The dagaa supply chain : a case study of Kisumu, Kenya

The dagaa supply chain

A case study of Kisumu, Kenya

Loes van der Plaat

MSc Thesis Environmental Geography (Human Geography)

Thesis supervision: Supervisor:

prof. dr. J.M. (Maarten) Bavinck

Faculty of Social and Behavioural Sciences University of Amsterdam

Nieuwe Achtergracht 166, Amsterdam J.M.Bavinck@uva.nl

Second reader: dr. J. (Joeri) Scholtens

Faculty of Social and Behavioural Sciences Nieuwe Achtergracht 166, Amsterdam J.Scholtens@uva.nl

Cover picture: 07:00hr in the morning. After a night of fishing, the vessels return at the landing site and traders obtain dagaa from fishermen. (Source: author)

Table of Figures

Figure 1: The fish supply chain ……… Figure 2: A pressure lamp ……… Figure 3: Early morning at the landing site ……….. Figure 4: Traders obtain dagaa from fishermen ……….. Figure 5: Whom traders buy dagaa from ………. Figure 6: Whom traders sell dagaa to ……….. Figure 7: Group membership ……….. Figure 8: Dagaa is spread out on top of nets to dry in the sun ……….. Figure 9: Traders at a market (1) ……… Figure 10: Traders at a market (2) ………..

Table of contents

Introduction ………1 Chapter 1: Theoretical framework ……… Introduction ……….. 1.1 (Food) Supply chains ……….. 1.1.1 The fish chain ……… 1.1.2 Chain network structures and information sharing

1.2 Fish chain governance ……… 1.2.1 External chain governance ……… 1.2.2 Internal chain governance ………. 1.3 Fisheries management and the challenge of legal pluralism ………. Research questions ……….. Outline ……… Chapter 2: Background ………... Introduction ……… 2.1 Geographical characteristics ……… 2.2 Lake Victoria fisheries ……….. 2.3 Political context: devolution ………. 2.4 Food security and the role of fish ………. 2.5 Kenya Marine and Fisheries Research Institute (KMFRI) ………... Conclusion ……….. Chapter 3: Research methodology ……….. Introduction ……… 3.1 Ontological position ……… 3.2 Research methods ……… 3.2.1 Semi-structured surveys ……… 3.2.2 Semi-structured in-depth interviews ………. 3.2.3 Non-participant observations ……… 3.2.4 Document analysis ……… 3.3 Sampling methods ……… 3.4 Ethical considerations ……….. 3.5 Limitations ……….

Conclusion ……… Chapter 4: The organization of the dagaa supply chain ………. Introduction ……… 4.1 Introducing the actors of the dagaa supply chain ……….. 4.2 Harvesting dagaa from Lake Victoria ……… 4.3 Processing, distribution, and marketing ……….. 4.4 Dagaa consumption patterns ……….. Conclusion ………. Chapter 5: Lake Victoria fisheries governance ……… Chapter 6: Conclusions and discussion ………. References ………. Appendices ……… Appendix A: List of all respondents ……….. Appendix B: Interview guide key informants ……….

Chapter 1: Theoretical framework

The aim of this research is to gain insight into the challenges that affect the supply chain of dagaa for human consumption. To achieve this objective, this chapter will provide a theoretical framework that forms the basis for the analysis in the next chapters. This chapter is organized as follows: in the first section, the supply chain and associated network structures will be discussed, as well as the relevance of information technology in trader networks for information sharing and money transfer. Networks are an important element to study, as they facilitate the flow of fish, and enable traders to get access to fish, employment, credit, markets, and information (Nunan et al., 2018). The section that follows discusses supply chain governance, reflecting the approach of Gereffi (2005), who makes a distinction between internal and external chain governance, and argues that value chain governance is intertwined with all the other chain elements, rather than being a separate element of the chain (Bavinck, 2018; Gereffi et al., 2005). ‘Governance’ is a central concept in supply chain analysis, as studying how a chain is governed allows for deeper understanding how the chain is controlled and coordinated, and which power relations exist in the relationships between different actors throughout the chain (Gereffi and Fernandez-Stark, 2011). After that, legal pluralism will be discussed, as this is considered a key challenge to governance (Bavinck & Gupta, 2014). In the concluding section, the research questions will be presented as well as an outline of the chapters this thesis consists of.

1.1 (Food) supply chains

Gereffi and Fernandez-Stark (2011) provide the following supply chain definition: “A chain represents the entire input-output process that brings a product or service from initial conception to the consumer’s hands” (2011: 5). Supply chains consist of vertically linked actors who collectively work to put products on a market. The scale of a supply chain can range from global to local; due to globalization, the production of a single good currently often spans multiple countries (Gereffi, 1994). However, as the dagaa supply chain in Kenya providing dagaa for human consumption only involves local actors, global supply chains will not be discussed.

of food chains can be affected by quality variation and perishability of products, as well as seasonal supply fluctuations. This has implications for the production and distribution management of food products (Bavinck et al., 2018; Trienekens et al., 2012).

The commercial outcome of a supply chain can be measured through acceptance of the end product by the consumer, which also implies that intermediate actors throughout the chain have accepted the product. According to Trienekens et al. (2012), acceptance of the end product in food supply chains depends mostly on the price, quality, and safety of the product, for which all actors throughout the chain are responsible. Consequently, improving product quality and safety and thereby creating extra value added, requires advanced chain governance mechanisms (idem). Accordingly, ‘value addition’ in food supply chains, which is understood as adding an extra feature to the product in order to increase its value, refers mainly refers to the safety and quality of the product. ‘Quality’ in this sense can either refer to the intrinsic features of the product itself, or to the extrinsic features related to the production process (e.g. the production of fair-trade products). The end-consumer’s willingness to pay determines the degree of value addition (Trienekens, 2011). The process of value addition is relevant to mention briefly as it plays a role in the competitiveness of food supply chains, however, as the dagaa supply chain only involves minimal value addition (value addition in dagaa trade is limited to sun-drying immediately after the fish is landed), it will not be discussed in more detail.

Coordination mechanisms in supply chains differ. Whereas Trienekens et al. (2018) argue that the power is usually in hands of large organizations downstream of the chain (such as retail or marketing and business organizations), and rarely lies with suppliers; Gereffi makes a distinction between demand-driven and supply-driven chains, in which power of control is reversed (Gereffi, 2001). Regarding supply chain coordination, business relationships in developing countries are often subject to uncertainties relating to inadequate physical and institutional infrastructures, uneven trade relationships, information asymmetry between chain actors, and adverse social and political circumstances. Therefore, developing country producers are navigating a landscape of risks and uncertainties. This requires more integrated forms of governance, meaning, more coordination and control (Trienekens, 2011). Trienekens (2011) and Gereffi (2005) state that this can be realized through standardization, monitoring, and improved communication systems. Supply chain governance will be discussed in section 1.2.

In order to be competitive, supply chain partners are also required to share information and coordinate, which demands integrated information systems (Trienekens et al., 2012). Section 1.1.2 pays attention to new developments in information systems.

1.1.1 The fish chain

The supply chain of aquatic products refers to the production, distribution, and consumption (Thorpe et al., 2005: 41). Thorpe et al. understand the ‘fish chain’ as a series of connected links, which are all influenced by and dependent on preceding and subsequent links in the chain (2005). As these authors argue, the fish chain is usually conceived of in a vertical sense, as it describes the movement of aquatic products from the ecosystem to the end-consumer, passing through several stages. However, they argue that this approach might not be the most comprehensive, and have created a model in which not only the flow of fish from the ecosystem to the end-consumer is displayed, but also the interactions that underpin the fish chain (idem). This model shows the movement of living resources from their natural environment to the human environment through processes of extraction, as well as the environmental consequences (such as changing ecosystems) of these processes. The model also addresses disruptions (e.g. pollution) affecting both the aquatic environment and the human environment, as these environments are interdependent; valuation of the extracted resource; the levels on which extraction and its consequences occur (international, regional, national, or local); and innovation and associated implications for both extraction and valuation of the resource (idem).

Explaining the human-induced ecological changes of Lake Victoria is not among the aims of the current research. However, such elements will be included in the analysis, as they play a central role in fish chain governance: a main objective of fisheries governance is to maintain a healthy ecosystem, whilst still economically benefiting from the resources obtained from it, in order to sustain livelihoods of those relying on the fisheries sector (Thorpe et al., 2005).

1.1.2 Chain network structures and information sharing

A fish chain consists of several connected links. This section will discuss chain network structures, as studying such structures is a means to gain insight in the nature of the connected links, and the power relations that exist within the chain. The focus will be on trader networks.

Chain-network structures have both a horizontal and a vertical dimension (Trienekens, 2011). Horizontal network structures refer to linkages between actors in the same chain link (e.g. fish processors or traders), which are mutually beneficial, as they are used to share information and improve the market position (Kaplinsky & Morris, 2000; Trienekens, 2011). Vertical chain network structures refer to the flow of products and services from the primary producer all the way to the end-consumer and the flow of money back, e.g. the relationship between wholesale and retail traders (Trienekens, 2011).

For actors in the fish supply chain, networks are essential in order to maintain a profitable business: networks determine the extent to which actors have access to resources and information, and are essential for accessing employment. It has also been shown that networks facilitate cooperation between actors governing resources (Crona & Bodin, 2006; Nunan et al., 2018). Within the context of Lake Victoria, networks have been shown to be conditional for the movement of fishermen between various landing sites, seeking more profitable fish stocks and better prices (Nunan et al., 2015).

In the broader literature on networks, social ties between people are often explained by homogeneity – whichever characteristics the members of the network have in common (McPherson et al., 2001). Within fishing communities, networks are often built around shared occupation, type of gear used for procurement, ethnicity, gender, and location (Nunan et al., 2018). Access to credit appears to be the most important benefit fish trades obtain from their social networks, as well as information sharing (idem). (Maarten: how about kinship?)

Mobile phones currently play a crucial part in trader networks, facilitating both communication and payments. Compared to the pre-cell phone era, new telecommunications technology has caused a reduction in information asymmetries, as well as in transportation and transaction costs (Overa, 2006). The transportation of products from the producer to the consumer requires cooperation, coordination, and information exchange between the various actors throughout the supply chain. In sub-Saharan Africa, transportation and transaction costs are two key factors determining traders’ margin of profit and, finally, also the availability and price of the product for the end-consumer (Overa, 2006; Minten & Kyle, 1999). According to Overa (2006), transportation costs (which include both the transportation of products from producer to consumer and the transportation of people for the purpose of information sharing) are linked to transaction costs, as information is expensive and “asymmetrically held by the parties of exchange” (North, 1995; 18). The costs thus increase when decent transportation means are lacking, roads are bad, or traders lack access to telecommunications. Traders therefore heavily depend on each other, which involves the risk

of making business arrangements with someone who turns out to be unreliable (Overa, 2006). Building a network of trustworthy individuals is therefore essential for traders to sustain their business. In recent decades, cell phones have played an important role in building and sustaining those networks, as this relatively new means of communication provides tools which ease networking and the exchange of information among traders (idem). Cell phones enable traders to access information about current prices, and communicate faster and more frequently with business partners. The rise of the use of cell phones thus caused a reduction in information asymmetries, and a rise in the predictability of transactions (idem).

Additionally, cell phones are currently to an increasing extent being used to transfer money, take out small loans, and save. The mobile money transfer system widely used in Kenya is M-Pesa, which is operated by Kenya’s largest mobile phone provider Safaricom. This relatively new cellular phone function has caused a shift in the organization of economic activity: within eight months after its emergence in 2007, over a million Kenyans had registered to use the financial service (Mbiti & Weil, 2015). Currently, M-Pesa has 30 million active subscribers (standardmedia.co.ke), and is the leading form of money transfer in Kenya (Mbiti & Weil, 2015). E-money services such as M-Pesa are safe and inexpensive, and therefore accessible to all groups within society (Gumba, 2018). Mobile money can thus be considered as a tool for financial inclusion of the poor, providing opportunities in the financial market to which access was previously limited to middle- and higher-income groups (idem). For fish trader networks, the mobile money transfer service has resulted in a proverbial shrinking of time and space, increasing efficiency, and decreasing risks of losing money to, for example, untrustworthy middlemen responsible for money transfer from one place to another.

1.2 Fish chain governance

The previous section discussed the fish chain and its associated chain network structures, and highlighted the role of new telecommunications technology in trader networks. This section will discuss chain governance, understood as the variety of chain coordination mechanisms.

Bavinck and co-authors (2018) argue that within literature on fisheries governance, there is a tendency of separating the “system-to-be-governed” (the fish chain) from the “governing system” (regulation, laws, etc.) (Bavinck et al., 2018: 277). They argue that as a result of this approach, perhaps governance activities within the supply chain have not received sufficient attention. The approach adopted in this research therefore builds on the

that determine how financial, material and human resources are allocated and flow within a chain” (1994: 97), and considers governance as coordination mechanisms that do not exclusively exist external to the value chain, but are embedded throughout all value chain interactions (Gereffi, 2005). In this respect, governance does not only affect the value chain externally through regulations and policies, but also through integrated governance forms within the value chain (‘intra-chain governance relations’) (Gereffi et al., 2005). This means that the chain network structures discussed in the previous section, are also considered as a form of governance. According to Bavinck (2018), this approach to governance corresponds with the legal pluralist approach of Turner (2016), who argues for a distinction between “chain normativity” (the whole of socio-legal aspects shaping the chain) and the “plural legal environment” in which the chain is rooted (Bavinck, 2018: 276). Section 1.3 will discuss this in more detail. The following sub-sections will discuss external and internal chain governance.

1.2.1 External chain governance

Government organizations, NGO’s, etc. Regulations, policies, laws [add later]. Internal chain governance

(Maarten: Je zou misschien Gereffi’s onderscheid tussen demand- and supply-driven chains willen aanhalen? Interessant ook in het geval van Kenya.)

Gereffi et al. (2005) developed a framework in which five types of chain governance are distinguished: 1) market governance, in which price is the deciding factor and the costs of switching to new business partners are low, but market relationships are not essentially temporary and can persist over time; 2) modular governance, in which suppliers typically provide turn-key services (this form of governance is not as relevant as the other forms in food-chains, as it applies mostly to industrial products (Bavinck, 2018)); 3) relational value chains, which are characterized by complex interactions between value chain actors, often creating mutual dependencies; 4) captive value chains, which are characterized by networks in which small suppliers transactionally depend on larger buyers and are subject to price fluctuations. Monitoring and control within these types of networks in often in the hands of lead firms; and 5) hierarchy, in which vertical integration is a key feature (Gereffi et al., 2005). The level of ‘explicit coordination’ and power asymmetry between supplier and buyer ranges in order of listing from low (market) to high (hierarchy) (idem).

Which type of governance emerges under which conditions depends on three factors: 1) the complexity of information and knowledge transfer needed to support inter-firm transactions; 2) the degree of codifiability of this information and knowledge, which determines the efficiency of transactions; and 3) the capability of the supplier regarding the requirements of the specific transaction (idem). The five types of governance combined with the three variables form the framework provided by Gereffi et al. (2005), which is displayed in table 1:

Table 1 “Key determinants of global value chain governance” (Gereffi et al., 2005: 87) Governance type Complexity of transactions Ability to codify transactions Capabilities in the supply-base Degree of explicit coordination and power asymmetry

Market Low High High Low

Modular High High High

Relational High Low High

Captive High High Low

Hierarchy High Low Low High

Note. Reprinted from “The Governance of Global Value Chains”, by Gereffi, G., Humphrey, J., & Sturgeon, T., 2005, Review of international political economy, 12(1), p. 78-104.

The hypothesis of this research is that market governance is the governance form that arises in dagaa trade for human consumption in Kenya; after all, market governance arises in the event of low asset specificity due to easily codified transactions, uncomplicated product specifications, and high supplier capabilities in the sense that suppliers need only little input from buyers to provide the products (Gereffi et al., 2005). Whether the dagaa supply chain is predominantly self-governing or mainly governed through coordination mechanisms external to the fish chain, will be analyzed in chapter [5].

1.3 Fisheries management and the challenge of legal pluralism 2 types: resource: pluralism between countries

and market: organizational structure within Kenya

In this section, the issue of legal pluralism will be discussed, as this phenomenon often forms a challenge in natural resource management (Bavinck et al., 2018). According to Bavinck &

Gupta (2014), legal pluralism is important to consider in the context of fish chains, as it can have implications for the quality of governance. Legal pluralism is understood as a condition in which multiple (formal and informal) legal systems apply to the same situation, which often occurs in natural resource management (Bavinck et al., 2018). Historically, oceans and lakes have been a source of crucial resources for societies all over the world. Scholars in the area of legal pluralism agree that sets of laws implemented to manage the resources from lakes and oceans are not exclusively formulated by governments (formal law), but are also rooted in custom and culture (informal law) (Bavinck & Gupta, 2014; Griffiths 1986; Merry, 1988; Twining, 2010). These various sets of rules are supported by different levels of power and authority, and imply different outcomes for the situations (or groups of people) to which they apply (Bavinck & Gupta, 2014). In the field of fisheries, for example, Bavinck & Gupta (2014) distinguish several governing actors: respectively organizations of fishermen at the beach who financially depend on fish, government actors formulating rules, NGO’s committed to conservation and sustainability of the fishing grounds, and market actors (2014).

Whereas most developed countries have reached a certain level of consistency in their law systems, this consistency is often still limited in developing countries. However, recently, new types of legal pluralism are arising, also in developing countries. This is due to the shift of power from centralized governance (national governments) to more decentralized forms of governance, in which a variety of state as well as non-state actors is involved in the governing process (Bavinck & Gupta, 2014).

Legal pluralism is argued to affect governance in various ways (ibid.: page). First, it causes incoherence, as mentioned above. Second, it causes a continuing struggle over power: as legal systems substantially differ, they accordingly have different perspectives on the content and definition of property rights, resulting in different legal systems assigning rights to different groups regarding the accessing, extracting, and selling of resources. Third, legal pluralism leads to competition between various rule systems. Consequently, as the most powerful actors often successfully force their rule system (representing their interests) upon the other actors in the field, this raises social justice issues, as it is often the rights of local communities that are suppressed. Lastly, as different legal systems point human behavior in different directions, thereby potentially causing social conflict, a state of legal pluralism can hinder the effectiveness of institutional frameworks. As Bavinck & Gupta (2014: page) point out: conflicts over resources obtained from oceans and lakes often form constraints for sustainable governance.

Research questions

The first section of this chapter discussed the fish supply chain and related network structures, and the section that followed discussed supply chain governance and legal pluralism, which is argued to form a severe challenge to governance. Together, these elements provide a theoretical framework for the analysis that will be carried out in the following chapters. The elements are displayed in figure 1, which provides a visual representation of the fish chain and the governance mechanisms underpinning the chain, both internally, and externally.

figure 1: The fish supply chain

Figure 1 visualizes the dagaa supply chain: the fish moves from the ecosystem to the end-consumer, passing through different steps (harvesting, processing, distribution, and marketing) (Thorpe et al., 2005). The process is governed through external governance, understood as laws, policies and regulations, as well as internal governance, which is understood as the coordination mechanisms located within the supply chain, such as price agreements and network structures between chain actors.

Maarten: Can you bring in legal pluralism? And maybe information technology too?

The following research questions will guide the analysis:

Main question: Lake Victoria ecosystem Harvesting Processing, distribution, marketing External governance Internal governance

Which social, political, economic and environmental factors affect the supply chain of dagaa for human consumption?

Sub questions:

- How is the supply chain of dagaa for human consumption organized? - How is Kenyan Lake Victoria fisheries governance currently organized? Outline (nog even kijken naar formulering)

The thesis is organized as follows: chapter 2 provides background and context information about the research location. The chapter that follows describes the research methodologies. Chapters 4 and 5 will discuss the findings of the study, organized according to the sub-questions. Chapter 4 describes the organization of the dagaa supply chain for human consumption. The chapter discusses the characteristics of the actors employed in the dagaa business; the means of harvesting dagaa, the ecological changes in the Lake affecting the availability of dagaa in the Lake; processing, distribution and marketing methods, and finally, consumer patterns and the changing preferences of consumer groups. The fifth chapter discusses the governance mechanisms of the chain, and challenges affecting governance.

Chapter 2: Background

This chapter introduces the context in which dagaa traders operate: the geographical characteristics of Kenya and the Lake Victoria region as well as the political landscape, with a particular focus on the newly established political administration: the devolved government. The first section provides information about the geographical characteristics, followed by sections on Lake Victoria fisheries, the political context and the devolution to county governments, and the (potential) role of fish regarding food security in the region. The chapter concludes with a brief description of the Kenya Marine and Fisheries Research Institute (KMFRI), which is connected to the project ‘SmallFishFood’. The institute introduced me to the fieldwork location, and facilitated guidance and assistance throughout the fieldwork.

2.1 Geographical characteristics

Kenya, located in East-Africa, is boarded by Sudan, Ethiopia, Somalia, Uganda, Tanzania, and the East Indian Ocean, and covers 580,370 km2_{. The population of Kenya consists of} approximately 50,039 million people (FAO, 2015).

Fieldwork has been carried out in and around Kisumu, located on the shores of Lake Victoria in western Kenya. Kisumu is Kenya’s third largest city, following after Nairobi and Mombasa. Kisumu is where the bulk of fish landed from Lake Victoria is traded, processed, and either transported to Kenya’s other major cities Nairobi or Mombasa, or exported (Export Processing Zones Authority, 2005).

The fisheries and aquaculture sector contributed 0.54% to Kenya’s GDP in 2015, and has expectedly grown in recent years (FAO, 2015). The fisheries and aquaculture sector consists of four sub-sectors: inland capture fisheries, marine capture fisheries, inland aquaculture and marine aquaculture. These four sub-factors all play a substantial role in the national economy: at least 80.000 people are directly employed in any of these sub-factors (e.g. fisherfolk and fish farmers), and the sector indirectly provides employment to approximately 2.3 million Kenyans, who are involved in the processing and trade of fish (idem).

2.2 Lake Victoria fisheries

Lake Victoria is the largest freshwater body in Africa. It covers 68,800 km2 and is shared between Tanzania (51%), Uganda (43%) and Kenya (6%). Tanzania, Uganda and Kenya share the responsibilities of managing the lake and its resources. Nowadays, the Lake Victoria species that are commercially most important are Nile perch (Lates niloticus), native cyprinid (Rastrineobola argentea), and introduced tilapia (Oreochromis niloticus) (Isaacs, 2016; Njiru et al., 2008).

Traditionally, fishing happened primarily by communities around the lake. However, the lake has become commercialized over a century ago, driven by factors such as colonialism and the development of modern technology, infrastructure, and foreign investments. These developments have led to ecological changes and declined fish stocks, due to the use of more destructive fishing methods such as trawling (Abila, 2003), and large-scale population movement to the lake regions. In the 1950’s, the Nile perch and Nile tilapia were introduced, mainly with the objective to relieve the commercial pressure on fish stocks. The introduction of the Nile perch transformed the Lake Victoria fisheries into a national and international capital investment industry, and has had severe ecological impacts. It resulted in a major reduction of biodiversity: 300 indigenous species became nearly extinct (Abila, 2003; Isaacs, 2016; Njiru et al., 2008). However, this species introduction seems to have left the small pelagics stocks unchanged, which is significant, as the fishmeal industry targets small pelagics. According to Isaacs (2016), the small pelagics stocks have even steadily increased over the past four decades in Lake Victoria, which implies that there is still room for expansion in this sub-sector. When ‘small pelagics’ are mentioned in the context of Lake Victoria, this usually refers to the small pelagic ‘silver cyprinid’ (Rastrineobola agrentea), locally known as dagaa. The silver cyprinid is also known by a different local name, omena,

which originates in the Luo language, widely spoken in western Kenya. In this research, the species will be referred to as dagaa.

2.3 Political context: devolution

Kenya has recently (2010) adopted devolution as a system of decentralization, shifting the power away from the national government (Khaunya et al., 2015). In post-independence Kenya, the political structure was highly centralized: political and economic power was centered in the hands of a few, resulting in an unequitable distribution of national resources and access to social services, governance lacked accountability and transparency, and communities only minimally participated in local development (Khaunya et al., 2015; The World Bank, 2012). The Constitution of 2010 changed the governance structure from one centralized political unit to 47 devolved sub-units, which are known as the county governments (Khaunya et al., 2015). The new Constitution was adopted after nearly a decade of failed transformation attempts. Objectives of the reform of the Kenyan state were focused on creating a more inclusive state, through new accountable and transparent institutions, a more inclusive approach to development, and a strong focus on a transparent and equitable distribution of social services to the citizens (idem). Kenya is a highly diverse country, with different needs among different ethnic groups and regions. A system of devolution would create the opportunity to more effectively address the diversity of local needs, making use of local knowledge (idem). Also, the new devolution structure would allow for the effective handling of the development challenges the centralized government had faced ever since the independence in 1963 (The World Bank, 2012).

The governance reform is very ambitious, and opinions are divergent on how fast the transition of responsibilities from national to county governments should have been implemented. According to Khaunya et al. (2015), indeed some of the areas of attention have been improved: for example, accountability and community participation increased, and resources are now more equitable distributed. Despite these achievements however, new challenges have arisen that stand in the way of these seemingly positive realizations. Examples are a lack of political goodwill at county governments, inadequate financing, the inability at county governments to perform some devolved functions, ‘devolved’ corruption and bureaucracy, etc. (idem).

The fairly new devolution structure is relevant for this research as it has far reaching implications for Kenyan fisheries governance. According to Horace Owiti, research scientist

introduced new complications in fisheries management, as more interests became involved, and the distribution of responsibilities became, at times, unclear. These developments and their implications will be discussed in detail in chapter 5.

2.4 Food security and the role of fish

With a gross national income per capita of USD 1.790 in 2018, Kenya is classified as a lower-middle-income country. Kenya is one of the Low-Income Food-Deficit Countries: in the early 2010’s, Kenya’s average daily protein intake (which includes animal protein) per capita was only 89% of the African average and 76% of the global average (FAO, 2018). In the mid-2010’s, one out of three Kenyan citizens was living below the food poverty line, meaning that one third of the population was lacking the financial means to meet the daily caloric requirements; 24% of the population was undernourished, and 26% of the children under 5 years old were hindered in their growth due to malnutrition (idem). Undernourishment and undernutrition challenging food security is thus a theme that requires attention in Kenya.

Fish could play an important role in addressing the food security challenges Kenya faces: fish (including other types of seafood) is a source of high-quality animal protein, and a rich source of macro- and micronutrients (FAO, 2018). Small pelagics such a dagaa, a type of forage fish, are generally cheap and play an important role in providing food security to the poor, not only in Kenya or elsewhere in Africa, but everywhere in the world (Tacon & Metian, 2009). Similar to the majority of African countries, Kenya has a much lower average animal protein intake than the global average. The fish share of the total animal protein intake however, is larger in many African countries (as well as Africa as a whole) than the world average. That said, Kenya is an exception. Whereas elsewhere in Africa the fish share of animal protein was approximately 20% in the early 2010’s, the percentage was 8% in Kenya (compared to milk and other dairy products that accounted for approximately 50% and meat that accounted for 36% of Kenya’s total animal protein intake) (FAO, 2018). The fish share of animal protein had declined since the early 1990’s, when it was approximately 10% (idem). The declined fish share between the early 1990’s and the early 2010’s reflected in the drop of per capita fish consumption, as a result of a decline in Kenya’s capture fisheries production (FAO, 2018). In the 1990’s, aquaculture only contributed minimally to the national fish production. However, due to government efforts, starting in the 2000’s, fish farms have become a substantial part of the fisheries sector. In the past decade, Kenyan nutrition education programs and projects have assigned an important role to fish as a vital

source for nutrients, and the benefits of a diet including fish is receiving increasing recognition (FAO, 2018). Kenya’s fish production from capture fisheries has been the main factor affecting the country’s fish consumption. The downward trend of capture fisheries between the 1990’s and the 2010’s indicates that capture fisheries are not able to maintain Kenya’s (already very low) fish consumption. Also, relying on imported fish to provide vital nutrients to the growing population has been proven not to be a workable solution: importing cheap fish products, such as frozen tilapia, results in competitive pressures on domestic fish production industries. It is thus currently widely acknowledged that, to be able to provide food security to the growing population (the population is expected to reach 67 million people by 2030 (United Nations, 2017)), an important solution lies in the rise of fish consumption, realized through the rapid development of aquaculture (FAO, 2018). Supported by government policies and large public investments, Kenyan aquaculture has grown 20% annually between the early 2000’s and 2017 (idem).

The growth of aquaculture is relevant for this research, as farmed fish often indirectly feed on small pelagics due to the use of fishmeal on fish farms (Isaacs, 2016). Globally, fish landings nowadays contain more small pelagics than any other type of fish (idem). Previous to the fieldwork in Kenya, it was expected that dagaa targeted to produce fishmeal, would be the same fish targeted for human consumption, therefore compromising food security. However, the research has shown that this is often not the case. Which factors determine whether dagaa is used for human consumption or for the production of animal feed will be discussed in chapter 4. However, dagaa can be considered as an important source of protein especially for Kenyans with a lower income, due to the combination of its low price and high nutritional value. In addition, dagaa is an important source of livelihood in terms of employment and income.

2.5 Kenya Marine and Fisheries Research Institute (KMFRI)

Chapter 3: Research methodology

(opmerkingen nog verwerken)

Introduction

As this study focuses on the challenges affecting dagaa trade for human consumption on both the level of traders and the level of stakeholders and policy makers, it requires research methods that allow gaining in-depth knowledge about these challenges from different perspectives. The following section describes the ontological position adopted in the study, followed by sections on research methods and techniques, a section on sampling methods, and a section on potential ethical considerations concerning the data collection in and around Kisumu, Kenya. The concluding section reflects on the limitations of the study in Kisumu. 3.1 Ontological position

The study requires an ontological position in which both the actor (a trader, a stakeholder, a policy maker, etc.) as an active agent with a view on the reality is acknowledged, as well as the broader society and the political landscape in which the actor operates. Accordingly, this research adapts a critical realist ontology. This ontological stance entails the assumption that there is an independent reality, without committing the researcher to the view that absolute knowledge of the way this presumed independent reality works is possible (Scott, 2005). In other words, the critical realist paradigm differentiates the ‘essence of things from their

appearance’ (Losch, 2009; p.86). Critical realism can be considered the ‘most encompassing’ (Losch, 2009; p.86) ontological position, as it entails a belief in an objective reality, and yet does not reject the constructivist perspective that social realities are constantly under construction through social actors. Critical realism thus provides some sort of intersection between a constructivist epistemology and a realist ontology, acknowledging the existence of an external reality, whilst not denying the subjectivity of the human vision and the belief that any attempts at understanding and explaining the world are inevitably fallible (Hassan, 2003; Scott, 2005).

In line with the foregoing, this empirical study aims to understand and describe a specific reality through the knowledge and experiences of the respondents, thereby accepting a certain level of subjectivity in the findings. In accordance to the critical realist perspective that it is rather difficult, if not impossible, to gain understanding about the absolute reality, the study explores the perspectives of dagaa traders on the ground and various actors on the level of policy-making, and does not aim to step beyond recording these perspectives to fully understand, and make statements about, the external reality which exists beyond the social reality of the respondents.

3.2 Research methods

In an attempt to gain in-depth understanding of the challenges in dagaa trade in and around Kisumu, a combination of several qualitative and quantitative research methods has been used. A quantitative strategy entails a deductive approach to the relationship between research and theory, meaning that the focus is on the testing of theory. Within quantitative strategies, the social reality is considered as an external, objective reality. Quantitative strategies can help showing patterns and allow for comparison. Qualitative strategies on the other hand, entail an inductive approach to the relationship between theory and research, and place the accent on the generation of theories. Within qualitative strategies, the social reality is seen as a construct that is constantly changing and subject to an individual’s creation (Bryman, 2012). Qualitative methods are helpful when exploring complex relations, and can help to better understand and explain the results obtained from a quantitative approach (Creswell & Plano Clark, 2018).

Qualitative and quantitative research strategies thus embody different views of social reality, and place the accent on different aspects of the relationship between theory and research. In the empirical study in Kisumu, both qualitative and quantitative methods have

the observed phenomena. This triangulation approach also allows for comparison between qualitative and quantitative results, in order to determine whether there is convergence, differences, or maybe some combination (idem).

In total, 75 semi-structured surveys were conducted with traders (wholesale and retail) and processors selling dagaa either on markets in the city of Kisumu or on landing sites located on the shores of Lake Victoria in the surrounding counties, and seven in-depth interviews have been conducted with key informants representing relevant institutions in Lake Victoria fisheries management. Also, observations were conducted at markets and landing sites. Lastly, two documents have been analyzed: the Frame Survey report of 2016, and the Kenya Fisheries Act of 2007 (the last updated version of the Act). These research methods are listed in table 2 and discussed in the following sections.

Table 2: Research methods Traders and processors Key informants Reports

 Semi-structured surveys  Semi-structured in-depth interviews

 Document analysis  Informal interviews and

fieldwork notes  Observations

3.2.1 Semi-structured surveys

Conducting the surveys with traders at markets and landing sites started mid-April, after getting familiar at the Kenya Marine and Fisheries Research Institute (KMFRI). In total, 75 traders participated in the semi-structured surveys, which were conducted at market stalls of different sizes where the traders were selling the dagaa, and at the landing sites where the traders were processing, packing, and selling the fish. Each survey took approximately 45 minutes, as the survey consisted of both open-ended and closed-ended questions. As the traders were working, we often took a small break for them to sell fish to a customer, and continued afterward. The majority of the traders did not speak English: surveys were

therefore conducted with the help of a research assistant for translation purposes. Respondents were never left to complete the survey on their own. Walking through the different questions together with the respondent allowed us to provide additional explanation when it was required, and to ask additional questions when we needed extra information on certain topics. The possibility of collecting additional data is an important advantage of a semi-structured survey, compared to a self-completion questionnaire (Bryman, 2012).

3.2.2 Semi-structured in-depth interviews

In-depth interviews have been conducted with seven representatives from five institutions playing a role in Lake Victoria fisheries management, in order to gain insight in the legal framework of governance and management of Lake Victoria, as well as the political landscape in which actors operate. Semi-structured in-depth interviews have been a useful method as this interview technique allows for some latitude to ask further questions in response to significant replies from the interviewee (Bryman, 2012).

3.2.3 Non-participant observations

In order to complement the semi-structured surveys and in-depth interviews and to triangulate the findings, also observations were conducted at markets and landing sites. Observations were non-participant, meaning that I did not take active part in the situations I was observing. During the observations at markets, I looked at what the traders were doing, who they were working with, what types of measurement units they used, what their stalls looked like, what the market place looked like, where they were keeping their dagaa stocks, etcetera. During observations at landing sites I looked at how the dagaa was processed, whether caradina or dagaa dominated the catches drying in the sun, what the traders were doing, how the traders obtained fish from the fishermen in the morning, etcetera. Observations helped me to gain a greater understanding of the daily activities of dagaa traders and processors, their interactions, their work hours, and the challenges they encounter doing their work.

3.2.4 Document analysis

Additional to the in-depth interviews with representatives of institutions, two reports have been analyzed: the Frame Survey report of 2016, and the Kenya Fisheries Act of 2007. Analyzing these documents has helped to gain knowledge about topics such as illegalities in fishing and trading, international and national quality standards for dagaa, international

provided relevant figures and statistics. I am however aware that written documents often reflect the particular point of view that the author wishes to get across.

3.3 Sampling methods

The participants in the empirical study in Kisumu were selected through non-random sampling. Non-random sampling entails gathering samples in such a way that it is relevant to the research question being posed (Bryman, 2012). A disadvantage of non-random sampling compared to random sampling, is that it does not give all the individuals in the population equal chances of being selected. However, representativeness of the population is not as big a problem in a – mainly – qualitative research framework as it would be in a quantitative research framework, wherein the aim is generalization (idem).

The participants were mainly selected using a purposive sampling method, which is a type of non-random sampling. This means of sampling helped us to ensure that the sample included wholesale traders as well as retail traders, both equally distributed over the markets within Kisumu and equally distributed within the markets, and dagaa traders and processors at the landing sites in the region around Kisumu, as dagaa traders and processors do not form a homogeneous group of people. The consequence of the choice for this type of sampling is that my personal judgement as well as the judgement of my research assistant affected the selection process. Also, to a lesser extent, snowball sampling was used. This type of sampling can be considered as a form of convenience sampling. In snowball sampling, the researcher initially makes contact with a small group of people who are relevant to the research question that is being posed, and then establishes contact with other people making use of the networks of the initial contacts (Bryman, 2012).

In total, 75 dagaa traders and processors participated in the empirical study. Their ages ranged from 18 to 61+, most of the respondents falling into the category of 31-45 years of age. The great majority of traders were female: only four out of 75 traders were male. The seven representatives of institutions were employed in various institutions in Kisumu.

3.4 Ethical considerations

Ethical issues need to be considered in social sciences, as they relate directly to the integrity of the research (Bryman, 2012). Diener and Crandall (1978) have broken down the great variety in possible ethical issues into four main themes: harm to participants, lack of informed consent, invasion of privacy, and deception.

The participants I interviewed rely on jobs provided by the fisheries sector in the Lake Victoria region. In recent decades, as the fisheries sector became increasingly commercialized, more effective (but also more destructive) fishing techniques have been developed (Abila, 2003; Tacon & Metian, 2009). As the issue of overfishing and unsustainable fishing methods has globally gained growing attention, controversy around fishing methods increased. In combination with declined fish stocks, this growing controversy possibly creates tension between actors relying on the Lake’s resources. Conducting research in the field of fisheries thus needs to be done with a certain level of awareness and sensitivity. Fundamental to ethical research is that participants of the study suffer no harm as a result of the research process and outcome of the study. To ensure this, first and foremost, respondents have been well informed, and consent has been gained from participants and interviewees. Also, people were given the choice to participate in the research process or not, and I have respected people’s decision not to participate. After indicating to be willing to participate in the survey, traders were given a soda, to compensate for giving up the time in which they could have earned money. This was an alternative to a financial imbursement, which would not have been ethically appropriate. All pictures were taken with consent. Furthermore, I have made my role as a researcher clear to those participating in the study, and I have explained the nature of the research process in order not to create false hopes and expectations. When conducting in-depth interviews, I have always asked permission before audiotaping the interview. Collected data has been confidentially treated, as information obtained could be sensitive. Lastly, data has been analyzed anonymously (unless the participant explicitly indicated he or she wanted to be mentioned), which has been made possible by the thematically coding of interview results.

As a final point of attention, my own positionality as a western female needs to be considered, as it has possible influenced and biased my understanding of the phenomena. During the fieldwork in Kisumu, I became increasingly aware of the importance of taking a reflective position in the research. In qualitative research, the researcher inevitably influences the collection, selection, and interpretation of data. Reflexivity is the process of reflecting on one’s behavior as a researcher, through a self-aware analysis of the research process (Finlay, 2002). Social reality cannot exist separated from the consciousness and language of people – including the respondents and the researcher – and separated from the context (Alvesson and Skoldberg, 2009). During the fieldwork, I have therefore continually attempted to take a reflective stance, by reflecting on how my presence and assumptions were potentially

beliefs and preconceptions. However, my position as a Western female is an ethical aspect that needs to be mentioned. During the fieldwork at markets and landing sites, I have constantly been aware of my own presence as a tall, white woman. However, through the presence and help of my research assistant, I have been able to gain trust from the people participating in the study.

3.5 Limitations

One of the main limitations I have faced while conducting surveys at markets and landing sites, was the language barrier. Along the eastern shores of Lake Victoria in Kenya, the main language spoken among dagaa traders is the Luo language, also known as Dholuo. The Swahili language, also known as Kiswahili, is widely spoken as well, all over eastern Africa. It was not possible for me to learn these languages in the short amount of time I was able to spend in the field, and I therefore always worked with a research assistant, who translated for me. I have worked with several assistants, who were assigned to me by KMFRI. All assistants I have worked with were very familiar with the region, the culture, the religion, the practices, and the languages, and were enthusiastic, friendly, patient, and persistent. Without the assistance from KMFRI, I would not have been able to visit all the places I have visited during the fieldwork. However, interacting with traders through a translator, has inevitably directed some of the answers from respondents, due to the subjective nature of interpretation.

The second limitation I encountered during the fieldwork relates to the fishmeal industry, which plays a rather important role in the dagaa trade in the Lake Victoria region. As mentioned in the previous chapter, prior to the fieldwork, it was expected that in Kenya, similar to Tanzania, much of the dagaa catches would be directed towards processing factories producing fishmeal. This was found not to be the case, however, it would have been interesting to interview an agent working on behalf of a processing company, to gather information coming from a different angle. Unfortunately, I have not been able to do so. Establishing contact with these agents is difficult, and time was limited.

The rather short period of time that was available for fieldwork was the last, and probably most influential, limitation I have encountered. Initially I would spend 6 weeks in the field, which is a relatively short period to collect an adequate amount of data. However, my time to collect data was limited even more, due to long-lasting, bureaucratic difficulties in Tanzania, where the fieldwork would take place initially. As a result of this, I have spent a month in Tanzania attempting to acquire the documents needed to start the fieldwork in Mwanza, also located on the shore of Lake Victoria, before changing the plan and crossing the border to

Kenya. The change of research location required far-reaching adjustments in the specific topic of the study, the research design, and the research questions, which has naturally been time-consuming. Together, these events have resulted in a short fieldwork period in Kisumu, covering only four weeks. However, through the support of KMFRI and by deciding to join our efforts and carry out data collection collectively, my fellow student Lola van der Voort and I have been able to collect an adequate amount of data within the four weeks available for fieldwork.

Conclusion

This chapter described the methodology that was used in the study in Kisumu, as well as the ontological position adapted in the study, the sampling methods, the ethical considerations and my own positionality as a Western female, and the limitations of the empirical study. The following 2 chapters describe the results of the study.

Chapter 4: The organization of the dagaa supply chain

In the following two chapters the findings of the empirical study that was conducted in Kisumu, Kenya, will be presented. The theoretical framework was used to structure and analyze the findings, and to present coherent and contextual information on the way the dagaa chain is organized and governed, internally as well as externally.

This chapter discusses the organization of the dagaa supply chain, and identifies the challenges affecting the chain. The first section introduces the people employed in dagaa fishing and trade and their socioeconomic characteristics, and discusses the highly gendered division of labor throughout the chain. The way that the three sections following after that are organized corresponds with the several connected fish chain links identified by Thorpe et al. (2005): dagaa moves from the ecosystem to the end-consumer, passing various stages. Accordingly, section 4.2 describes dagaa harvesting methods, followed by section 4.3 which describes dagaa processing methods, distribution, and marketing. The final section of this

chapter focusses on consumption patterns affecting the demand for dagaa, which entails consumer preferences and the perception of dagaa among consumer groups.

4.1 Introducing the actors of the dagaa supply chain

Traditionally, most people employed in dagaa trade and fishing are not highly educated and have a fairly low economic status. The survey results pointed out that the great majority (60%) of traders and/or processors completed primary education, whilst 26,7% completed secondary education, and 10,7% was not educated at all. The low economic status of dagaa operators is related to the perception of dagaa, which is historically considered ‘poor men’s food’, as the key informants emphasized. Horace Owiti, [research scientist at KMFRI], provided the following explanation: first of all, the capital requirements for dagaa trading are very low. Second, the cost of dagaa is relatively low, and the fish is divisible, making it possible to sell and buy it in small quantities. Third, sun-drying does not require expensive facilities and the beaches are free to use, and after the fish has dried, it can be stored and used up until a year later. These features make dagaa business accessible to everybody, regardless people’s economic status. Christopher Aura, [functie at KMFRI], emphasized that people employed in dagaa trade were until recently often marginalized groups, such as relatively old people and people with no alternative means of livelihood. Positions in the business were therefore predominantly occupied by widows and single women.

According to Owiti, the high prevalence rates of HIV/AIDS in the late 1990’s and early 2000’s in the Kisumu region, which were among the highest in Africa, are related to this in various ways, as high numbers of men died and widows were left to find new sources of livelihood. This was problematic from a cultural perspective: until a change in the constitution in 2011, it was impossible for a married woman to inherit land from her parents. In the event that she would be widowed whilst still relatively young, she would be left with very few options: going back to her own family would be a disgrace, and would be even worse with children involved, as they would be considered competitors. Staying at the family of her former husband would however be problematic as well: culturally, she should informally be ‘inherited’ as a kind of mistress by a male relative of her husband, which the wife of the male relative would not approve of, as it is ambiguous whether polygamy is a culturally accepted practice. This would leave the woman in a vulnerable position. The culture of inheritance contributed to the rapid spreading of HIV, as the widows were often HIV positive as well. In the current constitution, which was established in 2011, this ‘cultural mismatch’ is partly corrected, as it is now possible for women to inherit land from their

families. However, according to Owiti, operationalizing these newly established women’s rights is still a challenge. There is a second reason why HIV/AIDS is important to mention in relation to the dagaa chain: as the majority of people entering into dagaa trade are widows, single mothers, and divorced or separated women, this group of people is vulnerable to sexual exploitation by fishermen. The widely spread practice of jaboya, in which sex is traded for fish, also contributed to the spreading of HIV. In recent years, the practice of inheritance has declined, as awareness about the risks has risen among people involved in dagaa fishing and trade.

According to Owiti and Aura, the characteristics of the group of people employed in dagaa business are currently slowly changing, as people are realizing that dagaa trade provides opportunities to earn money: dagaa currently dominates the catches from Lake Victoria, and as Kenyan aquaculture grows, dagaa fishing has started to attract entrepreneurs. Another reason directing people into getting engaged in dagaa business is the growing pressure of maintaining living conditions, as the cost of living is increasing, due to a combination of population growth and limited employment opportunities. This means that currently, also people from stable families are engaging in dagaa trade, often investing in boats and gear, and hiring fishermen to work on their behalf. This shift was supported by the survey results, which indicated that 34,7% of the participating processors and/or traders was widowed and 6,6% was either single or separated, whilst 58,7% was married. However, the survey results also indicated that 52% of the traders and/or processors were head of their household. Within the Kenyan context, according to the interviewees, a man is automatically the head of the household in case of a marriage. It is therefore unsure to what extent these percentages are reliable.

Throughout the dagaa supply chain, there has historically been a highly gendered division of labor, which was supported by my observations during the field visits to markets within Kisumu and landing sites in the surrounding counties. I observed that fishermen and transporters are without any exception male, while dagaa processors are all female, as well as the great majority of traders operating in the markets. Four out of 75 traders and/or processors who participated in the survey were male, the rest was all female. Owiti acknowledged that in the lake region, dagaa trading is fully considered a female job. In distant markets there are some male traders operating – however, very few. Trade related to the fishmeal industry forms an exception. Owiti explains this by the starting capital needed to get involved in this type of fish trade, which women generally do not have access to. This

opportunities. In general, the respondents believed that it is easier for men to start a business than it is for women, as women often do not have access to starting capital.

According to the key informants, the gendered division of labor is rooted in cultural traditions: men are believed to be more efficient than women when it comes to finding food, whereas women are believed to be more efficient than men when it comes to handling food. Butcheries being an exception, when in Kenya trade is linked to food, it is female dominated. Among the participants, it was a widely held belief that fishing is a typical men’s job because it is a very physical job which takes place in precarious circumstances (during night hours, at open water), and men are believed to have more physical strength than women, whereas processing (sun-drying) and selling of fish is perceived as a typical women’s job, as women are believed to have more patience than men. Joyce Okwara, [functie at institute] however emphasized that this is not unique in dagaa fishing and trade, but widespread in other types of fisheries in Kenya as well. The transportation of dagaa, which often happens on motorcycles, is only done by men, as from a cultural perspective, women are not supposed to drive motorcycles.

4.2 Harvesting dagaa from Lake Victoria

This section discusses dagaa harvesting methods and associated challenges regarding sustainability of dagaa fishery.

Traditionally, dagaa fishing happened not far from the shore, using wooden dug-out paddle canoes. Currently however, most boats are motorized, which has increased the catches and allows vessels to move faster and go into deeper waters. Each boat is manned by approximately four fishermen, optionally including the owner of the boat. Fishermen operate at night and land the catches between six and seven in the morning. The catches include large quantities of caradina, a small, shrimp-like species that is found in the shallow waters of the Nyanza Gulf, on which shores Kisumu is located. Caradina is widely used to produce animal feed and is by the key informants argued to be unfit for human consumption. Upon arrival of the fishing vessels at the landing site, female processors and traders are already awaiting the days’ catch at the beaches.

A pressure lamp, used to attract dagaa to the surface.

07:00hr. Women carry dagaa (still mixed with caradina) from the boats to the nets spread out on the ground in order to sun-dry the fish.

Dagaa fishing vessels are usually equipped with seine nets and pressure lamps (figure 2) to attract the fish to the surface. This is a legal practice as it assures selectivity in the catch: dagaa is attracted by the light, whereas other species are not. As dagaa fishing relies on light attraction, dagaa catches strongly fluctuate according to the lunar cycle: during full moon, there is almost no dagaa fishing activity in the Lake. According to Rodrick Kundu (functie at institute), there are currently attempts to introduce lamps powered by solar batteries, but this practice is still in the testing phase: it is yet unclear whether solar lamps can produce sufficient light. The introduction of solar lamps would however be more environmentally friendly, as kerosene pressure lamps emit CO2 and often spill fuel into the water. During my morning visits at landing sites, I did however observe large batteries being carried away to be charged on solar power. This indicates that to some extent, solar lamps are already in use.

Regarding the harvesting of dagaa from Lake Victoria, several issues have been brought to light in the key informant interviews. First and foremost, Owiti and Kundu pointed out that water pollution in the Gulf is problematic in several ways. The issue of pollution can be seen as an overarching theme, as the declining water quality has multiple implications on the ground as well as on policy level. The first issue related to pollution, is the increase of conflict between fishermen over resources. Pollution in the Gulf has caused eutrophication of the water, which has led to proliferation of the water hyacinth (chapter 5 will discuss this in more detail), and decreasing oxygen levels. These ecological changes are driving dagaa (and other species) into deeper, cleaner waters, outside the Kenyan territory, as dagaa breeds in

section, most boats targeting dagaa are currently motorized (Frame Survey Report, 2016), which allows fishermen to move faster and further than previously. The ability of fishermen to follow the fish into distant waters has increased conflicts, mostly transboundary, as fishermen have started to operate in foreign waters. According to the key informants, as fishermen know where to find high concentrations of fish in the Lake, there are several areas where there is constantly ongoing conflict between Ugandan and Kenyan fishermen.

The second challenge relates to illegal, unsustainable fishing practices driven by water pollution. As mentioned earlier, the most common harvesting method for dagaa is the use of small seines in combination with pressure lamps. Due to water contamination however, this practice has become unsuitable for dagaa harvesting in the Gulf: turbidity in the polluted water limits the range of light from the lamps, strongly decreasing their effectiveness. This issue drives fishermen towards illegal, unsustainable practices: currently, at any landing site along the Gulf, dagaa is harvested through the use of boat seines without any type of light attraction. Compared to the initial method, the current harvesting practice entails less selectivity and accordingly less sustainability, as the catches currently often include by-catch of Nile perch juveniles. Even though the use of boat seines without light attraction is officially illegal, it is presently the most commonly used dagaa harvesting method along the Gulf. The policy implications of pollution in the lake will be discussed in the next chapter.

In line with the foregoing, the third issue also relates to unsustainable fishing practices. As recorded in the Frame Survey Report, the great majority of small seines targeting dagaa uses mesh sizes smaller than the recommended legal size of 8-10 mm: in 2016, 51% of the vessels targeting dagaa were small seines with mesh size 6-7 mm, approximately 38% fell into the mesh size category of ≤5 mm, and only a minority of approximately 11% of the vessels was using the recommended legal mesh size of 8-10 mm. According to Okwara, these small mesh size nets are originally produced for Nile perch or tilapia fishery, but found their way into dagaa fishery as the Nile perch and tilapia stocks were declining and people were seeking alternative sources of livelihood. Mesh sizes smaller than the recommended 8-10 mm increase the by-catch of immature fish. Besides the use of nets meant for other types of fishery, another illegal, unsustainable practice that was often mentioned in the key informant interviews, was the use of mosquito nets for dagaa harvesting. According to Kundu, a mesh size smaller than the recommended 8-10 mm impacts the environmental state of the lake and the future availability of fish, as catches often include juvenile dagaa. A third illegal practice was mentioned by Ken Luga (functie @ institute), who argued that multiple ≤6-7 mm nets are often sewed together to increase their