Agriculture global practice technical assistance Paper 94228 GHANA: AGRICULTURal SECTOR RISK ASSESSMENT Risk Prioritization Vikas Choudhary, Garry Christienson, Henri Josserand, and Stephen D’Alessandro WORLD BANK GROUP REPORT NUMBER 94228-GH June 2015 Agriculture Global Practice Technical Assistance Paper GHANA: AGRICULTURal SECTOR RISK ASSESSMENT Risk Prioritization Vikas Choudhary, Garry Christienson, Henri Josserand, and Stephen D’Alessandro © 2015 World Bank Group 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org Email: 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 volume do not necessarily reflect the views of the Executive Directors of 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 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 Clearance 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 Photos from left to right: 1. Neil Palmer (CIAT), Failed maize crops in Ghana’s Upper West Region. 2. Nana Kofi Acquah (CGIAR), Using a petrol pump to pump ground water for watering plants. Buying petrol is a more expensive way to farm. 3. IITA, Infected dried maize cobs in a farm store in Ghana. 4. Lava Kumar (IITA), Field training on yam virus disease indexing at CRI, Kumasi Ghana. CONTENTS Acronyms and Abbreviations vii Acknowledgments ix Executive Summary xi Chapter One: Introduction 1 Chapter Two: Overview of Agricultural Systems in Ghana 5 Agro-Climatic Conditions 5 Rainfall Patterns and Trends 7 Crop Production Systems 8 Agricultural Markets and Producer Price Trends 11 Livestock Production 12 Principal Constraints to Agricultural Production 14 Chapter Three: Agricultural Sector Risks 15 Production Risks 15 Market Risks 22 Enabling Environment Risk 26 Multiple Shocks 28 Chapter Four: Adverse Impact of Agricultural Risk 31 Conceptual and Methodological Basis for Analysis 31 Aggregate Crop Production Risks 33 Impact of Livestock Diseases 36 Chapter Five: Assessment of Stakeholder Vulnerability 37 Rain-Fed Agriculture 38 Irrigated Agriculture 39 Agro-Pastoralism 40 Commercial Farmers 41 Traders and Processors 41 Ranking of Stakeholder Risk Perceptions 42 Chapter Six: Risk Prioritization and Management 43 Risk Prioritization 43 Risk Management Measures 44 Description of Priority Risk Management Measures 47 Filtering and Prioritizing Interventions 51 Conclusion 53 References 55 Risk Prioritization iii Appendix A: Regional Risk Profiles 59 Appendix B: Commodity Risk Profiles 71 Appendix C: Rainfall Patterns and Implications for Crop Production 79 Appendix D: Climate Change Impact Assessment of Agriculture in Ghana 87 Appendix E: Stakeholder Vulnerability Analysis 95 Appendix F: Agricultural Risk Financing and Insurance for Ghana: Options for Consideration 101 Appendix G: Indicative Losses 105 Appendix H: Economic Indicators 113 Appendix I: Timeline of Events 115 Appendix J: Assessing Vulnerability in Northern Regions 119 Appendix K: Irrigation Development in Ghana 123 BOXES BOX 3.1: Rainfall Patterns and Crop Production 18 BOX 6.1: Veterinary Services in Ghana 48 FIGURES Figure ES.1: Agriculture Sector Growth (%), 1980–2012 xii Figure 1.1: Agriculture Sector Performance, 2007–12 2 Figure 1.2: Agricultural Sector Risk Management Process Flow 3 Figure 2.1: Administrative Regions and Agro-Ecological Zones 6 Figure 2.2: Monthly Rainfall Patterns by Region 7 Figure 2.3: Composition of Crop Production 8 Figure 2.4: Cocoa Production, 1990–2011 9 Figure 2.5: Maize Production, 1990–2011 10 Figure 2.6: Cassava Production, 1990–2011 10 Figure 2.7: Trends in Real Cereal Prices, 1991–2010 11 Figure 2.8: Trends in Real Prices for Root Crops and Plantain, 1991–2010 12 Figure 2.9: Producer Prices for Cocoa and Groundnuts, 1991–2010 12 Figure 3.1: Adverse Crop Production Events, 1980–2011 16 Figure 3.2: Impact of Risks on Cereal Production and Yields, 1980–2011 18 Figure 3.3: Cocoa Prices and Production, 1991–2010 24 Figure 3.4: Nominal Exchange Rates, 1980–2012 25 Figure 3.5: Commercial Bank Interest Rates, 2004–12 26 Figure 3.6: Trends in Maize Production and Prices, 1995–2011 27 Figure 3.7: Ghana/Côte d’Ivoire Cocoa Producer Price Differential 28 Figure 4.1: Frequency and Severity of Adverse Production Events by Crop 34 Figure 4.2: Crop Production Shocks by Region, 1992–2009 35 iv Ghana: Agricultural Sector Risk Assessment Figure 4.3: Frequency and Severity of Different Crop Risks 35 Figure C.1: Weather Station Distribution with Region Centroids 80 Figure C.2: Monthly Rainfall Patterns by Region 81 Figure C.3: Correlation Matrix Plot 84 Figure C.4: Factor Loadings Plot 85 Figure C.5: Mean Factor Scores by Region 86 Figure D.1: Changes in Mean Precipitation by 2030 (left) and Changes in Mean Precipitation by 2050 (right) 89 Figure D.2: Changes in Mean Annual Temperature 2030 (left) and Changes in Mean Annual Temperature 2050 (right) 90 Figure D.3: Current Suitability of Cocoa Growing Area (left) and Future Suitability of Cocoa Growing Area (right) 91 Figure D.4: Yield Changes 2010–50 92 Figure E.1: Crop Yield Sensitivity Indexes (left ) and Regional Vulnerability Indexes (right ) 98 Figure E.2: Livelihood Zones 98 Figure E.3: Food Consumption 98 Figure E.4: Mean Vulnerability Indexes of Upper East Districts (top left ), Upper West Districts (bottom left ), and Northern Districts (bottom right ) 99 Figure H.1: Agriculture, Value Added (2007–12) 113 Figure H.2: Growth in Gross Domestic Product (2006 Constant Prices), 2007–12 113 Figure H.3: Agriculture, Value Added (Annual % Growth), 1980–2012 113 Figure H.4: Agriculture, Value Added (Annual % Growth), 2000–12 114 Figure K.1: River Basins in Ghana 124 Figure K.2: Distribution of Irrigation System Typologies in the Regions of Ghana 124 TABLES Table 1.1: METASIP’s (2011–15) Six Program Areas and Agricultural Risks 2 Table 2.1: Agro-Ecological Zones of Ghana (North to South) 6 Table 2.2: Trends in Crop Production, 1990–2011 8 Table 2.3: Coefficients of Variation for Crop Production, 1990–2011 9 Table 2.4: Domestic Food Supply and Demand for Food Staples 10 Table 3.1: Frequency of Low Rainfall Events by Region, 1981–2010 17 Table 3.2: Frequency of Excess Rainfall Events by Region, 1981–2010 19 Table 3.3: Pests and Disease Risks for Ghanaian Agriculture 21 Table 3.4: Inter-Annual Crop Price Variability, 1991–2001 23 Table 3.5: Seasonal Price Variability for Food Crops, 2004–08 24 Table 4.1: Severity and Cost of Adverse Events for Aggregate Crop Production 33 Risk Prioritization v Table 5.1: Risk Ranking, Rain-Fed Farming 38 Table 5.2: Risk Ranking, Irrigated Farming 39 Table 5.3: Risk Ranking, Agro-Pastoralists 40 Table 5.4: Risk Ranking, Commercial Farmers 41 Table 5.5: Risk Ranking, Grain Traders 41 Table 5.6: Stakeholders’ Risk Perceptions and Rankings 42 Table 6.1: Ranking of Risks by Sub-Sector 44 Table 6.2: Ranking of Risks and Vulnerability by Region 45 Table 6.3: Indicative Risk Management Measures 46 Table 6.4: Relative Benefits of Risk Management Interventions 51 Table 6.5: Decision Filters for Risk Management Measures 52 Table 6.6: Integration with METASIP 54 Table C.1: Standardized Cumulative Rainfall 82 Table C.2: Impact of Rainfall Parameters on Crop Yield 83 Table C.3: PCA Analysis: Three Eigen Values and Proportion of Variance Explained 84 Table C.4: Correlation of Components 84 Table E.1: Food Insecurity and Vulnerability by Region 96 Table E.2: Vulnerable Groups 97 Table G.1: Indicative Losses (US$ Million) for Adverse Crop Production Events by Crop, 1991–2011 (Constant Prices = 2004–06) 106 Table G.2: Indicative Losses (% Gross Agric. Output) for Adverse Crop Production Events by Crop, 1991–2011 (Constant Prices = 2004–06) 107 Table G.3: Indicative Losses (US$ Million) for Adverse Producer Price Movements by Crop, 1991–2010 (Real Prices 2010 = 100) 108 Table G.4: Indicative Losses (% Gross Agric. Output) for Adverse Producer Price Movements by Crop (Real Prices 2010 = 100) 109 Table G.5: Indicative Losses (US$ Million) for Adverse Crop Production Events by Region (Constant Prices = 2004–06) 110 Table G.6: Indicative Losses (% Gross Agric. Output) for Adverse Crop Production Events by Region (Constant Prices = 2004–06) 111 Table J.1: Household Cropping Activity 120 Table J.2: Distribution of Household Farm Size, by Region (Acres) 120 Table J.3: Type of Access to Land, by Region 120 Table J.4: Average Yield for Major Crops in the Upper West Region, 2010 120 Table J.5: Average Yield for Major Crops in the Upper West Region, 2011 121 Table J.6: Weather Impacts on Key Crops, 2011–12 121 vi Ghana: Agricultural Sector Risk Assessment ACRONYMS AND ABBREVIATIONS Acronym Definition Acronym Definition AAGDS Accelerated Agricultural Growth and GOG Government of Ghana Development Strategy GFDRR Global Facility for Disaster Risk and Response AfDB African Development Bank HA Hectares ARMT Agriculture Risk Management Team HPAI Highly Pathogenic Avian Influenza ASF African Swine Fever ICT Information and Communication CABI Centre for Agricultural Bioscience Technology International IFPRI International Food Policy Research Institute CBPP Contagious Bovine Pleuri-Pneumonia IMF International Monetary Fund CFA Communauté Financière Africaine IPCC AR4 Intergovernmental Panel on Climate Change CFSVA Comprehensive Food Security and Fourth Assessment Report Vulnerability Assessment IPM Integrated Pest Management CIAT Center for Tropical Agriculture LEAP Livelihood Empowerment Against Poverty CNRM National Centre for Meteorological Research program COCOBOD Ghana Cocoa Board METASIP Medium Term Agricultural Sector CPUE Catch per Unit Effort Investment Plan CRI Crop Research Institute MIROC Model for Interdisciplinary Research on CSIR Council for Scientific and Industrial Climate Research MoFA Ministry of Food and Agriculture CSIRO Commonwealth Scientific and Industrial mt Metric Ton Research Organisation NADMO National Disaster Management CSSVD Cocoa Swollen Shoot Virus Disease Organization DFID Department for International Development NAFCO National Food Buffer Stock Company DSSAT Decision Support System for Agrotechnology NCAP Netherlands Climate Assistance Programme Transfer NCCSAP2 Netherlands Climate Change Studies DVO District Veterinary Officers Assistance Programme Phase 2 ECHAM European Centre Hamburg Model NGO Nongovernmental Organization EWB Engineers Without Borders NRM Natural Resource Management FAO Food and Agriculture Organization PPP Public Private Partnership FAOSTAT Food and Agriculture Organization Statistics PPR Peste des petits ruminants Division PPRSD Plant Protection and Regulatory Services FFW Food for Work Directorate GAO Gross Agricultural Output SECO Swiss Secretariat of Economic Affairs GCM Global Climate Model SST Sea Surface Temperature GDP Gross Domestic Product UNFCCC United Nations Framework Convention on G-8 Group of Eight Climate Change GSGDA Ghana Shared Growth and Development USAID United States Agency for International Agenda Development Monetary amounts are Ghanaian cedi (GH¢) unless otherwise indicated. VSD Veterinary Services Directorate Risk Prioritization vii ACKNOWLEDGMENTS This report was prepared by a team led by Vikas Choudhary (Senior Economist, Task Team Leader) and Stephen D’Alessandro (Agricultural Specialist, Co-Task Team Leader), and consisting of Garry Christenson (Consultant, Agricultural Economist) and Henri Josserand (Consultant, Crop and Livestock Specialist). Luis Arturo Castellanos (Consultant, Weather Risk Management) conducted weather analysis for the report, while Victor Antwi (Consultant) and Samuel Sarpong (Consultant) provided local coordi- nation and data collection support. The team is grateful to the Government of Ghana, in particular to the Policy Planning and Monitoring Directorate (PPMD) of the Ministry of Food and Agriculture and their staff for their collaboration and contributions to the work. Special thanks are due to Angela Danson (Deputy Director, PPMED) for her support and guidance through- out the assessment process. The team would also like to express its gratitude to all the stakeholders who participated during the fieldwork and workshop exercises for sharing their valuable time and perspective and discussing the findings. Their insights obliged the team to be realistic and practical. The team would like to thank Fenton Sands of the U.S. Agency for International Development (USAID) who provided valuable support, guidance, and feedback throughout the risk management engagement in Ghana. The team would also like to thank Yusupha B. Crookes (Country Director, Ghana), Martien Van Nieukoop (Sec- tor Manager), and Marc Sadler (Advisor) for their valuable guidance and support. Christopher Paul Jackson (Lead Economist), and Charles Annor-Frempong (Senior Economist) were peer reviewers of the document and the team would like to thank them for their valuable feedback and comments. The team would like to thank Hans Jansen (Senior Agriculture Economist), Mr. Waqar Haider (Sector Leader, Sustainable Development Department, World Bank), and Rose Ampadu (Program Assistant). The final report also benefited from the work of independent consulting editor Damian Milverton. The authors gratefully acknowledge the generous contributions from USAID, Ministry of Foreign Affairs of the Government of the Netherlands and State Secre- tariat for Economic Affairs (SECO) of the Government of Switzerland. Risk Prioritization ix EXECUTIVE SUMMARY In Ghana, the agricultural sector remains a backbone of the economy. Nearly two dec- ades of productivity growth, beginning in the early 1990s, has helped put Ghana back on a path to recovery following more than a decade of economic uncertainty. With the exception of millet and sorghum, output for most crops has increased at a faster rate than population growth. During the 18-year period between 1993 and 2010, the sector experienced only 1 year (2007) of negative growth. During the same period, it recorded 3 years when growth exceeded 7 percent. The sector’s remarkable recovery, facilitated in part by sustained public and private sector investments, has helped pull thousands of rural households out of extreme poverty. In the early 1990s, nearly two out of every three (63.6 percent) rural Ghanaians lived below the national poverty line. By 2006, the ratio had dropped to roughly two in five (39.2 percent), according to the National Statistical Service. Ghana is now well on track to reach the first Millennium Development Goal to halve poverty by 2015. Sustaining the sector’s growth trajectory is a top priority for the recently elected admin- istration of President John Dramani Mahama. Success will depend, in part, on the government’s ability to manage the country’s ongoing transition to a more diversified economy while ensuring that the country’s smallholder farmers, food processors, and other sector actors have what they need to remain competitive. It also hinges upon the ability of all stakeholders to recognize, respond, and adapt to a changing landscape: one characterized by climate change, increasing weather variability, increasing threats from pests and diseases, and higher food price volatility, among other risks. The cata- strophic flooding of 2007 and more recent food price shocks served as stark reminders of the importance of effective risk management. The government recognizes more than ever the need to strengthen existing risk management systems not to only ensure continued sector growth, but also, and more important, to protect the most vulnerable communities and strengthen their resilience to future shocks. Improved agricultural risk management is one of the core enabling actions of the Group of Eight’s (G-8’s) New Alliance for Food Security and Nutrition. The Agricul- tural Risk Management Team (ARMT) of the Agriculture and Environment Services Department of the World Bank conducted an agricultural sector risk assessment to Risk Prioritization xi FIGURE ES.1. AGRICULTURE SECTOR GROWTH (%), 1980–2012 12 10 9.7 8 7.5 7.4 7.2 5.8 6.1 6 5.2 5.1 5.3 4.4 4.5 4.3 4.3 3.9 4.0 0 4.1 4 3.6 3.7 3.3 2.2 2.31.9 2.1 2 1.3 0.7 0.8 0.0 0 0 1980 1985 1990 99 1995 2000 2005 2010 –2 –1.2 –2.0 –1.7 –2.6 –4 –6 –5.5 –7.0 –8 Source: Bank of Ghana 2013. better understand the dynamics of agricultural risks and output and growth is relatively low at the broader, identify appropriate responses, incorporate agricultural sector level. In the 1980–2012 period, agriculture risk perspective into decision-making, and build capacity sector growth was positive in 24 out of 31 years of local stakeholders in risk assessment and management. (figure ES.1). Certain inherent strengths reduce This activity was requested by the G-8 and principally the sector’s overall vulnerability to risk while financed by the United States Agency for International limiting associated losses. First, the diversity of Development (USAID) and Feed the Futures programs. agro-climatic conditions in Ghana, of production Contributions were also received by the Multi-Donor systems, and of the crops and seeds used within Trust Fund on risk management, financed by the Dutch those systems lowers the level of aggregate risk Ministry of Foreign Affairs and the Swiss Secretariat of for the agricultural sector as a whole. Second, this Economic Affairs (SECO). diversity also reduces impacts on livelihoods when production shocks occur. However, it also means The objective of this assessment was to assist Ghana’s that the causes, frequency, and severity of risks government to 1) identify, analyze, quantify, and prior- vary between regions, commodities, and years, itize the principal risks facing the agricultural sector (that with strong implications for risk management. is, production, market, and enabling environment risks); 2. Disaggregated analysis by region and by crop 2) analyze the impact of these risks on key sector stake- showed a higher frequency of adverse production holder groups (for example, farmers, vulnerable popula- and price events. The indicative losses were also tions, food processors, government); and 3) identify and proportionally much higher than losses at the sec- prioritize appropriate risk management interventions tor level. Adverse events occur in most years for (that is, mitigation, transfer, coping) that will help improve some regions and commodities. However, these stability, reduce vulnerability, and increase the resilience events are usually offset by above-trend produc- of agricultural systems. The analysis covers priority crops tion in other regions and other crops, so reducing (and livestock) that are most important to farming families the overall impact of risk. and other stakeholders in Ghana. This report presents a 3. Whereas the adverse impact of agricultural risk at summary of the assessment’s key findings. the broader sector level is low, its frequent occur- 1. The analysis shows that although risk is a perma- rence causes significant income volatility, espe- nent feature of agriculture in Ghana, its impact on cially for low-income rural households engaged xii Ghana: Agricultural Sector Risk Assessment in rain-fed agriculture. It is also the principal in recent years have contributed to higher levels cause of transient food insecurity, especially in the of price volatility in domestic food markets. In northern regions. addition to maize, plantain, cassava, and yams 4. Multiple shocks cause the greatest losses, particu- are among the crops most susceptible to adverse larly when they are precipitated by drought or other impacts from price variability. weather-related risk events. For example, wide- 9. Among enabling environment risks, the assess- spread wildfires in 1983 following a severe, multi- ment calls attention to concerns over weak capac- year drought (1981–82) caused colossal crop losses ity among state-level institutions tasked to manage across the country, including 60,000 hectares (ha) and respond to the most important risks facing of cocoa trees. Catastrophic flooding in 2007 follow- the agricultural sector. First, the analysis calls ing prolonged drought conditions resulted in nega- into question the Ghana Cocoa Board’s (COCO- tive sector growth for the first time since 1994. BOD’s) ability to move forward to effectively man- 5. Low-income, rural households, especially in the age both production and price risk for cocoa; this northern regions, are most susceptible to produc- is occurring within a context of declining inter- tion and price shocks. With scant coping capacity, national prices and current budget shortfalls. they are also the most vulnerable to the impacts of Second, the assessment raises questions over the such shocks. Regional risk analysis (see appendix National Disaster Management Organization’s G) further showed that Upper East, Upper West, (NADMO’s) operational funding and its capacity and Northern regions are most prone to drought to respond to multiple risk events. and flooding, whereas the Eastern Region is rela- tively susceptible to fluctuations in maize and cas- This assessment offers the following preliminary recom- sava production. mendations for consideration based upon its analysis of 6. Given Ghana’s heavy reliance on rain-fed agricul- risks to various commodities, the regional distribution of ture, drought causes the highest level of cumula- vulnerability to risks, and the filtering of potential risk tive losses with the greatest impact on livelihoods, management measures: particularly in the northern savannah zones. 1. Promoting improved farming practices (for exam- Drought events include the late onset of rains, the ple, integrated pest management, or IPM), espe- early cessation of rains, and low cumulative rain- cially in the south, and conservation agriculture fall, and are most likely to affect sorghum, millet, measures (especially in the north). maize, and groundnuts. In addition, flash flood- 2. Strengthening improved seed (that is, drought, ing resulting from excessive rainfall occurs with pest, and disease resistant) development and dis- relative frequency across Ghana, but rarely causes tribution systems. widespread destruction. Crops most affected 3. Upgrading information systems to ensure avail- include cassava, rice, yams, and groundnuts. Exist- ability of timely and relevant weather, prices, and ing capacity among stakeholders to mitigate such pest and disease information to farmers, traders risks or cope in their aftermath is severely limited. and other stakeholders, coupled with relevant 7. Posing a constant threat to both crops and live- technical advice and knowledge. This also includes stock, pests and diseases constitute the second most market information about production, stocks, and important production risk after drought. Cassava, trade of different commodities. cocoa, and plantain are among those crops most 4. Promoting improved water management (for susceptible to attack (see appendix H). However, example, soil and water conservation measures) current control measures, in some cases with cocoa and irrigation (especially micro-level) and drain- and cassava, have been relatively effective. age infrastructure (especially, in flood-prone areas). 8. Price volatility poses the most important mar- 5. Strengthening extension systems (for example, ket risk facing agricultural stakeholders. This is face-to-face, information communications tech- especially true for maize; growing maize exports nology based, peer to peer) to ensure that farmers Risk Prioritization xiii have better access to technology, agronomic (MoFA), and the wider development community of the advice, and other resources needed to put in place most important risks facing the agricultural sector in new mitigation measures and improve existing Ghana. It is expected that the outputs of this assessment methods. will serve to inform the Medium-Term Agricultural Sec- 6. Improving infrastructure (on-farm and off-farm tor Investment Plan (METASIP) and its various compo- storage, warehouses, roads, and so on) to improve nents to ensure sustainability of agricultural investments productivity, reduce post-harvest losses, and help and enhanced agricultural resilience. It is also hoped that manage the risk of price volatility. the findings of this report will lead to improved decision 7. Considering recent news about phased with- making and successful implementation over time of a drawal of COCOBOD from a centralized disease comprehensive, coordinated, and ultimately effective risk control system and its potential consequences on management framework. pest and disease outbreak and cocoa production, a more systemic approach of pest and disease man- Many of the recommendations suggested in the report are agement is required to 1) ensure a smooth transi- already being considered or implemented and are having tion from a centralized system of pest and disease positive impacts, albeit at a lower, localized level. Greater control to an effective decentralized model that emphasis should be placed on scaling up these interventions is managed at the farmer and community levels: to the national level to make a more meaningful impact 2) improve farmer access to affordable and quality on the agricultural sector in Ghana. This would require fungicides and insecticides; 3) strengthen effective- understanding the landscape of these interventions, assess- ness of fungicide applications against black pod; ing their relative efficacy, understanding principal barriers and 4) facilitate improved insecticide application and challenges to success and scale up, and identifying lev- techniques against capsids and mirids, with an erage points and necessary interventions to increase access emphasis on combining good spray coverage with to a majority of agricultural sector stakeholders. Assessing minimal spray volumes. solutions to help prioritize specific interventions, scaling up priority programs, and putting in place a risk management It is hoped that this study will contribute to a better road map will be the next steps in the process of build- understanding among policy makers, government offi- ing resilience and reducing the vulnerability of households cials, including at the Ministry of Food and Agriculture adversely affected by agricultural risks. xiv Ghana: Agricultural Sector Risk Assessment CHAPTER ONE INTRODUCTION By most measures, agriculture remains a vital sector to the Ghanaian economy. The sector in 2012 accounted for 23 percent of gross domestic product (GDP), 56 percent of the labor force, and 35 percent of foreign exchange earnings.1 Following more than a decade of economic instability (1980–1992) punctuated by 5 years of negative growth, the sector has since grown at an average annual rate of 4.0 percent. Although falling well short of growth targets under the Comprehensive Africa Agriculture Development Programme (CAADP), the sector has nonetheless made an important contribution to economic growth and poverty reduction in recent decades. Moving forward, raising agricultural productivity—particularly among smallholder farmers who dominate the agricultural landscape—remains central to the govern- ment’s rural sector growth and overall economic development strategy, as outlined in its Medium-Term Agricultural Sector Investment Plan (METASIP). Continued growth is expected, driven by new investments in productivity-enhancing technologies and yield gains. The above narrative, however, masks uncertainties that pose a notable challenge to continued sector growth. The agricultural sector’s performance and share in most key socioeconomic indicators has been declining in recent years amid strong expansion in other sectors of the economy (figure 1.1; see also appendix H). Yields have mostly stagnated with increases in output mainly due to the expansion of cultivated area. The share of agriculture raw materials exports among total merchandise exports has dropped by roughly half since the mid-1990s, whereas imports have increased nearly fivefold during the same period. This slide is partly due to a seemingly unbreakable cycle of inadequate input supplies, inappropriate technology, low levels of savings and on-farm investment, and low output and productivity growth. Compounding these challenges is the high level of uncertainty that characterizes all things agricultural. Owing to a strong reliance on rain-fed, small-scale produc- tion systems that predominate, the sector is susceptible to downside risks. It is also 1 Statistical Review, Bank of Ghana, June 2013. Risk Prioritization 1 FIGURE 1.1. AGRICULTURE SECTOR TABLE 1.1. METASIP’S (2011–15) SIX PROGRAM PERFORMANCE, 2007–12 AREAS AND AGRICULTURAL RISKS Agriculture, value added (% of GDP) Agriculture, value added (annual % growth) METASIP Relevance for 40 (2011–16) Agricultural Risk 35 Program Areas Management 30 1. Food security Crop failures (due to droughts, pest/ 25 and emergency disease outbreaks, flood, and so on) 20 preparedness and price spikes are two principal 15 causes of transient food insecurity. 10 2. Increased growth Agricultural risk causes income 5 in incomes volatility for agricultural households. 0 3. Increased Risk management is crucial for 00 2007 2008 2009 2010 2011 2012 –5 competitiveness sustained competitiveness and market Source: Bank of Ghana; World Development Indicators Database 2014. and enhanced integration may increase exposure to market market risks. integration due to variations in markets and to other events outside 4. Sustainable Sustainable management of land and the ambit of agriculture. Smallholder farmers, market management of water resources is one of the important land and water instruments for managing production traders, agro-dealers, and other agricultural stakeholders risks. often have limited capacity to manage such risks or cope 5. Science and Many risk management solutions with resulting losses when shocks occur. Setting the sector technology require application of science and firmly on a path for future growth will thus require effec- application technology. tive ways to manage risks within Ghana’s agriculture sys- 6. Improved Integrated risk management will tems. It will also require strengthening the resilience of all institutional necessitate improved institutional stakeholders and ensuring that appropriate risk manage- coordination coordination. ment mechanisms (that is, mitigation, transfer, and cop- Source: Ministry of Food and Agriculture (MoFA); authors’ notes. ing) and related institutions are in place to support them. Furthermore, attaining METASIP objectives will require an explicit focus on agricultural risk since it cuts across all and Nutrition and in close partnership with partner coun- METASIP program areas (see table 1.1). tries. The objectives of this study are 1) to analyze the frequency and severity of different types of agriculture Improved agricultural risk management is one of the core risk (that is, related to production, market, enabling envi- enabling actions of the Group of Eight’s (G-8’s) New Alli- ronment) in Ghana; 2) to determine the indicative cost of ance for Food Security and Nutrition. In 2012, the G-8 these adverse events; and 3) to develop recommendations highlighted the need for conducting national agricultural on how best to manage the risks of greatest importance to sector risk assessments in close partnership with the New Ghana’s agricultural economy. Alliance countries (Ghana, Ethiopia, Tanzania, Mozam- bique, Ivory Coast, and Burkina Faso) to provide a robust Owing to the diversity of agro-climatic conditions and analytical underpinning to the countries’ agricultural related production systems in Ghana, the risk analysis development strategies and investment plans. required a combination of regional and commodity- specific approaches (see appendixes A and B). The study It is within this context that the World Bank, with support focuses on all 10 of Ghana’s administrative regions and from the G-8 and the United States Agency for Interna- a select basket of priority crops: cocoa, cassava, maize, tional Development (USAID), commissioned the present yams, groundnuts, plantain, sorghum, millet, and rice. study. It is one of a series of agricultural sector risk assess- These crops accounted for approximately 81 percent of ments that the World Bank agreed to conduct within the the area cropped and 76 percent of the value of gross framework of the G-8’s New Alliance for Food Security agricultural output in 2011 (FAOSTAT). Risks to livestock 2 Ghana: Agricultural Sector Risk Assessment FIGURE 1.2. AGRICULTURAL SECTOR RISK MANAGEMENT PROCESS FLOW Development of Implementation Risk Solution risk management and risk assessment assessment plan monitoring Desk review Desk review Development of Implementation plan by stakeholders In-country In-country assessment assessment Monitoring risks mission mission Incorporation into Dissemination Stakeholder existing govt. workshop workshop programs and development plans Training Source: Agricultural Risk Management Team of the World Bank. production were also analyzed but to a lesser extent due chapter 6 with an assessment of the priorities for risk man- to the limited availability of suitable statistics. The rela- agement and a discussion of risk management measures. tive effectiveness of existing risk management measures was also assessed via: 1) an appraisal of public interven- The prescribed methodology contains logical steps within tions in the rural sector, 2) discussions with rural stake- two consecutive phases (figure 1.2). Phase I, for which this holders directly involved in risk management, and 3) a study is the primary deliverable, has focused on identi- technical consultation on the relative benefits of risk miti- fying and prioritizing the major risks that cause adverse gation interventions (for example, scalability, sustainabil- shocks to the sector. Following in-depth analysis of base- ity, impact on poverty reduction). line data, the team conducted broad-based, in-country consultations with stakeholders in May–June 2013. These The study draws on, among other resources: rainfall data included individual farmers, farmer groupings, input sup- for the period 1981–2011 from the Ghana Meteorologi- pliers, market traders, food processors, and representa- cal Service; national crop production data for the period tives of the government and of research institutes. The 1991–2011 and national producer price data for the results of this assessment will provide the conceptual basis period 1991–2010 from FAOSTAT; regional crop pro- for Phase II, during which a team of specialized experts duction data for the period 1992–2008 from the Ministry will be fielded to deepen the analysis and develop a multi- of Food and Agriculture (MoFA); archives of the National tiered strategy for managing the priority risks. Disaster Management Organization (NADMO); and qualitative data collected through direct consultations By the end of this activity, the World Bank in coordina- with stakeholders. tion with the government of Ghana (GOG) and sector stakeholders will have developed and validated a matrix The report begins with an overview of agriculture in of priority interventions related to risk mitigation, trans- Ghana in chapter 2, followed by an assessment of the fer, and coping, within a comprehensive risk management main agricultural risks in chapter 3. Chapter 4 analyzes framework. The outputs of this assessment will serve to the frequency and severity of the major risks identified inform the ongoing METASIP and its various compo- and assesses their impact. Stakeholder perception of these nents to ensure sustainability of agricultural investments risks is examined in chapter 5. The study concludes in and enhanced agricultural resilience over time. Risk Prioritization 3 CHAPTER TWO OVERVIEW OF AGRICULTURAL SYSTEMS IN GHANA Providing context for analysis and discussion of agricultural sector risk, this chap- ter presents an overview of the agricultural sector in Ghana. Sector characteristics most pertinent to risk are thus given particular attention. Analysis primarily covers the period 1991–2010 to assess the frequency and severity of the most important risks. The agriculture resource base is characterized by an abundance of land and diverse agro-ecological conditions. Of the 13.7 million hectares of agricultural land, only 7.85 million hectares (58 percent) are under cultivation. Owing to the diversity of agro-ecological conditions, crop production ranges from millet and sorghum in the semi-arid north, to maize, cassava, and other root crops in central Ghana, and cocoa, plantain, palm oil, and rubber in the forest zones of the south. These conditions also facilitate surplus production of most crops. Livestock production is of lesser impor- tance, representing 7.5 percent of agricultural GDP (including cocoa). The high proportion of unused agricultural land also highlights some of the major constraints that the sector faces: low levels of mechanization, low soil fertility, and limited access to water for irrigation. Roughly 90 percent of farms in Ghana are small (< 2 ha) and rely on manual labor or animal traction. Much of the land in the north and center of Ghana (approximately two-thirds of the total land area) consists of highly weathered soils with low fertility and low water-holding capacity. Only 30,000 hectares are irrigated, equivalent to 0.2 percent of total agricultural land. These constraints limit the ability to raise output and increase vulnerability to drought. AGRO-CLIMATIC CONDITIONS There are six agro-ecological zones (figure 2.1), of which five are important for agri- culture. They range from the hot, dry savannah conditions in the north to tropical and deciduous forests in the south and southwest (table 2.1). The northern savan- nah regions are hot and dry with a uni-modal rainfall distribution, and a growing season of 200 to 240 days. Agriculture is demanding in these regions. Agro-climatic Risk Prioritization 5 FIGURE 2.1. ADMINISTRATIVE REGIONS AND prone to drought. This region is of limited importance AGRO-ECOLOGICAL ZONES for agriculture. The savannah and transitional zones are mostly flat to undulating, broken only by the shallow drainage basins of the Volta river system in the center and to the west. Veg- etation is light savannah forest. The soils are light, highly weathered loams or sandy loams with low organic mat- ter, low mineral fertility, and low water-holding capacity. Topography, vegetation, and soil types then change when moving south into the forest zones. The land becomes more undulating and deciduous forests predominate in most areas except for the rain forest zone in the south- west. Soil fertility improves due to higher organic matter and mineral fertility and the soils are more friable and bet- ter suited to agriculture. The large deciduous forest zone is highly suited to production of cocoa, other tree crops (palm, rubber), plantain, root crops, and high-value fruit and vegetable crops for export. These characteristics have three important implications Source: Adapted from World Food Programme 2009. for agricultural sector risk. First, the diversity of agro- climatic conditions significantly reduces the level of conditions improve gradually moving from north to covariate risk for the sector as a whole. Second, the wide south, with increasing rainfall and the emergence of a diversity of crops grown enhances the level of variabil- bi-modal rainfall distribution. The central and southern ity in the frequency, severity, and causes of production regions have a longer growing season (250–330 days) and risk between regions and between years. Drought and greater potential for double cropping. The exception is fire risks are much higher in the northern regions, for the small coastal savannah region in the south, including example, owing to lower rainfall and the uni-modal rain- greater Accra, which has very low rainfall and is highly fall distribution. The drier conditions in the north also TABLE 2.1. AGRO-ECOLOGICAL ZONES OF GHANA (NORTH TO SOUTH) Zone Rainfall (mm) Production System Area (km2) Sudan savannah 800–1,200 Sorghum, millet, groundnut, cattle, small 2,200 (0.9%) (unimodal) ruminants Guinea savannah 800–1,200 Sorghum, millet, maize, groundnut, cattle, small 147,900 (61.9%) (unimodal) ruminants Transitional zone 1,100–1,400 Maize, cassava, yam, small ruminants 8,400 (3.5%) (bi-modal) Deciduous forest 1,200–1,600 Cocoa, cassava, maize, plantain, small ruminants 66,000 (27.8%) (bi-modal) Rain forest 800–2,800 Cassava, yam, plantain, small ruminants 9,500 (4.0%) (bi-modal) Coastal savannah 600–1,200 Not applicable 4,500 (1.9%) (bi-modal) Source: MoFA 2010. 6 Ghana: Agricultural Sector Risk Assessment FIGURE 2.2. MONTHLY RAINFALL PATTERNS BY REGION 300 Zuarungu Station - Upper East Region 250 Babile Station - Upper West Region 200 Nyankpala Station - Northern Region 180 250 200 160 Cumulative rainfall Cumulative rainfall 140 Cumulative rainfall 200 150 120 150 100 100 80 100 60 50 40 50 20 0 3 10 43 99 138 173 280 189 58 5 2 2 6 25 64 106 134 184 233 195 65 8 5 2 8 32 76 109 135 156 168 179 71 6 2 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 250 Ejura Station - Brong Ahafo Region 250 Kwadaso Station - Ashanti Region 200 Begoro Station - Eastern Region 180 200 200 160 140 Cumulative rainfall Cumulative rainfall Cumulative rainfall 150 150 120 100 100 100 80 60 50 50 40 20 7 25 99 150 174 181 132 90 214 169 38 20 28 51 114 151 169 209 127 86 177 195 96 41 23 53 122 145 158 184 124 100 172 159 49 24 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 250 Akaa Station - Volta Region 250 Breman Asikuma Station - Central Region 250 Sefwi Wiawso Station - Western Region 200 200 200 Cumulative rainfall Cumulative rainfall Cumulative rainfall 150 150 150 100 100 100 50 50 50 8 28 91 126 151 190 177 193 175 151 49 24 19 63 119 135 164 214 118 72 136 180 114 48 23 40 127 128 193 213 125 72 155 208 74 28 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 160 Afienya Station - Greater Accra Region 140 120 Cumulative rainfall 100 80 60 40 20 15 21 60 70 137 148 56 25 72 92 60 34 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Ghana Meteorological Service; World Bank. increase the risk of insect pests such as armyworm and variegated grasshopper. Drought risk falls moving from RAINFALL PATTERNS north to south, but the risk of pests and diseases such AND TRENDS as stem borer, capsid, black pod, sigatoka, and fruit fly Analysis of rainfall patterns for the period 1981–2010 increases because of higher temperature and humidity. confirms the regional differences in rainfall distribu- Third, diversified livelihood systems and income levels tion (figure 2.2). In the savannah zones, most rainfall make some regions less prone and less vulnerable to risks occurs during the summer months (June–September), than other regions. followed by a prolonged low rainfall period from Risk Prioritization 7 FIGURE 2.3. COMPOSITION OF CROP TABLE 2.2. TRENDS IN CROP PRODUCTION, PRODUCTION 1990–2011 Production Area Yield 2,000,000 Sorghum 70% 6.5% 62% 1,800,000 Cocoa beans 1,600,000 Millet 131% 8.7% 110% 1,400,000 Maize Maize 133% 84% 28% Area (ha) 1,200,000 Cassava 1,000,000 Yams+T Y Taro Rice 287% 151% 54% 800,000 Groundnuts pa Oil palm Cassava 217% 106% 58% 600,000 Plantai Plantain g table Vegetabless Fr F ui it Yam 243% 125% 63% 400,000 Sorghum s Pulses 200,000 Millet Plantain 261% 118% 67% 0 Other Other Ot r Rice Groundnuts 449% 207% 82% 00 01 20 20 02 03 4 20 20 0 5 06 Cocoa 155% 133% 9% 20 00 07 2 20 20 08 09 Year Y 20 20 10 1 20 20 1 Source: FAOSTAT; average of 1990–91 vs 2009–11. Source: FAOSTAT 2011. important. As rainfall increases, maize gradually replaces November–March. A bi-modal rainfall pattern is observed sorghum and millet as the major food crop in the Guinea in the transitional and forest zones with the main rainfall savannah, the largest agro-climatic zone. from March–July, followed by a minor rainy season from September–November. Moving south into the transition zone, cassava and yam begin to replace maize as the major food crops and live- Analysis of the main growing season (May–July) rainfall stock production becomes less important. Cocoa domi- data for the period 1970–2008 shows that there has been nates land use in the deciduous forest zone, with cassava, no secular trend in rainfall for any of the 10 regions (see maize and plantain as the main food crops. These food appendix I). crops also predominate in the high rainfall forest zone, which is less suited to cocoa production. The small coastal CROP PRODUCTION savannah region is unsuited to agricultural production. SYSTEMS Most farmers grow a range of food and cash crops. Pro- Three commodities dominate production: cocoa, with duction risks are reduced as a consequence, both at the 24 percent of total area, maize with 15 percent, and cassava farm level and for the agricultural sector as a whole. Farm- at 13 percent (figure 2.3). Cocoa accounts for the largest ers’ ability to diversify also allows them to change their area and the bulk of agriculture export earnings, whereas crop composition quite readily in response to changes in maize and cassava are the main food staples. The remain- the profitability of any given crop. ing land is planted to more than 40 other food and cash crops, with yams, oil palm, groundnuts, vegetables, and PRODUCTION TRENDS plantain the most important. With the exception of cocoa Crop production has grown steadily, with an average and groundnuts, where the area has fluctuated in the past, annual increase in the crop production index of 12.9 per- the composition of crop production is fairly stable. This cent from 1990–2011 (see figure 3.1), according to the combination of diversity and stability ensures an adequate World Development Indicators. This growth has been supply of staple foods at aggregate level. driven largely by area expansion, with the total cultivated area increasing from 2.9 million ha in 1990 to 6.76 million There is less diversity at the regional level, especially in ha in 2011 (FAOSTAT). Crop yields increased little for the north where agro-climatic conditions are less favora- most of this period, but have risen since the mid-2000s. ble. Sorghum and millet are the main food crops in the Yields remain relatively low, despite this increase. The Sudan savannah, with groundnuts as the main cash crop area expansion appears to be largely a result of popu- (table 2.2). Livestock production, especially cattle. is also lation increase in the rural areas. This has resulted in a 8 Ghana: Agricultural Sector Risk Assessment TABLE 2.3. COEFFICIENTS OF VARIATION FOR crops that predominate in the drier, savannah zones CROP PRODUCTION, 1990–2011 (sorghum, millet, groundnuts), as would be expected. Cassava, yam, and plantain exhibit the lowest lev- Production Area Yield els of variability, consistent with the higher drought Sorghum 0.20 0.13 0.16 resistance of root crops and the higher rainfall zones Millet 0.23 0.10 0.18* in which these crops predominate. The variability of Maize 0.13 0.08* 0.07* cocoa and maize production is intermediate between Rice 0.15* 0.11* 0.22* these two groups. Cassava 0.09* 0.06* 0.06* Yam 0.13* 0.11* 0.10* Trends in production and production variability are illus- Plantain 0.07* 0.05* 0.06* trated further for the three main crops (that is, cocoa, Groundnuts 0.24* 0.25* 0.15* maize, and cassava), and in figures 2.4, 2.5, and 2.6. The Cocoa 0.17 0.15* 0.14 higher levels of variability of cocoa relative to maize and Source: FAOSTAT. cassava are evident, and variability in both area and yield *Adjusted for trend using the Cuddy-Della Valle Index. are responsible for the interannual variation of cocoa production. Yield variation appears to be the main deter- limited overall change in the nature and composition of minant of variability in maize production. Cassava pro- production, as the small-scale subsistence farmers who duction is characterized by low levels of variation in both dominate production tend to retain a diversified crop mix area and yields and hence in overall production. It is also when they expand. notable that production drops for the three crops occur Production of root crops, plantain, groundnuts, and in different years; evidence of the generally low levels of rice has increased the most, whereas that of traditional covariate risk in Ghana as a result of its agro-ecological cereal crops has grown less rapidly. This gradual shift to diversity. root crops has improved the stability of the food supply and resulted in a more varied food diet—both of which FOOD SUPPLY AND DEMAND improve food security. Ghana currently produces a surplus of most food crops (table 2.4). Rice is the only food staple for which PRODUCTION VARIABILITY there is a structural deficit, with imports accounting for Comparisons of production variability, as measured approximately 55 percent of total consumption. Ghana by coefficients of variation, show that most of the also imports about half of its meat requirements, both main crops exhibit moderate to low levels of interan- through imports of meat from the world market and sub- nual variation (table 2.3). Variability is highest for the stantial (and under-reported) flows of live animals from FIGURE 2.4. COCOA PRODUCTION, 1990–2011 2,500,000 Area (ha) Prod (tonnes) Yield (t/ha) 0.500 0.450 2,000,000 0.400 Area/Production 0.350 1,500,000 0.300 Yield 0.250 1,000,000 0.200 0.150 500,000 0.100 0.050 0 0.000 1990 1991 1992 1993 1994 1995 1996 1997 1998 2099 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 11 19 Source: FAOSTAT. Risk Prioritization 9 FIGURE 2.5. MAIZE PRODUCTION, 1990–2011 Area (ha) Prod (tonnes) Yield (t/ha) 2,000,000 2.00 1,800,000 1.80 1,600,000 1.60 1,400,000 1.40 Area/Production 1,200,000 1.20 Yield 1,000,000 1.00 800,000 0.80 600,000 0.60 400,000 0.40 200,000 0.20 0 0.00 19 0 19 1 19 2 19 3 94 19 5 19 6 97 19 8 20 9 20 0 20 1 02 20 3 20 4 20 5 20 6 20 7 20 8 20 9 20 0 11 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 1 19 19 19 20 Source: FAOSTAT. FIGURE 2.6. CASSAVA PRODUCTION, 1990–2011 Area (ha) Prod (tonnes) Yield (t/ha) 16,000,000 18.00 14,000,000 16.00 12,000,000 14.00 Area/Production 12.00 10,000,000 10.00 Yield 8,000,000 8.00 6,000,000 6.00 4,000,000 4.00 2,000,000 2.00 0 0.00 19 0 19 1 92 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 08 20 9 20 0 11 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 1 19 19 20 Source: FAOSTAT. TABLE 2.4. DOMESTIC FOOD SUPPLY AND DEMAND FOR FOOD STAPLES Domestic Available for Human Estimated Production Consumption Aggregate Demand Deficit/Surplus Commodity (000 mt) (000 mt) (000 mt) (000 mt) 2009 2010 2009 2010 2009 2010 2009 2010 Maize 1619.6 1871.7 1197.7 1310.2 1052.1 1060.9 145.6 249.3 Rice (milled)* 234.9 294.9 204.3 256.6 576.5 581.4 −372.2 −324.8 Millet 245.5 218.9 213.6 190.5 24.0 24.2 189.6 166.3 Sorghum 350.5 324.4 304.9 282.3 12.0 24.2 292.9 258.1 Cassava 12,230.6 13,504.1 8,561.4 9,452.9 3,672.9 3,703.7 4,888.6 5,749.2 Yam 5,777.8 5,960.5 4,622.2 4,768.4 1,006.5 3,027.9 3,615.7 1,740.5 Cocoyam 1,504.0 1,354.8 1,428.8 1,287.1 960.9 968.9 467.9 318.2 Plantain 3,562.5 3,537.7 3,028.1 3,007.1 2,030.0 2,054.1 991.12 953.0 Groundnut 204.9 530.9 174.2 477.8 120.1 290.7 54.1 187.1 Sources: MoFA, 2009 Annual Progress Report; Agriculture in Ghana: Facts and Figures (2010). Note: mt = metric ton. *Sixty percent of paddy rice. 10 Ghana: Agricultural Sector Risk Assessment Burkina Faso, Mali, Niger, and other countries. Ghana’s FIGURE 2.7. TRENDS IN REAL CEREAL agricultural sector is also closely linked to major regional PRICES, 1991–2010 flows of primary commodities. Maize flows across the Maize Millet Rice (paddy) Sorghum 700 border according to the year and season. In addition to 600 live animals, inflows of cowpeas may also be significant. 500 As in neighboring countries, significant shifts in regional Cedi/Ton food production directly impact prices and other market 400 T dynamics in Ghana. 300 200 100 AGRICULTURAL MARKETS 0 AND PRODUCER PRICE 19 1 92 19 3 19 4 19 5 96 19 7 19 8 20 9 00 20 1 20 2 20 3 20 4 20 5 06 20 7 20 8 20 9 10 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 19 19 20 20 TRENDS Source: FAOSTAT. There are active markets for all major commodities, including a strong demand from international mar- with domestically produced white maize purchased inter- kets for Ghanaian cocoa. Surplus food commodities are mittently for animal feed. Sorghum and millet are grown exported to neighboring countries, particularly to the food in the northern regions, as these crops are more tolerant deficit countries north of Ghana such as Burkina Faso, of moisture stress. Both commodities are actively traded, Niger, and Mali. Trade with the border regions of Côte including export to neighboring countries. d’Ivoire and Togo is also active, but on a much smaller scale. Private traders buy surplus production at the vil- Real producer prices for cereals have increased steadily lage level for sale at regional markets throughout Ghana. since 1991, accompanied by an increase in price vari- They appear to collude in some cases (for example, yams) ability (figure 2.7). Note also the extent to which cereal to keep producer prices low, and to limit the number of prices track each other, indicating how close they are as traders to preserve their monopsony powers. Domestic food substitutes. The price spike in 2001 is due to the markets have become deeper and more efficient, nev- combined impact of high inflation and devaluation of the ertheless, in response to improved infrastructure (roads, cedi (World Food Program 2002), and in 2005 to localized communications) and the growth of numerous regional droughts and bushfires in northern Ghana. In contrast, market centers in northern and central Ghana. the price spike in 2008 was exogenously driven, reflect- ing the impact of the global food price crisis. In general, The analysis of price trends was based on national prices rice prices exhibit lower interannual variability as the reli- for the period 1991–2010, using FAOSTAT data. Real ance on imports results in a more stable supply. This pat- prices are used for analysis (deflated by the consumer price tern was broken by the global food crisis, however, which index), as very high inflation during this period makes it resulted in sharp price changes both during and after the difficult to draw useful conclusions from trends in nominal crisis. prices. Producer prices of the main root crops also tend to move FOOD CROPS together (figure 2.8) as they are close substitutes for con- The cereals market is dominated by maize and rice, but sumption and are grown in similar agro-climatic zones. the markets for all cereals are active and competitive. Root crop prices have become more stable since 2000 Maize is the second most important food staple after in response to increased supply and greater potential for cassava and is sold on domestic markets and for export exports. The higher variability of plantain production, as to neighboring countries. Demand for maize for feed is a result of frequent storm damage, results in more varia- also increasing for the poultry industry. The poultry feed tion in producer prices. Plantain is also sold on domestic market is dominated by imported yellow maize, however, markets and for export. Risk Prioritization 11 FIGURE 2.8. TRENDS IN REAL PRICES FOR FIGURE 2.9. PRODUCER PRICES FOR COCOA ROOT CROPS AND PLANTAIN, AND GROUNDNUTS, 1991–2010 Cocoa beans (real) Groundnuts 1991–2010 2500 Cocoa beans (nominal) 600 Cassava Y Yams Plantains 2000 500 Cedi/Ton 400 1500 T Cedi/Ton T 300 1000 200 500 100 0 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 03 20 4 05 20 6 20 7 20 8 20 9 10 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 20 20 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 20 9 10 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 19 Source: FAOSTAT; COCOBOD. Source: FAOSTAT. The increasing volume of food crop exports has intro- locally and it remains an important cash crop for farmers duced an additional source of price volatility to domestic in the north. Groundnut prices are set freely, with minimal food markets. This is particularly true for cereal crops, first increase in real prices since 1991. because production is concentrated in the more drought- prone transition and northern regions, which results in Of the other main cash crops, most palm fruit is sold substantial variation in the size of the marketable surplus. for processing for the domestic market. Four large-scale, Second, the level of demand for cereal imports from Bur- privately owned, corporate farming and processing enti- kina Faso, Niger, and Mali is also highly variable because ties meet the domestic demand for refined oil (170,000– of the even more drought-prone conditions in which cere- 180,000 tons per year) and small-scale informal processors als are produced in these countries. supply a similar volume of lower quality (unrefined oil). There is no government intervention in the palm oil markets. Cotton is also sold on international markets but CASH CROPS prices are set freely, with no government intervention. The markets for cash crops differ, depending on the reli- ance on domestic versus export markets and the extent to which government intervenes in these markets. There LIVESTOCK PRODUCTION is a high demand for Ghanaian cocoa on international The livestock sub-sector is estimated to account for about markets. Producer prices for cocoa beans are set by the 7 percent of the nation’s agricultural gross domestic Ghana Cocoa Board (COCOBOD) in local currency, product. However, anecdotal evidence suggests that offi- based on international prices. Nominal producer prices cial livestock projections are overly conservative, given have never been reduced, although they frequently fall in large inflows and settling in Ghana of livestock from real terms when the annual adjustment is insufficient to neighboring countries (large and small ruminants). The compensate for inflation (figure 2.9). Despite these varia- proportion is likely somewhat higher (approximately 8.6 tions, real producer prices have increased by more than percent, according to authors’ estimates). Nonetheless, the 170 percent since 1991. The risk associated with variation sub-sector is a significant source of income, meat, milk, in international prices is assumed by COCOBOD. organic fertilizer, and means of savings for rural house- holds, especially in the northern part of the country. A high proportion of groundnuts were exported to west- ern markets until the mid 2000s when this market was lost Livestock distribution and production systems have due to high alfatoxin levels. Most production is now sold been strongly influenced by geography and climate. 12 Ghana: Agricultural Sector Risk Assessment The greatest numbers of large and small ruminants pigs.2 With an estimated 800,000 swine in the north- have historically been found in the Guinea and Sudan ern regions, the national total is also likely well above savannah ecological zones spanning the Northern, official estimates. Upper West, and Upper East regions (together mak- ing up just over 40 percent of Ghana’s land area). The poultry sector is sharply divided between family Conditions there remain very favorable to extensive holdings and small- to large-scale industrial poultry farms animal husbandry; 10-year cumulative rainfall aver- (for broilers, layers, and even a few chick or guinea fowl ages in these regions range from 1,200 mm per year in hatcheries).3 Smallholder production is highest in the for- the Northern Region to about 940 mm per year in the est agro-ecological zone and in northern savannah areas. Upper East and West. Nationwide, chicken production (ranging from the exten- sive village systems to semicommercial groups) represent According to the 1996 national livestock census, the spa- between 11 percent and 13 percent of rural household tial distribution of livestock was well established, with incomes.4 nearly three-quarters of all cattle concentrated in the Northern, Upper West, and Upper East regions. Cat- Most poultry (an estimated 80 percent) is traditionally tle density is highest in the Upper East, where land is raised; the remainder is produced commercially, especially less suitable for agricultural production relative to other in the Ashanti and Greater Accra regions. Commercial northern regions. birds are primarily raised for eggs, as domestic produc- tion for poultry meat currently faces stiff competition from Pastoralism is the dominant form of livestock system, U.S., Brazilian, and European imports. Most commercial especially in the northern part of the country. Large- poultry operations are located around the urban areas scale north-south-north migration of livestock con- of the Greater Accra and Ashanti regions. There are an stitutes an important vector for contagious diseases. estimated 380 large-scale farms, each stocking more than Farmers with farmland in or near the traditional graz- 10,000 birds. Most are egg producers, although a limited ing corridors also complain widely of damage to their number raise exotic breeds of broiler chickens, guinea crops from cattle that graze or trample cropland in fowl, and turkeys for meat. Such operations manage their their path. Itinerant herders are typically Fulani herds- own feed mills, and some maintain hatcheries and par- men who bring their cattle south from Niger, Mali, ent stocks. In addition, there are nearly 1,000 small- to and Burkina Faso each year. This migration follows a medium-scale facilities (consisting of 50 to 10,000 birds) tradition of transhumance grazing that extends from that rely on external suppliers for day-old chicks and feed. the Sahel to the northern reaches of the forest zones Currently, there are an estimated 11 million chickens in in Ghana, Côte d’Ivoire, Togo, and Benin. They move industrial operations. Their annual feed consumption can south after the rainy season as the pasture in the Sahel amount to 600,000 tons of maize per year. Some portion dries out, selling many of their cattle for slaughter in of this is imported as yellow maize,5 a major source of the larger urban markets as they go. They then return feed for commercial poultry producers, but the majority north when the rainy season starts again. Their herds comes from domestic production (white).6 are typically large, often with several hundred cattle, many of which are owned by farmers who contract the 2 CFSVA 2012, WFP-MoFA. 3 Approximately 87 percent of chicken producers are in rural areas, of which 97 Fulani to herd them. percent are smallholders with less than 500 birds (IFPRI/DFID 2008). 4 IFPRI/DFID 2008. Swine production is widespread; most pig stock is 5 Annual maize imports in the past 10 years have ranged between 10,000 and 60,000 tons. held at the household level but there is some indus- 6 Aside from Nigeria (50 percent of West Africa’s total), Ghana is by far the larg- trial production in peri-urban areas. Swine owner- est regional producer of maize, but the relative share of maize in total human ship is also widespread in northern regions where the cereal consumption in Ghana is only 37 percent (FAO 2008, 2009 commodity average livestock-owning rural household keeps two balances). Risk Prioritization 13 access to inputs and financial services further contributes PRINCIPAL CONSTRAINTS to low adoption of productivity-enhancing technologies. TO AGRICULTURAL Underdeveloped road networks, especially rural feeder PRODUCTION roads, constrain farmers’ access to markets. Inadequate Ghana relies heavily on rain-fed agriculture and low- storage infrastructure further reduces farmer incentives input, low-output smallholder systems (90 percent, or less to invest in modern inputs. For the livestock sub-sector, than 2 ha). Soils are coarse with low water-holding capac- chief constraints include low-performing breeds; insuffi- ity. In the absence of good water management (less than cient feeding; high cost of poultry feed; poor husbandry 0.2 percent, or 30,000 ha, of agricultural land is irrigated), management; strong competition from imports; and poor crops are often subject to water stress during the growing post-production management. These constraints hinder season. Low agricultural productivity in Ghana is largely sector growth by limiting producer ability to raise out- attributed to low soil fertility and limited farmer use of fer- put. They can also amplify the impacts of adverse shocks tilizers, improved seeds, and agro-chemicals (for example, when they occur by weakening the capacity of various insecticides). There is a high reliance on family labor in stakeholders to manage their exposure and recover from the absence of mechanized equipment and services. Poor resulting losses. 14 Ghana: Agricultural Sector Risk Assessment CHAPTER THREE AGRICULTURAL SECTOR RISKS The main sources of agriculture risk are reviewed in this chapter: production risk, market risk, and a general set of risks associated with the enabling environment for agriculture. The incidence and implications of multiple or successive shocks are also considered. PRODUCTION RISKS Drought, floods, bushfires, and pests and diseases are the main sources of production risk. The incidence of these and other adverse events is shown in figure 3.1, based on reports of adverse events for 1980–2011. Drought emerges as the most common source of major production shocks, followed by pests and diseases and floods. Related risk events may occur in isolation, but can also present as multiple, overlapping shocks, with far greater impacts and higher associated losses. WEATHER VARIABLES Drought An agricultural drought occurs when a deficit of soil moisture significantly reduces crop yields. It can occur in response to low overall annual rainfall or to abnormalities in the timing and distribution of annual rainfall. Inadequate rainfall at key periods during the crop production cycle (seeding, flowering, and grain filling) affects crop yields, even when overall rainfall is comparable to long-term norms. During these peri- ods, a soil moisture deficit as short as 10 days can have a major impact on crop yields. Drought is typically defined relative to some long-term average balance between precipitation and evapotranspiration, which is considered “normal” for a particu- lar location at a particular time of year. Drought is thus a relative concept in that sub-optimal soil moisture levels and crop yields in one agro-climatic area may be acceptable in another. For purposes of analysis, standardized cumulative rainfall for the period March–October for each region was calculated for the period 1981– 2010, with drought defined as rainfall less than one standard deviation from the mean and severe drought as rainfall less than two standard deviations from the mean. The results are presented in table 3.1. The authors also conducted an in- depth analysis of rainfall patterns and crop production using data collected from Ghana’s weather stations (box 3.1). Risk Prioritization 15 FIGURE 3.1. ADVERSE CROP PRODUCTION intensity rainfall or river valley flooding along the main EVENTS, 1980–2010 waterways. Both can cause severe damage to property Crop production index (2004–06 = 100) and livelihoods, although usually with a localized and 140 therefore limited impact on aggregate crop production. 120 F Localized flash flooding occurs in most years but it was 100 L D marked in 1989, 1991, 1994, and 1999. Areas located 80 R within the Volta River basin in the Northern Region and D & 60 R C P D & P C D in the southwestern river system in the Western and Cen- 40 B D D DB F R tral Regions are particularly prone to seasonal flooding. 20 R R F 0 The most severe recent flood occurred in 2007, when 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 heavy late rains in September led to the inundation of 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 Source: FAOSTAT. vast areas across West Africa (see table 3.2). In Ghana, the Note: DR = Drought; BF = Bushfire; PD = Pests and Disease; CC = Civil Conflict; FL = Flood. problem was exacerbated by the release of water from the Bagre Dam in Burkina Faso into the White Volta River. A MoFA and United Nations Joint Preliminary Assess- In absolute terms, the risk of drought increases from ment Report estimated that the floods, which followed a south to north, as agro-climatic conditions change from prolonged drought, destroyed about 70,500 hectares of the higher rainfall, bi-modal rainfall distribution in the farmland. Approximately 160,000 metric tons of crops south and center to the lower rainfall and uni-modal rain- (including corn, sorghum, millet, peanuts, cowpea, yams, fall pattern in the north. All regions experience drought cassava, and rice) were lost. A subsequent assessment measured in relative terms, however, as indicated in estimated a 7 percent decline in the national harvest, table 3.1. Since 1980, severe, countrywide droughts have primarily due to a drop of approximately 15 percent in occurred in 1982, 1983, 1990, and 1998. The 1982–83 the drought- and flood-affected northern regions.7 This drought was particularly severe and was followed by huge shortfall resulted in acute food shortage in affected com- bushfires. Further, regional droughts occurred in 1986, munities. 1992, and 2005. In Ghana, droughts result in a severe, but not catastrophic, impact on cereal production. Figure 3.2 River flooding has subsequently assumed a higher profile highlights the adverse impact of droughts and flood on because of the continued release of surplus water from national cereal production and yield, particularly in 2007. the Bagre Dam, which causes flooding downstream in northern and eastern Ghana. This also occurred in 2009 A closer analysis of the relationship between rainfall and and 2010, resulting in the loss of many lives and exten- crop yield, described in detail in appendix I, shows that the sive property damage. Ghana is currently trying to reach volume of rainfall, starting date of the rainy season, and an agreement with Burkina Faso to time and manage this the intensity of rainfall are most closely related to yield. release so that it is not a risk to lives and property. The impact of these three parameters on yield varies by region and by crop. But they are never a major determi- nant of yield, even in the savannah regions where their Bushfires impact is strongest. In the southern regions, the variance Fire is widely used by rural people to clear land for culti- of rainfall explains very little of the variation in yield, vation, improve grazing, and facilitate hunting. Burn-offs suggesting that factors such as pests and diseases may be also help to control pests and diseases such as grasshop- more important determinants of yield. pers, locusts, ticks, anthrax, and livestock parasites. In for- ested areas, fire affords an easy (low-labor) way to open up new land, facilitate mechanized cultivation, and provide Floods Flooding also poses a risk for agriculture production, in 7 Ghana Grain and Feed Update 2008, USDA Foreign Agricultural Service, the form of flash flooding as a result of sudden, high- GAIN Report, Global Agriculture Information Network, January 22, 2008. 16 Ghana: Agricultural Sector Risk Assessment TABLE 3.1. FREQUENCY OF LOW RAINFALL EVENTS BY REGION,* 1981–2010 Upper Upper Brong- Greater Number Year West East Northern Ahafo Volta Ashanti Eastern Western Central Accra of Events 1981 −0.75 −1.56 0.01 −0.50 0.17 0.27 0.10 0.55 1.03 0.52 1 Risk Prioritization 1982 −0.07 −0.03 −0.78 −1.33 −1.83 −1.73 −1.77 −0.98 −0.05 0.62 4 1983 −3.07 −0/81 −1.55 −1.24 −2.09 −1.99 −1.90 −1.58 −2.57 −2.39 9 1984 −1.91 −1.53 −0.26 −0.17 0.95 1.78 0.77 1.72 0.50 0.84 2 1985 −0.09 −0.98 0.28 0.57 −0.20 0.85 0.74 −0.04 −0.14 0.50 0 1986 0.69 −0.42 −0.66 −0.26 −0.88 0.73 −1.07 −1.31 −1.79 −1.18 4 1987 0.11 −0.46 −0.27 0.75 0.49 0.83 1.38 2.15 1.40 0.86 0 1988 −0.57 −0.20 0.29 −0.35 0.47 −0.90 −0.06 −0.12 −0.06 0.40 0 1989 0.12 1.39 2.23 1.35 1.30 0.25 0.40 0.92 0.80 0.80 0 1990 — −1.22 −1.00 −0.86 −0.30 −1.51 −1.02 0.06 −1.38 −0.68 5 1991 −0.38 0.58 2.06 0.14 1.03 −0.26 1.85 −0.17 1.62 2.45 0 1992 −0.23 0.57 −1.20 −1.27 −0.97 −0.86 −0.85 −1.05 −0.91 −0.96 3 1993 0.28 −0.31 0.04 -0.26 −0.09 −0.35 1.16 −0.31 −0.35 −0.42 0 1994 0.22 1.18 0.05 −1.51 −0.41 −0.46 −0.04 −1.64 0.24 −0.88 2 1995 1.19 −0.72 0.31 0.37 1.76 0.48 1.71 0.03 0.77 0.47 0 1996 −0.81 1.07 0.60 0.01 −1.00 −0.71 −0.46 0.66 0.21 0.92 0 1997 0.62 −0.18 −0.01 0.36 0.09 −0.42 −0.58 −0.31 0.28 0.91 0 1998 −0.52 0.45 −2.08 −0.01 0.16 −0.51 −0.30 −1.28 −1.01 −1.17 4 1999 1.45 1.78 1.43 0.63 0.74 1.50 −0.09 −0.11 0.52 0.51 0 2000 1.85 −0.09 0.01 −0.07 0.71 −0.56 −0.35 0.07 −1.07 −1.53 2 2001 −0.57 0.10 −0.94 −0.61 −1.39 −0.34 −0.46 0.19 −0.71 0.08 1 2002 0.36 −0.68 −0.13 1.03 0.49 1.22 0.44 1.76 0.62 0.65 0 2003 0.45 0.45 0.05 −0.88 0.58 −0.35 0.00 −0.03 0.08 −0.31 0 2004 — −1.07 −0.24 0.93 −0.66 −0.75 −0.19 −0.74 −0.30 −0.98 1 2005 0.46 -0.42 −0.72 0.02 −2.04 −0.71 −1.97 −0.60 −1.07 −0.99 3 2006 1.01 −0.19 −0.62 −0.44 0.22 0.73 −0.21 0.06 0.61 −0.09 0 2007 — 2.62 −0.39 0.80 1.25 1.84 1.47 1.13 1.44 0.83 0 2008 — −1.05 1.49 −0.25 1.02 1.39 0.90 0.98 0.85 0.54 1 2009 0.10 1.00 0.50 −0.51 0.46 0.36 −0.12 −1.17 −0.55 −0.89 1 2010 0.06 0.72 1.51 3.59 −0.04 0.19 0.52 1.16 0.98 0.56 0 Frequency of Low Rainfall Events Moderate 1/30 5/30 3/30 4/30 2/30 3/30 5/30 6/30 5/30 3/30 37 Severe 1/30 0/30 1/30 0/30 2/30 0/30 0/30 0/30 1/30 1/30 6 Source: Ghana Meteorological Service; annual rainfall for March–October. 17 *Light shading indicates rainfall of more than one standard deviation below normal; dark shading indicates rainfall of more than two standard deviations below normal. FIGURE 3.2. IMPACT OF RISKS ON CEREAL PRODUCTION AND YIELDS, 1980–2011 Cereal production (metric tons) Cereal yield (kg per hectare) 2011-Drought 3,500,000 2,000 2001-Drought 2001-Floods 1,800 3,000,000 1992-Drought 1,600 2,500,000 1,400 1990-Drought 1,200 2,000,000 1983-Drought 1,000 1,500,000 800 2008-Fertilizer 1,000,000 600 reforms initiated 400 500,000 200 0 0 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 20 9 20 0 11 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 0 1 19 Source: FAOSTAT; authors’ notes. BOX 3.1. RAINFALL PATTERNS AND ruminants. Hunters use fire to drive game out of the bush CROP PRODUCTION into open areas where it is easier prey. The risk assessment included an in-depth analysis of Burning poses a risk when fires get out of control. This rainfall data collected from 99 weather stations located risk is highest when vegetation is dry and the Harmat- throughout the country. (Appendix I provides a summary tan is blowing. The incidence and severity of bushfires is of the analysis.) It provided useful information on the level and distribution of rainfall by region, highlighting impacts thus highest in the savannah zones of northern and cen- of six different rainfall characteristics on crop yields. The tral Ghana, where the rainy season is short (3–5 months) influence of rainfall on yield was examined using regional and drought is more frequent. The low population density production data for maize, rice, millet, groundnuts, cas- in these areas makes it more difficult to control fires set sava, and yams for 1992–2009. The analysis showed that by hunters and herdsmen, and means that uncontrolled the volume of rainfall, starting date of the rainy season, fires tend to burn over larger areas. The immediate eco- and intensity of rainfall are most closely related to yield nomic losses are generally limited, as few people live in but they are never a major determinant of yield, even in the savannah regions where their impact is strongest. Sec- such areas, but the longer-term environmental costs can ond, the impact of individual rainfall parameters is most be high as frequent burning changes the vegetative cover. apparent for the production of rice and groundnuts in the Where bushfires occur in farmed areas, they are a major Upper East Region. Maize and yam yields are particularly risk to late crops such as rice, particularly when planting vulnerable to drought in the Brong-Ahafo Region; yam has been delayed by late rains. yields are vulnerable to drought in the Central Region; and maize yields are vulnerable to excess rainfall in the Ashanti In the forest zones, most bushfires result from uncontrolled Region. land clearing and tend to be smaller in terms of the area affected. They can have a significant impact on people’s livelihoods, nevertheless, because they are more likely to nutrients (ash) for crops. In the drier savannah zones, it is occur in inhabited, farmed areas. In addition to the loss of used to burn off old vegetation and promote the growth surface crops, farmers also lose cocoa trees (which do not of younger trees and grasses for fodder for cattle and small regenerate after fire damage and must be replaced), and 18 Ghana: Agricultural Sector Risk Assessment TABLE 3.2. FREQUENCY OF EXCESS RAINFALL EVENTS BY REGION,* 1981–2010 Upper Upper Brong- Greater Number Year West East Northern Ahafo Volta Ashanti Eastern Western Central Accra of Events 1981 −0.75 −1.56 0.01 −0.50 0.17 0.27 0.10 0.55 1.03 0.52 1 Risk Prioritization 1982 −0.07 −0.03 −0.78 −1.33 −1.83 −1.73 −1.77 −0.98 −0.05 0.62 0 1983 −3.07 −0/81 −1.55 −1.24 −2.09 −1.99 −1.90 −1.58 −2.57 −2.39 0 1984 −1.91 −1.53 −0.26 −0.17 0.95 1.78 0.77 1.72 0.50 0.84 2 1985 −0.09 −0.98 0.28 0.57 −0.20 0.85 0.74 −0.04 −0.14 0.50 0 1986 0.69 −0.42 −0.66 −0.26 −0.88 0.73 −1.07 −1.31 −1.79 −1.18 0 1987 0.11 −0.46 −0.27 0.75 0.49 0.83 1.38 2.15 1.40 0.86 3 1988 −0.57 −0.20 0.29 v0.35 0.47 −0.90 −0.06 −0.12 −0.06 0.40 0 1989 0.12 1.39 2.23 1.35 1.30 0.25 0.40 0.92 0.80 0.80 4 1990 — −1.22 −1.00 −0.86 −0.30 −1.51 −1.02 0.06 −1.38 −0.68 0 1991 −0.38 0.58 2.06 0.14 1.03 −0.26 1.85 −0.17 1.62 2.45 5 1992 −0.23 0.57 −1.20 −1.27 −0.97 −0.86 −0.85 −1.05 −0.91 −0.96 0 1993 0.28 −0.31 0.04 −0.26 −0.09 −0.35 1.16 −0.31 −0.35 −0.42 1 1994 0.22 1.18 0.05 −1.51 −0.41 −0.46 −0.04 −1.64 0.24 −0.88 1 1995 1.19 −0.72 0.31 0.37 1.76 0.48 1.71 0.03 0.77 0.47 3 1996 −0.81 1.07 0.60 0.01 −1.00 −0.71 −0.46 0.66 0.21 0.92 1 1997 0.62 −0.18 −0.01 0.36 0.09 −0.42 −0.58 −0.31 0.28 0.91 0 1998 −0.52 0.45 −2.08 −0.01 0.16 −0.51 −0.30 −1.28 −1.01 −1.17 0 1999 1.45 1.78 1.43 0.63 0.74 1.50 −0.09 −0.11 0.52 0.51 4 2000 1.85 −0.09 0.01 −0.07 0.71 −0.56 −0.35 0.07 −1.07 −1.53 1 2001 −0.57 0.10 −0.94 −0.61 −1.39 −0.34 −0.46 0.19 −0.71 0.08 0 2002 0.36 −0.68 −0.13 1.03 0.49 1.22 0.44 1.76 0.62 0.65 3 2003 0.45 0.45 0.05 −0.88 0.58 −0.35 0.00 −0.03 0.08 −0.31 0 2004 — −1.07 −0.24 0.93 −0.66 −0.75 −0.19 −0.74 −0.30 −0.98 0 2005 0.46 -0.42 −0.72 0.02 −2.04 −0.71 −1.97 −0.60 −1.07 −0.99 0 2006 1.01 −0.19 −0.62 −0.44 0.22 0.73 −0.21 0.06 0.61 −0.09 1 2007 — 2.62 −0.39 0.80 1.25 1.84 1.47 1.13 1.44 0.83 6 2008 — −1.05 1.49 −0.25 1.02 1.39 0.90 0.98 0.85 0.54 3 2009 0.10 1.00 0.50 −0.51 0.46 0.36 −0.12 −1.17 −0.55 −0.89 1 2010 0.06 0.72 1.51 3.59 −0.04 0.19 0.52 1.16 0.98 0.56 3 Frequency of Excess Rainfall Events Severe 4/30 5/30 3/30 2/30 5/30 5/30 5/30 4/30 4/30 0/30 37 Catastrophic 0/30 1/30 2/30 1/30 0/30 0/30 0/30 1/30 0/30 1/30 6 Source: Ghana Meteorological Service; annual rainfall for March–October. 19 *Light shading indicates rainfall of more than one standard deviation above normal; dark shading indicates rainfall of more than two standard deviations above normal. root crops such as yams, because the tubers are rendered projected to fall markedly, diminishing a major source of inedible by high soil temperatures. income and food for producers in the far north. Studies also suggest that cocoa production would likely be the The most severe bushfires occurred in 1983 and again in major victim of any reduction in precipitation, which 1984–85. The bushfires of 1983 followed the nationwide could compound the impact of higher temperatures on drought of 1982 and burned throughout the country. evapotranspiration and access to soil moisture and exac- The social and economic consequences were immense, erbate the impact of increased temperature on pests and not only due to the direct loss of food crops but also disease. because 60,000 ha of cocoa trees were destroyed—a con- siderable loss which deepened and prolonged the collapse In addition to the scientific community, farmers and live- of Ghana’s cocoa sector. No bushfires of this magnitude stock keepers in Ghana have their own perceptions about have occurred since 1984–85, probably due to public climate change based on direct experience. Anecdotal action to educate rural people on the dangers of uncon- evidence collected for this study suggests that erratic trolled burning and on how to manage burning. Less rainfall patterns in recent years have made it increasingly severe bushfires still occur, however, particularly in the difficult for farmers to predict optimal planting times. northern and transition zones. Bushfires were reported Rains arrive much earlier, or later, than expected. Rain- as a problem in these regions in 1997, 2003, 2005, 2008, fall during the season is poorly distributed. Also, once the and 2009. growing season is underway, rains can stop for extended periods at critical stages in the crop development cycles, leading to poor yields, or in extreme cases, crop failure. Windstorms At times, when the rains return, heavy torrents cause Storm damage is a constant risk for plantains, an impor- flooding, inundating crops and compounding earlier tant food crop, as the trees break easily in high winds. losses. Some farmers have tried to adapt by staggering The damage is highly localized, however, and is only their planting or switching to alternative, more drought- apparent at regional level (see appendix F). At national tolerant crops. level, plantain production exhibits a very smooth trend, indicating that storm damage in one region is offset by above average production in others. PESTS AND DISEASES Pests and diseases are a permanent feature of Ghana- ian agriculture, for both crop and livestock production. Most can be controlled but farmers do not always use the Climate Change control techniques available, due either to lack of infor- There is no shortage of analysis on the probable impacts mation, access to needed inputs or financial resources, of climate change on agriculture production systems in or acceptance of the losses as a cost of agricultural Ghana and within the region (see appendix D for a synop- production. sis). Broadly, the research suggests that if climate change does reduce rainfall, its impact is likely to be greater in the center and south of the country than in the north. Crop Pests and Diseases Another shared belief is that temperatures will likely The main pests and diseases, the crops they damage, and increase more in the north than in the south. A recent available information on the incidence of major outbreaks study by the International Food Policy Research Institute is summarized in table 3.3. (IFPRI 2013) finds that food crop production could be affected by these changes in precipitation and temper- Of the numerous pests and diseases that affect non-cocoa ature. Models show a general decrease in maize yields crops, the four most serious are classified as “agricultural in 2050, compared to the 2000 baseline. Yield losses calamity” pests by government. Farmers who suffer crop are lower for rice, with some models showing extensive damage from these four (locusts, variegated grasshopper, areas of yield gains. Groundnut yield and production is African armyworm, and oil palm leaf minor) receive 20 Ghana: Agricultural Sector Risk Assessment TABLE 3.3. PESTS AND DISEASE RISKS FOR GHANAIAN AGRICULTURE Pest and Diseases Crops Damaged Incidence African armyworm Cereals, root crops, First, countrywide outbreak in 1987. Further outbreaks in 1994, 1997, vegetables, pasture 1999, 2001, 2002, 2005, 2006, 2009, 2010. Variegated grasshopper Cereals, root crops, vegetables Major outbreak in 1994. Minor outbreaks since in 1992, 1997, 2009. Oil palm leaf mite Oil palm Outbreaks in 1970, 1987, 1993–97, 2005–08. Locusts Cereals, root crops, vegetables No outbreaks since 1980. Black pod Cocoa Outbreaks since 1996. Capsid Cocoa Annual outbreaks. Cocoa swollen shoot Cocoa Outbreaks since 1993. virus disease (CSSVD) Stem borer Maize No information. Large grain borer Maize Under control since 1990s. Sigatoka Plantain First outbreak in 1992. Minor outbreaks in 1997, 2003. Striga Cereals Papaya mealybug Papaya First outbreak in 2008. Has decimated production. Fruit fly Papaya, pineapple, mangoes Recent outbreaks. Prevents exports. Rice blast Rice No information. White fly Vegetables No information. Rosette Groundnuts No information. Coconut wilt Coconuts First outbreak in 1982. Minor outbreaks since. Source: MoFA Annual Reports; Plant Protection and Regulatory Services Division, MoFA. assistance from government in the form of free chemicals, 2008–10, blackpod infected an average of 1.1 million advice and assistance with control and free materials for hectares per year (World Bank 2012). Caused by a fungus replanting. (Phytophtora) that attacks and eventually destroys the pods, it can be controlled if treated with three to six sprays of Overall, armyworm poses the greatest risk in terms of fungicide. However, the quality and timing of spraying is both the incidence and severity of crop damage. But as critical to successful control. the capacity of the Plant Protection and Regulatory Ser- vices Division of the Ministry of Food and Agriculture Cocoa pests such as mirids (or capsids) constitute another (MoFA) to control armyworm it is now quite strong, even major threat to cocoa production. They infect approxi- the worst attacks damage no more than 2,000 hectares mately 2 million hectares annually, with estimated losses of crops in a given year. Less severe attacks damage no of 83,400 tons in 2010. These insects feed on tree sap, more than 10 hectares, as in 1994. However, armyworm destroying growing shoots and often the tree itself. Regu- is moving further and further south, with recent out- lar spraying with insecticides and good tree maintenance breaks being highest in Volta region, Ashanti, and Brong- afford effective control. Ahafo. Armyworm outbreaks also tend to occur when the rains are late for planting, thus compounding the effect Yet another notable threat, swollen shoot virus (CSSVD) of drought. is a highly infectious virus spread by the mealybug, which infected 15,000–25,000 ha annually during 2007–10, Pests and disease are the major production risk for cocoa. with crop losses ranging from US$10 million to US$20.8 The control and eradication of these pests and disease is million. Control is based on removing and replacing the administered by COCOBOD. Blackpod is the main cause infected trees, which adds significantly to the costs of this of crop loss among cocoa producers. During the period disease as farmers income losses are compounded for Risk Prioritization 21 4 to 5 years until the new trees reach maturity. The breed- meat. Other common diseases include trypanosomiasis, ing of disease-resistant trees is viewed as the only long- tick-borne diseases and endoparasites. Official records are term solution. far from comprehensive, but most disease related mortali- ties of cattle are attributed to anthrax and CBPP. To reduce production risks, COCOBOD has also assumed wide-ranging responsibility for controlling pests Among small ruminants, the peste des petits ruminants (PPR), and disease. This program was initiated in 2000 and mange, and internal parasites remain major diseases, with now involves the monitoring and spraying of more than PPR being the main cause of reported mortalities. The 3 million hectares for capsid and blackpod, together relevant literature suggests that a PPR outbreak can cause with support for the rehabilitation of cocoa plots that up to 50 percent mortality among small ruminants in are old or infected with disease. Expenditure on these any affected area; during the past 5 years reported losses programs amounted to 524.7 million cedi in 2009/10 ranged from 100 to 800 annually. Consequently, the num- (US$366.7 million), equivalent to 20 percent of total cocoa ber of vaccinations has been relatively high, amounting to revenues (IFPRI, 2012 op cit.). While these measures have several hundreds of thousands per year, with a peak of 1 yet to provide complete protection, they have neverthe- million in 2008. PPR vaccines require cold chain handling less facilitated a 67 percent increase in the area planted and cost about GH¢5 per dose. to cocoa and a doubling of cocoa production since 2002 (see figure 3.3). Notwithstanding, COCOBOD’s recent For swine, mange and internal parasites are widespread. announcement of cost-cutting plans to phase out its flag- However, the highest numbers of losses appear to be ship spraying program raises significant concerns over the caused by African Swine Fever (ASF). In 1996, ASF killed future of cocoa production in Ghana, though some stud- about 25 percent of the pig population in neighboring Côte ies suggest that the withdrawal of public-led spraying ser- d’Ivoire, with mortality and eradication costs estimated vices could lead to improvements in service delivery and between US$15–30 million. The disease spread to Ghana farmer access (World Bank 2012). in the following years. In 2004, the Director of Veterinary Services reported that since the 1999 outbreak of ASF, the government had paid nearly 850 million cedis (approxi- Livestock Pests and Diseases mately US$120,000) as compensation to farmers whose Traditional agro-pastoralists, commercial farmers, pigs had been destroyed in efforts to control the spread of government officials and others who participated in the disease. There is no vaccine against ASF yet and there stakeholder consultations all highlighted diseases as have been two recent outbreaks in 2002 and 2007. among the top four risks facing livestock producers. Livestock pest and disease risk is difficult to assess For poultry, Newcastle and Gumboro diseases are domi- for two main reasons. First, incidence is not precisely nant, with reported losses equally attributed to each. known because of poor reporting. Secondly, livestock Over the last five years, between 6 and 10 million poultry diseases have multiple effects; in addition to mortality have been vaccinated annually against Newcastle disease. of young and adult animals, there are economic losses Reported losses have been a few thousand a year, with due to reduced calving rates, decreased milk produc- a maximum of nearly 19,000 in 2011. Several outbreaks tion, lower animal offtake rates and market prices, and of highly pathogenic avian influenza (HPAI) occurred in lesser draught efficiency. Direct losses can already be 2007 (near Tema and in the Volta region). While there quite substantial. have been no outbreaks since, risk assessment studies indi- cate that the threat remains high. Rinderpest has been eradicated, but such diseases as contagious bovine pleuropneumonia (CBPP), tuberculo- sis, brucellosis, anthrax, foot-and-mouth disease (FMD) MARKET RISKS and blackleg remain endemic which mainly affect cattle. Among common market risks are domestic and interna- Anthrax is a particularly serious threat given reported deaths tional price variability, exchange rate and interest rate in northern Ghana from consumption of contaminated volatility and counterparty risk. 22 Ghana: Agricultural Sector Risk Assessment PRICE VARIABILITY TABLE 3.4. INTER-ANNUAL CROP PRICE The analysis of producer price variability is based on VARIABILITY, 1991–2001 inter-annual price variability for 1991–2010, measured Coefficients of Variation by coefficients of variation (CV). Where necessary, these Root Crops CVs are adjusted for trend using the Cuddy-Delle Valle Cereal Crops and Plantain Cash Crops Index. Real prices,8 in cedis/ton, are used for the analy- Maize 0.18* Cassava 0.26* Cocoa (dom) 0.20* sis of domestic producer prices as high inflation during Sorghum 0.13* Yams 0.24 Cocoa (int)** 0.24* this period precluded any meaningful analysis of nomi- Millet 0.15* Plantain 0.31* Groundnuts 0.14 nal prices. The variability of international cocoa prices is Rice (paddy) 0.16* based on nominal prices in U.S. dollars. Annual producer price data are drawn from FAOSTAT and the interna- Source: FAOSTAT. *Adjusted for trend using Cuddy-Della Valle Index. tional cocoa price data is drawn from the World Bank **International prices in US$. commodity data series. The inter-annual variability of real cereal prices is rela- prices declined from 2003–07, and have increased slowly tively low, due probably to the slower 3–5 year cycle of since, while international prices have increased quite rap- prices that has occurred since the late 1990s (table 3.4). idly. This general upward trend in nominal and real prices Variability is also similar between cereal crops as they has also been facilitated by COCOBOD’s commitment to are close substitutes and are grown in similar (mostly raise producers’ share of world market prices. This share savannah) agro-climatic conditions. The higher over- averaged 54 percent of the free on board (FOB) price all price variability of cassava and yams is attributed from 2006–10 (IFPRI 2012).9 to high price volatility from 1995–99 (figure 2.8). Real prices have been relatively stable since 2000 and have The risk of adverse movements in international cocoa moved in unison. This is consistent with their produc- prices is assumed by COCOBOD. To mitigate this risk, tion in similar agro-climatic conditions and their role COCOBOD forward sells 60–80 percent of the expected as substitutes for consumption. Note also that while crop to international buyers. The value of these forward prices of both cassava and yams rose in 2001 and 2005 contracts then provides the basis for fixing the producer in response to domestic production and market shocks price at the beginning of each season. Forward selling (as for cereals), they were not affected by the global food does not remove all risk. A lower-than-expected harvest crisis. Plantain exhibits quite high inter-annual price would force COCOBOD to buy cocoa beans elsewhere to variability due to its vulnerability to storm and insect fulfill its contract, and the residual (non-contracted) har- damage, although prices have been more stable since vest would remain subject to the vagaries of international the early 2000s. This has also been the case for the other markets. Forward contracts also expose COCOBOD to food crops. counterparty risk, in that the international buying agents may not honor their contracts. Adverse outcomes from The variability of both domestic and international prices these risks have been minimal in the 1991–2010 period for cocoa is examined given that these risks are assumed by used for analysis. COCOBOD appears to have met all different actors. Comparison of international cocoa prices of its forward contracts, and nominal producer prices (in and the nominal cocoa price set by COCOBOD (figure U.S. dollars) have always been lower than international 3.3) shows the extent to which COCOBOD protects prices (figure 3.3). One international buyer defaulted on a domestic producers from volatility in international mar- forward contract in 1991/92, with a loss of £856,278— kets. Real producer prices fluctuate nevertheless, although equivalent to US$10 million in 2010 prices. COCOBOD inter-annual price variability is moderate and still less than the variability of international prices. Note also that real 9 “The Partially Liberalized Cocoa Sector in Ghana: Producer Price Determi- nation, Quality Control and Service Provision.” IFPRI, Development Strategy 8 Deflated by the consumer price index. and Governance Division. Discussion Paper 01213 (September 2012). Risk Prioritization 23 FIGURE 3.3. COCOA PRICES AND PRODUCTION, 1991–2010 Nom price (Cedi/T) Nom price ($US/T) World price ($US/T) Real price (Cedi/T) Production (T) 3,500 800,000 3,000 700,000 600,000 2,500 Production 500,000 2,000 Prices 400,000 1,500 300,000 1,000 200,000 500 100,000 0 0 19 1 19 2 19 3 94 19 5 19 6 97 19 8 20 9 00 20 1 20 2 03 20 4 05 20 6 20 7 20 8 20 9 10 9 9 9 9 9 9 9 0 0 0 0 0 0 0 19 19 19 20 20 20 Sources: FAOSTAT; COCOBOD; World Bank Commodity Reports; authors’ calculations. continues to seek ways to mitigate these risks, neverthe- TABLE 3.5. SEASONAL PRICE* VARIABILITY less, including forward contracts to multiple buyers as a FOR FOOD CROPS, 2004–08 means to reduce counterparty exposure. There is also 2004 2005 2006 2007 2008 scope to use futures contracts to mitigate the price risk incurred by that portion of the crop not covered by for- Coefficients of Variation ward contracts. Maize 0.17 0.18 0.27 0.12 0.24 Sorghum 0.28 0.14 0.07 0.10 0.23 Seasonal price variability is also reviewed for food crops, Millet 0.17 0.18 0.27 0.12 0.24 using nominal monthly retail prices for 2004–08 from Local rice 0.14 0.08 0.03 0.07 0.19 MoFA monthly retail price data (see table 3.5). There is Imported rice 0.11 0.08 0.02 0.18 0.17 no discernible pattern to seasonal price variability, either Cassava 0.13 0.08 0.09 0.12 0.23 across years or by commodity. It is probably this erratic Yam 0.12 0.34 0.10 0.17 0.16 pattern of seasonal price variability that leads farmers to Plantain 0.16 0.20 0.13 0.31 0.27 perceive it as a major source of price risk. Source: Ministry of Food and Agriculture. * Based on nominal monthly prices. MARKET ACCESS The abrupt loss of access to certain international mar- of a new variety that was highly sought after in European kets has also adversely impacted the sector during the past markets but not produced in Ghana (although this variety 10 years. Groundnut exports to the European Union were is now being grown in Ghana). decimated after 2004–05 when shipments were rejected due to unacceptable levels of aflatoxins. Total exports of groundnuts (with and without shell) have since fallen, EXCHANGE RATE RISKS from 14,583 mt in 2004 to 837 mt in 2010 (FAOSTAT). The variability of exchange rates is highly important Domestic production continues but now focuses on lower- for an export-oriented agricultural sector. This applies value commodities that are sold on the domestic market. not only to dollar-denominated exports such as cocoa An important source of export earnings has thus been and cotton, but also to food crop exports in franc CFA lost. Similarly, pineapple exports fell from 56,094 mt in (Communauté Financière Africaine) to countries such as 2004 to 9,971 mt in 2010 as a result of the introduction Burkina Faso, Niger, and Mali. The franc CFA exchange 24 Ghana: Agricultural Sector Risk Assessment FIGURE 3.4. NOMINAL EXCHANGE RATES, 1980–2012 Ghana Cedi/$US F Ghana Cedi/FCFA 2.000 0.0040 1.800 0.0035 1.600 0.0030 1.400 0.0025 Cedi/FCFA 1.200 Cedi/$US F 1.000 0.0020 0.800 0.0015 0.600 0.0010 0.400 0.200 0.0005 0.000 0.0000 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 Source: World Bank Development Indicators Database 2014. rate is also important for trade with Côte d’Ivoire, for both This can lead to an appreciation of the exchange rate and legitimate trade and the smuggling of cocoa and fertilizer. a fall in the price that domestic producers receive for agri- culture exports and for products competing with imports. Ghana effected a gradual realignment of its (highly over- This loss of trade competitiveness would also lead to a valued) exchange rate from 1983–1990, under a program decline in exports and loss of an important source of for- with the International Monetary Fund (IMF). It subse- eign exchange. quently benefited from a floating exchange rate policy and a period of exchange rate stability until 1999–2000 (figure INTEREST RATE VOLATILITY 3.4). This included adjustment to a 50 percent devalua- Both nominal and real interest rates are high in Ghana, tion of the franc CFA in 1994. A sharp exchange rate although analysis of quarterly interest rates for commer- depreciation occurred in 1999–2000 due to a terms-of- cial banks for 2005–12 shows that interest rate volatility is trade shock caused by a simultaneous increase in oil prices, relatively low (figure 3.5). The coefficient of variation for fall in cocoa prices, and decline in donor receipts. World interest for both agricultural loans and export loans is 0.10 gold prices were also very unstable during this period and and the coefficient of variation for base interest rates for domestic inflation and interest rates increased sharply. commercial banks is 0.16. A marked increase in interest Exchange rates then stabilized in 2004, only to experience rates occurred from 2008–09 as a result of high inflation another sharp depreciation in 2009 in response to another and the impact of the global food crisis, but interest rates terms-of-trade shock, this time associated with the global returned to their longer-term level by late 2010. Access to food crisis. A more gradual depreciation has occurred, credit and high underlying interest rates thus appear to be against both major trading currencies, since 2009. a bigger constraint than short-term volatility in the costs Moving forward, ongoing development of Ghana’s oil of this credit. and gas sector could have strong implications for the future growth of its agricultural sector. Any resulting COUNTERPARTY RISK appreciation of the exchange rate linked to inflation Producers and agricultural commodity traders in Ghana, could expose the agricultural sector to Dutch Disease. as elsewhere, face counterparty risks. This refers to the risk This phenomenon commonly refers to the adverse conse- to each party participating in a transaction that the coun- quences of a large increase in a country’s wealth resulting terparty will not live up to its obligations. In the absence of from a boom in a natural resource sector of an economy. commodity exchanges, well-developed warehouse receipts Risk Prioritization 25 FIGURE 3.5. COMMERCIAL BANK INTEREST Agriculture sector policy and regulation are a source of RATES, 2004–12 risk when public involvement in sector activities has unex- Base rate Agriculture Export pected, adverse consequences. 35 30 25 Interest % 20 POLICY MAKING 15 Ghana began the 1980s with a highly interventionist 10 policy stance, based on high levels of public expenditure 5 and budget support to strategic sectors (including agri- 0 culture), extensive price controls, and a fixed exchange 2 04 4 5 2 5 4 6 2 6 4 7 2 7 4 8 2 8 4 9 2 9 4 0 2 0 4 1 2 1 12 rate. These policies resulted in high inflation, an overval- Q 00 Q 00 Q 00 Q 0 Q 00 Q 00 Q 00 Q 0 Q 00 Q 0 Q 01 Q 01 Q 01 Q 01 20 20 20 20 20 2 2 2 2 2 2 2 2 2 2 2 4 ued exchange rate, a shortage of foreign exchange, and Q Q Source: Bank of Ghana. poor economic performance. An IMF structural adjust- ment program was initiated in 1983 to address these systems, and reliable contract enforcement mechanisms, issues. By the early 1990s, this program had achieved a traders have limited means to effectively manage default realignment of the exchange rate, the liberalization of risks. Given this environment and to minimize exposure, prices, the termination of subsidies (including fertilizer the vast majority of farmers and traders prefer to operate subsidies) and minimum prices for agricultural commodi- exclusively on a cash-and-carry basis. ties, and a large reduction of public employment. It was further consolidated in 1995 when Ghana joined the As the biggest exporter of agricultural commodities in World Trade Organization (WTO), setting modest levels Ghana, COCOBOD faces considerable counterparty of import protection. This included tariffs of 10 percent risks and has accordingly developed effective mecha- or 20 percent on most agricultural commodities. The nisms to manage them. First, in selling the crop forward, initial social costs of these changes were high, however, COCOBOD enters into fixed price contracts with a with high unemployment, and inflation that has yet to be select number of buyers (that is, international mer- brought under control. chants, processors, and chocolate manufacturers). In the event of a high volume default by one of these buyers, The more liberal, market-oriented polices introduced by COCOBOD’s financial results could be severely under- these reforms remain largely in place. Trade policy was mined. In order to limit to the greatest extent possible its temporarily modified during the global food crisis by exposure to any one individual buyer, COCOBOD typi- removing the 20 percent import tariff on rice and other cally tries to allocate available volumes among a wide imported foods in 2008. But these tariffs were restored in range of buyers. To a lesser extent, COCOBOD also December 2009. Fertilizer subsidies (50 percent subsidy faces default risks on the seed funding that it extends to for urea and compound fertilizers) were reintroduced in licensed buying companies (LBCs) at the beginning of 2009, and the government has subsidized the price of the season for the purchase of the cocoa crop. However, tractors since 2006. In 2000, commercial farmers also it has been largely able to manage such risks via local benefited from a reduction of corporate tax from 75 per- bank guarantees. cent to 25 percent. Policy changes pose a risk when they are made quickly and ENABLING ENVIRONMENT erratically, giving farmers, business agents, and consumers RISK little time to adjust. Fortunately, this does not appear to be Other sector risks arise from changes in the broader politi- the case (largely) in Ghana. The painful reduction of pub- cal and economic environment in which agriculture oper- lic support for agriculture during structural adjustment, ates. These changes can be both internal and external. including exchange rate realignment, was implemented 26 Ghana: Agricultural Sector Risk Assessment FIGURE 3.6. TRENDS IN MAIZE PRODUCTION AND PRICES, 1995–2011 Area (ha) Prod (tonnes) Nom price Real price 2,000,000 700 1,800,000 600 1,600,000 Nominal and real prices 1,400,000 500 Production/Area 1,200,000 400 1,000,000 800,000 300 NAFCO 600,000 200 Buffer stock policy 400,000 100 200,000 0 0 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 Sources: FAOSTAT; authors’ calculations. gradually. The government also responded to the 2008 In 2010, the government reestablished an agency for food price crisis in a balanced way. The removal of import public food reserves with formation of the National Food taxes on imported rice was coordinated with traders and Buffer Stock Company (NAFCO). The broad objectives businesses, as was the subsequent reintroduction of these of this agency are to 1) ensure that any “excess cereal taxes. production,” resulting from current measures to increase cereal production through the subsidization of fertiliz- ers and improved seeds, can be sold by farmers at guar- FOOD SECURITY POLICY anteed prices; 2) to stabilize the demand and supply for Government involvement in the production and market- cereals; and 3) to ensure a stock of food reserves to meet ing of cereals is based on the operation of public food emergency needs in the event of food crises or national reserves. This began in 2001 with the creation of a public disasters. The target for operational stocks for 2012 buffer stock in response to a perceived need for emergency was 15,000 mt white maize, 15,000 mt yellow maize, food reserves. The announcement of this policy led pro- 15,000 mt paddy rice, and 1,000 mt soya. For emergency ducers to increase the area planted to maize by 32 percent stocks, the targets were 10,000 mt white maize, 10,000 mt (figure 3.6). In combination with favorable growing condi- milled rice, and 1,000 mt soya. The impact of NAFCO’s tions this resulted in a 49 percent increase in production for activities on the stability of cereal prices and production the 2001/02 crop year. But as government only purchased has yet to be ascertained. 10,000 mt of maize for the buffer stock, this increase in production saturated the market and maize prices fell by INSTITUTIONAL RISKS: COCOBOD 22 percent in real terms. Farmers then reduced the area The government retains wide-ranging control of the under maize production for two successive years, and internal and external marketing of cocoa through the switched to other, more profitable crops. This policy- Ghana Cocoa Board. The policies and regulations used induced “shock” to production and prices was ultimately to exert this control have a significant influence on the the result of inadequate information to producers. Their sources and level of risk associated with cocoa produc- expectation was that all maize would be purchased by tion and marketing, including the level and stability of the government at guaranteed prices, rather than just the producer prices, the response to pests and disease, and 10,000 mt required for the buffer stock. This approach to the capacity to export cocoa profitably on international public food reserves was subsequently discontinued. markets. Risk Prioritization 27 FIGURE 3.7. GHANA/CÔTE D’IVOIRE COCOA PRODUCER PRICE DIFFERENTIAL 500 400 300 Price incentive - CDI to Ghana 200 100 $US/mt 0 02/03 03/04 04/05 05/06 06/07 07/08 08/09 09/10 10/11 –100 –200 –300 Price incentive - Ghana to CDI –400 –500 Source: World Bank 2012b. Cocoa Smuggling Measures to prevent smuggling through increased border Smuggling was for many years the main source of risk control were difficult to enforce given the difficulty of mon- to profitable operation by COCOBOD. It was the result itoring the border between the two countries. A reduction of the differential between producer prices in Ghana of smuggling risk will thus ultimately rely on measures and Côte d’Ivoire and the difficulty of controlling illicit to reduce the disparities between Ghanaian and Ivorian trade along the long, porous border separating the two producer prices. The Government of Côte d’Ivoire intro- countries. Until 2012, Ivoirian prices were set according duced cocoa sector reforms in 2012 and created a system to spot prices in international markets. Thus, they fluctu- offering a state-issued minimum price guarantee, backed ated above and below the fixed prices in Ghana. As shown by forward sales in international markets. This is likely to in figure 3.7, these price differentials created the incen- bring more certainty over prices and mitigate the level of tives for smuggling from Côte d’Ivoire to Ghana from cocoa smuggling between the two countries. 2003/04–2006/07 and in 2008/09; and from Ghana to Côte d’Ivoire in 2002/03, 2007/08, and 2009/10. DOMESTIC CONFLICT Internal conflict has been a significant source of risk Provided that it met COCOBOD quality standards, for agriculture, notably civil unrest in the Northern cocoa smuggled from Côte d’Ivoire to Ghana would gen- Region during 1994–95. In addition to the human suf- erate the same profit as Ghanaian cocoa. The resulting fering caused, these events led to a significant decline in higher volume of exports would also increase COCO- both regional crop production and regional trade, which BOD’s ability to cover its fixed costs and increase returns. had a wide-ranging impact on agricultural markets and The risks incurred by inward smuggling stemmed from access to food. A further, smaller civil conflict occurred the increased liquidity requirement generated by the pur- in 2002. chase of additional cocoa, and the increased consequent difficulty of making timely payment to Ghanaian produc- ers. Outward smuggling from Ghana to Côte d’Ivoire MULTIPLE SHOCKS incurred greater costs and higher risks as it reduced the The most severe shocks usually involve a succession or volume of exports. This had the potential to compromise combination of adverse events. Private households and COCOBOD’s capacity to honor its forward contracts public agencies typically have sufficient resources to and reduce overall profitability. respond to a single shock and effect a partial recovery, but 28 Ghana: Agricultural Sector Risk Assessment these resources are seldom adequate to cope with multiple 1992: A combination of localized floods and pest out- shocks. Multiple shocks to agriculture occurred in the fol- breaks in various regions throughout the country; plus lowing periods: the political and economic uncertainty associated with national elections. 1981–85: The catastrophic drought of 1981–82 was fol- lowed in 1983 by massive bushfires that destroyed crops 1994: Civil conflict, devaluation of the franc CFA, out- throughout the country, including 60,000 ha of cocoa breaks of variegated grasshopper and armyworm. trees. The impact of these natural disasters was com- pounded in 1983 by the expulsion of more than 1 million 1997: Drought, bushfires, and pest outbreaks in central Ghanaian migrant workers from Nigeria and the initia- and northern Ghana. tion of a structural adjustment program with large-scale reduction of public employment. A further severe out- 2007: Drought and severe flooding in the northern and break of bushfires occurred in 1984–85. central zones, and a collapse in world cotton prices. Risk Prioritization 29 CHAPTER FOUR ADVERSE IMPACT OF AGRICULTURAL RISK The frequency, severity, and costs of adverse events are analyzed in this chapter as the basis for prioritizing the various sources of risk. The conceptual and methodological basis for analysis is outlined first and then applied to production, market, and enabling environment risks. The various sources of risk are then reviewed to discern the most critical. 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 are unpredictable. Risk thus implies exposure to substantive losses, over and above the normal costs of doing business. In agriculture, farmers incur moderate losses each year because of unex- pected events such as suboptimal climatic conditions at different times in the produc- tion cycle or modest departures from expected output or input prices. Risk refers to the more severe and unpredictable adverse events that occur beyond these smaller events. This concept differs from the common perception of “risk” by farmers and traders, based on the year-to-year variability of production and prices. It should also be distin- guished from constraints, which are predictable barriers or limitations to agricultural production that farmers face every year. In Ghana, these constraints include poor access to farm inputs, limited access to markets, limited credit, poor infrastructure, and so on. LOSS THRESHOLDS As agricultural production is inherently variable, the immediate step for analysis is to define “loss thresholds,” which distinguish adverse events from smaller, interannual variations in output. This is achieved by first estimating a time trend of “expected” production in any given year, based on actual production, and treating the downside difference between actual and expected production as a measure of loss. Loss thresh- olds are then set for these downside deviations from trend, to distinguish between losses due to adverse events and those that reflect the normal costs of doing business. Two Risk Prioritization 31 thresholds are used to represent differing levels of sever- irrespective of actual production conditions. Hence, live- ity—severe losses and catastrophic losses. These below stock production losses due to drought or disease are not threshold deviations from trend allow estimation of the adequately captured. frequency, severity and cost of loss for a given time period. For price risks, the trend level of production for the rel- For purposes of analysis the threshold for severe losses evant crop was used as the point of reference. The total was set at 0.33 standard deviations from trend, and cat- loss due to a price fall was then calculated in cedi at real astrophic losses at 0.66 standard deviations from trend. prices (2010 = 100) as the difference between GAO at These thresholds captured the differing levels of adversity trend prices minus GAO at actual prices, and the remain- of known adverse events during the period of analysis. der of the calculation was derived as for production risks. The frequency and severity of losses derived in this man- This use of “normalized” (trend) production (rather than ner were also checked against historical records to ensure actual production) as the basis for analysis allows the loss consistency with actual adverse events. due to adverse price events to be captured more inde- pendently of losses due to an adverse change in produc- tion. Although this approach does not fully remove the THE INDICATIVE VALUE OF LOSSES joint endogeneity of prices and production, it was con- Available data on actual losses due to adverse events are sidered a reasonable proxy for the impact of price risk not always accurate or sufficiently consistent to facilitate given that factors other than production appear to cause comparison and ranking of the costs of adverse events. most (downward) price shocks, measured in real terms. Analysis was thus based on estimates of the “indicative” The available data suggest that in most cases changes in value of losses, which provide a more effective basis for international prices and exchange rates, inflation, policy, comparison. Indicative loss values are also compared with and externally driven changes in demand (for example, by the value of agricultural GDP in the relevant year to pro- importing countries) were more powerful determinants of vide a relative measure of the magnitude of loss. Although prices than were changes in supply. these estimates draw on actual data as much as possible, it is emphasized that they represent indicative rather than actual losses. Indicative losses were calculated as follows: DATA SOURCES Analysis of this nature requires a consistent set of data For production risks, the total value of gross agricultural on both production and prices for an extended period output (GAO) “lost” for each event was first calculated in of time. Of the various sources of data available, FAO- cedi as the difference between the actual and trend value STAT’s data series on the value of gross agricultural pro- of the relevant crop or crops, using constant producer duction (1991–2011), crop production (1991–2011), and prices (2004–06). The proportion of this total loss value crop producer prices (1991–2010) was considered the in excess of the threshold for trend production losses was most suitable. These data allow the analysis of risk over a deemed to represent the loss attributable to the adverse 20- to 21-year period. The analysis of risk at the regional event. The resultant value was converted into U.S. dollars level was based on data provided by MoFA for the nine at 2010 exchange rates, and also expressed as a percent- major crops for 1992–2009. age of the value of gross agricultural output to indicate the magnitude of losses in relative terms. This methodol- ogy was applied to each of the nine major crops and then CORROBORATION OF ANALYSIS to each region (based on trends in the aggregate value of The below-trend “adverse events” derived using this the nine crops at constant prices). methodology were not all unambiguously due to adverse conditions. Each identified adverse event was thus checked Production risks were analyzed for crops only as the avail- against climatic data, production and price data, infor- able livestock data were considered inadequate. Reported mation from annual MoFA reports, and other sources annual livestock numbers and production are based on to confirm that it was due to adverse conditions. Events a series of coefficients, which remain fairly constant not consistent with this information were excluded from 32 Ghana: Agricultural Sector Risk Assessment further analysis. Some were due to anomalies in the data, TABLE 4.1. SEVERITY AND COST OF ADVERSE particularly for the early 1990s. In other cases, a drop in EVENTS FOR AGGREGATE CROP production for a particular crop resulted from a decision PRODUCTION to switch to other crops. These cross-checks also facili- Indicative Loss tated the attribution of adverse events to specific causes Year Severity a Valueb,c (2010) Context (drought, floods, and so on), although a full attribution was not possible. Gaps in the corroborating evidence cedi US$ % (missing MoFA reports for 1993, 1995, 2000–03; and the (m) (m) GAO limited detail of the MoFA reports prior to 2000) plus the Production Risk (measured at constant prices) difficulty of accurately recalling past events precluded a 2007 Severe −137 −96 1.0% Major floods, full attribution. localized drought In some instances, adverse events were followed by sev- Sources: FAOSTAT; authors’ calculations. eral years of below-trend production, even though the Note: m = millions. a Severe: Production more than 0.33 standard deviation below trend. immediate cause of the shock was no longer extant. These b Calculated as the value of actual minus trend production, less the threshold post-shock years were not considered when determining for “normal” losses. the frequency and cost of the initial adverse event, as c In 2010 values based on real cedi prices (2010 = 100), and US$ to cedi ex- change rates for 2010. the below-trend production could not be unambiguously attributed to the initial shock. The absence of catastrophic events is consistent with the perception that agricultural sector risk is low at the AGGREGATE CROP aggregate level because of Ghana’s broad range of agro- PRODUCTION RISKS climatic conditions and the associated ability to produce a wide range of food and cash crops. Measured in terms of gross agricultural output (GAO) at constant prices,10 the volume of crop production was substantially reduced only once by adverse events during PRODUCTION RISKS FOR MAJOR CROPS 1991–2011—a low frequency of 0.05 (table 4.1). This Analysis of adverse production events by crop provides occurred in 2007, a year of multiple shocks including insight into crop-specific risk. Year-specific adverse drought, floods, and crop pest and disease outbreaks. The events were derived first and categorized into severe or indicative cost of this production shock was modest rela- catastrophic events (see appendix G, tables G.1 and G.2). tive to the total value of agricultural output (equal to just These events were then used to derive the probability of 1.0 percent of GAO). There were no catastrophic crop a production shock during the 21-year period analyzed production events at aggregate level. (1991–2011), and the average indicative cost of these shocks. The results are presented in figure 4.1. The frequency and severity of adverse events increased slightly when the combined impact of changes in pro- Adverse production events occur with a low to medium duction and producer prices was measured, with the frequency, with cassava and plantain exhibiting no pro- loss threshold exceeded in 1999 and 2007—a frequency duction shocks at all during the period of analysis. Yams of occurrence of 0.10. The indicative cost increased to and maize also exhibit a low frequency of production 1.9 percent of gross agricultural output for 2007—a mod- shocks, although with a medium level of indicative loss. erate cost to the sector. This suggests that price instability As cassava, yams, plantain, and maize together account increases both the frequency and cost of adverse events, for around 60 percent of crop GAO, the relative stability as would be expected, but not sufficiently to cause cata- of production for these crops accounts for much of the strophic shocks. observed stability in aggregate crop production. Rice, sorghum, millet, and groundnuts exhibit a higher 10 FAOSTAT: Constant producer prices calculated as average for 2004–06. frequency of adverse production events, as would be Risk Prioritization 33 FIGURE 4.1. FREQUENCY AND SEVERITY events, which are dominated by the low frequency–high OF ADVERSE PRODUCTION cost production shocks for cocoa. Millet exhibits the high- est incidence of catastrophic production shocks, but the EVENTS BY CROP 140 indicative costs are low. Overall, the results confirm the important stabilizing role of root crops and plantain for Indicative loss US$ (million) 120 100 aggregate production. Cocoa faces the biggest production 80 risk, although this too is moderate. Maize Maize Cocoa 60 Yam 40 Rice Sorghu Sorghum orghum Groundnut Ground PRODUCER PRICE RISK FOR 20 Millet et t MAJOR CROPS 0 A crop-specific analysis of price shocks was used to derive 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Probability of shock a similar set of data as that for production shocks. Detailed Sources: FAOSTAT; authors’ calculations. results are presented in appendix G, tables G.3 and G.4. No price shocks are observed for cocoa, as would be expected given that COCOBOD sets prices and ensures expected given that they are produced in the more risk- (nominal) price stability. Price shocks occur with a low fre- prone savannah regions. Their risk profile is moderate quency for most other crops, although the indicative losses nevertheless, with a low to medium frequency and sever- of these shocks vary markedly. Low frequency–higher ity of risk. They are more prone to catastrophic produc- cost shocks occur for cassava and yams, although even tion events, although the indicative costs are generally the largest of these shocks (for cassava) incurred losses of low (see appendix G, tables G.1 and G.2). But as these only 1.6 percent of GAO (see appendix G, table G.4). It is crops account for only 10 percent of aggregate crop out- worth noting here that prices for cassava and yams have put, production shocks have a limited impact on aggre- been much more stable within the most recent decade. gate production. Note, however, that the higher incidence The higher cost of these shocks is commensurate with of production shocks for these crops has a major impact their contribution to GAO. Maize, sorghum, and millet in the three northern provinces where they predominate. exhibit a medium frequency–low loss risk profile and rice, The catastrophic production shocks for these crops in groundnuts, and plantain exhibit a low frequency–low 2007 also show that these crops are more prone to covari- loss profile. ate production risk. A comparison of “severe” versus “catastrophic” price Cocoa is the most prone to production shocks, with a shocks shows that all crops (except cocoa) experience medium frequency of 0.2 (2 years in 10), and the highest severe price shocks, with maize as the most volatile. Cata- average indicative loss. The higher indicative loss is due strophic shocks are observed for sorghum, millet, and rice. to the high value of cocoa production, and the impact of Traditional cereal crops (maize, sorghum, millet) are the a catastrophic event in 2002. The indicative loss for this most likely to experience price shocks. The price shocks catastrophic event was low in relative terms, nevertheless, for these cereals also tend to occur simultaneously, sug- at 1.7 percent of GAO (see appendix G, table G.2). The gesting that they are covariate. other adverse production events for cocoa incurred low- moderate losses, suggesting that in most years, cocoa acts as a stabilizing influence on aggregate crop production. REGIONAL CROP PRODUCTION RISKS Production of the major crops was aggregated for each All crops (except cassava and plantain) experience “severe” region to facilitate analysis of the impact of adverse crop production shocks, although the frequency of these production events at the regional level. Analysis was based shocks is low to moderate. This is true even for the more on constant national prices (average for 2004–06) for each risk-prone crops (groundnuts, sorghum, and millet). A crop, as regional prices were not available. It thus measures different pattern is evident for “catastrophic” production production changes only, independent of price changes in 34 Ghana: Agricultural Sector Risk Assessment FIGURE 4.2. CROP PRODUCTION SHOCKS FIGURE 4.3. FREQUENCY AND SEVERITY BY REGION, 1992–2009 OF DIFFERENT CROP RISKS 140 300 Indicative loss US$ (million) Indicative loss US$ (million) 120 250 Floods F Drought/F (20077) 7 100 200 80 Northern Eastern 150 60 s st Upper Wes Upper Eas Upper s st 100 40 er-Ann l Inter-Annua shan Ashant Ashanti Ashan Other cocoa coa risks ocoa e varia price a ability Volta 50 20 s st Wes t tern a smuggling Cocoa sm zed d Localiz drought ong ng A Brong Ahafo 0 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Probability of shock Probability of stock Sources: MoFA; FAOSTAT; authors’ calculations. Sources: MoFA; FAOSTAT; authors’ calculations. regional markets. Detailed results are presented in tables and floods in 2007 (figure 4.3). This shock has recently G.5 and G.6 of appendix G and summarized in figure 4.2 been reported by the MoFA as the only adverse event to in this section. have significantly affected agricultural sector growth in the past 10 years. Its low frequency (once every 20 years), No production shocks were observed in the Central multiple causes, and high impact suggest that shocks of Region. Each of the other regions experienced one or this nature are best addressed by a combination of emer- more production shocks during the period of analysis, gency and mitigation measures. although in most cases the indicative costs were low to moderate in both absolute terms and measured as a per- The other shocks all incur low to moderate indicative centage of GAO. Upper East, Northern, Volta, and East- losses, although with differing frequencies. On aver- ern regions were the most prone to production shocks, in age, adverse (interannual) producer price shocks affect terms of both frequency and severity. Production vari- at least one of the major crops every second year. Some ability in the Upper East, Northern, and Volta regions is crops are more prone to these shocks than are others, attributed to the risks associated with drought and floods. with maize, sorghum, and millet as the most vulnerable. Fluctuations in maize and cassava production explain Adverse shocks from pests and diseases occur with regu- both the frequency and severity of losses in the Eastern lar frequency, and with some crops result in substantial Region. The frequency of adverse production shocks is losses year after year. However, such losses are not well low in the remaining regions, although the severity of loss documented and so are difficult to quantify. Localized differs. Medium-level losses were observed in the Upper droughts also occur with a high frequency, particularly in West owing to the high-impact shock that occurred in the savannah regions. These droughts result in relatively 2007. Losses in the Ashanti, Western, and Brong-Ahafo low indicative losses, as the crops they impact (sorghum, regions were low. Catastrophic losses were observed in the millet, groundnuts) account for a small component of Upper East, Upper West, Northern, and Volta regions. GAO. They have a major impact at the local and district This is commensurate with their vulnerability to drought levels, however, and continue to impede growth and pov- and floods. erty reduction in the northern regions. SOURCES OF RISK Shocks to cocoa production occur with medium fre- The analysis of different types of risk provides useful quency (four times in 21 years). Losses due to smuggling insight into their relative importance. The high loss–low are identified as a separate category as they are inade- frequency impact of catastrophic events caused by mul- quately reported, although they do not account for all of tiple shocks is evident from the consequences of drought the indicative losses observed. Residual cocoa production Risk Prioritization 35 losses are thus reported as a separate “undefined” category of livestock to GH¢610 million (or 8.6 percent of agri- as there is not enough information to attribute them to a cultural GDP, assuming no upward adjustment in other specific cause. Production shocks due to pests and diseases sub-sectors). account for part of this undefined loss, particularly in the period before mass control programs were introduced in Based on the information provided in chapter 3, in a the early 2000s. normal year, diseases affecting various species can be expected to result in losses of 25 percent in value overall.11 An occasional significant outbreak (for example, HPAI or IMPACT OF LIVESTOCK ASF) could lead to additional losses of 20 million cedis, DISEASES whereas a severe drought event affecting ruminants in In 2010, Ghana’s real agricultural GDP was calculated northern regions could imply losses amounting to twice at GH¢7.1 billion, including some GH¢474 million (or that amount.12 However, the impact of diseases and other 6.7 percent) for the livestock sector. The latter can be adverse impacts on livestock could be much greater in the recalculated by component species and adjusting for the northern parts of the country, where poorer households aforementioned estimated increases in national herd depend on livestock to a larger extent, as discussed in numbers. This would raise the value of the contribution chapter 5. 11 Through mortalities, lower market prices, and reduced milk or offspring pro- ductivity, and so on. 12 Assuming an additional 10 percent loss in value to ruminants in northern regions through mortalities, reduced market prices, and reduced milk production. 36 Ghana: Agricultural Sector Risk Assessment CHAPTER FIVE ASSESSMENT OF STAKEHOLDER VULNERABILITY As highlighted earlier, Ghana’s agricultural systems vary within the five different agro- ecological zones and across the country’s 10 regions (see appendix A). As a result, the nature and severity of agricultural risks can vary greatly from one area to the next. The ability to deal with those risks among different stakeholders also varies, based on myriad factors, including prevailing production systems, household income levels, and the relative diversity of income sources. Assessing levels of vulnerability among spe- cific groups thus requires an understanding of their level of exposure as well as their risk management capacity. This includes their capacity to cope with and recover from resulting losses. During the field portion of the assessment, the team met with a range of stakehold- ers13 and sought individual as well as focus group assessments of production, market, and enabling environment risks. Discussions covered main risks, impact, relative rank- ing in terms of severity, and mitigation measures undertaken as well as their relative effectiveness. Based on this input, this chapter identifies common agricultural systems, profiles some of the key stakeholders and their perceptions of agricultural risks, and evaluates levels of vulnerability to risks (see appendix E for a broader discussion of vulnerability among various livelihood groups in Ghana). It also presents some com- mon strategies they employ to manage these risks. Stakeholder groups included government and nongovernmental organization (NGO) technicians, agriculturalists operating under rain-fed or a mix of irrigated and rain-fed conditions, agro-pastoralists, commercial farmers, traders, and processors. As usual, stakeholder groups were not homogeneous; for instance, some adjustments had to be made to reflect the fact that a “government technicians” group often includes people 13 The information covered in this chapter is largely based on interviews with 28 rain-fed farmers in two locations, 15 agro-pastoralists in three locations, 24 traders in Wa, 2 commercial farm enterprises in Tamale, and 3 processors in Wa and Tamale as well as meetings with nearly 20 government technicians from MoFA, SADA, the Pong-Tamale livestock station, and the Wa RDA. Risk Prioritization 37 with a lifelong interest and experience in agriculture as TABLE 5.1. RISK RANKING, RAIN-FED well as others similarly vested in animal husbandry issues. FARMING* Commercial farmers also face some of the same risks that 5 smallholders face. Still, it was possible to obtain clearly 4 differentiated viewpoints on the types of risks affecting various production systems (see table 5.1). The analysis 3 highlights that rain-fed farms in Ghana’s northern regions are by and large the most vulnerable to agricultural risks. 2 1 RAIN-FED AGRICULTURE 0 This production system applies to the majority of rural Drought Uncertainty Crop pests Livestock Theft crops populations in Ghana, the vast majority of which culti- in product and diseases or livestock pricing diseases vate food crops and cash crops. About 2.74 million house- Source: Authors’ notes. holds either operate small family farms (most of them * Number of respondents citing risk type as most and second most important. less than 2 hectares) or keep livestock (MoFA 2008). Most households in northern regions keep backyard livestock, small ruminants, and sometimes cattle, in addition to cul- CSIR-Omankwa). Other effective but localized drought- tivating cereal and other crops. According to the 2009 related measures cited by respondents included soil mois- Comprehensive Food Security and Vulnerability Assess- ture conservation (half-moons, bunds, ridging); water ment (CFSVA), food crop farmers have the lowest annual harvesting (for example, small pond or levy construction); per capita income. Seventy-two percent of them cultivate and small-scale irrigation. less than 2 hectares and are primarily reliant on rainfall for water. Almost half of the heads of households have Mitigation measures against uncertainty in product pric- no educational background, and women head 22 percent ing included more and better on-farm storage; switching of these households. Among cash crop farmers, the most to crops for which there appears to be stronger effective vulnerable are in the Upper West Region where they con- demand (for example, soya beans, yams, groundnuts); and stitute 17 percent of the population. The majority of cash establishing better linkages to markets. Most respondents, crop farmers live in the forest zone. Their second income however, agreed that their capacity to implement these source is food crop production. They have the highest measures was very limited. With regard to mitigating the annual per capita income, but more than half are in the impact of crop pests and diseases, there was generally poorest wealth quintile; women head 18 percent of these broad consensus among respondents on the application households. of chemicals as well as recognition that few had the finan- cial means to do so. The importance of crop rotation was Risks mentioned by individuals, farmer groups, and gov- also widely cited as a way to reduce the risk of striga. ernment technicians linked to this system and who pro- vided feedback for this assessment and their ranking are Given relatively limited incomes, meager savings, and lack depicted in table 5.1. of alternative income sources, households that depend on rain-fed production are by and large the most vulnerable Having focused on the worst risks, farmer groups were to adverse shocks such as drought, excessive rainfall, pest/ then asked about mitigation measures and their relative disease outbreaks, and price volatility. Not only are they effectiveness. Mitigation measures against drought or least equipped to manage the risks before they are real- weather uncertainty included crop diversification, espe- ized, they are also the least able to cope with losses in the cially with tubers/root crops; combining short-cycle and wake of adverse shocks. Depending on the severity of the longer-cycle crops (for example, early maturing millet, shock event (or series of events), the loss of income and “Chinese” groundnuts); and integrating drought-tolerant assets can have myriad and devastating impacts on live- maize varieties (for example, CSIR-Aburohemaa and lihoods, both immediately following the event and over 38 Ghana: Agricultural Sector Risk Assessment the medium to long term. Immediate impacts can include TABLE 5.2. RISK RANKING, IRRIGATED loss of crops and related income that finances household FARMING* needs such as clothing, education, health services, and Uncertain cost of chem. inputs food. The worst affected households are often forced to Bushfires turn to emergency savings or sell off assets to secure basic Uncertainty in product foodstuffs and other essentials, responses that weaken pricing Grain-eating birds their resilience to future shocks. Similarly, owing to lack of storage options, adverse price shocks can often leave Drought such households with significantly reduced income or Crop pests/Diseases even losses on their farming investments if they are forced Uncertain access to machinery to sell. For these reasons, investments should be prioritized Floods that strengthen the capacity of the most vulnerable com- 0% 10% 20% 30% 40% 50% 60% munities to manage these risks and build their resilience Source: Authors’ notes. to withstand and recover from shocks in their aftermath. * Share of respondents citing risk type as most important or second most important. IRRIGATED AGRICULTURE The vast majority of agriculture in Ghana is rain-fed. structure or technology involved, and source of power for With abundant cultivable land and sufficient surface abstracting, conveying, and distributing water. Common water and groundwater resources, Ghana offers ample types are public-owned surface irrigation systems; public- scope for irrigation-based intensification. However, at private partnership commercial systems; small reservoir- present, roughly 30,000 hectares, or just 0.2 percent of and dugout-based systems; group-based river-lift systems; total agricultural land,14 is irrigated. Approximately one- groundwater-based irrigation systems; and lowland an third of this lies within 22 public schemes developed by inland valley rice water capture systems. Figure App.E.2 the government and various nongovernmental organiza- in appendix E illustrates the geo-distribution of various tions. The remaining two-thirds are made up of informal types of irrigation systems in Ghana. irrigation schemes, the location, development, and man- agement of which are not well documented. These are for The largest of Ghana’s public irrigation schemes, totaling the most part developed and run by private entrepreneurs about 2,490 hectares, is located in the Upper East Region. and farmers. Such schemes are thought to be expanding On some 1,500 ha of irrigated land about 2,500 farmers at a rapid rate, fueled in part by improved access to newer grow mostly rice and soya beans during both the rainy and cheaper pumping technologies. and dry seasons. Many also cultivate rainy-season ground- nuts and maize on their off-perimeter land holdings. Dry- Informal systems include tube well irrigation, small motor- season crops include onions, tomatoes, peppers, and leafy based irrigation, and out-grower systems. Surface-water, vegetables. All farmers associated with this irrigation pumping-based private and communal irrigation systems scheme own some livestock as well. Although there are are widely dispersed across all of Ghana’s 10 administra- as yet limited areas under irrigation, it is often mentioned tive regions, and are particularly abundant in the Eastern, as a way to mitigate the risk of drought. Thus, it is worth Ashanti, Brong-Ahafo, and Volta regions. Sub-surface and considering risks that affect irrigation directly. groundwater-based irrigation systems are not evenly dis- tributed across the regions but are fast spreading beyond Key risks highlighted by a farmer group and government traditional enclaves such as the Volta region’s Keta strip technicians regarding this system and their ranking are (IFPRI 2011). IFPRI classified Ghana’s irrigation systems depicted in table 5.2. into several typologies based on several criteria including ownership/management, source of water, type of infra- Mitigating measures against flooding consisted of improved drainage, but all respondents recognized 14 World Development Indicators. their limited ability in this regard. There were no clear Risk Prioritization 39 mitigation measures identified against uncertainty in TABLE 5.3. RISK RANKING, AGRO- access to mechanized equipment, which was regarded PASTORALISTS* as a serious risk, especially because labor costs are either 100% unpredictable or likely to be high at the same time. Miti- gation measures highlighted against crop pests and dis- 80% ease risk were the same as they were for rain-fed farms. 60% Finally, to mitigate uncertainty in product pricing, one large group of rain-fed farms had tried to contract with 40% large institutional clients (for example, schools, the World Food Programme), but with limited success. 20% Irrigation enables farmers to make better use of available 0 Drought Livestock Theft of Pricing water resources. It also facilitates year-round production. (pasture, fires) diseases livestock uncertainty Better control of resources and higher output means that Source: Authors’ notes. households practicing irrigated agriculture have both a * Share of respondents citing risk type as most important or second most important much higher capacity to manage production risks and to recover from shocks in their aftermath. Higher and more predictable revenue streams empower such households was one way to mitigate drought by ensuring that well- to invest in risk mitigating (for example, on-farm storage, fertilized fields could do reasonably well, even with poor pest-resistant varietals), and coping measures (for example, rainfall. Transhumance also was quoted as a way to deal agro-chemicals, borrowing), thus reducing overall impacts with uncertainty in weather. Regarding mitigation strate- from adverse shock events. For these reasons, investments gies against animal diseases, respondents complained that in irrigation development, particularly in areas that are access to vaccines is difficult, and that the reliability (that prone to extreme weather events, could go a long way in is, quality) of both vaccines and other medicines bought reducing levels of community vulnerability and strength- locally constitutes a risk in itself. Respondents had simi- ening resiliency. lar responses with regard to possible measures to reduce theft of livestock. Vaccinations as well as medical control of endo- and exo-parasites were deemed generally effec- AGRO-PASTORALISM tive in mitigating losses from livestock diseases. However, For many households, particularly in the northern part these measures were considered to be not entirely reliable of the country, livestock production contributes largely and not always accessible. With regard to limiting losses toward meeting food needs. It also provides draft power, from theft of crops and livestock, respondents highlighted manure to maintain soil fertility, and cash income. Rumi- a range of measures that were generally reckoned to have nant livestock more widely play a major role in the socio- a positive but limited impact. These included community cultural life of rural communities. It acts as a source of watches, ear tags, and cold branding for cattle. household savings and insurance in times of difficulty. According to surveys conducted in 2008, 59 percent of Livestock production offers myriad benefits. It provides agro-pastoralists live in the Northern Region and 21 per- supplemental and diversified income, a supplemental cent in the Upper East Region. Sixty percent of their food source, and valuable household assets. It enhances average income is derived from livestock and animal hus- household food security. By acting as a source of fertilizer bandry. Food crop production accounts for another fifth. and labor, it also contributes to more resilient crop pro- Four out of five households (88 percent) were identified as duction systems. Given these and other benefits, house- poor (CFSVA 2009). Table 5.3 depicts the main perceived holds owning livestock are often better equipped to face risks among these households and their rankings. and overcome adversity than those households who rely exclusively on rain-fed agriculture. For example, well- Some agro-pastoralists, especially in the Upper East, fertilized crops can better withstand stress from water argued that a better integration of agriculture and livestock or pests and diseases, resulting in fewer losses. Thus, 40 Ghana: Agricultural Sector Risk Assessment promoting livestock ownership among vulnerable com- TABLE 5.4. RISK RANKING, COMMERCIAL munities, particularly in the northern savannah regions, FARMERS* could be an effective measure in building resiliency 5 against risk events. 4 COMMERCIAL FARMERS 3 Larger commercial farmers, who account for roughly 2 10 percent of agriculture production in Ghana, high- lighted risks similar to those faced by the more traditional 1 smallholders. However, their ranking gave more promi- 0 nence to product pricing and uncertainty in the timing of Drought Floods Uncertain Uncertain Livestock Theft of access to agro-chemical inputs (see table 5.4). In decreas- product fertilizer diseases crops/ pricing policy livestock ing order of severity, they cited drought, floods, and Source: Authors’ notes. uncertainty in product pricing (as equally serious), uncer- * Number of respondents citing risk type as most important or second most tainty in access to fertilizer (time and price), and livestock important. diseases. In addition to the mitigating measures men- tioned against drought by smallholders, they also cited the TABLE 5.5. RISK RANKING, GRAIN practice of increasing organic matter content in soils by TRADERS* plowing under crop residues or manure from livestock. 5 Commercial farmers are also more likely to build small water reservoirs. Commercial farmers were better able 4 than smallholders to rely on storage and contract sales to 3 mitigate product price uncertainty, but with limited effec- tiveness. To mitigate the risk of uncertainty in access to 2 fertilizer, commercial farmers had greater financial means 1 than did smallholders, allowing them to partly circumvent any delays in the delivery of subsidized fertilizers. When 0 adversity strikes, impacts among commercial farmers are Incertainty in Insecurity Floods Fire product often much less significant, given their relatively limited pricing exposure to risks and their ability to respond quickly and Source: Authors’ notes. * Number of respondents citing risk type as most important or second most effectively to minimize losses. important. TRADERS AND PROCESSORS to their stocks such as flooding (of warehouses) and fire. In declining order of importance, main risks cited by Other risks, such as exchange rate fluctuations or weevil market traders included uncertainty over market prices attacks, were mentioned but deemed relatively minor. For (especially in 2013), when the prices of several basic their part, processors were mostly concerned about uncer- commodities—including maize and groundnuts—have tainties in quantity and quality of supply, and the reliabil- sharply fallen from harvest time to planting season),15 ity of electric power supply (table 5.5). and insecurity. These risks were distantly followed by risks Among group members who participated in the infor- 15 Maize fell from an average GH¢52 a bag in late 2012 to GH¢32 in early June mal surveys, more and better storage was also viewed as 2013. This unusual trend was attributed by most observers to large shifts into the most effective way to deal with uncertainty in prod- maize after the bad 2011/12 growing season and relatively abundant local and uct pricing. Interestingly, traders did not feel that value regional supplies. Good production in 2012 also allowed some producers to store maize until the next planting season, at which time a great quantity came chain improvements could reduce this risk significantly. on the market to finance the purchase of agricultural inputs. Measures designed to mitigate the rising risk of insecurity Risk Prioritization 41 TABLE 5.6. STAKEHOLDERS’ RISK PERCEPTIONS AND RANKINGS* Agriculturalists Agriculturalists Government Commercial Technicians Pastoralists Processors Rain-Fed Irrigated Farmers Traders Agro- Risk/Stakeholder Group Drought/long dry spells, late start 1 4 1 1 1 — — Uncertainty in product pricing 2 6 4 3 2 1 — Crop pests/diseases 3 3 — — 3 — — Livestock diseases 4 — 2 5 4 — — Excessive rainfall, flooding x 1 — 2 — 3 — Thefts of crops or livestock 5 x 3 6 6 — — Bush fires (or in storage) x — x — 5 4 — Grain-eating birds x 5 — — — — — Uncertain access to machinery — 2 — — — — — Insecurity/conflicts — — — — x 2 — Uncertainty in fertilizer policy — — — 4 — — — Uncertain electricity supply — — — — — — 1 Uncertain cost of chemical inputs — x — — — — — Uncertainty in labor costs — x — — — — — World market price risk — — — — x — — Exchange rate risk — — — — — x — Storage losses — — — — — x — Source: Authors’ notes. Numbers indicate the number of responses citing each risk type as the most important. * consisted of reducing amounts of cash transported, either figure prominently, just before floods and crop pests/ by bartering goods at the point of purchase, or relying diseases. more on bank transfers. Neither was judged to be fully effective. Many respondents mentioned that specific risks often cluster together, and this was taken into account in the overall ranking. Drought, for instance, is often associated RANKING OF STAKEHOLDER with increased severity of bushfires, because of much drier RISK PERCEPTIONS vegetation. The same phenomenon can be associated with Table 5.6 summarizes risk perception, by group, indi- attacks by grain-eating birds, because their usual sources cating the ranking by type of stakeholder. Price risk of food are more limited under arid conditions. Drought and drought are widely shared perceptions, the latter and floods are often mentioned together because weather because of its severity and of a widespread perception patterns in the recent past have often led to a series of dry that droughts are becoming more frequent. Livestock spells in the first half of the season followed by spates of diseases and thefts of crops and animals (also on the rise) torrential rainfall in the latter part. 42 Ghana: Agricultural Sector Risk Assessment CHAPTER SIX RISK PRIORITIZATION AND MANAGEMENT RISK PRIORITIZATION This assessment has highlighted key risks facing the agricultural sector in Ghana. These risks are both myriad and complex. They manifest with varying levels of fre- quency and severity, resulting in losses to crops and livestock and leading to income volatility. To identify an effective risk management strategy that maximizes available resources, it is necessary to prioritize these risks. This requires an understanding of which risks occur most frequently and which cause the biggest financial losses. Chapter 4 identifies priority risks, using quantitative measures, for the crop and live- stock sub-sectors. Owing to the paucity of data, some of the risks could not be quan- tified. Chapter 5 further highlights key risks based on anecdotal evidence collected directly from stakeholders. Based on the team’s combined quantitative and qualitative assessment, table 6.1 prioritizes the most important risks for each crop, for livestock, and for poultry. Overall, this prioritization identified 1) drought, 2) pests and diseases, and 3) price volatility as the most important risks. Flooding from excessive rainfall and bushfires associated with drought were also deemed important, but to a lesser extent. Commodity profiles in appendix B offer a more detailed sub-sector analysis of risks. It is worth noting that a single type of risk can affect one or several commodities (for example, armyworms impacting maize, rice, and millet yields, or drought damaging both crops and livestock pasture). However, as this report highlighted earlier, Ghana at the aggregate level is not particularly vulnerable to agriculture risk while related costs to the economy are relatively low. This is due to the diversity of agro-climatic conditions and agricultural systems (that is, crops, seed varieties) across the country’s 10 regions. The probability that a single risk event could adversely impact a large num- ber of commodities and regions at once is extremely low. Despite this diversity, the analysis highlights that Ghana’s agricultural sector is highly susceptible to downside risks associated with multiple shocks. This refers to, for example, times when extensive pest attacks and bushfires cause further losses to crops already damaged by prolonged drought conditions. Although less frequent, such Risk Prioritization 43 TABLE 6.1. RANKING OF RISKS BY SUB-SECTOR Risk Commodity Priority #1 Priority #2 Priority #3 Cocoa Pests and diseases Price volatility (International) Smuggling Cassava Diseases and pests Flooding/excessive rainfall Drought Maize Drought Price volatility Pests/diseases Yams Diseases/pests Flooding/excessive rainfall Price volatility Groundnuts Drought Diseases and pests Flooding/excessive rainfall Plantain Winds and storms Diseases and pests Price volatility Sorghum Drought Pests and diseases Price volatility Millet Drought Pests and diseases Price volatility Rice Drought Flood Price volatility Livestock Diseases Drought Theft/conflict Poultry Regulatory risks (imports) Diseases Price volatility Aggregate for sector Drought Pests and diseases Price volatility Source: Authors. multiple shock events cause the greatest financial losses to their impacts. The assessment also highlighted consider- livelihoods across the sector. This is especially true when able levels of vulnerability among smallholder producers they are associated with drought. These findings will ide- across the country’s cocoa belt. This is especially true in ally have a direct bearing on future risk management light of growing concerns over COCOBOD’s finances, interventions. punctuated by its recent announcement of cost-cutting measures. This includes a 5-year planned phaseout of its flagship spraying program, which for more than a decade RISK PRIORITIZATION BY REGION has covered the costs of fungicide and pesticide applica- As noted earlier, regional diversity helps mitigate aggre- tions aimed at controlling blackpod disease and capsid gate-level risk. However, it also implies that each region infestations. faces a different set of risks, with varying levels of vul- nerability and strong implications for risk management. Northern regions, for example, are more prone to drought RISK MANAGEMENT events, which can impact a large share of households; this MEASURES occurs within a context in which relative poverty makes it There is no single measure to manage all risks; effective more difficult for people to mitigate risks and cope with risk management requires a combination of coordinated shocks (see the vulnerability analysis in appendix E). Thus, measures. Some are designed to remove underlying con- it was necessary to assess key risks at the regional level straints; others are designed to directly address a risk or (see annex F). Table 6.2 details the team’s ranking of key a subset of risks. Available resources are often a limiting agricultural risks (among priorities identified above) for all factor, but integrated risk management strategies are often 10 regions and its assessment of the level of regional vul- more effective than one-off or stand-alone programs. Risk nerability to each risk type (see appendix A for a detailed management measures can be classified into the following regional risk profile). three categories: 1. Risk mitigation (ex ante). Actions designed The region-level analysis and filtering reveal that low- to reduce the likelihood of risk or to reduce the income rural households in the northern regions (Upper severity of losses (for example, water harvesting West, Upper Eastern, Northern) are most prone to key and irrigation infrastructure, crop diversification, production and price shocks and most vulnerable to extension). 44 Ghana: Agricultural Sector Risk Assessment TABLE 6.2. RANKING OF RISKS AND VULNERABILITY BY REGION Region Priority Risks Level of Vulnerability Ashanti 1. Crop pests and diseases Medium 2. Flooding and excessive rainfall Medium 3. Drought Low Brong-Ahafo 1. Crop pests and diseases Medium 2. Drought Low 3. Flooding and excessive rainfall Low Central 1. Drought Low 2. Crop diseases and pests Medium 3. Flooding and excessive rainfall Medium Eastern 1. Drought Low 2. Crop diseases and pests Medium 3. Flooding and excessive rainfall Low Greater Accra 1. Drought Low 2. Crop, livestock, and poultry diseases Medium to high 3. Flooding and excessive rainfall Medium Northern 1. Drought/flooding High 2. Price volatility Medium 3. Crop and livestock pests and diseases Medium Upper East 1. Drought and flooding High 2. Price volatility Medium 3. Crop and livestock pests and diseases High Upper West 1. Drought and flooding High 2. Price volatility Medium 3. Crop and livestock pests and diseases High Volta 1. Drought High 2. Crop pests and diseases Medium 3. Flooding and excessive rainfall Medium Western 1. Crop pests and diseases Medium 2. Flooding and excessive rainfall Medium 3. Windstorms Low to medium Source: Authors. 2. Risk transfer (ex ante). Actions that will trans- government assistance to farmers, debt restructur- fer the risk to a willing third party. These mecha- ing, contingent risk financing). nisms usually will trigger compensation in the case of a risk-generated loss (for example, purchasing Table 6.3 highlights some potential interventions that insurance, reinsurance, financial hedging tools). could help address key risks identified by this assessment. 3. Risk coping (ex post). Actions that will help These are classified under three types: 1) risk mitigation, the affected population to overcome crises and 2) risk transfer, and 3) risk coping measures. The list is by build their resilience to future shocks. Such inter- no means exhaustive, but is meant to illustrate the type ventions usually take the form of compensation of investments that, based on the analysis, have good (cash or in-kind), social protection programs, potential to improve agricultural risk management in and livelihood recovery programs (for example, Ghana. Unlike drought or livestock diseases, which have a Risk Prioritization 45 TABLE 6.3. INDICATIVE RISK MANAGEMENT MEASURES Mitigation Transfer Coping Promoting crop diversification (for example, tubers Macro-level Use of weather index to trigger early warning and root crops, seed varieties) crop insurance and response Combining short-cycle (for example, millet and Farm-level cropDecentralized disaster contingent fund for Chinese groundnuts) and longer-cycle crops insurance rapid response to local emergencies Promoting adoption of soil and water conservation Contingent financing and other instruments and NRM techniques to support coping measures Developing community-level food and fodder and Cash-for-work and Food-for-Work (FFW) Drought forage (that is, livestock) banks programs to support soil and water conservation Improved farming techniques (for example, Facilitating temporary migration or intercropping, conservation tillage) permanent relocation Promoting development of small-, medium-, and Promoting development of social safety net large-scale irrigation; water harvesting programs (for example, food aid, FFW) Community outreach programs on tree cutting Promoting household and community and reforestation savings Strengthening early outbreak detection and Crop insurance Developing social protection programs response systems (yield-based) Promoting crop rotation and transition to more Use of savings and borrowing pest- and disease-resistant crops Pests and Promoting IPM techniques, including biological control Direct compensation to affected farmers diseases Diversifying seeds varietals within crops (crops) Strengthening of pest- and disease-tolerant seed development and distribution systems Improving farmer access to high-quality and affordable agro-chemicals Strengthening early detection and response Direct Strengthening quarantine measures systems compensation Improving access to vaccination services and Direct compensation to farmers supplies Developing improved exo- and endo-parasite Disposal of carcasses to avoid contagion Pests and control measures diseases Vaccination or prophylactic treatment (for Borrowing money (food, feeder stock) (livestock) example, CBPP, blackleg, foot/mouth, PPR) Quarantining of animals or culling to reduce risk Selling animals to buy medicine, fodder of contagion Developing animal tracking systems (for example, via branding) Promoting farmer adoption of crops with strong Social safety net programs (for example, food market (domestic, export) demand aid, FFW) Strengthening value chain linkages Direct cash payments to affected households Improving market information systems Promoting household savings Price Improving trade facilitation and promoting cross- Substitutions or reductions in household diet volatility border trade Improving access to quality marketing infrastructure (that is, storage, roads) Promoting enhanced role of private sector in marketing services and policy making Source: Authors. 46 Ghana: Agricultural Sector Risk Assessment generally negative impact on almost everyone, price risk soil and water management measures, including the con- may be more specific to certain stakeholder groups. For struction and use of small-scale dams, gravity irrigation example, the atypical drop in maize prices in 2013 ahead schemes, semicircular or contour stone bunds, permeable of the new planting season can be considered both a risk to zai holes, and hand-dug trenches, can help make farm- traders with large inventories and a windfall for consumers. ers better prepared to cope with and build their resilience It is also worth noting that many of these interventions, if to weather-related shocks, particularly droughts. Training implemented concurrently, can help address multiple risks in mulching, composting, zero tillage and other conserva- with positive spillover effects across the sector. tion agriculture techniques can also help. Other measures may focus on the development of public-private partner- ships designed to improve smallholder farmers’ access to DESCRIPTION OF PRIORITY appropriately sized and affordable (for example, micro) RISK MANAGEMENT irrigation technologies. While helping farmers to increase MEASURES productivity and overcome climate uncertainty, irrigation development can also help to improve drainage of flood- The following section provides a brief description of prone areas. select interventions highlighted in table 6.3. IMPROVED FARMING PRACTICES STRENGTHENING EXTENSION Strengthening farmers’ capacity, knowledge, and self- Agricultural extension, whether delivered via face-to-face reliance through training in improved agricultural prac- demonstrations or via information and communication tices not only increases productivity, but also reduces technology (ICT), is generally recognized as a strong con- farmers’ vulnerability to agricultural risks. For example, tributor to agricultural development. It can also play a promoting seed varietal diversity within mixed-cropping useful role in improving agronomic practices for manag- systems, the use of short-cycle and longer-cycle crops ing risks. The primary focus of extension has traditionally and drought-tolerant varieties, and staggered or succes- been the transfer of technology and agronomic advice to sion planting are all interventions that could help mitigate farmers, as well as information and access to resources, risks associated with increasingly erratic rainfall patterns. which are all essential components for improved on-farm Improved farmer use of integrated pest management agricultural risk management. However, ensuring effective practices, such as 1) crop rotation and crop sequencing agricultural extension can be both resource intensive and to provide disease breaks for susceptible crops; 2) use of challenging in a multi-stakeholder, resource-constrained resistant cultivars and varieties; and 3) application of non- environment. According to a 2011 survey of MoFA exten- chemical control practices (for example, thermal), would sion officers in the Upper East Region’s Bongo District, help maximize biological prevention of pests and diseases as much as 71 percent of their working time was spent without the need for high-cost, synthetic agro-chemicals. facilitating farmer access to donor resources (EWB 2011). Where agro-climatic conditions permit, these measures There is need to refocus their efforts on improving farm- might also include encouraging farmers to transition to ers’ agronomic practices, facilitating technology transfer, more pest-resistant crops, for example, from millet and and enhancing information access to help them improve sorghum to maize, which has a relatively high yield as well productivity and reduce risks. Deployment of ICT can as resistance to pre-harvest pest attacks. supplement face-to-face interaction and reduce transac- tion costs associated with extension. SOIL AND WATER CONSERVATION As elsewhere, decreasing soil fertility, land use pressures, IMPROVING VETERINARY SERVICES and highly variable rainfall patterns (often associated with The provision of quality livestock health services is criti- climate change) pose a significant threat to farmers’ liveli- cal to safeguarding animal health. Such services also serve hoods and incomes. These challenges also expose farm- as the primary system for early detection and emergency ers to considerable risk. Extension training on improved response in the event of disease outbreaks to combat their Risk Prioritization 47 spread and reduce related losses. They are also crucial unit; and 3) the contracting out of selected VSD activities for managing disease prevention and eradication efforts. to the private sector and the mobilization of community- In Ghana, veterinary services are currently centralized based delivery systems (Diop, Daborn, and Schneider under MoFA’s VSD (see box 6.1). A 2008 performance 2011). Community-based participation often involves assessment of veterinary services in Ghana highlighted 1) training community-selected and community-based rep- concerns over the capacity of veterinary professional and resentatives in basic animal health care and livestock pro- paraprofessional staff to deliver critically important clinical duction techniques. Several studies of the approach and services; and 2) significant weaknesses, notably inadequate its role in animal health services delivery have concluded transport and insufficient operational funds, adversely that community-based approaches offer viable alternatives affecting VSD staff capacity to undertake effective and sus- to the resource-constrained and poorly functioning public tainable epidemiological-surveillance and disease control veterinary services. Experiences from these programs indi- activities. Key recommendations included 1) the restruc- cate that such programs encourage the participation of turing of VSD extension to make it more independent and the local communities in the design and delivery of animal to separate it from other extension services; 2) the return of health care services. They can also empower individuals to Animal Production and Animal Health into one technical determine the type of animal health services they receive. In some areas, conditions may exist that would permit full privatization using this approach. BOX 6.1. VETERINARY SERVICES IN GHANA The Veterinary Service Directorate (VSD) is mandated to check meat and live animal imports for contagious diseases, ENSURING QUALITY INPUTS and conduct surveillance (that is, inspections, surveys) and Relative to its regional neighbors, Ghana boasts properly prevention (that is, preparations of vaccines, vaccinations). functioning markets for agricultural inputs. For most crops It also controls the quality of services provided by private veterinarians, animal health assistants, community live- and in most production zones, farmers can easily obtain stock workers, and so on. Most districts in the northern improved seeds, fertilizers, agro-chemicals (for example, regions reportedly have one or several technical agents, but herbicides, insecticides), and farming equipment. With not every district has a District Veterinary Officers (DVO). government support as well as help from the Alliance There are currently four DVOs in Upper West region (nine for a Green Revolution in Africa and other development districts with an average population of 12,500 animals), partners, agro-dealer networks have expanded in recent and three or four in Upper East Region (nine districts with an average population of nearly 33,000). The Northern years, penetrating deeper into rural markets and making Region, on the other hand, has an average of one DVO inputs and related services more readily available to small- per district for an average population of 12,000. holder farmers. These investments were designed to help farmers raise productivity by increasing yields. They were A common complaint in the VSD is that MoFA institu- tional reforms have complicated and lengthened the chain also meant to improve farmers’ access to more affordable of command: DVOs and regional veterinary officers do not crop protection products to help them control threats from report directly to the VSD, but through the Regional Direc- pests and diseases. However, as markets have expanded, tor of Agriculture. When prompt resource allocation and so too have illicit practices in the production and market- action is required (for example, to confront an anthrax or ing of inputs. Although reliable statistics are not available, CBPP outbreak) reaction time can therefore be much lon- anecdotal evidence would suggest that counterfeited and ger. Another issue flagged to the team in northern regions adulterated products have proliferated in recent years. is that due to the late release of MoFA operational funds, the central veterinary laboratory requests up-front pay- This trend poses a significant risk to cash-strapped farmers ment from regional offices in exchange for vaccines. This and agro-pastoralists who are looking for ways to reduce also risks delaying critical interventions. Finally, another pest- and disease-related losses. This situation engenders common complaint heard from both VSD staff and some distrust among farmers and can discourage farmer invest- producers is that poorly labeled veterinary supplies and ments into productivity-enhancing inputs. Improved medicines of questionable effectiveness are readily found product tracking systems and stronger monitoring and on local markets. enforcement mechanisms could help tamp down abusive 48 Ghana: Agricultural Sector Risk Assessment behavior by input product manufacturers, importers, Upgrading information management systems at district wholesalers, and retailers. Private-sector-funded training and regional levels would better enable the National and awareness-building programs targeting input retail- Meteorological Service, NADMO, the Plant Protection ers and farmers and instructing them in how to spot and and Regulatory Services Directorate (PPRSD), the Vet- report adulterated products might also help. erinary Services Directorate, the Crop Research Institute (CRI), and allied institutions to better monitor, analyze, and share information and mount more rapid, more coor- MICRO-IRRIGATION DEVELOPMENT dinated, and more effective relief efforts when adversity Reliance on rain-fed agriculture for the majority of farm- strikes. It would also help ensure that available resources ing households in Ghana makes them highly vulnerable to reach the most affected communities. Resulting improved weather-related risks. As noted earlier, this exposes them flows of information would also strengthen linkages with to droughts and other types of unseasonable weather. international centers for better weather forecasting and Micro-irrigation expansion in Ghana holds promise in pest and disease surveillance. In addition, broader input some areas with highly variable rainfall but with sufficient by stakeholders in the development and implementation access to water resources (for example, surface, ground) of weather forecast programs would encourage decentral- to help farmers better manage such risks. It also can help ized (that is, based on local agro-ecological conditions), them to improve their yields, facilitate yearlong produc- demand-driven, and more effective weather forecast tion, increase household incomes, and strengthen food production and information dissemination. All such ini- security. The development of better and cheaper pump- tiatives would ideally include systematic institutional ing technologies in recent years is opening up new oppor- capacity strengthening and awareness building to link tunities for manufacturers and distributors to tap into climate information to multimedia communication sys- growing demand among farmers in micro- and small- tems (for example, cellphones, radio, television) and tailor scale irrigation. Any irrigation development strategy information to different audiences. would greatly benefit from the identification of appropri- ate finance mechanisms to improve smallholder access to new technologies. INFRASTRUCTURE DEVELOPMENT Ghana currently ranks well behind the best-performing African countries in terms of infrastructure quality (AfDB UPGRADING INFORMATION SYSTEMS 2012). This deficiency remains a critical constraint to The ability to respond to potential threats before they agricultural sector growth. It also is an important source manifest or crises as they unfold depends on ready access of price risk for stakeholders. In addition to better access to to reliable and timely information and effective commu- knowledge and information, improved access to marketing nication. Improving farmers’ access to more and better infrastructure such as on-farm storage for farmers, ware- weather forecasting, market prices, technical farming housing and covered market stalls for traders, and well- advice, and other critical information strengthens their maintained rural feeder roads can help attenuate variability capacity to make optimal, productivity-enhancing deci- in market prices. By making it less costly and easier to store, sions while increasing their resiliency. It can also help transport, and process agricultural products, improved mitigate inter-seasonal price volatility by enabling farmers infrastructure can enable more efficient arbitration and to respond more directly and readily to shifting weather price discovery, fewer ruptures in supply and demand, and patterns and market signals. Similarly, applying local- more predictable pricing. Potential gains can also have sig- ized pest and disease forecasting techniques could greatly nificant positive spillovers in terms of enhanced food secu- reduce the response time needed to ward off large-scale rity and lower retail food prices for consumers. crop and livestock losses. Ghana’s recent introduction of field-based, farmer-monitored pheromone traps for the monitoring of armyworm infestations is one example of COMMODITY EXCHANGES low-cost but effective early warning systems that could be By reducing transaction costs, exchange trading (physical scaled up and replicated. and futures) could benefit counterparties by guaranteeing Risk Prioritization 49 for sellers payment for what is sold while concurrently existing regulations that govern the import and distribu- assuring buyers the delivery of goods. This guarantee by tion of fungicides and insecticides crucial to the control the exchange, based on guarantees by warehouse opera- of blackpod and capsid and mirid infestation. Anecdotal tors, reduces the risk of nonperformance of trade con- evidence suggests that these regulations may be hindering tracts. The greater security in trade transactions leads to farmer access (due to poor market availability and high significantly lower costs associated with contract enforce- cost) to these critical inputs, thereby compromising farm- ment, especially in markets in which litigation is time con- ers’ risk management capacity. suming and expensive. It also opens up new opportunities in the area of trade finance, and can also help in improving CROP INSURANCE price discovery and provide a platform for hedging price Agricultural insurance could be a useful tool to transfer risk. With an eye to successful models developed beyond the risk of low frequency–high impact events (for example, the region, the government has promoted the develop- drought) in Ghana. A number of agricultural insurance ment of warehouse receipts systems and has been study- initiatives (largely weather index insurance) have been ing the feasibility of establishing a national commodities piloted; however, these are experiencing severe challenges exchange. However, successful development of such an and there are questions about their sustainability. The issue exchange requires a number of prerequisites (homoge- of multiplicity of risks, basis risks, unavailability of robust neous and standardized commodities; presence of ware- yield data, limited access to agricultural credit (agriculture, house infrastructure; transparent commodity policies and forestry, and fishing loans making up only 5.9 percent of regulations; limited government interventions; and so on), the total lending portfolio in Ghana16 and only 10 percent many of which currently do not exist in Ghana. It might of rural households have access to credit17), lack of rural be helpful to address the fundamental building blocks first distribution channels, and limited affordability due to to help improve the chances of success of any potential high poverty (especially in the northern regions) are some commodity exchange. of the challenges for scaling up agricultural insurance. Appendix D provides further details on agricultural risk INSTITUTIONAL REFORM financing and insurance options for Ghana. AND STRENGTHENING Institutional transformation within NADMO designed to strengthen its operational management, staff awareness SOCIAL SAFETY NET PROGRAMS Social protection programs are typically designed to build and technical know-how, access to resources, and commu- the assets of poor households to withstand shocks and nication systems (as noted previously) could go a long way provide support when widespread shocks occur. Given to improving Ghana’s ability to prepare for and respond high levels of vulnerability and weak coping capacity to natural hazards and other threats. It would also help among low-income households, particularly those inhab- to improve the country’s vulnerability (and resiliency) to iting Ghana’s northern savannah zones, provision of multiple shocks. Such reform would require that national social safety nets is typically an integral part of effective disaster management policies and strategies be coordi- risk management. Such measures can take many forms, nated with sector programs in terms of policy making, from programs that promote household savings to direct related legislation, and processes. It would also need to ex-post cash payments and food aid delivery to affected better exploit existing synergies and ensure mutual rein- communities. The Livelihood Empowerment Against forcing measures across ministries and agencies at the Poverty (LEAP) program, for example, is a social cash national, regional, and district levels. Such efforts would transfer program that provides cash and health insurance also greatly benefit from community outreach and engage- to extremely poor households across Ghana. Its objective ment mechanisms. Similarly, COCOBOD and other is to alleviate short-term poverty and encourage long-term cocoa stakeholders could potentially benefit from the development of a more systemic pest and disease man- agement approach, especially in light of ongoing bud- 16 Bank of Ghana annual report. get constraints. This might include, inter alia, reforming 17 GFDRR Country DRM Plan for Ghana. 50 Ghana: Agricultural Sector Risk Assessment human capital development. Other social safety net ini- It is important to highlight that almost all of the meas- tiatives can include Food-for-Work programs that provide ures described in table 6.3 are complementary in nature relief while facilitating the recovery of affected communi- and will contribute to improved risk management in the ties and enhancing their future resiliency. Community-level short, medium, and long terms. However, decision mak- food and fodder banks can greatly enhance community ers are compelled to find the quickest, cheapest, and most access to food and livestock feed in times of emergency effective measures among myriad policy options. Ideally, while speeding up food distribution and relief efforts. a detailed, objective, and exhaustive cost-benefit analy- Other possible measures include micro- and meso-level sis will help in selecting the most appropriate interven- crop insurance schemes tailored to the needs of small- tion options. But conducting a cost-benefit analysis of so holder farmers. Appendix D provides a more in-depth many different options in itself is often costly and time look at the potential for the development of such schemes consuming. in Ghana, which the analysis suggests may be limited at present. An alternative approach, using decision filters to evalu- ate and prioritize among a list of potential interventions, can aid decision makers in making rational resource allo- FILTERING AND PRIORITIZING cation decisions in lieu of a detailed cost-benefit analysis. INTERVENTIONS The filters described in tables 6.4 and 6.5 are indicative Some mitigation, transfer, and coping strategies fall only and incomplete. Nonetheless, they present a useful within the purview of the state (for example, financing a first step in the right direction. The government and its countrywide animal disease eradication campaign, pro- development partners could choose other criteria as fil- viding infrastructure), whereas others are implemented ters, but it is important to ensure clarity, consistency, and at the household level, often with support from govern- objectivity while using them to evaluate decision options. ment institutions. Effective risk management requires The following decision filters were developed and used that priorities be identified and solutions implemented by the World Bank team. The study team applied these within the framework of a comprehensive risk manage- filters to facilitate a rapid assessment to obtain a first ment strategy. order of approximation, based on its assessment of the TABLE 6.4. RELATIVE BENEFITS OF RISK MANAGEMENT INTERVENTIONS Reduces Addresses Climate Climate Reduces the Compensates Multiple Improves Change Change the Risk Losses after the Loss Risks Yields Mitigation Adaptation Strengthening extension support 1 1 2 1 1 3 3 Improving farming practices 1 1 2 1 1 3 1 Soil and water conservation 1 1 2 3 1 1 1 Strengthening seed systems 1 1 2 1 1 1 1 Improving veterinary services 1 1 2 2 1 2 2 Ensuring quality inputs 3 1 2 3 1 2 3 Irrigation development 1 1 2 1 1 1 1 Upgrading information systems 1 3 2 1 3 3 3 Infrastructure development 1 1 2 1 2 2 2 Commodity exchanges 1 3 3 1 2 2 2 Crop insurance 2 2 1 1 2 2 2 Social safety net programs 2 2 1 2 2 2 2 Source: Authors. Risk Prioritization 51 TABLE 6.5. DECISION FILTERS FOR RISK MANAGEMENT MEASURES Return Sustainability Implementation Time for Sustainability of Scalability Replicability Cost Difficulty Impact of Benefits Intervention Strengthening extension High High Medium Medium Medium Medium to Medium high Improving farming practices High High Medium Medium Short High Medium Soil and water conservation High High Medium Medium Short High Medium Strengthening seed systems High High Medium Medium to high Medium Medium Low Improving veterinary services High High High Medium Short Medium Low Ensuring quality inputs High High Medium Medium Medium High Low to high Irrigation development Medium Medium High High Short High Low to medium Upgrading information High Medium to High Medium Short Medium Medium systems high Infrastructure development Low High High High Short to High Low medium Commodity exchanges Low Low Medium Medium to high Medium High Low Crop insurance High High Medium High Short High Medium Source: Authors. situation on the ground. The team presented prelimi- pests and diseases, flood, price volatility) or would nary results to government officials and other stakehold- it address only a single risk? ers at a roundtable in Accra in early June, during which » Improve yields: Would the activity also lead to yield it solicited feedback that was subsequently incorporated improvement in normal years? into the final results. They are illustrated in tables 6.4 » Mitigate climate change: Would the activity help miti- and 6.5. gate climate change (that is, by reducing greenhouse gas emission or facilitating carbon sequestration)? In table 6.4, interventions are rated according to the fol- » Climate change adaptation: Would this activity help lowing: 1) Yes, 2) No, 3), Maybe (it depends), and 4) Not stakeholders better adapt to changing climate (for Certain, based on the following criteria: example, increasing heat or extreme weather events)? » Reduce the risk: Would this activity lead to reduced exposure (that is, probability of negative event hap- For table 6.5, interventions were rated based on the fol- pening and its impact), thereby reducing risk? lowing criteria: » Reduce the losses: Would this activity lead to reduc- » Scalability (Low, Medium, High): What is the potential tion of losses (that is, financial, crops, livestock), if outreach (possibility of reaching scale or reaching a risk event were to occur? 50–60 percent of the farming population)? » Compensate after the loss: Would this activity lead to » Replicability (Low, Medium, High): How replicable is compensation to the affected stakeholders after this intervention to the wider Ghanaian agricul- they have suffered losses? ture? Is this a niche intervention with relatively » Address multiple risks: Would this activity concur- limited applicability in a selected area or can it rently address multiple risks (for example, drought, be replicated widely throughout the country? 52 Ghana: Agricultural Sector Risk Assessment » Cost (Low, Medium, High): What is relative cost of 6. Improving infrastructure (on-farm and off-farm this intervention (in comparison with all the other storage, warehouses, roads, and so on) to improve interventions listed in table 6.5)? productivity, reduce post-harvest losses, and help » Difficulty of implementation (Low, Medium, High): In manage the risk of price volatility. general, how difficult is this intervention to imple- 7. Strengthening institutional capacity of NADMO, ment? Is this a complicated and technically sophis- COCOBOD, and other relevant agencies to man- ticated intervention or is this a relatively simple age agricultural risks, especially multiple shocks in intervention to implement? the same year. » Return time for impact (Short, Medium, High): How long does it take to see results from the intervention? It is worth noting here that northern regional depart- » Sustainability of benefits (Short, Medium, Long): How ments of agriculture extend technological packages very long do the benefits continue to accrue once the much in line with many of the risk management measures intervention has taken place? mentioned by respondent stakeholders and outlined in » Sustainability of intervention (Low, Medium, High): How chapter 5. These measures include: sustainable is the intervention over time? » Use of improved certified seeds » Proper farmer use of pesticides and other agro- chemicals Based on the prioritization of risk in table 6.5 and inter- » “Fodder banks”—harvesting and storage of crop vention measures in table 6.4, the following interventions residue for livestock have been identified as having significant potential to help » Use of agro-industrial by-products to fatten live- confront the most important risks facing Ghana’s agricul- stock tural sector, namely, drought, crop and livestock pests and » Adoption of hybrid maize varieties diseases, and price volatility: » Appropriate use of inorganic fertilizers 1. Improved farming practices (for example, pro- » Water-harvesting techniques moting integrated pest management, especially in » Soil conservation practices the south) and conservation agriculture measures » Composting (particularly in the north) to manage risks. » Improvement of post-harvest storage (yams, legumes, 2. Strengthening tolerant seed (drought, pest, and dis- grains) ease tolerant) development and distribution systems. » Crop rotation, inter-cropping (grain/legume) 3. Upgrading information systems to ensure availabil- » Reforestation ity of timely and relevant weather, prices, and pest » Training of agro-chemical input dealers in safe and disease information to the farmers, traders, and product management other stakeholders, coupled with advice and knowl- edge disbursement on ways to manage them. This This assessment highlights a need both to improve and to also includes market information about production, scale up these and other risk mitigation measures already stocks, and trade of different commodities. in place as well as integrate new and complementary 4. Improved water management (for example, soil measures that will empower more communities, particu- and water conservation measures and irrigation, larly those identified as the most vulnerable, to access and especially micro-level irrigation) and drainage benefit from them. infrastructure in flood prone areas. 5. Strengthening extension systems (face-to-face, ICT-based, peer-to-peer, and so on) to enable CONCLUSION farmers to gain access to technology, agronomic This document aims to contribute to and enrich the exist- advice, and resources to put in place risk mitiga- ing knowledge base of the agricultural sector in Ghana. tion measures. It systematically analyzes agricultural risks and impacts Risk Prioritization 53 over time (1980–2012). It helps place drought and other Greater emphasis should be placed on scaling up these natural hazards within the context of other agricultural interventions and looking at systemic changes on the risks. It prioritizes the most important agricultural risks national level to make a meaningful impact on agricul- for the country based on objective criteria. It offers a tural sector risks in Ghana. framework for the development of a more comprehen- sive, integrated risk management strategy to strengthen Scaling up of these approaches would require understand- existing mitigation, transfer, and coping measures. Finally, ing the landscape of interventions, assessing their relative it provides a filtering mechanism to select an appropri- efficacy, understanding principal barriers and challenges ate set of best possible interventions for agricultural risk to success and scale, and identifying leverage points and management. necessary interventions to increase their access to a wide majority of agricultural sector stakeholders. Assessing Most of the proposed intervention areas are already cov- solutions to help prioritize specific interventions to scale ered under various components of METASIP (see table up priority programs and putting in place a roadmap will 6.6) and are being implemented, albeit on a much smaller be the next steps in the process of building resilience and scale, by government agencies and development partners. reducing vulnerability of stakeholders and the agricul- tural sector in Ghana more broadly. TABLE 6.6. INTEGRATION WITH METASIP Overall, this assessment highlights that Ghana’s agricul- Indicative tural sector, from a risk standpoint, rests on sound footing. Interventions METASIP Components The diversity of Ghana’s agro-climatic conditions, farming Improved farming Program 1.1, Program 1.4, systems, and productive assets within those systems shields practices (for example, Program 1.6, Program 2.1, the sector in the aggregate from massive debilitating shocks. IPM, on-farm Program 2.3, Program 4.1, Sectorwide vulnerability to risks is thus limited. However, a practices, crop Program 5.1 deeper analysis of crop- and region-specific risks reveals a rotation) number of insights with important implications for agricul- Improving information Program 1.1, Program 1.4, tural risk management in Ghana. Among these, interven- systems (for example, Program 1.5, Program 2.2, tions should aptly focus on reducing the vulnerability of price, weather, early Program 2.3, Program 3.1, communities in the country’s three northern regions to pro- warning, extension) Program 5.1, Program 5.2, Program 6.1, Program 6.3 duction shocks such as drought and floods while increasing their resiliency to recover in their aftermath. Other priori- Infrastructure Program 1.1, Program 1.3, improvement Program 1.4, Program 1.6, ties include improving systems for pest and disease manage- program 1.7, Program 2.2, ment and strengthening farmers’ capacity to manage these Program 2.3, Program 2.5, risks and improving information systems and infrastructure Access to quality inputs Program 1.1, Program 1.4, to help manage price volatility. In addition, strengthening (that is, agro-chemicals, Program 2.1, Program 2.3, the capacity of government institutions at the national, vaccines, tolerant Program 5.1 regional, and district levels to withstand and respond to seeds) multiple shock events will have the biggest impact on sector Source: Authors. growth, rural livelihoods, and food security. 54 Ghana: Agricultural Sector Risk Assessment REFERENCES AfDB (African Development Bank). 2012. “Republic of Ghana: Country Strategy Paper, 2012–2016.” Country Operations Department (April). Amissah, L., B. Kyereh, and V. K. Agyeman. 2010. Wildfire Incidence and Management in the Forest Transition Zone of Ghana: Farmers Perspective. CSIR-Forestry Research Insti- tute of Ghana. Anim-Kwapong, G. J., and E. B. Frimpong. 2004. Vulnerability and Adaptation Assessment under the Netherlands Climate Change Studies Assistance Programme Phase 2 (NCCSAP2): Vulnerability of Agriculture to Climate Change—Impact of Climate Change on Cocoa Produc- tion (Ghana: Cocoa Research Institute of Ghana). Antwi-Agyei, Phillip, Evan D. G. Fraser, Andrew J. Dougill, Lindsay C. Stringer, and Elisabeth Simelton. 2011. “Mapping the Vulnerability of Crop Production to Drought in Ghana Using Rainfall, Yield and Socioeconomic Data,” Centre for Climate Change Economics and Policy Working Paper No. 55, Sustainability Research Institute Paper No. 25. Aryeetey, Ernest, and Ravi S. Kanbur. 2004. Macroeconomic Stability, Growth and Poverty Reduction. Accra, Ghana: Institute of Statistical, Social and Economic Research (ISSER). Baba, A., N. Musah, E. Mumuni, O. Abayomi, and M. B. Jibrel. 2013. “Effects of Floods on the Livelihoods and Food Security of Households in the Tolon/Kum- bumgu District of the Northern Region of Ghana.” American Journal of Research Communication 1 (8): 160–171. Bank of Ghana. 2012. Statistical Bulletin (June). IDPS Department, Bank of Ghana. Accra: Ghana. ———. 2013. Statistical Bulletin (June). IDPS Department, Bank of Ghana. Accra: Ghana. Ben Ltaifa, Nabil, Stella Kaendera, and Shiv Dixit. 2009. “Impact of the Global Financial Crisis on Exchange Rates and Policies in Sub-Saharan Africa.” African Department 09/3. Washington, DC: International Monetary Fund. Biderlack, Lisa, and Jonathan Rivers. 2009. Comprehensive Food Security & Vulnerability Analysis (CFSVA) Ghana (VAM World Food Programme Food Security Analysis). Birol, E., and D. Asare-Marfo. 2008. “Impact of HPAI on Ghanaian Rural Chicken Producers’ Incomes.” DFID, IFPRI, ILRI. ———. 2008. Impact of HPAI on Ghanaian Rural Chicken Producers’ Incomes. IFRPRI/DFID. http://r4d.dfid.gov.uk/PDF/Outputs/HPAI/IFPRI_ILRI_rbr01 .pdf. “Bushfires in Ghana.” 1996. International Forest Fire News (IFFN) 15 (September): 24–29. CIAT. 2011. “Predicting the Impact of Climate Change on the Cocoa-Growing Regions in Ghana and Cote d’Ivoire” (September). Risk Prioritization 55 Comprehensive Food Security and Vulnerability Assessment (CFSVA), MoFA/WFP, 2012. Diop, B., C. Daborn, and H. Schneider. 2011. “PVS Gap Analysis Report.” World Organization of Animal Health (August). Dontwi, Joyce, I. K. Dontwi, F. N. Buabeng, and S. Ashong. 2008. Vulnerability and Adaptation Assessment for Climate Change Impacts on Fisheries (Ghana: Netherlands Cli- mate Assistance Programme [NCAP]). EWB (Engineers Without Borders). 2011. “Agricultural Extension in Bongo District (UER).” Survey data collected for IFPRI, June–July. Environmental Protection Agency. 2002. National Action Program to Control Drought and Desertification (April). Accra, Ghana. FAO/WFP Crop and Food Supply Assessment Mission to Northern Ghana. March 13, 2002. http://www.fao.org/docrep/005/y6325e/y6325e00.htm#P1318_35461. FAOSTAT. 2008. Commodity Balances. and FAOSTAT. 2009. Commodity Balances. FAO (Food and Agriculture Organization). 2005. Fertiliser Use by Crop in Ghana. Rome, Italy: FAO. ———. 2013. Current Worldwide Annual Meat Consumption per capita, 2013. Frederick, A., O. David, T. Genesis, O. Justice, and Ernest K. A. Afrifa. 2010. “Impact of Floods on Livelihoods and Vulnerability of Natural Resource Dependent Com- munities in Northern Ghana.” University of Cape Coast, Cape Coast, Ghana. Ghana Statistical Service. 2009a. “Digest of Macroeconomic Data.” Statistical Abstract 2009. Accra. ———. 2009b. “Ghana’s Economic Performance.” Accra. IFPRI (International Food Policy Research Institute). 2011. “Irrigation Development in Ghana: Past Experiences, Emerging Opportunities, and Future Directions.” GSSP Working Paper No. 27 (March). IFPRI, Washington, DC. ———. 2012. “The Partially Liberalized Cocoa Sector in Ghana: Producer Price Determination, Quality Control and Service Provision.” Discussion Paper 01213 (September). IFPRI, Washington, DC. ———. 2013. West African Agriculture and Climate Change: A Comprehensive Analysis. Mono- graph (April). Washington, DC: IFPRI. “Improved Animal Health for Poverty Reduction and Sustainable Livelihoods, Animal Production and Health.” 2002. Paper 153. Food and Agriculture Organization of the United Nations. Josserand, H. 2013. “Volumes and Values of Regionally Traded Staple Commodities— Current and Prospective Analysis.” Paper presented at the Food Across Borders Conference, Accra, Ghana, January. USAID. Laderach, Peter. 2011. Predicting the Impact of Climate Change on the Cocoa-Growing Regions in Ghana and Cote d’Ivoire (CIAT, Climate Change, Agriculture, and Food Security). Masters, G., P. Baker, and J. Flood. 2010 “Climate Change and Agricultural Com- modities,” CABI Working Paper 2 (38 pp.). McSweeney, C., M. New, and G. Lizcano. 2008. UNDP Climate Change Country Profiles: Ghana (The United Nations Development Programme). MoFA (Ministry of Agriculture and Food). Annual Progress Reports. 1992, 1994, 1996, 1997, 1998, 1999, 2003, 2004, 2005, 2006, 2007, 2008, 2009. Accra, Ghana: MoFA. 56 Ghana: Agricultural Sector Risk Assessment ———. 2008. “Ghana National Investment Brief.” High Level Conference on Water for Agriculture and Energy in Africa: The Challenges of Climate Change. Sirte, Libyan Arab Jamahiriya, December 15–17. ———. 2010. “Agriculture in Ghana: Facts and Figures.” Statistics Research and Information Directorate (April). Accra, Ghana: MoFA. ———. 2011. Annual Progress Report, Northern Region. Accra, Ghana: MoFA. Namara, R. E., L. Horowitz, and B. Nyamadi. 2011. “Irrigation Development in Ghana: Past Experiences, Emerging Opportunities, and Future Directions.” Ghana Strategy Support Program (GSSP), Working Paper No. 0026. Interna- tional Food Policy Research Institute (IFPRI), Accra, Ghana. OECD/Sahel and West Africa Club. 2008. “The Future of Livestock in the Sahel and West Africa: Potentialities and Challenges for Strengthening the Regional Mar- ket.” Paris, France: OECD. Ofori-Sarpong, E. 1986. “The 1981–1983 Drought in Singapore.” Singapore Journal of Tropical Geography 7 (2). Oteng-Darko, P., N. Kye-Baffour, and E. Ofori. 2012. “Simulating Rice Yields under Climate Change Scenarios Using the CERES-Rice Model,” African Crop Science Journal (CSIR-Crops Research Institute, Department of Agricultural Engineering, KNUST, Kumasi, Ghana) 20, no. S2: 401–408. de Pinto, Alessando, Ulac Demirage, Akiko Haruna, Jawoo Koo, and Marian Asa- moah. 2012. “Ghana Strategy Support Program: Climate Change, Agriculture, and Foodcrop Production in Ghana.” IFPRI Policy Note No. 3. Republic of Ghana. 2011. “Second National Communication under United Nations Framework Convention on Climate Change.” (Ghana: Environmental Protection Agency). Rushton, J., D. Pilling, and C. L. Heffernan. 2002. “A Literature Review of Livestock Diseases and Their Importance in the Lives of Poor People.” Nairobi, Kenya: International Livestock Research Institute. Sagoe, Regina. 2006. Climate Change and Root Crop Production in Ghana: A Report Prepared for Environmental Protection Agency (EPA), ACRA-Ghana (Kumasi, Ghana: The Crops Research Institute). Sarpong, Daniel Bruce, and Samuel Asuming-Brempong. 2004. “Responding to Shocks in Ghana: The Agricultural Sector as a Social Safety Net.” eJADE. Elec- tronic Journal of Agricultural and Development Economics 1 (1): 117–137. 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?). Schmitz, C., and D. Roy. 2009. “Potential Impact of HPAI on Ghana: A Multi-Market Model Analysis.” DFID, IFPRI, ILRI. Seini, A. Wayo. 2002. “Agricultural Growth and Competitiveness under Policy Reforms in Ghana.” Technical Publication No. 61. Institute of Statistical, Social and Economic Research (September). Sudharshan, Canagarajah, and Claus C. Porter. 2003. “Evolution of Poverty and Wel- fare in Ghana in the 1990’s: Achievements and Challenges.” Africa Region Work- ing Paper Series No. 61, World Bank, Washington, DC. Risk Prioritization 57 Vabderpuye-Orgle, Jacqueline. 2004. “Economy and Poverty in Ghana in the 1990’s: A Review.” Discussion Paper No. 29, Institute of Statistical, Social and Economic Research, University of Ghana, Legon. Wheeler, David. 2011. “Quantifying Vulnerability to Climate Change: Implications for Adaptations Assistance,” CGD Working Paper 240 (Washington, DC: Center for Global Development). World Bank. 2005. “Managing Agricultural Production Risk: Innovations in Develop- ing Countries.” Agriculture and Rural Development Department. Washington, DC: World Bank. ———. 2006. “Managing Food Price Risks and Instability in an Environment of Market Liberalization.” Agriculture and Rural Development Department. Wash- ington, DC: World Bank. ———. 2010. Disaster Risk Management Programs for Priority Countries: Ghana. Washington, DC: World Bank. ———. 2012a. “Republic of Ghana: Cocoa Sector Policy Brief ” (May 24). Washing- ton, DC: World Bank. ———. 2012b. “Supply Chain Risk Assessment: Cocoa in Ghana.” Agriculture and Rural Development Department. Washington, DC: World Bank. 58 Ghana: Agricultural Sector Risk Assessment APPENDIX A REGIONAL RISK PROFILES18 18 Source: Authors’ analysis based on meteorological data from weather stations in Ghana and historical data from MoFA. Risk Prioritization 59 60 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 61 62 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 63 64 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 65 66 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 67 68 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 69 APPENDIX B COMMODITY RISK PROFILES19 19 Sources: FAOSTAT; MoFA; World Development Indicators Database 2014; authors’ calculations. Risk Prioritization 71 72 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 73 74 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 75 76 Ghana: Agricultural Sector Risk Assessment Risk Prioritization 77 78 Ghana: Agricultural Sector Risk Assessment APPENDIX C RAINFALL PATTERNS AND IMPLICATIONS FOR CROP PRODUCTION BACKGROUND An analysis of rainfall data provided useful information on the level and distribu- tion of rainfall by region and the impact of various rainfall characteristics on crop yields. Ghana has 99 weather stations located throughout the country, although some regions have a higher density than others. Analysis was based on daily rainfall data from 1981 to 2010. Figure C.1 shows the distribution of the weather stations (orange diamonds). As the weather stations do not have information on the region to which they belong, distance from the centroid of each region (i) was calculated for each station (j) using the Euclidean Distance formula: 2 Dist xi xj yi yj ) Where Dist = Euclidean Distance xi = longitude from region’s i centroid xj = longitude from station j yi = latitude from region’s i centroid yj = latitude from station j Each station was assigned to the region whose distance to the centroid was the small- est. Reference marks for the centroids of each region are indicated by the navy dots in figure C.1. Risk Prioritization 79 FIGURE C.1. WEATHER STATION DISTRIBUTION WITH REGION CENTROIDS 11 Vea G Manga Bawku Bin Bi Binduri u Bolgatanga atanga Zuarungu ngu gu Upper rE as East a s Babile Upper West Walewale 10 Pong Tamale Nyankpala Northern Damongo 9 Bui Salaga Kpandae Mankango Krachi Nkwanta 8 Kintampo Latitude Prang Atebubu Brong Ahafo Berekum Akaa Ejura Volta Hohoe Bechem Mampong Ashanti 7 Barekese Effiduasi Agogo Forifori Kpandu Amedzofe Goaso Owabi Kwad Kwadaso K d so Ashant Konongo ni ahu Tafo Kwah Kpeve Mprae Mpraeso Mpra wo es a i Bibian Bekwai aiAshanti Nkawka Begoro Eastern Asesewa ter ternn Bo e Kroso Krom om Tsito Ts Pankese Akrokerri College Bunso Farms k Akosom bo Ahunda Adaklu Kan Kanayereb K an o Tafo Ta Sefwi Wiawso Obuasi Kibi Huhuny H a Akokoaso Kp pon ong Kpong Ohawu Kade m nya A Soma s Asutsuar e A A Addid idome ife if e Weta Afife-Weta 6 Enchi est We Western Abes er ewa Gy n am an n Dunkwa Agric Kusi Asamankese Ak Akro Akropong opong Pomadze Nsawam g Akwapi Ak Akw m Aburri Avey v ime Keta Af A ienya si Assin Nyankumas Pokoase a Greater Accra Twif ku Atieku Praso fo P n As Breman Asiku Ce Centra C en e ntra n aw Ag l d ko wanyako A ona Swedru Bogoso Tarkwa Asuansi n a nne Winneb Apam Asebu Benso C Cape Coast 5 H lf Assini Ha Esiama Pretsea nd Komenda Princes Town 4 –3.5 –3.0 –2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 Longitude Source: Authors’ analysis based on meteorological data from weather stations in Ghana. cumulative rainfall variable was calculated for each region, RAINFALL DISTRIBUTION according to the formula: Most rain occurs during the summer months from June to September, followed by a dry winter from November to i mar Prec i − ( ∑ oct t ) StdRaini = i March. In the south, it is also common to find a dry period si during August, which is usually referred to as the “dog Where days of summer,” due to its relationship to the Dog Star StdRain standardized cumulative rainfall of Sirius in the Canis Major constellation. A uni-modal Pre daily rainfall rainfall pattern is thus observed in the lower rainfall north- μ mean yearly rainfall ern regions and a bi-modal pattern in the central and σ standard deviation of yearly rainfall southern regions. Figure C.2 shows the monthly distribu- i year tion of rainfall for each region. This variable makes it easier to discern drought and As shown by these charts most rainfall occurs in the excess rainfall events. Table C.1 shows the standard- March–October period, with an average of more than ized cumulative rainfall by year and region, with red 100 mm per month, followed by a dry season from signifying a drought event and green an excess rainfall November–February. The ensuing analysis focuses on event. observed rainfall during the period March–October. This analysis shows that drought typically affects numer- DROUGHT AND EXCESS ous regions simultaneously. RAINFALL Cumulative rainfall for all stations was calculated for the Drought years: 1982, 1983, 1986, 1990, 1992, 1998, and March–October period and the average of all stations 2005. During these years, rain was more than one stand- within a region was used as the basis for analysis. To deter- ard deviation below average in at least three regions. mine whether a year was dry or humid, a standardized Drought was particularly severe and widespread in 80 Ghana: Agricultural Sector Risk Assessment FIGURE C.2. MONTHLY RAINFALL PATTERNS BY REGION 300 Zuarungu Station - Upper East Region 250 Babile Station - Upper West Region 200 Nyankpala Station - Northern Region 180 250 200 160 Cumulative rainfall Cumulative rainfall 140 Cumulative rainfall 200 150 120 150 100 100 80 100 60 50 40 50 20 0 3 10 43 99 138 173 280 189 58 5 2 2 6 25 64 106 134 184 233 195 65 8 5 2 8 32 76 109 135 156 168 179 71 6 2 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 250 Ejura Station - Brong Ahafo Region 250 Kwadaso Station - Ashanti Region 200 Begoro Station - Eastern Region 180 200 200 160 140 Cumulative rainfall Cumulative rainfall Cumulative rainfall 150 150 120 100 100 100 80 60 50 50 40 20 7 25 99 150 174 181 132 90 214 169 38 20 28 51 114 151 169 209 127 86 177 195 96 41 23 53 122 145 158 184 124 100 172 159 49 24 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 250 Akaa Station - Volta Region 250 Breman Asikuma Station - Central Region 250 Sefwi Wiawso Station - Western Region 200 200 200 Cumulative rainfall Cumulative rainfall Cumulative rainfall 150 150 150 100 100 100 50 50 50 8 28 91 126 151 190 177 193 175 151 49 24 19 63 119 135 164 214 118 72 136 180 114 48 23 40 127 128 193 213 125 72 155 208 74 28 0 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 160 Afienya Station - Greater Accra Region 140 120 Cumulative rainfall 100 80 60 40 20 15 21 60 70 137 148 56 25 72 92 60 34 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Authors’ analysis based on meteorological data from weather stations in Ghana. 1983, with nine regions affected—including several with Excess rainfall years: 1987, 1989, 1991, 1995, 1999, cumulative rainfall more than two standard deviations 2002, 2007, and 2010. Rainfall was more than one stand- below the average. The most recent dry year was 2005 ard deviation above average during these years, meaning when the Eastern and Volta regions suffering from very that it was more than adequate. The most humid year low rains. These data suggest that there is a 23 percent occurred in 2007, affecting six regions—with rainfall in probability (7 out of 30 years) that drought will occur in the Upper East Region more than two standard devia- at least one region. tions above the average. Risk Prioritization 81 TABLE C.1. STANDARDIZED CUMULATIVE RAINFALL Up Up G Dry Exc Year East West North Brong A Ashanti Eastern Volta Central Western Accra Regs Regs Conclusion 1981 –1.56 –0.75 0.01 –0.50 0.27 0.10 0.17 1.03 0.55 0.52 1 1 Neutral 1982 –0.03 –0.07 –0.78 –1.33 –1.73 –1.77 –1.83 –0.05 –0.98 0.62 4 0 Dry 1983 –0.81 –3.07 –1.55 –1.24 –1.99 –1.90 –2.09 –2.57 –1.58 –2.39 9 0 Dry 1984 –1.53 –1.91 –0.26 –0.17 1.78 0.77 0.95 0.50 1.72 0.84 2 2 Neutral 1985 –0.98 –0.09 0.28 0.57 0.85 0.74 –0.20 –0.14 –0.04 0.50 0 0 Neutral 1986 –0.42 0.69 –0.66 –0.26 0.73 –1.07 –0.88 –1.79 –1.31 –1.18 4 0 Dry 1987 –0.46 0.11 –0.27 0.75 0.83 1.38 0.49 1.40 2.15 0.86 0 3 Excess 1988 –0.20 –0.57 0.29 –0.35 –0.90 –0.06 0.47 –0.06 –0.12 0.40 0 0 Neutral 1989 1.39 0.12 2.23 1.35 0.25 0.40 1.30 0.80 0.92 0.80 0 4 Excess 1990 –1.22 –1.00 –0.86 –1.51 –1.02 –0.30 –1.38 0.06 –0.68 5 0 Dry 1991 0.58 –0.38 2.06 0.14 –0.26 1.85 1.03 1.62 –0.17 2.45 0 5 Excess 1992 0.57 –0.23 –1.20 –1.27 –0.86 –0.85 –0.97 –0.91 –1.05 –0.96 3 0 Dry 1993 –0.31 0.28 0.04 –0.26 –0.35 1.16 –0.09 –0.35 –0.31 –0.42 0 1 Neutral 1994 1.18 0.22 0.05 –1.51 –0.46 –0.04 –0.41 0.24 –1.64 –0.88 2 1 Neutral 1995 –0.72 1.19 0.31 0.37 0.48 1.71 1.76 0.77 0.03 0.47 0 3 Excess 1996 1.07 –0.81 0.60 0.01 –0.71 –0.46 –1.00 0.21 0.66 0.92 0 1 Neutral 1997 –0.18 0.62 –0.01 0.36 –0.42 –0.58 0.09 0.28 –0.31 0.91 0 0 Neutral 1998 0.45 –0.52 –2.08 –0.01 –0.51 –0.30 0.16 –1.01 –1.28 –1.17 4 0 Dry 1999 1.78 1.45 1.43 0.63 1.50 –0.09 0.74 0.52 –0.11 0.51 0 4 Excess 2000 –0.09 1.85 0.01 –0.07 –0.56 –0.35 0.71 –1.07 0.07 –1.53 2 1 Neutral 2001 0.10 –0.57 –0.94 –0.61 –0.34 –0.46 –1.39 –0.71 0.19 0.08 1 0 Neutral 2002 –0.68 0.36 –0.13 1.03 1.22 0.44 0.49 0.62 1.76 0.65 0 3 Excess 2003 0.45 0.45 0.05 –0.88 –0.35 0.00 0.58 0.08 –0.03 –0.31 0 0 Neutral 2004 –1.07 –0.24 0.93 –0.75 –0.19 –0.66 –0.30 –0.74 –0.98 1 0 Neutral 2005 –0.42 0.46 –0.72 0.02 –0.71 –1.97 –2.04 –1.07 –0.60 –0.99 3 0 Dry 2006 –0.19 1.01 –0.62 –0.44 0.73 –0.21 0.22 0.61 0.06 –0.09 0 1 Neutral 2007 2.62 –0.39 0.80 1.84 1.47 1.25 1.44 1.13 0.83 0 6 Excess 2008 –1.05 1.49 –0.25 1.39 0.90 1.02 0.85 0.98 0.54 1 3 Neutral 2009 1.00 0.10 0.50 –0.51 0.36 –0.12 0.46 –0.55 –1.17 –0.89 1 1 Neutral 2010 0.72 0.06 1.51 3.59 0.19 0.52 –0.04 0.98 1.16 0.56 0 3 Excess Dry Years 5 2 4 4 3 5 4 6 6 4 Exc Years 6 4 5 3 5 5 5 4 5 1 Source: Authors’ analysis based on meteorological data from weather stations in Ghana. 1. Cumulative rainfall (cumrain). The sum of rain- THE IMPACT OF RAINFALL fall from March to October, it measures the total ON CROP YIELD amount of rain that accumulates yearly from RAINFALL PARAMETERS March to October. It is expressed in millimeters. Crops are sensitive to rainfall in different ways. Low cumu- 2. Onset date (onset). The time of the year in which lative rainfall is the main determinant of yield, but crops the rainy season starts, defined as the first day of can also be affected by late onset of the rainy season or an the year with 20 mm or more. It is measured as the early cessation of rains. Prolonged periods without rain number of days from the start of the year. can also reduce yields, as can excess rainfall. The follow- 3. Cessation date (cessation). The day the rainy sea- ing variables were thus calculated for each weather station son ends is defined as the day in which 90 percent and year as the basis for closer analysis of the relationship of total rainfall period occurs. It is measured as between rainfall and crop yield: the number of days from the start of the year. 82 Ghana: Agricultural Sector Risk Assessment TABLE C.2. IMPACT OF RAINFALL PARAMETERS ON CROP YIELD Upper Upper Brong- Parameter East West Northern Ahafo Ashanti Eastern Volta Central Western Cum rain Rice (−) Maize (+) (32%) (24%) GNuts (−) Yams (+) (35%) (23%) Onset of rain GNuts (+) GNuts (+) (27%) (39%) Number of rainfall days Rice (−) (26%) Cessation date Length of rainy season Dry spell Yams (−) (46%) Max days Rice (−) Maize (−) (42%) (26%) GNuts (−) (28%) Source: Authors’ analysis. Note: GNuts = groundnuts. 4. Length (length). The length of the rainy season is Yield = β0 + β5 length defined as the difference between the cessation date Yield = β0 + β6 drysp and the onset date. It is measured as number of days. Yield = β0 + β7 max 10 days 5. Rain days (events). The number of days in the period when rainfall was higher than 1 mm. Results are reported in table C.2 for those crops and 6. Dry spell (drysp). The longest number of consecu- regions where the regression coefficient was statistically tive days without rain. significant (at 5 percent). The coefficient of determina- 7. Extreme excess rainfall (max 10 days). The yearly tion, which measures the proportion of the variability in maximum amount of cumulative rainfall in any yield explained by each rainfall variable, is also reported 10 consecutive days. (in brackets) to indicate the magnitude of this impact. The influence of rainfall on yield was examined using The short time period for analysis and limited variabil- regional production data for maize, rice, millet, ground- ity of some of the data limited the explanatory power nuts, cassava, and yams for 1992–2009. of these regressions, although some general trends are apparent. REGRESSION ANALYSIS The rainfall parameters described above were averaged The impact of individual rainfall parameters is most across the weather stations in each region and regressed apparent for the production of rice and groundnuts in on yield, as described below. the Upper East Region. Excess rainfall is the major risk, rather than drought, as shown by the negative signs for the Yield = β0 + β1 cumrain impact of cumulative rainfall, number of rainfall days, Yield = β0 + β2 onset and the maximum rainfall in any 10-day period. Ground- Yield = β0 + β3 events nut yields respond positively to the earlier onset of rain in Yield = β0 + β4 cessation both the Upper East and Upper West regions. Risk Prioritization 83 FIGURE C.3. CORRELATION MATRIX PLOT Correlations (ghanamix 16v cumrain onset ces90 durac drysp Source: Authors’ analysis. The impact of rainfall in other regions is limited. Maize Some variables are closely related, such as the length of and yam yields are vulnerable to drought in the Brong- the rainy season (durac) and the cessation date (ces90)— Ahafo Region, yam yields are vulnerable to drought in the the higher the cessation date, the longer the rainfall period. Central Region, and maize yields are vulnerable to excess These correlations can also be highly negative—such as the rainfall in the Ashanti Region. This suggests that factors correlation between the length of the rainy season (durac) other than rainfall may be more important determinants and the onset date, because the later the rainy season starts of yield in these regions. the shorter the duration of the rainy season will be. PRINCIPAL COMPONENT Owing to this high correlation, principal component anal- ysis was used as a further means to analyze the impact of ANALYSIS these parameters on crop yield. Table C.3 shows the first As the variables used for analysis are all different attributes three eigen-values and the corresponding proportion of of the same weather phenomenon (rainfall), some will be variance explained. correlated. Figure C.3 shows the correlation matrix for the first six variables.20 The first component explains 52 percent of the total vari- ance for which the six variables account, and the second 20 The excess rainfall variable was not used in the Principle Component Analysis component explains an additional 25 percent of the vari- as it was introduced at a later stage. ance so that the cumulative variance explained is more 84 Ghana: Agricultural Sector Risk Assessment TABLE C.3. PCA ANALYSIS: THREE EIGEN TABLE C.4. CORRELATION OF VALUES AND PROPORTION COMPONENTS OF VARIANCE EXPLAINED Variable Factor 1 Factor 2 % Total cumrain 0.114 0.893 Eigen-value Variance Cumulative % onset −0.746 −0.263 1 3.1297 52.16 52.16 events 0.228 0.895 2 1.5165 25.27 77.44 ces90 0.861 0.053 3 0.6384 10.64 88.08 durac 0.984 0.174 drysp −0.113 −0.777 Source: Authors’ analysis. Source: Authors’ analysis. than 77 percent. This reduces the dimensionality of the original problem from six variables into two compo- nents with a reduction of variability of only 23 percent FIGURE C.4. FACTOR LOADINGS PLOT Rotation: varimax raw (100 percent−77 percent). A third component would add Extraction: principal components another 10 percent of the variability explained. The usual practice is to retain as many components as eigen values 1.0 cumrain 0.8 events are higher than one, which suggests retaining the first two 0.6 components. Table C.4 shows the correlation (factor load- 0.4 durac Factor 2 ings) of each component, with each of the variables in the 0.2 ces90 0.0 two-factor solution (retaining the first two components, –0.2 onset meaningful loadings marked in red): –0.4 –0.6 drysp –0.8 Figure C.4 shows that the length of the rainy season –1.0 –1.0 –0.8 –0.6 –0.4 –0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 (durac), cessation date (ces90), and onset date (onset) are Factor 1 highly correlated among themselves, and so constitute Source: Authors’ analysis. the first factor together with the negatively correlated onset date. This factor can be taken to represent the length of the rainy season given that the length is high, Based on this two-factor solution, it is possible to derive the onset date is low, and the cessation date is also high. factor scores, which are the transformations of the origi- Hence, when this factor is large, the rainy season was nal six variables into the two new variables (factors). These very long. scores are standardized, so that the mean is equal to zero and the standard deviation is equal to one. Figure C.5 Factor 2 consists of cumulative rainfall (cumrain), number shows the mean scores for each factor by region. of rainy days (events), and length of the dry spell (drysp), which are positively correlated, together with dry spell, The length of the rainy season is shorter than normal in which is negatively correlated. This means that when the the Upper East and Upper West regions, as their mean dry spell is very long, cumulative rainfall and number of factor scores are smaller than −1. By contrast, the Ashanti events will be low. This second factor represents the inten- Region has the highest mean score for factor 1, mean- sity of rainfall during the year. ing that the season is usually longer. For the intensity factor, the Volta and Western regions seem to have the These two factors (or principal components) by definition most intense rainfall because their mean score is almost are built orthogonally, meaning that they are independent one standard deviation above the mean (0.8), whereas the between each other. They suggest that rainfall in Ghana Greater Accra Region has the lowest intensity of rainfall. has two main attributes: the length of the rainy season The Brong-Ahafo, Ashanti, Eastern, Volta, and Western (factor 1) and the intensity of rainfall (factor 2). regions have similar rainfall conditions. Risk Prioritization 85 FIGURE C.5. MEAN FACTOR SCORES BY REGION 1 Factor 1, length Factor 2, intensity 0.5 0.8 0.8 0.6 0.4 0.4 0.1 0 0.4 0.3 0.3 0.3 0.2 0.2 1 0.1 0 –0.3 3 –0.3 0.0 Factor score –0.4 –0.5 –1.2 –1.2 –1.8 –1 –1.5 –2 Upper Upper Northern Brong Ashanti Eastern Volta V Central Western Greater east west ahafo accora Source: Authors’ analysis. The relative impact of these two rainfall factors was then confirm the vulnerability of cereal and groundnut examined for years of below average yield, with the results yields to a range of adverse rainfall patterns and summarized below. Although consistent with the results events in the lower rainfall zones in which they obtained from the regression analysis, they did not add predominate. substantially to an understanding of the impact of rainfall » Cassava and yam yields are vulnerable to shorter on yield. rainfall seasons and lower intensity rainfall, » Cereal and groundnut yields are affected by although neither set of factors had a substantive adverse patterns and events in both the length impact. This may result from the higher and more of the rainy season and the intensity of rain- reliable rainfall patterns in the transition and forest fall. Moreover, yields can be adversely affected zones where these crops predominate. when the rainy season is both too long and too » The impact of both sets of factors was higher short. Both high and low intensity rainfall can in the drier savannah zones, particularly in the also reduce yields, but excess rainfall appears Upper East and Upper West regions, as would be to pose the highest risk. Together, these results expected. 86 Ghana: Agricultural Sector Risk Assessment APPENDIX D CLIMATE CHANGE IMPACT ASSESSMENT OF AGRICULTURE IN GHANA INTRODUCTION Agriculture is vulnerable to climate change in Ghana, although the effects are heteroge- neous based on model assumptions and also across regions, socioeconomic groups, and crops and livestock. Agriculture accounts for 25.6 percent of the GDP and 56 percent of the labor force are involved in agriculture.21 The agricultural sector is composed of crops (primarily rain-fed), livestock, and fisheries. The single most important cash crop is cocoa. Cassava and maize are the primary food crops. The agricultural sector is largely composed of smallholders (more than 85 percent of holdings are 2 hectares or less) (Masters, Baker, and Flood 2010). Cocoa in particular seems to be adversely affected by climate change and will have negative implications on development strate- gies and the overall economy if left unaddressed. In the Mapping the Impacts of Climate Change index under “Agricultural Productivity Loss,” the Center for Global Development ranks Ghana 106th out of 233 countries globally for “direct risks” due to “physical climate impacts” and 68 out of 233 for “overall vulnerability” due to “physical impacts adjusted for coping ability” (Wheeler 2011). The impacts of climate change on agriculture in Ghana vary widely based on what assumptions are made, and which scenarios are played out. There are direct impacts, such as changes in crop yields due to precipitation changes, and indirect impacts, such as rising food prices due to production changes, and conflict over land tenure due to shifting agro-climatic zones. The newest installment of the IPCC did not narrow expected results from climate change, but rather widened the frame of variability. This in combination with various approaches to impact studies makes it difficult to 21 “Ghana,” CIA Fact Book (January 26, 2013). Risk Prioritization 87 generalize regarding the effects of climate change on agri- culture in Ghana. This appendix will discuss the various BRIEF HISTORY OF possible outcomes. CLIMATE CHANGE IMPACT ASSESSMENTS PRINCIPAL FINDINGS Ghana has been involved with climate change assess- » The agricultural sector in Ghana is highly vulner- ments since the early stages, ratifying the United Nations able to climate change, in large part due to depen- Framework Convention on Climate Change (UNFCC) in dence on rain (dryland farms are particularly 1995. Ghana hosted the 6th Working Group III session sensitive). Climate change will create water and for the IPCC assessment in 2001. Ghana has submitted heat stress, resulting in pest and disease outbreaks; two national communications (the most recent in 2011) to ecosystem deterioration, resulting in the loss of the conference of parties to the UNFCC, and has com- productive lands; and increased burdens to supply pleted a National Climate Change Adaptation Strategy. chains (from post-harvest losses in storage and dis- tribution) (De Pinto et al. 2012). METHODOLOGIES » Consequences for the agricultural sector include yield reductions, decreased livestock values, post- AND TEMPERATURE/ harvest losses, reduced food accessibility, and PRECIPITATION PROJECTIONS reduced consumption levels. The Netherlands Climate Assistance Programme (NCAP) » Across all models and projects there are signs of completed several climate change assessments for Ghana. warming, usually within the range of a 1.5°C– In its assessment of the impact on fisheries, it used regres- 3°C temperature increase by 2080 (De Pinto et sions of historical rainfall data and sea surface tempera- al. 2012). ture scenarios (SST) combined with dynamic production – Warming appears to be most rapid and occur- models (both a Cost per Unit Effort [CPUE]-based model ring to a greater degree in the north (McSwee- and r-based models) (Dontwi et al. 2008). ney, New, and Lizcano 2008). » Climate change will result in increased pressure on IFPRI’s study is based on the four downscaled global cli- water, soil, and other inputs. mate models (GCMs) from the IPCC AR4—the CNRM, » The agricultural sector is expected to see shift- ECHAM, CSIRO, and MIROC models. Based on these ing agro-climatic zones, and generally decreasing models, the IFPRI study uses the Decision Support Sys- yields due to climate change. tem for Agrotechnology Transfer (DSSAT) crop modeling – Ghana’s primary cash crop, cocoa, will be nega- software projections for crop yields, comparing yield pro- tively affected by climate change, resulting in a jections for 2050 against real 2000 yields. The CNRM decrease in national revenue. model predicts little change in annual precipitation, and – Reduction in productivity and yield is also a uniform temperature increase of 2°C–2.5°C across the expected in root and tuber crops. country. ECHAM also predicts little change in annual – Rice, maize, and groundnuts will also generally precipitation (with an increase in the southeastern part decrease in yield. of Ghana), but an increase in temperature of 1.5°C–2°C » These decreases may lead to increased poverty and across the country. CSIRO predicts a general reduc- food insecurity (Republic of Ghana 2011). tion in annual rainfall (100–200 mm in the middle belt, » Changes in temperature and precipitation also 50–100 mm in the northern savannah, and an increase at affect the fishing system, directly affecting the or above 50 mm in the southwestern corner), a tempera- productivity, catchability, and growth rates— ture increase of 1.5°C–2°C in the north, and an increase varying from species to species. Saltwater fish of 1°C–1.5°C in the south. Finally, the MIROC model were more affected by sea surface temperature, predicts decreased precipitation in the south and increased whereas freshwater fish were more affected by precipitation in the north, and an increase in temperature precipitation. of 1°C–1.5°C across the country. Based on precipitation 88 Ghana: Agricultural Sector Risk Assessment FIGURE D.1. CHANGES IN MEAN PRECIPITATION BY 2030 (LEFT) AND CHANGES IN MEAN PRECIPITATION BY 2050 (RIGHT) Source: International Center for Tropical Agriculture (CIAT) 2011. projects, CSIRO and MIROC appear to present signifi- The International Center for Tropical Agriculture (CIAT) cant challenges for agriculture (IFPRI 2013). published an analysis of the impact of climate change on cocoa-growing regions for the Bill & Melinda Gates Foun- The study on roots and tubers conducted for the Ghana dation. CIAT combined current climate data with future Environmental Protection Agency also used DSSAT, climate change predictions from 19 GCMs for 2030 and version 4, to evaluate root crops for their vulnerability 2050 (emissions scenario SRES-A2). These data were then and implications for impacts. Its scenarios were built on used in MAXENT, a crop prediction model. The model national climatic data from 1960–1990, and the approach finds that temperatures will increase by 1.2°C in 2030 and assumed mono-cropping (Sagoe 2006). by 2.1°C by 2050 (see figure D.2). Rainfall decreases only slightly, down 12 millimeters by 2050 (Laderach 2011). Figure D.1 shows the change in mean precipitation across The Netherlands Climate Change Studies Assistance Ghana and the Ivory Coast by 2030 and 2050. Programme Phase 2 (NCCSAP2) looked at the impacts of climate change across various crops in helping with Other models used to assess the impact of climate change the preparation of Ghana’s submission of the second on agriculture include CERES-Rice (Crop Environment national communication to the conference of parties to Resource Synthesis Model, version 4.0), using data from the UNFCCC. For the assessment of impact on cocoa, the Anum Valley Irrigation Project, and the Centre for NCCSAP2 used climate change scenarios for the semide- Agricultural Bioscience International (CABI), which ciduous forest and evergreen rain forest zones of Ghana assumes a 2.5°C–3.2°C increase across the country, with based on process-based methods relying on the General a decrease in annual rainfall by 9–27 percent by 2100. Circulation Models and Simple Climate Models. The projected mean annual rainfall in the semideciduous for- est zone would be projected to decline by 2.8 percent in GENERAL FINDINGS 2020, 10.9 percent in 2050, and 18.6 percent in 2080. Climate in Ghana is influenced by the Inter-Tropical Con- Similarly, in the evergreen rain forest zone, mean annual vergence Zone, and its interaction with the West African rainfall will decrease by 3.1, 12.1, and 20.2 percent. Pro- monsoon. Generally, there has been a warming trend in jected increases in temperature for 2020, 2050, and 2080 Ghana, with change occurring more rapidly in the north. were, respectively, 0.8°C, 2.5°C, and 5.4°C for the sem- Annual temperatures have risen by 1°C since 1960, and ideciduous forest, and 0.6°C, 2.0°C, and 3.9°C for the the frequency of “hot” days and nights has increased evergreen rain forest. (McSweeney, New, and Lizcano 2008). Climate change Risk Prioritization 89 FIGURE D.2. CHANGES IN MEAN ANNUAL TEMPERATURE 2030 (LEFT) AND CHANGES IN MEAN ANNUAL TEMPERATURE 2050 (RIGHT) Source: International Center for Tropical Agriculture (CIAT) 2011. will affect the various regions in Ghana differently. The Global Climate models (GCMs) broadly agree that there COCOA will be an increase in mean temperatures, but precipitation Cocoa is the principal cash crop in Ghana. It accounts changes are highly variable throughout the models. for 60–70 percent of agricultural foreign export earn- ings, or 20–25 percent of total foreign export earnings. An IFPRI Policy Note reviewed 15 different models, and CABI claims that more than 800,000 smallholder families found the mean annual temperature increase to be 1°C– (mainly in the western region) depend on cocoa produc- 3°C by 2060, and 1.5°C–5.2°C by 2090. It also noted the tion for their livelihood, and do not use significant tech- various negative impacts of climate change on produc- nology or inputs. tion, supply chains, and vulnerable population (particu- larly in the north) (De Pinto et al. 2012). The later IFPRI Cocoa is highly susceptible to climate change (particu- impact assessment itself found general losses in yield over larly temperature and intense dry seasons), and changes the vast majority of Ghana for rain-fed maize, rice, and in production will have a large impact on Ghana’s over- groundnuts (IFPRI 2013). all economy. The appropriate temperature for cocoa is The Netherlands Climate Change Assistance Programme 18°C–32°C, and the trees are highly sensitive to light in 2008 showed wide variability, and great vulnerabil- variations. CABI concludes that there will be shifts in the ity to climate change in Ghana. In the north, increased geographic distribution of cocoa and related pests, overall aridification and exposure to intense rainfall are expected, crop yields will decrease, and there will be a greater inci- which would result in decreased agricultural productiv- dence of crop loss, in turn affecting farm income, liveli- ity, flooding, and migration. In the south, cocoa produc- hoods, and farm-level decision making. (Four months of tion is projected to decline, and as the primary cash crop dry weather alone will lead to seedling mortality, reduced on which both the entire economy and individual small- bean size, and increased pest attack.) Not least of the holders alike depend, the impact will be great. Sea level cocoa industry’s problems will be the spread of black pod rise of 1 m by 2100 would result in the loss of more than disease (Phytophthora megakarya), which thrives in humid 1,000 km2 of land (displacing 132,000 people), particu- conditions, as a result of changing precipitation (Masters, larly along the east coast.22 Baker, and Flood 2010). 22 African Adaptation Programme, Ghana, http://www.undp-aap.org/countries The NCCSAP2 study found that projected climatic /ghana#Pro Doc. changes would exacerbate soil moisture conditions during 90 Ghana: Agricultural Sector Risk Assessment FIGURE D.3. CURRENT SUITABILITY OF COCOA GROWING AREA (LEFT) AND FUTURE SUITABILITY OF COCOA GROWING AREA (RIGHT) Source: International Center for Tropical Agriculture (CIAT) 2011. the dry season, thereby increasing the vulnerability of the western regions, and Brong-Ahafo in the south (see cocoa trees. Using a process-oriented computer model, figure D.3). By 2050, production will be concentrated in CASE2 (CAcao Simulation Engine 2), this regression two areas, in the mountain ranges of the Kwahu Plateau model used to estimate the production of dry cocoa beans (between the Eastern and Ashanti regions), and between in the Koforidua/Tafo cocoa district was extrapolated for the Central and Ashanti regions. Rising above the 2°C the national production. Results showed a 14 percent and mark increases in temperature puts cocoa in Ghana in 28 percent decrease in yield for 2020 and 2050, respec- severe jeopardy (Laderach 2011). tively, based on a year 2000 baseline. The model also pro- jects that moisture levels in 2080 would “not be adequate A study published in the British Journal of Environment and for profitable cocoa production.” The study asserts that Climate Change found a significant shift in the Wenchi “since cocoa is highly sensitive to drought in terms of Municipality in the forest and savannah transitional agro- growth and yield, it is reasonable to anticipate consistent ecological zone from cocoa to maize cropping systems decrease in output from 2020 to 2080” (Anim-Kwapong (humid to drought tolerant). This shift stemmed from and Frimpong 2004). decreases in the yield of cocoa attributed by the study pri- marily to changing rainfall patterns, but it was also the Problems with cocoa are exacerbated by difficulties in result of other factors such as declining producer price, reestablishment and replanting worsened by climate land tenure, and declining soil fertility. Interestingly, the change, the fact that about 25 percent of current cocoa same paper suggests that prevailing climatic conditions tree stocks are 30+ years old, and more than 60 percent and deforestation in Wenchi will prevent future shifts of the farmers are older than 50 years. (Adjei-Nsiah and Kermah 2012). CIAT adds that increased temperatures will increase ROOT AND TUBER CROPS evapotranspiration of the cocoa trees. With overall cli- In Ghana, important root and tuber food crops include mates becoming less seasonal, increased temperature and cassava, yams, and cocoyams. In a report prepared for less seasonal precipitation, CIAT finds that current cocoa- Ghana’s Environmental Protection Agency, the Crops growing areas in Ghana will decrease quite seriously by Research Institute in Kumasi used projected climate 2050. This shift is primarily attributed to the increase in scenarios and crop models (CROPSIM-cassava and temperature. According to their predictions, by 2030 suit- CROPCRO-tanier) and found negative impacts on able cocoa growing areas will start shifting, primarily in yields of cassava and cocoyams. Cassava is expected to Risk Prioritization 91 FIGURE D.4. YIELD CHANGES 2010–50 varieties, farmer training, and other agronomic practices (Oteng-Darko, Kye-Baffour, and Ofori 2012). IFPRI found that there would be a moderate yield decrease for rain-fed rice, but there was variation between models. The Centre National de Recherches Météorologiques was the most optimistic out of models reviewed, whereas the Microindustry Credit Rural Organization’s model showed a decrease in yield up to 25 percent in many areas (De Pinto et al. 2012). GROUNDNUTS IFPRI found high rates of decrease in yield for rain-fed groundnuts. Their literature review noted several varia- tions on results in the north. The European Centre Ham- burg Model (ECHAM5) projected high loss rate, whereas other models saw some increases. When the Common- wealth Science and Industrial Research Organization Source: IFPRI 2012. ran both the CSIRO and MIROC models, projections of more than a 25 percent loss were found in the central and see a reduction in productivity or yield by 3 percent in southern regions (De Pinto et al. 2012). 2020, 13.5 percent in 2050, and 53 percent by 2080. Reductions in cocoyam are projected to be 11.8 percent in 2020, 29.6 percent in 2050, and 68 percent by 2080 FISHERIES (Sagoe 2006). The impact of climate change on fisheries will be felt along the coast in Ghana, as almost 25 percent of MAIZE the population lives in the coastal zone, with around 10 percent dependent on fishing for their livelihood. The Under the IFPRI projections, rain-fed maize will suffer impacts of climate change will result in generally warmer across the country, but particularly in the south (IFPRI air and sea surface temperatures, along with decreased 2013) (see figure D.4). There may be limited areas in precipitation, affecting the industry. When combined the north that will see an increase in yield. Most of the with overfishing and population growth, a triangle of decreases will be small, under 25 percent. Other mod- production constraints appears (McSweeney, New, and els, such as from the Centre National de Recherches Lizcano 2008). When analyzed, most studies suggest that Météorologiques, reflect this prediction of a possible a rise in SST and changes in precipitation will correlate increase in the Upper East, Upper West, and Northern with an increase or decrease in catch rates dependent on regions (De Pinto et al. 2012). the variety of fish. However, there is wide variability, and multiple factors outside of climate change make it hard RICE to generalize. A study run by the Crops Research Institute in Ghana concluded that an increase or decrease in temperature Saltwater fish studied seem to show change in catch rates of 4°C from the maximum or minimum would decrease due to temperature (anchovies and the Round Sardinella). rice yields by 34 percent in relation to the year 2006 as a According to the NCAP vulnerability assessment for cli- base scenario. The study concluded that planting dates mate change impacts on fisheries, the anchovy shows are an instrumental tool in increasing rice yields under cli- the highest sensitivity to climatic changes. Owing to a mate change, along with more temperature-tolerant rice temperature increase of 1°C, the estimates for catch rate 92 Ghana: Agricultural Sector Risk Assessment increases vary from 2.8 percent under the CPUE-based drive productivity (with harvested land only increasing by model, to around 169 percent under the r-based model. around 10 percent). Precipitation had minimal effects (McSweeney, New, and Lizcano 2008). The model found smaller increases in productivity (yield) for cassava at 30 percent. Projections also included har- For the Round Sardinella, the SST negatively affected vested land growing by 7 percent, and production rising CPUE. An increase in temperature of 1°C resulted in a by one-third, but demand surpassing supply around 2025. 4.2 percent catch rate decrease under the CPUE-based The IMPACT model shows average increases of 54 per- model, and a 102 percent decrease under the r-based cent for sweet potatoes and yams, but differences between model. Precipitation also seemed to have a minimal climate models in the intermediate scenario. effect.23 Conversely, precipitation did seem to matter for the CONCLUSION freshwater or brackish water fish studied under the On a general level, a review of the literature suggests that NCAP assessment (tilapias, cat fish, and Flat Sar- there will be a decline in agricultural production based on dinella). The Flat Sardinella showed less response climate change, which in turn will affect various compo- to SST, and is more tolerant to changes in SST and nents of the national GDP. The results vary across crops salinity, but more sensitive to precipitation. Overall, (and fish varieties), and by region. however, the study concluded that the distribution and catchability of Flat Sardinella are “hardly affected by changes in the ocean climate” (McSweeney, New, and LIMITATIONS Lizcano 2008). There are many variations between climatic models and regions. These assessments could benefit from more in- BEYOND CROP IMPACT depth regional impact assessment and further research on food crops. Finally, there are several studies that assess STUDIES the impact of climate change based on land management International Food Policy Research Institute (IFPRI) took techniques. They have not been included in this discussion its assessment a step further, using the IMPACT global but may offer important insights into possible adaptation model for food and agriculture to estimate the impact of techniques, and risk intervention strategies, for example, future GDP and population scenarios on crop production and staple consumption, which “can be used to derive com- D. S. MacCarthy and P. L. G. Vlek, “Impact of Climate modity prices, agricultural trade patterns, food prices, cal- Change on Sorghum Production under Different Nutri- orie consumption, and child malnutrition” (IFPRI 2013) ent and Crop Residue Managment in Semi-Arid Region The IMPACT model projects maize yields increasing by of Ghana: A Modeling Perspective,” African Crop Science almost 60 percent between 2010 and 2050, but suggests Journal (African Crop Science Society, Uganda) 20, no. S2 that consumer demand and technological improvements (2012): 243–259. 23 Ibid. Risk Prioritization 93 APPENDIX E STAKEHOLDER VULNERABILITY ANALYSIS INTRODUCTION 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 they lack the capacity 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, as it allows for determination of impacts on populations, and who will be most affected. Vulner- ability is discussed here particularly in the context of food security. Ghana saw the number of people living in poverty halved between 1998/99 and 2005/06, but the depth of poverty has increased, and there are significant regional differences. According to the Ghana Living Standard Survey in 2005/06, 18 percent of the population has an income less than the costs of the minimum food basket, mak- ing them extremely vulnerable to food price shocks (Biderlack and Rivers 2009). There is a significant amount of information available in Ghana, including the Ghana Living Standard Survey, the Multiple Indicator Cluster Survey, and the Demographic Health Survey. Ghana also has a Food Security Monitoring System operated jointly by the World Food Program, the Ministry of Food and Agriculture, and the Ministry of Health, providing monthly updates on food security in three northern regions as well as a frequent Comprehensive Food Security and Vulnerability Analysis. There- fore, there are strong information systems related to vulnerability and food security in Ghana. Risk Prioritization 95 TABLE E.1. FOOD INSECURITY AND VULNERABILITY BY REGION Food Insecure Vulnerable to Food Insecurity Regions No. of People % Pop No. of People % Pop Western rural 12,000 1% 93,000 6% Central rural 39,000 3% 56,000 5% Greater accra rural 7,000 1% 14,000 3% Volta rural 44,000 3% 88,000 7% Eastern rural 58,000 4% 116,000 8% Ashanti rural 162,000 7% 218,000 10% Brong ahafo rural 47,000 3% 152,000 11% Northern rural 152,000 10% 275,000 17% Upper east rural 126,000 15% 163,000 20% Upper west rural 175,000 34% 69,000 13% Urban (acca) 69,000 2% 158,000 4% Urban (other) 297,000 4% 572,000 8% Total 1,200,000 5% 2,007,000 9% Source: Ghana CFSVA, 2009. contribute 80 percent of Ghana’s agricultural output. GENERAL TRENDS Table E.2 highlights sources of vulnerability for various IN VULNERABILITY groups involved in Ghana’s agricultural sector. » The national averages for food insecurity and related indicators mask drastic regional differences. UNDERLYING FACTORS OF » Food insecurity is concentrated in the areas prone to extreme weather events and with the highest lev- FOOD SECURITY IN GHANA els of poverty. MACRO-LEVEL FACTORS » Generally, the people most vulnerable to becoming » High food price volatility (particularly as 80 per- food insecure live in the Upper West, Upper East, cent of all households rely on markets as their and Northern regions (see table E.1). Other vulner- main source of food). able populations are spread out in the rural and » The lingering impact of the global financial crisis, urban areas of the other seven regions. particularly in terms of diminished export demand and declines in ODA and remittances. VULNERABLE GROUPS » Natural hazards such as floods and droughts, which could reduce resilience, leading into a downward (BIDERLACK AND RIVERS spiral. 2009) Certain populations have characteristics that make them HOUSEHOLD-LEVEL FACTORS more vulnerable to shocks than others, particularly » Lack of education. agricultural workers. More than half of Ghana’s work- » High dependency on agricultural livelihood activi- force is engaged in agricultural activities, and accord- ties as the primary source of income. ing to the 2000 census, more than 90 percent of farms » Lack of access to output markets. are smallholdings of less than 2 ha in size. These farms » Poverty and malnutrition. 96 Ghana: Agricultural Sector Risk Assessment TABLE E.2. VULNERABLE GROUPS Food crop farmers • Food crop farmers have the lowest annual per capita income, and 72% of them cultivate less than two hectares and are primarily relient on rain for water. Almost half of the heads of households have no educational background, and 22% are headed by women. • Location: 48% of the population in the Northern Savannah zone, primarily in the Upper East region. Cash crop farmers • The most vulnerable cash crop farmers are in the Upper West region where they comprise 17% of the population. (The majority live in the Forest zone.) • Their second income source is food crop production. • Among agriculturalists, they have the highest annual per capita income, but more than half are in the poorest wealth quintile. • 18% of the households are headed by women. Agro-pastoralists • 59% live in the Northern region, and 21% in the Upper East region. • 63% of their income comes from livestock and animal husbandry (primarily cattle and poultry), with 1/5th coming from food crops. • 83% of the heads of households have not received any schooling, 88% of households were poor, and 9% are headed by women. Food processors (millers, • Their second income source is food crop production. brewers, and shea nut • 56% of households are poor, and have the largest share of households with loans or debt (46%). producers) • 41% of the households are headed by women. Unskilled laborers • The majority live in urban areas, and urban poor spend about 67% of their income on food (15% higher than the national average). The unskilled laborers who live in rural areas are concentrated in Ashanti and the Upper East. • Their second income source is food crop production. • The average income per capita is the second lowest among livelihoods. • 33% of households have single heads, and 22% are headed by women. Source: Ghana CFSVA 2009. LIVELIHOOD AS CROP VULNERABILITY AN INDICATOR OF TO DROUGHT IN GHANA VULNERABILITY The vulnerability of millet and sorghum crops to drought Figures E.1 and E.2 illustrate how livelihood activities are is shown below, disaggregated by district. A crop’s vulner- intimately linked to geography and agro-climatic condi- ability depends on yield sensitivity, geographic exposure tions in Ghana. Producers in zones 1, 2, and 3 rely heav- to drought, and adaptive capacity. In northern Ghana, ily on cereal and livestock for their livelihoods, whereas sorghum is more vulnerable to drought than is millet in producers farther south may engage in more diversified 10 of 13 districts. agricultural activities. People living in zones 1, 2, and 3 are also the most likely to experience inadequate food Vulnerability = [(crop yield sensitivity index + exposure consumption, as shown in figure E.3, and are extremely index) − adaptive capacity] (Antwi-Agyei 2011). vulnerable to agricultural shocks and exhibit the highest degree of crop failure. Risk Prioritization 97 FIGURE E.1. CROP YIELD SENSITIVITY INDEXES (LEFT ) AND REGIONAL VULNERABILITY INDEXES (RIGHT ) Source: Centre for Climate Change Economics and Policy 2011. FIGURE E.2. LIVELIHOOD ZONES FIGURE E.3. FOOD CONSUMPTION Source: CFSVA 2009. Source: FAO 2010. 98 Ghana: Agricultural Sector Risk Assessment FIGURE E.4. MEAN VULNERABILITY INDEXES OF UPPER EAST DISTRICTS (TOP LEFT ), UPPER WEST DISTRICTS (BOTTOM LEFT ), AND NORTHERN DISTRICTS (BOTTOM RIGHT ) Sorghum Millet Maize Millet 1.4 1.3 1.4 1.3 Vulnerability 1.2 Vulnerability 1.2 1.1 1.1 1 1.0 0.9 0.9 0.8 0.8 Builsa Kassena Bongo Bolgatanga Bawku Bawku go Bi di us la am u ve a Sa u Ta a e Sa le ba al n Za ale u G isg Sa ag g g al W lo ug n G bi Bo on lu la bo ew Ye m To m East West b bz h am Districts D 1.4 Sorghum Millet Districts 1.3 1.2 Vulnerability 1.1 1 0.9 0.8 Wa Lawra Tumu Jirapa Nadowli Districts Source: Centre for Climate Change Economics and Policy 2011. Risk Prioritization 99 APPENDIX F AGRICULTURAL RISK FINANCING AND INSURANCE FOR GHANA: OPTIONS FOR CONSIDERATION MICRO-LEVEL OPTIONS: Public-Private Partnerships (PPP) in Agricultural Insurance for Farmers 1. Developing a viable PPP in agricultural insurance would require significant public investments in data and would most likely need to be accompanied by substantial reform in credit utilization. Developing a PPP in agricultural insurance is a long-term objective, which requires long-term leadership and engagement as well as high levels of invest- ment from both sectors. In addition, to achieve sustainable and meaningful uptake, it relies on several key pillars. One key pillar is an effective distribution channel to rural farmers, through which insurance can be sold. These distri- bution channels can take many forms, such as input suppliers or social wel- fare payment systems; however, the most commonly used, and which have the highest potential, are rural lending institutions. Despite this, there are several significant challenges for this distribution channel in Ghana. Rural credit levels are low, with agriculture, forestry, and fishing loans making only 5.9 percent of the total lending portfolio in Ghana24 and only 10 percent of rural households having access to credit.25 Rural finance has also experienced challenges in the past in terms of high rates of nonrepayment.26 Agricultural data make up another key pillar. Experience from other countries suggests that yield data are required to provide farmers with reliable protection, which can be enhanced by weather and satellite data. Currently, a limited amount of yield data is col- lected in Ghana by the Statistics, Research and Information department at the Ministry of Food and Agriculture, and significant increase in resources 24 Bank of Ghana Annual Report. 25 GFDRR Country DRM Plan for Ghana. 26 World Bank 2009 analysis of commercial bank lending by sector showed that the poorest performing sector was agriculture, forestry, and fishing with 23.1 percent of loans classified as nonperforming. Risk Prioritization 101 would be required to collect such data on a large- schemes. The insurance product covered only scale basis. Although the Ghana meteorological drought risk for rural lending institutions. Discus- department has an established weather station sions with such institutions indicated that drought network, the majority of the weather stations are risk is only one of the many risks they face (others manual, which need to be upgraded to automatic include flood, pests and diseases, bushfires, inabil- stations for insurance purposes. Moreover, the ity to enforce contracts, changes to agricultural audit procedures for these types of data are not policy, and so on). In addition, practitioners noted in line with international reinsurance standards. there was considerable basis risk with the product, Thus, significant financial and human resources because weather data were used to trigger insur- would be required to develop the agricultural data ance payments. These two significant issues led to an acceptable standard. institutions to believe that the use of the product 2. Given the significant fiscal burden of did not cover the cost. developing a PPP in agricultural insur- 4. To develop products that cover more of the ance, other risk management options may risks faced by rural banks and that mini- be more cost effective at this time. The mize basis risk, investments in yield data investments in data mentioned previously, as well would be required, which would face the as the intuitional and market investments required issues mentioned previously. Again, given the to reform and develop distribution channels to a high levels of investments required, other options level that could achieve critical mass, would be available to government could be considered at this high relative to other investment options in agri- stage. cultural risk management currently available to Ghana. In the future, any investments in agricul- Catastrophe Weather Index–Based Insurance tural insurance would be coupled with other ini- 5. Catastrophe weather index–based insur- tiatives; for example, should the Government of ance (WII) products could be considered Ghana aim to increase rural productivity levels, for large-scale commercial farmers. How- this could be achieved by providing farmers access ever, the development impact of any such to better information, improved extension ser- products may be low. vices, enhanced inputs (improved seeds, for exam- ple), and access to the credit required to purchase MACRO-LEVEL OPTIONS them. This would require development of mul- Sovereign Agriculture Risk Financing and tiple markets in Ghana, one of which is the rural Insurance credit market. Agriculture insurance would be an 6. It is unclear what the government con- excellent partner for such a venture as, if effec- siders to be its contingent liability to the tively developed, it can protect vulnerable farmers agricultural sector, and what it consid- against shocks as well as rural lending institutions ers to be the responsibility of donors or against covariate risks that can lead to bankruptcy, farmers. However, if government con- increasing their resilience. siders its contingent liability to be rela- tively low, then developing a sovereign MESO-LEVEL OPTIONS agriculture risk financing strategy may Meso-level Agriculture Risk Financing have lower impact than other agricul- 3. Weather index–based meso-level agricul- tural risk management investments. The tural insurance products have been piloted government has taken several steps to increase in Ghana; however, the pilots are expe- the protection it offers to vulnerable farmers riencing severe challenges and there are against agricultural shocks. Disaster risk reduc- questions about the sustainability of these tion has its main institutional home within the 102 Ghana: Agricultural Sector Risk Assessment National Disaster Management Organization most vulnerable households in Ghana. Currently, in the Ministry of the Interior. NADMO was there is a 3-year plan to increase the number of established in 1996 under a National Secu- families who receive social cash payments from rity Council, chaired by Ghana’s president. It 70,000 to 1,000,000. works with other organizations and ministries 8. Using insurance principals to automati- to ensure such responses are as effective as pos- cally scale up social welfare payments in sible. That said, given the fact that the govern- the aftermath of an agricultural shock ment uses grant funding for its annual budget,27 based on a pre-defined set of rules could its contingent liability to the agricultural sector be considered in conjunction with LEAP. could be considered minimal. Were the govern- An insurance type of product could be established ment to plan to increase its fiscal expenditure in where if there was an adverse climatic shock the aftermath of shocks, then a sovereign agri- of a given magnitude in a given location a pay- cultural risk financing strategy may become a ment would be triggered. This payment would be more attractive option. directly linked to cash transfer system to families in the effected region, thus immediately transferring Index-Based Social Safety Mechanisms funds to those who are in need. Using insurance 7. The government is looking to increase the principals in the scaling up of social safety nets has number of households that receive social numerous benefits: i) it brings efficiency to scheme, welfare payments through the Livelihood developing a money trail that minimizes leakages; Empowerment Against Poverty Program. ii) it increases transparency, enabling both recipi- The LEAP is a social cash transfer program that ents of the benefit and government to have a bet- provides cash and health insurance to extremely ter understanding of when and how much benefit poor households across Ghana to alleviate short- will be paid, thus enabling better planning; and term poverty and encourage long-term human iii) it disciplines the government to comply with capital development. It is targeted at the 18 percent the rules set under the scheme. Grants account for 9.2 percent of government revenues in 2012—GOG state 27 budget, 2012 Risk Prioritization 103 APPENDIX G INDICATIVE LOSSES Risk Prioritization 105 TABLE G.1. INDICATIVE LOSSES (US$ MILLION) FOR ADVERSE CROP PRODUCTION EVENTS BY CROP, 1991–2011 106 (CONSTANT PRICES = 2004–06) Number Events Aggregate Severe to Year (All Crops) Cocoa Cassava Yam Plantain Maize Sorghum Millet Rice Groundnut Catastrophic US$ Million (2010 Exchange Rates) 1991 1992 1993 1994 −45.24 1/0 1995 1996 1997 −10.01 −4.27 −5.49 3/0 1998 1999 −11.45 1/0 2000 −5.45 1/0 2001 −54.17 −52.76 −13.57 1/2 2002 −150.69 0/1 2003 −2.52 1/0 2004 −1.08 −11.55 1/1 2005 2006 2007 −96.05 −60.71 −31.73 -76.97 −79.70 0/4 2008 2009 2010 −34.68 1/0 2011 −4.44 1/0 Number Events per Time Period Severe 1/21 3/21 0/21 1/21 0/21 0/21 2/21 1/21 1/21 3/21 na Catastrophic 0/21 1/21 0/21 0/21 0/21 1/21 1/21 3/21 1/21 1/21 na Average Indicative Loss (US$ Million) Severe −96.05 −32.95 0 −45.24 0 0 −3.27 −4.27 −2.52 −10.46 na Catastrophic 0 −150.69 0 0 0 −52.76 −60.71 −18.95 −76.97 −79.70 na All Events −96.05 −62.39 0 −45.24 0 −52.76 −22.41 −15.28 −39.75 −27.77 na Source: FAOSTAT; authors’ calculations. Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. Ghana: Agricultural Sector Risk Assessment TABLE G.2. INDICATIVE LOSSES (% GROSS AGRIC. OUTPUT) FOR ADVERSE CROP PRODUCTION EVENTS BY CROP, 1991–2011 (CONSTANT PRICES = 2004–06) Number Events Aggregate Severe to Risk Prioritization Year (All Crops) Cocoa Cassava Yam Plantain Maize Sorghum Millet Rice Groundnut Catastrophic % Gross Agricultural Output (Crops Plus Livestock) 1991 1992 1993 1994 0.70 1/0 1995 1996 1997 0.13 0.06 0.07 3/0 1998 1999 0.14 1/0 2000 0.07 1/0 2001 0.64 0.63 0.16 1/2 2002 1.64 0/1 2003 0.03 1/1 2004 0.01 0.12 1/1 2005 2006 2007 0.97 0.61 0.32 0.78 0.81 0/4 2008 2009 2010 0.28 1/0 2011 0.11 1/0 Number Events per Time Period Severe 1/21 3/21 0/21 1/21 0/21 0/21 2/21 1/21 1/21 3/21 na Catastrophic 0/21 1/21 0/21 0/21 0/21 1/21 1/21 3/21 1/21 1/21 na Source: FAOSTAT; authors’ calculations. Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. 107 TABLE G.3. INDICATIVE LOSSES (US$ MILLION) FOR ADVERSE PRODUCER PRICE MOVEMENTS BY CROP, 1991–2010 108 (REAL PRICES 2010 = 100) Number Events Severe to Year Cocoa Cassava Yam Plantain Maize Sorghum Millet Rice Groundnut Catastrophic US$ Million (2010 Exchange Rates) 1991 1992 1993 1994 −6.41 −4.13 2/0 1995 1996 −7.86 1/0 1997 1998 1999 −65.10 −69.15 −12.59 −5.01 3/1 2000 2001 2002 2003 −29.06 −37.57 −9.33 −5.01 4/0 2004 2005 −4.71 1/0 2006 −64.75 1/0 2007 −184.34 −17.17 −9.92 1/2 2008 −49.57 1/0 2009 −16.00 1/0 2010 -29.79 −43.01 1/1 Number Events per Time Period Severe 0/20 2/20 1/20 2/20 4/20 2/20 3/20 1/20 1/20 na Catastrophic 0/20 0/20 0/20 0/20 0/20 2/20 1/20 1/20 0/20 na Average Indicative Loss (US$ Million) Severe 0 −106.70 −65.10 −27.14 −50.32 −7.87 −4.72 −16.00 −7.86 na Catastrophic 0 0 0 0 0 −14.88 −9.92 −43.01 0 na All Events 0 −106.70 −65.10 −27.14 −50.32 −11.38 −6.02 −29.51 −7.86 na Source: FAOSTAT; authors’ calculations. Ghana: Agricultural Sector Risk Assessment Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. TABLE G.4. INDICATIVE LOSSES (% GROSS AGRIC. OUTPUT) FOR ADVERSE PRODUCER PRICE MOVEMENTS BY CROP (REAL PRICES 2010 = 100) Number Events Severe to Risk Prioritization Year Cocoa Cassava Yam Plantain Maize Sorghum Millet Rice Groundnut Catastrophic % Gross Agricultural Output (Crops Plus Livestock) 1991 1992 1993 1994 0.10 0.07 2/0 1995 1996 0.15 1/0 1997 1998 1999 1.07 0.79 0.21 0.08 3/1 2000 2001 2002 2003 0.40 0.36 0.13 0.07 4/0 2004 2005 0.05 1/0 2006 0.46 1/0 2007 1.63 0.15 0.09 1/2 2008 0.38 1/0 2009 0.12 1/0 2010 0.14 0.30 1/1 Number Events per Time Period Severe 0/20 2/20 1/20 2/20 4/20 2/20 3/20 1/20 1/20 na Catastrophic 0/20 0/20 0/20 0/20 0/20 2/20 1/20 1/20 0/20 na Source: FAOSTAT; authors’ calculations. Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. 109 TABLE G.5. INDICATIVE LOSSES (US$ MILLION) FOR ADVERSE CROP PRODUCTION EVENTS BY REGION (CONSTANT 110 PRICES = 2004–06) Number Events Aggregate Upper Upper Severe to Year (All Crops) West East Northern Volta Ashanti Brong-Ahafo Eastern Western Central Catastrophic US$ Million (2010 Exchange Rates) 1992 1993 −23.95 0/1 1994 1995 1996 −22.26 1/0 1997 −8.67 −56.60 2/0 1998 −4.83 1/0 1999 −47.08 0/1 2000 −49.85 1/0 2001 2002 −18.27 1/0 2003 −21.73 1/0 2004 −3.71 −11.89 2/0 2005 −43.89 0/1 2006 2007 −96.05 −51.52 −126.15 −118.81 0/3 2008 2009 Number Events per Time Period Severe 1/18 0/18 2/18 2/18 1/18 1/18 1/18 1/18 1/18 0/18 na Catastrophic 0/18 1/18 2/18 1/18 2/18 0/18 0/18 0/18 0/18 0/18 na Average Indicative Loss (US$ Million) Severe −96.05 0 −6.19 −39.43 −11.89 −21.73 −4.83 −49.85 −18.27 0 na Catastrophic 0 −51.52 −86.62 −118.81 −33.92 0 0 0 0 0 na All Events −96.05 −51.52 −46.40 −65.89 −26.58 −21.73 −4.83 −49.85 −18.27 0 na Source: FAOSTAT; authors’ calculations. Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. Ghana: Agricultural Sector Risk Assessment TABLE G.6. INDICATIVE LOSSES (% GROSS AGRIC. OUTPUT) FOR ADVERSE CROP PRODUCTION EVENTS BY REGION (CONSTANT PRICES = 2004–06) Number Events Aggregate Upper Upper Severe to Risk Prioritization Year (All Crops) West East Northern Volta Ashanti Brong-Ahafo Eastern Western Central Catastrophic % Gross Agricultural Output (Crops Plus Livestock) 1992 1993 0.36 0/1 1994 1995 1996 0.30 1/0 1997 0.12 0.76 2/0 1998 0.06 1/0 1999 0.57 0/1 2000 0.60 1/0 2001 2002 0.20 1/0 2003 0.23 1/0 2004 0.04 0.12 2/0 2005 0.44 0/1 2006 2007 0.97 0.52 1.28 1.20 0/3 2008 2009 Number Events per Time Period Severe 1/18 0/18 2/18 2/18 1/18 1/18 1/18 1/18 1/18 0/18 na Catastrophic 0/18 1/18 2/18 1/18 2/18 0/18 0/18 0/18 0/18 0/18 na Source: FAOSTAT; authors’ calculations. Note: Severe: Production more than 0.33 standard deviation below trend; Catastrophic: Production more than 0.66 standard deviation below trend. 111 APPENDIX H ECONOMIC INDICATORS FIGURE H.1. AGRICULTURE, VALUE FIGURE H.2. GROWTH IN GROSS ADDED (2007–12) DOMESTIC PRODUCT (2006 Agriculture, value added (% of GDP) CONSTANT PRICES), 2007–12 Agriculture, value added (annual % growth) 40 Agriculture Industry Services 16,000 35 14,000 30 12,000 25 Millions of GHC 20 10,000 15 8,000 10 6,000 5 4,000 0 2,000 00 2007 2008 2009 2010 2011 2012 –5 0 Source: World Development Indicators 2013. 2007 2008 2009 2010 2011 2012 Source: Bank of Ghana 2013. FIGURE H.3. AGRICULTURE, VALUE ADDED (ANNUAL % GROWTH), 1980–2012 12 10 9.7 8 7.5 7.4 7.2 5.8 6.1 6 5.2 5.1 5.3 4.4 4.5 4.3 4.3 3.9 4.0 0 4.1 4 3.6 3.7 3.3 2.2 2.31.9 2.1 2 1.3 0.7 0.8 0.0 0 0 1980 1985 1990 99 1995 2000 2005 2010 –2 –1.2 –2.0 –1.7 –2.6 –4 –6 –5.5 –7.0 –8 Source: Bank of Ghana 2013. Risk Prioritization 113 FIGURE H.4. AGRICULTURE, VALUE ADDED (ANNUAL % GROWTH), 2000–12 10 8 6 4 % 2 0 2000 2002 2004 2006 2008 2010 2012 –2 –4 Source: Bank of Ghana 2013. 114 Ghana: Agricultural Sector Risk Assessment APPENDIX I TIMELINE OF EVENTS Year Region Events Affecting Agricultural Production 1992 National Widespread localized flash flooding. First major outbreak of sigatoka disease (plantain). 1992 National Private traders allowed to buy cocoa for first time in competition with Cocoa Board. 1992 National National election. Considerable internal instability. Widespread strikes by public sector and civil service. Civil service pay increased by 80%. 1993 Brong-Ahafo Fall in cassava production. Caused by a decline in area planted in response to low returns in previous year, and good opportunities to plant other crops. Storm damage to plantain. 1993 Greater Accra Fall in cassava production. 1994 General Fifty percent devaluation of franc CFA. 1994 Ashanti Fall in cassava, maize, and plantain production. 1994 Brong-Ahafo Fall in cassava production due to variegated grasshoppers and localized drought. High black sigatoka. 1994 Northern Fall in cassava, maize, yam production. Inter-ethnic conflicts. 1994 Upper East Serious drought from September–October reduced millet, sorghum production. 1995 Northern Inter-ethnic conflicts. 1996 Northern Inter-ethnic conflicts, high fertilizer costs reduced production of maize, rice, sorghum. 1996 Upper East Drought in June–July delayed planting of groundnuts and grain filling of millet. 1997 General Bushfires in northern and transition zones. 1997 Brong-Ahafo Drought and erratic rainfall reduced cassava production. 1997 Greater Accra Drought and erratic rainfall reduced cassava production. 1997 Northern Drought, erratic rainfall, and high fertilizer costs reduced millet and rice production. 1997 Upper East Drought reduced production of millet and sorghum. 1998 Upper West Drought (El Niño year) reduced production of millet. 1998 Central Drought reduced crop production—all crops. 1999 General Sharp exchange rate depreciation in response to Russian ruble crisis. 1999 Upper East Flooding during production period affected production of groundnuts, rice. 2000 Northern Drought (La Niña) reduced production of maize, millet, sorghum. 2000 Eastern Big drop in area cultivated, especially for cassava. But appeared to be an adjustment to higher planting in previous year. Not an adverse event. 2001 National Localized droughts in northern areas. 2002 Greater Accra Reduced production of rice relative to previous year (but not an adverse event). (continued ) Risk Prioritization 115 Year Region Events Affecting Agricultural Production 2002 Upper East Bushfires, late rains, low prices reduced production of rice. 2002 Upper West Flooding caused reduction in area planted to rice. 2002 Northern Inter-ethnic conflicts. 2003 General Bushfires in northern and transition zones. 2003 Ashanti Erratic rainfall during major and minor seasons. Sigatoka disease hurt plantain; stem borer infestation hurt maize. 2003 Brong-Ahafo Four-week dry spell during major season hurt crops. NPK* fertilizer prices up 20%. 2003 Central Newcastle disease affected poultry production. Generally favorable weather. 2003 Eastern Irregular rainfall hampered fertilizer and herbicide and pesticide application. Cassava price down 33%. Input prices rose significantly. 2003 General Cassava and gari priced down more than 10%. Prices of most other crops up or little change. 2003 Greater Accra Four-week dry spell during major season hurt crops. Unfavorable rainfall. Gari prices down 35%. NPK fertilizer, ammonia, and urea prices increased considerably. 2003 Northern First quarter hot, dry with Harmattan winds. Vigorous rains second–fourth quarters. Localized flooding hurt rice and maize. NPK fertilizer and ammonia prices up more than 25%. 2003 Upper East Disease destroyed 454 hectares of irrigated tomato fields along the Pwalugu River basin and Tono and Vea irrigation project sites. 2003 Upper West Rainfall distribution more favorable than previous year. 2003 Volta Dry spell of 4 weeks during major season adversely affected crop production. 2003 Western Crop price generally up, except plantain, which exhibited no change. Crop production unchanged from previous year. 2004 Ashanti Favorable weather for livestock and crops. Maize and legumes did poorly owing to continuous rains during growing period. Armyworm and capsid affected some crops. 2004 Brong-Ahafo Good distribution of rainfall, boosting crop production and pasture availability. High prices for inputs and agricultural services. 2004 Central Erratic rainfall during the first half of year was not good for crop production. 2004 Eastern Weather favored timely land preparation and production of maize, vegetables, plantain, cassava, and cocoyam. Private agencies continued to make inputs available. 2004 Greater Accra Total rainfall was poor, but well distributed. 2004 Northern Localized droughts and poor distribution of rainfall. A few districts suffer minor floods, others short droughts. Preparation costs up significantly. Various minor outbreaks of livestock disease. 2004 Upper East Localized droughts and poor rainfall affected all crops except maize. African swine fever caused 585 deaths in Boltanga district. 2004 Upper West Localized droughts. Harmattan winds during first quarter. Early rainfall. 2004 Volta Rainfall of higher intensity but lower frequency. Input prices up appreciably. 2004 Western Maize, rice, yam, cocoyam, and cassava prices up 30%–60%. Emergence of cocoa purple bean disease caused concern. Newcastle disease major cause of mortality in fowl. 2005 General Bushfires in northern and transition zones. Loss of access to European markets for groundnuts (alfatoxin). 2005 Northern Below average rainfall. 2005 Northern The 2005 MoFA annual report provided less detail regarding production losses than in previous years. 2005 Volta Below average rainfall. 2006 Ashanti Erratic major season rains delayed planting 3–5 weeks. Drought in November–December reduced minor season maize yields 50%. Armyworm outbreak. 2006 Brong-Ahafo Late major season rains, early end to minor season effected maize. Erratic main season rains delayed planting 3–5 weeks. Armyworm outbreak. 116 Ghana: Agricultural Sector Risk Assessment Year Region Events Affecting Agricultural Production 2006 Central Erratic main season rains delayed planting 3–5 weeks. Poor fourth quarter rains reduced minor season maize production. Armyworm outbreak. 2006 Eastern Low July–August rainfall, but quick recovery. Erratic main season rains delayed planting 3–5 weeks. Early end to rain adversely affected minor season maize. 2006 Greater Accra Late main season rains delayed planting 3–5 weeks. High rainfall caused localized flooding. Dry late third–fourth quarters reduced minor season maize yields 50%. 2006 Northern Rainfalls generally favorable. 2006 Upper East Poor rainfall distribution in July and August hurt crops. 2006 Upper West Intermittent drought in July and August favored cowpeas and early planting of maize. 2006 Volta Erratic main season rains delayed planting 3–5 weeks. Reduced minor season rains affected maize. Armyworm outbreak. 2006 Western Poor rainfall in fourth quarter caused maize and other crop failures. 2007 General Localized droughts in northern areas of country. Collapse of world cotton prices, Redenomination of currency. Flash flooding in many areas. 2007 General Major floods in Northern, Upper East, and Upper West regions in September 2007. Many killed. Disaster area declared. 2007 Northern Early drought followed by severe floods, which washed out replanted crops. 2008 General Sharp exchange rate depreciation and terms of trade shock due to global food price crisis. 2008 Ashanti Late first quarter rainfall delayed planting, causing most farmers to plant in May. February–May rainstorms destroyed plantain fields. 2008 Brong-Ahafo Late rains delayed major season planting. Localized flooding. July–November drought stressed maize, wilted vegetables. African swine fever killed one out of nine pigs. 2008 Central Weather in third and fourth quarters favored planting of oil palm, citrus, cocoa, coconuts, maize, plantain, and cassava. 2008 Eastern Late first quarter rains delayed planting. Rainfall declined in November–December, reducing yields. Landslide in October destroyed 105 households. 2008 General Bushfires in northern and transition zones 2008 Greater Accra Late first-quarter rains delayed planting. Poor distribution of rainfall during first quarter. Farmers plowed during second quarter. 2008 Northern Rainfall distribution normal, enhancing availability of pasture for livestock. Some flooding in Tolon-Kumbungu and West Mamprusi districts. 2008 Upper East Rainfall conditions normal. 2008 Upper West Good rainfall, below normal humidity during second quarter. Flood in June–September destroyed crops and bridges. 2008 Volta Rainfall distribution normal, but some flooding in valley bottoms. 2008 Western Heavy June–September rainstorms destroyed most plantain farms, raising prices; caused flooding that destroyed dams. 2009 Ashanti Fewer bushfires than previous year. Late planting caused wilting and stunting. Rainstorms caused lodging of plantain, reduced yields. 2009 Brong-Ahafo Sudden decline in rainfall in last quarter affected maize and rice during tasselling and milking stages. Patches of bushfires in November. 2009 Central Rainfall distribution poor; concentrated at end of May and early June. 2009 Eastern Drought when maize in tasselling stage. Heavy rainfall in late June resulted in good harvest in the Manya, YiloKrobo, and Asuogyaman districts. 2009 General Fruitfly menace throughout the country. 2009 General Outbreak of suspected new species of mealybugs. (continued ) Risk Prioritization 117 Year Region Events Affecting Agricultural Production 2009 General Bushfires in northern and transition zones. 2009 Greater Accra Low rainfall, resulting in dryness and localized bushfires. 2009 Northern Occasional thunderstorms. 2009 Upper East Climatic conditions normal. No floods or drought. 2009 Upper West Early part of the year dry. Localized flooding during rainy season, aggravated by opening of dam in Burkina Faso. 2009 Volta Dryness as a result of high temperatures, strong winds. These events favored productivity. 2009 Western Low humidity reduced post-harvest crop losses and livestock disease. Source: MoFA Annual Reports (1993–2009); interviews with MoFA officials, farmers, and traders; background reports (see References). * NPK = nitrogen, phosphorus, potassium. 118 Ghana: Agricultural Sector Risk Assessment APPENDIX J ASSESSING VULNERABILITY IN NORTHERN REGIONS The three northern regions account for just over 40 percent of the country’s land area but less than 17 percent of its population (see appendix E). Located in the savannah agro-ecological zones, in a typical year they produce all of the nation’s millet and sorghum, 90 percent of its groundnuts, 68 percent of its rice, 30 percent of its yams, and 17 percent of its maize. They also account for approximately 75–80 percent of the national cattle herd. According to the latest 10-year average data, the Northern Region receives about 1,200 mm of yearly precipitation. The Upper East and Upper West regions receive less, approximately 940 mm. As shown below, however, rainfall distribution over the uni-modal growing season is at least as important as cumulative levels. Northern regions have consistently recorded higher incidences of poverty, food inse- curity, and malnutrition. They are more rural, with household sizes larger than the national average. Compared with a national income per capita figure of GH¢397 in 2008,28 figures were GH¢106 in the Upper West, GH¢124 in the Upper East, and GH¢296 in the Northern regions. Among the 8,400 households surveyed in 38 districts within the three northern regions for the 2012 Comprehensive Food Security and Vulnerability Assessment, almost half (46 percent) derived their income from crop cultivation, whereas nearly one-third (29 percent), as agro-pastoralists, relied on a combination of income from livestock (49 percent) and crops. Surveyed households described their main cropping activities as summarized in table J.1. Most households manage diversified farms that extend over 11 acres of land that belong, for the most part, to household members. The vast majority of remaining households use land that is provided by extended family members. Yields for main 28 2008 Ghana living standards survey. Risk Prioritization 119 TABLE J.1. HOUSEHOLD CROPPING food crops are modest (maize averaging 1.6 tons per ha; ACTIVITY millet and sorghum ranging from 1.2 to 1.7 tons per ha) and are highly sensitive to weather conditions. For most Northern Maize Yam Groundnuts Rice agriculturalists and agro-pastoralists, the main sources of (75%) (38%) (28%) (25%) food consist of their own production (33 percent) and cash Upper East Millet Maize Sorghum Groundnuts Region (57%) (55%) (44%) (37%) purchases (60 percent). Upper West Maize Groundnuts Sorghum Rice and Region (84%) (53%) (27%) millet (13%) Finally, a ranking of most common agricultural problems reported by CFSVA respondent households showed that Source: Ghana CFSVA 2012. inadequate rainfall (64.5 percent) was considered the pri- mary challenge. More than 40 percent also complained TABLE J.2. DISTRIBUTION OF HOUSEHOLD of low soil fertility, whereas over half mentioned lack of FARM SIZE, BY REGION (ACRES) funds to buy agricultural inputs (for example, fertilizer, Large Medium Small (5 or pesticides) and other basic goods. Nearly 10 percent (11+) (6–10) Fewer) reported a lack of household labor for farming. Northern 50% 28% 21% Upper East Region 60% 24% 16% ENABLING ENVIRONMENT Upper West Region 84% 12% 4% RISKS FOR NORTHERN Source: Ghana CFSVA 2012. REGIONS TABLE J.3. TYPE OF ACCESS TO LAND, A number of respondents, corroborated by official reports, indicated that the following risks affect the enabling envi- BY REGION ronment in northern regions: Other » The general aging of the farming population From (Permission Extended from Chief implies that labor available for land clearing or Ownership Family and the Like) preparation has decreased while become more Northern 53% 27% 20% expensive. This means that some people increas- Upper East 77% 18% 5% ingly rely on herbicides to clear land, implying sig- Region nificant potential risks to groundwater resources Upper West 48% 29% 23% and public health. Region » Some members of the local research community Source: Ghana CFSVA 2012. are concerned that the shift to short-cycle, early TABLE J.4. AVERAGE YIELD FOR MAJOR CROPS IN THE UPPER WEST REGION, 2010 (Figures in MT/Ha) Crops Districts Maize Rice Millet Sorghum Yam G/Nuts Cowpea Soyabean Wa west 1.40 2.20 0.50 1.00 11.52 1.60 0.90 1.60 Wa east 1.50 2.08 0.90 1.00 21.00 1.25 0.83 1.30 Wa municipal 1.30 1.40 1.20 1.20 23.89 1.40 1.20 1.40 Lawra 1.00 1.60 1.40 1.10 0.00 1.60 1.10 0.96 Sissala east 2.00 2.08 1.90 1.60 13.00 1.50 1.00 1.93 Sissala west 1.70 2.10 1.00 0.95 15.81 1.90 1.00 1.90 Jirapa-Lambussie 1.40 1.35 0.60 0.70 13.90 1.50 1.05 0.81 Nadowli 1.50 1.66 0.95 1.10 23.00 1.20 1.10 1.00 Average yield 1.70 1.60 0.98 1.06 20.30 1.54 1.17 1.42 Source: Statistics, Research and Info. Directorate (SRID), MoFA, January 2011. 120 Ghana: Agricultural Sector Risk Assessment TABLE J.5. AVERAGE YIELD FOR MAJOR CROPS IN THE UPPER WEST REGION, 2011 (Figures in MT/Ha) Crops Districts Maize Rice Millet Sorghum Yam G/Nuts Cowpea Soyabean Wa west 1.20 1.60 0.40 0.70 11.00 1.40 1.00 1.30 Wa east 1.20 1.50 0.75 0.80 20.00 1.10 0.90 1.00 Wa municipal 1.25 1.30 1.04 0.90 23.80 1.20 1.30 1.20 Lawra 0.70 1.30 1.20 0.88 0.00 1.40 1.30 0.80 Sissala east 1.40 1.80 1.70 1.20 12.50 1.30 1.10 1.50 Sissala west 1.30 1.00 0.90 0.80 15.00 1.60 1.10 1.60 Jirapa-Lambussie 1.00 1.00 0.40 0.54 14.00 1.30 1.00 0.60 Nadowli 1.40 1.20 0.90 0.85 23.00 0.90 1.20 0.80 Regional average yield 1.23 1.35 0.85 0.80 17.96 1.22 1.12 1.13 Source: Statistics, Research and Info. Directorate (SRID), MoFA, January 2012. TABLE J.6. WEATHER IMPACTS ON KEY CROPS, 2011–12 Maize Rice Millet Sorghum Yam Groundnuts Northern −19.2 −9.1 −18.6 −10.5 −27.2 −4.9 UER −20.9 −19.1 −21.2 −19.7 −23.4 −30.2 UWR −13.9 −10.5 −15.4 −34.8 −9.4 −17.5 Source: CFSVA 2012. maturing crops and to certain cash crops (a way The season was characterized in the Northern Region as to trade price risk for drought risk) may reduce the the worst in the past 15 years, affecting both crop and regional crop genetic stock. livestock output. Most farmers reduced acreage and had » Most important, growing uncertainty about the lower yields, with a few giving up rainy season produc- timing and amount of operational funds for MoFA tion entirely. As a result, the prices of basic foodstuffs rose activities makes it very difficult to properly plan sharply; local rice and maize prices, for instance, started extension activities and to carry them out at the rising in June, ending up in December at 70 percent over optimal time. This affects vaccine production and their January–April levels.29 Table J.6 summarizes respon- delivery, epidemiological surveillance and action, dents’ assessments of the impact of this weather on key and the delivery of agricultural inputs (especially crops, in terms of the percentage change between 2011 fertilizer) to block farms and other producers. and the previous year. The data strongly support the con- clusion that the 2011 disruption of rainfall patterns had a significant impact on production and incomes. DROUGHT RISK AND IMPACT ON NORTHERN As mentioned previously, 46 percent of households sur- HOUSEHOLDS IN 2011–12 veyed in the 2012 CFSVA are agriculturalists, and 29 percent are agro-pastoralists. Most of them have a net The 2011 growing season is the most recent example of deficit in food production, but produce cash crops and unfavorable weather conditions. There were irregular rains and long dry spells from May through July, leading to poor germination, poor crop development, and low yields. 29 2011 Annual Progress Report, MoFA/Northern Region. Risk Prioritization 121 have enough other income to rely on the market for some most severe for the Upper West region households (nearly 60 percent of their basic food supply. In addition, produc- twice as much as the average for the other two regions).31 tion systems in the Northern Region are globally more In terms of crop failure, the Upper East appeared much diversified. The Upper East, in contrast, which is more less affected than the Upper West, and even less than the densely populated and suffers from poor soils and smaller Northern Region, partly because the proportion of mil- average farm sizes, relies more on a combination of pas- let and sorghum in production and food intake is higher toralism and more effective integration of livestock and there than in other northern areas. Globally, whereas agriculture.30 The Upper West appears more vulnerable 70 percent of households in the Upper East reported that to combined shocks, as discussed below. they had managed the crisis well, the percentages were 64 percent in the Northern and 48 percent in the Upper The 2012 CFSVA provides the best impact assessment West regions. of the negative 2011 season at the household level and its results are consistent with this simple characteriza- The fact that Upper East respondents appeared to have tion. Globally, one-third of households faced difficulties weathered the crisis relatively well does not mean that (reduction in production related to drought, basic food they are better off than people in other regions in all prices, or both) that were severe enough to reduce their respects. They have, after all, the highest proportion of food access for some time during the marketing year. The food-insecure households. The shock may have set them combined impact of crop failure and high food prices was back relatively less, but from a low base at the outset. 31 One will recall that UWR has the lowest per capita income of all three north- Upper East has also received significant assistance from the Northern Growth 30 ern regions. In addition, “lack of rainfall” was more often quoted as a problem Project. (37 percent) in UWR than in UER (29 percent) and NR (24 percent). 122 Ghana: Agricultural Sector Risk Assessment APPENDIX K IRRIGATION DEVELOPMENT IN GHANA Ghana is drained by three main river systems: the Volta, Southwestern, and Coastal river systems (see figure K.1): » The Volta river system consists of the Oti and Daka rivers, the White and Black Volta rivers, and the Pru, Sene, and Afram rivers—the basin covers 70 percent of the country’s area. » The southwestern river system comprises the Bia, Tano, Ankobra, and Pra riv- ers and covers 22 percent of the country’s area. » The coastal river system comprises the Ochi-Nakwa, Ochi Amissah, Ayensu, Densu, and Tordzie rivers, covering 8 percent of the country’s area.32 See R. E. Namara, L. Horowitz, and B. Nyamadi, “Irrigation Development in Ghana: Past Experiences, Emerging 32 Opportunities, and Future Directions,” Ghana Strategy Support Program (GSSP), Working Paper No. 0026 (Accra, Ghana: International Food Policy Research Institute [IFPRI], 2011. Risk Prioritization 123 FIGURE K.1. RIVER BASINS IN GHANA Source: IFPRI 2011. FIGURE K.2. DISTRIBUTION OF IRRIGATION SYSTEM TYPOLOGIES IN THE REGIONS OF GHANA Source: IFPRI 2011. 124 Ghana: Agricultural Sector Risk Assessment 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 94228-GH 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org/agriculture