Carbon Rights on Ancestral Domain

A community says yes. The memorandum of agreement is signed inside the ancestral domain, read aloud in the language the community speaks, thumb-marked and notarised. A scan of the signature page reaches an investment committee under the subject line “we have the land”.

You do not have the land, and you certainly do not have the carbon. You have one document at one point in a chain, and the chain is what an auditor will walk. Under Philippine law that signature is evidence that a consent process reached a particular stage. It is not carbon title. It is not authority to operate. It is not registration. And it is not durable: two of the instruments it leads to expire on a clock, and the consent behind it can be reopened.

None of which is a criticism of the framework. The Philippines is now more explicit about ancestral-domain forest carbon than most jurisdictions in the region. The difficulty is that the commercial mental model — secure the community, register the project — maps badly onto the statutory one, and the mismatch surfaces where it costs most to fix.

Start with the instrument, not the hectares

Establish which tenure instrument the area sits under before anything else. That single fact determines who may be the project proponent, whether a DENR tenure instrument is needed at all, which consent mechanism applies, and how long any commercial position can run.

DENR Administrative Order No. 2026-05, governing voluntary forest carbon credit projects in forest lands, NIPAS protected areas and ancestral lands and domains, was signed on 24 February 2026 and published in the Manila Standard on 28 February 2026. It takes effect fifteen days after publication in a newspaper of general circulation and upon submission of three copies to the Office of the National Administrative Register at the UP Law Center (§19); the order records that acknowledgment as 2 March 2026. It sets different specific requirements for each land category (§6.2).

Instrument	Who may be proponent	Duration ceiling
CADT or CADC, recognised under IPRA and registered with the Register of Deeds (§52(j)–(k)). The order also admits GALT and CALC holders (DAO §6.2.c.i)	The ICC/IP community through its IPO or indigenous political structure; another entity only by formal resolution of community consensus (DAO §6.2.c.i–ii)	Title does not expire. A non-member’s participation in resource development is capped at 25 years, renewable for not more than 25 (IPRA §57)
CBFMA, an agreement between DENR and a People’s Organisation over forest land. SFLMA is separately recognised as a valid tenure basis (DAO §6.2.a.iv)	Registered PO holding valid tenure and in good standing (DAO §6.2.a.i–ii)	CBFMA: 25 years, renewable for another 25 (DAO 96-29, Art. IV §1(a)). For any other instrument, read the term off its own face
NIPAS protected area, state land under a management plan and board	Requires a PAMB resolution, alignment with the management plan, and a tenure instrument issued by the Department (DAO §5), plus an ECC where required (§6.2.b.i)	Set by the tenure instrument issued
Ancestral domain overlapping forest land or a protected area, where both regimes apply	Subject to the exercise of priority rights by the concerned ICCs/IPs first; others considered only if the community declines (DAO §3)	As above, plus PAMB clearance and ECC (DAO §6.2.c.vi)

Two lines do most of the work. A CADT or CADC holder needs no separate DENR tenurial instrument (§6.2.c.iii): the instrument carrying the strongest claim is also the one over which the state has least discretion to act in your favour. And the community is the proponent by default. A developer is a technical service provider (§6.3.c.iv), and any transfer or sharing of participation rights or carbon entitlements needs the proponent’s prior written consent and notice to the Department.

That is a structural instruction, not a formality. Your position is contractual and derivative. It is not proprietary, and it is not yours to sell alone.

Carbon rights are defined, then left to you to construct

DAO 2026-05 defines carbon rights, and the definition repays close reading: the legal or beneficial entitlement to carbon credits generated from a project, distinct from the credits themselves, and determined by land tenure, project agreements, or applicable laws (§4(e)).

Notice what that does not do. It does not allocate carbon rights to a class of holder. It does not create a register of carbon title. It does not say the tenure holder owns the carbon. It says the entitlement is determined by three things and leaves you to assemble it from them: tenure you must verify, agreements you must draft, and law still being written. The order defers the minimum benefit-sharing ratio to later Forest Management Bureau guidelines with the NCIP (§8.2.a.iii), the capitalisation manual (§7), accreditation of project developers (§6.1.c.iii), and NDC eligibility and Article 6 transfers (§2(g)). No document exists that says “this entity holds the carbon rights to this area”. There is only a constructed chain.

The standards have moved to test that construction directly. VCS Standard v5.0, published on 16 December 2025 with most requirements effective 1 January 2027, retired project ownership and replaced it with two things a proponent must hold at project start and throughout implementation: the right to operate, and the right to reductions and removals. Each must be “substantiated by an analysis of applicable laws, regulations, and legal instruments” (§3.6.1). On ancestral domain the realistic bases are land or resource rights granted by a competent authority, or an enforceable and irrevocable agreement with the holders of those rights (§3.6.2).

Where a project is likely to affect land or resource rights, a separate rights analysis is required before the project start date, run by a team with knowledge of tenure regulation, customary rights, decision-making systems, international human rights law and existing conflict; a proponent without that expertise must partner with organisations that have it (§§3.6.3–3.6.5). It must locate overlapping claims, competing claims and violent conflict, and where competing claims exist, name the claimants and give an estimated timeline to secure the rights or resolve the dispute (§3.6.6). “We will resolve that later” is now a dated commitment. At every verification the analysis is updated to establish whether the proponent still retains both rights (§3.6.7). The Philippine order says carbon rights are determined by tenure, agreements and law; the standard says prove it, before you start and again each time you verify.

Initial FPIC and Comprehensive FPIC

DAO 2026-05 permits consent to be phased and names the phases: Initial FPIC for feasibility assessments and pilot activities, Comprehensive FPIC for full implementation, credit registration and long-term monitoring (§6.3.a.iii). Where the community is itself the proponent, a Resolution of Community Consensus replaces FPIC altogether (§6.3.a.ii).

Three consequences follow, and all three are commonly got wrong.

• Phasing is not automatic. It is drafted. The operative FPIC rules remain NCIP Administrative Order No. 3, Series of 2012, under which, unless the MOA specifically states otherwise, FPIC is exercised separately for each major phase of the activity (§20). If the MOA is silent, every major phase is a fresh process — which may be what the community wants, but is a consequence to choose rather than inherit.
• Carbon has been on the large-scale list since 2012. Section 19 classifies “carbon trading and related activities” as extractive, intrusive or large scale (§19(i)), alongside large-scale agricultural and forestry management projects (§19(h)) and the issuance of any land tenure or resource use instrument (§19(l)). Section 19 activities run the two-assembly process under §22; §24, the lighter two-meeting route, covers feasibility studies not embraced in the preceding sections. Whether a carbon feasibility study is a §19(a) study “related thereto” or a §24 study changes the timeline materially. Put that to the NCIP regional office in writing before you budget — not to an assumption inside a Gantt chart.
• The structure changes the instrument. If the community is the proponent you are not securing FPIC. You are supporting a community resolution and then negotiating a partnership, which has its own architecture: a declaration of the exercise of priority rights, a validating team, a community assembly to weigh the proposed partner’s capability and track record, a consensus-building period, a partnership MOA validated with the community at the partner’s expense, and confirmation by the Commission En Banc (§§50–55). That acknowledgment is expressly not a grant of a CADT (§56).

Where a domain coincides with forest land or a protected area, DAO 2026-05 puts the priority-rights gate in front of everything (§3). An external proponent whose first act is to propose itself has begun in the wrong order.

The one-year feasibility clock

DAO 2026-05 sets an eight-step pathway for new and pipeline projects (§9.2.a). Step three is a Notice to Proceed, which authorises preparatory activities and which the order is careful to say does not constitute approval or registration (§9.2.a.iii). Step four sets the clock: within one year of receiving it the proponent must complete a comprehensive feasibility assessment covering environmental, social and technical aspects, including safeguard screening and costing — on ancestral domain, expressly including NCIP coordination, consultations, and FPIC or a Resolution of Community Consensus consistent with the ADSDPP (§9.2.a.iv).

A year is generous for a technical study and tight for a consent process, because consent does not run on your calendar. Set the statutory sequence against it and the arithmetic is visible.

• The application for a Certification Precondition is endorsed by the relevant regulatory agency to the NCIP regional office with jurisdiction, with a project profile, indicative map, preliminary assessment of effects and operational plan (AO 3 §§6–7).
• A pre-FBI conference settles the work and financial plan and the fee (§§10, 12). The Field-Based Investigation commences within ten days of the fee being deposited and completes within ten working days, force majeure excepted; the report is sworn (§§13–14).
• Notices for the First Community Assembly are posted, and served personally on the domain representatives named in the CADT recognition book or ADSDPP book, seven days beforehand. That assembly validates the investigation report and the affected areas, takes the census, identifies and validates elders and leaders, determines the decision-making process, and fixes the date of the Second Assembly (§22).
• At the close of the Second Assembly the community is left alone to set its own decision schedule. Consensus-building must begin not less than ten days after that assembly and finish within a reasonable time, not more than two months (§22). The applicant and non-members are strictly enjoined from participating.
• On a favourable consensus, terms are negotiated, the MOA is drafted in the vernacular and in English, a validation assembly inside the domain explains it clause by clause, and signing happens inside the domain before a notary (§§22, 31, 34). The Regional Review Team then reviews within five days and the record is endorsed to the Ancestral Domains Office (§28); the Certification Precondition is a separate certificate, issued by the NCIP and signed by the Chairperson (§5(e)).

Every day count there is an administrative minimum, not an estimate, and none of them is the community’s calendar. Planting and harvest. Rituals and funerals. The rainy season and the roads it closes. The internal question of who is authorised to sign. A leadership succession mid-process. A typhoon. Those set the duration, and they are not delay in any culpable sense. They are the process working.

The clock runs at the other end too. Inaction for six months from the last FPIC activity, without justifiable reason, grounds termination of the process by the Regional Director (§29). If no activity follows within a year of the permit’s issuance the Certification Precondition goes dormant, the Chairperson issues a show-cause order, and failure to respond within fifteen working days leads to cancellation by Commission resolution (§30). A resolution of non-consent bars any similar proposal for six months (§27).

And the rule is moving. The NCIP announced in May 2026 that it had approved revised FPIC guidelines. As reported by Human Rights Watch on 4 August 2026, it requires a community assembly within ten days of notification and allows thirty days for consensus-building on large-scale projects; for small-scale projects, four days to convene and seven to build consensus. The 2012 position is no fixed deadline to convene and up to two months for consensus. The NCIP has defended it as the product of more than a decade of review, citing an FPIC transparency server, stricter applicant disclosure, a precautionary principle, and a continuing obligation to comply with FPIC throughout implementation. In the same reporting, Amnesty International Philippines objects that fixed day counts are laid over decision-making the guidelines themselves say should follow customary law.

We could not locate a published text, order number or effectivity date for that revision as at the date of this note. The operative published instrument remains AO 3, Series of 2012. For anyone modelling a consent process over the next twelve months, that is the real exposure: not that the timeline will tighten, but that you are budgeting against a rule you cannot yet read.

What the MOA is, and the four things it is not

Under AO 3 the MOA embodies the terms agreed after consent and is executed among the ICCs/IPs, the applicant, the NCIP, and any other necessary party (§31). The NCIP signs as third party through the Chairperson, or the Regional Director where that officer is authorised to issue the certification (§33(c)). Its required contents run to twenty-four heads, among them detailed benefit-sharing, development projects based on the community’s own priorities, redress mechanisms, monitoring paid for by the company, and remedies extending to customary law and sanctions (§32).

It is a serious instrument. It is also none of the following.

It is not the Certification Precondition. That certificate’s weight comes from IPRA §59, which enjoins every department and agency from issuing, renewing or granting any concession, licence or lease, or entering any production-sharing agreement, without prior NCIP certification — and no certification issues without the free and prior informed and written consent of the ICCs/IPs concerned. The same section bars any new concession, licence, lease or production-sharing agreement while a CADT application is pending: a live constraint across many upland landscapes, and one of the first facts worth establishing.

It is not registration. That is a DENR act at the end of the pathway: the Forest Management Bureau reviews for completeness, tenure legitimacy and safeguard compliance, and the Secretary issues a Certificate of Registration, a Conditional Certificate, or a Certificate of Non-Compliance (DAO §9.2.a.viii). The conditional route is narrower than it sounds: minor, readily correctible gaps such as clerical errors or missing signatures, expressly excluding deficiencies related to tenure, safeguards or financial integrity, and lapsing if uncorrected within ninety days. Tenure problems get no conditional pass.

It is not transferable, and it is not final on its own terms. The MOA must contain a clause on its own non-transferability, a clause for renegotiating the economic provisions, a statement of whether the community will require a further FPIC on merger, reorganisation, transfer of rights, acquisition or joint venture, and a provision on the effects of a dormant Certification Precondition (§32(f)–(h), (v)). Any project expecting to be sold, refinanced or moved into a special-purpose vehicle should assume the community has a say, and find out at drafting stage what it will cost.

It is not the end of consent. IPRA §59 gives ICCs/IPs the right to stop or suspend a project that has not satisfied the consultation requirement. The NCIP’s stated reform adds a continuing obligation through implementation. VCS v5.0 requires the proponent to keep customary rights holders updated at the frequency set in the project implementation agreement, to act where major changes occur to the conditions on which consent was granted, and to retest at each verification whether it still holds its rights (§§3.17.17–3.17.19, 3.6.7). Consent here is a maintained state, not a captured asset.

The fifty-year ceiling

Then the duration problem, the one most likely to be missed because it reads like boilerplate.

A non-member of the ICCs/IPs concerned may be allowed to take part in the development and utilization of the natural resources for a period of not exceeding twenty-five (25) years renewable for not more than twenty-five (25) years. — IPRA §57

DAO 2026-05 mirrors it: project development agreements on ancestral domain shall observe the permanence principle of twenty-five years plus a twenty-five-year renewal (§6.2.c.v), and a CBFMA runs the same twenty-five plus twenty-five (DAO 96-29, Art. IV §1(a)). The statutory maximum for a non-member is therefore fifty years, while crediting periods, buffer obligations and reversal-risk commitments are routinely written longer and the policy interest in the forest is indefinite.

DAO 2026-05 answers the DENR-side version with a tripartite agreement among the Department, the PO or IPO and the developer to carry permanence and monitoring obligations where tenure nears expiry (§6.2.a.ii), and with State assumption of management responsibility — expressly not ownership — where no eligible entity takes over at completion (§13.3).

The structuring conclusion is the one the order implies throughout. The permanent party is the community. The developer is a term-limited service provider whose obligations must be designed to survive its own exit. Built the other way round, a project will not survive contact with either IPRA §57 or a serious rights analysis under VCS §3.6.

What to diligence instead of the signature

None of this makes ancestral-domain carbon unbankable. It makes the bankable version identifiable, which is more useful. These are the questions that discriminate, and almost all can be answered before significant money is committed.

• Is there a CADT, and is there a pending application anywhere near? A registered title (IPRA §52(j)–(k)) and a pending application are different worlds, and §59 treats the pending one as a bar on new instruments.
• One domain, or several? Where two or more domains are affected the owners exercise FPIC separately; where one domain is held by two or more subgroups, any validly existing agreement among them governs and otherwise all subgroups participate (AO 3 §20), while the domain itself is recognised and treated as one undivided unit (§4(i)). A boundary question with a consent consequence attached.
• Does an ADSDPP exist, when was it last updated, and what does it exclude? The Revised ADSDPP Guidelines (NCIP AO No. 2, s. 2018) require the plan to be updated every five years (§12) and to carry the community’s own benefit and responsibility sharing schemes (§7.3.3). Decisively, where it conflicts with another plan the ADSDPP as approved by the community takes precedence (§14), so a benefit-sharing plan drafted against a developer template is drafted to lose. AO 3 §25 separately excludes sacred grounds and burial sites, heritage sites, community-reserved critical areas, and any other areas named in the ADSDPP — a hectarage question before it is an ethical one.
• Who owns the data you are about to publish? Information obtained in preparing the ADSDPP is community intellectual property, private and confidential unless FPIC is secured (AO 2, s. 2018 §16), and indigenous knowledge systems and practices are documented subject to community rules (§7.1.3). A project design document is a public artefact. Decide early, with the community, what goes into one.
• Does the map survive the walk? Titled hectarage, remote-sensed forest cover and ground-verified project area are three different numbers, and only the third is bankable. We have seen that divergence run wide on the coast, for reasons that had nothing to do with tenure (the case is set out here) — but the size of the gap is what decides what a tenure instrument is actually worth. The number you put in front of a buyer should be the number you walked.
• Who lodges the boundary first? The DENR Forest Carbon Credit Database may tag applications by status to prevent overlapping project areas (DAO §14.1(d)) — permissive on the face of the order, so do not treat the database as a defence of your boundary. Whoever lodges first still sets the reference the next applicant argues against, which is an argument for getting the geometry right and lodging early, not for lodging fast.

The test

The useful question is not whether a community has said yes. It is whether the yes is attached to an instrument that runs long enough, recorded in the chain the state recognises, renegotiable on terms someone will still accept in twenty years, and robust enough to survive being asked again. Because under both the Philippine framework and the current version of the standard, it will be asked again, at every verification, for the life of the project.

A signed MOA is the beginning of that test, not the end of it. It is a good day. It is not a title.