Climate Scenario Analysis for AASB S2: A Practical Guide

The most expensive line item in every Group 1 ASRS report we've seen wasn't emissions data. It was scenario analysis. Companies spent $150,000 to $300,000 on consultant-built scenario models that auditors picked apart in a week. Oxford Economics flagged this exact problem before the first reports even landed — generic global scenarios don't survive assurance. And they were right.

Here's the thing. Climate scenario analysis for AASB S2 sounds like it requires a PhD and a dedicated modelling team. It doesn't. What it requires is two well-chosen scenarios, localised to your actual operations, connected to specific financial line items, with documented assumptions your assurance provider can trace. That's the bar. Most Group 1 entities didn't clear it because they outsourced the thinking instead of just the technical work.

If you're a Group 2 entity starting your reporting period on 1 July 2026 — or a Group 3 company getting ready for 2027 — this guide covers exactly what scenario analysis requires, what it costs, and how to get it right the first time. We've also written about the broader Group 2 reporting requirements if you need the full picture.

What the law actually requires (it's less than you think)

Start here, because the gap between what people think scenario analysis requires and what it actually requires is enormous.

AASB S2 paragraph 22 says an entity must use climate-related scenario analysis to assess its climate resilience, "using an approach that is commensurate with its circumstances." That last phrase matters a lot. It's the proportionality principle — AASB published a whole paper on it in September 2025 — and it means a mid-market property company isn't held to the same standard as BHP.

The Corporations Act 2001 (section 1707E) adds a specific Australian requirement that goes beyond the ISSB standard. You need at least two scenarios referencing temperature increases from the Climate Change Act 2022. Specifically: one scenario aligned with a 1.5°C increase above pre-industrial levels, and one aligned with a temperature increase well exceeding 2°C — the ASIC guidance says 2.5°C or higher.

That's it. Two scenarios. One low-warming, one high-warming. You're not required to run five. You're not required to use quantitative modelling if you genuinely don't have the skills and resources (though you need to justify that). You're not required to build a proprietary climate model.

What you are required to do is show how your financial position, performance, and cash flows would shift under each scenario across short, medium, and long-term time horizons. And your assumptions need to be documented well enough that an assurance provider can follow the logic from global pathway to your specific P&L impacts.

Pick your scenarios: IEA and NGFS, not bespoke models

The biggest waste of money in Group 1 was companies paying consultants to build bespoke scenario narratives from scratch. You don't need to invent scenarios. You need to pick established ones and localise them.

For your 1.5°C scenario, use the IEA Net Zero by 2050 (NZE) pathway. It's the most widely recognised 1.5°C-aligned scenario globally, it's updated regularly through the World Energy Outlook, and your auditor will recognise it immediately. The NGFS Net Zero 2050 scenario works equally well — it's designed by central banks, uses carbon prices of roughly US$200 per tonne by 2030 in advanced economies, and maps directly to financial system impacts.

For your high-warming scenario, use the NGFS Current Policies pathway. It models what happens if governments stick to existing policies with no additional ambition. Global warming reaches roughly 3°C by 2100. Transition risk is minimal (because nobody's transitioning), but physical risk is severe. Under this scenario, the NGFS assumes effectively zero carbon pricing — which makes it a useful stress test for companies that think their only climate exposure is regulatory.

Why these two and not others? Three reasons. First, they're from authoritative bodies that auditors and ASIC will recognise. Second, they have publicly available data you can download and reference. Third, they sit at opposite ends of the spectrum — maximum transition risk versus maximum physical risk — which is what scenario analysis is designed to test.

We've heard from sustainability managers who feel pressure to run a third "middle of the road" scenario. You can if you want. But two is the legal minimum, and two well-executed scenarios will always beat three poorly done ones.

Localising to Australian conditions (this is where Group 1 failed)

This is the part that separates scenario analysis that survives assurance from the kind that doesn't. And it's where Group 1 reporters got caught out.

A global IEA or NGFS scenario tells you what happens to world GDP or global energy demand under a 1.5°C pathway. It doesn't tell you what happens to your revenue from coal haulage contracts in the Hunter Valley when the NSW government accelerates mine closures under a net-zero pathway. It doesn't tell you what a 3°C physical risk scenario means for your property portfolio in North Queensland, where cyclone intensity projections are materially different from the global average.

Oxford Economics identified this as a core pitfall: "aggregated GDP impacts overlook industry and region-specific risks." That's exactly what happened. Group 1 entities used the IEA NZE scenario, cited the global GDP impact, and wrote two pages of narrative around it. Then the assurance provider asked a simple question: how does this global pathway map to your specific financial statements? And nobody could answer.

To localise your scenarios, you need three layers of Australian-specific data.

Carbon pricing assumptions. Don't just use the NGFS global carbon price. Australia has its own carbon market. ACCUs traded at around $35-36 per tonne through mid-2025, with the Safeguard Mechanism cost containment measure capped at $82.68 for 2025-26. EY's January 2026 analysis projects prices staying flat at $30-35/tonne until 2028, then rising gradually to around $70/tonne by 2035. Under a 1.5°C scenario, you'd assume policy tightening pushes these significantly higher — perhaps $100-150/tonne by 2035. Under Current Policies, prices stay where they are. Your scenario analysis needs to use these Australian figures, not the NGFS US$200/tonne global estimate.

Physical risk data. Australia's first National Climate Risk Assessment (NCRA), released in September 2025, gives you the data you need. At 1.5°C warming, total annual economic costs from flood, bushfire, storm, cyclone, and hailstorm are projected at approximately $40 billion nationally by 2050. Heat exposure alone could reduce economic output by $135-423 billion by 2063. Sea level rise puts 1.5 million additional people in high-risk coastal areas by 2050. Use these numbers — and their regional breakdowns — for your high-warming physical risk assumptions. They're government-published, peer-reviewed, and exactly what an assurance provider will expect to see referenced.

Energy transition trajectory. Australia's grid emission factors are already falling. The NGA Factors 2025 show South Australia at 0.22 kg CO2-e/kWh and Tasmania at 0.20 — compared to Victoria at 0.78. Under a 1.5°C scenario, you'd model accelerated decarbonisation of the grid, which could actually reduce your Scope 2 costs. Under Current Policies, grid decarbonisation slows. If your company operates across multiple states, these differences matter — a lot. We wrote about how emission factors vary by state and why getting the right factor matters.

What financial impacts to model

ERM's guidance on financial quantification under AASB S2 is the most practical thing published on this topic. Their core point: you need to connect scenario assumptions to specific line items in your financial statements, not just produce a standalone "climate impact" number that floats in a separate report.

Here's what that looks like in practice. For each scenario, work through four categories.

Revenue impacts. Under the 1.5°C scenario, how does demand for your products or services change? If you're in construction, does a carbon price of $100/tonne shift the cost profile of concrete and steel enough to change procurement patterns? If you're in property, does mandatory energy efficiency regulation reduce the rental premium gap between green-rated and unrated buildings? Model these as percentage shifts against your current revenue baseline, across short (to 2030), medium (to 2035), and long-term (to 2050) horizons.

Cost impacts. Insurance premiums are the easiest place to start for physical risk. If you hold assets in flood-prone or cyclone-exposed regions, your insurer already has this data — ask them for forward projections under different warming scenarios. Operating costs shift too: energy costs under a 1.5°C scenario with higher carbon prices, maintenance costs under a high-warming scenario with more extreme weather events. Don't model these as precise numbers. Express them as ranges. ERM's advice on this is worth quoting directly: financial quantification "should be transparent, supported by sensitivity and uncertainty analysis, and expressed as ranges rather than with false precision."

Asset impacts. This is where CFOs pay attention. Under a transition scenario, assets tied to fossil fuel infrastructure face potential impairment. Under a physical risk scenario, coastal or flood-exposed property faces value decline. The question isn't whether the impairment happens — it's at what discount rate and over what timeframe. Work with your finance team to model these against your existing impairment testing methodology. Don't build a separate model. Embed climate assumptions into your existing asset valuation process.

Capital allocation. What additional investment does each scenario require? Under 1.5°C, you might need to accelerate building retrofits, electrify vehicle fleets, or invest in renewable energy procurement. Under Current Policies, you might need to spend more on physical resilience — flood barriers, cyclone-rated construction, backup cooling systems. Put rough numbers on both.

We're not going to pretend this is easy. Quantifying the revenue impact of a carbon price that doesn't exist yet, on a product mix that might change, over a twenty-year horizon — that's inherently uncertain. But AASB S2 doesn't ask you to predict the future. It asks you to show you've thought about it, with documented assumptions, and that your board understands the implications.

Time horizons: define them and stick to them

AASB S2 requires short, medium, and long-term time horizons but doesn't prescribe specific years. You define them, and then you have to disclose why you chose those periods.

For most Australian businesses, a workable approach is: short-term to 2030 (aligned with Australia's 43% reduction target), medium-term to 2035 (when carbon prices are projected to start rising materially and Australia's 2035 NDC kicks in), and long-term to 2050 (aligned with the net-zero target and the endpoint of the IEA NZE scenario).

One thing we've noticed: companies often make the short-term horizon too short. If your "short-term" is two years, you're basically describing the current environment with minor adjustments. That's not scenario analysis. Push it out to at least 2030 so you're actually modelling policy and market changes.

The long-term horizon is where it gets uncomfortable. Nobody can predict what their business looks like in 2050 with any confidence. And that's fine. Your long-term analysis should be more directional — order of magnitude impacts, not precise numbers. Auditors understand this. The AASB standard itself says that "as the time horizon increases and the availability of detailed information decreases," entities should adjust their approach accordingly. Just be transparent about the increasing uncertainty.

Documenting methodology for assurance (the part everyone skips)

Here's what killed Group 1 entities: they produced a nice PDF with scenario narratives and summary tables, but there was nothing underneath. When the assurance provider lifted the bonnet, there was no model to examine.

Your documentation needs to cover five things. Write this up as a methodology note — ten to twenty pages, not a hundred — and keep it alongside your scenario model.

First, explain why you chose the scenarios you chose. "We selected the IEA NZE 2050 pathway for our 1.5°C scenario because it is the most widely referenced 1.5°C-aligned pathway and is recognised by the NGFS, TCFD, and ASRS guidance" is fine. You don't need three pages of literature review.

Second, document every assumption you localised. What carbon price trajectory did you use for Australia, and where did you get it? What physical risk data came from the NCRA? What insurance premium projections came from your insurer? List them. Cite them.

Third, show the bridge from global pathway to your P&L. This is the single most important document for assurance. The auditor needs to see: "Under the IEA NZE scenario, the carbon price in Australia reaches $X by 2030. Our Safeguard-covered facility emits Y tonnes above baseline. Cost impact: $X × Y = $Z per annum." That chain of logic. Not a narrative. A calculation.

Fourth, document your sensitivity analysis. What happens if the carbon price is 20% higher or lower than your central estimate? What if physical risk costs arrive five years earlier than your baseline assumption? Show a couple of sensitivities. It demonstrates rigour and gives the auditor confidence that you understand your own model.

Fifth, record who did the work and when. Date-stamp your analysis. Note whether it was done internally or with external support. If you used a consultant, record what they did and what your internal team owns. The assurance provider needs to know there's someone inside the business who can explain and defend every number — not just refer them back to a consultant who may have moved firms.

Budget and timeline: what this actually costs

If you outsource the entire scenario analysis to a consulting firm — from scenario selection through financial modelling to documentation — expect to pay $50,000 to $100,000 for a mid-market company. That range assumes two scenarios, three time horizons, and a documented methodology. For companies with complex operations across multiple sectors or geographies, it goes higher. We've seen quotes above $200,000 for large diversified entities.

The smarter approach is to do significant work internally and bring in consultants for specific gaps. Hire a consultant to run a two-day workshop with your finance and sustainability teams to design the scenario framework and set up the model structure — that's $10,000 to $20,000. Then your team fills in the data, runs the sensitivities, and writes the methodology note. Bring the consultant back for a half-day review before you submit to assurance — another $5,000 to $8,000. Total: $15,000 to $28,000. And your team actually understands the model, which matters a lot more in year two when nobody's starting from scratch.

Timeline? Allow twelve to sixteen weeks from kick-off to assurance-ready documentation. The first four weeks are scenario selection, assumption setting, and data gathering. Weeks five through ten are financial modelling and sensitivity analysis. Weeks eleven through fourteen are documentation and internal review. Weeks fifteen and sixteen are for pre-assurance dry runs with your auditor — which, as we've argued in our ASRS Group 2 preparation guide, you should absolutely do.

Don't leave this until three months before your reporting deadline. The companies that struggled in Group 1 all had one thing in common: they started scenario analysis after their emissions data was "ready" — treating it as a late-stage drafting exercise rather than a core analytical workstream that needs to run in parallel with data collection.

The proportionality escape valve (use it honestly)

AASB S2 paragraph 22 includes a proportionality mechanism that many companies don't fully understand. It says your approach to scenario analysis should be "commensurate with the entity's circumstances" — considering your exposure to climate risks and the skills, capabilities, and resources available to you.

This means if you're a mid-market services company with limited physical asset exposure and no Safeguard Mechanism obligations, a largely qualitative scenario analysis — structured narratives with directional financial impacts rather than detailed quantitative modelling — may be acceptable. Not preferable. But acceptable.

The catch? You need to justify why your circumstances warrant a simpler approach. And you need to show a roadmap for building capability over time. Auditors expect to see progression. A qualitative approach in year one that becomes semi-quantitative in year two and quantitative by year three is defensible. A qualitative approach that never improves isn't — especially as assurance escalates from limited to reasonable.

But we'd be dishonest if we didn't say this: we're still not entirely sure where the line falls between "acceptable proportionality" and "insufficient effort" for mid-market companies. Nobody is. The guidance exists, but the case law doesn't. Early assurance outcomes from Group 2 will set the precedent. If you're going to rely on proportionality, document your reasoning carefully, get your assurance provider's input before you commit to the approach, and budget for upgrading next year.

What happens when you get it wrong

Scenario analysis sits within the modified liability window that runs until 31 December 2027. During this period, only ASIC can take action on protected disclosures — and even then, only for injunctions and declarations, not penalties. That's real protection. For now.

But there are two things the protection doesn't cover. First, it doesn't prevent investors from drawing conclusions about a company whose scenario analysis is visibly thin compared to peers. Institutional investors — particularly superannuation funds with their own ASRS obligations — are already comparing disclosures across portfolios. A generic, two-page scenario section signals that you haven't thought seriously about climate risk. That has capital allocation consequences.

Second, the liability window expires. From 1 January 2028, scenario analysis disclosures carry the same liability as any other part of your sustainability report. And by then, reasonable assurance is on the horizon (financial years beginning 1 July 2030). The scenario model you build now is the one you'll need to defend at that higher assurance standard. Build it properly from the start.

What we've seen from Group 1 is that the assurance conversation is where bad scenario analysis costs the most. Not in penalties — not yet. In audit fees. When your assurance provider can't follow the logic from global scenario to financial impact, they spend more time asking questions, requesting documentation, and running their own checks. That time gets billed. We've heard of mid-market assurance engagements running $20,000 to $30,000 over budget specifically because scenario analysis documentation was inadequate.

The one-page version

You need two scenarios: one at 1.5°C (use IEA NZE or NGFS Net Zero 2050), one at 2.5°C or above (use NGFS Current Policies). Localise both with Australian carbon price trajectories, NCRA physical risk data, and state-level grid factors. Connect each to your actual revenue, costs, assets, and capital needs across three time horizons. Document your assumptions in a methodology note. Budget $15,000-$28,000 if you do most of the work internally, $50,000-$100,000 if you outsource it. Start at least twelve weeks before you need assurance-ready documentation.

Build the model so your team can rerun it next year without starting over. That's the lesson every Group 1 reporter learned the hard way. Don't repeat it.

Related reading: