Safeguard Mechanism Compliance Tracking: Monitoring Emissions Against Declining Baselines

The 2024-25 preliminary Safeguard Mechanism data landed in November, and the headline number should worry every covered facility in Australia. Total exceedance across the scheme hit 13.7 million tonnes — up 48.9% from 9.2 million tonnes the year before. More than 140 facilities are now operating above their declining baseline.

That's not a rounding error. That's the ratchet doing exactly what it was designed to do.

Baselines dropped another 4.9%. Emissions didn't drop fast enough to keep up. And now the compliance deadline for 2024-25 — 31 March 2026 — is days away, with ACCU spot prices sitting around $36.50. For a facility 25,000 tonnes over its baseline, that's roughly $912,500 in surrender obligations. Six months ago, the same ACCUs would have cost $250,000 less.

I spent 18 years at BHP, Rio Tinto, and Senex Energy. I've seen Safeguard compliance from both sides — as someone producing the emissions data and as someone trying to make sense of it for reporting. The single biggest problem isn't the regulation itself. It's the timing of information. Most facilities discover their exceedance position months after they could have done something about it. By then, the only option left is buying credits at whatever the market charges.

That's why we built Safeguard Mechanism compliance tracking into Carbonly — not as an annual reconciliation tool, but as a month-by-month position monitor.

The real-view-mirror problem

Here's how Safeguard compliance works at most covered facilities today. Environmental managers collect emissions data through the year. Sometime in September or October, they compile the NGER report. That's when they find out — for the first time with any precision — whether their facility exceeded its baseline for the year that ended three months earlier.

If they did exceed, they have until 31 March to surrender ACCUs or SMCs. That gives them roughly five months to enter the carbon market, negotiate purchases, and settle — all while every other facility in the same position is doing the same thing.

The first year of the reformed scheme showed this pattern clearly. ACCU surrenders jumped from 1.2 million units in 2022-23 to 7.1 million in 2023-24, according to the Carbon Market Institute. Meanwhile, 8.3 million SMCs were issued to 62 facilities that beat their baselines. But SMC usage was less than 20% of issuance — meaning most credits sat in registries rather than being traded.

That's a market where supply exists but information asymmetry drives bad purchasing decisions. If you don't know you're 15,000 tonnes over baseline until October, you can't start sourcing SMCs in July when they're cheaper. You're a price-taker, not a price-maker.

What monthly position tracking actually looks like

The Safeguard Mechanism covers Scope 1 emissions only. That's a narrower scope than full NGER reporting, but it still requires facility-level granularity — and facility-level data collection is where things fall apart.

Consider a facility with a baseline of 120,000 tonnes for FY2025-26 (after the 4.9% decline from the prior year). By the end of each month, you should know three things: your cumulative emissions year-to-date, your pro-rata baseline position at that point, and the gap between them.

In Carbonly's Safeguard module, that gap drives everything. We calculate the 4.9% declining baseline trajectory for each facility and overlay your actual Scope 1 emissions as they flow in from fuel records, gas measurements, and process data. The status indicators are blunt on purpose:

On track: your cumulative emissions are at or below the pro-rata baseline. No action needed yet.
At risk: you're within 5% above baseline. This is the warning zone — close enough that a production spike or a measurement correction could push you over.
Exceeding: you're 5% or more above baseline. You need a plan. Now.

The point isn't precision to the tonne in month three. The point is knowing your trajectory early enough to act. If you're trending 8% above baseline by December, you've got three months to source ACCUs before the March 31 deadline — not five panicked weeks.

The ACCU purchasing window

This is where the financial argument lives. ACCU prices don't move in a straight line. EY's Australian Carbon Market Outlook (January 2026) projects spot prices staying in the $30-35 range through 2028, rising gradually toward $70 by 2035. But within any given compliance year, prices swing.

The cost containment price for 2025-26 is $82.68 per ACCU — that's the government's ceiling, where the CER will release units from the Commonwealth's holdings to prevent prices spiking above that level. The secondary market sits well below that today. But we've already seen intra-year volatility: ACCUs traded as high as $42 in late 2024 when Safeguard demand compressed against the compliance window, then settled back to the mid-$30s.

If you're running a cement plant 30,000 tonnes above baseline, the difference between buying ACCUs at $34 in August versus $42 in February is $240,000. That's real money. And it's entirely a function of when you had the information to act.

Our ACCU surrender tracking shows you two numbers side by side: how many credits you're likely to need based on your current trajectory, and the estimated cost at current market prices. It's not a crystal ball — we're explicit about that. But it turns "we might need to buy some offsets" into "at current run rate, we need 27,400 ACCUs by March 31, which will cost approximately $986,400 at today's spot price."

That's a number a CFO can work with.

SMCs: the credit you earn by outperforming

Safeguard Mechanism Credits are the other side of the ledger. When your facility's emissions come in below baseline, the CER issues you SMCs for the difference. In 2023-24, 62 facilities earned a total of 8.3 million SMCs.

Here's what's interesting: the preliminary 2024-25 data shows SMC eligibility dropping 15.7% as baselines tighten. Fewer facilities are beating their baselines by enough to generate credits. That's simple maths — the baseline falls 4.9% every year, so the bar keeps rising.

We track SMC generation potential the same way we track exceedance. If your facility is trending 5,000 tonnes below baseline, our module shows the estimated SMC issuance and lets you think about whether to bank them for future compliance years or sell them to other Safeguard facilities.

Banking matters more than most people realise. The 2026-27 review will reconsider whether SMCs remain valid past 2030. If you're generating credits now, the smart move might be to hold them rather than sell at today's price — but only if you believe the post-2030 rules will honour them. That's a genuine strategic uncertainty. We show you the data. We don't pretend to know the policy outcome.

Flexibility measures and when to use them

The Safeguard Mechanism isn't just "stay under baseline or buy credits." There are four compliance pathways beyond direct surrender, and most facilities we talk to don't fully understand when each one makes sense.

Borrowing from next year's baseline lets you take up to 10% of your following year's baseline to cover this year's excess. The catch: it costs you. Interest is 2% for FY2024-25 and FY2025-26, jumping to 10% from FY2026-27 onwards. That 10% rate makes borrowing punishingly expensive from next year — essentially a five-to-one cost disadvantage versus buying ACCUs at current spot prices for most facilities.

Multi-year monitoring periods (MYMPs) let you spread compliance across two to five years, as long as your cumulative emissions stay below cumulative baselines over the full period. Six facilities used this in 2023-24. It's useful if you have a major abatement project coming online in year two or three — say, an electrification of process heat — but you can't show reductions yet. The hard stop: MYMPs can't extend past 30 June 2030.

Trade-exposed baseline-adjusted (TEBA) determinations give eligible manufacturing facilities a slower decline rate — as low as 1% per year instead of 4.9%. Seventeen facilities used this in year one. If you're producing aluminium or cement and competing with imports from jurisdictions without carbon pricing, this is a lifeline. But it requires an independent audit and a cost-impact demonstration.

Exemption declarations only apply when excess emissions resulted from natural disasters or criminal activity. Very narrow. Very rarely used.

The decision tree matters. In Carbonly, we don't automate the choice — these are strategic decisions that need board-level input. What we do is show you the financial comparison: "surrender ACCUs now at $36/tonne, borrow at 10% interest, or apply for an MYMP and defer the obligation." That comparison only works if your emissions data is current, not nine months stale.

Coal mining: the method transition wildcard

One sector-specific wrinkle deserves its own discussion, because it affects the single largest source of Safeguard exceedance. Coal mining fugitive emissions measurement is undergoing a forced method transition.

Under Method 1, facilities used a default gas content factor of 1.21 m3/t of coal. The problem: actual observed gas content in the Bowen Basin runs 5-12 m3/t — four to ten times higher. The result was systematic undercounting at some mines and overcounting at others.

The outcomes were dramatic. Caval Ridge's reported emissions dropped 90% after switching from Method 1 to Method 2. Hail Creek's went up 3.5 times — they'd been massively undercounted. As of July 2025, mines producing more than 10 million tonnes ROM must use Method 2 or higher. The rest must switch by July 2026.

For mining companies tracking Safeguard compliance, this creates a forecasting nightmare. Your baseline was set using one measurement method. Your actual emissions might be reported using a different one. If your true emissions are significantly higher than Method 1 suggested, you could go from "comfortably under baseline" to "deeply in exceedance" in a single reporting year — not because anything changed operationally, but because the measurement got more accurate.

This is exactly the kind of scenario where monthly tracking against baseline catches problems before they become $5 million ACCU bills.

The 2026-27 review: planning for uncertainty

Everything I've described above operates under the current rules. Those rules are about to be reviewed.

The 2026-27 Safeguard Mechanism review — scheduled now that two years of post-reform data are available — will examine several foundational questions:

Whether the 4.9% annual decline rate is sufficient to meet Australia's 62-70% reduction target by 2035
How cost containment should work post-2030 (the current $82.68 ceiling is already under scrutiny)
Whether EITE arrangements remain appropriate as international carbon border mechanisms take shape
Whether SMCs should be valid beyond 2030
Whether the scheme's scope should expand (the Productivity Commission has estimated that extending coverage could deliver over $900 million in national benefits by 2035)

We don't know the outcomes. Nobody does. But we do know that building a reduction plan around the current 4.9% trajectory is the minimum — and you should be modelling scenarios where it accelerates.

In Carbonly's Safeguard module, the year-by-year baseline trajectory runs forward to 2030 at 4.9%, then defaults to 3.285% per year after that. But you can adjust the decline rate to model what-if scenarios. What if the review sets a 6% decline? What if it introduces a hard cap instead of intensity-adjusted baselines? Those scenarios produce different ACCU demand forecasts and different cost profiles. Having the data infrastructure to run them in hours instead of weeks is worth something.

What we're honest about

We'd be lying if we said this module solves every compliance problem. It doesn't. Here's what remains genuinely hard.

NGER data lags. The Safeguard Mechanism runs on NGER-reported Scope 1 emissions, and NGER reporting is annual with an October deadline. Your final verified emissions number won't exist until months after the financial year ends. Our monthly tracking uses the best available data — fuel records, production logs, process measurements — but it's an estimate until the NGER report is lodged. The gap between our monthly estimate and the final NGER figure depends entirely on your data quality.

Baseline adjustments for production changes are complex. Production-adjusted baselines move with your output. If you ramp up production mid-year, your baseline rises — but by how much depends on your calculated emissions intensity and the CER's assessment. We model the trajectory using the published emissions-intensity values, but the CER's final determination might differ.

ACCU price forecasting is inherently uncertain. We show you cost exposure at current spot prices and let you model at different price assumptions. We don't predict where prices will be in six months. EY projects $30-35 through 2028. RepuTex has a slightly different view. The honest answer is that nobody knows, and anyone selling you price certainty is selling you something else.

The 2026-27 review could change the rules. If the decline rate changes, if scope expands, if cost containment is restructured — all bets are off on forward modelling. We'll update the module when the rules are finalised. Until then, we model the rules as they exist today.

The $330-per-tonne penalty you want to avoid

One number concentrates the mind. Under the Safeguard Mechanism, the civil penalty for excess emissions that aren't managed by the compliance deadline is $330 per tonne (one penalty unit, currently $330 as of November 2024) plus $33,000 per day the facility remains in excess.

A facility 10,000 tonnes over baseline faces $3.3 million in penalties plus daily accruals. Compare that to buying 10,000 ACCUs at $36.50 — $365,000. The economics of proactive compliance aren't subtle.

In the first two years of the reformed scheme, no facility has actually been hit with the maximum penalty. The CER has taken a graduated approach — enforceable undertakings, like the one Beach Energy accepted in July 2025, are the preferred tool. But the regulator has been clear that patience has limits. The "training wheels" period, as the Carbon Market Institute called it, won't last forever.

Start tracking monthly. Not annually

If your facility emits more than 100,000 tonnes CO2-e and you're still reconciling your Safeguard position once a year when the NGER report is due, you're flying blind into a compliance environment that gets 4.9% harder every July.

The maths: by FY2029-30, your baseline will be roughly 28% lower than FY2023-24. Your production probably won't be 28% lower. Something has to give — either your emissions intensity drops through genuine abatement, or you're buying credits. The question is whether you buy them strategically with six months of lead time, or desperately with six weeks.

Get your Scope 1 data flowing monthly. Map it against your declining baseline. Know your position before the market does.

Related Reading: