Six first-wave IFRS S2 disclosures: what they skipped

TL;DR

  • Five of six deeply-read first-wave IFRS S2 reporters do not disclose a Para 29(c) amount; only Rio Tinto goes meaningfully beyond narrative, with annualised damage thresholds.
  • Where Para 29(c) is quantified, no two reporters use the same unit: sector concentration, AD percentage, hazard-action count, deferred-to-appendix.
  • Para 22 disclosure splits sharply: IAG and Bank of America name three or more scenarios with horizons; CLP shows “Nil” for six of seven sub-elements.
  • Insurance: actuarial PML capacity exists internally for treaty negotiation; the public Para 29(c) disclosure does not follow. This is the strongest empirical lever for year-two pressure.
IFRS S2 first-wave disclosures: 4-tier classification ladder mapping six reporters across OMITTED, QUALITATIVE-ONLY, PARTIAL, QUANTIFIED tiers
Six first-wave IFRS S2 reporters mapped onto the four-tier Para 29(c) classification ladder. Source: Continuuiti.

What we read and what the pattern looks like

We pulled every public IFRS S2, AASB S2, HKFRS S2, and SGX-aligned climate disclosure we could find as a full PDF. Eight reporters across six sectors and six jurisdictions. Six of the eight we read deeply, with verbatim Para 22 and Para 29(c) extraction against page references; these six are the analytical spine of this piece. The cohort table below also names Sembcorp Industries (alignment-claim verified at page 51 of its 2024 Annual Report; the full Para 29(c) extraction lives outside the document we hold) and Wesfarmers (agent-extracted partial covering the methodology language but not the full verbatim cell quotes).

A practitioner reading the four-question interpretive walkthrough of Para 29(c) is left with a clean question: how did the first wave actually answer those four questions? This piece is the empirical answer for the six we read end-to-end. After verifying each Para 29(c) treatment against the source PDF with page references, the pattern is what NZ FMA, KPMG FAST 30, PwC, and Deloitte have all flagged independently in their own cohort reviews. Para 29(c) is the universal weak point. Only one of the six discloses a Para 29(c)-relevant quantification beyond narrative, and even that one stops at risk categorisation rather than a specific amount-and-percentage figure. This piece names which reporter sits where, with verbatim language and exact page citations.

Methodology

Eight reporters across six sectors (banking, insurance, mining and metals, electric utilities, energy and utilities, industrial and logistics) and six jurisdictions (US, SG, AU, ZA, IN, HK). Six of the eight are deeply read with verbatim Para 22 and Para 29(c) extraction: Bank of America, Standard Bank, IAG, Rio Tinto, Tata Steel, and CLP (via the 2023 IFRS S2 Content Index). The pile is AU-skewed (six of eight are Australian under mandatory or voluntary AASB S2); EU CSRD reporters disclosing under ESRS E1 are a different cohort and not in this analysis. Real estate, REITs, oil and gas, healthcare, telecoms, technology, and asset management are absent from our pile.

Three alignment classifications matter for reading the cohort: Full alignment at the consolidated reporting level (CLP 2024 Annual Report, per CLP’s own claim, page 14 of 2024 Sustainability Report); voluntary “with reference to” or “informed by” with explicit transitional carve-out (Bank of America, IAG, Rio Tinto); TCFD-aligned with optional IFRS S2 reference (Sembcorp Industries explicit TCFD-only at page 51, Standard Bank TCFD-organised, Tata Steel claims aligned with both IFRS S2 and TNFD via the four-pillar structure).

The four published cohort reviews (NZ FMA / KPMG FAST 30 / PwC / Deloitte) are the external anchors. Verbatim discipline: every Para 29(c) classification below is anchored in a verbatim quote with a specific PDF page reference. Where a reporter does not disclose Para 29(c) at all, the absence is itself recorded.

The cohort on one page

Cohort overview, all eleven reporters. Read this table and the rest of the piece is the long-form unpacking of the six deeply-read rows.

Reporter (jurisdiction) Alignment claim Para 22 scenarios Para 29(c) classification Para 29(c) unit (if disclosed) Proportionality invocation
Bank of America (US, voluntary) “with reference to” ISSB 3 NGFS Phase IV scenarios + 1 acute physical (pages 16-17) QUALITATIVE-ONLY Sector concentration table at page 14 (mixes physical and transition; sourced from 10-K) EXPLICIT (page 11)
Standard Bank (ZA, voluntary) TCFD-organised; no IFRS S2 measurement-uncertainty language located 2 NGFS scenarios (Net Zero 2050 + Current Policies, page 14) QUALITATIVE-ONLY (sector exposure quantified, physical vulnerability not segmented) Real estate book R 512.3bn on-balance (page 27); no flood/heat segmentation NOT located
IAG (AU, voluntary FY25 transitional) “with reference to” AASB S2; “transitionary only” 3 NGFS Phase IV scenarios (Orderly / Disorderly / Hot House) with three time horizons defined (pages 41-42) QUALITATIVE-ONLY in AR proper CAP Target 6: 1.7m natural-hazard risk reduction actions (a target performance count, not a Para 29(c) figure, page 44) EXPLICIT (page 42)
Rio Tinto (AU/UK, mandatory AU first wave) TCFD-primary with IFRS S2 reference (page 41, page 75 TCFD Index) 3 macroeconomic + 2 physical-risk scenarios (SSP2-4.5 + SSP5-8.5), 3 time horizons (pages 43, 68) PARTIAL Annualised damage % thresholds (low <0.2%, medium 0.2-1%, high >1%); 9-region heat map at page 69; no specific carrying-amount-vulnerable figure SOFT (page 44)
Tata Steel (IN, voluntary; 6-pp summary in pile) Aligned “with IFRS S2 (formerly TCFD) and TNFD recommendations” via four-pillar structure (page 1) NOT NAMED in the 6-pp summary; references “TCFD-aligned independent third-party Climate Risk assessment” on three steelmaking sites (page 4) OMITTED in the 6-pp summary n/a; fuller content in Integrated AR (not in pile) NOT located in summary
CLP (HK, voluntary IFRS S2 early adopter; 2023 Content Index) “based on” ISSB IFRS S2 (Content Index page 1) 6 of 7 Para 22(b)(i) sub-elements show explicit “Nil” in 2023 AR/SR column (Content Index page 6) OMITTED in 2023 AR; Content Index page 9 explicit “Nil” for Para 29(b)-(e) n/a; Content Index points to separate Climate Vision 2050 appendix SOFT, repeated commitment-language pattern across multiple “Remarks” rows
Sembcorp Industries (SG) TCFD-only (page 51) Referenced (IPCC + NGFS) but verbatim and scenario count not extracted in our read TBD (alignment claim is TCFD-only; Para 29(c) extraction deferred) n/a NOT located in pages read
Wesfarmers (AU, voluntary FY25; agent-paraphrased) TCFD-aligned in FY25; first mandatory FY26 1.5°C / current policies / delayed transition; horizons 2030 + 2050 (page 62) likely PARTIAL per agent (carrying amounts not isolated for climate-vulnerable assets in the climate section) Store and facility location mapping referenced; counts not disclosed NOT detected in agent extraction

The six deeply-read reporters split as follows on Para 29(c): one PARTIAL (Rio Tinto); three QUALITATIVE-ONLY (Bank of America, Standard Bank, IAG); two OMITTED (Tata Steel six-page summary, CLP 2023 Content Index). Wesfarmers per the agent extraction is provisionally PARTIAL on physical-risk methodology, though the unit and verbatim quote are not captured. Sembcorp Industries carries an alignment-claim only at the document level we hold.

Finding 1: Five of six do not disclose a Para 29(c) amount

The dominant pattern in the deeply-read cohort is non-quantification of the Para 29(c) metric. Five of the six sit in the OMITTED or QUALITATIVE-ONLY tier. Only Rio Tinto goes meaningfully beyond narrative, and even Rio Tinto’s disclosure stops at risk categorisation rather than naming a single carrying-amount-vulnerable figure.

The most explicit non-disclosure in the cohort comes from CLP Group. CLP publishes a separate 2023 IFRS S2 Content Index (12 pages) that maps every IFRS S2 paragraph to where the disclosure lives in CLP’s other corporate reports. For Para 29(b), 29(c), 29(d), and 29(e) on page 9 of the index, the column for the 2023 Annual Report and Sustainability Report shows “Nil.” The disclosure points instead to “Appendix – Our scenario analysis and exposure to climate-related risks and opportunities” in the separately-published Climate Vision 2050 (2024 edition) document. CLP is the only reporter we have read that has self-classified its Para 29(c) coverage paragraph by paragraph, and its own classification places Para 29(c) in the deferred-to-appendix column.

The second-most-explicit non-disclosure comes from Bank of America. Page 11 of the ISSB IFRS S2 Disclosure 2025 invokes the proportionality relief provided under IFRS S2 to disclaim quantitative measurement: “Based on current capabilities and resources, we have determined the level of measurement uncertainty involved in estimating quantitative information about the financial effects and priority of identified climate-related risks is so high that it would not be useful… This approach is consistent with the relief provided under IFRS S2…” (Bank of America ISSB IFRS S2 Disclosure 2025, page 11). Pages 12-15 then describe physical risks qualitatively. Page 14 shows a sector concentration table for commercial credit ($1,286.6 billion total committed exposure as of December 2024) sourced from Bank of America’s 2024 Form 10-K. The table does not separate physical from transition risk and is not framed as a Para 29(c) compliance disclosure.

IAG sits in the QUALITATIVE-ONLY tier, which is more surprising than the other QUALITATIVE-ONLY reporters because IAG has the deepest physical-risk modelling capability in the cohort (an actuarial peril-modelling history, a third-edition Severe Weather in a Changing Climate report co-developed with the US National Center for Atmospheric Research, an internal Climate Risk and Opportunity process described as “both qualitative and quantitative in nature” on page 47). The page 38 Material climate-related risks table identifies two physical risks (increasing natural perils costs; government intervention in insurance markets) with qualitative “Potential effects on business model and value chain” bullets and “Current mitigations in place” bullets. There is no disclosed amount or percentage of assets or business activities vulnerable to physical risks at the asset level. Page 38 also discloses on a residual basis: “Our assessment is undertaken on a residual basis after considering the current mitigations already in place that are assessed to remain effective.” The Climate Action Plan target performance table on page 44 reports 1.7m natural-hazard risk reduction actions against a 1.0m FY25 interim target. That figure is a count of customer adaptation actions taken against a CAP transition-plan target, not a Para 29(c) amount of assets vulnerable.

Tata Steel’s six-page Climate Change Report extract addresses physical risk qualitatively at page 4: “The physical climate risk is an emerging risk for the Company and mitigation strategies are being developed as per the ERM framework. Climate and nature-related risks assessed by Tata Steel span across operational, financial, reputational, regulatory, and legal risks.” Page 6 directs the reader to the fuller Integrated Annual Report 2024-25 for detail. The Para 29(c) disclosure in our pile is OMITTED at the document level we hold; the IAR content lives outside the pile.

Standard Bank’s Climate-Related Financial Disclosures Report 2024 sits in the QUALITATIVE-ONLY tier despite quantifying the underlying portfolio exposure carefully. Page 27 reports residential real estate exposure of R 512.3 billion on-balance plus R 41.6 billion off-balance. Pages 17-25 disaggregate the banking book to R 2,147.3 billion total (2024) with climate-sensitive sector breakdowns (real estate 29.75%, oil and gas 5.33%, agriculture 4.19%, non-renewable power 0.35%, thermal coal 0.04%). The physical-risk vulnerability of those sector exposures is not separately quantified. Page 18 names physical risks for real estate verbatim: “Identified physical risks: extreme weather events such as hurricanes, floods, wildfires, and in the longer-term, rising sea levels in coastal areas.” The reader gets sector concentration and physical-risk identification; the Para 29(c) join is not made.

The five non-quantifying reporters span four sectors (banking, insurance, mining, utilities) and five jurisdictions (US, ZA, AU, IN, HK). The non-quantification is sector-agnostic and jurisdiction-agnostic in our read. This is consistent with the cohort reviews. KPMG’s FAST 30 commentary on Australia’s first wave reports that financial-quantification of climate impacts splits roughly two-thirds quantified to one-third did not. PwC’s review of 22 first-wave Australian Group 1 reporters (https://www.pwc.com.au/assurance/sustainability-reporting-and-assurance/aasb-s2-unpacked.html) reports the same split with different framing. Deloitte’s Wave 1 review (https://www.deloitte.com/au/en/services/audit-assurance/analysis/early-insights-wave-1-australian-climate-reporting.html) records reports averaged 33 pages in length (range 7-82) with high variability in physical-risk quantification depth. The NZ FMA / SLR Consulting review of New Zealand’s first mandatory CRD statements (https://www.slrconsulting.com/us/insights/what-companies-can-learn-from-the-first-statements-under-new-zealand-s-mandatory-climate-related/) flagged the amount and percentage of assets vulnerable to physical and transition risks as the systematic gap in the NZ first wave. Four independent cohort reviews, one shared finding. Our deeply-read cohort is consistent with all four.

Finding 2: Where Para 29(c) is quantified, no two reporters use the same unit

Of the six deeply-read reporters, only Rio Tinto produces a Para 29(c)-relevant quantification beyond narrative. Wesfarmers per the agent extraction has location-based physical-risk assessments mapping store and facility exposure to flooding, heatwave, and cyclone hazards (Wesfarmers 2025 Annual Report pages 63-64), but the agent did not capture a specific carrying-amount or asset-count figure verbatim.

Rio Tinto’s 2025 Climate Action Plan (extracted from pages 41-75 of the 2024 Annual Report) carries the most sophisticated physical-risk methodology in our read. Page 68 sets out the annualised damage (AD) framework: “Annualised damage (AD): The output of the modelling is calculated for each asset under various climate scenarios, time horizons and hazards. AD, expressed as a percentage, represents the expected average annual damage to an asset attributable to climate-related hazards relative to a fixed value (eg $1 million). As such, an AD of 0.5% would mean that for every $1 million of exposure, $5,000 could be damaged, on average, in any given year.” Risk categorisation thresholds are disclosed verbatim at page 68: low AD risk <0.2%, medium 0.2-1%, high >1%. The methodology covers eight climate hazards (riverine and surface water flooding, coastal inundation including sea level rise, extreme heat, cyclonic wind, extreme wind, forest fire, freeze-thaw) across two emissions scenarios (SSP2-4.5 and SSP5-8.5) and three time horizons (2030 medium term, 2040 and 2050 long term).

The Group-level result at page 68 is specific: “At the Group level, present day AD losses fall within the initial range of the medium AD risk category (0.2-1%). Considering projected future emission scenarios by 2050, we expect increases in AD. This places the Group’s AD in the intermediate range of the medium AD risk category, potentially exceeding a two-fold rise from present values.” Page 69 provides a heat-map table showing the AD risk class for the Rio Tinto Group plus nine regions across present day plus three future time horizons under both SSP scenarios. Page 68 also discloses the gross-of-adaptation framing: “Estimates consider a stationary ‘do nothing’ approach for our operating assets and do not consider present or future controls or adaptation or resilience projects that will likely materially impact our AD cost.”

This is the most rigorous physical-risk methodology disclosure in our read. Even so, the disclosure stops short of the literal Para 29(c) form. The reader knows that the Group’s AD is in the medium category (0.2-1%) and is expected to rise, that nine regions split low-medium-high, and that two regions (eastern Australia and New Zealand under both future scenarios by 2050) become high-AD-risk with a four-fold projected increase. The reader does not get a specific dollar figure or specific percentage of total Group property, plant and equipment classified as vulnerable. Rio Tinto’s disclosure is closer to “what is our Group risk class and how does it shift across regions and scenarios” than to “what amount of assets is vulnerable.” Hence PARTIAL, not QUANTIFIED.

The unit-choice variance across the cohort is itself a finding. Bank of America’s commercial-credit sector concentration ($1,286.6 billion at December 2024, page 14) groups physical and transition risk together. Standard Bank’s banking-book sector breakdown (R 2,147.3 billion at December 2024, page 25) is sector-concentration only with physical-risk identification narrative attached. IAG’s 1.7m hazard reduction actions count is a Climate Action Plan target performance metric. Rio Tinto’s annualised damage percentage is a methodology metric tied to risk categorisation thresholds. CLP’s Content Index defers Para 29(c) entirely. No two of the six deeply-read reporters use the same unit, and none of them publishes a single number that would let an investor compare physical-risk vulnerability across two reporters on the same axis.

This is the unit-of-disclosure problem the interpretive companion to this piece walked through under its first definitional question (“amount of what. Carrying value at amortised cost? Gross book value? Fair value? Replacement cost? Exposure-at-default? Insured value? Notional?”). The cohort does not converge. The implication is straightforward. A practitioner preparing a first-wave or second-wave Para 29(c) disclosure cannot pattern-match to a single peer norm. The choice of unit is itself the disclosure, and the choice has to be defended on its own terms.

Finding 3: Para 22 is widely under-disclosed against its own plain text

Para 22(b)(i) of IFRS S2 has seven sub-elements asking the entity which scenarios it used, whether the analysis included a diverse range, whether scenarios are physical or transition, whether at least one scenario aligns with the latest international agreement on climate change, why the chosen scenarios are relevant, the time horizons used, and the scope of operations covered. CLP’s Content Index disclosure on page 6 of the 12-page index is the cleanest empirical test of how reporters do against this plain-text checklist. CLP’s 2023 Annual Report and Sustainability Report column shows “Nil” for sub-elements (1), (2), (4), (5), (6), and (7). Only sub-element (3), the physical-versus-transition classification, is mapped to the Annual Report Risk Management Report. The detail for the other six sub-elements lives in CLP’s Climate Vision 2050 (2024 edition) appendix. CLP’s own paragraph-by-paragraph index is the most useful empirical artefact in the cohort because it forces the reporter into a yes-or-no on each sub-element.

IAG sits at the opposite end. The 2025 Annual Report Sustainability Report at pages 41-43 provides the strongest scenario disclosure in our read. Three named scenarios (Orderly Net Zero 2050, Disorderly Delayed Transition, Hot House World Current Policies) with explicit warming targets, RCP values (1.9 / 2.6 / 7.0 with note that the Hot House RCP is increased from NGFS guidance from 4.5 to 7.0 to capture broader physical impacts), and three time horizons defined (Short Term 2025-2027 aligned to three-year financial planning; Medium Term 2028-2035; Long Term 2036-2050). The primary scenario source is named verbatim at page 42: “Network for Greening Financial Sector (NGFS) Climate Scenarios for Central Banks and Supervisors Phase IV.” Four sector-specific sub-sources are also named (Australian Energy Market Operator Integrated System Plan 2024; CSIRO Electric Vehicles Projection 2023; Insurance Council of New Zealand climate scenarios for the New Zealand general insurance sector 2022; Climate Change Commission scenarios from the 2021 Draft Advice for Consultation). Para 22(b)(i)(2) “diverse range” claim is well-supported. Para 22(b)(i)(4) Paris-aligned scenario claim is supported via Orderly. Sub-element (5) rationale is set out at page 42 verbatim “selected to provide a broad range of future outcomes against which to test potential impacts on our business model.” Sub-element (6) horizons and (7) scope are explicit.

Bank of America’s three-NGFS-Phase-IV-scenarios (Net Zero 2050, Nationally Determined Contributions, Delayed Transition) plus the additional acute physical-risk scenario “events impacting the U.S., Europe and Asia within a two-year period” (page 16) is the second-strongest disclosure. The IPCC scenario framework reference for acute physical risks at page 16 supports the diverse-range claim. Standard Bank’s two-NGFS-scenarios (Net Zero 2050, Current Policies, page 14) is on the floor of the diverse-range claim with two scenarios. Rio Tinto uses a different structure: three macroeconomic scenarios for transition resilience (Conviction, Resilience, Aspirational Leadership) plus two physical-risk-specific scenarios (SSP2-4.5 and SSP5-8.5) at pages 43 and 68. Rio Tinto satisfies the diverse-range claim for physical risk through the SSP pair and for transition through the three-scenario macroeconomic set.

Tata Steel’s six-page summary names no climate scenarios at all. Page 4 references “a TCFD-aligned independent third-party Climate Risk assessment focusing on its key steelmaking sites in India, the Netherlands, and the UK” but the scenario inputs are not surfaced in the summary. Para 22(b)(i)(1) (which scenarios) and (b)(i)(2) (diverse range) are not addressable from the document we hold. The fuller Integrated Annual Report may carry the detail; the summary does not.

The pattern across the deeply-read cohort is that Para 22 disclosure lives at two ends of a quality spectrum. At one end, IAG and Bank of America name three or more scenarios with explicit time horizons, scenario sources, and methodology rationale. At the other, CLP defers the detail to an appendix in a separate document and shows “Nil” for six of seven sub-elements in the consolidated Annual Report column, and Tata Steel’s summary does not surface scenario inputs at all. Standard Bank, with two NGFS scenarios, sits on the floor of compliance; the diverse-range claim is technically satisfied at two but the Big-4 cohort reviews flag two-scenario disclosures as the boundary of defensible practice. Rio Tinto’s structural choice of separating macroeconomic from physical scenarios is unusual in the cohort and is the kind of transparency the standard reads as good practice when the rationale is disclosed (which Rio Tinto does at page 43).

Finding 4: The proportionality clause is doing more work than it should

Para 30 of IFRS S2 carries the proportionality clause: “In preparing disclosures to meet the requirements in paragraph 29(b)–(d), an entity shall use all reasonable and supportable information that is available to the entity at the reporting date without undue cost or effort.” The same language repeats at Paras 11 and 18. Two of the six deeply-read reporters invoke the clause explicitly. Two more invoke it softly through commitment-language patterns. The pattern across the cohort is not that proportionality is over-used at the citation level (it is not) but that the equivalent disclaim-and-defer language appears in many forms.

Bank of America’s invocation is the most complete and the most-cited example in the broader cohort. Verbatim at page 11: “Based on current capabilities and resources, we have determined the level of measurement uncertainty involved in estimating quantitative information about the financial effects and priority of identified climate-related risks is so high that it would not be useful. Accordingly, this disclosure provides qualitative descriptions of the nature of these risks and the processes in place to monitor and manage them. This approach is consistent with the relief provided under IFRS S2 and helps drive disclosures that remain decision-useful while proportionate to the Corporation’s circumstances.” The clause does three things in one sentence. It cites the relief by name. It pre-emptively reframes “qualitative descriptions” as the chosen-and-defensible output. It anchors the choice to the Corporation’s circumstances rather than to the underlying capability or skill question. Each of those moves is reasonable in isolation. Together, they leave the reader with a Para 29(c) treatment that does not name an asset, a unit, or a percentage anywhere in the disclosure.

IAG’s invocation is cleaner and narrower. Page 42, in the section on identification and assessment of climate-related risks: “The identification and assessment of climate-related risks and opportunities that could reasonably be expected to affect our future prospects requires several aspects of judgment. These include judgement as to how the potential future macro-system, competitor and internal organisational factors may interact under each climate-related scenario, using reasonable and supportable information available without undue cost or effort – including information about past events, current conditions and forecasts.” The clause is cited as an input qualification rather than a deferral. The forward-looking-statements disclaimer on page 37 does the heavier work of disclaim-and-defer: “To the maximum extent permitted by law, IAG makes no representation, assurance or guarantee in connection with, and disclaims all responsibility for the accuracy, completeness or likelihood of fulfilment of any forward-looking statement…”

Rio Tinto’s invocation is softer and applies to a specific scope. Page 44 verbatim: “There are no portfolio adjustments made to the Group’s medium- to long-term plan under the various scenarios. Additionally, as our macroeconomic modelling involves a range of variables, isolating and measuring the impact of specific climate risks and opportunities is challenging. Therefore, the potential quantitative financial impacts are not disclosed.” The proportionality logic is implicit (measurement difficulty justifies non-disclosure). The scope is narrow (macroeconomic financial impacts of scenario variables; the bottom-up physical-risk AD analysis at pages 66-69 is disclosed).

CLP’s pattern is the most distinctive. The 2023 IFRS S2 Content Index uses a recurring “Remarks” formulation rather than a single Para 30 invocation. Page 2, against IFRS S2-9(d) (anticipated effects on financial position): “This year’s disclosures consider current effects only. CLP is reviewing methodologies and processes that would help to further enhance identification and measurement of relevant anticipated effects.” The same Remarks language appears verbatim against IFRS S2-15(b) on page 5 and substantively against 16(b) and 16(d). Page 7, against IFRS S2-25(a)(iv) and 25(b): “While CLP has prioritised climate-related risks as part of the materiality assessment process and the overall risk management process, detailed disclosures on how CLP has undertaken the prioritisation process are not published. CLP will consider how to further enhance this disclosure in future reports.” The pattern is consistent. CLP discloses what is currently disclosed, declines to disclose what is not, and commits to enhancement in future periods. This is the proportionality clause in commitment-language form.

The implication for the practitioner is that the proportionality clause is interpretable narrowly or broadly, and the cohort interprets it broadly enough that Para 29(c) frequently lands in the qualitative-only or commitment-to-enhance bucket. B17 of the IFRS S2 Appendix carries the constraint that the cohort reviews are flagging: entities with high exposure to climate-related risks AND the skills, capabilities or resources are “required to apply a more advanced quantitative approach to climate-related scenario analysis.” Our read does not produce evidence that any reporter is failing the B17 test (the test is hard to apply from the disclosure alone). It produces evidence that the language Para 30 makes available is being used at the boundary of the defensible interpretation. The Big-4 commentary across the four cohort reviews flags the same trajectory: tolerance is narrowing year over year.

Finding 5: Where the actuarial capacity exists, the disclosure still does not follow

Insurance is the sector where Para 29(c) should be most attainable. The Probable Maximum Loss triangulation methodology under IBG vol-17 FN-IN-450a.1 (Probable Maximum Loss of insured products from weather-related natural catastrophes) is the actuarial standard for catastrophe-modelled exposure. Insurers compute PML by return period, by hazard, gross and net of reinsurance, every year as part of the reinsurance treaty renewal cycle. The data exists. The skill exists. The metric maps directly to the standard’s amount-and-percentage form for business-activity exposure under Para 29(c). The disclosure should follow as a surfacing of existing actuarial output.

It does not follow. IAG, the cohort’s strongest insurance reporter, does not publicly disclose specific PMLs in either the 2025 Annual Report or the Severe Weather in a Changing Climate third edition. The Annual Report Climate-related disclosures section at pages 37-49 carries strategy-level natural-perils risk narrative, qualitative material-risk identification (pages 38), and the Climate Action Plan target performance count of 1.7m natural-hazard risk reduction actions delivered against a 1.0m FY25 interim target (page 44). The Severe Weather report, co-developed with the US National Center for Atmospheric Research, carries directional-change attribution per peril (qualitative directional indicators across seven hazards) rather than numeric PML triangulation. The actuarial output exists internally for treaty negotiation purposes. The disclosed output is qualitative and target-count.

The gap matters because IAG describes the underlying process as “both qualitative and quantitative in nature” on page 47. The internal Climate Risk and Opportunity (CRR&O) Process “adopts a quantitative-first assessment approach to determine the materiality of the potential set of climate-related risks that were identified.” The process is quantitative; the disclosure is not. The standard asks for the disclosure. In the wider general-insurance cohort, the same pattern is reported by KPMG’s FAST 30 commentary and PwC’s AASB S2 review for the Australian first wave. The actuarial capacity exists; the publication discipline does not.

This is the strongest empirical lever for the year-two pressure that the four cohort reviews are flagging. B17 of the standard requires entities with high exposure and the skills to apply a more advanced quantitative approach. Insurance is the sector where the test is most easily applied because the actuarial capacity is documented in the regulatory filings (APRA CPS 220 in Australia, equivalents elsewhere) and the underlying PML output is computed as a matter of operational necessity. A Para 29(c) treatment that says “continues to mature” at the public disclosure level while the internal PML output is computed annually and triangulated for treaty negotiation is a treatment that auditors will press in years two and three. The lever is not a regulator-imposed stretch; it is the gap between internal capability and external publication.

For practitioners working on this gap, Continuuiti’s site-level climate risk product provides catastrophe-modelled physical-risk exposure across 12 hazards and multiple scenarios. This is the same shape of data underwriters use internally for PML triangulation, surfaced in a form that maps to a Para 29(c) disclosure (methodology).

IFRS S2 Practitioners
Defensible IFRS S2 Disclosures Across 12 Hazards
Site-level catastrophe-modelled physical risk for Wave 2 reporters preparing Para 29(c) and Para 22.

Book a Demo

Sector observations

Four sector-specific observations from the deeply-read cohort.

Banking sub-portfolio variance. The lending book is the natural Para 29(c) scope for banks because own operations are typically immaterial relative to financed assets. Sub-portfolio disaggregation (residential mortgage, commercial real estate, project finance) is the natural unit for physical-risk vulnerability in banking; neither Bank of America nor Standard Bank gets there in the documents we read. This is the disclosure form that IBG vol-19 metric FN-MF-450a.1 (mortgage loans in 100-year flood zones) and FN-MF-450a.2 (expected loss attributable to weather catastrophes) anticipate. A worked banking IFRS S2 disclosure shows how the sub-portfolio form lands when fully built out.

Mining under-disclosure of tailings vulnerability. Tailings storage facilities are typically the largest single PPE concentration on a mining balance sheet and the most exposed to chronic and acute physical hazards. Rio Tinto addresses TSF vulnerability in its page 67 risk register and Global Industry Standard on Tailings Management resilience assessment (page 69); Tata Steel’s six-page summary does not surface tailings at all.

Real estate is the clean missing sector. Our pile contains zero pure-play real estate or REIT reporters; KPMG’s FAST 30 dashboard surfaces AU REITs (Goodman, Stockland, GPT, Mirvac, Lendlease likely candidates) but none are in our pile. Real estate is the cleanest sector for Para 29(c) because the unit of analysis (the property) maps one-to-one onto the asset-vulnerability question, and the IBG IF-RE-450a.1 metric (area of properties in 100-year flood zones, by sector) is hazard-threshold-explicit.

Insurance three-lens scope expectation versus practice. Defensible Para 29(c) insurance disclosure surfaces three lenses: underwriting business activity exposure (PML triangulation per FN-IN-450a.1), investment portfolio assets vulnerable to physical risks, and operational property, plant and equipment. None of the three appears with a quantified amount in IAG’s Annual Report. The cohort pattern, per the four cohort reviews, is for insurance reporters to disclose qualitatively and reserve quantitative PML for non-public actuarial reporting. A worked insurance IFRS S2 disclosure shows what a three-lens defensible disclosure looks like when each lens is surfaced.

What the cohort reviews say about year two

The four published cohort reviews converge on a single direction: the proportionality clause is doing work that auditors and regulators expect to do less of in years two and three. KPMG Australia’s AASB S2 First Impressions (FAST 30 dashboard tracker) reports the AU first wave split roughly two-thirds quantified to one-third not, with tolerance narrowing year over year. PwC Australia’s AASB S2 Unpacked (22 first-wave Group 1 reporters) records the same split and recommends year-two reporters commit to explicit quantification roadmaps where year-one is qualitative-only (what gets quantified by when, with what methodology, sourcing what data). Deloitte Australia’s Wave 1 Review (reports averaged 33 pages, range 7-82) reads the year-two narrative as the move from transitional to fully-compliant disclosure quality, not from voluntary to mandatory. NZ FMA / SLR Consulting’s NZ CRD review was the earliest and most explicit on the amount-and-percentage-of-assets-vulnerable gap; the NZ regulatory expectation for year two is that the gap is closed, not committed-to-be-closed in year three.

The implication for the four-tier classification: the OMITTED tier (Tata Steel six-page summary, CLP 2023 Content Index) faces the highest year-two pressure. The QUALITATIVE-ONLY tier (Bank of America, Standard Bank, IAG) faces pressure to commit to quantification roadmaps with methodology and data sourcing identified. The PARTIAL tier (Rio Tinto with the AD analysis) faces pressure to extend from risk-categorisation to specific carrying-amount-vulnerable figures. The QUANTIFIED tier faces pressure on consistency across years (horizon decay, scenario re-baselining, threshold updates).

The interpretive companion to this piece set the four definitional choices Para 29(c) hides; the empirical record above shows how the first wave answered them. The next wave can read both, pick positions, and be ready.

Sources cited

Standard, Basis for Conclusions, Accompanying Guidance

  • IFRS S2 Climate-related Disclosures, June 2023 (IFRS Foundation). Paragraphs cited: 11, 18, 22(b)(i)(1)-(7), 29(b)-(g), 30, 31, 32. Appendix B paragraphs cited: B11, B17, B65.
  • Basis for Conclusions on IFRS S2. BC paragraphs cited: BC75, BC77, BC131, BC132.

Public disclosures cited by name with verbatim page references (six deeply-read reporters plus Sembcorp alignment claim)

  • Bank of America Corporation, ISSB IFRS S2 Disclosure 2025, December 2025. Pages 11 (proportionality opt-out verbatim), 12-15 (qualitative physical-risk descriptions), 14 (commercial-credit sector concentration table sourced from 2024 Form 10-K), 16-17 (NGFS Phase IV scenarios + acute physical scenario).
  • Standard Bank Group, Climate-Related Financial Disclosures Report 2024. Pages 14 (NGFS two-scenario disclosure), 17-23 (sector physical risk identification), 25 (banking book sector breakdown), 27 (real estate carrying amounts), 18 (real estate physical-risk verbatim).
  • Insurance Australia Group, 2025 Annual Report (Sustainability Report at pages 35-51). Pages 35-37 (alignment claim and disclaimer), 38 (Material climate-related risks table), 39 (Climate Action Plan summary), 41-43 (Climate-related scenario assessment with three NGFS Phase IV scenarios and three time horizons), 44 (Performance to transition plan targets, including Target 6 hazard reduction count of 1.7m vs 1.0m interim target), 47 (CRR&O Process description), 48-49 (Climate-related governance).
  • Insurance Australia Group, Severe Weather in a Changing Climate, 3rd edition, November 2025 (NCAR partnership). Cited for directional-change attribution methodology (qualitative per-peril direction across seven hazards).
  • Rio Tinto, 2025 Climate Action Plan (extracted from 2024 Annual Report pages 41-75). Pages 41 (alignment claim), 43-45 (macroeconomic scenarios), 44 (proportionality-adjacent language on quantitative financial impacts), 47 (KPMG assurance scope), 66-69 (Physical climate risk and resilience: 4-pillar methodology, 8 material physical risks, annualised damage methodology, Group and regional AD risk heat-map), 75 (TCFD Index).
  • Tata Steel, FY25 Climate Change Report (6-page summary; fuller content in Integrated Report 2024-25). Pages 1 (alignment claim with IFRS S2 and TNFD via four-pillar structure), 4 (TCFD-aligned third-party Climate Risk assessment for India / Netherlands / UK steelmaking sites; physical-risk qualitative narrative), 6 (referral to fuller IAR for detail).
  • CLP Group Holdings, 2023 IFRS S2 Content Index. Pages 1 (alignment claim “based on” ISSB IFRS S2), 2 (IFRS S2-9(d) Nil + commitment-language Remarks), 5 (IFRS S2-15(b) and 16(b)(d) Nil + commitment-language Remarks), 6 (Para 22 sub-elements showing six of seven Nil entries in 2023 AR/SR column), 7 (IFRS S2-25(a)(iv) and (b) Remarks), 9 (Para 29(b)-(e) explicit Nil entries in 2023 AR/SR column).
  • Sembcorp Industries, 2024 Annual Report. Page 51 (TCFD-only alignment claim verbatim and SGX requirement reference).

Public disclosures with partial agent-extracted evidence (one reporter)

  • Wesfarmers, 2025 Annual Report. Pages 62 (scenario analysis methodology language), 63-64 (physical-risk identification and store/facility location mapping).

Cohort reviews (publicly cited; not in local pile, paywalled or dashboard-gated)

Industry-Based Guidance referenced

  • IBG vol-17 FN-IN-450a.1 (Probable Maximum Loss of insured products from weather-related natural catastrophes); IBG vol-19 FN-MF-450a.1 (mortgage loans in 100-year flood zones), FN-MF-450a.2 (expected loss attributable to weather catastrophes); IBG vol-36 IF-RE-450a.1 (area of properties in 100-year flood zones, by property sector).

Continuuiti companion references

Frequently asked questions

Which IFRS S2 first-wave reporters quantify Para 29(c)?

Of the six deeply-read first-wave reporters in this analysis, only Rio Tinto produces a Para 29(c)-relevant quantification beyond narrative. Rio Tinto’s 2025 Climate Action Plan discloses an annualised damage methodology with risk-categorisation thresholds (low under 0.2 percent, medium 0.2 to 1 percent, high above 1 percent) across eight hazards, two SSP scenarios and three time horizons. Even so, the disclosure stops short of naming a specific carrying-amount-vulnerable figure. The other five reporters split as follows: Bank of America, Standard Bank and IAG sit in the QUALITATIVE-ONLY tier; Tata Steel’s six-page summary and CLP’s 2023 Content Index are OMITTED.

What is annualised damage and why does Rio Tinto use it?

Annualised damage (AD) is the expected average annual damage to an asset attributable to climate-related hazards, expressed as a percentage of a fixed asset value. Rio Tinto defines it verbatim at page 68 of the 2025 Climate Action Plan: an AD of 0.5 percent means that for every $1 million of exposure, $5,000 could be damaged on average in any given year. The Group-level present-day AD sits in the medium category (0.2 to 1 percent) and is projected to rise toward the high tier by 2050 under both SSP2-4.5 and SSP5-8.5. The methodology is gross-of-adaptation by design, on a stationary do-nothing assumption.

Why is insurance Para 29(c) hard despite PML data?

Insurers compute Probable Maximum Loss by return period and hazard, gross and net of reinsurance, every year as part of the reinsurance treaty renewal cycle. The actuarial capacity exists. The metric maps directly to the standard’s amount-and-percentage form for business-activity exposure. The disclosure should follow as a surfacing of existing actuarial output. It does not. IAG, the cohort’s strongest insurance reporter, does not publicly disclose specific PMLs in either the 2025 Annual Report or the Severe Weather in a Changing Climate third edition. The gap between internal capability and external publication is the year-two pressure point, not a regulator-imposed stretch.

Is the IFRS S2 proportionality clause being overused in first-wave disclosures?

Para 30 of IFRS S2 allows entities to use only the reasonable and supportable information available without undue cost or effort. Two of the six deeply-read reporters invoke the clause explicitly (Bank of America, IAG); two more invoke it softly through commitment-language patterns (Rio Tinto narrowly, CLP broadly via recurring Remarks formulations). B17 of the standard’s appendix carries a constraint the cohort reviews are flagging: entities with high exposure and the skills are required to apply a more advanced quantitative approach. The Big-4 commentary across the four cohort reviews flags the same trajectory: tolerance is narrowing year over year.

What does year-two pressure look like under IFRS S2 and AASB S2?

The four published cohort reviews (KPMG FAST 30, PwC AASB S2 Unpacked, Deloitte Wave 1, NZ FMA / SLR Consulting) converge on a single direction. The proportionality clause is doing work auditors and regulators expect to do less of. OMITTED-tier reporters face the highest pressure to disclose. QUALITATIVE-ONLY reporters face pressure to commit to explicit quantification roadmaps with methodology and data-sourcing identified by year two. PARTIAL-tier reporters face pressure to extend from risk-categorisation to specific carrying-amount-vulnerable figures. The NZ regulatory expectation is that the gap is closed in year two, not committed-to-be-closed in year three.

Govind Balachandran
Govind Balachandran

Govind Balachandran is the founder of Continuuiti. He writes extensively on climate risk and operational risk intelligence for enterprises. Previously, he has worked for 7+ years in enterprise risk management, building and deploying third-party risk management and due diligence solutions across 100+ enterprises.

Related Posts