AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER KENYA AGRICULTURAL SECTOR RISK ASSESSMENT Stephen P. D’Alessandro, Jorge Caballero, John Lichte, and Simon Simpkin WORLD BANK GROUP REPORT NUMBER 97887 NOVEMBER 2015 AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER KENYA Agricultural Sector Risk Assessment Stephen P. D’Alessandro, Jorge Caballero, John Lichte, and Simon Simpkin Kenya: Agricultural Sector Risk Assessment © 2015 World Bank Group 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved This volume is a product of the staff of the World Bank Group. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank Group or the governments they represent. The World Bank Group does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. World Bank Group encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clear- ance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone: 978-750-8400, fax: 978-750-4470, http://www.copyright .com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank Group, 1818 H Street NW, Washington, DC 20433, USA, fax: 202-522-2422, e-mail: pubrights@worldbank.org. Cover photos left to right: 1. Gathering corn—Curt Carnemark/World Bank. 2. Man with livestock—Curt Carnemark/World Bank. 3. Irrigating fields near Mount Kenya—Neil Palmer (CIAT). 4. A farmer in the Kibirichia area of Mount Kenya—Neil Palmer (CIAT). CONTENTS Acronyms and Abbreviations ix Acknowledgments xi Executive Summary xiii Chapter One: Introduction 1 Chapter Two: Agriculture Sector in Kenya 7 Agroclimatic Conditions 8 Rainfall Patterns and Trends 8 Crop Production Systems 9 Livestock Production Systems 15 Variability in Crop Production 15 Food Supply and Demand 16 Agricultural Markets and Price Trends 18 Livestock Production 22 Food Security 23 Constraints to Agricultural Growth 23 Chapter Three: Agriculture Sector Risks 25 Production Risks 25 Pests and Diseases 31 Market Risks 34 Enabling Environment Risks 40 Multiplicity of Risks 44 Chapter Four: Adverse Impacts of Agricultural Risks 45 Conceptual and Methodological Basis for Analysis 45 Crop Production Risks 46 Livestock Risks 48 Chapter Five: Stakeholder Vulnerability Assessment 49 General Trends in Vulnerability 49 Livelihoods and Agroclimatic Conditions 49 Poverty and Vulnerability 50 Vulnerability Among Livelihood Groups 52 Vulnerability and Risk Management 52 Risk Management Capacity 52 Vulnerability in ASALs 53 Chapter Six: Risk Prioritization and Management 55 Risk Prioritization 55 Risk Management Measures 56 Illustrative Risk Management Measures 59 Prioritization of Risk Management Measures 64 Conclusion 66 Kenya: Agricultural Sector Risk Assessment iii References 67 Appendix A: Climate Change Impacts on Agriculture in Kenya 73 Introduction 73 Principal Findings 74 Climate Change and Severe Weather Events 74 Methodologies 74 Crop Predictions 75 Crops Resistant to Climate Change 77 Conclusions 77 Appendix B: Stakeholder Risk Profiles 79 Case Study 1: Philip Mutua Mbai—Smallholder Maize Farmer, Machakos County 79 Case Study 2: Mrs. Maraba—Agro-input Dealer, Eldoret Uasin Gishu County 81 Case Study 3: Leshamon Olekoonyo—Wheat Farmer, Narok 82 Case Study 4: Marcel Wambua—Head of Finance, Lesiolo Grain Handlers Limited 84 Case Study 5: Michael Waigwa—Agricultural Underwriter, Cooperative Insurance Company 85 Case Study 6: Wilson Murunya—Livestock Herder, Kajiado County 87 Case Study 7: Yusuf Khalif Abdi—Livestock Herder, Garissa County 88 Case Study 8: Fresha Dairy—Milk Processors, Githunguri County 90 Appendix C: Stakeholder Vulnerability Analysis 91 General Trends in Vulnerability 91 Vulnerability, Livelihoods, and Agroclimatic Conditions 91 Poverty Status and Vulnerability 92 Vulnerable Groups 93 Pastoralists 94 Female-Headed Households (FHHs) 94 Unskilled/Casual Wage Laborers 94 Appendix D: Rainfall Analysis 95 Appendix E: Weather and Yield Impact Analysis 99 Background 99 Summary and Key Findings 100 Weather Information 100 Annual Rainfall Distribution in Kenya 101 Drought and Excess Rainfall Analysis 102 Rainfall—Yield Regressions 102 Appendix F: Crop Production Trends 107 Appendix G: Livestock Terms of Trade Analysis 109 Appendix H: Options for Scaling Up Livestock Insurance in Kenya 111 Appendix I: Results of Solutions Filtering Process 113 Food Crops 113 Cash Crops 114 Livestock 114 iv Agriculture Global Practice Technical Assistance Paper BOXES Box 3.1: Kenya’s Dairy Sector—A Case Study of Market and Enabling Environment Risk 42 FIGURES Figure ES.1: Historical Timeline of Major Agricultural Production Shocks in Kenya, 1980–2012 xiv Figure ES.2: Estimated Losses to Aggregate Crop Production from Risk Events, 1980–2012 (US$, millions) xv Figure 1.1: Agricultural GDP versus National GDP Growth (% change), 1968–2012 2 Figure 1.2: Agricultural Value Added (annual % growth), 1980–2013 3 Figure 1.3: Agriculture Sector Risk Management Process Flow 5 Figure 2.1: Average Cumulative Rainfall (mm) by Rainfall Zone, 1981–2011 9 Figure 2.2: Composition of Crop Production (area harvested, in thousand ha), 1990–2012 10 Figure 2.3: Food Crop Production (thousand MT), 1990–2012 11 Figure 2.4: Industrial Crop Production (thousand tons), 1990–2012 13 Figure 2.5: Coffee Production (tons), 1980–2012 14 Figure 2.6: Cereal Production Trends (thousand tons), 1990–2012 16 Figure 2.7: Maize Production versus Demand (thousand MT), 2003/04–2013/14 17 Figure 2.8: Trends in Cereal Prices (K Sh/ton), 1991–2011 19 Figure 2.9: Trends in Cash Crop Prices (K Sh/ton), 1991–2011 20 Figure 2.10: Coffee Price Comparison ($/kg), 2005–13 21 Figure 2.11: Trends in Producer Prices (K Sh/ton) for Fruits/Vegetables, 1991–2011 22 Figure 3.1: Historical Timeline of Major Agricultural Production Shocks, 1980–2012 26 Figure 3.2: Average Monthly Wholesale Market Prices (K Sh/90 kg), 2005–13 35 Figure 3.3: Price of Tea at Mombasa Auction ($/kg), 1980–2012 36 Figure 3.4: International Coffee Prices ($/lb), 1988–2013 36 Figure 3.5: Weekly Beef Cattle Prices (K Sh/kg) in Various Markets, 2006–11 37 Figure 3.6: Beef Cattle versus Maize TOT in S Major Markets, 2006–11 38 Figure 3.7: Cattle versus Maize TOT in Isiolo Market, 2006–11 38 Figure 3.8: Domestic Fertilizer Prices, 1998–2007 39 Figure 3.9: Exchange Rates ($/K Sh), 1995–2013 39 Figure 3.10: Commercial Banks’ Interest Rates (%), 1992–2013 39 Figure B3.1.1: Milk Production in the Formal Sector (millions of liters), 1984–2008 42 Figure 3.11: Humanitarian Assistance to Kenya ($ millions), 2000–11 42 Figure 4.1: Indicative Production Losses and Frequency for Key Crops, 1980–2012 47 Figure 4.2: Indicative Crop Losses for Maize, 1980–2012 48 Figure 4.3: Prioritization of Risks to Kenya’s Livestock Sector 48 Figure 5.1: Human Development Index Scores, by Province 50 Figure 5.2: Map of Kenya’s Livelihood Zones 51 Kenya: Agricultural Sector Risk Assessment v Figure 5.3: Percent of Severely Food Insecure, Non-WFP Beneficiary Households by Livelihood Zone 51 Figure 6.1: Prioritization of Key Agricultural Risks in Kenya 56 Figure A.1: Current Suitability of Tea Production Areas 76 Figure A.2: Future Suitability of Tea Production Areas 77 Figure A.3: Suitability Change for Tea Production in 2050 77 Figure C.1: Human Development Index Scores, by Province 92 Figure C.2: Map of Kenya’s Livelihood Zones 92 Figure C.3: Household Food Security by Livelihood Zone 93 Figure D.1: Agro-Ecological Zones 95 Figure D.2: Mean Annual Rainfall (mm) 95 Figure D.3: Monthly Cumulative Rainfall Patterns by Rainfall Zone (mm), 1981–2011 96 Figure D.4: Location of Regional Weather Stations in Kenya 97 Figure E.1: Provinces in Kenya before 2010 99 Figure E.2: Rainfall Pels Superimposed on a Map of Kenya 100 Figure E.3: Monthly Rainfall Pattern by Region 101 Figure E.4: Calendar for Main Crops in Kenya 105 Figure E.5: Map of Average Cumulative Rainfall, by Pixel 105 Figure F.1: Maize Production, 1990–2012 107 Figure F.2: Wheat Production, 1990–2012 107 Figure F.3: Dry Bean Production, 1990–2012 108 Figure F.4: Tea Production, 1990–2012 108 Figure F.5: Coffee Production, 1990–2012 108 Figure F.6: Sugarcane Production, 1990–2012 108 Figure G.1: TOT of Individual Markets in Northern Kenya, 2006–11 (number of 90-kg bags of maize exchanged for 1 beef cow) 110 Figure I.1: Prioritization of Risk Mitigation Solutions for Food Crops 113 Figure I.2: Prioritization of Risk Transfer Solutions for Food Crops 113 Figure I.3: Prioritization of Risk Coping Solutions for Food Crops 113 Figure I.4: Prioritization of Risk Mitigation Solutions for Cash Crops 114 Figure I.5: Prioritization of Risk Mitigation Solutions for Livestock 114 Figure I.6: Prioritization of risk Coping Solutions for Livestock 114 TABLES Table 2.1: Agro-Ecological Zones and Rainfall Characteristics in Kenya 8 Table 2.2: Livestock Population in Kenya, 2009 and 2012 10 Table 2.3: Trends in Crop Production, 1990–2012 10 vi Agriculture Global Practice Technical Assistance Paper Table 2.4: Value of Horticultural Production 12 Table 2.5: Coefficients of Variation for Crop Production, 1980–2012 15 Table 2.6: Cereal Supply/Demand Balance (thousand tons), 2013/14 16 Table 2.7: Value of Agricultural Exports (US$ thousands), 2010–12 19 Table 2.8: Livestock Populations in Kenya 22 Table 3.1: Frequency of Major Drought Events in Kenya, 1981–2011 27 Table 3.2: Dates and Impacts of Drought Events in Kenya, 1980–2011 28 Table 3.3: Frequency of Surplus Rainfall Events, 1963–2012 30 Table 3.4: Principal Pest and Disease Risks in Kenyan Agriculture 32 Table 3.5: Frequency and Impact of Livestock Disease Outbreaks in Kenya, 1980–2013 34 Table 3.6: Interannual Crop Price Variability, 1991–2011 35 Table 4.1: Cost of Adverse Events for Crop Production, 1980–2012 46 Table 4.2: Cost of Adverse Events by Crop, 1980–2012 47 Table 5.1: Household Characteristics According to Position in the Maize Market, 1997, 2000, and 2004 (nationwide sample of small-scale households in Kenya) 53 Table 5.2: Households’ Prioritization of Risks in ASAL Counties 53 Table 6.1: Indicative Risk Management Measures 57 Table 6.2: Filtering Criteria for Risk Management Solutions in Kenya 64 Table C.1: Poverty Transitions by Livelihood Group 93 Table E.1: Rainfall Anomalies for the 18 Weather Stations 103 Table E.2: Simple Linear Regression Results for Maize 106 Table E.3: Multiple Linear Regression Results for Wheat 106 Table E.4: Simple Linear Regression Results for Wheat 106 Kenya: Agricultural Sector Risk Assessment vii ACRONYMS AND ABBREVIATIONS ACCI Adaptation to Climate Change and Insurance IBLI Index Based Livestock Insurance Programme ICT information and communication technology AIDP Agriculture Insurance Development Program IFAD International Fund for Agricultural ASAL Arid and Semi-arid Land Development ASDS Agricultural Sector Development Strategy IFPRI International Food Policy Research Institute CAADP Comprehensive Africa Agriculture IITA International Institute of Tropical Agriculture Development Programme ILRI International Livestock Research Institute CBD Coffee berry disease IMF International Monetary Fund CBK Coffee Board of Kenya IPCC Intergovernmental Panel on Climate Change CBOK Central Bank of Kenya IPM Integrated Pest Management CBPP Contagious bovine pleuropneumonia ITCZ Inter-Tropical Convergence Zone CFA Cash-for-Assets KARI Kenya Agricultural Research Institute CIAT Information Center for Tropical Agriculture KCC Kenya Cooperative Creameries CIMMYT International Maize and Wheat Improvement KCPTA Kenya Coffee Producers and Traders Center Association CLR Coffee leaf rust KEPHIS Kenya Plant Health Inspectorate Services COMESA Common market for Eastern and Southern KESREF Kenya Sugar Research Foundation Africa kg Kilogram CSAE Centre for the Study of African Economies KNBS Kenya National Bureau of Statistics CV Coefficient of variation KRA Kenya Rainwater Association DAP Diammonium phosphate K Sh Kenyan shillings DTMA Drought Tolerant Maize for Africa KTDA Kenya Tea Development Authority EAC East African Community LGP Length of the growing period ECF East Coast fever LMD Livestock Marketing Division EU European Union MDTF Multi Donor Trust Fund FAO Food and Agriculture Organization of the MHH Male-headed household United Nations MLND Maize lethal necrosis disease FAOSTAT FAO Corporate Statistical Database mm Millimeters FFA Food-for-Assets MoA Ministry of Agriculture FHH Female-headed household MoALF Ministry of Agriculture, Livestock and FMD Foot and mouth disease Fisheries FPEAK Fresh Produce and Exporters Association of MT Metric ton Kenya MTP-I Medium Term Plan I GCM General circulation model MTP-II Medium Term Plan II GDP Gross domestic product NASEP Kenya’s National Agricultural Sector GM Genetically modified Extension Policy GoK Government of Kenya NCCRS National Climate Change Response Strategy ha Hectare NCPB National Cereals and Produce Board HCDA Horticultural Crops Development Authority NDDCF National Drought and Disaster Contingency HDI Human Development Index Fund HSNP Hunger Safety Net Program NDMA National Drought Management Authority Kenya: Agricultural Sector Risk Assessment ix NDVI Normalized Difference Vegetation Index SACCO Savings and Credit Cooperative NERICA New Rice for Africa SDL State Department of Livestock NGO Nongovernmental organization SECO Swiss Secretariat of Economic Affairs NIB National Irrigation Board TBK Tea Board of Kenya NSNP National Safety Nets Programme TOT Terms of trade OIE World Organization for Animal Health USAID U.S. Agency for International Development PFS Probability of a failed season WDI World Development Indicators PPP Public-private partnership WFP World Food Programme PPR Peste des petits ruminants WHFSP Water Harvesting for Food Security RPLRP Regional Pastoral Livelihoods Resilience Project Programme Note. All dollar amounts are U.S. dollars unless otherwise indicated. x Agriculture Global Practice Technical Assistance Paper ACKNOWLEDGMENTS This report was prepared by the Agricultural Risk Man- of Livestock for their invaluable support and contributions agement Team of the World Bank’s Global Food and throughout the assessment process. The team would like to Agriculture Practice (GFADR). The assessment team was thank all those who participated in the consultative process led by Stephen D’Alessandro and consisted of World Bank and who shared their invaluable time, perspective, and per- consultants Jorge Caballero, John Lichte, Simon Simpkin, sonal experiences. Their inputs greatly enriched the analysis Jeremy Swift, Jonathan Nzuma, Alice Mirage, and Eric and the study’s findings. The team is also grateful to Ade- Njue. Traci Johnson and Srilatha Shankar (GFADR) also mola Braimoh, Ladisy Komba Chengula, Vikas Choudhary provided valuable analytical and logistical support. (GFADR), and Daniel Clarke (GFMDR) for providing feed- back, guidance, and support during the report’s preparation. The authors would like to thank all the technical special- ists working across Kenya’s Ministry of Agriculture, For- Finally, the authors would like to highlight the generous estry and Fisheries (MALF) who contributed their expertise, contributions from USAID, Ministry of Foreign Affairs insights, and time to the study. The team is especially grate- of the Government of the Netherlands, and State Secre- ful to Kenneth O. Ayuko of the State Department of Agri- tariat for Economic Affairs (SECO) of the Government culture and Vincent Githinji Ngari of the State Department of Switzerland. Kenya: Agricultural Sector Risk Assessment xi EXECUTIVE SUMMARY Agriculture remains vital to Kenya’s economic growth. It is also vital to the country’s food security and poverty reduction efforts. Because the vast majority of Kenya’s poor depend on smallholder agriculture for their livelihood, increasing their productivity can contribute at once to improving food availability, increasing rural incomes, lower- ing poverty rates, and growing the economy. Putting more and better seeds, fertilizers, and other inputs into the hands of farmers and pastoralists and finding ways to link them more directly to markets are among the key thrusts of current sector develop- ment policies. More broadly, Kenya’s Vision 2030 aims in part to transform the coun- try’s agriculture from subsistence to a more competitive and commercially oriented sector, one that can meet the country’s food needs, expand exports, and become a key engine for forward growth. Despite Kenya’s strong commitment to agriculture, sectoral growth remains well below the 6 percent target, and meaningful gains in productivity and in rolling back rural poverty have been slow in coming. The Economic Survey 2014 shows that the agricul- ture sector grew by a mere 2.9 percent in 2013, down from 4.2 percent a year earlier. Moreover, Kenya continues to rely heavily on imports to feed its growing population1 amid a widening structural imbalance in key food staples. Key constraints limiting sector growth are well documented, as are associated response measures. Less well understood is how risk dynamics associated with production, mar- kets, and policy adversely impact sector performance, in terms of both influencing ex ante decision making among farmers, traders, and other sector stakeholders and causing ex post losses to crops, livestock, and incomes—destabilizing livelihoods and jeopardizing the country’s food security. The present study was commissioned in part to bridge this knowledge gap. It is the first step in a multiphase process designed to integrate a stronger risk focus into sector planning and development programs. It seeks to learn from and build on a range of 1 At 2.7 percent, Kenya has one of the highest rates of population growth in the world, according to World Development Indicators (WDI). The country’s population has tripled in the past 35 years. Kenya: Agricultural Sector Risk Assessment xiii FIGURE ES.1. HISTORICAL TIMELINE OF MAJOR AGRICULTURAL PRODUCTION SHOCKS IN KENYA, 1980–2012 14 Agriculture, value added (annual % growth) 20M kg 12 (US$11.4M) of green leaf 10 damaged by Drought; Violence follows frost; floods, 8 floods, elections; 2012 RF fever drought, 6 2006 2007 4 2 Drought, 0 Commodity 2011 price shock, –2 La Nina 2008 drought, –4 1999–2000 Drought Drought El Niño floods; Erratic rains, Prolonged –6 1983–84 1991–93 1.5 m affected; floods, drought, RV fever, 2002 –8 1997–98 2008–2009 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: WDI; authors’ calculations. broad initiatives by the government of Kenya (GoK) and that Kenya has experienced an extreme rainfall event its development partners purposed to enhance Kenya’s during two of every three years on average. It also sug- resilience and response to natural disasters. The ultimate gests the increasing frequency of severe droughts affecting objective is implementation of a holistic and systematic large swaths of the country in the last decade and rising risk management system that will reduce the vulnerability levels of year-on-year rainfall variability. The combination and strengthen the resiliency of Kenya’s agricultural sup- of frequent severe droughts, high dependence on rainfed ply chains, and the livelihoods that depend on them. This agriculture, and high poverty rates among smallholder sector risk assessment is the primary output of Phase One. farmers and pastoralists makes Kenya particularly vulner- The study’s main objective is to identify, assess, and prior- able to the effects of droughts. Erratic rainfall, punctu- itize principal risks facing Kenya’s agriculture sector by ated by severe droughts, is the biggest risk facing Kenya’s analyzing their impacts via quantitative2 and qualitative agriculture sector, with profound impacts on both crop measures. Based on this prioritization, the study identi- and livestock production. In addition to extreme weather fies key intervention areas for improved risk management. events, the global financial and economic crisis, high food The review encompasses the 33-year period 1980–2012. and fuel prices, and a tense and at times uncertain politi- cal environment in recent years have repeatedly disrupted Figure ES.1 depicts a historical timeline of the most nota- agricultural supply chains and markets, jeopardizing ble risk events to adversely impact sector performance growth and the sector’s ability to provide food security during the period under review. The study’s main findings and reduce poverty. Other key findings of the assessment highlight an agriculture sector increasingly vulnerable to are presented below. extreme weather variability. An analysis of cumulative annual rainfall during the period 1980–2011 indicates CROP AND LIVESTOCK PRODUCTION RISKS 2 A more extensive quantification of risk impacts was hampered by notable Extensive livestock systems and pastoralists in Kenya’s inconsistencies and gaps in production, weather, and other time-series data, underlining the need for future investments in improved data collection, man- northern rangelands are particularly vulnerable to agement, and dissemination. the effects of drought. Estimated losses to livestock xiv Agriculture Global Practice Technical Assistance Paper FIGURE ES.2. ESTIMATED LOSSES TO AGGREGATE CROP PRODUCTION FROM RISK EVENTS, 1980–2012 (US$, millions) 450 400 350 300 250 200 150 100 50 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT; authors’ calculations. populations from droughts that have occurred within the banana, and dry beans also experienced notable losses most recent decade alone amount to more than $1.08 over the period. Sugarcane represented for nearly half billion. Ancillary losses related to production assets and (46 percent) of aggregate indicative losses by volume but future income and the costs of ex post response measures less than 6 percent by value (figure ES.2). are likely several times that figure. The increased inci- dence of droughts across Kenya’s arid and semiarid lands Relative to most other crops, maize is highly susceptible to (ASALs) in recent years means that affected communities moisture stress. Kenya’s strong reliance on rainfed maize have less time to recover and rebuild their assets. This has production in meeting its food needs and growing con- weakened traditional coping mechanisms, handicapping solidation of production toward maize (and dry beans) household resilience against future shocks. has rendered the country increasingly vulnerable to sup- ply disruptions and food shortages. Amid declining yields, Key select crops in Kenya experienced significant pro- productivity gains have come largely through land expan- duction losses3 in 13 years as a result of adverse risk sion into marginal areas that receive lower and more vari- events during the period 1980–2012, or once every three able rainfall. This trend coupled with Kenya’s increasingly years on average (figure ES.2). All of these crop loss erratic rainfall has made the country’s maize production events resulted in a drop in agricultural gross domestic more susceptible to moisture stress and year-on-year yield product (GDP) of 2 percent or more. Losses ranging variability, with significant implications for the country’s from 3 to 4.2 percent occurred in six years. Indicative food security. losses were substantial for these events, totaling nearly $5.10 billion, or roughly $154.5 million on an aver- Beyond weather risks, the analysis highlights the impor- age annual basis, during the 33-year period. Maize tant threat that pests and diseases pose to Kenya’s farm- accrued by far the biggest losses measured in production ers. Left unchecked, crop pest and diseases regularly value over the period, accounting for nearly one-fifth cause considerable pre- and postharvest losses that (19.8 percent) of total indicative losses. Coffee, tea, dampen yields and incomes. The most common crop threats are armyworms, thrips, aphids, mealybugs, and nematodes, which are all a permanent fixture of 3 Measured in terms of gross agricultural value, or the total value of volume of Kenya’s agricultural landscape, as elsewhere. Parasitic production for each crop multiplied by the producer price. Crops covered in the weeds such as Striga are another common threat, affect- analysis include maize, wheat, paddy rice, sorghum, Irish potatoes (1980–2006), cowpea, dry beans, tea (1988–2012), coffee, sugarcane, bananas, and green ing large swaths of Kenya’s prime cropland. Maize is beans. particularly susceptible to a range of fungal (e.g., rust, Kenya: Agricultural Sector Risk Assessment xv spot, blight, smut) and viral diseases (e.g., maize streak high mortality rates, especially in improved breeds. Vacci- virus). The most noteworthy is maize lethal necrosis dis- nation is effective, but existing coverage is limited (roughly ease (MLND), which is considered the greatest threat 10 percent). Widespread outbreaks were recorded every to maize production as no definitive or resistant varie- third year on average during the review period. One ties have yet to emerge from research since the disease severe FMD flare-up in the early 1980s resulted in losses was first detected in June 2011. Incidence in the field valued at an estimated K Sh 230 million. Other notable ranges from 40 to 100 percent of the crop, and crop diseases include small ruminant pest (PPR [Peste des petits losses of over 80 percent have been reported. Among ruminants]), contagious bovine pleuropneumonia (CBPP), Kenya’s industrial crops, coffee is particularly suscep- and catarrhal fever. The risk associated with animal dis- tible to coffee berry disease (CBD) and coffee leaf rust ease is especially acute during drought when even com- (CLR), two major diseases of Arabica coffee that, left mon day-to-day levels of infection or internal or external untreated, can cause up to 50–80 percent losses. Chron- parasites can be fatal. ically low—farm-gate prices offer poor incentives to farmers to invest in control measures for these diseases, MARKET RISKS aggravating their impact. At the market level, the analysis highlights price volatility as the most significant risk. Producer prices in Kenya for Tea production in Kenya has long benefitted from favora- key crops are subject to moderate to high levels of inter- ble growing conditions that limit pest and disease threats. annual price variability. Rice paddy, coffee, sorghum, and Good crop husbandry is also an enabling factor. For the to a lesser extent, cowpea exhibit the highest levels of period 1988–2012, tea yield variability was by far the low- year-on-year producer price volatility. In the case of rice est among the crops analyzed. However, given the high and coffee, this suggests that domestic price fluctuations market value of tea, even small yield drops can amount are influenced by imports and/or changes in international to considerable dollar losses, as happened in 2009, 2011, market prices. It also suggests that rice and coffee produc- and 2012, when aggregate losses for the three years ers in Kenya are exposed to significant swings in farm- totaled $376.7 million. Although it remains unclear from gate prices from one year to the next. the analysis, the factors that may be driving higher levels of observed yield volatility of late, some industry officials Although public support programs manage to keep cited the effects of shifts in rainfall and temperature pat- producer prices for maize relatively stable, wholesale terns, with extended dry periods possibly linked to climate prices are among the most volatile, a critical issue for change. In addition, farmer groups interviewed for this the GoK given maize’s importance to household con- study highlighted increased incidences of frost. Notably, sumption and food security. Sharp increases during 20 million kg ($11.4 million) of green leaf was reportedly 2008–2009 and then again in 2011 and 2012 coincided damaged by frost in January 2012. with domestic and external shocks. For example, maize prices jumped by 145 percent during the first six months For Kenyan livestock, diseases pose a significant threat, of 2011 following a sharp increase (39 percent) in the though due to a paucity of data, related impacts are dif- commodity food price index and a near doubling of ficult to measure. East Coast fever (ECF) is perhaps the U.S. maize prices in 2010.4 In general, domestic maize most noteworthy threat. Tick-borne, ECF can kill large prices tend to be more volatile than international maize numbers of calves in pastoralist herds. The presence of prices, as domestic prices are highly sensitive to uncer- ECF in neighboring countries severely handicaps effec- tainty and constant speculation in projected and real tive control. Rift Valley fever in Kenya is similarly hard to annual output. The GoK’s active role in cereal markets, control but is more predictable due to its positive correla- while designed to increase productivity, stabilize prices, tion with heavy rainfall and flooding. During outbreaks, and ensure food availability, can also discourage private animal losses are often high, as treatment by vaccination frequently leads to abortion in pregnant animals. Foot and mouth disease (FMD) is endemic in Kenya and can cause 4 According to Index Mundi at indexmundi.com. Data accessed May 2014. xvi Agriculture Global Practice Technical Assistance Paper sector investment in input supply, storage, and other ser- With regard to Kenya’s sugar industry, the unpredict- vices due to the added uncertainty over the timing and ability of current policy5 related to import regulations scale of public interventions. and ongoing exceptions to the COMESA rules pose con- siderable risk to mills, cane producers, and other stake- holders. Unpredictability also impedes investments and ENABLING ENVIRONMENT needed industry reforms, including the planned privati- RISKS zation of remaining government-owned mills. Sizable Among notable risks within the sector’s enabling environ- unrecorded imports of refined sugar from outside the ment are Kenya’s growing cereal imports, which bring region pose additional risks to the industry. Prices can added uncertainty to the country’s food security situa- fall precipitously when the market becomes saturated tion. Imports today make up a much higher proportion and mills are unable to compete, as happened in 2002 (37 percent) than they did a decade ago. This exposes the when industry assumed massive debts. A more recent country to external pressures that can adversely impact surge in sanctioned and unsanctioned imports in 2013 domestic food prices, availability, and access. Moreover, resulted in sizable government payouts to a number of amid recurrent maize shortages, uncertainty exists about mills to stave off their bankruptcy. whether rising Kenya maize imports will be able to fill the gap in light of Kenya’s 50 percent ad valorem tariff for Finally, the political uncertainty and associated insecurity non-COMESA (common market for Eastern and South- that have disrupted agricultural production and markets ern Africa) sourced maize, its import ban on genetically in recent years have declined markedly since the new modified (GM) maize, and inadequate supplies of non- constitution was enacted in 2010. Moving forward, the GM exportable maize in the COMESA region. This is restructuring, consolidation, and reorganization of the especially true in light of episodic export bans for maize sector’s legal and regulatory frameworks and ministerial in Tanzania, Malawi, and Zambia during production functions and the devolution of policy planning, deci- shortfalls. Supply markets have also thinned out due to the sion making, and administration to the county level will growing attractiveness of the South Sudan market and of continue to have major consequences for the sector. Such markets in the Democratic Republic of Congo for Ugan- seismic change imparts uncertainty and significant and dan and Tanzanian maize exports. myriad institutional risks in the short to medium term. These include potential for increased inefficiencies, dis- The increasing frequency of shocks and the country’s grow- ruptions, and breakdown of critical public services such as ing dependence on emergency aid are also noteworthy. In extension, data collection, and management information addition to an estimated half million Somalian and Suda- systems (MIS) and higher volatility of producer, whole- nese refugees in Kenya’s Dadaab and Kakuma camps, an sale, and retail prices. estimated 1.5 million Kenyans are chronically food inse- cure and in need of assistance, according to the World MANAGING Food Programme (WFP). In drought years, that number AGRICULTURAL RISKS can grow exponentially, as it did in 2011 when 4 million While hindering growth, unmanaged risks are also a sig- Kenyans in the northern rangelands needed food aid. Dur- nificant factor contributing to chronic poverty in Kenya. ing 2006–11, Kenya received $1.92 billion in emergency Shocks to agricultural production and markets adversely aid, up from $150 million during the prior five-year period impact household well-being in a variety of ways: (2000–04). As evidenced elsewhere, frequent crises coupled by limiting food availability, weakening food access, with an overreliance on food aid can lead to a breakdown and negatively affecting future livelihoods through of household resilience. Although emergency food aid can income disruption and depletion of productive assets. help address immediate food needs, it does little to help rebuild household resilience and may induce higher rates 5 In February 2014, COMESA approved the extension for a further year of of dependency and chronic malnutrition. As such, it also Kenya’s special safeguard arrangement for sugar, thus allowing Kenya to main- can increase the cost of managing future crises. tain a 350,000-ton ceiling on duty-free sugar imports from COMESA. Kenya: Agricultural Sector Risk Assessment xvii Chronically vulnerable groups with high exposure to systems on which their livelihoods and the country’s food risks experience a disproportionately large impact from security depend: adverse events and typically lack coping mechanisms » To better optimize rainfall and soil moisture available to other groups. Understanding these and in marginalized production areas, promoting com- other risk dynamics is key to developing appropriate risk munity-driven investments in improved management responses that can help reduce production soil and water management measures such volatility, safeguard livelihoods, and put the sector and as terracing, water harvesting pans, roof and rock the broader economy on a firmer footing for growth. catchment systems, subsurface dams, and micro- Effective strategies can also make a meaningful contri- irrigation systems bution to poverty reduction efforts. » To curb soil erosion, increase soil fertility and water retention, and enhance the produc- Management of agricultural risk is not new to Kenya, tivity6 and biodiversity of smallholder systems and the GoK has a long track record of investment in across Kenya, promoting broader awareness and risk mitigation, transfer, and coping mechanisms. Mov- adoption (via farmer field schools and other par- ing forward, Kenya’s Vision 2030 recognizes the need to ticipatory extension approaches) of conservation strengthen existing risk management systems, and the agriculture practices such as zero tillage, mulch- GoK has launched a range of new initiatives to confront ing, composting and use of organic fertilizers, crop the most severe threats facing the country. In 2011, it diversification and rotation, intercropping, and established the Drought Risk Management Authority to integrated pest management (IPM) better coordinate preparedness and speed up response » To strengthen certified seed production and measures. It also launched the Disaster Risk Reduction distribution systems, build their credibility, and Program, the National Climate Change Action Plan, stimulate demand for improved seeds and fertiliz- and the National Hunger Safety Net Program. These ers by smallholders, investing in capacity building and other initiatives by the GoK and its development and training to strengthen monitoring and partners are already helping to safeguard livelihoods, enforcement of quality standards and reduce promote adaptation, and strengthen resilience against incidences of counterfeiting, adulteration, impacts from natural disasters and a changing climate. and other abuses that dampen farmer demand and Yet as revealed in this report, Kenya’s agricultural sup- productivity ply chains remain highly vulnerable to myriad risks that » To reverse degradation of water, soil, and disrupt the country’s economic growth, cripple poverty vegetation cover, safeguard the long-term reduction efforts, and undermine food security. The cur- viability of Kenya’s arid and semiarid range- rent study highlights the need for a more targeted and land ecosystems, and ensure access to suf- systematic approach to agricultural risk management in ficient grazing land, promoting: (1) use of contour Kenya. erosion and fire barriers, cisterns for storing rain- fall and runoff water, controlled/rotational graz- Based on an analysis of key agricultural risks, an evalua- ing, grazing banks, homestead enclosures, residue/ tion of levels of vulnerability among various stakeholders, forage conservation, and other sustainable and the filtering of potential risk management measures, land management practices; and (2) innova- this assessment makes the following recommendations tive rangeland comanagement (state and local for the GoK’s consideration. The proposed focus areas community) approaches that leverage customary of intervention encompass a broad range of interrelated, mutually supportive investments that together—aligned with Livelihoods Enhancement goals within Vision 2030—hold 6 Conservation agriculture allows yields comparable with modern intensive agri- strong scope to strengthen the resilience of vulnerable culture but in a sustainable way and with lower production costs (time, labor, farming and pastoralist communities and the agricultural inputs). Yields tend to increase over the years with yield variations decreasing. xviii Agriculture Global Practice Technical Assistance Paper forms of collective action and economic instru- market prices (input/output), agricultural research ments to reward sound pasture management and advice, and so on » To strengthen drought resilience among » To further objectives of the devolution process, pro- vulnerable pastoral communities in target moting institutional and organizational capacity ASAL counties and better safeguard the viability building and technical training at county and of animal herds during shortages, supporting the national levels to promote standardized collec- development of feed/fodder production and tion and management of agricultural data storage systems, animal health, market (in line with recently developed national guidelines) and weather information, and other critical services The purpose of this study was to help policy makers, » To mitigate growing pressures on rangelands in Ministry of Agriculture, Livestock and Fisheries (MoALF) ASALs and increasing vulnerability of smaller live- and other GoK officials, and the wider development com- stock (<50 animals) owners, in particular, putting munity better understand the most important risks fac- in place supportive policies and livelihood devel- ing Kenya’s agriculture sector. It is hoped that the study’s opment programs (targeted credit schemes, findings will inform ongoing and future policy planning skills training, public sector investments in labor- and programming to ensure sustainability of agricul- intensive infrastructure projects, cash for work) to tural investments and enhanced agricultural resilience facilitate their engagement in alternative liveli- over time. It is also hoped that the findings will lead to hood and income-generating activities improved decision making and successful implementation » To strengthen fiscal management and over time of a comprehensive, integrated, and ultimately reduce the GoK’s budget volatility (and diver- responsive risk management framework for Kenya’s agri- sion of development resources caused by ex post culture sector. crisis response), better safeguard rural liveli- hoods, and increase resilience, deepening Several of the recommendations proposed in the report investments in agricultural insurance mechanisms are already being considered or presently undergoing and markets (in partnership with the private sec- implementation. Some may well already constitute an tor), with an initial focus on asset protection integral part of existing risk management systems. Once (via early warning triggers and expedited payouts) the GoK has defined its priorities, Phase II will focus on among vulnerable pastoralist communities identifying: (1) pathways for scaling-up successful inter- and area yield index insurance for smallholder ventions to encompass a greater number of beneficiaries maize farmers and (2) existing gaps that need to be addressed. This will » To facilitate improved, evidence-based deci- entail an in-depth review of Kenya’s current risk manage- sion making among farmers, pastoralists, and pol- ment landscape to assess the effectiveness of various inter- icy makers and to mitigate price volatility, investing ventions, principal barriers, and challenges, and leverage in integrated data and information systems points to enhance more stakeholders’ access to risk man- build-out for more robust, cost-effective, and agement mechanisms. Phase II’s anticipated outcome will reliable collection, management, and dis- be development and implementation of a risk manage- semination (via terrestrial surveying, geographic ment implementation and monitoring road map, one that information systems [GIS], information and com- will reduce the vulnerability and strengthen the resilience munications technology [ICT], short message of Kenya’s agriculture sector and the millions of house- service [SMS]) of crop production, agro-weather, holds that depend on it for their livelihoods. Kenya: Agricultural Sector Risk Assessment xix CHAPTER ONE INTRODUCTION Despite myriad challenges, Kenya has emerged in recent years as one of Africa’s “frontier economies,” with headline growth in the most recent decade propelling the country toward middle-income status. Poverty rates have declined, while gross domes- tic product (GDP) per capita ($943 in 2012) has more than doubled. Average real GDP growth was 5.1 percent during 2010–13. Spurred by its dynamic business commu- nity, strong communication and transport links, and relatively well-developed financial and services sectors, Kenya is today among the top five destinations for foreign direct investment in Sub-Saharan Africa.7 Yet despite recent gains, poverty remains a major challenge. Moving ahead, the Government of Kenya (GoK) has set ambitious eco- nomic growth and poverty reduction targets; the economy is expected to expand by 6.3–6.5 percent during 2014–16 (IMF 2014a). Achieving these targets will depend to a large extent on the future performance of Kenya’s agriculture sector. A principal source of employment and major contributor to GDP, agriculture remains vital to the Kenyan economy. Nearly three in four Kenyans live in rural areas and are actively engaged in the production, processing, and marketing of crop, livestock, fish, and forest products. The sector accounts for an estimated 75 percent of informal employment and is the principal source of rural income and livelihoods. It also generates nearly two- thirds (65 percent) of merchandise exports and roughly 60 percent of foreign exchange earnings (World Bank 2013). In the five years ending in 2012, the sector’s annual contri- bution to GDP averaged 27.3 percent (29.9 percent in 2012). Not surprisingly, Kenya’s GDP growth is highly correlated with agriculture sector growth (figure 1.1). Launched in 2008, Kenya’s Vision 2030 strategy identifies agriculture as one of six priority sectors critical to delivering on the GoK’s economic growth target of 10 per- cent per annum. The second Medium Term Plan (MTP-II) of Vision 2030 covers the period 2013–2017 and looks to build on successes achieved in MTP-I (2008–12). Under MTP-II, objectives include maintenance of a stable macroeconomic environ- ment, modernization of infrastructure, and diversification and commercialization 7 Kenya was ranked as the fifth and fourth most popular destination for foreign direct investment in terms of new projects in 2011 and 2012, respectively, according to Ernst & Young (2013). Kenya: Agricultural Sector Risk Assessment 1 FIGURE 1.1. AGRICULTURAL GDP VERSUS NATIONAL GDP GROWTH (% change), 1968–2012 12 Agriculture GDP growth GDP growth 10 8 6 4 2 0 68–72 73–77 78–82 82–87 88–92 92–97 98–02 02–07 08–12 Source: WDI 2014 (http://data.worldbank.org/data-catalog/world-development-indicators). Note: Based on five-year averages. of agriculture. The five-year framework targets average sistence into a more competitive and commercially ori- annual real GDP growth of 8.2 percent between 2013 ented sector. Covering the period 2010–20, the ASDS is and 2017, with double-digit growth by 2017. Achiev- anchored in two strategic thrusts: (1) increasing produc- ing these targets will require a significant acceleration in tivity, commercialization, and competitiveness of agri- agricultural growth, which averaged 3.5 percent during cultural commodities and enterprises; and (2) developing 1997–2012 (IMF 2014a). and managing key factors of production (GoK 2010b). In addition to boosting growth, near-term targets include Considering the notable variability in year-on-year sec- reducing the share of the population living below the tor performance (figure 1.2), new and better ways must be absolute poverty line to less than 25 percent and cut- found to strengthen agricultural supply chains and make ting food insecurity by 30 percent. The ASDS calls for them more resilient to downside risks. Extreme volatility increased investments to, inter alia, promote the uptake of characterized Kenya’s agriculture sector’s annual growth new technologies, exploit irrigation potential, undertake over the period 1980–2012, particularly during the most crucial sector policy reform, improve institutional govern- recent two decades. Shifting weather patterns, popula- ance, and ensure more sustainable management of natu- tion growth, changing demographics, increasing market ral resources. integration, political instability, and other domestic and external pressures are making Kenyan agriculture more Between 2009 and 2013, Kenya allocated an average of vulnerable while exposing it to higher levels and incidences 4.6 percent of its national budget to agriculture—less of risk. Adverse impacts from droughts, floods, pest, and than half of the Maputo9 target. Looking ahead, MTE-II disease outbreaks, and other shocks repeatedly disrupt sec- commits the GoK to increasing public spending on agri- tor activities, jeopardizing incomes, hobbling sector growth, culture to 8 percent of the budget by 2020 (IMF 2014b). and handicapping livelihoods. They also contribute to food For the current fiscal year, the government scaled back its deficits and diversion of development resources to ex post agriculture budget by 29 percent, or $447 million versus emergency response and recovery measures. $627 for the previous year.10 A portion of this was to be The GoK’s Agricultural Sector Development Strategy (ASDS)8 incorporates the growth objectives of Vision 2030 9 Within the framework of the Comprehensive Africa Agriculture Development by seeking to transform Kenyan agriculture from sub- Programme’s (CAADP) Maputo Declaration, Kenya has committed to spend- ing 10 percent of its national budget on agriculture. 8 ASDS is a revision of the Strategy for Revitalizing Agriculture (2004–2015), 10 See “Kenya falls short of Maputo Declaration on Agriculture,” by Kibiwott incorporating Vision 2030 objectives. Koross, The Star, September 3, 2013. 2 Agriculture Global Practice Technical Assistance Paper FIGURE 1.2. AGRICULTURAL VALUE ADDED (annual % growth), 1980–2013 14 12 10 8 6 4 2 0 –2 –4 –6 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: WDI; 2013 growth estimate of 2.9 percent taken from 2014 Economic Review. reallocated from development expenditure to meet emer- The ASDS recognizes that farmers’ high risk exposure gency food needs. impedes sector growth. In efforts to enhance the resilience of the agriculture sector, safeguard food security, and pro- In recent years, the GoK has had to channel increas- tect livelihoods, the GoK placed drought management and ingly more resources into emergency response measures climate change mitigation and adaptation at the center amid mounting concerns over food security. In 2013, of its agricultural and economic development strategy. Kenya ranked 79 out of 107 countries, lagging behind Among key initiatives, it established the National Drought countries like Ghana, Uganda, and Cote d’Ivoire, on the Management Authority (NDMA) in November 2011 to Global Food Security Index, which measures the afford- better coordinate drought mitigation, contingency plan- ability, availability, and quality of food (Alarcon, Joehnk, ning, and response activities and resources at the national and Koch 2013). The country faces a structural deficit in level. In March 2013, the GoK launched the National some basic food staples, including maize, wheat, rice, and Climate Change Action Plan (NCCAP). NCCAP’s pri- sugar. Stagnant productivity, the high cost of farm inputs, mary objective is to implement Kenya’s National Climate and poorly developed storage are often cited as common Change Response Strategy (NCCRS), which seeks to drive causes. Addressing these and other growth constraints investments in, inter alia, water harvesting, early warning has long been the focus of sector development programs. systems, food storage facilities, broader use of drought- In addition, adverse shocks such as drought, disease out- tolerant crops such as millet and cassava, and promotion breaks, and volatile market prices continue to disrupt of conservation agriculture. To support the livestock sec- and debilitate increasingly vulnerable crop and livestock tor, NCCRS recommends breeding animals better able to production systems and the livelihoods they support. In cope with drought stress, improving vaccination programs response, public spending on emergency food aid has and disease surveillance, and establishing emergency fod- increased markedly in recent years. The number of Ken- der banks, among other initiatives. yans requiring food assistance rose from 650,000 in late 2007 to almost 3.8 million in late 2009 and early 2010. Although responding to threats posed by climate change Recurrent food insecurity remains an ongoing challenge and natural disasters is important, the GoK also recog- in Kenya.11 nizes the need to better manage other risks that adversely impact agriculture. A better understanding of risk occurrences and their frequency and impacts is essential 11 During 2007–11, Kenya received roughly $933 million in emergency food aid. This compared to an estimated $466.2 in emergency response assistance for developing appropriate strategies, interventions, and during the entire 16-year period 1990–2005. policies for improved agricultural risk management. Kenya: Agricultural Sector Risk Assessment 3 It is within this context that the World Bank, with sup- resiliency of agricultural supply chains and the liveli- port from the G-8 and the U.S. Agency for International hoods they support. Development (USAID) and in collaboration with Ken- ya’s Ministry of Agriculture, Livestock and Fisheries The analysis presented in this report is based on a meth- (MoALF), commissioned the present study. It is one of odology for assessing risks in agricultural supply chains a series of agriculture sector risk assessments that the designed by the Agricultural Risk Management Team World Bank agreed to conduct within the framework of the World Bank’s Agriculture and Environmental of the G-8’s New Alliance for Food Security and Nutri- Services department. It offers a conceptual framework tion and in close partnership with partner countries. and set of detailed guidelines for conducting a more The objective of this assessment was to assist the GoK system-wide assessment of risk, risk management, and to (1) identify, analyze, quantify, and prioritize principal vulnerability within agricultural supply chains. The risks (i.e., production, market, and enabling environ- methodology contains logical steps within four consecu- ment risks) facing Kenya’s agriculture sector; (2) analyze tive phases (figure1.3). Phase I, for which this study is the the impact of these risks; and (3) identify and prioritize primary deliverable, focuses on identifying, quantifying, appropriate risk management (i.e., mitigation, transfer, and prioritizing the major risks that cause adverse shocks coping) interventions that might contribute to improved to the sector. Key steps of the analysis include (1) identi- stability, reduced vulnerability, and increased resilience fying and characterizing risks across production systems of agricultural production and marketing systems in and market systems and within the enabling environ- Kenya. This report presents a summary of the assess- ment; (2) prioritizing these risks based on a quantifica- ment’s key findings. tion of their indicative impacts over time; (3) assessing stakeholder vulnerability and the effectiveness of exist- The study focuses on a select basket of priority crops and ing risk management strategies and instruments; and livestock products including maize, wheat, dry beans, tea, (4) identifying priority investments and needed policy coffee, sugarcane, cut flowers, and meat and dairy. These and institutional changes that can strengthen agricul- together accounted for roughly four-fifths of the value of tural systems’ resilience. gross agricultural output in 2012 (FAOSTAT). The rela- tive effectiveness of existing risk management measures Following in-depth analysis of baseline data, the study was also assessed via (1) an appraisal of public interven- team conducted broad-based, in-country consultations tions in the rural sector; (2) discussions with rural stake- with stakeholders in January–February 2014. These holders directly involved in risk management; and (3) a included individual farmers, farmer groups, input suppli- technical consultation on the relative benefits of various ers, market traders, food processors, and representatives risk mitigation interventions. of the government and research and academic institutes in and around Nairobi and in major production zones Chapter 2 provides an overview of Kenya’s agricul- and markets across Kenya’s eastern, northeastern, cen- ture sector and a discussion of key growth constraints. tral, Rift Valley, and western regions. The mission team Chapter 3 assesses the main agricultural risks (produc- organized a wrap-up roundtable consultation at the Min- tion, market, and enabling environment). Chapter 4 istry of Agriculture, Livestock and Fisheries (MoALF) on analyzes the frequency and severity of the major risks February 7, 2014, to share preliminary results and solicit identified and assesses their impact. Chapter 5 pre- feedback. Participants were asked to prioritize possible sents some stakeholder perceptions of these risks and future interventions by ranking a long list of risk manage- the potential to improve their management. Chapter 6 ment solutions. concludes with an assessment of priorities for risk man- agement and a broad discussion of possible risk man- The results of this assessment provide the conceptual agement measures that could help to strengthen the basis for Phase II, which will focus on identifying the 4 Agriculture Global Practice Technical Assistance Paper FIGURE 1.3. AGRICULTURE SECTOR RISK MANAGEMENT PROCESS FLOW PHASE I PHASE 2 PHASE 3 PHASE 4 Client demand Risk Solution Development of risk Implementation and assessment assessment management plan risk monitoring RM plan development Desk review Desk review Implementation by stakeholders Stakeholder In-country Monitoring risks consultations assessment mission Incorporation into existing govt. programs and Stakeholder development plans Finalize analysis Refining RM strategy workshop Source: World Bank. solution areas and related risk management interven- comprehensive risk management framework. It is hoped tions best suited to manage the priority risks identified. that the outcome of this assessment will serve to inform By the end of this activity, the World Bank, in close col- ongoing and future GoK agricultural policy and plan- laboration with the GoK and sector stakeholders, will ning, help ensure sustainability of agricultural invest- develop and validate a matrix of priority interventions ments, and enhance long-term agricultural resilience related to risk mitigation, transfer, and coping within a and growth. Kenya: Agricultural Sector Risk Assessment 5 CHAPTER TWO AGRICULTURE SECTOR IN KENYA To inform the analysis and discussion of agricultural risk in Kenya, this chapter presents an overview of the country’s agriculture sector.12 Sector characteristics most pertinent to risk are given particular attention. Analysis primarily covers the 33-year period from 1980 to 2012 to assess the frequency and severity of the most important risks. Kenya is endowed with diverse physical features, including its low-lying arid and semi- arid lands (ASALs), an extensive coastal belt, plateaus, highlands, and the lake basin around Lake Victoria. Yet Kenya’s agricultural resource base is best characterized by the limited availability of productive land. An estimated 17 percent of the country receives average annual rainfall of more than 800 mm, the minimum required for rainfed agriculture. The remaining land (83 percent) is arid or semiarid, generally unsuitable for rainfed farming or intensive livestock production. Cropland occupies approximately 31 percent, with grazing land (30 percent), forests (22 percent), and game parks, urban centers, markets, homesteads, and infrastructure accounting for the rest (GoK 2010b). Three main land tenure systems exist in Kenya, each of which influences produc- tion systems in different ways: communal lands, government trust lands, and privately owned lands. The communal land ownership system is based on traditional customary rights, in which individuals have a right to use but not sell land. Privately owned lands are registered; the owner holds the title under a freehold or leasehold system. In pasto- ral areas, trust land is the dominant tenure arrangement. Agriculture in Kenya covers small-, medium-, and large-scale farming, with small- holder farmers accounting for more than three-quarters of total production. Produc- tion is heavily reliant on rainfed systems. An estimated 7 percent is irrigated. 12 Broadly, the sector comprises six subsectors: cash crops, food crops, horticulture, livestock, fisheries, and forestry. This study focuses on cash crops, food crops, horticulture, and livestock. Kenya: Agricultural Sector Risk Assessment 7 TABLE 2.1. AGRO-ECOLOGICAL ZONES AND RAINFALL CHARACTERISTICS IN KENYA Agroclimatic Moisture Index Annual Rainfall Land Area Land Area Zone Classification (%) (mm) (%) (km2) I. Agro-Alpine Humid >80 1,100–2,700 II. High Potential Subhumid 65–80 1,000–1,600 12 68,297 III. Medium Potential Semihumid 50–65 800–1,400 IV. Transitional Semihumid to 40–50 600–1,100 5 28,457 semiarid V. Semiarid Semiarid 25–40 450–900 15 85,371 VI. Arid Arid 15–25 300–550 22 125,211 VII. Very arid Very arid <15 150–350 46 261,804 Source: Modified from Sombroek, Braun, and van der Pouw 1982. Livestock production plays an important socioeconomic livestock are concentrated in the ASALs. These house- role in many areas across Kenya. The livestock subsec- holds depend mainly on extensive livestock production tor accounts for roughly 40 percent13 of agricultural gross systems (ranching and pastoralism), often supplemented domestic product (GDP) and as much as 13 percent of by low-input, low-output cropping. Kenya’s high- to national GDP (GoK 2012a), and employs about 50 per- medium-potential areas, which receive more than 1,200 cent of the national agricultural workforce. In the coun- mm of rainfall annually, produce a large variety of crops try’s ASALs, it accounts for as much as nine-tenths of such as tea, coffee, sugarcane, maize, wheat, potatoes, employment and family income. The key livestock subsec- fruits, and vegetables. Figure C.2 in Appendix C pro- tors are beef and dairy cattle, sheep, goats, camels, pigs, vides a breakdown of Kenya’s major farming systems and poultry. and livelihood zones. AGROCLIMATIC CONDITIONS RAINFALL PATTERNS Factors such as climate, hydrology, and terrain determine AND TRENDS Kenya’s agricultural productivity. Climatic conditions The country’s climate is influenced by proximity to the in Kenya vary from humid, tropical regions along the equator, topography, the Indian Ocean, and the Inter- coast, to very humid highlands in the central and west- Tropical Convergence Zone (ITCZ). The ITCZ’s influ- ern regions, to arid inland areas in the north and east. ence is modified by the country’s diverse topography, Kenya has a total land area of 569,140 km2 (excluding which contributes to high spatial variance in seasonal surface water). Of this, 83 percent is classified as ASAL, rainfall distribution due to the altitudinal differences. lying in agro-ecological zones V to VII (table 2.1; see Annual rainfall in Kenya follows a strong bimodal sea- also figure D.1 in Appendix D). Predominantly located sonal pattern. Figure 2.1 and figure D.3 in Appendix D in the northern and eastern portions of the country, the illustrate average cumulative rainfall amounts and sea- ASALs are generally unsuitable for rainfed agriculture sonal patterns in 12 rainfall zones across Kenya during the due to low and erratic rainfall. Roughly 20–30 percent of period 1981–2011. Generally, the long rains occur from Kenya’s population and 50–70 percent of the country’s March to May, while the short rains occur from October to December. Mean annual rainfall ranges from approxi- 13 A joint Intergovernmental Authority on Development/Kenya National mately 200–300 mm in the north and northeast to nearly Bureau of Statistics study completed in 2011 demonstrated that livestock’s 1,400 mm in the central and southwestern highlands. The contribution to Kenya’s agricultural GDP was more than two-and-a-half times larger than the official estimate for 2009, or about $4.54 billion versus $5.25 onset, duration, and intensity of rainfall vary considerably billion for arable agriculture (ICPALD 2013). from one year to the next. However, analysis of rainfall 8 Agriculture Global Practice Technical Assistance Paper FIGURE 2.1. AVERAGE CUMULATIVE RAINFALL (mm) BY RAINFALL ZONE, 1981–2011 1,600 1,400 1,200 1,000 800 600 400 200 – Lodwar Mandera Garissa Voi Makindu Nyahururu Narok Dagoretti Malindi Eldoret Mombasa Kisumu Source: Kenya Meteorological Department. data since 1960 does not show statistically significant cent, respectively, of total production. This growing consol- trends (McSweeney, New, and Lizcano 2012). idation of production toward maize and dry beans makes Kenya increasingly vulnerable to food insecurity. CROP PRODUCTION In Kenya’s heavily populated, high rainfall areas—mainly SYSTEMS in the west—farmers grow a wide range of rainfed food Kenyan agriculture is predominantly carried out on a and cash crops, including cereals, pulses, coffee, tea, fruits, small scale and mainly in high-potential areas. Average and vegetables. In Kenya’s transitional and semiarid areas, farm sizes are 0.2–3 hectares (ha). Small-scale production which cover roughly a fifth of the country and where rain- represents roughly 75 percent of the total agricultural fall is more variable, cropping diversity is less. In these output and 70 percent of the marketed agricultural pro- areas, maize, pulses, roots, and tubers are important, with duce. Smallholders account for over 70 percent of maize, many farming households raising livestock, mostly small 65 percent of coffee, 50 percent of tea, 70 percent of beef, ruminants, in mixed crop/livestock systems. In the arid and 80 percent of milk production (GoK 2013a). Large- to very arid regions that cover roughly 68 percent of the scale farming is practiced on farms averaging about 50 ha country, the land is not suitable for rainfed agriculture. for crops and 30,000 ha for livestock ranches. The large- In these regions, extensive pastoralism is the main source scale farming subsector, which accounts for 30 percent of of livelihoods, centered on cattle, small ruminants (mostly marketed agricultural produce, mainly involves growing sheep and goats), and camels. commercial crops such as tea, coffee, maize, sugarcane, and wheat. Reliable statistics on livestock populations are difficult to obtain. The last comprehensive livestock census was done Agricultural production in Kenya is dominated by in 1969. As in many other African countries, livestock maize (38.2 percent) and dry beans (18.7 percent), which populations in Kenya are estimated, and actual losses are together cover well over half of total cropped area in 2012 difficult to calculate. The 2009 Kenya Population and (figure 2.2). The remainder comprises more than 150 Housing Census included questions on livestock owner- other food and cereal crops, with sorghum (3.9 percent), ship. Table 2.2 highlights considerable differences between cowpea (3.8 percent), tea (3.4 percent), coffee (2.8 percent), FAO figures and the 2009 Census data, especially for spe- wheat, potatoes, pigeon peas, and millet among the most cies commonly kept in the more remote ASAL regions. important (FAOSTAT). This crop composition has been Even for dairy cattle, it was estimated that Kenya’s actual fairly stable over time, with the exception of maize and dry cattle population in 2003/04 could be as many as three beans, which in 1990 comprised 24.4 percent and 12.2 per- times the government’s estimated number (FAO 2011). Kenya: Agricultural Sector Risk Assessment 9 FIGURE 2.2. COMPOSITION OF CROP PRODUCTION (area harvested, in thousand ha), 1990–2012 2,500 Maize 2,000 1,500 Beans, dry 1,000 Whea Whe Wheat Wh h at Te Cowpea Tea orgh orgh Sorghum So hum hum 500 Potato ot to Po to Sugar Sugar ssava Suga Cas Cassava Cass c r cane 0 Ba Banana anana Ban 1990 1992 Tomato T To t Rice Rice, paddy e, pad 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. TABLE 2.2. LIVESTOCK POPULATION IN KENYA, 2009 AND 2012 Chickens Year Cattle Sheep Goats Camels Pigs Indigenous Commercial 2009 National Census data 17,467,774 17,129,606 27,740,153 2,971,111 334,689 25,756,487 6,071,042 2009 FAO data 17,467,800 9,903,300 13,872,300 947,200 334,689 31,827,000 Percentage (%) difference — –42% −50% −68% — — 2012 FAO data 19,129,800 18,171,000 29,409,100 3,065,400 354,600 32,865,000 Source: FAOSTAT, 2009 Kenya Population and Housing Census. Note: —, not available. TABLE 2.3. TRENDS IN CROP PRODUCTION, 1990–2012a Area Production Yield PRODUCTION TRENDS Food Change Change Change Crops (%) (%) (%) Crop production overall has grown steadily (see figure 2.2), with an average annual increase in the crop production Bananas 1.4 11.7 88.7 index of 6.2 percent from 1990 to 2012. This growth was Beans, dry 55.6 22.2 −23.0 largely driven by area expansion, with the total cultivated Coffee 2.8 −61.2 −62.3 area increasing by 34.7 percent, or from 4.19 million ha Cowpea 273.8 236.8 123.2 in 1990 to 5.65 million ha in 2012 (FAOSTAT). Table Maize 39.6 17.3 −16.4 2.3 summarizes changes in area cropped, crop produc- Potatob 39.3 23.6 −11.7 tion, and yields for key food and cash crops. The area Rice, 58.9 60.8 −3.6 paddy cultivated for food crops increased from as little as 10.7 Sorghum 51.5 9.6 −29.6 percent (bananas) to as much as 274 percent (cowpea), Sugarcane 49.4 23.6 −17.5 with the exception of wheat, for which the area cultivated Tea 96.4 114.8 9.9 decreased slightly. In the case of industrial crops, tea cul- Tomatoes 110.3 399.5 142.6 tivation expanded the most rapidly, eclipsing that of cof- Wheat –6.4 44.6 51.6 fee, which remained virtually unchanged. Yields for some food crops increased, some considerably (cowpea, toma- Source: FAOSTAT. a Five-year average, 1986–90 versus 2008–12. toes), while that of five others declined, by more than 10 b For potato, Ministry of Agriculture, covering period 1990–2006, 5-year average, percent, with the exception of rice. 1986–90 versus 2002–06. 10 Agriculture Global Practice Technical Assistance Paper FIGURE 2.3. FOOD CROP PRODUCTION (thousand MT), 1990–2012 4,000 Maize Wheat Potato Beans Rice, paddy 3,500 3,000 2,500 2,000 1,500 1,000 500 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. FOOD CROPS 70 percent of total maize output in Kenya on an aver- Kenya’s principal food staples are maize, wheat, Irish age of 2 hectares of land. They retain 60 percent of their potatoes, and dry beans. Rice is an increasingly impor- output for home consumption and contribute roughly tant food staple, particularly among urban households. a third to total marketed surplus. Almost three in five Production growth for these crops was notably modest (58 percent) smallholder maize producers are net buyers in recent decades. In fact, according to available data,14 of maize (Kirimi et al. 2011). The remaining 30 percent potato is the only food staple for which the production of marketed maize is produced by a relatively small num- increase exceeded the increase in the population during ber of commercial farmers, who operate over 20 hectares the period 1990–2012. The area cropped to potato nearly on average and contribute 30 percent of total production. doubled, and yields were more than four times higher in recent years than in the earlier period. During the same Kenya boasts relatively high levels of adoption of hybrid period, production of Kenya’s other principal food crops seeds (up to 80 percent), but a much smaller percentage increased 17.3 (maize), 44.6 (wheat), and 22.2 (dry beans) of farmers regularly use fertilizer (41 percent). Commer- (figure 2.3). During the same period, the Kenyan popula- cial farmers make good use of hybrid seeds as well as fer- tion grew approximately 84 percent. tilizers and mechanization to attain high yields (Smale and Olwande 2014). Yields among smallholders, who generally Maize: According to FAO, maize production accounts forego investments in fertilizers and other improved produc- for approximately four-fifths (80.3 percent) by volume tion practices, are significantly lower (1.6 tons/ha in 2012). of Kenya’s total grain output. It is also the fourth most important commodity (after milk, potato, and cattle meat) Maize production in Kenya, as elsewhere, is mainly by value. Kenya produces between 37 and 40 million bags dependent on rainfall. As such, it is vulnerable to drought per year against a national requirement of approximately and to year-to-year yield fluctuations. Average maize 42 million bags.15 Production is dominated by an esti- yields increased about 10 percent from the early 1980s to mated 3 million smallholder farmers who produce roughly the mid-1990s, but have been declining since. Increases in production since 1990 have been driven by a nearly 40 percent growth in land area under cultivation. Much 14 The analysis highlights some notable discrepancies in both national and FAO of this expansion has been into marginal areas, where data. According to data from the MoALF, potato yields and production in 2007 jumped by 215 percent and 255 percent, respectively, compared to the previous soils and rainfall are less favorable to maize production. year, while area harvested increased by only 12 percent. According to official La Rovere et al. (2014) estimate that nearly one-fifth data, annual output has since remained well above historical averages. During (19.5 percent) of Kenya’s maize production takes place in the six-year period 2007–12, average output was 2.5 MT versus 0.96 MT dur- areas with high rainfall variability, rated with a probability ing the period 2001–06. Due to these inconsistencies, the analysis considers only the period 1980–2006 for potato production. of a failed season (PFS) of between 40 and 100 percent. 15 Maize is packed and marketed in 90-kg bags. Average yield in these areas is 1.08 tons/ha versus the Kenya: Agricultural Sector Risk Assessment 11 TABLE 2.4. VALUE OF HORTICULTURAL PRODUCTIONa Area Quantity Value Share by Value (Thousand has) (Thousand tons) (K Sh Million) (%) Vegetables 297 5109 95407 47 Fruits 168 3618 57582 28 Flowers 4 378 42086 21 Nuts 98 166 5524 2.7 Maps 9 57 1804 <1 Total 576 9328 202403 100 Source: Directorate of Crops. a Average 2010–12. national average of 1.62 tons/ha. Another one-quarter rainfed conditions, according to the National Irrigation (26.1 percent) of production is grown in areas rated with Board (NIB). Rice production is expected to increase a PFS of 20–40 percent. These trends have contributed to in response to ongoing GoK initiatives to rehabilitate higher levels of production variability, further amplifying and expand national irrigation schemes and growing Kenya’s structural deficit in maize. Production is also con- adoption of New Rice for Africa (NERICA), a rela- strained by underlying drawbacks such as soil acidification tively new, high-yielding seed variety (USDA 2013). due to year-in, year-out usage of diammonium phosphate Despite anticipated productivity gains, however, Kenya (DAP) fertilizer (USDA 2014) and a general decline in soil will continue to rely heavily on imports given expected fertility. Much emphasis has been placed on the use of demand growth. purchased inputs such as fertilizer and improved seeds, but adoption has not been sufficient to maintain the high yields achieved 30 years ago. HORTICULTURE CROPS Comprising a range of product categories including veg- Wheat: After maize, wheat and rice are Kenya’s most etables, fruits, flowers, nuts, and herbs/spices, Kenya’s important cereal crops. Wheat is predominantly grown in horticultural subsector continues to expand. Among these areas above 1,500 meters in the south and upper Rift Val- categories, vegetable production is the most important ley (e.g., Narok, Nakuru, Uasin Gishu Counties) and in in terms of share of total agriculture output by value Meru County in Eastern Province. Traditionally, Kenyan (38 percent in 2012), followed by fruits (22 percent) and wheat has been grown by large- and medium-scale com- cut flowers (18 percent). The subsector directly and indi- mercial farms using capital-intensive technology such as rectly employs an estimated 4 million people and makes tractors, tillage equipment, and combines. Wheat is the a substantial contribution to household food needs. It also only crop for which area under cultivation has dropped contributes substantially to Kenya’s agricultural export in recent decades (table 2.3), and yield has become more earnings. variable. This trend may be partly due to an ongoing shift in the epicenter of production away from large farms in Vegetables account for nearly half (47 percent) of total Upper Rift Valley to smaller-scale production in Narok production value (table 2.4). The leading vegetables by County. Wheat stem rust, poor yields, the high cost of production volume and value are Irish potatoes, tomatoes, farm inputs, and the shift in the 1990s toward more lib- and cabbage, all of which are widely consumed by rural eralized markets are also likely to be among contributing and urban households. The bulk of vegetables are pro- factors (Chemonics 2010; FAO 2013a). duced by smallholder farmers (estimated at 1.8 million). Vegetables are grown in a wider range of areas across the Rice: Irrigation schemes grow about 95 percent of all country than any other horticultural subgroup (World rice produced in Kenya while the rest is grown under Bank 2012). 12 Agriculture Global Practice Technical Assistance Paper Irish potato: In Kenya, Irish potato is the second most security and rural incomes. Farming households consume important food item after maize, with its importance about 24 percent of total output; the rest (76 percent) is growing along with urbanization. It is grown mainly sold to markets (World Bank 2012). by small-scale farmers as a cash and food crop, and is therefore important for rural income and food security. INDUST RIAL CROPS Potatoes are typically produced under rainfed condi- Tea, coffee, sugarcane, and cut flowers are among Kenya’s tions during two seasons. Farmers intercrop potatoes with principal cash crops. Among these, tea is by far the most maize and beans, and some plant potatoes after maize, important in terms of Kenya’s agricultural export earnings. wheat, or barley. In places like Meru, Kiambu, and Nyeri counties, where average farm sizes are smaller than 1 hec- Tea: At roughly 370,000 tons per year, Kenya stands as tare, farmers grow potatoes on up to 40 percent of their the world’s third largest tea producer after China and cropland, without rotation, which favors the emergence India. The highland tea-growing regions on either side of pests and diseases (World Bank 2012). According to of the Great Rift Valley are endowed with the ideal cli- MoALF statistics, potato production has grown by nearly mate for tea production. Production goes on year-round, 260 percent since 1990, while yields have more than dou- with two main peak seasons between March and June bled (115 percent). However, the expansion of crop area and October and December, coinciding with the rainy and yield is hampered by insufficient availability of high- seasons. Kenyan tea is grown without the use of insecti- quality planting seed. cides or herbicides because at 1,500–2,700 meters above sea level, the growing conditions act as a natural deter- Banana: Fruit production (28 percent value share in rent to pests. 2012) is second only to vegetable production in terms of the total value of horticultural production in Kenya. Tea production in Kenya has grown steadily over the Within the fruit segment, banana is Kenya’s most impor- most recent decade mainly because of expansion in land tant product, representing 37.6 percent of the total value area under cultivation (table 2.3). Kenyan tea is produced of domestic fruit production (MoA 2013). The crop is under two distinct production systems: smallholder pro- mainly grown by smallholder farmers under rainfed duction and commercial production by vertically inte- conditions. According to Africa Harvest, approximately grated multinationals. The latter benefit from higher 390,000 banana farmers operate in Kenya, of which yields but lower-quality output due to more extensive use approximately 84 percent are smallholder farmers (culti- of machinery for harvesting. According to the Tea Board vating <0.2 hectares). Relatively affordable for the average of Kenya (TBK), the smallholder sector is growing in rural and urban household, bananas are widely consumed importance and today accounts for roughly three-fifths across Kenya. The crop is important in terms of both food (59 percent in 2012) of national tea production. A state FIGURE 2.4. INDUSTRIAL CROP PRODUCTION (thousand tons), 1990–2012 450 Tea Coffee Sugar cane 7,000 400 6,000 350 5,000 300 Coffee/Tea Sugarcane 250 4,000 200 3,000 150 2,000 100 1,000 50 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. Kenya: Agricultural Sector Risk Assessment 13 FIGURE 2.5. COFFEE PRODUCTION (tons), 1980–2012 90,000 Estate (tons) Co-operative (tons) 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 – 1 6 1 6 1 6 1 /8 /8 /9 /9 /0 /0 /1 80 85 90 95 00 05 10 Source: CBK. corporation prior to 2000, the Kenya Tea Development $540 million to the country’s GDP.16 It employs more than Agency (KTDA) is now a farmer-owned limited liability 250,000 smallholder farmers who supply over 92 percent company that procures, processes, and markets all small- of the sugarcane processed by nearly a dozen domestic holder production in the country. It manages 67 process- sugar mills. The remainder is produced by factory-owned ing factories serving over 600,000 growers organized in nucleus estates (KSB 2010; KSI 2009). Sugar production Savings and Credit Cooperatives (SACCOs). The planta- is concentrated in four major areas, primarily located in tion subsector operates 39 tea factories and employs about southern and southwestern Kenya. Increases in produc- 33,000 outgrowers. tion during the most recent decade were largely the result of increases in total land planted while yields remained Coffee: Coffee remains important to Kenya’s agricul- stagnant. Widespread use of poor-quality sugarcane tural economy, but its importance is waning (figure 2.4). varieties, poor agricultural and land management prac- Since production peaked in 1988 at nearly 128,000 tons, tices, and delayed harvesting of mature sugarcane (due to yields and output have dropped by nearly half. Among weather and/or transportation problems) contributed to contributing factors are Kenya’s aging tree stock (with poor yields over time. high susceptibility to plant diseases) and declines in world coffee prices during 1986–1992 and 1998–2002. These Cut flowers: Kenya’s floriculture industry was worth an trends have had a substantial impact, particularly on estimated $490 million in 2012. Cut flowers are predomi- smallholders. Figure 2.5 shows the performance of estates nantly cultivated under modern farming systems and are and smallholders during 1980–2012. The latter suffered produced for export markets. Roughly 160 flower growers a drastic reduction in output, from around 70,000 tons exist in Kenya. The majority of producers are medium- to during the mid-1980s (before the coffee price crisis) to large-scale agribusinesses. However, 20–25 of these grow- less than 30,000 tons currently. Over 600,000 smallholder ers are large to very large commercial enterprises that producers are organized into about 550 cooperatives together account for roughly 75 percent of total flower and about 3,300 large-scale, vertically integrated coffee exports. Such operations are highly capital intensive, best estates. Smallholders account for 75 percent of the land characterized by their managerial and marketing sophis- under coffee production but only slightly over half of pro- tication and sizable investments in advanced technology duction, according to the Coffee Board of Kenya (CBK). and cultivation techniques (Hortiwise 2012). The leading Average yields on the estates are nearly 1.5 times higher counties in horticultural production are Kiambu, Nakuro, due to their more intensive use of fertilizers, pesticides, Meru, Nyandaru, Murang’a, Bungoma, and Makueni, herbicides, and fungicides as well as irrigation. which together account for more than 57 percent of total Sugarcane: Kenya’s sugar industry supports an esti- 16 See “Kenya: Poaching sugarcane” by Katrina Manson, Financial Times, mated 2 million people and contributes an estimated January 21, 2014. 14 Agriculture Global Practice Technical Assistance Paper output value, with the first three counties accounting for Emerging from a significant downturn in the 1990s, Ken- more than 30 percent (HCDA 2012). ya’s milk production sector is growing again. Valued at $800 million, it contributes 7 percent to national GDP and LIVESTOCK PRODUCTION 19 percent of agricultural GDP (KNBS 2009 in Macha- ria 2013). Over 1 million households produce milk, with SYSTEMS 80 percent of the 4 billion liters produced by small-scale Several different livestock production systems are com- farmers. The sector provides more than 850,000 jobs mon across Kenya, most notably extensive pastoralism, (FAO 2011). Government services were relatively effective mixed crop/livestock farming systems, and intensive poul- in regulating production and trade until mid-1980s, but try, pig, and dairy production. Each system faces different milk production started failing in the 1990s and collapsed constraints and risks. Vulnerability to risks is considered in early 2000 due to corruption in the management of the greater within extensive systems than within intensive cooperative sector. It was reinvigorated after being taken ones for myriad reasons—a lower level for capacity to over by the new Kenya Cooperative Creameries (KCC) in mitigate risks among pastoralist communities, punctu- 2004, and more than 30 registered processors are now in ated by declining mobility and unpredictable access to operation (Macharia 2013). and availability of water, pastureland, and other factors of production. Thus, analysis of risks to extensive systems was prioritized in this study. VARIABILITY IN CROP PRODUCTION Approximately 50–70 percent of the country’s live- An analysis of production variability suggests that several stock is produced under extensive systems, mostly in of Kenya’s main crops exhibit moderate to high levels of the ASALs, where the subsector accounts for roughly interannual variation (see table 2.5). These crops include nine-tenths of employment and animals provide the potatoes, rice paddy, coffee, bananas, sorghum, and cow- vast majority of household income. This system mostly pea. Fluctuations in yield rather than area planted across comprises indigenous races of cattle, camels, sheep, and the time series largely explain notable production variabil- goats, which graze or browse natural forage. Land is usu- ity for potatoes and coffee, while fluctuations in both area ally communally owned, although private or group ranch planted and yield account for observed variability in dry systems also exist. Some feed supplementation occurs bean and cowpea production. where there is mixed farming or irrigation using crop residues and weeds. Mixed farming, a system involving various food and cash crops integrated with a few cattle TABLE 2.5. COEFFICIENTS OF VARIATION or sheep and goats, stretches from the southern parts of FOR CROP PRODUCTION, the ASALs into Kenya’s high-potential agricultural pro- 1980–2012 duction areas. In highland areas, the animals are mostly Production Area Yield dairy cattle and sometimes pigs. These systems are often accompanied by medium- or small-scale “backyard” Bananas 0.30 0.14 0.24 Beans, dry 0.28 0.27 0.23 poultry production. Coffee 0.35 0.10 0.41 Intensive livestock production systems mostly consist of Cowpea 0.37 0.25 0.26 commercial poultry and pig production and a more lim- Maize 0.19 0.17 0.13 ited number of dairy farms. Commercial enterprises using Potatoes 0.77 0.30 0.55 intensive systems, especially for poultry, are normally Rice, paddy 0.38 0.32 0.20 established after undertaking their own risk assessments Sorghum 0.32 0.27 0.24 and are operated under high sanitary and biosecurity lev- Sugarcane 0.12 0.19 0.16 els with in-built mechanisms to avert or avoid risk. As a Tea 0.38 0.27 0.18 result, such systems were not considered as part of this Wheat 0.26 0.13 0.22 risk assessment. Source: FAOSTAT; MoALF. Kenya: Agricultural Sector Risk Assessment 15 FIGURE 2.6. CEREAL PRODUCTION TRENDS (thousand tons), 1990–2012 4,000 Maize Wheat Sorghum Rice, paddy Beans 3,500 3,000 2,500 2,000 1,500 1,000 500 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. Figure 2.6. illustrates historical trends in production for TABLE 2.6. CEREAL SUPPLY/DEMAND five of Kenya’s major staple crops. Figures F.1 through BALANCE (thousand tons), F.6 in Appendix F show individual production trends for 2013/14 six crops (i.e., maize, wheat, dry beans, tea, coffee, and sugarcane). Higher levels of variability in maize and wheat Total Wheat Rice Maize cereals production over the last decade are apparent, as is a cor- relation in some years with extreme drops in production, Cereal supply (‘000 tons) suggesting covariance of shocks. In both cases, fluctua- Previous year 442 122 3,922 4,486 production tions in yields are the strongest determinant of output Previous five years 356 83 3,300 3,739 variability from one year to the next. Historical patterns average production for dry bean production suggest that both changes in yield Previous year imports 1,000 405 506 1,911 and acreage planted have a strong influence on output. Domestic 405 85 3,616 4,106 Among cash crops, variations in tea and coffee produc- availability tion are most directly affected by fluctuations in yields. 2013 Production 390 190 3,489 3,994 For sugarcane production, changes in both yield and area Possible stock 15 10 127 152 planted appear to have a strong influence. drawdown Utilization 1,505 505 4,456 6,466 FOOD SUPPLY AND DEMAND Food use 1,205 467 3,850 5,522 Nonfood use 300 25 606 931 Kenya currently suffers from a structural deficit in the Imports 1,100 420 840 2,360 production of key staples, including maize, wheat, and requirement rice (table 2.6). Over the last decade, annual imports for maize in particular fluctuated significantly, accounting Source: FAO 2014. for between 6.2 and 32 percent of consumption depend- ing on the year, with a 10-year average of 14.5 percent, or 624,000 tons (FAOSTAT). Domestic production over yan households, with per capita consumption about 99 the last decade accounted on average for roughly half kg/year (FAOSTAT). It also accounts for roughly 40 to (48.5 percent) and one-fifth (21 percent) of wheat and rice 50 percent of household food expenditures. While grow- consumption, respectively. ing, maize production has not kept pace with population growth over the last 30 years. During the period 1980– Maize is the principal food staple in Kenya. It accounts 2012, production increased roughly 60 percent while the for nearly one-third of the calories consumed by Ken- population grew by 153 percent. To address the deficit, 16 Agriculture Global Practice Technical Assistance Paper FIGURE 2.7. MAIZE PRODUCTION VERSUS DEMAND (thousand MT), 2003/04–2013/14 4,500 Domestic production ('000 MT) Domestic utilization ('000 MT) 4,000 3,500 3,000 2,500 2,000 2003/4 2004/5 2005/6 2006/7 2007/8 2008/9 2009/10 2010/11 2011/12 2012/13 2013/14 Source: FAOSTAT. the country continues to depend heavily on imports. In Rising consumer demand for wheat is largely fueled by a 2012–13, for example, Kenya imported 425,000 tons of growing preference among burgeoning urban consumers maize for commercial and relief purposes. Maize con- who view wheat products as a convenience food. Conse- sumption will continue to grow despite efforts to diver- quently, wheat imports are expected to remain above 1 sify to other foods. Limited volumes of lower-grade maize million metric tons (MT) annually. The Russian Federa- go into the livestock feed industry. Deficits are offset by tion, Ukraine, Pakistan, Brazil, and Argentina remain the imports from within the East African Community (EAC), largest suppliers to the Kenyan wheat market. and imports from outside the EAC are subject to a steep external tariff (currently at 50 percent ad varolem). Rice is the third most important cereal food crop after maize and wheat. Irrigation schemes grow about three- Projections by FAO for current-year imports were 800,000 quarters (74 percent) of all rice produced in Kenya (FAO tons based on expected domestic output of 3.2 million 2012); the rest grows under rainfed conditions. The NIB tons and strong, continued growth in demand (figure 2.7). estimates per capita rice consumption will rise to 11 kg USDA (2014) forecasts were considerably lower at 2.8 mil- by 2015, up from 7 kg in 2013. The Ministry estimates lion tons. The decline in production is attributed to poor annual consumption is increasing at a rate of 12 per- yields due to below average rainfall. It is also due in part to cent compared to 4 percent for wheat and 1 percent for delayed and inadequate supply of subsidized fertilizers17 maize. It is expected to more than double to 495,000 tons and certified seeds, the spread of the MLND, widespread in 2014–15 (October 2014–September 2015) from about infestation by the parasitic weed Striga, and declining soil 237,000 tons consumed in MY 2004–05 (Gitonga and fertility. Shortages have also been aggravated by increased Snipes 2014). postharvest losses linked to poor drying and storage prac- tices and early sales of green maize. The former contrib- This trend is attributed to a progressive change in eating utes to high incidences of aflatoxin contamination. habits, particularly among Kenya’s urban households. Pakistan, Vietnam, Thailand, and India supply Kenya Growing consumption of wheat and wheat-based prod- with most of the rice imports. Tanzania supplies a sub- ucts far outstrips domestic production. Kenya’s wheat stantial amount through unrecorded cross-border trade. imports grew at an annual average rate of 13.25 percent The ad valorem tariff for rice coming from outside the between 2003/04 and 2012/13. In 2012/13, imports met region currently stands at 35 percent but can be as high roughly 44.2 percent of national wheat requirements. as 75 percent. Consumption of maize and wheat is expected to increase 17 Since 2007/08–2011/12, the GoK has allocated more than $117 million for distribution of subsidized fertilizers to smallholders in the Rift Valley and west- because of population growth, increased urbanization, ern Kenya via the National Cereals and Producer Board (NCPB). and growth in the food service sector. Imports to Kenya Kenya: Agricultural Sector Risk Assessment 17 are subject to external tariffs that range from 10 percent 60 percent of Kenya’s tea exports go to only three consum- for wheat, to 35 percent for rice, to 50 percent for maize. ing countries (Pakistan, Egypt, and the United Kingdom). Minimal exports of these staple commodities occur Kenyan horticultural exports to the European market through cross-border trade. have dropped in recent years due to the economic crisis in the European Union (EU) and difficulties among Kenya’s Although consumption is relatively diversified, production smallholder farmers in adhering to strict EU regulations is much more so. Maize, wheat, sugar, milk, and palm oil over agrochemical residues. together make up nearly two-thirds of daily per capita consumption. Maize accounts for the largest share of FOOD CROPS total daily staple food intake (65 percent) and total caloric Due to Kenya’s open trade regime and highly integrated intake (32 percent), with per capita consumption of markets, domestic prices in Kenya for agricultural com- about 99 kg/year. Wheat accounts for another 9 percent. modities, including major staple foods, are relatively Sugar, milk, and palm oil each contribute approximately sensitive to both internal and exogenous pressures and 7–8 percent of calories and round out the top five con- shifts in supply and demand. The following analysis of tributors to calorie consumption. price trends for key commodities (figure 2.8.) was based on time-series data of producer prices in local currency In Kenya, domestic beef consumption has more than (K Sh) for the period 1992–2011 doubled over the past two decades, and Kenyans’ rate of milk consumption, one of the highest in the world for Overall, domestic cereal prices over the last two decades developing countries (100 kg/capita/year), is still growing. were characterized by moderate levels of volatility. How- Such local demand, and a growing export market for live ever, prices became more volatile in recent years due to animals and products, is leading to increased intensifica- both external and internal dynamics. First, the surge in tion of production. Intensification itself, particularly in a international food prices in 2006–07, and in 2009–11 farming sector largely dominated by arid areas, can lead (as measured by FAO’s Food Price Index), had a notable further to increased risk. impact on domestic prices for rice and sorghum in par- ticular. Domestic unrest surrounding the 2007 elections AGRICULTURAL MARKETS likely contributed to increased levels of volatility. A ton of sorghum cost nearly 86 percent more in 2009 than AND PRICE TRENDS just two years earlier. A ton of rice paddy was nearly Agricultural markets in Kenya are highly integrated due 71 percent higher, and maize, 53 percent higher. Domes- to the country’s relatively well-developed road and com- tic prices for wheat followed similar patterns, with more munications networks and sea ports and airports and its pronounced variability in recent years. Wheat and rice open trade regime. Kenya benefits from robust and grow- are both routinely imported from world markets, and ing trade with its regional neighbors, especially within the tariffs are generally effective in keeping domestic prices framework of the EAC, and steady international demand high for producers. for some of its key exports. Tea and coffee have tradition- ally been Kenya’s top two agricultural export commodi- Maize and sorghum, on the other hand, are typically ties. In recent years, cut flowers have overtaken coffee to imported duty-free from countries within the EAC and become Kenya’s second most valuable export crop. Dur- COMESA (Common Market for Eastern and Southern ing 2010–12, cut flowers accounted for 8.94 percent of Africa) regions and are only imported from world markets total export value on average, after tea (20.27 percent) and under exceptional circumstances. Consequently, tariffs are coffee (4.46 percent). Vegetables, fruits, and related prod- not always effective in keeping maize prices high for pro- ucts accounted for an additional 4.5 percent (table 2.7). ducers. Given the importance of maize to Kenya’s food The country relies on a limited number of export prod- security, the GoK intervenes in markets to regulate prices ucts and trade partners, which makes Kenyan exports and ensure sufficient surplus stock. In low production vulnerable to external pressures. For example, more than years, the government often suspends the 50 percent tariff 18 Agriculture Global Practice Technical Assistance Paper TABLE 2.7. VALUE OF AGRICULTURAL EXPORTS (US$ thousands), 2010–12 2010 2011 2012 Average share (US$000) (US$000) (US$000) 2010–12 (%) Tea 1,163,630 1,176,308 942,101 20.27 Cut flowers and flower buds for bouquets 396,239 454,349 597,716 8.94 Coffee 207,473 223,509 291,937 4.46 Leguminous vegetables, shelled or unshelled 75,037 152,903 188,834 2.57 Vegetables, fresh or chilled 150,251 57,652 36,696 1.51 Fruit and vegetable juices, unfermented 26,997 27,876 15,588 0.44 Other — — — 61.80 All products 5,169,112 5,853,310 5,169,142 100 Source: International Trade Centre. Note: —, not applicable. FIGURE 2.8. TRENDS IN CEREAL PRICES (K Sh/ton), 1991–2011 90,000 Maize Rice, paddy Sorghum Wheat 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 Source: FAOSTAT. on grain imported from outside the EAC to hold prices local or the export market. They are also influenced by down. During such periods, price volatility increases as the extent of government intervention and participation. imported maize competes with domestic maize. In addi- The vast majority of tea and coffee produced is exported, tion, the National Cereals and Produce Board (NCPB) while virtually 100 percent of the sugarcane produced in purchases maize at fixed prices from large-scale farmers Kenya is refined and consumed domestically. Domestic tea and from some smallholders in a few major surplus zones, and coffee prices are set via the major auctions in Mom- particularly in the Rift Valley. It also distributes subsidized basa and Nairobi, respectively. Prices in both auctions, in fertilizer to smallholder farmers in parts of the Rift Valley turn, are heavily influenced by prevailing prices in external and western Kenya.18 markets. These include tea auction prices in Colombo and Calcutta and other major tea-producing countries, and CASH CROPS the New York “C” contract market for coffee. Since mar- Domestic prices, supply chain governance, marketing, and ket liberalization in the early 1990s, the GoK has assumed other market dynamics for Kenya’s key cash crops vary only a limited regulatory role in domestic tea and coffee depending on whether the end product is destined for the industries through the TBK and the CBK. By compari- son, Kenya’s sugar industry remains highly regulated, with 18 Some evidence shows that overreliance on DAP has contributed to excessive domestic prices directly influenced via import tariffs and soil acidity, and hence, low yields. quota protections. Kenya: Agricultural Sector Risk Assessment 19 FIGURE 2.9. TRENDS IN CASH CROP PRICES (K Sh/ton), 1991–2011 700,000 4,000 Coffee Tea Sugarcane 600,000 3,500 Coffee/Tea (Khs/ton) 3,000 500,000 2,500 Sugarcane 400,000 2,000 300,000 1,500 200,000 1,000 100,000 500 0 0 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 Source: FAOSTAT. Tea: The volume of Kenya’s tea exports increased over problem, as Kenyan coffee production has continued to the last decade, with some relatively modest fluctuations, slide amid a rebound in global prices. Kenya’s aging trees while the average value per kilogram in Kenyan shillings are increasingly susceptible to coffee leaf rust (CLR) and (K Sh) has increased steadily since 2007 (figure 2.9). These coffee berry disease (CBD), serious biological threats that trends led to rising export proceeds for the GoK over necessitate the use of costly fungicides. The rising value of the most recent decade. In 2012, tea exports accounted land due in coffee-producing areas and competition with for nearly one-fifth of Kenya’s total agricultural exports, other crops further contribute to this decline. valued at $942 million. Exports are highly dependent on three markets: Pakistan, Egypt, and the United King- The authority to regulate coffee sales and marketing in dom. Together, these account for more than 65 percent of Kenya has been vested in the CBK. The Kenya Coffee national tea exports. Pakistan alone imports 24 percent of Producers and Traders Association (KCPTA) manages Kenya’s total tea exports. The loss or significant reduction the auction through which nearly all coffee marketed in of demand from one or more of those markets is an ongo- Kenya is sold, with a small proportion sold through pri- ing risk to the industry, as happened during 2005–06 when vate contract arrangements.19 Estates and cooperative tea exports to Pakistan fell drastically. Slowing demand in societies employ one of eight licensed marketing agents Egypt and Pakistan, and globally, and a glut in global pro- to represent them at the coffee auctions. Around 50 duction resulted in weaker prices for Kenyan teas in 2014. licensed coffee dealers purchase coffee from the auction for export. However, a handful of buyers account for the Coffee: Despite the popularity of and strong apprecia- vast majority of transactions. In this situation, these buy- tion for Kenyan coffee globally, Kenya’s coffee industry ers exercise strong market power in maintaining favora- is crumbling under the weight of mismanagement. The ble auction prices, while agents and others are paid on importance of Kenya’s coffee crop as a major export has a fixed fee basis. The result is consistently low farm-gate declined drastically since production and exports hit their prices that discourage on-farm investments. Figure 2.10 peak during the mid-1980s. Today, it accounts for less compares Nairobi auction prices and internal prices with than 5 percent by value of Kenya’s agricultural exports. those received by the Roret Farmers Cooperative Society An important factor explaining the decline in coffee 19 The Coffee Act (2001) was amended through the Finance Act of 2005 to production and exports is the decline in world coffee allow for direct marketing of green coffee beans by producers. The direct sales prices between 1986 and 1992. Prices recovered partially window, commonly referred to as the “second window,” allows estate grow- ers, cooperative societies, and cooperative unions to self-market their coffee and between 1993 and 1997, but declined again between access foreign markets. However, since its inception in 2005, negligible volumes 1998 and 2002. After that, they increased consistently have been sold outside the auction due to stringent regulations and bank guar- until 2014. The price decline only partly explains the antee requirements. 20 Agriculture Global Practice Technical Assistance Paper FIGURE 2.10. COFFEE PRICE COMPARISON ($/kg), 2005–13 8 Nairobi coffee exchange 7 International prices 6 Roret farmers co-operative society ltd. 5 4 3 2 1 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: Nairobi Coffee Exchange; International Coffee Organization; Roret Farmers Cooperative Society. from 2005 to 2013. Farmers were paid about one-tenth of of which are privately owned. The largest mill, Mumias, the Nairobi auction prices on average. is government owned. Outgrowers sell their product to sugar mills that process the sugarcane into raw sugar, These and the other dynamics outlined above have con- which is then sold to the local food industry and house- tributed to the decline of Kenya’s coffee industry. The holds through wholesalers and retailers. Imported sugar smallholder subsector, in particular, faces significant dis- is transported to major wholesale markets and retailers, ease-related losses and limited incentives to invest in cof- where it competes directly with locally produced sugar. fee production under Kenya’s current marketing system. Kenya’s current sugar deficit is addressed through imports, Sugarcane: Kenya’s sugar industry is closely linked to both formal and illicit. Significant volumes of refined the government and is regulated via the Kenya Sugar sugar from outside COMESA countries are reportedly Board (KSB). As a member of the COMESA Free Trade regularly smuggled into Kenya and can at times cause Area, Kenya is obligated to allow duty- and quota-free significant distortions in the domestic market. Although access for sugar and other products for member states. imports are regulated through quotas and tariffs, insuf- Since 2000, however, the country has maintained a mar- ficient administration and high local retail prices allow ket-access safeguard that was extended until March 2015. importer “syndicates” to obtain profit margins that can This has allowed a range of protective measures to help be more than double those of local producers (Millen- ease the sugar industry’s transition to full market liber- nium Cities Initiative [MCI] 2008). Poor administration alization. Measures include tariffs and quotas20 under the of the quota system in years past resulted in heavy losses COMESA quota protection protocol that are applied to to processors unable to compete with significant volumes imports, effectively barring open competition between of cheaper imports, as happened in 2012. Kenyan and COMESA sugar producers. These protec- tions have kept domestic prices artificially high, benefit- Horticulture: Despite rapid growth in recent years, ting local producers but making raw sugar and sugar fruit and vegetable exports currently account for less than products more expensive for consumers. 5 percent of the total value of Kenya’s agricultural exports. The EU is Kenya’s biggest market for vegetables, import- The sector consists of more than 250,000 smallholder ing about 90 percent of all vegetables destined for export. farmers, who supply over 90 percent of the sugarcane More than 90 percent of all fruit and vegetable produc- processed by sugar companies, while the remainder is sup- tion is consumed domestically, either on-farm or through plied by factory-owned nucleus estates and 11 mills, six domestic markets. Price trends across product segments were relatively stable until 2007, when the food crisis and other events led to increasing levels of volatility (figure 20 Under the Safeguard Clause, Kenya was allowed to impose (1) a quota of 200,000 tons annually on sugar imported from COMESA countries and (2) a 2.11.). In recent years, Kenya’s small-scale farmers in par- tariff of 120 percent for any amount that exceeds the quota amount. ticular have been hit by rising production costs and the Kenya: Agricultural Sector Risk Assessment 21 FIGURE 2.11. TRENDS IN PRODUCER PRICES (K Sh/ton) FOR FRUITS AND VEGETABLES, 1991–2011 60,000 Bananas Cabbages Potatoes Tomatoes 50,000 40,000 30,000 20,000 10,000 0 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: FAOSTAT. EU’s stringent food safety regulations concerning agro- Kenya produces approximately 410,000 tons of beef chemical residues. EU demand has also slumped, partly annually, worth approximately $1.1 billion, amid strong as a result of the Euro-zone crisis. According to the Fresh growth in consumption, which doubled over the past Produce Exporters Association of Kenya (FPEAK), over- two decades. Goats and sheep play a key role in the food all exports of vegetables declined by 2.6 percent between security and incomes of pastoral households due to their 2011 and 2012, from $379 million to $369 million. Bean short-generation intervals, high adaptability, and versa- sales dropped by 25 percent in January 2013 compared to tile feeding habits. The country is home to an estimated January 2012, according to FPEAK. 29.4 million goats and 18.2 million sheep, which produce about 50,000 tons of chevon and mutton annually, worth LIVESTOCK PRODUCTION an estimated $128 million. As noted earlier, a comprehensive livestock census has not Milk production in Kenya is largely driven by the infor- been done since 1969, so actual animal population figures mal milk sector. An estimated 800,000 small-scale farm- are not known.21 According to FAO data, total cattle stocks ers and 350,000 small-scale milk vendors dominate the were an estimated 19.1 million heads in 2012 (table 2.8). milk marketing chain. Following more than a decade of decline, the industry has rebounded in recent years since 21 The 2009 Kenya Population and Housing Census encompassed an account- the GoK restructured it. New dairy regulations since 2004 ing of livestock ownership and stock at household level. enabled small-scale milk traders to gain licenses and to TABLE 2.8. LIVESTOCK POPULATIONS IN KENYA Year Annual Growth Rate (%) 1980 1990 2000 2010 1980–1990 1990–2000 2000–2010 Livestock Group (‘000s) Cattle 10,000 1,3793 11,444 17,862 3.79 −1.70 5.61 Sheep 5,000 9,050 7,939 17,562 8.10 −1.23 12.12 Goats 8,000 10,186 10,004 28,174 2.73 −0.18 18.16 Pigs 74 128 311 347 7.32 14.23 1.19 Camels 608 850 718 3,031 3.98 −1.56 32.24 Poultry 16,400 25,228 26,291 30,398 5.38 0.42 1.56 Source: FAOSTAT. 22 Agriculture Global Practice Technical Assistance Paper enroll for training in milk handling, processing, and mar- sufficient availability at stable, affordable prices for Ken- keting. In 2012, milk production was roughly 5.1 billion yan consumers, especially in Nairobi, the major deficit liters, with an estimated value of $1.16 million. Although market. NCPB is the primary implementing agency. The Kenya remains self-sufficient in milk, output has declined board procures and maintains a strategic grain reserve in recent years. Estimated annual per capita milk con- on the government’s behalf to buffer against food short- sumption ranges from 19 kg in rural areas to 125 kg in ages, and facilitates the procurement, storage, mainte- urban areas (FAO 2011). nance, and distribution of food aid to deficit areas under the National Famine Relief Program. The GoK has also Kenya has an estimated 31.8 million chickens, 80.2 per- provided input subsidies on a continuous basis, mainly cent of which are indigenous while 19.8 percent are com- for fertilizer, in the form of direct payment to farmers, mercial layers and broilers (GoK 2010a). Other poultry free distribution, or voucher programs. The longer-term types (duck, turkey, pigeon, ostrich, guinea fowl, and policy focuses on increasing production through upgrades quail) are becoming increasingly important. Annually, to research and extension. the country produces about 21,600 tons of poultry meat worth $39.6 million and 1.3 billion eggs worth approxi- CONSTRAINTS TO mately $110 million. AGRICULTURAL GROWTH Agricultural production in Kenya is handicapped by a FOOD SECURITY range of factors that limit producers’ ability to invest in A series of poor cropping seasons in recent years has con- their farming enterprise and raise output. These con- tributed to deterioration in Kenya’s national food security straints are well documented elsewhere and overcom- status. The number of Kenyans requiring food assistance ing them has long been the focus of the GoK’s sector rose from 650,000 in late 2007 to almost 3.8 million in late development planning and investments. They include 2009 and early 2010 (GoK 2011). In July 2011, an esti- the decreasing size of landholdings, limited access to pro- mated 2.4 million Kenyans required food and nonfood ductivity-enhancing technology (including affordable and aid assistance (KFSSG 2011). Aid agencies, the United timely inputs and input credit), declining soil health, weak Nations, and the GoK indicated that more than 3.5 mil- extension services and low technology adoption, and poor lion Kenyans faced starvation as the country struggled smallholder access to markets. These and other constraints with what is believed to be its worst drought in 60 years. In are notable within the context of vulnerability to risks and response, GoK policies and interventions focused increas- agricultural risk management. While dampening income ingly on stop-gap emergency measures such as safety net growth and agricultural supply chains’ competitiveness, programs (e.g., food distribution, food for work), short-term these constraints can also amplify the impacts of adverse export bans or import tariff reductions, and input subsidies. shocks (e.g., drought, disease outbreak) by weakening farmers’ and other stakeholders’ ability to manage risk Agricultural policies in recent years are best character- events and recover in their aftermath. It is worth noting ized by strong GoK presence and control of produce and here that many interventions that address risks can have input prices for producers and a sustained focus on stimu- positive spillovers in addressing growth constraints. For lating productivity. Perhaps the most prominent example example, improved soil and water management used to is the price stabilization and producer support prices for mitigate drought risks can catalyze productivity and gains maize. The major policy objective for maize is ensuring in farmers’ income. Kenya: Agricultural Sector Risk Assessment 23 CHAPTER THREE AGRICULTURE SECTOR RISKS The main sources of agricultural risk in Kenya are reviewed in this chapter. These include production risks, market risks, 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 Based on analysis of available quantitative and qualitative data, the most common risks to agricultural production in Kenya are drought, flooding, and crop and live- stock pest and disease outbreaks. The incidence of these and other adverse events is indicated in figure 3.1, based on reports of adverse events for the period 1980– 2012. Drought emerges as by far the most common source of production shocks, fol- lowed by floods, which have a much lower impact on crop and livestock production. Related risk events (e.g., pest/disease outbreaks, bushfires) may occur in isolation but can also present as multiple, overlapping shocks, with far greater impacts and higher associated losses. DROUGHT An agricultural drought22 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. Table 3.1 and table 3.2 are based on analysis of annual rainfall data collected from 12 weather stations for which consistent and reliable information was available for the period 1981–2011. It is worth noting that these weather stations are well distributed 22 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 periods, a soil moisture deficit during a period 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 particular location at a particular time of year. Drought is thus a relative concept in that suboptimal soil moisture levels and crop yields in one agroclimatic area may be acceptable in another. Kenya: Agricultural Sector Risk Assessment 25 FIGURE 3.1. HISTORICAL TIMELINE OF MAJOR AGRICULTURAL PRODUCTION SHOCKS, 1980–2012 14 Agriculture, value added (annual % growth) 12 10 Violence follows Drought; elections; 8 floods,RF drought, fever 2007 6 2006 4 2 Drought, 0 2011 Commodity –2 La Nina drought, price shock, 1999–2000 2008 –4 Drought El Niño floods; Drought Erratic rains, 1983–84 1.5 m affected; floods, Prolonged –6 1991–93 RV fever, 2002 drought, 1997–98 2008–2009 –8 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT; authors’ notes. across the country, and thus provide a reasonably indica- The combination of frequent severe droughts, high tive footprint of rainfall at the national level. An analysis dependence on rainfed agriculture, and high poverty of standardized cumulative rainfall provides insights into rates among smallholder farmers and pastoralists makes the frequency and severity of rainfall events during the Kenya particularly vulnerable to the effects of droughts. 31-year period. For the purpose of this analysis, drought is A 2010 survey covering the country’s humid, temper- defined as rainfall more than one standard deviation from ate, semiarid, and arid agro-ecological zones found that the mean and extreme drought as rainfall more than two more than 80 percent of all households interviewed had standard deviations from the mean. experienced drought over the last five years, regardless of agro-ecological zone (Bryan et al. 2011). In addition Based on the analysis, the country experienced 13 years to the immediate impacts, drought typically has long- of widespread drought during the period under review; term consequences. It generally takes more than one three of these were categorized as extreme droughts season for farmers to recuperate from seasonal droughts, (1983, 1984, 2005). This equates to a drought event in as resources, including seeds, are not available for the one out of every three years on average. The frequency following, nondrought season. of more widespread and severe droughts has increased since 2000, while less severe drought events have occurred Large numbers of people, mainly in the ASALs, have in other years, impacting one or more regions simultane- personal knowledge of the impacts of drought, which ously. Table 3.2 lists the years during which Kenya experi- have been widely documented (Aklilu and Wekesa 2002; enced a drought event, with details on associated impacts Republic of Kenya 2012; Zwaagstra et al. 2010). In referenced in media reports, GoK assessments, and the Kenya, drought has profound effects on the agriculture literature. It suggests that drought is a nearly constant sector and is by far the biggest risk facing agricultural dynamic that affects Kenya’s agricultural landscape with production, as its effects are typically felt over a wide varying levels of severity. area impacting both crop and livestock production sys- tems. For instance, the drought of 2009 affected nearly The impacts of drought depend on three factors: the one-quarter of the population. Drought conditions frequency of droughts, the degree to which the country also favor the emergence of other risks such as pests depends on rainfed agriculture, and the ability of the pop- or diseases, while weakening plant and animal resist- ulation to prepare for and adapt to drought conditions. ance to such threats. Resulting losses from crop failure 26 Agriculture Global Practice Technical Assistance Paper TABLE 3.1. FREQUENCY OF MAJOR DROUGHT EVENTS IN KENYA, 1981–2011 Year Lodwar Mandera Eldoret Nyahururu Garissa Kisumu Narok Dagoretti Makindu Voi Malindi Mombasa 1981 −1.03 1.14 −0.24 0.28 −0.33 −1.53 −1.16 0.67 0.40 0.92 −0.20 0.20 Drought 1982 3.31 0.50 0.86 1.48 0.42 0.57 0.18 0.65 1.69 0.69 2.52 1.81 Ext Excess 1983 −0.69 −0.87 1.63 −1.47 −1.34 −1.36 −0.50 −0.36 −1.49 −1.32 0.22 0.14 Ext Drought 1984 −1.08 −0.92 −2.18 0.40 0.17 −0.82 −1.50 −2.29 0.53 −0.22 0.67 0.20 Ext Drought 1985 0.01 −0.60 −0.60 −0.39 −1.02 −0.04 0.59 −0.80 −0.33 −0.62 −0.87 −0.63 Drought 1986 −0.32 −0.56 −1.43 0.10 −0.07 0.11 −1.30 −0.20 0.20 −0.08 0.13 0.05 Normal 1987 −0.15 −0.32 0.46 −1.26 −1.01 −0.41 0.69 −0.81 −1.26 −1.20 −0.10 −0.60 Drought 1988 1.40 −0.86 −0.28 0.50 0.16 0.41 0.49 1.28 0.63 0.22 −1.17 −0.24 Normal 1989 0.49 0.50 0.98 1.70 1.54 0.06 1.88 1.20 1.92 −0.45 −0.17 −0.11 Excess 1990 −0.89 0.68 −0.29 0.82 0.74 −1.13 −0.53 0.49 0.98 0.97 −0.80 −0.15 Normal 1991 −0.50 −0.11 −0.24 −0.51 −0.28 −0.52 −1.05 −0.39 −0.45 0.14 −0.32 0.17 Normal 1992 −0.97 −0.52 0.33 0.00 −0.21 −0.67 −1.48 −0.07 0.10 0.18 −1.16 −0.51 Drought 1993 −0.61 1.18 −1.84 −0.61 −0.14 −0.67 0.36 −0.42 −0.56 −0.17 −0.43 −0.81 Normal 1994 −0.51 −0.15 0.59 0.78 0.37 0.92 0.27 0.06 0.93 1.08 1.45 1.42 Excess 1995 −0.93 −0.08 −0.39 −0.99 0.27 0.71 −0.35 −0.19 −0.98 −0.73 −1.08 0.12 Normal 1996 −0.03 −0.92 0.17 −0.62 −1.26 1.15 0.45 −1.40 −0.57 −0.49 −0.71 −0.14 Drought 1997 1.26 4.08 0.79 0.91 3.23 1.60 1.45 1.02 1.07 0.79 2.41 3.51 Ext Excess 1998 −0.55 −0.54 2.42 1.72 1.69 −1.57 0.93 1.79 1.94 2.97 0.85 0.16 Ext Excess 1999 −0.73 −0.37 0.01 0.60 −0.33 0.97 −0.40 −0.11 0.74 −0.70 0.67 0.07 Normal 2000 −0.92 −0.39 −0.99 −0.37 −1.13 −1.07 −1.38 −1.58 −0.30 −0.33 −0.25 −0.28 Drought 2001 −0.75 −1.27 −0.18 0.40 −0.56 0.63 −0.02 1.80 0.53 −0.37 −0.68 −0.68 Drought 2002 0.95 0.18 −0.76 −0.38 0.95 1.72 1.85 0.37 −0.32 1.21 0.06 −0.07 Excess 2003 −0.03 −0.20 −0.56 −1.02 0.21 −0.42 0.37 −0.26 −1.01 −1.94 −0.42 −1.37 Drought 2004 −0.10 −0.20 0.06 −0.36 −0.88 0.41 −0.09 0.48 −0.29 1.10 −0.87 −0.90 Normal 2005 −0.18 −0.72 −0.49 −1.61 −1.25 −1.23 −1.26 −0.74 −1.65 −1.55 −0.99 −0.47 Ext Drought 2006 1.23 0.82 1.36 1.21 1.41 1.51 1.79 1.11 1.39 1.04 1.72 1.86 Ext Excess 2007 1.36 −0.15 0.44 −0.56 −0.29 −1.15 −0.15 −0.88 −0.51 −0.79 1.29 0.57 Normal 2008 −0.52 −0.80 −0.32 −0.90 −0.44 −0.82 −0.63 −0.96 −0.88 −0.73 −0.28 −0.70 Drought 2009 −0.30 −0.11 −1.03 −1.20 −0.43 0.42 −0.80 −0.97 −1.20 0.10 −0.59 −1.24 Drought 2010 0.43 0.61 1.44 −0.75 0.25 0.85 0.22 1.39 −0.71 −0.41 −0.47 −0.35 Normal 2011 1.37 0.95 0.28 2.08 −0.44 1.37 1.11 0.12 −0.58 0.67 −0.44 −1.02 Excess Source: Kenya Meteorological Department; authors’ calculations. and animal mortality can be substantial, especially in ern and central Kenya, the country’s agricultural heartland, extreme drought years. severely impacting tea and coffee production. Tea output dropped by an estimated 15 percent during the three-year As an example, the GoK estimated that total damages and period (Rice 2006). This prolonged drought is estimated losses resulting from the 2008–11 drought were $12.1 bil- to have cost the Kenyan economy around $2.5 billion lion, equivalent to a drag on economic growth of 2.8 per- (CERF 2008), accounting for approximately 20 percent of cent per year on average (GoK 2012a). An estimated the country’s GDP at the time (IMF 2008). In 2005–06, 85 percent of damages and losses were to agriculture another severe drought affected 3.5 million people, mostly (13 percent) and livestock (72 percent). GoK estimates of nomadic pastoralists in northeast Kenya. An estimated resources required for recovery and reconstruction were an 70 percent of the livestock in affected areas died (CERF additional $1.77 billion. During 1999–2002, an estimated 2009). The same drought reportedly caused heavy losses to 23 million people were affected by severe drought in west- the tea industry. The TBK reported that black tea produc- Kenya: Agricultural Sector Risk Assessment 27 TABLE 3.2. DATES AND IMPACTS OF DROUGHT EVENTS IN KENYA, 1980–2011 Year Regions Affected Impacts 1980–81 Lodwar, Kisumu, Narok 400,000 people affected 1983–84 Nyahururu, Garissa, Kisumu, Makindu, Voi, 50–75 percent cattle mortality in the northern rangelands; Lodwar, Eldoret, Narok, Dagoretti, severe food shortages in Eastern 1987 Nyahururu, Garissa, Makindu, Voi province 1991 Lodwar, Nyahururu, Kisumu, Narok, Lodwar, 1991–92: 1.5–2.7 million people affected; pastoralists forced to Mandera, Narok, Malindi, Mombasa move out of ancestral lands. Substantial numbers of livestock lost 1995–96 Lodwar, Nyahururu, Makindu, Voi, Malindi, 1995–96: Est. $2.8 billion in damages from the loss of crops Mandera, Garissa, Dagoretti, and livestock, etc. 1999–2000 Lodwar, Voi, Garissa, Kisumu, Narok, Dagoetti 23 million people affected; est. $80 million from animal losses; 3.3 million affected households; maize harvest drops by one-third; maize/bean prices 30–50 percent above five-year average 2003–04 Eldoret, Nyahururu, Makindu, Voi, Garissa, Long rains began late and were poor in many areas; est. Malindi, Mombasa 3 million households need food assistance 2005 Nyahururu, Garissa, Kisumu, Narok, Makindu, 3.5 million people and 37 of 78 districts affected; Voi est. $450 million in losses; $197 million in GoK and international humanitarian aid 2008–09 Mandera, Nyahururu, Kisumu, Dagoretti, 3.8 million people affected; $423 million in GoK and Makindu, Voi, Malindi, Mombasa, Eldoret international humanitarian aid; cattle and sheep mortality rates in six ASAL districts ranged from 57 to 65 percent. 2011 Mombasa 4.3 million people affected; food prices 130% higher than normal; from 2008 to 2011, est. $630 million in animal losses, other losses valued at $7.22 billion Source: EMDAT (Emergency Events Database); UN International Strategy for Disaster Reduction ; media and GoK reports; authors’ notes. Note: ASAL, arid and semiarid land; GoK, government of Kenya. tion fell by 20 percent in the first half of 2006 compared down of traditional coping mechanisms. In recent years, to the same period in 2005. Extended dry periods mostly these trends have contributed to an increasing reliance on affect crops in marginal tea production areas, and some emergency aid in the ASALs. Drought can also exacer- losses can occur when the dry period coincides with frost. bate market risks related to price volatility and the afford- ability of concentrate feeds and fodder. According to the literature and anecdotal evidence collected for this study, the main risk for northern pasto- The impacts of past drought years on the livestock sector ralists remains drought (see Chapter 5). Livestock herders are summarized below23: used to anticipate major droughts once every 10 years. » 1983/84: 50–75 percent of cattle reported to have This cycle allowed farmers to recover and rebuild their died in the northern rangelands livestock and crops before the next drought. In recent » 1999/01: Death of animals led to direct losses of K years, however, the frequency of drought has increased Sh 6 billion ($80 million) (Aklilu and Wekesa 2002) to once in every three to four years, leaving less time » 2005/06: Drought caused losses of $450 million for recovery and for rebuilding stocks of food and live- » 2008–11: Drought caused death of animals valued stock. The impact of weather variability is likely much at K Sh 56.1 billion ($630 million) greater in recent times due to rapid population growth and demographic change, contributing to an increasing 23 More details on mortality rates are available in Aklilu and Wekesa (2001, loss of mobility and access to grazing areas, and a break- 2002); Zwaagstra et al. (2012); and Fitzgibbon (2012). 28 Agriculture Global Practice Technical Assistance Paper In addition to a perceived increase in the frequency of Although certain global climate models predict increasing dry years, pastoralists believe that the rains themselves rainfall trends for the region as a whole, the risks of future are getting shorter; the long rains used to last three droughts should not be underestimated given anticipated months, but now last only two to three months. Droughts increases in climate variability linked to climate change. result in reduced pasture and reduced recharge of wells, A 2011 rainfall analysis by the International Food Pol- and thus to water shortages and reduced feed for ani- icy Research Institute (IFPRI) highlighted a noteworthy mals. Drought obliges pastoralists to use boreholes increase in the frequency and persistence of dry events instead of shallow wells; the boreholes are not only in recent years (IFPRI 2013). The same study also found quickly overloaded but the surrounding grazing land is through crop water production modeling that water stress rapidly depleted and eroded. As animals weaken and caused by insufficient rainfall plays a significant role in lose value, the price of staple food items has a tendency rainfed maize production, a finding that has substantial to increase. In the 2008–11 drought, terms of trade implications for Kenya’s food security situation. (TOT) for pastoralists dropped to 50–60 percent of the five-year average (GoK 2012a). An analysis of market FLOODS risks (see section, “Market Risks”) finds similar adverse Extreme rainfall causing localized flooding of cropland movements in TOT. and/or pastureland is a common weather phenomenon in Kenya. Perennial floods affect low-lying regions of the Beyond the problem of drought, erratic rainfall (includ- country such as river valleys, swampy areas, lakeshores, ing late rain onset, rains ending early, and extended dry and the coastal strips that are unevenly distributed in the periods during the rainy season) has become a significant country’s five drainage basins. Geographically, the west- problem for Kenyan farmers. Historically, farmers could ern, northern, eastern, central, and southeastern parts of count on rains arriving the last 10 days in October (short the country are quite susceptible to seasonal floods during rains) and the last 10 days of March (long rains). In recent the two rainy seasons. The Lake Victoria Basin in western years this has been less certain, and farmers find they do Kenya is the most flood-prone region, while the country’s not know when they will have enough moisture for the ASALs are also prone to floods, despite their low average planted seed to survive. Planting late results in poor yields rainfall of only 300–500 mm. or outright crop failure, while planting too early results in the need to replant, perhaps several times, and higher Table 3.3 shows years during which Kenya was affected costs. Farmers also report a higher incidence of years by excessive rainfall during the period 1981–2011. Based when rains end early or a dry spell occurs during the rainy on the analysis, the country experienced eight years dur- season and compromises yield. Maize farmers reported ing which the amount of rainfall was significantly higher significant losses due to this phenomenon in 2011 in than the norm; four of these were categorized as extreme particular. In some areas, such as the Eastern Province, rainfall events (1982, 1987, 1988, 2006). This suggests that farmers indicated that they could only get a good maize floods occur roughly once every four years on average. harvest once every six to eight seasons. Yet many continue to plant maize every season. Despite the relatively frequent occurrence of flooding in many regions, resultant losses to crops or livestock are Significant amounts of rainfall in the dry season (Janu- rarely extensive at an aggregate level as impacts tend be ary–February) can cause losses in coffee quality because isolated locally. An exception to this was the 1978/98 El of anticipated flourishing. Likewise, extended rainfall Niño event, which resulted in severe floods after major after the long rains (March–May) can encourage higher rivers in the country attained record peaks. Flooding incidences of coffee-related pests and diseases, necessitat- caused loss of lives and significant damage to infrastruc- ing treatment, lowering yields, and upping the costs of ture and other assets, with one estimate placing losses at production. As stated by farmers during the mission field 11 percent of national GDP. Widespread floods in 2006 visits, erratic rainfall can cause yield losses in coffee of up affected large swaths in Coast Province and parts of to 20–40 percent. North Eastern Province, in which the most affected dis- Kenya: Agricultural Sector Risk Assessment 29 TABLE 3.3. FREQUENCY OF SURPLUS RAINFALL EVENTS, 1963–2012 Year Lodwar Mandera Eldoret Nyahururu Garissa Kisumu Narok Dagoretti Makindu Voi Malindi Mombasa 1981 −1.03 1.14 −0.24 0.28 −0.33 −1.53 −1.16 0.67 0.40 0.92 −0.20 0.20 Drought 1982 3.31 0.50 0.86 1.48 0.42 0.57 0.18 0.65 1.69 0.69 2.52 1.81 Ext Excess 1983 −0.69 −0.87 1.63 −1.47 −1.34 −1.36 −0.50 −0.36 −1.49 −1.32 0.22 0.14 Ext Dry 1984 −1.08 −0.92 −2.18 0.40 0.17 −0.82 −1.50 −2.29 0.53 −0.22 0.67 0.20 Ext Dry 1985 0.01 −0.60 −0.60 −0.39 −1.02 −0.04 0.59 −0.80 −0.33 −0.62 −0.87 −0.63 Normal 1986 −0.32 −0.56 −1.43 0.10 −0.07 0.11 −1.30 −0.20 0.20 −0.08 0.13 0.05 Normal 1987 −0.15 −0.32 0.46 −1.26 −1.01 −0.41 0.69 −0.81 −1.26 −1.20 −0.10 −0.60 Drought 1988 1.40 −0.86 −0.28 0.50 0.16 0.41 0.49 1.28 0.63 0.22 −1.17 −0.24 Normal 1989 0.49 0.50 0.98 1.70 1.54 0.06 1.88 1.20 1.92 −0.45 −0.17 −0.11 Excess 1990 −0.89 0.68 −0.29 0.82 0.74 −1.13 −0.53 0.49 0.98 0.97 −0.80 −0.15 Normal 1991 −0.50 −0.11 −0.24 −0.51 −0.28 −0.52 −1.05 −0.39 −0.45 0.14 −0.32 0.17 Normal 1992 −0.97 −0.52 0.33 0.00 −0.21 −0.67 −1.48 −0.07 0.10 0.18 −1.16 −0.51 Normal 1993 −0.61 1.18 −1.84 −0.61 −0.14 −0.67 0.36 −0.42 −0.56 −0.17 −0.43 −0.81 Normal 1994 −0.51 −0.15 0.59 0.78 0.37 0.92 0.27 0.06 0.93 1.08 1.45 1.42 Excess 1995 −0.93 −0.08 −0.39 −0.99 0.27 0.71 −0.35 −0.19 −0.98 −0.73 −1.08 0.12 Normal 1996 −0.03 −0.92 0.17 −0.62 −1.26 1.15 0.45 −1.40 −0.57 −0.49 −0.71 −0.14 Normal 1997 1.26 4.08 0.79 0.91 3.23 1.60 1.45 1.02 1.07 0.79 2.41 3.51 Ext Excess 1998 −0.55 −0.54 2.42 1.72 1.69 −1.57 0.93 1.79 1.94 2.97 0.85 0.16 Ext Excess 1999 −0.73 −0.37 0.01 0.60 −0.33 0.97 −0.40 −0.11 0.74 −0.70 0.67 0.07 Normal 2000 −0.92 −0.39 −0.99 −0.37 −1.13 −1.07 −1.38 −1.58 −0.30 −0.33 −0.25 −0.28 Drought 2001 −0.75 −1.27 −0.18 0.40 −0.56 0.63 −0.02 1.80 0.53 −0.37 −0.68 −0.68 Normal 2002 0.95 0.18 −0.76 −0.38 0.95 1.72 1.85 0.37 −0.32 1.21 0.06 −0.07 Excess 2003 −0.03 −0.20 −0.56 −1.02 0.21 −0.42 0.37 −0.26 −1.01 −1.94 −0.42 −1.37 Drought 2004 −0.10 −0.20 0.06 −0.36 −0.88 0.41 −0.09 0.48 −0.29 1.10 −0.87 −0.90 Normal 2005 −0.18 −0.72 −0.49 −1.61 −1.25 −1.23 −1.26 −0.74 −1.65 −1.55 −0.99 −0.47 Ext Dry 2006 1.23 0.82 1.36 1.21 1.41 1.51 1.79 1.11 1.39 1.04 1.72 1.86 Ext Excess 2007 1.36 −0.15 0.44 −0.56 −0.29 −1.15 −0.15 −0.88 −0.51 −0.79 1.29 0.57 Normal 2008 −0.52 −0.80 −0.32 −0.90 −0.44 −0.82 −0.63 −0.96 −0.88 −0.73 −0.28 −0.70 Normal 2009 −0.30 −0.11 −1.03 −1.20 −0.43 0.42 −0.80 −0.97 −1.20 0.10 −0.59 −1.24 Drought 2010 0.43 0.61 1.44 −0.75 0.25 0.85 0.22 1.39 −0.71 −0.41 −0.47 −0.35 Normal 2011 1.37 0.95 0.28 2.08 −0.44 1.37 1.11 0.12 −0.58 0.67 −0.44 −1.02 Excess Source: Kenya Meteorological Department; authors’ calculations. tricts were Mombasa, Kwale, Kilifi, Isiolo, Turkana, and 2012. Floods affecting only a single region were recorded Moyale. Even though potatoes are highly susceptible to in another two years. Floods typically do not impact as flooding and water logging, these risks are not commonly much area or as many farmers as a drought, although faced by growers because the crop is predominantly cul- the impact on those directly affected may be quite severe. tivated in the highland areas of Rift Valley, Central, and Adverse impacts from floods were not evaluated for this Eastern Provinces, which are not prone to these risk fac- study as the agricultural damages associated with them tors. Tomatoes and some other vegetables are susceptible, are not as significant as those compared to drought. In but farmers generally do not consider flooding to be a the ASALs, other climatic events such as cold or out-of- major risk. season heavy rain and flash floods can cause severe losses to herds, especially those weakened by disease or lack of Based on observable records, an estimated nine major feed. These events and their impacts are often localized, flood events affected various regions between 1980 and however. 30 Agriculture Global Practice Technical Assistance Paper FROST rains will make it increasingly difficult to plan sowing and Frost mostly impacts crop production at higher altitudes. harvest times, causing lower maize yields in some major Tea is most susceptible to frost. Tea farmers near Keri- production zones, and greater food insecurity. Also, incre- cho who provided input for this study reported that frost mental changes in temperature and rainfall patterns are events have become more frequent in recent years. Frost expected to contribute to biodiversity loss and emergence exposure does not typically result in plant death but can of new pests and diseases. reduce productivity for several months while affected plants recover. Although damages from frost exposure at Some crops are expected to experience more favorable the aggregate level have been negligible historically, in growing conditions as a result of climate change, whereas January 2012 the industry reportedly lost an estimated 20 others will find future climatic conditions intolerable. million kg of green leaf, valued at US$11.4 million, from Equally, some regions (the mixed rainfed temperate and frost in what the KTDA reported as the worst case of frost tropical highlands) are projected to experience an increase to ever hit the country. in crop yield, whereas others (the ASALs) are projected to witness a significant decline in crop yield and livestock numbers as water resources become increasingly scarce. OTHER WEATHER-RELATED RISKS In addition to droughts, floods, and frost, Kenyan agricul- ture is affected by weather events such as hailstorms and PESTS AND DISEASES windstorms (often accompanying heavy rain or hail). How- As in other countries, pests and diseases are a permanent ever, such weather-related risks tend to affect smaller areas fixture of both crop and livestock production systems in with only negligible impacts on aggregate production. Kenya. The majority of pest and disease threats are man- ageable, but farmers and livestock herders do not always practice prevailing control measures or avail themselves CLIMATE CHANGE of available technologies, due to lack of information, Kenya is ranked as one of the countries (#13 of 169) access to needed inputs, or financial resources. This sub- most vulnerable to physical climate impacts from extreme section presents a discussion of some of the most notable weather, according to the Center for Global Develop- pest and disease risks in Kenyan agriculture. The main ment. In all of Africa, only Somalia (#7), Sudan (#9), biological threats and the crops they affect are summa- Malawi (#11), and Ethiopia (#12) are ranked higher. rized in table 3.4. Across Kenya’s economic landscape, the agriculture sec- tor is by far the most vulnerable to impacts from climate change. CROP PESTS AND DISEASES Outbreaks of African armyworm (Spodoptera exempta) are The climate predictions of IFPRI (2013) and others for commonplace across Kenya. The armyworm attacks all Kenya highlight a number of risks and impacts for the graminaceous crops and is a significant and perennial agriculture sector.24 These include more frequent extreme concern for farmers and livestock herders. Uncontrolled events such as prolonged drought and flooding, leading to outbreaks can cause total crop loss, with millions of hect- a decrease in reliable cropping days and higher incidences ares of farmland and pastureland affected in bad years. of crop failure. Increased frequency of drought will likely Normal rainfall following drought often precipitates large- contribute to more frequent water shortages for domestic scale infestation. According to CABI Africa, outbreaks of use and crop and livestock agriculture. Unpredictable pre- armyworm in mid-2008 were reported in 24 districts in cipitation during both the short and long rains, together Kenya, damaging 10,324 ha of crops and 41,435 ha of with extreme events, particularly increased frequency of pasture. Existing control measures are generally effective. drought, may cause a decline in agricultural productiv- These are managed via a national forecasting unit that ity. In addition, changes in the timing of long and short monitors previous outbreaks and meteorological data to predict broadly where outbreaks might occur in the near- term and an early warning system network of more than 24 See Appendix A for a synopsis of recent climate change impact analyses. Kenya: Agricultural Sector Risk Assessment 31 TABLE 3.4. PRINCIPAL PEST AND 400 pheromone traps operated by extension agents and DISEASE RISKS IN KENYAN the Plant Protection Services. AGRICULTURE Maize: Maize is particularly vulnerable to a wide range Pest/Disease Crops/Animals Affected of pests and diseases. Practically speaking, weeds are a Crops constant threat to maize production. Effective control Maize Streak Virus (MSV) Maize requires the use of significant labor or expensive herbi- Disease cides. Striga is a parasitic weed reported to infest 210,000 Maize Lethal Necrosis Disease Maize (MLND) hectares in western Kenya alone (AATF 2006). Accord- Large Grain Borer Maize ing to the African Agricultural Technology Foundation, Maize Weevil Maize Striga costs African farmers across the continent about $1 Stem/Stalk Borer Maize, wheat, sugarcane billion per year. Ratoon Stunting Disease Sugarcane Sugarcane smut Myriad insects are also a constant threat to maize produc- Termites Maize, sugarcane tion (e.g., stemborer) and storage (e.g., larger grain borer, Armyworm, bullworm Cereals, root crops, sugarcane, weevils), while others pose a more sporadic threat. Com- vegetables, pasture grasses mon insect pests are categorized into three general groups: Thrips, Aphids, Mealybugs, Maize, coffee, tea, sugarcane, (1) moths, which include cutworms, earworms, stemborer, Nematodes vegetables, fruits and grain moths; (2) beetles, including rootworms, wire- Coffee Leaf Rust (bacterial Coffee worms, grubs, grain borers, and weevils; and (3) disease vec- blight) tors, most notably leaf hoppers, thrips, and aphids. Many Coffee Wilt Disease pests and diseases can be controlled with good crop hus- Coffee Berry Disease bandry and chemical treatments, but these are often costly, Cassava Mosaic Disease (CMD) Cassava and farmers hesitate to pay the cost of treating for pests or Cassava Brown Streak Disease diseases not expected to be a serious problem. Some farm- (CBSD) ers’ reluctance to respond to low-level threats also contrib- Banana Xanthomonas Wilt (BXW) Bananas utes to periodic outbreaks of known pests and diseases. Black Sigatoka Leaf Spot (BSLS) Panama Disease Maize is susceptible to a long list of fungal (e.g., rust, spot, Yellow Sigatoka blight, smut) and viral diseases (e.g., maize streak virus). Bacterial Wilt Potato, tomato Pests and diseases will typically lower yields but not cause Late/Early Blight Potato, tomato substantial losses, as most farmers are aware of and know Potato Leaf Roll Potato how to manage them. The real problem is the emergence Weevils, beetles Sweet potato of a new threat. A recent example is the appearance of Red Spider Mite Tomato the maize lethal necrosis disease (MLND), first reported Striga Cereals in June 2011 in Bomet, Naivasha, and Narok Counties Livestock in the Southern Rift Valley (Wangai et al. 2012). Since East Coast Fever Cattle then, additional outbreaks of MLND have been reported Rift Valley Fever in parts of the North Rift Valley as well as in the south. Anthrax Cattle According to KEPHIS (Kenya Plant Health Inspectorate Foot and Mouth Disease Cattle, pigs Services), the disease is now widespread in Chepalungu, Contagious Bovine Sotik, Transmara, Bureti, Nakuru, Konoin, South Narok, Pleuropneumonia (CBPP) Mathira East, Imenti South Districts, and Nyeri. Inci- Pestes des Petits Ruminants (PPR) Goats, sheep dence in the field ranges from 40 to 100 percent of the Newcastle Disease Poultry crop, and over 80 percent crop loss has been reported in Source: Authors’ notes. some cases. 32 Agriculture Global Practice Technical Assistance Paper MLND has been the greatest pest/disease threat in Sugarcane: Sugarcane smut, caused by the fungus Usti- recent years in Kenya because there is no cure and lago scitaminea, is considered the most important disease resistant varieties have yet to definitively emerge from impacting sugarcane production in Kenya. It is endemic research. The main transmission route is insect vec- across Kenya’s sugarcane production zones. Yield losses tor (thrips and beetles), but transmission via seeds also of 21–38 percent were documented recently through seems likely. Chemical treatment to limit disease vectors field research by the Kenya Sugar Research Foundation is believed to help control its spread, but many farm- (KESREF). These and other findings suggest that varie- ers in parts of the Southern Rift Valley have reportedly ties previously rated as resistant or immune are becom- switched to other crops after suffering severe crop losses. ing increasingly susceptible and that new strains of the The GoK has ramped up research and distributed sor- fungus may have evolved (KESREF 2011). Ratoon Stunt- ghum, finger millet, cassava, and sweet potato seeds to ing disease is another common threat, but yield losses are farmers from areas previously affected with MLND to thought to be much lower. Among pests, stock borer and grow as alternatives to maize (Kamau 2013). Kenya termites are common threats. Agricultural Research Institute (KARI), KEPHIS, and the International Institute for Topical Agriculture (IITA) LIVESTOCK PESTS AND DISEASES are leading the search for effective ways to combat the East Coast fever (ECF) is considered the most serious disease, but the process of breeding resistant varieties livestock disease and is present on several of Kenya’s bor- will likely take three to six years. ders. Tick-borne, ECF can kill large numbers of calves in pastoralist herds. Spraying the ticks can be an effective Potato: The biggest threat to potato production in Kenya method of control if maintained, but this is expensive. is bacterial wilt (caused by Ralstonia solanacearum) The dis- With ECF present in neighboring countries, controlling ease is prevalent in all potato-growing areas in Kenya, ECF is difficult given Kenya’s open borders; emergency affecting over 70 percent of potato farms and causing fodder provision and climate change have also expanded yield losses of between 50 and 100 percent (World Bank the areas affected by ECF as the tick specie responsible 2012). Late blight is another common threat affecting an has spread. estimated two-thirds of all potato crop farms. Rift Valley fever is similarly hard to control in Kenya. It Coffee: Unlike tea, which is relatively resistant to pests could be considered a constraint rather than a risk as its and diseases, Kenya’s coffee industry is threatened by two strong, positive correlation with heavy rainfall and flood- major diseases: CBD (Colletotrichum kahawae) and CLR ing makes it relatively predictable. The risk is that very few (Hemileia vastatrix). Both are major diseases of Arabica cof- animals are vaccinated, because vaccination frequently fee that, left untreated, can cause significant losses. Severe leads to abortion in pregnant animals; even if mortality rust incidence may lead to loss of foliage (up to 50 per- is relatively low, the losses are high when outbreaks occur cent) and berries (up to 70 percent) (Alwora and Gichuru and vaccination takes place. 2014). CBD infects all stages of the crop, from flowers to ripe fruits, and can cause up to 50–80 percent yield loss Foot and mouth disease (FMD) is endemic in Kenya and if conditions are favorable and no control measures are can cause high mortality rates, especially in improved adopted (Gichimu and Phiti 2012). Both CBD and CLR breeds. Vaccination is effective and provides short-term are manageable via adoption of good cultural practices, immunity, but since cost recovery was introduced in the such as planting resistant varieties and applying contact late 1980s, coverage has fallen to around 10 percent. FMD and systemic fungicides. Chemical control of these dis- is especially damaging when it coincides with drought and eases is expensive (up to 30 percent of total production animals are weak and stressed. Over a 93-day quaran- costs), however. Control measures also focus on the devel- tine and observation period due to FMD in 2001, a study opment and dissemination of coffee varieties resistant to on a large Kenyan dairy farm recorded costs and losses CBD and CLR, but farmer access and the replanting of that included milk losses (42.0 percent), purchase of addi- improved varieties remains limited. tional feeds (13.6 percent), culling of milk cows that devel- Kenya: Agricultural Sector Risk Assessment 33 TABLE 3.5. FREQUENCY AND IMPACT OF LIVESTOCK DISEASE OUTBREAKS IN KENYA, 1980–2013 Year Description Early 1980s Animals worth K Sh 230 million lost to FMD 1996 1.47 million, 2.48 million, and 1.15 million animals vaccinated after FMD, Rinderpest, and CBPP outbreaks, respectively 1997 105 and 106 reported outbreaks of FMD and Rift Valley fever 1999 139 reported outbreaks of FMD; 0.65 million animals vaccinated 2000 95 cases of FMD; 0.46 million animals vaccinated; 16 cases of CPB and 1.1 million animals vaccinated 2001 54 FMD outbreaks, 0.76 million animals vaccinated; 1.96 million animals vaccinated against CPB (18 outbreaks); 11 cases of Newcastle disease 2002 48 FMD reported cases; 19 cases of CBPP; 21 cases of lumpy skin disease; 10 cases of Newcastle disease 2003 87 reported FMD outbreaks; 21 reported CPB outbreaks; 16 reported cases of Infectious bursal disease (Gumboro disease) 2004 95 FMD outbreaks; 46 cases of fowl typhoid; 24 cases of infectious bursal disease 2007 First outbreak of PPR, causing 1,500 animal deaths in Rift Valley; 37 cases of Rift Valley fever across 29 of 69 administrative districts in six of eight provinces 2011 4 reported severe outbreaks of African Swine fever in 16 districts Source: OIE database; HANDISTATUS II; media reports; GoK reports. Note: FMD, foot and mouth disease; contagious bovine pleuropneumonia (CBPP). oped chronic mastitis (12.5 percent), extra labor inputs during the period 1980–2013. Incidences of unreported (8.9 percent), veterinary fees (3.3 percent), transport outbreaks are undoubtedly considerably higher. The lack (3.0 percent), deaths (3.0 percent), drugs (2.9 percent), abor- of information on losses associated with these outbreaks tions (1.4 percent), and chemicals (0.5 percent). Quarantine makes it difficult to quantify their impacts. and lack of sales of other livestock (pigs) and commodi- ties (hay) on the farm led to overall short-term, farm-level MARKET RISKS direct and indirect losses of approximately $16,026.25 An Among the most common market risks presented in this earlier FMD outbreak in the 1980s was estimated to have section are price variability for crops and inputs, exchange caused K Sh 230 million in losses (GoK 2009a). rate and interest rate volatility, counterparty risks, and livestock theft. Other animal diseases that are potentially most serious during a drought include small ruminant pest (PPR), con- tagious bovine pleuropneumonia (CBPP), and catarrhal CROP PRICE RISK fever. Anthrax is a serious, yet localized threat, and a new Price fluctuations are inherent in agricultural markets, respiratory disease in camels is a source of concern. Animal and some level of variability is to be expected. This is disease is especially dangerous when drought and disease partly due to supply and demand dynamics and the are covariant, as is often the case, as even common day-to- unpredictability of weather patterns and harvest yields. day levels of infection by normally mild diseases (e.g., orf, However, extreme price volatility deters producers from pox) or internal or external parasites can become fatal. making productivity-enhancing investments and can weaken food access among poorer households. It can Table 3.5 provides some details on reported pest and dis- also lead to lost income. The analysis of producer price ease outbreaks affecting Kenya’s livestock populations variability is based on interannual price variability for the period 1991–2011, measured by CVs. Nominal prices in $/ton are used for the analysis of domestic 25 The Kenya Veterinarian (2001); see http://www.ajol.info/index.php/kenvet /article/view/39523 and http://www.flockandherd.net.au/other/reader producer prices. Annual producer price data are drawn /fmd%20kenya.html from FAOSTAT. 34 Agriculture Global Practice Technical Assistance Paper FIGURE 3.2. AVERAGE MONTHLY WHOLESALE MARKET PRICES (K Sh/90 kg), 2005–13 8,000 Maize Cowpeas Sorghum Irish potatoes Tomato Wheat 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 Jan-05 May-05 Sep-05 Jan-06 May-06 Sep-06 Jan-07 May-07 Sep-07 Jan-08 May-08 Sep-08 Jan-09 May-09 Sep-09 Jan-10 May-10 Sep-10 Jan-11 May-11 Sep-11 Jan-12 May-12 Sep-12 Jan-13 May-13 Sep-13 Source: MoALF. Note: Potato price is based on 110 kg unit; tomato price is based on 64 kg unit. TABLE 3.6. INTERANNUAL CROP PRICE remain relatively stable (19 percent). Maize price volatil- VARIABILITY (CV), 1991–2011 ity, in particular, is a critical issue for the GoK, given the importance of maize to household consumption and to Coefficient of Variation food security. In Kenya, maize prices increased sharply in Cereal Crops Other Cash Crops 2008–2009, fell in 2010, and rose again in 2011 and 2012. Maize 0.33 Cowpea 0.42a Tea 0.29 During the first half of 2011, maize prices jumped by Wheat 0.33 Dry beans 0.32a Coffee 0.53 145 percent. This followed a sharp increase (39 percent) Sorghum 0.49 Potato 0.28 Sugarcane 0.23 in the commodity food price index and a near doubling of Rice (Paddy) 0.75 Banana 0.28 U.S. maize prices during the period June 2010–February Source: FAOSTAT. 2011.26 In general, domestic maize prices tend to be more a Prices for cowpea and dry beans cover 1999–2011. volatile than international maize prices, as domestic prices are highly sensitive to constant speculation in projected and real annual output. Based on CV analysis, producer prices in Kenya for key Government procurement programs regularly exert pres- crops are subject to moderate to high levels of inter- sure on normal market price developments and contrib- annual price variability (table 3.6). Among crops, rice ute to higher intra-annual price volatility for maize and paddy, coffee, sorghum, and to a lesser extent, cowpea, other major staples. For farmers, maize prices invariably exhibit the highest levels of year-on-year price volatility. collapse at the peak of the harvest season. The GoK inter- In the case of rice and coffee, this suggests that domes- venes by announcing a pan-territorial price for all maize tic prices are highly influenced by imports and changes bought by the NCPB for replenishment of the strategic in international market prices. It also suggests that rice grain reserve. The intervention invariably pushes up and coffee producers in Kenya are highly exposed to prices in the short term. In 2013–14, the pan-territorial significant swings in farm-gate prices from one year to price for maize was K Sh 33,000 ($385) against an aver- the next. age market price of K Sh 30,000 ($347) per ton. When significant shortages occur, the Kenyan government can Figure 3.2 shows monthly fluctuations in wholesale prices also waive the 50 percent duty on extra-COMESA maize for six key crops during the period 2005–13. Tomato, maize, and Irish potato exhibit the highest levels of vari- ability, with CVs of 35–38 percent, while wheat prices 26 According to Index Mundi at indexmundi.com. Data accessed May 2014. Kenya: Agricultural Sector Risk Assessment 35 FIGURE 3.3. PRICE OF TEA AT MOMBASA AUCTION ($/kg), 1980–2012 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: TBK. FIGURE 3.4. INTERNATIONAL COFFEE PRICES ($/lb),* 1988–2013 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Jan-88 Oct-88 Jul-89 Apr-90 Jan-91 Oct-91 Jul-92 Apr-93 Jan-94 Oct-94 Jul-95 Apr-96 Jan-97 Oct-97 Jul-98 Apr-99 Jan-00 Oct-00 Jul-01 Apr-02 Jan-03 Oct-03 Jul-04 Apr-05 Jan-06 Oct-06 Jul-07 Apr-08 Jan-09 Oct-09 Jul-10 Apr-11 Jan-12 Oct-12 Jul-13 Source: International Coffee Organization. Note: *New York cash price, ex-dock New York; ($/lb). imports, as happened in 2008. When implemented, this of the average tea selling price at the auction (excluding measure exerts considerable and rapid downward pres- marketing, processing, and transport costs). Total farmer sures on prices. payments are typically above 70 percent of the auction price, which is considered to be a relatively good incen- No universal reference market exists for tea prices as for tive to maintain good crop husbandry and carry out farm other major commodities. Instead, domestic prices are investments. Multinational companies operating in Kenya influenced by prevailing auction prices in other major make just one payment to their outgrowers. During cer- tea-producing countries, such as those in Colombo and tain seasons, their buying practices promote side-selling Calcutta. The Mombasa Auction average yearly prices among KTDA farmers who prefer immediate payment in have increased since 2002, with notable inflexions during full. Whatever the extent of volatility, price risk is primar- 2007–09 and 2011–12 (figure 3.3). The interannual vari- ily borne by individual tea farmers and cooperative socie- ations are not very pronounced. In effect, the CV of the ties, and KTDA’s dual payment system has come under average annual auction price is 27 percent (1980–2013), increasing pressure in recent years as farmers demand relatively modest compared to the average annual inter- better price transparency. national price of coffee (41 percent CV in the period 1988–2013) based on the New York market. Coffee prices in the international market have historically been highly volatile. During the period 1988–2013, inter- Smallholder suppliers to KTDA factories are paid a fixed annual fluctuations in coffee prices were subject to a CV price during the whole year per green leaf kilogram (K of 0.43 (figure 3.4). The level of variability has decreased Sh 14/kg in 2013–14). In addition, farmers receive a markedly as prices have slid from their record peak in supplemental year-end bonus, determined on the basis April 2011. Price risk faced by buyers of Kenya coffee is 36 Agriculture Global Practice Technical Assistance Paper FIGURE 3.5. WEEKLY BEEF CATTLE PRICES (K Sh/kg) IN VARIOUS MARKETS, 2006–11 40,000 Garissa Isiolo Moyale Dagoretti 35,000 30,000 25,000 20,000 15,000 10,000 5,000 – Jan-06 Jul-14 Jan-07 Jul-14 Jan-08 Jul-08 Jan-09 Jul-09 Jan-10 Jul-10 Jan-11 Source: MoALF. largely managed via hedging on futures. Agents and other Figure 3.6 and figure 3.7 compare real prices and TOT downstream actors in the supply chain mostly work on between maize and beef cattle in six major market cent- commission and fixed fee–based rates, so face little to no ers, including Nairobi and Mombasa, for the period price risk. For the country’s farmers, their major concern January 2006 to January 2011. The TOT are calculated is consistently low prices at farm-gate rather than intra- as the number of 90-kg bags that can be purchased by or interannual variability of prices. During periods where selling one cow. Variations in average monthly prices and coffee prices are low, these households are often highly TOT ranged between 40 and 77 percent within years. vulnerable to food insecurity due to limited resources The sample years include the drought years of 2008–11. employed for food production. A depression in prices and especially in TOT is clearly seen as drought occurs. However, the highest varia- tion in TOT occurred in nondrought years rather than LIVESTOCK PRICE RISK drought years, when cattle prices often bottom out. Mar- Seventy percent of Kenya’s cattle are located in the ket dynamics in Nairobi, the major cattle market sink, ASALs, and the income they generate is essential to resi- show more stability than those in pastoral areas, espe- dents’ livelihoods. Livestock marketing has been liberal- cially in Isiolo Market (figure 3.7) where price drops are ized, and no livestock trade policies or regulations directly most visible. In other centers such as Garissa and Moyale, affect domestic prices (FAO 2013b). However, this analy- the variation may be less, as these are border towns with sis showed that cattle producers receive substantially less greater flows of cattle from neighboring countries. The than equivalent world market prices. The first reason for data clearly show how vulnerable livestock owners are to this is the nature of the markets themselves. price and market dynamics. Livestock prices are variable across both seasons and years; cattle prices are the most variable (see figure 3.5). Livestock marketing is poorly regulated and quite disor- Drought triggers adverse trends in prices both of animals ganized. Sellers are often exploited by traders and mid- and of the commodities that pastoralists consume, espe- dlemen, and significant inefficiencies exist. Producers face cially food items, such as maize and beans, that they do substantial market price disincentives despite Kenya’s not produce themselves. At early signs of low rainfall, pas- status as a net exporter of cattle (Makooha et al. 2013). toralists start to sell animals and market prices fall because These arise from market structure (traders’ high profit few buyers exist. At the same time, farmers and traders margins and heavy government fees and taxes imposed on begin storing food in the expectation of future price rises, cattle trekkers). The distortions are in part due to informa- pushing up the price of food staples. Combined with the tion asymmetry. Additional graphs showing the variation falling price of animals, pastoralists are caught in a classic in TOT for each of the individual markets are provided in price scissors situation. Appendix E (figures E.2–E.7). Kenya: Agricultural Sector Risk Assessment 37 FIGURE 3.6. BEEF CATTLE VERSUS MAIZE TOT IN SIX MAJOR MARKETS, 2006–11 Garissa Isiolo Moyale Dagoretti Wajir Mombasa 25 No. bags (90kg) maize 20 15 10 5 0 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 May-10 Jul-10 Sep-10 Nov-10 Jan-11 Source: MoALF. FIGURE 3.7. CATTLE VERSUS MAIZE TOT IN ISIOLO MARKET, 2006–11 18 16 14 12 10 8 6 4 2 0 Jan-06 Mar-06 May-06 Jul-06 Sep-06 Nov-06 Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 May-10 Jul-10 Sep-10 Nov-10 Jan-11 Source: MoALF. INPUT PRICE VOLATILITY EXCHANGE RATE VOLATILITY As elsewhere, high prices inhibit broad farmer utilization Kenya’s heavy reliance on cereal and other agricultural of fertilizers in Kenya. Since 2008, the GoK has imported imports means that movements in exchange rates between roughly 60,000 MT of planting and top-dressing fertiliz- the Kenyan shilling and major trading currencies represent ers annually through NCPB to stimulate production and a potential source of market risk. Any volatility also affects enhance food security. Yet fluctuating prices can impose exporters of agricultural products such as cut flowers, tea, added risk on farmers, making it difficult for them to plan coffee, and horticultural crops Available data from 1995 ahead. An analysis of average annual prices across key to 2013 show that significant monthly nominal exchange agricultural fertilizers during the period 1998–2007 sug- rate fluctuations occurred between May 2007 and April gests that while prices increased, producers faced only 2009, and then again from January 2011 to March 2012 moderate levels of year-on-year price volatility. Among (figure 3.9). The exchange rate actually rose from a low of nine fertilizers for which historical price information was K Sh 61.96 per U.S. dollar in May 2008 to a high of K available, only one (Urea, with a CV of 34 percent) exhib- Sh 101.16 per U.S. dollar in October 2011. This drastic ited higher than normal levels of price variability over the currency depreciation was due in part to Kenya’s growing 10-year period (figure 3.8). A further analysis of interan- trade imbalance and its sizable current account deficit, nual prices for six fertilizers during 2008 showed limited which was above 10 percent of GDP in 2011, one of the variability, with CVs ranging from 15 to 24 percent. The highest in the world. The gradual depreciation has damp- analysis suggests that farmers face limited risks from input ened Kenya’s capacity to import essential food and energy price risk volatility. commodities, while making its exports more competitive. 38 Agriculture Global Practice Technical Assistance Paper FIGURE 3.8. DOMESTIC FERTILIZER PRICES, 1998–2007 2,500 SSP TSP DAP MAP ASN SA UREA NPK* CAN 2,000 1,500 1,000 500 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Source: Agri-Business Directorate. FIGURE 3.9. EXCHANGE RATES ($/K Sh), 1995–2013 110.0000 100.0000 90.0000 80.0000 70.0000 60.0000 50.0000 40.0000 January 1995 January 2000 January 2005 January 2010 Source: OANDA. FIGURE 3.10. COMMERCIAL BANKS’ INTEREST RATES* (%), 1992–2013 35 30 25 20 15 10 5 0 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: Central Bank of Kenya. * Weighted monthly average lending rate. INTEREST RATE VOLATILITY During the same period, volatility was relatively high, Analysis of monthly interest rates on commercial bank with a CV of 30 percent. Variability declined markedly lending for the period 1992–2013 shows that rates have over the last decade (2004–13), with the exception of a declined markedly since their peak in 1994 (figure 3.10). sudden spike beginning in late 2011 when inflationary Kenya: Agricultural Sector Risk Assessment 39 pressures, tighter monetary policy, and unexpected vola- tility in exchange rates drove up lending rates by nearly ENABLING ENVIRONMENT one-third (31.7 percent) over a two-month period. This RISKS analysis suggests that volatility in the cost of credit could Other sector risks arise from both internal and external pose a risk. Considering Kenya’s narrow agricultural changes in the broader political and economic environment credit markets, which accounted for a mere 3.6 percent in which agriculture operates. Agriculture sector policy and of total lending in Kenya in 2013 (CBOK 2014), any ex regulation are a source of risk when public involvement post impacts are considered negligible at the aggregate in sector activities has unexpected, adverse consequences. level. Other risks include general insecurity as a result of domes- tic social unrest or transboundary conflict that disrupts agricultural production systems and livelihoods. COUNTERPARTY AND DEFAULT RISKS Counterparty risk refers to the risk that one or more POLICYMAKING parties participating in a transaction will not live up Ongoing implementation of Kenya’s 2010 Constitution to, or will otherwise default, on their obligations. Most will continue to have major consequences for agriculture stakeholders across Kenya’s agricultural supply chains sector policies and programs. It requires a drastic reduc- (including producers, wholesale traders, processors, and tion by at least half in the number of ministries and a banks) have only limited means to effectively manage substantial consolidation and reorganization of ministe- such risks. With few alternatives, many actors prefer to rial functions. It also calls for consolidation of more than minimize their exposure by operating on a limited vol- 100 overlapping pieces of legislation into four new laws: ume, cash-and-carry basis. For rural banks, microcredit the “Agriculture, Fisheries and Food Authority Act,” the institutions, and other credit lenders, repayment failure “Livestock and Fisheries Act,” the “Crops Act,” and is a constant problem and a significant disincentive to the “Kenya Agricultural and Livestock Research Act.” lend to farmers, especially smallholders, who represent a The first three were passed in January 2013 and are substantial and hard-to-manage default risk. In fact, the now undergoing implementation. The new constitution inability to manage such risks is a principal factor limit- also mandates devolution of sector service delivery to ing farmers’ access to credit and driving up the cost of Kenya’s 47 counties. agricultural credit. Such drastic changes raise the spectre of uncertainty over Side-selling is another form of nonpayment risk that is planning and support to the sector in the near to medium particularly common among sugarcane and, to a lesser term. In many cases, experienced District Agricultural extent, tea producers in Kenya. Come harvest time, Officers are being replaced by new graduates from the small-scale sugarcane farmers often sell their crops to local area, a process that will potentially reduce the qual- impromptu roadside buyers rather than honor contracts ity of the civil service officers available and the extension with millers who regularly provide inputs in exchange services they provide. Anecdotal evidence suggests that for the crops they help finance. The practice is known counties have a growing tendency to raise taxes collected locally as “cane poaching.” Mumias Sugar Company, on local agricultural production and to collect taxes again Kenya’s largest miller, reportedly made pre-tax losses of on any products in transit across their territory. This may $26 million in 2013 as a result of the combined impact significantly increase the cost of transporting grain and of illegal imports and cane poaching. In the same way, other staples from surplus to deficit areas, especially if smallholder farmers who are members of KTDA often they must cross several counties to get there. sell to estate companies, attracted by their spot payment arrangements and, depending on the market, attractive Another source of risk within the sector’s enabling envi- prices. This has been a problem for some KTDA facto- ronment stems from the government’s active role in ries, which have operated well below their potential due regulating domestic food markets. These risks are most to an insufficient supply of green leaf tea. notable in the maize and sugar subsectors. In the case 40 Agriculture Global Practice Technical Assistance Paper of maize, uncertainty over the scope and timing of GoK in the fertilizer market has created a number of challenges interventions in grain markets poses a considerable risk to including (1) uncertainty over the timing of delivery and maize producers, traders, and other stakeholders across year-on-year support; (2) poor targeting of subsidies; (3) the maize supply chain. The common wisdom is that a late planting and high farmer dependency; and (4) lack of lack of storage prevails at the farm level and in rural com- a clear exit or sustainability strategy. munities. An alternative view is that strong disincentives exist to storing grain. This is because government policy In Kenya’s sugar industry, the unpredictability of cur- can cause unexpected adverse price shocks, such as tem- rent policy27 with regard to import regulations and porary suspension of the 50 percent tariff on non-EAC exceptions to the COMESA rules poses considerable maize imports or the release of grain from the Strategic risk to mills, cane producers, and other stakeholders. It Grain Reserve. In part because farmers and first buyers also impedes investments. Sizable unrecorded imports have no interest in storing grain that will enter commer- of refined sugar from outside the region pose yet another cial channels, little consideration is given to moisture con- risk to the industry. Imported sugar slated for industrial tent and quality. Rather, value chain participants each try use has been known to find its way into domestic markets to sell the grain before it deteriorates, leaving the responsi- for household consumption. In any case, prices can fall bility for drying and quality to the next buyer in the chain. precipitously when the market becomes saturated and Thus, much of the grain entering commercial channels is mills are unable to compete. For example, in 2002, the unfit for storage and subject to significant losses and qual- industry suffered considerable losses due to import surges ity deterioration (including aflatoxin contamination). and the failure among sugar millers to make payments to cane farmers and other suppliers. As a result, the entire Moving forward, it is unclear how Kenya will be able to sugar sector accumulated heavy debts. The high level of overcome recurrent maize production shortages, which indebtedness of state-owned mills, reportedly five times can jeopardize food security. Amid increasing import vol- more than their current assets combined, has helped stall umes, uncertainty exists over whether imports will be able needed reforms.28 to fill the gap in light of Kenya’s 50 percent ad valorem tariff for non–COMESA sourced maize, its import ban on Overregulation, lack of transparency, and asymmetric genetically modified (GM) maize, and inadequate supplies governance in Kenya’s coffee supply chain threaten the of non–GM-exportable maize in the COMESA region. coffee industry’s long-term viability. International buyers This is especially true in light of episodic export bans for and marketing agents wield excessive market power while maize in Tanzania, Malawi, and Zambia. Supply markets farmers are forced to absorb an oversized share of mar- have also been thinned in recent years by the growing ket as well as production risks. As a consequence, farm- attractiveness of the South Sudan market for Ugandan gate prices are consistently low, leaving farmers with scant maize exporters and of DRC markets for Tanzanian incentives to maintain proper crop husbandry within the maize exporters. context of aging trees, widespread disease, and declining productivity. Another major risk is government intervention in input Kenya’s livestock subsector has long been underfunded markets, such as for fertilizer. Kenya’s fertilizer market and faces a range of challenges ahead. Since most of was liberalized during the early 1990s when price and Kenya’s livestock herds are found in the ASALs, which marketing controls, licensing arrangements, import per- are both more exposed to impacts from natural disasters mits, and quotas were eliminated. The bulk of fertilizers (e.g., droughts, conflict) and more vulnerable than other are imported and distributed by the private sector. Since 2008, however, the government through the fertilizer sub- sidy program has procured about 494,000 MT of fertilizer 27 In February 2014, COMESA approved the extension for a further year of Kenya’s special safeguard arrangement for sugar, thus allowing Kenya to main- in support of the agriculture sector. In 2015, the govern- tain a 350,000-ton ceiling on duty-free sugar imports from COMESA. ment is projected to import and distribute 143,000 MT of 28 See “COMESA approves 1-year extension of Kenyan sugar safeguards,” Agri- fertilizers through NCPB. This government involvement trade, May 11, 2014. Kenya: Agricultural Sector Risk Assessment 41 BOX 3.1. KENYA’S DAIRY SECTOR—A CASE STUDY OF MARKET AND ENABLING ENVIRONMENT RISK Kenya’s dairy sector has only recently fully recovered from a 15-year crisis caused in part by mismanagement. Government ser- vices to the sector were reliable until the mid-1980s, but started failing in the 1990s and collapsed in early 2000 due to corruption and mismanagement in the cooperative sector. This caused a decline in milk handling of 266 million liters (estimated value $43 million). A failure to pay producers for milk in 1994 alone caused a loss in excess of $16 million. Amid failing public services, privatization of the sector and service delivery began in 1993, including artificial insemination. With privatization and recession in the sector, concentrate feed production also dropped during 1993–94. Counterparty risk in the milk sector due to manipulation of Kenya Cooperative Creameries (KCC) board members led to delays and failures in payment. In 2008, livestock concentrate and fodder (hay) prices increased by 40 percent and 100 percent, respectively, due to postelection violence and impacts from the global financial crisis, and in 2010, a substantial surplus in milk production caused market glut and farmer distress. FIGURE B3.1.1. MILK PRODUCTION IN THE FORMAL SECTOR (millions of liters), 1984–2008 450 Government intervention 400 350 300 250 Nonpayment 200 150 100 50 0 84 85 86 87 89 91 93 94 95 96 97 98 99 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 Source: FAO 2011. livelihood zones, public support to the sector often comes test and graze quarantined animals no longer exist or have in response to emergencies rather than to modernizing and become dysfunctional. A review of such policies and the restructuring the livestock sector. Thus, Kenya’s northern broader legal and regulatory framework is required. rangelands continue to suffer from insufficient infrastruc- ture, low education levels, and poor delivery of health care Another enabling environment risk to Kenya’s agriculture and other services. And some policies in place remain out- sector is linked to the country’s growing dependence on dated. Health quarantine laws, for example, hinder live- food aid (figure 3.11). During 2006–11, Kenya received stock trade as many of the facilities originally in place to $1.92 billion in emergency response aid, up from $150 FIGURE 3.11. HUMANITARIAN ASSISTANCE TO KENYA (US$, millions), 2000–11 $500 $400 $300 $200 $100 $0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: Central Bank of Kenya. Note: Weighted monthly average lending rate. 42 Agriculture Global Practice Technical Assistance Paper million during the prior five-year period (2000–04). As contest the rights of the first group to exclude strangers. evidenced elsewhere, frequent crises coupled with overre- Disputes of this sort have arisen and continue between liance on food aid can lead to a breakdown of household Somali and Boran pastoralists in northern Kenya, and resilience. This happens as household capital (i.e., savings, can give rise to extended conflict. Such conflict is most assets) is depleted to finance emergency coping mecha- likely to arise in a drought or other period of stress on nisms. When affected communities have insufficient time people and animals, when access to water and pasture is to rebound and replenish their resources, subsequent essential, and herds have to leave their normal grazing shocks further weaken their capacity to cope and with- areas to find them. stand future shocks. Although emergency food aid can help address immediate food needs, it does little to help CONFLICT AND INSECURITY IN ASALS rebuild household resilience and may induce higher rates Kenya’s low-density, sparsely populated northern coun- of dependency and chronic malnutrition. In this way, it ties and long borders provide both opportunities and can also increase the cost of managing future crises, and risks. Conflicts over natural resources, normally con- is thus not a sustainable way to manage food crises. For tained by customary rules, become much more difficult example, the cost of humanitarian aid needed to respond to settle when government staff, members of the armed to the famine in Niger in 2010 was twice that needed in forces, and wealthy businessmen become involved. The 2005 (Michiels, Blein, and Egg 2011). presence of jihadist fighters on the border with Somalia and nearby across the Gulf of Aden in Yemen adds a INSTABILITY complicating factor. Internal and cross-border conflict is a significant source of risk for Kenyan agriculture. The violence and civil In a regional drought, cattle prices are low but variable, unrest surrounding the December 2007 presidential elec- providing traders with an opportunity to provide a useful tions resulted in large-scale displacement and crop losses arbitrage service. Such arbitrage occurs for animals from as farming communities abandoned their fields. It also Somalia, Djibouti, Ethiopia, and northeastern Kenya. delayed land preparation and planting in the next season. The main problem for such traders is insecurity. Traders Partly as a result, losses in maize production alone were who buy animals cheaply in Somalia need an escort to an estimated $101 million, or more than 1.0 percent of get them to Kenya. This escort needs to be a local per- the value of agriculture GDP in 2008. According to a son from each place—locality or larger area—who is well study by the Centre for the Study of African Economies known and can persuade local people in each area to (CSAE) at Oxford University, impacts from the postelec- guarantee safe passage, provide protection, and provide tion violence in 2007–08 on the Kenyan flower industry contacts for the sale of the animals, all of which adds sub- reduced flower exports across the country by an estimated stantially to the cost. 24–38 percent. Political instability and conflict in neigh- boring Somalia can adversely impact agricultural markets, Interethnic conflict has occurred at a large scale during particularly livestock production and trade across Kenya’s elections since the early 1990s and up to those of 2007– northeast region. Since Kenya’s military intervention in 08. Clashes also occur regularly over water, grazing, and Somalia in 2011, Kenya has seen a rising incidence of land control in various counties, particularly in the ASALs attacks by the Somalia-based Al-Shabaab group, which but also in parts of the more fertile Rift Valley. Clan and claimed more than 14 bombings or armed attacks in 2012 tribal conflicts are frequent along borders (e.g., Pokot and alone. Kenya also witnessed a resurgence of interethnic Turkana districts, and Moyale in 2013). violence between August and December 2012. The insecurity triggered by jihadist movements in Soma- Customary land tenure systems are another source of lia has created a serious risk for all members of society in conflict, particularly in the ASALs, either when disagree- Kenya. The direct risk to life and livelihoods from jihad- ment arises within the system between different rights- ists is as likely to affect city dwellers in Nairobi and Mom- holders or when members of a different ethnic group basa as it is livestock owners in northern Kenya. In border Kenya: Agricultural Sector Risk Assessment 43 areas, animal health services can be disrupted due to Cattle rustling has long existed among pastoralists in the threats to Kenyan government staff, carjacking, and so on, region. Losses can be large or small, and local mechanisms and some traders, suppliers, and transporters may avoid for recovering stolen livestock and the resulting fines are well traveling to border areas. Livestock owners and traders developed within pastoralist communities. The risk of cat- are affected by the impact of antiterror operations and tle rustling has grown as it has become more commercial- tensions along the Somali border due to the Al-Shabab, as ized and protected by politically connected elites, rendering transaction costs, restrictions, and harassment by security traditional resolution mechanisms obsolete and ineffective. personnel are likely to increase. Well-organized, “commercial” cattle rustling has emerged since the early 2000s and was a notable problem especially during the election periods in the Rift Valley. CROP THEFT AND CATTLE RUSTLING Farmers and traders face the risk of both crop and cattle theft. The increase in market prices in recent years has MULTIPLICITY OF RISKS encouraged growing incidences of crop theft, particularly An important feature of agricultural risk is the high degree in maize-growing areas across Kenya. The risk is greatest of covariance of the main risk components. For example, around harvest time when harvested maize is left in the drought and animal disease commonly occur together fields for drying, but many farmers have reported losing when animals weakened by lack of feed and water subse- their crops to thieves even before harvest. Partly to save quently travel long distances and become more susceptible harvesting, drying, and storage costs, but also to protect to infection. In so doing, they end up spreading infections themselves from theft risk, some farmers sell green maize. over long distances. This happened, for example, in 2008– As happened in 2012, this practice can result in significant 10, when rinderpest returned to Kenya from South Sudan drops in farm-gate prices for both fresh and dried maize. and Somalia, where internal conflict had halted disease It also contributes to maize shortages by reducing the vol- control. Similarly, drought and price volatility and market ume of maize reserves. and enabling policy risks are often highly interrelated. 44 Agriculture Global Practice Technical Assistance Paper CHAPTER FOUR ADVERSE IMPACTS OF AGRICULTURAL RISKS 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 meth- odological basis described below is applied to production, market, and enabling environment risks. The various sources of risk are then reviewed to discern the most critical ones. CONCEPTUAL AND METHODOLOGICAL BASIS FOR ANALYSIS For the purposes of this study, risk is defined as an exposure to a significant financial loss or other adverse outcome whose occurrence and severity is 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 due to unexpected events such as suboptimal climatic conditions at different times in the production cycle and/or modest departures from expected output or input prices. Risk within the con- text of the current analysis refers to the more severe and unpredictable events that occur beyond these smaller events and that result in substantial losses to assets and livelihoods at the aggregate, sector level. This concept differs from the common perception of “risk” by farmers and traders based on year-to-year variability of production and prices. It should also be distin- guished from constraints, which are predictable and constant limitations to productiv- ity and growth faced by farmers and other agricultural stakeholders. In Kenya, these constraints include poor access to farm inputs, limited access to markets, limited credit, poor infrastructure, and the decreasing size of landholdings. LOSS THRESHOLDS As agricultural production is inherently variable, the immediate step for analysis is to define loss thresholds that 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 Kenya: Agricultural Sector Risk Assessment 45 TABLE 4.1. COST OF ADVERSE EVENTS FOR CROP PRODUCTION, 1980–2012 Indicative loss value US$, Percentage Year Description millions (%) GDP 1980 Regional droughts; 400,000 people affected; 728 MT maize loss −304.9 −3.2 1984 Drought in Lodwar, Eldoret, Narok, Dagoretti regions; 274 MT maize loss −258.3 −2.7 1993 Regional droughts in Eldoret; low rainfall in Lodwar, Nyahururu, Kisumu, Mombasa −243.2 −2.6 1996 Widespread drought in ASALs; 1.4 million people food insecure; 323K MT maize lost −291.9 −3.1 1997 El Nino floods; Rift Valley fever outbreaks; 1.5 million people affected (1997–98); −383.2 −4.0 157,000 MT maize loss 1998 El Nino floods; Rift Valley fever outbreaks; 1.5 million affected (1997–98) −266.6 −2−.8 2000 La Nina drought hits Garissa, Kisumu, Narok, Dagoretti; maize harvest drops by one-third −359.3 −3.8 2001 Drought in Mandera; poor rainfall in Lodwar, Malindi, Mombasa −223.2 −2.4 2002 Erratic rainfall; floods −242.6 −2.6 2004 Regional droughts in Gariss, Malindi, Mombasa; estimated 3 million households require food aid −255.9 −2.7 2009 Widespread drought across Eldoret, Nayhururu, Makindu, Mombasa; 592,000 MT maize loss −395.3 −4.2 2011 Regional drought in region of Mombasa −231.5 −2.4 2012 −284.0 −3.0 Source: FAOSTAT; authors’ calculations. Note: Cowpea losses were included from 1989 to 2012 due to data availability. Potato losses were calculated from 1980 to 2004 due to inconsistencies in data thereafter. ASAL, arid and semiarid land; MT, metric ton. between actual and expected production as a measure DATA SOURCES of loss. A loss threshold of 0.33 standard deviations Analysis of this nature requires a consistent set of data on from trend is then set to distinguish between losses due both production and prices for an extended time period. to adverse events and those that reflect the normal costs Of the various sources of data available, FAOSTAT’s of doing business. Those below threshold deviations from data series on the value of gross agricultural production trend allow estimation of the frequency, severity, and cost (1980–2012) and crop production (1980–2012) were con- of loss for a given time period (see Appendix H for illus- sidered the most suitable. These data allow the analysis of trations of indicative crop loss estimates). The frequency risk over a 33-year period. and severity of losses derived in this manner were also verified against historical records to ensure consistency with actual adverse events. CROP PRODUCTION RISKS Measured in terms of gross agricultural value,29 crop pro- duction in Kenya was significantly reduced by adverse THE INDICATIVE VALUE OF LOSSES events 13 times during the period 1980–2012, for an over- Available data on actual losses due to adverse events are all frequency rate of 39 percent (table 4.1). All of these not always accurate or consistent enough to facilitate the crop loss events on aggregate resulted in a drop in agri- comparison and ranking of losses. The analysis was thus cultural GDP of 2 percent or more. Losses ranging from based on estimates of the “indicative” value of losses, 3 to 4.2 percent occurred in six years. Indicative losses which provide a more effective basis for comparison. were substantial for these events, as would be expected, Indicative loss values are also compared to the value of whether measured in value of lost production or as a per- agricultural GDP in the relevant year to provide a relative centage of agricultural GDP. They amounted to nearly measure of the magnitude of loss. These estimates draw on actual data as much as possible, but it is emphasized Gross aggregate value is the total value of volume of production for each crop 29 that they represent indicative, not actual, losses. multiplied by the producer price. 46 Agriculture Global Practice Technical Assistance Paper FIGURE 4.1. INDICATIVE PRODUCTION LOSSES AND FREQUENCY FOR KEY CROPS, 1980–2012 1,200 Maize 1,000 800 Banana Tea 600 Coffee Potato Bea 400 Sugarcane Sorghum Wheat 200 Ric Cowpea 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 –200 Source: FAOSTAT; authors’ calculations. Note: Cowpea calculations were made using data from 1989 to 2012 due to data availability. Losses for tea and potatoes were calculated for the periods 1998–2012 and 1980–2004, respectively, due to inconsistencies in available data. TABLE 4.2. COST OF ADVERSE EVENTS BY their frequency. Maize accrued by far the biggest losses by value over the period, accounting for roughly one-fifth CROP, 1980–2012 (19.8 percent) of total indicative losses. Coffee and tea also Total incurred substantial losses due to their high market val- Total Losses Frequency Losses (US$, ues. Sugarcane and maize recorded the highest losses by Crop Rate (tons) Millions) volume, followed by banana and Irish potato. In terms of frequency, wheat production was adversely impacted on Maize 0.33 −3,752,514 1012.4 average every two to three years but in only two of these Rice, paddy 0.24 −140,325 115.1 years (1993, 2009) did associated losses amount to the Wheat 0.39 −691,113 250.9 equivalent of a 3 percent loss to agricultural GDP. Other Sorghum 0.33 −293,452 155.3 crops such as maize, sorghum, and cowpea were affected Beans, dry 0.36 −986,993 651.1 by adverse events on average every third year, with the Cowpeaa 0.33 −129,216a 83.9a Potato+ 0.40 −1,360,331 556.5 exception of rice paddy, potato, and coffee, which experi- Tea 0.39 −233,408 663.7 enced notable shocks every four years on average. Coffee, green 0.24 −132,596 630.4 Given the importance of maize production and the diverse Banana 0.33 −1,923,262 668.1 agroclimatic conditions in which Kenyan maize is grown, Sugarcane 0.36 −8,195,675 310.9 it is not surprising that maize is so vulnerable to produc- Total −17,435,580 5,098.3 tion shocks, or that aggregate losses are substantial. How- Source: FAOSTAT; authors’ calculations. ever, of the 11 risk events observed, three catastrophic a Cowpea calculations were made using data from 1989 to 2012 due to data availability. Losses for tea and potatoes were calculated for the periods 1998– event years (1980, 2008, 2009) accounted for nearly half 2012 and 1980–2004, respectively, due to inconsistencies in available data. (47 percent) of the total aggregate losses for maize over the 33-year period (figure 4.2). Estimated indicative losses $5.10 billion, or roughly $154.5 million on an average in 1980 (723,133 tons) and 2009 (605,926 tons) were likely annual basis, during the 33-year period. caused in large part by the severe droughts that affected Table 4.2 and figure 4.1 show the indicative costs of major maize production areas across Kenya in those years, adverse events by crop for the period 1980–2012 and while the postelection violence in 2008 is undoubtedly a Kenya: Agricultural Sector Risk Assessment 47 FIGURE 4.2. INDICATIVE CROP LOSSES FOR MAIZE, 1980–2012 2.5 Yield (MT/ha) Trend 0.3 trend 2.0 Yield (MT/ha) 1.5 1.0 0.5 Linear (Yield (MT/ha)) 0.0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT; authors’ calculations. FIGURE 4.3. PRIORITIZATION OF RISKS TO KENYA’S LIVESTOCK SECTOR Drought Severity of impact Price volatility ct, Disease theft, rustling Erratic rainfall Floods Frequency/Probability Source: Authors’ calculations. major factor behind the 386,792-ton observed loss in that Figure 4.3 provides a risk profile based broadly on the year. assessment team’s estimation of how frequently such vari- ous risks events have occurred over the period 1983–2011, LIVESTOCK RISKS and the severity of their impacts. Severity is measured on a hypothetical scale of 1–5 (where 5 is very high impact) The key risks in terms of probability and impact to and estimated based on how widespread the impact might Kenya’s livestock sector were identified as severe drought, be in terms of geographic spread or losses, bearing in price volatility, and cattle rustling. Animal diseases (such mind that cattle rustling and theft might be highly local- as FMD, Rift Valley fever, and anthrax) and floods were ized, while price volatility may occur frequently but can important but considered a lesser priority. Some of these have both positive and negative impacts among different risks are more relevant to ASAL production systems than stakeholder groups. The dearth of information on the to Kenya as a whole. Drought-induced price shocks, dis- impact of diseases, floods, off-season or erratic rains, and ease and resultant quarantine restriction movements, cattle rustling unfortunately limits a more precise prioriti- theft, and counterparty risk are more likely to threaten zation of risks based on actual financial losses. highland and mixed farming systems. 48 Agriculture Global Practice Technical Assistance Paper CHAPTER FIVE STAKEHOLDER VULNERABILITY ASSESSMENT Agricultural shocks are one important factor driving chronic poverty and food inse- curity in Kenya. Shocks impact household well-being by limiting food availability, weakening food access, and negatively affecting monetary well-being through the depletion of productive assets. Chronically vulnerable groups with high exposure to hazards experience a disproportionately large impact from adverse events and lack coping mechanisms available to other groups. In this context, vulnerability is a useful lens through which to examine agricultural shocks because it allows policy makers to determine which groups are most affected and to target risk management solutions accordingly. GENERAL TRENDS IN VULNERABILITY Some general considerations and trends concerning human development and vulner- ability in Kenya are as follows: » Levels of human development, poverty, and food insecurity vary widely between regions. » Exposure to extreme weather events is highly correlated with being poor and being food insecure. » About 70 percent of Kenya’s poor live in the central and western regions, in areas that have medium to high potential for agriculture (IFAD 2013). » Poverty and food insecurity are acute in the country’s ASALs, which have been severely affected by recurrent droughts. LIVELIHOODS AND AGROCLIMATIC CONDITIONS Especially in rural areas, patterns of livelihood activities are strongly influenced by the prevailing agroclimatic conditions, which determine planting calendars, soil qual- ity, and crop suitability. Approximately 80 percent of Kenya’s land area lies in the Kenya: Agricultural Sector Risk Assessment 49 FIGURE 5.1. HUMAN DEVELOPMENT INDEX SCORES, BY PROVINCE 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Nairobi Central Rift Valley Eastern Coast Western Nyanza North Nation Eastern Source: UNDP 2009. ASALs, home to more than 30 percent of the population and minimal nonfood items. Of these 17 million people, and 75 percent of the country’s livestock (GoK 2011). more than 85 percent live in rural areas. Spatial dispari- ASAL districts have the highest incidence of poverty in ties in both the incidence and depth of poverty are pro- the country, contain 18 of Kenya’s 20 poorest constituen- nounced: Poverty incidence at the district level ranges cies, and are predominantly inhabited by pastoralists and from 94 to 12 percent, and the poverty gap ranges from agro-pastoralists (GoK 2009c). Pastoralist districts con- 70 to 2 percent. According to an econometric analysis of sistently rank below the national average in terms of the district-level poverty data, stark spatial variations in the Human Development Index (HDI), as well as on other incidence and depth of poverty arise from differences in indicators of well-being (figure 5.1). These communities agroclimatic conditions and income-earning opportuni- are among the most chronically food insecure in the coun- ties, as well as unobserved factors (World Bank 2008). In try and typically experience the highest rates of severe other words, household location is an excellent predictor malnutrition. Several underlying factors increase pastoral of livelihood activity, poverty status, and household con- communities’ vulnerability, including land fragmentation, sumption level (figure 5.2). population growth, low literacy and education provision, poor infrastructure, and weak market integration. These Districts with high levels of poverty and food insecurity chronic weaknesses undermine pastoralist groups’ capac- are also characterized by a high frequency of extreme ity to respond to shocks like drought and livestock disease weather events. Households in the bottom expenditure outbreaks, which occur frequently in the ASALs. In turn, quintile are the most likely to experience a weather- the increasing frequency and simultaneous occurrence related shock. By virtue of their location, poorer house- of multiple shocks erode the effectiveness of traditional holds experience a variety of natural hazards more coping mechanisms, creating a vicious cycle of crisis and frequently compared to better-off and richer households, underdevelopment. and are less able to mobilize productive resources to respond to shocks. POVERTY AND Figure 5.3 shows the percentage of severely food insecure VULNERABILITY households as of May 2013 in areas where the World Food In 2005–06, approximately 17 million Kenyans, or 47 per- Programme (WFP) operates. The graph reflects the food cent of the population, were too poor to buy enough food security status of nonbeneficiary households, as opposed to meet the recommended daily nutritional requirements to WFP-beneficiary households. 50 Agriculture Global Practice Technical Assistance Paper FIGURE 5.2. MAP OF KENYA’S LIVELIHOOD ZONES KE01 - Northwestern Pastoral Zone KE02 - Turkwell Riverine Zone KE03 - Northwestern Agropastoral Zone KE04 - Lake Turkana Fishing Zone KE05 - Northern Pastoral Zone KE06 - Marsabit Marginal Mixed Farming Zone KE07 - Northeastern Agropastoral Zone KE08 - Mandera Riverine Zone KE09 - Northeastern Pastoral Zone KE10 - Grasslands Pastoral Zone KE11 - Southeastern Pastoral Zone KE12 - Coastal Low Potential Farming Zone KE13 - Coastal Marginal Agricultural Mixed Farming Zone KE14 - Tana Delta - Irrigated Zone KE15 - Coastal Medium Potential, Mixed Farming Zone KE16 - Southern Pastoral Zone KE17 - Southeastern Marginal Mixed Farming Zone KE18 - Southeastern Medium Potential, Mixed Farming Zone KE19 - Southern Agropastoral Zone KE20 - Central Highlands, High Potential Zone KE21 - Western Medium Potential Zone KE22 - Western High Potential Zone KE23 - Western Lakeshore Marginal Mixed Farming Zone KE24 - Lake Victoria Fishing Zone KE25 - Western Agropastoral Zone Source: FEWSNET 2011. FIGURE 5.3. PERCENT OF SEVERELY FOOD INSECURE, NON-WFP BENEFICIARY HOUSEHOLDS BY LIVELIHOOD ZONE 62 45 41 35 32 21 21 16 4 Coastal Eastern Grassland NE Northern NW SE Southern Western marginal pastoral pastoral pastoral pastoral pastoral marginal pastoral agropastoral Source: WFP 2013. that preclude access to resources for individuals, house- VULNERABILITY AMONG holds, or livelihood groups. The groups identified below LIVELIHOOD GROUPS are especially vulnerable to agricultural shocks: Certain population groups and certain types of house- holds are more vulnerable to agricultural shocks than others, depending on their level of exposure to risks, sus- PASTORALISTS ceptibility, and capacity to respond and/or recover from » Pastoralist households are more likely to be poor adverse events. In many cases, patterns of vulnerability and more likely to be food insecure than nonpasto- reflect underlying inequalities and social marginalization ralist households. Kenya: Agricultural Sector Risk Assessment 51 » Up to a fifth (15–20 percent) of households in the Northern Pastoral Zone engage in begging, a rate RISK MANAGEMENT much higher than in any other livelihood zone. CAPACITY » The highest rates of global acute malnutrition are The capacity to manage risks among different stakehold- in the Northeastern and Northwestern Pastoral ers varies widely, based on myriad factors. These include Zones (WFP 2013). prevailing production systems, household income levels, and diverse income sources. Assessing levels of vulnerabil- ity among specific groups thus requires an understand- FEMALE-HEADED HOUSEHOLDS ing of their level of exposure and their risk management (FHHS) capacity. This includes their capacity to cope with and » FHHs are 13 percent less likely to be food secure recover from resulting losses. than male-headed households (MHHs). » FHHs, on average, have smaller farm sizes and About 84 percent of Kenyan farmers own less than 3 ha lower education levels compared to their male of land, and 45 percent own less than 1 ha (World Bank counterparts. 2012). This statistic is supported by findings in panel sur- » Roughly 49 percent of the total cultivated land veys by the Tegemeo Institute, which found that 69 per- owned by MHHs is good to medium fertile land cent of smallholder farmers cultivate 1.5 or less (Tegemeo compared to 39 percent of land owned by FHHs 2010). It is difficult to grow enough food crops on these (Kassie et al. 2012). small farms to feed a family for an entire year and/or produce enough income to meet the household’s basic UNSKILLED/CASUAL WAGE LABORERS needs. Most families on these small farms do not achieve » Casual wage laborers are considered particularly subsistence, but must sell their labor or find other sources vulnerable to food price, production, and labor of income. Many end up selling a portion of their food shocks since they purchase almost all of their food crops to obtain income for critical needs (sickness, school from the market. fees, debts, and other social obligations). Many maintain » During the 2008 food crisis, labor demand and livestock as an additional source of revenue. If they have wage rates stagnated as food prices rose by up to access to a reliable market, many grow some horticultural 50 percent (KFSSG 2008). crops, which can be sold at prices that tend to be higher than those for cereals. VULNERABILITY AND RISK With regard to subsistence, most smallholder farmers in MANAGEMENT Kenya are net buyers of maize (Buy only + Net buyer in table 5.1). Since maize is the most important food crop, The distribution of losses among stakeholders within a this strongly supports the contention that the majority of supply chain is to a great extent a function of supply chain smallholder households do not achieve subsistence. A very governance and stakeholders’ capacities and opportunities high proportion of Kenyan farmers (possibly as high as for risk management. The way stakeholders are affected 80 percent) are resource-poor, low-input, low-output crop depends greatly on their level of vulnerability, as defined and livestock producers. Thus a majority of smallholder by their socioeconomic situation, gender, and location, farmers are net buyers of food staples and are highly among other factors. All actors along a supply chain are dependent on the market for the purchase of food items. exposed to the variability in primary farming production. However, smallholder farmers and their families are par- ticularly and very strongly affected by production risks. VULNERABILITY IN ASALS They are the weakest segment in the supply chain, and Pastoralist communities in arid lands long ago devel- the prevalence of risks contributes to the tightening of the oped their own drought-coping and adaptation strate- vicious cycle of poverty. gies, but those are no longer effective. One reason is 52 Agriculture Global Practice Technical Assistance Paper TABLE 5.1. HOUSEHOLD CHARACTERISTICS ACCORDING TO POSITION IN THE MAIZE MARKET, 1997, 2000, AND 2004 (nationwide sample of small-scale households in Kenya) Household market position (% of households) Neither Buys Sell Only Buy Only Net Seller Net Buyer Net Equal nor Sells Total Characteristic (n = 781) (n = 2052) (n = 467) (n = 242) (n = 18) (n = 412) (n = 3972) % of total sample 19.7 51.7 11.8 6.1 0.5 10.4 100 Land Cultivated (Ha) 3.0 1.1 1.9 1.2 1.0 1.5 1.6 Source: Kirimi et al. 2011. that drought intensity and frequency have increased, TABLE 5.2. HOUSEHOLDS’ while political marginalization and chronic underde- PRIORITIZATION OF velopment of pastoralist communities, characterized by RISKS IN ASAL COUNTIES lack of basic education, infrastructure, and health, have Importance greatly reduced their capacity to adapt and their resil- Threat (weighted score) ience to shocks. Drought 21 Pastoralists’ and other livestock keepers’ ranking of their Animal disease 11 main risks is largely shared by administrators. A rapid Market disruption 17 informal survey by the National Drought Management Conflict 8 Authority (NDMA) in four pastoral counties (Turkana, Predation 3 Kitui, Kajiado, and Laikipia) in 2013 showed that some Wild fire 2 variant of the four most important threats—drought, ani- Floods 2 mal disease, market disruption, and conflict—was cited in Policies and institutions 2 all ASAL counties, although their priorities differed (GoK Source: NDMA; authors’ calculations. Note: Impacts weighted as follows: catastrophic, 5; critical, 4; consid- 2013b). A simple weighting procedure shows how these erable, 3; moderate, 2; negligible, 1. Maximum possible score is 40. risks are assessed against each other (table 5.2). Kenya: Agricultural Sector Risk Assessment 53 CHAPTER SIX RISK PRIORITIZATION AND MANAGEMENT RISK PRIORITIZATION This assessment has highlighted the Kenyan agriculture sector’s key inherent risks, which are both numerous and complex. They manifest with varying levels of fre- quency and severity and can result in substantial losses to crops and livestock, which can have profound short- and long-term impacts on income and livelihoods. Putting in place effective risk management measures can help to reduce agricul- tural stakeholders’ vulnerability and the impacts of related shocks to production and marketing systems. In resource-constrained environments, however, it is virtu- ally impossible to address all risks at once. Thus, a prioritization of interventions is needed to address the risks that occur most frequently and that cause the greatest financial losses. Chapters 3 and 4 identified priority risks using quantitative measures and anecdo- tal evidence collected directly from crop and livestock subsectors stakeholders. Due to the paucity of data, some risks could not be quantified so the assessment team relied more on qualitative measures. Based on the team’s combined quantitative and qualitative assessment, figure 6.1 prioritizes the most important risks for focus crop and livestock subsectors. These were validated at a roundtable at MoALF in Nairobi on February 7, 2014. The figure shows a summary of the agricultural risks sorted on the basis of the probability of each event and its anticipated financial losses. The grayer the area, the more significant is the risk. Overall, this prioritiza- tion identified the most important risks facing Kenya’s agriculture sector to be (1) severe drought, affecting both crop and livestock production; (2) price volatility; and (3) crop and livestock diseases. Erratic rainfall, floods, cattle rustling, and conflict were also deemed important, but to a lesser extent. Kenya: Agricultural Sector Risk Assessment 55 FIGURE 6.1. PRIORITIZATION OF KEY AGRICULTURAL RISKS IN KENYA Severity of impact Negligible Moderate Considerable Catastrophic q y probable – Striga (M) – Erratic rainfall – Power outages (C) – Stemborer (M) p – Foot & Mouth Disease Hiqhly – Cattle rustling – Hailstorms (T) – Erratic inputs (i.e., quality) – Severe drought – Stalk borer (S) – Unpredrictability of – Price volatility Probability of event Probable – Termites (S) SGR release (M) – Crop/livestock diseases – Theft (M, C, T) – Lifting import tariff (M,S) – Foot & Mouth Disease (L) – CLR/CBD (C) – Policy risk (S) – Regional drought – Frost (T) – Flooding – Maize lethal necrosis Occassional – Untimely input credit (C) – Sugarcane mosaic virus – Maize streak virus – Power outage (C) – Maize chlorotic virus – Ratoon stunting disease – Maize dwarf – Theft (M,L) mosaic virus – Price shock – Drought (S) – Conflict Remote – Windstorms (M, W, S, B) – Anthrax (L) – Thrips (tea) Key: Maize, Wheat, Beans, Cowpeas, Irish Potato, Tea, Coffee, Sugarcane, Cut flowers, Bananas, Livestock Source: World Bank. nisms typically trigger compensation in the case of RISK MANAGEMENT a risk-generated loss (e.g., purchasing insurance, MEASURES reinsurance, financial hedging tools). No single measure can manage all risks. Effective risk 3. Risk coping (ex post): Actions that will help management requires a combination of coordinated affected populations overcome crises and build measures. Some are designed to remove underlying con- their resilience to future shocks. Such interven- straints. Others are designed to directly address a risk or tions usually take the form of compensation a subset of risks. Available resources are often a limiting (cash or in-kind), social protection programs, and factor, but integrated risk management strategies are often livelihood recovery programs (e.g., government more effective than one-off or stand-alone programs. Risk assistance to farmers, debt restructuring, contin- management measures can be classified into the following gent risk financing). three categories: 1. Risk mitigation (ex ante): Actions designed Table 6.1 highlights potential interventions that could to reduce the likelihood of risk or to reduce the help address the key risks identified by this assessment, severity of losses (e.g., water harvesting and irriga- classified by the three types of risk management meas- tion infrastructure, crop diversification, extension). ures described above. The list is by no means exhaustive, 2. Risk transfer (ex ante): Actions that will trans- but it is meant to illustrate the type of interventions that, fer the risk to a willing third party. These mecha- based on the analysis, have good potential to improve 56 Agriculture Global Practice Technical Assistance Paper TABLE 6.1. INDICATIVE RISK MANAGEMENT MEASURES Mitigation Transfer Coping Drought Promote diversification toward more Macro-level crop Facilitate temporary migration or permanent (crops) drought-tolerant crops (e.g., cassava) insurance relocation Establish community-level food banks Farm-level crop Develop cash-for-work and food-for- insurance work programs to support soil and water conservation Promote adoption of soil and water Expand social safety net programs (e.g., food- conservation/natural resource management for-work) techniques Improve availability of existing drought- Use weather index for triggering early resistant seed varieties warning and response Strengthen input delivery systems and Use a decentralized disaster contingent fund ensure quality inputs for rapid response to local emergencies Improve farming techniques (e.g., Promote household/community savings and conservation agriculture, intercropping) informal credit Promote water harvesting and irrigation Strengthen early warning systems and response Drought Develop effective and environmentally Sovereign Buy fodder, crop residues; supplementary (livestock) appropriate systems of water harvesting, agriculture risk feed, emergency stores management, and irrigation financing Link early warning system to rapid reaction Index-based Ensure emergency water supply; use fuel and relevant response (e.g., tracking strategy livestock subsidies and repair boreholes and LEGS1 programming) insurance Improve access to emergency grazing Build water pans (via safety net programs) Invest in climate proof livestock sector Support exceptional livestock movements infrastructure Promote haymaking and storage, reserve Support conditional parks grazing/wildlife/ grazing/standing pastures; irrigated fodder livestock coexistence (and food) production incorporating stakeholder and pastoralist ownership and interests Enforce tougher screening at international Develop multiyear food and cash mechanisms borders based on early warning and food security data Develop plan for long-term subsector support Support livestock destocking–commercial and and new livelihood growth program GoK purchases for a fixed price, with animals slaughtered and meat distributed among most needy households/communities Support conditional parks grazing/wildlife Reconstruct destroyed assets with improved, livestock coexistence climate-resilient standards Promote herd diversification (continued) Kenya: Agricultural Sector Risk Assessment 57 TABLE 6.1. INDICATIVE RISK MANAGEMENT MEASURES (continued) Mitigation Transfer Coping Pests and Where relevant, adapt policy to arid lands and Warehouse Conduct strategic livestock vaccination in diseases ensure appropriate, affordable service delivery receipts systems response to outbreaks (crops and livestock) Establish private, quality, comprehensive Strengthen quarantine measures/mechanisms animal health care facilities Intensify and strength disease surveillance Implement proven and approved veterinary response interventions (LEGS, OIE) Improve animal health through increased Promote farmer group-operated storage uptake of vaccination campaigns centers Train (farmers and local officers) on IPM, fumigation, and pre- and postharvest management Apply/enforce moisture and grain quality standards Promote use of hermetically sealed storage sacks and silos Price Improve producers’ access to market Commodity Improve efficiency of emergency grain reserve volatility information hedging (crops and livestock) Raise the levels of strategic food reserves to Warehouse Promote market subsidies and commercial stabilize maize prices receipt systems destocking Develop and expand livestock markets Use transport subsidies Improve market infrastructure Destock livestock Develop policy on livestock marketing to fit global standards and local conditions (farm to fork) Train and build capacity of producers and officers towards market orientation and opportunity Conduct international and regional planning Exploit value chain niche markets and develop cross-sector linkages Link different sectors of the value chain Cattle Promote good governance and implement Restock livestock rustling existing laws Address conflict: reinforce customary Support social safety nets mechanisms and create joint customary/ formal mechanisms Support community peacekeeping programs Source: Authors’ notes. 1 The Livestock Emergency Guidelines and Standards (LEGS) provide a set of international guidelines and standards for the design, implementation, and assessment of livestock interventions to assist people affected by humanitarian crises. Established in 2005, the LEGS Project is overseen by a Steering Group of individuals from the EU, FAO, the International Committee of the Red Cross, the Feinstein International Center at Tufts University, the World Society for the Protection of Animals, and Vetwork UK. 58 Agriculture Global Practice Technical Assistance Paper agricultural risk management in Kenya. Unlike drought In 2009, Kenya pioneered a drought management contin- or livestock diseases, which have a generally negative gency planning system, managed by the newly established impact on almost everyone, price risk may affect certain NDMA, which consists of (1) an early warning system; stakeholder groups differently. For example, the release (2) a rapid reaction capability, including contingency of maize stocks by NCPB can be considered both a risk funding for urgent mitigation activities such as destock- to traders with large inventories and a windfall for rural ing and emergency food aid; and (3) a recovery program and urban households. It is also worth noting that many to assist those most affected to get back on their feet and of these interventions, if implemented concurrently, can strengthen their resilience to future droughts. The NDMA help address multiple risks at once, with positive spillover provides a platform for long-term planning and action, effects across the sector. as well as a mechanism for coordination across GoK agencies and with other stakeholders. For financing early mitigation efforts to reduce the time between warning of ILLUSTRATIVE RISK drought stress and responses at district and national levels, MANAGEMENT MEASURES the GoK and its donor partners established a multidonor The following section provides a brief description of basket fund, the National Drought and Disaster Contin- broad areas of intervention (encompassing some of the gency Fund (NDDCF). measures highlighted above) with scope to address the most important risks impacting Kenya’s agriculture sector. The economic case for early reaction to impending drought in the ASALs is strong. In a modeling exercise, STRENGTHENING RESPONSE AND various mitigation activities were costed against the ben- RESILIENCE TO DROUGHT efits that would accrue (Venton et al. 2012). The results Drought sets off a vicious cycle of adverse socioeco- were encouraging. For example, the benefit-cost ratio for nomic impacts. It begins with crop-yield failure, unem- destocking was 390:1; that is, every U.S. dollar spent on ployment, income disruptions, depletion of assets, commercial destocking generated $390 in benefits (meas- worsening of living conditions, and poor nutrition. It ured in food aid and animal losses avoided). The ratio for often ends in decreased coping capacity among affected building resilience—via ensuring that pastoralists have communities, and thus, increased vulnerability to sub- access to functioning livestock markets, veterinary care, sequent shocks. Early warning and early response cou- and adequate feed and water—was also positive, although pled with effective coordination and coherence in both to a much lesser extent (2.9:1). According to the study, a short- and long-term efforts to safeguard livelihoods full package of livestock interventions that build resilience and promote future resilience are critical to effective would result in $5.5 of benefits for every $1 spent. Addi- drought risk management. tional benefits such as improved functioning of livestock markets and more effective animal health would likely Many drought mitigation activities have been tried over accrue in nondrought times as well. the years in northern Kenya, beginning in the colonial era when government-sponsored destocking of vulnerable Scope now exists to review Kenya’s recent experience in stock took place on a large scale through the activities of responding to drought emergencies under the NDMA the Livestock Marketing Division (LMD). More recently, regime. The objective of such a review would be to identify this approach was superseded by more limited, ad hoc existing operational, institutional, and financial barriers responses designed to mitigate the impact of drought via that impede more rapid and effective response measures. the provision of ex post emergency assistance to affected It could also identify potential synergies with other pro- communities. Such measures are often poorly coordinated grams and avenues for more effective data monitoring, at the local or national level. Until recently, few measures information sharing, and coordination of interventions in have been taken to improve pastoral communities’ self- the future and innovative ex ante approaches to building sufficiency or to assist development of community-man- resilience. Efforts to strengthen existing drought manage- aged drought mitigation activities. ment systems might include development and adoption of Kenya: Agricultural Sector Risk Assessment 59 a common approach to using triggers to better ensure that In northern Kenya, customary systems of land use persist. decision makers know exactly what they ought to be doing These can be quite simple, with a minimum of rules to as the situation deteriorates and the consequences if they be followed, but all customary grazing systems have rules fail to act. Ideally these triggers should be developed with about who may graze what area of land in what season, input from all stakeholders, be context specific to account and under what conditions. The most sophisticated of for different livelihood zones, and avoid facilitating inter- these sets of rules is probably that of the Boran pastoral- ventions that undermine communities’ existing and future ists in northcentral Kenya; it includes detailed rules for capacity to cope. pasture use in different seasons, the preservation of emer- gency drought pastures, and access to wells and water use, Future initiatives might also expand on existing Food-for- as well as rules for dealing with strangers and passers-by. Assets (FFA) and Cash-for-Assets (CFA) projects, currently The Boran have been able to preserve the key features led by the WFP, designed to promote food security and of this system of natural resource management, but the reduce levels of vulnerability. These activities range from system is under constant pressure from the rest of Kenya, rainwater harvesting for human and livestock use and soil where customary resource management systems of that and water conservation, to rehabilitation of degraded complexity are rare and not well understood, and where agricultural land and production of drought-tolerant land privatization is extensive. Similarly, large water and crops. Through new or rehabilitated assets and develop- oil resources have recently been found in pastoral areas; ment of relevant skills, communities can improve their it is vital that exploitation of these resources takes into resilience to weather-related shocks and invest in a more account local needs and that stakeholder participation is sustainable future. included in future planning. IMPROVING LIVESTOCK MOBILITY Finally, as cross-border movement and trade are sig- Drought’s effects are exacerbated by a number of fea- nificant and vital to Kenya’s livestock sector, regional tures associated with the local livelihood system and approaches and mechanisms for policy implementation national policy. Land tenure is among the most impor- must more effectively ensure that livestock owners have tant of these. Policy on land ownership and tenure, and ready access to cross-border markets and services. The user rights, must take note of the likely continued need for proposed Regional Pastoral Livelihoods Resilience Pro- mobility in ASAL livestock systems. The erosion, and in ject (RPLRP) is a step in the right direction. In Kenya, places destruction, of traditional grazing systems creates the project is being implemented in 14 ASAL counties new risks where there were fewer before. In the Maasai (Lamu, Isiolo, Laikipia, Mandera, Marsabit, West Pokot, areas of southern Kenya, subdivision of previously com- Turkana, Tana River, Garissa, Baringo, Samburu, Narok, munally held land, and in many places sale of land, has Samburu, and Wajir), all of which have transboundary had a powerful influence on the way livestock are grazed. stock routes linking pastoral communities on either side Some Maasai have been able to preserve seasonal pastures of the border with Somalia, Ethiopia, and the Sudan. which they use in rotation, including dry season reserves. Meetings of officials from both sides and technical peo- In areas where fencing of subdivisions of former Maasai ple, together with pastoralist leaders, can help stakehold- grazing land has gone quite far, a new risk has been cre- ers avoid problems and conflicts. ated: because of the fences, herders are unable to move their animals away at the first sign of pasture or water STRENGTHENING PASTORALISTS’ shortage. Movement away from a risk is a key response in LIVELIHOODS drought; pastoralists move to their own reserve pastures, Livelihood development initiatives in the ASALs should to the land of kin and allies, and to government land. The be based on proven, evidence-based research carried places pastoralists move to in times of crisis need to be out over a reasonable time frame (three to five years) to better protected and managed. Stronger recognition of ensure sustainability and economic viability. Too often, the importance of mobility is a key part of any pastoral academic or modern innovations for temperate or tropi- drought management and development policy. cal animal production are imposed on extensive mobile 60 Agriculture Global Practice Technical Assistance Paper systems, and most tend to fail as they are not suitable. produces quality products for value-added market oppor- Governments, organizations, and farmers have all suf- tunities. Innovative ways to protect the TOT of livestock fered losses due to failure to test new approaches in and staple foods and services in the pastoralist system must the pastoral sector context. To ensure successful policy be investigated. Price variations regularly reach 100 per- making and wide uptake at the grassroots level, it is rec- cent even in normal years due to seasonality and market ommended that customary law be considered in policy supply and demand. TOT can drop by 300 percent or mechanisms. Full stakeholder participation and owner- more in shock years, which regularly occur every three to ship are encouraged in the formulation and implementa- four years, and preventative measures are needed to avoid tion of policies and strategies. price collapse and food insecurity. Market information systems (MISs) for enhanced price transparency could be developed. Effective actions include collective cereal or IMPROVING ANIMAL HEALTH AND commodity storage, credit systems, insurance, and early VETERINARY SERVICES marketing. Years of underinvestment and neglect mean Private sector provision of affordable, quality veterinary that much more needs to be done to establish or improve services is economically infeasible, especially in exten- the infrastructure essential to markets, especially road, sive production systems in remote arid regions, due to a transport, and communication systems. This includes number of well-documented constraints. Efforts must be ensuring adequate resources and qualified personnel are made, supported by reforms where necessary, to ensure availed to the livestock marketing and animal production nascent private sector initiatives are allowed to prosper. departments of MoALF and innovative but proven qual- To succeed, any approach will have to include the decen- ity services are provided. tralized Community-Based Animal Health Worker sys- tem. Efforts should be explored to support and strengthen the system with the use of vouchers or via partnerships IMPROVING SECURITY IN ASALS with the private sector in emergency relief. The veterinary Laws and legal frameworks already exist for dealing with department must be assured of adequate resources and theft and cattle rustling. Due to corruption and ineffi- qualified personnel so that quality and timely services are ciency, however, a sense of impunity and a breakdown available when needed. In remote areas, it is often difficult of law and order prevail in Kenya. Reported incidences for a young graduate veterinarian or livestock professional in which members of the security services and armed to operate at the level at which he or she has been trained, forces have been involved in livestock theft and reports as local resources and basic services are limited. This is of political protection facilitating large-scale or “commer- a source of demotivation and leads to lack of field pres- cial” cattle raiding during past elections do little to con- ence, access to farmers, and service provision. As needs vince livestock owners that a solution or response to theft are great and government services are likely to remain is imminent. The solution lies in good governance and overstretched and underfunded in many remote counties, ensuring that police and security forces are adequately the GoK should establish an enabling environment to equipped and motivated to fulfill their duties and pro- incorporate proven and recognized animal health service vide services as intended. As the majority of government providers (e.g., SIDAI franchisees) to provide services at services are city or town based, it is recommended that scale even during emergencies. community involvement in ensuring security and peace- building is incorporated and expanded in more remote or rural areas where feasible. Kenya’s northern borders IMPROVING LIVESTOCK MARKET with other ASAL countries are far too long to be effec- INFRASTRUCTURE tively policed by the army or the police force. They can Markets are increasingly important for livestock enter- be watched, however, by pastoralists living in border areas prises. An efficient pastoral livestock marketing system and legitimately occupying territory that they know better needs to be developed, where stock can be finished on than anyone else. In the process of rehabilitating pastoral feedlots or ranches and dividends paid to producers to livelihoods, it would be easy to design this role, for which encourage a more commercial market orientation that pastoralists could be compensated. Kenya: Agricultural Sector Risk Assessment 61 STRENGTHENING EXTENSION SERVICES production risks drives their ex ante decisions and dis- Reliable farmer access to extension services is an integral courages them from investing in fertilizers, improved part of any agricultural risk management strategy. It can seeds, and better crop husbandry practices. Irrigation also produce positive spillovers. Adoption of improved infrastructure build-out is costly and not suitable for many practices (e.g., conservation agriculture, IPM), drought- areas where long-term access to ground-water is uncer- and disease-resistant seeds, and other innovations can at tain. Alternatively, water harvesting and improved soil once help farmers lower their risk exposure and their costs management offer a sustainable and cost-effective way to while enhancing their productivity. Figures vary on the favor investments in nutrients and other yield-enhancing proportion of Kenyan farmers accessing public extension practices. Water harvesting alone—via water pans, roof services. The World Bank estimates that 50 percent of and rock catchment systems, subsurface dams, and other farmers now have access, but anecdotal evidence collected means—has been reported to provide between a 1.5- to during this study suggests that in some districts, at least, 3-fold increase in yields in Kenya, as elsewhere in Burkina access is much lower. The government currently allocates Faso and Tanzania (Hatibu et al. 2006; Kayombo, Hat- roughly 25 percent of its agriculture budget to extension ibu, and Mahoo 2004; Ngigi et al. 2005; Rockstrom, Bar- services—a relatively high amount compared to other ron, and Fox 2002). African countries. Yet service delivery is still recognized as inadequate. The increased reliance on private extension The Kenya Rainwater Association (KRA) has been work- providers means that extension services are often skewed ing to promote rainwater harvesting and complemen- toward well-endowed regions, bypassing poorer farm- tary technologies since the mid-1990s. These and other ers, and can lack sufficient state funding to ensure their initiatives such as MoALF’s Water Harvesting for Food effectiveness in meeting farmers’ needs. Paying for ser- Security Programme (WHFSP) should be supported and vices is beyond the reach of most poor farmers. Moving expanded to assist more farmers in mitigating risks asso- to “demand-driven” services requires further state invest- ciated with erratic rainfall and drought. Low-head drip ment in building farmer’s organizations’ capacity, because irrigation offers farmers a flexible system that is relatively the poorest farmers are not currently organized and are easy and affordable to install, operate, and maintain. This poorly positioned to demand and/or pay for services. and other systems can be scaled up in size to accommo- date larger dimensions and enable farmers to gradually Strong scope exists within the framework of Kenya’s increase their crop production over time. Through the National Agricultural Sector Extension Policy (NASEP) to Kenya Horticulture Competitiveness Project, KRA with explore broader use of new information and communica- its partners has trained 2,200 growers in eastern Kenya tion technologies (ICTs) to disseminate targeted information in water-harvesting techniques and has established 60 and knowledge more cost-effectively and provide needed water ponds that are proving reliable water sources for training to more farmers. The ICTs can amplify the efforts more than 4,750 farmers in the region. Likewise, better of extension and advisory services providers in disseminat- soil management through increased use of organic mat- ing agricultural information to remote locations and diverse ter, composting, demi-lunes, zero tilling, and other conser- populations. They can greatly facilitate the delivery of near vation agriculture techniques can help to increase water real-time information on weather, market prices, disease retention capacity while restoring soil nutrients and soil and pest outbreaks, and the availability of services, allowing health. These and similar initiatives should be promoted farmers to make more informed decisions on what to grow more widely via public and private extension to benefit and how best to grow it. In doing so, they can also help build more farmers. farmers’ capacity to manage production and other risks. STRENGTHEN SEED DISTRIBUTION IMPROVED WATER AND SOIL SYSTEMS MANAGEMENT Farmer adoption of improved drought- and other stress- In Kenya, where access to irrigation remains limited, tolerant maize varieties can go a long way toward reduc- farmers are at the mercy of rainfall. Perception of high ing weather-induced production risks, but farmers should 62 Agriculture Global Practice Technical Assistance Paper also be encouraged to invest in other yield-enhancing consider them too risky. This study notes how agricultural inputs and practices. More broadly, these practices could insurance, when combined with other, more traditional help improve Kenya’s food security situation by lowering risk mitigation and coping measures, can greatly reduce the country’s year-on-year maize production variabil- the immediate losses and long-term development setbacks ity. Broad farmer adoption depends on access, however. farmers absorb from agricultural risks. Insurance can also Maize seed research in Kenya as elsewhere in the region help to lower borrowing costs, thereby enhancing farm- is ongoing, and new drought-resistant varieties are find- ers’ access to needed credit. ing their way into the market. The Drought Tolerant Maize for Africa (DTMA) Project, funded by the Bill and The GoK is committed to expanding farmers’ access to Melinda Gates Foundation, promotes the development agricultural insurance. Recognizing the importance of and dissemination of drought-tolerant, high-yielding, Kenya’s livestock sector, MTPII (2013–17) calls specifi- locally adapted maize varieties in Kenya and a dozen cally for establishment of a National Livestock Insurance other countries in Africa. During 2007–12, participants Scheme. This initiative will build on the experience of marketed or otherwise made available 60 drought-tol- two innovative insurance programs already underway in erant hybrids and 57 open pollinated varieties to small- Kenya. Kilimo Salama is an insurance scheme that pro- holder farmers. Such efforts should be further supported tects farmers’ investments in seeds, fertilizers, and other and scaled up. inputs via payouts when experts monitoring local weather conditions and rainfall determine that crops have become The International Wheat and Maize Improvement unviable. In northern Kenya, the ILRI-led Index Based Center reports that the drought-tolerant maize germ- Livestock Insurance (IBLI) uses satellite images of veg- plasm developed for Africa in collaboration with IITA etation to determine when scarce pasture is likely to lead allows a yield increase of 40 percent over commercial to animal mortality, triggering automatic payments to varieties, under severe stress, and an equal yield level insured livestock keepers. under optimal cropping conditions (Cenacchi and Koo 2011). In addition to drought tolerance, the new varieties The GoK has already expressed interest in setting up a and hybrids also possess desirable traits such as resistance public-private partnership (PPP) platform to scale up pro- to major diseases (e.g., maize streak virus, Turcicum leaf grams for livestock insurance that will enhance the resil- blight, and gray leaf spot) and superior milling or cook- ience and reduce the vulnerability of small-scale pastoral ing quality. Despite these advances, adoption rates among farmers. With support from the World Bank’s Agriculture Kenya’s maize farmers remain low due to limited avail- Insurance Development Program (AIDP), IRLI, and oth- ability and farmer awareness. To have an impact, seeds ers, the State Department of Livestock (SDL) is explor- must be available at the right time, at the right place, and ing scope for the development of a macro-level livestock at the right price. Ways to incentivize new investments Normalized Difference Vegetation Index (NDVI) insur- in seed duplication, marketing, and training services, ance program in the ASALs. Under the proposed scheme, coupled with initiatives to stimulate farmer demand, are targeted beneficiaries would receive fully supported insur- needed to strengthen seed supply networks and improve ance, purchased on their behalf by the GoK, while wealth- farmers’ access. ier households would be able to purchase the product on a voluntary basis. It is envisaged that the macro livestock LIVESTOCK INDEX INSURANCE insurance product will aim at offering asset protection Today, few Kenyan farmers have access to risk transfer (i.e., covering the impact of pasture degradation on risk instruments to help them mitigate their exposure to price, reduction expenditures such as relocation, destocking, or weather, and other risks. Kenya’s rural credit and agricul- purchase of fodder) versus asset replacement (i.e., cover- tural insurance markets are as yet underdeveloped. Aside ing livestock mortality). Payouts would be made at the from the adverse effects of weather shocks on farmers’ onset of severe drought, thus reducing livestock mortality livelihoods, farmers’ high risk exposure limits their access and asset depletion. Appendix E provides more details on to credit as banks and other formal lending institutions the proposed initiative. Kenya: Agricultural Sector Risk Assessment 63 REVIEW OF COFFEE SECTOR Decision filters can be used as an alternative approach to Kenya’s coffee sector is in decline, in large part due evaluate and prioritize among a lengthy list of potential inter- to aging trees, falling productivity, a weak cooperative ventions. This can aid decision makers in making rational sector, and competition from other economic activities resource allocation decisions in lieu of a detailed cost-benefit offering higher returns. The decline is also due to inef- analysis. The following decision filters were developed and ficient regulation and a marketing structure that handi- used by the World Bank team. The study team applied these caps smallholder farmers’ and cooperatives’ ability to filters to facilitate a rapid assessment to obtain a first order of cope within the context of low returns and high risks. approximation, based on its assessment of the situation on Under Kenya’s marketing system, price risk and mar- the ground. The team presented preliminary results and the keting costs are transferred back upstream to farmers filtering approach to MoALF officials at a roundtable in Nai- in the form of consistently low farm-gate prices. The robi in early February 2014. During the exchange, the team result is growing divestment in coffee production among solicited feedback that it subsequently incorporated into the Kenya’s farmers, who have scant capacity or resources final results. Appendix 8 presents the results of the filtering to combat plant diseases, price shocks, and other risks. process for proposed mitigation, transfer, and coping strate- A review of the coffee sector’s structure could aid iden- gies. Whatever the filtering process and criteria adopted to tification of opportunities for streamlining how coffee evaluate decision options, it is important to ensure their clar- is bought and sold and for opening up Kenya’s coffee ity and consistency. markets to more competition and increased efficiencies. A more open marketing system holds scope to rebal- Table 6.2 describes the filtering criteria the assessment ance the way in which value and costs are shared across team used to rate each intervention. the coffee supply chain. It could also create a more enabling environment that incentivizes farmers—via higher farm-gate prices and better access to information and needed technologies—to invest in risk reduction TABLE 6.2. FILTERING CRITERIA FOR RISK and productivity-enhancing measures. This includes MANAGEMENT SOLUTIONS IN replanting old trees with new varieties resistant to CBD KENYA and CLR via improvements in extension and seed and input distribution services (see above). Any initiatives Criteria Description to address these challenges would undoubtedly require Applicability to Public sector: Is the proposed strong political will on the part of the Kenyan govern- current agricultural solution in line with current/ policy/programming existing agricultural policy/ ment and would greatly benefit from broad engagement or business objectives programs/priorities, and so on? with the cooperative and private sectors. Private sector: Is the proposed solution in line with current/existing business objectives, and so on? PRIORITIZATION OF RISK Feasibility of Is the proposed solution “easy” to MANAGEMENT MEASURES implementation implement in the short to medium Most of the measures outlined above are complemen- term? tary in nature and have strong potential to contribute to Affordability of Is the proposed solution affordable implementation to put into action/implement? improved risk management in the short, medium, and Scalability of Is the proposed solution easy to long term. However, decision makers are compelled implementation scale up/make available to an to find the quickest, cheapest, and most effective mea- increased number of beneficiaries? sures among myriad policy options. Ideally, a detailed, Long-term Is the proposed solution sustainable objective, and exhaustive cost-benefit analysis would sustainability in the long term? help in selecting the most appropriate intervention Source: World Bank. options, but such an analysis is often costly and time Note: The team answered the question posed in each criteria’s description consuming. using a scale of 1–5 (1, No; 2, Marginally; 3, Somewhat; 4, Yes; 5, Absolutely). 64 Agriculture Global Practice Technical Assistance Paper Management of agricultural risks is nothing new to Kenya, » To curb soil erosion, increase soil fertility and the GoK has a long track record of investment in risk mit- and water retention, and enhance the produc- igation, transfer, and coping mechanisms. Moving forward, tivity31 and biodiversity of smallholder systems Kenya’s Vision 2030 recognizes the need to strengthen exist- across Kenya, promoting broader awareness ing risk management systems, and the GoK has launched a and adoption (via farmer field schools and other range of new initiatives to confront the most severe threats participatory extension approaches) of conserva- facing the country. In 2011, it established the Drought Risk tion agriculture practices such as zero tillage, Management Authority to better coordinate preparedness mulching, composting and use of organic fertil- and speed up response measures. It also launched the Disas- izers, crop diversification and rotation, intercrop- ter Risk Reduction Program, the National Climate Change ping, and IPM Action Plan, and the Hunger Safety Net Program.30 These » To strengthen certified seed production and and other initiatives by the GoK and its development part- distribution systems, build their credibility, and ners are already helping to safeguard livelihoods, promote stimulate demand for improved seeds and fertiliz- adaptation, and strengthen resilience against impacts from ers by smallholders, investing in capacity building natural disasters and a changing climate. And yet as high- and training to strengthen monitoring and lighted by this report, agricultural supply chains in Kenya enforcement of quality standards and reduce remain highly vulnerable to myriad risks that disrupt the incidences of counterfeiting, adulteration, country’s economic growth, cripple poverty reduction and other abuses that dampen farmer demand and efforts, and undermine food security. The current study productivity highlights the need for a more targeted and systematic » To reverse degradation of water, soil, and approach to agricultural risk management in Kenya. vegetation cover, safeguard the long-term viability of Kenya’s arid and semiarid range- Based on an analysis of key agricultural risks, an evalua- land ecosystems, and ensure access to suf- tion of levels of vulnerability among various stakeholders, ficient grazing land, promoting (1) use of contour and the filtering of potential risk management measures, erosion and fire barriers, cisterns for storing rain- this assessment makes the following recommendations fall and runoff water, controlled/rotational graz- for the GoK’s consideration. The proposed focus areas ing, grazing banks, homestead enclosures, residue/ of intervention encompass a broad range of interrelated, forage conservation, and other sustainable mutually supportive investments, that together—aligned land management practices; and (2) innova- with Livelihoods Enhancement goals within Kenya’s Vision tive rangeland comanagement (state and local 2030—hold strong scope to strengthen the resilience of community) approaches that leverage customary vulnerable farming and pastoralist communities and the forms of collective action and economic instru- agricultural systems on which their livelihoods and the ments to reward sound pasture management country’s food security depend: » To strengthen drought resilience among vul- » To better optimize rainfall and soil moisture nerable pastoral communities in target ASAL in marginalized production areas, promoting com- counties and better safeguard the viability of ani- munity-driven investments in improved mal herds during shortages, supporting the devel- soil and water management measures such opment of feed/fodder production and as terracing, water harvesting pans, roof and rock storage systems, animal health, market catchment systems, subsurface dams, and micro- and weather information, and other critical irrigation systems services 30 The Hunger Safety Net Program is one of five cash transfer programs under the National Safety Nets Programme (NSNP). HSNP is implemented by 31 Conservation agriculture allows yields comparable with modern intensive NDMA and targets the poorest and most vulnerable households in four ASAL agriculture but in a sustainable way and with lower production costs (time, labor, counties (i.e., Turkana, Mandera, Wajir, and Marsabit). inputs). Yields tend to increase over the years with yield variations decreasing. Kenya: Agricultural Sector Risk Assessment 65 » To mitigate growing pressures on rangelands in the ASALs and increasing vulnerability of smaller live- CONCLUSION stock (<50 animals) owners in particular, putting in This Phase I assessment analyzes agricultural risks and place supportive policies and livelihood devel- impacts incurred in Kenya over the period 1980–2012. By opment programs (targeted credit schemes, documenting and analyzing how Kenya’s agriculture sector skills training, public sector investments in labor has been affected in the past by risk events, the study gen- intensive infrastructure projects, cash for work) to erates insight into which sources of risk are most likely to facilitate their engagement in alternative liveli- impact agricultural production systems and livelihoods in the hood and income-generating activities future. It prioritizes the most important agricultural risks for » To strengthen fiscal management and the country based on objective criteria. It offers a framework reduce the GoK’s budget volatility (and diver- for development of a more comprehensive, integrated risk sion of development resources caused by ex post management strategy to strengthen existing risk mitigation, crisis response), better safeguard rural liveli- transfer, and coping measures in Kenya. Finally, it provides a hoods, and increase resilience, deepening filtering mechanism to select an appropriate set of best pos- investments in agricultural insurance mechanisms sible interventions for agricultural risk management. and markets (in partnership with the private sec- Many of the proposed intervention areas are covered to tor), with an initial focus on asset protection varying degrees under the GoK’s Vision 2030 and ASDS (via early warning triggers and expedited payouts) development frameworks. Some may currently be in the among vulnerable pastoralist communities process of implementation by either government agen- and area yield index insurance for smallholder cies or their development partners. Moving forward, maize farmers stronger emphasis should be placed on scaling up these » To facilitate improved, evidence-based deci- interventions to reach a larger number of beneficiaries. sion making among farmers, pastoralists, and A greater emphasis should also be placed on ensuring a policy makers and to mitigate price volatility, more coordinated, integrated approach to risk manage- investing in integrated data and information ment in Kenya to ensure more effective and meaningful systems build-out for more robust, cost- risk reduction and resilience building across the sector. effective, and reliable collection, manage- ment, and dissemination (via surveying, GIS, It is hoped that this assessment’s findings and conclusions will ICT, SMS) of crop production, agro-weather, contribute to the existing knowledge base regarding Kenya’s market price (input/output) information, and agricultural risk landscape. To be certain, Kenya’s agricul- agricultural research and advice ture sector faces myriad risks. By prioritizing them, the study » To further objectives of the devolution pro- can help the GoK focus attention and resources on a smaller cess, promoting institutional and organizational set of key risks that have the most adverse impacts on pro- capacity building and technical training at duction yields, incomes, and livelihoods. It is also hoped that county and national levels to promote standard- the study will inform a dialogue between the GoK, the World ized collection and management of agri- Bank, and the GoK’s other development partners that will cultural data (in line with recently developed lead to concrete interventions toward improved agricultural national guidelines) risk management and livelihood resilience in Kenya. 66 Agriculture Global Practice Technical Assistance Paper REFERENCES African Agricultural Technology Foundation (AATF). 2006. Empowering African Farmers to Eradicate Striga from Maize Croplands. Nairobi, Kenya: The African Agricultural Technology Foundation. Ainsworth, E. A., and J. M. McGrath. 2010. “Direct Effects of Rising Atmospheric Carbon Dioxide on Crop Yields.” In Climate Change and Food Security: Adapting Agri- culture to a Warmer World, edited by D. Lobell, M. Burke. New York, NY: Springer. Aklilu, Y., and M. Wekesa. 2002. Drought, Livestock and Livelihoods: Lessons from the 1999–2001 Emergency Response in the Pastoral Sector in Kenya. London, UK: HPN, Overseas Development Institute, 40: 109–130. Alarcon, Diane, T. F. Joehnk, and B. Koch. 2013. Global Food Security Index, 2013: An Annual Measure of the State of Global Food Security. London, UK: The Economist Intel- ligence Unit. Alwora, G. O., and E. K. Gichuru. 2014. “Advances in the Management of Coffee Berry Disease and Coffee Leaf Rust in Kenya.” Journal of Renewable Agriculture 2(1): 5–10. Bryan, E., C. Ringler, B. Okaba, et al. 2011. Coping with Climate Variability and Adapting to Climate Change in Kenya: Household and Community Strategies and Determinants. Report to the World Bank Report. Cenacchi, N. and J. Koo 2011. Effects of Drought Tolerance on Maize Yield in Sub- Saharan Africa. Article prepared for the conference: “Increasing Agricultural Productivity & Enhancing Food Security in Africa: New Challenges and Opportu- nities”. November 13, 2011, Africa Hall, UNECA, Addis Ababa, Ethiopia. Inter- national Food Policy Research Institute. Central Bank of Kenya. 2012. Annual Report. Nairobi, Kenya: Central Bank of Kenya. Central Emergency Response Fund (CERF). 2008. African Review Report on Drought and Desertification. Addis Ababa, Ethiopia: United Nations Economic Commission for Africa. ———. 2009. CERF Funds Jump Start Emergency Aid Operations in Kenya. Addis Ababa, Ethiopia: United Nations Economic Commission for Africa. Chemonics. 2010. Staple Food Value Chain Analysis; Kenya Country Report. Washington, DC: The Competitiveness and Trade Expansion Program, USAID. Christensen, J. H., B. Hewitson, A. Busuioc, et al. 2007: Regional Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Climate Change Risk Assessment Report, Government of Kenya. Adaptation Technical Analysis, Technical Report 1: Risk Assessment Report, produced by the Climate and Development Knowledge Network. Nairobi: Republic of Kenya. Accessible at http://www.google.com/url?sa=t&rct=j&q=&esrc=s&s Kenya: Agricultural Sector Risk Assessment 67 ource=web&cd=5&cad=rja&uact=8&ved=0CDoQFjAE&url=http%3A%2 F%2Fwww.kccap.info%2Findex.php%3Foption%3Dcom_phocadownload% 26view%3Dcategory%26download%3D157%3Atr1-nccap-risk-assessment- report%26id%3D30%3Aadaptation-analysis&ei=w_d1Vc2oKqnmsASC0oAY& usg=AFQjCNF0NcPBZxB-7YRpT3to4D90QcnM7Q. The Contribution of Livestock to the Economies of Kenya, Ethiopia, Uganda, and Sudan [ICPALD 8/SCLE/8/2013]. IGAD Centre for Pastoral Areas and Livestock Devel- opment (ICPALD). Retreived from: http://igad.int/attachments/714_The%20 Contribution%20of%20Livestock%20to%20the%20Kenya,%20Ethiopia,%20 Uganda%20and%20Sudan%20Economy.pdf. Ernst & Young. 2013. Ernst & Young’s Attractiveness Survey, Africa 2013: Getting Down to Business. London, UK: Ernst & Young. Famine Early Warning Systems Network (FEWSNET). 2011. “Kenya Livelihood Zones.” Available at http://www.fews.net/east-africa/kenya/livelihood-zone- map/march-2011. Fitzgibbon, C. 2012. Economics of Resilience Study—Kenya Country Report. London, UK: Department for International Development (DFID). Food and Agriculture Organization of the United Nations (FAO). 2011. Dairy Develop- ment in Kenya, by H.G. Muriuki. Rome, Italy: FAO. ———. 2012. Analysis of Incentives and Disincentives for Rice in Kenya. Monitoring African Food and Agricultural Policies. Rome, Italy: FAO. ———. 2013a. Cereal Supply/Demand Balances for Sub-Saharan Africa. Global Information and Early Warning System on Food and Agriculture. Rome, Italy: FAO. ———. 2013b. Policy Brief: Improving Price Incentives for Cattle Producers in Kenya. MAFAP. Policy Briefing No 10. Rome, Italy: FAO. Food and Agricultural Organization of the United Nations Corporate Statistical Database (FAOSTAT). 2014. Kenya Food Balance data. Rome, Italy: FAO. ———. 2014. FAOSTAT Agricultural Data, available at http://faostat.fao.org. Gichimu, B. M., and N. A. Phiri. 2012. “Response of Newly Developed and Introduced Arabica Coffee Genotypes to Coffee Berry Disease (Colletotrichum kahawae) in Kenya.” Gitonga, K., and K. Snipes. 2014. GAIN Report: Kenya Corn, Wheat and Rice Report. Grain and Feed Annual. USDA Foreign Agricultural Service. Available at: http://gain .fas.usda.gov/Recent%20GAIN%20Publications/Grain%20and%20Feed%20 Annual_Nairobi_Kenya_4-8-2014.pdf. Government of Kenya (GoK). 2009a. Disaster National Policy for Disaster Management in Kenya. 2009. Nairobi, Kenya: Republic of Kenya. ——— (Ministry of Agriculture). 2009b. Assessment of Costs of Maize Production, Market- ing and Processing In Kenya: A Maize Grain-Maize Meal Value Chain Analysis. Nairobi, Kenya: Republic of Kenya. ——— (Office of the Prime Minister). 2009c. Releasing Our Full Potential: Draft Sessional Paper on National Policy for the Sustainable Development of Northern Kenya and Other Arid Lands. Nairobi, Kenya: Republic of Kenya. ———. 2010a. 2009 Kenya Population and Housing Census Vol. II. Population and Household Distribution by Socioeconomic Characteristics. Nairobi, Kenya: Republic of Kenya. ———. 2010b. Agriculture Sector Development Strategy, 2010–2020. Nairobi, Kenya: Republic of Kenya. 68 Agriculture Global Practice Technical Assistance Paper ———. 2011. Vision 2030 Development Strategy for Northern Kenya and Other Arid Lands. Nairobi, Kenya: Republic of Kenya. ———. 2012a. Kenya Post-Disaster Needs Assessment (PDNA): 2008–2011 Drought, with support from the European Union and the ACP-EU Natural Disaster Risk Reduc- tion Program. Nairobi, Kenya: Republic of Kenya. ———. 2012b. National Climate Change Action Plan 2013–2017. Nairobi, Kenya: Repub- lic of Kenya. ——— (Ministry of Agriculture). 2013a. Economic Review of Agriculture. Nairobi, Kenya: Central Planning and Project Monitoring Unit, Ministry of Agriculture, Republic of Kenya. ———. 2013b. Public Expenditure Review. Nairobi. Nairobi, Kenya: Republic of Kenya. Hatibu, N., K. Mutabazi, E. M. Senkondo, and A. S. K. Msangi. 2006. “Economics of Rainwater Harvesting for Crop Enterprises in Semi-arid Areas of East Africa.” Agricultural Water Management 80 (1–3): 74–86. Herrero, M., C. Ringler, J. van de Steeg, P. Thornton, T. Zhu, E. Bryan, A. Omolo, J. Koo, and A. Notenbaert. 2010. Climate Variability and Climate Change and Their Impacts on Kenya’s Agriculture Sector. Nairobi, Kenya: International Livestock Research Institute (ILRI). Horticultural Crops Development Authority (HCDA), Ministry of Agriculture, USAID (U.S. Agency for International Development). 2012. Horticulture: Validated Report. Nairobi, Kenya: Republic of Kenya. Hortiwise. 2012. A Study on the Kenyan-Dutch Horticultural Supply Chain. The Hague, Neth- erlands: The Dutch Ministry of Economic Affairs, Agriculture and Innovation. International Fund for Agricultural Development (IFAD). 2013. Enabling poor rural peo- ple to overcome poverty in Kenya. Rome, Italy: IFAD. International Monetary Fund (IMF). 2008. World Economic Outlook Database 2008. Avail- able at http://www.imf.org/external/pubs/ft/weo/2008/01/weodata/index .aspx (accessed September 25, 2009). ———. 2014a. Kenya: Poverty Reduction Strategy Paper—Joint Staff Advisory Note. Country Report No. 14/74. Washington, DC: IMF. ———. 2014b. World Economic Outlook—Recovery Strengthens, Remains Uneven. Washington, DC: IMF. Intergovernmental Panel on Climate Change (IPCC). 2012. “Managing Interna- tional Livestock Research Institute the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change.” Edited by C.B. Field, V. Barros, T.F. Stocker, et al. Cambridge, UK: Cambridge University Press. Retreived from: https://books.google.com/books?id=nQg3SJtkOGwC&dq= Managing+International+Livestock+Research+Institute+the+Risks+of+Extreme +Events+and+Disasters&source=gbs_navlinks_s. Kamau, Carol N. 2013. Kenya Corn, Wheat, and Rice Report. Washington, DC: Global Agriculture Information Network (GAIN), USDA Foreign Agricultural Service. Kamau, Carol N., and Kate Snipes. 2013. Kenya Corn, Wheat and Rice Report. Grain and Feed Annual. USDA Foreign Services Kenya, Office of Agricultural Affairs, Nairobi. Kenya: Agricultural Sector Risk Assessment 69 Kassie, M., S. W. Ndiritu, and B. Shiferaw. 2012. “Determinants of Food Security in Kenya, a Gender Perspective.” 86th Annual Conference, April 16–18, Warwick University, Coventry, UK (No. 135124). Agricultural Economics Society. Katungi, A., J. Farrow, L. Chianu, L. Sperling, and S. Beebe. 2009. Common Bean in Eastern and Southern Africa: A Situation and Outlook Analysis. International Centre for Tropical Agriculture, Cali, Colombia. Kayombo, B., N. Hatibu, and H. F. Mahoo. 2004. “Effect of Micro-catchment Rain- water Harvesting on Yield of Maize in a Semi-arid Area.” ISCO 2004- 13th Inter- national Soil Conservation Organisation Conference. Retreived from: http:// www.tucson.ars.ag.gov/isco/isco13/PAPERS%20F-L/KAYOMBO.pdf. Kenya Food Security Steering Group (KFSSG). 2008. The Impact of Rising Food Prices on Disparate Livelihood Groups in Kenya. Nairobi, Kenya: Government of Kenya. ———. 2011. The 2011/12 Short Rains Season Assessment Report. Nairobi, Kenya: Government of Kenya. Kenya Sugar Board (KSB). 2010. The Kenya Sugar Industry Value Chain Analysis: Analysis of the Production and Marketing Costs for Sugarcane and Sugar Related Products. Nairobi, Kenya: KSB. Kenya Sugar Industry (KSI). 2009. Kenya Sugar Industry Strategic Plan 2010–2014. Nairobi, Kenya: KSI. Kenya Sugar Research Foundation (KESREF). 2011. Assessment of Yield Loss Due to Sug- arcane Smut Infection in Kenya, by H.S. Nike and J.E. James. Nairobi, Kenya: KESREF. Kirimi, Lilian, Nicholas Sitko, T. S. Jayne, et al. 2011. “A Farm-Gate-to-Consumer Value Chain Analysis of Kenya’s Maize Marketing System.” MSU International Development Working Paper No. 111. East Lansing, MI: Michigan State University. La Rovere, R., T. Abdoulaye, G. Kostandini, et al. 2014. “Economic, Production, and Poverty Impacts of Investing in Maize Tolerant to Drought in Africa: An Ex-Ante Assessment.” Journal of Developing Areas 48(1): 199–225. Lobell, D., M. Burke, C. Tebaldi, et al. 2008. “Prioritizing Climate Change Adapta- tion Needs for Food Security in 2030.” Science 319(5863): 607–610. Luedeling, E. 2011. “Climate Change Impacts on Crop Production in Busia and Homa Bay Counties, Kenya.” Produced by Adaptation to Climate Change and Insur- ance (ACCI), World Agroforestry Centre, Nairobi, Kenya: Republic of Kenya. Macharia, G. 2013. “Financing Dairy sector.” Equity Bank. Presentation made in Nairobi, Kenya: Republic of Kenya. Makooha, S., and M. Witmer. 2013. “Analysis of Incentives and Disincentives for Cattle in Kenya.” Technical Note Series, MAFAP. Rome, Italy: Food and Agricul- ture Organization. McSweeney, C., M. New, and G. Lizcano. 2012. “UNDP Climate Change Country Profiles: Kenya.” Available at: http://country-profiles.geog.ox.ac.uk/. Michiels, D., R. Blein, and J. Egg. 2011. “Évaluation des systèmes de réponses à la crise 2010 au Niger.” Délégation de l’Union européenne au Niger, COWI/IRAM. Millennium Cities Initiative (MCI). 2008. Sugar In Kisumu, Kenya. New York, NY: The Earth Institute at Columbia University. Retrieved from http://www.vcc.columbia .edu/pubs/documents/SugarWorkingPaper.pdf. Ngigi, S. N., H. H. G. Savenije, J. Rockstrom, and C. K. Gachene. 2005. “Hydro- economic Evaluation of Rainwater Harvesting and Management Technologies: 70 Agriculture Global Practice Technical Assistance Paper Farmers’ Investment Options and Risks in Semi-arid Laikipia District of Kenya.” Physics and Chemistry of the Earth 30(11–16): 772–782. Republic of Kenya. 2012. Kenya Post Disaster Needs Assessment (PDNA): 2008–2011 Drought. Nairobi, Kenya: Republic of Kenya, with technical support from the European Union, United Nations, and World Bank. 17 Oxfam. Rice, Xian. 2006. “Kenya’s tea industry wilts in the sun.” The London Times, February 27. Rockstrom, J., J. Barron, and P. Fox. 2002. “Rainwater Management for Increased Productivity among Small-holder Farmers in Drought Prone Environments.” Physics and Chemistry of the Earth 27(11–22): 949–959. Smale, M., and J. Olwande. 2014. “Demand for Maize Hybrids and Hybrid Change on Smallholder Farms in Kenya.” Agricultural Economics 45: 1–12. Sombroek, W. G., H. M. H. Braun, and B. J. A. van der Pouw. 1982. “Exploratory Soil Map and Agro-Climatic Zone Map of Kenya, 1980. Scale: 1:1’000’000.” Explor- atory Soil Survey Report No. E1. Nairobi, Kenya: Kenya Soil Survey Ministry of Agriculture–National Agricultural Laboratories. Tegemeo Institute. 2010. Pathway Into and Out of Poverty: A Study of Rural Household Wealth Dynamics in Kenya.” Nairobi, Kenya: Tegemeo Institute. The Kenya Veterinarian. 2001. Vol. 22: 76–78. Available at http://www.ajol.info/index .php/kenvet/article/view/39523. Thornton, P. K., P. G. Jones, T. M. Owiyo, et al. 2006. “Mapping Climate Vulnerabil- ity and Poverty in Africa.” Report to the Department for International Develop- ment, International Livestock Research Institute, Nairobi, Kenya. Thornton, P. K., P. G. Jones, G. Alagarswamy, J. Andresen, and M. Herrero. 2009. “Adapting to Climate Change: Agricultural System and Household Impacts in East Africa.” Agricultural Systems (in press), online at http://dx.doi.org/10.1016 /j.agsy. 2009.09.003. Thornton, P. K., P. G. Jones, P. J. Ericksen, and A. J. Challinor. 2011. “Agriculture and Food Systems in Sub-Saharan Africa in a 4°C+ World.” The Royal Society 369(1934): 117–136. United Nations Development Programme (UNDP). 2010. Kenya National Human Development Report 2009: Youth and Human Development: Tapping the Untapped Resource, Kenya. Retrieved from: http://hdr.undp.org/sites/default/files/kenya_nhdr_2009 _en.pdf. Venton, Courtney, C. Fitzgibbon, T. Shitarek, L. Coulter, and O. Dooley. 2012. The Economics of Early Response and Disaster Resilience: Lessons from Kenya and Ethiopia. London, UK: Department for International Development (Df ID). Waithaka, M., G. C. Nelson, T. S. Thomas, and M. Kyotalimye, Eds. 2013. East African agriculture and climate change: A comprehensive analysis. Washington, D.C.: International Food Policy Research Institute (IFPRI) http://dx.doi .org/10.2499/9780896292055. Wangai, A. W., M. G. Redinbaugh, Z. M. Kinyua, et al. 2012. “First Report of Maize Chlorotic Mottle Virus and Maize Lethal Necrosis in Kenya.” Plant Disease 96(10): 1582–1583. Available at http://apsjournals.apsnet.org/loi/pdis. World Bank. 2008. Kenya Poverty and Inequality Assessment. Washington, DC: World Bank. Kenya: Agricultural Sector Risk Assessment 71 ———. 2009a. Kenya Economic Update Still Standing: Kenya’s Slow Recovery from a Qua- druple Shock. Edition No. 1. Poverty Reduction and Economic Management Unit Africa Region. Washington, DC: World Bank. ———. 2009b. Eastern Africa: A Study of the Regional Maize Market and Marketing Costs. Agriculture and Rural Development Unit (AFTAR) Report No. 49831–AFR. Washington, DC: World Bank. ———. 2012. Kenya Horticulture Supply Chain Risk Assessment. Washington, DC: World Bank. ———. 2013. Kenya—Comprehensive Public Expenditure Review 2013: Eye on Budget—Spending for Results. Public Expenditure Review (PER). Washington, DC: World Bank. World Food Programme (WFP). 2013. Kenya Food Security and Outcome Monitoring Consoli- dated Report. Rome, Italy: WFP. Zwaagstra, L., Z. Sharif, A. Wambile, et al. 2010. An Assessment of the Response to the 2008–2009 Drought in Kenya. A Report to the European Union Delegation to the Republic of Kenya. Nairobi, Kenya: International Livestock Research Institute (ILRI). APPENDIX A CLIMATE CHANGE IMPACTS ON AGRICULTURE IN KENYA INTRODUCTION Like those of many of the countries in Sub-Saharan Africa, Kenya’s agriculture sector is highly vulnerable to the effects of climate change. The country’s climate is already characterized by high temperatures and low but highly variable annual precipitation, factors that negatively affect the productivity of heat-sensitive crops. Climate change is a long-term trend that will exacerbate natural resource constraints on agricultural production in Kenya by making weather patterns more variable and by increasing the frequency and intensity of severe weather events. As a result, climate change will directly affect the incidence of some agricultural risk events and indirectly affect the incidence of others. Understanding how climate change trends affect farm productiv- ity is essential to formulating an agricultural risk management plan that maximizes the use of scarce resources. Regardless of the future extent of global warming, identifying and implementing risk management strategies that address agricultural risks, includ- ing those exacerbated by climate change, can reduce volatility and improve sustain- ability in the sector. Due to the importance of the agriculture sector in Kenya’s national economy, cli- mate change impacts on crop yields and land suitability will have far-reaching effects. Agriculture accounts for 24 percent of national GDP and 65 percent of all export earnings (GoK 2012a). Agriculture also plays a key role in poverty reduction and food security through its contribution to livelihood security. The sector employs more than 75 percent of the workforce and generates most of the country’s food require- ments. Within the sector, smallholder farms account for 85 percent of employment and 75 percent of total agricultural output (GoK 2012a). In the Mapping the Impacts of Climate Change index under “Agricultural Productivity Loss,” the Center for Global Development ranks Kenya 13 out of 233 countries globally for “direct risks” due to “extreme weather” and 71 out of 233 countries for “overall vulnerability” to climate change when adjusted for coping ability. Kenya: Agricultural Sector Risk Assessment 73 Climate change impacts agriculture through temperature minant of land suitability for production, will decrease increases, changes in precipitation, and increases in the in many places. Some regions (the mixed rainfed tem- frequency and severity of extreme weather events. There perate and tropical highlands) are projected to experi- are direct impacts, such as changes in land suitability for ence an increase in crop yield. Other regions, especially crops due to temperature changes, and indirect impacts, the ASALs, are projected to experience a significant such as changes in food prices that ultimately affect food decline in crop yields and livestock productivity as water demand and well-being. Models predicting the effects of resources become increasingly scarce. These patterns climate change on agriculture vary across regions and are largely driven by regional variability in future pre- crop/livestock sectors, and depend heavily on the under- cipitation and geographic exposure to extreme events, lying assumptions. The projected effects of changes in particularly drought frequency. precipitation are particularly difficult to reconcile, given the vast regional variation in annual rainfall and limited Key uncertainties surrounding the impact of climate district-level data. Rising temperatures are also expected change on agriculture include to increase evapotranspiration, offsetting productivity » Extent of crop yield increases due to CO2 fertiliza- gains. Although a large degree of uncertainty exists about tion the magnitude of the impacts, this brief synthesizes exist- » Effect of ozone damage on crop yields (Ainsworth ing crop forecasts, highlights areas of consensus between and McGrath 2010; Iglesias et al. 2009) different studies, and identifies areas of disagreement. » Extent of crop damage caused by the evolution of pests and diseases PRINCIPAL FINDINGS Several structural vulnerabilities in the sector are likely to CLIMATE CHANGE AND exacerbate the impact of rising temperatures and changes SEVERE WEATHER EVENTS in precipitation on crop and livestock production, includ- The frequency of severe weather events has already ing dependence on rainfed agriculture; overcultivation and increased, and the intensity of weather events like land degradation; lack of technologies to improve produc- drought, extreme heat, and floods is likely to increase tion; and high levels of poverty among smallholder farmers. as temperatures rise (IPCC 2011). In Kenya, recurrent drought in particular has profound effects on the agri- The studies cited in this report conclude that temperature culture sector. Intense droughts occurred in 1991–92, increases will have a significant impact on water avail- 1995–96, 1998–2000, 2004–05, and 2008–11, result- ability and soil quality, and thus will likely exacerbate ing in severe crop losses, livestock deaths, spikes in food drought conditions. Precipitation is projected to increase insecurity, and population displacement. For instance, between 0.2 and 0.4 percent per year in Kenya, but the the government of Kenya (GoK) imported 2.6 million direction and magnitude of change vary considerably bags of maize worth K Sh 6.7 billion between 2008 and across regions, and warming-induced increases in evapo- 2009. In the future, any increase in the frequency and/ ration rates are likely to offset the benefits of precipitation or severity of drought conditions will have profound far- increases in some regions. In addition, an increase in the reaching effects on national food security and the viabil- intensity of high rainfall events is expected in Kenya and is ity of livelihood activities, especially for people living in already underway: the number of extremely wet seasons the ASALs. is increasing to roughly 1 in 5, compared to 1 in 20 in the late twentieth century (Christensen et al. 2007; Herrero/ IFPRI 2010). In the arid and semiarid lands (ASALs) of METHODOLOGIES Kenya, fragile soils are particularly vulnerable to flash- Data analyses from the literature reviewed in this flooding and erosion during high rainfall events. brief draw from downscaled general circulation mod- els (GCMs). The studies use multiple GCMs, simulate Temperature and precipitation changes suggest that between one and four greenhouse gas emissions scenar- the length of the growing period (LGP), a key deter- ios, and incorporate crop prediction models. As a result, 74 Agriculture Global Practice Technical Assistance Paper conclusions vary depending on the underlying model commodities. Crop yield projections for maize vary assumptions: widely, depending on the region and the specified climate » Based on these models, the IFPRI study uses the scenario. Although the magnitude of the change in Decision Support System for Agrotechnology yields varies under each scenario, most models predict Transfer (DSSAT) crop modeling software projec- declining yields in large parts of the ASALs and in the tions for crop yields, comparing yield projections lowlands, and yield increases in the temperate central for 2050 against real 2000 yields. and western highlands. » Thorton et al. (2006) combine projected climate change scenarios with vulnerability data to iden- According to the IFPRI analysis, four out of six climate tify high-risk regions and population groups. The scenarios predict an overall decrease in rainfed maize analysis uses LGP as a proxy for agricultural yields. Kenyan maize yields drop by 51–55 percent under impacts. Predictions on changes in LGP vary con- the NCAR 369, CSIRO 369, and CSIRO 532 A2 sce- siderably across Kenya depending on the underly- narios, compared to 2050 yields with historic climate lev- ing assumptions. Across a range of future climate els. In contrast, yields increase by 25 percent under the scenarios, however, many parts of Kenya are pre- Hadley 369 A2a scenario. dicted to experience a decrease in LGP, and some a severe decrease. This suggests a need to focus on the development and dissemination of short- Thorton et al. (2009) predict a maize production decline season cultivars, as well as water management of 8.4 percent and 9.8 percent in the semiarid areas and strategies. humid/subhumid areas, respectively. The same study » Thorton et al. (2009) examines the spatial varia- predicts a 46.5 percent increase in maize production in tion of climate change impacts on crop yields, the mixed rainfed systems in the temperate areas. Under using GCMs, crop models CERES-Maize and this scenario, however, total production would still decline BEANGRO, and soil and crop management given the relatively small contribution of the temperate data. areas to total production. » A study undertaken by the Adaptation to Climate Change and Insurance (ACCI) Programme evalu- Beans: Beans are an important food security crop in ates the current and future suitability of the Lake Kenya, accounting for 17.9 percent of the total harvested Victoria region (Busia and Homa Bay counties) for area. Beans are grown in every region in the country, major agricultural crops using a range of GCMs, with about 75 percent of total production concentrated soil data, and crop models. in three regions: Rift Valley, Nyanza, and Easter Province » The Information Center for Tropical Agricul- (Katungi et al. 2009). Like maize, crop yield projections ture (CIAT) analyzes future climatic suitability for beans vary depending on the region and the specified for tea-growing areas in Kenya. The study com- climate scenario. Although the magnitude of the change bines current climate data with future climate in yields varies under each scenario, most models project change predictions from 19 GCMs for 2030 and yield declines. 2050 (emissions scenario SRES-A2). These data are then used in MAXENT, a crop prediction Thorton et al. (2011) use an ensemble mean of three model. emissions scenarios and 14 GCMs to run crop simula- tions for conditions in a 4°C warmer world by 2090. For East Africa, a mean yield loss of 47 percent is projected CROP PREDICTIONS for beans. However, the disaggregated analysis predicts Maize: Maize is the principal food crop in Kenya. substantial yield increases for beans at higher eleva- Grown in every region, maize accounts for 37.5 per- tions in Kenya’s western highlands, up to average tem- cent of the total harvested area, and contributes 17.9 peratures of about 20–22°C, after which yields decline percent to the total value of production of agricultural (DFID 2012). Kenya: Agricultural Sector Risk Assessment 75 FIGURE A.1. CURRENT SUITABILITY OF TEA PRODUCTION AREAS Lobell et al. (2008) use 20 GCMs to run crop simulations tion. Recent declines in tea production have already been in East Africa. At least 75 percent of the projections are directly linked to erratic rainfall patterns and drought associated with Phaseolus bean yield losses. (Herrero et al. 2010). Tea: Tea is the most important agricultural export crop According to a suitability analysis conducted by the CIAT, in Kenya, accounting for 33 percent of total agricultural some areas will become unsuitable for tea (Nandi, Keri- exports and 3.5 percent of GDP32. Although few rigorous cho, and Gucha), while some will remain suitable for tea estimates of future changes in yield exist for the tea sec- (Bomet, Kisii, Nyamira) if farmers adapt agricultural tor, several authors have analyzed the impact of climate management practices to new climate conditions. Suit- change on future land suitability for tea production. Most ability for tea increases in some current growing areas studies find declines in suitability for land currently under (Meru, Embu, Kirinyaga, Nyeri, Murangá, Kiambu), and tea production and increases in suitability for new areas at new areas will become suitable for tea (especially higher higher altitudes. As a result of higher temperatures, all of altitudes around Mount Kenya). However, many of the the models predict major shifts in the geographic distribu- potential new areas for tea are located in protected areas tion of tea production. and forested lands (figures A.1 to A.3). According to maps provided by UNEP-GRID, a 2°C Preliminary results from an FAO study indicate that climate increase in temperature would render much of the cur- change is expected to increase suitability of tea-growing rent tea-growing area in Kenya unsuitable for produc- areas by 8 percent by 2025, and then negatively impact suitability as mean air temperatures rise above the 23.5°C 32 Data available at http://faostat.org. threshold, dropping by 22.5 percent by 2075 (FAO 2013). 76 Agriculture Global Practice Technical Assistance Paper FIGURE A.2. FUTURE SUITABILITY OF TEA PRODUCTION AREAS FIGURE A.3. SUITABILITY CHANGE FOR TEA PRODUCTION IN 2050 Source: CIAT 2011. Kenya: Agricultural Sector Risk Assessment 77 hide vast regional differences in crop performance under CROPS RESISTANT TO climate change. Many of the existing studies on crop CLIMATE CHANGE yield and land suitability under future climate scenarios A regional study undertaken by the ACCI Programme (Lue- lack detailed, regional impact assessments. Regional and deling 2011) analyzes the future suitability of the Lake Vic- district-level crop analyses could provide a better under- toria region for major agricultural crops. It should be noted standing of the aggregate impact of climate change on that the results of the study are limited to Busia and Homa agricultural systems and food security in Kenya. Bay counties. Using a range of GCMs, soil data, and crop prediction models, the study found that sorghum, ground- The studies all agree that rising temperatures are likely nuts, and fava beans were moderately resistant to the effects to exacerbate drought conditions in some regions, espe- of climate change. The ACCI study also identified cassava, cially the ASALs, due to highly variable rainfall patterns, sweet potato, mango, banana, and pineapple as crops with changes in seasonal water availability, and poor soil qual- the potential to thrive under warmer temperatures. ity. Even with increases in annual precipitation, extreme heat and drought conditions will negatively affect yields in parts of the current production area for crops like maize CONCLUSIONS and tea. Thus, as climate variability and uncertainty As discussed previously, the projected impacts of climate increase, there is an urgent need to identify and imple- change on agriculture vary considerably depending on ment risk management solutions to mitigate agricultural the climatic model specified, and national-level estimates losses and increase stakeholders’ coping capacity. 78 Agriculture Global Practice Technical Assistance Paper APPENDIX B STAKEHOLDER RISK PROFILES CASE STUDY 1: PHILIP MUTUA MBAI— SMALLHOLDER MAIZE FARMER, MACHAKOS COUNTY INTRODUCTION Philip Mbai is a 68-year-old small-scale farmer in Machakos County. He bought his 23-acre farm in 1978 while still working as an administrative clerk with Gailey and Roberts, Ltd, an engineering firm, and settled down to full-time farming in 2000. He practices mixed farming and grows maize and beans on 8 and 4 acres, respectively, and commits another 2 acres each to green grams and cowpea and approximately another 2 acres to fruits (mangoes, bananas, and oranges) and vegetables (kales, onions, toma- toes, and cabbages). The remainder of his farm is used for growing pasture to feed his four exotic cattle (Guernsey) and another six indigenous cattle. A small portion is also dedicated to bee farming, with an apiary of about 50 beehives featuring traditional log hives, Kenya top bar hives, and modern Langstroth bee hives. OVERVIEW OF KEY FARM ACTIVITIES Except for the production of vegetables, Mr. Mbai’s crop production is purely rainfed, with low attainable yields of below 0.5 ton/acre (about 10 90-kg bags per acre) largely attributed to low and unreliable rainfall. Located about 1,600 meters above sea level, his farm receives less than 800 mm of rainfall annually on average, although in the last 10 years, average annual rainfall has been below 600 mm and is falling. The farm serves as a meteorological monitoring site for rainfall data and has a fitted rain gauge that Mr. Mbai monitors on a daily basis. In a good year, Mr. Mbai makes about K Sh 500/bag, which translates to K Sh 40,000 a year. On average, Mr. Mbai derives 30 percent of his income from maize farming and the rest from the other farm activities. Kenya: Agricultural Sector Risk Assessment 79 IMPACT OF RISK EVENTS a drought once every two years. For instance in the last 10 Mr. Mbai encounters the following major agricultural years we had droughts in 2001, 2003, 2005, 2008, 2011 risks, in decreasing order of importance: drought, pests and now in 2014” (Philip Mbai). The severity of these and diseases, postharvest losses, and volatile market droughts has also increased, with yield losses estimated to prices that are exacerbated by lack of markets. “Overall, have increased from about 20–50 percent to 100 percent dependence on rainfed agriculture is the most important in the recent past. risk in this area” (Philip Mbai). The drought problem in Machakos has been worsened by changes in the onset RISK PRIORITIZATION and cessation of rains. Ten years ago the onset of rains In decreasing order of importance, the key risks encoun- was predictable, today rains typically begin late and cease tered in Machakos include early, which has shortened the length of the growing sea- » Drought son. The frequency and severity of droughts in the recent » Pests and diseases past has also increased. “While 30 years ago we experi- » Postharvest losses enced a drought once every 5 years, today we experience » Price volatility RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Crop Frequency Impact 1. Drought Maize Once every two years Yield losses of 30–50% 2. Pests and diseases Maize Occasionally Yield losses of up to 10% 3. Postharvest losses Maize Each season Yield losses of up to 30% 4. Price volatility Maize Occasionally Income losses of up to 20% RISK MITIGATION STRATEGIES ing, planting Napier grass along terraces, enterprise diversi- To mitigate against drought, Mr. Mbai undertakes a num- fication, and supplementary irrigation when possible). Even ber of agronomic practices. These include growing drought- though relief food supplies are the most widely used drought- resistant maize varieties, planting early, and using soil and coping strategy in the area, Mr. Mbai does not rely on relief water conservation practices (terracing, road water harvest- food, but instead sells livestock in times of severe drought. CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Drought Early planting None Sale of livestock Not effective Drought-tolerant varieties Soil and water conservation Enterprise diversification Supplementary irrigation Pests and diseases Use of chemicals None Use of indigenous Somewhat effective knowledge in disease and pest control Postharvest losses Adoption of improved None Use of indigenous Somewhat effective storage technologies knowledge Price volatility Storage None Sourcing alternative Somewhat effective markets 80 Agriculture Global Practice Technical Assistance Paper around seed and fertilizer, respectively. In addition, agro- CASE STUDY 2: inputs are regulated by the Poisons and Pharmacy Board MRS. MARABA—AGRO-INPUT along with the Kenya Veterinary Board. Operations in DEALER, ELDORET UASIN the agro-input outlet typically peak around January to GISHU COUNTY April when farmers in the region are planting maize and again in June when they are top dressing. In the other months, animal feed and day-old chicks are the main INTRODUCTION inputs sold. Mrs. Maraba operates an agro-input shop in Eldoret Town of Uasin Gishu County that is located in Kenya’s IMPACT OF RISK EVENTS main maize-growing zone. The agro-input shop doubles Agro-input dealers face the following main risks, in both as a wholesale and a retail outlet selling to other declining order of importance: government policy agro-input dealers and farmers. She has been in this busi- (NCPB subsidies), erratic seed quality, adulteration of ness for the last 15 years and stocks farm inputs such as fertilizers, foreign exchange, theft, and health risks. seeds, fertilizers, chemicals, livestock feed, and veterinary Government policy in the recent past has negatively vaccines. Her major customers are small-scale farmers affected input dealers since the subsidized inputs pres- who are unable to access subsidized fertilizers from the ent an unfair competition to other industry players. state-run National Cereals and Produce Board (NCPB) This makes it difficult for input dealers to plan their that in the last three years has been stocking subsidized stocking rates because they fear the price undercutting fertilizers. Mrs. Maraba sources her inputs from seed associated with subsidized fertilizers. When an input companies such as Kenya Seed Co. Ltd, FreshCo Kenya dealer stocks fertilizers before the government’s pro- Ltd, East Africa Seed Co. Ltd, and Western Seed Com- nouncement of subsidized fertilizer prices, he or she is pany Ltd. Unlike those of other seed companies, prices saddled with stocks that cannot be sold after the arrival for seeds from Kenya Seed Company, which accounts of government-subsidized fertilizers. Besides subsidized for more than 80 percent of the market share, are regu- fertilizers, another key risk is associated with stocking of lated, with the company controlling prices at both the poor-quality seeds, especially from Kenya Seed Co. In wholesale and retail levels. Fertilizer supplies are sourced the last two years, farmers have complained about the from the seven major fertilizer importers in Kenya: Yala, poor germination rate of maize seed from Kenya Seed MEA East Africa Ltd, Export Trading Company, Africa Co. Given that agro-input dealers source substantial Ventures, Devji Megji Brothers, Sharkaji (SKL), and amounts of seed from there, this issue presents a real Eldoret Packers. business constraint. OVERVIEW OF KEY ACTIVITIES RISK PRIORITIZATION Mrs. Maraba employs about five permanent staff and In decreasing order of importance, the key risks faced by another five casual workers on a daily basis. The major agro-input dealers include activity in the outlet revolves around stock management » Unpredictable government policy even though her staff also provide extension advice to » Erratic seed quality farmers seeking to purchase inputs. Agro-input dealers » Adulterated fertilizers in Kenya are regulated by both the Kenya Plant Health » Insecurity Inspectorate Service (KEPHIS) and the Kenya Bureau » Foreign exchange fluctuations of Standards (KEBS), whose regulatory services revolve » Health risks Kenya: Agricultural Sector Risk Assessment 81 RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Crop Frequency Impact 1. Unpredictable Maize Occasionally 20–30% reduction in profits government policy 2. Erratic seed quality Maize Occasionally 10% loss in market share 3. Adulterated Maize Occasionally 5% loss in market share fertilizers 4. Insecurity Maize Occasionally Scares away workers 5. Foreign exchange Maize Occasionally 6. Health risks Maize Occasionally Scares away workers RISK MITIGATION STRATEGIES This agro-input dealer’s key risk mitigation strategy is the procurement of insurance against theft, fire, and burglary. CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Unpredictable Stocking operations None Not effective government policy Erratic seed quality Forward contracts with suppliers None Adulterated Forward contracts with suppliers None fertilizers Insecurity Adherence to safety standards Insurance Safety at the workplace Somewhat effective Foreign exchange Hedging Forward contracts Health risks Medical insurance Insurance Observance of work safety K Sh 2,500/bag, Mr. Olekoonyo earned K Sh 3.6 million CASE STUDY 3: LESHAMON from his wheat enterprise last year. OLEKOONYO—WHEAT FARMER, NAROK OVERVIEW OF KEY FARM ACTIVITIES The key farm activities on Mr. Olekoonyo’s farm revolve INTRODUCTION around wheat production. The period February to April Mr. Olekoonyo is a wheat farmer in lower Narok, where is used in land preparation, where the main challenge is he grows wheat on 400 acres. He also chairs the Narok the high cost of tractor hire, largely driven by the high Wheat Farmers’ Association, a farmers’ organization cost of diesel. After planting, the farm engages in top formed to lobby wheat millers to provide better prices dressing, chemical control of pests and diseases, and to wheat farmers. Through the farmers’ association and later on harvesting, which occurs between June and July. in collaboration with the Cereal Growers Association The area has two rainy seasons and therefore produces (CGA), wheat farmers in Kenya were able to sign an two wheat crops each year. The major challenge in har- agreement in 2008 that ensures that wheat millers pur- vesting is achieving a moisture content of 13 percent, chase all wheat produced in Kenya before resorting to because most of the wheat is harvested at high moisture imports. On average, Mr. Olekoonyo attains a yield of 15 content, and the cost of drying at the National Cereals bags/acre, equivalent to 1.35 MT/acre. At the current and Produce Board (NCPB) is K Sh 26 per percentage price of K Sh 3,100/bag and given a production cost of drop in moisture. 82 Agriculture Global Practice Technical Assistance Paper IMPACT OF RISK EVENTS The quality of both seed and fertilizer inputs has become Risks are inherent in wheat production right from plant- extremely erratic. In the recent past, wheat seed quality has ing through harvesting. Wheat farmers in Narok face the deteriorated to the extent that large-scale wheat farmers have following major risks, in decreasing order of importance: begun to import seed from as far as South Africa. Moreo- drought, pests and diseases, erratic input quality, volatile ver, wheat marketing has become a challenge because mill- prices, hailstorms, and high wind speed, which leads to log- ers, which are local producers’ only buyers, have not signed ging. Droughts in particular have become a major challenge a new agreement since 2008; that agreement is outdated and in lower Narok, with their frequency estimated at once every does not reflect the current cost of production. Given this two years. In upper Narok however, the major risk to wheat state of affairs, wheat millers dictate the prices paid to farm- production is hailstorms, which hit once every three to four ers, and prices have been changing from one year to the next. years, although their incidence has increased in the recent Farmers would prefer to negotiate prices each year. past. On average, drought reduces wheat yield from 15 bags/ acre to 10 bags/acre, a loss of 30 percent; at times drought RISK PRIORITIZATION leads to a total loss. Given the current prices of maize and In decreasing order of importance, the key risks faced by wheat and the ongoing threat of drought and hailstorm, wheat farmers in Narok include the following. some farmers are shifting to maize production. “The only reason why I continue to produce wheat is because it is fully mechanized unlike maize” (Mr. Olekoonyo). RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Crop Frequency Impact 1. Drought Wheat Once every two years Yield losses of up to 30% 2. Pests and diseases Wheat Occasionally Yield losses of up to 10% 3. Erratic input quality Wheat Each season Yield losses of up to 30% 4. Price volatility Wheat Occasionally Income losses of up to 20% 5. Hailstorms Wheat Occasionally 6. High wind speed Wheat Occasionally RISK MITIGATION STRATEGIES vast majority of the indigenous Maasai wheat farmers in A few large-scale farmers in Narok have begun to purchase Narok have not adopted any drought mitigation strategies. crop insurance from Cooperative Insurance Company Mr. Olekoonyo says he would rather let God be his insur- (CIC). Typically CIC insures the cost of wheat produc- ance. Mr. Olekoonyo’s failure to insure his crop is due to the tion, estimated at K Sh 24,000/acre or an equivalent yield experiences of some wheat farmers who had to go to court of 8 bags/acre. Insurance premiums are currently set at to claim compensation from the insurance companies. about 6 percent, or an equivalent of K Sh 1,450/acre. A CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Drought Drought-tolerant varieties Insurance Not effective Soil and water conservation Pests and diseases Use of chemicals Insurance Use of indigenous knowledge Somewhat effective in disease and pest control Erratic input quality Imports None Use of own seed Somewhat effective Price volatility Storage None Sourcing alternative markets Somewhat effective Hailstorms Insurance Planting edge trees Wind Insurance Planting wind breaks Kenya: Agricultural Sector Risk Assessment 83 grain moisture content, protein, bushel weight, levels of CASE STUDY 4: MARCEL aflatoxin, and all grain parameters in grain grading. WAMBUA—HEAD OF FINANCE, LESIOLO GRAIN The company’s average annual grain handling turnover HANDLERS LIMITED is 110,000 MT. Its drying charges are K Sh 28 per per- centage moisture drop per bag. In collaboration with the Eastern Africa Grain Council (EAGC) and other part- INTRODUCTION ners, LGHL participates in EAGC’s regulated Warehouse Marcel Wambua is the Head of Finance at Lesiolo Receipting System. This allows clients to deposit at least Grain Handlers Limited (LGHL), a medium-sized grain 10 tons of maize or wheat into certified silos during the handler located in Lanet of Nakuru County. Established harvest season when prices are low. The client is given in 2003 through funding from the International Finance a warehouse receipt that can be used in participating Cooperation (IFC) and the East Africa Development banks for financing. The banks provide financing up to Bank (EADB), LGHL’s existence stems from the belief 60–80 percent of the crop value, allowing clients to hold that agriculture in Kenya demands both modern and their crop until the prices are better. For example, farmers efficient grain handling services. LGHL’s management who deposited grain in January 2012 when the price was believes that these demands are best served by private K Sh 2,200 per 90-kg bag sold their maize in May 2012 at sector entities like itself. LGHL provides a compre- K Sh 3,400. They paid a total of 120 per bag for the five hensive solution for grain storage and related services. months of storage, earning a K Sh 1,080/bag margin. In The core services offered include grain drying, clean- addition, LGHL buys wheat, maize, sorghum, beans, and ing, fumigation, storage, and seed dressing, among a soybeans from farmers and sells them to local processing host of other services aimed at allowing customers the companies (e.g., flour millers, feed manufacturing plants, opportunity to realize the highest prices for their grain. breakfast cereal manufacturing firms, and humanitar- LGHL currently handles maize, wheat, and barley. Its ian relief organizations). The company also buys grain main customers are large-scale wheat farmers, primar- on behalf of millers, which reduces their logistical costs ily located in Nakuru, Narok, Moiben, Timau, and and frees their finances for production until they actu- Naivasha. The company was contracted by East Africa ally require the crop. LGHL offers competitive rates to Breweries Limited (EABL) to handle barley grown in farmers for their crops to enable them to have access the region under contract. to markets. OVERVIEW OF KEY ACTIVITIES IMPACT OF RISK EVENTS The major risks faced by the company emanate from gov- LGHL has an installed storage capacity of over 30,000 ernment policy, especially the operations of the National MT in a configuration of 16 silos and 8 wet bins. The Cereals and Produce Board (NCPB). The board’s buying equipment includes two Cimbria 54-ton dryers, two operations negatively impact LGHL’s buying operations Cimbria Delta 120 cleaners, a belt conveyor system, and since the board’s prices are always fixed at rates higher a 12-ton mini-dryer to cater to smaller customers. The than the market. The other risks faced emanate from farm company has the capability to handle bulk or bagged operations and include high moisture content in grains, grain and has a bagging unit to accommodate customers pests and diseases, and drought. with specific needs. LGHL also has 2 mobile dryers with a 12-ton holding capacity that are used to dry farmers’ grains at farm gate. These mobile dryers can be operated RISK PRIORITIZATION by tractor or using 3-phase electricity. In addition, the In decreasing order of importance, LGHL faces the fol- company runs a laboratory where it is able to determine lowing key risks: 84 Agriculture Global Practice Technical Assistance Paper RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Crop Frequency Impact 1. Government interference Wheat, maize All the time Profit losses of about 10% 2. High moisture content Wheat Occasionally Reduces farmer incomes 3. Pests and diseases Wheat Each season Reduces farmer yields 4. Price volatility Wheat Occasionally Income losses of up to 10% 5. Operational hazards Wheat Occasionally Reduces staff productivity (fire, injury, etc.) RISK MITIGATION STRATEGIES LGHL is fully insured against fire, theft, and burglary and maintains work insurance for all of its staff. In addition, it employs the following tools and mechanisms: CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Government policy Forward contracts Stock management Not effective High moisture content Pest and diseases Fumigation Insurance Fumigation Somewhat effective Price volatility Storage None Stock management Somewhat effective Operational hazards Maintenance of work Insurance safety standards the company has offered a Multi-peril Crop Insurance CASE STUDY 5: MICHAEL (MPCI) that covers risks associated with drought, excess WAIGWA—AGRICULTURAL rain, flooding, hail, and frost; over the period 2011–2012, UNDERWRITER, it paid out claims amounting to K Sh 130 million to wheat COOPERATIVE INSURANCE farmers in Narok alone. CIC’s loss ratio (claims divided by premiums) in 2011 was 170 percent, indicating that the COMPANY company incurred a loss from its MCPI business. INTRODUCTION OVERVIEW OF KEY ACTIVITIES Michael Waigwa is an agricultural underwriter at the The MCPI is reinsured by Swiss Reinsurance, which cov- Cooperative Insurance Company (CIC), a general insur- ers 80 percent of the risks while CIC covers the rest. Pre- ance company that provides crop insurance to wheat miums are shared in the same ratio as claims. Insurance farmers in Kenya. Mr. Waigwa has been in the crop coverage is provided subject to crop inspections conducted insurance business for the last four years. On average, the periodically by CIC. Three inspections are undertaken farms covered are in the range of 10–7,000 acres. In col- during the crop life: at germination, mid-season, and laboration with Swiss Reinsurance, an international rein- shortly preharvest. CIC provides farmers with a 65 per- surance company, CIC offers crop insurance to wheat, cent yield guarantee based on their long-term average barley, maize, and barley farmers. For the last four years, yield; for example, assuming a long-term average yield Kenya: Agricultural Sector Risk Assessment 85 of 12 bags/acre, the company would insure the equiva- insured; the majority want to be compensated for produc- lent of roughly 8 bags/acre. The figure of 8 bags/acre tion revenues rather than for the cost of production, even is estimated to be the cost of production for wheat. CIC though their insurance covers the latter. For example, CIC compensation enables farmers to recoup production costs currently faces a court case in which a farmer is claim- in case of a shock. For instance, if a farmer has bought ing K Sh 600,000 even though his yields were above the insurance and his yield falls to 6 bags/acre, CIC compen- 65 percent trigger threshold. Other challenges faced by sates the farmer for the difference of 2 bags/acre. The insurance companies include diversion of product, fraud, premiums charged are 6 percent of the sum insured, or and diversion of inputs in cases of contract farming, such an equivalent of K Sh 1,450/acre assuming a production as for barley. cost of K Sh 24,000/acre for wheat. RISK PRIORITIZATION IMPACT OF RISK EVENTS CIC insures the following major risks: Crop insurance in Kenya is a new phenomenon that many » Drought farmers have not yet taken advantage of, but Mr. Waigwa » Excessive rain/flooding and logging feels adequate capacity exists to undertake cereals insur- » Hailstorms ance in Kenya from CIC and other insurance companies, such as Union and Provisional (UAP). Crop insurance in These risks reduce the insurance company’s profits by an Kenya is a risky business, as evidenced by CIC’s loss ratio estimated 30 percent. On the other hand, CIC faces the in 2011. One key impediment to successful crop insurance following key risks: in Kenya is the lack of legislation that makes crop insur- » Fraud ance mandatory, as it is for motor vehicle insurance. In » Crop/input diversion addition, farmers have difficulties understanding the sum » Excessive litigation by clients RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Crop Frequency Impact 1. Fraud Wheat Occasionally Profit losses of about 10% 2. Crop/input diversion Wheat/barley Occasionally Reduces profits 3. Excessive litigation Wheat Occasionally Reduces profits RISK MITIGATION STRATEGIES Kenya Reinsurance in this scheme and to share the pre- CIC’s key risk mitigation strategy is reinsurance with miums and claims across CIC, Kenya Reinsurance, Africa Swiss Reinsurance. There is a new initiative to include Reinsurance, and Swiss Reinsurance. CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Fraud Reinsurance Crop diversion Reinsurance Excessive litigation Hire lawyers 86 Agriculture Global Practice Technical Assistance Paper Mr. Murunya’s main source of income is sale of live ani- CASE STUDY 6: WILSON mals. On average, a mature zebu steer raises K Sh 30,000 MURUNYA—LIVESTOCK during the normal season sales. Mr. Murunya acknowl- HERDER, KAJIADO COUNTY edges that in a year he can sell up to 20 steers depend- ing on financial need. The money acquired is used to buy food for his family, drugs for the livestock, supplementary INTRODUCTION feed, and sometimes children’s school fees. Wilson Murunya is a 30-year-old herder from Kajiado County with a primary-level education. Like many other young Maasai men, Mr. Murunya dropped out of school to RISK EVENTS start herding. He started keeping livestock at the age of 20 According to Mr. Murunya, drought is the main risk after inheriting 10 cows from his father. Mr. Murunya comes affecting herders in the region. During such times, herd- from a polygamous family; his father owns approximately ers are forced to migrate and look for pasture in other 200 acres of land that are shared among the extended fam- areas, sometimes trekking hundreds of kilometers. “In ily. He concentrates purely on herding, whereby he keeps 2007, there was an extensive drought and we had to move indigenous zebu and a few crosses with the dual-purpose our livestock in search of pasture. I moved up to Magadi sahiwal. He states that the dual-purpose sahiwal is good for where I lost over 70 cows while looking for pasture. Since milk production and faster growth. However, the crosses then, I have not been able to restock again and my cur- are less resistant to diseases and succumb more quickly to rent herd size is 50. During the same year, we went to herd drought than the zebus. Nevertheless, Mr. Murunya asserts our animals in the Tsavo National parks where some of that the sahiwal is more in demand despite these limita- our colleagues were killed by the lions. Drought not only tions. He managed to raise his herd to 120 heads of cattle, brings us livestock deaths but also loss of human life, and but currently has only 50 heads, all of which are indigenous our families back home also suffers” (Mr. Murunya). zebu, having lost almost 70 animals to drought. He also keeps a few goats, which are mainly raised by women and children. His herd includes 12 mature females, which when RISK PRIORITIZATION milked produce up to 1 liter per cow per day. According to » Drought the Maasai tradition, milk belongs to women and is either » Animal diseases consumed at home or sold in the market. » Unsustainable milk supply RISK EVENT MATRIX RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Livestock Frequency Impact 1. Drought Cattle, sheep, goats Every 10 years for a major drought, and Loss of pasture every 5 years for a minor drought Lack of water Low prices of livestock (50% drop) Livestock deaths. Mr. Murunya lost 70 out of 120 heads of cattle in 2007 2. Animal Cattle, sheep, goats Most of the diseases are endemic and Livestock deaths diseases occur every time there is a trigger. Most High costs of drugs common are FMD, mastitis, lumpy skin disease, anthrax. Occurrence of FMD and anthrax was very probable, while the other were rare. 3. Unsustainable Milk market prices When there is drought, the supply of Lack of stable markets milk supply milk is unstable; hence, most traders exit Low prices of milk the market. Kenya: Agricultural Sector Risk Assessment 87 RISK MANAGEMENT STRATEGIES Mr. Murunya relies on the following risk management strategies: CURRENT RISK MANAGEMENT PRACTICES Risk Type Risk Mitigation Risk Transfer Coping Strategy Effectiveness Drought Reserve grazing areas None Buying hay Not effective since hay has to during the dry season Outmigration be gotten from very far and is Increase the number expensive and sizes of water pans Migration leads to more disease Standing hay spread and animal deaths Animal diseases Vaccination campaigns None Regular Effective during the rainy season, Cattle dips vaccination but during the dry spells, animals are too weak to be sprayed Unsustainable Upgrading zebus with None Sourcing milk Not effective milk supply sahiwal to increase from other villages milk supply to the to sustain the market market demand 3. Low financial stability CASE STUDY 7: YUSUF 4. Insecurity and theft KHALIF ABDI—LIVESTOCK 5. Clan conflicts over pasture HERDER, GARISSA COUNTY 6. Social pressure 7. Price risks 8. Floods along the rivers INTRODUCTION Yusuf Khalif is a 65-year-old full-time herder from Modo- Mr. Khalif notes that “In 2010–2011, a major drought gashe in Garissa County. Although Mr. Khalif did not occurred, I moved with my 700 goats from Modogashe enumerate the number of animals he currently owns, he to Fafi, but I only returned back with 14 goats , all the said he keeps cattle, goats, and sometimes camels. His sole others died on the road due to lack of pasture.” Before source of income is sale of live animals. He occasionally outmigrating, Mr. Khalif wanted to sell his animals in a selects livestock from his herd and brings them to Garissa move to destock, but the community did not allow this; market for sale. community members believe that if an animal is meant to die, it will, so herders are pressured not to sell. A major RISK PRIORITIZATION drought occurred in 1987, and another in 2000. These According to Mr. Khalif, the major types of risks are droughts resulted in extensive livestock deaths; the com- 1. Recurrent drought munity was afraid to speak about them in case “It heard 2. Animal diseases and came back.” 88 Agriculture Global Practice Technical Assistance Paper RISK EVENT MATRIX RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Livestock Frequency Impact 1. Recurrent Cattle, goats, sheep, camel Every 10 years for a major Varied. In 2010–2011, Mr. Khalif drought drought, and every 2–3 years for lost 700 goats, or over 90% of his a minor drought stock 2. Animal Cattle, goats, sheep, camel Occasionally In 2007, farmers lost over 50% of diseases their stock from Rift Valley fever 3. Low financial Herders lack financial Often Herders’ inability to destock their stability capacity due to lack of animals when there is an early Sharia-compliant credit warning institutions 4. Insecurity Cattle, goats, sheep, camel Occasionally 5. Clan conflicts Grazing pastures Occasionally during drought Sometimes loss of lives as clans over pasture fight for grazing pasture Social Affects efforts to destock Often Farmers end up losing their pressure animals as the community puts pressure not to destock Price risks Market prices for cattle, Often The price of animals drops up to goats, sheep, camel 50% during a drought RISK MANAGEMENT STRATEGIES Mr. Khalif employs the following tools and mechanisms to manage risks: CURRENT RISK MANAGEMENT PRACTICES Risk Risk Type Risk Mitigation Transfer Coping Strategy Effectiveness Recurrent Establish irrigation None Livestock migration Not effective: Leads to drought schemes for pasture. A case Supplementary feeds like hay livestock deaths and disease in comparison is the Ewaso spread Nyiro irrigation scheme Animal Regular vaccination and None Vaccination Effective diseases vaccination campaigns Low financial Provision of Sharia- None Herders sell their animals and keep the Sometimes the herders lose the stability compliant credit schemes money in the house money due to lack of secure measures Insecurity Community surveillance None Community surveillance Relatively effective but the problems are always recurrent Clan conflicts Community peacekeeping None Clan elders solve major conflicts with Relatively effective but the over pasture programs the clan problems are always recurrent Social Community peacekeeping None Herder currently succumbs to social pressure programs pressure. For example, when Mr. Khalif was denied selling his goats due to drought, he decided to migrate and look for pasture, and in the process, some goats died Price risks Proper structures for None Herders are price takers, and have low livestock trade bargaining power Kenya: Agricultural Sector Risk Assessment 89 credit and artificial insemination services at a subsidized CASE STUDY 8: FRESHA rate. The Fresha processing plant has a good risk man- DAIRY—MILK PROCESSORS, agement strategy; while it supports its farmers to produce GITHUNGURI COUNTY more milk, the factory ensures that all milk is purchased from the farmers. The cooperative also ensures a steady and rising price per liter paid to its members. Fresha man- INTRODUCTION agement stated their farmers are not adversely impacted Fresha Dairy, opened in 2004 with a capacity of 50,000 by risk because liters per day, now has a capacity of 200,000 liters per » When there is less fodder, the cooperative buys hay day. The dairy cooperative owns the whole value chain, in bulk on behalf of its farmers, which they can including production, processing, and distribution. The pay for with the milk they supply. dairy processing plant is owned by the Githunguri Dairy » Fresha supplies other feed on credit, ensuring a Society, which started in 1961 with a membership of 34 steady supply of milk from farmers. and has grown to the current 23,120 members, of which » In addition to processing milk, the cooperative has 19,000 are active dairy farmers. It has 75 milk collection embarked on value addition and making long-life centers, at which 200,000–218,000 liters of milk are col- milk, which allow staggering of sales and hence lected from members daily. Fresha has 65 stores where more steady prices. members can access animal feed and other inputs on RISK EVENT MATRIX RISK FREQUENCY AND SEVERITY Rank Risk Type Affected Livestock Frequency Impact 1. Low-quality milk/spoilage Occasionally Low volume of milk processed 2. Drop in milk supply Occasional Low operational capacity 3. Oversupply of milk during the rainy season Milk spoilage RISK MANAGEMENT STRATEGIES Fresha Dairy employs the following risk management strategies: CURRENT RISK MANAGEMENT PRACTICES Risk Risk Type Risk Mitigation Transfer Coping Strategy Effectiveness Low-quality milk/ Milk quality checks at the farm level None Training farmers Effective spoilage Training farmers on clean milk production on clean milk Quality checks at the laboratory production Drop in milk supply Support to farmers to maintain regular supply of milk; None for example, buying hay and concentrates for farmers, which the farmers pay for after delivering the milk Oversupply of milk Milk value addition—process yoghurt and other dairy None during the rainy products season Establishment of a long-life milk processing plant to handle excess milk 90 Agriculture Global Practice Technical Assistance Paper APPENDIX C STAKEHOLDER VULNERABILITY ANALYSIS Agricultural shocks are one important factor driving chronic poverty and food insecu- rity in Kenya. Shocks impact household well=being in a variety of ways, by limiting food availability, weakening food access, and negatively affecting monetary well-being through the depletion of productive assets. Chronically vulnerable groups with high exposure to hazards experience a disproportionately large impact from adverse events and lack coping mechanisms available to other groups. In this context, vulnerability is a useful lens through which to examine agricultural shocks because it allows policy makers to determine which groups are most affected and to target risk management solutions accordingly. GENERAL TRENDS IN VULNERABILITY » Levels of human development, poverty, and food insecurity vary widely between regions. » Exposure to extreme weather events is highly correlated with being poor and being food insecure. » About 70 percent of Kenya’s poor live in the central and western regions, in areas that have medium to high potential for agriculture (IFAD 2013). » Poverty and food insecurity are acute in Kenya’s arid and semiarid lands (ASALs), which have been severely affected by recurrent droughts. VULNERABILITY, LIVELIHOODS, AND AGROCLIMATIC CONDITIONS Especially in rural areas, patterns of livelihood activities are strongly influenced by prevailing agroclimatic conditions, which determine planting calendars, soil quality, and crop suitability. Approximately 80 percent of Kenya’s land area lies in the ASALs, home to more than 30 percent of the population and 75 percent of the country’s live- stock (GoK 2011). ASAL districts have the highest incidence of poverty in the country, contain 18 of Kenya’s 20 poorest constituencies, and are predominantly inhabited by pastoralists and agro-pastoralists (GoK 2009). Pastoralist districts consistently rank below the national average in terms of Human Development Index (HDI) scores Kenya: Agricultural Sector Risk Assessment 91 FIGURE C.1. HUMAN DEVELOPMENT INDEX (figure C.1), as well as on other indicators of well-being. SCORES, BY PROVINCE These communities are among the most chronically food 0.7 insecure in the country and typically experience the high- 0.6 est rates of severe malnutrition. Several underlying factors 0.5 increase pastoral communities’ vulnerability, including land fragmentation, population growth, low literacy and 0.4 education provision, poor infrastructure, and weak market 0.3 integration. These chronic weaknesses undermine pasto- 0.2 ralist groups’ capacity to respond to shocks, like drought 0.1 and livestock disease outbreaks, which occur frequently in 0 the ASALs. In turn, the increasing frequency and simul- bi l y n st rn a n n tra taneous occurrence of multiple shocks erode the effective- lle er nz er io ro oa te at en st st va ya ai es C N Ea Ea N C N W ift ness of traditional coping mechanisms, creating a vicious R th or N cycle of crisis and underdevelopment. Source: Kenya National Human Development Report 2009. FIGURE C.2. MAP OF KENYA’S LIVELIHOOD ZONES KE01 - Northwestern Pastoral Zone KE02 - Turkwell Riverine Zone KE03 - Northwestern Agropastoral Zone KE04 - Lake Turkana Fishing Zone KE05 - Northern Pastoral Zone KE06 - Marsabit Marginal Mixed Farming Zone KE07 - Northeastern Agropastoral Zone KE08 - Mandera Riverine Zone KE09 - Northeastern Pastoral Zone KE10 - Grasslands Pastoral Zone KE11 - Southeastern Pastoral Zone KE12 - Coastal Low Potential Farming Zone KE13 - Coastal Marginal Agricultural Mixed Farming Zone KE14 - Tana Delta - Irrigated Zone KE15 - Southern Pastoral Zone KE16 - Southeastern Marginal Mixed Farming Zone KE17 - Southeastern Medium Potential, Mixed Farming Zone KE18 - Southern Agropastoral Zonevvvv KE19 - Central Highlands, High Potential Zone KE20 - Western Medium Potential Zone KE21 - Western High Potential Zone KE22 - Western Lakeshore Marginal Mixed Farming Zone KE23 - Lake Victoria Fishing Zone KE24 - Western Agropastoral Zone Source: FEWSNET 2011. incidence at the district level ranges from 94 percent to POVERTY STATUS AND 12 percent, and the poverty gap ranges from 218–230K. VULNERABILITY According to an econometric analysis of district-level In 2005–06, approximately 17 million Kenyans, or poverty data, stark spatial variations in the incidence and 47 percent of the population, were too poor to afford the depth of poverty arise from differences in agroclimatic cost of buying enough food to meet the recommended conditions and income-earning opportunities, as well as daily nutritional requirements and minimal nonfood unobserved factors (World Bank 2008). In other words, items. Of these 17 million people, more than 85 percent household location is an excellent predictor of livelihood live in rural areas. Spatial disparities in both the inci- activity, poverty status, and household consumption level dence and depth of poverty are pronounced: poverty (figure C.2). 92 Agriculture Global Practice Technical Assistance Paper TABLE C.1. POVERTY TRANSITIONS BY LIVELIHOOD GROUP Percentage Percent Percent Point Livelihood Remained Escaped Became Remained Poor at the Poor at the Change in Group Poor Poverty Poor Non-Poor Beginning End Poverty Mix farm, high 31 0 11 48 41 42 1 Farm/fish, low 24 7 14 55 31 38 7 Farm/ 43 11 13 33 54 56 2 livestock, low Pastoral 38 3 22 36 42 61 19 Urban 42 14 15 29 56 57 1 Average 35 9 14 42 44 50 5 Source: Mango, et al (2007). FIGURE C.3. HOUSEHOLD FOOD SECURITY BY LIVELIHOOD ZONE % of severely food insecure households 62 45 41 32 35 21 21 16 4 l al al al al al l al al na na or or or or or or or gi gi st st st st st st st ar ar pa pa pa pa pa pa pa m m ro n nd E n W rn al SE er er ag N he N st la st th oa ss ut rn Ea or So ra te C N es G W Source: World Bank 2008. Districts with high levels of poverty and food insecurity are Figure C.3 shows the percentage of severely food inse- also characterized by a high frequency of extreme weather cure households, as of May 2013, in areas where the events. Households in the bottom expenditure quintile are World Food Programme (WFP) operates. The graph the most likely to experience a weather-related shock. By reflects the food security status of nonbeneficiary house- virtue of their location, poorer households experience a holds, as opposed to WFP-beneficiary households. variety of natural hazards more frequently compared to better-off and richer households, and are less able to mobi- lize productive resources to respond to shocks. Table C.1 VULNERABLE GROUPS shows how the poverty status of different types of house- Certain population groups and certain types of house- holds changed between 1997 and 2005–06.33 At the end holds are more vulnerable to agricultural shocks than of the study period, the poverty rate was highest among others, depending on their level of exposure to risks, pastoralists (61 percent), and a higher percentage of pasto- susceptibility, and capacity to respond and/or recover ralists had slid into poverty than any other livelihood group from adverse events. In many cases, patterns of vul- (22 percent), indicating a substantial drop in well-being. nerability reflect underlying inequalities and social marginalization that preclude access to resources for individuals, households, or livelihood groups. The 33 “Mix farm, high” is a mixed crop area with high potential. “Farm/fish, low” is a mixed crop and fishing area with low or marginal potential, and “Farm/ groups identified below are especially vulnerable to livestock, low” is an agro-pastoral area with low or marginal potential. agricultural shocks: Kenya: Agricultural Sector Risk Assessment 93 » On average, FHHs have smaller farm sizes and PASTORALISTS lower education levels compared to their male » Pastoralist households are more likely to be poor counterparts. and more likely to be food insecure than nonpasto- » Roughly 49 percent of the total cultivated land ralist households. owned by MHHs is good to medium fertile land » Up to a fifth (15–20 percent) of households in the compared to 39 percent of land owned by FHHs Northern Pastoral Zone engage in begging, a rate (Kassie et. al. 2012). that is much higher than in any other livelihood zone. » The highest rates of global acute malnutrition are UNSKILLED/CASUAL WAGE in the Northeastern and Northwestern Pastoral LABORERS Zones (WFP 2013). » Casual wage laborers are considered particularly vulnerable to food price, production, and labor FEMALE-HEADED shocks since they purchase almost all of their food from the market. HOUSEHOLDS (FHHS) » During the 2008 food crisis, labor demand and » FHHs are 13 percent less likely to be food secure wage rates stagnated as food prices rose by up to than male-headed households (MHHs). 50 percent (KFSSG 2008). 94 Agriculture Global Practice Technical Assistance Paper APPENDIX D RAINFALL ANALYSIS FIGURE D.1. AGRO-ECOLOGICAL ZONES FIGURE D.2. MEAN ANNUAL RAINFALL (mm) ANNUAL RAINFAL F L LEGEND (in mm) Zone I 2000+ Zone II 1600–2000 Zone III 1200–1600 Zone IV Zone V 800–1200 Zone VI 600–800 400–600 Zone VII 200–400 200 or less Source: Adapted from Kenya Soil Survey 2009. Kenya: Agricultural Sector Risk Assessment 95 FIGURE D.3. MONTHLY CUMULATIVE RAINFALL PATTERNS BY RAINFALL ZONE (mm), 1981–2011 Dagoretti region Eldoret region Garissa region 250 250 212 250 200 200 164 200 158 152 152 150 134 150 150 118 102 101 1 101 106 100 66 100 75 73 62 69 100 77 52 63 51 50 37 50 33 37 37 50 16 35 34 16 22 25 2 16 5 5 6 7 0 0 0 Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Kisumu region Lodwar region Makindu region 250 250 250 202 200 200 200 163 165 1 156 150 135 150 150 121 10698 101 88 100 91 87 89 100 100 66 63 67 49 50 50 29 32 25 50 27 33 6 4 16 13 16 10 10 24 16 25 2 1 1 2 0 0 0 Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Malindi region Mandera region Mombasa region 350 250 265 311 300 250 200 250 200 150 157 200 165 150 128 150 143 96 101 100 81 70 100 94 94 67 57 68 100 57 47 60 39 46 48 51 42 30 50 22 50 25 50 9 4 15 10 2 5 2 1 1 3 0 0 0 Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Narok region Nyahururu region Voi region 250 250 250 200 200 200 163 150 150 133130 150 126 119 100 83 66 95 90 75 72 100 65 88 100 78 80 48 47 50 27 16 24 26 31 50 26 27 35 50 25 32 20 7 4 9 6 4 3 6 10 0 0 0 Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb March Apr May Jun Jul Aug Sep Oct Nov Dec Source: Kenya Meteorological Department. 96 Agriculture Global Practice Technical Assistance Paper FIGURE D.4. LOCATION OF REGIONAL WEATHER STATIONS IN KENYA Source: GFDRR 2012. Kenya: Agricultural Sector Risk Assessment 97 APPENDIX E WEATHER AND YIELD IMPACT ANALYSIS BACKGROUND The World Bank is conducting a study on the effect of several climatic events on dif- ferent crops’ yield in Kenya. The purpose of the study is to determine whether and by how much yield is affected by climatic events. Figure E.1 shows the political division of Kenya prior to implementation of the 2010 Constitution; at that time, Kenya had eight provinces. FIGURE E.1. PROVINCES IN KENYA BEFORE 2010 Provinces: 1 = Central, 2 = Coast, 3 = Eastern, 4 = Nairobi, 5 = North Eastern, 6 = Nyanza, 7 = Rift Valley, 8 = Western. Kenya: Agricultural Sector Risk Assessment 99 Similarly, in the Rift Valley Province, many rainy events SUMMARY AND KEY during 1998 explain low production there. FINDINGS Rainfall distribution in Kenya is highly variable. In very Wheat production in the Eastern Province is affected by dry years, like 1984, 2000, and 2009, little precipitation fell rainy events during the growing season. For example, in throughout most of the country. But 1968, 1997, and 2006 2004, higher than expected rainy events catastrophically were very humid years. This high variability in rainfall dis- affected wheat production. tribution affects the yield of Kenya’s main food crops. It is worth noting that the geographical resolution of the crop Rainfall occurs mostly between March and May; April is production information is very low, because it is only avail- the most humid month of the year. But a second rainy able at a provincial level; this might mask some relationships period also exists: November receives a lot of precipita- that could be found with a finer geographical resolution. tion in some regions of the country. Main crops like maize and beans are sown in March to take advantage of the WEATHER INFORMATION first rainy season. The dry season during July–September, Two data sources were used to analyze Kenya’s rainfall when crops are in the middle of their growing stage, can patterns. The Kenya Meteorological Department pro- affect their yield, however. vided a database with annual cumulative rainfall at 18 different weather stations from 1963 to 2012. This infor- In Kenya’s Central and Coast Provinces, statistical evidence mation is only available at an annual level. shows that drought during the growing stage greatly affects maize production. For example, maize production was low A gridded database from the Global Precipitation Climate in the Central Province during dry years like 2001, 2004, Project (GPCP, see http://precip.gsfc.nasa.gov/) was also and 2008, and in the Coast Province in 2011 and 2012. used. The resolution of the grid is 1 degree, so there are many pixel points with data from January 1, 1997, to Bean production is affected by too many rainy days dur- December 31, 2013, for the whole country. Figure E.2 ing the harvesting season in the Coast Province; for exam- provides a scope of the grid superimposed on a map of ple, in 1997, a very wet year, bean production was low. Kenya. Pixels in blue were considered for the analysis. FIGURE E.2. RAINFALL PIXELS SUPERIMPOSED ON A MAP OF KENYA 5 4 3 2 3 Latitude degrees 1 7 5 8 0 –1 1 –2 6 –3 4 2 –4 –5 33 34 35 36 37 38 39 40 41 42 Longitude degrees east 100 Agriculture Global Practice Technical Assistance Paper The bimodal pattern (long rains, short rains) is clear in ANNUAL RAINFALL most provinces, although rain falls more uniformly in DISTRIBUTION IN KENYA Western and Nyanza Provinces, the provinces farthest Rain in Kenya follows two seasonal patterns through- from the coast. April is the wettest month in all provinces, out the year: the “long rains” season, which occurs from with average cumulative rainfall fluctuating between 86 March/April to May/June; and the “short rains” season, and 148 mm, while August seems to be the driest month, which occurs from October to the end of the year. Figure E.3 with only 10 mm on average, except in the westernmost shows the average monthly distribution of rain for a pixel provinces, where cumulative rainfall ranges between 35 within all regions. and 102 mm. FIGURE E.3. MONTHLY RAINFALL PATTERN BY REGION Central province Coast province Eastern province Monthly average cumulative rainfall (mm.) Monthly average cumulative rainfall (mm.) Monthly average cumulative rainfall (mm.) 148 150 150 150 130 122 130 130 115 110 110 98 110 87 83 90 90 80 90 70 74 70 66 65 70 70 59 57 60 51 50 50 44 50 39 36 36 30 22 30 22 20 30 22 21 20 18 11 13 13 14 18 8 9 4 7 7 10 10 10 –10 –10 –10 br ry M ry ch M l ay ne Se Aug ly em t N cto r D em r em r r br ry M ry ch M l ay ne Se Aug ly em t N cto r D em r em r r br ry M ry ch M l ay ne Se Aug ly em t N cto r D em r em r r ri ri ri pt us pt us pt us O be ov be ec be be O be ov be ec be be O be ov be ec be be Ju Ju Ju Ap Ap Ap Fe ua ua Fe ua ua Fe ua ua ar ar ar Ju Ju Ju n n n Ja Ja Ja North eastern province Nyanza province Monthly average cumulative rainfall (mm.) Monthly average cumulative rainfall (mm.) 146 150 137 150 130 130 111 110 110 102 92 90 84 90 83 79 70 64 70 59 57 52 57 50 45 50 39 43 50 26 30 29 30 24 30 16 16 9 10 10 –10 –10 br ry M ry ch M l ay ne Se ug ly em t N cto r D em r em r r br ry M ry ch M l ay ne Se ug ly em t N cto r D em r em r r ri ri pt us pt us O be ov be ec be be O be ov be ec be be Ju Ju Ap Ap Fe ua ua Fe ua ua ar ar Ju Ju n n A A Ja Ja Rift valley province Western province Monthly average cumulative rainfall (mm.) Monthly average cumulative rainfall (mm.) 150 150 130 130 110 99.4 110 90 90 86 70 62.5 63 70 52.1 70 56 61 48.1 35.4 46 50 40.3 50 40 40 32.6 35.1 31.5 28 30 30 22.7 23.1 30 24 1 15.2 18 10 10 –10 –10 br ry M ry ch M l ay ne Se ug ly em t N cto r D em r em r r br ry M ry ch M l ay ne Se ug ly em t N cto r D em r em r r ri ri pt us pt us O be ov be ec be be O be ov be ec be be Ju Ju Ap Ap Fe ua ua Fe ua ua ar ar Ju Ju n n A A Ja Ja Source: Kenya Meteorological Department. Kenya: Agricultural Sector Risk Assessment 101 drought events, although 2009 was also generally dry. The DROUGHT AND EXCESS Narok station (about 150 km west of Nairobi) had con- RAINFALL ANALYSIS secutive extreme drought events in 2002 and 2003, with The Kenya Meteorological Department database was 0 mm and 50 mm of cumulative rain, respectively, but it used to analyze the differences among years, since its time is odd that no other station experienced a similar event. series is longer (50 years) than that of the gridded data- base, and it only includes annual cumulative rainfall. To Excess rainfall years were 1963, 1967, 1968, 1977, 1978, determine whether a year was dry or wet, the standard- 1988, 1998, 2006, and 2012. During these nine years, ized cumulative rainfall was calculated for each region rainfall was more than one standard deviation above aver- according to the following formula: age for at least five stations, meaning that rainfall was gen- erally plenty during these years. The year 1968 was the ( eci − ) most humid, with 10 stations showing a positive anomaly ii= StdRain i si in rainfall, and seven showing an extreme event. Nanyuki where station seems to be an outlier or to have a measurement StdRain, Standardized cumulative rainfall error, because its observation for 1968 is 9,895 mm, a fig- Prec, yearly rainfall ure completely out of line. Still, more than 2,000 mm fell m, mean yearly rainfall at several stations. The year 2012 was also wet, with seven s, standard deviation of yearly rainfall stations showing excess rainfall, particularly the Kisii and i, year Kericho stations (both close to Nyanza Province), which saw more than 2,000 mm of cumulative rainfall. Using the standardized cumulative rainfall makes it easier to identify drought and excess rainfall years. Table E.1 shows the standardized cumulative rainfall by year and RAINFALL—YIELD weather station, color coded as follows: red means an REGRESSIONS extreme drought event (StdRain < −2); orange means a The NASA database was used to determine the relation- drought event (StdRain < −1); light blue mean a light excess ship between rain and yield for the various crops since it rainfall event (StdRain > 1); and navy blue means an excess has daily data, allowing for a deeper analysis of dates and rainfall event (StdRain > 2). events. The only caveat is that these data do not cover the whole 50 years for which crop data exist. Several conclusions can be drawn from table E.1. First, there were 33 normal years, 7 drought years, and 8 excess It is worth noting that the geographical resolution of the rainfall years. Extreme drought and extreme excess rain- two data sets is not the same. Rainfall data are available fall years each occurred once. Second, extreme events on point estimates for pixels on a 1 × 1 degree grid, while seem to have occurred less frequently in recent times, a yield data are available by province. It is therefore nec- conclusion that might be at odds with the general concept essary to make equivalent the geographical resolution of of climate change. both data sets. Because no information exists regarding the sowing zones within each province, all available pixels Drought years were 1965, 1975, 1976, 1980, 1984, 1987, within the province were considered to match the yield 1993, and 2000. During these eight years, rain was more information of each province. Thus, the average of the than one standard deviation below average for at least five available pixels within a province was used as a proxy for stations. The year 1984 was particularly extreme, with 11 each province’s rainfall. stations showing drought, of which five had an extreme drought shock. For several stations, cumulative rainfall Figure E.4 depicts the calendar used to determine the was less than 500 mm for the whole year. The most recent sowing, growing, and harvesting seasons for the main generally dry year was 2000, when six stations showed crops in Kenya. 102 Agriculture Global Practice Technical Assistance Paper TABLE E.1. RAINFALL ANOMALIES FOR THE 18 WEATHER STATIONS Ext Ext Year Eldoret Embu Garissa Kajiado Kakamega Kericho Kiambu Kisii Kisumu Kitale Machakos Meru Mombasa Nakuru Nanyuki Narok Nyahururu Nyeri Dro Drought Normal Excess Excess Comment 1963 1.72 1.14 0.60 −1.42 0.30 0.71 2.33 −0.40 −0.16 −0.98 2.45 −0.03 0.46 0.22 0.15 2.12 0.93 0 1 11 5 3 Excess 1964 0.79 −0.69 −0.12 1.19 −0.02 0.40 −0.35 −0.40 0.65 −0.20 −0.22 0.79 −0.55 −0.17 0.08 0.82 0.59 0 0 16 1 0 Normal 1965 −0.88 −0.99 −0.26 −0.41 −0.42 −0.86 −0.77 −0.26 −0.89 −1.64 −0.44 −1.91 0.20 −2.00 −0.25 −1.69 −1.04 0 5 12 0 0 Drought 1966 0.80 0.11 −0.58 0.55 −0.98 −0.60 0.92 1.78 1.66 0.39 0.38 0.85 0.37 0.28 0.01 −0.53 −0.54 0 0 15 2 0 Normal 1967 1.38 1.56 0.88 −0.12 −0.53 −0.27 1.42 −0.51 −0.66 −3.33 0.89 2.22 1.09 −0.51 0.04 0.00 0.15 1 1 11 5 1 Excess 1968 −0.04 2.22 3.49 −0.10 0.22 0.79 2.17 −0.48 2.03 −0.40 2.38 2.65 1.39 0.44 6.85 1.60 1.39 0 0 7 10 7 Ext Excess 1969 −0.61 −0.24 −0.47 0.00 −0.73 −2.39 −1.44 0.22 −0.91 −0.56 0.22 1.08 −0.28 −0.95 0.03 −0.76 −0.94 1 2 14 1 0 Normal 1970 1.60 −0.65 −0.71 1.66 0.84 0.57 −0.17 1.25 −0.61 0.23 −0.17 1.16 −0.64 0.76 −0.12 0.30 0.18 0 0 13 4 0 Normal 1971 0.01 0.49 0.01 0.33 −0.47 −1.56 0.52 4.12 0.28 −0.10 0.25 0.03 −0.99 −0.39 −0.01 0.51 −1.08 0 2 14 1 1 Normal 1972 0.16 1.60 −1.04 −0.59 0.48 −1.41 −0.43 0.66 0.60 1.69 −0.05 2.59 0.79 −0.47 −0.12 −0.47 −0.11 0 2 12 3 1 Normal 1973 −0.91 −1.14 −0.55 0.29 1.50 −0.83 −1.28 −0.53 −1.75 −0.84 −0.14 0.11 −0.04 −0.85 −0.16 −0.85 −0.43 0 3 13 1 0 Normal 1974 −1.11 −0.11 −0.69 0.20 −0.98 −1.19 −0.46 −0.47 −0.93 −0.81 0.38 1.12 −0.83 −0.03 −0.03 0.22 −0.62 0 2 14 1 0 Normal 1975 1.18 −1.18 −0.20 0.33 0.07 0.81 −1.12 −0.03 −0.60 −0.31 −0.35 −1.03 −0.50 1.02 −0.09 0.58 −1.06 0 4 11 2 0 Drought 1976 −0.43 −0.47 −0.98 −1.45 −0.51 −1.21 −0.60 0.24 −0.52 −1.13 −0.56 −1.00 −0.43 −0.51 0.04 −0.78 −0.58 0 4 13 0 0 Drought 1977 2.49 0.63 0.37 1.85 1.33 −0.51 1.38 1.56 1.32 2.01 0.63 0.76 0.19 0.86 0.04 1.90 −0.15 0 0 9 8 2 Excess 1978 −0.12 0.85 2.32 0.59 0.09 1.64 1.40 0.52 2.06 1.36 1.28 1.64 0.38 0.97 0.01 1.53 0.35 0 0 9 8 2 Excess 1979 0.24 0.37 1.13 −0.07 −0.47 −0.58 −0.47 −1.52 0.40 0.45 2.21 0.59 0.97 −0.12 −0.13 0.30 1.41 0 1 13 3 1 Normal 1980 −0.90 −0.94 −1.51 −0.59 0.07 −1.21 −0.21 −1.07 −1.48 −1.10 0.90 −1.01 −0.48 −0.61 −0.40 −0.78 −1.18 0 7 10 0 0 Drought 1981 0.53 0.74 −0.28 −0.41 1.04 −0.83 0.32 −0.97 −1.81 0.80 0.70 0.38 0.05 −0.21 −0.04 0.37 −1.06 0 2 14 1 0 Normal 1982 −0.15 0.37 0.52 0.03 1.52 −0.27 −0.12 −1.52 0.40 1.80 −0.56 0.70 1.31 0.30 −0.16 0.69 0.89 0 1 13 3 0 Normal 1983 0.00 −0.42 −1.40 −0.61 −0.29 1.04 −0.80 −0.39 −1.37 −0.11 −1.44 −0.87 0.07 −0.25 −0.29 −0.33 0.26 0 3 13 1 0 Normal 1984 −2.95 −1.10 0.92 −1.27 −2.08 −2.54 −2.65 −1.14 −0.77 −1.49 −0.87 −0.21 −0.82 −1.61 −0.29 −2.67 −1.24 5 11 6 0 0 Ext Dro 1985 −0.67 0.76 −0.78 −0.94 3.01 1.40 −0.42 0.00 −0.12 0.35 0.03 −0.13 −0.94 −0.06 −0.26 0.45 −0.17 0.00 0 0 16 2 1 Normal 1986 −1.08 −0.54 0.03 −0.65 −0.76 −0.80 −0.53 −0.75 −0.35 −1.64 −0.44 −0.59 −1.97 −0.22 −0.21 −0.83 −0.09 0.84 0 3 15 0 0 Normal 1987 0.22 −0.87 −1.78 −2.30 −0.19 0.51 −0.25 1.27 −0.39 −0.30 −1.97 −1.49 −2.04 −1.12 −0.30 0.37 −2.28 −1.21 3 8 9 1 0 Drought 1988 −0.39 2.08 0.26 1.67 1.48 1.18 0.53 0.20 0.06 0.00 −0.70 −1.48 0.97 −0.25 0.35 0.59 1.18 0 1 11 5 1 Excess 1989 −0.70 −0.44 −0.19 −0.10 0.60 1.62 −1.12 −0.31 0.04 −0.19 −0.55 −1.53 3.94 −0.06 2.13 −0.85 −0.62 0 2 12 3 2 Normal 1990 −0.31 1.20 1.03 2.80 0.11 0.76 0.43 −0.85 −0.28 0.69 0.06 0.36 −0.05 −0.16 0.06 0.19 0.62 0 0 14 3 1 Normal 1991 −0.39 −1.05 −0.07 1.40 0.26 0.42 −0.40 −1.42 −0.69 0.06 −0.72 −0.18 0.38 2.39 −0.50 −1.02 0.38 0 3 12 2 1 Normal 1992 0.02 −1.96 −0.21 −0.15 −0.80 −0.03 0.04 −1.17 −0.65 0.08 −0.77 −0.79 −0.32 −0.34 −0.40 −0.89 −0.80 −0.67 0 2 16 0 0 Normal 1993 −1.76 −1.19 −0.55 1.08 −1.54 −0.64 −0.69 −1.05 −1.12 0.28 −0.48 −0.65 −0.47 −1.11 −0.44 0.30 −0.33 −0.18 0 6 11 1 0 Drought 1994 0.51 −0.39 0.50 1.59 −0.22 0.38 −0.15 0.44 0.66 0.13 0.28 −0.39 2.53 −0.21 −0.32 1.26 0 0 13 3 1 Normal 1995 −0.50 0.51 0.37 −0.07 −0.55 1.15 0.23 0.42 0.61 −0.70 −0.59 −0.01 −0.29 −0.32 −0.12 −0.05 0.38 0 0 16 1 0 Normal 1996 0.00 −1.03 0.19 0.51 −1.53 0.56 0.86 −0.32 −0.91 −0.28 −0.23 −0.20 0.38 −0.49 −0.99 0 2 13 0 0 Normal (continued) TABLE E.1. RAINFALL ANOMALIES FOR THE 18 WEATHER STATIONS (continued) Ext Ext Year Eldoret Embu Garissa Kajiado Kakamega Kericho Kiambu Kisii Kisumu Kitale Machakos Meru Mombasa Nakuru Nanyuki Narok Nyahururu Nyeri Dro Drought Normal Excess Excess Comment 1997 0.55 −1.19 −0.26 0.34 −0.21 1.27 −0.01 0.06 2.84 0.18 0.11 0.97 0.48 3.06 0 1 10 3 2 Normal 1998 1.93 −0.33 1.01 0.91 −0.69 −1.03 0.83 0.36 2.34 −0.01 0.07 0.53 1.14 1.24 0 1 8 5 1 Excess 1999 −0.06 −0.79 −0.13 0.08 0.27 0.70 −0.27 −0.55 0.52 −1.10 −0.28 −0.22 −0.75 −1.58 0 2 12 0 0 Normal 2000 −1.02 −1.79 −0.31 −0.58 0.07 −0.87 −0.89 −1.16 0.88 −1.47 −0.39 −0.92 −1.12 −1.50 0 6 8 0 0 Drought 2001 0.32 0.06 2.11 0.50 0.42 0.60 0.85 −0.78 −0.26 0.53 −0.06 0.08 0.90 0.37 0 0 13 1 1 Normal 2002 −1.00 0.11 0.71 0.26 0.62 1.62 −0.19 −0.75 0.03 0.35 −0.15 −2.57 −0.09 0.42 1 1 12 1 0 Normal 2003 −0.49 0.66 0.19 −1.23 0.73 −0.62 0.50 −0.77 0.04 0.53 −0.04 −2.39 −0.17 0.68 1 2 12 0 0 Normal 2004 −0.15 −1.31 −0.32 0.55 0.07 0.15 −0.51 −0.67 −0.51 −0.31 −0.10 −0.27 −0.26 −0.44 0 1 13 0 0 Normal 2005 −0.65 0.14 −0.25 −0.38 −1.16 0.72 −0.94 −0.97 −0.30 −0.41 −0.13 0.24 −1.16 0 2 11 0 0 Normal 2006 1.45 1.73 1.66 0.93 1.42 1.33 0.42 1.15 0.94 1.46 1.19 0.89 −0.06 −0.16 1.26 0.36 1.78 0 0 7 10 0 Excess 2007 0.25 −0.08 −0.24 −0.02 0.60 0.24 −0.85 −1.07 1.30 −0.94 0.03 0.32 0.86 −0.10 −0.05 1.46 0.15 0 1 14 2 0 Normal 2008 −0.42 −0.62 −0.41 −0.18 −0.06 −0.93 −0.22 −0.79 −0.14 −1.35 −0.91 −0.39 −0.74 −0.26 −0.38 −0.17 −0.79 0 1 16 0 0 Normal 2009 −1.05 −0.48 −0.40 −0.74 −0.66 −0.92 0.01 0.25 0.02 −1.31 0.11 −0.75 −1.05 −0.33 −0.49 −1.14 −0.98 0 4 13 0 0 Normalw 2010 1.13 −0.32 0.36 0.15 0.96 0.49 0.83 0.60 1.05 −0.21 0.15 −0.17 1.46 −0.07 0.20 1.27 0.07 0 0 13 4 0 Normal 2011 0.11 −0.04 −0.41 −0.05 0.63 −0.70 0.26 1.03 1.10 −0.86 0.47 −0.61 0.41 0.04 0.80 0.30 −0.15 0 0 15 2 0 Normal 2012 1.33 0.56 −0.60 0.16 1.27 −0.33 0.87 1.71 1.54 −0.39 0.46 −0.80 0.79 0.03 1.02 1.34 1.69 0 0 10 7 0 Excess Extreme 1 0 0 1 1 2 1 0 0 1 0 0 1 0 0 2 2 0 Drought Drought 6 7 4 4 5 7 6 9 7 6 4 6 4 7 0 2 6 10 Ext Dro 1 Normal 35 27 31 19 38 35 35 36 34 36 31 36 40 38 49 21 35 32 Drought 7 Excess 9 7 5 6 7 8 9 5 9 8 5 8 6 4 1 3 8 8 Normal 33 Extreme 1 2 2 0 2 1 2 1 2 1 3 3 3 2 1 1 1 1 Excess 8 Excess Prob Ex 2% 0% 0% 3% 2% 4% 2% 0% 0% 2% 0% 0% 2% 0% 0% 8% 4% 0% Ext 1 Drought Excess Prob 12% 17% 10% 14% 10% 14% 12% 18% 14% 12% 10% 12% 8% 14% 0% 8% 12% 20% Drought Prob 70% 66% 78% 66% 76% 70% 70% 72% 68% 72% 78% 72% 80% 78% 98% 81% 71% 64% Normal Prob 18% 17% 13% 21% 14% 16% 18% 10% 18% 16% 13% 16% 12% 8% 2% 12% 16% 16% Excess Prob Ext 2% 5% 5% 0% 4% 2% 4% 2% 4% 2% 8% 6% 6% 4% 2% 4% 2% 2% Excess Source: Kenya Meteorological Department. FIGURE E.4. CALENDAR FOR MAIN The following rainfall parameters were estimated for each CROPS IN KENYA crop season (sowing, growing, and harvesting): Kenya » Cumulative rainfall (cumrain)—The sum of Crop calendar (*major foodcrop) daily precipitation in millimeters for each season described above. Barley (Long rains) » Number of rainy events (events)—The number of days in the season in which rain is greater than 5 mm. Barley, Maize, Millet & Sorghum (short rains) Figure E.5 illustrates the average cumulative rainfall dur- Beans (Long rains) ing 1997–2012, which covers the NASA rainfall database Beans (Short rains) for the March–November timeframe. Maize (Long rains)* Figure E.5 shows that Western and Nyanza Provinces are Millet (Long rains) the most humid, getting more than 1,000 mm of rain- fall on average. North and Coast Provinces follow with Sorghum (Long rains) more than 800 mm of rainfall on average. But most of the Wheat (Long rains)* center of the country is very arid, with less than 400 mm J J A S O N D of rainfall on average for this period of the year. Lean period (N, S & E pastoral areas)- FEWSNET To determine the relationship between yield and rain, lin- Lean period (N, S & E ear regression models were run using both rain parame- pastoral areas and ters during each stage of the crop cycle as the explanatory SE marginal cropping areas)- variable for yield. FEWSNET Yield = β0 + β1 cumrainsow Lean period (central Yield = β0 + β2 cumraingrow cropping areas)- FEWSNET Yield = β0 + β3 cumrainharvest Sowing Yield = β0 + β4 eventsow Growing Yield = β0 + β5 eventgrow Harvesting Yield = β0 + β6 eventharvest FIGURE E.5. MAP OF AVERAGE CUMULATIVE RAINFALL, BY PIXEL 1,600 1 600 Cumulative rainfall (mm.) 1,400 1,200 1,000 800 600 400 200 5 3 Latitude 1 1 –1 –3 41 2 42 –5 39 40 36 37 38 34 35 Longitude 0-200 200-400 200 400 400 400-600 600 600 800 600-800 800 1,000 800-1,000 1,000-1,200 1,000 1,200 , , 1,200-1,400 1,4 1,400-1,600 Kenya: Agricultural Sector Risk Assessment 105 TABLE E.2. SIMPLE LINEAR REGRESSION RESULTS FOR MAIZE ProvNo Province CumRain1 Cumrain2 CumRain3 Events1 Events2 Events3 1 Central 2.0% 60.0% 0.0% 4.3% 59.6% 0.0% 2 Coast 8.0% 44.3% 0.4% 5.3% 20.7% 0.0% 3 Eastern 0.3% 6.6% 0.2% 1.8% 4.1% 0.6% 6 Nyanza 2.8% 0.1% 0.0% 2.3% 0.0% 1.3% 7 Rift Valley 1.1% 0.8% 3.3% 5.0% 0.0% 3.5% 8 Western 9.4% 13.1% 12.8% 5.4% 5.0% 4.2% TABLE E.3. MULTIPLE LINEAR REGRESSION RESULTS FOR WHEAT ProvNo Province CumRain1-CumRain3 Events1-Events3 1 Central 10.1% 31.1% 3 Eastern 18.7% 36.3% 7 Rift Valley 28.5% 28.4% TABLE E.4. SIMPLE LINEAR REGRESSION RESULTS FOR WHEAT ProvNo Province CumRain1 CumRain2 CumRain3 Events1 Events2 Events3 1 Central 7.3% 0.0% 0.7% 4.4% 3.4% 0.8% 3 Eastern 3.7% 17.0% 1.2% 1.0% 32.6% 1.1% 7 Rift Valley 4.4% 22.1% 1.7% 12.1% 20.3% 2.3% The main objective of the regression analysis is to calcu- Wheat: Wheat production data were only available for late the determination coefficient (R2) for each variable. three provinces. Table E.3 summarizes the regression The determination coefficient is a measure of the propor- determination coefficient for a multiple linear regression tion of the variability in yield explained by each rainfall analysis using the three seasons of the cumulative rain- variable. Therefore, the higher the R2, the more likely the fall and the three seasons of the rainy event variables by particular rain parameter and yield are related. Regres- province. sion analyses for each crop and province follow. As table E.3 shows, generally speaking, the number of Maize: Table E.2 summarizes the regression determina- rainy events in the three seasons (sowing, growing, and tion coefficient for each rain parameter by province. harvesting) explains more variability in wheat yield than cumulative rainfall does. Table E.4 summarizes the simple Table E.2 shows that cumulative rainfall seems to bet- linear regression analysis performed using each variable ter explain maize yield than the number of rainy events. as a regressor. Cumulative rainfall during the growing season explains a significant amount of the variability in maize yield for the Once each season is analyzed separately, it can be seen Central and Coast Provinces. The number of rainy events that as with maize, the growing season explains more during the growing season also explains a significant pro- variability in wheat yield than the other seasons. But portion of variability in maize yield in these provinces, the number of rainy events during the growing season is but slightly less than cumulative rainfall. Rain during the the most significant variable, explaining between 19 and harvest season is not significant in explaining maize yield. 33 percent of the variability in yield. 106 Agriculture Global Practice Technical Assistance Paper APPENDIX F CROP PRODUCTION TRENDS FIGURE F.1. MAIZE PRODUCTION, 1990–2012 4,000 25,000 Production (tons) Area harvested (Ha) Yield (Hg/Ha) 3,500 Production/area (in 000s) 20,000 3,000 2,500 15,000 Yield 2,000 1,500 10,000 1,000 5,000 500 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. FIGURE F.2. WHEAT PRODUCTION, 1990–2012 Production (tonnes) Area harvested (Ha) Yield (Hg/Ha) 600 35,000 Production/area (in 000s) 500 30,000 25,000 400 20,000 Yield 300 15,000 200 10,000 100 5,000 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. Kenya: Agricultural Sector Risk Assessment 107 FIGURE F.3. DRY BEAN PRODUCTION, 1990–2012 1,200 25,000 Production (tonnes) Area harvested (Ha) Yield (Hg/Ha) Production/area (in '000) 1,000 20,000 800 15,000 Yield 600 10,000 400 200 5,000 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. FIGURE F.4. TEA PRODUCTION, 1990–2012 450 Production (tonnes) Area harvested (Ha) Yield (Hg/Ha) 30,000 Production/area (in 000s) 400 25,000 350 300 20,000 250 Yield 15,000 200 150 10,000 100 5,000 50 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 0 Source: FAOSTAT. FIGURE F.5. COFFEE PRODUCTION, 1990–2012 200 Production (tones) Area harvested (Ha) Yield (Hg/Ha) 8,000 Production/area (in 000s) 180 7,000 160 6,000 140 120 5,000 Yield 100 4,000 80 3,000 60 2,000 40 20 1,000 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. FIGURE F.6. SUGARCANE PRODUCTION, 1990–2012 7,000 Production (tones) Area harvested (Ha) Yield (Hg/Ha) 90,000 Production/area (in 000s) 6,000 80,000 70,000 5,000 60,000 4,000 Yield 50,000 3,000 40,000 30,000 2,000 20,000 1,000 10,000 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. 108 Agriculture Global Practice Technical Assistance Paper APPENDIX G LIVESTOCK TERMS OF TRADE ANALYSIS Kenya: Agricultural Sector Risk Assessment 109 110 –2 0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 Av Jan 06 Av Jan 06 Av Jan 06 Av May 06 Av May 06 Av May 06 Av Sep 06 Av Sep 06 Av Sep 06 Av Jan 07 Av Jan 07 Av Jan 07 Source: MoALF 2014. Av May 07 Av May 07 Av May 07 Av Sept 07 Av Sept 07 Av Sept 07 Av Jan 08 Av Jan 08 Av Jan 08 Av May 08 Av May 08 Av May 08 Av Sept 08 Av Sept 08 Av Sept 08 Wajir Mombasa Moyale av Jan 09 av Jan 09 av Jan 09 Av May 09 Av May 09 Av May 09 Av Sept 09 Av Sept 09 Av Sept 09 for 1 beef cow) Av Jan 10 Av Jan 10 Av Jan 10 Av May 10 Av May 10 Av May 10 Av Sept 10 Av Sept 10 Av Sept 10 Av Jan 11 Av Jan 11 Av Jan 11 0 5 10 15 20 25 0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 Av Jan 06 Av Jan 06 Av Jan 06 Av May 06 Av May 06 Av May 06 Av Sep 06 Av Sep 06 Av Sep 06 Av Jan 07 Av Jan 07 Av Jan 07 Av May 07 Av May 07 Av May 07 Av Sept 07 Av Sept 07 Av Sept 07 Av Jan 08 Av Jan 08 Av Jan 08 Av May 08 Av May 08 Av May 08 Av Sept 08 Av Sept 08 Isiolo Av Sept 08 Garissa av Jan 09 Dagoretti av Jan 09 av Jan 09 Av May 09 Av May 09 Av May 09 Av Sept 09 Av Sept 09 Av Sept 09 Av Jan 10 Av Jan 10 Av Jan 10 2006–11 (number of 90-kg bags of maize exchanged Av May 10 Av May 10 Av May 10 Av Sept 10 FIGURE G.1. TOT OF INDIVIDUAL MARKETS IN NORTHERN KENYA, Av Sept 10 Av Sept 10 Av Jan 11 Av Jan 11 Av Jan 11 Agriculture Global Practice Technical Assistance Paper APPENDIX H OPTIONS FOR SCALING UP LIVESTOCK INSURANCE IN KENYA In the recent past, livestock insurance has gained a lot of interest in Kenya as a viable solution to addressing covariate risks like those associated with drought. International Livestock Research Institute (ILRI) has piloted index based livestock insurance (IBLI) in arid and semiarid land (ASAL) regions of Kenya. The pilot started in Marsabit County in 2010 is now in three counties (Marsabit, Wajir, and Isiolo), and ILRI has plans to expand to all 14 counties34 in ASAL regions. Through IBLI, farmers are able to cushion themselves against the impact of droughts, which have increased in frequency and severity, a phenomenon associated with climate change. In Kenya, 28 severe droughts have been recorded in the last 100 years, including four droughts in the last 10 years. In the major drought of 2000, major animal losses occurred. Analysis undertaken by ILRI on IBLI has shown that providing access to insurance creates an effective safety net for vulnerable-but-non-poor pastoralists. The importance of agricultural insurance in addressing food security seems to have been embraced by Kenya’s current crop of politicians. In the 2013 presidential elec- tion, provision of agricultural insurance was one of the key pledges of candidates. The coalition that won the election included livestock insurance as part of its party mani- festo. Kenya’s agricultural/livestock insurance policy is clearly spelled out in Executive Order No. 2, which outlines what the current government intends to achieve. Medium Term Plan II (MTP-II), which covers 2013–2017, recognizes the importance of live- stock insurance and talks of establishing a National Livestock Insurance Scheme. To keep its political promise and to implement the projects spelled out in MTP-II, the government of Kenya (GoK) through the Ministry of Agriculture, Livestock and Fisheries (MoALF) sought support from the World Bank Agriculture Insurance Devel- opment Program (AIDP) to assist in formulating a large-scale national agricultural insurance program as a public-private partnership (PPP). Starting in January 2014, Lamu, Isiolo, Laikipia, Mandera, Marsabit, West Pokot, Turkana, Tana River, Garissa, Baringo, Samburu, Narok, 34 Samburu, and Wajir Counties. Kenya: Agricultural Sector Risk Assessment 111 the AIDP team worked closely with State Department of Turkana, and Wajir Counties and use of HSNP’s livestock Livestock (SDL) to think through an appropriate livestock census and classification system for targeting. Beneficiar- insurance PPP for the government. In partnership with ies of the SDL-paid macro coverage will be “vulnerable the World Bank’s Financial Sector Development unit and pastoralists” immediately above HSNP target beneficiar- ILRI, AIDP will assist the MoALF in implementing a ies (the 100,000 poorest households) in these counties. macro-level livestock Normalized Difference Vegetation AIDP has agreed to target households above those already Index (NDVI) insurance program to address the SDL’s receiving a nonconditional cash transfer from HSNP and objective of enhancing resilience and reducing vulner- to provide insurance coverage for five Tropical Livestock ability of small-scale pastoral farmers. Units (TLUs) for selected households. Under the SDL-driven livestock insurance initiative, a HSNP counties were selected to introduce the macro- graduated approach is envisaged, with government pur- level insurance product because household censuses have chasing an insurance product at a macro level for targeted already been done in them, their poverty levels have been vulnerable households, while wealthier households will be determined, and the infrastructure to pay herders (in the able to purchase the product on an individual, voluntary event of a payout) has been established. HSNP’s infra- basis. It is envisaged that the macro livestock insurance structure will be used to register and make payouts35 to product will offer asset protection, having early payouts the SDL beneficiaries. ILRI-supported IBLI will also be with the onset of drought to empower pastoralists to pro- available and accessible to those who want to top-up (e.g., tect their herds. The voluntary component is expected to those covered by the government-paid insurance) or those initially be an asset replacement-type product, paying out who want to voluntarily purchase it (e.g., those who are at the end of the drought season; however, in the medium not under macro-level coverage). The voluntarily pur- term, the SDL with the support of AIDP will develop an chased coverage will be available in more counties than asset protection-type product to be offered to this group as those covered under HSNP (because it will not require well. The asset protection cover would use the ILRI NDVI HSNP infrastructure to operate and there would be no database and methodology to make timely payouts to tar- need to target). geted vulnerable pastoralists (beneficiaries) at the onset of severe drought, thus reducing livestock mortality and Strong synergies would be gained from working closely asset depletion. This product assumes that investments with the World Bank–supported Regional Pastoral Live- will strengthen availability of animal feed, destocking, and lihood Resilient Project (RPLRP) to promote livestock other critical market services. insurance in the targeted counties. RPLRP’s objective is to enhance the resilience of pastoral and agro-pastoral The macro-level NDVI index insurance cover would be communities in drought-prone areas through regional purchased by SDL-GoK as part of their national drought approaches. The project will be implemented in 14 coun- risk reduction and risk financing strategy for pastoral- ties, and one of its key components is pastoral risk man- ists in ASAL regions; SDL would be the insured party, agement, which corresponds well with the objective of responsible for payment of the premium for the macro- promoting livestock insurance. level coverage. SDL has also requested that AIDP design: (1) voluntary top-up coverage for targeted beneficiaries; and (2) a micro-level individual livestock producer policy. 35 HSDL is working with Equity Bank to ensure that every beneficiary house- AIDP has proposed to SDL linkage of the macro-level hold of the cash transfer program has a bank account. It is hoped that a simi- livestock index insurance product with the Hunger Safety lar agreement can be reached with Equity Bank for the separate AIDP-SDL Net Program (HSNP) program in Mandera, Marsabit, macro-level livestock index insurance program. 112 Agriculture Global Practice Technical Assistance Paper APPENDIX I RESULTS OF SOLUTIONS FILTERING PROCESS FOOD CROPS FIGURE I.1. PRIORITIZATION OF RISK MITIGATION SOLUTIONS FOR FOOD CROPS Average score (max. score 25) Increase predictability of government interventions 18.1 Training on post-harvest management 21.1 Fumigation 16.0 Apply moisture and quality standards 18.5 Hermetically sealed silos 16.7 Safety net interventions 17.5 Drought tolerant seeds 21.8 Irrigation 16.8 Water harvesting 20.5 Conservation farming 20.1 Source: Authors’ notes. FIGURE I.2. PRIORITIZATION OF RISK TRANSFER SOLUTIONS FOR FOOD CROPS Average score (Max. score 25) Warehouse receipt system 19.8 Weather insurance 17.1 Source: Authors’ notes. FIGURE I.3. PRIORITIZATION OF RISK COPING SOLUTIONS FOR FOOD CROPS Average Score (Max. Score 25) Promote farmer association operated storage centers 18.8 Cash for work 17.6 Food for work 16.6 Programs to strengthen livelihoods & social capital 17.9 Source: Authors’ notes. Kenya: Agricultural Sector Risk Assessment 113 CASH CROPS FIGURE I.4. PRIORITIZATION OF RISK MITIGATION SOLUTIONS FOR CASH CROPS Average Score (max. score 25) Tea: Work on a branding policy to assure more extensive 20.9 participation in niche markets and diversified export destinations Sugar: Maintain and enforce a long term market policy including 19.8 indicators about meeting COMESA agreements. Coffee: Open the market to ensure greater price transmission to 21.2 farmers and more incentives to invest in production. Source: Authors’ notes. LIVESTOCK FIGURE I.5. PRIORITIZATION OF RISK MITIGATION SOLUTIONS FOR LIVESTOCK Average score Capacity building (farmers and local officers) 20.8 Irrigated fodder production 13.3 Strengthening community customary governance 19.8 Community peace keeping programs 19.2 Intensification and strengthening of disease surveillance 20.2 Increase water conservation pans 18.7 Conditional parks grazing /Wildlife /livestock coexistence 15.5 Destocking 19.3 Increased vaccination campaign 22.2 Controlled livestock movement 19.7 Reserve grazing pastures /standing pasture 16.9 Institutional reform 21.5 Livestock micro-finance 18.5 Animal feed: haymaking and storage, irrigated fodder 19.8 Invest in livestock sector infrastructure 19.1 Address conflict: reinforce customary mechanisms & create joint… 22.3 Source: Authors’ notes. FIGURE I.6. PRIORITIZATION OF RISK COPING SOLUTIONS FOR LIVESTOCK Exceptional livestock movement 15.8 Livestock vaccination 20.7 Building water pans 18.8 Crop residues 18.5 Supplementary feed, emergency stores 15.8 Buying hay 14.5 Source: Authors’ notes. 114 Agriculture Global Practice Technical Assistance Paper 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 97887 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org/agriculture Twitter: @WBG_Agriculture