VM0047 ARR in the Cerrado.

Why this is the binding constraint

The Cerrado is the most active jurisdiction in the world today for new ARR project pipeline. Verra's registry shows acceleration in VM0047 development activity from this biome through 2024–25, culminating in the December 2024 ICVCM CCP approval and the May 2025 first-issuance approval for Project 5511. Principal carbon buyers, Microsoft Carbon Removal at 45 Mt of fiscal-2025 contracting, Frontier at $1B+ AMC, the Symbiosis Coalition with its 20 Mt by 2030 ambition, are taking offtakes. Farmland funds with regenerative-conversion theses (Manulife FCF, BTG Pactual TIG, Nuveen Natural Capital) are pricing carbon-side optionality against the same evidence base. Commodity-finance banks are structuring sustainability-linked loans against project KPIs.

The common operational problem: the project IRR is calculated by the developer. The credit risk has to be calculated by the underwriter. Translating between the two is the work, and it is the work most off-the-shelf vendors skip.

What VM0047 actually changes

The methodology, published in 2023 and now in v1.1 (effective from the 1 October 2025 ICVCM-approved minor revision), introduces four architectural changes from the legacy AR-AM ACM0003 and prior Verra ARR methodologies.

Change 01, Dynamic baselines. Baselines are not fixed at project start; they are updated at each verification period using a performance-benchmark approach. The benchmark is constructed from a stratified control population in the same biome with similar land-use history. This produces credits whose issuance volume reflects actual outperformance against a contemporaneous comparator, not a static historical assumption.

Change 02, Performance-benchmark crediting. Credit issuance is calibrated to the project's performance relative to the benchmark population, not to a counterfactual scenario the developer specifies. This is the substantive integrity change. It reduces the bias toward over-crediting that affected prior ARR methodology designs.

Change 03, Area-based and census-based quantification. The methodology supports both stratified-area and census-based quantification, with explicit precision-and-accuracy requirements per stratum. Practical effect: smaller projects can use census methods that were previously impractical under sampling-only quantification.

Change 04, End-of-period buffer cancellation. The Verra buffer pool sits between the issuance and the ultimate value-at-risk. At the end of the (typically 30-year) crediting period, Verra cancels all remaining buffer credits to cover future reversal risk. During the crediting period, buffer credits are cancelled on a reversal-event basis.

The combined effect is that VM0047 credits have a more accurate ex-ante issuance volume, a more conservative reversal-risk treatment, and a clearer audit trail than the prior methodology generation. For underwriters, this means the methodology-level risk is reduced, but the project-level reversal economics are tighter and more explicit. The buffer pool sizing is binding, not theoretical.

Project 5511, the worked example

Brazil Cerrado 1, the first VM0047 issuance, anchors the worked example below.

• Project area. Cerrado biome, distributed across Mato Grosso do Sul, Mato Grosso and Minas Gerais.
• Restoration target. Up to 25,000 ha of degraded land restored to mixed native vegetation.
• Expected removal. ~7.0 Mt CO₂e over the 20-year first crediting period, approximately 351,328 tCO₂e annually.
• First issuance. 230,120 VCUs approved under VM0047 v1.0.
• Buffer pool sizing. Determined by the project-specific risk assessment under VM0047, materially above the Verra portfolio-default minimum.

The NGFS Phase V translation

NGFS Phase V scenarios, published November 2024, are the current canonical reference for credit-risk translation under nature-related disclosures. They are the substantive update from Phase IV (October 2023) and supersede the earlier Phase III (2022) framing. Two operational changes from Phase IV matter.

Change 01, Updated damage function. Phase V's headline result is that chronic physical-risk losses are now 2–4× higher globally by 2050 versus Phase IV, driven by an updated macroeconomic damage function. NGFS itself notes that the Kotz et al. (2024) Nature paper underpinning the damage function was retracted and resubmitted; the Phase V scenarios continue to use the framing pending the resubmission outcome.

Change 02, Phase III/IV credit-risk translation continuity. The Phase III credit-risk-translation logic (PD/LGD adjustment under climate scenarios) is preserved in Phase V; the substantive change is the damage-function calibration, not the translation mechanic. For underwriters this means existing Phase III-anchored PD/LGD overlays carry forward with updated stress magnitudes.

Translating Project 5511 to PD/LGD

For a representative Cerrado ARR project, anonymised, structurally similar to Project 5511 in biome, size and operator profile, the credit-risk translation runs as follows.

Step 01, Baseline credit metrics. Sponsor profile: a mid-tier Brazilian agribusiness operator with prior Verra projects issued. PD baseline (sponsor-level): 2.1%. LGD baseline on an unsecured covenant: 45%. These are the inputs from the operator's credit profile, not the project.

Step 02, Permanence-reversal probability. Computed in the permanence module as a function of biome-specific perils. For a Cerrado parcel: fire return-period 8–14 years (this biome carries one of the highest fire-return rates of any forest biome). Drought stress: rising under RCP4.5; substantially more under RCP8.5. Encroachment pressure: distance-to-frontier-dependent. Counterparty execution-failure: priced from the sponsor's prior project track record. The 20-year reversal probability for a typical Cerrado VM0047 project lands at 7–11% (median), with the 90th percentile at 14–17%.

Step 03, Buffer-pool recovery. The buffer-pool size is the project-specific output of VM0047's required risk assessment. For Project 5511's structure, the buffer-pool sizing sits above the 15% Verra minimum, likely in the 18–22% range given the Cerrado fire-return-period priors. The LGD adjustment is the inverse: an 18% buffer pool reduces effective LGD by ~18% on the permanence-failure leg.

Step 04, Horizon extension. VM0047's 100-year permanence commitment requires translating a 20-year reversal hazard to a 100-year exposure. This is not a simple multiplication. The hazard curve has non-stationarity (climate projections worsen post-2050) and counterparty-failure compounds. We model this as a piecewise hazard curve with breakpoints at 2040, 2060 and 2080. The 100-year cumulative reversal probability for a Cerrado VM0047 project lands in the 18–28% range depending on RCP assumption.

Step 05, Combined credit metric. The adjusted PD on an SLL covenanted to this project: 9.4–11.2% (median band), against 2.1% baseline. The adjusted LGD: 36–38% (after buffer-pool recovery). Expected loss (year 1, 20-year horizon): 3.4–4.2% of notional. The same calculation on a 100-year horizon, with the hazard-curve non-stationarity: 8.2–10.7%.

These are not abstract numbers. They are the inputs the CRO sees on a Cerrado ARR-linked deal. The substantive read for the deal team is that the 100-year exposure is materially higher than the 20-year exposure, and the buffer-pool recovery is the single largest sensitivity in the model.

What this changes operationally

Three implications for deal teams.

Implication 01, The covenant structure has to price horizon. A standard sustainability-linked-loan covenant priced against the 20-year exposure understates the position. The cleaner structure is a tiered covenant: tighter pricing for the first 20 years, with explicit re-marking at 20, 40, 60 and 80 years against the updated hazard curve. Where the deal is structured as a hold-period rather than a 100-year commitment, the structure can be simpler, but the documentation should make the horizon explicit.

Implication 02, Buffer-pool sizing should be parcel-specific, not portfolio-default. The 15% Verra minimum is a portfolio-average safe value; VM0047's mandatory project-specific risk assessment makes buffer-pool sizing a parcel-specific output. For a fire-return-period-of-8 Cerrado parcel, the 15% minimum is materially too low; for a fire-return-period-of-50 Atlantic-forest parcel, 15% may be materially too high. Parcel-specific monte-carlo is the right reference.

Implication 03, The IRR-to-credit-risk translation is not optional. The 14% stated developer IRR is correct, given the developer's assumptions. The 3.4–4.2% expected loss on the SLL is correct, given the credit-risk translation. Both numbers are true; they answer different questions. The investment committee needs to see both, and the gap between them is the substantive risk premium.

How Kyroq runs the working

The permanence monte-carlo is a single workstream in the broader memo. For a typical Cerrado ARR deal, the workings (PD adjustment, LGD adjustment, horizon extension, sensitivity table) ship as a 6-page appendix to the investment-committee pack. The buffer-pool recommendation is computed and journaled per vintage. The sustainability-linked-loan covenant draft is delivered with the appendix.

For an in-flight position, the same module re-runs annually. Permanence-buffer drift, hazard-curve update under the latest NGFS Phase V vintage, and counterparty-track-record refresh all feed the re-mark. The fund's re-mark cycle is the lowest marginal cost on the platform.

Beyond Project 5511: pipeline outlook

VM0047's December 2024 ICVCM CCP approval and the v1.1 October 2025 minor revision establish a stable methodology floor for the second-wave Cerrado pipeline. Additional ICVCM-approved ARR methodologies, including ACR's Afforestation and Reforestation of Degraded Lands (approved July 2025), broaden the supply side. As of end-November 2025, ICVCM has approved 7 major carbon-crediting programs and 36 methodologies per the Impact Report 2025; ACR-issued CCP-labelled credits total 7,792,791 to date.

The second-wave pipeline matters for buyers because it is the volume that clears the AMC and principal-buyer demand. Microsoft Carbon Removal's 45 Mt fiscal-2025 contracting cycle included re.green's expansion to ~6.5 Mt across Brazilian biomes, including Cerrado parcels, and Chestnut Carbon's ~7 Mt expansion in the US. Symbiosis Coalition first-round offtakes named Mombak in the Brazilian Amazon. The Cerrado is not yet the named winner in the major AMC offtake disclosure cycles, but VM0047's first-issuance approval establishes the methodology basis for the next 24 months of disclosure.

The reading list

Verra · "Verra Approves First Credits under VM0047 ARR Methodology." Project 5511 issuance approval.

Verra · "Verra Publishes Minor Revision of ARR Methodology (VM0047)." The v1.1 minor revision.

ICVCM · CCP-Approved Methodologies registry.

ICVCM · "Assessment Status" page and "Impact Report 2025."

Carbon Herald · "Verra Registers Landmark Carbon Project Under ICVCM-Approved Methodology."

BeZero Carbon · "Assessment of Verra's new VM0047 ARR Methodology."

EP Carbon · "How the Recent Updates to Verra's VM0047 Methodology Improves Project Quality."

NGFS · "NGFS publishes latest long-term climate macro-financial scenarios climate risks assessment" (Phase V, November 2024).

NGFS Scenarios Portal. The substantive download for Phase V data.

NGFS · "Statement regarding physical risk estimates in Phase V." The Kotz et al. retraction note.

Sylvera · "Policy consultation response: Verra's Long-Term Reversal Monitoring System."

QCIntel · "Verra updates VM0047 ARR methodology following feedback."