Gender relations in the millet value chain in western Kenya

(1)

Gender relations in the millet value

chain in western Kenya

Wai-Leng Chon

Student Number: 12532150

Email: wai_chon@hotmail.com

Date of Submission: Monday 17 August 2020

Institution: Graduate School of Social Sciences

Programme: MSc International Development Studies

Thesis Supervisor: Dr Verena Bitzer, Royal Tropical Institute (KIT)

Local Supervisor: Dr Osca Ingasia Ayuya, Egerton University

(2)

Wai-Leng Chon 12532150 Gender relations in the millet value chain in western Kenya

Abstract

Undernutrition continues to be a prevalent issue in western Kenya, and despite research into gender- and nutrition-sensitive value chains (NSVC), there is a research gap concerning gender relations in value chains extending beyond the economic impacts of agricultural production. This is significant because women are important stakeholders in agricultural value chains, playing a vital role at the nexus of agriculture and nutrition. Therefore, in collaboration with the EaTSANE project (Education and Training for Sustainable Agriculture and Nutrition in East Africa), this research aims to examine the gender relations in the finger millet value chain in Teso South, western Kenya. This study considers how nutrition-sensitive programmes like EaTSANE can leverage these findings in order to improve nutrition outcomes. It builds upon existing theory regarding women in agriculture to consider how gender relations shape participation in the value chain and the implications for strengthening NSVCs. It also contributes to the social debate surrounding whether models of gender-and nutrition-sensitive value chains can address undernutrition, particularly in scenarios shaped by local culture, religion and tradition. The research used qualitative methods, predominantly semi-structured interviews and focus group discussions with male and female millet value chain actors and experts. Results show that millet is an inherently gendered crop: cultural norms favour men in terms of land inheritance; the size of the land allotted to growing millet dictates who manages the crop; and who manages the crop impacts how the benefits (financial and non-financial) are generated and shared. Gender relations impact the value chain directly: women conduct labour-intensive productive and reproductive roles, lack access to resources and benefits and have limited intra-household bargaining power, which collectively weaken their productive and decision-making capacity, dampening the potential for participation in the millet value chain to translate into nutrition improvements. There exists a ‘can-do’ gap between what participants profess women are able to do, and what they actually do. These gender relations contribute to broader value chain phenomena of asymmetric information and a lack of coordination resulting in missed opportunities to enhance economic and nutritional impact, such as institutional markets and value addition products. To address this, EaTSANE should collaborate with extension services, research institutions and NGOs to 1) Improve the efficiency of the NSVC to increase availability and accessibility of diversified nutrient-dense foods; 2) Empower women to integrate in value chains; and 3) Improve nutrition- and gender-sensitive education to encourage behavioural change. EaTSANE should also be aware of the limitations of their ability to address this issue, given wide-ranging and deep-rooted socio-cultural norms. Nutrition-sensitive programmes such as EaTSANE are in prime position to engage with stakeholders and deliver interventions that improve women’s situations and increase nutrition, which is all the more important in the increasingly fragile, post-COVID world. Keywords: Gender relations, finger millet, nutrition-sensitive value chain, Kenya

(3)

Acknowledgements

First and foremost, I would like to extend my gratitude to my supervisor, Dr Verena Bitzer, for her continued support, guidance and critical eye throughout the research and thesis-writing process. I would also like to thank Dr Osca Ingasia Ayuya at Egerton University in Kenya, for acting as my local supervisor and for providing local expertise to ground my research approach. Many thanks also go to Juliet Sakaya, for her support as research translator and facilitator in Teso South. I would also like to thank Dr Lydiah Waswa, from the EaTSANE project in Kenya, and Mona Dhamankar at KIT, for their guidance during fieldwork. Lastly, I extend my gratitude to Dr Esther Miedema, the Second Reader of this thesis and Lecturer of the Gender, Intersectionality and Development course that provided much inspiration for this research.

(4)

1.3 Problem statement ... 3 1.4 Research question ... 4 1.5 Thesis structure ... 4 2. Theoretical framework ... 5 2.1 NSVC and programmes ... 5 2.2 Gender relations ... 6 2.3 Socio-cultural norms ... 9 3. Research Methodology ... 11 3.1 Conceptual scheme ... 11 3.2 Unit of analysis ... 13 3.3 Research location ... 13

3.4 Data collection methods ... 14

3.5 Sampling ... 16

3.6 Data analysis methods ... 16

3.7 Methodological reflection ... 16

3.8 Ethical reflection ... 18

4. Context ... 20

4.1 Gender and policy ... 20

4.2 Iteso culture ... 20

4.3 EaTSANE ... 20

Chapters 5-8 Discussion of findings ... 21

(5)

5.1 Value chain mapping and characterisation ... 21

5.2 Economic, nutritional and social value ... 24

5.3 Broad power relations in the millet VC ... 26

5.4 VC opportunities ... 27

5.5 Conclusion ... 28

6. How the dynamics of gender relations shape the millet value chain ... 29

6.1 Finger millet as a gendered food crop ... 29

6.2 Gendered division of labour and the ‘can-do’ gap ... 30

6.3 Unequal access to/control over resources and benefits ... 33

6.4 Decision-making ... 36

7. The influence of socio-cultural norms ... 39

7.1 Iteso culture ... 39

7.2 Religion ... 42

7.3 Education ... 43

8. How nutrition-sensitive programmes can enhance nutrition outcomes through these findings .... 48

8.1 Opportunities identified from research findings ... 48

8.2 Constraints ... 54

8.3 Other opportunities – out of scope for EaTSANE ... 56

9. Conclusion ... 57

9.1 Summary of findings ... 57

9.2 Theoretical reflection ... 59

9.3 Recommendations and suggestions for further research ... 62

10. Appendices... 64

10.1 References ... 64

10.2 Interview participants ... 69

10.4 Operationalisation ... 73

(6)

List of Tables

Table 3.1 FGD locations and participants ...14

Table 3.2 Interviewee VC role and gender ...15

Table 3.3 Expert interviewee role and gender...15

Table 6.1 Gendered division of labour in millet VC ...30

Table 8.1 Key findings and opportunities ...49

Table 9.1 Gender and value chain reflections on the agriculture-nutrition pathways ...60

List of Figures

Figure 3.1 Conceptual scheme ……...12

Figure 3.2 Research location ...13

Figure 5.1 Millet VC process map ...22

Figure 7.1 Level at which participant left education. n =20 ...43

(7)

List of Acronyms

DFID Department for International Development

EaTSANE Education and Training for Sustainable Agriculture and Nutrition in East Africa FAO Food and Agriculture Organization of the United Nations

FGD Focus group discussion

FNS Food and Nutrition Security

GBC Gender-based constraint

GDP Gross domestic product

HSNP Hunger Safety Net Programme

IFAD International Fund for Agricultural Development KALRO Kenya Agricultural & Livestock Research Organisation

KES Kenyan shilling

KSC Kenya Seed Company

NGO Non-government organisation

NSVC Nutrition-sensitive value chain

OECD Organisation for Economic Co-operation and Development PoU Prevalence of undernutrition

SDG Sustainable Development Goal

SIGI Social Institutions and Gender Index SME Small and medium-sized enterprise TIPs Trials of Improved Practices

UN United Nations

UNDP United Nations Development Programme UNEP United Nations Environment Programme

UNICEF United Nations International Children's Emergency Fund

VC Value chain

WEAI Women's Empowerment in Agriculture Index

(8)

1. Introduction

1.1 Undernutrition is a critical development problem

Undernutrition leads to premature deaths, illness and impairment for hundreds of millions of people (Anim-Somuah, Henson, Humphrey, & Robinson, 2013). In 2012, the Copenhagen Consensus ranked improving nutrition as the highest priority development intervention (DFID, 2012). In 2018, Eastern Africa had the highest prevalence of undernourishment (PoU) in the world at 30.8%, and in Kenya, drought and economic slowdown has contributed to increased undernourishment (FAO, 2019). However, evidence shows that increasing agricultural production and incomes is insufficient to address undernutrition (Anim-Somuah, Henson, Humphrey, & Robinson, 2013). Improving agricultural income does not guarantee access to nutrient-dense foods, nor equal access for all household members to improved nutrition (DFID, 2012). A review of 23 agricultural interventions reported mixed results and was unable to conclude whether these interventions have a positive impact on nutritional status (Masset, Haddad, Cornelius, & Isaza-Castro, 2012).

Therefore, rather than focusing solely on increasing agricultural production, a value chain (VC) approach has established a place in development discourse. It provides an analytical framework that focuses on the key actors and processes – from production through marketing and, eventually, consumption – that make the biggest difference to delivering foods to consumers (Anim-Somuah, Henson, Humphrey, & Robinson, 2013). Relevant to this research is the recent focus on the nutrition-sensitive value chain (NSVC), which identifies the most important hotspots in the flow of agricultural resources, goods, money and information which can have a positive impact on nutrition (IFAD, 2018). While a VC approach might focus on the “farm to fork” concept, NSVCs consider how to maximise the nutritional impact at every stage and process of the VC. Therefore, NSVCs expand the scope of food systems to encompass all elements from inputs and food production, to processing and consumption, with a specific view to nutrition (Jaenicke & Virchow, 2013). A 2011 cross-country study found that “a nutrition-sensitive growth strategy is undoubtedly the best means of sustainably eradicating malnutrition” (Headey, 2011, p. 4).

Key to understanding how a NSVC approach can address undernutrition is examining the links between agriculture and nutrition. Ruel and Alderman (2013) identified six pathways through which agriculture impacts nutrition: 1) As a source of food; 2) As a source of income; 3) Food prices affecting income, purchasing power and budget choices; 4) Women’s social status and empowerment through resources, decision-making and intra-household resource allocation; 5) Women’s time; and 6) Women’s own health and nutritional status (Ruel & Alderman, 2013). These impact pathways highlight the potential impact of agriculture on nutrition beyond food and income. Importantly, three of these six pathways are dependent on women, who are recognised as a critical consideration in the link between agriculture and nutrition.

(9)

1.2 The role of women in VCs

Women are important stakeholders in agricultural VCs, making up 43% of the global agricultural labour force (FAO, 2011). Although the estimated figures vary, it has been asserted that in Africa, women account for 70% of agricultural labour and 80% of food processing labour (Ruel & Alderman, 2013). They have an important role at the nexus of agriculture and nutrition, through their roles as agricultural producers and marketers, and as the household member typically responsible for food preparation and consumption, as well as childcare and feeding practices (Kurz, 2013). Studies indicate that within households, the question of who controls the income from farm sales and other means of employment has a bearing on household nutrition (Malapit & Quisumbing, 2016). Income-generating activities for women are found to have positive effects on nutritional status (Newton, Verhart, & Bake, 2018). A 2018 study in Kenya found positive associations between women’s empowerment and dietary diversity at individual and household level (Kassie, 2018). Gender inequalities tend to exacerbate undernutrition during periods of economic slowdown (FAO, 2019). However, dynamic gender relations throughout the VC often dictate what crops men and women grow, how resources are allocated and who partakes in marketing activities. If female farmers had the same access to resources as men, women’s yields and total agricultural output could increase, reducing the number of food insecure people globally by approximately 150 million (UN, 2019), or 12-17% (FAO, 2011). Closing the gender gap could reduce poverty and improve nutrition both directly (those working in agriculture) and indirectly (those employed in sectors linked to agriculture) (UN Women, UNDP, UNEP and World Bank, 2015). Although women are identified as a key determinant for the impact of agriculture on nutrition, it is through the lens of gender – the “roles, behaviours, activities, attributes and opportunities that any society considers appropriate” for men and women (WHO, 2020) – that we may shed more light on the impact pathways. Yet gender often remains overlooked in agricultural development programmes. Women have more difficulties accessing productive resources and participating in and benefitting from agri-food VCs. Women are also disadvantaged disproportionately from reduced natural resource availability (FAO, 2017). Interventions in agricultural VCs can be “highly disruptive and damaging” without sound gender analysis (Coles & Mitchell, 2011, p. 11) and are most effective when they incorporate contextual factors such as women’s social status and empowerment (Ruel, Quisumbing, & Balagamwaba, 2017). Existing gender-sensitive value chains (GSVC), which integrate gender concerns into VC development projects, do not holistically assess the opportunities for men and women in the VC (FAO, 2017). Gender-equitable VCs aim to go beyond gender sensitivity to achieve “Gender-equitable gender relations in terms of labour division, income generation, and overall livelihood and business decisions” (Stoian, Donovan, Elias, & Blare, 2018, p. 497). However, a review of guides for gender-equitable VC development highlighted conceptual and methodological blind spots including treating economic growth and gender equality as mutually supportive goals, and failing to acknowledge the impact of women’s bargaining positions at the household level in enhancing strategic decision-making and negotiating their roles within a VC (Stoian, Donovan, Elias, & Blare, 2018). This narrow focus on economic benefits measured in terms of productivity, income and sales is inadequate in assessing the outcomes of interventions in women’s collective action. In addition, current perceptions of agricultural VCs tend to overstate roles in

(10)

farming, and understate roles in marketing, creating a skewed perception of the relative importance of roles, which has a gender impact (Baden, 2013).

1.3 Problem statement

There is a scholarly research gap in studying the gender dynamics of NSVCs using holistic approaches that explore gender relations beyond economic factors. Specific gaps include studies at the intra-household level (FAO, 2017) and where changes to particular crops’ economic success may disrupt traditional gender dynamics and have a broader bearing on well-being (Baden, 2013). Literature suggests that research is needed to understand how gender dynamics in different contexts affect the mechanisms that drive nutritional and health outcomes (Malapit & Quisumbing, 2016), and in particular any unintended negative impacts of agriculture on nutrition (Malapit, 2019). Addressing this gap would enable nutrition-sensitive programmes to understand the dynamic nature of gender relations and equip them with insights that could improve their eventual impact on nutrition outcomes.

In order to address these issues and to maximise its practical, theoretical and social relevance, this study is embedded in EaTSANE (Education and Training for Sustainable Agriculture and Nutrition in East Africa), an ongoing action research project that aims to improve household diets in Kenya and Uganda by diversifying the food system through a NSVC (EaTSANE, 2019). EaTSANE’s activities are aimed at addressing its main research question: ‘What food system innovations in farming practices, VCs and nutrition behaviour lead to improved diets and more sustainable farming in Uganda and Kenya?’

In western Kenya, the finger millet VC is particularly important because this traditional crop is a nutritious and viable income alternative to cash crops such as maize. It is better adapted to agro-ecological conditions and is therefore more resilient to climate change and soil degradation. It has lower transaction costs and can be consumed by the household in times of unfavourable markets or food shortage, which increases food security (Handschuch & Wollni, 2013). As one of the main crops included in the EaTSANE project, further research into finger millet will contribute directly to the project’s interventions to diversify local food systems. The objective of the study is to understand the gender-based drivers and constraints to men and women’s participation in the millet VC, and their access to the entailing rewards, which is correlated with nutritional status. Through examining the millet VC in western Kenya, the study aims to understand the role that gender relations, intra-household dynamics and the wider community setting could play in supporting or hindering the nutritional impact of the project’s interventions. These findings will enable nutrition-sensitive programmes to more accurately and meaningfully design interventions that consider these gender relations in order not to exacerbate them, and that can address the gender relations directly to further improve nutrition. The study aims to uncover the crop- and area-specific gender dynamics in order to expand the existing body of research and create a practical platform upon which nutrition-sensitive programmes act.

This study explores the nexus between three elements: gender, nutrition and the VC, building upon existing research, which typically focuses on two of these elements, most often women and agriculture. This research better positions arguments relating to gender relations in the VC, and the subsequent link

(11)

between VC integration and nutrition. The theoretical relevance of the study is strengthened by a focus beyond economic drivers and benefits of participation in the millet VC, to provide a more comprehensive analysis on VC dynamics than is common in market-driven approaches to VC analysis. In addition, it discusses how intra-household dynamics affect the efficacy of nutrition-sensitive programmes.

The study will inform the burgeoning social debate around the role of women not just in agriculture, but in VCs on the whole, building upon the debate around the proportion of women in agriculture. The research builds upon broader discussions in international development studies, such as the extent to which standardised models for gender- or nutrition-sensitive VCs (FAO, 2017; IFAD, 2018) can apply to social contexts that are specifically shaped by culture, religion and tradition.

This study does not focus on measuring the eventual nutritional impact, as this would require more time and data to build a nutrition baseline. Rather, it aims to understand the complex interplay of gender dynamics which the study assumes may have an eventual impact on nutrition.

1.4 Research question

How do gender relations shape the millet agricultural value chain in western Kenya, and what lessons can be applied to nutrition-sensitive programmes?

The sub-questions beneath this over-arching question are:

1. How is the traditional millet agricultural value chain in western Kenya structured?

2. What are the dynamics of gender relations in the millet value chain regarding division of labour; access to/control over resources and benefits; and decision-making?

3. How are the above dynamics influenced by the socio-cultural norms of Iteso culture, religion and education?

4. How can nutrition-sensitive programmes leverage gender-related opportunities and address constraints in order to improve nutrition outcomes?

1.5 Thesis structure

This thesis comprises ten chapters. Chapter 1 introduces the research, problem statement and research questions. Chapter 2 posits existing theories and interprets relevant concepts. Chapter 3 presents the research methodology including conceptual scheme. Chapter 4 provides context to the research. Chapters 5-8 provide research discussion, addressing each research sub-question in turn. Chapter 9 concludes the thesis, summarising the findings and providing a theoretical reflection. Chapter 10 includes the appendices, such as participant information and coding network structure.

(12)

2. Theoretical framework

This chapter presents the central debates regarding the research topic and clarifies the selected conceptualisations. First, it explores the concept of the NSVC, and the subsequent application in nutrition-sensitive programmes. Second, it discusses gender relations and how they apply to NSVCs. Third, it explores conceptualisations of socio-cultural norms, in particular: Iteso culture, religion and education.

2.1 NSVC and programmes

A value chain is defined as “the full range of activities which are required to bring a product or service from conception, through the different phases of production ... to final consumers, and final disposal after use” (Kaplinsky & Morris, 2002, p. 4). In this definition, the concept of the “chain” is the sequential linking of different phases that a commodity goes through from production to consumption. (IFAD, 2018). Food VCs are classified into: traditional, modern, modern-to-traditional and traditional-to-modern. In traditional VCs, traders buy primarily from smallholder farmers and sell to consumers in local markets. Modern VCs are where domestic and multinational food manufacturers procure from commercial farms and sell through modern supermarkets. The modern-to-traditional VC is where domestic and multinational food manufacturers sell through a network of traditional traders and retailers. The traditional-to-modern VC is where supermarkets and food manufacturers procure food from smallholder farmers and traders (Gómez & Ricketts, 2013). This research focuses on the traditional VC, upon which rural communities primarily depend to access adequate quantities of calories and micronutrients, and where EaTSANE focuses project activities. Modern VCs are thought to have little nutritional impact (Gómez & Ricketts, 2013). VCs have been identified as a potential strategy to improve nutrition through agriculture, by acting as an entry point to stimulate supply or demand for micronutrient-rich foods. This market-based approach commonly perceives the concept of value as economic value that is cumulatively added to the commodity at each stage along the VC (Ruel & Alderman, 2013). However, others argue that the chain should be viewed as a more complex network of activities and relationships between different actors, with the chain acting as a metaphor for their multi-directional connectedness (Humphrey & Navas-Alemán, 2010). Likewise, the NSVC urges us to view the concept of value differently, beyond economic value, to consider how nutrition value can be added or taken away by different processes and through different interactions in this network. The NSVC approach therefore identifies the most important hotspots in the flow of agricultural resources, goods, money and information that can have a positive impact on nutrition (IFAD, 2018).

By considering the activities, stakeholders, and economic and nutritional value at each stage, from inputs and production to consumption, the NSVC approach is better positioned to address undernutrition, a form of malnutrition that includes wasting, stunting, underweight and deficiencies in vitamins and minerals (WHO, 2020). It also shifts away from considering agriculture growth and food production as the key interventions to reduce undernutrition. While agriculture growth has been shown to have an impact on reducing undernutrition, it may be insufficient to address the problem of scarcity

(13)

of access to nutritious and diverse diets (Ruel & Alderman, 2013). A cross-country analysis found that the impact of agriculture growth on nutrition decreases as incomes and calorie consumption rise (Headey, 2011). Beyond a certain point, addressing undernutrition may be less about how much is produced and consumed, and more about what specific nutrient-rich foods are consumed. Nutrition-sensitive agriculture takes a systems approach to addressing undernutrition by shifting the focus from producing sufficient calories to considering vitamins, minerals and micronutrients. It aims to sensitise the agricultural sector to the importance of nutrition and health (Jaenicke & Virchow, 2013).

Nutrition-sensitive programmes can address the potential to improve nutrition through both nutrition-sensitive agriculture and the VC by implementing specific interventions at nutritional hotspots. These interventions may be agricultural or relating to behaviours and consumption patterns that also impact nutritional status (Ruel & Alderman, 2013). The programmes address the underlying causes of undernutrition, particularly for vulnerable groups such as women and children. They also serve as delivery platforms for nutrition-specific interventions that address the immediate determinants of nutrition and development, such as adequate nutrition intake and feeding practices. Nutrition-sensitive policy interventions play a key role in reducing undernutrition (Anim-Somuah, Henson, Humphrey, & Robinson, 2013), because “although addressing general deprivation and inequity would result in substantial reductions in undernutrition and should be a global priority, major reductions in undernutrition can also be made through programmatic health and nutrition interventions” (Black, et al., 2008, p. 5). However, the impact of nutrition-sensitive programmes is not yet defined, with researchers calling for theory- and cost-based effectiveness assessments, as well as research to understand any potential barriers to participation in the programmes (Ruel & Alderman, 2013).

2.2 Gender relations

It is important to distinguish between how the literature conceptualises both ‘women’ and ‘gender’ in the VC. Three of the six agriculture-nutrition pathways identified by Ruel and Alderman (2013) involve women: women’s social status and empowerment through resources and assets, decision-making power and intra-household allocation of resources; women’s time through participation in agriculture, whose effect may be positive or negative depending on the trade-off between income-generating activities and household maintenance; and women’s own health and nutritional status through participating in agriculture (Ruel & Alderman, 2013). This framework identifies causality between women’s roles in agriculture and nutritional outcomes. The interplay between these pathways and the different elements of gender relations represent a complicated web of correlation, causality and dependence. For example, women’s social status and empowerment (or lack thereof) influences the division of labour across the agriculture VC, where women typically undertake manual work in hazardous or arduous conditions. This, in turn, impacts women’s health and nutrition status – and subsequently that of their children as a result of their reproductive role (Malapit & Quisumbing, 2016). At each stage of the VC, these gender relations come into play and impact the causal link between agriculture and nutrition. This is particularly pertinent in the context of nutrition-sensitive programmes, because these pathways are not mutually exclusive, and may cause positive and negative impacts on nutrition.

(14)

The Women’s Empowerment in Agriculture Index (WEAI) measures women’s empowerment through domains including “(1) decisions about agricultural production, (2) access to and decision making power about productive resources, (3) control of use of income, (4) leadership in the community, and (5) time allocation” (Alkire, Meinzen-Dick, Peterman, Quisumbing, & Seymour, 2013, p. 73). The index also measures the relative empowerment gap between men and women in their households, as well as discussing the connection between nutrition-sensitive programmes and women’s empowerment. Nutrition-sensitive programmes often identify women’s roles as agricultural producers and household gatekeepers of nutrition security (Malapit, 2019). This dynamic is perceived as reciprocal: not only can nutrition-sensitive programmes help to empower women, but, among other things, “the nutrition-sensitivity of programmes can be enhanced by … focusing on improving women’s physical and mental health, nutrition, time allocation, and empowerment” (Ruel & Alderman, 2013, p. 537). Existing research tends to focus on women, and women’s empowerment is identified as a key mediator of nutrition impact (de la Peña & Garrett, 2018).

While acknowledging the important role that women play in linking agriculture and nutrition, considering women alone is insufficient. Many VC programmes mistakenly equate targeting women with working towards gender equality (FAO, 2017). A gender focus should look at the “roles, behaviours, activities, attributes and opportunities that any society considers appropriate” for both women and men (WHO, 2020). Viewing gender as one of several social relations that impact an individual’s social positioning implies taking a dynamic view towards the norms, roles and responsibilities of men and women. Connell (2009) defines the dominant constructs of gender relations as labour, cathexis, symbolism and power – reflecting how gender shapes people’s actions, thoughts, choices to represent themselves, and their environment. The gender regime has an impact on the division of labour, and this allocation affects social practices. For example, the labour split along a productive-reproductive line impacts not only the types of work that men and women do but also how the characteristics associated with that work are perceived. Social norms shape the patterning of desire, impacting the so-called laws of attraction. Gender also impacts the imagery surrounding men and women, style of dress and portrayals in the media. Surrounding and compounding this is a state hierarchy that legitimises gender-based violence through the gender order, including through laws and structural violence (Connell, 2009). Agarwal describes gender relations as the “socially-constructed relations of power between women and men”, spanning both the material and ideological spheres (Agarwal, 1997, p. 1). Due to the scope and VC topic, this research will consider gender relations through the dimensions of: (1) division of labour; (2) access to and control over resources and benefits; and (3) decision-making; all underpinned and shaped by socio-cultural gender norms (van Eerdewijk & Danielsen, 2015). Rather than viewing men and women as distinctly separate, static groups, viewing gender as relational invites exploration of the different and ever-changing ways that men and women are conceptualised in society.

The research focuses primarily on intra-household gender relations because of the role that nutrition-sensitive programmes have in potentially altering household resource allocations and the relative bargaining power of men and women in the household. It incorporates EaTSANE’s Gender Strategy by applying the agriculture-nutrition pathways and key decision-making moments in the VC (EaTSANE,

(15)

2018). This Strategy recommends gathering sex-disaggregated data, which better enables the research to capture the dynamic, socially-constructed nature of gender relations in the VC. Rather than creating absolute measures, the research considers the distinction between how men and women conceptualise and practise societal norms while participating in the VC. While it is not possible for the research to measure the WEAI due to time and practical constraints, the research considers the below three elements of gender relations:

Division of labour

Assessing women’s involvement in the VC is key for creating positive impacts on women’s and household nutrition and identifying necessary mitigation measures such as labour-saving technologies in order to ensure a “do-no harm” approach (de la Peña & Garrett, 2018). This requires considering the division of labour in farming and other aspects of the VC (van Eerdewijk & Danielsen, 2015). Women play a major part in the agriculture VC, with a productive role that often encompasses food processing and weeding. This combines with reproductive responsibilities, which traditionally fall within women’s domain in most societies (Komatsu, Malapit, & Theis, 2018). This implies that there are two levels where gendered division of labour can come into effect. In most communities in sub-Saharan Africa, the productive and reproductive roles are such that women produce subsistence crops for household consumption and perform household activities, and men provide land, grow cash crops for sale and take responsibility for payment of school fees (Quisumbing, 1994). It is therefore valuable to examine this gendered division of labour in the millet VC and the related impacts on nutrition and time.

Access to and control over resources and benefits

In the FAO’s guiding framework on developing gender-sensitive VCs, gender-based constraints (GBC) are defined as “restrictions on men’s or women’s access to resources or opportunities that are based on their gender roles or responsibilities” (USAID, 2009). They “highlight the many disadvantages faced by women in their efforts to engage in the economy” (USAID, 2009). That USAID and the FAO consider access to resources to be central to gender-based constraints is meaningful, yet this view of GBCs predominantly focuses on economic impacts and therefore excludes other linkages between agriculture – and associated access to resources and benefits – and nutrition. Others argue for a broader, more holistic view of resources and benefits outside of a narrow focus on economic benefits (Baden, 2013). Studies have found that women lack access to technological resources that could be labour- and energy-saving (van Eerdewijk & Danielsen, 2015). Understanding of women’s control over resources, as well as their social status and roles in and outside the household, is critical to ensuring that nutrition-sensitive programmes promote women’s empowerment in ways that are conducive to positive nutrition outcomes (IFAD, 2018). The IFAD guide to NSVC considers commodities to have a high potential to improve nutrition through women’s empowerment if women significantly control certain functions of the VC and benefit from participating without negative consequences for their own health and nutrition or for that of the household (IFAD, 2018).

(16)

Decision-making

Intra-household decision-making reflects the different bargaining power held by men and women in the household (Agarwal, 1997). This can have immediate and secondary effects on nutrition. Studies in sub-Saharan Africa found gender relations do not allow women to be effective decision-makers (Farnworth, Sundell, Nzioki, Shivutse, & Davis, 2013). Improvements in women’s decision-making power can have significant positive impacts on both their own nutrition and that of other household members (IFAD, 2018). Criticism of bargaining-power models within households include the suggestion that incomes are rarely pooled together, and that a rural person’s bargaining strength depends on several aspects including ownership and control over assets, access to employment and social norms (Agarwal, 1997).

2.3 Socio-cultural norms

Values and assumptions are discussed by van Eerdewijk & Danielsen (2015) as a subset gender relations, however this research identifies the concept of socio-cultural norms as one that underpins and shapes gender relations – and the two concepts as inextricably linked. Definitions of social norms mainly agree that they “encompass ‘rules’ that govern social behaviour and expectations” (Markel, Gettliffe, Jones, Miller, & Kim, 2016, p. 10). As a result, social norms cover many broad aspects, though this research will focus on the socio-cultural norms that impact and incorporate gender. The gender norms discourse highlights that gender norms are learned in childhood through socialisation (Tenenbaum & Leaper, 2002) and reinforced through family and the social context: schools, religion, media and social institutions (Cislaghi & Heise, 2020). These gender norms are reproduced through social interactions from the individual level through to how institutions function (West & Zimmerman, 1987), and this structure means inequitable gender norms reflect and perpetuate inequitable power relations that are often disadvantageous towards women (Connell, 2014). As such, this research operationalises the gender norms related to Iteso culture, religion and education – categories that are particularly pertinent to how gender relations are reproduced in the context of the millet VC and the institutions surrounding it. Gender norms theory discusses concordant and discordant norms and attitudes, which require different strategies for adjusting personal attitudes and practices (Cislaghi & Heise, 2020).

Cultural norms and traditions may restrict women’s ability to inherit land (Kameri-Mbote, 2005). Perceptions of who makes an effective marketer impact prices and nutritional purchasing power as a result. A study found that female sellers received substantially lower selling prices than male sellers unless they participated in a finger millet group (Handschuch & Wollni, 2013). Cash cropping is considered a male domain, therefore if the categorisation of finger millet changes from a traditional food crop to cash crop because of economic and environmental factors, this may have an impact on the gendered roles in this VC (Handschuch & Wollni, 2013). It is important to explore these socio-cultural norms as they shape the gender relations observed in the millet VC and can provide opportunities for nutrition-sensitive programmes to direct their efforts in light of these norms. Socio-cultural norms also reflect the limitations of nutrition-sensitive programmes in improving nutrition where the programmes

(17)

do not explicitly include exercises such as improving women’s empowerment. Norms based on long-held tradition, religion or culture may be beyond the scope of such programmes, and they should be designed such that they do not cause harm within the context of these existing norms.

(18)

3. Research Methodology

This chapter explains the methodology used to conduct the research. Firstly, I share the conceptual scheme, then describe the methods used to collect and analyse data. Lastly, I share methodological and ethical reflections on the research.

3.1 Conceptual scheme

Figure 3.1 reflects how concepts in scope of the research (denoted by the dotted line) are linked to the research questions. Socio-cultural norms influence gender relations, which impact directly the millet VC, as well as indirectly through nutrition-sensitive programmes. The eventual impact on nutrition is out of scope due to practical reasons, as this requires the development of a baseline.

(19)

Figure 3.1 Conceptual scheme

In scope of research

Nutrition-sensitive millet

value chain

Nutrition-sensitive

programmes

(e.g. EaTSANE)

Influence

1. How is the traditional millet agricultural value chain in western Kenya structured? 2. What are the dynamics of

gender relations in the millet value chain regarding division of labour; access to / control over resources and benefits; and decision-making?

3. How are the dynamics of gender relations influenced by the socio-cultural norms of Iteso culture, religion and education?

4. How can nutrition-sensitive programmes leverage gender-related opportunities and address constraints in order to improve nutrition outcomes?

Nutrition outcomes

Which conduct interventions in Which shape

Gender relations

Gendered division of labour Access to / control over resources and benefits Unequal decision-making power Which create opportunities or barriers for

Socio-cultural gender norms

Iteso tribal

culture Religion Education

(20)

3.2 Unit of analysis

The unit of analysis is the women and men who participate in the millet VC in Teso South. This coincides with, but is not equal to, the unit of analysis for EaTSANE. Millet is one of the three crops included in EaTSANE, selected for its nutrition content and potential to improve diversity in diets.

3.3 Research location

Figure 3.2 Research location1

The research took place between January and March 2020 in Kenya, where the prevalence of undernutrition (29.4%) has been rising since 2014 (FAO, 2019). Agriculture contributes 26% of Kenya’s

1_{Adapted from https://commons.wikimedia.org/w/index.php?curid=7410073 and}

https://commons.wikimedia.org/w/index.php?curid=38843970. Accessed 10 December 2019.

Busia County denoted in red

Kenya

Uganda

Tanzania

Somalia Ethiopia

(21)

Gross Domestic Product (GDP) and another 27% indirectly through linkages with other sectors. The sector employs 40% of the total population and over 70% of rural people (FAO, 2019).

The research was conducted in Teso South, one of seven constituencies in Busia County, western Kenya. It has a population of 168,116 in an area of 302.9 km2_{(Republic of Kenya, 2019). It being an}

EaTSANE site maximises the utility of research findings. I was based in Busia, a town bordering Uganda. Sections of the VC are conducted in Uganda, outside the research scope. As an important trade hub, it also offered a bustling setting for research.

In Busia County, the largest ethnic groups are the Luhya tribe, Iteso and Luo, and the Iteso ethnic group is the largest in Teso South (Miguel, 2002). The population speaks Ateso and/or Swahili and so a translator was required.

I conducted research in six villages, of which three are villages in which EaTSANE operates: Adungosi, Asinge and Goria. The non-EaTSANE villages are Alomodoi, Alupe and Apegei.

3.4 Data collection methods

Recognising that gender categories are socially constructed and relational, this piece of gender-informed research2_{used qualitative approaches to understand gender relations in the millet agricultural}

VC. Qualitative methods are well-suited to exploring context and cultural differences, opening embedded practices up for discussion and potentially change (Rubin, 2016). Qualitative research explores why behaviours and beliefs exist, probes diversions between beliefs and actual practices, and captures participants’ views in their own words. I used the following qualitative data collection methods:

Focus group discussions (FGD)

I conducted five FGDs, both single-gender and mixed-gender, in different villages, outlined below.

FGD Topic Village Men Women Total

1

Millet VC Adungosi 3 4 7

2 Asinge 4 4 8

3

Gender relations and norms

Goria 0 7 7

4 Apegei 5 4 9

5 Alomodoi 7 0 7

Total 19 19 38

Table 3.1 FGD locations and participants

The first two FGDs were on the topic of the millet VC structure: key stages, stakeholders, processes and activities. Subsequent FGDs covered the norms and values that shape socially-constructed notions

2_{Research that aims to apply information about men and women’s behaviours, values, opportunities}

(22)

of gender within the millet VC context. I conducted one FGD per week at safe, neutral locations, such as the Ward Administrator Office.

Semi-structured interviews

I conducted 36 interviews in total – 32 with millet VC actors and four with experts, detailed in Tables 3.2 and 3.3 below. I conducted the interviews following each week’s FGD, to follow up with individuals and elaborate further on the FGD discussion. In this research, farmer refers to finger millet producers; processor relates to an individual who transforms millet from its raw, seed form to another, for sale; retailer is someone who sells their millet product in the market; and trader refers to an individual who buys millet from farmers to sell on. Interviews are referenced using the form (I5), which indicates interview reference 5. FGD references take the form (FGD1:2) which indicates FGD1, reference 2. These references link to the Appendix (10.2 and 10.3).

VC actor role Men Women

Farmer 7 4

Farmer and Retailer 2 9

Farmer and Processor 0 2

Farmer, Processor and

Retailer 1 0

Trader/Buyer 2 2

Farmer and other 2 1

Sub-total 14 18

Total 32

Table 3.2 Interviewee VC role and gender

Expert role Men Women

Government Agricultural Officer 1 1 Research Practitioner (EaTSANE) 0 1 Research Practitioner (non-EaTSANE) 0 1 Sub-total 1 3 Total 4

Table 3.3 Expert interviewee role and gender

Interviews with VC actors related to their potentially multiple roles in the millet VC. I used an interview guide but allowed flexibility for participants to discuss topics relevant to their roles. I identified VC actor roles by discussing with a local facilitator and government extension officers at the beginning of data gathering. I defined experts according to their role as Government Extension Officers i.e. coordinators across the VC, or experience in research and nutrition-sensitive programmes. I conducted expert interviews at different stages in the data collection process to maximise knowledge-gathering of the millet VC and gender relations.

Participatory methods

In FGDs, I provided opportunities for participants to actively contribute and provide creative insights to topics that might otherwise be challenging to explore. In FGD1-2 (millet VC), I conducted VC mapping exercises with participants, where we drew a VC diagram and participants provided information related to stakeholders, inputs, processes and timelines for each stage of the VC. In FGD3-5 (gender relations and norms), the groups explored notions of what it means to be a ‘good’ man or woman, and linked characteristics to roles in the value chain.

(23)

Wai-Leng Chon 12532150 MIDS Thesis: Gender relations in the millet value chain

Direct observations

I kept detailed field notes based on direct observation (e.g. at daily markets), which helped me to understand common conditions or practices and identify specific topics for interviews and FGDs. However, observation of the different stages of the millet VC was limited because I conducted the research in the dry season, meaning there were few opportunities to observe farming activities, in particular those conducted by women.

3.5 Sampling

I used data from previous EaTSANE work to identify the initial set of FGD participants and interviewees in each village. I worked with Agricultural Officers and local facilitators who had previously worked with EaTSANE to select villages and participants with roles in various VCs (e.g. beer brewing, milling) and at the different stages of the VC(s). I used snowball sampling to conduct further interviews to ensure that I engaged with stakeholders across as many parts of as many VCs, including those of which I had less knowledge before entering the field.

3.6 Data analysis methods

As the interviews and FGDs were translated in real time by my translator, I transcribed the recordings in English. I cross-referenced between interviews and FGDs and identified the preliminary themes and codes while in the field, discussing these findings with my supervisor, local supervisor and translator, where appropriate. Once back from the field, I created a colour-coded map of the key themes and codes identified to create a high-level visual. I then conducted detailed qualitative analysis of the transcripts using the Atlas.ti programme to create a relational understanding of the NSVC, gender relations and socio-cultural norms. I created codes, categories, themes and eventually concepts in a hierarchical order. Through an iterative process of combining and grouping codes, I created a network diagram that defines how the codes are linked by relationships such as causality and association. I used a combination of deductive and inductive approaches to analysis. Interview and FGD questionnaires are based on the conceptual framework and operationalisation I defined pre-fieldwork, which led to generating codes by matching observations with this framework. I also conducted inductive coding, defining codes as I identified them from the data. The code network diagram is included in the Appendix (10.5).

3.7 Methodological reflection

Given my constructionist ontological position and my interpretivist epistemological position, applying quality criteria helps to minimise researcher bias and assess research credibility in a “‘post-positivist’ world – one that is infinitely complex and without a defined ‘truth’” (O’Leary, 204, pg. 58).

(24)

3.7.1 Trustworthiness Credibility

I triangulated by comparing FGD and interview findings, enabling me to explore the concepts in more detail and see if responses were consistent outside of the group setting. My supervisors and EaTSANE practitioners acted as auditors, so that I could adjust my approach to gather data that was most reflective of the social reality. After five FGDs and 36 interviews, I reached a point of data saturation as I was not gathering any new information. I worked with a translator/facilitator who translated into both Swahili and Ateso, enabling me to attain more credible responses from participants who could respond in their preferred language. Although we worked to mitigate against risks of inaccuracy, limitations exist where participants’ responses cannot be directly translated into English.

Transferability

Though transferability may be difficult to achieve for small-scale qualitative research, the use of thick descriptions enables other researchers and practitioners to judge which elements of my research may apply to their context. This may make the research findings applicable to comparable crops or research locations.

Dependability

I adopted an ‘auditing’ approach to reflect on subjectivities that are inherently apparent in qualitative research. I shared a comprehensive research update weekly with my supervisors and EaTSANE practitioners, creating a transparent audit trail regarding my decisions, methods, risks and mitigations. I recorded every interview and FGD conducted (with permission) and anonymised interview transcripts. The Transparency Document and Appendix (10.2 and 10.3) include information on interviews and FGDs conducted, excluding any identifying details.

Confirmability

Viewing gender through the lens of socially-constructed relations renders objectivity difficult to achieve. However, it is necessary to reflect on, and minimise, instances where my personal values may impact research methods and findings. In interviews and FGDs, I adopted open and simple lines of questioning. While thesis-writing, I reviewed my findings with supervisors and peers, who audited possible subjectivities.

3.7.2 Authenticity Fairness

I demonstrate the varied value systems and viewpoints of research participants by selecting a representative sample, and conducting interviews and FGDs with men and women in different VC roles, and same-sex and mixed-gender groups.

(25)

Ontological and educative authenticity

I believe helping research participants “arrive at a better understanding of their social milieu” (Bryman, 2012, p. 393) was a natural result of the research in which we discussed the participants’ beliefs about roles in the millet VC. In FGDs, I asked open questions and when a participant shared their opinion, I asked the rest if they agreed or disagreed, and why. By discussing different opinions during FGDs, participants could better appreciate where their opinions differed from others.

Catalytic/tactical authenticity

This research aims to enhance participants’ decision-making power, acting as an impetus to bring about change in the researched social setting. It will inform EaTSANE, and other nutrition-sensitive programmes, of ways to harness the research findings, enable sustainable agriculture and reduce undernutrition. EaTSANE can act upon participants’ recommendations to enable and empower them, and avoid enacting undesired change. This helps to transition away from considering participants as ‘subjects’ to active social agents and achieve a degree of social justice.

3.8 Ethical reflection

Ethical considerations are guided by the principle of doing no harm (Bryman, 2012). In line with the GSSS Ethical Guidelines for Student Research, I informed participants that they were not obligated to participate and adapted questions so as not to probe participants who showed reluctance. I shared a Research Information Sheet and Research Consent Form with all participants, whose contents were translated in person by my translator into Swahili and Ateso. Though deception may be inherently harmful, it is accepted as widespread in research to generate more honest and natural responses (Bryman, 2012). I did not reference ‘gender’ explicitly, because the term may be interpreted differently and may elicit socially desirable answers. I achieved a balance by providing sufficient information for informed consent and omitting some details in order to generate truthful responses, insofar as this did not harm participants. I tried to maintain participants’ safety by addressing sensitive topics initially through expert interviews, and then validating through interviews and FGDs once I understood the topics’ sensitivity. I aimed to conduct interviews in private, convenient locations, ideally in participants’ homes.

To achieve confidentiality, I use pseudonyms in transcripts to anonymise participants and do not include identifiable details in anecdotes. I typically conducted an FGD and then, once participants were familiar with me and my research approach, I invited them to take part in an individual interview. I found this an effective way to build rapport and trust. However, at times it was challenging to maintain an appropriate researcher-participant relationship, being a western researcher. For example, some interviewees explicitly asked for money, which would have been inappropriate and unfeasible. Advised by my local facilitator, I provided bars of soap to participants which, while inexpensive, are highly valued and yet often neglected during times of hardship. I also explained how EaTSANE might use the research, providing sustainable benefits.

(26)

3.8.1 Positionality

My positionality impacts my research objectives and methods, my perception of the participants, and how I interpret and analyse the data collected. I tried to refrain from embodying a ‘white feminist’, who acts as a saviour in the modern development context (Syed & Ali, 2011). Despite being a non-white woman, due to my British nationality and education, I may portray the same whiteliness as my white colleagues. I applied a social relations approach to gender research, whose operating principles include resisting the notions that all rural areas are the same, and their inhabitants are neutral actors with no gender, age, class or other identities (Okali, 2012). This implies reflecting on intersectionality, understanding individuals’ experiences at the intersections of race, gender, disability and socio-economic status (Crenshaw, 1991) so that I do not perceive all women as facing the same “single-issue struggle” (Lorde, 1982, p. 1). I aimed to examine the dominant narratives about men’s and women’s roles, understanding that men and women renegotiate dynamic gender relations to meet their own interests. I used sex-disaggregated data to understand men and women both separately and in relation to each other, rather than considering only women, which is a common shortcoming of gender research (Doss, 2014). I consider men and women within each household as individual and the dynamics between the two ever-changing (Agarwal, 1997).

(27)

4. Context

4.1 Gender and policy

Kenya’s Vision 2030 states enhancing Food and Nutrition Security (FNS) is one of its Big Four initiatives, addressed through high-impact nutritional interventions (Kenya, 2018). The Hunger Safety Net Programme (HSNP) provides social protection during times of income hardship (FAO, 2019). However, it is not nutrition-sensitive because it provides unconditional cash transfers (Hunger Safety Net Programme, 2020). Kenya have committed to promoting gender mainstreaming in their constitution, national laws, and intergovernmental and regional policies (Republic of Kenya, 2016). The Social Institutions and Gender Index (SIGI), which measures discrimination against women in social institutions, ranks Kenya 83rd_{out of 120 countries and 10}th_{out of 29 countries in sub-Saharan Africa}

(OECD, 2020). In 2016, 5% of Kenyan Chiefs – a position particularly important in rural communities – were women, below the constitutional provision of at least one third women in elective and appointive positions (Ministry of Devolution and Planning, 2017). A report outlining how Kenyan policy can enable the achievement of the UN Sustainable Development Goals (SDG) by 2030, identified such challenges to SDG5 (Gender equality) as the “harmful and prohibitive socio-cultural traditional practices and beliefs such as female genital mutilation” and the “high level of tolerance of Gender Based Violence… [due to] socio-cultural norms” (Ministry of Devolution and Planning, 2017, p. 31).

4.2 Iteso culture

All research participants belonged to the Iteso tribe, an ethnic group with pastoralist traditions and the largest ethnic group in Teso South (Miguel, 2002). The Iteso live in the predominantly Bantu-speaking area of western Kenya and eastern Uganda (Karp, 1978). Literature suggests the “Iteso have an undeserved reputation in Kenya for cultural conservatism” and share many customs with neighbouring ethnic groups (Karp, 1995, p. 127). Catholic missionaries have had considerable influence among the Iteso, almost all of whom had been baptised by 1990 (Karp, 1995).

4.3 EaTSANE

EaTSANE is a three-year project, scheduled to complete in 2021. While the findings from this research can be applied throughout EaTSANE, they are most relevant to work package 2 (WP2): ‘Value chains and enabling environments for diversified food systems’. This work package aims to “identify conditions and incentives for farmers, farmer organisations and SMEs in value chains to engage in sustainable and diversified nutrient-dense crop production, processing, trading and marketing” (EaTSANE, 2019). Finger millet was selected for EaTSANE due to its potential to improve and diversify nutrition in the research area. To date, millet VC-mapping exercises have been conducted, but no other activities which have been conducted for other crops, such as Trials of Improved Practices (TIPs).

(28)

Chapters 5-8

Discussion of findings

The following chapters provide the research findings, addressing each sub-question in sequence.

5. Structure of millet agricultural value chain in western Kenya

This chapter provides an analysis of the millet VC through a nutrition lens (de la Peña & Garrett, 2018). First, I analyse the VC structure and actors. Second, I explore ‘value’ in terms of economic, nutritional and social value. Third, I discuss broad power relations in the millet VC which facilitates the examination of gender relations, a subset of the enabling environment. Finally, I discuss further opportunities to maximise the nutrition impact of the VC.

5.1 Value chain mapping and characterisation

This research focuses on the traditional finger millet VC, where traders buy from smallholder farmers and sell to consumers in local markets, as opposed to the larger-scale modern value chain, which involves commercial farmers, supermarkets and multinationals (Gómez & Ricketts, 2013). There are up to two millet harvests per year, with the larger harvest produced between February and June, and the smaller crop produced between August and December. Figure 5.1 is a VC process map, informed by participatory research findings, depicting nutritional hotspots and how actors’ interactions result in nutritional impact. While not exhaustive, it highlights the following:

Key stakeholders in the millet VC: farmers, processors, retailers and traders. Rather than being a chain of sequential activities, the millet VC is shaped more like a network between these stakeholders with different bargaining and power relations between stakeholder groups at different stages of the VC; Processes and activities (primarily) conducted by each stakeholder group, aligned to high-level process headings: Inputs/land preparation; Food production; Storage and processing; Distribution and transport; Trading and marketing; Preparation and consumption (IFAD, 2018);

Key decisions in the millet VC, such as whether to eat, store or sell harvested millet, or where and how to sell the millet, the results of which have a nutritional impact;

Different forms of millet such as grain, flour, yeast, staple food, snacks and local brew. Each has differing economic, social and nutritional value;

(29)

Figure 5.1 Millet VC process map

F a rm e r( s ) P ro c e s s o r( s ) R e ta il e r( s ) T ra d e r( s ) Prepare inputs Produce millet Process Buy grain At home At aggregation centre Transport grain Store Sell Grain Buy grain Sell to trader Processed Market Traders Sell to: Local market Regional market Institutional market Cross-border (Uganda) Source inputs: Previous crop Agrovet KSC Trader Market Cross-border (Uganda) Sell to: Local market School students Process into: Flour Yeast Staple food Snacks Local brew Sell to: Local market Regional market Institutional market Cross-border (Uganda) Preparation and consumption Food production Inputs / land

preparation Storage and processing Trading and marketing

Distribution and transport Consume as: Flour Yeast Staple food Snacks Local brew Transport grain Processor Eat, store or sell? Processed or grain? Sell in market or directly to traders? Which market? Processed how? Eat N N N N Key: Source of inputs Consumer market Finger millet form Out of research scope

Decision

(30)

VC structure

The VC(s) range from simple (smallholder farmers grow millet for consumption or to sell to locals in the neighbourhood market) to more complex (buyers purchase seeds from one market to sell at others, incorporating cross-border trade and institutional markets). The majority of participants were farmers who either sold their crop at local (nearby) markets or to traders from their homestead. Men predominantly prepare the land and transport millet, while women conduct the remaining activities of producing, harvesting, processing and marketing the crop. Women are also the household member responsible for cooking the finger millet for the family’s consumption. The gender division of these roles is further explored in Chapter 6.

The VC is both sequential and cyclical, because most farmers plant using seeds from a previous crop, and some farmers produce seeds for the sole purpose of planting and producing more seeds for sale (i.e. seed producers). Other sources of seeds are retailers, agrovets or institutions such as the Kenya Seed Company (KSC), whose certified seeds claim to offer a higher rate of germination but come at a higher cost. Using these costlier seeds depends on affordability and necessity. The production and marketing of millet uses resources such as land, seeds, oxen and plough, fertiliser, pesticides, tarpaulin for drying, and sacks for storage. Although millet is considered to be a hardy crop that requires fewer pesticides and fertilisers (Handschuch & Wollni, 2013), research participants all discussed using these products to enhance their yields. Access to these resources differs, depending on the actor’s level of capital and the plot location. Those with greater capital can hire people to help in more labour-intensive, or time-bound, activities such as harvesting.

Decision-making

A key decision is whether farmers keep millet for consumption, store it until it can fetch a higher market price, or sell it immediately. This decision is dependent on nutritional and financial need and flexibility. One male farmer said: “At times, I look at the market: if millet is in high demand, I call a buyer to collect the millet from my homestead. This helps me to reduce transportation costs and the market tax. In times of low demand, I keep my millet in the store and wait for prices to go high” (I26). Interviewees indicated that they sell on average 64% of their harvest, consume 24% and store 12%. However, previous studies have found that smallholder farmers are net food buyers (Gómez & Ricketts, 2013). If true, this indicates that the quantity and price of millet the farmer can attain is crucial to how nutritious a diet they achieve. Figure 5.1 highlights how other value chain actors make decisions about the way millet is processed and sold, such as whether to sell the millet in seed or processed form, whether to sell from their homestead or at a market, and which market to sell at. Decisions have nutritional impacts, either directly, or indirectly through income and purchasing power.

Information sharing

Information is transferred within actor groups, but not often between the stages of the VC or between different types of actors. Farmers share price information and knowledge of different markets with each other at group gatherings. Some farmers access price information by ascertaining the price at local

Gender relations in the millet value chain in western Kenya

Gender relations in the millet value

chain in western Kenya

Wai-Leng Chon

Student Number: 12532150

Email: wai_chon@hotmail.com

Date of Submission: Monday 17 August 2020

Institution: Graduate School of Social Sciences

Programme: MSc International Development Studies

Thesis Supervisor: Dr Verena Bitzer, Royal Tropical Institute (KIT)

Local Supervisor: Dr Osca Ingasia Ayuya, Egerton University

Abstract

Acknowledgements

Contents

List of Tables

List of Figures

List of Acronyms

1.

Introduction

2. Theoretical framework

3. Research Methodology

Nutrition-sensitive millet

value chain

Nutrition-sensitive

programmes

(e.g. EaTSANE)

Nutrition outcomes

Gender relations

Socio-cultural gender norms

4. Context

Chapters 5-8

Discussion of findings

5. Structure of millet agricultural value chain in western Kenya