Madagascar's GELOSE resource-management transfers

A contract gives a community legal standing without giving it ownership

Law 96-025 created GELOSE—Gestion Locale Sécurisée, or secured local management—in September 1996. It covers renewable resources in the domain of the state or territorial collectivities: forests, terrestrial and aquatic wild fauna and flora, water, and grazing territories. A voluntary local group can receive legal personality and apply to manage resources within its territory. The law calls the resulting authority autonomous, but makes it subject to the negotiated contract and specifications.¹

The arrangement does not privatize the resource. The state or collectivity remains the owner; the local community receives management of access, conservation, exploitation, and valorization within the contract. The commune participates in the transfer, the competent administration grants approval, and the initial term is three years. After evaluation it may be renewed for ten years, changed by agreement, or withdrawn for failure to perform the community's obligations.¹ GELOSE therefore transfers a bundle of decisions, not sovereignty over land or resources.

That bundle can still matter. A recognized organization can negotiate with officials, hold funds and equipment, adopt enforceable internal rules, and defend contracted use against an outsider. It can also become the local bearer of a conservation plan whose valuable decisions remain subject to ministerial approval, technical inventories, or outside finance. A signature records formal authority; the next disputed use reveals practical authority.

The legally recognized association is more specific than the Fokonolona

The implementing decree makes the communauté de base, often called a VOI or COBA, a voluntary association rather than an automatic synonym for everyone in a village. Any resident within the territory may apply, but the general assembly decides admission. The assembly adopts the statute and dina, sets an annual work plan, elects a president, vice-president, treasurer, and secretary, approves accounts, and allocates funds. Decisions begin with consensus and can move to specified majority and quorum rules. Written records, a receipts-and- expenses book, and an elected accounts commissioner make the association legible to administrative partners.²

Those provisions create inspectable member powers. They also create a boundary between member and nonmember, and they presume that someone can convene a meeting, maintain lists, write minutes, keep accounts, and speak to officials. The decree's model rules require every member to participate in annual-plan tasks. They permit monetary reparation, repair work, and eventual exclusion, after the accused member can argue before the assembly.² Formal equality in the statute does not establish equal time, literacy, customary standing, income, or bargaining power.

A community-network account supplies a different institutional vocabulary. Writing on behalf of TAFO MIHAAVO and FANONGA FOKONOLONA, Vololoniaina Rasoarimanana describes a network created by community representatives in 2012 that had grown by 2021 to nearly 600 Fokonolona. Some members act through GELOSE; others govern resources in practice. The network's stated 2050 vision centers autonomous and sovereign Fokonolona living well with nature, and its account names land appropriation and the use of office or money to dominate vulnerable people among the hardest problems.³

Community representatives from Madagascar's regions made the representation problem explicit in 2019 workshops reported by Natural Justice. They asked for greater decision rights, workable obligations and evaluation, protection for people reporting infractions, land security, economic viability, and a way to recognize customary rules without changing their meaning. The report calls the required association form controversial because representation, structure, and operation can displace a community's own governance.⁴ GELOSE can recognize custom, but recognition is itself an exercise of power.

A dina joins customary authority to public review

The dina is the hinge. Law 96-025 says relations among association members are governed through a dina approved under the community's customary rules. It cannot violate public order, legislation, or recognized and uncontested communal usage. It becomes executable only after the mayor's visa and has the force of law among members; a member may face vonodina sanctions, while judicial action follows exhaustion of the internal procedure.¹ The decree adds a model and places adoption, amendment, enforcement, and financial accountability inside the association.²

This hybrid can convert a rule known locally into one an official will recognize. It can also freeze one contested version of custom, especially when the people writing, translating, approving, or funding the instrument do not experience the resource in the same way. The law's own mediation procedure tries to surface interests, territorial claims, long-term visions, access, sales, benefit allocation, and sanctions. Yet a neutral mediator and a signed text cannot by themselves decide whose knowledge counts.

Fieldwork along the Onilahy River shows both sides. In May and June 2006, researchers interviewed residents in five south-western sites. At Ampany-Bevilany they found villagers using a GELOSE contract to preserve collective grazing and gathering against newly arrived sapphire miners. In other villages, clan leaders described the contract as recognition of their authority, migrants' access depended on COBA approval, and leaders still wanted the state to intervene where they could not stop cultivation inside the contracted area.⁵ This is evidence from a small set of places, not a national portrait. It demonstrates why local control can protect a claim and reproduce a hierarchy at the same time.

Real authority appears in the next valuable decision

An early academic analysis used the distinction between formal and real authority to explain GELOSE. It argued that delegation becomes effective when the parties' interests converge and that exclusion rights can create that convergence: communities may value the ability to keep outside permit holders or harvesters from using the resource, while the state values local control. The article also shows how the contract's information demands and the state's reluctance to lose control can narrow practical discretion.⁶ One coauthor then headed the government's GELOSE department, so the article is both informed by implementation and not wholly external to it.

Daniela Raik's later fieldwork tested perceptions across the operating coalition. She conducted 55 open-ended interviews with community members in eight Menabe villages, forestry staff, and two NGOs, then surveyed 621 people associated with 12 forest-management contracts in 2006. Her conclusion was deliberately narrower than “empowerment”: respondents generally perceived good governance outcomes, but decentralization changed institutional relationships more visibly than individuals' capacity to act.⁷ The contract sites were purposively selected, 73.8 percent of survey respondents were men, and almost all had not completed primary school. The study is strong evidence about a varied participant sample and weak evidence for a current national average.

Scale does not resolve that gap. On 15 July 2026, the administrative SIG-TGRNR dashboard displayed 1,580 COBAs, 2,189 COBA contracts, 1,273 expired contracts, and 46 percent of COBAs without a listed technical and financial partner.⁸ Those live administrative counts may change and the page does not fully explain deduplication or completeness. They nevertheless show why creation counts and active capability are different measures. A contract can exist in the register while renewal, finance, enforcement, or partner support has lapsed.

The practical test follows the next timber request, grazing conflict, fishing rule, sanction, leadership election, account inspection, or proposed amendment to the dina. Who can obtain the information, convene the decision, authorize valuable use, enforce it against a powerful outsider, appeal an internal sanction, and pay the cost? GELOSE has transferred capability to the extent that the community organization can complete that sequence without an outside actor silently deciding each consequential step.

Participation follows valuable rights and credible support

Participation is not a free input. Meetings, minutes, patrols, fire control, monitoring, dispute handling, and restraint consume time that can otherwise produce food or income. In sixteen communities around Ambohilero Forest, members associated with development-oriented and community-logging support were more involved in management than members associated with a conservation project. The researchers connected lower participation to difficult access and the lack of livelihood incentives after the conservation contract was implemented.⁹

A 2007 survey of 220 households in twelve south-western villages found neither simple rejection nor simple consent. Of households answering the relevant questions, 60 percent favored the contracts, often citing improved relations with forest officials; 46 percent wanted more control, 30 percent reported satisfaction, and about one third expressed interest in continuing traditional, unregulated access. Some villages described cutting permissions as strict and costly.¹⁰ The authors flag inconsistent denominators, small cells, and the possibility that respondents told interviewers what they wanted to hear. The result supports a bounded point: legal recognition and a better official relationship may be valuable even when direct livelihood gains are weak.

A 2022 independent evaluation of selected GEF projects found the same conditional pattern in more recent community-based conservation. In two villages, residents said livelihood inputs had reached only 11 people out of about 400 and 20 households out of about 4,000. Forest patrols continued in some places through voluntary work or modest payments from later partners. In Ambavarano, a women's association reported benefits from mangrove restoration; the evaluators used imagery to corroborate regrowth near the village and recorded continuing links to crab harvest and markets.¹¹ These cases belong to selected GEF projects, not a representative sample of GELOSE contracts, but they show what contract text leaves contingent: finance, market access, partner continuity, and who receives the livelihood activity.

National results reject automatic success and automatic failure

The strongest national studies usually pool GELOSE with the later, simplified GCF forest-contract mechanism under “community forest management.” They test the broader practice of forest transfer, not Law 96-025 alone, and they cannot represent water, fisheries, wildlife, or grazing contracts.

For 2000–2010, a matched national analysis detected no average deforestation reduction for community-managed forests relative to comparison forests, even after restricting the analysis to places with indications of implementation. It also detected no effect where commercial use was allowed, while community forests prohibiting commercial use showed lower deforestation than their comparators.¹² The design controls observed differences and improves on a raw before-and-after comparison; it does not randomize contracts or prove that every mapped association operated as its agreement specified.

A separate national study estimated a positive but small effect on per-capita household consumption that was not statistically distinguishable from zero. It could reject a large average negative effect, and estimates were more positive for households nearer forests and with more education.¹³ Zero statistical distinction is not evidence that every household was unchanged. Consumption also misses tenure security, voice, conflict, unpaid work, cultural value, and distribution within a household or association.

Political crisis supplies another comparison. A 2024 study followed 362 community forests created before 2005 and 45 Madagascar National Parks areas from 2005 through 2020. Both performed poorly during the 2009–2013 crisis. After matching and controlling observed time-varying factors, community forests had significantly more deforestation than comparable park forests in 2014–2017; the difference remained negative but was no longer statistically significant in 2018–2020.¹⁴ The authors caution that remotely sensed cover can be biased, designations overlap, and their matched park areas are the accessible subset comparable to community forests, not the whole park system.

These findings do not cancel the Ambohilero participation difference or the Ambavarano mangrove regrowth. A site study, a national average, a household consumption estimate, and a crisis comparison answer different questions. The consistent lesson is heterogeneity: contract purpose, commercial rights, resource value, location, member composition, partner support, political conditions, and enforcement capacity change what the same legal form can do.

The organizational profile follows a contract inside a coalition

State bureaucracy grants approval and retains ownership; local-federated authority comes through the association and its assembly; the dina is a commons protocol; and professional expertise enters through mediators, foresters, conservation staff, researchers, and technical partners. Decisions therefore sit at a frontline-local assembly, in a rule-bound public hierarchy, and in professional cells that produce maps, inventories, plans, and evaluations.¹²

State ownership coexists with a member-governed association and a partnership network. Rule and ritual organize the dina and contractual recognition; standards organize legal and technical compliance; mutual adjustment organizes negotiation; and planning appears in the work plan and resource plan. Embedded practice, two-way negotiation, and specialist staff are the main knowledge flows. Operational, informal, and mission measures respectively track work and resource use, social acceptance and compliance, and conservation or livelihood purpose.⁶⁷

Apprenticeship, formal research, and local experimentation describe how skills can develop, while adaptation occurs through site-level iteration and slow legal or administrative change. Communities, members, public institutions, nonhuman life, and ecosystems are plausible beneficiaries; naming them does not assert that each benefited. Capture, suppressed voice, externalized livelihood or ecological harm, and fragility after contracts or support expire are the principal failure risks.³⁸¹¹¹⁵

Six themes define the case

Authority, legitimacy, and acceptance, delegation, decentralization, and responsibility, decision-making, judgment, and bounded rationality, knowledge, expertise, and professional autonomy, innovation, entrepreneurship, and renewal, and governance, stewardship, and accountability receive the highest emphasis. GELOSE recombines state law, customary rule, a voluntary association, and negotiated contract; its central questions are where authority actually moves, how bounded actors decide, whose knowledge enters the plan, and who can hold each party accountable.¹⁶

Purpose, coordination, cooperation, learning, and culture receive medium emphasis because conservation purpose, multi-actor negotiation, participation incentives, accumulated capability, and informal authority materially change implementation. Structure, measurement, and strategy receive lower emphasis: they illuminate hierarchy, evaluation, and adaptation without defining the case. Work design, executive attention, and organizational ignorance score zero because production-system design, a single executive sensing mechanism, and ignorance as an independently demonstrated strategy are not developed enough in the evidence to organize the account.¹⁵

Related paths compare distributions of valuable authority

Cameroon's community-forestry system, Nepal's community forest user groups, and ACOFOP's community forest concessions offer other legal forms for asking whether a local organization controls valuable use, exclusion, revenue, revision, and enforcement or mainly carries obligations designed elsewhere. The comparison concerns institutional design, not a claim that one program caused another.

Zimbabwe's CAMPFIRE provides a further comparison between benefits assigned to a local constituency and practical control that may remain with public or intermediary bodies. The benefit-for-all-life lens keeps human livelihood, nonhuman life, ecosystems, and time horizons in the same inquiry without assuming that one aggregate score can resolve their tradeoffs.

The defining idea links above are interpretive lenses. In particular, delegation, decentralization, and responsibility asks whether authority, resources, review, and liability travel together. None of these links asserts a documented historical or intellectual influence.

Effects and evidence boundaries

The formal record is unusually clear about legal personality, member rules, contract duration, resource scope, and retained public powers. It cannot show who attended a village meeting, how a forestry officer exercised discretion, or whether an expired contract still organized behavior. The administrative dashboard supplies scale and status counts without a public audit of coverage, definitions, duplicates, or outcomes.¹²⁸

Affected-community evidence is present but uneven. TAFO MIHAAVO speaks through a national community network but not for every Fokonolona. The 2019 workshop record aggregates priorities rather than publishing a full transcript. The Onilahy and south-western studies preserve resident accounts yet concern a small region and older fieldwork. Raik's larger survey spans twelve contracts but purposively selected sites and underrepresented women.³⁴⁵¹⁰⁷ Public evidence located for this review does not support separate national estimates for women, poorer nonmembers, migrants, customary authorities, pastoralists, fishers, or people sanctioned by a dina.

Outcome evidence is strongest for forest cover and per-capita consumption. Those studies pool different forest-transfer instruments, periods, and contract purposes. They do not measure every ecological process, species population, customary value, conflict, distributional effect, or non-forest resource. The future-generations record remains research-needed because none of the studies follows descendants and resource users across repeated renewal cycles.

Paths into deeper study

• Publish a versioned, deduplicated contract register with resource type, boundaries, members and nonmembers affected, signatories, renewal status, partner support, management rights, restrictions, sanctions, appeals, and reasons for expiry or withdrawal.
• Follow representative cohorts of GELOSE and GCF sites from negotiation through renewal, recording who attends, speaks, votes, keeps accounts, performs unpaid work, receives revenue, is sanctioned, and exits membership.
• Compare the written dina with living customary practice in Malagasy and local speech, under community control, and document which rules changed in translation, mayoral review, or technical drafting.
• Pair satellite and ecological data with species, water, grazing, fishery, and livelihood measures; distinguish resource protection inside the boundary from displacement outside it.
• Measure completed and appealed enforcement events, not only patrols or rules, and test whether communities can act against politically or economically powerful outsiders without retaliation.
• Fund women-, migrant-, pastoralist-, fisher-, poorer-household-, and nonmember-led research that can contest association records and define outcomes in locally meaningful terms.
• Track public, donor, NGO, market, and community costs through partner exit and contract renewal so that transferred authority can be distinguished from transferred labor and liability.