ESTABLISHING DEMOCRATIC DECENTRALIZATION FOR FORESTRY

Democratic decentralization entails that representative local authorities (individual or collective bodies) are entrusted with significant powers. Such a transfer of powers to local authorities is theorized to improve local decision-making efficiency and equity (Agrawal & Ribot Reference Agrawal and Ribot1999; Smoke Reference Smoke2003). Popular (by which we mean broad based or inclusive of the whole population) participation advocates make similar promises (Chambers Reference Chambers, Nelson and Wright1994). In decentralization theory, efficient and equitable outcomes are expected when powers that are relevant to local people are devolved to representative local bodies (see Agrawal & Ribot Reference Agrawal and Ribot1999). Following Manin et al. (Reference Manin, Przeworski and Stokes1999), an authority is representative when it is responsive to local needs. Representation is democratic when responsiveness is driven by accountability of that authority ‘downwardly’ toward the concerned population. It is benign dictatorship when it is driven by good will. We focus on its democratic form. By relevant powers, we mean powers that are sufficient to enable local bodies to be responsive to local needs and aspirations, and by accountability we mean counter power (for discussion of accountability mechanisms see Ribot Reference Ribot2004; Rihoy & Maguranyanga Reference Rihoy and Maguranyanga2007). Downward accountability refers to situations where local bodies can be held to account by the populace through an ensemble of positive or negative sanctions (Agrawal & Ribot Reference Agrawal and Ribot1999; Ribot Reference Ribot2004; Ribot et al. Reference Ribot, Chhatre and Lankina2008).

The distinction between participatory approaches and democratic decentralization is that the latter is specifically about including whole populations in decision making based on representative authority (Wily no date). Democratic decentralization involves the transfer of powers to democratically elected local governments. The community concerned is composed of the citizens, who are usually those who live in the jurisdiction and endowed with certain rights of belonging. Further, decentralization locates powers in permanent local government institutions (we use the term institutions narrowly to refer to organizations in general rather than ‘rules of the game’; as in North Reference North1990). Hence, decentralization can be expected to be more durable than most other interventions and ad hoc arrangements. The World Bank's community-driven development (CDD), and other participatory or community-based natural resource management (CBNRM), define the community for each intervention, such as the user group, the stakeholders, the people near the forest or the fishers. They also define the mode of representation of each ‘community’ as they go, creating or choosing to work with appointed committees, elected committees, stakeholder fora, participatory processes, ‘customary’ chiefs, project personnel, and so on (Ribot & Mearns Reference Ribot and Mearns2008). These temporary means of participation rarely constitute democratic decentralization. Further, many transfers to customary and other private bodies, neither support nor follow the logic of democratic decentralization. We focus on democratic decentralization, which is the institutionalized form of popular participation.

Attempts to establish decentralized forestry are shaped by a multi-layered political economy (Rihoy & Maguranyanga Reference Rihoy and Maguranyanga2007; Wittayapak & Vandergeest Reference Wittayapak and Vandergeest2009; see also K. Eaton, K. Kaiser & P. Smoke, personal communication 2009). They are partly about demand for change from below and largely about the institutional infrastructure to enable those demands to take shape (Rihoy & Maguranyanga Reference Rihoy and Maguranyanga2007). We focus on the institutional infrastructure or state creation of (dis)enabling spaces. In producing these spaces there are constant struggles over what kind of powers should be transferred to which authorities. The outcomes expected from decentralization depend on the resulting distribution of power. The vertical division of power between centre and local shapes the degrees of freedom and accountability of local institutions (Mawhood Reference Mawhood1983; Føllesdal Reference Føllesdal1998; Rocher & Rouillard Reference Rocher and Rouillard1998; Manor Reference Manor1999). The horizontal distribution of functions and powers among different kinds of local institutions (different in their relations of accountability), such as administrative, representative, customary, private or third sector, also shapes local decisions and outcomes (Ribot et al. Reference Ribot, Chhatre and Lankina2008).

The quality of local and national democracy also shapes democratization in specific sectors, including forestry (Rihoy & Maguranyanga Reference Rihoy and Maguranyanga2007; Wiley no date). Wily (no date, p. 18), among others, observed that ‘where devolved governance is poorly developed, empowerment of local forest management in new legislation is demonstrably constrained. . ..’ She argues that decentralized forestry requires establishment of decentralized government institutions. This is not a great surprise. Decentralization in any sector requires that some institutional infrastructure of decentralization is in place. Yet, in practice, most emerging elected local governments across Africa have constitutions that compromise their democratic or decentralized nature. Policy makers and other deciders make elected local leaders upwardly accountable to political parties and create budgets making them accountable to line ministries and legislatures. They have little discretion and can hardly be said to represent their local population. Further, most development agencies and forestry line ministries avoid local democratic governments. Even where there are strong elected local authorities, they choose to work through participatory, stakeholder or CBNRM approaches (as with CDD). Finally, line ministries create and strengthen their local deconcentrated offices. In short, forestry, by and large, avoids democratic local institutions (Ribot & Oyono Reference Ribot and Oyono2006; Ribot et al. Reference Ribot, Chhatre and Lankina2008).

Local power without representation is not democratic decentralization. Nor is representative local authority without powers. Accordingly, we should not expect improved equity, efficiency or local enfranchisement from these two common institutional configurations (powerless representatives and powerful autocrats), despite that they are often created in the name of democratic decentralization. Power transfer and institutional choice are equally important. Below, we explore why and how international and national agencies and front-line professionals avoid choosing to empower democratic local institutions.

DOES DECENTRALIZATION IMPROVE FOREST MANAGEMENT?

Empirical research on the effect of decentralized forestry on management and condition in Africa is scarce. In their review of CBNRM in Africa, Roe et al. (Reference Roe, Nelson and Sandbrook2009) found a number of studies documenting positive effects of these approaches to wildlife populations, whereas there is a striking absence of similar well-documented studies on forest impacts. Most studies on forestry report perceived changes in forest condition only, with limited hard data. Most evidence on impacts of decentralization and other participatory processes of forest management in Africa comes from Tanzania (Roe et al. Reference Roe, Nelson and Sandbrook2009), the bulk reporting that decentralized forest management conserves forest resources. The results from Tanzania's joint forest management processes are, however, ambiguous. In these co-management processes (a form of collaboration between state and communities distinguished from decentralization since powers are not formally transferred to local authorities), communities are managing under restrictive regulations set by the Forest Division that generally allow some extraction of dead-wood and low value non-timber forest products but rarely allow harvest of timber or non-timber forest products for commercial purposes. Where Kajembe et al. (Reference Kajembe, Nduwamungu and Luoga2006) and Persha and Blomley (Reference Persha and Blomley2009) found that communities managing jointly with the state did no better than the state on its own, Meschak et al. (2006) and Blomley et al. (Reference Blomley, Pfliegner, Isango, Zahabu, Ahrends and Burgess2008) found conversely that community involvement improved state management. The results from the last two studies, however, stemmed from studies of old project sites, implying that project-based high-intensity implementation may underlie the observed effects.

In relation to community-based forest management in Tanzania, where communities receive wide ranging powers over forests, the results are altogether positive (Sauer & Abdallah Reference Sauer and Abdallah2007; Lund & Treue Reference Lund and Treue2008; Persha & Blomley Reference Persha and Blomley2009; Vyamana Reference Vyamana2009). These results do, however, stem from special circumstances. While Persha & Blomley (Reference Persha and Blomley2009) report from a relatively small forest area given to a cooperative of former workers on a private estate in 1982, the three other studies report on villages that have received major attention from donors, policy makers and researchers in the late 1990s and early 2000s. Notwithstanding, however, the results indicate that the Tanzanian process, in the cases where the state has surrendered powers to the villages, has created an environment that is conducive to community engagement and conservation. One of the elements that stand out in the Tanzanian context is that forest management committees are vested in the democratically elected village government system.

Outside Tanzania, there are few studies. One study from Malawi found deteriorating forests under community management and attributed this to competition between the community forest committees and customary authorities as well as a lack of accountability of all leaders at the community level (Zulu Reference Zulu2008). A recent study from Mali showed how differences in local government support of community management initiatives were directly translated into differences in management performance and the resulting biophysical changes in locally managed forests (Benjamin Reference Benjamin2008). Like Wardell & Lund (Reference Wardell and Lund2006), this study highlighted the importance of relationships between local government and community-level institutions, and how legal pluralism creates a sphere where rights are continuously transformed through negotiation. In the Mali case, the communities that managed to negotiate good terms with local government were successful in managing their forest according to their purposes.

In Senegal, Wurster (Reference Wurster2010) found that forests exploited for charcoal managed under two decentralized forest management projects (World Bank and USAID) were ecologically no different (in natural regeneration or species changes) than charcoal production areas with no management at all. While the projects claimed to be decentralized, the forest service was still making almost all decisions (codifying them in management plans). Hence, results cannot be attributed to decentralization. Nevertheless, the case indicates that management planning, whether or not decided on by local populations, does not necessarily improve forest quality. The existence of management plans cannot be taken as an indicator of improved use.

Although management plans may not always improve outcomes, whether communities will conserve forests or not depends at least partly on rules. Autonomy guarantees little. So-called sustainable management is usually posed as a prerequisite on behalf of the state for establishing or supporting local peoples’ rights to forests. In standard practice, decentralized forest resources can be recentralized if certain environmental criteria are not fulfilled (Hobley Reference Hobley1996; Ribot Reference Ribot2004). While potentially protecting forests, such prerequisites are often used as a backdoor through which state agencies recentralize control of forests under a veil of concern for the resource status (Oyono & Ribot Reference Oyono, Ribot and Larson2005). Central forestry departments have wide discretion to retain or recapture forest management on grounds of concern for ecological (not social) sustainability. This discretion is often used, with little justification, to override decisions devolved to local authorities under decentralization (Lund et al. Reference Lund, Helles and Treue2007). There is plenty of room for abuse because (1) local managers’ access to information and decision-making fora is generally limited, and (2) the criteria for judging when recentralization is ecologically necessary, to meet higher-scale societal objectives, are often vague and ambiguous.

WHEN DOES DECENTRALIZATION IMPROVE LOCAL LIVELIHOODS?

Few studies have examined whether sustaining forests is good for local livelihoods. In general, profits from alternative land uses, such as agricultural, captured by the immediate land users, are higher than the returns obtained from forests. Whereas agriculture usually provides products that benefit the landowner or tenant only, forests provide multiple products and services that benefit a variety of actors from the local to the international level. If rural people are denied access to the most-lucrative aspects of forestry (timber and wildlife), it is unlikely that they will perceive forest as the more lucrative land use. If national governments benefit more from agriculture, horticulture or timber plantations, they are also unlikely to maintain natural forests. Thus, rules regulating who has access to benefits from different land uses underlie land-use decisions.

Decentralized forests are often sources of public revenue for local governments or user groups. Increasing revenues at community level is one of the most prominent outcomes of decentralized forest management in Africa. Although local government remains sidelined as a manager of forests in Uganda, they have become empowered through their legal entitlement to 40% of the revenues from the management of National Forest Reserves (Muhereza Reference Muhereza2006; Turyahambwe et al. Reference Turyahambwe, Geldenhuys, Watts and Obua2007). In Cameroon and Tanzania, revenues have also increased for rural communities in community forestry areas (Oyono & Efoua Reference Oyono and Efoua2006; Oyono & Nzuzi 2006; Lund Reference Lund2007). Such revenues are generally used to cover the direct costs of forest management and for public infrastructure or services, like roads, bridges, gutters, schools, health and clinics, which indirectly contribute to improved local livelihoods. Still little is known about how the benefits of improved public infrastructure are distributed among social strata in rural communities.

Empirical research suggests that whereas the wealthy in rural communities have a higher absolute forest income, the poor are more dependent upon forest income (relative to total income) (Cavendish Reference Cavendish2000; Campbell et al. Reference Campbell, Jeffrey, Kozanayi, Luckert, Mutamba and Zindi2002). In addition, forest resources have an important function in providing safety nets for poor rural households in times of hardship (Pattanayak & Sills Reference Pattanayak and Sills2001). So, forest returns to certain, often poorer, segments of communities, in the form of income and, in particular, livelihood security functions may be higher than their expected alternative income, should forest land be converted to other uses.

Conversely, however, it is often the poor and marginalized forest-dependent people who are hit hardest when centralized forest management is decentralized, and restrictions on forest use as a result become more effective (Saito-Jensen & Jensen Reference Saito-Jensen and Jensen2010). People living on the forest border are more prone to crop damage and loss of livestock owing to forest wildlife, and can experience constraints in expanding their farmland if the forest border is suddenly fixed (Schreckenberg & Luttrell Reference Schreckenberg and Luttrell2009). Further, it has been observed in both Kenya and Tanzania, that forest decentralization results in a net transfer of forest revenues from the poorest households to communities’ common funds (which are not necessarily spent on pro-poor activities) and to more well off households. This happens through locally devised taxation and licensing systems for extraction of commercial products like timber and charcoal (Lund & Treue Reference Lund and Treue2008; Schreckenberg & Luttrell Reference Schreckenberg and Luttrell2009; Vyamana Reference Vyamana2009).

Local forest committees could modify management rules to favour the poor, but the poor are often already marginalized groups with interests in forests that run counter to those of the majority and community leaders (Lund & Treue Reference Lund and Treue2008). In many cases, the livelihood improvement potential of decentralized forest management is severely restricted by the type of forests over which powers are devolved or the restricted nature of powers devolved. In Tanzania, joint forest management signifies a form of decentralized forest management where democratically elected village councils and the Forest Division cooperate to manage forest reserves. In the majority of cases, however, the powers devolved to the village councils are confined to (unfunded) management responsibilities and rights to low value non-timber forest products mainly for subsistence uses. This means that there is virtually no livelihood improvement opportunities as the costs associated with management outweigh the benefits (Topp-Jørgensen et al. Reference Topp-Jørgensen, Poulsen, Lund and Massao2005; Meshack et al. Reference Meshack, Adhikari, Doggart and Lovett2006). This is because the majority of joint forest management agreements are in montane rainforests, which host national and international biodiversity values that prohibit use, whereas joint management of production forest reserves has not yet been implemented. Likewise, the majority of functional community-based forest management regimes in Tanzania (where village councils have almost total autonomy) are situated in low-value and degraded forest areas, which implies that the immediate livelihood improvement potential is limited (Lund Reference Lund2007; Mustalahti & Lund Reference Mustalahti and Lund2010).

A plethora of rules and regulations formulated at more-central administrative levels tend to affect the distributional outcomes of decentralized forest management. Larson and Ribot (Reference Larson and Ribot2007) described how policies originally designed to favour foreign (colonial) and local elites remain biased and create double standards on a playing field already slanted towards the rich, powerful and well-connected. A common example of such policies are licensing and taxation systems that effectively exclude people without access to financial capital and good relations to the administrative system from capturing more than a diminutive share of commercial forest profits (Ribot Reference Ribot1998; Lund & Treue Reference Lund and Treue2008).

Existing social stratifications underlie many adverse effects observed in decentralized forest management processes. In Zimbabwe, Nemarundwe (Reference Nemarundwe2004) described how stratification influences who participates in committees and how some local people are able to bring their problems to authorities who will be more favourable (what von Benda Beckmann Reference von Benda-Beckmann1981 called ‘forum shopping’). Bandiaky (Reference Bandiaky2007) showed how decentralized forestry follows the contours of existing gender inequalities of local social hierarchies, transforming the intended equity outcomes into reinforced inequalities. Other studies came to similar conclusions concerning the reproduction of inequality in Zimbabwe, Uganda and Ethiopia (Bazaara Reference Bazaara2006; Feyissa Reference Feyissa2006; Mandondo & Kozanayi Reference Madondo and Kozanayi2006). Assembe Mvondo (Reference Assembe Mvondo2006) also noted that decentralization to multi-ethnic villages resulted in reinforcement of existing hierarchies, whereas access to benefits was observed to be more equitable in communities composed largely of a single ethnic group.

In general, pre-existing social relations influence who can maintain access through manipulation of the authorities around them and negotiate continued access either through reciprocal relations or payments (Nemarundwe Reference Nemarundwe2004; Blaikie Reference Blaikie2006). Such negotiated access is ‘. . .influenced by factors such as descent, social stratification, character of person involved, and longevity of residence in the area’ (Nemarundwe Reference Nemarundwe2004, p. 289), among others (Ribot & Peluso Reference Ribot and Peluso2003). The implication is that the structures and processes of decentralized forest management must ensure that the poorest and most vulnerable groups are identified and that rules are constructed that specifically favour these groups.

ANALYTICAL MISCONCEPTIONS

In current forestry discourses almost everything is called decentralization. Without careful reading, it is difficult to distinguish participatory approaches (any consultation, mobilization or involvement of local people) from co-management (a more contractualized form of ‘participation’) or democratic decentralization (involving power transfers to representative authorities). The confusion adds to an institutional amalgam an analytic nightmare in which no policy or project labelled decentralized can be taken at face value. Each intervention must be analysed according to its substantive elements, which include the degree of transfer of discretionary powers to local actors who are downwardly accountable to the population. The extent to which decentralization is being established and what the resulting outcomes are should, at a minimum, be analysed at the national, intermediary and local level to provide an overview of and linkages between (1) the political and legislative framework, (2) the official implementing agencies and (3) the environmental and economic (including distributive) outcomes at local level.

Misconceptions in analyses are multiple. Tacconi et al. (Reference Tacconi, Siagian and Syam2006) argued that studies fail to question the core assumptions of decentralization, which advocate that (1) democratic decentralization is an institutionalized and scaled up form of CBNRM, (2) people will engage in conservation due to the benefits that will follow (also see Hailey 1938) and (3) benefit-driven conservation will lead to less deforestation. The first item, however, is definitional, not an assumption and no amount of research will resolve it. Rather it is an ‘if-then’ statement: if there is institutionalized representation in the form of democratic decentralization and the representative authorities hold management responsibilities over natural resources, then it is an institutionalized form of CBNRM. As for the last two statements, Tacconi et al. (Reference Tacconi, Siagian and Syam2006) showed that these assumptions hold only under certain circumstances owing to the presence of multiple intervening variables, as reflected throughout this review. After their analysis of 95 cases in India, Agrawal and Chhatre (Reference Agrawal and Chhatre2006, p. 164) added, ‘variations in how the same factors operate and should be operationalized in different micro-contexts should make us pessimistic about the possibility of a universal theory of the commons.’ Indeed, there may be no generalizable theory of the commons. There may, however, be a generalizable theory of decentralization (a small subset of what happens in and around commons), or even a theory that allows that when the basic conditions of democratic decentralization of forests are established, the outcomes are far from certain.

Decentralization of forest management and use is not expected to produce positive outcomes in a vacuum. It occurs within a set of management and use rules and is shaped by the values attributed to the concerned forest. In the event that there are rules to protect forest values, under what conditions will decentralizing both responsibility for and benefits from working within these rules result in better management/enforcement? Are positive ecological and social outcomes likely when the costs are greater than the benefits of local forest governance? Probably not. Can market-based mechanisms, which constantly shift the costs and benefits, really ensure conservation over the long run? If the growth rate of a forest provides lower financial returns than elsewhere, why not cut it down now and invest the profit in a fund that has a greater return? The market or use value of forests cannot save forests from destruction, and the fickle nature of markets may mean that today forests are safe but tomorrow prices rise and cutting begins. There is no inherent reason to believe that local (or any other) people will not sell or convert forests if those are the most-lucrative options. There are only reasons to believe that collective local decisions may differ from individual local decisions or outsider decisions (since collectives internalize a broader range of values in decisions). Standards and rules set at a higher level are needed (see Ribot Reference Ribot2004). These rules can (and should) be used to create spaces of local discretion, but rarely are.

Differences between those wishing for conservation and those bearing the costs reshape motives for implementation. While higher-level institutions promote decentralization in search of forest conservation for values at higher levels of social aggregation, such as carbon storage or watershed regulation, the opportunity and direct costs of ensuring all of these values are largely born at the local level. Devolving responsibility for values that do not accrue to local people who bear the cost of conservation measures is not likely to lead to success unless complemented by funds so that local people have incentives to provide these services (Tacconi et al. Reference Tacconi, Siagian and Syam2006). Conservation is often viewed as something that local people will do for the returns they will gain (Wily no date). But those returns are public while people's unpaid labour is private. There is no more reason to believe that a village should pay or labour for a bridge so a national highway can cross a river near them, than to believe they should pay for forest management to ensure national or global benefits of conservation.

Analysts (Agrawal & Ribot Reference Agrawal and Ribot1999; Kulipossa Reference Kulipossa2004; Wily no date) posit that democratic decentralization can work for specified objectives (for example conservation or livelihoods enhancement) only if (1) the appropriate powers for the context are transferred to locally accountable authorities and (2) appropriate rules establish incentives for the specified objectives to be achieved (Larson Reference Larson2003). Fines for cutting too many or protected trees are among such rule-based incentives. Of course, any rights of this nature also require some mechanisms by which elected local councils can defend, negotiate and even expand their rights as national or regional or even global decisions bear down on and restrain their domain of discretion.

In short, the analytic questions are (1) Does democratic decentralization exist? Why or why not? (2) How does democratic decentralization interact with the incentives on the ground (the rules of engagement and shifting market values) concerning particular objectives? And (3) What is the relation between decentralization, the incentive structure and observed outcomes within specific levels of political-administrative or social organization?