requires access to shelter, feed, and medication, whether affected by three primary factors: wheat-feed prices, alter- antibiotics or vaccinations. Virtually all inputs (rations, native meat prices (that is, beef), and disease outbreaks. Each concentrates, medicines, vaccines, veterinary instruments, of these makes for volatility in the marketplace (figure E.1). Risk Prioritization 83 SORGHUM Sorghum, a major cereal grain, accounts for 2.56 per- grasshoppers, and locusts also constrain production. The cent of national agricultural GDP (Pauw et al. 2012) with most important threat however comes from weeds. In par- over 2.7 million hectares planted, and is generally used ticular Striga, a parasitic weed that attaches itself to the for household food security compared with maize and sorghum roots from which moisture and nutrient require- rice. Because of a lack of seed and poor distribution of ments are drawn and inhibits plant growth, reduces yields improved varieties, farmers continue to use local varieties, and, in severe cases, causes plant death. which have low productivity (0.2–0.6 ton/ha25). In terms of actual observed losses, in 1998/99 locusts attacked sorghum, among other crops, in the coastal PRODUCTION RISKS regions of Nampula and Cabo Delgado.27 DROUGHT Sorghum thrives in arid and semiarid conditions and is con- sidered relatively drought resistant.26 It is commonly grown MARKET RISKS on upland fields and is more drought tolerant than is maize. DOMESTIC PRICE VOLATILITY During the planting season and before the rainy season at Nearly all sorghum is produced by small farmers, mainly about November, it is often intercropped with maize and for subsistence. Few farmers sell any significant quantities is thus a way for farmers to ensure they procure some pro- of sorghum to market on a regular basis. Seasonal price duction from a field in the event that weather and fertility variations and price collapses in good years pose impor- conditions are inadequate for successful maize production. tant risks to farmers as they typically sell all their harvest at the same period, driving down prices (see figure E.2). PESTS AND DISEASES It should be noted that with the presence of growing The major diseases affecting sorghum are stalk borer ethanol demand, sorghum is slowly becoming an impor- complex, Busseola fusca, Sesamia calamistis, anthracnose, tant substitute for feed and human consumption and this charcoal rot, downy mildew, ergot, and leaf blight. Insect will likely have an impact on prices, both domestic and pests including borers, shoot fly, chafer grubs, armyworms, international. FIGURE E.2. DOMESTIC PRICE VOLATILITY, SORGHUM 250 Producer price, sorghum (USD/ton) 200 150 USD/ton 100 50 0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 25 http://intsormil.org/smimpacts/IIAM.pdf. Annual Report on the Agricultural Campaign 1998/99 and Evaluation of a 27 26 http://africaharvest.org/sorghum.php. Technical Mission Team Report. 84 Mozambique: Agricultural Sector Risk Assessment TOBACCO Tobacco is the most exported agricultural commodity but PESTS AND DISEASES it is produced by only 10.24 percent of farming house- Tobacco pests include beetles, crickets, aphids, wee- holds in Mozambique (over 73,725 hectares28) and it rep- vils, cutworms, and suckfly. One particular disease often resented just 1.17 percent of national agricultural GDP observed in Mozambique in tobacco is rattle virus, which (Pauw et al. 2012). The provinces of Niassa and Tete causes necrosis and leaf molting that stunts the cultivated account for 90 percent of this cultivated area. parts of the plant. Mozambique was the 16th largest exporter globally in 2010.29 Tobacco exports were valued at US$204 million in MARKET RISKS 2011, accounting for 6.2 percent of total exports.30 There DOMESTIC PRICE VOLATILITY are several large commercial operations with downstream Tobacco is a crop with established off-takers and whose integration with farmers along the value chain through prices are published well before the start of trading. Pri- contract farming. This leads to about 52 percent of the vate companies are promoting contract growers’ schemes total area planted by small farms.31 with smallholder (0.5 to 1 ha) producers. These producers receive credit for inputs (seeds, fertilizers, and pesticides) PRODUCTION RISKS that are then deducted at the end of the campaign (for marketing). Producers are organized into small groups DROUGHT, DELAY IN RAINS (6–15 members per group). Approximately 20–30 groups Tobacco is typically a transplanted crop, whose seedlings are then assisted by extension agents who work for private are cultivated for 60 days and then moved into fields. It companies. is during this phase in particular that the plant is vulner- able to the effects of drought. Typically, the crop requires average rainfall of 625–1,000 mm. For example, a delay INTERNATIONAL PRICE VOLATILITY in rains in 2008/09 during the seeding phase led to a sig- Given that tobacco is a primary cash crop for export, nificant drop in reported yield. domestic farmers are heavily influenced by international market movements. Real prices of tobacco have fallen in most countries, but much less rapidly than prices of other FLOODS agricultural crops and basic commodities.34 In the absence Conversely, in waterlogged soil, diffusion of gases is of volume sales growth, the tobacco industry relies on strongly inhibited and this adversely affects growing roots pricing strength and product mix improvements via inno- of the plant while also retaining the broken-down gases vation to keep profits rising. However, pricing strength is such as carbon dioxide that it produces. This can result threatened by issues such as illicit trade and tobacco con- in loose or dead roots, rot, and stunting.32 Yield is reduced trol measures such as plain packaging, which is causing by 23 percent, 43 percent, 76 percent, and 82 percent the commodification of cigarettes and eviscerating profit because of 24, 48, 72, and 96 hours of waterlogging, growth.35 From 1975 to 1998, world tobacco production respectively.33 has increased by nearly 60 percent. This increase, how- ever, is not evenly distributed across production countries; nearly all growth in production comes from developing 28 http://www.drumcommodities.com/assets/33/Tobacco_Project_Executive _Summary_April_2012.pdf. countries. 29 http://www.drumcommodities.com/assets/33/Tobacco_Project_Executive _Summary_April_2012.pdf. 30 Ibid. 31 Dropbox under Benfica name. 32 http://www.plantstress.com/articles/waterlogging_i/waterlog_i.htm. 34 http://www.fao.org/docrep/006/y4997e/y4997e0l.htm#bm21.1. 33 http://link.springer.com/article/10.1007%2Fs40502-013-0008-0. 35 http://www.researchandmarkets.com/research/393vsz/pricing_in_the. Risk Prioritization 85 In 2008/09, the sector also experienced what can best be ENABLING ENVIRONMENT described as execution risk—there was a delay because of All tobacco produced is exported to Europe and Asia after rains and a systemic delay in the distribution of inputs, processing (separation of the ribs of the leaf blade). In particularly fertilizer, because a commitment by a com- Mozambique, there are single concession areas where pri- pany in Nampula was not upheld.36 vate companies operate for a period of 10 years. Annually, the promoter company must submit a business plan to the Provincial Directorate of Agriculture of the province. 36 Balanco de PES 2009. 86 Mozambique: Agricultural Sector Risk Assessment VEGETABLES Vegetables are often included as a component of horticul- In January 2012, 0.8 percent of the planted area of tural commodities that also comprise fruits and ornamental maize, beans, and horticulture in Cabo Delgado, Zam- plants. The focus in Mozambique is on the more common, bezia, Sofala, Inhambane, Gaza, and Maputo were lost highly perishable vegetables, including tomatoes, squash, because of excessive rain and the tropical storms Dando peas, green beans, garlic, onions, cabbage, lettuce, Irish and Funso. potatoes, and peppers. Most cultivation occurs in Manica and Sofala provinces by smallholder farmers—accounting PESTS AND DISEASES for 99 percent of output—and is likely to remain primar- Pests and diseases are the major risk to vegetable produc- ily for domestic consumption as it makes up less than tion and can lead to rapid, systemic, and catastrophic 0.1 percent of export value.37 Of the 36 percent of farm- damage that can occur to the harvest. There are several ing households that produce vegetables, only 8 percent pest species that affect vegetables including fruit flies, is sold and taken to market.38 As calculated by IFPRI, in mites, bollworms, ants, aphids, cutworms, worms, leaf 2009, vegetables accounted for 8.21 percent of national miners, and grasshoppers.41 agricultural GDP share and from 5 percent to 33 percent in inter-regional agricultural GDP share. These pests and diseases cause significant damage by affecting the growth process, leading to stunted growth It is important to note that given the perishability of the in addition to breaching plant skin and rot. Given the produce, price volatility, and the high cost of necessary sensitivity of fresh vegetable produce, pests and diseases inputs given current scale, the poorest 60 percent of sell- need to be addressed immediately in order not to lose the ers account for only 7 percent of the total value of sales.39 entire production or to prevent the risk of spread of the contamination. PRODUCTION RISKS DROUGHT In the 2001/02 agricultural year, both nematodes and Most vegetable production takes place under irrigation birds affected tomato growing in the south, so much so during the drier winter season, when growing conditions that it was noted in reports that year.42 A significant drop are more favorable for these commodities. Normally, in production also occurred in 2009/10 linked to adverse 8–10 gallons of water per minute per acre are required climatic conditions, but particularly tomato virus.43 for horticultural crops.40 Fifty-two percent of horticul- ture producers state that they use manual or mechanical Farmer interviews suggested that the risks present in irrigation systems. This has led to a control of the dam- regard to the tomato virus could be easily overcome with age incurred because of drought. Hot, humid conditions, the correct application of pesticides and the like once they though, do stress most vegetables and irrigation cannot were identified. totally offset this. MARKET RISKS EXCESSIVE RAIN AND TROPICAL INPUT PRICE VOLATILITY STORMS The high cost of fertilizer and pesticides is a barrier for As with other commodities, excessive rain and tropical many farmers and is a risk that hedges how much cultiva- storms are always a risk in certain areas of Mozambique. tion is actually undertaken. The use of fertilizer inputs 37 Smallholders Involvement in Commercial Agriculture. 38 IIAM. 41 http://www.nri.org/projects/adappt/docs/Nyirenda.pdf. 39 Ibid. 42 RA DINA 2002. 40 http://aggie-horticulture.tamu.edu/archives/parsons/drought/cropmgmt.html. 43 Balanco de PES 2010. Risk Prioritization 87 can increase yields upward of 67 percent if prices stabilize DOMESTIC PRICE VOLATILITY at low levels. High price variability in domestic markets is the result of the local production and market dynamics of the crops. Risks associated with the enabling environment exist but Surplus production tends to enter the market at specific are generally less significant and mainly apply to a minor- times every year, accentuating the volatility of domestic ity of producers who specialize in production for market. prices. 88 Mozambique: Agricultural Sector Risk Assessment GROUNDNUTS In Mozambique, groundnuts are cultivated solely by small delays the start of the period of rapid pod growth by about farming households for several uses but are most commonly 15 days and hence extends the time required to reach used as edible oil seed in local dishes or for confectionary maturity.47 Harvesting should be performed when most exports. Current groundnut production is on the order (more than 80 percent) of the pods show signs of maturity, of 130,000 MT a year, of which 78 percent are grown that is, by the darkening of the inside of the pods.48 in Nampula, Inhambane, Zambezia, and Cabo Delgado along the coast.44 Although there is limited productivity— PESTS AND DISEASES on average 200–300 kg/ha in the south and somewhat The major pests and diseases include pests such as aphids, larger yields in the north (450–600 kg/ha)—for a house- termites, grasshoppers, leaf beetles, larvae, weevils, and hold, it is heavily constrained by agronomic practices that diseases like leaf spot, rosette virus, mosaic, aflatoxin, and rely on saving seed for planting the following season. thuku. In particular, Mozambique groundnuts are shown to have higher levels of aflatoxin than export quality stan- PRODUCTION RISKS dards allow—especially for product destined for South DROUGHT Africa.49 Many factors can contribute to the levels of afla- Drought exacerbates issues with groundnut production in toxin in groundnuts and, in terms of risk, end-of-season Mozambique, which also suffers from conditions of low drought is a particular issue. Contamination is exacerbated soil fertility and moisture content. This water stress affects by slow evacuation and processing of the crop postharvest, the crop at different growth stages during the season. which is an issue that needs to be taken into consideration Drought stress during the flowering and the pod filling as far as dealing with constraints is concerned.50 stages is critical for yield and other agronomic character- istics. This results in a drastic reduction in crop yield with the magnitude of degree of reduction depending on the MARKET RISKS varieties. It is accepted that there must be adequate water DOMESTIC PRICE VOLATILITY during the flowering period, the peg formation (about 6–8 Producer prices vary between provinces, reflecting weeks after planting), and the pod formation and filling.45 transportation cost differentials.51 In the southern part of the country, most of the harvest is used for home Not only the yield but also the quality of groundnuts consumption and only a small portion (usually not more decreases under drought conditions.46 A water deficit than 30 percent) is sold at market. In the north, it has FIGURE E.3. DOMESTIC PRICE VOLATILITY, GROUNDNUTS 550 Producer price, groundnuts, with shell (USD/ton) 450 USD/ton 350 250 150 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: FAOSTAT. 47 http://www.sciencedirect.com/science/article/pii/037842909390032I. 48 External Market Taskforce. 44 External Market Taskforce document on Dropbox. 49 Ibid. 45 Ibid. 50 Technical report—aflatoxin on Dropbox. 46 http://www.ajebs.com/vol-4/39.pdf. 51 Ibid. Risk Prioritization 89 become an important cash crop where local aggregators, National prices have historically been flat because of lim- large companies, and NGOs work to collect, process, ited refining capacity, but they have decreased over the and market groundnuts.52 However, the graph in figure past year. They are increasingly influenced by traders and E.3 indicates the volatility in this market; for example, processors aiming to export to South Africa and Europe.53 the spike in 1998 matches the insufficient rainfall in the With this comes the influence of price seasonality with southern coastal belt. established commercial relationships. 52 Ibid. 53 Ibid. 90 Mozambique: Agricultural Sector Risk Assessment CASHEW NUTS Cashew nuts are the primary cash crop for over tree losses. In 2000, thousands of cashew trees were 1.3 million smallholder farmers with average household knocked down by tropical storm Delfina,60 and in 2003 production of 100 kg generating about US$50 per sea- cyclone Japhet downed 12,325 trees.61 Another cyclone in son.54 Although these smallholder farmers account for 2008, Jokwe, hit Nampula and destroyed 1.47 million of 98 percent of national output, production is centered the 10 million trees in the province,62 and cyclone Funso in Nampula (48 percent of sales) with intensive plant- downed 53,130 trees in 2012.63 ing in Inhambane, Cabo Delgado, Gaza, and Zambezia. Once a large exporter in international markets, contrib- WILDFIRES uting 40 percent to global trade in the 1970s, Mozam- Wildfires damage 30–50 percent of trees because of the bique now makes up just 2 percent of the US$2.4 billion lack of proper crop management and an absence of global industry.55 Furthermore, only 18 percent of the weeding and burning areas to clean them.64 value added within the value chain is contributed within Mozambique itself.56 PESTS AND DISEASES65 The fungus Colletotrichum gloecosporioides is one of the most PRODUCTION RISKS common pathogens found in cashews. Initial symptoms DROUGHT show the development of reddish-brown, shiny, water- Irregular rainfall during flowering and at the beginning soaked lesions, followed by resin oozing from the affected of fruit formation is detrimental to cashew tree health. parts. The affected nuts and apples decay and shrivel, and This can result in stunted tree growth and longer matu- the flowers turn black and fall off. There may be associ- ration periods.57 Droughts occur in all cashew-produc- ated pitting of the surface of the nut. This serious disease ing areas but with a higher incidence in the central and requires the affected branches to be pruned and sprayed southern regions. In particular, drought has affected with fungicides. the production in the provinces of Sofala, Inham- bane, and Gaza, which account for one-third of total national production.58 In 2009, for example, “cashew MARKET RISKS rains” needed during flowering and fruiting did not Given that Mozambique is primarily an exporter of raw occur after the last rainfall in March 2009, affecting cashew kernels as opposed to processed nuts, producers production.59 are vulnerable to domestic value chain inefficiencies and international market fluctuations. CYCLONES AND TROPICAL STORMS Nampula province in the north, which has the largest In 2009/10, the government started a program to distrib- concentration of cashew trees, has suffered from regular ute cashew seedlings to schools under its Let’s Plant More cyclones and/or high winds, which regularly cause cashew Cashew Trees initiative. Although the price effect of this 54 Dropbox Doc. 60 IRIN. 55 PREM document on Dropbox. 61 ReliefWeb. 56 Africa Cashew Initiative on Dropbox. 62 ReliefWeb. 57 http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1677-0420200700 63 AllAfrica.com. 0400012. 64 Africa Cashew Initiative. 58 Africa Cashew Initiative on Dropbox. 65 http://www.fao.org/inpho_archive/content/documents/vlibrary/ac306e/ 59 Balanco de PES 2009. ac306e03.htm. Risk Prioritization 91 FIGURE E.4. DOMESTIC PRICE VOLATILITY, CASHEW NUTS (WITH SHELL) 400 Producer price (USD/ton), cashew nuts, with shell 350 300 USD/ton 250 200 150 100 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: FAOSTAT. project will not be evident until the trees are at maturity by 2013/2014, the 66,252 trees planted may cause local- ENABLING ENVIRONMENT ized effects.66 There was a significant increase in price in RISKS 2010/11 were prices went from 26 to 42 cents/kg to over As a primary export commodity, the history of the cashew 50 cents/kg, causing producers to flood the market (see sector is a prime example of the potential negative impact figure E.4).67 of government policies. Mozambique’s liberalization and privatization of the cashew sector in the early 1990s— eliminating export controls and processor protections— INTERNATIONAL PRICE VOLATILITY increased capital intensity in production and output but Foreign processor demand, changing produce qual- resulted in negative shocks to farm gate prices and urban ity standards, and consumer demand drive fluctuations employment. These risks are now somewhat mitigated in international market prices. Given the global value through strong government support of the private sector chain, these international factors are often controlled by and market liberalization policies that directly affect the traders and roasters in Asia before proceeding to devel- demand of raw cashew nuts. The parastatal body of the oped markets. Consequently, ultimately local nut cultiva- National Cashew Institute (INCAJU) has the responsibil- tors receive only 10 percent of what final consumers pay ity for collecting this tax and other permit revenues to pro- abroad.68 vide extension services and establish nurseries.69 66 Balanco de PES 2009. 67 Balanco de PES 2011. 68 Africa Cashew Initiative. 69 World Bank Cashew Reforms Revisited on Dropbox. 92 Mozambique: Agricultural Sector Risk Assessment COTTON The cotton sector currently accounts for 20 percent can cause serious damage, as much as 30–50 percent or of Mozambique’s agricultural GDP and benefits over more of production losses in the affected areas. A farm- 300,000 households, or 1.5 million inhabitants.70 In 2008, based, multiyear study has shown that timely application cotton was the third most valuable export crop, behind of pesticides by farmers improves productivity by more tobacco and sugar. Production is primarily concentrated than 100 percent.72 in the regions of Cabo Delgado and Nampula and is vari- able annually (figure E.5), largely because of risks that Controlling pests—particularly American bollworm—is reinforce low productivity, prices, and returns.71 among the most important activities to improve yields. First sprays each season are particularly important in PRODUCTION RISKS order to effectively kill sucking pests. Second and third pesticide applications kill pests at hatching, before they DROUGHT AND FLOODS invade the cotton squares, flowers, and bolls. Further- Cotton copes relatively well with weather variability more, studies show that many farmers spend up to 50 per- compared with other crops (for example, maize and ses- cent of their cotton labor time weeding around the crop ame). However, various stakeholders in the cotton sup- to prevent stunted plant growth. ply chain largely attribute yield variability to weather factors, particularly severe drought. The country was affected by regional droughts in 1994, 1996, and 2001; MARKET RISKS these droughts negatively affected crop production in DOMESTIC PRICE VOLATILITY some provinces. Drought is more of a relevant risk in the Mozambique has a national minimum pricing system southern region than in the cotton-growing central and for cotton that applies to all ginners who are awarded northern provinces. geographic concessions. In this indicative system, ginners and farmers agree on a bale price as far as 7 to 8 months Stakeholder interviews in provinces near river areas esti- before procurement of cotton begins. The aim of this mate flood-related losses suffered by cotton production indicative price is to assist farmers in deciding on their run at about 10–12 percent nearly every year. In some of cotton production plans and as an incentive to minimize the border regions (such as the border with Malawi), the crop substitution. Agreeing on a price with farmers, in low altitudes in cotton-growing areas increase its vulner- advance of their procuring seed cotton from farmers, ability to flooding because of rains coming in from neigh- places the greatest risk of price volatility on the cotton gin- boring countries. The causes of yield variations shown in ners. Therefore, credit default risks in the cotton industry figure E.5 are difficult to pinpoint without long-term his- are moderate because farmers who do not deliver cotton torical weather and yield data for various regions, though seed to the ginneries leave the ginneries unable to pro- it appears at first glance that some droughts and floods cure repayment for the inputs they provided to farmers in can partly explain the behavior of yield volatility. the sowing season. Although this system protects farmers from intra-seasonal price volatility, they remain exposed to price volatility. PESTS, WEEDS, AND DISEASES Insect infestations are an important risk in cotton produc- In 2009, prices paid to producers decreased by 35.6 per- tion yields in Mozambique: aphids, jassids, and American cent, causing them to cease production. This develop- bollworms are considered the most critical. If not con- ment was largely attributed to a lack of understanding trolled in time, an outbreak of red or American bollworm in regard to application of phytosanitary treatments.73 70 http://www.icac.org/tis/regional_networks/documents/africa_10/business _meeting/5_mozambique.pdf. 72 ARMT Risk Assessment. 71 Ibid. 73 Balanco de PES 2010. Risk Prioritization 93 FIGURE E.5. COTTON YIELDS, 1965–2007 (KG/HA) 800 Recovery from recent weather events? 700 Drought Floods in (7% drop) the northern 600 region Drought (localized (33% drop) event) 500 kg/ha 400 Drought 300 (keeps Severe floods yields (22% Floods in low) 200 drop) the central region (keeps yields low) 100 0 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 Source: World Bank 2010b. Another 30 percent drop occurred in 2011/12 compared FIGURE E.6. INTERNATIONAL COTTON with 2010/11 despite increases in international prices.74 PRICES, 2005–10 Cotton, A index (cents/kg) 180 INTERNATIONAL PRICE VOLATILITY The minimum price systems for cotton buffer farmers from intra-seasonal price volatility; however, it places 160 the greatest risk of price volatility on to the cotton gin- Cents/kg ners and traders. The system of minimum prices and 140 fixed prices places the ginners and traders at risk of fall- ing world cotton prices, as prices are set up to 8 months 120 prior to purchasing and processing the physical cotton. Although world cotton prices play a significant role in 100 setting the minimum price, there is significant volatility Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 that is currently borne by ginners. Unexpected losses can Source: World Bank 2010b. occur when the world price (which is part of the mini- mum price formula) falls below the agreed minimum price for seed cotton. The capacity of ginners and traders from farmers is by far the highest risk facing the cotton to manage such price risk varies markedly depending on industry today, which is becoming increasingly erratic as their expertise, size, and scale (see figure E.6). a result of crop substitution by farmers. Several other crops (for example, sesame, pigeon peas/cow peas, soy- Crop substitution: Given the concentration of activi- beans) have become more remunerative and are compet- ties around ginneries, the major risk facing the indus- ing with cotton for acreage. Depending on comparative try today is related to guaranteeing the supply of seed price of these different crops, farmers are making their cotton from farmers as the central operators of the cot- decision, creating huge uncertainty of seed cotton sup- ton supply chain. Securing input supply of seed cotton ply for the ginners. 74 Balanco de PES 2011. 94 Mozambique: Agricultural Sector Risk Assessment SUGARCANE Sugarcane contributes 25 percent to total agricultural FLOODS exports, the second largest export commodity after Generally, sugarcane is not considered a flood-tolerant tobacco.75 Unlike tobacco or cotton, sugarcane is grown crop and there is a reported mean yield reduction of 8.3 as a plantation crop by large industry stakeholders and to 25 percent under a 15 cm water table.80 At the same through their out-grower schemes.76 It is forecast that time, previous research by USDA-Agriculture Research sugarcane production will increase to almost 4 mil- Service (USDA-ARS) indicates that sugarcane can tol- lion tons in 2012/13 based on an increase in hectares erate repeated floods of 7–14 days without yield loss, if planted. This growth trend originated with the sugar this short-duration flooding occurs just before or during industry witnessing average production increases of 30 harvest. However, the biomass of recently planted sugar- percent annually between 2000 and 2006.77 The planted cane is increasingly reduced by flood durations of zero to area of sugarcane is expected to increase to about 50,000 6 weeks and flood onsets after 2 or 4 weeks at a 12-inch hectares, with new investments and increases in the pro- water-table depth are similarly detrimental.81 cessing efficiencies of Mozambique’s four commercial sugar mills.78 PESTS AND DISEASES African armyworms, stem borers, sugarcane scale, mealy PRODUCTION RISKS bugs, termites, and rats are common pests for sugarcane DROUGHT crops. These pests attack the stems and leaves of the plant Plants are constituted of about 85 to 95 percent water, and result in yield losses of between 18 percent from and any losses in water content produce changes in phys- termites and 30 percent from sugarcane scale.82 Sugar- iological reactions. Plants suffering from a water deficit cane diseases such as the mosaic potyvirus, rust, red rot, change their metabolism in order to protect themselves Ratoon stunting disease, and scald are equally devastat- against this stress. This results in stunted inter-nodes, ing, often requiring the complete destruction of harvests drying of bottom leaves, and inward rolling of top leaves. and repeated application of pesticides.83 Drought can reduce elongation by about 30 percent and yield loss can be even more. Early stages of growth, par- ticularly the active tilling period, are critical in drought MARKET RISKS situations. 79 DOMESTIC PRICE VOLATILITY Domestic price volatility has closely mirrored gen- Although the majority of sugarcane is irrigated, which eral food inflation in Mozambique, resulting in higher minimizes the problems caused by irregularities in rain- prices and decreased general consumption by 11 percent fall, there was an 8.3 percent reduction in production in (figure E.7).84 2005/06 because of a deficiency in the irrigation sys- tem, which failed to smooth out the lack of required ENABLING ENVIRONMENT RISK rainfall. Raw and processed sugars are subject to import surtaxes, in addition to the basic duty of 7.5 percent. The two 75 GAIN report on dropbox or http://www.thebioenergysite.com/reports /?id=585. 80 http://www.wsrjournals.org/download.php?id=477277168993247722.pdf 76 http://www.africanwaterfacility.org/fileadmin/uploads/awf/publications &type=application/pdf&op=1. -reports/COFAMOSAREPORTNOV07_0.PDF. 81 https://www.certifiedcropadviser.org/story/2012/may/mon/study-examines 77 Smallholder Sugarcane Production on Dropbox. -effects-of-flooding-on-young-sugar-cane. 78 Xinavane and Maragra in Maputo province and Sena and Mafambisse in 82 http://www.infonet-biovision.org/default/ct/134/crops. Sofala province. 83 Ibid. 79 http://cdn.intechopen.com/pdfs/26982/InTech-Sugarcane_responses_at 84 GAIN report on dropbox or http://www.thebioenergysite.com/reports _water_deficit_conditions.pdf. /?id=585. Risk Prioritization 95 FIGURE E.7. DOMESTIC PRICE VOLATILITY, SUGARCANE 17 Producer price, sugar cane (USD/ton) 15 USD/ton 13 11 9 7 5 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 Source: FAO PriceSTAT. Note: 2003–07 are FAO forecasts. variable surtaxes are set on a monthly basis and depend implemented through a variable import tax system to on the difference between the Mozambican minimum absorb world market price fluctuations. The authorities prices (US$385/ton for raw sugar and US$450/ton for claimed that, given distorted international sugar prices, processed sugar) and world market reference prices.85 such a system provided greater certainty to the investors The sugar surtaxes have been in place since 1999,86 that rehabilitated the estates devastated in the civil war.87 85 Ministerial Diploma No. 56/2001 of 30 March 2001 sets out the mecha- nism. For February 2008, Order of Service No. 002/DGA/2008 of January 28, 2008, sets the applicable reference prices (per ton) for raw (US$347.18) and processed sugar (US$388.09), and the associated surtaxes on raw sugar (10 per- cent) and processed sugar (15 percent). 87 WTO Trade Policy Review Mozambique 2009, https://www.wto.org/english 86 Decree No. 74/99 of 12 October 1999. /tratop_e/tpr_e/tp309_e.htm. 96 Mozambique: Agricultural Sector Risk Assessment CASSAVA Cassava is a subsistence commodity produced across over a harvest season, negative effects can begin to be 2.5 million farms and is a staple crop with 70 percent of seen within three days of there being standing surface national harvests utilized as household food.88 Production water.91 is concentrated along the northern and southern coasts, with the provinces of Zambezia, Nampula, Cabo Delgado, In terms of actual recorded losses, in 2000, tropical storm and Niassa accounting for 85 percent of production. The Delfina destroyed 2,000 ha of beans and cassava92 and crop is versatile with high levels of drought tolerance and the 2001/02 season saw losses in the cassava sector as a efficient growth without the intensive application of fertil- result of irregular and insufficient rain.93 Furthermore, in izers or pesticides. However, it is exposed to various risks 2003 flooding led to losses of 3,000 ha in Nampula and because of inadequate agronomic practices, geographic Zambezia.94 vulnerabilities, and limited market linkages and support. PESTS AND DISEASES PRODUCTION RISK There are several types of pest and disease threats to DROUGHT cassava including cassava brown streak disease (CBSD), Although mature cassava is drought resistant, the crop cassava mosaic disease (CMD), cassava bacterial blight, is especially vulnerable during the planting period. The green spider mites, mealy bugs, white flies, fungal diseases, use of stem cuttings allows planting any time of year yet and locusts. typically farmer’s plant near the end of the dry season in September to take advantage of the oncoming rains.89 Although it is often difficult to differentiate between them, As sprouting happens as early as one week after plant- these diseases result in increased root rot and decreased ing there is particular vulnerability in the first month. leaf size and stem size.95 Although not resulting in total Drought and irregular rainfall disrupt planting timing, crop failure in many cases if caught early and treated, reduce soil moisture, and increase the likelihood of cut- contamination and nontreatment with pesticides can tings not developing.90 From 1992 to 2012, the northern result in future crop issues because cuttings taken con- and southern coastal regions have been victims of eight tinue to be diseased. Furthermore, varieties bred to be major drought events. resistant to cassava mosaic disease are being observed as susceptible to new strains of CBSD.96 In 2002, a CBSD outbreak led to decreases in output prices and thus to FLOODING price spikes of up to 45 percent in Nampula (McSeen et During the rainy season, there are continued threats al. 2006). Cassava is particularly important in Mozam- from flooding. In particular, there is a high probability of bique regarding food security, thanks to the “natural flash flooding, which severely threatens cassava harvest- storage” offered because it needs to be harvested for ing. Research shows that although mature cassava can up to two years before it can be used. However, CBSD remain in the ground upward of two years, this does not can render roots unusable when left in the ground for necessarily protect the plant from adverse flood effects, reducing root yields by 24 percent, reducing nutritional value by 30 percent of starch content, increasing flesh 91 “Cassava Breeding Report: Proceedings of the Fourth Regional Workshop,” cyanide concentration, and causing root rot. Although November 2–6, 1993. the latter is particularly likely given periodic flooding 92 IRIN. 93 RA DINA 2002. 94 ReliefWeb. 95 http://www.apsnet.org/publications/apsnetfeatures/Pages/cassava.aspx. 88 Sub-Sector Strategic Study on Cassava 2007). 96 This has been observed in Uganda, inland areas of Tanzania, and West- 89 Ibid. ern Kenya. “Cassava Brown Streak Diesease,” ASRECA—National Crops 90 FAO http://www.fao.org/docrep/x5032e/x5032E01.htm. Resources Research Institute (NaCRRI) Namulonge Paper, 1. Risk Prioritization 97 FIGURE E.8. INTERNATIONAL PRICE VOLATILITY, CASSAVA Nampula, manioc (cassava), retail 500 Nampula, manioc (cassava), wholesale 400 USD/ton 300 200 100 0 Sep-06 Dec-06 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Mar-10 Jun-10 Sep-10 Dec-10 Mar-11 Jun-11 Sep-11 Dec-11 Mar-12 Jun-12 Sep-12 Dec-12 Mar-13 over 9 months97 or cause root yield loss of 100 percent by postharvest deterioration. Additionally, the character- in the worst, untreated cases. The threat continues with istics of cassava are such that degradation begins within upward of 40 percent of crops exhibiting incubated signs 48 hours and so losses are magnified because of an inabil- of the disease (McSween et al. 2006). In terms of actual ity to get the product to market in time.102 Figure E.8 pro- recorded losses, in 2000 it was recorded that locusts vides an indication of the price volatility since 2006 of destroyed about 160,000 hectares of cassava,98 whereas producer prices, both retail and wholesale. in 2001/02 root rot99 was recorded as negatively affecting production and in 2001/02100 and 2004101 a brown streak The lack of support for cassava-based product (that is, cas- disease epidemic was recorded. sava flour, cassava chips) producers not only inhibits cassa- va’s market potential but also makes production victim to MARKET RISK low prices and volatility, and it removes the incentives for With this domestic market orientation of cassava con- farmers to invest in better agronomic techniques. Cassava sumption and sales, its exposure to market risk is based in particular is very responsive to fertilizer application, yet on price volatility along with common constraints such as its low utilization by households means there is little short- value chain inefficiencies, crop loss, and spoilage caused term volatility in input prices. 97 Cassava Brown Streak Diesease, ASRECA—National Crops Resources Research Institute (NaCRRI) Namulonge paper, 1. 98 IRIN. 99 RA DINA 2002. 100 Ibid. The Cassava+ project, 110921 case description for Agri-ProFocus knowledge 102 101 Ibid. agenda on BDS, DADTCO. 98 Mozambique: Agricultural Sector Risk Assessment MAIZE Maize is Mozambique’s most cultivated agricultural com- FLOODS modity, produced by 78 percent of all rural households, Maize is also very sensitive to flooding, particularly in the and the staple dietary crop, with only 18 percent sold into early vegetative stages, and can survive for only two to four the market (Donovan and Tostao 2010). Maize is inten- days underwater.108 Waterlogged soil reduces the diffusion sively farmed in the northern provinces, which account for of oxygen, gradually suffocating plant roots and leading about 50 percent of national production and 70 percent to an accumulation of toxic nutrients. Studies have shown of sales (Tschirley, Abdula, and Weber 2006). Whereas that maize is more susceptible during early seedling and output has increased over the past 60 years, this has pre- the ensuing tasseling stages, resulting in visible leaf rolling, dominately been caused by expansion of cultivated areas. wilting, and root decay.109 The lack of productivity increases can be partly explained via several risks and inefficiencies in the value chain result- In terms of actual recorded losses, about 45,000 hectares ing in average actual maize yields of 1.4 tons/ha com- were destroyed in 1996,110 whereas the devastating floods pared with 5 tons/ha potential (Zavale, Mabaya, and of 2000 saw about 140,000 hectares lost in total; maize and Christy, 2006). rice111 made up most of the losses. In 2009, a national loss of 13 percent was recorded,112 whereas the most recent floods in 2013 led to estimated losses of up to 154,000 hectares PRODUCTION RISKS total, including a large proportion of maize.113 Excessive rain and the storms Dando and Funso have been recorded DROUGHT as leading to losses of planted areas in Cabo Delgado, Maize is produced under rain-fed conditions. This makes Zambezia, Sofala, Inhambane, Gaza, and Maputo.114 for an especially vulnerable crop because it thrives on high fertility and in soils with good moisture-retention PESTS AND DISEASES capabilities.103 Water or heat stress at the time of a maize Given maize’s high prevalence throughout the country, plant’s flowering can directly reduce yields by disrupt- it is regularly exposed to various pests and diseases. The ing pollination and the formation of maize grain; at major maize diseases can be grouped into four catego- 8 inches in height, the numbers of kernel rows on the ries: leaf blights, stalk rots, ear rots, and viral diseases. ear have already been determined.104 Yield reductions The main pests include cutworms, ground weevils, maize by as much as 7 percent per day of plants under stress chafer, beetles, and aphid disease.115 Because maize is usu- have been observed, making this development stage the ally intercropped and cultivated from earlier harvest gen- most sensitive to drought and increased temperatures.105 erations, there are many vectors for disease transmission. Furthermore, the 14-day periods before and after silk- Given the importance of the early vegetative stage of the ing—or the emergence of the ear shoots—are crucial to plant, the application of herbicides and other pesticides final yield.106 Recorded losses indicate that in the 2010/11 completed within the first two months of corn develop- season, a lack of rain in January and February led to pro- ment will have a significant impact on output.116 duction declines of 22 percent in Tete and 39 percent in Inhambane and Gaza.107 108 http://www.ag.ndsu.edu/procrop/env/fldwhb07.htm. 109 Ibid. 110 FAO. 103 USAID Maize Value Chain. 111 World Bank. 104 http://dirp3.pids.gov.ph/ACIAR/relatedresources/Impact%20of%20drought 112 IRIN. %20on%20corn%20productivity.pdf. 113 FAO. 105 USAID and USGS. 114 Balanco de PES 2012. 106 http://dirp3.pids.gov.ph/ACIAR/relatedresources/Impact%20of%20drought 115 http://www.biodiversityexplorer.org/plants/poaceae/zea_mays_pests.htm. %20on%20corn%20productivity.pdf. 116 http://dirp3.pids.gov.ph/ACIAR/relatedresources/Impact%20of%20drought 107 Report on Monitoring of Food Security and Nutrition in Mozambique 2011. %20on%20corn%20productivity.pdf. Risk Prioritization 99 FIGURE E.9. DOMESTIC AND INTERNATIONAL PRICE VOLATILITY, MAIZE Nampula, maize (white), retail Maize (corn), U.S. no.2 yellow, FOB Maxixe, maize (white), retail Gulf of Mexico, U.S. price Maputo, maize (white), retail 350 600 300 500 250 USD/MT 400 200 USD/ton 300 150 200 100 100 50 0 0 Dec-94 Aug-96 Apr-98 Dec-99 Aug-01 Apr-03 Dec-04 Aug-06 Apr-08 Dec-09 Aug-11 1980M01 1982M08 1985M03 1987M10 1990M05 1992M12 1995M07 1998M02 2000M09 2003M04 2005M11 2008M06 2011M01 Source: FAO GIEWS Food Price Data and Analysis Source: IMF Primary Commodity Prices. Tool. In 1998/99, it was observed that locusts, birds, and wheat- system not capturing the losses resulted in limited supply worm negatively affected maize production,117 and locusts and unexpectedly high prices. Traders, lacking informa- and lizards brought about the destruction of the second tion, were unable to import quantities to fill the resulting season crop in Gaza in 2006.118 demand gaps. MARKET RISKS INTERNATIONAL PRICE VOLATILITY Another notable aspect is that the high prices in Mozam- DOMESTIC PRICE VOLATILITY bique seen in the recent period have not declined as world The limited volatility in Mozambique’s maize prices can market prices declined (figure E.9).119 Commodity prices, be largely attributed to the government’s trade policies. set by maize futures contracts, are intimately tied to energy Administrative procedures imposed to ensure food secu- prices and weather predictions. The alternative use of rity, product safety, and the addressing of environmen- maize as a source of fuel ethanol affects prices directly tal issues has distorted the trade of goods, conditions for and indirectly. Maize prices peaked at US$310.24 per retail, and factors of production. This environment has ton (US$7.88/bushel) in June 2008 with the after effects fostered an uncertain investment climate, often creating of floods in the United States, then fell in late 2008 and a disincentive for private sector actors to make significant early 2009 as commodity prices, led by oil prices, declined investments. The trading environment is characterized by during the 2008 financial crisis. Analysis from the Inter- a lack of harmonization in cross-border trade with deficits national Food Policy Research Institute estimates that ris- and surpluses moderated by informal and illegal trading ing demand for ethanol caused 40 percent of the rise in with neighboring states. maize prices from 2000 to 2007. The United States, the world’s major exporter, had a federal mandate to devote Maize prices show distinct seasonal fluctuations, as pro- approximately 27 percent of the 2011 maize crop to etha- duction is based on rain-fed irrigation systems and there is nol production.120 a single, main wet season. A particular example of domes- tic volatility was the price spike of 2005. A combination of a bad production year and of the crop forecasting 119 Donovan et al (2010) Staple food prices in Mozambique, Prepared for the Comesa policy seminar, Comesa-MSU-IFPRI African Agricultural Marketing Project (AAMP). 117 Annual Report Agricultural Campaign 1998/99. 120 http://www.unctad.info/en/Infocomm/AACP-Products/Commodity-Profile 118 RNDD 2000. ---Corn/. 100 Mozambique: Agricultural Sector Risk Assessment APPENDIX F RISK TRANSFER AND FINANCING OPTIONS IN MOZAMBIQUE MICROLEVEL OPTIONS THE PUBLIC-PRIVATE PARTNERSHIP (PPP) IN AGRICULTURAL INSURANCE FOR FARMERS Developing a viable PPP in agricultural insurance would require signifi- cant public investments in data and would probably need to be accompa- nied by substantial growth in credit or input utilization. Developing a PPP in agricultural insurance is a long-term objective that requires appropriately sustained engagement and high levels of investment from both the public and private sectors. In addition, to achieve sustainability and meaningful uptake, it relies on several key pil- lars, one of which is an effective distribution channel to rural farmers through which insurance can be sold. These distribution channels can take many forms, such as rural lending institutions, input suppliers, or social welfare payment systems. The main dis- tribution channels in Mozambique currently are rural credit and input suppliers, but only 2.3 percent of the rural population has access to finance (World Bank 2012b); 3.7 percent has access to credit (World Bank 2012b); and only 10 percent of maize farm- ers and 4 percent of rice farmers use improved seeds (Government of Mozambique 2008). These low levels of outreach in effect mean that currently there is not an effec- tive distribution channel to provide insurance to the vast majority of rural farmers. Experience from other countries also suggests that significant investments in weather, yield and satellite data collection, and auditing would be required if agricultural insur- ance products were to offer reliable protection to farmers at low cost. Currently, a very limited amount of yield data are collected in Mozambique and the weather station network is limited. Moreover, the audit procedures for these types of data are not in line with international reinsurance standards. Risk Prioritization 101 Given the significant fiscal burden of developing lending is unlikely to achieve sufficient uptake (and thus a PPP in agricultural insurance, other risk man- generate sufficient premium volume) to be sustainable. agement options may be more cost effective at Again, in the medium term, should the government look this time. The investments in data mentioned above, as to develop the rural credit market, this option could be well as the institutional and market investments required considered. to develop distribution channels to a level that would achieve critical mass, would be high relative to other CATASTROPHE WEATHER INDEX-BASED investment options in agricultural risk management INSURANCE currently available to Mozambique. In the future, any Catastrophe weather index-based insurance (WII) investments in agricultural insurance would be coupled products could be considered for large-scale com- with other initiatives; for example, should the govern- mercial farmers. However, the development impact of ment of Mozambique want to develop rural productiv- any such products may be low. ity, it could look to develop the rural credit market to increase rural lending to farmers. Agriculture insurance is an excellent partner for such a venture; if effectively developed, it can protect vulnerable farmers against MACROLEVEL OPTIONS shocks as well as rural lending institutions against covar- SOVEREIGN AGRICULTURE iate risks that can lead to bankruptcy, increasing their RISK FINANCING AND INSURANCE resilience. It is unclear as to what the government considers its contingent liability to the agricultural sector, and what it considers to be the responsibility of MESOLEVEL OPTIONS donors or farmers. However, if the government considers its contingent liability to be relatively MESOLEVEL AGRICULTURE RISK low, then developing a sovereign agriculture FINANCING risk financing strategy may have lower impact Given the low levels of rural lending in Mozam- than other agricultural risk management invest- bique and the lack of any direct regulatory ments. In recent years, the government has taken incentive for lenders to transfer their portfolio- steps to increase the protection it offers to vulnerable level agricultural risk, implementing any form farmers against agricultural shocks. It has established a of unsubsidized mesolevel agricultural insur- platform for disaster risk management in Mozambique, ance program would have limited uptake and the Coordinating Council for Disaster Risk Manage- thus be unsustainable. Insurance companies require ment (CCGC) under the Council of Ministers, com- high levels of premium volume from the products they prising representation across sectors. It also established sell to be sustainable in the long run.121 This becomes a The INGC, the executive arm of the CCGC, which is key challenge when, for example, insurance companies responsible for the coordination of disaster risk man- look to sell microinsurance; because the premiums are agement activities, including disaster response and so much lower, they must sell a large number of policies reduction. Disaster risk management, including disas- to be sustainable. Given the low levels of rural credit ter risk reduction, is also integrated into strategic plan- in Mozambique, any insurance program targeting rural ning documents including; the National Action Plan for Poverty Reduction, and the 2010–14 5-Year Plan. It 121 This is to cover overheads, such as fixed costs, variable costs, cost of capital, also has several initiatives in place to assist vulnerable and so on. farmers in the aftermath of agricultural shocks, such as 102 Mozambique: Agricultural Sector Risk Assessment distribution of free seeds. That said, given the fact that INDEX-BASED SOCIAL SAFETY the government is reliant on donor funding for both its MECHANISMS budget122 as well as disaster response,123 it has a limited Although the government is looking to increase the out- contingent liability to the agricultural sector. Were the reach of the social welfare scheme, currently there do not government to plan to increase its fiscal expenditure in appear to be immediate plans or interest in developing the aftermath of disasters, then a sovereign agricultural a scheme that would automatically expand social welfare risk financing strategy may become a more attractive payments in the aftermath of an agricultural shock based option. on a predefined set of rules. 122 Foreign aid accounted for 46 percent of the state budget in 2010, according to the AfDB in 2011. 123 The government funds between 20 percent and 30 percent of the total esti- mated annual expenditure of expected disaster scenarios, with the remaining funds provided by donors. Risk Prioritization 103 A G R I C U LT U R E G L O B A L P R A C T I C E T E C H N I C A L A S S I S TA N C E P A P E R W O R L D B A N K G R O U P R E P O R T N U M B E R 96289-MZ 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org/agriculture