UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Strengthening food security in rural Ethiopia Cochrane, Logan 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_may_cochrane_logan.pdf [ 30.52MB ]
Metadata
JSON: 24-1.0343434.json
JSON-LD: 24-1.0343434-ld.json
RDF/XML (Pretty): 24-1.0343434-rdf.xml
RDF/JSON: 24-1.0343434-rdf.json
Turtle: 24-1.0343434-turtle.txt
N-Triples: 24-1.0343434-rdf-ntriples.txt
Original Record: 24-1.0343434-source.json
Full Text
24-1.0343434-fulltext.txt
Citation
24-1.0343434.ris

Full Text

  Strengthening Food Security in Rural Ethiopia  by  Logan Cochrane  MA, Staffordshire University, 2013 BA, University of Victoria, 2006   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF   DOCTORATE OF PHILOSOPHY  in  THE COLLEGE OF GRADUATE STUDIES  (Interdisciplinary Studies)  THE UNIVERSITY OF BRITISH COLUMBIA   (Okanagan)  March 2017  © Logan Cochrane, 2017  	   ii EXAMINATION COMMITTEE The undersigned certify that they have read, and recommend to the College of Graduate Studies for acceptance, a thesis entitled:   Strengthening Food Security in Rural Ethiopia   Submitted by   Logan Cochrane   in partial fulfillment of the requirements of the degree of   Doctor of Philosophy    Dr. John Wagner, Associate Professor, UBC (Okanagan) Supervisor Dr. Jon Corbett, Associate Professor, UBC (Okanagan) Supervisor Dr. Kevin Hanna, Associate Professor, UBC (Okanagan) Supervisory Committee Member Dr. Mary Stockdale, Sessional Instructor, UBC (Okanagan) Supervisory Committee Member Dr. Adam Jones, Professor, UBC (Okanagan) University Examiner Dr. Paul Shaffer, Associate Professor, Trent University External Examiner  Date submitted to College of Graduate Studies: March 31, 2017   	   iii ABSTRACT Food insecurity in rural areas of southern Ethiopia is widespread; in recent years over half of all communities in this region have been reliant upon emergency support. However, food security status varies significantly from year to year, as the region experiences variations in rainfall patterns. Research is required to better understand how food security can be strengthened. To do so, this research was driven by three research questions. First, what makes smallholder farmers in southern Ethiopia vulnerable to food insecurity. Second, according to the literature, the adoption of programs and services is low, and thus a community-based assessment was undertaken to understand why. The third question reflected on the methodology – a participatory, co-produced approach, evaluating whether this form of engaged research enabled positive change.  The findings suggest that vulnerability to food insecurity differs by scale. At the community level, access to irrigation infrastructure strengthened food security, and was the most transformative difference between the communities. Within communities, food security distribution was complex and few generalizations can be made. The participatory processes identified that research often makes invisible the purposeful and insightful choices farmers make. When surveyed, they are asked to provide generalizations about input use, crop choice and practices, when in reality each crop, input and practice varies. Similarly, some commonly used measures of vulnerability can also be expressions of security; aggregated averages obfuscate localized inequality. For some programs and services, adoption was found to be quite high – it was only when all services were analyzed as a package that adoption was low. However, not all programs and services served the food insecure households, and the reasons for this are explored in detail. The participatory, co-produced approach enabled unique research questions and metrics and added significant value to the research process, which may also enable long-term positive change to programs and services. Keywords: Food Security, Agriculture, Rural Development, Ethiopia, Co-production, Vulnerability, Adoption, Smallholder agriculture  	   iv PREFACE  This research project was approved the Behavior Research Ethics Board (BREB) at the University of British Columbia (Okanagan campus), Certificate H14-03358. I am solely responsible for the design and conduct of the research project, the analysis of the data and the writing of the dissertation.    	   v Table of Contents Examination Committee ..................................................................................................... ii Abstract .............................................................................................................................. iii Preface ................................................................................................................................ iv List of Tables ..................................................................................................................... vii List of Figures ..................................................................................................................... ix Acronyms ............................................................................................................................ x Glossary of Amharic Terms ............................................................................................... xi Acknowledgments ............................................................................................................. xii Dedication ......................................................................................................................... xiv Chapter 1. Introduction ...................................................................................................... 1 1.1 Ethiopia	  ...............................................................................................................................................................	  6	  1.2 Research Questions	  ........................................................................................................................................	  8	  1.3 Research Area	  ...............................................................................................................................................	  11	  1.4 Thesis Structure	  ............................................................................................................................................	  15	  Chapter 2. Research Area ................................................................................................ 18 2.1 Ethiopia	  ............................................................................................................................................................	  19	  2.2 SNNPR	  ............................................................................................................................................................	  34	  2.3 Wolaita	  .............................................................................................................................................................	  44	  2.4 Damot Gale	  ....................................................................................................................................................	  53	  Chapter 3. Development, Power and Politics ................................................................... 61 3.1 Encountering Food Security	  .....................................................................................................................	  61	  3.2 Power and Politics	  ........................................................................................................................................	  66	  Chapter 4. On Food Security ........................................................................................... 71 4.1 Framing Food Security	  ...............................................................................................................................	  72	  4.2 Measuring Food Security	  ...........................................................................................................................	  79	  4.3 Theoretical Approach in this Research	  ................................................................................................	  84	  Chapter 5. Methods .......................................................................................................... 90 5.1 Stages of Food Security	  ..............................................................................................................................	  94	  5.2 Limitations	  ....................................................................................................................................................	  114	  5.3 Risk Mitigation	  ............................................................................................................................................	  115	  5.4 Ethics Approval	  ...........................................................................................................................................	  117	  5.5. Timing	  ..........................................................................................................................................................	  120	  Chapter 6. Vulnerability to Food Insecurity ................................................................... 121 6.1 Overview	  .......................................................................................................................................................	  121	  	   vi 6.2 Smallholder Farmers Vulnerable to Food Insecurity	  .....................................................................	  139	  6.3 Final Remarks on Vulnerability to Food Insecurity	  .......................................................................	  190	  Chapter 7. Adoption of Extension Services and Programs ............................................ 192 7.1 Overview of Services and Programs	  .....................................................................................................	  192	  7.2 Adoption of Programs and Services	  .....................................................................................................	  209	  7.3 Final Remarks on the Adoption of Extension Services and Programs	  .....................................	  223	  Chapter 8. Impact of Participatory Engagement ........................................................... 225 8.1 Participation & Change	  ............................................................................................................................	  225	  8.2 Theorizing Change	  ....................................................................................................................................	  231	  8.3 Reflections	  .....................................................................................................................................................	  238	  Chapter 9. Conclusion .................................................................................................... 244 9.1 Findings	  ..........................................................................................................................................................	  244	  9.2 Options / Recommendations	  .................................................................................................................	  247	  9.3 Future Research	  ..........................................................................................................................................	  266	  9.4 Final Words	  ..................................................................................................................................................	  272	  References Cited ............................................................................................................. 273 Appendices ...................................................................................................................... 325 Appendix A: Letter of Informed Consent – Individuals (English)	  .....................................................	  325	  Appendix B: Letter of Informed Consent – Individuals (Amharic)	  ...................................................	  327	  Appendix C: Script of Informed Consent – Focus Groups (English)	  ................................................	  329	  Appendix D: Script of Informed Consent – Focus Groups (Amharic)	  .............................................	  331	  Appendix E: Script of Informed Consent – Survey (English)	  ..............................................................	  333	  Appendix F: Script of Informed Consent – Survey (Amharic)	  ............................................................	  335	  Appendix G:  Individual Interview Script	  ...................................................................................................	  337	  Appendix H: In-depth Interview Script	  .......................................................................................................	  339	  Appendix I: Focus Group Script	  ...................................................................................................................	  341	  Appendix J: Preliminary Survey	  .....................................................................................................................	  342	  Appendix K: Community Survey	  ..................................................................................................................	  345	  Appendix L: Community Survey (Amharic)	  ..............................................................................................	  347	  Appendix M: Confidentiality Agreement (English)	  .................................................................................	  349	  Appendix N: Confidentiality Agreement (Amharic)	  ...............................................................................	  350	  Appendix O: Letter of Support from Dr. Yishak Gecho, Wolaita Sodo University	  ....................	  351	  Appendix P: Ethics Approval from the Ethiopian Public Health Institute	  ......................................	  352	  Appendix Q: Scientific and Ethical Review Committee Approval	  .....................................................	  353	  Appendix R: Letter of Support from SNNPR Health Bureau	  ............................................................	  354	  Appendix S: Letter of Support from Woliata Zone Health Department	  .........................................	  355	  Appendix T: Letters of Support from Damot Gale Agricultural District Office to Sub-Districts (Adeaaro, Adea Ofa and Buge)	  .....................................................................................................................	  356	   	   vii LIST OF TABLES Table 6.1 Demographic and Economic Dependency Ratios….……………. 123 Table 6.2 Community Demographics, Government Data….……………… 124 Table 6.3 Spectrum of Food Security Factors, in States……………………. 130 Table 6.4 Household Assets by Community……..…………………………. 148 Table 6.5 Intra-community Asset Differences….…………………………... 150 Table 6.6 Poverty Proxy Measures by Community..……. ………………… 151 Table 6.7 Intra-community Poverty Proxy Measure Differences..….……… 152 Table 6.8 Average Number of Fruit Trees by Community…….…………... 153 Table 6.9 Average Number of Fruit Trees by Food Security Status….…….. 154 Table 6.10 Average Livestock Holdings by Community….………………... 154 Table 6.11 Average Livestock Holdings by Food Security Status...…..…….. 155 Table 6.12 Highest Level of Education in Household (%)……...…………... 159 Table 6.13 Educational Enrolment Rates in Wolaita….…………………… 160 Table 6.14 Migration Levels by Community…….…………………………. 164 Table 6.15 Migration Levels within Communities………..………………… 165 Table 6.16 Time Spent Collecting Water and Firewood (minutes)….……... 168 Table 6.17 Tracking Time by Activity and Gender (Averaged by Community)..…………………………………………………....  169 Table 6.18 Off-farm and Non-farm Activities (% of Community Engaged)………………………………………………………...  173 Table 6.19 Off-farm and Non-farm Activities (% by Food Shortage Status).. 174 Table 6.20 Livelihoods by Community…………………..…………………. 176 Table 6.21 Livelihoods by Food Shortage Status……………..…………….. 176 Table 6.22 Sources of Debt, % of Households by Community.……….…… 178 Table 6.23 Reasons for Borrowing…………….………………………….… 179 Table 6.24 Percent of Community Members Receiving Remittances..…..… 182 Table 6.25 Household Land Size……………….…………………………... 184 Table 6.26 Relative Food Security Change: Compared to 10 Years Previous (%)……………………………………………………..  187 	   viii Table 6.27 Relative Food Security Change: Compared to 25 Years Previous (%).................................................................................  189 Table 7.1 Agricultural Practice Prevalence, % Engaged In….………….…. 211 Table 7.2 Agricultural Extension Support….………………………………. 217    	   ix LIST OF FIGURES Figure 1.1 Famine in Ethiopia…..………………………………………………. 2 Figure 2.1 Productivity per Hectare of Teff and Maize….……………………… 32 Figure 2.2 Administrative Zones of SNNPR…………………………………….. 35 Figure 2.3 Population Density by Districts, SNNPR…………………………….. 36 Figure 2.4 Fertilizer Distribution in Wolaita Zone in the 1970s and 2000s…....... 39 Figure 2.5 Fertilizer and Improved Seed Distribution in Wolaita Zone (2010s).... 39 Figure 2.6 Average Teff and Maize Yields in Wolaita Zone in the 1970s and 2010s, by 50 kg unit per hectare……..…………………………………  40 Figure 2.7 Enset Cultivation and Ethnicities……….……………………………... 42 Figure 2.8 Administrative Districts of Wolaita Zone……………………………... 45 Figure 2.9 Land Allocation (ha) by Crop in Wolaita Zone…….…………………. 49 Figure 2.10 Livestock Population in Wolaita Zone………………………………. 51 Figure 2.11 Administrative sub-Districts of Wolaita Zone……………………….. 54 Figure 5.1 Flow Diagram of Research Activities…………………………………. 97 Figure 6.1 Location of Communities within Damot Gale………………………... 122 Figure 6.2 The Distribution of Stages of Food Security, by Factor, in Adeaaro…. 132 Figure 6.3 The Distribution of Stages of Food Security, by Factor, in Adea Ofa... 133 Figure 6.4 The Distribution of Stages of Food Security, by Factor, in Buge.……. 135 Figure 6.5 Average Rainfall in Wolaita Sodo (2003-2013)……………………….. 140 Figure 6.6 Rainfall Variability in Wolaita Sodo (2003-2013), selected years…….. 141 Figure 6.7 Selected Months, Rainfall 1970-2014……………….………………… 142 Figure 6.8 Seasonal Malnutrition in Wolaita Zone (New Intake of Out-Patient Child Malnutrition Cases) ………...……………………………………  144 Figure 6.9 Impact of Market Access on Child Malnutrition………….…………... 145 Figure 6.10 Irrigation Water Reservoir in Buge………………………………….. 147 Figure 6.11 Highest Average Educational Attainment in Household (%)….…….. 157 Figure 6.12 Percent of Households Farming less than 0.5 ha & Population Growth………………………………………………………………….  183 Figure 7.1 Safety Net (blue) and Reports of Emergency Conditions (red)………... 203 	   x ACRONYMS AIDS Acquired Immune Deficiency Syndrome CSA Central Statistics Agency DDT Dichlorodiphenyltrichloroethane DFID Department for International Development (UK) ECX Ethiopian Commodity Exchange EPHI Ethiopian Public Health Institute EPRDF Ethiopian People’s Revolutionary Democratic Front ETB Ethiopian Birr FAO Food and Agriculture Organization FDI Foreign Direct Investment FEWS NET Famine Early Warning System Network FTC Farmer Training Center GM Genetically Modified GDP Gross Domestic Product GNP Gross National Product GoE Government of Ethiopia GPS Global Positioning System HIV Human Immunodeficiency Virus INGO International Non-governmental Organization MSF Medecins Sans Frontieres (Doctors Without Borders) NGO Non-governmental Organization ODA Official Development Assistance PSNP Productive Safety Net Program SIDA Swedish International Development Agency SNNPR Southern Nations, Nationalities and Peoples’ Region TPLF Tigrayan People’s Liberation Front USAID United States Agency for International Development WHO World Health Organization   	   xi GLOSSARY OF AMHARIC TERMS Dega High-altitude Kebele Sub-District Kola Low altitude Marasha Traditional plow Orominya Oromo language Quantal 50 kg unit Tigrinya Tigray language Wayna dega Mid-altitude Wolaitenya Wolaita language Woreda District    	   xii ACKNOWLEDGMENTS  I am motivated by the people I met throughout Wolaita for this research. I wish to acknowledge them first, and the fact that they believed in me and this research. They need not discuss for hours. But, they did. They would be justified in tossing aside yet another researcher to pass through their communities asking questions. But, they did not. My gratitude is great and my debt deep.  I am would not be where I am without family. My wife has always been there to help. My kids continue to be supportive, even if what it is that I do is only partially clear. My parents have encouraged me from the beginning. My grandparents were always willing to lend a hand and share a story. I wish to thank: the late John Johnson, Colin Welch and Janice Gladish for opening my eyes to new worlds; Dr. Mark Pinkoski for pointing me in the direction of social justice; Dieudonne Amisi Mutambala and Bilombele Asukulu Philbert for showing me the true meaning of commitment and self-sacrifice; Dr. John Wagner and Dr. Jon Corbett for the time, direction and mentorship as my doctoral supervisors and Dr. Mary Stockdale and Dr. Kevin Hanna as my doctoral committee members and teachers. I wish to thank the Government of Canada and the Social Sciences and Humanities Research Council of Canada for awarding me a Vanier Canada Graduate Scholarship and the University of British Columbia for the support it has provided. I stand upon shoulders of giants, without whom there would be no dissertation to write. There are many people I have worked with, or whose works I have read, that have inspired me. Of these I wish to single out one: Dessalegn Rahmato. Ethiopia is not a place where critical social science research is always welcome. Despite the challenges, he has been speaking and writing for five decades. When he speaks and when he writes, it is evident that critical research is more than a career of publications and awards. Dessalegn Rahmato believes Ethiopians deserve better and that Ethiopia can be better. The criticism has a purpose and is given with passion. He has set a path for researchers to follow – a commitment to critical engagement, active participation and rigorous research. 	   xiii Not done comfortably from afar; he moves from the farms to government offices and back to Ethiopia’s foremost independent social science research institute where he works, which he founded two decades ago. It has been honor to know Dessalegn Rahmato, and I look forward to continued collaboration in the future.   	   xiv DEDICATION  To those we routinely neglect and forget.  I have been given opportunities to see the world from your eyes. The injustice you experience demands revolutionary change. I have fallen short in acting upon that and calling for it; neglecting and forgetting. For this, I seek your pardon.    	   1 CHAPTER 1. INTRODUCTION   On October 23rd, 1984, the British journalist Michael Burek reported on the Ethiopian famine. The images altered the way Ethiopia and Ethiopians would be viewed for decades (Gill, 2010). The report began: “Dawn, and as the sun breaks through the piercing chill of night on the plain outside Korem it lights up a biblical famine, now, in the 20th century. This place, say workers here, is the closest thing to hell on earth” (BBC, 1984). When I first started working in Ethiopia in 2006, the messages I sent to friends and family reflected how I was influenced by these representations; I wrote about being amazed with how much rain there was and how green things were. While these images imprinted perceptions around the world, within Ethiopia concerns about food insecurity have deep historical and political roots. After the passing of Prime Minister Meles Zenawi in 2012, Ethiopian television stations proudly proclaimed that Ethiopia was no longer the example given in the Oxford Dictionary for the entry on ‘famine.’ It was a victory, of sorts.  The recent experience of relatively improved food security in Ethiopia has its share of successes. These ought not overshadow the significant challenges that remain, but also not be forgotten or dismissed. As Africa’s second most populous nation, and as a country that has experienced famine events, on average, once per decade for millennia (Pankhurst, 1985), extreme food insecurity is not new; improved prevention and management is. In the last century the population has grown rapidly – eighteen million in 1950, thirty five million in 1980, sixty five million in 2000 and ninety seven million in 2016 – and a series of large scale famine events have been experienced: 1888-1892, 1958, 1966, 1973 and 1984 (de Waal, 1991; Graham, Rashid and Malek, 2012; Sen and Dreze, 1999; UN, 2011; 2015; Wolde Giorgis, 1989). In the most recent three decades, however, the trend shifted, and deaths due to famine began to decline (see Figure 1).    	   2  Figure 1.1 Population Growth and Death due to Famine in Ethiopia  Source: de Waal, 1991; Devereux, 2009; Dorosh and Rashid, 2012; Gill, 2010; Graham, Rashid and Malek, 2012; Sen and Dreze, 1999; Wolde Giorgis, 1989.  Notable examples of this shift are that the droughts of 1999-2000 and 2002-2003, each of which affected millions of people, did not result in significant losses of life. A study of the latter of these two events found no measurable increase of child mortality (de Waal, Taffesse and Carruth, 2006). Improved management of drought and prevention of famine-related death is an important success, yet it was based on unsustainable and costly humanitarian interventions, often made possible by international support. In 2005, Ethiopia launched Africa’s second largest safety net program to support the most food insecure households with predictable, long-term support so that the reliance upon emergency aid could be significantly reduced (Coll-Black et al, 2012). Since its launch, the program has supported nearly eight million people to reduce food insecurity and has enabled farmers to retain assets during challenging years. This was another major success in the effort to strengthen food security.  100,000 250,000 300,000800,000100,000 3001958 1966 1973 1984 1999-­‐‑2000 2002-­‐‑2003Major	  Famine	  Events	  &	  Estimated	  Death	  Tolls	  (Since	  1950)Population	  (hundreds)	   3 Despite widespread coverage of the safety net, there were concerns that it was not enabling households to become food secure, rather that it was stabilizing households from losing assets but still leaving them vulnerable to extreme food insecurity (Maxwell et al, 2013; Rahmato, 2013; Siyoum, 2013). When the rains failed in 2015 in connection with El Niño, these concerns materialized: the government determined that the poorest remain vulnerable to food insecurity despite the safety net program having operated for ten years. Due to the drought, an additional ten million people required emergency food assistance in 2015 and 2016, beyond the almost eight million people already being served by the safety net at the time (OCHA, 2016). That almost one in five Ethiopians required emergency food aid during 2015/16 demonstrates that the transition from emergency responses to sustained and targeted support is ongoing, wherein much more progress is required. Independent studies on the impact of the 2015/16 drought are not yet available, but the loss of life is expected to be lower than that of 2002/03 (Davison, 2015). Food insecurity impacts the lives of people in many different ways, and assessments ought not to be limited to reducing famine-related deaths. Malnutrition and micronutrient deficiencies for infants and children can result in life-long developmental consequences (Gibson, 2012; Martins et al, 2011; Rivera et al, 2003). In this regard, Ethiopia has also made some progress. Stunting due to malnutrition for children under the age of five was reduced from an extremely high rate of fifty seven percent in 2000 to forty four percent in 2011 (UNICEF, 2013). The ‘silent famine’ of chronic malnutrition due to food insecurity remains far too common, and its consequences are severe: one in every eleven children dies before reaching the age of five, thirty five percent of children are moderately underweight, fourteen percent of children are severely underweight, forty four percent of children suffer from moderate stunting and twenty eight percent suffer from severe stunting (CSA, 2011; Evans, 2012; UNICEF, 2013).  Small scale agriculture is the primary livelihood practice for the vast majority of families who experience food insecurity, malnutrition and micronutrient deficiencies. Paradoxically, it is also these smallholder farmers who are the foundation of the national economy and who are the main source of Ethiopia’s exports: agriculture accounts for nearly half of the gross domestic product (Loening, Durecall and Birru, 2009) and 	   4 agricultural products account for fifty five percent of all exports (OEC, 2014). It is smallholder farmers, as opposed to commercial operations, who farm more than ninety percent of all cultivated land – more than sixty percent doing so on less than one hectare (Taffesse, Dorosh and Gemessa, 2012).  Food insecurity is best understood in these rural agricultural contexts as a seasonal experience that reflects a dependence upon rain-fed practices vulnerable to unpredictable rainfall. Each year, during the lean season when saved yields run out, there is a spike in children diagnosed as malnourished (Cochrane and Gecho, 2016). Because the vast majority of smallholder farmers rely entirely upon rainfall for their agricultural livelihoods (CSA, 2009), rainfall variability (too much, too little or at the wrong time) can result in failed yields and cause significant increases in food insecurity, which the failure of two consecutive rains in 2015 demonstrated. As household assets are depleted, and finances limited for investing in future crops, these events can have multi-year impacts (FEWS NET, 2012b).  Progress made in reducing mortality, malnutrition and nutrient deficiencies will not necessarily continue. In fact, the trends suggest that existing programs and services will be insufficient as rainfall becomes more unpredictable due to climate change and land holding size decreases due to inheritance and fragmentation, dropping below levels that are able to meet the basic needs of households (Barker, 2007; Eriksen, 2008; UNEP, 2014; Vervoot et al, 2013; Wegner and Zwart, 2011). In some areas the average landholding size has already dropped below half a hectare, which is what some argue is necessary for basic self-sufficiency (e.g. Rahmato, 2007).  Preventing the negative impacts of food insecurity is one motivation for conducting research that aims to support the strengthening of food security. However, there are also many positive ways to frame the justification or rationale, such as positive impacts on health, strengthening of immune systems, or on education, as families are better able to send their children to school and children are better able to learn. The economic impacts of improved food security means that assets are not eroded in years of insufficient or irregular rainfall, and it means that farmers are more financially secure, as they do not need to take high-interest loans to meet their basic household needs.  	   5 Even more important, I argue, is that strengthening food security is a means to establishing and protecting the right to adequate food, which was recognized in the 1948 Universal Human Rights Declaration, stating: “Everyone has the right to a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control.” (Article 25, Universal Declaration of Human Rights) This right has been re-emphasized in subsequent international agreements, including the International Covenant on Economic, Social and Cultural Rights (Article 11), which defined the right to food as: “The right to adequate food is realized when every man, woman and child, alone or in community with others, has the physical and economic access at all times to adequate food or means for its procurement” (FAO, 2016).  The focus of this research addresses only one component, of many, that require change in order for the right to food to be realized. Aspects of concern beyond this research include international trade regulation, unfair competition due to subsidies and profiteering in agricultural investments affecting commodity price variability, to name just three. While I am optimistic about the chances for positive change, the transformations required of the global marketplace and the restructuring of the global community necessitate radical reformation that appears unlikely in the foreseeable future. One cannot cover all relevant issues within a single research endeavor. I undertook this project recognizing its limitations while also hopeful of its potential to support the strengthening of food security by improving rural programs and services, and in the process supporting smallholder farmers to enhance their livelihoods in a sustainable way.    	   6 1.1 ETHIOPIA  Ethiopia is located in Eastern Africa, within the region known as the Horn of Africa, between the Equator and the Tropic of Cancer. Much of the nation is mountainous; the capital of Addis Ababa is more than 2,300 meters above sea level, one of the most elevated capital cities in the world. In most of the agricultural areas there are two growing seasons associated with the two rainy periods, the meher and belg. The former is the main production season, with harvesting generally lasting from September until February, while the latter runs from March until August. This generalization holds true for much of the highlands, but excludes others. For example, the Afar and Somali regions, in the east of the country, have low elevation, warmer temperatures and much less precipitation, while some of the western parts of the country have tropical rainforest environments.  Ethiopia covers an area as large as France and Spain combined, with great variation in climate, temperature, elevation and terrain (Pankhurst, 1990). The country borders six nations: Djibouti, Eritrea, Kenya, Somalia, South Sudan and Sudan. Following the independence of Eritrea in the early 1990s, Ethiopia became a landlocked nation and much of the exports are transported via ports in Djibouti. Although exceeding a million square kilometers in size, eighty eight percent of the population live in the highlands located between 1500 and 3500 meters above sea level; this area is also home to seventy five percent of all livestock and ninety five percent of total cultivated land (Dalelo and Stellmacher, 2012). As a result of these geographic differences, livelihoods and vegetation vary from region to region. The highland areas are cereal breadbaskets and are thought to be the original locations for plant domestication of teff (Eragrostis teff), nug / Niger seed / blackseed (Guizotia abyssinica) and dagusa / finger-millet (Eleusine corocana) (Pankhurst, 1998). Indigenous crops that are important for national consumption include enset (ensete ventricosum), the stimulant khat (Catha edulis) and coffee, the latter two of which being primary export commodities. Other important cash crops include pulses, oilseeds and cereals. In the last ten years the flower industry has developed into one of the largest 	   7 agricultural exports in the country. Livestock populations in Ethiopia are amongst the highest in Africa, and pastoral livelihoods are primary in the east and south. Due to the diversity of livelihoods, crops and practices, few generalizations can be made about agriculture in Ethiopia. Agricultural practices and crop types are strongly influenced by ecological zones, as the country ranges from less than 500 meters above sea level to more than 3700 meters above. Below 500m there are low levels of rainfall and agriculture of any type is only possible with irrigation; from 500-1500m, sorghum, teff and pulses/oilseeds; from 1500-2300m wheat, teff, maize, sorghum, oilseeds, barley and enset; from 2300-3200m barley, wheat, pulses/oilseeds; from 3200-3700m barley; and, above 3700m no regular crops are grown (Chamberlin and Schmidt, 2012). Although in some parts of the ‘developing’ world, or ‘global south,’ livelihoods are becoming detached from farming (Rigg, 2006), agricultural practices continue to be important for the majority of Ethiopians (Mengistu, 2006).  Some studies of ‘traditional’ agricultural practices suggest they are inefficient or harmful and require change (Coppock, 1993; Dubale et al, 2014; Mintesinot et al, 2004; Temesgen et al, 2007), while others suggest these practices are suited to the contexts within which they are practiced and may be more sustainable than their modern counterparts (Ciampalini et al, 2008; Ciampalini et al, 2012; Lemenih et al, 2004; Mesfin and Obsa, 1994; Nyssen et al, 2000; Tesfahunegn, Tamene and Vlek, 2011; Teshome et al, 1999).  Yet, others find that no simple conclusions can be drawn; traditional practices may be better suited and more productive in some settings, while commercial operations and chemical inputs can be more appropriate in others (Kassie et al, 2010). This study suggests that such generalizations do not align with the experiences of farmers, whereby smallholder farmers selectively, and purposefully, integrate traditional and modern practices into different aspects of their agricultural system. I use the terminology ‘smallholder farmers’ throughout as it references the individuals and their livelihood practices, as opposed to other terms, such as peasants, which primarily focus on the relationship individuals have with the government.  	   8 1.2 RESEARCH QUESTIONS  The research presented in the chapters that follow is founded on three primary research questions, each of which focus on local processes with national implications: (1) what makes households vulnerable to food insecurity, (2) why does the literature indicate that levels of service and program adoption are low, and (3) can a participatory, co-produced research approach facilitate positive change in programs and services? A case study approach with three communities within one district was utilized in order to draw comparisons within and between them. Comparing data within communities allows for a detailed study of the dynamics and relative differences that exist when access to services and infrastructure are reasonably equal, while the cross-community comparisons provide insight about the impact that access to markets, irrigation, transportation, healthcare and education have. Recognizing the challenges of comparability in studies of this nature, the three communities were purposefully selected as sharing livelihood practices, ethnicity, language and agroecology and are located within a single district wherein programs and services are, at least theoretically, the same.  In addition to answering the research questions, this research aims to develop a methodology that enables a fully contextualized understanding of food security. National governments and the United Nations have a wealth of data about food insecurity, providing insight into the extent of the challenges and trends. The methodology developed here, and its findings, complement this data by providing an approach to identify effective and appropriate means to strengthen food security. In developing this methodology, which is based upon work by Dr. Krishna at Duke University, I emphasize the experiences, ideas and priorities of community members in understanding vulnerability to food insecurity. A participatory approach was used to co-create quantitative surveys with smallholder farmers, with the objective of identifying opportunities, strengths and challenges that may not be sufficiently addressed in the existing data. 	   9 In order to enhance and expand existing knowledge on vulnerability to food insecurity the first research question asks: What makes smallholder agricultural households vulnerable to food insecurity? While this question is not new, the process of answering it resulted in unique findings. As is detailed in Chapter 5, typical surveys conducted on food security, and those used in Ethiopia, draw upon data collection tools, household survey questions and metrics based on assumptions about vulnerability that are not embedded within or reflective of the lived experiences of those encountering food insecurity. Using broad metrics, national surveys miss relative differences within and between communities, and its questions often prioritize export crops. As a result, common results make invisible important aspects of food security and therefore the recommendations may not be appropriate, and in particular may not meet the needs of the most vulnerable. Thus, this question was approached using participatory approaches that co-created data collection tools, including determining the most relevant questions and most appropriate metrics. The methodology of which these processes are a part, is called Stages of Food Security, which has been adapted from Krishna’s (2004; 2005; 2010) Stages of Progress methodology that focused upon poverty. The second research question seeks to understand program and service adoption. In doing so, it builds upon the assessment of how smallholder farmers define food insecurity and assess their own vulnerabilities. In the process of identifying the ways in which vulnerabilities manifest, community members reflected on their experiences with the programs and services. Specifically, community members identified how their diverse practices are not captured by typical surveys. For example, they are asked to generalize about fertilizer or improved seed use, but in actuality they make choices on a crop by crop basis. Analyzing each crop and input allowed for a better assessment of the programs and services offered to smallholder farmers. The results identify key areas where programs do, in fact, meet farmer needs, as well as components that are not serving them. The co-analysis of survey results indicates how individuals identify opportunities and navigate barriers, highlighting the need for programs and services that are much more localized, adaptive and responsive.  	   10 The first and the second research questions allow for an evaluation of the effectiveness and appropriateness of rural agricultural extension programs and services. Due to the politicization of direct conversations that provide feedback on governmental programs, the evaluation of these programs and services draws upon the qualitative and quantitative data used to answer questions one and two. For some governmental programs and services, household survey data was collected on adoption (e.g. fertilizer, pesticide and improved seed), in other instances on coverage (e.g. agricultural extension worker training provision), while others draw on proxy measures (e.g. metrics related to poverty). The first two research questions enabled community members and I to co-analyze the results to offer recommendations as to how programs and services could be improved, as a means to strengthen food security. While this research is based on specific case studies, and the details cannot be generalized broadly, the insights derived on why and how programs are not meeting the needs of smallholder farmers, and the process used to arrive at these conclusions, can be widely applied. The identification of structural, institutional, design-related and implementation-based barriers reflect broader systems, and are applicable beyond the communities and district studied.  The third research question assesses the impact of the research process, and asks specifically if a co-produced approach that uses participatory processes can facilitate the strengthening of food security through engagement in collective problem definition, data collection, analysis and action. In answering the third research question I analyze theories of change wherein participatory action is foundational to change and discuss what can be learned about how change happens, reflecting on the entire research process undertaken in this study. Based upon the results and experiences, I explore other theories of change and what insight they offer into understanding the process of change within the context of rural agricultural programs and services.  These research questions were developed in 2014, before the recent wave of large-scale protests began. I first started working in Ethiopia in 2006, which followed on the heels of a controversial election and the imprisonment of large numbers of Ethiopians driven by ethnic divisions. As I gained experience in different regional states of Ethiopia I became increasingly aware of tensions that are often not spoken about openly. The description 	   11 made of Ethiopia included: increasing ethnic inequality, institutionalized, state-driven patronage, centralized political control with strong disincentives for citizens to engage, arbitrary arrests and severe restrictions on freedom of the press and freedom of speech. In many ways, the situation resembled that of Rwanda before the genocide (Uvin, 1999). Similar to other contexts, the international community, non-governmental organization and researchers have largely been unwilling or unable to openly and directly address the increase of political crises and political discontent (Autessere, 2010; Starn, 1991; Uvin, 1999). During the last three years hundreds of large-scale protests have occurred throughout rural and urban Ethiopia, and the third research question is timely and relevant to issues well beyond the challenges of rural agricultural programs and services.  1.3 RESEARCH AREA  The research presented in this thesis is centered upon fieldwork that was conducted in the Southern Nations, Nationalities and Peoples’ Region (SNNPR) of southern Ethiopia. Twenty percent of Ethiopia’s ninety seven million people live in this region, making it amongst the most populous regions (along with Oromia and Amhara). The reason SNNPR was selected was the convergence of two unique factors, both of which play a significant role in food insecurity. The first factor is that SNNPR has the highest rural population density in the country (CSA, 2007), and in many ways what is happening in parts of SNNPR now may indicate what will happen elsewhere as population continues to increase.  The second factor that makes SNNPR unique, or at least the central part of the region, is its rainfall situation. Central SNNPR is neither rain secure, as the highlands tend to be, nor does it consistently lack rainfall, as is common in the arid Somali and Afar regions. Rainfall is particularly important in SNNPR as the vast majority of smallholder farmers practice rain-fed agriculture. Year-to-year variability provides unique insight into the dynamics of inequality, population, land size, seasonality, rainfall, climate change and the 	   12 impact of interventions designed to strengthen food security. In years when rainfall is too little, too late or at the wrong time the impact can be devastating. For example, consecutive seasons of low agricultural production resulted in emergency situations in 2011 and 2012 (FEWS NET, 2012b) and in the latter year fifty five percent of the districts in SNNPR were chronically food insecure (FEWS NET, 2012b).1 Difficult years such as 2012 result in multifaceted, negative impacts, that include the loss of assets and significant increases of child malnutrition. In other years, such as 2013 and 2014, the region experienced relatively higher levels of food security, and relatively low levels of child malnutrition (Cochrane and Gecho, 2016). Even in years when harvests are strong and food security increases for SNNPR, a significant minority remain chronically food insecure. Thus, unique environmental and demographic factors make SNNPR a particularly challenging context wherein more research is needed. Within SNNPR, research was conducted in three sub-districts (kebeles), within the Wolaita Zone. These areas were selected due to high population density within SNNPR, high levels of chronic food insecurity and their respective differences within a similar agroecological setting: one rural and remote, another rural near to a market town, and the third rural with irrigation infrastructure. The root-crop based agricultural system, described in Chapter 2, is representative of the majority of the Wolaita Zone. The exceptions include parts of Humbo and Diguna Fango districts, which are lowland areas and where root crops are less common. These two districts are not included in this study, and their unique agricultural contexts beyond its scope.  The selection of these districts for study was in response to the complex and overlapping layers of vulnerability experienced within them (Husmann, 2016; Rahmato, 2007). The vulnerabilities experienced are an expression of regional, national and local contexts, opportunities, limitations and barriers. One of Ethiopia’s foremost scholars who has studied issues related to food security, and conducted research in Wolaita, explained that positive change “must first be based on a clear and in-depth understanding of the lives and livelihoods of the people and the farming systems that they have evolved over the 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  1 The use of chronic food insecurity in this dissertation refers to the dimension of time. As opposed to a short-term, transitory or emergency of insufficient food, chronic food insecurity refers to a long-term or persistent inability to meet minimum food requirements.  	   13 generations” (Rahmato, 2007: 34). Despite the severity of food insecurity within Wolaita Zona, Rahmato (2007: 23) writes that “no attempt has been made to estimate the magnitude of poverty from the point of view of food insecurity.” This research attempts to address part of this knowledge gap by presenting detailed data about food security using in-depth, contextualized research. In addition to the need for research to inform policy making, programs and services specific to this region, the institutions and systems that operate within SNNPR are common in nearly all parts of Ethiopia and thus this research offers broad insight into the nature of food security and the ways in which programs and services impact individuals and communities throughout the nation. There are findings and recommendations that are specific to Wolaita, and even the districts within Wolaita wherein the study took place, and these ought not be overgeneralized. At the same time, the research provides new knowledge on broader questions within the food security discourse that are applicable for audiences in Ethiopia, East Africa and beyond. The methodology is applicable globally, and provides new avenues for assessing and understanding food security. The three primary research questions, while specific to Ethiopia, take place within a global context wherein food insecurity continues to be one of the world’s greatest challenges. The United Nations’ World Food Program estimates that more than 870 million people are chronically hungry (WFP, 2014). The burden of this problem is greater in developing countries, and particularly in sub-Saharan Africa, which has the highest prevalence of people facing chronic hunger. Furthermore, the challenges of food insecurity are experienced disproportionately by smallholder farmers, who constitute the majority of the global population experiencing poverty and food insecurity (Gibson, 2012). Despite gains made during the 1980s and 1990s, in the most recent decade the number of people who are chronically hungry has steadily risen (WFP, 2014). The smallholder farmers within Wolaita Zone, and Ethiopia more broadly, exist within a world wherein over 400 million of 525 million (76%) farms are managed by smallholder farmers on less than two hectares of land (Gibson, 2012). These smallholder farmers, moreover, occupy 	   14 about a third of all arable land globally and “these smallholders effectively contribute about half of the world’s total food supply” (Gibson, 2012: 316). Yet, it is these same people who experience the greatest vulnerability to food insecurity. Strengthening food security in Wolaita has the potential to offer pathways for positive change for smallholder farmers around the world. Globally, the rise of people who are food insecure has not been a result of having insufficient food; rates of agricultural production have risen faster than population growth (FAO, 2012a). The distribution, however, is unequal. Agroecological settings affect the production potential of individuals and countries, such as rainfall, soil types, land size and water availability. For example, while much of North and South America, Europe and South East Asia have per capita food production above 8,000 kcal/day, per capita production in much of Africa, the Middle East and Asia is less than half or a quarter of that level (FAO, 2012a). Food insecurity is not solely related to geography, however, there are limitations and opportunities related to different supportive systems and infrastructure. At the individual and household levels, barriers to obtaining food primarily revolve around localized challenges, such as accessibility, production, seasonality, poverty and inequality. Regions that currently have lower per capita food production, such as East and Central Africa, the Middle East and North Africa and parts of Asia (specifically from Iran to Bangladesh), are also ones that have high population growth rates, and where much of the projected global population growth is projected to occur (FAO, 2012a; UN, 2011). Additionally, these regions face the greatest overall vulnerability to the negative impacts of climate change (CGD, 2014). Compounding issues of regional availability are the ways in which existing resources are utilized, such as shifting dietary composition to ones that include more meat and dairy products, as well as shifting land use to grow crops for non-food use, such as biofuels (Brown, 2012; Cotula, 2013; FAO, 2012a). High levels of chronic hunger, and expectations that the situation may worsen (Hallegatte et al, 2016), have resulted in a great deal more focus on food security. This study highlights how the global trends of food insecurity may worsen, by exploring case studies in Ethiopia. In doing so, it emphasizes how definitions and metrics greatly affect the 	   15 understanding of food security, and thus the ability to appropriately and effectively design programs that reduce vulnerabilities and strengthen food security. 	  1.4 THESIS STRUCTURE  This opening chapter presents a broad overview of the research questions, area of study and justification for its importance. The second chapter of this thesis delves into the details of the research area. The sections progressively narrow down in their focus: starting with the country, then the regional state, the zone and then the district. In presenting the research area, this chapter highlights the political, environmental, historical, socio-cultural and livelihood contexts. Varying degrees of emphasis are given to each thematic area in the sub-sections by highlighting unique features that contribute to food security. Although this chapter is detailed, it is selective. For example, while relevant, the governmental and constitutional structure (formal and informal) are presented in brief, as the details require entire texts (e.g. Abebe, 2016; Kefale, 2014). Other specific details, such as the number of livestock holdings in the area of study, are presented as they offer insight into the livelihood options, opportunities and limitations. In my selectivity of data to present, the chapter aims to provide relevant information for a comprehensive contextualization of the research area.  Chapters 3 and 4 outline the theoretical framing and positionality on development and food security respectively. In analyzing the concept of development, I pay explicit attention to the role of power and politics, and specifically how development activities, including agricultural programs and services, can become highly politicized. While Chapter 3 outlines how development is politicized, this idea runs throughout the dissertation, and is first presented in Chapter 2 in discussing the quality of data and the existence of multiple, irreconcilable data sets. The chapter on food security specifies how the framing and measurement of food security significantly affects the understanding of it, responses to it, and broadly situates my own approaches. 	   16 The methodology developed for, and used within, this research is outlined in Chapter 5. As the Stages of Food Security methodology is an adaptation of an existing model that offers new approaches and foci, the chapter provides a detailed outline of the processes involved. It is hoped that this chapter, along with subsequent publications devoted specifically to the methodology (Cochrane, 2017a), will enable other researchers to utilize and further adapt this research approach. In addition to processes, this chapter presents the limitations of the methodology and this specific implementation of it. Although less commonly presented, I included information about the process of obtaining national ethics approval from Ethiopian authorities as well as situating the timing of the research within broader political and social events. The explanation of how ethics approval was obtained from Ethiopian authorities and the research timing offer insight into my positionality as a researcher and the research process. The ethics approval and letters of support from national, regional, zonal and district authorities are included as appendices (See Appendices P, Q, R, S and T).  Chapter 6 draws upon qualitative and quantitative data to analyze what makes smallholder farmers vulnerable to food insecurity. For the purposes of readability, the data was divided into themes (seasonality, rainfall, poverty, location, education, inequality, diversity, population and land size, change over time) but in reality these themes are interconnected. To the extent possible, I have attempted to reinforce the complexity of smallholder farmer realities, rather than present them as compartmentalized, technical and simplified factors. Chapter 7 draws upon the findings of vulnerability to food insecurity to analyze the programs and services offered to smallholder farmers and the reasons why the literature suggests that adoption for some programs and services is low and why discontinuation is relatively high. The methodology developed for this research is rooted in participation and co-production. Chapter 8 explores the assumptions embedded within the theory of change that informs approaches driven by participatory approaches and co-production, and examines to what extent this process enabled positive change. In doing so, I present alternative theories of change and reflect on how these different ideas can facilitate an understanding of why change did or did not occur. The chapter concludes with the 	   17 limitations of planned theories of change and the opportunities offered by emerging approaches that emphasize learning, adaptive management and complexity-based approaches. The final chapter of this dissertation outlines conclusions, primarily in relation to the three research questions posed at the outset. It offers recommendations as well as highlights areas for future research. A large number of appendices have been included as a means to enhance the transparency of the research process and to support researchers who wish to undertake research using the Stages of Food Security methodology. This includes informed consent forms, scripts, surveys and confidentiality agreements used in this research. Of note, however, is that with the exception of one household survey, these materials were written in advance of the research, as required by my host university ethics approval process. As outlined in the preamble of a number of these forms, a more iterative approach was used in the actual implementation of the research.  	   18 CHAPTER 2. RESEARCH AREA   Ethiopia is home to a diversity of ethnic, linguistic and religious groups. It also has diverse agroecological settings wherein myriad livelihoods are practiced, adapted and changed. This research took place within a single agroecological area, which is where a single ethnic and linguistic group live and where two sects of Christianity predominate, as detailed in Section 2.3. This area, however, exists within a nation and a regional state that influence the available options, opportunities and limitations, which are respectively introduced in Sections 2.1 and 2.2. Entire volumes have been devoted to Ethiopia’s general history (Pankhurst, 1998), its agricultural history (McCann, 1995), its social history (Pankhurst, 1990) and its peripheral areas (Pankhurst, 1997). Books have also been written on specific food insecurity events in Ethiopian history (de Waal, 1991) as well as on contemporary food insecurity challenges (Rahmato, Pankhurst and van Uffelen, 2013). This chapter summarizes and situates the components that are relevant to the research and that have direct impacts on the research site. Determining what is excluded, and what amount of background detail is required, is a challenging task; invariably some readers will feel important aspects are underrepresented, while others may find it overly detailed. I have attempted to find a balance that does not present a comprehensive background on Ethiopia, but a sufficient amount of context to provide a narrative around the conditions, policies, programs and services that exist within the research site.      	   19 2.1 ETHIOPIA  History The borders of Ethiopia are a relatively modern phenomenon. The empires of ancient history were largely based in the northern and highland areas, from the D’mt in the 10th century BCE to Solmonic Dynasty of the 13th century. It was not until Tewodros II in the 1850s, and Menelik II in the 1880s that ‘unification’ and expansion processes resulted in the forming of what would become the nation of Ethiopia. Of note, however, is that those who were conquered in this process do not view it as unification, but colonization. Since many regions are relatively recent additions, and its inhabitants faced marginalization once incorporated, tensions between loyalty to the nation and to one’s ethnic group continue to be one of the most challenging domestic issues. For example, during the run-up to the election in 2015 I was in Benishangul Gumuz regional state, wherein politicians from the majority ethnic group of that regional state promised that if elected they would kick out the “red” people, meaning the Amhara and Tigray people, and take back the land that had been stolen from them. The rate and scale of globalization that has emerged in recent decades is unprecedented. Yet, the international exchange of goods and ideas has long been practiced, and the lands that would become Ethiopia have always played an important role. Trade in ancient times occurred with the Pharaohs of Egypt, to areas in present day Sudan, to and from the Middle East and India (Pankhurst, 1998). In international trade markets, Ethiopia was known as a source of gold, ivory, myrrh and slaves, for which it would trade weaponry and luxury goods for the elite, such as Mediterranean wines (Pankhurst, 1998). International interactions were not limited to trade, however. The Aksumite Empire conquered southwestern Arabia and Sudan between the 3rd and 6th century (Pankhurst, 1998), making it one of the most important political empires of the world, along with Rome, Persia and China (Munro-Hay, 2002). It also embraced Christianity as a state religion in the 4th century, making it one of the first Christian nations (Sulas, Medella and French, 2009). The Aksumite Empire was the only African empire to mint its own 	   20 currency, which was valued on par with Roman and Byzantium coinage. The empire fell in the 10th century and was followed by the Solomonic line of rule, the leaders of which claimed linage from King Solomon of Israel and the Queen of Sheba, whose meeting is described in a Biblical account but whose supposed progeny are not (1 Kings 10:1-13). The rise and fall of empires is beyond the scope of this chapter. However, the historical experience of food security warrants some discussion. Agriculture played an important role in the early empires, with historical records suggesting that, at least for the land-holding class, significant relative wealth could be obtained from yields and livestock (D’Andrea et al, 2008; Munro-Hay, 1991; Pankhurst, 1990). At the same time, however, drought and famine have been recorded for at least a thousand years; between the 15th and 19th centuries, for which greater data is available, historian Richard Pankhurst suggests that a famine occurred, on average, once per decade (Pankhurst, 1985). A major famine occurred between 1888 and 1892, known as the ‘evil days,’ wherein a third of the population may have died (Sen and Dreze, 1999). Famine occurred in Tigray in 1958, and in Wollo in 1966, respectively resulting in the loss of an estimated 100,000 and 250,000 people (Graham, Rashid and Malek, 2012). Famine occurred again in Wollo in 1973 causing the death of 40,000 (Gill, 2010) to 300,000 (Graham, Rashid and Malek, 2012) people. This was one of the first famine events to be shown on international media. The 1984 famine resulted in the death of between 400,000 (de Waal, 1991) and 1.2 million (Wolde Giorgis, 1989). Politics played heavily into the death toll of the latter of these famine events, as the government sought to contain people from supporting and/or joining rebel movements, as it fought against these groups in the north and east, and then resettled massive numbers of people in a large-scale villagization scheme. Alex de Waal (1991) suggests that 50,000 died due to the resettlement process itself, while Doctors Without Borders (Medecins Sans Frontieres, MSF) suggest the figure was closer to 100,000 people (Gill, 2010). The modern Ethiopian state took its form during the reign of Menelik II, who ruled from 1889-1913. Under his lead, the nation expanded and conquered much of southern and eastern Ethiopia, developed currency and postage stamps, introduced piped water, established a railway and telegraph line and founded modern hospitals and schools 	   21 (Pankhurst, 1998). Menelik II defeated an Italian attempt of colonization in 1896. Menelik II was followed by Empress Zewditu in 1916 and then Emperor Haile in 1930, the latter of whom experienced a return of Italian forces, who occupied Ethiopia from 1936 to 1941. The Italian forces were defeated with allied support as a part of WWII.  The Solomonic dynasty came to an end in 1974, when it was overthrown by the Marxist-inspired Derg government. A coalition of rebel groups, largely led by the Tigrayan People’s Liberation Front (TPLF), took power in 1991. Meles Zenawi, chairman of the TPLF and the Ethiopian People’s Revolutionary Democratic Front (EPRDF), was the transition President of Ethiopia following the fall of the Derg, and Prime Minister of Ethiopia from 1995 to 2012. After his unexpected death, the constitutionally mandated successor, Hailemariam Desalegn, took over. This Prime Minister has continued to lead the party, which has won every parliamentary election since coming to power (along with its allied parties), including every single seat in the 2015 election (NEBE, 2015).   Politics & Policymaking The Federal Democratic Republic of Ethiopia is composed of the federal government and regional states. The lowest level of government is the sub-district, kebele, followed by the district, woreda, and then the zone, which are administrative levels under the regional state. Article 50 of the constitution outlines that the regional states are “responsible” to residents of that state, and that lower levels of government are granted “adequate power” to make decisions accordingly (GoE, 2014). The constitution gives the federal government power to “formulate and implement the country’s policies, strategies and plans in respect of overall economic, social and development matters” as well as to “enact laws for the utilization and conservation of land and other natural resources” (Article 51; GoE, 2014). While the federal government “shall formulate and implement the country’s policies”, the regional states have jurisdiction in some areas (those not “given expressly to the Federal Government alone”, Article 52 of the Constitution). Constitutionally, therefore, the regional states potentially have the power to create and implement policy. In practice, 	   22 however, the federal government continues to centralize power, even through its decentralization initiatives (Chinigo, 2013; Mezgebe, 2015).  Regional states have exercised their power through creating and implementing some development policy, such as the first pilot of the land certification scheme and unique regulations of land inheritance. Both of these processes, however, operate within the bounds of federal policies regarding land tenure. As such, the federal government remains the primary creator of policy, and delegates jurisdiction and responsibility, allowing regional states to tailor some of the details for their particular contexts.  When the EPRDF came to power in 1991 they “understood the role that famine had played in its victory” (Graham, Rashid and Malek, 2012: 263). The members of the new government had lived through, and fought amidst, famine. Its members had also witnessed two governments weakened, if not toppled, as a result of their lack of action on addressing emergency needs and ensuring food security. When in power, the EPRDF set about to support the majority rural population, placing them at the center of their major policy documents, including: Agricultural Development-Led Industrialization (1992), the National Policy on Disaster Prevention and Management (1993), the Sustainable Poverty Reduction Strategy (2002) the Plan for Accelerated, Sustained Development to End Poverty (2006), and the Growth and Transformation Plan (2010). The government also upheld, and created, a number of bodies to support this work, such as the Agricultural Input Supply Enterprise / Agricultural Inputs Supply Corporation, Emergency Food Security Reserve Administration, Ethiopian Grain Trading Enterprise (1992), Productive Safety Net Program (2005), the Ethiopian Commodity Exchange (2008), Disaster Risk Management and Food Security Service (2008), Household Assets Building Program (2009), and the Ethiopian Agricultural Transformation Agency (2010). Ethiopia has limited political freedom; in both the 2005 and 2010 elections key opposition party leaders were imprisoned (Abbink, 2006; Tronvoll, 2010). In the halls of government, there is a softly spoken debate that reflects the origin of democracy in the Athenian sphere. The ruling elite have, in practice, taken the Platonic approach of opting for rule by the self-determined wisest and best. Some parts of the population tacitly approve of these choices, believing the alternative options are worse, pointing to the 	   23 situation in neighboring Somalia, Eritrea and South Sudan as examples of why the stability of a non-democratic developmental state is better than the perceived alternative of a failed state. Government narratives reinforce this dichotomy by stifling political opposition, freedom of speech and freedom of the press, thus limiting the space for viable alternatives. However, there is discontent with the status quo, which is often driven by ethnic politics, but increasingly by a general discontent regarding the lack of inclusive democratic processes. An example of this is the activism of the sizable Muslim population who have, practically for the first time in this historically Christian state, attempted to utilize the democratic process to have their concerns heard and their rights upheld, to address their lack of representation and marginalization, as well as to challenge the arrest of their religious leaders without charge (Feyissa and Lawrence, 2014). The rise of mass collective action in 2016 resulted in the government instating a state of emergency in October, scheduled to last for six months, as a means to maintain power and control. Unless the democratization process moves beyond rhetoric, the fate of the current political elite may not be determined by the ballot box, but at the hands of the people.  Population The last national census took place in 2007, and since that time population data has largely been based on projections, resulting in significant discrepancies. For example, the Government of Ethiopia projected that the population was 87.9 million (CSA, 2013) while the World Bank projected 96.9 million (World Bank, 2016). This discrepancy amounts to a larger population than neighboring Djibouti and Eritrea combined, and almost as much as the entire population of neighboring Somalia or South Sudan. While this specific point is not an essential one with regard to the research questions, it highlights the problematic nature of data in Ethiopia; in many instances the figures provided by federal, regional, zonal, woreda and kebele level administrations are best viewed as approximations. Based upon available data it is clear that the national population has grown steadily over the last century, a trend that is expected to continue until 2050. In 1960, the national 	   24 population was estimated to be twenty two million, which had more than doubled by 1990, to forty eight million, and again doubled by 2014, rising to ninety seven million (World Bank, 2016). The United Nations projects that the population will double again by 2055 at 200 million, and will stabilize by the end of the century at around 240 million (UN, 2015). Based on global population growth, by 2050 Ethiopia will be amongst the top ten most populated countries in the world and will remain so throughout the rest of the century (UN, 2015). At present, the urban population is low (16%), and urbanization rates are relatively low compared to global and regional ones, yet these urbanization rates are affected by definitions and some suggest the urbanization rate may be as much as double the government listed rate of four percent (Chamberlin and Schmidt, 2012). In either case, the predominantly rural population who are engaged in smallholder agricultural livelihoods will encounter increasing pressure on land distribution as the population continues to grow.   Religion and Ethnicity Ethiopia is home to great religious and ethnic diversity. There are an estimated eighty ethnic groups, most of which are officially recognized by the Government of Ethiopia and recorded in national census data. The two largest ethnic groups are the Oromo (35%) and Amhara (27%), followed by Somali (6%) and Tigray (6%) (CSA, 2007). Ethnic division is often aligned to linguistic groupings, with multilingualism being common and linking smaller linguistic groups with larger ones. For example, an ethnic Harari living in the city of Harar will speak Harari (called Gey Sinan by its own speakers, a name rarely known by non-speakers) at home and with fellows of their ethnic group, but they will also speak the national language of Amharic, have a basic knowledge of English from the public school system, and the elder generation able to read and write Arabic (this is less common amongst the youth today). In Wolaita the linga franca is Wolaitenya, the local language of the ethnic group. Outside of the towns, there are few speakers of Amharic. More than the national language, people in Wolaita speak Oromiffa, an important regional language.  	   25 Religious affiliation is typically divided into three groups: Christianity, Islam and Traditional Faiths. However, the divisions within these groups are also significant, such as those between Ethiopian Orthodox, Protestant and Catholic. Residents in rural areas will often avoid intermarriage amongst these different Christian sects. In addition, syncretism is common. For example, someone classified as a Muslim may, in numerous aspects of their life, prioritize rites of traditional faiths over those of Islam. Syncretism between Christianity or Islam with traditional faiths is common in Ethiopia and is manifested in diverse ways (e.g. Braukamper, 1992; Vecchiato, 1993).   Historically Ethiopia was known as a Christian state, and governmental statistics continue to show that Ethiopian Orthodox Christians are the majority of the population. The statistics on religious demographics, however, are contested. Government data is the only available nationally-representative data, which suggests the percent of Muslims in 2000 was 33 percent (CSA, 2000) and 34 percent in 2007 (CSA, 2007). Although the source is not cited, a report commissioned by the United Nations in 2006 suggested that the Muslim population was 45 percent (Barnes, 2006), which was a figure also listed by the U.S. State Department (2007), making it the largest religious group. However, in recent reports the U.S. State Department lists the Government of Ethiopia data (e.g. U.S. State Department, 2014). While this figure does not directly affect the study, it highlights the politics of data. The intersections of religion, ethnicity and language, and their impact on daily life cannot be understated. Regional states are, largely, drawn upon ethnic boundaries, which are reinforced through regional language policy. Local languages are commonly the primary language of instruction, after which the language of instruction is English. The result is that many children do not become proficient in the national language. As a consequence, in many parts of the country, the national language of Amharic is not commonly spoken. In the research area of this study, for example, Amharic was not written or spoken by the majority of rural residents.  The ethno-linguistic grouping to which one is affiliated influences choices of day-to-day life, such as who to marry (and not marry) and where one choses to live (or not live), as well as components of life that might not often be associated with ethnicity or religion, 	   26 such as choice of bank and where one choses to shop (or not shop). These choices are purposeful and made at an individual level. Some of these choices are displays of power, such as when federal government personnel speak Tigrinya to each other in government offices (not the national language) and similarly when the regional Oromia government personnel only speak Oromiffa (in Amharic: Orominya), even when Amharic is known and the service-seeker is not a speaker of Oromiffa. Religion plays a significant, and divisive, role in Ethiopian society. In many cases, religion is perceived to be a part of ethnicity. While this is not always the case, Ethiopians make, and reinforce, relationships of this kind: Amhara are Orthodox, Somalis are Muslim, Wolaitans are Protestant, Hararis are Muslim, Gumuz practice a traditional faith, Agaw are Orthodox, Afaris are Muslim, and so forth. Without doubt, there are many exceptions to these generalizations; however they tend to persist since, by and large, these statements do represent the majority and reflect historical notions of identity. In many instances, religion and ethnicity overlap, and their impact on personal choices and societal engagement reflect this, thus establishing a religious-linguistic-ethnic nexus.   Developmental Context & Challenges Over the last decade, Ethiopia has experienced rapid economic growth, ranging between 8.6 percent and 12.6 percent annual gross domestic product (GDP) growth (World Bank, 2016). Despite its economic growth, however, Ethiopia has one of world’s lowest gross national income’s per capita, at US$550 (World Bank, 2016; using 2014 figures). During this period of growth, school enrolment has risen rapidly, reaching ninety five percent. But, while children throughout the country are gaining access to education, the national literacy rate is forty seven percent, which is significantly lower for women at thirty eight percent, and also disproportionately lower in rural areas (CSA, 2012). Life expectancy has increased to sixty four years, higher than the average for sub-Saharan Africa and the average for low income countries (World Bank, 2016). Poverty has declined from 45.5 percent in 1995 to 29.6 percent, an achievement that has occurred 	   27 amidst significant population growth (World Bank, 2016). However, it ought to be noted that the declines in poverty that are often touted by the government have been challenged as being inaccurate, or at best as only part of the story. Dereveux and Sharp (2006) find problems with the government methodology, cite studies showing the opposite trend, and highlight the neglect of seasonality, which as shown in Chapter 6 has a significant impact on the results. Research by Devereux and Sharp (2006) themselves identify high levels of poverty and that these numbers are increasing over time, not decreasing. Sundaram (2016) has shown that there are methodological problems with many of the assessments suggesting rapid declines of poverty around the world. Nonetheless, the government data, which is promoted by international agencies such as the World Bank and USAID, indicate significant reductions of the percent of people living in poverty. This study makes clear seasonality’s key role in vulnerability to food insecurity, and yet, as Chambers has stated the topic remains “grossly neglected” (2012a: xv). Hirvonen, Taffesse and Worku (2015: 2) state that despite the recognition of intra-annual shifts in health and nutrition “seasonality generally has received less research attention and has been largely neglected in the policy arenas.” The findings outlined in Chapter 6 show the impact of seasonality on child malnutrition diagnoses that result from insufficient food quantity (diagnosis requires significant wasting), but less is known about nutrient fluctuations, in other words the seasonality regarding the quality of diets. Evidence indicates that in rural Ethiopia there are seasonal drops in average per capita caloric intake (10%) and similar declines in average diet diversity (7%) (Hirvonen, Taffesse and Worku, 2015). The study of Hirvonen, Taffesse and Worku (2015) is important, but has a number of limitations. In Chapter 6, I show that the impact of seasonal changes varies significantly from year to year, and their study only draws upon one year (2010-2011). Additionally, the averaging of all rural households makes invisible the inequalities that exist between households, while quintile-based assessments of changes to diet quantity and quality would have been much more beneficial in understanding the impact. Further research is needed to better understand these trends using the Household, Consumption and Expenditure data, collected by the Ethiopian Central Statistical Agency. Based on the available data, and drawing upon research from other countries (Devereux, Sabates-Wheeler and Longhurst, 2012; Devereux, Vaitla and Hauenstein-Swan, 2008; Gill, 1991; 	   28 Sahn, 1989), it is evident that an understanding of seasonality is crucial if food insecurity is to be reduced, and must be taken into account in the design and implementation of programs and services. Health coverage has risen rapidly since 2006. For example, although HIV and AIDS is complex for health systems – from testing to calibrating diagnostic machines and adjusting treatment regimens – there has not been a single case of treatment interruption, and coverage has reached eighty percent, rising from less than ten percent in 2006 (Taddesse, Jamieson and Cochrane, 2015). Yet significant challenges remain. Although progress has been made in expanding healthcare coverage and providing services, one in every eleven children die before the age of five, and well over half of all births are not attended by a care provider; three quarters of children (12-23 months) are not vaccinated, only four percent of infants (6-59 months) are fed according to the WHO Infant and Young Child Feeding practices; almost half of all children (6-23 months) are anemic; and, maternal deaths account for thirty percent of all deaths of women aged 15-49 (CSA, 2011). As mentioned in Chapter 1, malnutrition levels are extremely high in Ethiopia: thirty five percent of children are moderately underweight, fourteen percent are severely underweight, fifty one percent suffer from moderate stunting and twenty eight percent from severe stunting (Evans, 2012). These health impacts are compounded by limited access to clean water and sanitation services. National level statistics indicate that almost half of all households do not have access to an improved source of drinking water, only eight percent have an improved toilet facility and three-quarters of homes lack electricity (CSA, 2011). Ethiopia has made impressive progress in reducing the loss of life due to famine. Whereas droughts and famines in the past resulted in tens or hundreds of thousands of lost lives, the droughts of 1999-2000 and 2002-2003, despite affecting millions of people, had no measurable increase in child mortality (de Waal, Taffesse and Carruth, 2006). Yet food insecurity remains chronic and the nation relies upon aid and trade to manage food deficits; between 1985 and 2001 food aid contributed ten percent or more of total food demand (Dalelo and Stellmacher, 2012). The Productive Safety Net Program (PSNP), started in 2005, is one of Africa’s largest efforts to provide multi-year transfers to the most 	   29 food insecure households to protect the loss of assets. While it is undoubtedly having a positive impact on income and health, there are also political implications to the implementation of this program (Cochrane and Tamiru, 2016; see Chapter 7). As indicated by the statistics on malnutrition and stunting, the improvements have yet to address the ‘silent famine’ of chronic food insecurity. Based upon health and education indicators, residents of the so-called ‘emerging’ regions of Ethiopia face significantly greater challenges than their fellow citizens - this category is used by the Government of Ethiopia (e.g. MFA, UNCDF and UNDP, 2007), and is defined as faring poorly compared to other regions, on the basis of measures of poverty, of basic services and availability of basic infrastructure. These regions have been largely excluded from nation-building, and for many of their residents the narrative of colonialism is better suited than that of national building. Weber’s description of the processes occurring in rural France in the 1800s seems a fitting description of governmental efforts in recent decades: “…the unassimilated rural masses had to be integrated into the dominant culture as they had been integrated into an administrative entity. What happened was akin to colonization and may be easier to understand if one bears that in mind” (1976: 486). In the Ethiopian context this includes the administrative description of residents of these areas as backward, as well as the residents themselves describing the ‘national’ language as a slave language and its speaking as a constant reminder of the dominance of northern highland peoples over the rest.2    	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  2 On one occasion in eastern Ethiopia some people refused to speak with me because I had learned and used the ‘slave language’ and not a local language. The gravity of these issues are not specifically linguistic, but also historical and cultural; many regions pass on detailed histories of the atrocities they endured as their people were brought under the control of the Ethiopian state (for many this occurred in the late 1800s, but for some it continues). The stories of oppression and injustice are reinforced in a variety of ways; one example of this comes from eastern Ethiopia where the interior of a certain part every house is painted red, representing the blood of their young men killed by the government. The 2015 election provided many examples of ethnic-based rhetoric and the perception of colonization, whereby people were encouraged to vote for someone from “their people” so that the invading northerners could be kicked out of their lands and their properties confiscated. 	   30 Agricultural Sector Ethiopia is primarily an agricultural economy built upon smallholder agriculture. Almost half of the GDP is agriculturally based, and smallholder farmers cultivate more than ninety percent of agricultural land (Taffesse, Dorosh and Gemessa, 2012). Exports are also primarily agricultural, including coffee, khat, oil seeds, fresh cut flowers, cereals and vegetables (Cochrane and O’Regan, 2016; NBE, 2014). Nearly eighty five percent of employment is within the agricultural sector, which is an area of the economy that continues to grow in importance with time (Loening, Durevall and Birru, 2009). The foundation of this sector, individual smallholder farmers, face vulnerabilities due to unpredictable rainfall and a lack of irrigation (Cochrane and Gecho, 2016). ‘Smallholder farmers’ are defined as those who cultivate less than 25.2 hectares of land and largely produce for their own consumption with the surplus for market sale (Taffesse, Dorosh and Gemessa, 2012). In practice, holdings are much smaller: sixty percent of smallholder farmers cultivate less than 0.9 hectares of land and forty percent less than 0.52 hectares (Taffesse, Dorosh and Gemessa, 2012). In the highland areas, per capita land holdings have dropped from 0.5 hectares in the 1960s to 0.2 hectares as of 2008 (Spielman, Mekonnen and Alemu, 2012).  Smallholder farming is almost entirely rain-fed (CSA, 2009) and due to declining landholding size as a result of population growth, productivity per household and average yields per capita are declining (ACCRA, 2011).  The Government of Ethiopia encourages the use of market-based inputs (fertilizer and pesticide) and improved seed varieties; it supports research and provides subsidies (when direct subsidies for inputs have not been applied the subsidization is indirect as the government offers credit services and uses the governmental agricultural extension system for promotion and distribution). However, uptake of improved varieties was mixed; as of 2008 seventy one percent of wheat crops were improved varieties but only twenty percent of maize, and adoption of improved varieties of other crops is lower still (Spielman, Mekonnen and Alemu, 2012). About a third of all smallholder farmers use fertilizers (the CSA reported 39% and the ERSS 32%), largely for teff, wheat and maize production (Spielman, Mekonnen and Alemu, 2012). Credit barriers and low and inconsistent levels of input supply prevent greater uptake. Services advocated by agricultural extension 	   31 workers, such as planting methods, experience mixed uptake, with a discontinuation rate potentially as high as a third of all these adoptions (Bonger, Ayele and Kuma, 2004; EEA/EEPRI, 2006).  Commercial farms, held by the state or by the private sector, are defined as being larger than 25.2 hectares, but in actuality average 323 hectares in size. The products of these farms are sold to local and international markets but are limited in quantity, making up about four percent of national production (Taffesse, Dorosh and Gemessa, 2012). Although they comprise a small share of the national agricultural picture, commercial farms account for large shares of specific crops, such as coffee (19.1%), fruit (19.4 percent), vegetables (23.7%), sugarcane (78.1%) and sesame (42.6%) (Taffesse, Dorosh and Gemessa, 2012). Commercial farms more commonly utilize mechanization, irrigation and external inputs, whereas these technologies are less common on smallholder farms. Yields can be as much as three times higher as a result (Taffesse, Dorosh and Gemessa, 2012). Over the last decade, steady and significant gains have been made in average yield per hectare in teff and maize, on the national, regional and zonal levels (see Chart 2.1). However, that data is questionable and highly politicized; Jerven’s (2013) research on the quality of statistical data in Africa suggests data quality issues are common, and examples of its problematic nature are discussed throughout this work. For example, during the 2007/08 season a higher yielding variety of taro was introduced, yet no increase was recorded in the years that followed.     	   32  Figure 2.1 Productivity (50 kg units) per Hectare of Teff and Maize Source: Central Statistics Agency, 2001-2015.  In addition, the CSA data on sweet potato for SNNPR from 2007 to 2011 are stable, but during this time sweet potato virus disease infection was extremely high and was affecting roots, weight and cuttings (Tefera, Handoro and Gemu, 2013). Furthermore, in the 2012/13 planting season the yields per hectare of taro and sweet potato, two crops of primary importance in southern Ethiopia, tripled, according to governmental data (in Wolaita Zone taro production rose from eighty six quintal per hectare to 327, and sweet potato rose from 106 quintal to 241 and then to 364 in the two following seasons). According to the Head of Agricultural Statistics at the Central Statistics Agency of Ethiopia, the reported improvements were the result of changes in methods for calculating yields in the annual surveys, not necessarily actual changes (Personal Communication 3 April 2016). Yield data is political and politicized. The trends in Figure 05101520253035Ethiopia	  (Teff) SNNPR	  (Teff) Wolaita	  (Teff)Ethiopia	  (Maize) SNNPR	  (Maize) Wolaita	  (Maize)	   33 2.1 are government data, which should be viewed as important components of a governmental narrative of growth and progression toward the government-mandated targets as much as they are reflections of actual agricultural output. Development narratives from the Government of Ethiopia, as are all narratives, are shaped by the inclusion and exclusion of information, the selection of metrics and the interpretation of the data (Cochrane and Skjerdal, 2015).  These are, however, averages. As Scott described, the “farmer rarely experiences an average crop, an average rainfall, or an average price for his crops” (1998: 46). Average yield per hectare does not highlight localized crop failures, due to unpredictable rainfall or disease. Nor do average yields result in improved or equitable access. For example, Sen (1981) found that in the 1973/74 famine there were limited food shortages, but significant shortages of purchasing power. Similarly, during the 1982-1984 famine period, national yields were stable, with regional spikes of prices (de Waal, 1997). As will be explored in this dissertation, average agricultural yields of smallholder farmers can increase while food security remains chronic due to poverty, declining land holding size and increasing inequality. These changes result in increased yields for the relatively better off smallholder farmers and increased vulnerability, poverty and food insecurity for the relatively poorer and landless households.  Being a country heavily dependent upon agriculture, climate change is considered a significant threat to the development of the country. Ethiopia has been ranked tenth out of 230 nations in terms of its vulnerability to climate change (ACCRA, 2011; CGD, 2014). The changes in rainfall, temperature and weather variability have already begun to negatively affect lives and livelihoods in parts of Ethiopia, and the projected changes are expected to continue, and worsen in impact due to the country’s limited capacity to adapt (Cochrane and Costolanski, 2013; Di Falco et al, 2011; Kassie et al, 2015; ND-GAIN, 2016; Wheeler and von Braun, 2013).   	   34 2.2 SNNPR  According to the 2007 national census, when Ethiopia’s population was estimated to be slightly under 80 million, the most populous regions were Oromia, Amhara and SNNPR, accounting for four-fifths of the entire population (CSA, 2007). The research sites of this study are located in SNNPR (see Figure 2.2), which has the highest rural population density in the country. This population density is particularly high in eastern and central SNNPR (see Figure 2.3), and all of the rural districts with a population density at or above five hundred persons per square kilometer are located in SNNPR (CSA, 2007).3 In addition to the pressures discussed above, this region encounters irregular rainfall patterns, lacks adequate infrastructure and is affected by crop disease and pests. Together, these factors result in SNNPR experiencing high levels of food insecurity. For example, more than half (55%) of the districts are reliant upon the Productive Safety Net Program  for their basic needs to be met (FEWS NET, 2012b).     	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  3 The districts per square kilometer are: Wenago 1121, Damot Gale 746, Aleta Wendo 705, Yirgachefe 677, Sodo Zuria 638, Kacha Bira 637, Angacha 624, Dara 633, Kedida Gamela 594, Shebedino 592, Boloso Sore 583 and Awasa 565 (Adugna, 2014). All of these districts are located within SNNPR. 	   35  Figure 2.2 Administrative Zones of SNNPR (Wolaita Zone identified) Source: UN OCHA (cited in Adugna, 2014)  Emergency situations can result from consecutive seasons of low agricultural production, as happened in 2011 and 2012, largely due to low levels and irregular patterns of rainfall (FEWS NET, 2012b). Periods of crises can result in 150 percent increases in admissions of malnourished children, complete loss of long-cycle crops (maize and sorghum), and up to 400 percent price increases of staple crops, such as maize (FEWS NET, 2011d). The Productive Safety Net Program operates when and where needed; in 2012 it operated in seventy eight districts, more than half the total districts in SNNPR4 (FEWS NET, 2012b) and supported more than three hundred thousand beneficiaries in SNNPR (FEWS NET, 2012a).  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  4 As of August 2014 there were 158 districts in SNNPR. Those districts exist within 19 Zones, however four of the districts are considered Special Districts, which are not included in the list of 158 and function administratively as Zones. 	   36   Figure 2.3 Population Density by District, SNNPR (Damot Gale District identified) Source: Adugna, 2014  The food security situation within SNNPR is complex and generalizations cannot be made about all districts. In the western parts of the region, food security is stronger, due to relatively consistent rainfall (FEWS NET, 2013), while the central and eastern areas of the region experience moderate food insecurity that can fluctuate from minimal stress to generalized crises (FEWS NET 2010; 2012a; 2012b; 2014). In addition to annual 	   37 fluctuations, food security is seasonal, resulting in food insecurity existing for specific months of each year, in addition to the significant portion of the population in SNNPR that is chronically food insecure (FEWS NET, 2011b). Root crops are of primary importance to the agricultural system in SNNPR, and as such water stress – too little, too much or at the wrong time – can significantly impact yields. Sweet potato is notable is this regard because it is an important crop for poorer households to bridge the food gap that typically occurs from April to June. Livelihoods in the south and southwest of Ethiopia tend to depend on a combination of crop cultivation and livestock husbandry. The traditional crops include enset, maize, teff, barley, sweet potato and taro. Root crops form the basis of the agricultural complex in southern Ethiopia, with enset having important socio-cultural value as well as an important role in assuring food security in the lean season. All the root crops (enset, sweet potato and taro) are high yielding per hectare, when compared to cereals. Teff is also notable in Ethiopia; it is the most important and valuable cereal, and also stores well and provides good fodder for livestock (McCann, 1995). A large number of fruits are grown, such as avocado, banana, orange and papaya, with cash crops of coffee, khat, bamboo and eucalyptus (Tesfaye, 2008).  Livestock include cattle, sheep, goats, horses, donkeys and chickens. Animal products are important as the enset-based diet is low in protein. At the same time, enset cultivation requires significant amounts of fertilizer, which is provided by the animals (Tesfaye, 2008).  Agriculture Based upon many of the newly developed agricultural policies, strategies, plans and agencies of the Government of Ethiopia (GoE), it might be assumed that the push for increased inputs in the agricultural sector is relatively new, and an action of the current government. It is not. In fact, inputs have been distributed in rural parts of Ethiopia, including the research areas, dating back to the early 1970s (Rahmato, 2007).  	   38 Despite the long-term advocacy by the government for farmers to utilize fertilizer, amongst other inputs (e.g. pesticides and improved seeds), the uptake has been moderate. Amongst the range of innovations provided to farmers, including the above agricultural inputs, as well as new planting methods and credit services, fertilizer uptake is arguably the greatest success, in terms of farmer adoption (Taffesse, Dorosh and Gemessa, 2012). However, according to national data only thirty two to thirty nine percent of smallholder farmers use fertilizers (Spielman, Mekonnen and Alemu, 2012). To varying degrees, all of these services have been promoted for decades (Rahmato, 2007) and international research projects seeking to understand their poor levels of adoption success have been conducted since at least the 1980s (Kebede, Gunjal and Coffin, 1990). Given almost fifty years of advocacy, this is a dismal failure. One anecdote about fertilizer use by smallholder farmers, provided by a faculty of a local university, is that even when they are highly encouraged (i.e. forced) to buy it by governmental workers, they re-sell it on the market for a reduced price as they cannot afford it (Personal Communication 4 April 2015).  One of the reasons farmers are reluctant to take up these innovations is that once they begin relying upon inputs, they want to be sure that such inputs will be available year-to-year; some studies indicate that the percentage of households receiving full packages of fertilizer and seed can be as low as twenty two percent (Tadesse, 2014). Concerns about availability and access are supported by data on fertilizer distribution in Wolaita Zone in the 1970s, which indicates that the region experienced significant fluctuations in availability, shown in Figure 2.4. Furthermore, the variability and uncertainty in the input supply continued into the 2000s (Rahmato, 2007: 15) and 2010s (Figures 2.4 and 2.5). Additionally, fertilizer is not the primary determining factor of yields, and thus the high costs may not be the best use of limited funds.    	   39  Figure 2.4 Fertilizer Distribution in Wolaita Zone in the 1970s and 2000s, by 50kg bags (x axis are years of the respective decades; e.g. 2 is 2002) Source: Rahmato, 2007  Figure 2.5 Fertilizer and Improved Seed Distribution in Wolaita Zone (2010s) Source: Wolaita Zone Agricultural Office on May 14th, 2015 15,227 19,96040,13022,143 16,67215,65331,56343,54526,930 32,949 21,93312,00017,00022,00027,00032,00037,00042,0002 3 4 5 6 7 8 9 02000s 1970s152,18852,847161,477 184,41879,64854,741 72,61691,413040,00080,000120,000160,000200,0002011 2012 2013 2014Fertilizer Improved	  Seed	   40 For farmers investing in fertilizer, actualized impact upon yield is an important indicator of the value of that investment, however fertilizer distribution data shown in Figure 2.4 and yield data shown in Figure 2.6 during the 1970s highlights that a 101 percent increase in fertilizer distribution from 73/74 to 74/75 was associated with a 26.5 percent yield increase for maize and a 12.5 percent loss in teff yields (Rahmato, 2007). From 2004/05 to 2006/07, fertilizer distribution increased by 178 percent (as shown in Figure 2.4), while maize production increased by forty seven percent and teff production increased by fifty seven percent (Central Statistics Agency 2005-2007). When fertilizer distribution dropped by 188 percent from 2011 to 2012 season, and then increased by 206 percent from 2012 to 2013 (Figure 2.5), maize and teff production, according to governmental data, remained relatively stable, as shown in Figure 2.6 (Central Statistics Agency 2012-2015).    Figure 2.6 Average Teff and Maize Yields in Wolaita Zone in the 1970s and 2010s, by 50 kg unit per hectare5 Source: Rahmato, 2007; Central Statistics Agency 2012-2014 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  5 Fertilizer distribution is listed as specific years (e.g. 2011) as it relates to the time when distributions take place before planting, which occur within a single Gregorian calendar year. Yield data from the main growing season, however, is presented as covering two Gregorian calendar years (e.g. 2011/12) as the harvesting season extends into the beginning of the year that follows it. This is also the reporting approach taken by the Central Statistics Agency of the Government of Ethiopia. 8 7 5 6 520 25.3 21.7 20.7 19.29.07 10.6 11.6424.1 25.73 25.24010203073/74 74/75 75/76 76/77 77/78 2011/12 2012/13 2013/14Teff	  Yield Maize	  Yield Teff	  Yield	   Maize	  Yield	  	   41  These findings suggest that inputs are a relatively poor indicator of agricultural yields. In addition to the variability of provision and the cost of purchasing inputs, there appears to be a weak correlation of fertilizer input with improved yields, which might explain the apparent low interest on the part of farmers. As will be explored in greater detail below, fifty years of agricultural extension activity might have been much more effectively channeled into low-cost, locally-managed irrigation systems. Figure 2.6, on teff and maize yield in Wolaita, is noteworthy because the year of greatest production was the same year as the severe famine in Wollo (1974), which may have taken the lives of up to 300,000 people (Graham, Rashid and Malek, 2012). The relative isolation of famine-affected areas at the time highlights the political and market-based drivers of famine (Sen and Dreze, 1990; 1999). Wolaita has also experienced its share of environmental stresses, including droughts that have resulted in famine. The 1984/85 famine that took the lives of between 400,000 and 1,200,000 Ethiopians greatly affected Wolaita (de Waal, 1991; Wolde Giorgis, 1989). Shortly thereafter, in 1987/88, a localized food insecurity situation emerged due to enset disease combined with insufficient rains (Rahmato, 2007). Serious food insecurity situations occurred in 1990/91, 1994/95, 1998/99, 2000 and 2003. The common causes were irregular and insufficient rainfall as well as crop disease (Rahmato, 2007). Planted crops have significantly different yield per hectare, and can be grouped into three categories: (1) 5 to 11 quantal per hectare for lentils: chick pea, haricot bean, peas and faba bean, (2) 7.5 to 14.6 quantal per hectare for cereals: maize, sorghum, teff, wheat and barley, and (3) 25 to 106 quantal per hectare for root crops: enset, sweet potato and taro (Central Statistics Agency, 2012). While cereals are an important cash crop, and are closely monitored by the government, the most important crops in SNNPR are root crops, which play a crucial role in regional food security, with enset being one of the most important (Olango et al, 2014).  Enset is a crop native to Ethiopia that is grown throughout the south of the country (see Figure 2.7). Enset was domesticated thousands of years ago and its use is one of few 	   42 examples of long-term sustainable agricultural systems in Africa (Brandt et al, 1997). Enset cultivation is considered the most sustainable indigenous system for the highly populated region where it is grown (Tsegaye and Struik, 2002). This historically sustainable, and currently most appropriate, system is largely controlled by women, who manage cultivation, harvesting and processing (Tsegaye and Struik, 2002). Although the plant produces fruit and seeds, they are inedible; what is consumed is the underground stem and the bases of the leaf sheaths (Tesfaye, 2008). In addition to food, enset is also an important source for livestock feed. According to farmers, enset is “the enemy of hunger” (Tsegaye and Struik, 2002: 292).  Figure 2.7 Enset Cultivation and Ethnicities (Wolaita identified) Source: Brandt et al, 1997: 4 	   43  Typically, enset is planted around other crops, resulting not only in a food crop of its own but also an effective means of supporting the prevention of erosion, enhancement of water retention and improvement of soil quality. Enset is often inter-cropped with perennial tree species, such as coffee, avocado and guava, as well as annual crops, such as maize, kale and yams (Olango et al, 2014). Local farmers have immense knowledge about enset varieties, which number upwards of fifty, and the different uses for them. Diversity is valued and households maintain at least two varieties, with at least twenty five varieties maintained in Wolaita (Tsegaye and Struik, 2002). Some of the reasons for keeping multiple varieties include: meeting different needs, providing flexibility, and enhancing the yield (Tesfaye, 2008).  Enset is affected by disease and pests, and traditional practices, such as intercropping, were developed to prevent and control the spread of these problems (Mulualem and Walle, 2014). Enset also has significant cultural importance in many parts of south and southwestern Ethiopia, and is used as an expression of identity as well as status. For this reason it is said that ‘we were born and grew up on enset, we are the people of enset’ (Olango et al, 2014:12). Another root crop common in SNNPR is sweet potato, which is sensitive to insufficient or too much moisture (FEWS NET, 2011a). Even a slight reduction of potato production can result in a drastic deterioration of food security because sweet potato is the primary crop that bridges the food gap for the poor and very poor households between April and June (FEWS NET, 2011d). Low availability of cuttings for planting sweet potato can have long-term impacts. For example, the poor rains of 2008 continued to negatively affect sweet potato cutting availability in 2012 (FEWS NET, 2012b). When important transition crops, such as sweet potato, are lacking, consumption substitution may take place with enset. However, enset requires five years to mature, and its overconsumption may allow a short-term nutritional gain while creating long-term vulnerabilities (FEWS NET, 2012b). Taro is a third crucial root crop common in the south and southwest of Ethiopia. This crop was introduced at least two thousand years ago to Africa after having been domesticated in Asia (Fujimoto, 2009; Mulualem, WeldeMichael and Belachew, 2013). 	   44 Taro is cultivated for its starchy roots, and like enset, farmers maintain its diversity for specific purposes, having multiple varieties on their farms at one time. 	  2.3 WOLAITA6  As a socio-cultural, linguistic and identity-based area, Wolaita has existed for as long as records of the area exist (Pankhurst, 1997). However, the administrative and political classifications have not been stable. During the time of Haile Salessie’s Imperial government (1930-1974) the area was called Wollamo, and was located within Sidamo Province, which covered a territory similar to the current Wolaita Zone. During the early years of the Derg government (1974-1991)7 there was no change. However, a restructuring and administrative reorganizing of the country in the late 1980s resulted in the merging of districts, creating the North Omo Region. When the Derg government was overthrown and the Ethiopian Peoples’ Revolutionary Democratic Front (EPRDF) regime began, Wolaita did not benefit from the new powers of self-determination given to the regional states because it was embedded within the North Omo Region. Although the desire for self-determination and administrative independence was not eliminated, that sentiment did not result in political action until the 1990s. The spark of the change was driven by language, with Wolaita language (Amharic: Wolaitenya) being intimately tied to ethnic identity. The government planned to merge three regional languages, with Wolaita being one of them, into one, which would be used as a language of instruction in schools. The people of Wolaita started a struggle to regain their political and administrative control, which included violent protests. While protest is not unheard of in rural areas, the governmental response was somewhat uncommon: after years of 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  6 “Wolaita” is spelled in a number of ways (e.g. Wolayta, Wollaita) and was previously named Wollamo, Wolaita is used as this is the version used by the Zonal Administration. 7 The leader of the Derg, Mengistu Haile Mariam, officially abolished the Derg government in 1987 and established the People’s Democratic Public of Ethiopia, which he led until his overthrow in 1991. Academics and non-academics tend to label this entire period (1974-1991) as being ruled by the Derg government. 	   45 activism the government responded to the demands of the people of Wolaita. In 2000, Wolaita Zone became its own administrative zone within SNNPR. In 2005/6 some of the districts within the zone were divided, expanding the seven districts that had existed since the 1960s to twelve districts and three administrative towns. Within each district and administrative town there are sub-districts. As of 2014, there were a total of 352 sub-districts, 56 (16%) of which are considered urban sub-districts (see Figure 2.8).    Figure 2.8 Administrative Districts of Wolaita Zone (Damot Gale District identified) Source: Wolaita Zonal Administration on February 28th, 2015. The Administration was not able to provide an electronic version.  	   46 Wolaita was one of the predominant kingdoms ruling in southern Ethiopia before the expansionist Ethiopian state overtook its territory. Wolaita strongly resisted incorporation, but was defeated by Menelik II in 1894 (Aalen, 2011). Defiance is a tradition that has continued in different forms over the decades, as described above. Although Menelik II and the rulers that followed attempted to dismantle the traditional power structures that enabled a strong resistance to exist, it also incorporated Wolaita elites into its system, and thus offered a degree of continuity for powerful lineages and families within Wolaita until the Derg land reform of 1975 (Chinigo, 2015).  The ethnicity of the zone is almost entirely Wolaita; according to the 2007 census the figure was over ninety six percent. Ethnic homogeneity is a product of the 2000 political restructuring along ethnic lines, thus general linguistic, cultural, and ethnic homogeneity is expected. The SNNPR region, however, is uniquely home to a diversity of fifty six ethnicities. While Wolaita Zone is ethnically homogenous its residents interact with a greater diversity of ethnicities when compared to other regional states wherein ethnic homogeneity exists throughout. While Christianity is the dominant religion, there are two major sects, which can result in significant tension; fifty five percent are Protestant and forty percent Ethiopian Orthodox, while the remaining five percent are Muslim (CSA, 2007). These dynamics differ significantly from national religious affiliation, being 43.5 percent Ethiopian Orthodox, 33.9 percent Muslim, 18.6 percent Protestant, 2.6 percent ‘traditional’, 0.7 percent Catholic and 0.6 percent ‘others’ (CSA, 2007). As is common in Ethiopia, religious adherence is important in daily life in Wolaita, and it influences who one interacts with and how. Wolaita zone ranges from 500 to 3,000 meters above sea level and is classified into three agroecological zones: high-altitude (dega), mid-altitude (wayna dega) and low-altitude (kola). The majority of the land, sixty percent, is located in mid-altitude areas, with a small percentage of high altitude areas (Rahmato, 2007). The mid-altitude and high-altitude areas are where three-quarters of the population live, and account for four-fifths of the food crops grown in Wolaita (Rahmato, 2007). The areas of lower elevation, accounting for a third of the land area in Wolaita, such as parts of Humbo and Diguna Fango districts, do not practice root crop agriculture and their unique contexts will not be 	   47 explored in this research. Root crops are not grown in the low elevation areas, where different cash crops, such as cotton and tobacco, are grown. Settlement (and in some instances re-settlement) into low altitude areas has been slow due to the presence of animal diseases (e.g. trypanosomiasis) not found in the mid- and high-altitudes, and higher levels of malaria. Efforts to control disease in recent decades have facilitated new flows of migration and resettlement (Rahmato, 2007).  Historical population data on Wolaita Zone is unavailable until the 1960s, around which time the administration began to recognize over-population as a challenge (Rahmato, 2007). At the time, the population was estimated to be 600,000 people (CSO, 1966; cited in Rahmato, 2007). By the 1994 national census the population had almost doubled to 1.13 million (CSA, 1996), in the 2007 census it had risen to 1.5 million (CSA, 2007) and by 2014 it had risen to 1.9 million.8 The population is almost entirely rural and not experiencing urbanization at the same rate as other parts of the country. In 2005 only eight percent of the population was urban, a figure that had only risen one percent since 1994; in contrast the national urban population was fourteen percent in 1994 and sixteen percent in 2005 (CSA, 2011; Rahmato, 2007). In the Imperial times of the 1960s, when population pressure was identified as a problem by governmental officials, the proposed solution was resettlement, including to low elevation areas within Wolaita Zone, which was done at the time (Rahmato, 2007), an approach advocated by both the Derg and EPRDF governments.  International non-governmental organizations (INGOs) began to have a more significant presence in Wolaita in the 1980s, previous to which a limited amount of services were offered by Catholic and Protestant churches (Rahmato, 2007). The expansion of INGOs into Wolaita aligns with the rapid increase of INGOs globally and the funding they received during this period.9 The entry of INGOs to Wolaita also aligned with the international media attention that brought the 1984/85 famine in Ethiopia into the homes of people throughout the world. The Derg government, which ruled at this time, was suspicious of INGO activity and heavily regulated the location of programs and the 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  8 Based upon data from the Zonal Administration Office, provided on May 14th, 2015. 9 In 1970 INGOs received US$ 860 million and 1980 INGO funding has rapidly risen to US$ 2.3 billion (Riddell, 2007).  	   48 types of activities they engaged in. The control of INGOs changed in the 1990s under the early years of the EPRDF government, restricting their ability to operate in the country. The result of this was an expansion of community-based and national organizations, with which (or through which) international organizations continued to offer goods and services. However, the government has again become increasingly suspicious of INGO activity and introduced strict regulations in 2009. In the traditional Wolaitan system, when a sufficient amount of land was available, households divided their land into sections: (1) enset around the home, (2) the darkua area with mixed root crops and non-root crops, (3) the shoqa field for cereals, and occasionally (4) an outa for trees and grass (Rahmato, 2007). The amount of land allocated to root and cereal crops still somewhat reflects these patterns, as shown in Figure 2.9 below. The utilization of space as designed in the traditional Wolaita system is remarkably similar to the models advocated in contemporary research about permaculture, which takes into account the distribution of organic material as well as required labor (e.g. Altieri, 1995; Holmgren, 2002; Mollison, 1991). However, this traditional system was disrupted by declining land size per capita as well as advocacy by the government to shift to cereal crops, which are primarily for market sale and export (Eyasu, 2000; 2002). Rahmato writes that the “strategy of changing the cropping system pursued by WADU [Wolaita Agricultural Development Unit] by encouraging a shift from emphasis on root crop cultivation to cereal cultivation was, under the prevailing circumstances, ill advised” (2007: 33).  	   49  Figure 2.9 Land Allocation (ha) by Crop in Wolaita Zone (2011/12) Source: CSA, 2012  While the shift of land use and crop choice negatively affected traditional agricultural systems, arguably the greatest change was that the required size of landholdings to implement this system no longer existed. Rahmato has described Wolaita as “a land of micro-holdings” wherein land “holdings have always been small relative to other parts of the country” but have “been growing smaller through the decades” (2007:3). Rahmato’s differentiation between ‘smallholder’ and ‘microholder’ is based on land size less than 0.5 of a hectare, a plot that, he argues, can no longer sustain those who farm it and who experience “collapse under even minimum pressure” (Rahmato, 2007: 10). The majority of smallholder farmers in Wolaita farm micro-plots, which is the result of decades of high population growth and land fragmentation. Rahmato cites a report conducted in 1976 that found the average landholding size was 0.7 hectare. A survey Rahmato conducted in 1989 found that forty five percent of households had less than 0.5 of a hectare (Rahmato, 1992). In 2003, 53.3 percent farmed less than 0.5 ha, with only 4.4 percent farming more than two ha (CSA, 2003; cited in 24,1531,09720,1994,0342,435 23,3221,0219771,236 7,4017,3467,6436,9600 5,000 10,000 15,000 20,000 25,000 30,000MaizeSorghumTeffWheatBarleyHaricot	  beanPeaFaba	  BeanChickpeaSweet	  potatoTaroEnsetCoffee	   50 Rahmato, 2007). In the 2012-13 season, 57.4 percent held less than 0.5 ha, and the average land holding in Wolaita Zone was 0.58 ha, with only 1.7 percent holding more than two ha (Central Statistics Agency, 2013). In 2015, the average land holding size in Damot Gale district (where this study took place) had fallen to 0.25 hectares.10  Cattle play an important role in plowing fields and providing manure, as well as meat and milk, from which products are derived and sold on the market, such as butter. Yet, plots have become too small, and the economic pressures too high, that maintaining cattle has become less viable. In 1976, fifty eight percent of households used oxen for cultivation, but by 2007 only nineteen percent owned two oxen and twenty five percent owned one (Rahmato, 2007). In addition, many families do not have transportation livestock, but these are essential for smallholder farmers to bring their goods to regional markets for sale. This transportation also enables individuals to buy goods at the district markets and return them to their communities for sale or trade. However, without transport livestock households rely upon traders for purchasing goods and selling yields. The Zonal Administration Office data on livestock populations (shown in Figure 2.10 below) is suggestive of error, as the variation of livestock populations in this time period is unlikely to increase this dramatically, particularly for cattle, but nonetheless indicates general livestock population trends. In this study, land fragmentation and the decline of communal grazing areas are suggested as contributing to a decrease of livestock holdings. However, the data in Table 2.10 shows increases. There are two potential explanations for this: the trend in Wolaita Zone differs from that of Damot Gale district, and the micro-trends do not appear at this level, or the data is inaccurate. My experience suggests the latter (similar challenges exist with agricultural data).   	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  10 Data provided on 12 June 2015 by the Damot Gale District Agricultural Office, based upon 2015 data. 	   51  Figure 2.10 Livestock Populations in Wolaita Zone in 2011 and 2013 Source: Wolaita Zone Finance and Economic Development Department, 2012; 2015  In addition to the challenges of fragmentation of micro-plots, smallholder farmers are vulnerable to unpredictable rainfall – too much, too little, too early or too late - because only 0.4 percent of the land in Wolaita is irrigated (Rahmato, 2007). The result is greater demand from smaller plots, which has pushed farmers away from the “sound and sustainable” traditional practices of land use and crop choice, to those that provide the greatest benefit in the short term (Rahmato, 2007: 9). Rahmato concludes that agriculture in Wolaita “has exhausted its potential and is becoming increasingly unviable for the great majority” (2007: 17). The percentage of the population reliant upon food aid reflects these changes. In 1994, 17.8 percent of the population required food aid (CSA, 1996; Rahmato, 2007) and based upon available district level data, the percentage of the population enrolled in the Productive Safety Net Program, which serves rural food-1,146765156 250617489850717216050530 100 200 300 400 500 600 700 800 900 1,0001,1001,200CattlePoultrySheepGoatsDonkeysHorsesMuelsThousands20112013	   52 insecure households, ranged from fourteen to thirty one percent (Cochrane and Vercillo, 2017).11 While the land situation is dismal, Rahmato’s research indicates that root-crop based agriculture can provide sufficient yields for household consumption with plots ranging from 0.1 to 0.8 of a hectare (Rahmato, 1995). However, the potential for self-sufficiency using root crops must be considered in light of the decline of cattle holdings, which is an essential contributor of the manure fertilizer required by root crops. It must also be considered in light of the low protein, carbohydrate-based diets that result from the consumption primarily of these root crops, which can result in nutritional deficiencies. The figures from Rahmato are a potential based upon average output, but farmers rarely experience an average rainfall and average yield. A single year of poor yields can require several years of good or above-average yields to recover from asset loss (FEWS NET 2012b).  Rahmato (2007) appropriately draws attention to a segment of society in an even more difficult situation than those farming micro-plots of land: the landless. For this segment of the population, Rahmato states that almost no data is available, but he suggests that the proportion of the population may be as high as fifteen percent. The livelihood of the landless revolves around their labor; as farm laborers or as share-croppers – engaging in migration by necessity (Cochrane and Vercillo, 2017). Off-farm and non-farm activities that are commonly engaged in include: trading of small goods, unskilled wage labor, handicraft production, collection and sale of wood, charcoal or grass, as well as individual service provision, such as care work. One of the internal pressures keeping the population attached or semi-attached to Wolaita Zone is land inheritance (specifically land use inheritance), which is commonly sub-divided amongst numerous heirs. All land is owned by the government, and only recently have certifications affirming and protecting the right to use land been issued (discussed in Chapter 7). The land outlined in such certificates cannot be sold or transferred, and can only be inherited by direct descendants. 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  11 Food aid, in the past and present, refers to food distributions in response to specific emergency needs. The PSNP provides regular, multi-year transfers to households for six-months of the year. In most regional states, the transfer is made in the form of cash payments, for which labor contributions are required (see Cochrane and Tamiru, 2016). 	   53 If the land is unused, it returns to the governmental land bank for redistribution. As a result, seasonal migration is common, but permanent migration is rare due to the expectation of inheriting land, while semi-permanent migration is done only by a portion of the household to ensure the land remains used and therefore retained by the family.   2.4 DAMOT GALE  The district within which the research sites were selected for this dissertation, Damot Gale (see Figure 2.11), has the second highest rural population density in the country: 746 persons per square kilometer (CSA, 2007). The district (woreda) of Damot Gale is broken down into thirty four sub-districts (kebeles), three of which are considered urban that are all located near to Boditi Town Administration.12  However, despite three sub-districts being classified as urban, the entire population of Damot Gale is considered to be rural by the national definition.13 Based on information from the Zonal Administration Office, fifty seven percent of the land in Damot Gale district is cultivated, eleven percent is used for grazing, nineteen percent covered by forest and bush and the remaining three percent is covered by water or residential areas. While a third of Wolaita Zone is low altitude, wherein significantly different agricultural systems are practiced, Damot Gale is a district that is representative of the root crop agricultural system, with only a small fraction of low altitude land. In Damot Gale, sixty two percent of the district is located in the mid-altitude (woyna dega) agroecological zone, thirty three percent in the high-altitude (dega) zone and five percent 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  12 Sub-districts are divided by population (as opposed to geographical area) and new sub-districts are introduced regularly. For example, in 2012 Damot Gale had thirty-one sub-districts, and in 2014 that has risen to thirty-four. The geographical area of districts ranges significantly. In Damot Gale, Humbo sub-district covered 859 km2 while Damot Pulasa covered only 165 km2, and the Administrative Towns cover smaller areas, such as Boditi covering only 20 km2. This data was provided by the Wolaita Zone Administrative Office on May 14th, 2015. 13 The classification of sub-districts as urban appears to be related to administrative classification and nearness to Town Administrations, rather than having a certain percentage of the population considered urban according to the national definition. 	   54 in the low-altitude (kola) zone. The sub-district is home to the highest point in the Zone, Damot Mountain, at 2,958 meters above sea level. The mean annual rainfall in Damot Gale district ranges from 1,000 to 1,400 mm, and the mean annual temperature from 12.6 to 22.5 degrees Celsius.14   Figure 2.11 Administrative sub-Districts of Wolaita Zone (three study kebeles identified: Adeaaro, Adea Ofa and Buge) Source: Wolaita Zonal Administration on May 14th, 2015  Based on 2014 data provided by the Zonal Administration, there were 149,823 people in Damot Gale, living in 29,115 households, making the average household size 5.2 persons. Of that population, 38,258 (26%) are enrolled in school; and of the school enrolled 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  14 This data was provided by the Zonal Administrative Office on May 14th, 2015. 	   55 children 25,092 (66%) are in grades 1-4, 11,591 (30%) in grades 5-8, 1,575 (4%) in grades 9-10, and 0 in grades 11-12.15 As these figures indicate, rates of enrolment significantly decline with age. One of the primary reasons for this decline is access. In Damot Gale there are two schools offering grades 1-4, thirty four schools offering grades 1-8, one school offering grades 9-10 and no schools offering grades 11-12. Of note: according to governmental data, there is only a gender disparity of a few percentage points throughout all grade levels. However, enrollment data reflects the existing options; although some students from Damot Gale may attend secondary school in other sub-districts, which does not appear in the currently available data. Opportunities to attend school, although they remain limited, are largely a new experience. In 1962 there was only one public school and one non-governmental school serving the entire Wolaita Zone (Rahmato, 2007).  Accessibility to these schools is greatly influenced by distance and transportation infrastructure. As a result, there are significant barriers for those living in sub-districts without asphalt road access to attend secondary school. As of 2013, there were only twenty seven kilometers of asphalt road and twenty six kilometers of gravel road in Damot Gale district, and the Zonal Administration reports that there is no public transportation coverage from Damot Gale to the Zonal Capital, Sodo.16 However, private transportation is present, largely in the form of motorbikes on dirt roads and minibuses on asphalt.  Wolaita Zone has two hospitals, seventy one health centers and 348 health posts.17 Within Damot Gale there are seven health centers and thirty one health posts. Despite these facilities, Damot Gale is home to 0 physicians, 125 nurses (1:1,199), 18 pharmacists (1:8,324), 19 laboratory technicians (1:7,885) and 64 health extension workers (1:2,341).18 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  15 This data was provided by the Zonal Administrative Office on May 14th, 2015. Its validity is highly questionable. The same administration office provided three different data sets on enrollment, which significantly conflicted with each other (see Sandefur and Glassman (2015), Carletto, Jollife and Banerjee, 2015, and Jerven (2013) for more detailed explorations of the problems of statistical data in developing countries). As a result, these figures should be understood as indicators of general trends, rather than exact data.  16 This data was provided by the Zonal Administrative Office on May 14th, 2015. 17 Non-governmental data from Management Sciences for Health Ethiopia, an organization I was a consultant for from 2013 to 2015, supports the government data for health posts and hospitals, however lists only seventeen health centers.  18 This data was provided by the Zonal Administrative Office on May 14th, 2015. 	   56 Throughout the entire district of Damot Gale there are only eight pharmacies. However, the level of current healthcare service coverage requires contextualization: in 1962 there was only one hospital and one charitable clinic in the entire Zone of Wolaita (Rahmato, 2007). While the rapid expansion of health posts, including in rural areas, is impressive (see Taddesse, Jamieson and Cochrane, 2015) the services offered by health posts are limited because they are staffed by nurses or health extension workers and have to cover large geographical areas. As a result, any serious medical situation requires travel to a health center or hospital. For those living near to asphalt roads, this poses fewer challenges. However, individuals living in rural areas may not have access to vehicles, and have limited network coverage to call for vehicles, resulting in extremely high costs for emergency transportation. To put those costs into context, the cost of emergency transportation to towns from rural areas can be the equivalent to 200 kilograms of maize. Considering families in Damot Gale have, on average, 0.25 hectares of land for cultivation, these costs consume a significant portion of their income and, for many, pose insurmountable barriers for accessing healthcare services. The important role that transportation infrastructure plays for rural residents is not a recent discovery in the development sphere. Weber’s (1976) history of the modernization of France outlines the changes that occurred in the 1800s and early 1900s and almost all of the socio-cultural, political and economic changes were founded upon transportation infrastructure. Changes in communication, migration, taxation, military, and industry would not have occurred without transportation infrastructure. Weber states, “until roads spread, many rural communities remained imprisoned in semi-isolation, [and its members] limited participants in the economy and of the nation” (1976: 195). In rural Ethiopian contexts, like Damot Gale, we can also add: limited access to education beyond primary school, limited access to healthcare except basic services and almost no access to improved services, such as safe drinking water, or access to infrastructure, such as for irrigation.  Agricultural support services have rapidly expanded in recent decades, but remain few and far between. For example, Damot Gale is home to almost 150,000 people and each sub-district has one Farmer Training Center (FTC), that collectively have trained under 	   57 2,000 people (less than 2% of the population). In many communities in Damot Gale the FTCs have limited functionality. One of the FTCs in the communities within which this research took place was rarely used because the agricultural extension worker lived in the town and rarely visited the area covered by this FTC. According to the job requirements, the agricultural worker is supposed to live within the community, but this is not always the case. The entire Wolaita Zone only has one Agricultural Research Center, based in Areka, located forty to fifty kilometers from the research areas. In addition, there is only one animal health laboratory for the entire Damot Gale district and it has no livestock breeding centers. The Zonal Administration data suggests that the 110 agricultural extension workers employed in Damot Gale have complete coverage of all households. While this may technically be the case, these 110 agricultural extension workers cover a geographical area of 235.5 km2 and provide services to a large population. In addition, few live in close proximity to FTCs and due to the time-sensitive nature of plowing, planting and harvesting, it is not practically possible for extension workers to reach farmers when they are in need of extension support. Depending on which figures are used from the Zonal Administration, the ratio of agricultural extension workers to households ranges from 1:265 to 1:348. Animal health extension services have an even larger area to cover, with the entire district of Damot Gale having seven veterinary technicians and one veterinarian, who provide service to large livestock populations, including 93,770 cattle 60,018 goats and 115,970 chickens. Even if livestock figures are inflated, without livestock health support, such as vaccinations, animal disease results in significant losses and create disincentives for future livestock investments (see Chapter 6). Due to the national and export importance of cereals, crop production data in Wolaita Zone focuses upon maize, sorghum, teff and wheat. In Damot Gale the crop grown in the largest area is teff (3,686 ha), followed by wheat (2,081 ha), maize (1,426 ha) and sorghum (299 ha). Based on the 2014 total production and land coverage data, the average yield per hectare for teff is twenty quantal per hectare, wheat is forty quantal per hectare, maize is forty five quantal per hectare and sorghum is eighteen.19 These figures either suggest 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  19 This data was provided by the Zonal Administrative Office on May 14th, 2015. 	   58 that yield per hectare has risen two to three fold between 2012 and 2014, that Damot Gale is significantly more productive than all other regions, or the figures are incorrect. My fieldwork suggests the latter. The Zonal Administration does have data on non-cereal production, although some categories are non-specific, such as ‘root crops.’ Similar to cereals, the Zonal data on enset, vegetables and fruits are significantly above national averages. The national and zonal data is further made questionable when district and zonal governments do not have accurate yield data, resulting in large data changes (such as the national 200-300% increases in a single season discussed earlier in this chapter due to methodological changes). Also of note is that the Zonal Administration does not have data on root crops, but somehow the national CSA has specific figures for all root crops. Some of the potential political reasons why data may be incorrectly reported are explored by Sandefur and Glassman (2015). However, it may also be the result of poor data and problematic data collection methodologies, as Jerven (2013) and Carletto, Jollife and Banerjee (2015) have shown to be the norm in many developing country statistical offices. The Zonal Administration reports that thirty seven percent of Damot Gale has access to safe water. That access, however, does not specify distance from available sources, which include hand pump wells and springs. According to Zonal data, one-fifth of hand dug wells with hand pumps are not functional, more than half of shallow wells with hand pumps are not functional, sixty nine percent of deep wells are not functional, and eleven percent of springs with distribution points are not functional. The lack of access means family members have to walk long distances to acquire water, wait several hours in line to fill fifty liter containers and have no ability to irrigate their crops as a result. In addition to the consumption of significant amounts of time, insufficient water access poses challenges for sanitation and hygiene, resulting in greater disease burdens. A second set of data, provided by the same Zonal Administration, suggests that potable water access in Damot Gale has risen from thirty five percent in 2008 to ninety eight percent in 2014.  Damot Gale, like Wolaita Zone as whole, is home to an array of non-governmental organizations. In 2013 this included: Wolaita Development Association, Inter AID France, Rural Community Based Development Initiative Association, Mennonite Economic Development Association, Centro-Auitipar, World Vision, Orbis, Initiative 	   59 Ethiopia International Children’s Association, Mossy Treatment Prevention Association, and Tesfa East Africa Water Charity Organization. Their reported budgets for 2013 surpassed 93,926,844 ETB (~$US 4,900,000). Damot Gale is also home to a government-run microfinance institution, South Omo Microfinance. In this district, it serves 1,383 people (847 male, 536 female); if it is assumed only one member of the household is accessing credit, this means the coverage is less than five percent of all households. This research project purposefully selected three sub-districts in Damot Gale for comparative purposes. Adeaaro was selected due to its proximity to an asphalt road and gravel roads, and hence greater access both for input supply and market access, as well as its nearness to Boditi town, which has a health clinic and secondary education. Adea Ofa, a neighboring sub-district, has a shared agroecological and socio-cultural milieu, but has no access to asphalt or gravel roads, making it challenging and costly to access services located in the town. These two sub-districts are of comparable size, with Adeaaro having an approximate population of 5,333 and 1,011 households while Adea Ofa has an approximate population of 4,000 and 771 households. In Damot Gale district there are significant differences in the number of households within the kebeles, so this similarity of size is not always the case (from 424 in Adea Sibeye and 1,957 in Buge). The third community, Buge, was selected because of the existence of an irrigation project, built ten years before the research took place. The part of Buge sub-district with regular and daily access to irrigation for their agricultural land was selected for research to assess the impact of irrigation access. Further details about these sub-districts are presented in Chapter 6. Having outlined the national, regional, zonal and district contexts, the following chapter analyzes the concept of development, with specific to reference to Ethiopia, as well as the role of power and politics that occur at multiple scales of governance. While Chapter 2 highlighted some of the ways in which power and politics influences activity, such as the politicization of data, Chapter 3 engages with these topics in a more explicit fashion. The role of power and politics, and my understanding of them in relation to research 	   60 approaches and data analysis, set an important foundation for the findings presented in Chapters 6 and 7.  	   61 CHAPTER 3. DEVELOPMENT, POWER AND POLITICS   Analyzing vulnerability to food security, adoption of programs and services and the impact of participatory methodologies necessitates a contextualization of development, power and politics. This is because vulnerability is reduced, managed or increased by development activity, as much as by political action and power relations (Watts, 1983). This chapter will review some of the key literature on development, power and politics generally, and make specific references to food security in Ethiopia. While this chapter does not aim to contribute new knowledge to a well-studied field, it sets a foundation for how development, power and politics relate to food security and the ways in which these terms are utilized throughout this dissertation. Drawing upon relevant literature, the following section explores the contested nature of the concept and practice of development. While brief, it situates my own positionality within the development discourse. The second section explores power and politics, and their relation to food security. Throughout each of these sections I offer reflections on how these concepts and practices influence this research process and analysis. 	  3.1 ENCOUNTERING FOOD SECURITY  This section heading draws upon two books, each offering significantly different perspectives on development as a concept, and as a practice. The first is Encountering Development (Escobar, 1994) and the second is Encountering Poverty (Roy et al, 2016). The former takes a strong oppositional position against development, as a concept and practice that have failed. Escobar argues that development has caused underdevelopment, famine, poverty, malnutrition and violence, and that development is a tool of control akin to colonialism (Escobar, 1994: 4). Alternatively, Roy et al argue that 	   62 one must position themselves within the development discourse, academically or as a practitioner, “to be engaged in the battle of ideas. Instead of positioning critics as those situated outside of development, we seek to explore how those within the system can participate in such struggles” (2016: 46). One might suggest that on a spectrum of framing development positively or negatively, Roy et al (2016) offer a middle ground of critical engagement. Others, such as Kenny (2011), present a positive and optimistic promotion of activity in the name of development. The positionality of opposing or advocating does not, however, address the nuances of the ways in which these authors have defined what they respectively mean by ‘development’ in their works. While Escobar (1994) is an advocate of civil society action for reframing development based on local realities and priorities, he also refers to the common metrics used by Kenny (2011), such as income, health, food security, stability and peace. Roy et al similarly draw upon these metrics but add “dignity, voice and power” (2016: 31), not as measures per se, but as ways in which development activity can be assessed as successful, or not. The diverse means of conceptualizing and assessing development are part of the reason such divergent opinions about it exist. More problematic, however, is that ‘development’ is often not explicitly defined when used, and thus the nuances may not be immediately obvious (Bellu, 2011; Sumner and Tribe, 2008). The framing and dynamics of the development discourse change over time, as new ideas emerge. For some, development, as a concept and practice, emerged in the pre-WWII period as a colonial enterprise (Riddell, 2007). As Eyben (2014) notes, the transition from colonialism into something recognizably different was a slow process. Indeed, many processes are still in this transformation. Following the war and as countries gained independence, international development efforts largely focused upon macro-economic growth (e.g. Lewis, 1955; Millikan and Rostow, 1957; Rostow, 1960). This continued throughout the ‘Development Decade’ of the 1960s, and framing development in this form continues to be common, such as in the works and projects of Sachs (2005).  Although income distribution or inequality were not on the agenda in the immediate post-WWII period, the Universal Declaration of Human Rights, of 1948, set the groundwork for some different directions that would emerge in the following decades. It 	   63 was in the 1970s that poverty alleviation became focal, but that was limited to raising income per capita and employment (Riddell, 2007). It was not until the 1980s that development encompassed health, education and living standards, largely based on the work of Sen (1981; 1983; 1985). It was Sen again, in the 1990s, who was the driving force behind another shift: incorporating opportunities and capabilities in conceptualizing poverty (Sen, 1999) and the emergence of the annual United Nations Human Development Reports, which institutionalized the broader definition of development. Many of these metrics would be used in developing the Millennium Development Goals (2000-2015). The Millennium Development Goals period, however, was influenced by increasing ties between development initiatives and military action in the War on Terror, akin to the politicization of development and humanitarian activity during the Cold War.  For Soubbotina (2000) and Barder (2012), the definition of development must encompass lasting change, and should not be limited to a measurement in one moment of time. The focus on durability of change emphasizes the important role of permanent actors, notably governmental institutions and the ways in which economic, political and social systems contribute to, or negate, development that is sustained (Acemoglu and Robinson, 2006; 2012). Thus the most effective and/or appropriate implementing body for development varies based on what objectives are prioritized. With these additional layers of complexity emerging, development activity is increasingly difficult to define. As a response, Chambers (1995) opted to refer simply to ‘good change,’ and allow the objectives, actions, outcomes and impact specify the details. Development is, therefore, not only a product of the individual describing it and the theoretical focus, but also the time period within which it is written about and the objectives sought. In reflecting upon the idea of ‘good change’ proposed by Chambers (1995) it appears that the critiques waged by Escobar (1994) and the enthusiasm offered by Kenny (2011) can be brought together wherein nuanced analyses may assess for whom good change occurs and in what forms, and for whom negative impacts result and in what forms; when, and for how long. Such an assessment would need to take into account the complex ways that development activity can result in both positive and negative change for a single individual. As this dissertation demonstrates, development activity can 	   64 result in good change, but that is not experienced equally by all, nor in the same ways. It also has the potential to result in negative change, often affecting those already marginalized.  The Government of Ethiopia positions agriculture and food security as development issues. The objectives of the programs and policies it implements in these fields represent the diversity of definitions given to the term. As explored in Chapter 7, interventions made in the name of development within the agricultural sector do not necessarily aim to enhance the food security of all people equally, including those most vulnerable to food insecurity. In some instances, such as with the agricultural extension program, the primary focus has been on high potential agricultural areas and export crops, excluding the more marginal areas and those livelihoods not geared toward export markets. Government-supported cooperatives tend to benefit certain segments of society, not only those with more land but also specific ethnic and religious groups. Microcredit services are only accessed by a small minority, often those with the assets that can ensure repayment regardless of rainfall. Other programs that are designed to support the most food insecure, such as the Productive Safety Net Program, work to entrench elite control and disempower citizens in their implementation.  Development is almost always presented with the assumption that it occurs primarily for the purpose of creating positive change, however development may be engaged with for reasons that are other than its stated objectives. The current Government of Ethiopia is well aware of this dissonance of intentions, as development activity was a primary reason for it coming to power. When the Tigrayan People’s Liberation Front (TPLF) fought to overthrow the Derg government, foreign agencies acted as “the relief wings of the rebel movements, and no realistic distinction could be made between food that fed guerrillas and food that fed civilians” (Gill, 2010: 68). Although a violation of the sovereignty of the then Ethiopian government, the United States began offering food aid via Sudan to the rebel controlled territory, who were actively fighting against the Soviet-influenced Derg government. The decision to send food aid to rebel held territory was a political one, and it is political interests such as these that can influence how and why humanitarian and development activity occurs. Political neutrality is an ideal many adopt, but it is 	   65 complicated in practice (Carothers and de Gramont, 2013; Donini, 2012). Zinn (2002) argues that action and non-action speak volumes with regard to political positions and priorities. The support of TPLF fighters, as opposed to other actors, or the sovereign Ethiopian government of the day, is an example of the political nature of development priorities and the rationale that drives its practice. The influence of political objectives upon development is not limited to the way in which international agencies and/or states interact with one another; the politics of power also influences action within the state. The 1999-00 and 2002-03 famines in Ethiopia occurred in minority-dominated regions, which have been given far fewer resources than other regions (Lautze and Maxwell, 2007) and largely remain excluded from programs and services, such as the Productive Safety Net Program (FAO, 2008a). During the Derg government, relocation was used as a tool to control the eastern regions of Ethiopia and, in the case of government-promoted resettlement in the famine-hit north, food aid was withheld due to a lack of “volunteers” for the program (Terry, 2002). The motivations that drive the Government of Ethiopia may be purposefully hidden behind the stated objectives. For example, policies might be designed to appease international donors, with little intention to see them through (Andrews, 2013). Programs may be well designed, but implemented to meet unstated objectives (Berhanu and Poulton, 2014; Cochrane and Tamiru, 2016; Planel, 2014). According to the late Prime Minister of Ethiopia, Meles Zenawi, who led the country for two decades, democracy can hinder development (Zenawi, undated). The vision of development and democracy of Meles is reflected in the ways programs and policies have been waged in the name of development, have opposed democratic processes and inhibited the transition to inclusive institutions (Abbink, 2006; Kebede, 2013; Tronvoll, 2010).  This research project begins its assessment with an understanding of the diverse, and at times divergent, meanings of development being employed. With this foundation, it is insufficient to explore the relationship between food insecurity and land size, examinations must also assess who is gaining access to government supported services. For example, the findings of this research show that those who have not been trained by an agricultural extension worker have an average land size of 1.2 temut while those who 	   66 have been trained have an average land size of 1.6 temut. Those gaining access to fertilizer have significantly larger average land holdings than those who do have access (1.5 temut versus 0.7). Similarly, those who have gained access to government supported and distributed improved seed have larger average land holdings (1.6 temut versus 0.9). As Tefera (2015) has explained, the impact of agricultural extension has been the marginalization of poor, rural households, despite being created with the stated intention to do otherwise. It comes as no surprise that, in an assessment of satisfaction, forty five percent of farmers said they were dissatisfied with this freely provided service (Elias et al, 2015). How is it that these services, which are supposed to be offered equally and in the name of development, disproportionately benefit those with more assets and exclude those who most need them? The answer, I argue, is that any framing of development activity be understood within the context of politics and power.  3.2 POWER AND POLITICS  The expression of power in order to maintain political control can take the form of physical force. The Government of Ethiopia has frequently utilized this tactic. For example, it is estimated that as many as 500 protests occurred in response to a federal proposal to expand city administration planning into Oromia Regional State between November 2015 and March 2016. The government responded with lethal force and mass arrests, resulting in up to 400 protester deaths (HWR, 2016). In July and August of 2016 there were large-scale protests about the rezoning of districts from Amhara Regional State to Tigray Regional State, resulting in 97 protester deaths (Amnesty International, 2016). Between 2011 and 2014 regular protests occurred in response to the government seeking to mandate individuals for religious leadership positions. In response to these protests, the government conducted mass arrests with reports of mistreatment of those detained (HRW, 2012b). Prominent leaders of the Muslim community were detained without charge in 2012 and held until convicted under anti-terrorism legislation in 2015 with sentences ranging from seven to twenty two years (Fasil, 2015). Until recently, the 	   67 heavy-handed use of force has restricted collective action. However, as protests expand in scope, scale and frequency, and as divergent protest groups begin to work together, the government appears to recognize this strategy will no longer be viable. For example, it has called upon diasporic and domestic opposition groups to meet at the African Union (Addis Standard, 2016). While force is an important component as an expression of power, there are other means which are less confrontational but can be sustained and transformational. The expression of power and control can be ideological (Gramsci, 1971). Expressed in this form, power can be normalized within mundane, regular activities and practices (Foucault, 1977). The concept of governmentality was proposed and developed by Foucault (1979) to assess how power is expressed and control established. Foucault describes governmentality as the “ensemble formed by the institutions, procedures, analyses and reflections, the calculations and tactics, that allow the exercise of this very specific albeit complex form of power” (1979: 20). Drawing upon governmentality as a framework for assessing power, the (non)actions, policies, programs and statements of a government can be evaluated as a means to shape individuals within society to align with the government’s objectives. Governing can therefore be viewed as an assertion of power and an exercise of control. The ways in which programs have been used to service government control and strengthen elite power have been well documented in the development literature (de Waal, 2015; Ferguson, 1990; Li, 2007; Scott, 1985). However, these studies have tended to be anthropological in nature, and limited progress has been made with regard to integrating these perspectives and findings into development practice, or confronting development practice itself (Carothers and de Gramont, 2013). A number of researchers have reflected on how development actors have been unable or unwilling to engage with the politics of power (Autessere, 2010; Starn, 1991; Uvin, 1999). This study approaches the questions of vulnerability to food insecurity and low adoption rates while remaining cognizant of the ways in which development is politicized, acting with dual purposes of achieving a particular development outcome as well as an expression of power and establishment of control. It is informed by the fact that development has been used as a 	   68 means to centralize control and that much activity done in the name of development was done to achieve alternative objectives (Uvin, 1999). Even if the politicization of policies and programs becomes normalized and routine, in the long-term it can foster opposition, and have negative impacts on the wealth generation that the elites are attempting to capture. Acemoglu and Robinson argue that politicization and patronage of this nature can create a negative cycle: When extractive institutions create huge inequalities in society and great wealth and unchecked power for those in control, there will be many wishing to fight to take control of the state and institutions. Extractive institutions then not only pave the way for the next regime, which will be even more extractive, but they will also engender continuous infighting and civil wars. These civil wars then cause more human suffering and also destroy even what little state centralization these societies have achieved. This also often starts a process of descent into lawlessness, state failure, and political chaos, crushing all hopes of economic prosperity (2012: 366-367). Although the above prediction is slightly too deterministic, the historical study conducted by Acemoglu and Robinson (2012) demonstrates that this cycle has often repeated. The World Bank has begun to advocate for citizen power to “select and sanction leaders who have the political will and legitimacy to delivery public goods needed for development” as opposed to the traditional development assistance, which “can contribute to the persistence of government failures” (Devarajan and Khemani, 2016: 1). The forthcoming flagship World Bank report, World Development Report 2017: Governance and the Law, will further highlight the need for greater reflexivity on power within state-citizen relations, and those mirrored or reinforced by development activity. Viewing political action from a perspective that is attentive to power in Ethiopia assists in answering the above-raised question: how and why is it that the services designed to support all people, or the most vulnerable, disproportionately benefit the relatively better off? One explanation is that the programs and services are politicized, expressing power and asserting control. Within Ethiopia, Berhanu and Poulton (2014) and Planel (2014) 	   69 have found this to be the case in the implementation of the agricultural extension program. Chinigo (2013) has also found it in rural land reform, Cochrane and Tamiru (2016) have identified it within the Productive Safety Net Program, and de Waal (2015) notes that the politicization of access to services and the provision of goods has been commonplace in rural Ethiopia for decades. In fact, the roots can be traced much deeper, to the Imperial period, as noted by Acemoglu and Robinson (2012: 358, 361), highlighting the broader historical context within which these practices, and cycles of practices, exist. This broader contextualization of qualitative and historical knowledge was emphasized by Mintz (1985), and to the extent possible, has been integrated into this work. Understanding the ways in which power and control are embedded within development activity is not only a matter of improving implementation or enhancing effectiveness, as Acemoglu and Robinson (2012) point out. The current situation within Ethiopia resembles that of Rwanda before the genocide, which Uvin described as: “Ethnic inequality; institutionalized, state-organized racism; regional politics; lack of dignity and self-respect; the generalized presence of impunity and fear of the absence of justice; human rights violations; the oppressive presence of the state” (Uvin, 1999: 45). Worrisomely, the trajectory is similar: agricultural crises, followed by economic crises and then political crises and “a rise of political discontent within the country” (Uvin, 1999: 53). In August 2016 an opposition politician outlined that the rising tensions, increasing frequency and scale of protests, and rise of anti-government, often ethnic-based, sentiment could result in civil war (Hayden, 2016). Former opposition leader Berhanu Nega, and current leader of a rebel movement based in Eritrea that is fighting the Government of Ethiopia, anticipates that the resistance movement will topple the current government within the coming decade (Hammer, 2016). Addressing these concerns is crucial for the future stability of Ethiopia (Adugna, 2011; Feyissa and Lawrence, 2014; Tache and Oba, 2009). The participatory methodology of this study has been designed to contribute, however modestly, to that goal.  Just as Chapter 2 progressively narrowed down in scale in presenting the research context (form national to regional, zonal and district), Chapters 3 and 4 narrow down the 	   70 theoretical foundations of this research. This chapter analyzed the broad concept of development and how activities done in the name of development interact with politics and power. Chapter 4 looks specifically at the concept of food security, its framing and its measurement. After analyzing the broad (Chapter 3) and specific (Chapter 4) conceptual components of this research, the final section of Chapter 4 outlines the theoretical foundations that I draw upon and utilize, while attempting to explore my biases in a process of critical reflexivity.    	   71 CHAPTER 4. ON FOOD SECURITY   Food insecurity is a complex, global chal