NOTE: edits since KIP-26 highlighted in bold
Summary
Allocate a limited amount of treasury BCT to remove the problematic HFC-23 credits from the BCT pool and properly dispose of them.
Background on HFC-23 Credits
In order to ensure all community members can make a fully informed decision justified by a nuanced understanding of the situation, the original text of KIP-26 has been amended to include the following background material on the history of HFC-23 disposal as a carbon credit methodology, as well as data on the latest activity of HFC-23 credits.
This background material serves to justify the proposal to remove these credits from the BCT pool and destroy them without claiming the underlying environmental credit, despite them being associated with properly issued Verra Verified Carbon Units that were approved by Toucan for tokenization via their one-way bridge.
- HFC-23, a hydrofluorocarbon, is gas created as a byproduct of refrigerant gas production (HCFC-22) which has a colossal global warming potential.
- HFC-23 projects generated over half a billion UN Clean Development Mechanism (CDM) credits, and made significant profits. It costs about 1 Euro to destroy a tonne of HFC-23, but credits sold up to 25Euro (an example of asymmetric market information).
- Some factories ramped up production of the refrigerant HCFC-22, purely to profit from issuing credits by destroying the HFC-23 byproduct.
- Credits were eventually banned in compliance markets; this ban was extended to VCS in 2014.
- Two projects (Yingpeng and Honeywell) had been registered prior to the ban; with only one issuance ever made on VCS (Yingpeng) for 4.5mn VCS for activity in 2008 and 2009.
- Approximately 95k tonnes retired ever; for Chinese banks. It is not known who holds the remaining issued but unretired credits (Verra account is private).
- This paper by Resources for the Future explores the relationship between emissions reduced in 12 HFC-23 projects in China and Mexico between 2006 and 2013, and the number of credits issued by the CDM registry.
- The analysis examined projects that have earned credits under the third, fourth and fifth versions of the CDM’s HFC-23 methodology, installations that account for the vast majority of issued units.
- Main finding: under-credited emissions reductions from HFC-23 nearly match the supply of over-credited tonnage. That is, the CDM HFC-23 projects studied maintain their environmental integrity (on average).
- There had been substantial over-crediting of some projects, which had ramped up their HCFC production during the initial CDM baseline-setting period of 2001-2004, there had also been significant under-crediting of other projects.
- Under-crediting occurred because the baselines evolved to become relatively conservative compared to the business-as-usual output from the HFC-23 projects, which continued to climb globally. Because of the massive profits on offer, factories still sought to participate despite only getting credited for emission reductions below this conservative baseline. Meaning that in reality, the paper found, the facilities were making more substantial reductions below their Business-as-Usual (BAU) levels.
- “If policymakers only think about over-credited offsets when they are considering the emissions impact of HFC-23 projects, then they have failed to look at the far side of the moon. For this project type, under-credited emissions reductions seem to play quite an important role”
- According to the data we received from Allied Offsets as part of their analysis of the BCT pool, the last retirement of credits from an HFC-23 project was in April of 2020. While the beneficiary field was left blank on that retirement, most previous retirees of these credits were Chinese financial institutions.
- Because of the particularly low marginal mitigation costs for HFC-23, this overissuance situation does not necessarily extend to industrial and chemical mitigation methodologies in general.
To sum up: on one hand, the specific HFC-23 project whose credits are in the BCT pool was properly verified and issued legitimate VCUs for actually decomposing HFC-23 and preventing its release into the atmosphere, prior to the methodology being deprecated. On the other hand, because of the misaligned incentives outlined above, compliance markets have explicitly prohibited use of HFC-23 decomposition credits for almost a decade, and major voluntary registries deprecated the methodology in 2014 - so no new projects will have credits issued under it, and there have not been any traditional retirements in Verra using these credits in over two years.
This nuanced situation is why we took additional time to gather input and data even after the original KIP-26 was met with resounding approval by the community.
Motivation
Per the KlimaDAO Tokenized Carbon Dashboard’s BCT page, there are approximately 672 thousand tonnes of TCO2 in the BCT pool bridged from Verra’s deprecated HFC-23 methodology, AM0001:

While many have rightly pointed out that these credits should never have been retirable within Verra’s registry - which would have prevented them from being bridged - they met BCT’s initial vintage and methodology criteria and were de-facto allowed into the pool. Once the problem was identified, Toucan promptly excluded the methodology from the BCT pool requirements.
The ReFi community has taken efforts to prevent the use of HFC-23 credits by excluding them from the no-fee automatic redemption function on Toucan’s TCO2 contracts, so anyone trying to consume the HFC-23 credits would have to pay the steep selective redemption fee. They have also been disabled in the KlimaDAO retirement aggregator. However, the presence of these credits is still a limiting factor for the integrity, demand and adoption of BCT as a ReFi building block.
Methodology
While we had hoped to partner with Toucan to equally split the burden of this cleanup initiative, as indicated by Toucan’s official response to our RFC on their governance forum, they have decided not to allocate the capital required to acquire the necessary amount of BCT for their half.
However, they have agreed to collaborate in a more limited manner:
- Toucan has implemented a new “redeemAndBurn” function to facilitate pool cleanup. No fees are charged for redemptions made using this function, which saves KlimaDAO roughly 33% in tonne terms (about 335k BCT) on the operation.
- They will utilize the accumulated fees in the regenerative BCT selective redemption fee wallet for destroying HFC-23 tonnes (about 14,330 tonnes as of writing, approx. 2% of the HFC-23 total).
If this proposal is approved, the DAO multisig would be used to withdraw at most 672,249 tonnes of BCT from the treasury to be used for permanently destroying the HFC-23 TCO2 tokens.
This amount is quoted as a maximum limit rather than an exact value, because Toucan will be burning the small amount of BCT accumulated in their fee wallet. The exact amount of BCT withdrawn from the treasury and burned will be determined by subtracting the amount burned via the Toucan redemption fee multisig.
NOTE: For community members who would like to participate in the cleanup effort:
- any selective redemption of tonnage from the BCT pool will contribute to Toucan’s regenerative fee wallet,
- or, you can simply send BCT directly to that fee wallet’s address and it will be burned by Toucan: 0x222aAdfAF8FfA6CC7D89F7E99cf94963b121ADbf
Implications
Destroying treasury BCT to clean up the HFC-23 tonnage is expected to have two primary impacts:
- Reduce the amount of excess reserves held in the treasury for backing KLIMA tokens by about 7%, which will reduce staking reward runway by about 140 days at the current reward rate, and reduce CC/KLIMA by about 4%.
- Reduce BCT supply by about 3.5% and increase integrity of the BCT pool (currently the treasury’s largest holding).
Proposal
- Allocate at most 672,249 tonnes of treasury BCT for the HFC-23 cleanup operation.
- Coordinate with Toucan Protocol to:
- burn the accumulated BCT fees designated for this purpose,
- then utilize the allocated treasury BCT to destroy all remaining HFC-23 tonnes.
Links
Polling Period
The informal forum poll begins now and will end in approximately 2 days on September 8th at 18:00 UTC. Assuming in favor, this vote will go to Snapshot.