Article 6.4 Cookstove Methodology: How Many Credits a Stove Keeps

At a Glance · Where Things Stand

Climate Decode expects the Article 6.4 cooking methodology to pay fewer credits per stove than Gold Standard, but not because it doubts the fuel savings. The charcoal emission factors are within 9% of each other on a common basis. What cuts the credits are the deductions Article 6.4 applies after the net reduction: the reversal buffer, a rising downward adjustment, 7% mechanism deductions and, for stoves measured by kitchen performance tests, a Hawthorne default from the first year.

Metering is the dividing line. A metered induction stove keeps 60% of its MECD credits with no Hawthorne adjustment. An improved charcoal stove measured by kitchen performance tests carries both the Hawthorne default and the buffer, and keeps 58% of its RECH credits.

Most of the gap can be narrowed. Developers of biomass stoves who fit stove use monitors to the kitchen performance test sample from the first campaign raise credits from 0.38 to 0.53 tCO2e a year, and a host country that qualifies for a lower buffer takes them to 0.58.

The Article 6.4 Supervisory Body adopted the mechanism methodology for cooking energy transitions (A6.4-SBM023-A09, version 01.0) at its October 2026 meeting. It entered into force on 9 October 2026 for improved biomass, electric and LPG cookstoves and is valid until 8 October 2031.¹

This review compares it with the two Gold Standard methodologies published on 5 May 2026, Metered and Measured Energy Cooking Devices (MECD) v2.0 and Reduced Emissions from Cooking and Heating (RECH) v5.0, and with Verra’s VM0050 v1.0, which covers both stove types.

Table 1 · Methodologies and the stoves they are compared on. Sources: A6.4-SBM023-A09; Gold Standard MECD v2.0 and RECH v5.0; Verra VM0050 v1.0.

RECH v5.0 no longer covers metered electric cooking, which has moved to MECD v2.0, and MECD requires every device to be metered, so it does not apply to an improved charcoal stove. The induction stove is therefore compared across MECD, VM0050 and Article 6.4, and the improved charcoal stove across RECH, VM0050 and Article 6.4.

Both examples use the same urban sub-Saharan household, with crediting from 2027 over the life of each stove: five years for the induction stove and three for the improved charcoal stove. The only thing that changes is how each methodology turns the same field data into credits.

Table 2 · Example household assumptions. Sources: Climate Decode model; methodology defaults as cited.

The comparison uses a 4:1 wood-to-charcoal ratio because that is the basis that meets the CCP labelling condition. VM0050 is modelled on its conversion route at its 4:1 default, which converts charcoal to wood and applies the wood factor of 112 tCO2/TJ, so its CO2 factor equals the Gold Standard 4:1 value.

Climate Decode models crediting under MECD, RECH, VM0050 and Article 6.4 side by side on a programme’s own data.

All four methodologies share the same baseline structure: the CO2 factor is multiplied by fNRB and the non-CO2 factor is added in full. They differ in what goes into the charcoal emission factor and in what is deducted after the net reduction. Table 3 sets out each parameter and its effect on credits per stove-year in the example households.

Table 3 · Parameter comparison and effect on credits, tCO2e per stove-year. Effects are rounded and may not sum exactly. Sources: methodology documents as cited in the notes; Climate Decode model.

Charcoal CO2 factor

MECD and RECH derive the charcoal CO2 factor from the wood burned in the kiln: 236.91 tCO2/TJ at a 4:1 ratio, and 355.36 tCO2/TJ at the 6:1 ratio both set as the sub-Saharan default. Article 6.4 uses measured factors of 89 tCO2/TJ for combustion and 98 tCO2/TJ for kiln production, 187 tCO2/TJ in total at an implied 5:1 ratio.^1,2

VM0050 offers a direct factor of 165.22 tCO2/TJ including production, or a conversion route that multiplies the wood factor of 112 tCO2/TJ by the wood-to-charcoal ratio, giving the same 236.91 tCO2/TJ as Gold Standard at 4:1.³ On the common basis the Article 6.4 factor is about 21% lower, which cuts credits by 0.20 tCO2e a year for induction and 0.07 for the improved stove.⁴

Charcoal non-CO2 factor

Methane and nitrous oxide from the kiln and the stove are added in full rather than scaled by fNRB, so this factor carries a lot of weight at an fNRB of 0.30. Gold Standard applies 61.74 tCO2e/TJ at 4:1 and Article 6.4 applies 65.157 tCO2e/TJ, about 5% more, which adds 0.04 and 0.02 tCO2e a year.^1,2

VM0050 allows the production non-CO2 factor on the conversion route only where a project-specific, regional or national value exists; the IPCC default does not qualify. A project relying on defaults applies the combustion-only 5.865 tCO2e/TJ and loses 0.73 tCO2e a year for induction and 0.27 for the improved stove against Gold Standard. For induction, that is more than the whole difference between the two.³

Downward adjustment

Gold Standard applies a country factor by period, 2.76% for Ghana from 2026 to 2030 in this example. Article 6.4 applies the higher of a Monte Carlo 10th-percentile result and 10% of the first-year emission reduction ratio, rising one point a year.⁵ On the minimum rule alone, the adjustment starts at 5.4% for induction and 4.0% for the improved stove and costs 0.10 and 0.04 tCO2e a year. A Monte Carlo result, not run here, is likely to be higher given the uncertainty around fNRB and the non-CO2 factors. VM0050 has no equivalent, which adds 0.04 and 0.03 tCO2e a year relative to Gold Standard.

Hawthorne effect

How use is measured decides the Hawthorne treatment. MECD exempts metered activities, and Article 6.4 applies its adjustment to activity emissions derived from kitchen performance tests, so a metered induction stove carries none under either.⁶

For the improved stove, RECH phases the factor in at 0.90 for 2026 and 2027, 0.85 for 2028 and 2029 and 0.75 from 2030. Article 6.4 applies 0.75 from the first year, valid until 31 December 2027, after which the factor must be measured with stove use monitors, mandatory from 1 January 2029. Applying the default earlier costs 0.08 tCO2e a year against RECH over a three-year life. VM0050 v1.0 has no Hawthorne adjustment, worth 0.06 tCO2e a year relative to RECH, although draft v2.0 introduces a 0.75 default for kitchen performance tests and exempts metered devices.⁶

Leakage

MECD and RECH deduct 2% for market leakage and spread embodied emissions of 0.017 tCO2e per device over five years. VM0050 applies a 0.95 factor to net reductions. Article 6.4 deducts 0.0128 tCO2e per industrially made stove once, in the first year, with no market leakage deduction.⁷ That adds 0.02 and 0.01 tCO2e a year relative to Gold Standard, while the VM0050 factor costs about 0.01 for both stoves.

Mechanism deductions

Article 6.4 sends a 5% share of proceeds to the Adaptation Fund and cancels at least 2% for overall mitigation in global emissions before credits reach the activity participants. Neither Gold Standard nor Verra has an equivalent, and the deductions cost 0.07 and 0.03 tCO2e a year.⁸

Reversal buffer

Article 6.4 sends a share of the fNRB-related CO2 reductions to its Reversal Risk Buffer Pool at a rate set by the reversal risk assessment tool. The tool went to the Supervisory Body in October 2026 as a draft, with a human-induced factor of 20% to 45% and reduction factors of 20% to 80%, so the model runs cases of 11%, 26% and 49%.⁹

MECD treats cooking under a no-control exemption and requires no buffer, and VM0050 sets none. At the mid case the buffer costs 0.18 tCO2e a year for induction and 0.08 for the improved stove; across the three cases, 0.08 to 0.34 and 0.03 to 0.15.

Baseline caps and sampling

Every methodology caps back-calculated baseline consumption, and the example household, at 0.13 t of charcoal per person-year, sits within every cap.¹⁰ Article 6.4 also requires sampling at 95/10 confidence and precision, against 90/10 under Gold Standard and VM0050, and makes stove use monitors the sole basis for usage from 1 January 2029. These affect sample size and monitoring cost rather than credits per stove and are not modelled.¹¹

Crediting period and technical life

MECD and RECH allow a five-year crediting period, renewable twice, which may not exceed the technology’s technical lifetime unless replacement and maintenance are in place. Article 6.4 refers to its technical lifetime tool and requires a lock-in analysis only for non-electric stoves lasting more than ten years.¹² The model credits the induction stove over five years and the improved stove over three, the lower end of the 3.1 to 6 years found for improved charcoal stoves in Kenya.¹³

The induction stove earns 1.20 tCO2e a year under MECD, 0.50 under VM0050 and 0.57 to 0.83 under Article 6.4, depending on the buffer. Electricity emissions of 0.54 tCO2e a year are the same under every methodology, so a lower charcoal factor reduces the net reduction proportionately more than it reduces the baseline.

Table 4 · Induction stove emission reductions. Source: Climate Decode model; generic urban sub-Saharan household, fNRB 0.30, grid 0.40 tCO2/MWh, MECD at its 4:1 factors.

Figure 1 · Induction stove credits per year, MECD v2.0 to Article 6.4 at the mid buffer. Metered stove, no Hawthorne adjustment. Steps are rounded and may not sum exactly. Source: Climate Decode model.

Over its three-year life the improved charcoal stove earns 0.65 tCO2e a year under RECH, 0.45 under VM0050 and 0.31 to 0.42 under Article 6.4. The Hawthorne adjustment applies under both RECH and Article 6.4, because a biomass stove’s savings are measured through kitchen performance tests.

Table 5 · Improved charcoal stove emission reductions. Source: Climate Decode model; generic urban sub-Saharan household, fNRB 0.30, three-year stove life, RECH at its 4:1 factors.

Figure 2 · Improved charcoal stove credits per year, RECH v5.0 to Article 6.4 at the mid buffer. Source: Climate Decode model.

Most of the parameters that separate Article 6.4 from Gold Standard can be influenced by an activity developer or a host country. Table 6 shows the effect of each on Article 6.4 credits at the mid buffer.

Table 6 · Effect of developer and host-country measures on Article 6.4 credits. Source: Climate Decode model.

For activity developers

1. Meter every electric stove and credit from the meter data, which carries no Hawthorne adjustment under Article 6.4.
2. Fit stove use monitors to the kitchen performance test sample of biomass stoves from the first campaign. It replaces the 25% default with a measured value, and it becomes mandatory from 2029 in any case.
3. Run kitchen performance tests at 95/10 and document every uncertainty input, since the Monte Carlo downward adjustment rewards measured precision.

For host countries

1. Qualify for a lower buffer. The draft tool’s reduction factors apply where the NDC covers land use and is tracked in the biennial transparency report, a long-term low-emission strategy commits to halting deforestation, and all four REDD+ elements are in place.
2. Set a national fNRB under the Article 6.4 fNRB tool and grid factors under its electricity tool, since these set the starting point for every activity.

Host-country support matters for buyers too. Sovereign programmes such as Norway’s Article 6 programme pay only for authorised reductions, so the national choices above feed directly into what a programme can buy.

1. Gold Standard sets 6:1 factors of 355.36 tCO2/TJ and 89.68 tCO2e/TJ as the sub-Saharan default, with 4:1 reserved for developers seeking CCP label alignment. At 6:1 the induction stove earns 2.03 tCO2e a year under MECD and the improved stove 0.96 under RECH, but these credits do not meet the CCP labelling condition.
2. VM0050 on its direct charcoal factors of 165.22 tCO2/TJ and 44.83 tCO2e/TJ, which include production, gives 0.73 tCO2e a year for induction and 0.56 for the improved stove. It allows up to 6:1 on the conversion route where government-endorsed or peer-reviewed values support it.
3. The Article 6.4 buffer cases of 11% and 49% move induction credits to between 0.57 and 0.83 tCO2e a year, and improved stove credits to between 0.31 and 0.42.
4. A Monte Carlo downward adjustment above the 10% minimum would cut Article 6.4 credits further for both stoves, so the Article 6.4 figures are an upper estimate.

1. The Article 6.4 Hawthorne adjustment applies to activity emissions measured through kitchen performance tests. An electric stove credited from meter data and controlled cooking tests carries none; the text is less explicit where an electric stove is credited through activity kitchen performance tests. For the improved stove the model applies 0.75 in every year, since the default is valid until 31 December 2027 and stove use monitors become mandatory from 2029.
2. The reversal risk assessment tool remains a draft, with ranges for its human-induced and reduction factors. Natural risk values have been reported only for some countries, for example Kenya 7.4%, Tanzania 8.7% and Nigeria 12.8%.
3. Every methodology uses an fNRB of 0.30 here, although each prescribes its own source, so values in practice will differ by country and methodology.
4. No published technical lifetime was found for induction cookers, so five years, one crediting period, is used.
5. The 90% usage of the improved stove is a household assumption applied equally under every methodology.

Climate Decode Platform

Cookstove supply, ratings and labels are tracked in Canopy Market Watch across the main registries.

Climate Decode · Advisory

Article 6.4 and cookstove portfolios

Methodology choice, Article 6 authorisation and crediting outcomes, modelled on programme data.

Notes

Sources

This review models a generic household to compare methodologies. It is not a crediting estimate for any project, and the reversal risk tool it relies on is still a draft.