Bulk of emissions reduction from Safeguard Mechanism is smoke and mirrors
· Michael West
The 2nd year of Safeguard Mechanism data shows only one fifth is genuine emissions abatement. David McEwen finds the devils in the detail.
Visit truewildslot.com for more information.
Behind the government’s headline claims of falling emissions at Australia’s largest industrial facilities lies a more complicated reality, one driven more by closures, accidents and accounting than by genuine decarbonisation.
On a “net” basis, the Safeguard Mechanism seems to be achieving its objective of reducing most large industrial facility emissions by 4.9% per year. That’s how it’s set up. But most of the decline is done with offsets and other accounting. The real work of reducing emissions has hardly begun.
How the Safeguard Mechanism works
The Safeguard Mechanism, reformed in 2023, is Australia’s primary policy for reducing emissions at large industrial facilities. It covers around 220 facilities emitting more than 100,000 tonnes of CO₂-equivalent annually – predominantly in mining, oil and gas, manufacturing, transport and waste sectors.
Together, these facilities account for nearly 30% of Australia’s total national emissions, so it’s important that it does what it says on the packet.
Under the current scheme, for which the second year’s worth of data was released last week, each facility has an emissions baseline that generally must decline by 4.9% per year through to 2030. The policy is designed to contribute to Australia’s commitment to reduce emissions by 43% below 2005 levels by 2030 and reach net zero by 2050.
But there’s a lot of smoke and mirrors when you delve into the detail.
If a facility exceeds its baseline, the operator must ‘offset’ the excess by surrendering Australian Carbon Credit Units (ACCUs) or Safeguard Mechanism Credits (SMCs) purchased from facilities that beat their targets. These baselines are production‑adjusted, meaning they rise and fall with output – a design choice that becomes important later.
The offset problem
In 2024-25, a striking 147 out of 228 facilities – nearly two-thirds – exceeded their emissions baselines. To achieve compliance, these facilities surrendered 10.5 million ACCUs and 2.6 million SMCs – a significant increase from the 7.6 million ACCUs and 1.4 million SMCs surrendered the previous year.
This means the vast majority of ‘compliance’ with declining baselines is being achieved not through actual emissions reductions at facilities, but through the purchase of carbon credits.
While offsets have a legitimate role in the transition, their escalating use raises questions about whether the scheme is driving the at-source abatement it was designed to deliver. And ACCUs vary in integrity – for example, firms are still permitted to surrender previously issued “avoided deforestation” credits, despite this type of credit being discontinued following the Chubb Review.
Unlike Australia, the EU Emissions Trading System does not allow companies to meet their obligations with external offsets. All compliance must come from within the scheme,
which is why EU carbon prices are four times higher
– and why the system drives real abatement.
Offsets aside, what’s really happening?
The Clean Energy Regulator’s latest data release shows actual emissions from Australia’s largest industrial facilities actually fell only 4% in the two years the scheme has been in operation (from 138.6 Mt CO2-e in 2022-23 to 132.8 in 2024-25).
But a closer examination of the numbers reveals even this modest figure significantly overstates the success of the Safeguard Mechanism in driving genuine emissions reductions.
Analysis of the facility-level data reveals a troubling picture.
Of the reported 5.8 million tonne decline in actual emissions, only around a fifth can be attributed to verified genuine emissions abatement, and most of that comes from a single project: Santos’s Moomba carbon capture and storage (CCS) facility.
You may recall that a model of the Moomba facility was prominently displayed in the Australian stand at the COP26 climate conference in Glasgow: it’s now operational and making a dent in that gas project’s on-site emissions, as well as generating a record number of ACCUs for Santos.
Even then, Moomba’s benefit applies only to the extraction site.
The far larger emissions from burning the exported gas remain untouched.
Woodside and the baseline paradox
Perhaps the most revealing insight comes from Woodside’s North West Shelf Project – Australia’s largest LNG processing facility. Despite its emissions falling by 17% (from 6.94 million to 5.75 million tonnes), this year it still exceeded its government mandated Safeguard baseline by 921,000 tonnes and had to surrender that amount in carbon credits to achieve compliance.
How is this possible?
Under the reformed Safeguard Mechanism, baselines are ‘production-adjusted’ – meaning lower production earns a lower baseline allocation. As the North West Shelf’s mature gas fields deplete, production has declined. But LNG facilities have significant fixed energy requirements for compression and liquefaction that don’t scale down proportionally.
The result: emissions intensity (emissions per unit of output) actually increases as production declines.
The 40‑year extension granted last May applies only to the LNG export terminal. It does not approve any new gas fields. Without new upstream approvals – particularly the Browse Basin – production continues to decline, which in turn lowers the baseline and raises emissions intensity.
Back to the Safeguard Mechanism, and this reveals a design challenge: facilities with declining production may face increasing difficulty meeting their baselines – even as their absolute emissions fall. Combined with the 4.9% annual baseline reduction,
mature facilities may need more offsets over time, not fewer.
This somewhat counterintuitive outcome means the scheme may inadvertently penalise facilities that are naturally winding down.
Closures and incidents have driven recent reductions
The list of facilities included in the Safeguard Mechanism changes from year to year as projects start, enter the scheme when they reach the threshold of 100,000 tonnes of emissions annually, expand, and then contract towards the end of the facility’s life.
Over the past two years, up to three dozen new facilities entered the Safeguard Mechanism – including new coal mines. Meanwhile, the number of genuine facility exits has been modest. Amongst them was Alcoa’s Kwinana Alumina Refinery (WA). It permanently closed in September 2025 after 60 years of operation, citing age, costs and market conditions. Emissions fell 87% – not from efficiency gains, but from shutting down.
Meanwhile, unexpected incidents have had a downward effect. For example, Anglo Coal: its Grosvenor Coal Mine (QLD) was sealed since a methane explosion in June 2024, reducing Safeguard emissions by around a million tonnes!
Its reported emissions dropped 97% because the mine isn’t operating.
Meanwhile Anglo’s Moranbah North Mine (QLD) has experienced various safety incidents, including an underground coal fire in March 2025. Reported emissions fell 38% compared to the previous year.
Clearly, closures and accidents are not climate policy.
Gas and coal not pulling their weight
Unsurprisingly, gas and coal facilities – which together account for over half of total Safeguard emissions – contributed only about 13% of the net decline over the last two years.
In the coal sector, 42 facilities actually increased their emissions, while only 20 decreased. The sector’s modest net reduction was largely offset by new coal mines entering the scheme, including Pembroke Resources’ Olive Downs mine in Queensland.
In the gas sector, increases at facilities like Chevron’s Gorgon (+830,000 tonnes), Inpex’s Ichthys LNG (+414,000 tonnes) and Chevron’s Wheatstone (+231,000 tonnes) partially offset reductions elsewhere. Gas’s share of total emissions actually increased from 30.3% to 31.0% between 2022-23 and 2024-25.
Where’s the decarbonisation investment?
Announcements of major decarbonisation projects at Safeguard facilities reveal a troubling pattern.
Both BHP and Rio Tinto have significantly scaled back their decarbonisation budgets.
Rio Tinto slashed its budget from US$7.5 billion to US$1-2 billion and disbanded its dedicated decarbonisation division. BHP cancelled its 50MW Jimblebar solar farm and significantly cut decarbonisation spending from US$2 billion announced in 2023. Capital expenditure cuts today mean higher emissions later,
On the flipside, Fortescue’s ambitious ‘Real Zero by 2030’ program seems to remain on track, but most benefits won’t appear in the data until 2026-2030. It’s worth reflecting that genuine emissions reduction projects at scale do take time:
those that aren’t investing had best get their skates on.
The bottom line
The Safeguard data’s headline figure of a 4% decline in actual facility emissions in the two years the declining (but production-adjusted) baselines have been in place tells a selective story.
When you strip away the closures, accidents, and natural production declines, genuine emissions abatement from proactive investment amounts to barely a fifth of the reported reduction.
The Safeguard Mechanism review scheduled for 2026-27 will need to grapple with these uncomfortable realities.
If Australia is serious about its 2030 targets, the scheme needs to drive actual at-source abatement – not just measure the incidental emissions benefits of mine explosions and refinery closures. Phasing out the use of ACCUs would be a good first step.
KEY DATA SOURCES:
- Clean Energy Regulator, 2024-25 Safeguard Mechanism data (published 15 April 2026)
- Clean Energy Regulator, 2023-24 and 2022-23 Safeguard facility data
The dog ate my homework! The missing pieces of the climate plan