4. DIVERSIFICATION AND NATURAL RESOURCE ACCESS: POLICY ISSUES ARISING

Little systematic thought has hitherto been given to the linkages between diversification and access to natural resources, although some suggestive strands emerge from the preceding discussion of the dynamics of diversification. There are some relevant distinctions and categories that need to be kept in mind before giving a view of such linkages, and as a precursor to considering the natural resource policy implications of diversification. One such distinction is between natural resources that are mainly under private control as against common pool or common property resources; another is that livelihood diversification can to some degree occur within a natural resource based context, as well as outside it.

For most purposes, the distinction between private and common property natural resource regimes comes down to considering cultivated land access separately from that of resources such as grazing, forests, fishing and wildlife that nearly always take the form of social rather than private access^[2]. This distinction is of course by no means watertight - irrigation systems, for example, typically combine common property features with private land access and grazing land can be either communal (rangelands) or private (ranches) - however, it provides a useful first distinction. Even though tending to be utilized privately, cultivated land access involves multiple different forms of tenure in different settings, with customary tenure remaining widely prevalent in Sub-Saharan Africa. Tenure systems constitute the institutional rules that govern access to land, and, as such, they form part of the institutional and policy context in the livelihoods framework.

There has been a worldwide trend over the past fifteen years to change the management regime surrounding certain types of common property from state control to so-called community-based natural resource management (CBNRM). This may take the form either of co-management, implying joint responsibilities by government and community-based organizations (CBOs), or of the more wholesale devolution to communities themselves. The resources involved have been forests, fisheries and natural parks (wildlife and tourism), formerly mainly state owned and administered. Other common property resources e.g. grazing, watersheds, watercourses etc have in the past tended to be under local jurisdictions in any case, although increasing pressures imposed upon them have often resulted in CBNRM being advocated to replace previous social regulatory systems that are considered impaired or disintegrated. Like land tenure systems, CBNRM regimes may be characterised as institutions; although in this instance they are new institutions, often created by outsiders, that sit uneasily or supplant pre-existing social conventions for dealing with conflict over access to common resources. Proponents of CBNRM argue that pre-existing or ‘traditional’ institutions for common pool resource management may not be adequate to ensure sustainable resource use in the face of increasing population pressure and monetization of the rural economy, making their replacement with more relevant new institutions justifiable.

Livelihood diversity can clearly occur via direct utilization of land and other natural resources, as well as from their indirect use (e.g. trading or processing NR products) in addition to non-NR diversification. Indeed, it is essential to consider diversification as covering this entire spectrum, since in practice diverse rural livelihoods make full use of all the opportunities available. Land may be used for one crop or multiple crops; low value crops can be replaced by high value ones. The use made of different common property resources will depend on other components of the livelihood system (e.g. livestock ownership), as well as on coping strategies at times of crisis (intensive collection of wild foods is observed as a response to acute food insecurity). The linkages between resources accessed by a single household and those accessed by multiple households are especially important, as illustrated earlier for hillside agricultural systems in Nepal (see Box 1).

Whereas the opportunities represented by non-NR diversification are theoretically unlimited (even though constrained in practice), the increased income that can be derived from an existing area of land, or patch of forest, or grazing zone is limited by current technology, resource depletion, market constraints, or some combination of all of these. Nor is ‘human carrying ‘capacity’ static; in most poor rural areas rural population growth on its own ensures increasing land sub-division, or expansion of the cultivated area into formerly protected areas, or increased off-take of trees for charcoal etc., despite rapid rates of urbanization. Diversification thus occurs in part as a response to natural resource scarcity, and in such a context may be considered beneficial to resource sustainability.

An example of diversification and its drivers comes from Western Kenya, where, on the shores of Lake Victoria, livelihoods have traditionally combined farming, livestock herding and fishing (Geheb and Binns, 1997). Declining land availability and removal of subsidies on agricultural inputs under structural adjustment, together with new opportunities in Nile Perch fishing have combined to increase pressure on Lake Victoria’s fishery resources as lakeshore households invest more in this sector. As profit margins in fishing diminish under heavy exploitation, further diversification, in the form of migration to urban areas or non NR-based activities is occurring (Freeman, Ellis & Allison, 2004).

It is tempting in the light of the foregoing to see diversification as running in linear fashion from natural resource scarcity to rising reliance on non-NR livelihood components. However, this is manifestly not always the case, as demonstrated by degrees of reliance on natural resources and yield gradients across income classes as observed in Tables 1 and 2 above. The rural poorest are typically both the most dependent on natural resources and those with least access to land. Meanwhile, the better off are both less dependent on natural resources and have best access to land, and, in addition, obtain the highest returns from a given unit of land. There are what Gunnar Myrdal (1968) years ago called ‘circular causations’ rather than linear relationships between diversification and natural resource access.

To take the links between diversification and natural resource access further it is necessary to disaggregate between different classes and types of natural resource. This is necessary anyway, since different institutions and policies apply to land as compared to common property resources, and within the latter, fisheries, for example, play a different role in the livelihoods of those with access, than do trees and forests. While disaggregation is inconvenient it should result in sharper perceptions, and patterns of contrast and similarity can emerge from looking at different cases in parallel.

Given the overall desirability of diversification, as concluded in the last section, the question needs to be posed how do land tenure arrangements promote or inhibit diversification, both of the on-farm type and the non-farm type. At least part of the answer is that insecurity of tenure, including insecurity in the hereditary transmission of tenure across generations, considerably inhibits both what is done with the land when it is under a particular stewardship, and the options open for its allocation as different livelihood alternatives present themselves. There is in addition a powerful gender dimension to all aspects of land management within diversified livelihood systems, since women often have a subordinate role in land use decision making despite being the members of the family most likely to be left behind when male members migrate to take up work opportunities elsewhere (Kothari, 2003). In addition, women rarely have ownership rights over land, and even more rarely are allowed to inherit land through customary or legislative forms of land inheritance.

Inflexible land tenure institutions typically act as a barrier to the optimum deployment of land as a resource in the widespread circumstances that have been described whereby people do best when they have the greatest ability to respond flexibly to new opportunities both near and far. Under customary tenure systems it is not at all unusual to see land lying idle adjacent to other areas that are intensively farmed. This occurs due to the absence or poor development of a rental market in land, so that a family that has lost its interest or ability to till the land is unable to make that land available to another family due to the fear of losing it altogether. Other reasons for the same phenomenon (idle or poorly-utilized land in the presence of acute land shortage) include family members working on other farms for immediate food security reasons, family members engaging in seasonal or circular migration, and inability of women to make full use of the land when male household heads are absent due to not having the decision-making capability to do so. These and related situations would occur considerably less often in the presence of a workable legislative framework that ensured security of land ownership at the same time as facilitating the development of a rental market in land.

There is, of course, also a much longer understood reason for land to be underutilized, and this occurs when land is held for purposes other than production by the (often absent) wealthy in rural areas exhibiting highly skewed land ownership distributions (Berry & Cline, 1979). The problem here differs from that in the preceding paragraph in that instead of thwarted motivation to make land available to others through rental, there is lack of motivation on the part of owners to cede or rent out land. A huge literature on land reform deals with the complex policy issues of changing land ownership distributions, and this is not entered into here. The immediate problem of leaving land idle can, at any rate in principle, be tackled by taxation penalties on idle land, and provision for such penalties exists in the land legislation of many countries, although how often they are applied in practice is beyond the scope of this paper.

Land tenure reform must chart a perilous course between, on the one hand, improving the access and security of the poor to land, and, on the other, not opening up the floodgates of ‘land grabs’ by the local rich and powerful when the opportunity to do so arises. Recent legislative changes put in place in the countries that were part of the livelihoods study in four African countries on which this paper has already drawn have tended to be too timid to help much with the land security pre-occupations of the poor and women, while at the same time just opening up the freehold option enough to permit highly unequal ownership distributions to emerge for ‘new’ land made available through the sale of former state-held land in plantation agriculture (Cross, 2003). In particular, the weak status of women with respect to land access and inheritance persists in all new legislation, with policy makers and legislators unwilling to enshrine land inheritance by women in law.

For women, the inter-relationship between land access and livelihood diversification is a particularly disadvantageous one in the Sub-Saharan African context in which male migration and male non-farm work opportunities are more prevalent than female ones. The position would doubtless be different in economies experiencing rapid growth in the demand for female labour in the manufacturing sector. In SSA, women tend to be left behind on the farm, but with little real scope for decision-making over farm resources and limited ability to generate cash on own account. Households with female heads (especially those resulting from widowhood or divorce) are often discovered to be amongst the poorest households, and in some rural societies both widowhood and divorce results in immediate reclaim of land by the male side of the family. Of course, this is not always the case, and of course, as much of the gender literature points out, women do exhibit agency within the constraints they confront. Nevertheless the structural circumstances of women with respect to both land and livelihood diversification is particularly debilitating for them in SSA.

At the risk of over-generalization, the links between livelihood diversification and land access may be summarised as follows (Ellis, 2003a):

(1) land continues to play a critical role as a key asset within a diversified rural livelihood, and indeed diversification can enable improvements to the quality of land and its productivity that are not possible in the absence of the financial resources that it makes available;

(2) nevertheless, the nature of this link varies according to differing circumstances and across income groups, with the poorest typically least able to achieve this virtuous synergy due either to lack of complementary assets (for example, low human capital) or poor security of land tenure, or some combination of both these factors;

(3) poorly functioning or outmoded land tenure institutions that inhibit the development of a rental market in land hamper people’s diversification options and can result in inefficient utilization of land as a resource;

(4) in many rural areas characterized by widespread poverty and vulnerability, women are particularly disadvantaged jointly by diversification (absentee male family members) and customary constraints on land use decision making, despite at the same time often being responsible for ensuring the food security of the household;

(5) approaches to land tenure reform often take a myopic view of land security and the development of a rental market in land, providing little or no scope for improving the land security of women or for creating workable land rental regulatory institutions at local levels.

Some different and some similar arguments apply to the links between livelihood diversification and common property resources, including former state-owned resources that are devolved to community management or co-management under CBNRM policy initiatives. Again here, the general question must be posed whether CBNRM regimes on balance encourage or inhibit individuals and families from deploying their resources flexibly in order to construct less vulnerable and improving livelihoods. All common property type resources represent potential assets on which rural livelihoods can be built, although the significance of such assets within livelihoods in general, and for the poor in particular, vary considerably across resources and locations. The chief motivating forces behind CBNRM regimes have been to do with equity of access (ensuring that the poor have access as well as the better off) and sustainability, linked to resource conservation concerns.

A great deal has been written on CBNRM and space does not permit here more than the briefest outline of some critical features and arguments. The CBNRM approach originates from two main directions: a positive stance taken on the capability of communities to manage common property resources argued in an earlier literature (Wade, 1987; 1988; Ostrom, 1990); and the search for institutional alternatives to costly and failing top-down state management regimes over resources such as gazetted forest reserves and national parks. Mixed in with these precursors are a host of subsidiary arguments concerning the ability of local actors to participate in and take control of their own development destiny in the face of the manifest failings of bureaucratic central authority to do so (Blaikie, 2004). This has included conservationist arguments to the effect that giving communities a stake in the conservation of their own resources will result in more sustainable utilization than the ‘fences and fines’ approach of centralised management.’’

In the academic literature CBNRM has exhibited a cycle comprising initial enthusiasm (Western & Wright, 1994), followed by critical analysis (Agrawal & Gibson, 1999; Leach, Mearns and Scoones, 1999; Kellert et al., 2000), followed in turn by empirical evidence of substantive failures in practice (Campbell et al., 2001; Kumar, 2002). Policy lags far behind the debate, with enthusiasm by NGOs and other practitioners peaking in the mid-1990s just when a critical literature was starting to emerge, and governments redesigning their forestry, wildlife and fisheries policies along CBNRM or co-management lines in the 2000s just when evidence is beginning to accumulate of failures of CBNRM to achieve its stated objectives. At the centre of both the critical literature and observed failures of implementation lies the false notion of community as a homogeneous group of people with a single identity of interest (Agrawal & Gibson, 1999; Leach et al., 1999; Allison, 2004). Although some (e.g. Wilson, 2003) doubt that such simplified conceptions of community can continue to exist in the face of decades of social science scholarship on the complexities and subtleties of human relations that make up communities, evidence from the design of many donor and NGO-promoted CBNRM regimes suggest otherwise. If indeed the promoters of CBNRM programmes realised the problems with their conception of ‘community’, then they regarded such problems as being more easily surmountable than they have proved to be. It is common, for example, to promote or legislate for the inclusion of women on CBNRM committees without addressing the underlying reasons for women’s exclusion in the first place, often leading to ‘token’ participation (Agarwal, 2000).

In the context of poverty reduction, the most important failing of CBNRM regimes is that they have been prone to so-called ‘elite capture’ which means their benefits are skewed towards the better off in rural society (Platteau & Gaspart, 2003). Often promoted within a conservation-led agenda, the benefits of some CBNRM projects (especially wildlife and tourism ones) have been, to say the least, peripheral to poverty reduction in rural areas at large (Songorwa, 1999; Ashley & Mdoe, 2002). Even CBNRM initiatives explicitly intended to benefit the poor, such as Joint Forest Management in India, have been judged to have largely failed in this objective (Kumar, 2002).

From a livelihood diversification viewpoint, different common property or CBNRM type resources play different livelihood roles for different wealth groups and thus have differing implications for poverty eradication. Communal grazing is not typically subject either to former state control or CBNRM regimes, and communities therefore resolve conflicts of access by custom rather than by recent design. Forest reserves, on the other hand, have tended to be public or state owned, sometimes as conservation areas, sometimes as relatively inaccessible and unexploited resources. The problem here has been resolving the cumulative pressures on access due to the multiple roles forest resources can play in livelihoods (wood fuel, building materials, charcoal for sale etc) with sustainability and conservationist objectives. CBNRM regimes have perhaps been more popular in forestry than for any other resources, sometimes building genuinely on pre-existing institutional arrangements governing access at community level, and with a mixture of outcomes in terms of the relative success or failure of recent institutional innovation.

Fishing resources represent yet a different case again with former mainstream policy being to increase catch efficiency (improved boats and gears) for artisanal fishermen while simultaneously maximising overall sustainable yield by state-enforced limitation of access (Allison & Ellis, 2001). This approach rested on the mistaken belief that artisanal fishermen have specialized livelihoods and that they depend only on the fishery resource. There has also been a widespread and empirically unverified belief that artisanal fishing families are amongst the ‘poorest of the poor’ (Béné, 2003). Livelihoods approaches show that many of the earlier suppositions about the small-scale sector are wrong. Artisanal fishing is often a part-time or seasonal occupation, and fishermen turn to different activities when costs rise due to scarcity of fish. Fishermen are often seasonal migrants. Fishing strengthens diversified livelihoods, and those with access to it are typically better off than those who depend on farming alone in the same locations. One reason for this is that fishing is a highly monetised activity, thus providing cash that can be flexibly utilised between different livelihood objectives (Allison, 2004).

CBNRM has become popular in fisheries as elsewhere, and is predicated on an ‘overfishing’ paradigm similar to the ‘deforestation’ paradigm in forestry. However, just as in forestry it is often large scale commercial logging activities that initially disrupt and open up previously relatively undisturbed forest areas; so in fishing it is often large scale vessels using the full array of modern technologies that are responsible for resource depletion, not the artisanal fishermen with their flexible entry and exit capabilities. CBNRM in fisheries also tends to invoke principles such as territorial exclusion that make little sense for a fugitive resource.

Some of these difficulties are illustrated by empirical investigation in Malawi, utilising a livelihoods perspective, of artisanal fishing communities in fresh water lakes. CBNRM was introduced to Malawian fishing villages, first by the German aid agency GTZ, in the late 1990s, and later in generalised legislation that promoted its widespread implementation in the sector (Allison & Mvula, 2002; Allison, 2004). The institutional device that was created was the Beach Village Committee (BVC) as a coordinating and regulatory body at the community level.^[3] The negotiation and implementation of BVCs in Malawi seems to have been founded on several initial implicit or explicit CBNRM propositions:

(a) that no pre-existing indigenous regulatory mechanisms regarding artisanal fishing existed in Malawian villages

(b) that artisanal fishermen were specialised single occupation enterprises, who comprised the poorest of the poor in rural Malawi and had few if any alternative livelihood sources

(c) that communities were ethnically, occupationally, and distributionally relatively homogeneous, resulting in a common interest in management of the resource

(d) that artisanal fishermen exploiting an open access resource were chiefly responsible for endangering the future productivity of Malawian lakes

(e) that territoriality (each village having command over its own fishing territory) could provide a useful principle for focussing community management efforts

(f) that fisheries department officials, formerly charged with policing top-down fishing regulations, would take a background advisory and coordinating role with respect to the BVCs

Livelihoods research revealed an entirely different picture of how artisanal fishing operated in Malawi by comparison to this blueprint. First, in many Malawian fishing villages (especially those in the southern arm of Lake Malawi), fishing is predominantly undertaken by visiting fishermen from other parts of the country (migrants of varying durations from months to years). The migrants are a different ethnic group from the resident villagers, and as outsiders they have no customary rights with respect to village decision-making or access to resources other than the fishery (although for longer duration stays some such rights may be negotiated). Second, migrant fishermen are not the poorest of the poor, nor is fishing their only option (most of them have farms and families in their villages of origin elsewhere in Malawi). In the event of declining catches and rising costs in the fishery, they turn to other activities or return home

Their responsibility for destabilising the renewal of the most important species making up the fishing resource is thus difficult to demonstrate, particularly given recent evidence that stock fluctuations are driven more by climate-induced productivity variations than fishing-induced depletion (Allison and Mvula, 2002; Jul-Larsen et al., 2003). Larger, high value fish, of disproportionate interest to larger-scale fishing enterprises and wealthier consumers, have indeed been depleted and will require management intervention to rebuild stocks, but these efforts may need to be more focused on the activities of the larger-scale sector, including the two or three large scale trawlers that ply Lake Malawi for which already sunk large-scale capital means that fishing is continued as long as the marginal cost of being on the lake is covered (and possibly beyond that point, due to subsidy arrangements).

Third, the fish species that are preferred for smoking and trading in Malawi, the usipa and allied shoaling species, do not obey territorial behaviours; they are fugitive species and can appear in abundance almost anywhere on the lake. Therefore the fishermen must move to the fish, and typically fishermen from up to a dozen different villages on both sides of the lake will converge on places where fish of this type are present.

Fourth, it becomes quickly apparent that pre-existing institutional arrangements exist for this type of artisanal fishery even though they are not codified and written down and parcelled up in such a neat concept as a Beach Village Committee. There is an implicit recognition on the part of resident villagers that migrant fishermen bring benefits to their village. Fish are landed there, and that attracts fish traders, as well as opening up the potential for fish trading as an occupation; fishermen and fish traders spend money in the village and purchase local agricultural output for sale; in the agricultural slack season work on migrant boats can provide an important seasonal livelihood activity for members of farming families. For this reason, migrants have historically been welcome: they are allocated land on which to build houses, or beach areas for temporary accommodation. Similarly, there are unwritten codes between fishermen with respect to territory; the general case is one of non-exclusion, on the basis that allowing fishermen from other villages into the waters of a particular village also ensures reciprocal access when the fish are somewhere else.

It can be seen that parachuting idealised, blue-printed CBNRM ideas into this negotiated pattern of social and economic inter-relationships risks adverse consequences for diverse rural livelihoods. BVCs risk being dominated by village authorities that represent the interests of some resident farmers, not those of migrant fishermen; the principle of territoriality if pursued by BVCs threatens the successful reciprocal exploitation of a fugitive resource by thousands of artisanal fishermen; complex patterns of complementarity in the livelihood patterns of artisanal fishermen and resident farmers that have arisen over decades are destabilised; and it is unclear whether the regulatory functions of BVCs will achieve any greater purchase on the depletion of certain fish stocks than the former top-down regime. For example, at Malawi’s Lake Chilwa former fisheries officers were found to have taken over BVCs and used them to designate exclusionary rights over particular fishing areas.

The Malawi fisheries example illustrates some of the difficulties that CBNRMs may represent for successful livelihood diversification with its attributes of flexibility, mobility and adaptability. More generally:

(1) CBNRM regimes tend to be spatially exclusionary; they identify the resource as a territory pertaining to a particular community, however, territorial exclusion may inhibit mobile livelihood strategies;

(2) CBNRM regimes tend to assume a homogeneity of interest in the resource across community members, however, individual livelihood strategies vary tremendously and some community members will have a lot less interest or reliance on the resource than others, and the exercise of communal access may not correspond to, or may even oppose, these variations in underlying reliance or needs;

(3) CBNRM functioning will tend to represent the interests of resident community members, not those that are involved in mobile lifestyles away from the community and are therefore unable to be involved in local decision making;

(4) In certain circumstances CBNRM regimes may exclude the interests of minority ethnic groups since decision making will tend to be dominated by majority groups; similar exclusions may occur along religious or caste lines;

(5) In practice CBNRM regimes are often male dominated, and in patriarchal societies they may even strengthen the control of men over resources to which women previously enjoyed relatively unimpeded access under poorly implemented state regulation;

(6) The foregoing points imply that the equity impacts of CBNRM are unclear, and it cannot just be assumed that equity goals are met through this institutional form of natural resource management.

This implied critique of the way that CBNRM initiatives have been designed and implemented does not mean rejection of the underlying principles of locally managed natural resources. Institutions that promote democratically negotiated, locally adapted, equitable access to natural resources, their sustainable use and have the possibility to empower previously marginalized people are self-evidently a good thing. CBNRM represents a potential way forward, despite past faults in conception and implementation. Many of the faults with CBNRM initiatives can be traced to misplaced assumptions about the sectoral nature of people’s livelihoods and misplaced optimism about the ease of transforming the hierarchical societies that comprise many rural ‘communities’ into the democratic governance instruments that CBNRM requires if it is to contribute to rural poverty eradication. Optimism is an admirable quality in the context of development intervention, but if it blinds CBNRM proponents to the political realities of elite capture, it is damaging to the interests of the poor.