University of Ghana http://ugspace.ug.edu.gh THE EFFECTS OF OUTGROWER SCHEME ON LIVELIHOODS OF SMALLHOLDER SORGHUM FARMERS IN NORTHERN GHANA BY CHARLES KWOWE KWAME NYAABA (10169660) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT OF REQUIREMENT FOR THE AWARD OF DOCTOR OF PHILOSOPHY DEGREE IN AGRIBUSINESS DEPARTMENT OF AGRICULTURAL ECONOMICS AND AGRIBUSINESS COLLEGE OF BASIC AND APPLIED SCIENCES UNIVERSITY OF GHANA, LEGON JULY, 2019 University of Ghana http://ugspace.ug.edu.gh DECLARATION I, Charles Kwowe Kwame Nyaaba, do hereby declare that except for the references cited, which have been duly acknowledged, this thesis titled, “The Effects of Outgrower Scheme on the Livelihoods of Smallholder Sorghum Farmers in Northern Ghana” is the result of my own research. This thesis has never been presented either in whole or in part for any other degree of this University or elsewhere. i University of Ghana http://ugspace.ug.edu.gh DEDICATION This thesis is dedicated to the members of Peasant Farmers Association of Ghana (PFAG). ii University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENTS I thank the Almighty God for giving me good health and protection throughout the four years of my studies. My profound gratitude goes to my supervisory team comprising of Prof. Daniel Bruce Sarpong, Prof. Irene S. Egyir and Dr. Freda Asem for their scholarly advice. I also wish to acknowledge all lecturers at the Department of Agricultural Economics and Agribusiness for their comments and suggestions that helped to refine the thesis. The thesis would not have been possible without funding from the Africa Climate Change Adaptation Initiative (ACCAI), Pan-Africa Doctoral Academy (PADA) and Danida Fellowship Centre (DFC). I am especially indebted to Prof. Yaa Ntiamoa-Baidu, the Director of PADA and Dr. Naalamle Amissah from ACCAI for their professional guidance. I also wish to acknowledge the contributions of my PhD colleagues from the Department of Agricultural Economics and Agribusiness during seminars and daily discussions. Mr. Dinko Hanaan Dinko and Prof. Joseph Awetori Yaro equally played important roles by providing technical support, I am most grateful. For the smallholder farmers I collected the data from and those I work with, I thank all of them for their cooperation. My special thanks go to the National President of the Peasant Farmers Association Ghana (PFAG) (Mr. Abdul-Rahman Mohammed) for his encouragement and moral support. The Executive Director of PFAG, Madam Victoria Adongo has contributed so much to my career development more than I could express in this thesis. May the Almighty God continue to guide and protect her. To all staff of PFAG, I am grateful for their moral support. Nobody has ever been so important in the pursuit of my career development than the members of my family. I thank my mother whose love and guidance is always with me in whatever decision I make in life. Most importantly, I thank my loving and supportive wife, Reina Dam-Balagwor, my two wonderful daughters Wesom and Chagewe and My son Wesei for being my inspiration. iii University of Ghana http://ugspace.ug.edu.gh ABSTRACT Outgrower scheme (OGS) is widely articulated as an ideal option that can deal with subsistence farming practices of smallholder farmers (SHF) to approach their farming as a business. For OGS to attract SHF participation and lead to livelihoods enhancement, this study argues for strengthening extension services, guaranteed market and promotion of FBO formation as part of the OGS support to farmers. The study also advocates for integration of climate change mitigation services as part of the OGS package. The study combined quantitative and qualitative research methods to analyse the effects of OGS on the livelihoods of smallholder sorghum farmers in Northern Ghana. Specifically, the study examines factors influencing SHF participation in the OGS, the effects of OGS on their productivity, profitability, postharvest loss (PHL) and their vulnerability to climate change. The multistage sampling procedure was used to collect quantitative data from 516 sorghum outgrower farmers (treatment) and non-outgrower farmers (control) groups in Garu and Jirapa districts in the Upper East and Upper West regions of Ghana respectively. Using the probit regression model to determine factors influencing SHF participation in OGS, the results pointed to belonging to FBO, access to market and access to extensions services as key determinants. The study also found average productivity of 1,207kg/ha, profitability of GHS 270/ha and post-harvest losses (PHL) of 14% for the treatment group. For control group, the average productivity was 820kg/ha, profit losses of GHS 92/ha and PHL of 27%. The study further found the treatment group to be relatively vulnerable to climate change than the control group with their overall aggregate livelihood vulnerability index (LVI) of 0.393 and 0.386 respectively. (LVI closer to 1 denotes highly vulnerable). Using endogenous switching regression model (ESRM) to establish treatment effect of OGS on SHF, the results suggest positive effects of OGS on productivity, PHL and profitability of resourced endowed farmers than ordinary SHF. On vulnerability to climate change, participation in OGS have minimal effect of climate change on SHF in the study areas. To stimulate SHF participation in OGS, the study recommend improvement in market access, extension services and establishing and strengthening the existing FBOs. Finally, to help improve SHF productivity, reduce their PHL and increase their profitability, the study recommends modification of the current OGS to make it more pro- poor and also, policies that will incentivize private sector to engage SHF on OGS that are pro-poor. For OGS to become more sustainable and contribute to reducing SHF vulnerability to climate change, the study suggests inclusion of climate change support variables as part of the OGS support to farmers. iv University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENTS DECLARATION ................................................................................................................. i DEDICATION .................................................................................................................... ii ACKNOWLEDGEMENTS .............................................................................................. iii ABSTRACT ...................................................................................................................... iv TABLE OF CONTENTS ................................................................................................... v LIST OF TABLES .............................................................................................................. x LIST OF FIGURES .......................................................................................................... xii LIST OF ACRONYMS .................................................................................................. xiii CHAPTER ONE ................................................................................................................. 1 INTRODUCTION .............................................................................................................. 1 1.1 Background of the Study .............................................................................................. 1 1.2 Problem Statement ........................................................................................................ 4 1.3 Objectives of the Study ............................................................................................... 11 1.4 Hypothesis of the Thesis ............................................................................................. 11 1.5 Relevance of the Study ............................................................................................... 11 1.6 Organisation of the Thesis .......................................................................................... 13 CHAPTER TWO .............................................................................................................. 14 LITERATURE REVIEW ................................................................................................. 14 2.1 Introduction ................................................................................................................. 14 2.2 Background of Sorghum Production .......................................................................... 14 2.2.1 History of Sorghum Outgrower Schemes in Ghana ......................................... 16 2.3 Agriculture and Economic Development in Africa .................................................... 17 2.3.1 Agriculture and Economic Development in Ghana ......................................... 19 2.3.2 Food and Agricultural Policies in Ghana ......................................................... 22 2.3.3 Medium Term Agricultural Sector Investment Plans in Ghana ....................... 24 v University of Ghana http://ugspace.ug.edu.gh 2.4 Ghana Government Agricultural Flagship Programmes ............................................ 25 2.5 Concept of Smallholder Farmers ................................................................................ 27 2.5.1 Smallholder Farmers Crop Productivity and Profitability ............................... 29 2.5.2 Postharvest Loss Among Smallholder Farmers ............................................... 31 2.5.3 Smallholder Farmers and Climate Change ....................................................... 35 2.5.4 Smallholder Farmers Vulnerability to Climate Change ................................... 37 2.6 Outgrower Scheme Debate ......................................................................................... 39 2.6.1 Types of Outgrower Schemes .......................................................................... 43 2.7 Summary and Knowledge Gap in the Literature ........................................................ 44 CHAPTER THREE .......................................................................................................... 46 METHODOLOGY ........................................................................................................... 46 3.1 Introduction ................................................................................................................. 46 3.2 Conceptual Framework of the Study .......................................................................... 46 3.3 Analytical Framework ................................................................................................ 50 3.4 Theoretical Framework for Factors Influencing Participation in Outgrower Scheme 52 3.5 Theoretical Framework for Effects of Sorghum Outgrower Scheme on Productivity, Postharvest Loss and Profitability ............................................................................. 54 3.5.1 Estimation of Treatment effects of Outgrower Scheme on Productivity, Postharvest Loss and Profitability .................................................................... 56 3.5.1.2 The Endogenous Switching Regression Estimation ...................................... 60 3.6 The Effects of Sorghum Outgrower Scheme on Vulnerability to Climate Change ... 64 3.7 Method of Data Analysis ............................................................................................ 65 3.7.1 Background of the Study Respondents ............................................................ 66 3.7.2 Factors Influencing Smallholder Farmers Participation in Outgrower Scheme .......................................................................................................................... 66 3.7.3 The Effects of Outgrower Scheme on Productivity, Postharvest Loss Reduction and Profitability .............................................................................. 73 vi University of Ghana http://ugspace.ug.edu.gh 3.7.4 Determining the effects of Outgrower Scheme on Smallholder Farmers Vulnerability to Climate Change ..................................................................... 77 3.8 Method of Data Collection ......................................................................................... 83 3.8.1 Sources of Data, Instruments and Interview Procedure ................................... 83 3.8.2 Sampling Procedure ......................................................................................... 84 3.8.3 Interview Procedure ......................................................................................... 87 3.9 Study Area .................................................................................................................. 91 3.10 Scope and Limitations of the Study .......................................................................... 94 CHAPTER FOUR ............................................................................................................ 95 RESULTS AND DISCUSSION ....................................................................................... 95 4.1 Introduction ................................................................................................................. 95 4.2 Background of the Study Respondents ....................................................................... 95 4.2.1 Household Characteristics of the Respondents ................................................ 95 4.2.2 Farm Characteristics of the Respondents ......................................................... 99 4.3. Socio-economic and Political Characteristics of the Respondents .......................... 102 4.4 Description of the Sorghum Outgrower Scheme in the Study Area ......................... 105 4.4.1 Type of Contracts Between Farmers and Buyers ........................................... 106 4.4.2 Perception of Smallholder Farmers on Price Determination .......................... 107 4.4.3 Kind of Support Received by Farmers from Buyers ...................................... 108 4.4.4 Specific Support Targeting Smallholder Farmers Vulnerability to Climate Change ............................................................................................................ 109 4.5 Factors influencing Smallholder Farmers Participation in Sorghum Outgrower Scheme .................................................................................................................... 111 4.6. Determining the Effects of Outgrower Scheme on Smallholder Farmers Productivity ................................................................................................................................. 116 4.6.1 Productivity Analysis ..................................................................................... 116 4.6.2 Average Treatment Effect on Sorghum Productivity ..................................... 118 vii University of Ghana http://ugspace.ug.edu.gh 4.7 Determining the Effects of Outgrower Scheme on Smallholder Farmers Profitability ................................................................................................................................. 123 4.7.1 Profitability Analysis ...................................................................................... 123 4.7.2 Average Treatment effects on Profitability .................................................... 123 4.8 Determining the effects Outgrower Scheme on Smallholder Farmers Postharvest Loss Reduction ........................................................................................................ 126 4.8.1: Analysis of Postharvest Situation in the Study Area ..................................... 126 4.8.2 Average Treatment Effects of Sorghum Outgrower Scheme on Postharvest Loss Reduction ............................................................................................... 130 4.9 Determining the Vulnerability Level of Smallholder Farmers to Climate Change .. 135 4.9.1 Results of Vulnerability of Smallholder Farmers to Climate Change ............ 135 4.9.2 Average Treatment effects on Smallholder Farmers Vulnerability to Climate Change ............................................................................................................ 142 4.9.3 Summary of Smallholder Farmers Vulnerability to Climate Change ............ 146 CHAPTER FIVE ............................................................................................................ 148 SUMMARY, CONCLUSION AND RECOMMENDATION ....................................... 148 5.1 Introduction ............................................................................................................... 148 5.2 Summary and Major Findings .................................................................................. 148 5.3. Recommendation ..................................................................................................... 152 REFERENCES ............................................................................................................... 155 viii University of Ghana http://ugspace.ug.edu.gh APPENDICES ................................................................................................................ 173 Appendix 1: Questionnaire ............................................................................................. 173 Appendix 2: Interview Guide ......................................................................................... 189 Appendix 3: Postharvest Loss Along the Various Postharvest Chain ............................ 194 Appendix 4: Propensity Score Matching ........................................................................ 198 Appendix 5: LVI Results in Percentage for Control and Treatment .............................. 201 ix University of Ghana http://ugspace.ug.edu.gh LIST OF TABLES Table 1.1: Actual and Achievable Productivity of Major Staple Crops in Ghana ............. 2 Table 2.1: Productivity of Major Staple Crops in Ghana ................................................. 30 Table 2.2: Trends of Productivity in Crops (MT/Ha) in Ghana ....................................... 31 Table 2.3: Mean Annual Change in Rainfall (%) in Ghana ............................................. 36 Table 2.4: Projected Mean Annual Temperature Changes in Ghana ............................... 36 Table 3.1: Treatment Effect .............................................................................................. 62 Table 3.2 :Variables Influencing Participation in Sorghum Outgrower Scheme ............. 69 Table 3.3: Major and Sub-Component for Natural Disasters and Climate Change ......... 79 Table 3.5: Sampling Procedure ......................................................................................... 86 Table 3.6: Communities and Number of Sorghum Farmers Sampled .............................. 87 Table 4.1: Household Characteristics of the Respondents ............................................... 96 Table 4.2. Farm Characteristics of the Respondents ...................................................... 100 Table 4.3: Market, Community and Socio-economic and Political Characteristics ....... 102 Table 4.4: Kind of Support Received by the Sorghum Farmers from Buyers ............... 108 Table 4.5: Kind of Support on Climate Resilient Received by Farmers ....................... 110 Table 4.6: Factors Influencing Participation in the Sorghum Outgrower Scheme ......... 112 Table 4.7: Comparison of Productivity, Revenue and Costs Across Treatment and Control Groups ............................................................................................. 117 Table 4.8: Results of Full Information Maximum Likelihood Estimates of Endogenous Switching Regression on Productivity .......................................................... 119 Table 4.9: Endogenous Switching Regression (ESR) Results on Productivity .............. 121 Table 4.10: Propensity Score Matching Results on Farm Productivity.......................... 122 Table 4.11: Full Information Maximum Likelihood Estimates Results on Profitability. ...................................................................................................................... 124 x University of Ghana http://ugspace.ug.edu.gh Table 4.12: Endogenous Switching Regression Results on Profitability ....................... 125 Table 4.13: Propensity Score Matching Results on Profitability ................................... 126 Table 4.14: Postharvest Loss for Treatment and Control Groups .................................. 127 Table 4.15: Full Information Maximum Likelihood Estimates Results on Postharvest Loss Reduction ............................................................................................. 132 Table 4.16: Endogenous Switching Regression Results on Postharvest Loss Reduction ...................................................................................................................... 133 Table 4.17: Propensity Score Matching Results on Postharvest Loss Reduction .......... 134 Table 4.18: Indexed Sub and Major Component of LVI for Water, Socio-demographic, Food and Social Network ............................................................................. 136 Table 4.19: Indexed Sub and Major Component of LVI for Livelihood Strategy, Natural Disaster and Climate Change and Health ..................................................... 138 Table 4.20: LVI-IPCC Contribution Factors to Climate Change ................................... 139 Table 4.21: Full Information Maximum Likelihood Estimates of effects of Sorghum Outgrower Scheme on Reducing Vulnerability to Climate Change ............. 142 Table 4.22: Endogenous Switching Regression Results of Treatment effects of Sorghum Outgrower Scheme on Vulnerability to Climate Change ............................. 144 Table 4.23: Treatment effects on Vulnerability to Climate Change ............................... 145 xi University of Ghana http://ugspace.ug.edu.gh LIST OF FIGURES Figure 1.1: Trend of Sorghum performance for the Period of 2005 to 2016 ...................... 7 Figure 2.1: Agricultural Sub-Sectors in Ghana ................................................................ 20 Figure 2.2: Real Annual GDP Growth Rate of the Agricultural Sector 2010 – 2017 ...... 22 Figure 3.1: Conceptual Framework of Sorghum Outgrower Scheme in Ghana ............... 49 Figure 3.2: Sustainable Livelihood Framework ............................................................... 51 Figure 3.3: Map of the Study Area ................................................................................... 93 Figure 4.1: Existing Contracts Between Farmers and Buyers ........................................ 106 Figure 4.2: Perception of Farmers on How Prices Price Determination ........................ 107 Figure 4.3: Percentage Losses in the Various Postharvest Loss stages .......................... 128 Figure 4.4: Vulnerability Triangle Diagram of LVI-IPCC for Control and Treatment.. 140 Figure 4.5: Vulnerability Spider Diagram of the Major Components of the LVI for Control and Treatment Groups ..................................................................... 141 xii University of Ghana http://ugspace.ug.edu.gh LIST OF ACRONYMS 1D1F One District One Factory 1D1W One District One Warehouse 1V1D One Village One Dam AAGDS Accelerated Agricultural Growth and Development Strategy AGRA Alliance for Green Revolution in Africa ATT Average Treatment Effect on the Treated ATU Average Treatment Effect on the Untreated CAADP Comprehensive Africa Agriculture Development Programme CAPI Computer-Assisted Personal Interviewing ECOWAP Economic Community of West African States Regional Agricultural Policy ECOWAS Economic Community of West African States EPA Environmental Protection Agency ESR Endogenous Switching Regression ESRM Endogenous Switching Regression Model EUCORD European Cooperation for Rural Development FAO Food and Agriculture Organisations of United Nations FASDEP I First Food and Agricultural Sector Development Policy one FASDEP II Second Food and Agricultural Sector Development Policy II FIML Full Information Maximum Likelihood Estimate GCAP Ghana Commercial Agricultural Programme GDP Gross Domestic Product GGBL Guinness Ghana Brewery Limited GhAIP Ghana Agricultural Investment Plan GLSS Ghana Living Standard Survey GOPDC Ghana Oil Palm Development Company Limited GPRS I First Ghana Poverty Reduction Strategy GSGDA Ghana Shared Growth and Development Agenda GSS Ghana Statistical Service IPCC Inter-Governmental Panel on Climate Change xiii University of Ghana http://ugspace.ug.edu.gh IPWRA Inverse Probability Weighted Ratio Adjusted ISSER Institute of Statistical, Social and Economic Research MDGs Millennium Development Goals METASIP I First Medium-Term Agricultural Sector Investment Plan METASIP II Second Medium-Term Agricultural Sector Investment Plan MoFA Ministry of Food and Agriculture NEPAD New Partnership for Africa Development NGO Non-Governmental Organisations OGS Out Grower Scheme OLS Ordinary Least Square PERD Planting for Export and Rural Development PFJ Planting for Food and Jobs PHL Post Harvest Loss PSIA Poverty and Social Impact Analysis PSM Propensity Score Matching SARI Savanna Agriculture Research Institute SDGs Sustainable Development Goals SHF Small Holder Farmer SLF Sustainable Livelihood Framework SOGS Sorghum Out Grower Scheme SSA Sub-Saharan Africa TH Transitional Heterogeneity TZ Tuozafi UNDP United Nations Development Programme UNEP United Nations Environment Programme WASVCD West African Sorghum Value Chain Development Project xiv University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE INTRODUCTION 1.1 Background of the Study There is global consensus on agricultural contribution to poverty reduction, job creation and overall economic development (Alexandratos & Bruinsma, 2012; Schaffnit- Chatterjee, 2014). The Food and Agriculture Organisations of the United Nations (FAO) (2018) identified strategic agricultural investment in Sub-Saharan Africa (SSA) as an important option that can increase incomes of smallholder farmers (SHF), reduce their poverty level and at the same time, guarantee enough food for the projected global population of 9.8 billion people by 2050. In analysing challenges and opportunities in agricultural investment in SSA, Saghir & Hoogeveen (2017) found effects of investing in the agricultural sector as eleven times effective in reducing poverty than similar investments in the other sectors of the SSA economy. In Ghana, the agricultural sector play a leading role in the overall economic development. The sector provides jobs for about 38.3% of the population and remain the major sources of income for majority of low income earners, largely, those in the rural areas (Ghana Statistical Service (GSS), 2019). According to GLSS 7 report, about 63.3% of rural folks are engaged as skilled agricultural workers compared to 11.4% for those in the urban areas. Several literature identified SHF constraints such as difficulty in accessing financial support; access to mechanization services; reliance on outmoded method of farming; poor extension services; poor market and storage infrastructure; high postharvest losses (PHL); limited irrigation facilities and effects of climate change as a major challenge for SHF development in Ghana (Awunyo-Victor & Al-hassan, 2014; Boateng & Nyaaba, 2014; Dittoh & Akuriba, 2018; FAO, 2017; GSS, 2014d; Intergovernmental Panel on Climate Change (IPCC), 2013; MoFA, 2017b; Villano, Asante, & Bravo-Ureta, 2019). 1 University of Ghana http://ugspace.ug.edu.gh Smallholder farmers productivity in Ghana as is the case for most SSA countries is generally low, production is limited to home consumption and surpluses for the market (Ecker, 2018). They are classified among the poor in the country (GSS, 2019). To improve the livelihoods of SHF will require a strategy that will transform their current subsistence farming practices to approach their farming as a business. Ghana government response to addressing SHF constraints is articulated in the Food and Agricultural Sector Development Policy I and II (FASDEP I &II) and associated first and second Medium-Term Agricultural Sector Investment Plans (METASIP I and II) (Allience for Green Revolution in Africa (AGRA), 2016; Ministry of Finance, 2018; MoFA, 2017b). Some programmes in METASIP II implemented in support of SHF include: Fertilizer and Seed Subsidy Programme, Creation of Agricultural Mechanization Centres and establishment of National Food Buffer Stock Company (Fearon & Adraki, 2015; Institute of Statistical, Social and Economic Research (ISSER), 2012; World Bank, 2012). With all these interventions, data from the Ministry of Food and Agriculture (MoFA) shows little improvement in SHF performance which reflect the stagnation of productivity of various staples in Ghana (Ministry of Food and Agriculture (MoFA), 2017a). Table 1.1 shows actuals yields and achievable yields of some selected staple crops of SHF in Ghana. Table 1.1: Actual and Achievable Productivity of Major Staple Crops in Ghana Staple crops Achievable yield Actual yield % Achieved (MT/ha) (MT/ha) Maize 5.50 2.05 37.27 Rice (Paddy) 6.00 3.01 50.17 Cassava 45.00 20.68 45.96 Yam 52 16.74 32.19 Sorghum 2.00 1.24 62.00 Soybeans 3.00 1.70 56.67 Source: MoFA, 2017 2 University of Ghana http://ugspace.ug.edu.gh 1.1.1 Smallholder Farmers and Sorghum Production in Ghana The concept of SHF varies from different literature. From a general standpoint, SHF are farmers operating under structural constraints such as access to sub-optimal amounts of resources, technology and markets (Dittoh & Akuriba, 2018; Dixon et al., 2004) summarize this idea when they say that “the term smallholders refers to limited resource endowment of farmers compared to other farmers in the sector”. In Ghana, Ministry of Food and Agriculture categorised SHF as those cultivating a land area below two hectares (MoFA, 2016). For the purposes of this study, sorghum smallholder farmers are sorghum farmers who cultivate less than two hectares, are constraint with production and marketing resources and hardly access quality extension services. Sorghum is a staple crop predominantly cultivated by SHF in Northern Ghana. It is largely grown in the Sudan and Guinea Savanna agro-ecological zones of Ghana (Akuriba & Asuming-Brempong, 2012). Sorghum farming is said to be convenient due to its low inputs’ requirements, less laborious and has the ability to withstand the unfriendly weather condition in Northern Ghana. Sorghum can be grown in any marginal lands without heavy fertilizer application (Angelucci, 2013). Every part of the sorghum plant has economic value. Sorghum can be used for preparation of light porridge and tuozafi (TZ). Tuozafi has become a staple food in Ghana, especially among the indigenes of Northern Ghana. The flour of sorghum serves critical needs such as curing illnesses and feeding lactating mothers. Sorghum leaves and stalks are used for fencing, weaving baskets, roofing, mat making and also fuel for cooking in most rural areas in northern Ghana (Angelucci, 2013; Ratnavathi et al., 2016). The fresh sorghum leaves and stalks as well as processed sorghum grains are good for feeding livestock (Angelucci, 2013). During festivals, funerals and other formal and informal social 3 University of Ghana http://ugspace.ug.edu.gh gatherings, sorghum artisanal beer popularly called pito is widely consumed in Ghana (Djameh et al., 2015). Guinness Ghana Brewery Limited (GGBL) has discovered sorghum as perfect substitute for barley since 2006 (Angelucci, 2013; Guinness Ghana Brewery Limited (GGBL), 2017). According to GGBL (2017), with better production and marketing arrangement, sorghum farming can create jobs in northern Ghana and subsequently, lead to reduction of migration of the youth to southern Ghana in search of non-existing jobs. Sorghum has been substituted for barley by GGBL for the brewing of alcoholic and non- alcoholic drinks. According to GGBL (2017), the company policy to source local raw materials including sorghum for their beverage production is expected to create market, create jobs and alleviate poverty among the producers who are prodominately SHF. Subsitituting imported barley to sorghum has also reduced the import bill and contributed to stabilizing the local currency (GGBL, 2017; Sarfo-Mensah, 2017) 1.2 Problem Statement Agriculture play significant role in socio-economic development of Ghana contributing about 21.2 per cent of gross domestic product (GDP) and providing a major source of income and employment for most households (GSS, 2019; Ministry of Finance, 2019). Ghana’s agricultural sector just like other SSA countries, is dominated by SHF who constitute over 80 percent of food crop farmers (Villano et al., 2019). Their land holdings are between one to two hectares (SRID, 2017). Most of them still depend on outmoded technologies in farming (AGRA, 2017; Chauvin et al., 2012; Ecker, 2018). Other challenges are their inability to access quality inputs, credit and extension services leading to low crop yields, low productivity reflecting in their low profitability. Smallholder 4 University of Ghana http://ugspace.ug.edu.gh farmers are also challenge with guaranteed market and profitable prices (Mariano et al., 2012; Tsinigo & Behrman, 2017; Al-Hassan et al., 2013; Ecker, 2018). Poor agronomic practices coupled with poor postharvest facilities, smallholder farmers are also face with high PHL (Kiaya, 2018). The high PHL is usually due to poor postharvest facilities such as, harvesters, dryers, appropriate transport facilities and storage infrastructure (Affognon et al., 2015; Sheahan & Barrett, 2016). Aside the above challenges, access to market is a major challenge facing several farmers in Ghana including those cultivating sorghum. Availability of market has become a key determinant of the choice of crop produced (Ebata & Hernandez, 2017; Opoku, 2012). In recent times, sorghum cultivation has attracted the interest of many SHF in northern Ghana due to huge demand of sorghum from GGBL for brewing of alcoholic and non-alcoholic beverages (Angelucci, 2013; Sarfo-Mensah, 2017). Whiles GGBL marketing arrangement contributed to addressing marketing challenges of sorghum in Ghana, the call for provision of market for sorghum is still high due to GGBL targeting farmers who produced under GGBL promoted outgrower schemes (Angelucci, 2013). According to Sarfo-Mensah 2017, “some sorghum farmers are still unable to access market from GGBL due to their inability to produce quality sorghum that meets GGBL quality requirements”. Access to market is not only available market, but also market that offers profitable prices (Ebata & Hernandez, 2017) Similar to the general problems confronting other food crop producers, sorghum farmers have to also struggle to access optimal inputs, quality extension services, quality postharvest facilities, credit and also, suffer with poor weather conditions (Azumah et al., 2016; Tsinigo & Behrman, 2017; Udimal et al., 2017). 5 University of Ghana http://ugspace.ug.edu.gh Majority of sorghum farmers use poor quality seeds for planting and hardly follow good agronomic practices leading to low crop yields. Harvesting is done manually using simple farm tools such as hoes and cutlasses. The harvested grains in most occasions are dried on the field and threshed manually (Akuriba & Asuming-Brempong, 2012; Statistics, Research and Information Directorate of Ministry of Food and Agriculture (MoFA-SRID), 2016). These practices apart from reducing the yield quality and quantity, also lead to high postharvest losses. In Ghana, high PHL loss, especially aflatoxin contamination in sorghum is of major concern. Available statistics put PHL in sorghum between 5% to 15% ( High Level Panel of Expert on Food Security and Nutrition (HLPE), 2014; MoFA, 2016). Another constrain to sorghum production is the effect of climate change. Intergovernmental Panel on Climate Change opines that climate related hazards affects the livelihoods of rural poor directly by affecting their crop yields and worsening their food insecurity situation (Intergovernmental Panel on Climate Change (IPCC), 2018). The changes are in the form of low and unpredicted rainfall, rising temperatures, flooding, drought and emergence of pest and diseases (Adu et al., 2018; Dickinson et al., 2017; IPCC, 2018; Jamshidi et al., 2019; Makate et al., 2019; Makuvaro et al., 2018; Shah et al,. 2013; Yaro, 2013). These challenges have drastically affected smallholder farmers productivity, PHL and their profitability. The combined effects are reduction in sorghum production, slow growth of sorghum production and reduction in sorghum contribution to agricultural GDP (MoFA-SRID, 2016). Figure 1.1 shows the trend of sorghum production, area planted, annual production and annual growth for the period of 2005 to 2016. The area planted, annual production and annual growth has experienced increment from 2007 to 2008 and declined after 2008 (Figure 1.1). 6 University of Ghana http://ugspace.ug.edu.gh 400 350 300 250 200 Area plated (000ha 150 Annual production (000mt) 100 Annual growth rate (%) 50 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 -50 -100 Year Figure 1.1: Trend of Sorghum performance for the Period of 2005 to 2016 Sources: SRID (2017) Given the important role of sorghum to socio-economic development in Northern Ghana and also the consistent decline in the overall performance, any government policy direction to support sorghum production and marketing could contribute to overall wellbeing of farmers. some programmes implemented by governments over the years to modernise the activities of SHF are captured in FASDEP II and METASIP II (MoFA, 2015). These programmes include: Fertilizer and Seed Subsidy Programme implemented to reduce cost of fertilizer and seeds for easy access. The programmes were expected to lead to improvement in application of fertilizer and improved seeds (Fearon & Adraki, 2015), establishment of Agricultural Mechanization Centres expected to reduce difficulty of SHF farmers access to mechanization services, establishment of National Food Buffer Stock Company expected to provide guaranteed prices and market for grains, construction of warehouses in every district to help address the problem of access to storage facilities and instituting Ghana Commodity Exchange programme to provide storage facilities, credit, market access and profitable prices (MoFA, 2007, 2015, 2017a). 7 Sorghum prodcution in MT University of Ghana http://ugspace.ug.edu.gh In 2017, the government of Ghana took another bold step to support SHF by introducing Planting for Food and Jobs (PFJ). The PFJ provided farmers with subsidised fertilizer, seeds, extension services and market access with the aim of modernizing their activities, addressing food security constraint and creating jobs (MoFA, 2017b). Even though limited assessment reports on these interventions exist, there is no evidence of SHF modernization and increased in crop productivity in Ghana. Inability of government to modernize the activities of SHF and to alleviate them from their current subsistence farming to market-oriented agribusiness call for alternative approach in handling government interventions. Several studies shows evidence of potentials of OGS to modernize the activities of SHF and improve their production efficiency ( Minot, 2011; Agbelengor, 2015; Maertens & Velde, 2017; Ton et al., 2017; Ragasa et al., 2018) . The benefits of OGS ranges from helping SHF to access credit, provision of quality inputs (fertilizer and seeds), provision of mechanization services, provision of extension services and guaranteed market (Barrett et al., 2012; Minot & Ronchi, 2014; Minot, 2015; Maertens & Velde, 2017). The extension services as part of OGS package also provide farmers with climate information, early warning system and climate resilient crop varieties (Feleke et al., 2017; Ghimire et al., 2015; Uaiene et al., 2009). Regardless of the optimism about the prospects of OGS in modernizing the activities of SHF and transforming them to market oriented agribusiness entrepreneurs (Barrett et al., 2012; Maertens & Velde, 2017; Minot, 2015; Minot & Ronchi, 2014; Paglietti & Sabrie Roble, 2012), other studies such as Oya (2012); Wang et al. (2014), Otsuka et al. (2016) and Vicol (2017) are pessimistic about the real benefits of OGS to SHF and raised concerns of buyers unilaterally determining quality standards, prices, cost of inputs and repayment terms leading to low benefits and high exit rate among most schemes. 8 University of Ghana http://ugspace.ug.edu.gh Vicol (2017) case study of potato outgrower scheme in Maharashtra, India for instance, argues that rather than an inclusive alternative to land grabbing, outgrower scheme represents another form of land grabbing in rural India. According to the study “while some individual households have improved their livelihoods through participation, the scheme acts to reinforce already existing patterns of inequality”. The study concluded that, “the unequal power relations between firms and farmers skew the capture of benefits towards the firms, and render participating households vulnerable to indebtedness and loss of autonomy over land and livelihood decisions”. Most OGS studies in Ghana such as Agbelengor (2015); Torvikey et al. (2016) and Paglietti & Sabrie Roble (2012) are all on high valued cash crops usually produced in southern Ghana by medium to large scale farmers. The sorghum crop has dual role as staple food crop and now industrial cash crop and is mainly cultivated by SHF in Northern Ghana. The few studies on staple crops in Ghana such as Ragasa et al. (2018) on limitation of maize contract farming as a pro-poor strategy in the Upper West Region of Ghana and Brigitte & Ragasa (2018) on effect of contract farming on development projects and private sector activities in Northern Ghana and also, Maertens & Vande (2017) on rice in Benin. Whiles these studies focus on productivity and profitability, information on postharvest losses and vulnerability to climate change is ignored. In terms of appropriateness of the methodology adopted, the methods used were robust and appropriate for the context. However, Ragasa et al. (2018) study on maize cannot be used for policy making for sorghum due to numerous government support and available market for maize in Ghana. Also, Brigitte & Ragasa (2018) study was comparing effect of development projects on private projects and did not analyse the effect on livelihoods. Concerning Maertens & Vande (2017) study on rice, apart from substantial differences between sorghum and rice in terms of utilization, government support and available market 9 University of Ghana http://ugspace.ug.edu.gh for rice, geographical location and country specific characteristics weaken any justification for policy making in Ghana based on the Maertens & Vande (2017) work. Whiles Maertens & Vande (2017) and Ragasa et al. (2018) work provided information on effects of outgrower scheme on profitability for rice and maize farmers respectively, information on effects on PHL and vulnerability to climate change was ignored. Other studies such as Azumah et al. (2016) to determine the factors influencing SHF decision to participate in OGS and its effect on farm income in the Northern Region of Ghana found that access to credit and extension services positively influenced farmers decision to participate (Azumah et al., 2016). The Azumah et al. (2016) study did not target specific crop which defeats the concept of OGS. Given the importance of sorghum production to the livelihoods of SHF in Northern Ghana and lack of empirical study on its effect and also, the inability of government interventions to improve smallholder sorghum farmer’s performance, information on programmes and interventions that can transform their subsistence farming practices to agribusiness will be relevant for policy making. This research therefore examines whether OGS has any effects on the livelihoods of smallholder sorghum farmers in Northern Ghana or not. Specifically, the study will answer the following research questions: i. What factors influence smallholder farmers’ decision to participate in sorghum outgrower scheme? ii. What are the effects of outgrower scheme on smallholder sorghum farmers’ productivity, postharvest loss and profitability? iii. What are the effects of outgrower scheme on smallholder sorghum farmers’ vulnerability to climate change? 10 University of Ghana http://ugspace.ug.edu.gh 1.3 Objectives of the Study The main objective of the study is to examine the effects of outgrower scheme on the livelihoods of smallholder sorghum farmers in Northern Ghana. The specific objectives are to: i. Determine factors influencing smallholder sorghum farmers’ decision to participate in the outgrower scheme. ii. Establish the effects of outgrower scheme on smallholder sorghum farmers’ productivity, postharvest loss and profitability. iii. Determine the effects of outgrower scheme on smallholder sorghum farmers’ vulnerability to climate change. 1.4 Hypothesis of the Thesis The main hypotheses of the thesis are: 1. Household characteristics, farm and market characteristic, socio-economics and political characteristics significantly influence smallholder farmers’ decision to participate in sorghum outgrower scheme. 2. Sorghum ourgrower scheme enhances smallholder farmers’ productivity, helps reduce their postharvest losses and increases their profitability. 3. Sorghum ourgrower scheme reduces smallholder farmers’ vulnerability to climate change. 1.5 Relevance of the Study This study is relevant because it provides information on factors influencing SHF decision to participate in SOGS in Northern Ghana. Knowledge on factors influencing SHF decision to join outgrower scheme is important information that can inform policy makers 11 University of Ghana http://ugspace.ug.edu.gh on the type of policies that can improve the sorghum industry. Given the uniqueness of sorghum as staple crop and now industrial crop and also the complex nature of SHF engaging in cash crop production, understanding the specific factors motivating their decision to participate is relevant for policy making. Secondly, information on effects of OGS on SHF farmers’ productivity, postharvest loss and their profitability is important information for both private sector and government investment decisions. Existing literature on OGS on staple crops focus only on productivity and profitability. Information on postharvest losses and SHF vulnerability to climate change has always been ignored of which this study provides. Another important dimension of this study is to understand the effects of outgrower scheme on SHF vulnerability to climate change. Climate change and its impact on livelihoods is of global concern and most previous studies focuses only on farm level adaptation strategies. Evidence on the effects of OGS on vulnerability to climate change will help policy makers to take investment decisions on OGS development in Ghana. Following the government policy decision to develop Ghana beyond, various flagship programmes were implemented in the agricultural sector. These programmes aimed at enhancing SHF performance, promoting government modernization agenda and consequently leading to food security and job creation. Given that similar projects such as the establishment of Agricultural Mechanisation Centres, National Food Buffer Stock Companies and Fertilizer Subsidy programmes in the past failed to achieve project objectives, evidence on effects of SOGS could serve as a learning curve to guide government on future investment decisions. Finally, there are limited empirical literature on SOGS in Ghana. The few existing literature are end of project evaluation reports. The methodology adopted by such studies 12 University of Ghana http://ugspace.ug.edu.gh are not robust and policy recommendations from such reports are not scientifically proven. Using information from such reports for policy formulation could be misleading and leading to weak policies. This study employed Endogenous Switching Regression Model (ESRM) which address heterogeneity and endogeneity problems leading to unbiased and consistent results. Recommendations based on consistent and unbiased results will generate debate that will resonate beyond academic cycles to ideal policies that have positive impact on the beneficiaries. 1.6 Organisation of the Thesis The thesis is broadly organised in five chapters. Chapter one contains background information of the study, problem statement, objectives of the study, relevance of the study and organization of the thesis. Chapter two reviewed literature on SHF and their livelihood situation, their constraints, their productivity, their PHL and their vulnerability to climate change. Literature on OGS and its livelihood implications to SHF were also reviewed. Other areas that were reviewed in chapter two were agricultural policies and investment plans and Ghana government agricultural flagship programmes. The last part of chapter two summarises the literature and present important issues identified in the literature. Chapter three is about the conceptual framework of the study, theoretical framework, data sources and sampling procedure. Information on the study area and methods of data analysis were also discussed. Chapter four contain the results and discussions. Chapter five is the summary of the study, conclusion, policy recommendations and suggestions for future research. 13 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO LITERATURE REVIEW 2.1 Introduction This chapter presents a review of the relevant empirical literature on outgrower schemes. The chapter is organized into sections as follows. Section 2.2 background of sorghum production and sorghum ourgrower scheme in Ghana. Section 2.3 reviews literature on agriculture and economic development in Africa and that of Ghana in food and agricultural policies. 2.4 Ghana government’s agricultural flagship programmes. Section 2.5 contains contextual issues surrounding smallholder farmers, their livelihood situation, their production, postharvest losses and climate change issues. Section 2.6 review literature on outgrower schemes Section while section 2.7 summarises the literature and identifies literature gaps. 2.2 Background of Sorghum Production Sorghum farming in recent times attracted global attention due to discoveries on its nutritional values and the comparative advantage it has in the brewing industry to other crops (Ramashia et al., 2019). Among cereal crops consumed by humans, sorghum ranked third only superseded by wheat and rice in Africa, China and India (Hariprasanna et al., 2015; Ramashia et al., 2019). Some of the north-eastern countries in China such as Jilin, Liaoning, and Heilong-jiang have rice-like sorghum food as their major diet (Upadhyaya et al., 2017). United States, Australia, and Argentina account for 20%-30% of the world sorghum and top the sorghum exporting countries (Mccuistion et al., 2019). According to Ratnavathi et al. (2016), the unique attributes of sorghum as a drought-tolerant crop, grown in any marginal lands, requiring lesser inputs in production and climate-resilient makes it suitable 14 University of Ghana http://ugspace.ug.edu.gh crop for the poor. Utilization of sorghum as snacks, cookies and noodles are on the increase in Japan, the United States of America and Vietnam. Consumption of gluten-free beers prepared from sorghum are currently on the increase in America to cater for celiac patients (Aruna & Visarada, 2018). According to Aruna & Visarada (2018), special sorghum varieties are produced for direct consumption and for preparation of various snacks and weaning foods in India. In Africa, several meals such as porridge, steam-sorghum and snacks are prepared from sorghum and are delicacy for most rural and urban populations in many countries in the continent (Djameh et al., 2015; Upadhyaya et al., 2017). Couscous (grain like meal, prepared from sorghum flour) for example, is a granulated and steamed traditional food preferred in many parts of SSA, especially Senegal but originated from North Africa (Ramashia et al., 2019). Injera (crepe-like spongy bread made from sorghum) of Ethiopia, Kisra from Sudan is thin fermented bread prepared with sorghum, Ogi from Nigeria is fermented cereal pudding made of sorghum and gluten-free breads from Nigeria are some of the important food prepared using sorghum in SSA (Djameh et al., 2015; Upadhyaya et al., 2017). Also, lager and stout beer known as clear beer is brewed through malting of sorghum in Nigeria (Aruna & Visarada, 2018; Ogunsakin et al., 2017; Ramashia et al., 2019; Upadhyaya et al., 2017) Sorghum ranked third among important cereal crops grown in Ghana after maize and rice, with a share of 12% of the total cereal production (Akuriba & Asuming-Brempong, 2012; Angelucci, 2013) The popular sorghum varieties grown in Ghana are Naga White, Naga Red, Kapaala, and Dorado. Sorghum is considered a climate tolerant crop, easy to cultivate, requires low fertilizer application and can be cultivated on any marginal lands. It is considered in recent times as a cash crop of which majority of the youth in northern 15 University of Ghana http://ugspace.ug.edu.gh Ghana are engaged as economic activity (Akuriba & Asuming-Brempong, 2012; Djameh et al., 2015). Apart from food security and income, the value of sorghum is associated with social, cultural and religious ceremonies. The leaves and stalks are used for fencing, weaving baskets, roofing, mat making and also fuel for cooking in most rural areas in northern Ghana. Sorghum also serves critical needs such as curing illnesses, feeding lactating mothers and serving as delicacies for many households in rural northern Ghana (Angelucci, 2013; Ratnavathi et al., 2016). The GGBL discovered sorghum as a perfect substitute for barley and as part of their local resource used policy, sorghum demand by GGBL for brewing has increased (Sarfo-Mensah, 2017). According to Sarfo-Mensah (2017), sorghum production has attracted majority of farmers in northern Ghana through the GGBL SOGS. Despite the importance of sorghum production, there is no empirical literature on its livelihood’s implications 2.2.1 History of Sorghum Outgrower Schemes in Ghana The sorghum OGS in Ghana has its roots from European Cooperation for Rural Development (EUCORD) project, titled “West African Sorghum Value Chain Development” (WASVCD) which was launched in 2006. The project’s aim was to develop high-quality sorghum supply chain in West Africa to become a substitute for barley. Ghana and Sierra Leone were the pilot countries with initial funding of US$ 2,897,000 for 60 months (European Cooperation for Rural Development (EUCORD), 2008; Paglietti & Sabrie Roble, 2012). According to EUCORD (2008), the stakeholders for the project in Ghana were Technoserve who played the lead role in managing and coordinating the activities of all the other stakeholders; GGBL the end-user providing guaranteed market; Nucleus 16 University of Ghana http://ugspace.ug.edu.gh Farmers, SHF, Outgrowers and Large-Scale Individual Farmers being the main producers; Credit Venture Capital Fund and Sinapi Aba Savings and Loans companies were the financial institutions providing monetary support; Dizengoff Ghana Ltd supporting with agro-inputs such as fertilizer and weedicides and training on safe use of agrochemicals; Savannah Agricultural Research Institute (SARI) for agronomic support and varietal release and Nasia Rice Company providing storage facilities. The project led to an increase in sorghum production in the target communities from 112, MT in 2005/6 to 904 MT in 2006/7 and again, to 1,272 MT in 2007/8. The number of participating farmers also increased from 900 in 2005/6 to 3,210 and 5,670 in 2006/7 and 2007/8 respectively. After exit of this project, many companies and individuals engaged sorghum farmers in the Upper East, Upper West and Northern Region with GGBL still being the end-user providing guaranteed market. The structure of SOGS today comprises of SHF and medium to large scale farmers whose primary role is production. There are buyers popularly called aggregators buying from groups of farmers and supplying to big suppliers also called off-takers. The off-takers comprises of limited liability companies and Non-Governmental Organization (NGO) who signed contracts with GGBL. These actors deal with farmers directly by supporting them with inputs and other technical services to improve the quantity and quality of sorghum to meet the requirements of GGBL. They also support farmers with weather information and supply them with early maturing sorghum seeds to reduce the impact of climate change on their farming activities (Angelucci, 2013; Paglietti & Sbrie Roble, 2012). 2.3 Agriculture and Economic Development in Africa Agriculture plays an important role in the socio-economic development of almost all countries in Africa (FAO, 2015; Saghir & Hoogeveen, 2017; AGRA, 2018; Ecker, 2018). 17 University of Ghana http://ugspace.ug.edu.gh Its role ranges from poverty reduction, food security, job creation, boosting intra-Africa trade, rapid industrialization, sustainable resource and environmental management, economic diversification, shared prosperity and human security (AGRA, 2018; FAO, 2018; New Partnership for Africa Development (NEPAD), 2013). In reviewing prospects and challenges of Africa agriculture performance, FAO (2018) reported 53% of the 718 million total rural population in Africa as earning direct employment in the agricultural sector. The sector contributes 15% to Gross Domestic Product (GDP) on average, but the contribution varies widely across countries from under 3% in Botswana and South Africa to over 50% in Chad (NEPAD, 2013). Majority of the sector’s players are SHF who produce most of the food for domestic consumption and for export (AGRA, 2018). About thirty-three million farms of less than 2 hectares per household exist in Africa (FAO, 2018). Irrespective of their numbers, SHF often do not harvest enough food for home consumption and for the market due to inadequate access to productive resources. Catalysing the activities of SHF has greater potentials to enhance rural development which is fundamental for the attainment of the Sustainable Development Goals (SDGs) in Africa (AGRA, 2018; FAO, 2017). There are several empirical literatures linking market imperfection to the inefficiency of SHF (AGRA, 2017, 2018). AGRA (2017) reported general poor crop yields among SHF in SSA and associated agricultural growth in the last three decades to area expansion. This type of growth is said to be unsustainable due to consistent decline in farm lands as a result of population growth (AGRA, 2017). According to FAO (2017), yield improvement base on crop intensification practiced in the developed countries is more sustainable and should be adopted by policy makers in SSA. 18 University of Ghana http://ugspace.ug.edu.gh To improve their household incomes and mitigate consumption risks, SHF in most SSA countries rather adopted the option of diversifying their economic activities to include non- farm activities such as trading, animal rearing and agro-processing (Snyder et al., 2015). The production constraints of SHF is likely to impact on their nutrition, food security and undernourishment among many households. Despite being reduced from the 1990-92 figure of 33% to 23% in 2014-16, the absolute number of undernourished people in SSA remains relatively high (FAO, IFAD, & WFP, 2015). In absolute numbers, the undernourished people in SSA has increased from 44 million in 2014 to 218 million in 2018 (AGRA, 2018; FAO, 2018). The huge opportunities in the agricultural sector to change the fortunes of the continent has not been fully harnessed. Apart from the few commercial agriculture projects covering a relatively small share of crop production that practiced an improved method of farming, proper application of improved farming practices among SHF is much lower (Gandhi, 2014; Koira et al., 2014). Increasing population leading to limited land for a fallow period which has been the normal practice of Africa farmers to replenish their depleted soils and also a limited public investment in SHF threatened the sustainability of SHF crop productivity (AGRA, 2018; Ecker, 2018; FAO, 2015, 2018; Saghir & Hoogeveen, 2017). Despite the importance of SHF in ensuring food security in SSA, the focus of policies are usually on large scale industrial agriculture (FAO, 2018). Deeper thinking to unmask unexploited approach of transforming the activities of SHF is required for agricultural transformation that can impact on the SDGs in Africa. 2.3.1 Agriculture and Economic Development in Ghana The agricultural sector plays an important role in the socio-economic development of Ghana. The sector leads in provision of food, raw materials for industry, job creation and 19 University of Ghana http://ugspace.ug.edu.gh foreign exchange earnings(GSS, 2019). According the GLSS 7 report, the agricultural sector engages about 38.3% of the employed labour in Ghana compare to the services and industry sector of 43.5% and 18.2% respectively. More rural dwellers of about 65.2% are employed in the agricultural sector compare to the urban dwellers of 11.8% (GSS, 2019). The agricultural sector in Ghana is divided into five sub-sectors. The crops, fisheries, livestock, cocoa and forestry/logging (figure 2.1). Figure 2.1: Agricultural Sub-Sectors in Ghana Source: MoFA (2017) The Ministry of Food and Agriculture (MoFA) is the lead ministry responsible for livestock and the crops sub-sectors which happened to be the largest among all the sub- sectors (Ministry of Finance, 2018). The MoFA is also the lead government agency responsible for coordinating agricultural activities among the various ministries and other non-governmental organisations and is given the mandate to develop and execute policies and programmes within agriculture development agenda. 20 University of Ghana http://ugspace.ug.edu.gh Ministry of Fisheries and Aquaculture Development is responsible for the fisheries sector (Ministry of Finance, 2018). For cocoa, MoFA provides technical support and the cocoa board oversees all operational issues (Ministry of Finance, 2018). The Ministry of Lands and Natural Resources (MLNR) is in charge of the forestry and logging sector (Ministry of Finance, 2018). Other ministries such as Roads and Highways; Environment, Science, Technology and Innovation; Health; and Trade & Industry also undertake related activities that support agricultural development in the country (MOFA, 2015). The interconnectivity of various ministries suggests that the sector needs effective and holistic coordination of all the ministries for effective operations. Performance of Agricultural Sector in Ghana: The Agricultural Sector growth rates for the years 2010 to 2017, was projected at 6.0% (MoFAa, 2017). According to the 2017 progress report of MoFA and the 2018 budget statement by the Ministry of Finance, between 2010 and 2016, the average growth rates for the agricultural sector was 3.5%. The lowest growth rate of 0.8% was recorded in 2011 and 5.7% in 2013. The low growth rate in 2011 was attributed to the poor performance of the fisheries (-8.7%) and forestry/logging (-14%) coupled with poor rainfall patterns. The sector’s performance, however, experienced significant improvement in 2017. The agricultural Gross Domestic Product (GDP) in 2017 was 8.4%, exceeding the 2016 growth rate by 5.4 percentage points. The crops sub-sector was the highest with an average growth rate of 9.4% in 2017 (Ministry of Finance, 2018; MoFA, 2017). The introduction of government flagship programme “Planting for Food and Jobs (PFJ)” which provides farmers with subsidized fertilizer, high yielding seeds and extension services contributed to improvement in the agricultural performance in 2017 as presented in Figure 2.2. 21 University of Ghana http://ugspace.ug.edu.gh The inconsistencies in the sector’s performance over the years as shown in figure 2.2 call for strategies that will attract consistent public and private investment due to its critical role in ensuring food security, poverty reduction and socio-economic development of the country. 9 8 7 6 5 4 3 2 1 0 2010 2011 2012 2013 2014 2015 2016 2017 Year Figure 2.2: Real Annual GDP Growth Rate of the Agricultural Sector 2010 – 2017 Sources Ghana Statistical Service (2018) 2.3.2 Food and Agricultural Policies in Ghana The Food and Agriculture Sector Development Policy (FASDEP I & II) is the main government policy document governing the activities of MoFA and other related ministries in Ghana. The FASDEP I was formulated in 2002 as a holistic policy document targeting the private sector as the engine of growth (MoFA, 2007). The FASDEP I was formulated in line with the Accelerated Agricultural Growth and Development Strategy (AAGDS) and agricultural modernization agenda of the government. After four years of implementation of FASDEP I, Poverty and Social Impact Analysis (PSIA) was conducted and concluded that FASDEP I cannot achieve its objectives due to the following reasons:  Lack of proper targeting of the poor for the modernization agenda. 22 Agricultural GDP Growth University of Ghana http://ugspace.ug.edu.gh  Weak analysis of needs and priorities of the poor leading to wrong agenda setting.  No clear road map for the Ministry of Food and Agriculture to execute agricultural interventions that were outside the domain Ministry of Food and Agriculture (MoFA, 2007) Based on lessons learned from FASDEP I and upon broader consultation of stakeholders, the FASDEP II was formulated targeting fewer commodities and supporting all categories of farmers. It was also aimed to enhance the productivity of the agricultural commodity value chain by applying science and technology in production (MoFA, 2015). Six areas were identified as relevant for intervention under FASDEP II (MoFA, 2007). They are “Food Security and Emergency Preparedness; Improved Growth in Income; Increased Competitiveness and Enhanced Integration into Domestic and International Markets; Sustainable Management of Land and Environment; Science and Technology Applied in Food Production and Enhanced Institutional Coordination” (Boateng & Nyaaba, 2014; MoFA, 2007). Apart from the various areas targeted under the FASDEP II objectives, there were also specific policies and strategies for the various sub-sectors. These include “crop development policy, fisheries development policy, livestock development policy, and cocoa development policy”. Strategies developed for service delivery were “irrigation development strategy, extension services delivery strategy, agricultural mechanization strategy, plant protection strategy, financial services delivery strategies, inputs distribution strategy, gender mainstreaming strategy, youth in agriculture and human resource development strategy”(MoFA, 2007) According to MoFA (2007) inadequate allocation of funding, low prioritisation of food and agriculture sector by District Assemblies, inadequate response of other Ministries, 23 University of Ghana http://ugspace.ug.edu.gh Departments and Agencies (MDAs) to agriculture sector policy initiatives, inadequate response of private sector to policy initiatives and inadequate response of producers to policy initiatives were the risk identified as constraint to the success of FASDEP II. Other risk areas identified were adverse dynamics in international trade regimes for agricultural commodities, down-turns in world prices of key agricultural export commodities, poor rainfall patterns, an outbreak of pests and diseases and low commodity prices on the domestic market. Foreseeing these risks and lack of mitigation strategy in the FASDEP II was a serious omission that might have affected the achievement of FASDEP II objectives. Low interest in private sector participation in agriculture and financial institutions engaging SHF was identified as a major setback to the success of FASDEP II (Awunyo-Victor & Al-hassan, 2014; Mustapha et al., 2016). In conclusion, the FASDEP II objectives appear to contain broader range of issues in the various sectors that requires attention. The most importance issue identified by FASDEP II was the identification and targeting specific commodities with their potentials for investment. However, the challenge that limited the success of FASDEP II has to do with limited funding for investment areas identified, poor implementation of projects, low interest of key stakeholders, low technology uptake by the targeted farmers and poor institutional coordination were identified by literature as barriers to the success of FASDEP II (Abdul-Razak & Kruse, 2017; Fearon & Adraki, 2015) 2.3.3 Medium Term Agricultural Sector Investment Plans in Ghana To operationalise the FASDEP II, the METASIP I & II were developed as investment plans for the medium term (MoFA, 2018). The METASIP I was designed for the period of 2011 to 2015 with the aim of stimulating private sector investment in the medium-term 24 University of Ghana http://ugspace.ug.edu.gh (Boateng & Nyaaba, 2014). The aim was also to increase agriculture growth by at least 6% annually with government increasing national budget allocation of at least 10% within the investment period (MoFA, 2015). METASIP II was formulated to replace METASIP I for the period of 2014 to 2018 aiming at modernising agriculture. The investment under the METASIP I & II were largely executed as independent projects and treated as part of a range of projects. Even though the programme areas were designed as compliments, the interventions failed to meet the design requirements (MoFA, 2018). 2.4 Ghana Government Agricultural Flagship Programmes Modernising the activities of SHF is government’s policy priority (MoFA, 2015, 2018). Before 2017, notable flagship programmes seeking to modernise SHF activities and improve their productivity were the Ghana Commercial Agricultural Project (GCAP); Multinational NERICA Rice Dissemination Project; Agricultural Mechanization Centres, Youth in Agriculture with Block Farming concept, National Food Buffer Stock Company, the Fertilizer and Seed Subsidy Programme (Diao et al.,, 2014; Fearon & Adraki, 2015; Safo, 2016; Samrat, 2013; Udimal et al., 2017; World Bank, 2012) As part of campaign to remove the bottlenecks stifling the growth of the private sector and to provide enabling environment for growth, job creation, and prosperity for all, the New Patriotic Party in their 2016 Manifesto outlined flagship programmes for agricultural transformation in Ghana (New Patriotic Party (NPP), 2016) These programmes include: “Planting for Food and Jobs, One District One Warehouse (1D1W), One District One Factory (1D1F), One Village One Dam (1V1D), Planting for Export and Rural Development (PERD) and Rearing for Food and Jobs” (Akoto, 2019; Ministry of Finance, 2018; New Patrotic Party (NPP), 2016). 25 University of Ghana http://ugspace.ug.edu.gh The PFJ which was launched in April 2017 appears to be the government’s main agricultural modernization programme targeting SHF (MoFA, 2017b). The PFJ campaign is anchored on five pillars: Provision of improved seeds, the supply of fertilizers, provision of dedicated extension services, marketing and e-Agriculture (MoFA, 2017b). According to the 2017 monitoring report of MoFA, the implementation of PFJ in 2017 led to an increased in crop yields, improved extension services and creation of jobs (MoFA, 2017a). The 1D1W aim was to provide storage facilities to help address high postharvest losses that may be associate with expected bumper harvest due to the implementation of the PFJ. The 1D1F programme aimed was to establish at least, one factory in each of the 216 districts in Ghana to create economic growth, create jobs and transform the economic structure from raw material export to private sector value-added industrialized economy (Ministry of Finance, 2018). The 1V1D targeted northern Ghana, part of Volta region and the northern part of Brong Ahafo region where dams are expected to be dug in every village to conserve water for domestic and agricultural activities” (New Patriotic Party (NPP), 2016) The PERD was a decentralised tree crop project to develop nine tree value chains in the various districts. The trees under consideration were: Cashew, Citrus, Coffee, Cotton, Coconut, Mango, Rubber, Oil Palm and Shea nut trees. The aim of PERD was expected to provide raw materials to feed the IDIF that was to be established. It was also to promote rural economic growth and improve household incomes of rural farmers through the provision of certified improved seedlings, extension services, business support and regulatory mechanisms (Farmer Helpline, 2018) whiles these bold steps by the government was necessary, there is no proper documentation of detail implementation plan, funding sources, actual implementation commencement and 26 University of Ghana http://ugspace.ug.edu.gh exit plan. Information of the role of OGS on SHF productivity will be relevant to guide government implementation of its flagship programmes due to its private led modernization concept. 2.5 Concept of Smallholder Farmers The conceptualization of smallholder farmers (SHF) varies across countries and context. While the SDGs Monitoring Framework refers to it as “food producers”, other literature calls it small farmers, family farms, peasant farmers or subsistence farmers (Dittoh & Akuriba, 2018; Garner & de la O Campos, 2014; Khalil et al., 2017; MoFA, 2007, 2017b; Murphy, 2012) Dixon et al. (2004) summarises SHF when they say “smallholder farmers refers to limited resource endowment of farmers compared to those of other farmers in the sector”. In the same vein, World Bank, (2003) explained the concept of smallholders as “those farms with low asset base and operating in less than 2 hectares of land”. For Brooks et al., (2009), smallholders are “farm households which struggle to be competitive, either because of their endowments of assets compare unfavourably with those of more efficient producers or because they have to content with under-developed markets”. Similarly, in Murphy (2012), smallholder farmers are “characterized by marginalization in terms of their access to credit, information technology and capital”. The Ministry of Food and Agriculture in Ghana (2016) maintains that, a farmer is SHF based on their level of resource endowment and landholdings of less than two (2) hectares. This study is much associated with the concept presented in the report of High-Level Panel of Experts on Food Security and Nutrition (HLPE) in 2019. According to the HLPE 2019 report, “ smallholders are agricultural holding run by a family using mostly (or only) their own labour and deriving from that work a large but variable share of its income, in kind 27 University of Ghana http://ugspace.ug.edu.gh or in cash. The family relies on its agricultural activities for at least part of the food consumed, be it through self-provision, non-monetary exchanges or market exchanges. The family members also engage in activities other than farming, locally or through migration. The holding relies on family labour with limited reliance on temporary hired labour, but may be engaged in labour exchanges within the neighbourhood or a wider kinship framework”. As stated by Narayanan (2002), one of the reasons why the sole consensus around the concept of SHF may lack an agreed definition is the wide variety of farm structures and characteristics across different contexts and geographical locations. Much literature mentions the absence of such agreement, but few papers venture proposing acceptable definition. In the policy debate, the notion of “smallholder farmers” goes hand in hand with the idea of disadvantage, risk of poverty, lack of opportunities, and need of support. Hence an ideal definition should be consistent with the concepts of absolute poverty, severe food insecurity and access of optimal-productive resources and must be based on a criterion that does not necessarily depends on outcomes that have to be measured. Following the various debates Brooks et al. (2009); Garner & de la O Campos (2014); HLPE (2019); Khalil et al. (2017); Murphy (2012); Narayanan (2002) and World Bank (2003), this study summarises the definition of SHF as the one who is disadvantaged in access to optimal productive resources, lack of market access, marginalized in policy space, food insecure, absolute poverty and dependent on less than five (5) hectares of land for farming activities. Sorghum SHF are therefore defined as sorghum farmers who are dependent on not more than five hectares of land for cultivating sorghum and other crops, 28 University of Ghana http://ugspace.ug.edu.gh are disadvantaged in access to optimal productive resources, lack of market access, food insecure, absolute poverty and are marginalized in the policy space. 2.5.1 Smallholder Farmers Crop Productivity and Profitability The term productivity has largely been used to express different meanings and has provoked many conflicting interpretations. It is sometimes regarded as the overall efficiency with which a production system works, while on other occasions, it is defined as a ratio of output to the application of a given resource (Darku et al., 2016; Dharmasiri, 2012). Profitability, on the other hand, is the difference between the cost of production and yield per production area, usually, measured by kilograms per acre or per hectare (Devkota et al., 2019; Montgomery et al., 2017; Wünsch, Gruber et al., 2012). The main variable differentiating profitability and productivity is cost of production and price. While a farmer may obtain higher productivity, when the cost of production is high and selling price is low, that farmer is likely to make losses. On the other hand, lower productivity, low cost of production and higher prices can lead to higher profitability (Devkota et al., 2019; Montgomery et al., 2017; Wünsch et al., 2012). The table 2.1 shows average and actual productivity in Ghana for major crops for the 2016 farming season. 29 University of Ghana http://ugspace.ug.edu.gh Table 2.1: Productivity of Major Staple Crops in Ghana Staple crops Achievable yield Actual yield % Achieved Maize 5.50 2.05 37.27 Rice (Paddy) 6.00 3.01 50.17 Cassava 45.00 20.68 45.96 Yam 52 16.74 32.19 Sorghum 2.00 1.24 62.00 Soybeans 3.00 1.70 56.67 Source: MoFA (2017) In Ghana, SHF productivity and profitability is a reflection of the condition under which they operate. Most farm holdings of SHF are less than 5 hectares per household and dependent on traditional production technologies. There abounds considerable evidence of productivity gaps of SHF in Ghana (Villano et al., 2019). Table 2.1 contain 2017 national productivity of major staple crops in Ghana. Smallholder farmers face numerous challenges in production, ranging from limited or lack of access to technical assistance, modern inputs, access to credit, and mechanization services (Al-Hassan et al., 2013; Baiyegunhi et al., 2019; Donkor & Owusu, 2019; Feleke et al., 2017). These constraints led to their actual yields falling below their potential yields leading to low incomes and food insecurity. The efforts to address the yield gaps should be relentless if the contry can make significant in her support to achieving the SDG’s, especially, goals one and two. The table 2.2 contains trend of productivity in metric tonnes per hectare from 2013 to 2017. There was increase in productivity of some staple crops but the increment is still below the achievable yields for almost all the crops. The achievable yields for sorghum for instance was 2 tonnes per hectare (table 2.1), but the highest yield ever achieved was 1.24 in 2017 implying opportunity for sorghum farmers to make more income through the adoption of yield improvement programmes. 30 University of Ghana http://ugspace.ug.edu.gh Table 2.2: Trends of Productivity in Crops (MT/Ha) in Ghana Crop 2013 2014 2015 2016 2017* % Change Maize 1.72 1.73 1.92 1.99 2.05 3.0 Rice (Paddy) 2.64 2.69 2.75 2.92 3.01 3.1 Millet 0.97 0.96 0.97 1.16 1.05 -9.5 Sorghum 1.14 1.14 1.00 1.14 1.24 8.8 Cassava 18.27 18.59 18.78 20.25 20.68 2.1 Yam 16.78 16.63 16.96 17.42 16.74 -3.9 Cocoyam 6.50 6.48 6.49 6.53 6.79 4.0 Plantain 10.81 10.74 10.90 11.17 11.77 5.4 Groundnuts 1.24 1.28 1.24 1.30 1.37 5.4 Cowpea 1.24 1.21 1.25 1.41 1.37 -2.8 Soyabean 1.64 1.63 1.65 1.65 1.68 1.8 Source: MoFA (2017) 2.5.2 Postharvest Loss Among Smallholder Farmers Postharvest loss (PHL) is defined as a measurable decrease in the edible part of the food available for consumption but never made it to the consumer’s table (Chegere, 2018; Kiaya, 2014) Postharvest loss concerns have received global attention due to its effects on food security and the environment. The Food and Agricultural Organisation of The United Nations put the value of PHL in SSA as 4 billion USD a year (Sheahan & Barrett, 2017). This figure exceeds the total food aid received in SSA over the last decade. For the World Bank (2011), crop production constitutes about 70 percent of incomes in SSA of which 10 to 20 percent crops are loss through post-harvest. The annual value of the losses is estimated at USD 4 billion which is equivalent to calorific requirement of 48 million people annually in SSA (Wold Bank, 2011) Reducing PHL does not only improve SHF profitability, but also contribute to reduction in climate change, help stabilize the import bill due to limited resources invested in importation of food and improve labour productivity due to healthy living as a result of consuming less poisonous food (Affognon et al., 2015; Goldsmith et al., 2015; Rembold et al., 2014) 31 University of Ghana http://ugspace.ug.edu.gh Causes of PHL in developing countries is attributed to poor postharvest infrastructure, limited access and application of improved technology in food production and harvesting, high illiteracy among farmers to follow appropriate harvest and postharvest management practices, poor extension services and unfriendly weather conditions (Affognon et al., 2015; Chegere, 2018; Sheahan & Barrett, 2016; Wold Bank, 2011) In Ghana, most SHF depends on outmoded harvesting methods, stored their harvested produce in unhygienic conditions after harvest exposing them to pest, unfriendly weather leading to high PHL (Affognon et al., 2015). Postharvest loss and its implication to food security have received global attention but investment in PHL reduction in Ghana is still very low (Ministry of Finance, 2018). Some selected initiatives in Ghana to address PHL include the Warehouse Receipt System, Ghana Commodity Exchange program; One District, One Warehouse and establishment of National Food Buffer Stock Company (Ministry of Finance, 2018; MoFA, 2017b, 2018). Coulter & Onumah (2002) articulated the importance of regulated warehouse receipt system in addressing marketing challenges and emphasised the need to encourage stakeholder participation for effective implementation. Another major challenge to the PHL campaign is lack of comprehensive data on PHL loss for various crops. Limited knowledge in measuring PHL has led to oversimplification or overestimation of loss figures on most occasions. Sometimes estimates of a single national figure for a year are constantly being quoted for several years (MoFA, 2016). FASDEP II estimated PHL in 2007 as 20%-50% for fruits, vegetables, roots and tubers, and 20%-30% for cereals and legumes (MoFA, 2007). These figures are constantly being quoted and have now become an accepted figure for PHL in Ghana. The approach may be misleading since PHL in a given crop varies from other crops and varies across seasons (MoFA, 2016). At the 2018 biannual review meeting held in Gabon Brazzaville to assess Africa countries' 32 University of Ghana http://ugspace.ug.edu.gh progress towards achieving CAADP commitments, PHL was not among the indicators for Ghana due to lack of data (Kiaya, 2018). Uncertainty in estimating PHL and lack of reliable data of specific PHL loss in a given crop in Ghana could lead to poor policies to address losses and also, sub-optimal choices of strategies to reduce the losses. Various literature attempted to estimate the postharvest loss in cereals, specifically, in sorghum. They suggested seven stages where losses normally occur (Kiaya, 2018; Kitinoja & Kader, 2015; Sheahan & Barrett, 2016). These are a pre-harvest stage, harvesting, and initial handling, gathering and heaping, transportation/ carting, winnowing/threshing, drying, and storage. Pre-harvest sorghum loss occurs when there is damage to the grain in the field before harvesting. This could be due to biological and biotic factors such as poor control of weeds, insects, pests and diseases (FAO, 2011). Most SHF pre-harvest practices are poor due to limited investment, limited extension services and high level of illiteracy among SHF to follow good farming practices (MoFA, 2017). Harvesting and initial handling factors: Sorghum harvesting and handling are done manually by most SHF. During the harvesting stage, losses are mainly due to shattering and shedding of grains with the amount of loss largely dependent upon the duration of harvesting (FAO, 2014b). Over maturity and delayed harvesting is reported to be a major factor contributing to postharvest loss at harvesting stage (Huang et al., 2017; Jones et al., 2018). In Ghana, farmers leave harvested grains on the field to dry because they lack drying facilities (MoFA, 2017a). Gathering and Heaping: Gathering and heaping of sorghum is usually done by women and children (MOFA, 2016). After harvesting, the sorghum grain heads are gathered and heaped on the farm to either be transported to the farmhouse or to be threshed. Within the 33 University of Ghana http://ugspace.ug.edu.gh heaped period, substantial proportion of grains are either damaged, loss or contaminated. This is due to rainfall, high moisture content or pest infestation (Akuriba & Asuming- Brempong, 2012). Drying: Normally, grains are dried on bare grounds before carting for threshing (Kitinoja & Kader, 2015). Farmers who primarily depend on this method of drying are faced with challenges in the raining season. Also, the tendency of pests and rodents to attack during this period is high (Hengsdijk, 2017). Grains dried on bare floor could be contaminated with foreign materials, dirt, rainfall, pests, insects, livestock and bird attack (Kiaya, 2018). The solar driers are recommended for grains but highly expensive for SHF to afford. Threshing and winnowing: Threshing of sorghum is still a major challenge. Most SHF depends on manual threshing using sticks to separate grains from the thresh (MoFA, 2017). This process leads to losses due to splashing or incomplete threshing. When the harvest is threshed before it is fully dried, some grains will remain in the stalks and if the grain is threshed when is damp and immediately stored, it will be much more susceptible to micro- organisms and aflatoxin contamination (Africa Union, 2018). Limited availability of threshers for sorghum is a major cause of sorghum postharvest loss. Storage: Storage is an important activity in the postharvest chain as it allows farmers to keep their grains for good prices and for future consumption (Gitonga et al., 2013; Hengsdijk, 2017). Good quality grains can only be realized if the storage conditions are appropriate. Improper storage can cause stored sorghum to be contaminated by aflatoxin and also, pest and rodent infestation (Affognon et al., 2015; HLPE, 2014; Neme & Mohammed, 2017). Storage losses in Ghana are on-farm, household-level storage or market storage. According to FAO (2014), on-farm storage losses are between 4-10 while market storage is between 1-2%. For Rembold et al. (2014), on-farm storage is between 34 University of Ghana http://ugspace.ug.edu.gh 2-5% and 2-4% on market storage for grains. Limited warehouses compelled SHF to either store their sorghum on the farms or in their rooms leading to pest infestation leading to grain losses (Rembold et al., 2014). As part of the government’s agricultural flagship projects, warehouses are being constructed to help reduce storage losses. Transportation/carting: Transporting sorghum is another herculean task as in Ghana and many parts of Africa, women and children have to convey farm produce on their heads and shoulders to the nearest farmhouses (Africa Union, 2018). According to FAO (2014), grains losses due to poor transportation is between 1-2%. Produce are usually conveyed by women using head pans, baskets and on few occasions, animal tracking such as bullocks and donkey carts are commonly used (MoFA, 2007). Some farmers also engaged the services of motor tricycle popularly called “motor- king” for a fee. Farm produce are normally transported between the farmer's fields to where grains are threshed and the threshing floor to the storage centre and to the markets. Transportation could be made easier with motor trucks or cars but this is dependent on the availability of access roads and farmers’ ability and willingness to pay for the transport cost (Hell & Mutegi, 2011). The degree of grain loss during transportation is mostly proportional to the distance from the farm to the final destination (Rembold et al., 2014). 2.5.3 Smallholder Farmers and Climate Change Climate change and its impact on the environment and socio-economic activities of communities is a major concern for policy makers in SSA (Abdul-Razak & Kruse, 2017; Pandey et al., 2017; Thompson, Berrang-Ford, & Ford, 2010) . According to (Porter et al. (2014), climate change in the last 30 years has contributed to a declining agricultural performance between 1% to 5% per decade. The world poorest are said to suffer more from the impact due to their weak capacity to adapt (Adeyeri et al., 2019). 35 University of Ghana http://ugspace.ug.edu.gh Climate change data in Ghana from 1961 to 2000 shows a progressive increase in temperature and declined to mean annual rainfall pattern in all the six agro-ecological zones (UNEP & UNDP, 2013). The annual average temperature has increased by 1°C in the last 30 years. (UNEP & UNDP (2013) projected temperature to continue to rise, while rainfall is predicted to decline in all agro-ecological zones in Ghana (Table 2.3 and Table 2.4) Table 2.3: Mean Annual Change in Rainfall (%) in Ghana Year Sudan Guinea Transitional Deciduous Rainforest Coastal 2020 -1.1 -1.9 -2.2 -2.8 -3.1 -3.1 20150 -6.7 -7.8 -8.8 -10.9 -12.1 -12.3 2080 -12.8 -12.8 -14.6 -18.6 -201.2 -20.5 Source: Minia et al. (2004) Table 2.4: Projected Mean Annual Temperature Changes in Ghana Year Sudan Guinea Transitional Deciduous Rainfores Coastal 2020 0.8 0.8 0.8 0.8 0.8 0.8 20150 2.6 2.5 2.5 2.5 2.5 2.5 2080 5.8 5.4 5.4 5.4 5.4 5.4 Source: Minia et al. (2004) The climate change situation is more serious in northern Ghana which has increasing number of droughts, floods and bushfires. The SHF with minimal livelihoods alternatives are largely affected by the impact (Aniah et al., 2019; Dickinson et al., 2017; Etwire et al., 2013). Majority of SHF in northern Ghana rely on rainfed for their agricultural activities and any variations in the rainfall pattern translate to low farm productivity, lower-income and high vulnerability (Morton, 2013; Yaro, 2013). According to Abdul-Razak & Kruse (2017), since patterns and volumes of rainfall determines agricultural productivity in northern Ghana, people whose livelihoods depends on agriculture would be largely 36 University of Ghana http://ugspace.ug.edu.gh affected. According to Antwi-Agyei et al. (2018), crop yields are estimated to reduce in Ghana by 7% by 2020 as a result of the projected decline in rainfall which would negatively affect SHF more since their main economic activity is farming. It is important to acknowledge the efforts of researches seeking remedy to address the impact of climate change on vulnerable communities. Most of the studies centred on farm- level resilience and adaptation strategies (Adu et al., 2018; Dzanku, 2015; Etwire et al., 2013; Yaro, 2013). Limited studies examined specific interventions that increases incomes of SHF and indirectly empowers them to become less vulnerable. Knowing the level of vulnerability among various groups of farmers and the impact of a specific agricultural intervention that has demonstrated positive impact on livelihoods would be an ideal step for policymakers. Information on SOGS and its effect on SHF farmers' profitability and subsequently, their vulnerability is relevant. 2.5.4 Smallholder Farmers Vulnerability to Climate Change There are different definitions of climate change vulnerability. The most comprehensive definition commonly applied in literature is the one by IPCC 2009 authored by Hahn Riederer and Foster (2009). Hahn et al. (2009) defines climate change as “the degree in which a system is susceptible and unable to cope with adverse effects of climate change while vulnerability is a function of magnitude and rate of which a system is sensitive, exposed and adapt to changes in climate” (Hahn, Riederer, & Foster, 2009). Vulnerability is usually considered a function of three elements. “sensitivity to hazard, exposure to hazard and the capacity of that system to adapt to the hazard as a result of the climate change”(Adu et al., 2018; Chinwendu et al., 2017; Hahn et al., 2009; Jamshidi et al., 2019). Chinwendu et al. (2017) have argued that vulnerability explains the degree of risk and inability of a system to resist to the climate variations. The exposure is the degree 37 University of Ghana http://ugspace.ug.edu.gh to which a system is exposed to the effect of changes in climate and sensitivity is the extent to which a system is impacted adversely or positively by climate change. Smallholder farmers in Ghana face a wide range of climate-related risks ranging from weather failure, total crop failure, market risk, postharvest loss with negative effects on livelihoods (Jamshidi et al., 2019; Lahouar et al., 2016; Makate et al., 2019). How to withstand this climate shocks has been the concern of many researchers. Most studies focus on identifying adaptation and coping strategies by SHF to reduce the adverse impacts on their livelihoods (Antwi-Agyei, Dougill et al., 2018; Boansi et al., 2017). Other studies such (Abdul-Razak & Kruse, 2017; Etwire et al., 2013; Yaro, 2013) have explored strategies that local institutions adopt to promote climate change adaptation in northern Ghana. Some studies also focused on the role of indigenous knowledge in farming to address the adverse effect of climate change on livelihoods (Bhattacharjee & Behera, 2018; Pandey et al., 2017) There exist limited studies exploring interventions that link farmers and business entities such as agribusiness companies who support farmers to improve their productivity and profitability and how such interventions could empower farmers to indirectly become less vulnerability to climate change. It is against this backdrop that this study examines OGS to establish its appropriateness to deal with the diverse constraints inhibiting the performance of SHF and enhance their livelihood outcome. The possibility of SHF attaining higher income through out-grower schemes and by inference, improved livelihoods is reported by Minot (2015); Maertens & Velde (2017) and Ragasa et al. (2018). 38 University of Ghana http://ugspace.ug.edu.gh 2.6 Outgrower Scheme Debate In the era of globalization and market liberalization, there are globally concerns of the role of SHF in the market economy due to their subsistence nature (Almas & Obembeb, 2014; Koira et al., 2014; Gaffney et al., 2019). Smallholder farmers could be marginalised as the agribusiness companies may capitalise on the market opportunities to take over their lands (Eaton & Shepherd, 2001). This could lead to farmers abandoning farming activities for alternative livelihoods which has been the case in most developing countries (Bellemare & Bloem, 2018; Jayne et al., 2016). Attempts by policy makers to avert the situation is by providing rural communities with social support services such as subsidising farm inputs (FAO, 2014a; Saghir & Hoogeveen, 2017). Several impact assessments reports proved that government interventions in Ghana has minimal impact on the beneficiaries (Fearon & Adraki, 2015; ISSER, 2012). Following Maertens & Velde (2017) and Fawad et al. (2019), OGS is recommended as appropriate in dealing with these challenges. The OGS provides linkages and offers important pathway for SHF to obtain improved agronomic services and inputs and also, approach their farming with business mindset. These services and support systems apart from demographic characteristic of the farmers such as gender and age, are said to be the main motivations for SHF participation in OGS and also, modernising their farming activities (Gebrezgabher et al., 2015; Ghimire et al., 2015; Gover, 1990; Mariano et al., 2012; Udimal et al., 2017). Recent report by Ragasa, Lambrecht, & Kufoalor (2018) and Fawad et al. (2019) supported the earlier reports on benefits of OGS to SHF claimed that, the OGS also creates opportunities for farmers to access market and industries to obtain their raw materials from farmers. The OGS is defined as a business transaction between farmers and buyers for the purposes of production and supply of agricultural commodities based on formal or informal prior 39 University of Ghana http://ugspace.ug.edu.gh agreement (Minot & Nicholas, 1986; Eaton & Shepherd, 2001). Minot (2011) defined OGS as an existing ad hoc trade arrangement being replaced with well-structured commercial trade arrangement between farmers and traders leading to a vertical integration in a given agricultural value chain. Food and Agricultural Organisations of the United Nation in 2012 integrates the above definitional schisms to define OGS as supply agreements between agribusiness companies and farmers leading to mutual gains (Paglietti & Sabrie Roble, 2012) The OGS and contract farming are often used interchangeably in literature (Bellemare, 2012; Bellemare & Bloem, 2018; Ragasa et al., 2018; Wendimu et al., 2016). There are however some subtle differences between the two schemes depending on how the arrangement is made and who initiates and controls the scheme. According to Glover & Kusterer, (1990), contract farming is a private-led scheme whiles OGS are public led enterprises. This distinction appears to be losing relevance in recent times as private companies also engaged outgrowers on contractual bases other than the usual contract farming (Barrett et al., 2012; Maertens & Velde, 2017; Ton et al., 2017). This study found no differences between OGS and contract farming. The importance of OGS and its impact on participating farmers is viewed from different lenses. Among the proponents views, adoption of OGS has the potential to transform SHF subsistence farming to commercial oriented enterprise (Maertens & Velde, 2017; Minot, 2015; Minot & Ronchi, 2014; Paglietti & Sabrie Roble, 2012; Ton et al., 2015). They argued that OGS support farmers with new agronomic information, improved inputs, credit and guaranteed market on contractual bases. The conditions attached to these services compelled subsistence farmers to change their practices in fulfilment of the contractual obligations. 40 University of Ghana http://ugspace.ug.edu.gh Risk sharing argument was advanced by Glover & Kusterer (1990) that, OGS is a way of sharing production risk between farmers and buyers. According to them, the SHF on their own tend to be risk averse and are unlikely to invest in expanding production for uncertain market. Against this uncertainty, agribusiness companies linked capital to the farmers via contracts with guaranteed markets. Similar to Glover & Kusterer (1990) assession, Eaton & Shepherd (2001); Barrett et al. (2012) and Ton et al. (2017) also concede that, OGS linked SHF to guaranteed markets and create opportunity for them to participate in high- value markets. From the international trade point of view, Paglietti & Sabrie Roble (2012); Barrett et al. (2012); and Odunze et al. (2015) claimed OGS are encouraged by multinational donor agencies and governments as a channel that links SHF to global markets and at the same time, connecting them to yields improving technology. The literature refers to this kind of investment as responsible investment. Some recent studies confirming earlier studies on relevance of OGS concluded that, the OGS helped SHF overcome market imperfections, reduce market risk and reduce their vulnerability to climate shocks (Johanna et al., 2019; Rhebergen et al., 2018). Despite optimism expressed by various authors on the positive outcomes of OGS, others have contrary views on welfare gains by SHF for participating in the OGS (Adams et al., 2019; Fawad et al., 2019; Huang et al., 2018; Manda et al., 2018; Narayanan, 2014; Odunze et al., 2015; Oya, 2012; Vicol, 2017). According to Narayanan (2014), participation in OGS shows high level of dissatisfaction among farmers in southern India and high exit rate due to high debt they incurred. For Odunze et al. (2015) and Vicol (2017), power imbalance between SHF and buyers could subject the farmers to farm labourers. Vicol (2017) remarked that “In India, the politics of land and agriculture means that land acquisition is not a viable option for foreign or domestic agri-business capital 41 University of Ghana http://ugspace.ug.edu.gh searching for investment and accumulation opportunities. Agri-business is instead coming to control rural land through other means such as CF, while continuing to introduce new modes of production and accumulation into already highly unequal rural livelihood landscapes. Contract farming entangles farmers in new relationships of capital, debt and power that often do not result in the pro-poor rural development outcomes imagined by some”. Ragasa et al. (2018) study on maize OGS in Upper West Region of Ghana claimed the OGS has potentials to increase productivity and improve supply of high-quality maize to buyers, but farmers on the scheme were making low profit than non-scheme members due to high cost of production such as high input cost. The report claimed that the yield increase by participating farmers was not enough to offset the high cost of inputs provided by buyers. On gender perspective, Adams et al. (2019) study on gender balance and sugarcane outgrower scheme in Zimbabwe argued that, while the scheme gave bargaining power for participating women, it has created gender gap in land control, heightened demand for women’s farm labour and gave them limited leeway to demand appropriation of their unpaid labour from their husbands. The third set of researchers have neutral stands on impact of OGS and concluded that, it should be analysed base on context and content specific (Bellemare & Bloem, 2018; Fawad et al., 2019). They argued that, benefits of OGS is based on the type of contractual agreement between producers and buyers. In occasions where OGS failed, reports of poor communication among parties, unfavourable terms and dictating of prices by services providers were reported. 42 University of Ghana http://ugspace.ug.edu.gh 2.6.1 Types of Outgrower Schemes While some OGS are structured around formal contracts and may include provision of credit and inputs, other schemes are structured around informal contracts without credit, supply of inputs or provision of extension services (Eaton & Shepherd, 2001; Hobden & Sands, 2017; Paglietti & Sabrie Roble, 2012) Five different types of OGS were categorized by Hobden & Sands (2017) which is in conformity with FAO earlier classification published by Eaton & Shepherd (2001). The various types of outgrower schemes listed are as follows: Centralised Model of Outgrower Schemes: The centralized model is a form of vertical coordination of which buyers provide production and marketing services to SHF on their own land. Crop category in this model are those requiring high level processing such as coffee, cocoa and tea. The OGS operated by Blue Skies Company Limited in the Akwapim South Municipality of Ghana is a typical example of Centralised Outgrower Model (Agbelengor, 2015) Nucleus Estate Model: Buyers often own a plantation or estate and at the same time engage SHF in the neighbourhood to produce. According to Eaton and Shepherd (2001), the nucleus-estate model is suitable for perennial crops such as oil palm, rubber and coffee. Example of nucleus estate model in Ghana is the Ghana Oil Palm Development Company Limited (GOPDC) scheme. Under the GOPDC arrangement, the company has a plantation estate and also supporting SHF with inputs and extension services for a reliable supply of fresh oil palm (Loggoh, 2013) The Multipartite Estate Model: The multipartite estate model is a joint venture involving a public entity, a private firm and farmers. Baumann (2000) contend that most of the multipartite estate models are practiced in developing countries. Typical example is 43 University of Ghana http://ugspace.ug.edu.gh Rubber Outgrower Plantation Project comprising the financial institutions, Ghana Rubber Estate Limited and the outgrowers (Paglietti and Sabrie, 2012). The Intermediary Model: Under the intermediary model, firms do not contract directly with the farmers but does it through an intermediary. The firms subcontract intermediaries who deal directly with farmers. The intermediary could be farming committees or aggregators. The contracting firms have limited control over the farmers and how the produce is produced (Eaton & Shepherd, 2001) The current GGBL OGS mimic this model (Paglietti & Sabrie Roble, 2012; Sarfo-Mensah, 2017). 2.7 Summary and Knowledge Gap in the Literature In summary, the review found the role of SHF in Ghana’s agriculture development as crucial in poverty reduction, job creation and overall economic development. Their low productivity, high PHL, and low profitability is due to challenges ranging from difficulty in access to quality productive resources, knowledge gap, lack of market access and guaranteed prices and impact of climate change. To address these challenges, FASDEP I and II were formulated with METASIP I and II as medium-term investment plans to modernise the activities of SHF. While investment as part of the modernization agenda to improve productivity was lauded by many stakeholders, limited investment in the area of postharvest loss reduction and market access were identified as serious drawback to SHF modernisation agenda. The role of OGS in supporting SHF production and market access is recommended as having potentials to transform their subsistence farming practices to see farming as a business. Despite the optimism expressed on important role OGS play in transforming the subsistence farming activities of SHF, there is limited research on effect of OGS on SHF 44 University of Ghana http://ugspace.ug.edu.gh productivity, postharvest loss and profitability on indigenous crops such as sorghum in Ghana. On the impact of climate change on the livelihoods of SHF, the literature found severe effect on socio-economic activities and livelihoods of rural communities. Policy recommendations to mitigate the effects by different studies were in the area of adaptation and coping mechanisms. There is limited research in the area of OGS and its relationship to climate change mitigation among SHF. In conclusion, there is literature gap on the effect of OGS on smallholder sorghum farmer’s productivity, their postharvest loss reduction, increasing their profitability and reducing their vulnerability to climate change of which this thesis is to investigate. 45 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE METHODOLOGY 3.1 Introduction This chapter is about the methods used to achieve the research objectives. It begins with a section explaining the conceptual, analytical and theoretical frameworks of the study. The chapter also elaborates on the data collection processes including sources of data, sampling methods, interviews conducted, analysis of the data and the study area. The final section of the chapter presents the scope and limitation of the study. 3.2 Conceptual Framework of the Study Smallholder farmers face imperfect credit and input market (Aniah et al., 2019; Ebata & Hernandez, 2017; Gaffney et al., 2019). They lack extension services, they have difficulty accessing credit and agro-inputs, poor pricing and limited access to ready market (Al-hassan, 2014; Dittoh & Akuriba, 2018; Ecker, 2018; Baiyegunhi et al., 2019). They are also confronted with negative impacts of climate change which are associated with low and erratic rainfall patterns, floods, droughts, pests and disease infestation (Abdul-Razak & Kruse, 2017; Makuvaro et al., 2018). These issues negatively impact on their productivity, post-harvest management and increase their vulnerability to climate change. The combined effect is poor livelihoods (Aniah et al., 2019). In the context of climate change, vulnerability is conceptualised as fall in well-being leading to people’s inability to cope with changes in climate variables (Shah et al., 2013; Pandey et al., 2017; Jamshidi et al., 2019). Whiles there are interconnectivity between low income earning, asset holding and livelihoods empowerment (Dzanku, 2015; Pandey et al., 2017), vulnerability to climate change equally have replicate effect on livelihoods (Alobo, 2019; Aniah et al., 2019). Low income earners and low asset holders are vulnerable to climate change than high income 46 University of Ghana http://ugspace.ug.edu.gh earners (Shah et al., 2013). Vulnerability also weakened labour productivity and limit capacity of the affected person to obtain quality livelihood outcomes such as access to food, access to shelter and quality health care ( Travis, 2014; Adu et al., 2018). Any intervention that improves incomes will probably influence livelihoods and vulnerability to climate change (Yaro, 2013; Tshilidzi et al., 2016; Pandey et al., 2017). According to Barrett et al. (2012), Abebe et al. (2013), Maertens & Velde (2017) and Bellemare & Bloem, (2018), participation in OGS is expected to improve farmers’ productivity, reduce their postharvest losses and increase their profitability, increase their disposable income and increase their asset holding. Improved incomes and asset holding is an indication of ability to adapt to climate stress and become less vulnerable to climate change (Abdul-Razak & Kruse, 2017; Antwi et al., 2015; Boansi et al., 2017; Jamshidi et al., 2019) In the conceptual framework in Figure 3.1, smallholder sorghum farmers are experiencing low productivity, high postharvest losses, difficulty accessing guaranteed market leading to low profitability. Their activities are further worsened by impact of climate change. On the other hand, there is demand for sorghum by agribusiness companies and manufacturing companies. The Agribusiness Companies therefore enter into contractual arrangement with the manufacturing and processing companies and agreed on some package to support the smallholder farmers to produce sorghum of certain quality, quantity and an agreed price. In addition, the manufacturing and processing companies engages financial institutions to support the Agribusiness Companies to extend their support to the smallholder farmers. The agribusiness companies then engaged the smallholder farmers on contractual bases again, support them with inputs, credit and extension services to improve their productivity. They also engaged government and private Extension Officers to train the farmers on postharvest management leading to low postharvest losses. Enhanced productivity and low postharvest 47 University of Ghana http://ugspace.ug.edu.gh losses leads to improved profitability and increase disposable income. Due to guaranteed market and guaranteed supply, transaction cost of searching for market and price negotiation is also minimized for both producers and suppliers (Mwangi & Kariuki, 2015). With the improved profitability, the SHF can diversifies their investments into off-farm activities and livestock production and thereby are able to withstand climate shocks and become less vulnerable to climate change. When farmers are less vulnerable to climate change, they are more likely to have improved livelihoods (Figure 3.1). 48 University of Ghana http://ugspace.ug.edu.gh Financial Guinness Ghana Brewery Training Institutions/ Institution Limited MoFA/Meteo etc. Credit Facility NGO Agribusiness Com. FM OUTGROWERS C redit Facility Constraints Low Output GAP District MoFA Improved Productivity, Low Improve Livelihood C limate Resilience outcomes PHL and High Profit Figure 3.1: Conceptual Framework of Sorghum Outgrower Scheme in Ghana Source: Author’s conceptualization 49 Weather Info Regulation, Coordination, M&E Sales Sales Contracts Contract Investme Support nt University of Ghana http://ugspace.ug.edu.gh 3.3 Analytical Framework The general analytical framework of the thesis is based on Sustainable Livelihood Framework (SLF) (DFID, 1999; Knutsson, 2006; Pandey et al., 2017; Scoones, 1998). The concept of SLF was founded in 1990s by Robert Chambers and later developed by Chamber and Conway in the 1990s (Chambers & Conway, 1991). Sustainable livelihood is defined as “accumulation of both material and social assets by people for positive living” (Kemp et al., 2009; Pandey et al., 2017). Livelihood is said to be sustainable when “people can cope and recover from shocks, stresses and enhance their capabilities without undermining the importance of the natural resource base” (Shah et al., 2013; Pandey et al., 2017). The sustainable livelihood concept views poverty as a multifaceted concept, encompassing different variables beyond just economic growth (Pandey et al., 2017; Bhattacharjee & Behera, 2018). Other factors of relevance beyond economic variables include social network, empowerment and right of the poor to have meaningful living in all spheres of life. This is an important information to understand the need to involve the poor in livelihood empowerment activities such as the outgrower scheme activities. The SLF has been simplified with a model in Figure 3.2 for easy understanding of the different components and their interrelations. The context of vulnerability describes the external environment where poor people live (Antwi et al., 2015; Shah et al., 2013). This includes trends related to technology or population. It also encompasses shocks such as inflation or natural disasters and seasonality which refers to employment opportunities. All these factors influence the assets base of the poor. 50 University of Ghana http://ugspace.ug.edu.gh Figure 3.2: Sustainable Livelihood Framework Source: DFID (2000) The sustainable livelihoods model was developed on the assumption that people need different kinds of assets for meaningful livelihood outcomes (Chambers & Conway, 1991; Pandey et al., 2017). These assets were classified as “Human capital referring to the knowledge, skills, ability and good health that will ensure people obtain their desired outcomes. Human (H) capital is essential for combination of other kinds of capitals for higher productivity. Social (S) capital refers to the social resources such as membership of groups, networking, or mere trust that exists between people that make them help each other. Natural (N) capital covers tangible resources such as trees, land and intangible products such as biodiversity and atmosphere. Physical (P) capital describes the basic infrastructure and producer goods. Financial (F) capital is the financial resources. Transforming structure and processes includes the policies and the institutions that frame the livelihoods of the poor. These processes determine the opportunities people have to different kinds of assets. 51 University of Ghana http://ugspace.ug.edu.gh Livelihood strategies are defined by how people act to achieve the desired living standards. Livelihood outcomes are determined and achieved by strategies people adapt. The outcomes of SOGS that provide skills and information influence productivity, postharvest loss reduction, and profitability mimics the sustainable livelihoods framework in the sense that, skills and information can improve profitability and improve incomes which can influence assets accumulation leading to sustainable livelihood empowerment and less vulnerability to climate change. 3.4 Theoretical Framework for Factors Influencing Participation in Outgrower Scheme Factors influencing smallholder farmers participation in SOGS is founded on utility maximization theory. From the utility perspective, SHF choice to participate in an intervention is influenced by their expected utility. Smallholder farmers are said to maximize their utility function, subject to their expected benefit against none participation (Aleskerov et al., 2007) In the Equation 3.1, the difference between the utility of participation (𝑈𝑝𝑖) and utility of none participation (𝑈𝑛𝑖) in OGS may be denoted as 𝑂𝐺𝑆 ∗ such that utility maximizing farmer (i), will choose to participate, if the utility gained for participating is greater than the utility of not participating. This is expressed in linear function as 𝑂𝐺𝑆∗ = 𝑈𝑝𝑖 − 𝑈𝑛𝑖 > 0. (3.1) Since these utilities are unobservable, they can be expressed as a function of observable elements in the following latent variable model in equation 3.2: ∗ 1 𝑖𝑓 𝑂𝐺𝑆 ∗ > 0 𝑂𝐺𝑆 = 𝛽𝑋𝑖 + 𝑢𝑖 with 𝐺𝑖 = { (3.2) 0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 where OGS is a binary indicator variable that equals 1 if SHF participate in OGS and zero if otherwise 52 University of Ghana http://ugspace.ug.edu.gh 𝛽 is a vector of parameters to be estimated; X is a vector of explanatory variables; u is the error term. The utility maximization theory can also be expressed as in equation 3.4: 𝑀𝐴𝑋 (𝑈)= 𝑓(𝑥) (3.3) Utility, U, is determined by a set of variables X. These variables influence the farmer’s decision, ability and willingness to participate in SOGS and at the same time, influence the outcome variables where: U= utility (3.4) 𝑥= factors influencing decision to participate. Given that the net utility is represented by 𝑈∗, 𝑈∗= 𝑈𝑖𝑗𝑡>𝑈𝑖𝑚𝑡 (3.5) Farmer i will only participate in SOGS (j) at time (t) if the expected utility derived from participating (𝑈𝑖𝑗𝑡) is greater than the expected utility (𝑈𝑖𝑚𝑡) of not participating. Where: 𝑈∗ = benefits of participating in j as opposed to not participating (m) while 𝑈∗ is unobserved. To determine factors influencing SHF participation in SOGS following utility maximization theory, a probit model was adopted. Common models for analysing binary dependent variable include discriminant analysis, the linear probability model, the probit model, and the logit model (Maddala, 1983) Discriminant analysis is not suited for this analysis as it assumes that the explanatory variables are normally distributed. For the linear probability model, its limitation is the predictions could fall outside the interval (0,1) (Maddala, 1983). Results from probit or logit model are 53 University of Ghana http://ugspace.ug.edu.gh similar (Maddala, 1983). Following Maddala (1983), probit model was adopted to determine the factors influencing participation in SOGS. The Utility Maximization theory has been applied in many studies to explain technology adoption and contract farming schemes in many parts of Africa. For instance, it was recently used by Ragasa et al. (2018) to analyze the limitations of maize contract farming as a pro- poor strategy in the Upper West Region of Ghana; determining agricultural technology adoption in Mozambique; farmers’ response to adoption of commercially available organic fertilizers in Oyo state , Nigeria and determinants of adopting imazapyr-resistant maize for Striga control in Western Kenya (Ajewole, 2010; Mwangi & Kariuki, 2015; Ragasa et al., 2018; Uaiene et al., 2009). 3.5 Theoretical Framework for Effects of Sorghum Outgrower Scheme on Productivity, Postharvest Loss and Profitability The theoretical foundation determining the effect of sorghum outgrower scheme on productivity, postharvest loss reduction and profitability is production theory. The production process involves combination of different factors of production into an output using a given technology (Choi, 2009). Agronomics measure crop productivity as the crop yield generated from a unit area of land, usually one hectare. Econometricians analysed productivity using econometric models such as data envelopment analysis, stochastic frontier approach, linear regression, propensity score matching and endogenous switching regression (Donkor & Owusu, 2019). The stochastic frontier and data envelopment analysis are mostly applied to estimate productive efficiency, whereas linear regression, propensity score matching and endogenous 54 University of Ghana http://ugspace.ug.edu.gh switching regression are used to measure effect of a given variable of interest on a given outcome variable (Asfaw et al., 2012; Lokshin & Sajaia, 2018). Lokshin & Sajaia (2018) analytical approach, the sorghum productivity is examined with a generalised production function presented in equation (3.6): 𝑃𝑅𝑂𝐷𝑖 = 𝑓(𝑋1𝑖,𝑋2𝑖,𝑋3𝑖, … , 𝑋𝑗𝑛𝑆𝑂𝐺𝑆𝑖,𝑋𝑖,Ɛ)𝑖 = 1,2, … . , 𝑁 (3.6) Where 𝑃𝑅𝑂𝐷𝑖,= sorghum productivity per hectare 𝑋1𝑖,𝑋2𝑖,𝑋3𝑖, … … , 𝑋𝑗𝑛 = A set of factor inputs (seed, fertilizer, labour, seed, knowledge used for production) f = The relationship of the various output and the factor inputs. SOGS= denotes package of SOGS captured as a variable, 1 represents participation in SOGS and 0 not participating. For profitability, this study extends the analysis of profit margin presented by Wünsch et al. (2012) in Equation 3.7 estimate profit by accounting for the differences of product of price per kilo and total cost (fixed cost plus variable cost per one hectare). Following Wunsche eta al. (2012), Equation 𝜋𝑖𝑗𝑘= (𝐶𝑅𝑖 +𝐿𝑖𝑘 +DP-𝐷𝐶𝑖𝑗 -𝐹𝐶𝑖 ) 𝐻𝑖 (3.7) Where: 𝜋𝑖𝑗= average profitability 𝐶𝑅𝑖 = average revenue per hectare for cropping year which is the sum of product of price and yield for a given crop 55 University of Ghana http://ugspace.ug.edu.gh 𝐿𝑖𝑘 = is crop insurance indemnities per hectare for cropping system i and is the sum of indemnities for a given crops weighted by the share of area for the crop. The indemnity for a given crop is zero expect when the yield for the given year falls below the yield guarantee. DP= is direct payments per hectare, which are constant across cropping season, machinery complements. 𝐷𝐶𝑖𝑗 = Variable cost per hectare for cropping season i and farm size j. This include cost of fertilizer, seeds, labour, transportation, storage and land rental costs for cropping season i on a per hectare basis. 𝐹𝐶𝑖 = fixed cost per hectare (machine ownership, land ownership, farm insurance and utility). 𝐻𝑖 = The number of hectares for cropping season i. 3.5.1 Estimation of Treatment effects of Outgrower Scheme on Productivity, Postharvest Loss and Profitability In analysing the effect of SOGS on productivity, postharvest loss and profitability, impact evaluation assessment approach is adopted. Assessing the gains of using non-experimental observations has received criticisms for either over estimation or underestimation (Hausman, 1978; Heckman, 1979). The Criticisms are due to difficulty of controlling for observed and unobserved characteristics of the beneficiaries that might influence the outcomes of the intervention apart from the intervention itself. The reason being that, one cannot observe the gains of those who participated had they not participated (the counterfactual outcome) on most methodologies used. 56 University of Ghana http://ugspace.ug.edu.gh In experimental projects, these problems are addressed using randomization. The SOGS participants are not randomly selected but rather, the outgrowers self-select to participate based on available information. The outgrowers and non-outgrowers may have differences in terms of both exogenous and endogenous characteristics that may or may not influence participation. Some of the known econometric approaches employed include Propensity Score Matching (PSM), Inverse Probability Weighted Ratio Adjusted (IPWRA) and Endogenous Switching Regression method (ESRM). The essence of using PSM is to compare observable characteristics on treatment units with similar characteristics on the untreated units. Once the units are matched based on observed characteristics, PSM conclude no significant differences in their observable characteristics and therefore results obtained by comparing the two units is consistent and unbiased (Gitonga et al., 2013; Rosenbaum & Rubin, 1983; Wooldridge, 2002). The PSM can still generate biased results in the presence of mis-specification in the propensity scores. The potential remedy for such mis-specification is to use IPWRA. Regardless of matching techniques to adjust for mis-specification, it can only overcome selection bias due to observable characteristics. However, when the causes of selection bias are due to unobserved endogeneity such as farmer's inherent skill, the results based on matching techniques will still be biased (Asfaw et al., 2012). The ESRM approach that accounts for both observed and unobserved biases by comparing counterfactual outcomes is considered superior for impact evaluation (Asfaw et al., 2012; Donkor & Owusu, 2019; Lokshin & Sajaia, 2018). This study used ESRM to estimate treatment effects whiles PSM is used as complementary for robustness check. 57 University of Ghana http://ugspace.ug.edu.gh 3.5.1.1 Propensity Score Matching (PSM) The PSM is a non-parametric method which constructs statistical similar group by modelling the probability of participating in an intervention on the basis of similar observed characteristics of the treated and untreated units that are unaffected by the intervention (Hausman, 1978; Heckman, 1979). Once the units are matched based on observables characteristics, PSM assumed no significant differences in the observable characteristics after the match and therefore, measure the average differences in the outcomes of the two comparable units (Caliendo, 2005; Gitonga et al., 2013; Khandker et al., 2010; Rosenbaum & Rubin, 1983) Propensity score (P(X) = Pr (D = 1|X) = E (D |X)) is defined as the probability that a unit in the combined sample of treated and untreated receives the treatment, given a set of observed variables (French & Popovici, 2011; Rosenbaum & Rubin, 1983) There are two assumptions that underlie the implementation of PSM (Caliendo, 2005; Rosenbaum & Rubin, 1983). These include conditional independence or unconfoundness assumption and common support assumption. The unconfoundness assumption implies that potential outcomes are independent of the treatment status assuming set of observed X covariates are controlled. The unconfoundness assumption is denoted as (Y ,Y )  D / X (3.8) 1 0 Where  denotes conditional independence. Potential outcome (sorghum crop productivity, farmers’ postharvest loss reduction, profitability and vulnerability to climate change) of participation in the outgrower scheme is conditionally independent. This implies that the differences in outcome between participants in OGS and not participating is due to the effect of the SOGS (Caliendo, 2005). The common support condition implies that for individuals 58 University of Ghana http://ugspace.ug.edu.gh with the same value of X, there is a probability of either being in the SOGS or not being in the SOGS. This is specified as: 0 1 𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 { (3.10) 𝑦0𝑖 = 𝛽 ∗ 0𝑥0𝑖 + 𝜀0𝑖 𝑖𝑓 𝐷𝑖 ≤ 1 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 60 University of Ghana http://ugspace.ug.edu.gh In equation 3.9, 𝑦1𝑖 and 𝑦0𝑖 are the welfare gains for treatment and control groups respectively. The 𝑥1𝑖 and 𝑥0𝑖 represent a vector of exogenous factors that are thought to influence participation. The 𝛽1𝑎𝑛𝑑 𝛽0 are the parameters to be estimated. 𝜀0𝑖 𝑎𝑛𝑑 𝜀1𝑖 and 𝑢𝑖 (in the decision equation) are the stochastic disturbance from the outcome equations and selection equation. These variables are assumed to have a trivariate normal distribution, with mean vector zero and non- singular covariance matrix (Maddala, 1983). 𝜎2𝜀1 . 𝜎𝜇1𝜇 Cov (𝜀1𝑖,𝜀2𝑖,𝜇)= [ . 𝜎 2 𝜀2 𝜎𝜇2𝜇] (3.11) 𝜎 2𝜇𝜀1𝜇 𝜎2𝜇𝜀 𝜎𝜇 𝜎2𝜀1 and 𝜎 2 𝜀2 are variances of the stochastic disturbance terms in the treatment functions in equation 3.11. 𝜎2𝜇 is the variance of the stochastic disturbance term in the selection equation. 𝜎𝜀1𝜀 represents the covariance of the stochastic disturbance. while 𝜎𝜇𝜀1𝜇 is the covariance of 2 𝜀1𝑖and 𝑢𝑖.𝜎 is the covariance of 𝜀2𝑖 𝑎𝑛𝑑 𝑢𝑖 . The covariance between 𝜀1𝑖and 𝜀2𝑖 is not 𝜇2𝜇 defined. 𝑦1𝑖 and 𝑦0𝑖 from equation 3.10 are not determined simultaneously and it is assumed that 𝜎2𝑢 = 1 because is estimable only up to a scalar factor (Maddala, 1983). A useful implication of the error structure is that the stochastic disturbance terms from the equations (3.10) are correlated with the stochastic disturbance term in the selection equation. Therefore, expected values of the stochastic disturbance terms from the functions in equation (3.10) conditioned on sample selection are not equal to zero as shown below ∅(𝛽𝑥 𝐸(𝜀 / 𝐷 𝜎 𝑖) ∅(𝛽𝑥𝑖) 1𝑖 𝑖 =1 )= 𝜀1𝑢 = 𝜎𝜃(𝛽𝑥 ) 𝜀1𝑢𝛿1𝑖 where 𝛿1𝑖 = (3.12) 𝑖 𝜃(𝛽𝑥𝑖 ) ∅(𝛽𝑥𝑖) ∅(𝛽𝑥𝐸(𝜀1𝑖 / 𝐷 𝑖) 𝑖 =9 )= 𝜎𝜀2𝑢 = 𝜎𝜀2𝑢𝛿2𝑖 where 𝛿2𝑖 = (3.13) 1−𝜃(𝛽𝑥𝑖 ) 1−𝜃(𝛽𝑥𝑖 ) 61 University of Ghana http://ugspace.ug.edu.gh If the estimated 𝜎 𝜀1𝑢 and 𝜎 𝜀2𝑢 are statistically different from zero, the null hypothesis of absence of self-section is rejected. This suggests that the decision to participate in SOGS and the outcome variable are correlated, thus, existence of selection bias. Conditional expectations, treatment and heterogeneity effects Following Asfaw et al. (2012) and Donkor & Owusu (2019), the ESRM can compare expected effect of SOGS on productivity, postharvest loss reduction, profitability and vulnerability to climate of treatment group (a) in equation 3.14 with the effect on the control group (b) in equation 3.15 and to examine the expected outcomes of the same variables in the counterfactual hypothetical cases that the treatment group did not participate (c) in equation 3.16, and that the control group participated (d) in equation 3.17. The conditional expectations for the outcome variable in the four cases are presented in Table 3.1 and defined as follows: 𝐸(𝑌1𝑖|𝐷𝑖 = 1, 𝑋1𝑖)=𝛽1𝑥1𝑖+𝜎𝜀1𝑢𝛿1𝑖 (3.14) 𝐸(𝑌2𝑖|𝐷𝑖 = 0, 𝑋2𝑖)=𝛽2𝑥2𝑖+𝜎𝜀2𝑢𝛿2𝑖 (3.15) 𝐸(𝑌1𝑖|𝐷𝑖 = 0, 𝑋1𝑖)=𝛽2𝑥1𝑖+𝜎𝜀1𝑢𝛿2𝑖 (3.16) 𝐸(𝑌2𝑖|𝐷𝑖 = 1, 𝑋2𝑖)=𝛽1𝑥2𝑖+𝜎𝜀2𝑢𝛿1𝑖 (3.17) Table 3.1: Treatment Effect Sub sample Decision Stage Treatment effects Treatment Control Farmers who participated (a) (c) On the treated 𝐴𝑇𝑇𝑖 in SOGS 𝐸(𝑌1𝑖|𝐷𝑖 𝐸(𝑌2𝑖|𝐷𝑖 = 1) = 1) Farmers who do not (d) (b) On the untreated (𝐴𝑇𝑈𝑖) Participate in SOGS 𝐸(𝑌1𝑖|𝐷𝑖 𝐸(𝑌2𝑖|𝐷𝑖 = 0) = 0) Heterogeneity effects 𝐵𝐻1𝑖 𝐵𝐻2𝑖 TH Source: Asfaw (2012) 62 University of Ghana http://ugspace.ug.edu.gh The outcomes (a) and outcome (b) represent observed outcomes while outcome (c) and outcome (d) represent their respective counterfactual expected outcomes. 𝐷𝑖 = 1 if smallholder sorghum farmer i participate in SOGS and 𝐷𝑖 = 0 otherwise. 𝑌1𝑖 = sorghum farmers productivity, profitability, postharvest loss reduction or vulnerability of the sorghum farmer i participating in SOGS. 𝑌2𝑖= sorghum farmer productivity, profitability, postharvest loss or vulnerability to climate if the sorghum farmer i did not participate in SOGS. 𝐴𝑇𝑇𝑖= the effect of the treatment (outgrowers) on the treated (outgrowers) 𝐴𝑇𝑇𝑖=E(𝑌1𝑖-𝑌2𝑖 |𝐷𝑖=1)= 𝛽1𝑥1𝑖+𝜎𝜀1𝑢𝛿1𝑖- 𝛽1𝑥2𝑖-𝜎𝜀2𝑢𝛿1𝑖= 𝛽1(𝑥1𝑖-𝑥2𝑖)-( 𝜎𝜀1𝑢-𝜎𝜀2𝑢) 𝛿1𝑖 (3.18) In similar manner, the effect of SOGS of the untreated (ATU) (non-outgrowers) is calculated as the difference between (d) and (b), 𝐴𝑇𝑈𝑖=E(𝑌1𝑖-𝑌2𝑖 |𝐷𝑖=0)=𝑥2𝑖(𝛽𝑖-𝛽2)+(𝜎𝜀1𝑢-𝜎𝜀2𝑢) 𝛿2𝑖 (3.19) Heterogeneity effect: The farmers on SOGS may perform better than those who did not participate due to unobservable characteristics such as their skills, information and business orientation but not necessarily due to their participation in the SOGS (Asfaw et al., 2018; Asfaw et al., 2012; Carter & Milon, 2005; John Ng’ombe, 2013; Lokshin & Sajaia, 2018). Adapting Carter & Milon (2005) method in this case, “the effects based on heterogeneity” for participating farmers (𝐵𝐻1𝑖), can be defined as differences between (a) and (d) in Table 3.1. as: 𝐵𝐻1𝑖= E( 𝑌2𝑖 |𝐷𝑖=1)-E(𝑌1𝑖 |𝐷𝑖=0)= 𝛽1(𝑥1𝑖-𝑥2𝑖) + 𝜎𝜀1𝑢( 𝛿1𝑖-𝛿2𝑖) (3.20) 63 University of Ghana http://ugspace.ug.edu.gh For the effect on no-participating farmers (𝐵𝐻2𝑖) in the SOGS, the heterogeneity effect is the difference between (c) and (b) as: 𝐵𝐻2𝑖= E( 𝑌2𝑖 |𝐷𝑖=1)-E(𝑌2𝑖 |𝐷𝑖=0)= 𝛽2(𝑥1𝑖-𝑥2𝑖) + 𝜎𝜀2𝑢( 𝛿1𝑖-𝛿2𝑖) (3.21) Transitional Heterogeneity (TH) is calculated to understand the level of effect of SOGS on those who actually participated and those who did not or in the counterfactual case that they choose to participate. The TH is the difference between equations (𝐴𝑇𝑇𝑖) and 𝐴𝑇𝑈𝑖. 3.6 The Effects of Sorghum Outgrower Scheme on Vulnerability to Climate Change The study adopted the Livelihood Vulnerability Index (LVI) framework employed by the Intergovernmental Panel on Climate Change (IPCC) for analysing the vulnerability to climate change on both treatment and control groups. The IPCC define vulnerability as a function of exposure, sensitivity and adaptive capacity (Hahn et al., 2009). Vulnerability indicators offer useful means of recognising and monitoring vulnerability over time and identifying contributing factors to vulnerability and livelihoods outcomes (Bhattacharjee & Behera, 2018; Djalante, 2019; Hahn et al., 2009). They provides a useful suggestions on classification of vulnerability into: adaptive capacity, sensitivity, and exposure (Shah et al., 2013). Adaptive capacity defines ability of a system to adjust to expected or actual climate stresses and to cope with the consequential effect. Adaptive capacity is a function of wealth, information, infrastructure, technology, education, access to resources, skills, stability and management capabilities (Adu et al., 2018). Sensitivity denotes the degree to which a system responds to changes in climate variables. Exposure refers to the degree of climate related stress in a particular place which could be 64 University of Ghana http://ugspace.ug.edu.gh either a long-term change in climate conditions, or changes in the climate variability (Shah et al., 2013). The Livelihood Vulnerability framework is therefore relevant to understand vulnerability to climate change because it provides a framework for analysing both the sub-components and major components that make up livelihoods and the contextual factors influencing them. Following IPCC definition of vulnerability developed by Hahn et al. (2009), seven major components were listed. They are “socio-demographic profile, livelihood strategies, social networks, health, access to food, access to water and natural disasters and climate change”. Each of these components are made up of sub-components measured on a different scale. The LVI has been used by different authors to estimate SHF vulnerability to climate change (Adu et al., 2018; Bhattacharjee & Behera, 2018; Etwire et al., 2013; Hahn et al., 2009; Shah et al., 2013). Having established vulnerability of the treatment and control group to climate change, the study then adopted the ESRM 3.5.1.2 and the PSM method for robustness test in 3.5.1.1 above to determine the effect of SOGS on reducing the vulnerability to climate change on both groups. 3.7 Method of Data Analysis The data collected were analysed both quantitatively and qualitatively and integrated at the report writing stage in line with principles of mixing data. The quantitative data which was captured electronically using CAPI was transferred onto a platform and analysed using STATA. Qualitative data from interviews and focus group discussions were transcribed and analysed using NVivo Pro 11 software. In the report writing stage, quotations were used to emphasise key statements. 65 University of Ghana http://ugspace.ug.edu.gh 3.7.1 Background of the Study Respondents Descriptive statistics were used to explain the background of the study respondents. Variables describing the outgrowers and non-outgrowers were generated and mean differences estimated using cheque squire test for categorical variables and t-test for the continuous variables. The first set of analyses examined the differences in means of farmers on outgrower scheme and non-outgrower members. The main areas of interest were their demographic characteristics, socio-economic characteristics, farm characteristics and other observable features. Information generated was presented in tables and diagrams and the mean differences between the two groups explained. A t-test was conducted and the significance of their mean differences was based on p<0.05. Variables which were found to be significant were thoroughly discussed. 3.7.2 Factors Influencing Smallholder Farmers Participation in Outgrower Scheme Probit regression analysis was conducted to determine the factors influencing SHF participation in SOGS. The dependent variables which is participating in sorghum OGS is dichotomous (1 = Sorghum Outgrower, 0= Sorghum farmers not under the Outgrower Scheme). The probit model assumes existence of "latent" dependent variable, in this case the probability of participating in SOGS (Maddala 1988). Consider the dependent variable 𝑦𝑖, 1 𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑃𝑖 𝑦𝑖 = { (3.22) 0 𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 1 − 𝑃𝑖 The probability of participating in the sorghum outgrower scheme, 𝑃𝑖, is assumed for any observation i by the probit function. 66 University of Ghana http://ugspace.ug.edu.gh 𝑌∗=𝛽 𝑋𝑖 +Ɛ, Ɛ ~ N (0, 𝜎2 ) (3.23) Where, 𝑌∗= is the OGS 𝑋𝑖 is the independent variable, 𝛽 is the coefficient of estimation, Ɛ is the error term. The Empirical Model of the factors influencing sorghum farmers’ participation in SOGS is specified as: 𝑂𝐺𝑆 = 𝛽0 + 𝛽1(𝐺𝑒𝑛) + 𝛽2(𝐴𝑔𝑒) + 𝛽3(𝐸𝑑𝑢) + 𝛽4(𝐻𝐻𝑠) + 𝛽5(𝑀𝑎𝑟𝑖) + 𝛽6(𝑂𝑐𝑐𝑢) +𝛽7𝐸𝑥𝑝) + 𝛽8(𝐻𝐻𝑆𝑡𝑎𝑡𝑢𝑠) + 𝛽9(𝐶𝑟𝑒𝑑) + 𝛽10(𝐹𝐵𝑂) + 𝛽11(𝐸𝑥𝑡) + 𝛽12(𝐹𝑠𝑖𝑧𝑒) +𝛽13(𝐶𝑜𝑚𝑦𝑒𝑎𝑟) + 𝛽14(𝐷𝑖𝑠𝑡) + 𝛽15(𝐿𝑒𝑎𝑑𝑒𝑟) + 𝜀 (3.24) Where OGS= Participation in the sorghum outgrower scheme which is the dependent variable of the regression. 𝛽0 =The intercept (constant), 𝛽0 𝑡𝑜 𝛽15 are the Parameters to be estimated  = Disturbances term which is independent, identical normally distributed with zero (0) mean and constant variance Ɛ ~ N (0, 𝜎2). Having generated the results, only variables which 67 University of Ghana http://ugspace.ug.edu.gh were found to be statistically significant at P < 0.05 were discussed and their prior expectation interpreted. 3.7.2.1 Variables, Definition, Measurement Criteria and Prior Expectation Table 3.2 contain variables influencing participation in the SOGS. Gender of a farmer has influence on participation in new agriculture intervention. In SSA, women and men are engaged in agricultural activities but women appear to be disadvantage in access to necessary productive resources to pursue activities with high-returns such as cash crop farming due to economic, social, physical and cultural barriers (Ali et al., 2016; Anderson, 2019; Chigbu, 2019; Lambrecht, 2016; Lipton & Saghai, 2017; Wineman & Liverpool-Tasie, 2017). Given that sorghum is a cash crop in northern Ghana and also given that women are disadvantaged in family resources and decision making, men are more likely to dominate in the sorghum outgrower scheme (Lambrecht, 2016; Mignouna et al., 2011; Mwangi & Kariuki, 2015). This study anticipates less women to participate in OGS compared to men. Gender is measured as dummy. Educational level of a farmer increases his ability to obtain, process and use information relevant to adoption of a new technology (Mignouna et al., 2011; Mwangi & Kariuki, 2015). For instance, study by Okunlola, Oludare, & Akinwalere (2011) on adoption of new technologies by fish farmers and Ajewole (2010) on adoption of organic fertilizer found that the level of education had a positive influence on adoption. This study postulates that the influence of formal education on participation would be positive. Level of education was measured by years of formal school completed starting from primary school class six. 68 University of Ghana http://ugspace.ug.edu.gh Table 3.2 :Variables Influencing Participation in Sorghum Outgrower Scheme Variable Description Measurement Prior expectat ion Demographic characteristics Gender Sex of the farmer Dummy (male=1, 0=female +/- Age Years of the farmer Number Edu Educational status Years in formal school completed +/- Labor Adult household size people Years + age (16 to 60) Mari Married or not married Dummy (marry=1, 0=not marry +/- Occu The main occupation of farmer Farmer=1, 0=otherwise + Exp Number of years of farming Years + Institutional factors Cred Access to credit Amount of credit received in the last 3 years +/- FBO Belonging to farmer Dummy (1=yes, 0=no +/- organization EXT Number of extensions visit number of days in season + Farm characteristic Fsize Number of sorghum farms own Hectares + Locational effect Dist Distance from homestead to Kilometres + main market Other factors ComYear Number of years stay in the Years +/- community Leadership Leadership in social Dummy (yes= 1, no=0) + organization Source: Author compilation (2018) Age has negative influence in farmers’ decision to participate in OGS and was measured by how old the farmers are in years (Ajewole, 2010; Al-Hassan et al., 2013; John Ng’ombe, 2013; Koira Alemayehu Konde, 2014; Mauceri et al., 2005; Udimal et al., 2017). Older farmers are assumed to have gained knowledge and experience over time and are better able to evaluate technology information than younger farmers will do. This wealth of information placed them in a position to make well informed decisions (Mignouna et al., 2011). Al-Hassan, Egyir, & Abakah (2013); Mauceri et al. (2005) and Mignouna et al. (2011), studies suggest that as farmers grow older, their risk aversion increase and they develop less interest in new areas of investment. Younger farmers on the other hand are typically less risk-averse 69 University of Ghana http://ugspace.ug.edu.gh and are more willing to try new technologies. Adoption of genetically modified maize for instance increased with age for younger farmers as they gained experience and increased their stock of human capital but declined with age for those farmers closer to retirement (Mwangi & Kariuki, 2015). Al-Hassan, Egyir, & Abakah, (2013) report on impact of ICT based market information shows that, for a unit increase in age, the likelihood of participation decrease by 0.5. Age in this study was measured by how old the farmers is and is predicted to have either negative or positive influence on decision to participate in the SOGS. For the availability of farm labour, household size is used as a proxy. A larger household size is likely to have access to family labour (Ajewole, 2010; Mignouna et al., 2011; Udimal et al., 2017). It is hypothesized that farmers with more family labour have access to surplus labour. Household with surplus labour are more likely to participate in SOGS (Okunlola et al., 2011). Labour availability was measured by the active population in the household (age 16-60). Farmers who are married are perceived as being more responsible and reliable in rural communities in northern Ghana. Reliable people are likely to be attracted for support such as credit and positively influence their participation (Awunyo-Victor & Al-hassan, 2014). This will have positive influence on their participation than farmers who are single. Marital status is specified as a dummy variable. The main occupation of the respondent is explained by whether farming is the main occupation of the respondent or the respondent is only into farming to get additional income. According to Mariano et al. (2012), people who work in on-farm activities as their main occupation have a higher probability of participation in a new intervention than those on off- farm. This study expects positive relationship between farming as main occupation and probability of participation. 70 University of Ghana http://ugspace.ug.edu.gh A farmer’s experience may have negative or positive influence on participation. With more experience, a farmer can become more or less averse to the risk of OGS depending on the benefits or otherwise of OGS (Adesina & Baidu-Forson, 1995; Ahmed, 2014; Gebrezgabher et al., 2015; Sakha, 2019). Experience is measured by the number of years of farming. Household status refers to whether the respondent is the household head or otherwise. Some study found negative relationship between household head and participation in new intervention (Mustapha et al., 2016; Nkegbe et al., 2017). Due to responsibility of the household head in the study area to provide food for the entire family, this study anticipates negative relationship between household head and probability of participation. Access to credit has been reported to stimulate technology adoption (Awunyo-Victor & Al- hassan, 2014). For SOGS, farmers with access to credit are able to improve their input access and meet contract requirements such as cost of fertilizer and improved seeds (AGRA, 2016; Awunyo-Victor & Al-hassan, 2014). Access to credit is hypothesised to have either negative or positive influence on decision to join SOGS. The unit of measurement was amount of credit received in the last three years. FBO membership is likely to have positive influence on participating in SOGS (Ajewole, 2010; Feleke et al., 2017; ISSER, 2012; Maertens & Velde, 2017; Mwangi & Kariuki, 2015). Belonging to a social group enhances social capital allowing trust, ideas and information exchange (Mignouna et al., 2011; Mwangi & Kariuki, 2015). Mwangi & Kariuki (2015) suggest that, the benefit of social network should not be underestimated and indicated that in the particular context of agricultural innovation, farmers learn from each other and the information from peer group may influence their adoption decision (Mwangi & Kariuki, 2015). Studying the effect of community-based organization in adoption of corm-paired 71 University of Ghana http://ugspace.ug.edu.gh banana technology in Uganda, Mwangi & Kariuki (2015) indicated that farmers who participated in community-based organizations were more likely to adopt the technology. Belonging to FBO is a dummy variable and is expected to influence decision to participate in SOGS positively. Access to extension services has also been realise as key factor influencing farmers decision to adopt agricultural programme and for that matter, participating in OGS (Asfaw et al., 2012; Azumah et al., 2016; Feleke et al., 2017; ISSER, 2012; MoFA, 2017b). Farmers are usually informed about the existence of new project as well as the effectiveness of a programme through extension agents. Access to extension agent acts as a link between the supplier and the farmer in identifying farmers for new project (Azumah et al., 2016; Doss & Morris, 2001). The influence of extension agents can counter balance the negative effect of lack of years of formal education in the overall decision to adopt (Feleke et al., 2017). The number of visits per farming season was used as proxy to access to extension services in this study. The effect of farm size on adoption of a new programme is anticipated to have positive relationship with adoption decision (Baiyegunhi et al., 2019; Delbridge et al., 2013; Ghimire et al., 2015). Years of stay in a community has similar effect on participation as the age of the farmer. Long stay in the community could be associated with comparatively, older age. The longer the farmers stay in the community, the lesser he/she is likely to participate in the SOGS (Sakha, 2019) . Distance to the main market is used as a proxy for access to market. In a situation where there is ready market for sorghum without SOGS, farmers participation in OGS may reduce (Ebata & Hernandez, 2017b; Opoku, 2012; Villano et al., 2019). Ready market will be measured by distance of farm to the nearest market. It is hypothesized that farmers whose farms are far 72 University of Ghana http://ugspace.ug.edu.gh from the main market are likely to participate in SOGS than those closer (Kiaya, 2018; Okunlola et al., 2011; Tanellari et al., 2011). This study hypothesized positive relationship between distance to market and probability of participation in SOGS. Distance to market is measured by kilometres from homestead to the nearest market. Leadership in social organization is said to have positive relationship with participation. According to Ton et al. (2015), community leaders are usually the point of call for any new intervention and are sometimes, made part of the selection process in most rural communities and are more likely to participate. This study anticipates positive relationship between holding a leadership position in the community and the probability of participation. 3.7.3 The Effects of Outgrower Scheme on Productivity, Postharvest Loss Reduction and Profitability Two approaches were adopted to achieve this objective. Sorghum productivity, postharvest loss and profitability was first estimated for both treatment and control groups without controlling for possible selection biased. This results only provide information on the outcome of SOGS but relying on such information to conclude on effects of the SOGS without addressing selectivity bias could be misleading due to observed and unobserved heterogeneity and endogeneity of individual farmers that could influence their participation decision and the outcome variables. For instance, if participating farmers are naturally business oriented leading to their decision to participate in the SOGS and obtained higher profitability than non-participant, it is likely they would have equally performed better without being on the scheme. Concluding on the 73 University of Ghana http://ugspace.ug.edu.gh effects of the SOGS based on only the outcome variable could be misleading (Heckman, 1979). To control for possible selection biases, endogenous switching regression model (ESRM) was used to estimate average treatment effect. The ESRM results are interpreted based on counterfactual outcomes which addresses both endogeneity and selectivity bias while PSM is conducted as a robustness check. 3.7.3.1 Productivity and Profitability Analysis Following Wünsch et al. (2012), Delbridge et al. (2013), Montgomery et al. (2017 and Devkota et al. (2019), Sorghum productivity and profitability was calculated using total variable cost, area of production and price of sorghum in the 2017 farming season. The components of total variable cost ranges from cost of renting of land, cost of inputs and cost of hire labour and/or cost of family/communal labour (Montgomery et al., 2017; Devkota et al., 2019). Labour cost ranges from cost of land preparation, planting, fertilizer application, harvesting, carting, threshing/winnowing and cost of bagging and storage (Arslan et al., 2017). Information on the cost was obtained during focus group discussion, expert interviews and the survey interviews for members on the outgrower scheme and non-members of the schemes. The information provided was further confirmed from buyers and district Agricultural Extension Officers and from Guinness Ghana Brewery Limited. Unit of measurement was the new Ghana Cedis (GHS). The price of sorghum was calculated in GHS/kilogram. The reference year was 2017 major farming season since there is only one cropping season for sorghum in the study area. Yields were estimated in kilogram per hectare. 74 University of Ghana http://ugspace.ug.edu.gh Using STATA version 15, the Average revenue, variable cost, productivity and profitability was calculated separately and compared for farmers on the treatment group and those on the control group. 3.7.3.2 Estimating Postharvest Loss for the Various Stages The various postharvest loss stages are: losses during harvesting; during carting and heaping; during drying; during threshing/winnowing and during storage (Basavaraja et al., 2007; Sheahan & Barrett, 2016). Average quantity loss and percentage loss were calculated along the various stages. The losses in the various stages was accumulated as total percentage loss Estimating the Average Quantity Loss: in estimating the average quantity loss, quantity of sorghum (j) held (tq) and lost (qi) at each stage of the postharvest handling activities was specified by the respondents. Sorghum (j) held at the beginning of the postharvest management chain is expressed on per unit basis (50 kg bag of sorghum). The average quantity (TQij) lost, given n number of respondents at each ith stage of the chain is given by:  tq TQij  n 3.25 Where TQij = Average quantity of sorghum loss tq = Quantity of sorghum loss n = Number of respondents Average market value of Loss: The market value of sorghum loss is estimated along each stage of the postharvest handling activities. The market value of sorghum loss is estimated as the 75 University of Ghana http://ugspace.ug.edu.gh product of average per unit and average quantity loss of sorghum (j) at each ith level. The market value of sorghum loss at each ith level is given by: V  PiQij 3.26 Where: V = Market value of sorghum loss at the ith stage of the postharvest handling activities P = average price per unit of sorghum at the ith stage of the postharvest handling activities i Qij = Average quantity of sorghum lost at the ith stage of the postharvest handling activities Average percentage loss along the chain: Percentage loss of sorghum is estimated as the ratio of mean quantity lost to initial mean quantity held as a percentage at each stage of the postharvest handling activities.   Q ij %TQL   *100 TQ    ij  3.27 Where %TQL= Total percentage post-harvest loss of sorghum Qij = average quantity lost at the ith stage of the marketing channel TQij = Average quantity of sorghum held Following (Asfaw, 2012), we implored ESRM to estimate treatment effect of SOGS on postharvest loss reduction and used PSM for robustness check explained above. 76 University of Ghana http://ugspace.ug.edu.gh 3.7.4 Determining the effects of Outgrower Scheme on Smallholder Farmers Vulnerability to Climate Change The approach to calculate the effect of SOGS on SHF vulnerability to climate are in two folds. The first approach was estimating SHF vulnerability to climate change using the IPCC-LVI. The second approach applied ESRM to estimate the treatment effect and PSM technique used for robustness check as explained on estimating treatment effect on productivity and profitability above. 3.7.4.1 Estimating Vulnerability Level of Smallholder Farmers in the Study Area The Hahn et al, (2009) identified seven major components namely: “Socio-demographic profile, livelihood strategies, social networks, health, access to food, access to water and natural disasters and climate change” as contributing factors to vulnerability to climate change. Each component is made up of several indicators or sub-components, each of which is measured on a different scale (Hahn et al., 2009). The sub-components are standardized using equation 3.28: S Index  h Smin shi 3.28 SmaxSmin Where 𝑆ℎ is the observed sub-component of indicator for household, smin and smax are the minimum and maximum values respectively, after each is standardized, the sub-component indicators are averaged using the index of each major component: n  indexi1 shiM  3.29 h n 77 University of Ghana http://ugspace.ug.edu.gh According to Hahn et al., (2009) “𝑀ℎ represent one of the seven major components [Socio- Demographic Profile (SDP), Livelihood Strategies (LS), Social Network (SN), Health (H), Food (F), Water (W), or Natural Disaster and Climate Variability (NDCV)] for household, h ; 𝑖𝑛𝑑𝑒𝑥𝑠ℎ represents the sub-components, indexed by i , that make up each major component, and 𝑛 is the number of sub-components in each major component. Once values for each of the seven major components for a household are calculated, they are averaged to obtain the household-level LVI” as: 7  wM Mh LVI  i1 i i h 3.30 7  wMii1 This can also be expressed as w LVI  SDP SDPh  wLS LSh  wH Hh  wSN SNh  wF Fh  wWWh  wNDC NDCh h wSDP  wLS  wH  wSN  wF  wW  wNDC 3.31 The weights of each major component, 𝑤𝑀𝑖, are determined by the number of sub- components that make up each major component and are included to ensure that all sub- components contribute equally to the overall LVI. The LVI was scaled from 0 low vulnerable to 1 extremely vulnerable (Hahn et al., 2009). 78 University of Ghana http://ugspace.ug.edu.gh Table 3.3: Major and Sub-Component for Natural Disasters and Climate Change “Major Sub-components Measurement Components Water Percent of household reporting water conflict Percent Percent of household that utilize a natural water source Percent Average time to water source Minutes Percent of household that do not have a consistent water supply Percent Inverse of the average number of waters stored per household 1/litres Socio- Dependency ratio Ratio demographic profile Percent of female- headed household Percent Average age of female-headed household Years Percent of household where head of household has not attended school Percent Percent of households with orphans Percent Livelihood Percent of households with family member working in a different community Percent strategies Percent of households dependent solely on agriculture as a source of income Percent Average agricultural livelihood diversification index 1/# livelihood Social Network Average receive: give ratio Ratio Average borrow: lend money ratio Ratio Percent of households that have not gone to their local government for percent assistance in the past 12 months” Source: Hahn et al. (2009) 79 University of Ghana http://ugspace.ug.edu.gh Table 3.4: Major and Sub-Component for Natural Disasters and Climate Change “Major Sub-components Measurement Components Health Average time to health facility Minutes Percent of households with family member with chronic illness Percent Percent of households where a family member had to miss work or school in Percent the past 6 months due to illness Average malaria exposure*prevention Month*Bednet indicators Food Percent of households dependent solely on the family farm for food Percent Average number of months household struggle to find food Number Average crop diversity index 1/# crops Percent of households that do not save crops Percent Percent of households that do not save seeds Percent Natural Disaster Average number of flood and drought events since 2000 Count and Climate variability Percent of households that did not receive a warning about natural disaster Percent Percent of households with an injury or death as a result of flood or drought Percent since 2000 Mean standard deviation of monthly average maximum daily temperature Celsius since 1983 Mean standard deviation of monthly average minimum daily temperature Celsius since 1983 Mean standard deviation of monthly average precipitation since 1983 Millimeter” Source: Hahn et al. (2009) The climate variability is measured by the average standard deviation in monthly minimum and maximum temperatures and monthly rainfall over 30-year period (Hahn et al., 2009). When the LVI for the major component computed and the sub-components are calculated for 80 University of Ghana http://ugspace.ug.edu.gh the treatment and control groups, student t-test (2-tailed) was employed to test whether there is a significant difference between the means of the LVI or not 3.7.4.2 Using LVI-IPCC Framework to Estimate Vulnerability Hahn et al. (2009) further developed an alternative method for assessing vulnerability by estimating contributing factors to vulnerability in Table 3.5. This approach reorganised the seven major components of vulnerability into Exposure, Sensitivity and Adaptive Capacity. Exposure is measured by the number of natural disasters that have occurred and climate variability. Adaptive capacity is measured by the demographic profile, the types of livelihood strategies employed and the strength of social networks. Sensitivity is measured by the state of food, water and health status. Table 3.5: Factors Contributing to Vulnerability Contributing factors to climate change Major component I. Exposure Natural disasters and climate variability II. Adaptive Capacity 1. Socio-Demographic Profile 2. Livelihood Strategies 3. Social Networks III. Sensitivity 1. Health 2. Food 3. Water Source: Hahn et al., (2009). The equation 3.49 explains how LVI-IPCC is calculated. ∑𝑛 𝐶𝐹 = 𝑖=1 𝑊𝑀𝑖𝑀𝑑𝑖 𝑑 𝑛 3.32 ∑𝑖=1 𝑊𝑚𝑖 81 University of Ghana http://ugspace.ug.edu.gh Where 𝐶𝐹𝑑 is an IPCC-defined contributing factor (Exposure, Sensitivity and Adaptive Capacity) for community d. 𝑀𝑑𝑖 are the major components for community d indexed by i, 𝑊𝑀𝑖 is the weight of each major component, n is the number of major components in each contributing factor. Once Exposure, Sensitivity, and Adaptive Capacity were calculated, the three contributing factors are then combined using equation 3.50 𝐿𝑉𝐼 − 𝐼𝑃𝐶𝐶𝑑 = (𝑒𝑑 − 𝑎𝑑) ∗ 𝑆𝑑 3.33 Where 𝐿𝑉𝐼 − 𝐼𝑃𝐶𝐶𝑑 is the LVI for community d expressed using IPCC vulnerability framework, e is the calculated exposure score for community d (equivalent to the natural disaster and climate variability major component), a is the calculated adaptive capacity scores for community d (weighted average of the socio- demographic Livelihoods strategies, and social network major component), and s is the heath, food and water major components). The LVI-IPCC is then scaled between -1 (least vulnerable) to 1 (most vulnerable). When the LVI-IPCC are calculated for the treatment and control groups, student t-test (2- tailed) was employed to test whether there is a significant difference between the means or not. 82 University of Ghana http://ugspace.ug.edu.gh 3.7.4.2 The treatment effects of Sorghum Outgrower Scheme on Vulnerability to Climate Change Following Asfaw (2012), the ESRM explained in 3.5.1.2 was employed whiles PSM explained in 3.5.1.1. is used as robustness check to determine the effects of SOGS on smallholder sorghum farmers vulnerability to climate change. 3.8 Method of Data Collection 3.8.1 Sources of Data, Instruments and Interview Procedure The main sources of data for the study was secondary and primary data sources. The secondary data source focused mainly on documentary analysis of published and unpublished literature. Time series data from the Ghana Meteorological Agency was also sourced. The variables of interest were rainfall and temperature. This was needed to help establish the changes that have occurred overtime to help in determining vulnerability to climate change. The primary data source targeted stakeholders ranging from sorghum farmers, extension officers, MoFA officers, Sorghum Aggregators and institutions supplying sorghum to GGBL. In view of the limitations of dichotomous qualitative and quantitative approaches, the mixed- method of data collection approach was used to compensate for the inherent limitations of using single method and also, to broaden the scope of the study to include different actors in the sorghum value chain. Irrespective of the challenges associated with such a triangulation of methods such as time constraints, high research costs, and conflicting data from different sources, combination of qualitative and quantitative methods was considered most appropriate for this study. 83 University of Ghana http://ugspace.ug.edu.gh 3.8.2 Sampling Procedure Based on the existence of sorghum outgrower scheme, purposive sampling was first used to select Upper East and Upper West regions of Ghana. Eventhough sorghum is cultivated in Northern Region, Brong Ahafo Region and Volta Region, there are no SOGS actively operating in those regions. At the district level, one district each from the Upper East and Upper West regions were selected purposively based on practice of SOGS. Garu district in the Upper East Region and Jirapa district in the Upper West region were selected. Using the 2017 data of SOGS from Guinness Ghana Brewery Limited, communities in Garu district and Jirapa district that grow sorghum were listed and stratified. In Garu district, 10 communities were on the treatment stratum and 10 on the control stratum. In Jirapa district, 5 communities were on the treatment stratum and 15 communities on the control stratum. Simple random sampling based on lottery approach was then used to select two communities each from each of the stratums. Zari community and Guzeig community in Garu district were selected from the treatment stratum. Bugri and Tubong communities from the control stratum were also in Garu district. In the Jirapa district, Sabuli and Chapuri communities were selected from the treatment stratum and Tizza and Kuncheni communities were selected from the control stratum. A total of 516 households were sampled from both the treatment and control groups for the study using Yamane sampling formula. Having obtained four treatment communities and four control communities from treatment and control communities respectively, the systematic randomization was used to obtain 216 households from the list of 998 sorghum households from the treatment communities. For the control communities, 1,499 households growing sorghum were listed from the district agricultural extension officers in Garu and Jirapa 84 University of Ghana http://ugspace.ug.edu.gh districts combined. Systematic random sampling was used to sample 300 households from the control communities from the list provided. Note that Yamane (1967) formula had already been used to obtain the total sample size for the study. Using the 2010 Population and Housing Census District Analytical Reports in Garu district and Jirapa district, the combined total number of households in the study area was 28,226. Using the Yamane (1967) simplified formula to calculate the sample size at 95% confidence level and assuming maximum variability to be 0.5 (p=5) with a 5% level of precision, the equation is assumed as follows: 𝑁 𝑛 = 2 3.34 1+ 𝑁(𝑒) Where n is the sample size, N is the households in the study communities, and e is the level of precision. When this formula is applied the results is as follows: 28,226 𝑛 = 1 + 28,226(0.05)2 28,226 = 394 households 1+ 71.565 Considering the design of the study (quasi-experimental), the control group was oversampled. A total of 516 respondents were targeted considering the interview period of 60 days. Enumerators over interviewed by 16 in the treatment communities. The treatment sample size is 216 and Control sample size is 300. This sample is considered a good representative sample (Yamane, 1967). 85 University of Ghana http://ugspace.ug.edu.gh Table 3.4: Sampling Procedure Stages Sampling method Sampling approach and sample areas 1 Purposive Two regions in Northern Ghana (Upper East and Upper West) were selected based on sorghum production and existence of SOGS (MoFA, 2018). 2 Purposive Garu and Jirapa districts were selected based on information of sorghum production and existence of SOGS. 3 Stratification Communities were stratified into treatment and control communities in each district. 4 Lottery approach In Garu, Zari and Guzeig were randomly selected from the treatment stratum and Bugri, and Tubong from the control stratum. In Jirapa, Sabuli and Chapuri were selected from the treatment stratum and Tizza and Kuncheni from the control stratum. 5 Systematic random Total list of 998 sorghum households on OGS were obtained sampling from FBOs in the respective treatment communities. From the calculated sample size, systematic random sampling was then use to obtain 216 households 6 Listing For the control communities, 1,499 households were listed. Systematic random sampling was used to obtain 300 households after calculation of sample size was done. To ensure representativeness of sample size in each community, the proportional sampling was applied to obtain sample size for each community. The number of total households obtained was proportionally distributed based on the number of households for each community as shown in the table 3.8 below. 86 University of Ghana http://ugspace.ug.edu.gh Table 3.5: Communities and Number of Sorghum Farmers Sampled Treatment Status Community Sample Size Garu District Treatment Zari 62 Gozeig 58 Control Tubong 131 Bugri 50 Jirapa District Treatment Chapuri 45 Sabuli 51 Control Kuncheni 40 Tizza 79 Total 516 Source: Field Survey (2018) 3.8.3 Interview Procedure The data collection started on 5th August 2018 and ended on 30th October 2018. Prior to the field visit, community leaders (Assembly Members and Chiefs) in the study areas were consulted and informed about the purpose and content of the research. Engaging them helped to get their support in organizing the communities for the interviews. As customary demands, cola-nut and alcoholic drink popularly called, “Akpeteshie” were offered in each community as part of the community entry before explaining the purpose of the visit. Qualitative Interview. The purpose of the qualitative interviews was to understand the farmers’ farming experience; their perception and belief systems about the outgrower scheme; constraints and opportunities in sorghum farming; their relationship with other actors such as buyers and their everyday social worlds and realities. Information generated from the qualitative interviews helped to better understand the sorghum production and the outgrower 87 University of Ghana http://ugspace.ug.edu.gh arrangement which helped to review the survey questionnaire. The two main qualitative data collection method were Focus Group Discussions (FGDs) and in-depth interviews. The Focus Group Discussion (FGD): The purpose of the FGD was to provide natural setting for various categories of farmers to openly discuss issues surrounding sorghum production. In all, eight FGD were carried out. In each community, one FGD was held with each different groups of men, women and the youth. The participants in each group ranged between 7-12 farmers. Treatment groups were interviewed in the treatment communities comprising of only the farmers who participate in the SOGS while control groups were held in the control communities for farmers who grow sorghum but do not participate in the SOGS. The FGDs were held before the in-depth interviews to ensure that important issues were identified for further probing. The FGDs covered several issues including: types of farming practices in the community; the current architecture of agricultural investments; history of sorghum farming in the community; average farm size; knowledge in SOGS; SOGS package; contractual arrangement between farmers and buyers; their farming constraints; productivity issues; harvest and postharvest management practices; climate change experiences over the past 30 years; support given to farmers as part of the SOGS package to reduce vulnerability to climate change and resilience strategies. The role of SOGS in climate change adaptation were intensively discussed. Equally, the perception of non-participants and their view on the SOGS was also interrogated among the control groups. In-depth Interviews: Purpose of the in-depth interviews was to explore information and new themes to improve the survey questionnaire. Semi-structured flexible interview guides were used to solicit information from respondents. Most of the interviews were conducted in the 88 University of Ghana http://ugspace.ug.edu.gh local languages and recorded electronically. Interviews with agricultural staff and GGBL was in English language. All interviews were face to face. The respondents interviewed were officers of District Department of Agriculture, lead sorghum farmers, sorghum buyers and firms that were buying and supplying sorghum for GGBL. Specifically, five farmers who have expert knowledge in sorghum production in each community were interviewed. Two small size sorghum aggregators, two medium aggregators, one large aggregator and one supplier from each of the districts were interviewed. One officer from GGBL was also interviewed to understand the history and their perspectives of the SOGS. Quantitative Interviews: The survey questionnaires were administered by five trained research assistants who understand and could speak the language of the people. For Garu, the predominant languages spoken are Kusaal, Bimoba and Mampruli and for Jirapa, Dagaare and Wale. The well-structured close-ended questionnaires were first computed into Computer- Assisted Personal Interviewing (CAPI) software. The questionnaires were reviewed severally to ensure that, information generated addresses the research objectives. All the questions were thoroughly discussed with the research assistants and translations done in the respective local languages. Pre-testing of questionnaires was conducted to examine the adequacy of the questionnaire, the sequence and how the respondents understood the questions and to evaluate whether the questionnaire would pose any challenge. Pretesting of the Questionnaire: The Pre-testing of the survey instruments is one of the basic requirements for professional research. The experience from the pre-testing was used to improve the final survey questionnaire. The two days pretesting was conducted on the 20th and 21st August, 2018. One district (Garu) was purposively selected for the pre-testing. 89 University of Ghana http://ugspace.ug.edu.gh Worikambo community and Kpatia community were selected to represent treatment and control communities respectively. A total of thirty (30) households producing sorghum were interviewed. Fifteen (15) of the households were from Worikambo where SOGS is practiced and they were randomly selected from a list of outgrowers provided by the Garu-Tempani Farmers Associations. Fifteen (15) non-outgrower households were also selected from Kpatia community randomly according to the willingness to participate in the study. These communities eventhough were in Garu, they were not part of the targeted communities for the main research. The pretesting helped to modify some of the questions. For example, prior to the pretesting, question on labour used: “do you hire labour? The options were yes or no. There was no provision for communal labour. This was reviewed to include communal labour. Time Series Data: time series data on rainfall and temperature was collected from the Ghana Meteorological Agency for the last 30 years. This data was collected from Garu and Hang. Eventhough Hang is not one of the study districts, lack of weather station in Jirapa makes it impossible to get rainfall and temperature information from Jirapa. However, Hang is the nearest district to Jirapa that has weather station. 90 University of Ghana http://ugspace.ug.edu.gh 3.9 Study Area This study was carried out in the Upper East and Upper West regions of Ghana where sorghum is largely produced. While Faranaya, a subsidiary company of Presbyterian Agricultural Station and Akuafo Nketewa Company Limited, a subsidiary company of the Peasant Farmers Association Ghana leads in the aggregating of sorghum in the Upper East Region, Agriaccess Company Limited leads the aggregation activities in the Upper West region. There are other ten smaller aggregators supplying to GGBL. The two regions have the highest level of poverty in Ghana, with 70.4% and 44.4% of the population in Upper West and Upper East Region respectively living below the poverty line in 2013 (Cooke, Hague, & Mckay, 2016; GSS, 2014c, 2014d). Socio-economic Characteristics: Both regions are predominantly rural and depends largely on subsistence farming as their main economic activity (GSS, 2014b, 2014a). The main crops cultivated are maize, rice, groundnuts, millet, sorghum and vegetables. Livestock such as cattle, goats, sheep and pigs are also kept. Poultry, especially domestic fowls and guinea fowls are predominant in the regions (MoFA-SRID, 2016). Weaving, hunting and dry season gardening is popular activities of men during dried seasons, whiles the women engaged in petty trading, backyard gardening and basket weaving (GSS, 2014a, 2014b). Other areas that offer employment to the people are public service, food processing, textile and leather works (GSS, 2014b, 2014a). Geographic Characteristics: The regions fall within semi-arid Guinea and Sudan Savannah zones which correspond to the north-eastern part of Upper East region whilst the Guinea Savannah is made up of the Upper West region and Northern region (GSS, 2014b, 2014a, 2014c). The Guinea Savannah has a short unimodal tropical monsoon, with a mean annual 91 University of Ghana http://ugspace.ug.edu.gh rainfall of 1,100 mm and a rainy season lasting between 180–200 days (GSS, 2014b, 2014a). The Sudan Savannah has a lower mean annual rainfall of 1,000 mm and between 150–160 rainy days. This is followed by 180 to 210 dry days (November to mid-April) characterised by dusty harmattan winds between November and February and a high temperature of 35 degrees centigrade during the day time (GSS, 2014b, 2014a). The long dry season coupled with the inadequate number of irrigable dams contribute to the seasonal migration of the youth from these regions to southern sector of Ghana in search of menial jobs (Cooke et al., 2016). Garu and Jirapa districts have the highest sorghum production in Upper East and Upper West Region respectively (MoFA, 2017a). The treatment communities have the highest concentration of outgrower schemes based on 2016 official data from the district department of agriculture and Guinness Ghana Brewery Limited. 92 University of Ghana http://ugspace.ug.edu.gh Figure 3.3: Map of the Study Area Source: RS/GIS lab (2018) 93 University of Ghana http://ugspace.ug.edu.gh 3.10 Scope and Limitations of the Study The study determines the effects of SOGS on SHF profitability, postharvest loss reduction and their vulnerability to climate change. Treatment group and control group were targeted for the study and their performance in sorghum production were compared. A particularly limitation of this approach is selection bias or farmers may choose to participate on the basis of both observed and unobserved characteristics that could also influence their outcome variables apart from the SOGS. To address the selectivity bias, the ESRM used counterfactual outcomes to compare performances. More robust results could be achieved in future with experimental design of which farmers would be randomly selected to participate. Comparing performance of such farmers with non-participates could provide more robust results. The second limitation is using cross-sectional 2017 data for the study. Given that farmers in the study area relied on rainfall which is beyond their prediction, in the year of bad rains, the results might be different and present a different picture of sorghum performance under SOGS. Using panel data covering different years could provide more superior information. Given the time frame and the resources available, this study could not collect panel data. To compensate for lack of panel data, the qualitative interviews conducted allowed households to recall developments in the previous years. To determine the effects of SOGS on SHF vulnerability to climate change, data on SOGS package that addresses SHF vulnerability to climate change was solicited based on 2017 information by the respondents. Comparing the 2017 performance with changes over 30 years may not provide consistent information. To address the inconsistencies, recall information was solicited during FGD and expert interviews to compensate for lack of panel data. 94 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR RESULTS AND DISCUSSION 4.1 Introduction This chapter is divided into four main sections. The first section highlights the background of the study respondent, farmers perceptions about the outgrower scheme, contractual arrangement and the kind of support received from the SOGS. The second section presents the probit regression results on factors influencing SHF participation in the SOGS. The third section presents the results of sorghum productivity, postharvest loss reduction and profitability for participating in the SOGS and the effects of SOGS on the outcome indicators (sorghum productivity, postharvest losses and profitability). The final section present results of IPCC-LVI for treatment and control groups vulnerability to climate change and the effect of the SOGS on reducing their vulnerability level. 4.2 Background of the Study Respondents Background of the study respondents are in table 4.1 to table 4.3. The Table 4.1 contain household characteristics of variables for treatment and control groups. Table 4.2 contain farm characteristics and Table 4.3 present market, community and socio-economic and political characteristics. 4.2.1 Household Characteristics of the Respondents Table 4.1 contain results of household characteristics of respondents in the study area. The results show differences in means of gender, age, education, non-farm livelihoods and farming experience between treatment and control groups. On gender of the respondents, 68% of the respondents producing sorghum were male farmers. About 71% and 65 % respectively of the male respondents were in the treatment and control groups respectively. This result support 95 University of Ghana http://ugspace.ug.edu.gh several literature on male farmers being likely to engage in cash crop production compared to their female counterparts (Ali et al., 2016; Anderson, 2019; Chigbu, 2019; Lambrecht, 2016; Lipton & Saghai, 2017; Wineman & Liverpool-Tasie, 2017). The finding corroborates Tanellari et al. (2011), Mariano, et al. (2012) and Ragasa et al. (2018) studies which found more male farmers adopting new technology that require land, labour and other resources than female farmers. Table 4.1: Household Characteristics of the Respondents P-values (chi- Variable Treatment Control Overall Difference square/ N= 216 N= 300 N= 516 (means) t-test) Gender (Male %) 71.3 65.3 68.3 6.0 0.153 Age of farmer (number of years) 46.6 50.1 48.4 3.4 0.014** Educational level 0.780 No formal education (%) 63.9 77.7 70.8 -13.8 Basic (%) 19.0 14.3 16.7 4.7 Higher than Basic (%) 17.1 8.0 12.6 9.1 Marital status (% married) 87.0 83.3 85.2 3.7 0.640 Experience in sorghum farming (years in 5.8 9.1 7.5 3.3 0.000*** farming sorghum) Household Size (no. of people in the household) 5.6 6.16 5.9 0.6 0.039 Farm Dominant livelihood (%) 88.9 90.0 89.5 1.1 0.120 Non –farm dominant livelihoods (%) 38.9 13.7 26.3 25.2 0.000*** Length of stay in the Community (years) 0.000*** 28.9 34.6 31.8 5.7 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) 96 University of Ghana http://ugspace.ug.edu.gh According to Abdul-Razak & Kruse (2017), the land tenure systems in northern Ghana favours male farmers. Since sorghum is a cash crop and requires separate land, this could be a barrier that limits participation of the female farmers. Land is a crucial resource in the efforts to bridge gender disparity in most rural areas. In northern Ghana, cultural barriers restrict women from negotiation for land on their own without their husbands. A study by Lambrecht (2016) in Ghana found that women in some rural communities in the Upper East region cannot negotiate land for economic activities on their own without their husbands involvement. Land access by women has serious implication on agricultural activities of women and require policy review to ensure that women can independently own and use their lands in a way that benefit them. Majority of farmers on the treatment group are relatively younger with an average age of 46 years compared to 51 years for those in the control group. The result is similar to Al-Hassan et al. (2013); Maertens & Velde (2017) and Udimal et al. (2017) studies on older farmers less likely to adopt new technology compared to relatively younger ones. Udimal et al. (2017) further asserted that longer period of farming correlate with age and experience, hence, relatively older farmers are most likely to be content with their own ways of farming. As a 68-year old farmer from Bugri community in Garu districts characterises new schemes during an individual face-to-face interview “I am not interested in these things [sorhgum outgrower scheme]. We had similar projects in the 1990s in cotton and it collapsed. I lost a lot of money when the cotton company at the time refused to buy the cotton they themselves asked us to grow. At my age, my concern is how to feed my three wives and children. It is food that matters to me most not money”. 97 University of Ghana http://ugspace.ug.edu.gh Similar sentiments were also expressed by a 64-year old farmer from Tizza community in Jirapa districts during FGD. “I heard the support Pastor Anthony [Anthony is the CEO of Agri-access limited that lead the SOGS in the Jirapa district] is giving to farmers to produce sorghum for him to buy, but I am not interested. Today when you go to Sabuli, all their millet and groundnut farms have been converted for sorghum production. Eventhough sorghum will give you the money, what about food for the family? I have advised my children never to do that because it can lead to hunger in the future. A family is happy when there is enough food in the house. How much money will you get from sale of sorghum to be able to buy food for the family throughout the year?” She concluded. Old farmers’ lack of interest in new schemes could be explained mainly by two factors. First, they are risk averse. This risk aversion is often underpinned by experiences with similar schemes in the past. Perceived unknown long-term beneficial outcomes of new interventions are often ranked high among older people (Sakha, 2019). Secondly, older people perhaps would probably not adopt new approach of farming because they possibly believed that conventional ways of farming are still better (Gebrezgabher et al., 2015). On educational level, both members of the treatment and control groups have low level of education. Overall, 72% of the respondents have never been to school, while 64% and 78% of control and treatment groups respectively have never been to school. About 19% of control group have basic education while 14% of control groups had at least basic education; 17% and 8% of members in the treatment and control groups respectively had education higher than basic school. This results is similar to several literature on low level of education among SHF (Dzanku, 2015; Ghimire et al., 2015; Jolly et al., 2006; Makate et al., 2019; Mariano et al., 2012; Tsinigo & Behrman, 2017). Farmers on the treatment group are slightly less 98 University of Ghana http://ugspace.ug.edu.gh experienced in farming than farmers in the control group with years of farming between 6 years for treatment group and 9 years for the control group. Both groups have average household size of 6. The number of years of farming could correlate with experience in farming, hence confirming the earlier findings of relatively younger people being on the treatment group. On length of stay in the community, control group appears to stay in the community for longer period of 35 years compared with 28 years for the treatment group. Farmers who stayed in the community for long are less likely to participate compared with those who live in the community for relatively shorter period. It is possible those who live in the community for relatively shorter period migrated to those communities for farming activities. Majority of farmers (about 90%) depend mainly on agriculture as their primary source of livelihood. Even though 28% and 14% of treatment group and control group respectively diversified their income sources. This finding also support literature that most SHF depends on agriculture for their livelihoods (Darfour & Rosentrater, 2016; WFP, MoFA, 2012; Yaro, 2013). 4.2.2 Farm Characteristics of the Respondents Table 4.2 contain variables that explain farm characteristics for both treatment and control groups. The average land size reported is 0.8 hectares. The treatment group has relatively larger land size of 1hectare on average compared to 0.7 hectares for the control group. An average of 92% of respondents were farming on family land and 3% farm on land they have purchased. About 2% and 4% respectively from treatment and control groups farm on rented land and also land belonging to others without paying for it. 99 University of Ghana http://ugspace.ug.edu.gh On the type of sorghum seeds cultivated 66% and 33% of dorado and kapaala respectively was cultivated by the treatment group compared to 53% and 9% respectively for control group. About 1% and 2% respectively cultivate Naga white. About 36% of control group cultivate kadaga and no farmer in the treatment group cultivates kadaga. Dorado and Kapaala are new improved sorghum variety preferred by GGBL for brewing. They are also early maturing and high yielding variety, especially the dorado. Kadaga and Naga white are the old sorghum varieties with long maturity period. Both varieties are not preferred for brewing by GGBL. The high percentage of treatment group cultivating dorado and kapaala which is preferred by GGBL confirmed OGS condition of buyers determining the type of produce that farmers should cultivate (Brigitte & Ragasa, 2018). Table 4.2. Farm Characteristics of the Respondents Difference in P-value (T-test /chi- Variable Treatment Control Overall means square) Farm Size (Ha) 1.0 0.7 0.9 0.3 0.066 Land Ownership 0.382 Family land (%) 92.59 91.67 92.1 0.9 Purchased land (%) 1.39 3.33 2.4 -1.9 Rented land (%) 1.39 2 1.7 -0.6 Land owned by others (%) 4.63 3 3.8 1.6 Fertiliser application (%) 79.2 54.7 67.0 24.5 0.000*** Variety of seeds cultivated 0.000*** Naga white (%) 0.9 2.0 1.5 -1.1 Kadaga (%) 0.0 36.0 18.0 -36.0 Kapaala (% 33.3 9.0 21.2 24.3 Dorado (%) 65.7 53.0 59.4 12.7 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) 100 University of Ghana http://ugspace.ug.edu.gh With regards to the sources of seeds, an average of 86% of the respondents used their own saved seeds, 10% purchased seeds and 2% get seeds as part of SOGS package. Less than 1% get seeds from government and other sources such as friends and relatives; 18% of treatment group purchased seeds and 5% of control group purchased seeds. As part of government support through the Planting for Food and Job programme, sorghum seeds were provided to farmers (MoFA, 2017). About 1% of the treatment group received seeds from government and no member from control group received seeds from government. The findings suggest that farmers in the study area do not benefit from seed under the government Planting for Food and Jobs programme. Irrespective of dorado being improve variety, the results suggest that most farmers saved and replant dorado. On the perception about the quality of the seeds used, an average of 24% said the seeds they used were excellent and 63% indicated the seeds were very good. Only a small percentage (less than 1%) said the seeds were bad. This finding shows that many farmers do not buy certified seeds. This further support literature on SHF relying on saving seeds from their own harvest than buying cerified seeds despite low yields from farmer saved seeds (Kuivanen et al., 2016; Morris et al., 2019; Sheahan & Barrett, 2017). Contrary to views of low fertilizer application among SHF, this study found an average of 65% of the respondents applying fertilizer. About 79% of treatment group and 55% of control group respectively apply fertilizer. The relatively higher fertilizer application among treatment group support literature on input support for farmers on OGS as condition for scheme participation (Maertens & Velde, 2017; MoFA, 2017b; Paglietti & Sabrie Roble, 2012; Ragasa et al., 2018; Takeshima & Lee, 2012). 101 University of Ghana http://ugspace.ug.edu.gh 4.3. Socio-economic and Political Characteristics of the Respondents Table 4.3 contain results of market, community and socio-economic characteristics of respondents. Nearness to market was used as a proxy to market access. The average distance from home to the nearest market was 1.5 km and 1.2 km for treatment and control groups respectively. Given that more treatment group members were residing relatively far from the nearest market, access to market could be a problem and motivation factor for scheme participation. This finding is similar to other literature on access to market being determinant for the type of crop cultivated by SHF (Ebata & Hernandez, 2017; Opoku, 2012; Villano et al., 2019). About 22% of respondents who belonged to the treatment group hold a form of leadership position in social organisations while 10% of the control group hold leadership positions. Table 4.3: Market, Community and Socio-economic and Political Characteristics Difference P-value (T-test / Chi- Variable Treatment Control Overall (means/freq) square) Distance from home to 1.6 0.6 1.1 1.0 0.000*** the nearest market (Km) Farmer holds leadership 21.8 9.7 15.8 12.1 0.000*** position (%) FBO Membership (%) 62.5 0.7 31.6 61.8 0.000*** Access to credit (%) 8.3 2.3 5.3 6.0 0.000*** Extension Visit (%) 54.2 6.7 30.5 47.5 0.000*** *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) On FBO membership, 63% of the respondents from the treatment group belonged to FBO against 1% from the control group. Contrary to Ragasa et al. (2018) that FBO membership is 102 University of Ghana http://ugspace.ug.edu.gh not a requirement for participation in outgrower scheme, this results confirmed the argument ISSER (2012); Maertens & Velde (2017) and MoFA (2017) that belonging to a group or FBO increases the chances of participating in new schemes in farming. Ajewole (2010); Mwangi & Kariuki (2015) and Feleke et al. (2017) explained that being a member of a farmer association allows information sharing and peer education. Perhaps, during meetings, information on OGS are shared and might have influenced scheme participation. Access to credit was very low for both treatment and control groups. While 8% of members in the treatment group have access to credit, only 2% of members in the control group have access to credit. This result reinforces the fact that SHF generally faced difficulty in their effort to get credit from financial institutions (Awunyo-Victor & Al-hassan, 2014; Mustapha et al., 2016). Given that access to credit is part of OGS arrangement, access to credit for treatment group was expected to be at least higher than what is reported. Access to credit attracted a huge debate during FGD discussion in Zari in Garu district. While majority of farmers were disappointed with buyers for not supporting them with credit, there were few farmers who blamed their colleague farmers of failure to honour their part of obligation anytime buyers give them credit. As a farmer put it “the buyers are not helping us with credit. To farm sorghum well, you need tractor service, fertilizer and money to hire people for weed control and harvesting, but our buyers are not interested in supporting us with money. Any time you ask for support, they will tell you to be managing small-small”. This was also discounted by another farmer in the group. “Let us all be sincere to ourselves and tell the truth. I remember in 2016 we were given GHS 500.00 each by Presbyterian- Agricultural station to buy fertilizer, how many of you paid all the money back? For me I paid 103 University of Ghana http://ugspace.ug.edu.gh my part and I am still getting support from them. Who will continue to give you money when you failed to honour your obligation?”, he questioned. On the part of the aggregators, farmers are difficult to deal with and one has to be careful when giving them support. This was a response from an aggregator when questioned about the type of support they give to farmers as part of the OGS arrangement in Jirapa. “I have been doing this work (sorghum aggregation) since 2010. I worked with different groups of farmers and have learned my lessons. I have given money, fertilizer and weedicide to different groups of farmers in the past but I had it very tough in retrieving my money. Some farmers will even take your inputs and use it to produce for different buyers; will you believe that? Now I only give support to few farmers who demonstrated commitment and sincerity, majority of the farmers we work with are not faithful at all!!, If you are not diligent, they will collapse your business”. Number of extension visit was used as proxy for access to extension services. The package of extension services discussed with farmers include: training in agronomic practices, early warning system, weather information and planting varieties that are climate resilient. Whereas 54% of members in the treatment group experienced extension visit and training in the above issues in the last farming season, only 7% of members of the control group experienced extension visit. Thus, belonging to outgrowers scheme could influence access to extension services and at the same time, availability of extension services could also be a source of information for scheme participation. Buyers seeking to improve the quality of the sorghum, may organise private extension services or arrange with public extension at the district department of agriculture to train scheme members. 104 University of Ghana http://ugspace.ug.edu.gh A buyer explains their engagement with farmers on extension provision. “we are not agronomies, but what we do is to work with the Jirapa District Assembly to allow some of their extension officers to train our farmers. We also pay some allowance for their services, fuelled their motor bikes and provide them with phone credit. The additional information we give them is to teach the farmers on weather information and the type of seeds to plant. Our interest is “dorado” which is early maturing variety so that when the rains stop early, they can still harvest something”. During the FGD, a 43-year farmer explained “the agric officer who come here help me a lot. Last year, he brought some seeds for us to buy. He also told us to plant when the Easter Rain comes” (Easter rain is the early rain that normally experienced after Easter celebration). “The seeds take less than three months. When I planted it, in September, it was ready for harvesting. I was able to complete harvesting before the rain stopped in October”. Relatively higher extension visit to treatment group support the call for farmers to be in groups for easy extension delivery due to limited number of extension officers (Asfaw et al., 2012; Azumah et al., 2016; Feleke et al., 2017; ISSER, 2012; MoFA, 2017b). 4.4 Description of the Sorghum Outgrower Scheme in the Study Area This section present results of farmers perception of what explains sorghum outgrower scheme. Major factors of interest presented is the type of contracts between farmers and buyers, how prices are determined and the type of support system received from buyers. 105 University of Ghana http://ugspace.ug.edu.gh 4.4.1 Type of Contracts Between Farmers and Buyers The results on the type of contracts between farmers and buyers in Figure 4.1 show that 21% of treatment group have documented contract while 79% do not have documented contract with buyers. This result is similar to issues raised during the focus group discussions, that buyers are not interested in signing contracts with farmers, but only interested in the verbal contracts. Several literature on OGS in SSA shows that buyers do not sign formal contract with farmers (Barrett et al., 2012; Fischer & Wollni, 2018; Little & Watts, 1999; Šūmane et al., 2018; Watanabe et al., 2017). 21% Not Documented Contract 79% Documented Contract Figure 4.1: Existing Contracts Between Farmers and Buyers Source: Field Survey data (2018) 106 University of Ghana http://ugspace.ug.edu.gh 4.4.2 Perception of Smallholder Farmers on Price Determination Figure 4.2 show that 54% of respondents indicated aggregators determining prices while 32% claimed price determination is an arrangement between buyers and farmers. Only 12% said the market determined price while an insignificant number (one person) said farmers determined prices. Aggregators determined prices 1 12 Farmers determines the prices Arrangement between farmers 54 and buyers 32 Prevailing market price 1 others specify Figure 4.2: Perception of Farmers on How Prices Price Determination Source: Field Survey data (2018) The results on price determination support the complains from most farmers during the FGD and individual interviews who claimed aggregators determined prices without consulting them. According to them, they suspect aggregators get attractive prices from GGBL but pay them lower prices. During FGD in Zari, a 37-year male farmer indicated, “even though the sorghum farming is good, the buyers we work with don’t support us with anything. They only come to discuss how to improve on the quality of sorghum we supply to them and the prices they will offer for the season. We don’t benefit from the buyers. If we get another buyer, it will 107 University of Ghana http://ugspace.ug.edu.gh be good for us”. Contrary to this accession, the aggregators were also blaming GGBL for not supporting them with inputs, credit and attractive prices. They also expressed fear of non- sustainability of the SOGS due to monopoly power of GGBL and hoped more companies will join the sorghum market and compete with GGBL. These findings support similar reports on buyers imposing prices on farmers in most OGS (Jayne, 2012; Tang et al., 2016). 4.4.3 Kind of Support Received by Farmers from Buyers The nature of support received by farmers on the SOGS is presented in Table 4.4. This was a multiple-choice question for farmers to list all support received. Among the area of support, 73%, 58% and 54% of the treatment group respectively claimed they received extension services, guaranteed market and fertilizer. Higher percentage of farmers receiving extension services is a reflection on the higher adaptive capacity of treatment group since the training given by Extension Officers included training on weather information, agronomic practices, early warning signs and the need to plant early maturing crop varieties. Table 4.4: Kind of Support Received by the Sorghum Farmers from Buyers Kind of support received Frequency Percent of cases Extension Service 76 35.2 Guaranteed market 60 27.8 Fertilizer 56 25.9 Guaranteed price 43 19.9 Seeds 13 6.0 Credit (financial support) 8 3.7 Tarpaulin 2 0.9 Pesticide 1 0.5 Source: Field Survey data (2018) 108 University of Ghana http://ugspace.ug.edu.gh On access to market, the result suggests that majority of farmers on the SOGS achieved their expectation of market access which was consistently mentioned as a major problem facing farmers. Access to credit, access to seeds, and tarpaulin recorded low percentages of 6.1%, 13% and 2% respectively. Even though most farmers faced problems with access to credit and expressed their disappointment during the FGD, high default rate, bureaucracies and collateral requirements of financial institutions may be the reason for buyers not supporting farmers with inputs (Alobo, 2019; Al-hassan et al., 2014; Gitonga et al., 2013; Mustapha et al., 2016; Opoku, 2012; Ton et al., 2016; Uaiene et al., 2009) 4.4.4 Specific Support Targeting Smallholder Farmers Vulnerability to Climate Change Table 4.5 contain results of support that target at capacity building of SHF to withstand climate change. Multi choice questions were directed to both treatment and control groups. The results suggest that both groups received some form of training or support on climate resilient. For the treatment group the common support received by majority of farmers were information on planting early maturing plant varieties of 51%, planting period 46% and information on harvesting on time 47%. The kind of support received by the control group was information on when to harvest 29%, information on planting early maturing plant varieties 27% and information to engage in off-farm activities 25%. The results is consistent with literature on the role of OGS in modernising the activities of SHF by introducing them to new planting materials and agronomic information (Murphy, 2012; World Bank, 2003). 109 University of Ghana http://ugspace.ug.edu.gh Table 4.5: Kind of Support on Climate Resilient Received by Farmers Treatment=216 Control=300 Overall=516 Type of Support/Training Frequency % Frequency % Frequency % Early Maturing Variety 111 51.39 80 26.67 95.5 18.51 Change the Crop Type 50 23.15 10 3.33 30 5.81 Cultivated Change from Crops to 2 1 8 2.67 5 1 Livestock Combine Crop with 90 40.17 50 16.67 70 13.57 Livestock Information on 100 46.3 90 30 65 12.6 Planting Period Information on Harvest 100 46.6 86 28.67 93 18.02 Period Engage Off-farm 30 13.89 74 24.67 52 10.8 Activities Others 6 2.78 60 20 33 6.4 Source: Field Survey data (2018) This result suggest that the treatment group are likely to be aware of the type of farming practices that makes them less resilient to climate change than those on the control group. Dorado being sorghum variety promoted by GGBL and also being early maturing plant variety could be the reason for more treatment group having information on early maturing plant varieties. During the FGD, majority of farmers on the treatment group indicated their desire to plant the dorado sorghum variety due to its early maturing characteristics and its ability to withstand draught and high temperature. They further explained that GGBL advised them to plant the dorado between the middle of May to 20th of June. According to them, for other crops, they usually could plant until the end of July and due to short duration of the rainfall, most crops do not mature before the rain stops. This is what 50-year old male farmer indicated: “we normally plant dorado in late part of May or early June and harvest in September. Unlike other crops, you are able to harvest dorado before the rain stops. Also, if you do not harvest it early and 110 University of Ghana http://ugspace.ug.edu.gh the rain fall on it, creamy colour turned to black and you are likely not to get good market for such produce”. 4.5 Factors influencing Smallholder Farmers Participation in Sorghum Outgrower Scheme Probit regression results of factors influencing SHF participation in the SOGS are presented in Table 4.6. The Prob > chi2 implies that the model is statistically significant and can be used for the estimation. Out of the 15 explanatory variables examined, four variables lacked significant explanatory power. All the other 11 variables were statistically significant with 8 variables having the prior expected sign. Gender of the farmer is a significant factor influencing participation with probability of a male farmer joining the SOGS increased by 11%. This finding on gender is consistent with Ali et al. (2016); Lambrecht (2016); Mariano et al. (2012) and Abdul-Razak & Kruse (2017) studies indicating that male farmers have better access to economic resources such as land, credit and technological resources such as better information on the soil fertility which place them in better position to adopt new intervention. The results were expected and is consistent with findings by FAO (2015) and Tanellari et al., (2011) that, male farmers are more likely to adopt contract farming due to functional command over land for scheme participation. Lambrecht (2016) study found similar results that men unlike women have decision making powers over how family land is used in patriarchal societies such as northern Ghana. 111 University of Ghana http://ugspace.ug.edu.gh Table 4.6: Factors Influencing Participation in the Sorghum Outgrower Scheme Variable dy/dx Std. Err. Gender (Male=1) 0.11*** 0.04 Age of farmer -0.01*** 0.00 FBO Membership 0.50*** 0.05 Number of Extension visits received 0.10*** 0.02 Distance from community to main market 0.01** 0.04 Marital Status -0.06 0.04 Household Status -0.08* 0.04 Formal education completed Basic -0.04 0.04 Higher than Basic 0.08 0.06 Household size -0.01** 0.01 Main Occupation 0.03 0.05 Farming Experience -0.01*** 0.00 Length of Stay in the community 0.00*** 0.00 Farm Size 0.07*** 0.02 Access to Credit 0.10 0.07 Leadership in social organization 0.07* 0.04 Number of obs = 516 Prob > chi2 = 0.0000 Pseudo R2 = 0.5482 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) Contrary, Bellemare & Bloem (2018) found gender variable in adoption to be negligible. Their study argued that most household decisions in farm operations such as seed selection are jointly taken by husbands and wife in most rural areas. Since having access to farm land is 112 University of Ghana http://ugspace.ug.edu.gh prerequisite for participation in SOGS participation is likely to skew in favour of male farmers in Northern Ghana. The probability of participation decreases by 1% for a unit increase in both age and experience. Age of the farmer and farming experience have positive relationship. The older farmers are generally found to be more endowed with experience in farming. Compared to younger farmers, older farmers approached their farming activities as a way of life, content with the little harvest they obtained and are less likely to change from their farming practices to join new intervention (Ajewole, 2010; Udimal et al., 2017). The relatively younger farmers on the other hand appears to be less experienced and will adopt farming practice that generates more income (Al-Hassan et al., 2013). Given the high level of unemployment in Ghana and the fact that sorghum provides ready market relatively younger farmers participating was expected. This is positive development for the study area in the light of high rural urban migration in search of non-existing jobs in southern Ghana. The household status was included to explain the relationship between being the household head and probability of participating in SOGS. This variable seemed to play important role in the decision to participate with probability of participation decreased by 8% for a unit increase in being the household head. The status of household head in northern Ghana is generally characterised with old age, more responsibilities and more experience in farming (Mustapha et al., 2016; Nkegbe et al., 2017). The results were expected due to the cultural dynamics that obligates the household head to provide food for the family. Hence, household heads will rather cultivate staple food crops for family consumption than sorghum, which has become cash crop in the study area. 113 University of Ghana http://ugspace.ug.edu.gh Another unanticipated outcome is the probability of participation and household size. While several studies such as Mignouna et al. (2011); Ajewole (2010) and Udimal et al. (2017) postulate positive relationship between household size and OGS due to availability of family labour to meet labour requirement of OGS conditions, this study found the probability of participation decreased by 1% for a unit increase in household size. Even though the finding may not be similar to what would have happened in other jurisdiction, the results is still relevant and portrays the reality of limited lands for individual members in large families in northern Ghana. Perhaps, large family size could be associated with smaller per unit land size for each member of the family. Apparently, the requirement of separate plots for sorghum production could limit farmers from larger families to participate due to computing uses of the small piece of land for the family food crops production. Farm size was measured by the number of hectares cultivated in 2017 farming season. The probability of participation increased by 7% for a unit increase in farm size. Farm size is often used as a proxy indicator of wealth and social status in rural communities (Aniah et al., 2019). The positive relationship found between land access and participation in SOGS could be explained by the fact that such farmers have sufficient land for both staple and cash crops production. This finding validates similar observations by Ton et al., (2016). In addition, large farm size is an indication of wealth and available collateral to access finance (Ragasa et al., 2018; Ton et al., 2016). Farmer Based Organisation (FBO) membership play important role in technology adoption and for that matter, participation in SOGS (Maertens & Velde, 2017). Apart from 63% of the treatment group belonging to FBOs, the probability of participation increases by 50% for a unit increase in FBO membership. This findings also confirmed Barrett et al. (2012); Feleke 114 University of Ghana http://ugspace.ug.edu.gh et al. (2017) and MoFA (2017a) studies recommending farmers to form groups for easy information dissemination, access to credit, access to extension services and easy targeting by government. Working in groups also helped in price negotiation and marketing. To adopt new intervention sometimes is hampered not only by the inherent requirements that might restrict participation but also, lack of information on the benefits of such interventions. Belonging to FBO is a form of social capital that gives information to the farmer (Odunze et al., 2015). Also most agribusiness investors, government and NGOs prefer dealing with farmers in groups than individual farmers (MoFA, 2015, 2017b). The probability of participation increases by 10% for a unit increase in extension visit. According to Udimal et al. (2017), agricultural extension is a learning system that builds human capital of farmers by providing them information and exposing them to improved technologies. This leads to higher profitability and welfare gains. Essentially, farmers with frequent extension visits tend to be more progressive in terms of productivity and profitability (Mwangi & Kariuki, 2015; Udimal et al., 2017). The role of extension agents was frequently mentioned during the FGD and re-emphasised by aggregators and GGBL during the qualitative interviews. As an aggregator indicated in Jirapa “extension agents popularised the scheme by providing essential information, appropriate knowledge and capacity development in sorghum production. They provide training on quality requirement and standards, marketing and also equipped farmers with information for price negotiation”. Distance to the nearest market was used as a proxy for access to market. The decision to participate in SOGS has positive relationship with distance to the nearest market. The probability of participating increases by 1% for a unit increase in distance to the nearest market. Information generated during focus group discussion support this claim as the farmers 115 University of Ghana http://ugspace.ug.edu.gh staying far from market complained about distance to the nearest market and low interest of buyers to travel longer distance to buy sorghum. As a 53-year farmer indicated in Sabule “our roads here are bad, the market is far and getting a vehicle to convey your produce to the nearest market is difficult. But if you manage to send your produce to Jirapa or Wa, they will buy them. We are suffering with how to sell our produce in this community”. Leadership in social organisation has positive relationship with participation. The probability of participation increases by 7% for a unit increase in leadership position. This was expected since the entry point in rural community starts either with the Chief, Assembly Member, Chief Iman or the Community Pastor. These leaders sometimes participate in selection of beneficiaries of new interventions (Ton et al., 2015). Likelihood of participation of a community leader is higher than those without leadership position. 4.6. Determining the Effects of Outgrower Scheme on Smallholder Farmers Productivity 4.6.1 Productivity Analysis Table 4.7 contain results of sorghum productivity. Yield per hectare is used to explain sorghum yield in 2017 farming season. On average the treatment group obtained yield of 1,207kg/ha while control group obtained 820kg/ha. The higher yield for treatment group could be attributed to application of quality inputs and good agronomic practice being encouraged by SOGS. The treatment group invest more in hired labour, ploughing; weed control and application of pesticides. Good agronomic practices by treatment group is reflected in their higher expenditure in chemical fertilizer, organic fertilizer, improved seeds, pesticides, cost of hired labour and cost of transportation compared with expenditure by control group. While treatment group spent GHS 355.00, GHS 62.00 and GHS 11.00 respectively on chemical 116 University of Ghana http://ugspace.ug.edu.gh fertilizer, organic fertilizer and pesticides respectively, the control group spent GHS 280.00, GHS 3.00 and GHS 9.00 respectively on the same inputs. Table 4.7: Comparison of Productivity, Revenue and Costs Across Treatment and Control Groups Variable Treatment Control Overall Difference T-test N= 216 N= 300 N= 516 Yield/productivity (Kg/Ha) 1206.9 819.86 1013.4 387.03 0.00*** Average Price GHS/Kg 1.2 1.3 1.3 0.1 0.00*** Revenue (GHS/Ha) 1444.4 1027.9 1236.2 416.5 0.00*** Less Variable Cost Seed cost (GHS/Ha) 18.8 7.3 13.1 11.5 0.00*** Chemical Fertiliser cost (GHS/Ha) 354.5 280.1 317.3 74.3 0.05* Organic Fertiliser cost (GHS/Ha) 61.5 2.6 32.1 58.9 0.00*** Pesticide cost (GHS/Ha) 10.7 9.4 10.1 1.3 0.84 Hired labor cost (GHS/Ha) 406.2 218.3 312.3 187.8 0.00*** Family labor cost (GHS/Ha) 305.1 595.9 450.5 290.7 0.00*** Transportation Cost (GHS/Ha) 17.2 6.5 11.9 10.7 0.00*** Total Variable Cost (GHS/Ha) 1174.1 1120.3 1147.2 53.8 0.40** Profit (excluding family labor) (GHS/Ha) 575.4 503.5 539.5 71.9 0.38 Profit (including family labor) (GHS/Ha) 270.2 -92.4 88.9 362.7 0.00*** *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) 117 University of Ghana http://ugspace.ug.edu.gh Treatment group spent GHS 406.00 on hired labour while control group spent GHS 218.00 on hired labour. Hired labour cost in this study comprises of cost incurred in ploughing, clearing, gathering and heaping, weeding, harvesting, pesticide application, threshing and winnowing. It appears control group depend more on family/communal labour than scheme members. The control group spent GHS 596.00 on family/communal labour against GHS 305.00 for scheme members. The transportation cost for scheme and control group was GHS 17.00 and GHS 7.00 respectively. 4.6.2 Average Treatment Effect on Sorghum Productivity The estimation of average treatment effect of SOGS on productivity using ESR are presented in three folds. The results of the FIML in table 4.8, ESR results of treatment effects in table 4.9 and the PSM robustness check results in table 4.10. The Results of FIML: The results of FIML of the ESR that control for unobservable selection bias and the significance of the ESR results are reported in Table 4.8. The second and fourth column in the table contain the treatment and control equations respectively. To analyse the correlation between decision to participate in SOGS and productivity, a set of broad explanatory variables were included to determine their significance. Gender, household size, farming experience and FBO members are all statistically significant variables influencing participation decision and at the same time influencing productivity of the treatment group. While the probability of participation decreases with farming experience, the probability of participation increased for being a member of FBO. The positive sign of FBO membership suggests that, being a member of FBO influence participation decision. The result is consistent with earlier results and empirical literature of belonging to FBO 118 University of Ghana http://ugspace.ug.edu.gh influencing participation in OGS (Feleke et al., 2017). The only significant variable on the control equation is farm size. The positive sign suggests that farmers with large farm size who are in the control group are more likely to participate in the SOGS. This result is consistent with earlier results and existing literature that farmers with large farm size are more likely to participate in OGS. Implying that the OGS is not pro-poor farming scheme. Table 4.8: Results of Full Information Maximum Likelihood Estimates of Endogenous Switching Regression on Productivity Independent Variable Outgrowers Non Outgrowers Std. Coef. Err. Coef. Std. Err. Gender 173.917* 104.830 7.200 148.689 Age -0.162 3.362 4.918 4.155 HHstatus -105.538 126.843 141.759 182.970 Household Size 34.384** 17.015 27.510 20.775 Educated 135.735 99.531 63.474 149.575 Occupation -90.430 136.051 123.153 200.429 Farming Experience -28.893** 13.734 -12.139 7.489 Farm Size 209.144*** 56.782 456.876*** 132.082 FBO Membership 745.623*** 157.366 81.828 792.757 Access to Credit 25.958 159.429 -139.127 407.877 Number of Extension Visits received -5.660 29.647 -58.372 138.510 Constant 736.055** 244.611 646.683** 312.240 /lns 6.448*** 0.062 6.918*** 0.042 /r 0.960** 0.336 0.175 0.123 Sigma 631.561 39.194 1010.299 42.115 Rho 0.744*** 0.150 0.173 0.120 Number of observations = 516 Wald chi2 (11) = 45.46*** Log likelihood (χ2 (11) = -658.3*** prob>chi2 = 0.00 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) 119 University of Ghana http://ugspace.ug.edu.gh The covariance term for productivity of treatment equation is statistically significant at 1%, implying self-selection into SOGS and that the participating in SOGS may not exhibit the same effects on random sorghum farmers, if they choose to participate in the SOGS. The value of the chi-square statistics of 45.46 is statistically different from zero, suggesting that the independence assumption of participation decision and productivity must be rejected at 1% level justifying the appropriateness of using the ESR to jointly estimate participation decision and sorghum productivity. The likelihood ratio (LR) chi-square (χ2 (11) = -658.3) is statistically significant at 1% further confirming that the explanatory variables examined in the model jointly influenced the participation decision and sorghum productivity. The prob>chi2 is indication that the model is significant and can be used for the estimation. Finally, the significance and differences in coefficient of correlation between the participation equation and productivity is an illustration of self-selectivity independent variables of treatment and control equation signifies the presence of heterogeneity in the sample. The Results of ESR: The ESR results on effect of SOGS on productivity are presented in Table 4.9. Cases (a=1206) and (b=820) along the diagonal represent the actual expectations observed for both treatment and control groups respectively. The results suggest that the treatment group obtained higher productivity of 386 than the control group. Cases (c=374) and (d=831.7) represent counterfactual outcomes for the treatment and control groups respectively. This result suggest that had the treatment group decided not to participate, they would have realised less productivity per hectare and had the control group participated they would have realised higher productivity per hectare than their current productivity by 12.2. 120 University of Ghana http://ugspace.ug.edu.gh Table 4.9: Endogenous Switching Regression (ESR) Results on Productivity Decision Stage Treatment Treatment Status To participate Not to participate Effect Treatment group (a) 1205.81 (c) 374.2 831.6*** Control group (d) 831.7 (b) 819.5 12.2*** Heterogeneity effects 𝐵𝐻1= 374.11 𝐵𝐻2= -445.3 TH= 819.41*** *** denotes significance at 1% Source: Field Survey data (2018) The last column of table 4.9 contains treatment effect of SOGS (the difference between the results of actuals and counterfactual outcomes). The treatment effect is positive for both the treatment group and the control group. This result implies that SOGS has positive effect on sorghum productivity for both the treatment and the control groups. To gain further understandings of the level of effect, the heterogeneity effect is reported in the last row of the Table 4.9. The heterogeneity effects for treatment (𝐵𝐻1) = 374.11 and control (𝐵𝐻2) = -445.3. The differences in 𝐵𝐻1 and 𝐵𝐻2 illustrate existence of unobserved heterogeneity effects. The transitional heterogeneity effect (TH) is positive and significant at 1%. This suggest that SOGS has effect on both treatment and control groups. In conclusion, the SOGS has positive effect on random sorghum farmer who choose to participate in the OGS in the study area. The level of effect is not mainly due to participation in the SOGS but there exist unobserved characteristics of the treatment group that influenced their participation decision and their productivity. 121 University of Ghana http://ugspace.ug.edu.gh The Results of PSM: The PSM estimates is used for robustness check. The distribution of propensity scores, indicator of matching quality, sensitivity analysis and quality test of matching algorithm are attached in Appendix 4 Figure A1, Table A7, Table A8 and Table A9 respectively. The nearest neighbour matching results on common support generated was 383. This shows that 133 respondents lack similarity in terms of their observable characteristics. The PSM results suggest significant reduction in differences in observable characteristics of the treatment and control group after matching and that there are no significant differences in observed characteristics between the treatment and control groups under PSM estimation. The PSM results of the average treatment effect on the treated (ATT) and the untreated (ATU) are presented in Table 4.10. The ATT and ATU results in Table 4.10 are 858 and 348 respectively. The results further confirmed positive effect of SOGS on productivity. Comparing the treatment effect of ESR of 831.6 and 12.2 for the treatment and the control groups respectively with the PSM results of 858 and 348, it appears PSM overestimated the effect on both the treatment and control groups. The differences could be due to existence of unobserved characteristics that PSM could not addressed (Donkor & Owusu, 2019). The ESR results and the PSM results is consistent with empirical literature on positive effect of OGS on SHF productivity (Maertens & Velde, 2017; Ragasa et al., 2018). Table 4.10: Propensity Score Matching Results on Farm Productivity Sample Effect Std. Err T-stat ATT 858.0*** 138.0 4.8 ATU 348.0*** 130.4 2.7 ATE 454.6*** 110.3 4.1 *** denotes significance at 1% Source: Field Survey data (2018) 122 University of Ghana http://ugspace.ug.edu.gh 4.7 Determining the Effects of Outgrower Scheme on Smallholder Farmers Profitability 4.7.1 Profitability Analysis Table 4.7 above contain results of SHF profitability. The average profit received by treatment group was GHS 575.00/ha and that of the control group was GHS 504.00/ha. When family labour is computed as part of variable cost of which control group largely rely on but do not cost it, the treatment group profit reduces to GHS 290.00/ha while the control group incurred loss of GHS -92.00/ha. Even though the treatment group incurred high variables cost of GHS 1,174.00/ha, they still got high profit due to higher yields. The higher variable cost for treatment group is due to increase in fertilizer application, improved seeds used and the use of hired labour to ensure proper farm management. In terms of prices received, the treatment group and control group received GHS 1.20/kilo and GHS1.30/kilo respectively. The lower prices received by treatment group is attributed to contractual arrangement that pre-determined prices before production begins. Also, clauses in contractual agreement usually prevents treatment group from selling to open market irrespective of open market prices (Odunze et al., 2015). The flexibility of control group to be able to store and sell their produce during lean season may account for their higher prices relative to the treatment group. 4.7.2 Average Treatment effects on Profitability The Results of FIML: The results of FIML on profitability are presented in Table 4.11. Gender and FBO membership in the treatment equation are significant with positive signs. This suggest that male farmers and a member of FBO influenced the treatment group participation in the SOGS and their profitability. The finding confirmed earlier report and consistent with existing literature on positive relationship between male farmers and FBO 123 University of Ghana http://ugspace.ug.edu.gh membership on participating in OGS (Feleke et al., 2017). Access to extension services has negative sign and significant at 5% suggesting that extension services has influence on the treatment group members participation in the SOGS. This implies that the treatment group who participated will not have participated if they were to have access to extension services suggesting that expectation of getting extension services when join OGS is a significant factor influencing participation. For the control equation, age, being the head of household, having farming experience and access to credit are significant factors influencing participation. Whiles age, being household head and having more farming experience influence participation positively, access to credit has negative influence. Table 4.11: Full Information Maximum Likelihood Estimates Results on Profitability. FIML Endogenous Switching regression model Outgrowers Non Outgrowers Variable Coef. Std. Err. Coef. Std. Err. Gender 300.52* 148.07 -129.12 132.80 Age 4.67 4.71 -7.34** 3.71 HH Status 26.32 179.18 -277.18* 163.63 Household size -3.16 23.76 28.95 18.69 Educated 104.55 138.90 119.81 133.65 Occupation -43.21 191.96 94.07 179.02 Farming Experience -6.90 19.79 -15.61* 6.97 Farm Size -56.94 77.75 194.29 119.68 FBO Membership 411.60** 193.19 -890.09 754.03 Access to Credit 147.35 217.87 -822.93* 365.01 Number of extension visits received -85.72** 39.82 100.96 125.89 Constant -258.38 338.17 -847.82*** 282.32 /lns1 6.73 0.05 6.80 0.04 /r1 0.05 0.18 0.07 0.19 sigma_1 839.35 40.44 900.68 rho_1 0.05* 0.18 0.07 Number of obs = 516 Log likelihood = 1433.48*** Wald chi2(11) = -19.66*** Prob > chi2 = 0.05 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) 124 University of Ghana http://ugspace.ug.edu.gh Also, the chi-square statistics of 19.66 is statistically significant and different from zero, meaning that the independence assumption given to the selection and the outcome equations must be rejected at 1% level. The likelihood ratio (LR) chi-square (χ2 (11) = -1433.48) is statistically significant at 1% further confirming that the explanatory variables examined in the model jointly influenced the participation decision and sorghum profitability. The prob>chi2 is indication that the model is significant and can be used for the estimation. Finally, the significance and differences in coefficient of correlation between the participation equation and profitability is an illustration of self-selectivity and presence of heterogeneity in the sample. The Results of ESR: The ESR results of treatment effect of SOGS on profitability under actual and counterfactual conditions for treatment and control groups are presented in table 4.12. Cells (a=270.3) and cell (b=-92.5) represent the expected profitability observed. The results suggest that the expected profitability of the treatment group is higher than the control group. The counterfactual outcomes suggest that, had the treatment group decided not to participate, they would have realized losses of GHS 482.10/ha and had the control group decided to participate, they would have made profit of GHS 42.70 against their current losses of GHS 92.50/ha. Table 4.12: Endogenous Switching Regression Results on Profitability Decision Stage Farm Profitability (GHS/Ha) To Participate Not to Treatment Effect participate Treatment group (ATT) (a) 270.3 (c) -482.1 752.4*** Control group (ATU) (d) 42.7 (b) -92.5 135.2*** Heterogeneity effects 𝐵𝐻1= 227.6 𝐵𝐻2= = -389.6 TH= 617.2 *** *** denotes significance at 1% Source: Survey data (2018) 125 University of Ghana http://ugspace.ug.edu.gh The last column of the table 4.12 contain treatment effect on profitability. The results show positive effect on both treatment and control group with the effect being higher for the treatment group than the control group. The last row contains heterogeneity effects with 𝐵𝐻1 =227.6 and 𝐵𝐻2= -389.6 respectively. The higher value for 𝐵𝐻1 shows that SOGS has more effect on the treatment group than the control group. The differences in 𝐵𝐻1and 𝐵𝐻2 illustrate existence of unobserved characteristics that influence profitability apart from the SOGS. The transitional heterogeneity effect (TH) is positive effect of SOGS on profitability on both treatment and control groups. The Results of PSM: The PSM results presented in Table 4.13 show ATT and ATU results as 1159 and 352 respectively. This means that the SOGS has positive effect on profitability for both treatment and control group which is consistent with the ESR results except that the effect under PSM is higher than ESR indicating overestimation of treatment effects under PSM. The differences in the results may be attributed to inability of the PSM to account for unobservable factors. Table 4.13: Propensity Score Matching Results on Profitability Sample Effect Std. Err T-stat ATT 1159.3** 770.1 1.5 ATU 351.5*** 159.6 2.2 ATE 629.2** 425.3 1.5 ** and *** denotes significance at 5% and 10% respectively. Source: Survey data, (2018) 4.8 Determining the effects Outgrower Scheme on Smallholder Farmers Postharvest Loss Reduction 4.8.1: Analysis of Postharvest Situation in the Study Area Table 4.14 contains results of postharvest loss among treatment and control groups. The average postharvest loss for treatment and control groups are 14.2% and 27.4% and the value 126 University of Ghana http://ugspace.ug.edu.gh loss of GHS 193.00/ha and GHS 166.00/ha for the treatment and control groups respectively. The higher value loss of the treatment group is due to high quantity harvest of which small percentage loss represent high value of total harvest. Figure 4.3 contain percentage losses in the various postharvest stages. The highest losses for treatment group is recorded during storage of 4.4%, grading and bagging 3.8% and transporting from homestead to market 2.8%. Table 4.14: Postharvest Loss for Treatment and Control Groups Variable Treatment Control Overall Difference T-test N= 216 N= 300 N= 516 Loss during transportation from farm to homestead (Kg) 17.2 12.6 14.9 4.6 0.349 1.4 0.446 Losses during heaping 9.1 10.5 9.8 Losses during threshing and winnowing 11.1 9.5 10.3 1.6 0.349 30.9 0.082 Losses during grading and Bagging (Kg) 42.8 11.9 27.4 Losses during Storage (Kg) 49.5 64 56.8 14.5 0.004 Losses during transportation from homestead to the market (Kg) 31.3 18.9 25.1 12.4 0.147 Average Quantity lost (Kg) 161 127.4 144.2 33.6 0.849 Average value of Losses (GHS) 193.2 165.6 179.4 27.6 0.918 Total quantity harvested that is lost (%) 14.2 27.2 20.7 13.0 0.150 Average Quantity Harvested (Kg) 1130.8 468.1 799.5 662.7 0.000 Average Price (GHS/Kg) 1.2 1.3 1.3 0.1 0.000 Value of Quantity harvested (GHS) 1356.0 608.5 982.3 747.5 0.000 Source: Field Survey data (2018) 127 University of Ghana http://ugspace.ug.edu.gh For the control group, the highest losses were recorded during storage of 13.7%, transporting from homestead to the market of 4% and transporting from farm to homestead of 2.7%. The high storage loss for both groups was expected and is consistent with studies by Gitonga et al., (2013) and Hengsdijk (2017) of which limited warehouses and lack of storage facilities to protect grains from moisture, pest and insect’s infestation led to high postharvest loss among SHF in SSA. Also, the relatively low storage loss of treatment group compared to control group could be attributed to guaranteed market by SOGS for treatment group leading to shorter duration of storage period compared to control group who normally stored produce in anticipation of good prices during lean season. 30.00 25.00 20.00 15.00 10.00 5.00 - Treatment Loss (%) Control Loss (%) Overall Loss (%) Figure 4.3: Percentage Losses in the Various Postharvest Loss stages Source: Field Survey data (2018) 128 Percentage Loss University of Ghana http://ugspace.ug.edu.gh The need for appropriate storage facilities were emphasised during the focus group discussion in both treatment and control communities. According to a 47-year female farmer in Gozeig community in Garu, “I remember in 2015, after I harvested my sorghum in September, it rained consistently for three days. I left the un-threshed sorghum on the rain because there was no place to store them. By the time I started the threshing in somewhere January, about half of the sorghum went bad and the colour also changed to black. You won’t believe that Faranaya (company that buy sorghum for GGBL in Garu) rejected them when it was time for buying, claiming that GGBL will not accept this type of sorghum?”. In Kuncheni community in Jirapa, a 59-year-old male farmer also narrated how he lost his sorghum grains to rats and mice. According to him, “After harvesting, I stored my sorghum grains in a room I converted into a store room whiles waiting for good price to sell. One day I entered the room to check whether the sorghum was in good condition. To my surprise, I rather provided good feeding grounds for rats and Mice. Most of the bags were leaking and the grains were spread everywhere in the room. Some of the bags reduced in size after we removed them”. The situation of postharvest loss in Ghana is serious and is acknowledged by the government in the 2018 national budget of which budget allocation was made for construction of at least, one warehouse in every district to help address storage challenges facing farmers (Ministry of Finance, 2018). The high grading/bagging loss of 3.8% for treatment group was however not expected and its occurrence could be due to repackaging requirement by buyers. Another result is losses that occurred during transportation. Losses of 2.7% and 1.5% respectively is reported for transporting produce from homestead to market centre and farm to homestead respectively for treatment group. Similar results of 4.0% and 2.7% were reported 129 University of Ghana http://ugspace.ug.edu.gh for control group on transporting from homestead to market and from farm to homestead. The high losses recorded during transportation could be due to poor feeder roads in the study area which was expressed during the FGD. Some farmers claimed they sometimes produced for birds and rodents due to their inability to transport the produce to the nearest storage centres immediately after harvest. In appendix 3, results of causes of sorghum losses in the various postharvest loss stages are presented. Both treatment and control group reported poor road networks, lack of appropriate vehicles and long distance from farm to storage centres of 69%, 62% and 57% respectively as major causes of postharvest losses during transportation. This result is similar to information provided during the expert interviews and FGD of which poor feeders roads was consistently mentioned as a major constraint to sorghum production in the study area. Also several literature such as Affognon et al., (2015); FAO (2014b) and Rembold et el. (2014) all cited poor feeder roads and lack of appropriate vehicles in most rural areas as a major cause of postharvest loss in SSA. Also, the highest causes of storage losses for both treatment and control groups are high rainfall, aflatoxin contamination and insects attack of 12%, 5.2% and 7.6% respectively (appendix 3, table A6). 4.8.2 Average Treatment Effects of Sorghum Outgrower Scheme on Postharvest Loss Reduction Results of FIML: The results of FIML estimate on postharvest loss reduction are reported in Table 4.15. The second and fourth column in the table contain the treatment and control equations respectively. To analyse the correlation between decision to participate in SOGS 130 University of Ghana http://ugspace.ug.edu.gh and postharvest loss reduction, a set of broad explanatory variables were included in the model to establish their influence in participation decision and postharvest loss reduction. Age of the household head, age of the farmer and access to extension services have negative sign and are statistically significant variables influencing participation decision and postharvest loss reduction for the treatment group. The result is inconsistent with earlier results of older farmers; being household head and farmers who have access to extension services less likely to participate in new intervention. Belonging to FBO influencing participation in SOGS is also consistent with earlier results and empirical literature on FBO members likely to adopt new technology compared to those who are not. For the control equation, the household head, household size and access to extension services are statistically significant factors influencing participation decision. The negative signs are indications that the household head, household size and access to extension services negatively influence participation and postharvest loss reduction. 131 University of Ghana http://ugspace.ug.edu.gh Table 4.15: Full Information Maximum Likelihood Estimates Results on Postharvest Loss Reduction FIML Endogenous Switching Regression Model Dependent variable: SOGS and PHL Treatment Control Std. Std. Variable Coef. Err. Coef. Err. Gender 0.003** 0.007 -0.007 0.008 Age-Squared -0.004*** 0.000 -0.05 0.000 Age -0.004*** 0.001 -0.001 0.001 HH Status 0.002 0.008 -0.031** 0.010 Household size -0.0001 0.001 -0.002* 0.001 Educated 0.004 0.006 -0.006 0.008 Occupation 0.010 0.009 0.014 0.011 Farming Experience -0.0001 0.001 0.001 0.000 Farm Size 0.0001 0.003 0.020* 0.007 FBO Membership 0.013* 0.007 0.068 0.046 Access to Credit 0.008 0.010 0.009 0.023 Number of extension visits received -0.003* 0.002 0.013* 0.008 Constant 0.107*** 0.030 0.075** 0.032 - /lns -3.291* 0.048 2.888*** 0.042 /r -0.016 0.109 0.117 0.192 Sigma 0.037** 0.002 0.056 0.002 Rho -0.016 0.108 0.116 0.189 Number of obs = 516 Wald chi2(11) = 27.95*** Log likelihood = 672.76*** Prob > chi2 = 0.001 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source: Field Survey data (2018) The likelihood ratio (LR) chi-square (χ2 (11) = 672.76) suggest that the vector of explanatory variables examined in the models jointly influenced participation decision and its effect on postharvest loss reduction for the treatment and control groups. The chi-square statistics of 27.95 is statistically significant and different from zero meaning that the independence assumption given to the selection and the outcome equations must be rejected at 1% level. The significance of the coefficient of correlation between the participation equation and 132 University of Ghana http://ugspace.ug.edu.gh postharvest loss reduction is an illustration of self-selectivity and presence of heterogeneity in the sample which is controlled during the estimation of the treatment effect. The Results of ESR: The Table 4.16 contain results of ESR on the effect of SOGS on postharvest loss reduction under actual and counterfactual conditions. The (Cells a=14.24) and (cell b=27.22) represent the expected postharvest loss observed for treatment and control groups respectively. Cell (c=20.4) and (cell d=13.6) presents counterfactual outcome for treatment and control groups respectively. The treatment group had less postharvest loss than the control groups for actual and high postharvest loss than the control under counterfactual condition. Table 4.16: Endogenous Switching Regression Results on Postharvest Loss Reduction Decision Stage Crop Loses To Participate Not to enrol Treatment Effect Treatment (a)14.24 (c) 20.4 -6.16 Control Group (d)13.6 (b) 27.22 -13.62 Heterogeneity effects BH1= 0.64 BH2= -6.82 TH= 7.46 Source: Field Survey data (2018) The last column of the Table contains average treatment effect. The results suggest that, the treatment effect on the treatment and the untreated are -6.16 and -13.62. The results suggest that sorghum farmers who did not participate in the SOGS, if they had participated, they would have realised lower average sorghum postharvest loss than treatment group who participated. Also, had the treatment group decided not to participate, they would have realised low postharvest loss than control group Case (c) but higher than their current postharvest loss. This result suggests that participation in SOGS has positive effect on postharvest loss reduction for both treatment and control groups but the effect is more on the control group 133 University of Ghana http://ugspace.ug.edu.gh than the treatment group. This means that if the control group had participated in the OGS, their postharvest losses would have been lower than the treatment group. The last row of the Table 4.16 contains heterogeneity effects. The differences in the values of BH1 and BH2 of 0.64 and -6.82 respectively suggest that the effects of SOGS is higher on control group than the treatment group. The differences in the BH1 and BH2 illustrates existence of some unobserved heterogeneity effects (for example farming skills of individual households in managing postharvest losses) that makes control group better in controlling postharvest loss than treatment group if control group had participated in the SOGS. Finally, the positive value of transitional heterogeneity (TH) effect suggests that SOGS has positive effect on both treatment and control group. The Results of PSM: The PSM results presented in Table 4.17 show ATT and ATU as -0.07 and -0.01 respectively. The result is consistent with ESR results of SOGS having effect on postharvest loss reduction for both treatment and control groups. Unlike the ERS results, the effect on loss reduction is higher under PSM estimate. The differences in effect under ESR and PSM could be attributed to inability of PSM to account for unobservable factors and therefore over estimation the effect. Table 4.17: Propensity Score Matching Results on Postharvest Loss Reduction Effect Std. Err T-stat ATT -0.8*** 0.04 -8.7 ATU -0.14*** 0.03 0.5 ATE -0.16 0.03 -0.6 *** denotes significance at 1% Source: Field Survey data (2018) 134 University of Ghana http://ugspace.ug.edu.gh 4.9 Determining the Vulnerability Level of Smallholder Farmers to Climate Change This section present results of SHF vulnerability to climate change. The results are in two folds. The first fold contains LVI and LVI-IPCC results of the state of vulnerability of SHF to climate change among treatment and control groups. The second part present results of treatment effect of SOGS in reducing the vulnerability of SHF to climate change. 4.9.1 Results of Vulnerability of Smallholder Farmers to Climate Change Results of Major and Sub-component: Table 4.18 contain indices on sub and major components for treatment and control groups. The corresponding percentages are found in appendix 5 table A10. The vulnerability indices of the major component ranges from 0.228- 0.524. Indices closer to 0.228 suggest less vulnerable to climate change and indices closer to 0.524 indicate high level of vulnerability. Table 4.18 contain results of water, socio-demographic, food and social network sub and major components. The water major component has relatively lesser vulnerability index of (LVI=0.401) for control group than the treatment group of (LVI=0.417). The control group recorded 27.67% of households reporting conflicts over water resources compared to the treatment group of 36.28%. In terms of natural water source utilization, the treatment group recorded 86.51% while the control group recorded 74%. This result suggests that, the treatment group is more vulnerable in access to quality water as compared to the control group. The result is consistent with Adu et al. (2017) report that households depending on natural water sources such as lakes and dam are vulnerable to water borne diseases. In terms of consistent water supply, the control group recorded 46.33% and treatment group 59.07%. Due to over reliance on natural water source in both region the households are affected much 135 University of Ghana http://ugspace.ug.edu.gh during the dry season when most natural water sources tend to dry up. The results further suggest that most households in the two regions stored water to be used later. Table 4.18: Indexed Sub and Major Component of LVI for Water, Socio-demographic, Food and Social Network Sub-component Contro Treat Major Contro Treatme l ment component l nt Percent reporting water conflict 0.28 0.36 Water 0.401 0.417 Percent dependent on natural water sources 0.74 0.87 Time spend to source water 0.20 0.15 Percent that have no consistent water supply 0.46 0.59 Inverse of the average number of liters of water 0.32 0.11 stored per household Dependency ratio 0.21 0.18 Socio- 0.300 0.300 demograph ic profile Percent of female-headed households 0.29 0.29 Percent of households where the head has not 0.94 0.99 attended school Percent of households with orphans 0.08 0.19 Percent who dependent solely on family farm for food 0.99 0.99 Food 0.408 0.352 Average number of months strugging to find 0.44 0.33 food Average crop diversity index 0.21 0.36 Percent of households that do not save seeds 0.33 0.005 Percent of households that do not save crops 0.07 0.07 Average Receive: Give ratio 0.21 0.18 Social 0.478 0.485 Networks Average Borrow: Lend 0.29 0.29 Percentage that have not gone to their local 0.94 0.99 government for assistance in the past 12 months Source: Survey data (2018) Given that SOGS intervention do not focus on providing water as part of the SOGS package, no meaningful correlation could be drawn from scheme participation to impact on water utilization. During focus FGD in Sabule community, a 47-year-old woman expressed her disappointment in the area of support for access to quality water as she narrated her 136 University of Ghana http://ugspace.ug.edu.gh experience: “the buyers do not care whether we have drinking water or not, their only interest is on sorghum production and supply. Access to quality water is a serious problem in this community, especially during the dry season. Without water, you cannot be strong to farm”. In terms of the food major component, the treatment group was found to be more vulnerable (LVI = 0.352) than control group (LVI= 0.348). The average number of months households struggle to find food was found to be higher (3months) in the control than their treatment counterpart who recorded 4months. This result could be interpreted as farmers using all their food crops lands for producing sorghum leading to domestic food shortage. According to Drafor, Kunze, & Sarpong (2013), increasing food production improves food security outcomes and access to food improves household’s resilience to external stresses including extreme climatic events. This can be inferred that, as individuals, communities and countries improve their food production, there could be reduction in real prices for food which could results in improved real incomes and asset accumulation. This could motivate them to adapt climate change strategy. On sources of food, the result showed that about 99% of the households in the control depend solely on the family farm for food whereas about 98% of the households in the treatment group depend solely on the family farm for food. The average crop diversity index showed that the treatment group was more vulnerable (LVI= 0.36) than the control group (LVI = 0.22). This can be explained by the fact that, farmers in the treatment group use most of their lands for growing sorghum and use less for other crops whereas the farmers in the control group do mixed cropping. With sorghum, farmers usually do not intercrop with other crops. An officer from the district Directorate of Agriculture from Jirapa indicated: “sorghum will do well when it is grown as a mono crop. We advised the farmers not to intercrop sorghum with other crops, 137 University of Ghana http://ugspace.ug.edu.gh especially, the dorado and kapala variety”. When all the three sub-components under social network were aggregated, the control group was found to be more vulnerable (LVI = 0.485) than the treatment group (LVI = 0.478). Table 4.19 contain results of livelihoods strategies, natural disasters and health. Under the livelihood strategies major component, the control group showed greater vulnerability (LVI= 0.456) than the treatment group (LVI= 0.393). Table 4.19: Indexed Sub and Major Component of LVI for Livelihood Strategy, Natural Disaster and Climate Change and Health Sub-component Control Treat Major Contro Treatme ment component l nt Percent households who family members 0.19 0.12 Livelihood 0.456 0.393 work in different communities strategies Households dependent on agriculture as main 0.90 0.70 source of income Average agricultural livelihood diversification 0.28 0.36 index Percentage hat do not receive a warning about 0.96 0.92 Natural 0.488 0.524 the pending natural disaster Disaster and climate variability Households experienced injury and death due 0.003 0.005 to natural disaster Average number of floods, drought, bushfires 0.18 0.43 events in the past 6 years Mean monthly average minimum daily 0.55 0.55 temperature (years: 1983-2013) Mean, standard deviation of monthly average 0.53 0.53 maximum daily temperature (years: 1983- 2013 Mean, standard deviation of monthly average 0.71 0.71 precipitation (years: 1983-2013) Average time to health facility by walking 0.13 0.21 Health 0.228 0.242 Households having family member with 0.09 0.13 chronic illness Households who family member had to miss 0.38 0.33 work or school in the past 6 months Average malaria exposure prevention index 0.31 0.30 Overall LVI 0.386 0.393 Source: Survey data (2018) 138 University of Ghana http://ugspace.ug.edu.gh When all the sub-components under health were aggregated, the control group showed more vulnerability (LVI= 0.288) than the treatment group (LVI=0.242). control group are less vulnerable in natural disaster and climate change (LVI = 0.488) compared to the treatment group (LVI=0.524). From appendix 5 table 10A, about 95% of the households in the control group did not receive warning about pending natural disaster whereas 91% of those in the treatment group did not receive warning about impending disaster. The overall aggregated LVI computed showed that the treatment group was more vulnerable (LVI=0.393) in terms of climate change than their control counterpart (LVI=0.386). In table 4.20, the LVI-IPCC was also computed by grouping the seven major components into three categories, namely, exposure, sensitivity and adaptive capacity. The exposure indices were 0.488 and 0.524 for the control and treatment groups respectively. For sensitivity, the indices were 0.332 and 0.344 respectively for control and treatment groups. The results suggest that the treatment group are more vulnerable in the area of exposure and sensitivity than the control group. The index for adaptive capacity was 0.400 and 0.383 for the control and treatment groups respectively. The results suggest that treatment group can adapt to climate change better than the control group. Table 4.20: LVI-IPCC Contribution Factors to Climate Change IPCC Contributing Factors to Vulnerability Control Treatment Exposure 0.488 0.524 Adaptive Capacity 0.400 0.383 Sensitivity 0.332 0.344 LVI-IPCC 0.029 0.048 Source: Survey data (2018) 139 University of Ghana http://ugspace.ug.edu.gh Adaptive capacity is an important variable in determining the impact of climate change on livelihoods of the poor. While sensitivity and exposure are exogenously determined, the adaptive capacity could be enhanced through capacity building, income, assets accumulation, information and knowledge sharing. The treatment group being less vulnerable in adaptive capacity could be associated with the SOGS which provide support to farmers to improve their livelihood outcomes. The work of Aniah et al. (2019) on smallholder farmers' livelihood adaptation to climate variability and ecological changes in the savanna agro ecological zone of Ghana confirmed the important role of building the adaptive capacity of SHF farmers as an ideal approach to addressing their vulnerability. To further confirm contribution factors to vulnerability, IPCC vulnerability exposure, sensitivity and adaptive capacity is presented in the vulnerability triangle as shown in Figure 4.4. It ranges from 0 (low contributing factor) and 0.6 (high contributing factor). The vulnerability triangle further confirmed that the treatment group was more vulnerable to climate change in the area of exposure and sensitivity and less vulnerable in the area of their adaptive capacity. Exposure 0.6 0.4 0.2 0 Sensitivity Adaptive Capacity Control Treatment Figure 4.4: Vulnerability Triangle Diagram of LVI-IPCC for Control and Treatment Source: survey data, (2018) 140 University of Ghana http://ugspace.ug.edu.gh Finally, all the major components are summarized in vulnerability spider in Figure 4.5. The vulnerability spider diagram ranges between 0 (less vulnerable) and 0.6 (Extremely vulnerable). The control group was more vulnerable in terms of livelihood strategies whereas the treatment group was more vulnerable in terms of natural disaster and climate variability, food, health, water and social network. The two groups obtained the same level of vulnerability in terms of the socio-demographic profile. The overall LVI estimates for the control and treatment groups are 0.386 and 0.393 respectively. This implies that in the overall LVI computed, the treatment group was more vulnerable than the control group. Water 0.6 0.5 0.4 Socio-demographicHealth profile 0.3 0.2 0.1 0 Natural Disaster and Food climate variability Livelihood Strategies Social Networks Control Treatment Figure 4.5: Vulnerability Spider Diagram of the Major Components of the LVI for Control and Treatment Groups Source: Survey data, (2018) 141 University of Ghana http://ugspace.ug.edu.gh 4.9.2 Average Treatment effects on Smallholder Farmers Vulnerability to Climate Change The results of FILM: The results of FIML of the ESR that control for unobservable selection bias and the significance of the ESR model are reported in Table 4.21. The second and fourth column in the table contain the treatment and control equations of FIML respectively. To analyse the correlation between decision to participate in SOGS and effect of SOGS on reducing SHF vulnerability to climate change, broad set of explanatory variables were included in the model to establish their influence in participation and effect on reducing SHF vulnerability to climate change. Table 4.21: Full Information Maximum Likelihood Estimates of effects of Sorghum Outgrower Scheme on Reducing Vulnerability to Climate Change Treatment group Control group Variable Coef. Std. Err. Coef. Std. Err. Gender -0.001 0.005 0.001 0.0027 Age2 0.007 0.000 -0.000 0.000 Age -0.001 0.001 0.000 0.000 HH Status 0.002 0.007 0.003 0.003 Household size -0.002 0.001 -0.000 0.000 Educated -0.003 0.006 -0.001 0.003 Occupation -0.002 0.007 0.003 0.004 Farming Experience 0.001 0.001 0.000 0.000 Farm Size -0.001 0.003 0.005** 0.002 FBO Membership -0.013* 0.008 0.005 0.016 Access to Credit -0.029*** 0.011 -0.028*** 0.007 Number of extension visits received -0.010*** 0.004 -0.002 0.003 Constant 0.035 0.029 0.011 0.010 /lns -3.602*** 0.098 -4.111*** 0.046 /r -0.944*** 0.287 0.071 0.347 Sigma 0.027 0.003 0.016 0.001 Rho -1.474** 0.131 0.071 0.345 Number of obs = 516 Wald chi2(12) = 19.98*** Log likelihood = 816.73** Prob > chi2 = 0.067 *, ** and *** denotes significance at 10%, 5% and 1% respectively Source Survey data, (2018) 142 University of Ghana http://ugspace.ug.edu.gh Apart from FBO membership, access to extension services and access to credit on the treatment equation being significant, the rest of the explanatory variables are not significant and have no influence on participation decision and reduction in vulnerability to climate change. Access to credit is statistically significant at 1% with negative coefficient implying that access to credit influence farmers decision to participate negatively. Otherwise, farmers with access to credit participating in the SOGS are less likely to reduce their vulnerability to climate change. For the control equation, the only significant variable influencing participation is access to credit at 1% significance with negative coefficient. This implies that farmers in the control group who have access to credit are less likely to participate in the SOGS. The results is consistent with literature on difficulty SHF have in accessing credit and will not be willing to participate in intervention provided they already have access to credit ( Uaiene et al., 2009). The likelihood ratio (LR) chi-square (χ2 (11) = 816.73) is statistically significant at 5% level. This finding suggests that the explanatory variables in the models jointly influenced participation decision and vulnerability of the treatment and control groups to climate change. The chi-square statistics of 19.98 is statistically significant and different from zero meaning that the independence assumption given to the selection and the outcome equations must be rejected at 1% level. The Results of ESR: The ESR results on effect of SOGS are presented in table 4.22. Cases (a= 0.393) and (b=0.386) along the diagonal represent the actual expectations observed for both treatment and control group on SHF vulnerability to climate change respectively. This finding suggests that, the treatment group is more vulnerable to climate change than the control group. Cases (c=0.395) and (d=0.376) represent the counterfactual expected 143 University of Ghana http://ugspace.ug.edu.gh outcomes. The counterfactual outcome is lower than the actual outcomes for control group and higher than the actual outcome for the treatment group. this means that if the treatment group had not participated in the SOGS, their vulnerability to climate change would have been higher than their current level of vulnerability and if the control group had participated in the SOGS their vulnerability would have been lower than their current level of vulnerability. This implies that, SOGS has positive effect on vulnerability to climate change for both groups. Table 4.22: Endogenous Switching Regression Results of Treatment effects of Sorghum Outgrower Scheme on Vulnerability to Climate Change Treatment Status Decision Decision Not Treatment Effect to Participate to Participate Treatment Group (a)0.393 (c) 0.395 -0.002 Control Group (d) 0.376 (b) 0.386 -0.01 Heterogeneity effects BH1= 0.017 BH2= 0.009 TH= 0.008 Source Survey data, (2018) To further confirm the treatment effect, the last column of table 4.22 contain the treatment effects results. The treatment effect is -0.002 and -0.01 for the treatment and control groups respectively. The lower treatment effect value for treatment group than the control group suggest that the SOGS has more effect on treatment group. The last row of table 4.22 contain heterogeneity effects. The coefficient for heterogeneity effect on treatment and control groups are 𝐵𝐻1 =0.017 and 𝐵𝐻2 = 0.009 respectively. The differences in 𝐵𝐻1 and 𝐵𝐻2 suggests existence of unobserved characteristics of members in the treatment group and the control group which suggest that, the effect of SOGS on reducing SHF vulnerability to climate change is not limited to only participating in the SOGS but other unobserved characteristics such as individual skills of farmers. The transitional heterogeneity 144 University of Ghana http://ugspace.ug.edu.gh effect (TH) is 0.008 which is positive suggesting that participating in SOGS has positive effect on reducing SHF vulnerability to climate change for both treatment and control groups. The results are consistent with empirical literature that enhancing SHF asset based through capacity building, information and support system that improve their economic activities build their adaptive capacity and make them resilience to climate change (Abdul-Razak & Kruse, 2017; Jamshidi et al., 2019; Makate et al., 2019). Given that the SOGS support farmers with extension services, climate information and guaranteed market leading to increase productivity and profitability, they are more likely to be less vulnerable to climate change. The PSM results presented in Table 4.23 shows ATT and ATU estimates are -0.027 and - 0.033 respectively. The PSM results are similar to the ESR results on positive effect of SOGS on reducing SHF vulnerability to climate change. There are small variations on the ATT values and ATU values with that of the ESR results on effect on SOGS on reducing SHF vulnerability to climate change. Table 4.23: Treatment effects on Vulnerability to Climate Change Effect Std. Err T-stat ATT -0.027*** .003 -8.4 ATU -0.033*** .005 -7.0 ATE -0.031*** .004 -8.6 *** denotes significance at 1% Source Survey data (2018). The relatively lower values of ATT compared to treatment effect of ESR results suggest that PSM overestimated the effect of SOGS on reducing SHF vulnerability to climate change on the treatment group. The differences in the treatment effect under ESR and PSM could be attributed to inability of PSM to account for unobservable characteristics of farmers that could 145 University of Ghana http://ugspace.ug.edu.gh influence their ability to reduce vulnerability to climate change apart from participating in the SOGS. 4.9.3 Summary of Smallholder Farmers Vulnerability to Climate Change In summary, the differences in perceived vulnerability to climate change between the treatment and control observed in this study could be explained by the fact that vulnerability in itself is fluid and subject to multiple conceptualizations. Aniah et al. (2019) stated that farmers who have access to extension services are likely to better understand changes in the climate as extension services provide information about climate and weather. This emphasises the fact that the perception of people changes with the different services and information available to them. As the findings suggests, about 54% of farmers on the SOGS had access to extension services compared to only barely 7% of those who are not on the SOGS. For this reason, farmers on the SOGS are likely to be aware of what constitute vulnerability and can relate to climate change information that extension officers may have given them compared to those who are not and this could influence their responses during the data collection on their vulnerability to climate change. Secondly, farmers on the SOGS view farming as a business and are likely to relate climate change and variability to business loses. They are predisposed to having a heightened sense of vulnerability in comparison to control group who may not necessarily farm for business. The treatment group’s sense of vulnerability to climate change does not necessarily suggest that they are worse off with the impact of climate change. While the treatment group may appear vulnerable, they are more resilient to the impact of climate change. Resilience in general sense means, a system’s or a household’s ability to deal with stresses and disturbances and maintaining its basic structure and ways of functioning, capacity for self-organisation, 146 University of Ghana http://ugspace.ug.edu.gh and capacity to learn and adapt to change(UNEP & UNDP, 2013). Since treatment group have higher productivity and profitability as the data suggest, their resource banks are likely to be better than the control group. Even though the data suggest that they are more vulnerable to climate change, being less vulnerable to adaptive capacity suggest that they are most likely to be in a better position to cope with the changes in climate and become more resilient compared to the control group. 147 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE SUMMARY, CONCLUSION AND RECOMMENDATION 5.1 Introduction This chapter summarises issues raised in chapter one, the methods used to investigate them and the key findings. Conclusion of the study, recommendations for policy makers, research gaps and direction for future research are all presented in this chapter. 5.2 Summary and Major Findings The role of smallholder farmers in Ghana’s agricultural development is highly acknowledged by all stakeholders. Their contribution is however, limited by diverse constraints leading to low productivity, high postharvest losses and low profitability. Their livelihoods are further worsened by negative impact of climate which is making them vulnerable. Approaches adopted by policy makers to optimise SHF economic activities have not yielded much results. Smallholder farmers still produced far below their potentials and remained the poorest among the poor. The OGS concept is recommended globally as an ideal approach to transform SHF farmers activities as it support them with productive resources, knowledge and guaranteed market. Regardless of the optimism expressed by several literature on importance of OGS in transforming the activities of SHF, there is limited research on its effect on the livelihoods of SHF producing indigenous crops like sorghum in Ghana. This thesis analysed the effects of OGS to establish whether participation has any effects on the livelihoods of smallholder sorghum farmers in northern Ghana or not. The specific 148 University of Ghana http://ugspace.ug.edu.gh objectives of the thesis were to identify factors influencing SHF participation in SOGS; examine the effects of OGS on SHF productivity, postharvest losses, profitability and the effects on reducing their vulnerability to climate change. Using multistage sampling procedure, primary and secondary data was sourced qualitatively and quantitatively. The secondary data focused mainly on documentary analysis of published and unpublished literature and time series data from the Ghana Meteorological Agency. The qualitative data from primary sources were mainly from lead farmers, FGD and face-to-face interviews. The quantitative data from primary source was obtained from 516 SHF households on treatment and control groups. The data was analysed both quantitatively and qualitatively and integrated at the report writing stage. The quantitative data was captured electronically using CAPI and transferred onto a platform and analysed using STATA version 15. The ESRM was employed to address selectivity bias to determine the effects of SOGS on the livelihoods of SHF while PSM technique was adopted for robustness check. The following were findings of the thesis: 1.0 Out of 15 variables modelled using probit regression to determine the factors influencing SHF participation in SOGS, 11 variables were found to be statistically significant and influencing participation. The most important among the variables were age of the farmer, gender of the farmer, access to market, extension services and being a member of FBO. The study found older and female farmers less likely to participate relative to younger and male farmers respectively. While relatively younger farmers interest in participation signifies brighter future for sustainability of the SOGS, the likelihood of more male farmers participating compare to female farmers present a glooming picture for the efforts to bridge gender inequality between men and women on access to productive resources such as land. 149 University of Ghana http://ugspace.ug.edu.gh Smallholder farmers participation being influenced by market access explains the role of market in reorienting the mindset of SHF from their subsistence farming practices to farming as a business. Being a member of FBO was also found to positively influence SHF participation in SOGS. In conclusion, being a male farmer, SHF expectation of access to guaranteed market and belonging to FBO are the most important factors influencing their participation in the SOGS. This result is consistent with the hypothesis of the thesis. 2.0 The results on productivity suggest that the average productivity of SHF on the treatment group was 1,207kg/ha and control group was 820 kg/ha. The counterfactual outcome from the ESR shows that if the treatment group had decided not to participate, their productivity would have been lower and if the control group had decided to participate, their productivity would have been higher. The findings suggest a positive effect of SOGS on the productivity of both groups but the effect is higher for the treatment group than the control group. This implies that the OGS has unequal effect on random smallholder sorghum farmer who choose to join the scheme. The results illustrate existence of unobserved heterogeneity effects which influence their participation decision and at the same time make them more productive than a random smallholder farmer who choose to join the scheme. The result is inconsistent with the hypothesis of the thesis that SOGS increase productivity of all smallholder farmers who participate in the SOGS. In conclusion, the thesis argues that the outgrower scheme concept in its current form favours who are endowed with productive resources. For profitability, the study found profitability to be GHS 575/ha and GHS504/ha for treatment and control group respectively. The ESR on counterfactual outcomes suggest that had the 150 University of Ghana http://ugspace.ug.edu.gh treatment group choose not to participate in the SOGS, they would have made losses of GHS 482/ha and if the control group had decided to participate, they would have realised profitability of GHS 43/ha against their current losses of GHS 93/ha. The findings imply that SOGS has positive effect on profitability for both the treatment and the control group but the effect is higher for the treatment group than the control group. Inclusion, the smallholder sorghum farmers could increase their profitability by participating in the SOGS. But the margin of profitability will largely depend on intrinsic characteristic of the farmer of which farmers who are endowed with resources are likely to have more profitability. The finding is inconsistent with the hypotheses of the thesis that, OGS has effect on the profitability of any smallholder sorghum farmer who choose to join the OGS. On postharvest losses, the treatment and control groups loss 161 kg/ha and 127 kg/ha respectively. Average value losses are GHS193/ha and 166/ha, and average percentage losses are 14.24% and 27.22 % respectively for the treatment and the control groups. The result show that on real terms, the treatment group loss more produces due to PHL than the control group but on percentage terms, the control group loss more sorghum than the treatment group. The counterfactual outcomes show that if the treatment group had not participated in the SOGS, they would have recorded losses higher than their current losses and if the control group were to participate, they would have recorded lower than their current losses. In conclusion, random SHF who choose to participate in SOGS will reduce their postharvest losses compare to not participating. The finding is consistent with the hypothesis of the thesis that participating in OGS help to reduce SHF postharvest losses. 3.0 The LVI results on smallholder farmers vulnerability to climate change shows that both the treatment and control groups are vulnerable to climate change with the treatment group 151 University of Ghana http://ugspace.ug.edu.gh being relatively more vulnerable than the control group. For the LVI-IPCC results on contributing factors to vulnerability, the treatment group is more vulnerable in the area of sensitivity and exposure than the control group but less vulnerable in adaptive capacity than the control group. The ESR results on treatment effects of SOGS on reducing smallholder farmers vulnerability to climate change suggest that the SOGS has minimal effect on vulnerability to climate change. The limited effect of the SOGS to smallholder farmers vulnerability to climate change could be due to the buyers not providing support in the area of water, food and health major components as part of the OGS package to farmers. The result is inconsistent with the hypotheses of the study that SOGS scheme reduces smallholder farmers vulnerability to climate change. In conclusion, the current sorghum outgrower scheme promoted by Guinness Ghana Brewery Limited is largely on unequal demand and supply bases which skew the capture of benefits towards the buyers. The is not a pro-poor scheme and may suffer high exit rate leading to unsustainability in future if the concept is not modify to consider the welfare of farmers as part of the scheme package. 5.3. Recommendation Given that Ghana’s agricultural sector is dominated by older farmers and also given the low interest of the youth to take farming as a business, the probability of relatively younger people likely to participate in the sorghum outgrower scheme compared to relatively older people provide opportunity to attract more youth into farming through the SOGS. Also, the scheme giving preference to farmers who belong to FBO suggests that when the youth are encouraged to join FBO their participation in OGS would increase. Any government policy that promote 152 University of Ghana http://ugspace.ug.edu.gh the youth to join FBO could increase their participation in OGS leading to more youth developing interest in farming. Market access was another important factor influencing participation. The efforts to transform smallholder subsistence farming to market-oriented farming can be achieved through investing in projects that can provide guaranteed market for SHF. Increase investment in market infrastructure such as feeder roads and storage facilities will increase market access and attract more people to go into farming. To increase productivity, reduces postharvest losses and increase profitability of smallholder farmers, government could consider modifying the current outgrower scheme to make it pro- poor and adopt it in the implementation of the government agricultural flagship programmes such as the planting for food and jobs, rearing for food and jobs and planting for export and rural development which sought to increase smallholder farmer’s livelihood outcomes. Smallholder farmers will benefit more by increasing their productivity, their postharvest loss reduction and their profitability from government support such as fertilizer and seed subsidy programme, extension services and marketing arrangement through channelling the support to private investors under OGS arrangement with government playing a coordination and monitoring to ensure mutual benefit. On climate change, the results suggest that, smallholder farmers in the study area are vulnerable to climate change. The ESR results on effects of SOGS on reducing smallholder farmers vulnerability to climate change suggests that the SOGS has minimal effect on reducing vulnerability. For the current outgrower scheme to become sustainable, the thesis recommend to private sector promoting the scheme to consider including welfare variables such as access to water, food and access to health care as part of the outgrower scheme package. 153 University of Ghana http://ugspace.ug.edu.gh Given the global call for nations to increase their efforts to reduce the effect of climate change and the efforts by Ghana government to help smallholder farmers become less vulnerable to climate change, to achieve this goal, this thesis recommend to government to adopt and modify the current outrgrower scheme to include welfare variables such as access to water, food and health care as part of the outgrower scheme concept. Suggestion for Future Research: Given the importance of SOGS in addressing the constraints of SHF, future research should broaden the study areas to cover different ecological and socio-cultural locations that will allow for comparison of performance. This will help to understand how different ecological and cultural dynamics affect sorghum outgrower scheme outcomes. This research was limited to only two regions due to limited resources and time constraints. Secondly, the changes in climate has different implications on smallholder farmer’s productivity, profitability and their vulnerability to climate change in different years. In a year of stable weather, farmers may obtain higher productivity and profitability. Panel data that explains trend of productivity and profitability of farmers on SOGS across different years will provide more robust results. This study used cross-sectional data due to lack of existing time series data on SOGS and limited time frame for the study to collect time series data. Future research could focus on gathering time series data. On effect of SOGS on postharvest loss reduction, the study focuses on only on quantity and value losses. Future research could consider quality loss, especially aflatoxin contamination. Considering the background of the researcher, the study was limited to quantity and value losses. 154 University of Ghana http://ugspace.ug.edu.gh REFERENCES Abdul-Razak, M., & Kruse, S. (2017). The adaptive capacity of smallholder farmers to climate change in the Northern Region of Ghana. Climate Risk Management, 17, 104– 122. https://doi.org/10.1016/j.crm.2017.06.001 Abebe, G. K., Bijman, J., Kemp, R., Omta, O., & Tsegaye, A. (2013). Contract farming configuration : Smallholders ’ preferences for contract design attributes. Food Policy, 40, 14–24. https://doi.org/10.1016/j.foodpol.2013.01.002 Adams, T., Gerber, J., & Amacker, M. (2019). Constraints and opportunities in gender relations : Sugarcane outgrower schemes in Malawi. World Development, 122, 282– 294. https://doi.org/10.1016/j.worlddev.2019.05.029 Adesina, A. A., & Baidu Forson. (1995). Farmers ’ perceptions and adoption of new agricultural technology : evidence from analysis in Burkina Faso, 13, 1–9. Adeyeri, O. E., Lawin, A. E., Laux, P., Ishola, K. A., & Ige, S. O. (2019). Analysis of climate extreme indices over the Komadugu-Yobe basin , Lake Chad region : Past and future occurrences. Weather and Climate Extremes, 23(November 2018), 100194. https://doi.org/10.1016/j.wace.2019.100194 Adu, D. T., Kuwornu, J. K. M., & Anim-somuah, H. (2018). Application of livelihood vulnerability index in assessing smallholder maize farming households ’ vulnerability to climate change in Brong-Ahafo region of Ghana. Kasetsart Journal of Social Sciences, 39(May), 22–32. Affognon, H., Mutungi, C., Sanginga, P., & Borgemeister, C. (2015). Unpacking postharvest losses in sub-Saharan Africa: A Meta-Analysis. World Development, 66, 49–68. https://doi.org/10.1016/j.worlddev.2014.08.002 Agbelengor, B. M. (2015). Outgrower models and the modernisation of agriculture: the case of Blue Skies Company Limited in the Akwapim South Municipality. University of Ghana. https://doi.org/10.1038/253004b0 AGRA. (2016). Africa Agriculture Status Report 2016; progress towards agricultural transformation in Africa. AGRA. (2017). Africa Agriculture Status Report: The Business of Smallholder Agriculture in Sub-Saharan Africa. Alliance for a Green Revolution in Africa. Nairobi. AGRA. (2018). Africa Agriculture Status Report 2018; Catalying Government Cpacity to Drive Agricultural. Nairobi. Ahmed, R. (2014). Performance Evaluation of Draught Animal Power Cultivator. Asian Journal of Applied Science and Engineering, 3(8), 111–115. https://doi.org/10.15590/ajase/ABSTRACT Ajewole. (2010). Farmer ’ s response to adoption of commercially available organic fertilizers in Oyo state , Nigeria, 5(18), 2497–2503. Akoto, A. (2019). Akufo-Addo launches ‘Rearing for Food and Jobs.’ Retrieved July 20, 155 University of Ghana http://ugspace.ug.edu.gh 2019, from https://www.ghanaweb.com/GhanaHomePage/business/Rearing-for-Food- and-Jobs-will-develop-livestock-industry-Akufo-Addo-757765 Akuriba, M., & Asuming-Brempong, S. (2012). Productivity and Competitiveness Of Sorghum Production in Northern Ghana; A Policy Analysis Matrix Approach. Ghana Journal of Development Studies, 8(2). https://doi.org/10.4314/gjds.v8i2.7 Al-Hassan, R. M., Egyir, I. S., & Abakah, J. (2013). Farm household level impacts of information communication technology (ICT)-based agricultural market information in Ghana. Journal of Development and Agricultural Economic, 5(4), 161–167. https://doi.org/10.5897/ Aleskerov, F., Bouyssou, D., & Monjardet, B. (2007). Utility maximization, choice and preference. Utility Maximization, Choice and Preference, 1–283. https://doi.org/10.1007/978-3-540-34183-3 Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050. Land Use Policy, 20(4), 375. https://doi.org/10.1016/S0264-8377(03)00047-4 Ali, D. A., Collin, M., Deininger, K., Dercon, S., Sandefur, J., & Zeitlin, A. (2016). Small price incentives increase women’s access to land titles in Tanzania. Journal of Development Economics, 123, 107–122. https://doi.org/10.1016/j.jdeveco.2016.06.001 Almas, L. K., & Obembeb, O. (2014). Agribusiness model in Africa: A case study of Zambeef products PLC. International Food and Agribusiness Management Review, 17(SPECIALISSUEB), 111–116. Alobo, S. (2019). Household livelihood diversification and gender: Panel evidence from rural Kenya. Journal of Rural Studies, 69(February 2017), 156–172. https://doi.org/10.1016/j.jrurstud.2019.03.001 Anderson, J. E. (2019). Access to land and permits: Firm-level evidence of impediments to development in transition countries. Journal of Economics and Business, 101(September 2018), 38–57. https://doi.org/10.1016/j.jeconbus.2018.09.001 Angelucci, F. (2013). Analysis of incentives and disincentives for Sorghum in Kenya. Technical note series. Food and Agriculture Organization of the United NAtions. Rome. Retrieved from http://www.fao.org/mafap Aniah, P., Kaunza-Nu-Dem, M. K., & Ayembilla, J. A. (2019). Smallholder farmers’ livelihood adaptation to climate variability and ecological changes in the savanna agro ecological zone of Ghana. Heliyon, 5(4), e01492. https://doi.org/10.1016/j.heliyon.2019.e01492 Antwi-Agyei, P., Dougill, A. J., Stringer, L. C., Nii, S., & Codjoe, A. (2018). Climate Risk Management Adaptation opportunities and maladaptive outcomes in climate vulnerability hotspots of northern Ghana. Climate Risk Management, 19(April 2017), 83–93. https://doi.org/10.1016/j.crm.2017.11.003 Antwi, E. K., Boakye-danquah, J., Owusu, A., Seram, L., Mensah, R., Boafo, A., & Apronti, P. (2015). Community vulnerability assessment index for fl ood prone savannah agro- ecological zone : A case study of Wa West District , Ghana, 10, 56–69. 156 University of Ghana http://ugspace.ug.edu.gh https://doi.org/10.1016/j.wace.2015.10.008 Arslan, A., Belotti, F., & Lipper, L. (2017). Smallholder productivity and weather shocks: Adoption and impact of widely promoted agricultural practices in Tanzania. Food Policy, 69, 68–81. https://doi.org/10.1016/j.foodpol.2017.03.005 Aruna, C., & Visarada, K. B. R. S. (2018). Sorghum Grain in Food and Brewing Industry. Breeding Sorghum for Diverse End Uses. Elsevier Ltd. https://doi.org/10.1016/b978-0- 08-101879-8.00013-9 Asfaw, A., Simane, B., Hassen, A., & Bantider, A. (2018). Variability and time series trend analysis of rainfall and temperature in northcentral Ethiopia : A case study in Woleka sub-basin. Weather and Climate Extremes, 19(December 2017), 29–41. https://doi.org/10.1016/j.wace.2017.12.002 Asfaw, S., Shiferaw, B., Simtowe, F., & Lipper, L. (2012). Impact of modern agricultural technologies on smallholder welfare: Evidence from Tanzania and Ethiopia. Food Policy, 37(3), 283–295. https://doi.org/10.1016/j.foodpol.2012.02.013 Awunyo-Victor, D., & Al-hassan, R. M. (2014). Credit constraints and smallholder maize production in Ghana Dadson Awunyo-Vitor *. Int. J. Agricultural Resources, Governance and Ecology, Vol. 10,(November 2014), 239–256. https://doi.org/10.1504/IJARGE.2014.064007 Azumah, S. B., Donkoh, S. A., & Ehiakpor, D. (2016). Examining the determinants and effects of Contract Farming on Farm Income in the Northern Region of Ghana, (April). Baiyegunhi, L. J. S., Majokweni, Z. P., & Ferrer, S. R. D. (2019). Impact of outsourced agricultural extension program on smallholder farmers’ net farm income in Msinga, KwaZulu-Natal, South Africa. Technology in Society, 57(June 2018), 1–7. https://doi.org/10.1016/j.techsoc.2018.11.003 Barrett, C. B., Bachke, M. E., Bellemare, M. F., Michelson, H. C., Narayanan, S., & Walker, T. F. (2012). Smallholder participation in contract farming: Comparative evidence from five countries. World Development, 40(4), 715–730. https://doi.org/10.1016/j.worlddev.2011.09.006 Basavaraja, H., Mahajanashetti, S. B., & Udagatti, N. C. (2007). Economic analysis of post- harvest losses in food grains in india: a case study of karnataka. Agricultural Economics Research Review, 20(June), 117–126. Baumann, P. (2000). Equity and Efficiency in Contract Farming Schemes : The Experience of Agricultural tree Crops (No. 139). Bellemare, M. F. (2012). As You Sow , So Shall You Reap : The Welfare Impacts of Contract Farming. World Development, 40(7), 1418–1434. https://doi.org/10.1016/j.worlddev.2011.12.008 Bellemare, M. F., & Bloem, J. R. (2018). Does contract farming improve welfare? A review. World Development, 112, 259–271. https://doi.org/10.1016/j.worlddev.2018.08.018 Bhattacharjee, K., & Behera, B. (2018). Determinants of household vulnerability and 157 University of Ghana http://ugspace.ug.edu.gh adaptation to floods: Empirical evidence from the Indian State of West Bengal. International Journal of Disaster Risk Reduction, 31(July), 758–769. https://doi.org/10.1016/j.ijdrr.2018.07.017 Boansi, D., Tambo, J. A., & Müller, M. (2017). Analysis of farmers ’ adaptation to weather extremes in West African Sudan Savanna. Weather and Climate Extremes, 16(March), 1–13. https://doi.org/10.1016/j.wace.2017.03.001 Boateng, J., & Nyaaba, C. (2014). Perceptions on the Impact of METASIP on Food Security in Ghana Perceptions on the Impact of METASIP on Food Security in. Iiste, 4(15), 1– 12. Brigitte, I., & Ragasa, C. (2018). Do development projects crowd-out private sector activities ? Evidence from contract farming participation in Northern Ghana ☆. Food Policy, 74(November 2017), 9–22. https://doi.org/10.1016/j.foodpol.2017.11.001 Brooks, J., Cervantes-Godoy, D., & Jonasson, E. (2009). Strategies for smallholders in developing countries: commercialisation, diversification and exit. OECD, Trade and Agriculture Directorate, 5(June), 1–21. Retrieved from https://www.kent.ac.uk/economics/documents/research/ceas/2009/Contributed papers/128.pdf Caliendo, M. (2005). Discussion Papers Sabine Kopeinig ** Some Practical Guidance for the Implementation of Propensity Score Matching, (April). Carter, D. W., & Milon, J. . (2005). Price knowledge in household demand for Utility services. Land Economics, 81 (2)(2005), 265–283. Chambers, R., & Conway, G. (1991). Sustainable rural livelihoods: Practical concepts of the 21st century. Chauvin, N. D., Mulangu, F., & Porto, G. (2012). Food Production and Consumption Trends in Sub-Saharan Africa: Prospects for the Transformation of the Agricultural Sector. Working Papers 2012-11, (February), 76. https://doi.org/10.1016/j.foodpol.2013.10.006 Chegere, M. J. (2018). Post-harvest losses reduction by small-scale maize farmers : The role of handling practices. Food Policy, 77(May), 103–115. https://doi.org/10.1016/j.foodpol.2018.05.001 Chigbu, U. E. (2019). Masculinity, men and patriarchal issues aside: How do women’s actions impede women’s access to land? Matters arising from a peri-rural community in Nigeria. Land Use Policy, 81(October 2018), 39–48. https://doi.org/10.1016/j.landusepol.2018.10.033 Chinwendu, O. G., Sadiku, S. O. E., Okhimamhe, A. O., & Eichie, J. (2017). Households Vulnerability and Adaptation to Climate Variability Induced Water Stress on Downstream Kaduna River Basin, 247–267. https://doi.org/10.4236/ajcc.2017.62013 Choi, H. (2009). The Profitable Theory of Consumers’ and Producers’ Surplus: Positive Demand, Negative Supply? Ssrn.Com/Abstract=1520419. 158 University of Ghana http://ugspace.ug.edu.gh Cooke, E., Hague, S., & Mckay, A. (2016). The Ghana Poverty and Inequality Report. Coulter, J., & Onumah, G. (2002). The role of warehouse receipt systems in enhanced commdity marketing a rural livelihoods in Africa. Food Policy, 27(4), 319–337. https://doi.org/10.1016/S0306-9192(02)00018-0 Darfour, B., & Rosentrater, K. A. (2016). Agriculture and Food Security in Ghana. https://doi.org/10.13031/aim.20162460507 Darku, A. B., Malla, S., & Tran, K. C. (2016). Sources and Measurement of Agricultural Productivity and Efficiency in Canadian Provinces: Crops and Livestock. Canadian Journal of Agricultural Economics, 64(1), 49–70. https://doi.org/10.1111/cjag.12061 Delbridge, T. A., Fernholz, C., King, R. P., & Lazarus, W. (2013). A whole-farm profitability analysis of organic and conventional cropping systems. Agricultural Systems, 122, 1–10. https://doi.org/10.1016/j.agsy.2013.07.007 Devkota, M., Devkota, K. P., Acharya, S., & McDonald, A. J. (2019). Increasing profitability, yields and yield stability through sustainable crop establishment practices in the rice-wheat systems of Nepal. Agricultural Systems, 173(March), 414–423. https://doi.org/10.1016/j.agsy.2019.03.022 DFID. (1999). Sustainable Livelihoods Guidance Sheets. London UK. Dharmasiri, L. M. (2012). Measuring Agricultural Productivity Using the Average Productivity Index (API). Sri Lanka Journal of Advanced Social Studies, 1(2). https://doi.org/10.4038/sljass.v1i2.4855 Diao, X., Cossar, F., Houssou, N., & Kolavalli, S. (2014). Mechanization in Ghana: Emerging demand, and the search for alternative supply models. Food Policy, 48, 168– 181. https://doi.org/10.1016/j.foodpol.2014.05.013 Dickinson, K. L., Monaghan, A. J., Rivera, I. J., Hu, L., Kanyomse, E., Alirigia, R., … Wiedinmyer, C. (2017). Changing weather and climate in Northern Ghana: comparison of local perceptions with meteorological and land cover data. Regional Environmental Change, 17(3), 915–928. https://doi.org/10.1007/s10113-016-1082-4 Dittoh, S., & Akuriba, M. (2018). Africa’s looming food and nutrition insecurity crisis – a call for action. Ghana Journal of Agricultural Economics and Agribusiness, 1(1), 148– 170. Dixon, J., Tanyeri-Abur, A., & Wattenbach, H. (2004). Framework for analyzing impacts of globalization on smallholders. AGSF Occasional Paper 5. Djalante, R. (2019). Key assessments from the IPCC special report on global warming of 1.5 °C and the implications for the Sendai framework for disaster risk reduction. Progress in Disaster Science, 1(March), 100001. https://doi.org/10.1016/j.pdisas.2019.100001 Djameh, C., Saalia, F. K., Sinayobye, E., Budu, A., Essilfie, G., Mensah-Brown, H., & Sefa- Dedeh, S. (2015). Optimization of the sorghum malting process for pito production in Ghana. Journal of the Institute of Brewing, 121(1), 106–112. 159 University of Ghana http://ugspace.ug.edu.gh https://doi.org/10.1002/jib.191 Donkor, E., & Owusu, V. (2019). Mineral Fertiliser Adoption and Land Productivity : Implications for Securing Stable Rice Production in. MDPI, 59(March), 1–13. https://doi.org/10.3390/land8040059 Doss, C. R., & Morris, M. L. (2001). How does gender affect the adoption of agricultural innovations ? The case of improved maize technology in Ghana, 25. Drafor, I. V. Y., Kunze, D., & Sarpong, D. B. (2013). Food Security: How Rural Ghanaian Households Respond to Food Shortages in Lean Season. International Journal of Agricultural Management, 2(4), 199–206. https://doi.org/10.5836/ijam/2013-04-03 Dzanku, F. M. (2015). Transient rural livelihoods and poverty in Ghana. Journal of Rural Studies, 40, 102–110. https://doi.org/10.1016/j.jrurstud.2015.06.009 Eaton, C., & Shepherd, A. (2001). Contract farming Partnerships for growth. FAO Agricultural Services BullentinAgricultural Services Bullentin, 145, 1–182. Retrieved from http://www.fao.org/3/y0937e/y0937e00.pdf Ebata, A., & Hernandez, M. A. (2017). Linking smallholder farmers to markets on extensive and intensive margins: Evidence from Nicaragua☆. Food Policy, 73(September), 34– 44. https://doi.org/10.1016/j.foodpol.2017.09.003 Ecker, O. (2018). Agricultural transformation and food and nutrition security in Ghana : Does farm production diversity ( still ) matter for household dietary diversity ? Food Policy, 79(August), 271–282. https://doi.org/10.1016/j.foodpol.2018.08.002 Etwire, P. M., Al-Hassan, R. M., Kuwornu, J. K. M., & Osei-Owusu, Y. (2013). Application of Livelihood Vulnerability Index in Assessing Vulnerability to Climate Change and Variability in Northern Ghana. Journal of Environment and Earth Science, 3(2), 157– 170. European Cooperation for Rural Development (EUCORD). (2008). West Africa Sorghum Value Chain Development Project (Ghana and Sierra Leone), CFC/FIGG/34; : Mid- Term Evaluation Report. FAO. (2011). Continental Programme on Post-harvest Losses (PHL) Reduction Rapid Needs Assessment – Working Document - Ghana. Rome. FAO. (2014a). Developing sustainable food value chains-Guiding principles. FAO. (2014b). Food Loss Assessments : Causes and Solutions. Case Studies in Small-Scale Agriculture and Fisheries Subsectors, 92. Retrieved from http://www.fao.org/fileadmin/user_upload/save- food/PDF/Kenya_Food_Loss_Studies.pdf FAO. (2015). Regional Overview of Food Insecurity Africa: African food securiy prospects brighter than ever. Accra. FAO. (2017). The future of food and agriculture: Trends and challenges. Retrieved from http://www.fao.org/publications/card/en/c/d24d2507-41d9-4ec2-a3f8-88a489bfe1ad/ 160 University of Ghana http://ugspace.ug.edu.gh FAO. (2018). OECD-FAO Agricultural Outlook 2018-2027. Agricultural Outlook 2016- 2025, 181(November 1947). https://doi.org/10.1787/agr_outlook-2018-en FAO, IFAD, & WFP. (2015). The State of Food Insecurity in the World: Meeting the 2015 international hunger targets: taking stock of uneven progress. FAO, IFAD and WFP. https://doi.org/I4646E/1/05.15 Farmer Helpline. (2018). Planting for Export and Rural Development (PERD). Retrieved May 20, 2019, from http://agricinghana.com/2018/10/05/overview-planting-for-export- and-rural-development-perd-programme/ Fawad, M., Nakano, Y., & Kurosaki, T. (2019). Impact of contract farming on land productivity and income of maize and potato growers in Pakistan. Food Policy, 85(April), 28–39. https://doi.org/10.1016/j.foodpol.2019.04.004 Fearon, J., & Adraki, P. K. (2015). Fertilizer Subsidy Programme in Ghana : Evidence of Performance after Six Years of Implementation, 5(21), 100–107. Feleke, S., Haile, M. G., Abdoulaye, T., Olanrewaju, A., Wossen, T., Manyong, V., & Alene, A. (2017). Impacts of extension access and cooperative membership on technology adoption and household welfare. Journal of Rural Studies, 54, 223–233. https://doi.org/10.1016/j.jrurstud.2017.06.022 Fischer, S., & Wollni, M. (2018). The role of farmers’ trust, risk and time preferences for contract choices: Experimental evidence from the Ghanaian pineapple sector. Food Policy, 81(November), 67–81. https://doi.org/10.1016/j.foodpol.2018.10.005 Food and Agriculture Organisation of the United Nations (FAO). (2018). World Food and Agriculture - Statistical Pocketbook 2018. Rome. 254 pp. Licence: CC BY-NC-SA 3.0 IGO. Retrieved from http://www.wipo.int/amc/en/mediation/rules French, M. T., & Popovici, I. (2011). That instrument is lousy! In search of agreement when using instrumental variables estimation in substance use research. Health Economics, 20(2), 127–146. https://doi.org/10.1002/hec.1572 Gaffney, J., Challender, M., Califf, K., & Harden, K. (2019). Building bridges between agribusiness innovation and smallholder farmers: A review. Global Food Security, 20(September 2018), 60–65. https://doi.org/10.1016/j.gfs.2018.12.008 Gandhi, V. P. (2014). Growth and Transformation of the Agribusiness Sector. Indian Journal of Agricultural Economics, 69(1), 44–74. Garner, E., & de la O Campos, A. P. (2014). Identifying the “family farm.” ESA Working Paper, (14), 1–30. Gebrezgabher, S. A., Meuwissen, M. P. M., Kruseman, G., Lakner, D., & Oude Lansink, A. G. J. M. (2015). Factors influencing adoption of manure separation technology in the Netherlands. Journal of Environmental Management, 150, 1–8. https://doi.org/10.1016/j.jenvman.2014.10.029 GGBL. (2017). Annual report 2016. https://doi.org/10.1080/00344893.2017.1291131 161 University of Ghana http://ugspace.ug.edu.gh Ghimire, R., Huang, W. C., & Shrestha, R. B. (2015). Factors Affecting Adoption of Improved Rice Varieties among Rural Farm Households in Central Nepal. Rice Science, 22(1), 35–43. https://doi.org/10.1016/j.rsci.2015.05.006 Gitonga, Z. M., De Groote, H., Kassie, M., & Tefera, T. (2013). Impact of metal silos on households’ maize storage, storage losses and food security: An application of a propensity score matching. Food Policy, 43, 44–55. https://doi.org/10.1016/j.foodpol.2013.08.005 Glover, D., & Kusterer, K. (1990). Small farmers, big business: contract farming and rural development. Small Farmers, Big Business: Contract Farming and Rural Development. Goldsmith, P. D., Gaudencio, A. M., & Moura, A. D. de. (2015). The economics of post- harvest loss: a case study of the new large soybean??- maize producers in tropical Brazil. Food Security, 7(4), 875–888. https://doi.org/10.1007/s12571-015-0483-4 Gover, D. (1990). Contract farming and outgrower schemes in East and Southern Africa. JAE Journal of Agricultural Economics, 41(3), 3015–3315. https://doi.org/https://doi.org/10.1111/j.1477-9552.1990.tb00648.x GSS. (2014a). 2010 population & housing census; district analyticla report Jirapa district. GSS. (2014b). 2010 population and housing census; district analytical report Guru Tempane District. Accra. GSS. (2014c). Ghana Living Standards Survey Round 6 (GLSS 6) (Vol. 12). https://doi.org/10.1007/s13398-014-0173-7.2 GSS. (2014d). Ghana Statistical Living Standars Survey Round 6. Government of Ghana (Vol. 53). Accra. https://doi.org/10.1017/CBO9781107415324.004 GSS. (2018). Ghana Living Standard Survey Round 7 (GSS). Retrieved from http://www.statsghana.gov.gh/docfiles/publications/GLSS7/Poverty Profile Report_2005 - 2017.pdf GSS. (2019). Ghana Living Standards Survey round 7 (GLSS7). Retrieved from https://statsghana.gov.gh/gsspublications.php?category=MTAwMjg3Mzk3NC4zMDc= /webstats/1opr93rn57 Hahn, M. B., Riederer, A. M., & Foster, S. O. (2009). The Livelihood Vulnerability Index: A pragmatic approach to assessing risks from climate variability and change-A case study in Mozambique. Global Environmental Change, 19(1), 74–88. https://doi.org/10.1016/j.gloenvcha.2008.11.002 Hariprasanna K, Agte V, Elangovan M, Gite S, & Kishore A. (2015). Anti-nutritional factors and antioxidant capacity in selected genotypes of sorghum [sorghum bicolor l. (moench)]. International Journal of Agriculture Sciences, 7(8), 975–3710. Hausman, J. A. (1978). Specification tests in econometrics. Econometrica, 46(1978), 1251- 1272. 162 University of Ghana http://ugspace.ug.edu.gh Heckman, J. (1979). Sample selection bias as a specification error. Econometrica, 47(1979), 153–161. Hell, K., & Mutegi, C. (2011). Aflatoxin control and prevention strategies in key crops of Sub-Saharan Africa. African Journal of Microbiology Research, 5(5), 459–466. https://doi.org/10.5897/AJMR10.009 Hengsdijk, H. (2017). Post-harvest management and post-harvest losses of cereals in Ethiopia, 945–958. https://doi.org/10.1007/s12571-017-0714-y HLPE. (2014). Food losses and waste in the context of sustainable food systems. Rome. HLPE. (2019). Agroecological and other innovative approaches for sustainable agriculture and food systems that enhance food security and nutrition. A report by the High Level Panel of Experts on Food Security and Nutrition of the Committee on World Food Security. The High Level Panel of Experts on Food Security and Nutrition. Rome. Hobden, T., & Sands, T. (2017). Maximising the Impact of Outgrower Schemes : Opportunities , Challenges and Lessons from the AECF. Huang, T., Bingchan, L., Shen, D., Cao, J., & Mao, B. (2017). ScienceDirect Analysis of the grain loss harvest based on logistic regression Analysis of the grain loss in in harvest based on logistic regression. Procedia Computer Science, 122, 698–705. https://doi.org/10.1016/j.procs.2017.11.426 Huang, Z., Xu, Y., Zeng, D., Wang, C., & Wang, J. (2018). One size fits all ? Contract farming among broiler producers in China. Journal of Integrative Agriculture, 17(2), 473–482. https://doi.org/10.1016/S2095-3119(17)61752-0 Intergovernmental Panel on Climate Change (IPCC). (2013). Summary for Policymakers. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, 33. https://doi.org/10.1017/CBO9781107415324 IPCC. (2018). Summary for Policymakers. Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change,. https://doi.org/10.1017/CBO9781107415324 ISSER. (2012). an Impact Evaluation of the Mida Fbo Training. MiDA, (Institute of Statistical, Social and Economic Research). Jamshidi, O., Asadi, A., Kalantari, K., Azadi, H., & Scheffran, J. (2019). Vulnerability to climate change of smallholder farmers in the Hamadan province, Iran. Climate Risk Management, 23(June 2018), 146–159. https://doi.org/10.1016/j.crm.2018.06.002 Jayne, T. S. (2012). Managing food price instability in East and Southern Africa. Global Food Security, 1(2), 143–149. https://doi.org/10.1016/j.gfs.2012.10.002 Jayne, T. S., Chamberlin, J., Traub, L., Sitko, N., Muyanga, M., Yeboah, F. K., … Kachule, R. (2016). Africa’s changing farm size distribution patterns: the rise of medium-scale 163 University of Ghana http://ugspace.ug.edu.gh farms. Agricultural Economics (United Kingdom), 47, 197–214. https://doi.org/10.1111/agec.12308 Johanna, S., Gobien, S., & Kirk, M. (2019). Land Use Policy Socio-economic well-being , contract farming and property rights : Evidence from Ghana. Land Use Policy, 81(September 2016), 878–888. https://doi.org/10.1016/j.landusepol.2017.04.023 John Ng’ombe. (2013). Impact of Conservation Farming on Smallholder Farm. University of Zambia Lusaka. Jolly, P., Jiang, Y., Ellis, W., Awuah, R., Nnedu, O., Phillips, T., … Jolly, C. (2006). Determinants of aflatoxin levels in Ghanaians: Sociodemographic factors, knowledge of aflatoxin and food handling and consumption practices. International Journal of Hygiene and Environmental Health, 209(4), 345–358. https://doi.org/10.1016/j.ijheh.2006.02.002 Jones, M. S., Alexander, C. E., & Smith, B. (2018). Economic consequences of post-harvest insect damage in Rwandan common bean markets. Crop Protection, 104(November 2017), 92–100. https://doi.org/10.1016/j.cropro.2017.10.015 Kemp-Benedict, E., Bharwani, S., De La Rosa, E., Krittasudthacheewa, C., & Matin, N. (2009). Assessing Water-related Poverty Using the Sustainable Livelihoods Framework. Environment, 1–32. Khalil, A. C., Conforti, P., Ergin, I., & Gennari, P. (2017). Globalization and the smallholders: A review of issues, approaches, and implications. Markets and Structural Studies Division. Defining Small Scale Food Producers to Monitor Target 2.3. of the 2030 Agenda for Sustainable Development (Vol. 17). Retrieved from www.fao.org/3/a- i6858e.pdf Khandker, S. R., Koolwal, G. B., & Samad, H. A. (2010). Handbook on Impact. Kiaya, V. (2014). Post-Harvest losses and Strategies to Reduce Them. ACF International, (January), 1–25. Kiaya, V. (2018). Post-Harvest losses and Strategies to Reduce Them. ACF International, (January), 1–25. Kitinoja, L., & Kader, A. A. (2015). Measuring postharvest losses of fresh fruits and vegetables in developing countries. PEF White Paper 15-02, (September), 26. Retrieved from http://www.postharvest.org/PEF_White_Paper_15- 02_PHFVmeasurement.pdf Knutsson, P. (2006). The Sustainable Livelihoods Approach. Aquaculture Economics Management. https://doi.org/10.1080/13657300903156092 Koira Alemayehu Konde. (2014). Agribusiness in Sub-Saharan Africa: Pathways for developing innovative programs for youth and the rural poor, (August), 1–23. Retrieved from http://www.mastercardfdn.org/wp- content/uploads/2015/08/Agribusiness-in-Sub-Saharan-Africa-2014.pdf Kuivanen, K. S., Alvarez, S., Michalscheck, M., Adjei-Nsiah, S., Descheemaeker, K., 164 University of Ghana http://ugspace.ug.edu.gh Mellon-Bedi, S., & Groot, J. C. J. (2016). Characterising the diversity of smallholder farming systems and their constraints and opportunities for innovation: A case study from the Northern Region, Ghana. NJAS - Wageningen Journal of Life Sciences, 78, 153–166. https://doi.org/10.1016/j.njas.2016.04.003 Lahouar, A., Marin, S., Crespo-Sempere, A., Saïd, S., & Sanchis, V. (2016). Effects of temperature, water activity and incubation time on fungal growth and aflatoxin B1 production by toxinogenic Aspergillus flavus isolates on sorghum seeds. Revista Argentina de Microbiologia, 48(1), 78–85. https://doi.org/10.1016/j.ram.2015.10.001 Lambrecht, B. (2016). “As a Husband I Will Love, Lead, and Provide.” Gendered Access to Land in Ghana. World Development, 88, 188–200. https://doi.org/10.1016/j.worlddev.2016.07.018 Lipton, M., & Saghai, Y. (2017). Food security, farmland access ethics, and land reform. Global Food Security, 12(October 2015), 59–66. https://doi.org/10.1016/j.gfs.2016.03.004 Little, P., & Watts, Madison, M. . (1999). Contract Farming and Agrarian Transformation in Sub-Saharan Africa, 6(1994), 141–144. Retrieved from http://pdf.usaid.gov/pdf_docs/PNACL391.pdf Loggoh, B. (2013). The Contribution of Outgrower Scheme to Farmers Livelihood-A Case of Oil Palm Farmers in the Kwaebibirem District of Ghana. UNIVERSITY OF GHANA. Lokshin, M., & Sajaia, Z. (2018). Maximum Likelihood Estimation of Endogenous Switching Regression Models. The Stata Journal: Promoting Communications on Statistics and Stata, 4(3), 282–289. https://doi.org/10.1177/1536867x0400400306 Maddala, G. (1983). Limited-Dependent and Qualitative Variables in Economics. Maertens, M., & Velde, K. (2017). Contract-farming in Staple Food Chains : The Case of Rice in Benin. World Development, 95, 73–87. https://doi.org/10.1016/j.worlddev.2017.02.011 Makate, C., Makate, M., Mango, N., & Siziba, S. (2019). Increasing resilience of smallholder farmers to climate change through multiple adoption of proven climate- smart agriculture innovations. Lessons from Southern Africa. Journal of Environmental Management, 231(August 2018), 858–868. https://doi.org/10.1016/j.jenvman.2018.10.069 Makuvaro, V., Walker, S., Masere, T. P., & Dimes, J. (2018). Smallholder farmer perceived effects of climate change on agricultural productivity and adaptation strategies. Journal of Arid Environments, 152(February), 75–82. https://doi.org/10.1016/j.jaridenv.2018.01.016 Manda, S., Tallontire, A., & Dougill, A. J. (2018). Outgrower schemes, livelihoods and response pathways on the Zambian ‘sugarbelt.’ Geoforum, 97(February), 119–130. https://doi.org/10.1016/j.geoforum.2018.10.021 Mariano, M. J., Villano, R., & Fleming, E. (2012). Factors influencing farmers’ adoption of 165 University of Ghana http://ugspace.ug.edu.gh modern rice technologies and good management practices in the Philippines. Agricultural Systems, 110, 41–53. https://doi.org/10.1016/j.agsy.2012.03.010 Mauceri, M., Alwang, J., Norton, G., & Barrera, V. (2005). Adoption of Integrated Pest Management Technologies: A Case Study of Potato Farmers in Carchi, Ecuadorvided that this copyright notice appears on all such copies. In American Agricultural Economics Association Annual Meeting, Providence, Rhode Island. Providence, Rhode Island. Mccuistion, K. C., Selle, P. H., Liu, S. Y., & Goodband, R. D. (2019). Sorghum as a Feed Grain for Animal Production. Sorghum and Millets. AACCI. https://doi.org/10.1016/B978-0-12-811527-5.00012-5 Mignouna, D. B., Manyong, V. M., Mutabazi, K. D. S., & Senkondo, E. M. (2011). Determinants of adopting imazapyr-resistant maize for Striga control in Western Kenya : A double-hurdle approach, 3(11), 572–580. Ministry of Finance. (2018). The Budget Statement and Economic Policy of the Government of Ghana for the 2019 Financial Year. Accra: Republic of Ghana. Retrieved from https://www.mofep.gov.gh/budget-statements/2019 Ministry of Finance. (2019). The Budget Statement and Economic Policy for 2020 Financial Year. Accra. Minot, N. (2011). Contract Farming in sub-Saharan Africa : Opportunities and Challenges, (April). Minot, N. (2015). Contract Farming in sub-Saharan Africa : Opportunities and Challenges, (January). Minot, N., & Ronchi, L. (2014). Risks and Benefits of Partnership between Farmers and Firms. Minot, & Nicholas. (1986). Contract Farming and Its Effect on Small Farmers in Less Developed Countries. Food Security International Development Working Papers. Retrieved from https://ideas.repec.org/p/ags/midiwp/54740.html Miyata, S. (2009). Impact of Contract Farming on Income : Linking Small Farmers , Packers , and Supermarkets in China. World Development, 37(11), 1781–1790. https://doi.org/10.1016/j.worlddev.2008.08.025 Mmm, M. (2008). wrong document. Accra. Mmmmmmmmmmmmm. (2018). Ghana Agricultural Investment Plan (GhIP) 2018-2021. Accra. MoFA-SRID. (2016). Facts and Figures. Government of Ghana (Vol. 11). https://doi.org/10.1002/2013JB010448.Received MoFA. (2007a). Food and Agriculture Sector Development Policy. Accra, Ghana: Government of Ghana. MoFA. (2007b). Food and Agriculture Sector Development Policy (FASDEP II). Accra. 166 University of Ghana http://ugspace.ug.edu.gh MoFA. (2015). Medium term agricultural sector investment plan ( metasip ) ii , 2014 – 2017. Minisry of Food and Agriculture. MoFA. (2016). Agricultural sector progress report, 2016 Ministry of Food and Agriculture, Monitoring and Evaluation Directorate. MoFA. (2017a). Agricultural Sector Progress Report 2017. Accra. MoFA. (2017b). Planting for food and jobs: Strategic plan for implementation (2017-2020). Accra. MoFA. (2018). Investing for Food and Jobs: an Agenda for Transforming Ghana’s Agriculture (2018-2021). Accra. Montgomery, S., Guppy, C., Martin, R., Wright, G., Flavel, R., Phan, S., … Tighe, M. (2017). Productivity and profitability of upland crop rotations in Northwest Cambodia. Field Crops Research, 203, 150–162. https://doi.org/10.1016/j.fcr.2016.12.010 Morris, K. J., Kamarulzaman, N. H., & Morris, K. I. (2019). Small-Scale Postharvest Practices Among Plantain Farmers and Traders: a Potential for Reducing Losses in Rivers State, Nigeria. Scientific African, 4, 86. https://doi.org/10.1016/j.sciaf.2019.e00086 Morton. (2013). Ghana policy journal. Ghana Policy Journal, 5. Murphy, S. (2012). Changing Perspectives: Small-scale farmers, markets and globalization. International Institute for Environment and Development (IIED) and Humanist Institute for Development Cooperation (HIVOS). Mustapha, S., Tanko, M., & Alidu, A. F. (2016). Determinants of Household Food Insecurity in Northern Ghana: An Ordered Probit Approach. Journal of Economics and Sustainable Development, 7(16), 2222–1700. Retrieved from www.iiste.org Mwangi, M., & Kariuki, S. (2015). Factors Determining Adoption of New Agricultural Technology by Smallholder Farmers in Developing Countries. Issn, 6(5), 2222–1700. Retrieved from www.iiste.org Narayanan, S. (2002). Globalization and the smallholders. ResearchGate. Washington DC: t http://www.cgiar.org/ifpri/divs/mssd/dp.htm. Narayanan, S. (2014). Profits from participation in high value agriculture : Evidence of heterogeneous benefits in contract farming schemes in Southern India. JOURNAL OF FOOD POLICY, 44, 142–157. https://doi.org/10.1016/j.foodpol.2013.10.010 Neme, K., & Mohammed, A. (2017). Mycotoxin occurrence in grains and the role of postharvest management as a mitigation strategies. A review. Food Control, 78, 412– 425. https://doi.org/10.1016/j.foodcont.2017.03.012 New Partnership for Africa Development (NEPAD). (2013). Africulture in Africa, Transformation and Outlook. Johannesburg. New Patriotic Party (NPP). (2016). Change an Agenda for Jobs; Creating Prosperity and Equal Opportunity for All. New Patriotic Party Manaifesto for Election 2016. Accra. 167 University of Ghana http://ugspace.ug.edu.gh Nkegbe, P. K., Abu, B. M., & Issahaku, H. (2017). Food security in the Savannah Accelerated Development Authority Zone of Ghana : an ordered probit with household hunger scale approach. Agriculture & Food Security, 1–11. https://doi.org/10.1186/s40066-017-0111-y Odunze, D., Niekerk, V., & Ndlovu, S. (2015). Assessment of factors that impact on the viability of contract farming: a case study of maize and soya beans in Mashonaland west and Central Provinces in Zimbabwe. S.Afr. Tydskr. Landbouvoorl./S. Afr. J. Agric. Ext., 43(2), 359–390. Retrieved from http://kiss.kstudy.com/search/detail_page.asp?key=2234223 Ogunsakin, A. O., Vanajakshi, V., Anu-Appaiah, K. A., Vijayendra, S. V. N., Walde, S. G., Banwo, K., … Prabhasankar, P. (2017). Evaluation of functionally important lactic acid bacteria and yeasts from Nigerian sorghum as starter cultures for gluten-free sourdough preparation. LWT - Food Science and Technology, 82, 326–334. https://doi.org/10.1016/j.lwt.2017.04.048 Okunlola, J. ., Oludare, A. ., & Akinwalere, B. . (2011). Adoption of new technologies by fish farmers in Akure , Ondo state , Nigeria, 7(6), 1539–1548. Opoku, S. (2012). Logistic Analysis of f Factors Motivating Smallholder Farmers To Engage In Contract Farming Arrangements With Processing Firms In Ghana. Paper, 3208, 58–74. Otsuka, K., Nakano, Y., & Takahashi, K. (2016). Contract Farming in Developed and Developing Countries Annual Review of Resource Economics. Eonomics, Vol. 8(October 2016), 353–376. Oya, C. (2012). Contract Farming in Sub‐ Saharan Africa: A Survey of Approaches, Debates and Issues. Journal of Agricultural Science and Technology AAgrarian Change, 12(1)(December 2011), 1–33. https://doi.org/10.1111/j.1471- 0366.2011.00337.x Paglietti, L., & Sabrie Roble. Outgrower schemes: advantages of different business models for sustainable crop intensification (2012). Pandey, R., Kumar, S., Alatalo, J. M., Archie, K. M., & Gupta, A. K. (2017). Sustainable livelihood framework-based indicators for assessing climate change vulnerability and adaptation for Himalayan communities. Ecological Indicators, 79, 338–346. https://doi.org/10.1016/j.ecolind.2017.03.047 Porter, G., & Phillips-Howard, K. (1997). Comparing contracts: An evaluation of contract farming schemes in Africa. World Development, 25(2), 227–238. https://doi.org/10.1016/S0305-750X(96)00101-5 Porter, J., Challinor, A., Iqbal, Mohsin, M., Howden, M., Travasso, I., Chhetri, N., … Ziska, L. (2014). Food Security and Food Production System. Ragasa, C., Lambrecht, I., & Kufoalor, D. (2018). Limitations of Contract Farming as a Pro- poor Strategy : The Case of Maize Outgrower Schemes in Upper West Ghana. World Development, 102, 30–56. https://doi.org/10.1016/j.worlddev.2017.09.008 168 University of Ghana http://ugspace.ug.edu.gh Ramashia, S. E., Anyasi, T. A., Gwata, E. T., Meddows-Taylorm, S., & Jideani, A. I. O. (2019). Processing, nutritional composition and health benefits of finger millet in sub- saharan Africa. Food Science and Technology, 2061(June), 253–266. https://doi.org/10.1590/fst.25017 Ratnavathi, C. V., Komala, V. V., & Lavanya, U. (2016). Sorghum Uses-Ethanol. Sorghum Biochemistry: An Industrial Perspective. https://doi.org/10.1016/B978-0-12-803157- 5.00004-6 Rembold, F., Hosges, R., Bernard, M., Knipschild, & Leo, O. (2014). The African Postharvest Losses Information System. Aphlis (Vol. 2.2). https://doi.org/10.2788/40345 Rhebergen, T., Fairhurst, T., Whitbread, A., & Giller, K. E. (2018). Yield gap analysis and entry points for improving productivity on large oil palm plantations and smallholder farms in Ghana. Agricultural Systems, 165(February), 14–25. https://doi.org/10.1016/j.agsy.2018.05.012 Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.1093/biomet/70.1.41 Safo, K. (2016). Impact of Warehouse Receipt System on Access to Markets and Income of Smallholder Maize Farmers in the Northern Region of Ghana. https://doi.org/10.1038/253004b0 Saghir, B. J., & Hoogeveen, H. (2017). Transforming Sub Saharan Africa Agriculture Research to Practice Policy Briefs, 1–19. Sakha, S. (2019). Determinants of risk aversion over time: Experimental evidence from rural Thailand. Journal of Behavioral and Experimental Economics , 80(March 2018), 184– 198. https://doi.org/10.1016/j.socec.2019.03.008 Samrat, S. (2013). National Food Buffer Stock Company (15). Accra. Sarfo-Mensah, P. (2017). Guinness Ghana local raw material sourcing initiative generates over GH¢100m in income for farmers. Retrieved from https://www.ghanabusinessnews.com/2017/09/29/guinness-ghana-local-raw-material- sourcing-initiative-generates-over-gh¢100m-in-income-for-farmers/ Schaffnit-Chatterjee, C. (2014). Agricultural value chains in Sub-Saharan Africa. Deutsche Bank Research, 1–28. Retrieved from https://www.dbresearch.com/PROD/DBR_INTERNET_EN- PROD/PROD0000000000333152/Agricultural+value+chains+in+Sub- Saharan+Africa:+From+a+development+challenge+to+a+business+opportunity.pdf Scoones, I. (1998). Sustainable Rural Livelihoods a Framework for Analysis. Analysis, 72, 1–22. https://doi.org/10.1057/palgrave.development.1110037 Shah, K. U., Dulal, H. B., Johnson, C., & Baptiste, A. (2013). Understanding livelihood vulnerability to climate change: Applying the livelihood vulnerability index in Trinidad and Tobago. Geoforum, 47, 125–137. https://doi.org/10.1016/j.geoforum.2013.04.004 169 University of Ghana http://ugspace.ug.edu.gh Sheahan, M., & Barrett, C. B. (2016). Food loss and waste in Sub-Saharan Africa : A critical review. Sheahan, M., & Barrett, C. B. (2017). Food loss and waste in Sub-Saharan Africa: A critical review. Food Policy, 70, 1–12. https://doi.org/10.1016/j.foodpol.2017.03.012 Snyder, J., Dolislager, M., Reardon, T., Haggblade, S., & Goeb, J. (2015). Africa ’ s Unfolding Diet Transformation : Implications for Agrifood System Employment, (November). https://doi.org/10.1108/JADEE-01-2015-0003 Šūmane, S., Kunda, I., Knickel, K., Strauss, A., Tisenkopfs, T., Rios, I. des I., … Ashkenazy, A. (2018). Local and farmers’ knowledge matters! How integrating informal and formal knowledge enhances sustainable and resilient agriculture. Journal of Rural Studies, 59, 232–241. https://doi.org/10.1016/j.jrurstud.2017.01.020 Takeshima, H., & Lee, L. H. (2012). Agricultural inputs subsidy and their developmental impact: conventional wisdom. Tanellari, E., Kostandini, G., & Bonabana, J. (2011). Gender Impacts on Adoption of New Technologies: Evidence from Uganda. Tang, C. S., Sodhi, M. S., & Formentini, M. (2016). An analysis of partially-guaranteed- price contracts between farmers and agri-food companies. European Journal of Operational Research, 254(3), 1063–1073. https://doi.org/10.1016/j.ejor.2016.04.038 Thompson, H. E., Berrang-Ford, L., & Ford, J. D. (2010). Climate change and food security in Sub-Saharan Africa: A systematic literature review. Sustainability, 2(8), 2719–2733. https://doi.org/10.3390/su2082719 Ton, G., Vellema, W., D’Haese, M., & Desiere, S. (2015). Systematic Review of the Effectiveness of Contract Farming for Income and Food Security of Smallholder Farmers in Low- and Middle Income Countries. Campbell Collaboration, 13(August 2016), 1–64. https://doi.org/10.4073/csr.2017.13 Ton, G., Vellema, W., D’Haese, M., Desiere, S., Weituschat, S., Brouwer, A., & Hoeven, H. van. (2016). Systematic Review of the Effectiveness of Contract Farming for Income and Food Security of Smallholder Farmers in Low- and Middle Income Countries. The International Initiative for Impact Evaluation, Inc, 6. Ton, G., Vellema, W., D’Haese, M., Desiere, S., Weituschat, S., Brouwer, A., & Hoeven, H. van. (2017). The effectiveness of contract farming for raising income of smallholder farmers in low- and middle-income countries: a systematic review. CampbellCollaboration. https://doi.org/10.4073/csr.2017.13 Torvikey, G. D., Yaro, J. A., & Teye, J. K. (2016). Farm to Factory Gendered Employment: The Case of Blue Skies Outgrower Scheme in Ghana. Agrarian South: Journal of Political Economy, 5(1), 77–97. https://doi.org/10.1177/2277976016669188 Travis, W. R. (2014). Weather and climate extremes: Pacemakers of adaptation? Weather and Climate Extremes, 5(1), 29–39. https://doi.org/10.1016/j.wace.2014.08.001 Tshilidzi Madzivhandila, Sibanda, S., & Gwelo, F. A. (2016). Achieving Food Security and 170 University of Ghana http://ugspace.ug.edu.gh Nutrition. Africa Agriculture Status Report 2016, 19. Tsinigo, E., & Behrman, J. R. (2017). Technological priorities in rice production among smallholder farmers in Ghana. NJAS - Wageningen Journal of Life Sciences, 83(July), 47–56. https://doi.org/10.1016/j.njas.2017.07.004 Uaiene, R., Arndt, C., & Masters, W. (2009). Determinants of agricultural technology adoption in Mozambique. National Directorate of Studies and Policy Analysis, (Ministry of Planning and Development, Republic of Mozambique), 1–31. Retrieved from http://196.46.4.208/gest/documents/67E_AgTechAdoptionMoz.pdf Udimal, T., Jincai, Z., Mensah, S., & Ayamba, E. (2017). Factors Influencing the Agricultural Technology Adoption : The Case of Improved Rice Varieties ( Nerica ) in the Northern Region , 8(8), 137–148. UNEP, & UNDP. (2013). National Climate Change Adaptation Strategy. Global Environmental Change (Vol. 5). https://doi.org/10.1007/978-3-319-31499-0 Upadhyaya, H. D., Narsimha Reddy, K., Vetriventhan, M., Irshad Ahmed, M., Murali Krishna, G., Thimma Reddy, M., & Singh, S. K. (2017). Sorghum germplasm from West and Central Africa maintained in the ICRISAT genebank: Status, gaps, and diversity. Crop Journal, 5(6), 518–532. https://doi.org/10.1016/j.cj.2017.07.002 Vicol, M. (2017). Geoforum Is contract farming an inclusive alternative to land grabbing ? The case of potato contract farming in Maharashtra , India. Geoforum, 85(August), 157–166. https://doi.org/10.1016/j.geoforum.2017.07.012 Villano, R., Asante, B. O., & Bravo-Ureta, B. (2019). Farming systems and productivity gaps: Opportunities for improving smallholder performance in the Forest-Savannah transition zone of Ghana. Land Use Policy, 82(May 2018), 220–227. https://doi.org/10.1016/j.landusepol.2018.12.010 Wang, H. H., Wang, Y., & Delgado, M. S. (2014). The transition to modern agriculture: Contract farming in developing economies. American Journal of Agricultural Economics, 96(5), 1257–1271. https://doi.org/10.1093/ajae/aau036 Watanabe, K., Bánkuti, S. S., & Lourenzani, A. E. (2017). “Pingado Dilemma”: Is formal contract sweet enough? Journal of Rural Studies, 54, 126–137. https://doi.org/10.1016/j.jrurstud.2017.06.001 Wendimu, M. A., Henningsen, A., & Gibbon, P. (2016). Sugarcane Outgrowers in Ethiopia: “Forced” to Remain Poor? World Development, 83(October 2014), 84–97. https://doi.org/10.1016/j.worlddev.2016.03.002 WFP, MoFA, G. (2012). Comprehensive Food Security & Vulnerability Analysis Focus on Northern Ghana The north south divide. Wineman, A., & Liverpool-Tasie, L. S. O. (2017). Land Markets and Land Access Among Female-Headed Households in Northwestern Tanzania. World Development, 100, 108– 122. https://doi.org/10.1016/j.worlddev.2017.07.027 Wold Bank. (2011). Missing Food; the Case of Postharvest Grain Losses in Sub-Saharan 171 University of Ghana http://ugspace.ug.edu.gh Africa, (60371). Wooldridge Jeffrey. (2002). Econometric Analysis of Cross Section and Panel Data. In Econometric Analysis of Cross Section and Panel Data (pp. 1–740). World Bank. (2003). Reaching the rural poor; Renewed Strategy for Rural Development. Development Policy Review (Vol. 19). https://doi.org/10.1111/1467-7679.00153 World Bank. (2012). Project Apraisal; Ghana Commercial Agricultural Programme. Wünsch, K., Gruber, S., & Claupein, W. (2012). Profitability analysis of cropping systems for biogas production on marginal sites in southwestern Germany. Renewable Energy, 45, 213–220. https://doi.org/10.1016/j.renene.2012.03.010 Yamane, T. (1967). Statistics, An Introductory Analysis (Second). New York: Harper and Row. Yaro, J. A. (2013). Building Resilience and reducing Vulnerability to Climate Change : Implications for Food Security in Ghana, (August 2013). 172 University of Ghana http://ugspace.ug.edu.gh APPENDICES Appendix 1: Questionnaire SURVEY QUESTIONNAIRE My name is Charles Nyaaba, a PhD candidate at the University of Ghana, Legon and I am conducting a research on the topic “The effects of Outgrower Scheme on the Livelihoods of Smallholder Sorghum Farmers in Northern Ghana”. You have been identified as a knowledgeable informant on this topic and hence I will be very grateful if could spend some time to assist me complete this questionnaire. This study is purely an academic exercise as such all information provided will be treated strictly as confidential and for academic purposes only. Responses will be used anonymously and cannot be traced to the persons who provide them. Thank you in advance for your time and contribution to this research. Questionnaire Number Date of interview (mm/dd/yyyy): Region: 1. Upper East 2. Upper West District: Community Name Language Used for Interview 1. English 2. Ga-Adangbe 3. Nzema 4. Akan 5. Dagbani 6. Ewe 7. Hausa 8. Dagaare 9. Other __________ (Specify) Name of Enumerator: Name of Respondent Contact Number of Respondent 173 University of Ghana http://ugspace.ug.edu.gh MODULE A: DEMOGRAPHY Ask these questions for each of the members that have stayed with this Household for a period of at least six month over the last 12 months. (Household members are defined as all those who normally live and eat their meals together here. Include household members temporarily studying elsewhere or traveling, but who spent AT LEAST SIX continuous months living and share the same cooking arrangements). Reference Period: The Past 12 months Common What is Name's Answer Member Name What is the How Relationship Is name the What is religious [Name]’s If A05<3 yrs Is If A11=2, Can Can What is for ID & sex of old is to current head primary [Name]’s denomination Marital What is the [NAME] Why is [NAME] [NAME] the MAIN household Name [NAME]? Name See code respondent Nationality? Status highest level currently [NAME] read? write? occupation members (Start 1=male below 1=Yes (ask only of education attending not in 1=Yes 1=Yes of older than with the 2=female 2=No if A04> [Name]’s has school? school? 2=No 2=No [NAME]? 12 years. household 12) completed? Did [Name] head) see codes 1 = Did receive cash See codes below Yes>> [Name] or payment below A13 receive in kind from cash from salaried 2= No informal employment, /business wage activity? activities, Include remittances dividends or pensions? (in the (in the past past 12 12 months) months) 1=Yes 1=Yes 2=No 2=No A06 A07 A08 A10 A14 A15 A16 A17 A02 A01 A03 A04 A05 A09 A11 A12 A13 1 2 3 174 University of Ghana http://ugspace.ug.edu.gh 4 5 6 7 8 9 100 175 University of Ghana http://ugspace.ug.edu.gh Religious Denomination Reason for absence Marital Status from school (A12) Relation to head (A05 ) Nationality Education levels (A08) Main Occupation (A15) (A09) (A07) 1= head 10=other 1 = single 1.Ghanaian N one...................................00 1.No religion 1 0=Too young 1= Salary earner 14=Miller relative (e.g., teacher, 2= spouse 2 = 2. Nigerian K indergarten.......................01 2.Catholic 2 1=Cannot afford policeman, etc.) 15=Trading 11=non- monogamous expenses farm produce 3= own child relative married 3. Primary…………………...02 3.Protestant 3 2= Casual wage Burkinabe 2=Working earner 16=Trading 4= step child 12=brother 3 = JSS/JHS/Middle….……….03 4.Pentecostal/Charismatic food items /sister-in- polygamous 4. Nigerien 4=Pregnancy 3= Farm 5= parent SSS/SHS/O’level/A’level…04 law married labourer 17=Trading 5. Liberian 5.Other christian 5=Sickness/disability livestock 6= brother Voc/Tech/Commercial……05 13=parent- 4 = divorced 4=Transportation /sister 6. Malian 6.Islam 6=Refused to in-law Teacher business 18=Trading 5 = widowed continue firewood/timber 7= nephew 7. Training/Agric/Nursing 7.Traditional 14=Worker 5= Bicycle /niece 6 = separated Togolese C ert…….............................06 7=Completed repair/mechanics 19=Trading 8.Other 15=Adopted schooling non-food goods 8= 7=Cohabitation 8.Other Degree…………………….07 child 6= Brewing son/daughter- Ecowas 8=Too old to be in business 20=Farming in-law Post-graduate……………08 16=Grand school 9.Other parent Other……………………...09 7= Brick making 21=Tailor 9= African 9=Other grandchild (specify)___ Don't know……………….99 8=Butcher 22=Schooling 10.Other Continent 9=Carpentry 24=Food vendors 10=Charcoal burning 23=Other (specify) 11=Clothes business (trading) 12=Construction 13=General- kiosk owner 176 University of Ghana http://ugspace.ug.edu.gh SECTION A: SOCIOECONOMIC CHARACTERISTICS 1. How many children in your household are orphans ………….. 2. How long have you been farming……………. Number of years 3. What is your length of stay in this community …………… Number of years 4. How many people in your family go to a different community to work? .......................... 5. Do you belong to any Farmer Based Organisation (FBO)? a. Yes b. No 6. What is the name of the FBO……………………………… SECTION B: FARM CHARACTERISTICS 7. What is the total area of your land (include both owned and rented areas that the farmer farms in and around the village)? ....................Acre(s) 8. How many sorghum farms did you cultivate? …………..Number of plots 9. What is the size of your sorghum farm cultivated during the last production season? ……….acre(s) 10. What is the land tenure situation that best describe your farm? 1. Producer owns land 2. Producer pays money to rent all the land 3. The land is under communal ownership 4. Producer farms land owned by other without paying anything 5. Producer exchanges produce for rights to all land farmed (sharecropping) 11. What variety of sorghum is cultivated on your farm? 1. local (specify name of variety)…………………………………………………………. 2. Naga white 2. Framida 3. Kadaga 4. Kapaala 5. Dorado 6. Other improved variety (specify)…………………………………………………………. 12. What is the distance from your homestead to the farm? ...............Kilometers SECTION C: SORGHUM OUTGROWER SCHEME 13. Do you participate in sorghum outgrower scheme? 1. Yes 2. No >> 14. Do you have any contract with an aggregator to supply them with sorghum? 1. Yes 2. No >> 15. If yes, which aggregator are you in contract with ? a. Faranaya b. PFAG c. Agric Access d. other specify 16. What kind of agreement is it? a. Oral b. Documented 17. How many years have you been enrolled under the sorghum outgrower scheme…………….(Years) 18. Did you receive some support from the Outgrower scheme in the last production season? 1. Yes 2. No >> 19. What kind of technical assistance do you receive from the Outgrower scheme? Multiple response 1. Fertiliser 2. Pesticide 3. Training 4.credit (financial support) 5. Tools/equipment 6. Seeds 7. Other specify………………………………………………….. 20. How was the price determined? 1. Arrangement between the aggregator and farmer 2. Farmer determines the price 3. Prevailing market price 4. Aggregator determines the price 5. Other specify……………… 177 University of Ghana http://ugspace.ug.edu.gh SECTION D: TRAINING SERVICES 21. Did you receive extension visit in the last production year? 1. Yes 2. No >> 22. How many times did you receive extension visit in the last production………………? Number of times 23. What kind service did you receive from the extension agent? Circle all that apply 1. Improving sorghum farming 2. Improving record keeping 3. Marketing support Information practices or processing practices 4. Health and social issues 5. Environmental issues 6. Managing the farm's business or finances 7. Crop losses 8. Adult literacy 9. Other (specify) 24. Did you receive any other training other extension service? 1. Yes 2. No >> 25. Did you receive any other training other than extension service? 1. Yes 2. No (Skip to Q) 26. Who provided the training service? Circle all that apply 1. NGO 2. Aggregator/Buyer 3. Research Institute/University 4. Other (specify 27. What kind of training did you receive? Circle all that apply 1. Improving sorghum farming 2. Improving record 3. Marketing support Information practices or processing keeping practices 4. Health and social issues 5. Environmental issues 6. Managing the farm's business or finances 7. Adult literacy 8 Other (specify) 28. How many training sessions were attended? …………….... (Number) 29. How many minutes were spent travelling to attend this ……..………. training? Minutes 30. What was the cost of this travel? ……………................... . Amount in Ghana cedis 178 University of Ghana http://ugspace.ug.edu.gh SECTION E: ACCESS TO CREDIT 31. Did you try to get credit or loan for sorghum production in the last production season? 1. Yes 2. No >> Section F LN Sources from Did you try to Amount Amount Interest Have you What is which loans were get requested from received paid per paid (fully) the requested credit/loan this source (If from this anum back the loan outstandin from this several loans source (Percent) received? 1. g loan to source were requested Yes>>next be paid? 1.Yes from one source, source 2. Amount 2. No>> next show total of all No in GHS source requests) LN 42A 42B 42C 42D 42E 42F 42G 1 Friends 2 Relatives 3 Religious groups 4 Banks 5 Government lending Institutions 6 Non-Bank Financial Institution 7 Buyers/Aggregator directly 8 NGOs 9 Cooperatives 10 Farmer Based Organisation 60. Did you lend any money to relatives/friends in the past 12 months? [ ] 1=Yes 0 = No SECTION F: INPUTS USED IN FARM PRODUCTION 32. In the last production season, did you apply fertiliser or pesticide on your sorghum farm? 1. Yes 2. No >> Section G Input Did you use How did What Unit If some or What was the this input in you obtain quantity of 1. all the value of the the last the the chemical Gram chemical chemical production chemical fertiliser did 2. Litre fertiliser fertiliser used in season? fertiliser? you use?. was the last 1.Yes purchased, production 2. No >> what was season (including Next input the total chemical amount fertiliiser paid? 179 University of Ghana http://ugspace.ug.edu.gh subsidised or obtained as gift) 42A 42B 42C 42D 42E 42F 42G Chemical fertiliser Organic fertiliser Pesticide Code for 1. Purchased from Agro-input dealer 2. Purchased from the market 3.Borrowed (loan) 4. Gift 5. Organisation came to community 6. Aggregator 7. Government/MOFA/Govt. Organisation 8. Other specify………………………………. SECTION G: HARVEST AND HARVESTING ACTIVITIES 43. How do you harvest your produce? Tick the appropriate responses 1. Manual with/without cutlass 2. Employ additional labour using the sickle/knife 3. Use a machine harvester 4. Others (specify)………………………………………… 44. How do you check for the maturity of your produce before harvesting? Tick the appropriate responses 1. By hand feel 2. By visual observation 3. Uses an instrument to measure 4. Uses days after planting as index 5. Others (specify) ………………………………. 45. What harvesting method did you use in the last production season? 1. Mechanical (use of tractors) 2. Manual (labourers) 46. What quantity of sorghum was harvested from your farm in the last production season? ……………. Unit 1. Kg 2. Maxi bag 3. Mini bag 47. What quantity of sorghum did you expect to harvest from your farm?..................... Unit 1. Kg 2. Maxi bag 3. Mini bag 48. While still in the field, was any sorghum damaged? 1. Yes 2. No >> 49. What was the crop lost to? 1. Rotting 2. Diseased 3. Fire 4. Flood 5. Drought 6. Birds 7. Insects 8. Rodents 9. Sheep 10. Goat 11. Cattle 12. Other 50. How much of the sorghum did you lose in total during harvesting?........................... Unit 1. Kg 2. Maxi bag 3. Mini bag 180 University of Ghana http://ugspace.ug.edu.gh SECTION H: POST HANDLING ACTIVITIES AND POST HARVEST LOSS Stacking/Heaping 51. Did you heap/stack the harvested produce on the farm before transporting? 1. Yes 2. No >> 52. What quantity of sorghum was lost during heaping……………………. Unit 1. Kg 2. Maxi bag 3. Mini bag 53. What is the value sorghum lost during heaping or stacking …………………… GHS Transporting to the from Farm gate to homestead 54. Do you readily get transport to cart your agricultural produce to the market centres? 1. Yes 2. No >> 55. How do you usually transport purchased agricultural commodities to your point of sale? 1.by own Donkey 2. by appropriate rented/hired vehicle 3. by any available passing commercial vehicle 4. by own non-motorized truck 5. by appropriate rented/hired truck 6. by any available head-load porter 7. bicycle 8. Other………………………… 56. How much did you spend on transporting sorghum from the farm to homestead………………. GHS 57. What is the distance from the farm to your homestead………………. Km 58. Did you incur any sorghum loss from the farm gate (loading into the vehicle) to the homestead? 1. Yes 2. No >> 59. What quantity of sorghum was lost when transporting the produce from the farm gate to the homestead………….. Unit 1. Kg 2. Maxi bag 3. Mini bag 60. What is the value of sorghum lost when transporting from the farm gate to your homestead?.................... GHS Drying 61. Did you dry your produce? 1. Yes 2. No >> 62. Did you experience crop losses during sun-drying? 1. Yes 2. No >> 63. What was the sorghum lost to during drying 1. Wind 2. Over-heating 3. Microbial infestation 4. Birds 5. Insects 6. Other specify …………………... 64. What quantity of sorghum was lost during drying ……………… Unit 1. Kg 2. Maxi bag 3. Mini bag Winnowing/Threshing 65. Did you thresh your sorghum after harvesting? 1. Yes 2. No >> 66. How did you thresh your sorghum? 1. Mortar and pestle 2. Beating with sticks 3. Motorized thresher 4. Other specify……………………. 67. What was the expected quantity of sorghum before threshing……………………. Unit 1. Kg 2. Maxi bag 3. Mini bag 181 University of Ghana http://ugspace.ug.edu.gh What was the quantity loss of sorghum after threshing ……………. Unit 1. Kg 2. Maxi bag 3. Mini bag 68. During winnowing what quantity of the sorghum was lost? ……………. Unit 1. Kg 2. Maxi bag 3. Mini bag Grading and Bagging 69. Do you grade your produce immediately after threshing and winnowing? 1. Yes 2. No 70. What quality indices do you use for grading your produce? Tick the appropriate responses 1. Colour 2. Size 3. Weights 4. Shape 5. Physical blemishes 6. Others (specify)……………………… 24. What quantity of sorghum was lost during grading …………….. Unit 1. Kg 2. Maxi bag 3. Mini bag 25. What quantity of the sorghum was bagged for storage………….. Unit 1. Kg 2. Maxi bag 3. Mini bag SECTION I: STORAGE PRACTICES AND CROP LOSSES AT STORAGE 1. Did you store some of the produce after harvest? 1. Yes 2. No >> 2. What facility do you use for the storage of sorghum harvested? 1. Traditional crib 2. Improved crib 3. Room 4. Other (specify)…………………………… 3. How do you store produce before marketing? 1. On a platform uncovered (tarpaulin) 2. Off-farm own shed 3. On bare floor uncovered 4. Government silo 5. On platform covered 6. Barns 7. On bare floor covered 8. Cribs 9. In containers 10. In plastic and jute sacs 11. On-Farm own shed 12. Others (Specify) 23. During the last 12 months, has any of the stored crop been lost to rot, pest, or any other cause? 1. Yes 2. No >> 30. What has the sorghum stored been lost to? 1. Rotting 2. Aflatoxin/Mould 3. Fire 4. Birds 5. Insects 6. Rodents 7. Sheep 8. Goat 9. Cattle 10. Other 31. What quantity of the sorghum was lost during storage ……………… Unit 1. Kg 2. Maxi bag 3. Mini bag 32. What is the value of crop lost at storage? ...................................... GHS 33. What is the distance from your homestead to the nearest storage facility………….. Km Transporting to the from Homestead to the Sales point/Market Do you readily get transport to cart your agricultural produce to the market centres? 1. Yes 2. No >> How did you transport sorghum to your point of sale? 1.by own Donkey 2. by appropriate rented/hired vehicle 3. by any available passing commercial vehicle 4. by own non-motorized truck 5. by appropriate rented/hired truck 6. by any available head-load porter 7. bicycle 8. other 182 University of Ghana http://ugspace.ug.edu.gh How much did you spend on transporting sorghum from the farm to sales point/market………………. GHS What is the distance from the homestead to your nearest market………………. Km Did you incur any sorghum loss from while transporting to the point of sale? 1. Yes 2. No What quantity of sorghum was lost when transporting the produce from the home to the sales point………….. Unit 1. Kg 2. Maxi bag 3. Mini bag What is the value of sorghum lost when transporting from the homestead to the point of sale?................. GHS SECTION J: CROP SALES AND CONSUMPTION 34. What quantity of the harvested sorghum did you to sell? .................. Unit 1. Kg 2. Maxi bag 3. Mini bag 35. What quantity of the harvested sorghum did your household consume …………. Unit 1. Kg 2. Maxi bag 3. Mini bag 36. What quantity of the sorghum harvested was given out as gift ……….. Unit 1. Kg 2. Maxi bag 3. Mini bag 37. How much did you receive from sales of sorghum harvested in the major season? ………… GHS 38. Where does your family get most of its food? 1. Own harvest 2. Purchased 3. Gift from friends and family 4. Other specify………………… 39. Does your household have adequate food throughout the year? 1. Yes 0. No 40. If no, how many months in a year does this household experience food shortage [………] 183 University of Ghana http://ugspace.ug.edu.gh SECTION K: INFORMATION ON LABOUR Activity 30a. Days to do 30b. Hours 30c. 30d. How 30e. How 30f. How many 30g. How many How much this job in the per day to Number of many of the much was of the workers workers would you last production do this job workers workers were paid to hired involved in this involved in this have paid to year used hired labour job were job were unpaid household exchanged labourers members labourers (household & (Nnoboa) Exchanged labour) 1. Clearing 2. Ploughing 3. Planting 4.Planting of seeds 5. Fertiliser Application 6.Harvesting 7. Threshing 8. Winnowing 184 University of Ghana http://ugspace.ug.edu.gh SECTION L: OTHER CROPS CULTIVATED Do you cultivate other crops other than sorghum? 1. Yes 2. No >> What crops did Area cultivated Harvest of this Percentag Percentage Total value of you cultivate crop per planting e crops last season (hectares) season (100kg) sold (2013)? consumed Harvested (GHS) Does this household save seeds to cultivate in the next planting season/year? 1. Yes 2. No SECTION L: AVAILABILITY AND ACCESS TO INFRASTRUCTURE 1. Does any member of your household work outside this community? [……..] 1. Yes 2. No 2. What is the main source of water for drinking and for household chores? […….] 1. Pipe borne 2. Dam 3. Rain 4. River, lake, stream 5. Wells 6. Borehole 7. Other specify …..………………………………. 3. How long (in minutes) does it take to get to the water source on foot? [………….] 4. How long (in minutes) does it take to get to the water source on bicycle? [………..] 5. Has water availability been a problem? 1. Yes 0. No 6. Estimate the number of buckets (Size 34) of water stored everyday……… 7. In the past, have you heard about any conflicts over water in this community? [..........] 1. Yes 0. No 8. Does your household have adequate food throughout the year? 1. Yes 0. No 9. How long in (minutes) does it take to get to a health facility? […………] 10. Do any of the household members have a chronic illness? [……..] 1. Yes 0. No 11. Has any member of the household been very ill in the past 6 months that they had to miss work or school? […….] 1 = Yes 0 = No 12. How many months in a year is malaria particularly common? [……………….] 13. Which months of the year is malaria particularly bad? Select all that apply 1. January 2. February 3. March 4. April 5. May 6. June 7 July 8. August 9. September 10. October 11. November 12. December 14. How many mosquito nets does the household have? [………………] SOCIAL NETWORK 15. In the past month, did relatives or friends help you and your family (e.g., Get medical care or medicines, Sell animal products or other goods produced by family, Take care of children)? 1. Yes 2. No >> 16. How many times did you or relative receive this help in the last month……… 17. In the past month, did you and or your family help relatives or friends: (same choices as above) 1. Yes 2. No >> 18. How many times did you or your family offered this help in the last month………… 185 University of Ghana http://ugspace.ug.edu.gh 19. In the past 12 months, has any member of your household gone to your community leader for help (e.g., Chief, Assemblyman, Member of parliament etc.)? [………..] 1. Yes 2. No 20. Do you play a leadership role in any social organisation in this? 1. Yes 0. No 21. In the past 12 months, have you or someone in your family gone to your community leader for help? 1. Yes 2. No SECTION M: CLIMATE CHANGE 63. Have you noticed any changes in the weather pattern in the past 30years? 1. Yes 2. No 65. What changes have been observed in the temperature pattern? 1. Consistent 2. Do not understand 3. Decreased 4. Increased 66. What changes have been observed in the rainfall pattern? 1. Consistent 2. Do not understand 3. Decreased 4. Increased 67. Has your household suffered from any drought or flood since 2000? 1. Yes 2. No >> 68. Did you receive any warning about the flood or drought before it happened? 1. Yes 2. No 69. Did any member of your household sustain any injury or lost their life as a result of the flood or drought? 1. Yes 2. No >> 70. Indicate the number that got injured [.........]/ passed away [……..] 71. Did you lose any livestock as a result of the flood or drought? 1. Yes 2. No 72. Was there any loss in the value of your livestock as a result of the flood or drought? 1. Yes 2. No 73. Did you received any training/support on how to withstand the changes on climate? 1. Yes 2. No >> 74. who provided the training (Multiple Choice)? 1. Aggregators/ buyers 2. Outgrower Scheme officers 3. MoFA 4. NGOs 5. Ghana Metrological Agency 75. What type of training/support was given (multiple choice)? 1. Using early maturing seed varieties 2. Change the type of crops planted 3. Change from crop farming to livestock farming 4. Combine crop with livestock farming 5. Information on when to plant 6. Information on when to harvest 7. Engage in off- farm activities 8. Any other 76. Has the training or the support help you become resistant to climate stress? 1. Yes 2. No HOUSEHOLD HUNGER SCALE During the WORST MONTH, how often did you or any other HH member… 77. Have to go for a whole day and night completely without food due to lack of resources to get food? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 78. Have to sleep at night hungry because there was not enough food? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 186 University of Ghana http://ugspace.ug.edu.gh 79. Have to go a whole day and night without eating anything at all because there was not enough food? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 80. Have to limit the frequency of meals because of lack of resources? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 81. Have to eat a smaller meal than you felt you needed because of lack of resources/food? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 82. Have to eat food that you did not like to eat because of a lack of resources to obtain other types of food? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 83. Have to limit the variety of foods you ate because of lack of resources? 1. Never 2. Rarely (1-2 times) 3. Sometimes (3-5 times) 4. Often (5+ times) 187 University of Ghana http://ugspace.ug.edu.gh SECTION N: ASSETS Circle the items the household owns and provide the current value or amount paid Item Number N u m b e r 1 Radio ………………… Cedis…………… 2 Television ………………… Cedis…………… 3 Mobile Phone ………………… Cedis…………... 4 Bicycle …………………… Cedis…………… 5 Refrigerator …………………… Cedis…………… SECTION O: LIVESTOCK OWNED Do you own livestock? 1. Yes 2. No >> End Interview Livestock Do you own this How many do you own What is the value of livestock (Number) the livestock owned 1.Yes (Amount in Ghana 2. No >> Next cedis) livestock Cattle Draught animals (oxen, donkeys, horses etc.) Poultry Rabbits Sheep Goats Pigs Fish (estimate number) Beehives Others ( Please specify) 188 University of Ghana http://ugspace.ug.edu.gh Appendix 2: Interview Guide Interview guide for FGD and individual interviews Questionnaire for FGD Date of Interview: Interviewer: Location: Protocol: Duration: Translator: Section 1: Basic Info on group >General Status quo Analysis History of the group: Could you please give a brief history of your group? Year of registration? Reason for group formation? Number of group members (m/f)? Popular crop grow and reason for that crop? Average farm size? Do you have experiences with OGS? a. Yes b. no Are you currently in any OGS? a. Yes b. no proceed with section 3 Section 2: formality of contractual arrangements under the OGS 2a Could you please tell us more about your OGS, who the contractor is and how it came into being! What motivated your participation in the OGS? Probe for details and multiple explanation What is in the sorghum OGS that motivated your participation? Is there specific requirement to meet before consider to join the OGS? Who is your current contractor (buyer/aggregator)? How did it come into being, who was the initiator? Did you have any guidance in designing or negotiating contract terms? How much time did you have to discuss the contract with group members? For how many years have you been farming under the OGS arrangement? • With the current contractor? • With others? • Which products? Who was negotiating the current contract and how? 2b kindly list and explain more about specific elements of your current OGS arrangement such as the product, quantity, quality, delivery period etc. What is the contract duration? a. Year b. 6 months others: Produce 189 University of Ghana http://ugspace.ug.edu.gh Quantity/planting programme a. fixed at Prices b. minimum or range Quality, specified and how? a. Yes b. no a. Farmers b. special graders c. contractor Who does the grading? d. aggregator. Input supply: Who supply what? Farmer and buyer /contractor? Seeds a. Farmer b. buyer c. GGBL Fertilizer a. Farmer b. buyer c. GGBL Chemicals, pesticides, herbicides a. Farmer b. buyer c. GGBL Postharvest support (Harvesters, tarpaulin, sacks, a. Farmer b. buyer c. aggregator d. scales) GGBL In case you are paid by cheque: How much do you pay to cash the cheque? When/how is the loan deducted from your sales? What percentage goes into servicing credit? Is there an interest rate? External support: Could you please tell us what kind of external support you receive in the various areas of production and marketing and by whom! Production how often: � Chemical application how often: � Record keeping how often: � Water management how often: � Gap training how often: � grading how often: � Do you know where to get advice or information on OGS? Transport of produce: Could you please tell us more about the way the transport of your produce is organised? Where is the produce picked up? Distance from pick up point to next main road or factory. How often is the produce picked up? Who pays the transport costs? 190 University of Ghana http://ugspace.ug.edu.gh Are there documents, which have to accompany the produce (traceability)? To whom do the bags belong? Rejection of produce: What is your experience with rejections of your produce? Is there anything mentioned in the contract how to deal with rejections? (price reduction) Where does the rejection take place? How is it justified? What are reasons? Do you have a possibility to intervene? What happens with the produce when it has been rejected? Legal issues in the contract and its enforcement Please tell us whether the contract has an exit/termination clause for both parties? Are there clear sanctions to mitigate breach? Is an arbitrator specified? How to deal with force majeure/natural calamities/natural risks? Is there risk sharing? Who signed the contract? Any witnesses present (MoFA)? Group mechanisms/group management: Could you please tell us more about your mechanism and the way in which you manage your group! Are there charges for group management? Does the group have written by-laws (rules and regulations including sanctions)? How often do you meet? Standards: Could you please tell us your experience with standards such as aflatoxin? 2c Are there any things you would like to be added to/regulated by your contract? How should the ideal contract look like? 2d What kind of support do you wish for your future engagement in OGS? 2e Problems/challenges faced by the group: Please tell us the problems you were faced with since you entered into OGS? Was there contract breach? Who breached the contract and why? Pick up of produce Mode of payment Price 191 University of Ghana http://ugspace.ug.edu.gh Input supply Quality of seeds Efficacy of chemicals Extension service, support, training Quality of fertilizers 2e Advantages/disadvantages of OGS Which advantages do you see for you as a farmer group being engaged in OGS? Which disadvantages do you see for you as a farmer group being engaged into OGS Which advantages do you think does the buyer have when engaged in OGS? Which problems do you think does the buyer face when engaged in OGS? Section 3: Informal arrangements (no CF experiences) 3a You told us that you don’t have any experiences with OGS. So please tell us more about the way you market your produce! How do you market your produce at the moment? Where do you market your produce at the moment? Are you satisfied with your current marketing system? 4. What is you understanding of climate change? What factors explain climate change in your opinion? Have you observed any changes in the last 30 years? What is the impact on your farming activities (positive and negative)? probe for details Has the climate change impacted on your livelihoods? probe for details Do you receive any training/support on climate change by the OGS? What kind of training/support was given? Who provide the training/support? Has the training/ support received has any impact on your farming activities (positive/negative)? Probe for details Do you have any question, comment and contribution on climate change? Further guiding questions for expert interviews:  What is sorghum farming like (production, profitability, postharvest loss)?  What do you think the role of institutions like MoFA should be in order to improve the current situation in OGS?  What are the main driving forces for participating in OGS?  What are criteria for consideration when contracting farmer groups (size, volume delivered, establishment of the group, location in a certain area)?  How is the process/what are the steps in contracting new farmer groups? 192 University of Ghana http://ugspace.ug.edu.gh  Which are the challenges/problems you are faced with as a processor/buyer concerning farmers you are working with and buyers in Ghana?  Does all the produce you are processing/packing come from formally contracted “out-growers”?  What is the climate change situation like?  Were you given training/supported on how farmers could adapt to the climate change?  What kind of training/support was given?  Has the training or support made farmers resilient to climate change? 193 University of Ghana http://ugspace.ug.edu.gh Appendix 3: Postharvest Loss Along the Various Postharvest Chain Table A1: Causes of Sorghum Losses During Heaping Treatment Control Overall Percent of Percent of Percent of Freq. cases Freq. cases Freq. cases High rainfall 4 21.1 24 25.0 28 24.4 Delay heaping 6 31.6 20 20.8 26 22.6 Lack of good platforms to heap sorghum 15 79.0 68 70.8 83 72.2 Birds attack 2 10.5 50 52.1 52 45.2 Livestock attack 3 15.8 43 44.8 46 40.0 Fire outbreak 0 0.0 2 2.1 2 1.7 Other specify 2 10.5 12 12.5 14 12.2 Table A2: Causes of Sorghum Loss During Transportation Treatment Control Overall Percent Percent Percent Freq. of Cases Freq. of Cases Freq. of Cases Poor roads 16 80 24 63.2 40 68.97 Lack of appropriate vehicles 13 65 23 60.5 36 62.07 High cost of transport 3 15 13 35.2 16 27.59 Distance from farm to storage centres 13 65 20 52.6 33 56.9 Other specify 2 10 3 7.9 5 8.62 194 University of Ghana http://ugspace.ug.edu.gh Table A3: Causes of Sorghum loss During Drying Treatment Control Overall Freq. % Freq. % Freq. % Rodents attack 30 10.0 7 3.2 37 7.2 rainfall/high moisture 77 25.7 67 31.0 144 27.9 high temperature 7 2.3 0 0.0 7 1.4 Birds attack 13 4.3 9 4.2 22 4.3 Livestock 61 20.3 45 20.8 106 20.5 Other specify 5 1.7 3 1.4 8 1.6 Table A4: Causes of Sorghum losses during Threshing and Winnowing Treatment Control Overall Freq. % Freq. % Freq. % Wind 12 5.6 21 7.0 33 6.4 Grains remain in thresh due to improper or incomplete threshing 73 33.8 66 22.0 139 26.9 Grains remain in soil 8 3.7 5 1.7 13 2.5 Livestock 6 2.8 2 0.7 8 1.6 Birds 1 0.5 1 0.3 2 0.4 Other specify 1 0.5 4 1.3 5 1 Table A5: Causes of Sorghum losses during Grading and Bagging Treatment Control Overall Freq. % Freq. % Freq. % Grains spilling 19 8.8 52 17.3 71 13.8 Livestock 11 5.1 1 0.3 12 2.3 Other specify 1 0.5 4 1.3 5 1 195 University of Ghana http://ugspace.ug.edu.gh Table A6: Causes of Sorghum losses during Storage Treatment Control Overall Freq. % Freq. % Freq. % Rot due to high moisture or rainfall 29 13.4 33 11 62 12 Aflatoxin/Mould 1 0.5 26 8.7 27 5.2 Insects and other pest 6 2.8 33 11 39 7.6 Rodents 0 0.0 14 4.7 14 2.7 Livestock 0 0.0 1 0.3 1 0.2 Other specify 0 0.0 4 1.3 4 0.8 196 University of Ghana http://ugspace.ug.edu.gh Table A4: Causes of Sorghum loss During Threshing and Winnowing Treatment Control Overall Freq. % Freq. % Freq. % Wind 12 5.6 21 7.0 33 6.4 Grains remain in thresh due to improper or incomplete threshing 73 33.8 66 22.0 139 26.9 Grains remain in soil 8 3.7 5 1.7 13 2.5 Livestock 6 2.8 2 0.7 8 1.6 Birds 1 0.5 1 0.3 2 0.4 Other specify 1 0.5 4 1.3 5 1 Table A5: Causes of Sorghum Loss During Grading and Bagging Treatment Control Overall Freq. % Freq. % Freq. % Grains spilling 19 8.8 52 17.3 71 13.8 Livestock 11 5.1 1 0.3 12 2.3 Other specify 1 0.5 4 1.3 5 1 Table A6: Causes of Sorghum Loss During Storage Treatment Control Overall Freq. % Freq. % Freq. % Rot due to high moisture or rainfall 29 13.4 33 11 62 12 Aflatoxin/Mould 1 0.5 26 8.7 27 5.2 Insects and other pest 6 2.8 33 11 39 7.6 Rodents 0 0.0 14 4.7 14 2.7 Livestock 0 0.0 1 0.3 1 0.2 Other specify 0 0.0 4 1.3 4 0.8 197 University of Ghana http://ugspace.ug.edu.gh Appendix 4: Propensity Score Matching 0 .2 .4 .6 .8 1 0 .2 .4 .6 .8 1 Before Matching After Matching Untreated: Off support Untreated: On support Untreated: Off support Untreated: On support Treated: On support Treated: Off support Treated: On support Treated: Off support Figure A1: Distribution of Propensity Scores Before and After Matching Table A7: Indicators of Matching Quality Before and After Matching Outcome Pseudo-R2 Pseudo-R2 Std. Bias Std. Bias % reduct Variable (Unmatched) (matched) (Unmatched) (matched) |bias| Farm 0.361 0.344 (0.000) 158.7 158.2 30.3 Profitability (0.000) Source: Survey data, 2015 Table A8: Sensitivity Analysis Treated Control Critical Value (Γ) On Support Off Support On Support Off Support 1.32 131 1 251 1 198 University of Ghana http://ugspace.ug.edu.gh Table A9: Quality Test of Matching Algorithm Variable Matched Treated Control %bias bias t p>t Gender U 0.713 0.653 12.8 1.43 0.153 M 0.652 0.621 6.5 49.2 0.51 0.610 Age of farmer U 46.648 50.097 -22.1 -2.46 0.014 M 48.515 57.879 -60.1 -171.5 -5.34 0.000 Marital Status U 0.870 0.833 10.4 1.16 0.247 M 0.811 0.765 12.8 -22.7 0.9 0.368 Household Status U 0.824 0.800 6.2 0.69 0.493 M 0.826 0.841 -3.9 37.1 -0.33 0.742 Completed Basic U 0.190 0.143 12.5 1.41 0.159 M 0.159 0.136 6.1 51.1 0.52 0.604 Higher than Basic U 0.171 0.080 27.8 3.19 0.001 M 0.098 0.083 4.6 83.4 0.43 0.670 Household Size U 5.630 6.163 -18.7 -2.07 0.039 M 5.402 3.977 49.9 -166.9 4.9 0.000 Main Occupation U 0.889 0.900 -3.6 -0.41 0.685 M 0.886 0.939 -17.2 -377.3 -1.53 0.128 Farming Experience U 5.796 9.137 -51.7 -5.46 0.000 M 6.030 10.061 -62.3 -20.7 -5.12 0.000 Length of Stay in the community U 28.949 34.637 -31.2 -3.47 0.001 M 30.091 41.311 -61.6 -97.3 -4.37 0.000 Farm Size U 1.032 0.688 51.2 6 0.000 M 0.815 0.921 -15.7 69.2 -1.2 0.230 FBO Membership U 0.625 0.007 177.7 21.66 0.000 M 0.394 0.379 4.4 97.5 0.25 0.801 Access to credit U 0.083 0.023 26.9 3.16 0.002 M 0.045 0.106 -27.2 -1 -1.87 0.063 199 University of Ghana http://ugspace.ug.edu.gh Number of Extension visits received U 1.120 0.110 83.7 10.11 0.000 M 0.409 1.030 -51.5 38.5 -4.57 0.000 Distance from community to main market U 2.445 1.978 10.1 1.21 0.229 M 2.562 3.347 -17.1 -68.3 -1.26 0.209 Leadership in social organization U 0.218 0.097 33.6 3.87 0.000 M 0.159 0.174 -4.2 87.5 -0.33 0.742 200 University of Ghana http://ugspace.ug.edu.gh Appendix 5: LVI Results in Percentage for Control and Treatment Table A10: LVI Sub-component Values and Minimum and Maximum Sub-components in Percentage for Control and Treatment Major component Sub-component Max for Min. for Control Treatment Both Both Water Percent of household reporting water 27.67 36.28 100 0 conflict Percent of households that utilize a natural 74 86.51 100 0 water source Average time to water source 8 10 60 0 Percent of household that do not have a 46.33 59.07 100 0 consistent water supply Inverse of the average number of liters of 0.12 0.13 1 0.02 water stored per household Socio- Dependency ratio 0.94 0.88 7 0 Demographic Percent of female headed household 22.33 18.14 100 0 Percent of households where head has not 76.33 65.12 100 0 attended school Percent of household with orphans 8 19.07 100 0 Food Percent of households dependent solely on 99.33 98.60 100 0 family farm for food Average number of months households 4 3 9 0 struggle to find food Average crop diversity index 0.17 0.17 0.5 0.08 Percent of households that do not save 1.33 0.47 100 0 seeds Percent of households that do not save 7.33 7.44 100 0 crops Social-network Average Receive: Give ratio 1.19 1.13 6 0.05 Average Borrow: Lend 0.93 0.93 2 0.5 Percentage of household that have not gone 93.67 98.6 100 0 to their local government for assistance in the past 12 months Livelihood Percent of households with family member 18.67 12.09 100 0 strategies working in different community Percent of households dependent solely on 90 69.77 100 0 agriculture as a source of income Average agricultural livelihood 0.32 0.34 0.5 0.25 diversification index Natural Disaster Percent of households that do not receive a 95.67 91.63 100 0 warning about the pending natural disaster Percent of households with injury or death 0.33 0.47 100 0 as a result of recent natural disaster Average number of floods, drought, 3 3 17 0 bushfires events in the past 6 years Mean standard deviation of monthly 11.1 11.1 12.8 9 average minimum daily temperature (years:1988-2013) 201 University of Ghana http://ugspace.ug.edu.gh Mean standard deviation of monthly 17.83 17.83 19.97 15.38 average maximum daily temperature (years:1988-2013 Mean standard deviation of monthly 83.8 83.8 115.68 7.49 average precipitation (years: 1983-2013) Health Average time to health facility (foot) 24 26 180 0 Percent of households with family member 9.33 12.56 100 0 with chronic illness Percent of households where a family 38 33.02 100 0 member had to miss work or school in the past 6 months Average malaria exposure*prevention 1.83 2.17 6 0 index 202