MANAGING INEQUALITY: THE POLITICAL ECOLOGY OF A SMALL-SCALE FISHERY, MWERU-LUAPULA, ZAMBIA

This paper starts from the perspective on resource management approaches as based upon a body of environmental knowledge. By analysing fisheries management in Mweru-Luapula, Zambia, we argue that this body of environmental knowledge has (i) remained largely unchanged throughout the recent shift to co-management and (ii) is to a great extent based upon general paradigmatic conventions with regard to common property regimes. We therefore simultaneously studied the historical trajectories of both resource management as the political ecology of Mweru-Luapula’s fishing economy. Using a relational perspective – by looking at interaction of the local fishing economy with external developments, but also by examining socioeconomic relations between individual actors – this study exposes constraints and incentives within the local fishing economy that are not absorbed in the current comanagement regime. These findings challenge both policy goals as community-based resource management itself. We therefore argue that governance of small-scale fisheries – in order to close the gap between locally based understandings, policy and legislation – should always be built upon all dimensions (social, economic, ecological, political) that define a


Introduction
Many scholars assume that most small-scale inland fishery communities represent the poorest part of rural societies (Béné, Mindjimba, Belal & Jolley, 2000). This claim is often argued through what Béné calls the 'old paradigm' on poverty in inland fisheries: poverty is associated with natural factors like ecological effects of certain exploitation levels (Béné, 2003). The approach of inland fishing communities as the 'poorest of the poorest' does not directly imply that fisheries automatically lead to poverty, but is linked to the nature of many inland fishing areas as a common-pool resource (CPR) (Gordon, 2005). From this paradigmatic point of view, a common and open-access property resource is incapable of sustaining increasing exploitation levels caused by horizontal efforts (e.g. population pressure) and vertical intensification (e.g. technological improvement) (Gordon, 2003;Kapasa, Malasha & Wilson, 2005). The gradual exhaustion of the fishery resource due to 'Malthusian' overfishing was identified by H. Scott Gordon (1954) and called the 'tragedy of the commons' by Hardin (1968). This influential model explains that whenever individuals use a resource in common -without any form of regulation or restriction -this will inevitably lead to its environmental degradation. This link is exemplified by the prisoner's dilemma game where individual actors, by rationally following their self-interest, will eventually deplete a shared resource, which is ultimately against the interest of each actor involved (Haller & Merten, 2008;Ostrom, 1990). Summarized, the model argues that the open-access nature of a fisheries resource will unavoidably lead to its overexploitation (Kraan, 2011). The paradigm of Malthusian overfishing is related to poverty in two ways. For one thing it explains population pressure and increasing fishing efforts through the image of small-scale fishing as a last resort whereby a larger socioeconomic or political context pushes people into fishing activities (Béné, 2003(Béné, , 2009Onyango, 2011;Smith, Koa & Lorenzen, 2005). Increasing population pressure then has negative impacts on the level of fish stocks in the resource, which in turn contributes to further impoverishment of the fishermen (Agrawal, 2005). Overexploitation will only increase as fishermen's incomes decrease (Kraan, 2011).
These presumptions argue for an 'external Leviathan' to control and regulate shared natural resources and have therefore historically legitimized governmental management systems (Jentoft, Bavinck, Johnson & Thomson, 2009;Ostrom, 1990). As Haller (2007) explained, the model on the tragedy of the commons "seems to have encouraged many governments in the South to take rights and responsibilities related to resources out of the hands of local groups and to legitimize this by claiming that the overuse of CPRs such as forests, wildlife, fisheries and pastures was the result of this tragedy" (p. 1).
It is however widely acknowledged that conventional centralized management of natural resources has failed to improve resource development, due to a lack of enforcement capacity and insufficient legitimacy at the local level to ensure compliance with regulation measures (Evans & Andrew, 2011;Jul-Larsen & Van Zwieten, 2002). This realization has generally led to a shift from a centralized management approach of fishery resources to decentralized efforts where locally based institutional systems are included and participation at community level is promoted (Haller, 2007;Jul-Larsen & Van Zwieten, 2002). This decentralization process has been recommended by international organizations as the World Bank and FAO in order to address the issue of resource overexploitation. Thus opportunities were created to implement co-management regimes, which include partnership arrangements and power-sharing models that can instruct local communities but also transfer effective control by delegating power to community level (Pomeroy, 2011).
Every resource management approach is an application of environmental knowledge. This means governance is composed by understandings about "human societies and the environment, scientific findings and technologies, standard (accepted) management approaches, political and economic prerogatives, and location-specific understandings" (Goldman & Turner, 2001, p. 2). The body of environmental knowledge that shapes resource management has however been widely criticized for prioritising environmental sustainability over social and economic reflexions (Hauck, 2011). Regarding poverty this means that a direct relationship is being put between low fishermens incomes -due to low catch -and overexploited resources (a result of the open-access nature of fisheries) (Kraan, 2011). Many scholars hence argue for the inclusion of issues of power and ideology -essential elements to understand social and institutional change -in fisheries management (Béné, 2003;Haller & Merten, 2008; three main commitments: (i) a commitment to understanding biophysical processes and the ecological system and not solely using it as a background context for social conflict, (ii) a commitment to context and the ambition to connect local processes to broader social and ecological changes and (iii) strong commitments to social justice by focusing on mechanisms of political and socioeconomic subordination, marginalization and disempowerment (Goldman, 2011;Hauck, 2011). 2 Simply stated this means that three elements were central in every observation: nature, context and power. This approach is political in two ways: (i) it focuses on the political-economic processes in which environmental change is embedded and (ii) it is political in the sense that researchers are eager to change the situation they study, for example by improving standards of living or by reversing environmental degradation. This understanding should equally be present in the analysis of the political ecology in Mweru-Luapula. This means that one has to take into account the sociopolitical processes and ecological context that led to the production, application and circulation of Mweru-Luapula's knowledge base and the management's environmental knowledge. Furthermore the application of the fisheries management regime should be studied by focusing on enforcement and compliance by taking into account the (possibly unequal) distribution of costs and benefits, its sociopolitical effects and its ecological consequences.
This research on governance and the capitalization of Mweru-Luapula's fisheries is mainly composed by qualitative fieldwork data -collected through individual and key informant interviews and observations. Other data is constituted by governmental reports and legislative documents and documents published by NGOs. The fieldwork was carried out in villages, fishing communities and administrative centers in Nchelenge District in Luapula Province from September 2011 to November 2011. Thirty-five key informants were interviewed and the range of these informants spread through various backgrounds (e.g. fishermen, chisense fishermen, female fish mongers, urban fish traders, VMC members, community headmen, traditional authorities, government and NGO officials, Director of Fisheries, etc.). The interviews were semi-structured, in-depth and open-ended. Observations were made at the Nchelenge fish market and through patrolling on Lake Mweru with the Immigration Office and representatives of the Village Management Committee and the Department of Fisheries.  Walker (2005) explains, a debate currently takes place between those who argue political ecology should mainly focus on the biophysical aspect and between those who concentrate on the social and political aspect of ecological change. Without taking a methodological stand to firmly, this study tries to look at how all aspects -social, biological, ecological, institutional, economic -are intertwined and how they mutually influence each other.

Mweru-Luapula's biophysics: a story of resilience
The Mweru-Luapula fishery includes Lake Mweru, its islands, the Luapula River with its lagoons and swamps and flood plains from Mambilima Falls south to Mwense up to Lupiya at the Lunchinda River. The border between the Democratic Republic of Congo (DRC) and Zambia runs right through both the lake as the river and lagoons. As we will see later, the nature of Lake Mweru as a cross-border resource has until now been a determining factor for both management as the political ecology of Mweru-Luapula. Because of the Luapula River that flows into Lake Mweru, and the outflow of the Luvua River in the north, Lake Mweru is an allotrophic lake, which means that total fish production is highly dependent on the constant in-and outflow of nutrients (Mölsä, 2009;Van Zwieten, Goudswaard & Kapasa, 2003). This means that yield potential for a lot of fish species is high but also highly variable. An important consequence of this exists in the fact that the Lake's ecology offers biological counter pressure to the Tragedy of the Commons paradigm, because allotrophic lakes are known to have a low susceptibility towards increased fishing efforts and a high but variable recovery rate (Van Zwieten et.al., 2003). Mweru-Luapula therefore has two main biological characteristics that are essential within this study: it has a high variability of fish species and an ecological resilience that is rare among common-pool resources (Annear, 2010). Those two features are intertwined, as ecological resilience in this case means that certain species can sustain extensive fishing pressure but also that a reduction of fish levels of a certain species often goes combined with the rise of another. It is easy to notice the Lake's potential to avert environmental degradation, yet it depends on the adaptability and resilience of both fishermen, fish traders as the market towards these biological variations to evade impoverishment too.
Numerous works on Mweru-Luapula's fisheries described its system as an economy build-up by a dynamic population of fishermen using an ecological resilient resource (Annear, 2010;Gordon, 2003Gordon, , 2005Jul-Larsen & Van Zwieten, 2002;Van Zwieten et.al., 2003). Even the current management regime takes the view that -in principle -an ecological equilibrium could define Mweru-Luapula's fishing resource. From this perspective change is brought on by external factors as population pressure -e.g. caused by permanent and seasonal migration -that can be problematic as they disturb this balance and eventually cause biological depletion. Yet we argue that change is the interaction between the endogenous (e.g. institutions, ideology, organization) and exogenous (e.g. social and physical environment, population and technology) aspects of a society (Ensminger cited in Haller, 2007). The following chapters on the historical evolution of Mweru-Luapula's fishing economy therefore focus on the social constraints and incentives for individuals -within the broader context of socioeconomic and political change -as well.

Meanings of wealth and poverty: historical trajectories of governance and political ecology
To exemplify that wealth characterizes inland fisheries as much as poverty does, this study disputes two common assumptions that greatly influence the management regime of Mweru-Luapula. The approach of fisheries as an activity of last resort for impoverished households lays claim on clarifying the cause of both migration into as outside of fisheries (Cinner, Daw & McClanahan, 2008;Van Zwieten et.al., 2003). The second presumption concerns the linear and often simplified link between migration and the biophysical nature of a common-pool resource. Without denying environmental effects of population pressure, the next chapters argue for a more holistic approach that starts from the wider context of migration and incorporates socioeconomic and institutional consequences too.

Colonial Capitalism
The Mweru-Luapula region has historically been characterized by strong seasonal and permanent migratory patterns (Mölsä, 2009). This observation is connected with the role of the Luapula valley as a center for regional trade at least since the eighteenth century and with its history of conquests (Gordon, 2003). Various oral traditions narrate the legends of settlers and conquerors, among which the conquest of the Luapula valley by the eastern Lunda tribes under Mwata Kazembe -and the following establishment of the Kazembe Empire -is probably the most imaginative and well-known. Yet the most crucial and decisive story of conquest is left by Zambian and Congolese colonial history. In respect to Mweru-Luapula's fisheries, some important migratory movements took place under colonial domination.
During British and Belgian colonial rule European -mostly Greek -fishermen and fish traders were encouraged to start their economic activities on and around Lake Mweru (Gordon, 2005). The context of their presence -situated historically in the beginning of the twentieth century and especially the 1920s and 1930s -has to be understood through the efforts to industrialize and commercialize the fishing economy in Mweru-Luapula (Musambachime, 1995). Urban markets in the Zambian Copperbelt but especially in Elisabethville -near the booming copper mines controlled by Union Minière du Haut Katanga (UMHK) -demanded a stable supply of fish (Annear, 2010;Gordon, 2005). This encouraged expatriate merchants and fishermen who owned sufficient capital to start commercial activities in Mweru-Luapula (Jul-Larsen & Van Zwieten, 2002). Both the Belgian as Northern Rhodesian colonial administration shared the idea that such a form of capitalized entrepreneurship was desirable. The Northern Rhodesian administration however preferred indirect rule policies to reach this objective, while the Belgians seldom placed resources under the control of local tribal and traditional authorities (Nugent, 2004).
The coexistence of two different views on colonial management -a paternalistic and authoritarian system versus an organization whereby native authorities were given decision-making powers to some extent -was an important feature within Mweru-Luapula's economic and political system. For one it meant that fewer expatriate fishermen and traders were present on the Northern Rhodesian side of the Luapula. This was important as it slowed down the process towards a commercialized fishery in Northern Rhodesia. Furthermore, enforcement of unpopular legislation concerning resource management -like the implementation of a closed fishing season in the Luapula River to protect the mpumbu's spawning runwent much more difficult in Northern Rhodesia.
The difficult task for the local chiefs to mediate between communities and administration, to enforce colonial legislation on the one hand but to maintain authority and respect with local fishermen on the other hand, exposes not only a vital feature of the indirect rule system. At the same time it uncovers an ambiguous relationship between enforcement of fisheries legislation and local compliance. Within the Northern Rhodesian administration this duality was expressed by the linking together of unpopular conservation measures and development projects (Gordon, 2005). Regarding fisheries commercialization, such development projects actually stimulated the formation of African capitalized entrepreneurs above simply ensuring compliance. For instance, the construction of an ice plant in 1955-56 and the improvement of road infrastructure made the fresh fish trade a less lucrative undertaking and therefore encouraged participation of African traders. 3 Knowing this, one has to nuance a discourse that opposes traditional and commercial fishing by proclaiming that commercialization of fisheries in Northern Rhodesia solely discriminated local communities and pushed households out of fishing activities. Due to the indirect rule system, artisanal and small-scale fishing continued to coexist alongside more commercial fishing, and even partly converged as a number of small-scale African fishermen were able to transform themselves into more capitalist entrepreneurs towards the end of the fifties.
The nature of the colonial administration in Northern Rhodesia, together with the opportunities for social mobilization -an integral feature of capitalism -created decentralized political and economic networks wherein a local elite was able to form itself. It is important to notice that this was not only a formal process instigated by the granting of authority in colonial chiefs. A new civil society emerged, guided by commerce and capital. Aimed at monetary accumulation, new economic alliances based on credit and trust came into being. Such networks often began with the expatriate traders who extended goods to African traders, who in turn exchanged them for cash or used them as credit for fishermen in exchange for a predetermined amount of fish (Gordon, 2003). Economic growth in the Copperbelt and nearby cities also stimulated distribution of new goods and materials to rural areas like the Luapula valley. This incited the swift rise of small businesses in fishing villages as Kashikishi and Pweto. From this moment on Mweru-Luapula's economy would be largely driven by capitalist standards and commerce that -according to David Gordon (2003) -"swept away established hierarchies of political power, kinship patterns, and forms of social organization" (p. 103). In fact the relationship between both traditional and capitalist forms of organization up until now remains ambiguous. For one thing both patterns went on to coexist, as commercial fishing occurred for the most part on open waters, while subsistence fishing remained in the swamps, lagoons and shallow waters, were more traditional tenure rules are employed. Secondly, traditional production relations based on kinship were not just replaced by large-scale organizational forms, as they often continue to exist even when incorporated in a larger economic and commercial system (Platteau, 1989).
Yet one can not neglect the reality that the rise of a new capitalist entrepreneurial class will inevitably collide with a traditional elite based on custom and heritage. As elite formation is an essential feature of capitalist development, it necessarily includes mechanisms of exclusion and marginalization. In Mweru-Luapula's colonial fisheries this was visible from both a political as ecological perspective. Events of local opposition and rebellion against the colonial administration gradually increased, as colonial administration policy and legislation mainly served to protect the rights of capitalized expatriates. This had notable ecological effects as the expanding commercial efforts eventually led to the disappearance of the mpumbu fish, which had become a highly commercial commodity. Yet the mpumbu fish was also targeted for subsistence purposes through customary practices in traditional fishing communities. Its gradual depletion increased competition -among each other and with the recently installed European-driven commercial fishing economy -and forced many households to migrate out of Mweru-Luapula. This pattern of out-migration was furthermore driven by growing industrialization and urbanization in Northern Rhodesia and Belgian Congo, which incited rural-urban patterns of labor migration, meaning households became attracted by more profitable employment opportunities in other economic sectors (Annear, 2010;Ferguson, 1999;Gordon, 2005).

Independent nations: state patronage and rural development from a government perspective
During the 1960s and 1970s one could witness a decline in catches in Mweru-Luapula's fisheries, corresponding with a decrease in capital investment. Reasons for this can be found in both the local as regional context. Without mpumbu the fishery was less profitable, hence its disappearance eventually made European expatriate fish traders and fishers leave the Luapula valley. Catches of pale and other commercial species had also diminished. While from a government perspective this was presumably caused by harmful fishing techniques, the reality of a succession of droughts affecting breeding and spawning possibilities, seems -regarding Lake Mweru's allotrophic nature -a more plausible explanation for the decrease in fish levels (Gordon, 2005).
Yet regional economic outcomes affected fisheries investment too. Since the beginning of the 1970s, many mining companies in the Katangan Copperbelt ceased to provide their workers with food rations -which was a form of payment -and consequently ended signing contracts with capitalized fish traders. Almost simultaneously copper prices collapsed, which influenced the process of industrialization in Zambia and Zaire that -after a long period of rapid expansion -began to slow down. This had immediate effects on Mweru-Luapula's commercial fishery that had become highly dependent on urban and industrial demand.
Furthermore the postcolonial institutional and political context affected Mweru-Luapula's fishery to a great extent too. The Zaireanization decree of 1973 resulted in the confiscation of expatriate industries by President Mobutu Sese Seko, who installed allies who more than often lacked the necessary knowledge of fish production, processing and marketing, to run fishing companies in Zaire. In Zambia, development plans in the 1960s and 1970s were drawn up to actively support rural production through the formation of cooperatives, the provision of credit and the centralizing of product marketing. In Mweru-Luapula, this resulted in the take-over of the Kashikishi ice plant by a parastatal organization called Lake Fisheries of Zambia (LFZ), to become a central marketing board. At the same time fixed market sites were established at the lakeshore, and a pricing regulation sought to counteract proliferation of fish prices. The objective was preventing fishermen from being exploited by fish traders, and eventually even eliminating the need for those private traders. The provision of credit by the rural development program to fishing cooperatives should be analyzed as an attempt to unite fishermen in one or more controllable and monitorable organizations.
The story of Gabriel Kunda, a fisherman in Kashikishi, exemplifies the opportunities and weaknesses of these development programs (Mr. Kunda, personal communication, October 6, 2011). In the early 1960s, he moved out from Mansa -Luapula Province's provincial capital -to become a fisherman like his parents had been. Using government credit, he was able to buy two marine engines and seventy nets. He did not process his fish but sold it directly to Isabi Investment, the Kashikishi ice plant owned by LFZ. In 1964 he founded a fishing cooperative through which fishermen could look after and protect fishing nets and boats. He had forty-two fishermen who worked on his behalf and shared in his profit. The government's pricing policy however provided these fishermen with the incentive to sell portions of catched fish on the Lake itself to private traders, instead of selling them on the fixed marketing ponts to Isabi Investments. To prevent fish being stolen from him that way, Mr. Kunda utilized credit ties with his fishermen by giving them cash or fishing gear in advance. Yet networks of kin and credit existed between his fishermen and private traders too. Although Mr. Kunda made good profit in the first years, his business eventually slumped because of the decline in pale catches and the low prices the buying depot in Kashikishi offered. Mr. Kunda's story is not uncommon, as the development programs and the LFZ collapsed in 1979. This was bound to happen as socioeconomic relations -based on mutual gain and dependency -between private traders and fishermen persisted and even became stronger as the real income of fishers was being decreased by the government's pricing policy. The implemented system of state patronage was unable to compete with existing networks between fishermen and fish traders. Gordon (2005) argues that "by the late 1970s, only remnants of the colonial fishery were left" (p. 504). This applies to the institutional context too. Although in 1974 the Department of Fisheries (DoF) was created and the Fisheries Act decreed -which was an ordinance that provided for the development and control of fishing and the registration of fishers and boats -this centralization process did not succeed in replacing former colonial and traditional regulation systems by formal state policy (Malasha, 2007). For example, fishing restrictions in the Mifimbo breeding area were abolished after independence, and efforts to install a closed fishing season could not be enforced until 1986 (Annear, 2010;Gordon, 2005). Eventually, state absence and incapacity in the Luapula valley granted almost unhindered access for any individual with the means -i.e. capital and knowledge -to exploit the fishing resource.

Private victory: the emergence of the patrons
From the 1970s onwards, the fishing population in the Luapula valley grew dramatically. As Van Zwieten et.al. (2003) showed -based on their 1997 frame survey -one fifth of all fishermen were not born in the Luapula area. This shows that population growth largely resulted from a migration movement, a movement that was inherently linked to the regional and urban context in the Copperbelt. The crisis in the mining industry in the 1970s and the subsequent contraction of the formal sector -caused by the externally laid on reforms of the Structural Adjustment Programs -by the end of the 1980s and 1990s, pushed innumerous people into small-scale fishing that thus served as an informal safety net (Gordon, 2005;Jul-Larsen & Van Zwieten, 2002;Kapasa et.al., 2005;Malasha, 2007). A lack of employment opportunities made many households move back into the rural economy (Ferguson, 1999;Islam, 2011;Nugent, 2004). Yet it would be over-simplified to analyze population growth in Mweru-Luapula solely through the paradigm of fishing as a last resort. Although commercial fishing activities decreased, they did not disappear. The departure of European fishermen and traders in fact created a vacuum that offered further opportunities for both Zambian as Zairian entrepreneurship. Moreover, retrocession of the Zaireanization process towards the end of the 1970s led to the return of some business ventures to its original owners (Nugent, 2004). The story of Zairian businessman Katebe Katoto is exemplary for this. Through a lucrative with Gécamine -in former times called UMHK -he continued to supply the copper mines in Katanga. He dominated the entire Zairian side of the fishery and even offered Mobutu Sese Seko to buy Lake Mweru (Gordon, 2003(Gordon, , 2005. Although Katebe Katoto was probably the biggest patron Lake Mweru has ever seen, he was certainly not the only one. Many Zairian and Zambian fish traders used labor-and credit ties to elaborate an extensive economic network based on patron-client relationships. More often than not these patrons -who saw economic opportunities in the fish tradecame from elsewhere. This is understandable for the following reasons: (i) many of them 'returned' to Mweru-Luapula and/or already had relatives or acquaintances engaged in fishing activities, (ii) they had been able to generate capital through other economic activities, making it possible to invest directly in fishing gears and labor and (iii) their urban social network gave them a comparative advantage over other 'autochthonous' fish traders with respect to the fish commerce.
An economic system based on patronage was however primarily stimulated by the nature of pale fishing, which became the single most important commercial species after the disappearance of the mpumbu fish. Pale -also called green-headed bream -mostly lives in the open waters of the lake, while floodplains and beaches are used for breeding (Gordon, 2003). The catching of pale therefore demanded more capital-intensive equipment like expensive nylon nets and fishing boats. In consequence many fishermen depended on credit-relationships with Zambian and Zairian traders for the provision of this equipment. Private traders supplied nylon nets in exchange for a predetermined quantity of fish. Such economic transactions in that manner functioned as an insurance mechanism for private traders who established a base of dependent fishermen -hence a stable supply of fish. As a new road was constructed on the Zambian side of Lake Mweru, urban areas became far more accessible through Zambia than through Zaire. This advantage made many Zairian traders move their business from Pweto and Kasenga to growing Zambian fishing towns as Kashikishi and Chiengi (Musambachime, 1995).
Summarized, the migration movement towards the Luapula valley did not only result from a deteriorating external economic context but was also stimulated by the economic opportunities of pale fishing. In this fishing economy, private traders became essential key players as the need for capital to engage in pale fishing could not be met by state initiatives. More traditional fishing methods in the lagoons and at the shoreline targeted juvenile pale by the use of weirs and traps. Such fishery was -like the colonial mpumbu fishery -more locally based contrary to the more accessible nature of lake fishery. Yet the targeting of juvenile fish was also directly related to capital-based pale fishing. The promise of future catches within the patron-client relationships forced fishermen to target juvenile fish as well when other catches proved to be insufficient. Furthermore not all fishermen were able to obtain credit to buy the necessary fishing gears. Possibilities of receiving credit were for instance much lower in more remote areas were fresh fish trade to the urban markets was not evident.
The success of open lake pale fishing depends naturally on successful breeding periods, which makes the targeting of juvenile fish a hazardous occupation. This made lake fishers and patrons call for stronger restrictions to catching in the breeding areas. In 1974, the Zambian authorities therefore tried to enforce strict fishing regulations in the Mifimbo breeding area -which is an important place for the spawning of pale fish. However, several factors contributed to the constant failure of these regulations. Weak monitoring capacity of the Department of Fisheries and the ambiguous role of Traditional Authorities and local politicians -especially during election times -towards fishermen in the Mifimbo breeding area made enforcement of such measures weak. Furthermore efforts to prohibit the catching of juvenile fish to protect the commercial open-lake pale fishing falsely juxtaposed a subsistence way of fishing against a commercial fishing economy. The high prevalence of credit in the patronage networks forced fishermen -due to hazards of debt-recycling -into such illegal fishing methods as well. The distinction between traditional and commercial fishing is therefore partly invented, due to the nature of juvenile pale fishing as an undercapitalized commercial activity (Annear, 2010). Moreover the destructive impact of juvenile fishing in the 1970s and 1980s does not have to be exaggerated. As mentioned, Mweru-Luapula's waterbodies can sustain intensive fishing pressure in high fertility years. Despite growing exploitation -due to an increasing number of fishermen in open-lake pale fishing and shoreline juvenile fishing -total pale catches stayed relatively constant. Many argue that fluctuation levels were not so much determined by exploitation effort but by weather conditions (Annear, 2010;Gordon, 2003Gordon, , 2005Jul-Larsen & Van Zwieten, 2002;Van Zwieten et al., 2003). As Gordon (2003) argues, "a good rainy season allowed for stock replenishment" (p. 146).

Swimming gold? The rise of chisense fishing
Increasing exploitation levels -due to population pressure -did not seem to cause environmental degradation directly, yet total catch per capita decreased. A fishing effort became higher, competition grew, which limited profits for many. The rise of chisense fishing throughout the 1970s and early 1980s is therefore a much an example of the adaptability of Mweru-Luapula's fishing population as a consequence of its ecological resilience. Chisense are small lake sardines who live in large shoals in the upper layers of open waters. Its stock levels increased dramatically in the 1970s, probably due to changing nutrients levels in Lake Mweru -caused by a growing lakeshore population and the use of artificial fertilizers (Gordon, 2003). What makes chisense special is its incredible resilience to high exploitation levels, caused by its extremely high turnover rate. 4 Traditionally chisense was caught -for subsistence needs -in shallow waters close to the sandy shores of Pweto and Chiengi, by women using small seine nets or even clothing. The increasing amount of chisense fish and the ease of capture gradually made women dry the fish and sell it locally. The commercial potential for chisense was high on both sides of the fish chain. It was easy to catch and process, there was plenty, and the urban markets demanded a large supply of cheap fish. Its commercial potential attracted entrepreneurs who started using a more capital-intensive method of chisense fishing. By means of paraffin-lamps to attract these lake sardines by night and small mesh or meshless material to drag the fish, total catch per capita could be up to fifty kilograms per night (Gordon, 2005). This technological innovation did not only increase total production but commercialized chisense fishing to a capital-intensive occupation that required specialization through knowledge and experience.
The rise of the commercial chisense fishing in the early 1980s occurred alongside more customary net fishing. Since a lot of fishermen relied on patron-client credit relationships they were tied to their usual fishing practices and lacked alternative credit to diversify their economic activities. Therefore mostly migrant fishermen engaged themselves in chisense fishing. Thus a new class of semi-urban entrepreneurs was formed along the shores of Lake Mweru. Contrary to the patrons in pale fishing they were far more independent and lacked their extensive patronage network. This was partly due to the nature of chisense as a migratory species, which made credit relations more precarious as they had to follow the large shoals of fish from place to place. Secondly, the high level of competence needed for chisense fishing makes such patronage networks difficult to establish. Employers on chisense fishing boats mostly work one year for a gear owner to be able to buy their own gears and start their own business (chisense fisherman, personal communication, October 13, 2011) Relationships between gear owners and boat workers are therefore more temporary and contractual. Economic bonds between gear owners exist as well: often they use one large boat to transport the catched sardines altogether and fuel costs are shared. This system of cooperative labour is not based on kinship or custom, but is economic as it reduces costs for all stakeholders involved.
Contrary to chisense fishing the processing phase is far from capital-intensive. After chisense is dried in the sun it can simply be sold or transported to the urban markets. Therefore local women, who were involved in the processing of chisense and the small-scale trade, were able to earn an income independently. Gordon (2003Gordon ( , 2005 and Annear (2010) correctly postulate that from a gender perspective this was an essential impulse for women's empowerment. Until now, they remain to have a very strong position in the chisense fishing economy. Chisense traders who only intervene in the marketing stage, are forced to buy from these female traders -called abakukungwisha. Their success is further signalled by the fact that nowadays they provide starting chisense fishermen with money or fishing gears in exchange for fish. As we will see further, the role of women in Mweru-Luapula's fishing economy has become only more significant the last decades.

Competition and conflict: from state absence…
As we pointed out the importance of meteorological factors like rainfall for Lake Mweru's ecology, consequences are that good years can be followed by bad years and vice versa. The nature of the Mweru-Luapula fishery is therefore rather one of unpredictability -thus insecurity -than one of gradual environmental degradation.
Nonetheless, fluctuating catches did affect the nature of the pelagic fishery and had social and economic consequences too. Towards the end of the 1980s catches per unit in the gillnet fishery had decreased from 10kg/unit/night in 1980 to 4kg/unit/night in 1986 (Mölsä, 2009). The general response from the fishing communities surrounding Lake Mweru included technological change and an increase in scale. The former means that fishermen generally shifted to smaller mesh sizes while the latter comes down to the use of more nets per fisherman. Both responses did however require the necessary amount of capital to change and/or purchase fishing gear. That is why cases of net theft increasingly occurred towards the end of the 1980s. Even more, the use of self-improvised fishing gears like mosquito nets or even piece of clothing grew, which became heavily criticized for their destructive effects on juvenile fish (Gordon, 2005;Kapasa et al., 2005;Malasha, 2007;Wilson et al., 2004). This critique extended itself to illegal fishing techniques too, like beach seining -a technique whereby large amounts of small and juvenile fish and even fish eggs are dragged towards the shoreline using large small-meshed fishing nets. The use of such illegal fishing gears and methods, and the fishing activities on Kanakashi and Potolo Island -settlements based in the Mifimbo breeding area -were targeted as the causes for the continuing decrease in catch per unit. Accusations of illegal fishing were often connected to an autochthonous discourse whereby recently arrived migrant fishermen were blamed for destroying Mweru-Luapula's fish population (Wilson, 2004). Yet, residents from Kanakashi and Potolo Island utilised an indigenous discourse -arguing for the protection of their 'traditional' fishing practices -while the spreading of illegal fishing gears was many a time distributed through the same socioeconomic networks that had defined the pale fishing in the 1970s and 1980s in Mweru-Luapula (Annear, 2010). Exemplary for this are the accusations towards Congolese traders -who had often arrived in Zambia during the heydays of the pale industry -as the main providers for illegal fishing methods and gears.

…to state incompetence
The prevalence of illegal fishing but especially net theft generated many conflicts in the surrounding fishing communities. This led to important trends that would define the current comanagement fisheries regime. Firstly, these events in the late 1980s laid the true base for a discourse that linked decreasing catches directly to illegal fishing. Increased competition -caused by population pressure -gave many fishermen the incentive for the use of illegal fishing methods to maximize their profits. Without restrictions to these practices a Tragedy of the Commons would be unavoidable. In 1974 fishing regulations to cope with juvenile fishing were already introduced, but enforcement capacity was low -the DoF lacked the necessary resources, namely fuel, to patrol regularly. Throughout the 1980s the declining catches per capita made experts and policy makers call for the reintroduction of a closed season. The last successful closed season policies dated back to the colonial restrictions to protect the mpumbu spawning run. In 1986 the Zambian authorities agreed on a closed season from December to March, following Zairian success to restrict fishing in this period. 5 The closed season policy was supported by wealthier fishermen who were able to bridge this period financially and who could even support more small-scale fishermen through their patronage network, which in this case functioned as an informal insurance mechanism. Yet, independent small-scale fishermen -who saw their income cut off for three months -resented the closed season just like the fish traders did. The issue of compliance was inherently linked with low enforcement capacity of the DoF and state officials who often -in turn for bribes from wealthier traders -turned a blind eye for fishing in the closed season (Gordon, 2003). These practices, whereby some could circumvent regulations using financial pressure while others saw their gears confiscated by state officials, made many fishermen feel that the administrative measures were enforced unfairly. Many argued that regulations were only used by state agents to collect personal revenue, by accepting bribes and by keeping income from policing for themselves. The Local Government Act from 1991, which encouraged rural councils to collect their own finances, surely enhanced these practices. Soon policing collapsed due to a consensus within Mweru-Luapula's fishing communities that they should not comply with the conservation measures that benefited state agents more than the fishermen. Due to this collapse net theft only increased further. This called for a new form of local organization whereby control of fishing gears could be monitored. This led to the formation of Fishing Associations (FA) aimed at the controlling of fishing gear theft (Kapasa et al., 2005;Malasha, 2007;FA Chairman Kashikishi, personal communication, October 6, 2011). The first FA was founded in Chiengi District, close to the Zairian border where net theft was rampant. Eventually this Chipungu Fishing Association moved to Kashikishi in 1988. This form of local organization would later on be used as an example for a more participary approach towards fisheries management. Yet this was certainly not the only example of local institutional power in Mweru-Luapula. The role of Traditional Authorities and village headmen for instance has always been important in conflict resolution and local governance. Cases of net theft were often addressed to these local authorities who had the power to punish or even evict someone from their communities (village headman Nchelenge, personal communication, September 30, 2011).

The interpretation of history
These events in the 1980s and early 1990s formed the body of environmental knowledge that constitutes the actual policy perspectives on Mweru-Luapula's fishery. To recapitulate, this perspective starts from Lake Mweru as a common-pool resource that -after decolonization -ensured open access to anyone who wished to enter. This led to a migration movement of individuals seeking their luck in smallscale and artisanal fisheries. Population pressure consequently caused an increase in the use of illegal fishing methods and called for a policy aimed at environmental protection through law enforcement and monitoring, control and surveillance practices to counter illegal and unreported fishing.
By discussing the political ecology of Mweru-Luapula' fisheries from the colonial period onwards to the early 1990s these chapters questioned the premises on which this management regime is based. Firstly, the nature of pale and chisense fishing -the two most important commercial species after colonial domination-as a capital-intensive economic activity, does not support the idea of an open-access nature of Mweru's fishery. Access to knowledge and capital is essential. These assets are distributed through socioeconomic networks that define the pale and chisense fishing economy -albeit in two different ways. Wealth and knowledge is distributed through these networks, yet in an unequal way. The assumption that Mweru-Luapula's fishing economy consists of equal, rational and independent economic agents moves past the existence of socioeconomic relations that generate and sustain inequality and create networks of dependencies. The notion of inequality and interdependency challenges this dominant rationalist model based on an economic concept of community -characterized by perfect transactions of costs and information.
The next chapters will explore such economic networks in the current fishing economy in depth to uncover such social constraints and incentives, which influence people's actions. Regarding the comanagement regime -that focuses heavily on illegal, unreported and uncontrolled fishing -this implicates among others that policy has to make a distinction between those "who are maximising gains from illegal fishing and those who are simply relying on resources for their livelihood" (Hauck, 2011, p. 208).

Parallel transformations: institutional and socioeconomic change Towards decentralized resource management From CAMAP to PLARD: the institutional set-up
The inability of central government policy to enforce conservation regulations, together with the persistence of conflicts with regard to net theft, led to the establishment of the Conservation and Management Action Plan (CAMAP) by the DoF in 1992. This was a program aimed at elaborating a participatory approach towards fisheries and conservation management (Kapasa et al., 2005). In 1994 advice was sought from the UN Food and Agriculture Organisation (FAO) to revise the 1974 Fisheries Act. 6 This consultation led to recommendations to decentralize fisheries management obligations to local fishing communities (Malasha, 2007). During 1996 and 1997 a consultative process -with the help of Dutch aid organisation Stichting Nederlandse Vrijwilligers (SNV) -was executed through the Fishing Associations to create a new co-management outline around Lake Mweru, a structure that was established at three levels (Kapasa et al., 2005;Malasha, 2007). At the local level seventy-eight Village Management Committees (VMCs) were founded, which include grassroots stakeholders like fishers, farmers, businessmen and fish traders, and the village headmen. Zonal Management Committees (ZMCs) were raised at chiefdom level: they consist of the chief or sub-chief, a fisheries officer, FA members, a council representative and other interest groups as more large-scale fish traders. Ultimately the Mweru-Luapula Fisheries Management Committee (FMC) came into being, comprising the senior chiefs in the area and a DoF representative (Mölsä, 2009). Despite these restructuring efforts, decision-making power still remained above the created institutions, as the 1974 Fisheries Act had no legal provision on the participation of local stakeholders in the management of fisheries (Fisheries Act, 1974;Musumali, Heck., Husken, Wishart, 2009). As a result of this, the VMCs lacked formal authority and could only fulfill the role they were expected to play by the central government. Furthermore the heavy involvement of the Traditional Authorities (TA) in the co-management structure was notable. Their role is not uncontested: on the one hand it is shown that involvement of TAs is necessary for co-management initiatives to succeed, on the other hand they also have their own private interests that can thwart decision-making capacities from the fishermen themselves and counteract the interests of co-management programs (Kapasa et al., 2005;Malasha, 2007;Wilson et al., 2004).
In 2007 the Fisheries Amendment Act was proclaimed, which officially provided for the creation of Fisheries Management Areas that would be responsible for the development of Fisheries Management Plans. The Fisheries Amendment Act set the stage for the creation of the new 2011 Fisheries Act and for the management of Mweru-Luapula's fisheries (Mölsä, 2009). This new Act legally provides the opportunity for the creation of VMCs, ZMCs and a covering FMC -which consists of delegates of the local fishing communities, TAs and local authorities, delegates of non-governmental organizations, commercial fishing operators and the aquaculture industry (Fisheries Act, 2011).
In Mweru-Luapula new co-management structures were created in 2007 and 2008 through the support of the Programme for Luapula Agricultural and Rural Development (PLARD), which is a programme jointly funded by the Governments of Zambia and Finland. The implementation of this project -for which the 2007 Fisheries Amendment Act supported the set-up -led to the creation of 77 VMCs and 14 ZMCs. In 2010 the Fisheries Management Committee was put in place.

Evaluating compliance against enforcement
The main goal of the Mweru-Luapula Fisheries Management Plan (FMP) is promoting sustainable development by ensuring ecological sustainability, which means maintaining the base stock and species at acceptable levels (Mölsä, 2009). Ensuring sustainability of fish stocks is strived for by limiting and controlling overall fishing effort -principally by opposing illegal, unreported and uncontrolled (IUU) fishing. The monitoring of overall fishing effort stems from the idea that too many fishermen and fishing nets are putting high pressure on the Lake's ecology. As Hanna Mölsä (2009) -a PLARD officerindicates, the nature of fishing has become more investment-driven. This is supported by the fact that the number of nets per boat has increased from 30/50 nets per unit in the 1990s up to 200/400 nets per unit today. 7 Simultaneously the number of fishermen at the Zambian side of Lake Mweru has increased a fourfold in the 1970s.
To control the number of fishers, boat and gears, a system of fish licensing was reintroduced, whereby fishermen have to pay license fees according to the gears they use. Formally, this is called to be the main system restricting access to newly arrived fishermen. 8 However, at village level a customary mechanism exists as well, whereby newcomers have to bring along a reference letter from their former village headman underwriting one's character (Annear, 2010; Village headman Nchelenge, personal communication, September 30, 2011).
The licensing system also serves as a source of revenue for the VMCs, who legally have the obligation to collect license fees on behalf of the DoF. 9 However, weak patrolling and enforcement capacity -largely due to a lack of funds -still grants unlicensed fishing with impunity (Fish businessman, personal communication, October 7, 2011;DoF Officer, personal communication, September 30, 2011). This applies to the controlling of a maximum number of gillnets per unit as well. 10 Another controlling mechanism is foreseen in Mweru-Luapula's FMP, namely use of fish levies, which are taxes on transported fish. According to the Local Government Act (2004, Art. 69/1/b) this is a responsibility of the council. However they do not have the capacity to collect these levies in the fishing camps and on the islands, where most of the fish is transported to the urban markets (DoF Officer, personal communication, September 30, 2011). Finally, efforts are being put in the promotion of alternative sources of income, generated through for example cassava farming but also through the development of aquaculture -i.e. fish farming -to reduce pressure on the Lake's resources (Mölsä, 2009). Gordon (2003) and Annear (2010) already explained that historically a lot of households are used to combine farming and fishing activities. Although these livelihood diversification measures have potential to improve the local economy, living standards and nutritional health -and therefore can promote sustainable development -it does not target all households in Mweru-Luapula. After all, the most vulnerable households -living in the fishing camps -are full-time fishermen without any formal ownership of property and therefore without any land to cultivate for farming or aquaculture. 11 Mweru-Luapula's FMP also comprises monitoring, control and surveillance (MCS) of IUU fishing. This illegal fishing includes fishing activities during the annual fish ban -from December to March -and in the restricted breeding areas, and of course the use of illegal fishing gears and methods. An essential component of this policy comprehends sensibilization campaigns in the villages, fishing camps and on the islands. Such campaigns are proving to be successful, in the sense that illegal fishing methods and fishing when and where it is prohibited are almost unanimously identified as the main issues to be tackled in Mweru-Luapula's fisheries. This local consensus does not necessarily justify the strong focus of the FMP on IUU fishing. In this regard Agrawal (2005) points to the concept of 'environmentality' whereby decentralisation of regulation by creating governmentalized localities -in this case the formation of local management committees -can create environmental subjects who are self-regulating and who come to think and act in new ways in relation with the environment, ways that are integrated with the interests of the State. Secondly, as fishing yields vary during high and low fertility years and seasons -linked to periods of droughts and rains -one can argue that local focus on IUU fishing is higher in the dry season than during the rainy months after the annual fish ban. Nonetheless such a consensus is of essential importance for the FMP to ensure compliance and cooperation with the monitoring and surveillance of illegal fishing. For instance, the high cost of lake patrolling -due to the cost of fuel -makes the inclusion of multiple stakeholders to share patrolling costs interesting. This is why the Council, VMCs, FAs, DoF, the police and the Immigration Office, support current patrols. Yet, this still does not suffice to cover the costs for regular patrolling activities. This is mainly due to the currently malfunctioning system of license fee and fish levy collection and to the historically high cost of fuel. Enforcement capacity to restrict illegal fishing therefore stays low, as regular patrolling is only possible during the annual fish ban. Furthermore the removal of the illegal settlements on Kanakashi and Potolo Island requires the cooperation of the resettlement scheme office in Mansa. This has political consequences as politicians usually do not dare to take the unpopular measures of forced resettlement, out of fear of losing votes (DoF Officer, personal communication, September 30, 2011;Fisheries Management Advisor, personal communication, September 26, 2011). The TAs, who are heavily represented in the co-management structure to enforce measures but have to render account to their community as well, also fulfill an ambiguous role in this (Wilson et al., 2004). 12 The inability to stop illegal fishing in Mifimbo is potentially threatening for overall compliance with the FMP's conservation measures. As a DoF officer (personal communication, September 30, 2011) phrased: "if we fail to remove the people from Mifimbo, then why should others stop fishing illegally?"

Local representation and the illusion of community
Yet it would be over-simplified to address non-compliance solely to malfunctioning enforcement capacities. Within the FMP, the increase in illegal gear use and illegal fishing is linked directly with a trend in declining catch rates (Mölsä, 2009). However they admit as well that no systematic catch monitoring has taken place since the early 1990s. Yet illegal fishing is linked to decreased catches per unit, which can be as much a sign of increased competition as of biological overfishing. This study argues 9 Mölsä (2009): according to the FMP, fifty per cent of the licensing revenue has to be allocated to the FMC, who divides the money between the different sub-committees. However, this allocation system is still not in place as the money stays with the DoF. This causes a lot of irritation within the different committees (Fisheries Management Advisor, personal communication, September 26, 2011;Chairman VMC Kashikishi, personal communication, October 6, 2011). 10 Mölsä (2009): the FMP limits the number of nets per unit to eighty nets of fifty meters length. 11 DoF Officer, personal communication, September 30, 2011: in 1989 a Land Resettlement Scheme was however proclaimed, which makes it possible for anyone to acquire land from the government. Yet a lot of people are not aware of this or prefer to stay in the fishing camps as land allocation would force them to move out. 12 For a further elaboration on the problem of policy enforcement in the Mifimbo breeding area, see Annear (2010), Navigating Constricted Channels: Local Cooption, Coercion and Concentration under Co-management, Mweru that Mweru-Luapula's fisheries management is based on not only a rationalist model of societyconsisting of independent and economic agents -but also on a rationalist model of compliance and deterrence. This leads to a management analysis that goes as follows: if sanctions are higher than gains, compliance increases (Andrade & Midré, 2011). This means that co-management -in the case of Mweru-Luapula -is merely used as a management tool aimed at increasing compliance with policy goals that are still created above the decentralized institutions. This argument is supported by grieves of local community members who see the VMCs as nothing more than a prolonged arm of the DoF instead of decentralized institutions with decision-making powers.
However, although I believe the creation of co-management structure to be based on this underlying rationalist idea, it would nonetheless be over-simplified to ignore complexities of local politics and power that affect issues of local legitimacy and representation. The latter is possibly challenging to Mweru-Luapula's fisheries management.
Within PLARD, the local management committees strive to be a platform for a solid partnership between many different stakeholders (Mölsä, 2009). Membership of the VMC board is obtained through local elections, which is a system that encourages community involvement. However, the current VMCs are lacking such adequate community representation. This is due to several reasons among which the aforementioned low popularity of such committees that are often seen as DoF extension pieces. In 2010 the VMC office in Kashikishi was even burned down. The main reason however seems to be the local electoral system itself, which favours the most respected community members (Fisheries Management Advisor, personal communication, September 26, 2011). This turns out to exclude certain community layers based on profession (e.g. the VMC of Nchelenge currently has no fishermen present), age, gender and origin. The latter is an important factor as certain species -like chisense -demand a high level of mobility from fishermen who therefore do not really belong to any community and lack representation in the VMCs. 13 In fact, the electoral system is based on a false pretence of community. For instance, the national director of the DoF labels Mweru-Luapula's fishermen as indigenous, which moves right past above mentioned rural-urban linkages, the migration movement and fishermen's high mobility (Director DoF, personal communication, September 22, 2011).
The issue of unequal representation -e.g. practically all Nchelenge VMC members are elderly businessmen who grew up in Nchelenge -potentially challenges the co-management structure. As Svein Jentoft (2009) posits: "Stakeholder groups possess unequal powers, for instance due to their privileged representation in the co-management regime. This means that some groups are capable of making their views more forcibly than others, in some instances even to the point that they can control the management system" (p. 30).
This potential threat is based upon the role of power. Every human social system is after all characterized by more and/or less powerful stakeholders. Amartya Sen (1981) describes this by means of his capabilities approach, where access to different assets -human, social, physical, financial, naturaldefines one's wealth or poverty. This is never a static reality but rather a dynamic process that constantly (re)creates and distributes access to assets through socioeconomic and political networks. Such networks are therefore essential and "carry power relations in both the patterns and processes of connection" (Goldman, 2011, p. 215).
The role of such networks has proven to be essential within Mweru-Luapula's fishing economy, as was illustrated by the credit-and patron-client relations within both the colonial as postcolonial commercial fishery. The next chapter builds on the role of such networks within the current fishery. This is done by a fish chain analysis focusing on interactions between different economic stakeholders within the fishing economy. Such an analysis is useful to see how broader trends -like population pressure, growing investment and fluctuating fish prices -actually take place within a local setting. The works of Platteau (1989Platteau ( , 1987 on credit-relationships and the role of power in interactions between different agents -e.g. fishermen, fishmongers, fish traders -have proven to be an essential theoretical guideline for following chapters.

Social change: the political ecology of inequality Market, competition and the rationale of credit
Mweru-Luapula's fisheries are characterized by a high number of artisanal fishers (Mölsä, 2009). Although the nature of fishing has become more investment-driven, it is hard to analyze it in terms of a modernisation process at first sight. Small-scale organizational forms seem to persist. The nature of the fishery is however largely commercial -according to a PLARD survey, eighty-four percent of annual yield is sold -which makes the market a determining factor for Mweru's fishery. Price fluctuations -the last years one has witnessed an increase in fish price -are therefore explained as much through decreasing catches per unit as through the large urban demand that -according to simple economics -gives rise to prices (Musumali et al., 2009). Within the FMP, this economic logic of the market is yet a mere context for the set-up of their program, focused on ensuring environmental sustainability. 14 This chapter will, by focusing on economic relations between fishermen, dealers, processors and traders, analyze this economy of the market in depth. 13 A study of Lwenya, Yongo, Abila & Numan (2008) on the impact of movement of fishermen on fisheries management in Kenya suggests that this is an issue that challenges decentralized fisheries management in general as well. 14 As a community facilitator in Nchelenge explained, this is indicated by the change in denomination of Mweru-Luapula from a 'commercial fishing area' to a 'fisheries management area' (personal communication, October 7, 2011).
A common view on the commercial nature of small-scale fisheries witnesses a system where private traders get involved directly only at the later stages of the marketing-processing fish chain (Platteau, 1989). This means that their only direct concern is the buying, transporting and selling of already catched and/or processed fish. Their involvement with the fishing itself is however often indirect, e.g. through credit-links. In fact this system is a blueprint for the 1970 and 1980 pale fishing economy that relied heavily on credit ties between large trade patrons and small-scale fishermen.
The rationale of credit has multiple origins, linked to the institutional set-up and the economic and ecological context. The nature of Mweru-Luapula as an allotrophic lake that relies heavily on meteorological conditions makes catch predictions highly uncertain. Credit ties therefore act as an insurance mechanism for both parties -the individual receiving credit and the person lending out creditinvolved. For the fishermen unpredictability of catches means a high variability of fishing income, which exposes them to substantial risk and uncertainty. Reciprocal credit can therefore function as a hunger insurance mechanism that reduces the risks in fishing and avoids disaster. The same uncertainty stimulates fish traders to establish such 'credit-cum marketing' relations that serve as an output-securing device, meaning that contractual relations that are more lasting are preferred over open-market relations that are single-purpose and short-term arrangements. For the private traders this means that access to a stable supply of fish is guaranteed.
The context of credit relations in Mweru-Luapula is rather found in a fear of losing income than in a desire of maximizing gains. Regarding poverty, this means "it is scarcity and not sufficiency that makes people generous, since everybody is thereby insured against hunger" (Evans-Pritchard, cited in Platteau, 2000, p. 468). On the other hand, credit-cum marketing relations separate fishermen from market effects as well. Platteau and Abraham (1987) clarify that a stable source of credit makes fishermen lose possibilities of maximizing incomes ('hidden interest'), for example by selling under the formal price which would have been set by the market ('hidden costs').
The nature of credit-relations in Mweru-Luapula's fishing economy depends -besides on the allotrophic nature of the resource -on the type of fish species targeted as well. Within chisense fishing labour-tying credit relations between gear owners and crewmen exist besides output-securing arrangements between traders, fishermen and aforementioned abakukungwisha (Chisense fisherman, personal communication, October 13, 2011). Chisense fishing depends heavily on peak operations during the year, while pale fishing is a somewhat more constant economic activity. Craft owners try to enlist dependable crewmen in advance of the season by using credit contracts -e.g. cash advances -involving a long-term commitment (mostly one fishing season). Quick and ready availability of labor is essential in chisense fishing.
As explained, credit relations were already present in the colonial fishery but proliferated from the 1970s onward. As Gordon (2003) explains: "Patronage networks that rested on the distribution of goods in exchange for the promise of future labor remained a feature of the colonial and postcolonial economy through a new class of fish traders. Wealthy traders supplied fishers with nets and other goods and thereby created indebted dependents, who promised them their catch" (p. 20).
Reasons for this were not only ecological or environmental, but stemmed from an institutional context as well. With the Zambian government losing grip on the Luapula valley in the 1970s, the formal credit market in the region collapsed. Government loans that had helped Mr. Kunda's fishing business were no longer available. Credit availability was therefore to be found within the fish chain itself. Informal money lenders did exist but they gradually disappeared (Businessman and former moneylender, personal communication, October 7, 2011). This is mainly due to the increasing mobility of fishermen that affected mutual trust to engage in money-lending activities. Furthermore credit relations based on networks of family and kinship do exist as well, but these function more as a support mechanism than an economic alliance. Yet their role can not be underestimated, as most respondents in this research at one time or relied on financial aid from friends or relatives or provided them with money. Above mentioned creditcum marketing and credit-cum labour arrangements are thus the outcome of imperfections in -and even the absence of -alternative sources of credit as well. 15

The irrationality of credit
Those labor-and output-tying credit arrangements within the fishing economy do not only act as an indicator for growing economic competition, uncertainty and institutional change. Although smallscale organizational forms of fishing persist, they are incorporated within a larger economic system, which affects traditional production relations. After all, the credit arrangements do not only function as a mutual insurance mechanism. The economic alliances that originate through such contractual arrangements also channel market preferences, e.g. popularity of salted and dry fish, to the local fishermen. Furthermore these financial bonds to a great extent affect the local markets on the shore of Lake Mweru, because less fish is being sold to local merchants.
The role of power is essential in describing credit-relations in Mweru-Luapula. Mechanisms as subordination and marginalisation are present in every aspect of it. Its role as an informal insurance mechanism for both parties involved cannot omit the fact that it comprises negative dependencies too. Fishing catches depend not only on competence but also on luck (Platteau, 1989). A bad fishing season can lead to indebtedness for fishermen, which can result in a loss or selling of assets. This is a risk peculiar to credit arrangements with (semi-) urban traders: they can be mutually beneficial but the hazards are bigger for the debtor. Moreover not every fisherman is able to obtain credit. In remote areas like the fishing islands in Lake Mweru the presence of fish traders can be very limited, which means less competition between traders hence less incentive for those traders to establish credit-cum marketing relations. 16 Consequently it is more difficult for the fishermen on these islands to gather capital to purchase fishing nets and boats. These unequal proportions also exist on the other marketing side of the fish chain. More locally based fish traders are generally less well provided with capital. They even often have to delay payment to fishermen until they have sold the fish themselves. This reverses in fact such credit-cum marketing arrangements. Fishermen therefore usually prefer urban and more large-scale fish traders over local dealers.

Coping with moral hazard: from the emergence of a new class…
Mweru-Luapula's fisheries have evolved notably over the last decades. This has challenged creditcum marketing relations between large-scale traders and fishermen up to a point that they are almost nonexistent nowadays on Mweru's shoreline. This is due to several reasons. Firstly improvement of road infrastructure and the introduction of the mobile phone has improved regional integration of fishermen and especially increased their mobility. For large-scale dealers this is risky because possibilities of freeriding by fishermen consequently grow. This increases costs of business transactions -and makes them more risky -for the traders. Such an evolution is further strengthened by a growing competition between the traders as well. Platteau (1989) in this regard points to risks of 'moral hazard', meaning that external fish traders are not in a good position to assess the trustworthiness of producers, to establish social bonds with them and to pressure them.
These risks of moral hazard have only increased the last decades to a point where large-scale traders had to change their strategies. A class of local intermediaries between fishers and traders came into being. They are referred to as abanamakupwila, i.e. female fish traders who act as middlepersons between fishermen, local consumers and large-scale fish traders. These petty (headload) traders can generally be considered as 'autochthonous', meaning that their relations with fishermen are often founded on ties of caste, residence, kinship and custom (Platteau, 1989). Although these relations are very important, they are mostly beneficial in addition to credit-relations. Competition between these abanamakupwila is very high, so provision of credit -mostly through cash-in-advance arrangements for one bucket of raw fish -and even fish-for-sex arrangements with fishermen are not exceptional (Annear, 2010;female fishmonger, personal communication, September 30, 2011). As could be observed, when the fishermen arrive on shore, these female traders rush towards their boats to be the first to buy fish. Intensive discussion and conflict between them is a usual consequence. The advantage of pre-made arrangements can therefore be noticed easily.
Their existence and origin is however ambiguous. On the one hand their presence secures a certain supply of fish for the local communities surrounding the Lake. On the other hand they act as a necessary go-between for large-scale fish dealers with no local background. For many large-scale dealers it is safer to make arrangements with these local female intermediaries because they usually buy, process and sell their fish at the same local market. Yet it is also more expensive. Huge competition between the female fishmongers makes them raise the bid as well, which consequently affects the people buying from them. Therefore, fish merchants increasingly try to circumvent these middle(wo)men. One strategy is 'employing' a female trader, e.g. through marriage, or establishing social bonds with them. This can pressure prices down -even level it out in case of marriage -or ensure a stable supply, because all abanamakupwila respondents admitted they prefer selling their fish stock to familiar or friendly traders (abanamakupwila, personal communication, September 29, 30 & October 12, 2011). Yet another tactic consists of competing directly with these female fishmongers, by buying directly from the fishermen on the Lake. 17 This 'price competition' affects these female traders to a great extent and even influences food security around Lake Mweru. As many are no longer able to afford full-grown fish, alternatives have to be sought. As one respondent mentioned, "local people rather buy beef because it is cheaper now" (Chisense fisherman, personal communication, October 13, 2011). Even more, local preference for cheap fish is being channelled through the abanamakupwila to the fishermen. This stimulates the targeting of juvenile fish -also known as icibalebale -which is sold by these female traders in the local markets of the villages surrounding Lake Mweru. Illegal fishing methods are therefore not only an outcome of competition between fishermen. It is driven by the local merchants to secure a source of income and in general by households in nearby villages who can no longer afford to buy adult pale or breams.

…to its elimination: the vertical integration of large-scale dealers
Generally, connections between large-scale dealers, local intermediaries and fishermen are not reliable. One day fishermen may prefer to sell to external traders for a better price, the next day they can promise to supply local traders related to them by kinship or custom. Transaction costs are therefore usually high. This reality is currently stimulating these urban traders to enter the production phase of fishing themselves, instead of intervening at the processing or marketing phase. This has led to the observation that many individuals are combining fishing and trading activities nowadays. It is important to understand that this is a process of vertical integration instigated by the traders themselves, and not an expansion of occupation by the fishermen. Within the fish chain it is thus an evolution moving top-down and not bottom-up. In order to obtain sufficient supply, it is more profitable for the traders to invest in fishing boats and nets and to employ fishermen than to buy from them. This vertical integration affects traditional production relations. While usually relations between crew members and gear owners rested on family ties or kinship, more formal and contractual production relations -like in the chisense fishing economy -are emerging. Regarding the fishing unit this means that a fishing boat is nowadays rather considered as individual property than an instrument of cooperative labour (Platteau, 1989).
The proliferation of leleke fishing on Lake Mweru is a manifest illustration of this. Leleke fishing entails a manner of open water fishing and processing by means of anchor boats, called ifyombo. From these boats the fish is sold to urban markets in the Copperbelt, Lubumbashi and Lusaka. The boats are also used as bases for sleeping and cooking for the fishermen, who stay on the Lake for months. The ifyombo on the Zambian side of Lake Mweru are owned by both Zambian as Congolese businessmentraders, while most of the 'fish laborers' on the ifyombo are Congolese. The anchor boats are usually guarded by attendants who see to it that no fish is stolen from the vessel owner. Fishermen are paid fifty per cent of the catch per boat. Their relationship with the owner-trader is purely businesslike, but certainly not on equal footing. Living circumstances on the ifyombo boats are harsh: twenty to thirty fishermen have to live on a vessel not larger than thirty square meters, without any sanitary facilities. Outbreaks of cholera are consequently not uncommon on these boats. Moreover ifyombo are usually not well build or maintained. This has cost forty Zambian and Congolese lives in March 2011 (Online Article Zambian Watchdog, March 2, 2011). Since the implementation of the Fisheries Management Plan has started, leleke fishing is now formally prohibited, because of its semi-industrial nature -it happens more than a thousand nets are employed to provide one anchor boat. Yet, current patrolling units only target the illegal Congolese fishermen-employees while the ifyombo owners remain unaffected. One can notice easily that leleke fishing -as it currently takes place -contains strong mechanisms of subordination and marginalisation.
The vertical integration of large-scale traders further threatens the position of the small female headload traders, as a large quantity of fish moves directly passed traditional landing sites towards the urban markets. The absence of reliable statistics on this 'invisible' fishing economy is problematic as it challenges any specific environmental and developmental management policy. The FMP does plan on installing fixed landing sites to control price proliferation, but it is questionable that this policy will end the invisible and illegal trade.
Summarized, the increasing vertical integration into fisheries production phase by private traders resulted from problems on supply side, enforced by risks of moral hazard and high prices. Logically this increases overall fishing effort. As one Congolese large-scale fish traders and gear owner explains, if supply was sufficient and more affordable, vertical integration would stop (personal communication, October 7, 2011). He even states that "the more fish we have in the Lake, the less fishermen would be involved". From a deductive point of view one can understand this point of thought, yet it is almost certain that demand will always surpass supply in Mweru-Luapula's fishery, making it highly unlikely that fishing effort will decrease.

Assessing inequality: witchcraft and the defense of a moral community
If one places the -often told -story of reduced catch that threatens livelihoods for small-scale fishermen and local traders next to this process of vertical integration -comprising capital investment in fishing gears, units and crew members -one can notice that Mweru-Luapula's fishing economy still contains both aspects of vulnerability and opportunity. The narrative of a collapsed commercial economy -exemplified by the remnants of former ice plants in Nchelenge and Kashikishi -is simultaneously countered by the initiative a Congolese trader-owner has started to build his own ice plant for transport to the urban markets (personal communication, October 7, 2011). Even more, the contrast between the most impoverished and vulnerable households and such elite as the Congolese trader has only gotten bigger.
Currently no study offers quantifiable data to prove this emerging inequality, yet an interesting indicator does exist. It is found in the rapid increase in witchcraft accusations, an observation that I believe should be treated as a sort of social diagnostic rather than a traditional or customary belief system. This is not just a local phenomenon. The latest decades, research in Africa has reported an ongoing increase in witchcraft and sorcery accusations (Moore & Sanders, 2001). One can relate this to expanding opportunities for socioeconomic upward mobility and local responses to this. This is excellently described by Moore and Sanders (2001): "popular belief about witchcraft are seen as a form of political action from below, where witchcraft accusations are used as levelling mechanisms preventing social inequalities from developing beyond the point of community control" (p. 18). 18 The connection between wealth accumulation and witchcraft is extensively elaborated (Ashforth, 2001;Sanders, 2001;Rasmussen, 2001). Exposure to market economics and capitalist entrepreneurship created possibilities of wealth accumulation that lie beyond community control. Responses to these changes in wealth distribution, by means of witchcraft accusations, in that sense function as a defense of a 'moral community', based on senses of economic justice and responsibility (Gordon, 2003). Yet within this research witchcraft accusations are interpreted as a social response to change rather than a desire for a past moral order. Witchcraft accusations have historically in Mweru-Luapula been a symptom for change: the disappearance of mpumbu made African fishers accuse European fishermen of using witchcraft, the arrival of urban chisense fishermen was -and still is -accompanied by a cultural idiom that insinuates only sorcerers catch chisense (Gordon, 2003;chisense fisherman, personal communication, October 13, 2011). Changes in wealth and its distribution are explained through such narratives using witchcraft and invented traditions.
In Nchelenge and Kashikishi, where this research took place, the current relevance of bad magic and sorcery was confirmed all respondents -i.e. fishermen, small-scale traders, abanamakupwila, urban traders and institutional as well as traditional legal authorities. The role of the latter is in fact interesting and possibly challenging for the co-management regime regarding fisheries. As chiefs and village headmen use witchdoctors to verify if one is practicing bad sorcery, government policy sees witchcraft accusations as a legal offence (Village headman Nchelenge, September 30, 2011;Chief Kambwali, personal communication, October 12, 2011;Police officer, personal communication, October 5, 2011). Not only does this expose an ambivalent relation between the state and culture or tradition, invented or not, it is also capable of creating the perception that the state -by outlawing witchcraft -tries to sustain social inequality and protect those who allegedly used sorcery for their own enrichment (Moore & Sanders, 2001). The inclusion of those two legal approaches to witchcraft in the co-management structure can prove to be restrictive for the enforcement of fisheries policy.

Conclusion: closing the gap
Throughout the entire production chain -from the ecological environment to harvest and postharvest processes -one can find both challenges to fisheries as opportunities to improve fisheries governance (Chuenpagdee & Jentoft, 2011). An intersection approach regarding fisheries management should therefore be properly examined and understood, rather than merely focusing on the bio-physical environment and the harvest system.
Social and economic relations between actors within this fish chain contain aspects of power, whereby processes of subordination and marginalization occur. The use of credit arrangements is crucial in this. Indebtedness to middlepersons of traders may drive fishermen to use destructive fishing methods to increase catches, in their turn damaging fishing habitats and overall productivity. Moreover, demand for juvenile fish by households seeking a cheap supply of food is channelled through such economic networks. A management policy that focuses heavily on the MCS of IUU fishing therefore moves right past these social and economic constraints -and incentives -that lead to one fishing illegally. One can regard the removal of illegal fishers by patrolling as mowing the lawn: it does not prevent the grass from growing back. In other words it is a simplism to approach compliance merely as proportionally connected to enforcement capacity.
This research attempted to confront the environmental knowledge base -that constitutes the current management system -with a relational and historical perspective on the social and economic fisheries system. As mentioned, the vision of the FMP relates an open-access nature of Lake Mweru to population, resource availability and poverty. This perspective is relational only in the sense that one's wealth or poverty is directly related to the quantity of fish in the Lake and lagoons. From both a historical as contemporary analysis of the political ecology of Mweru-Luapula's fishery, one can however question if poverty is not rather a distributional problem than a consequence or cause of overfishing. It is essential to see how power works in small-scale fisheries: it is an instrument that defines socioeconomic relations in Mweru-Luapula and determines distribution of benefits and losses. No fisheries system is characterized by equal and independent agents whose wealth or poverty depends only on the amount of fish they catch, process or market: "one needs to be sensitive to the fact that one group may draw benefits from the other in ways that indicate subordination and marginalisation" (Jentoft & Midré, 2011). The relational perspective on fisheries taught us that growing competition throughout the whole fish chain has modified traditional production relations to an extent that inequality only increases more. This 'penetration of capital' -as Platteau (1989) described, is a vital factor in the creation and recreation of wealth and poverty (Bavinck, 2011). Briefly worded, the analysis of the political economy of exploitation and dependence illustrates that changing production relations due to growing competition -caused by the 'predatory incursion of open-access capitalism' -instigate and sustain processes of marginalisation, exclusion and subordination, which only increase inequality. 19 These insights do not only expose superficialities in policy priorities of Mweru-Luapula's fisheries management: they also challenge co-management in itself. As power is essential in the social fisheries system, it is important to see how power works in co-management too (Jentoft, 2007). As Jentoft & Midré (2011) elucidate, "co-management is about restructuring relationships between user-groups and government authorities to facilitate commitment and mobilise resources and interactive learning. Comanagement arrangements are about creating action space and about providing the individual and collective capabilities that users need in order to fill it" (p. 50). In Mweru-Luapula's fishery, this political action space for fishermen, traders and other agents is not a static but competitive space where confrontations between stakeholders occur. For example, the removal of the abanamakupwila is an item on the agenda of the local co-management institutions as well as the Mweru-Luapula Fishing Association (FA chairman, personal communication, October 6, 2011). Their exclusion will benefit large-scale traders but cause harm to local markets and small-scale fishermen who sell their fish exclusively to these middlepersons. Co-management arrangements therefore often benefit not the poor, but rather those who are already better off. In the context of inequality, more powerful users will always try to renegotiate institutional regulations in their favour (Haller, 2008).
Generally, this research argues that institutional governance in fisheries development should always be built upon, or start from, the many dimensions (social, ecological, political, economic) of a fishery system that are shaped not only locally but through a constant interaction with external factorsacting as both threats and opportunities. This is necessary to close the gap between locally based understandings, policy and legislation. Too often conventional management is based on general paradigms regarding common-property theory and shared resources. This one-size-fits-all approach characterizes a lot of small-scale fisheries management programs in the global South (Hauck, 2011). The trend for developed countries to assist developing countries with IUU fishing challenges -like PLARDdoes not seem to help a necessary differentiation process in global management of small-scale fisheries.