This analysis is able to be downloaded as a static PDF.

Key insights

(Click or tap to expand)

In more detail

Scheme-wide changes in emissions

• Before accounting for offset use or the entry and exit of facilities from the scheme between years, total on-site emissions from safeguard facilities declined marginally between the first and second year of the Safeguard Mechanism. In 2024-25 total emissions from facilities covered by the scheme amounted to 132.8 million tonnes, 2.4% below where emissions were in 2023-24. However, this top line figure significantly overstates the total benefit delivered.
• Almost all of this change in total emissions is attributable to facilities entering and exiting the scheme. Facilities exiting the scheme were responsible for 4.2 million tonnes in 2023-24 and facilities (re-)added to the scheme in 2024-25 were responsible for 1.4 million tonnes. The net difference of transfers in and out of the scheme amounts to 2.8 million tonnes, or 85% of aggregate on-site emissions reductions between the two years. The aggregate of emissions reductions at continuing safeguard facilities amounts to just under 500,000 tonnes, a reduction of less than four tenths of one percent (0.4%).

• The temporary shutdown of one underground coal mine – Grosvenor mine near Moranbah, Queensland – produced more than double the emissions reduction that occurred at ongoing safeguard facilities. In June 2024, a significant fire at the Grosvenor mine led to its closure for the full 2024-25 year. This also caused the mine to temporarily exit from the scheme as reported emissions fell below the safeguard threshold. In 2023-24, that mine reported 1.1 million tonnes of greenhouse gas emissions. As noted above, the aggregate of reported emissions from ongoing facilities under the Safeguard Mechanism declined by just 500,000 tonnes between years. Had the Grosvenor mine fire not occurred, then the aggregate of emissions from continuing facilities would have increased by 600,000 tonnes between years.

The impact of new facilities and future trends in safeguard emissions

• The start-up of the Barossa gas field and processing of this gas for export will likely outstrip any future inter-annual emissions reductions at safeguard facilities. In the second half of 2025, first gas was produced from the Barossa gas field north of Darwin. This gas is now being piped to the Darwin LNG terminal for export. According to Santos’s latest Production Operations Environment Plan for Barossa, the gas field will produce roughly 2.4 million tonnes worth of emissions per year when at full production. An average of roughly 1.6 million tonnes will be produced at the Darwin LNG terminal while processing this gas for export in the same period. Total emissions will therefore be over 4 million tonnes per year. Due to the need to ramp production from the field and terminal during 2025-26, and the fact that the field has only been in operation for part of the year, this total is unlikely to be hit in the third year of the reformed scheme. However, what is produced will almost certainly exceed the 500,000 tonne drop seen between the first and second years.

• While estimates vary based on subjective judgement, there are around 50 coal and gas projects in the pipeline – including new mines and basins, extensions and expansions – that might plausibly come online over the next ten years. These projects vary significantly in size. However, the combined impact of these new facilities, extensions and expansions is likely to either outpace or match any closures, contractions, and emissions reductions that occur at safeguard facilities. Safeguard Mechanism. While individual projects might see success, total on-site emissions at safeguard facilities are likely to stay high for the foreseeable future. Meaningful on-site emissions reductions are very unlikely to occur until after 2030, and possibly as late as 2035.
• Most of these projects will be counted by the Safeguard Mechanism as extensions or expansions to existing facilities – rather than new facilities in their own right – and so can avoid the more onerous obligations that apply to new safeguard facilities. There is a loose cohort of projects for whom the link to the existing facility is especially questionable. For example, the new Lake Vermont underground coal mine shares little other than an owner, a CHPP and a property boundary with the existing open cut mine. Projects like these have a more tenuous claim to being part of the existing operation, but will likely be able to claim it under existing safeguard rules.

Offset use in the first years of the reformed Safeguard Mechanism

• Between the first and second years of the reformed scheme, reliance on offset units – ACCUs and SMCs – increased by almost 50%. 13.1 million offset units were surrendered to meet obligations in the second year of the reformed scheme, while just 9.0 million were surrendered in the first. 80% of the offset units surrendered were ACCUs. 
• The demand for offset units in the second year of the reformed safeguard mechanism is very much higher than expected. The federal government’s 2024 emissions projections had total demand for offset credits in 2024-25 at just 8.7 million units.
• More than three-quarters of ACCUs surrendered were issued under methods that bear on-going, significant, and unresolved integrity issues. In the Safeguard Mechanism’s second year:
  • 35.0%, or 3.7 million units, were surrendered from human-induced regeneration projects.
  • 26.5%, or 2.7 million units, were surrendered from landfill gas projects.
  • 20.3%, or 2.1 million units, were surrendered from avoided deforestation projects.

These three methods cumulatively provided 82% of the ACCUs surrendered in 2024-25. All were all targeted for review in the 2022 Chubb Review of Australian Carbon Credit Units, discussed further below.

• It is undoubtedly true that these offset methods have driven some positive action. However, there is no plausible claim that each notional tonne of abatement delivered under these methods is linked to verifiable emissions reductions that would not have occurred otherwise. While the criticisms of these methods are nuanced and rely on the kind of detailed appraisal of the methodologies that cannot be properly explained here, in an oversimplified form:
  • Criticisms of the human-induced regeneration methods primarily relate to the degree to which the selected land is capable of hosting permanent vegetation growth, the degree to which the actions of the project operator can be shown to have led to any subsequent vegetation growth and the degree to which pre-existing vegetation has been accounted for in calculating the abatement delivered.
  • Criticisms of the landfill gas methods primarily relate to the question of whether the claimed abatement delivered by each project would have occurred anyway given the numerous revenue streams available, the strict environmental obligations placed on landfill operators, and the reality that many of these facilities have been in operation for over 20 years, generously subsidised by taxpayers for that entire time.
  • Criticisms of the avoided deforestation methods primarily relate to whether the often marginal lands that are notionally being protected were ever at serious risk of being cleared, and – given the scale – even whether the practical means to clear the notionally protected properties ever existed. Alongside systemic critiques, there are also concerning anecdotal stories that the revenue provided from the sale of avoided deforestation ACCUs has been used to purchase and clear other land, cancelling out the claimed benefit.
• Unfortunately, the federal government is yet to grapple with these well-founded critiques in an earnest way. While the 2022 Chubb Review of Australian Carbon Credit Units was tasked with reviewing these methods in light of the sustained criticism of them, it did not meaningfully engage with that criticism. This has left much of the necessary work to either improve upon or affirm the integrity of these methods undone. The final review report made essentially no effort to engage fully with the arguments of – or offer effective counter-arguments to – those who have raised concerns with these offset methods. Tellingly, the final review report’s entire summary of the status quo, discussion of the critiques, and counter-arguments to those critiques ran to just six pages, including large photos and re-statements of the panel’s terms of reference. No evidence of any kind was provided to support the review panel’s findings on these matters.
• This unfinished business leaves a deep pall hanging over these ACCU methods. Given the dominance of these methods in the scheme, that pall equally blankets the integrity of the Safeguard Mechanism as a whole. Precisely quantifying the scale of the impact is somewhere between impossible and intractably subjective. However, a plausible argument could be made that half or more of all ACCU credits surrendered under the Safeguard Mechanism come from projects that deliver little or no demonstrable environmental benefit.
• Any attempt to claim that the Safeguard Mechanism has delivered emissions reductions aligning with the sum of baselines under the scheme is strictly not true. So far, facilities are meeting their declining baselines. However, at this stage the scheme is almost exclusively operating as a clearinghouse for low-integrity offset units and on-site emissions reductions are not occurring under the scheme in a meaningful way. The delivery of these offset units does not reliably indicate that real world emissions reductions are occurring. Any public statements based on net emissions reductions under the scheme as determined by the aggregate of baselines are quite simply false. A spreadsheet is not necessarily a reliable indicator of reality. 

Offset use in the future

• Offset use under the Safeguard Mechanism is set to increase into the future. The federal government anticipates that on-paper emissions reductions delivered through the use of offset credits will be the main way that facilities meet their obligations under the safeguard mechanism until at least 2035. According to the most recent federal emissions projections, on-site emissions reductions will only pass the purchase and surrender of offset units as the main way that facilities meet their safeguard baselines in the second half of the 2030s.
• Long-term forecasting of emissions reductions is a mug’s game, but DCCEEW’s assessment is reasonable based on today’s knowledge. Third-party assessments of the future of the Safeguard Mechanism broadly align with DCCEEW’s assessment of its future. The ready availability of low-integrity – and therefore cheap – offset credits necessarily mutes demand for more robust offset units and on-site emissions reduction activities. There is only poor-quality evidence to indicate that the supply of low-integrity units might decline over time, and new low-integrity offset methods might be accredited at any time. This means that the incentive to reduce emissions on-site through meaningful emissions reduction actions is likely to be muted for the foreseeable future.
• This is the scheme working as intended. Those who would prefer to mute criticism of the Safeguard Mechanism are quick to point out that the heavy reliance on offsets over on-site emissions reductions is not evidence that the scheme is failing. They are correct. This is the scheme working precisely as it was designed.

Rather than asking whether the reformed Safeguard Mechanism is failing, it is better to ask whether the scheme should have been designed to do more, and whether the Review of the Safeguard Mechanism scheduled for later in 2026 is an opportunity to deliver it.

Fossil fuel production facilities under the Safeguard Mechanism

• In 2024-25, facilities that primarily produce, transport and process coal, gas and oil were responsible for over half (56.9%) of all emissions from covered facilities under the scheme. This is a small increase on the share in 2023-24 (56.6%), though cumulative absolute emissions from these sectors fell by 1.7% between years.
• While absolute emissions from coal, gas and oil production fell between years, total production of fossil fuels fell by more, meaning an overall increase in emissions intensity, despite the safeguard mechanism. Meanwhile:
  • Production of Australian black coal fell by 2.6%
  • Production of Australian gas fell by 2.7%.
  • Production of Australian liquid petroleum fell by 4.3%.
  • Production of naturally occurring Australian LPG fell by 4.9%.

These declines in production universally outpace the fall in emissions, meaning that between years the production of every tonne of coal, litre of oil, or cubic metre of gas produced more emissions than it did in the year before. This sector is going backward on emissions reduction, rather than forward.

• The aggregate backsliding of this sector is made even more remarkable by the on-paper success of Santos’s Moomba carbon capture and storage facility. Past experience with Chevron’s Gorgon CCS facility indicates that we should reserve final judgment on the merits of Santos’s Moomba CCS facility until it can reliably capture emissions over several years. On-paper, Gorgon was extraordinarily successful right up until it began to fail. In 2025 Chevron captured less carbon dioxide at Gorgon than ever before. We have no reliable indication that Moomba will deviate from Gorgon’s precedent. However, on paper Moomba should have captured sufficient carbon dioxide to tip the scales for the fossil fuel production industry as a whole. The fact that the sector’s emissions intensity has increased despite Moomba CCS indicates a very concerning early trajectory for the sector. The increase in emissions-intensity is an issue that will be made substantially worse in future as the Barossa – and later Browse – gas fields begin to show up in the data, with their comparably high carbon dioxide content.
• It is worth reiterating that the Safeguard Mechanism does not account for most of the emissions impact of the Australian coal, gas and oil producers. The aggregate greenhouse emissions of those Australian fossil fuel production, processing and transport facilities captured by the Safeguard Mechanism amounted to 76.2 million tonnes. However, this figure captures only those emissions that occur on-site. It does not consider the full climate impact. The cumulative impact of burning just those Australian fossil fuels that are exported overseas amounts to more than 1.1 billion tonnes of greenhouse gas, around 15 times more than those emissions that are subject to slowly increasing offset requirements under the Safeguard Mechanism.

Frequently Asked Questions

What is the Safeguard Mechanism?

The Safeguard Mechanism is an Australian greenhouse gas emissions trading scheme and the country’s most expansive climate policy. Under the scheme, Australia’s largest industrial facilities are bound to annually-declining emissions limits, known as baselines.

Each facility has its own individual baseline that is set using complex formulas that reference that facility’s own historical emissions intensity, the performance of other facilities in its sector, and its annual production in each year. Usually, facility baselines decline by 4.9% each year. However, facilities operating in identified ‘trade-exposed’ sectors may apply to slow this annual decline rate if the cost of meeting their obligations exceeds certain thresholds.

In 2024-25, 208 facilities were covered by the Safeguard Mechanism. This includes gas terminals, coal mines and metals mining and processing facilities, along with many other sectors. Combined, these facilities reported a total of 132.8 million tonnes worth of greenhouse gas emissions. This is just under the annual emissions from the Australian electricity sector and nearly one-third of Australia’s total net annual emissions.

The scheme does not cover emissions from farming, land clearing or the production of electricity supplied to one of Australia’s five largest grids.

Facilities bound by safeguard baselines can either reduce their reported emissions to below their annual emissions limit in each year or surrender offset credits to bring their annual emissions below the permitted level on paper. To date nearly all action to comply with the Safeguard Mechanism has occurred through the purchase and surrender of offset credits.

The Safeguard Mechanism commenced in 2015, but was heavily reformed in 2023. It is scheduled to be reviewed later in 2026.

What is an ACCU? What is an SMC?
The Safeguard Mechanism treats as substitutable two different offset units, the Australian Carbon Credit Unit (ACCU) and the Safeguard Mechanism Credit (SMC). Both units are notionally equivalent to one tonne of carbon dioxide equivalent greenhouse gas, and either unit can be surrendered to lower a facility’s net emissions to meet safeguard baselines without penalty.

ACCUs have existed since 2011. These credits are issued to registered projects around the continent that use approved methods to either draw greenhouse gas emissions from the atmosphere or avoid them in the first place. After projects are credited for meeting the requirements of their methods, these credits are frequently sold to major emitting facilities who surrender them under the Safeguard Mechanism. 

ACCUs have been the primary offset unit underpinning the Gillard-era Carbon Farming Initiative, the Abbott-era Emissions Reduction Fund and now the Safeguard Mechanism. The three largest approved methods have faced considerable criticism primarily relating to whether the promised abatement has occurred at all, has occurred in the quantities claimed, or is attributable to the ACCU scheme. 

SMCs are new offset units created under the reformed Safeguard Mechanism. These are issued to facilities that report emissions below their total allowances in a given year. Credits are issued on a one-to-one basis for any unused allowances and are transferable and tradeable between facilities and operators. 

While in theory SMCs might create an incentive for facilities to reduce emissions below their baseline, in practice almost all credits that have ever been issued and likely ever will be issued are granted to facilities whose emissions have fallen below their baseline for unrelated reasons. By design, no additionality tests are placed on the issuing of SMCs: meaning there are no checks for whether the creation of an SMC has delivered a concrete environmental benefit or is attached to any other beneficial actions from the operator. This element of the scheme has previously been criticised as ‘money for nothing’, particularly in the context of the Australian coal industry.

---

About Naru Research

Naru Research is a new climate and energy research consultancy founded by Tim Baxter, a respected expert in the field with a demonstrated track-record of success working with the Australian NGO sector. Tim prides himself on his willingness to assess climate and energy policy on its raw merits – without fear or favour – and an exclusive commitment to verifiable real-world outcomes and demonstrated benefit.